31st Annual Meeting and Associated Programs of the Society for Immunotherapy of Cancer (SITC 2016): part one

National Harbor, MD, USA. 9-13 November 2016
  • Andreas Lundqvist
  • Vincent van Hoef
  • Xiaonan Zhang
  • Erik Wennerberg
  • Julie Lorent
  • Kristina Witt
  • Laia Masvidal Sanz
  • Shuo Liang
  • Shannon Murray
  • Ola Larsson
  • Rolf Kiessling
  • Yumeng Mao
  • John-William Sidhom
  • Catherine A. Bessell
  • Jonathan Havel
  • Jonathan Schneck
  • Timothy A. Chan
  • Eliot Sachsenmeier
  • David Woods
  • Anders Berglund
  • Rupal Ramakrishnan
  • Andressa Sodre
  • Jeffrey Weber
  • Roberta Zappasodi
  • Yanyun Li
  • Jingjing Qi
  • Philip Wong
  • Cynthia Sirard
  • Michael Postow
  • Walter Newman
  • Henry Koon
  • Vamsidhar Velcheti
  • Margaret K. Callahan
  • Jedd D. Wolchok
  • Taha Merghoub
  • Lawrence G. Lum
  • Minsig Choi
  • Archana Thakur
  • Abhinav Deol
  • Gregory Dyson
  • Anthony Shields
  • Cara Haymaker
  • Marc Uemura
  • Ravi Murthy
  • Marihella James
  • Daqing Wang
  • Julie Brevard
  • Catherine Monaghan
  • Suzanne Swann
  • James Geib
  • Mark Cornfeld
  • Srinivas Chunduru
  • Sudhir Agrawal
  • Cassian Yee
  • Jennifer Wargo
  • Sapna P. Patel
  • Rodabe Amaria
  • Hussein Tawbi
  • Isabella Glitza
  • Scott Woodman
  • Wen-Jen Hwu
  • Michael A. Davies
  • Patrick Hwu
  • Willem W. Overwijk
  • Chantale Bernatchez
  • Adi Diab
  • Erminia Massarelli
  • Neil H. Segal
  • Vincent Ribrag
  • Ignacio Melero
  • Tara C. Gangadhar
  • Walter Urba
  • Dirk Schadendorf
  • Robert L. Ferris
  • Roch Houot
  • Franck Morschhauser
  • Theodore Logan
  • Jason J. Luke
  • William Sharfman
  • Fabrice Barlesi
  • Patrick A. Ott
  • Laura Mansi
  • Shivaani Kummar
  • Gilles Salles
  • Cecilia Carpio
  • Roland Meier
  • Suba Krishnan
  • Dan McDonald
  • Matthew Maurer
  • Xuemin Gu
  • Jaclyn Neely
  • Satyendra Suryawanshi
  • Ronald Levy
  • Nikhil Khushalani
  • Jennifer Wu
  • Jinyu Zhang
  • Fahmin Basher
  • Mark Rubinstein
  • Mark Bucsek
  • Guanxi Qiao
  • Cameron MacDonald
  • Bonnie Hylander
  • Elizabeth Repasky
  • Shilpak Chatterjee
  • Anusara Daenthanasanmak
  • Paramita Chakraborty
  • Kyle Toth
  • Megan Meek
  • Elizabeth Garrett-Mayer
  • Michael Nishimura
  • Chrystal Paulos
  • Craig Beeson
  • Xuezhong Yu
  • Shikhar Mehrotra
  • Fei Zhao
  • Kathy Evans
  • Christine Xiao
  • Alisha Holtzhausen
  • Brent A. Hanks
  • Nicole Scharping
  • Ashley V. Menk
  • Rebecca Moreci
  • Ryan Whetstone
  • Rebekah Dadey
  • Simon Watkins
  • Robert Ferris
  • Greg M. Delgoffe
  • Jonathan Peled
  • Sean Devlin
  • Anna Staffas
  • Melissa Lumish
  • Kori Porosnicu Rodriguez
  • Katya Ahr
  • Miguel Perales
  • Sergio Giralt
  • Ying Taur
  • Eric Pamer
  • Marcel R. M. van den Brink
  • Robert Jenq
  • Nicola Annels
  • Hardev Pandha
  • Guy Simpson
  • Hugh Mostafid
  • Kevin Harrington
  • Alan Melcher
  • Mark Grose
  • Bronwyn Davies
  • Gough Au
  • Roberta Karpathy
  • Darren Shafren
  • Jacob Ricca
  • Taha Merghoub
  • Jedd D. Wolchok
  • Dmitriy Zamarin
  • Luciana Batista
  • Florence Marliot
  • Angela Vasaturo
  • Sabrina Carpentier
  • Cécile Poggionovo
  • Véronique Frayssinet
  • Jacques Fieschi
  • Marc Van den Eynde
  • Franck Pagès
  • Jérôme Galon
  • Fabienne Hermitte
  • Sean G. Smith
  • Khue Nguyen
  • Sruthi Ravindranathan
  • Bhanu Koppolu
  • David Zaharoff
  • Gustavo Schvartsman
  • Roland Bassett
  • Jennifer L. McQuade
  • Lauren E. Haydu
  • Michael A. Davies
  • Hussein Tawbi
  • Isabella Glitza
  • Douglas Kline
  • Xiufen Chen
  • Dominick Fosco
  • Justin Kline
  • Abigail Overacre
  • Maria Chikina
  • Erin Brunazzi
  • Gulidanna Shayan
  • William Horne
  • Jay Kolls
  • Robert L. Ferris
  • Greg M. Delgoffe
  • Tullia C. Bruno
  • Creg Workman
  • Dario Vignali
  • Prasad S. Adusumilli
  • Ephraim A Ansa-Addo
  • Zihai Li
  • Andrew Gerry
  • Joseph P. Sanderson
  • Karen Howe
  • Roslin Docta
  • Qian Gao
  • Eleanor A. L. Bagg
  • Nicholas Tribble
  • Miguel Maroto
  • Gareth Betts
  • Natalie Bath
  • Luca Melchiori
  • Daniel E. Lowther
  • Indu Ramachandran
  • Gabor Kari
  • Samik Basu
  • Gwendolyn Binder-Scholl
  • Karen Chagin
  • Lini Pandite
  • Tom Holdich
  • Rafael Amado
  • Hua Zhang
  • John Glod
  • Donna Bernstein
  • Bent Jakobsen
  • Crystal Mackall
  • Ryan Wong
  • Jonathan D. Silk
  • Katherine Adams
  • Garth Hamilton
  • Alan D. Bennett
  • Sara Brett
  • Junping Jing
  • Adriano Quattrini
  • Manoj Saini
  • Guy Wiedermann
  • Andrew Gerry
  • Bent Jakobsen
  • Gwendolyn Binder-Scholl
  • Joanna Brewer
  • MyLinh Duong
  • An Lu
  • Peter Chang
  • Aruna Mahendravada
  • Nicholas Shinners
  • Kevin Slawin
  • David M. Spencer
  • Aaron E. Foster
  • J. Henri Bayle
  • Cristina Bergamaschi
  • Sinnie Sin Man Ng
  • Bethany Nagy
  • Shawn Jensen
  • Xintao Hu
  • Candido Alicea
  • Bernard Fox
  • Barbara Felber
  • George Pavlakis
  • Jessica Chacon
  • Tori Yamamoto
  • Thomas Garrabrant
  • Luis Cortina
  • Daniel J. Powell
  • Marco Donia
  • Julie Westerlin Kjeldsen
  • Rikke Andersen
  • Marie Christine Wulff Westergaard
  • Valentina Bianchi
  • Mateusz Legut
  • Meriem Attaf
  • Garry Dolton
  • Barbara Szomolay
  • Sascha Ott
  • Rikke Lyngaa
  • Sine Reker Hadrup
  • Andrew Kelvin Sewell
  • Inge Marie Svane
  • Aaron Fan
  • Takumi Kumai
  • Esteban Celis
  • Ian Frank
  • Amanda Stramer
  • Michelle A. Blaskovich
  • Seth Wardell
  • Maria Fardis
  • James Bender
  • Michael T. Lotze
  • Stephanie L. Goff
  • Nikolaos Zacharakis
  • Yasmine Assadipour
  • Todd D. Prickett
  • Jared J. Gartner
  • Robert Somerville
  • Mary Black
  • Hui Xu
  • Harshini Chinnasamy
  • Isaac Kriley
  • Lily Lu
  • John Wunderlich
  • Paul F. Robbins
  • Steven Rosenberg
  • Steven A. Feldman
  • Kasia Trebska-McGowan
  • Isaac Kriley
  • Parisa Malekzadeh
  • Eden Payabyab
  • Richard Sherry
  • Steven Rosenberg
  • Stephanie L. Goff
  • Aishwarya Gokuldass
  • Michelle A. Blaskovich
  • Charlene Kopits
  • Brian Rabinovich
  • Michael T. Lotze
  • Daniel S. Green
  • Olena Kamenyeva
  • Kathryn C. Zoon
  • Christina M. Annunziata
  • Joanne Hammill
  • Christopher Helsen
  • Craig Aarts
  • Jonathan Bramson
  • Yui Harada
  • Yoshikazu Yonemitsu
  • Christopher Helsen
  • Joanne Hammill
  • Kenneth Mwawasi
  • Galina Denisova
  • Jonathan Bramson
  • Rajanish Giri
  • Benjamin Jin
  • Tracy Campbell
  • Lindsey M. Draper
  • Sanja Stevanovic
  • Zhiya Yu
  • Bianca Weissbrich
  • Nicholas P. Restifo
  • Cornelia L. Trimble
  • Steven Rosenberg
  • Christian S. Hinrichs
  • Kwong Tsang
  • Massimo Fantini
  • James W. Hodge
  • Rika Fujii
  • Ingrid Fernando
  • Caroline Jochems
  • Christopher Heery
  • James Gulley
  • Patrick Soon-Shiong
  • Jeffrey Schlom
  • Weiqing Jing
  • Jill Gershan
  • Grace Blitzer
  • James Weber
  • Laura McOlash
  • Bryon D. Johnson
  • Simin Kiany
  • Huang Gangxiong
  • Eugenie S. Kleinerman
  • Michael Klichinsky
  • Marco Ruella
  • Olga Shestova
  • Saad Kenderian
  • Miriam Kim
  • John Scholler
  • Carl H. June
  • Saar Gill
  • Duane Moogk
  • Shi Zhong
  • Zhiya Yu
  • Ivan Liadi
  • William Rittase
  • Victoria Fang
  • Janna Dougherty
  • Arianne Perez-Garcia
  • Iman Osman
  • Cheng Zhu
  • Navin Varadarajan
  • Nicholas P. Restifo
  • Alan Frey
  • Michelle Krogsgaard
  • Daniel Landi
  • Kristen Fousek
  • Malini Mukherjee
  • Ankita Shree
  • Sujith Joseph
  • Kevin Bielamowicz
  • Tiara Byrd
  • Nabil Ahmed
  • Meenakshi Hegde
  • Sylvia Lee
  • David Byrd
  • John Thompson
  • Shailender Bhatia
  • Scott Tykodi
  • Judy Delismon
  • Liz Chu
  • Siddiq Abdul-Alim
  • Arpy Ohanian
  • Anna Marie DeVito
  • Stanley Riddell
  • Kim Margolin
  • Isabelle Magalhaes
  • Jonas Mattsson
  • Michael Uhlin
  • Satoshi Nemoto
  • Patricio Pérez Villarroel
  • Ryosuke Nakagawa
  • James J. Mule
  • Adam W. Mailloux
  • Melinda Mata
  • Phuong Nguyen
  • Claudia Gerken
  • Christopher DeRenzo
  • David M. Spencer
  • Stephen Gottschalk
  • Mélissa Mathieu
  • Sandy Pelletier
  • John Stagg
  • Simon Turcotte
  • Nicholas Minutolo
  • Prannda Sharma
  • Andrew Tsourkas
  • Daniel J. Powell
  • Nadine Mockel-Tenbrinck
  • Daniela Mauer
  • Katharina Drechsel
  • Carola Barth
  • Katharina Freese
  • Ulrike Kolrep
  • Silke Schult
  • Mario Assenmacher
  • Andrew Kaiser
  • John Mullinax
  • MacLean Hall
  • Julie Le
  • Krithika Kodumudi
  • Erica Royster
  • Allison Richards
  • Ricardo Gonzalez
  • Amod Sarnaik
  • Shari Pilon-Thomas
  • Morten Nielsen
  • Anders Krarup-Hansen
  • Dorrit Hovgaard
  • Michael Mørk Petersen
  • Anand Chainsukh Loya
  • Niels Junker
  • Inge Marie Svane
  • Charlotte Rivas
  • Robin Parihar
  • Stephen Gottschalk
  • Cliona M. Rooney
  • Haiying Qin
  • Sang Nguyen
  • Paul Su
  • Chad Burk
  • Brynn Duncan
  • Bong-Hyun Kim
  • M. Eric Kohler
  • Terry Fry
  • Arjun A. Rao
  • Noam Teyssier
  • Jacob Pfeil
  • Nikolaos Sgourakis
  • Sofie Salama
  • David Haussler
  • Sarah A. Richman
  • Selene Nunez-Cruz
  • Zack Gershenson
  • Zissimos Mourelatos
  • David Barrett
  • Stephan Grupp
  • Michael Milone
  • Alba Rodriguez-Garcia
  • Matthew K. Robinson
  • Gregory P. Adams
  • Daniel J. Powell
  • João Santos
  • Riikka Havunen
  • Mikko Siurala
  • Víctor Cervera-Carrascón
  • Suvi Parviainen
  • Marjukka Antilla
  • Akseli Hemminki
  • Jyothi Sethuraman
  • Laurelis Santiago
  • Jie Qing Chen
  • Zhimin Dai
  • Seth Wardell
  • James Bender
  • Michael T. Lotze
  • Huizi Sha
  • Shu Su
  • Naiqing Ding
  • Baorui Liu
  • Sanja Stevanovic
  • Anna Pasetto
  • Sarah R. Helman
  • Jared J. Gartner
  • Todd D. Prickett
  • Paul F. Robbins
  • Steven A. Rosenberg
  • Christian S. Hinrichs
  • Shailender Bhatia
  • Melissa Burgess
  • Hui Zhang
  • Tien Lee
  • Hans Klingemann
  • Patrick Soon-Shiong
  • Paul Nghiem
  • John M. Kirkwood
  • John M. Rossi
  • Marika Sherman
  • Allen Xue
  • Yueh-wei Shen
  • Lynn Navale
  • Steven A. Rosenberg
  • James N. Kochenderfer
  • Adrian Bot
  • Anandaraman Veerapathran
  • Aishwarya Gokuldass
  • Amanda Stramer
  • Jyothi Sethuraman
  • Michelle A. Blaskovich
  • Doris Wiener
  • Ian Frank
  • Laurelis Santiago
  • Brian Rabinovich
  • Maria Fardis
  • James Bender
  • Michael T. Lotze
  • Edmund K. Waller
  • Jian-Ming Li
  • Christopher Petersen
  • Bruce R. Blazar
  • Jingxia Li
  • Cynthia R. Giver
  • Ziming Wang
  • Steven K. Grossenbacher
  • Ian Sturgill
  • Robert J. Canter
  • William J. Murphy
  • Congcong Zhang
  • Michael C. Burger
  • Lukas Jennewein
  • Anja Waldmann
  • Michel Mittelbronn
  • Torsten Tonn
  • Joachim P. Steinbach
  • Winfried S. Wels
  • Jason B. Williams
  • Yuanyuan Zha
  • Thomas F. Gajewski
  • LaTerrica C. Williams
  • Giedre Krenciute
  • Mamta Kalra
  • Chrystal Louis
  • Stephen Gottschalk
  • Gang Xin
  • David Schauder
  • Aimin Jiang
  • Nikhil Joshi
  • Weiguo Cui
  • Xue Zeng
  • Ashley V. Menk
  • Nicole Scharping
  • Greg M. Delgoffe
  • Zeguo Zhao
  • Mohamad Hamieh
  • Justin Eyquem
  • Gertrude Gunset
  • Neil Bander
  • Michel Sadelain
  • David Askmyr
  • Milad Abolhalaj
  • Kristina Lundberg
  • Lennart Greiff
  • Malin Lindstedt
  • Helen K. Angell
  • Kyoung-Mee Kim
  • Seung-Tae Kim
  • Sung Kim
  • Alan D. Sharpe
  • Julia Ogden
  • Anna Davenport
  • Darren R. Hodgson
  • Carl Barrett
  • Jeeyun Lee
  • Elaine Kilgour
  • Jodi Hanson
  • Richard Caspell
  • Alexey Karulin
  • Paul Lehmann
  • Tameem Ansari
  • Annemarie Schiller
  • Srividya Sundararaman
  • Paul Lehmann
  • Jodi Hanson
  • Diana Roen
  • Alexey Karulin
  • Paul Lehmann
  • Mark Ayers
  • Diane Levitan
  • Gladys Arreaza
  • Fang Liu
  • Robin Mogg
  • Yung-Jue Bang
  • Bert O’Neil
  • Razvan Cristescu
  • Philip Friedlander
  • Karl Wassman
  • Chrisann Kyi
  • William Oh
  • Nina Bhardwaj
  • Svetlana Bornschlegl
  • Michael P. Gustafson
  • Dennis A. Gastineau
  • Ian F. Parney
  • Allan B. Dietz
  • Daniel Carvajal-Hausdorf
  • Nikita Mani
  • Vamsidhar Velcheti
  • Kurt Schalper
  • David Rimm
  • Serena Chang
  • Ronald Levy
  • John Kurland
  • Suba Krishnan
  • Christoph Matthias Ahlers
  • Maria Jure-Kunkel
  • Lewis Cohen
  • Holden Maecker
  • Holbrook Kohrt
  • Shuming Chen
  • George Crabill
  • Theresa Pritchard
  • Tracee McMiller
  • Drew Pardoll
  • Fan Pan
  • Suzanne Topalian
  • Patrick Danaher
  • Sarah Warren
  • Lucas Dennis
  • Andrew M. White
  • Leonard D’Amico
  • Melissa Geller
  • Mary L. Disis
  • Joseph Beechem
  • Kunle Odunsi
  • Steven Fling
  • Roshanak Derakhshandeh
  • Tonya J. Webb
  • Sigrid Dubois
  • Kevin Conlon
  • Bonita Bryant
  • Jennifer Hsu
  • Nancy Beltran
  • Jürgen Müller
  • Thomas Waldmann
  • Rebekka Duhen
  • Thomas Duhen
  • Lucas Thompson
  • Ryan Montler
  • Andrew Weinberg
  • Max Kates
  • Brandon Early
  • Erik Yusko
  • Taylor H. Schreiber
  • Trinity J. Bivalacqua
  • Mark Ayers
  • Jared Lunceford
  • Michael Nebozhyn
  • Erin Murphy
  • Andrey Loboda
  • David R. Kaufman
  • Andrew Albright
  • Jonathan Cheng
  • S. Peter Kang
  • Veena Shankaran
  • Sarina A. Piha-Paul
  • Jennifer Yearley
  • Tanguy Seiwert
  • Antoni Ribas
  • Terrill K. McClanahan
  • Razvan Cristescu
  • Robin Mogg
  • Mark Ayers
  • Andrew Albright
  • Erin Murphy
  • Jennifer Yearley
  • Xinwei Sher
  • Xiao Qiao Liu
  • Michael Nebozhyn
  • Jared Lunceford
  • Andrew Joe
  • Jonathan Cheng
  • Elizabeth Plimack
  • Patrick A. Ott
  • Terrill K. McClanahan
  • Andrey Loboda
  • David R. Kaufman
  • Alex Forrest-Hay
  • Cheryl A. Guyre
  • Kohei Narumiya
  • Marc Delcommenne
  • Heather A. Hirsch
  • Amit Deshpande
  • Jason Reeves
  • Jenny Shu
  • Tong Zi
  • Jennifer Michaelson
  • Debbie Law
  • Elizabeth Trehu
  • Sriram Sathyanaryanan
  • Brendan P. Hodkinson
  • Natalie A. Hutnick
  • Michael E. Schaffer
  • Michael Gormley
  • Tyler Hulett
  • Shawn Jensen
  • Carmen Ballesteros-Merino
  • Christopher Dubay
  • Michael Afentoulis
  • Ashok Reddy
  • Larry David
  • Bernard Fox
  • Kumar Jayant
  • Swati Agrawal
  • Rajendra Agrawal
  • Ghayathri Jeyakumar
  • Seongho Kim
  • Heejin Kim
  • Cynthia Silski
  • Stacey Suisham
  • Elisabeth Heath
  • Ulka Vaishampayan
  • Natalie Vandeven
  • Natasja Nielsen Viller
  • Alison O’Connor
  • Hui Chen
  • Bolette Bossen
  • Eric Sievers
  • Robert Uger
  • Paul Nghiem
  • Lisa Johnson
  • Hsiang-Fong Kao
  • Chin-Fu Hsiao
  • Shu-Chuan Lai
  • Chun-Wei Wang
  • Jenq-Yuh Ko
  • Pei-Jen Lou
  • Tsai-Jan Lee
  • Tsang-Wu Liu
  • Ruey-Long Hong
  • Staci J. Kearney
  • Joshua C. Black
  • Benjamin J. Landis
  • Sally Koegler
  • Brooke Hirsch
  • Roberto Gianani
  • Jeffrey Kim
  • Ming-Xiao He
  • Bingqing Zhang
  • Nan Su
  • Yuling Luo
  • Xiao-Jun Ma
  • Emily Park
  • Dae Won Kim
  • Domenico Copploa
  • Nishi Kothari
  • Young doo Chang
  • Richard Kim
  • Namyong Kim
  • Melvin Lye
  • Ee Wan
  • Namyong Kim
  • Melvin Lye
  • Ee Wan
  • Namyong Kim
  • Melvin Lye
  • Ee Wan
  • Hanna A. Knaus
  • Sofia Berglund
  • Hubert Hackl
  • Judith E. Karp
  • Ivana Gojo
  • Leo Luznik
  • Henoch S. Hong
  • Sven D. Koch
  • Birgit Scheel
  • Ulrike Gnad-Vogt
  • Karl-Josef Kallen
  • Volker Wiegand
  • Linus Backert
  • Oliver Kohlbacher
  • Ingmar Hoerr
  • Mariola Fotin-Mleczek
  • James M. Billingsley
  • Yoshinobu Koguchi
  • Valerie Conrad
  • William Miller
  • Iliana Gonzalez
  • Tomasz Poplonski
  • Tanisha Meeuwsen
  • Ana Howells-Ferreira
  • Rogan Rattray
  • Mary Campbell
  • Carlo Bifulco
  • Christopher Dubay
  • Keith Bahjat
  • Brendan Curti
  • Walter Urba
  • E-K Vetsika
  • G. Kallergi
  • Despoina Aggouraki
  • Z. Lyristi
  • P. Katsarlinos
  • Filippos Koinis
  • V. Georgoulias
  • Athanasios Kotsakis
  • Nathan T. Martin
  • Famke Aeffner
  • Staci J. Kearney
  • Joshua C. Black
  • Logan Cerkovnik
  • Luke Pratte
  • Rebecca Kim
  • Brooke Hirsch
  • Joseph Krueger
  • Roberto Gianani
  • Amaia Martínez-Usatorre
  • Camilla Jandus
  • Alena Donda
  • Laura Carretero-Iglesia
  • Daniel E. Speiser
  • Dietmar Zehn
  • Nathalie Rufer
  • Pedro Romero
  • Anshuman Panda
  • Janice Mehnert
  • Kim M. Hirshfield
  • Greg Riedlinger
  • Sherri Damare
  • Tracie Saunders
  • Levi Sokol
  • Mark Stein
  • Elizabeth Poplin
  • Lorna Rodriguez-Rodriguez
  • Ann Silk
  • Nancy Chan
  • Melissa Frankel
  • Michael Kane
  • Jyoti Malhotra
  • Joseph Aisner
  • Howard L. Kaufman
  • Siraj Ali
  • Jeffrey Ross
  • Eileen White
  • Gyan Bhanot
  • Shridar Ganesan
  • Anne Monette
  • Derek Bergeron
  • Amira Ben Amor
  • Liliane Meunier
  • Christine Caron
  • Antigoni Morou
  • Daniel Kaufmann
  • Moishe Liberman
  • Igor Jurisica
  • Anne-Marie Mes-Masson
  • Kamel Hamzaoui
  • Rejean Lapointe
  • Ann Mongan
  • Yuan-Chieh Ku
  • Warren Tom
  • Yongming Sun
  • Alex Pankov
  • Tim Looney
  • Janice Au-Young
  • Fiona Hyland
  • Jeff Conroy
  • Carl Morrison
  • Sean Glenn
  • Blake Burgher
  • He Ji
  • Mark Gardner
  • Ann Mongan
  • Angela R. Omilian
  • Jeff Conroy
  • Wiam Bshara
  • Omilian Angela
  • Blake Burgher
  • He Ji
  • Sean Glenn
  • Carl Morrison
  • Ann Mongan
  • Joseph M. Obeid
  • Gulsun Erdag
  • Mark E. Smolkin
  • Donna H. Deacon
  • James W. Patterson
  • Lieping Chen
  • Timothy N. Bullock
  • Craig L. Slingluff
  • Joseph M. Obeid
  • Gulsun Erdag
  • Donna H. Deacon
  • Craig L. Slingluff
  • Timothy N. Bullock
  • John T. Loffredo
  • Raja Vuyyuru
  • Sophie Beyer
  • Vanessa M. Spires
  • Maxine Fox
  • Jon M. Ehrmann
  • Katrina A. Taylor
  • Alan J. Korman
  • Robert F. Graziano
  • David Page
  • Katherine Sanchez
  • Carmen Ballesteros-Merino
  • Maritza Martel
  • Carlo Bifulco
  • Walter Urba
  • Bernard Fox
  • Sapna P. Patel
  • Mariana Petaccia De Macedo
  • Yong Qin
  • Alex Reuben
  • Christine Spencer
  • Michele Guindani
  • Roland Bassett
  • Jennifer Wargo
  • Adriana Racolta
  • Brian Kelly
  • Tobin Jones
  • Nathan Polaske
  • Noah Theiss
  • Mark Robida
  • Jeffrey Meridew
  • Iva Habensus
  • Liping Zhang
  • Lidija Pestic-Dragovich
  • Lei Tang
  • Ryan J. Sullivan
  • Theodore Logan
  • Nikhil Khushalani
  • Kim Margolin
  • Henry Koon
  • Thomas Olencki
  • Thomas Hutson
  • Brendan Curti
  • Joanna Roder
  • Shauna Blackmon
  • Heinrich Roder
  • John Stewart
  • Asim Amin
  • Marc S. Ernstoff
  • Joseph I. Clark
  • Michael B. Atkins
  • Howard L. Kaufman
  • Jeffrey Sosman
  • Jeffrey Weber
  • David F. McDermott
  • Jeffrey Weber
  • Harriet Kluger
  • Ruth Halaban
  • Mario Snzol
  • Heinrich Roder
  • Joanna Roder
  • Senait Asmellash
  • Arni Steingrimsson
  • Shauna Blackmon
  • Ryan J. Sullivan
  • Chichung Wang
  • Kristin Roman
  • Amanda Clement
  • Sean Downing
  • Clifford Hoyt
  • Nathalie Harder
  • Guenter Schmidt
  • Ralf Schoenmeyer
  • Nicolas Brieu
  • Mehmet Yigitsoy
  • Gabriele Madonna
  • Gerardo Botti
  • Antonio Grimaldi
  • Paolo A. Ascierto
  • Ralf Huss
  • Maria Athelogou
  • Harald Hessel
  • Nathalie Harder
  • Alexander Buchner
  • Guenter Schmidt
  • Christian Stief
  • Ralf Huss
  • Gerd Binnig
  • Thomas Kirchner
  • Shankar Sellappan
  • Sheeno Thyparambil
  • Sarit Schwartz
  • Fabiola Cecchi
  • Andrew Nguyen
  • Charles Vaske
  • Todd Hembrough
  • Jan Spacek
  • Michal Vocka
  • Eva Zavadova
  • Helena Skalova
  • Pavel Dundr
  • Lubos Petruzelka
  • Nicole Francis
  • Rau T. Tilman
  • Arndt Hartmann
  • Irena Netikova
  • Carmen Ballesteros-Merino
  • Julia Stump
  • Amanda Tufman
  • Frank Berger
  • Michael Neuberger
  • Rudolf Hatz
  • Michael Lindner
  • Rachel E. Sanborn
  • John Handy
  • Bernard Fox
  • Carlo Bifulco
  • Rudolf M. Huber
  • Hauke Winter
  • Simone Reu
  • Cheng Sun
  • Weihua Xiao
  • Zhigang Tian
  • Kshitij Arora
  • Niyati Desai
  • Anupriya Kulkarni
  • Mihir Rajurkar
  • Miguel Rivera
  • Vikram Deshpande
  • David Ting
  • Katy Tsai
  • Adi Nosrati
  • Simone Goldinger
  • Omid Hamid
  • Alain Algazi
  • Paul Tumeh
  • Jimmy Hwang
  • Jacqueline Liu
  • Lawrence Chen
  • Reinhard Dummer
  • Michael Rosenblum
  • Adil Daud
  • Tsu-Shuen Tsao
  • Julia Ashworth-Sharpe
  • Donald Johnson
  • Srabani Bhaumik
  • Christopher Bieniarz
  • Joseph Couto
  • Michael Farrell
  • Mahsa Ghaffari
  • Iva Habensus
  • Antony Hubbard
  • Tobin Jones
  • Brian Kelly
  • Jerome Kosmeder
  • Cleo Lee
  • Erin Marner
  • Jeffrey Meridew
  • Nathan Polaske
  • Adriana Racolta
  • Diana Uribe
  • Hongjun Zhang
  • Jian Zhang
  • Wenjun Zhang
  • Yifei Zhu
  • Larry Morrison
  • Lidija Pestic-Dragovich
  • Lei Tang
  • Takahiro Tsujikawa
  • Rohan N. Borkar
  • Vahid Azimi
  • Sushil Kumar
  • Guillaume Thibault
  • Motomi Mori
  • Edward El Rassi
  • Daniel R. Clayburgh
  • Molly F. Kulesz-Martin
  • Paul W. Flint
  • Lisa M. Coussens
  • Lisa Villabona
  • Giuseppe V. Masucci
  • Gary Geiss
  • Brian Birditt
  • Qian Mei
  • Alan Huang
  • Andrew M. White
  • Maribeth A. Eagan
  • Eduardo Ignacio
  • Nathan Elliott
  • Dwayne Dunaway
  • Lucas Dennis
  • Sarah Warren
  • Joseph Beechem
  • Dwayne Dunaway
  • Jaemyeong Jung
  • Chris Merritt
  • Isaac Sprague
  • Philippa Webster
  • Yan Liang
  • Sarah Warren
  • Joseph Beechem
  • Jessica Wenthe
  • Gunilla Enblad
  • Hannah Karlsson
  • Magnus Essand
  • Barbara Savoldo
  • Gianpietro Dotti
  • Martin Höglund
  • Malcolm K. Brenner
  • Hans Hagberg
  • Angelica Loskog
  • Matthew J. Bernett
  • Gregory L. Moore
  • Michael Hedvat
  • Christine Bonzon
  • Seung Chu
  • Rumana Rashid
  • Kendra N. Avery
  • Umesh Muchhal
  • John Desjarlais
  • Michael Hedvat
  • Matthew J. Bernett
  • Gregory L. Moore
  • Christine Bonzon
  • Rumana Rashid
  • Seung Chu
  • Kendra N. Avery
  • Umesh Muchhal
  • John Desjarlais
  • Matthew Kraman
  • Katarzyna Kmiecik
  • Natalie Allen
  • Mustapha Faroudi
  • Carlo Zimarino
  • Mateusz Wydro
  • Jacqueline Doody
  • Sreesha P. Srinivasa
  • Nagaraja Govindappa
  • Praveen Reddy
  • Aparajita Dubey
  • Sankar Periyasamy
  • Madhukara Adekandi
  • Chaitali Dey
  • Mary Joy
  • Pieter Fokko van Loo
  • Henrike Veninga
  • Setareh Shamsili
  • Mark Throsby
  • Harry Dolstra
  • Lex Bakker
  • Ajjai Alva
  • Juergen Gschwendt
  • Yohann Loriot
  • Joaquim Bellmunt
  • Dai Feng
  • Christian Poehlein
  • Thomas Powles
  • Emmanuel S. Antonarakis
  • Charles G. Drake
  • Haiyan Wu
  • Christian Poehlein
  • Johann De Bono
  • Rajat Bannerji
  • John Byrd
  • Gareth Gregory
  • Stephen Opat
  • Jake Shortt
  • Andrew J. Yee
  • Noopur Raje
  • Seth Thompson
  • Arun Balakumaran
  • Shaji Kumar
  • Brian I. Rini
  • Toni K. Choueiri
  • Mariangela Mariani
  • Laurence Albiges
  • John B. Haanen
  • Michael B. Atkins
  • James Larkin
  • Manuela Schmidinger
  • Domenico Magazzù
  • Alessandra di Pietro
  • Robert J. Motzer
  • Troels Holz Borch
  • Rikke Andersen
  • Per Kongsted
  • Magnus Pedersen
  • Morten Nielsen
  • Özcan Met
  • Marco Donia
  • Inge Marie Svane
  • Karim Boudadi
  • Hao Wang
  • James Vasselli
  • Jan E. Baughman
  • Jon Wigginton
  • Rehab Abdallah
  • Ashley Ross
  • Charles G. Drake
  • Emmanuel S. Antonarakis
  • Robert J. Canter
  • Jiwon Park
  • Ziming Wang
  • Steven Grossenbacher
  • Jesus I. Luna
  • Sita Withers
  • William Culp
  • Mingyi Chen
  • Arta Monjazeb
  • Michael S. Kent
  • William J. Murphy
  • Smita Chandran
  • Robert Somerville
  • John Wunderlich
  • David Danforth
  • James Yang
  • Richard Sherry
  • Christopher Klebanoff
  • Stephanie Goff
  • Biman Paria
  • Arvind Sabesan
  • Abhishek Srivastava
  • Steven A. Rosenberg
  • Udai Kammula
  • Brendan Curti
  • Jon Richards
  • Mark Faries
  • Robert H. I. Andtbacka
  • Mark Grose
  • Darren Shafren
  • Luis A. DiazJr.
  • Dung T. Le
  • Takayuki Yoshino
  • Thierry André
  • Johanna Bendell
  • Minori Koshiji
  • Yayan Zhang
  • S Peter Kang
  • Bao Lam
  • Dirk Jäger
  • Todd M. Bauer
  • Judy S. Wang
  • Jean K. Lee
  • Gulam A. Manji
  • Ragini Kudchadkar
  • John S. Kauh
  • Shande Tang
  • Naomi Laing
  • Gerald Falchook
  • Edward B. Garon
  • Balazs Halmos
  • Hui Rina
  • Natasha Leighl
  • Sung Sook Lee
  • William Walsh
  • Konstanin Dragnev
  • Bilal Piperdi
  • Luis Paz-Ares Rodriguez
  • Nabeegha Shinwari
  • Ziewn Wei
  • Michael P. Gustafson
  • Mary L Maas
  • Michael Deeds
  • Adam Armstrong
  • Svetlana Bornschlegl
  • Tim Peterson
  • Sue Steinmetz
  • Dennis A. Gastineau
  • Ian F. Parney
  • Allan B. Dietz
  • Thomas Herzog
  • Floor J. Backes
  • Larry Copeland
  • Maria Del Pilar Estevez Diz
  • Thomas W. Hare
  • Warner Huh
  • Byoung-Gie Kim
  • Kathleen M. Moore
  • Ana Oaknin
  • William Small
  • Krishnansu S. Tewari
  • Bradley J. Monk
  • Ashish M. Kamat
  • Joaquim Bellmunt
  • Toni K. Choueiri
  • Kijoeng Nam
  • Maria De Santis
  • Robert Dreicer
  • Noah M. Hahn
  • Rodolfo Perini
  • Arlene Siefker-Radtke
  • Guru Sonpavde
  • Ronald de Wit
  • J. Alfred Witjes
  • Stephen Keefe
  • Dean Bajorin
  • Justin Kline
  • Philippe Armand
  • John Kuruvilla
  • Craig Moskowitz
  • Mehdi Hamadani
  • Vincent Ribrag
  • Pier Luigi Zinzani
  • Sabine Chlosta
  • Seth Thompson
  • Arun Balakumaran
  • Nancy Bartlett
  • Chrisann Kyi
  • Rachel Sabado
  • Yvonne Saenger
  • Loging William
  • Michael Joseph Donovan
  • Erlinda Sacris
  • John Mandeli
  • Andres M. Salazar
  • Philip Friedlander
  • Nina Bhardwaj
  • John Powderly
  • Joshua Brody
  • John Nemunaitis
  • Leisha Emens
  • Jason J. Luke
  • Amita Patnaik
  • Ian McCaffery
  • Richard Miller
  • Ginna Laport
  • Andrew L. Coveler
  • David C. Smith
  • Juneko E. Grilley-Olson
  • Thomas F. Gajewski
  • Sanjay Goel
  • Shyra J. Gardai
  • Che-Leung Law
  • Gary Means
  • Thomas Manley
  • Brendan Curti
  • Kristen A. Marrone
  • Gary Rosner
  • Valsamo Anagnostou
  • Joanne Riemer
  • Jessica Wakefield
  • Cynthia Zanhow
  • Stephen Baylin
  • Barbara Gitlitz
  • Julie Brahmer
  • David F. McDermott
  • Sabina Signoretti
  • Wenting Li
  • Charles Schloss
  • Jean-Marie Michot
  • Philippe Armand
  • Wei Ding
  • Vincent Ribrag
  • Beth Christian
  • Arun Balakumaran
  • Patricia Marinello
  • Sabine Chlosta
  • Yayan Zhang
  • Margaret Shipp
  • Pier Luigi Zinzani
  • Yana G. Najjar
  • Lin
  • Lisa H. Butterfield
  • Ahmad A. Tarhini
  • Diwakar Davar
  • Hassane Zarour
  • Elizabeth Rush
  • Cindy Sander
  • John M. Kirkwood
  • Siqing Fu
  • Todd Bauer
  • Chris Molineaux
  • Mark K. Bennett
  • Keith W. Orford
  • Kyriakos P. Papadopoulos
  • Sukhmani K. Padda
  • Sumit A. Shah
  • A Dimitrios Colevas
  • Sujata Narayanan
  • George A. Fisher
  • Dana Supan
  • Heather A. Wakelee
  • Rhonda Aoki
  • Mark D. Pegram
  • Victor M. Villalobos
  • Jie Liu
  • Chris H. Takimoto
  • Mark Chao
  • Jens-Peter Volkmer
  • Ravindra Majeti
  • Irving L. Weissman
  • Branimir I. Sikic
  • David Page
  • Wendy Yu
  • Alison Conlin
  • Janet Ruzich
  • Stacy Lewis
  • Anupama Acheson
  • Kathleen Kemmer
  • Kelly Perlewitz
  • Nicole M. Moxon
  • Staci Mellinger
  • Carlo Bifulco
  • Maritza Martel
  • Yoshinobu Koguchi
  • Bernard Fox
  • Walter Urba
  • Heather McArthur
  • Magnus Pedersen
  • Marie Christine Wulff Westergaard
  • Troels Holz Borch
  • Morten Nielsen
  • Per Kongsted
  • Trine Juhler-Nøttrup
  • Marco Donia
  • Inge Marie Svane
  • Jayesh Desai
  • Ben Markman
  • Shahneen Sandhu
  • Hui Gan
  • Michael L. Friedlander
  • Ben Tran
  • Tarek Meniawy
  • Joanne Lundy
  • Duncan Colyer
  • Malaka Ameratunga
  • Christie Norris
  • Jason Yang
  • Kang Li
  • Lai Wang
  • Lusong Luo
  • Zhen Qin
  • Song Mu
  • Xuemei Tan
  • James Song
  • Michael Millward
  • Matthew H. G. Katz
  • Todd W. Bauer
  • Gauri R. Varadhachary
  • Nicolas Acquavella
  • Nipun Merchant
  • Gina Petroni
  • Craig L. SlingluffJr.
  • Osama E. Rahma
  • Brian I. Rini
  • Thomas Powles
  • Mei Chen
  • Yang Song
  • Markus Puhlmann
  • Michael B. Atkins
  • Sriram Sathyanaryanan
  • Heather A. Hirsch
  • Jenny Shu
  • Amit Deshpande
  • Arun Khattri
  • Jason Reeves
  • Tong Zi
  • Ryan Brisson
  • Christopher Harvey
  • Jennifer Michaelson
  • Debbie Law
  • Tanguy Seiwert
  • Jatin Shah
  • Maria Victoria Mateos
  • Morio Matsumoto
  • Hilary Blacklock
  • Albert Oriol Rocafiguera
  • Hartmut Goldschmidt
  • Shinsuke Iida
  • Dina Ben Yehuda
  • Enrique Ocio
  • Paula Rodríguez-Otero
  • Sundar Jagannath
  • Sagar Lonial
  • Uma Kher
  • Patricia Marinello
  • Jesus San-Miguel
  • Jatin Shah
  • Sagar Lonial
  • Moacyr Ribeiro de Oliveira
  • Habte Yimer
  • Maria Victoria Mateos
  • Robert Rifkin
  • Fredrik Schjesvold
  • Enrique Ocio
  • Paula Rodríguez-Otero
  • Jesus San-Miguel
  • Razi Ghori
  • Patricia Marinello
  • Sundar Jagannath
  • Anna Spreafico
  • Victor Lee
  • Roger K. C. Ngan
  • Ka Fai To
  • Myung Ju Ahn
  • Quan Sing Ng
  • Ruey-Long Hong
  • Jin-Ching Lin
  • Ramona F. Swaby
  • Christine Gause
  • Sanatan Saraf
  • Anthony T. C. Chan
  • Elaine Lam
  • Nizar M. Tannir
  • Funda Meric-Bernstam
  • Ulka Vaishampayan
  • Keith W. Orford
  • Chris Molineaux
  • Matt Gross
  • Andy MacKinnon
  • Sam Whiting
  • Martin Voss
  • Evan Y. Yu
  • Haiyan Wu
  • Charles Schloss
  • Mark R. Albertini
  • Erik A. Ranheim
  • Jacquelyn A. Hank
  • Cindy Zuleger
  • Thomas McFarland
  • Jennifer Collins
  • Erin Clements
  • Sharon Weber
  • Tracey Weigel
  • Heather Neuman
  • Greg Hartig
  • David Mahvi
  • MaryBeth Henry
  • Jacek Gan
  • Richard Yang
  • Lakeesha Carmichael
  • KyungMann Kim
  • Stephen D. Gillies
  • Paul M. Sondel
  • Vivek Subbiah
  • Ravi Murthy
  • Lori Noffsinger
  • Kyle Hendricks
  • Marnix Bosch
  • Jay M. Lee
  • Mi-Heon Lee
  • Edward B. Garon
  • Jonathan W. Goldman
  • Felicita E. Baratelli
  • Dorthe Schaue
  • Gerald Wang
  • Frances Rosen
  • Jane Yanagawa
  • Tonya C. Walser
  • Ying Q. Lin
  • Sharon Adams
  • Franco M. Marincola
  • Paul C. Tumeh
  • Fereidoun Abtin
  • Robert Suh
  • Karen Reckamp
  • William D. Wallace
  • Gang Zeng
  • David A. Elashoff
  • Sherven Sharma
  • Steven M. Dubinett
  • Nina Bhardwaj
  • Philip Friedlander
  • Anna C. Pavlick
  • Marc S. Ernstoff
  • Brian Gastman
  • Brent Hanks
  • Mark R. Albertini
  • Jason J. Luke
  • Tibor Keler
  • Tom Davis
  • Laura A. Vitale
  • Elad Sharon
  • Patrick Danaher
  • Chihiro Morishima
  • Martin Cheever
  • Steven Fling
  • Christopher R. Heery
  • Joseph W. Kim
  • Elizabeth Lamping
  • Jennifer Marte
  • Sheri McMahon
  • Lisa Cordes
  • Farhad Fakhrejahani
  • Ravi Madan
  • Kwong Tsang
  • Caroline Jochems
  • Rachel Salazar
  • Maggie Zhang
  • Christoph Helwig
  • Jeffrey Schlom
  • James L Gulley
  • Roger Li
  • John Amrhein
  • Zvi Cohen
  • Monique Champagne
  • Ashish Kamat
  • M. Angela Aznar
  • Sara Labiano
  • Angel Diaz-Lagares
  • Manel Esteller
  • Juan Sandoval
  • Ignacio Melero
  • Susannah D. Barbee
  • David I. Bellovin
  • John C. Timmer
  • Nebiyu Wondyfraw
  • Susan Johnson
  • Johanna Park
  • Amanda Chen
  • Mikayel Mkrtichyan
  • Amir S. Razai
  • Kyle S. Jones
  • Chelsie Y. Hata
  • Denise Gonzalez
  • Quinn Deveraux
  • Brendan P. Eckelman
  • Luis Borges
  • Rukmini Bhardwaj
  • Raj K. Puri
  • Akiko Suzuki
  • Pamela Leland
  • Bharat H. Joshi
  • Todd Bartkowiak
  • Ashvin Jaiswal
  • Casey Ager
  • Midan Ai
  • Pratha Budhani
  • Renee Chin
  • David Hong
  • Michael Curran
  • William D. Hastings
  • Maria Pinzon-Ortiz
  • Masato Murakami
  • Jason R. Dobson
  • David Quinn
  • Joel P. Wagner
  • Xianhui Rong
  • Pamela Shaw
  • Ernesta Dammassa
  • Wei Guan
  • Glenn Dranoff
  • Alexander Cao
  • Ross B. Fulton
  • Steven Leonardo
  • Kathryn Fraser
  • Takashi O. Kangas
  • Nadine Ottoson
  • Nandita Bose
  • Richard D. Huhn
  • Jeremy Graff
  • Jamie Lowe
  • Keith Gorden
  • Mark Uhlik
  • Laura A. Vitale
  • Thomas O’Neill
  • Jenifer Widger
  • Andrea Crocker
  • Li-Zhen He
  • Jeffrey Weidlick
  • Karuna Sundarapandiyan
  • Venky Ramakrishna
  • James Storey
  • Lawrence J. Thomas
  • Joel Goldstein
  • Henry C. Marsh
  • Tibor Keler
  • Jamison Grailer
  • Julia Gilden
  • Pete Stecha
  • Denise Garvin
  • Jim Hartnett
  • Frank Fan
  • Mei Cong
  • Zhi-jie Jey Cheng
  • Marlon J. Hinner
  • Rachida-Siham Bel Aiba
  • Corinna Schlosser
  • Thomas Jaquin
  • Andrea Allersdorfer
  • Sven Berger
  • Alexander Wiedenmann
  • Gabriele Matschiner
  • Julia Schüler
  • Ulrich Moebius
  • Christine Rothe
  • Olwill A. Shane
  • Brendan Horton
  • Stefani Spranger
  • Thomas F. Gajewski
  • Dayson Moreira
  • Tomasz Adamus
  • Xingli Zhao
  • Piotr Swiderski
  • Sumanta Pal
  • Marcin Kortylewski
  • Alyssa Kosmides
  • Kevin Necochea
  • Jonathan Schneck
  • Kathleen M. Mahoney
  • Sachet A. Shukla
  • Nikolaos Patsoukis
  • Apoorvi Chaudhri
  • Hung Pham
  • Ping Hua
  • Xia Bu
  • Baogong Zhu
  • Nir Hacohen
  • Catherine J. Wu
  • Edward Fritsch
  • Vassiliki A. Boussiotis
  • Gordon J. Freeman
  • Amy E. Moran
  • Fanny Polesso
  • Lisa Lukaesko
  • Andrew Weinberg
  • Emelie Rådestad
  • Lars Egevad
  • Jonas Mattsson
  • Berit Sundberg
  • Lars Henningsohn
  • Victor Levitsky
  • Michael Uhlin
  • William Rafelson
  • John L. Reagan
  • Loren Fast
  • Pottayil Sasikumar
  • Naremaddepalli Sudarshan
  • Raghuveer Ramachandra
  • Nagesh Gowda
  • Dodheri Samiulla
  • Talapaneni Chandrasekhar
  • Sreenivas Adurthi
  • Jiju Mani
  • Rashmi Nair
  • Amit Dhudashia
  • Nagaraj Gowda
  • Murali Ramachandra
  • Alexander Sankin
  • Benjamin Gartrell
  • Kerwin Cumberbatch
  • Hongying Huang
  • Joshua Stern
  • Mark Schoenberg
  • Xingxing Zang
  • Ryan Swanson
  • Michael Kornacker
  • Lawrence Evans
  • Erika Rickel
  • Martin Wolfson
  • Sandrine Valsesia-Wittmann
  • Tala Shekarian
  • François Simard
  • Rodrigo Nailo
  • Aurélie Dutour
  • Anne-Catherine Jallas
  • Christophe Caux
  • Aurélien Marabelle
Open Access
Meeting abstracts

Adoptive Cellular Therapy

O1 IL-15 primes an mTOR-regulated gene-expression program to prolong anti-tumor capacity of human natural killer cells

Andreas Lundqvist1, Vincent van Hoef1, Xiaonan Zhang1, Erik Wennerberg2, Julie Lorent1, Kristina Witt1, Laia Masvidal Sanz1, Shuo Liang1, Shannon Murray3, Ola Larsson1, Rolf Kiessling1, Yumeng Mao1

1Karolinska Institutet, Stockholm, Stockholms Lan, Sweden; 2Weill Cornell Medical College, New York, NY, USA; 3Nova Southeastern University, Cell Therapy Institute, Fort Lauderdale, FL, USA
Correspondence: Andreas Lundqvist (andreas.lundqvist@ki.se)

Background

NK cell-based immunotherapy is a potential therapeutic modality in patients with advanced cancers as transfer of haploidentical NK cells induces beneficial responses in patients with hematological malignancies; and leukemia clearance correlates with persistence and in vivo expansion of NK cells after infusion. Thus, sustained NK cell activity in vivo likely represents a therapy performance-limiting factor.

Methods

We performed genome-wide analysis of cytosolic and polysome-associated mRNA from interleukin (IL)-2 and IL-15 activated NK cells. Furthermore, the ability of IL-2 and IL-15 to sustain human NK cell activity following cytokine withdrawal as well as their effect on NK cells to resist tumor-induced immunosuppression was compared.

Results

After cytokine withdrawal, IL-15-treated NK cells maintained a higher level of cytotoxicity (p < 0.05) and showed lower levels of apoptosis (p < 0.05) compared with cells treated with IL-2. IL-15 augmented mTOR signaling, which correlated with increased expression of genes related to cell metabolism and respiration. Consistently, mTOR inhibition abrogated IL-15-induced cell function advantages. Moreover, mTOR-independent STAT-5 signaling contributed to improved NK cell function during cytokine activation but not following cytokine withdrawal. Upon co-culture with tumor cells or exposure to tumor cell supernatant, IL-15 activated NK cell maintained a significantly higher level of proliferation and cytotoxic activity (p < 0.05). Mechanistically, tumor-derived prostaglandin-E2 suppressed IL-2 cultured NK cells while IL-15 cultured NK cells remained activated. The superior performance of IL-15 stimulated NK cells was also observed using a clinically applicable protocol for NK cell expansion in vitro and in vivo.

Conclusions

This study adds to our understanding about establishment and maintenance of tumor-reactive NK cells and supports clinical implementation of IL-15 for adoptive NK cell therapy. More broadly, our studies suggest that a large aspect of cytokine-mediated gene expression programs and downstream cellular functions, including anti-tumor capacity, are overlooked if post-activation conditions are omitted. This is likely not limited to NK cells and should hence be considered in similar studies of other immune cells.

Biomarkers and Immune Monitoring

O2 ImmunoMap: a novel bioinformatics tool for analysis of T cell receptor repertoire data in model systems and clinical settings

John-William Sidhom1, Catherine A Bessell2, Jonathan Havel3, Jonathan Schneck4, Timothy A Chan3, Eliot Sachsenmeier5

1Johns Hopkins University School of Medicine, Baltimore, MD, USA; 2Immunology Program, Johns Hopkins University, School of Medicine, Columbia, MD, USA; 3Memorial Sloan Kettering Cancer Center, New York, NY, USA; 4Johns Hopkins Medical Institute, Baltimore, MD, USA; 5University of Rochester, Monrovia, MD, USA
Correspondence: John-William Sidhom (jsidhom1@jhmi.edu)

Background

There has been a dramatic increase in T cell receptor (TCR) sequencing spurred, in part, by the widespread adoption of this technology across academic medical centers and by the rapid commercialization of TCR sequencing. While the raw TCR sequencing data has increased, there has been little in the way of approaches to parse the data in a biologically meaningful fashion. The ability to parse this new type of ‘big data’ quickly and efficiently to understand the T cell repertoire in a structurally relevant manner has the potential to open the way to new discoveries about how the immune system is able to respond to insults such as cancer and infectious diseases.

Methods

Here we describe a novel method utilizing phylogenetic and sequencing analysis to visualize and quantify TCR repertoire diversity. To demonstrate the utility of the approach, we have applied it to understanding the shaping of the CD8 T Cell response to self (Kb-TRP2) and foreign (Kb-SIY) antigens in naïve and tumor bearing B6 mice. Additionally, this method was applied to tumor infiltrating lymphocytes (TIL’s) from patients undergoing Nivolumab (anti-PD-1) therapy in a clinical trial for metastatic melanoma to understand TCR repertoire characteristics between responders and non-responders.

Results

Analysis of the naïve CD8 response to SIY showed a lower clonality yet more closely structurally related response whereas CD8 responses to TRP2 were highly clonal yet less structurally related. Presence of tumor exhibited interesting differential effects on SIY vs. TRP2. We believe that differences in TCR repertoire suggest effects from central and peripheral tolerance on self vs. foreign antigens. In clinical trial data, the phylogenetic analysis revealed unique TCR repertoire signatures that differentiated responders from non-responders to anti-PD-1 therapy, including some that could be detected prior to initiation of therapy. Additionally, this analysis revealed that patients whose CD8 response had a larger contribution from novel and unique structural clones responded better to therapy.

Conclusions

In summary, we have developed and demonstrated a novel method to meaningfully parse and interpret TCR repertoire data and have applied it to yield a novel understanding of CD8 T Cell responses to different types of antigens as well as key characteristics in those who respond to anti-PD-1 therapy.
Fig. 1

(Abstract O2). a Weighted Phylogenetic Trees Comparing Kb-SIY vs Kb-TRP2 TCR Repertoire. Size of circles proportional to frequency of sequence. Color of circle corresponds to V-Beta Usage. b Dominant Motifs gathered from phylogenetic trees determined by homologous sequences and their contribution to the response c V-Beta Usage of Kb-SIY vs Kb-TRP2 Response

Fig. 2

(Abstract O2). a Weighted Phylogenetic Trees comparing Naive to Tumor-Bearing TCR Repertoire from spleen. Size of Circles proportional to frequency of sequence. Blue Circles = Naive Repertoire. Red Circles = Tumor-Bearing Repertoire. b V-Segment Usage of Kb-SIY vs Kb-TRP2 Responses for Naive vs Tumor-Bearing Response

Fig. 3

(Abstract O2). Examples of weighted phylogenetic trees from four cohorts of responders on anti-PD-1 therapy. Size of circles proportional to frequency of sequence. Blue = TCR repertoire prior to therapy. Red = TCR repertoire on therapy

O3 Increased STAT3 signaling and decreased suppressive function of regulatory T cells are biomarkers of positive patient outcome to nivolumab therapy

David Woods1, Anders Berglund2, Rupal Ramakrishnan2, Andressa Sodre1, Jeffrey Weber1

1NYU Langone Medical Center, New York, NY, USA; 2H. Lee Moffitt Cancer Center, Tampa, FL, USA
Correspondence: David Woods (david.woods@nyumc.org)

Background

Antibody-mediated blockade of the inhibitory receptor PD-1 on T cells has shown clinical efficacy in the treatment of various malignancies. However, biomarkers of response and mechanisms of resistance remain largely unidentified. To address this gap, we sought to identify the role(s) of regulatory T cells (Tregs) in metastatic melanoma patients treated with the PD-1 antibody nivolumab.

Methods

Pre and post-treatment Tregs were isolated from the peripheral blood of surgically resected stage III/IV metastatic melanoma patients treated with adjuvant nivolumab. Suppressive capacity was assessed in an allogeneic mixed lymphocyte reaction. Paired (pre vs. post-treatment) Tregs were assessed by flow cytometry for phosphorylated STAT3 (pSTAT3) expression. Finally, paired Treg samples were assessed for gene expression by RNA-sequencing.

Results

Tregs from non-relapsing patients demonstrated a significant decrease in suppressive capacity post-treatment (p < 0.05). However, suppressive capacity in relapsing patients did not decrease and their Tregs were significantly more suppressive post-treatment relative to non-relapsers (p < 0.01). Significantly increased levels of pSTAT3 post treatment were observed in non-relapsers (p < 0.05) but not in relapsers (p < 0.40). Significantly increased pSTAT3 was not seen in conventional T cells after nivolumab therapy. Culturing treatment-naïve T cells with PD-1 blocking antibodies in vitro resulted in increased levels of pSTAT3 in Tregs compared to IgG controls (p < 0.01). In vitro PD-1 blockade also significantly increased the number of Tregs (p < 0.01), and significant increases were seen in paired patient samples (p < 0.05). Paired analysis of Treg RNA-seq data using Panther and GeneGo. Metacore showed several significantly increased pathways associated with proliferation in non-relapsers. Changes in these pathways were absent in relapsers. Gene Set Enrichment Analysis of non-relapser Tregs showed significant (q=8.2e-18) overlap with known STAT3 target genes. Conversely, Enrichr analysis of relapsers showed significant upregulation of STAT1 and STAT2 target genes. No overlap of significantly changed gene expression or pathways in Tregs vs. conventional CD4+ T cells were observed.

Conclusions

These results highlight the potential importance of Tregs in mediating benefit with PD-1 blockade, demonstrating pSTAT3 induction and reduced suppressive capacity as biomarkers of clinical benefit. PD-1 blockade also increased the percentages of Tregs, consistent with the known roles of STAT3 in promoting cell survival and proliferation. RNA-seq data demonstrated increased STAT3 and proliferation associated gene expression. Intriguingly, Tregs from relapsing patients had increased expression of genes associated with STAT1/2 signaling, warranting further investigation of these pathways. In addition to highlighting STAT signaling as a biomarker of relapse, these results demonstrate distinct differences in the impact of PD-1 blockade in Treg vs. conventional T cells.

O4 Analysis of pharmacodynamic biomarkers in the first in-human trial of GITR co-stimulation with the agonist antibody TRX-518 in advanced solid cancer patients

Roberta Zappasodi1, Yanyun Li1, Jingjing Qi2, Philip Wong2, Cynthia Sirard3, Michael Postow4, Walter Newman3, Henry Koon5, Vamsidhar Velcheti6, Margaret K Callahan7, Jedd D Wolchok4, Taha Merghoub1

1Ludwig Collaborative Laboratory, Memorial Sloan Kettering Cancer Center, New York, NY, USA; 2Immune Monitoring Core Facility, Memorial Sloan Kettering Cancer Center, New York, NY, USA; 3Leap Therapeutics, Cambridge, MA, USA; 4Department of Medicine, Memorial Sloan Kettering Cancer Center, New York, NY, USA; 5Case Western Reserve University, Cleveland, OH, USA; 6Cleveland Clinic Main Campus, Cleveland, OH, USA; 7Memorial Sloan Kettering Cancer Center, New York, NY, USA
Correspondence: Roberta Zappasodi (zappasor@mskcc.org)

Background

GITR is a tumor necrosis factor receptor expressed at high levels on regulatory T cells (Tregs) and up-regulated on T cells upon activation. GITR stimulation abrogates Treg suppression and enhances T cell effector function. These observations suggest that GITR could be an attractive target for immunotherapy with agonist antibodies. GITR stimulation in tumor-bearing mice has shown therapeutic activity associated with both Treg reduction and modulation. Here we report results of pharmacodynamic analyses in the first in-human phase I trial with the fully humanized agonist anti-GITR antibody TRX518 as monotherapy in patients with advanced refractory solid tumors.

Methods

Patients were accrued to 9 cohorts (up to 6 patients/cohort) to receive a single dose of TRX518 (dose range: 0.0001-8 mg/kg). Pharmacodynamic analyses included flow cytometric evaluation of frequency and phenotype of circulating T cells and cytokine quantification in serum samples at different time points up to 12 weeks after treatment. Relevant changes observed with these analyses were monitored in pre- and post-treatment tumor biopsies by immunofluorescence staining.

Results

Here we report results obtained in 37 patients treated with ≥0.005 mg/kg TRX518 (cohorts 3-9), including 6 melanoma, 7 non-small cell lung cancer (NSCLC) and 7 colorectal cancer (CRC) patients and 17 patients with 11 other solid tumors. Among the T cell parameters analyzed, we found frequent reduction in circulating Tregs after treatment with TRX518 across all cohorts, with some exceptions. Importantly, this effect could be maintained over the 12-week observation period. When the analysis was performed by disease type, it revealed a pronounced TRX518 dose-dependent down-regulation of peripheral Tregs in both melanoma and CRC patients. Interestingly, in NSCLC cancer patients, Tregs did not always decrease after treatment. In a subset of patients (n=6; 2 melanoma, 2 CRC, 2 lung), for whom we had pre- and post-treatment tumor biopsies in addition to PBMCs, we tested whether intra-tumor Tregs were consistently affected. In melanoma and CRC patients, intra-tumor Foxp3+ Tregs were significantly reduced after treatment, in agreement with the peripheral Treg down-modulation observed in the same patients. In lung cancer patients, lack of circulating Treg reduction was consistently associated with stable or increased intra-tumor Treg infiltration after TRX518.

Conclusions

Circulating Treg reduction is a potential pharmacodynamic biomarker of TRX518 biological activity. This parameter may allow predictive correlation with changes in intratumoral Treg infiltration. We plan to further investigate this effect and its relevance for the association with clinical responses in our recently opened TRX518 multi-dose study.

Trial Registration

ClinicalTrials.gov identifier NCT01239134.

Consent

Written informed consent was obtained from the patient for publication of this abstract and any accompanying images. A copy of the written consent is available for review by the Editor of this journal.

Bispecific Antibodies

O5 Clinical responses in advanced pancreatic patients treated with bispecific antibody armed T cells (BATS)

Lawrence G. Lum1, Minsig Choi2, Archana Thakur1, Abhinav Deol3, Gregory Dyson3, Anthony Shields3

1University of Virginia Cancer Center, Charlottesville, VA, USA; 2Stony Brook University Medical Center, Stony Brook, NY, USA; 3Karmanos Cancer Institute, Detroit, MI, USA
Correspondence: Lawrence G. Lum (lgl4f@virginia.edu)

Background

Conventional chemotherapy (chemo) for locally advanced pancreatic cancer (LAPC) and metastatic pancreatic cancer (PC) is associated with dismal responses and poor survival rates. Arming activated T cells (ATC) with anti-CD3 x anti-EGFR bispecific antibody (EGFRBi) turns every ATC into a non-MHC restricted EGFR-specific cytotoxic T lymphocyte [1]. Engagement of CD3 on T cells and EGFR on Mia PACA-2 leads to cytokine secretion, proliferation, cytotoxicity by ATC and inhibition of tumor growth [2]. An earlier study using Infusions of anti-CD3 x anti-HER2 (HER2Bi) armed ATC in metastatic breast cancer provided encouraging survival (OS = 36 months) and evidence of anti-breast cancer immunity [3].

Methods

In this study, we used anti-CD3 x anti-EGFR bispecific antibody (EGFRBi)-armed T cells (EGFR BATs) to target EGFR in 5 metastatic PC patients and 6 colorectal cancer patients treated at Karmanos Cancer Institute on Protocol #2014-025 in a phase I dose escalation involving 3 weekly infusions of 10, 20, and 40 x 109 BATs/infusion followed by a booster infusion 3 months later.

Results

In the 5 PC patients, we report 1 patient was stable for 6.5 months and 2 patients in whom infusions of EGFR BATs may have “sensitized” the tumor to subsequent chemotherapy. The patient with stable disease had a near partial response. The median overall survival in 5 patients is 23.5 months with the median time to progression (TTP) of 7.0 months. Patient IT20102 received BATs and was stable (decreased marker lesion by 27%) at 6.5 mos. IT20091 had a remarkable clinical response to chemotherapy after progressing after immunotherapy at 4.6 months. After 3 BATs infusions, patient IT2010 had a “flare” or progression and subsequently had a complete response to Xeloda and remains in remission. This phase I study shows: 1) long-term stabilization in one patient; 2) a persistent complete responder after BATs "progression" followed by chemotherapy; 3) improved chemotherapy responsiveness after EGFRBi-BATs therapy; and 4) two patients with slow progressive disease who survived beyond 400 days. Survival for the 5 patients was 13.6, 14.5, 23.3 (alive in CR), 24.9 (alive, stable), and 31.0 months after enrollment, respectively (as of 7-20-16).

Conclusions

Targeting PC with EGFR BATs resulted in improved survival and remarkable post-immunotherapy chemotherapy responses in a small series of patients. The series provides evidence for anti-tumor activity of EGFR BATs as well as evidence that BATs infusions can sensitize tumors to subsequent chemotherapy.

Acknowledgements

Funding for this study was provided by Helen Kay Trust and Philanthropy and Startup Funds at KCI. We acknowledge the efforts of clinical coordinating staff, the clinical trials office staff, GMP laboratory staff, and clinical nursing support staff to making this study possible. The study was conducted at KCI.

References

1. Reusch U, Sundarum M, Davol PA, Olson SD, Davis JB, Demel K, et al: Anti-CD3 x anti-EGFR Bispecific Antibody Redirects T Cell Cytolytic Activity to EGFR-Positive Cancers In Vitro and in an Animal Model.CCR 2006, 12:183-190.

2. Grabert RC, Cousens LP, Smith JA, Olson S, Gall J, Young WB, et al: Human T Cells Armed with Her2/neu Bispecific Antibodies Divide, Are Cytotoxic, and Secrete Cytokines with Repeated Stimulation.CCR 2006, 12:569-576.

3. Lum LG, Thakur A, Al-Kadhimi Z, Colvin G, Cummings F, Legare R, et al: Targeted T cell Therapy in Stage IV Breast Cancer: A Phase I Clinical Trial.CCR 2015, 21:2305-2314.
Table 1

(Abstract O5). Clinical data

Pt

Age

Disease

Prior Tx

BATS (x 109)

TTP (mo)

OS (mo)

Comments

IT20087

58

Mets to liver

Folfirinox

47

6 mo

Died 13.6 mo

Progressed after Immunotherapy

IT20091

63

T3 N1Mets to liver. S/P Whipple

5FU, Leu/5FU Folfirinox

9 79

4.8 mo

Died 31 mo

Folfirinox induced CR after IT and responded a 2nd time to Folfirinox

IT20092

64

T2b Abd Nodes, S/P Whipple

Gemzar, 5FU, radiation

36

7 mo

Died 14.5 mo

Slowly progress with chronic diarrhea

IT20102

56

T4, Mets to liver, lungs

Folfirinox

74

6.5 mo

Alive 24.9 mo

Progressed after 6.5 mo

IT20104

51

T4, Abd Node

FOLFOX, X eloda

72

2.2 mo

Alive 23.3 mo

Chemo Induced CR after IT; On Xeloda

Combinations: Immunotherapy/Immunotherapy

O6 Reactivating the anti-tumor immune response by targeting innate and adaptive immunity in a phase I/II study of intratumoral IMO-2125 in combination with systemic ipilimumab in patients with anti-PD-1 refractory metastatic melanoma

Cara Haymaker1, Marc Uemura1, Ravi Murthy1, Marihella James1, Daqing Wang2, Julie Brevard2, Catherine Monaghan2, Suzanne Swann2, James Geib2, Mark Cornfeld2, Srinivas Chunduru2, Sudhir Agrawal2, Cassian Yee1, Jennifer Wargo1, Sapna P Patel1, Rodabe Amaria1, Hussein Tawbi1, Isabella Glitza1, Scott Woodman1, Wen-Jen Hwu1, Michael A Davies1, Patrick Hwu1, Willem W Overwijk1, Chantale Bernatchez1, Adi Diab1

1University of Texas MD Anderson Cancer Center, Houston, TX, USA; 2Idera Pharmaceuticals, Inc., Cambridge, MA, USA
Correspondence: Cara Haymaker (chaymaker@mdanderson.org)

Background

While checkpoint inhibitor (CPI) therapy has transformed metastatic melanoma (MM) treatment, many patients remain refractory. We reasoned that combining CPI with an agent that activates antigen presenting cells and improves T cell priming may result in improved response. Our approach is to modulate the tumor microenvironment through intratumoral (i.t.) injection of the TLR9 agonist, IMO-2125, in combination with ipilimumab (ipi). We hypothesize that this will result in dendritic cell (DC) activation and induction of tumor-specific CD8+T cells which will synergize with ipilimumab to overcome immune-escape. Based on this rationale we initiated a phase I/II clinical trial.

Methods

Adults with refractory MM despite up to 2 lines of CPI including PD-1 blockade therapy (with or without a BRAF inhibitor) are eligible. IMO-2125, in doses escalating from 4mg to 32mg, is given i.t. weeks 1, 2, 3, 5, 8, and 11 along with ipilimumab i.v. 3 mg/kg weeks 2, 5, 8, and 11. Dose-limiting toxicity (DLT) is evaluated using a modified Toxicity Probability Interval design. Primary endpoints are safety, tumor response, and PK. Blood and injected and distal tumor biopsies are obtained pre- and on-treatment. Immune analyses include DC subsets and their activation status as well as T cell activation, function and proliferation. T cell repertoire diversity will be evaluated by high throughput CDR3 sequencing.

Results

As of August 2, 2016, 11 pts have been enrolled. DLT has not been observed. Grade 3 hypophysitis (2 subjects) is the only immune-related AE observed to date. No other drug-related grade 3-5 AEs were documented and only 1 subject experienced a grade 2 fever. Five patients are evaluable for response - 2 PR, 2SD, 1PD per investigator assessment. Fresh tumor biopsies show maturation (upregulation of HLA-DR) of the myeloid DC1 subset (CD1c+CD303-) in the IMO-2125 injected tumor lesion 24 hrs post-treatment compared to pre-treatment biopsy. On-treatment biopsy results are consistent with a higher rate of proliferative (Ki67) effector CD4+ and CD8+ T cells in responders. Cytokine analysis shows a 2-3 fold increase in circulating IFNγ levels compared to pretreatment in responders.

Conclusions

Though preliminary, these results demonstrate that the combination of ipi and IMO-2125 is well tolerated with encouraging preliminary activity in a PD-1 refractory population. Dose escalation is ongoing and a phase II expansion will include IMO-2125 in combination with both ipi and anti-PD-1. Updated safety, antitumor activity, and biomarker data will be presented.

Trial Registration

ClinicalTrials.gov identifier NCT02644967.

O7 Clinical safety and efficacy assessment of the CD137 agonist urelumab alone and in combination with nivolumab in patients with hematologic and solid tumor malignancies

Erminia Massarelli1, Neil H Segal2, Vincent Ribrag3, Ignacio Melero4, Tara C Gangadhar5, Walter Urba6, Dirk Schadendorf7, Robert L Ferris8, Roch Houot9, Franck Morschhauser10, Theodore Logan11, Jason J Luke12, William Sharfman13, Fabrice Barlesi14, Patrick A Ott15, Laura Mansi16, Shivaani Kummar17, Gilles Salles18, Cecilia Carpio19, Roland Meier20, Suba Krishnan20, Dan McDonald20, Matthew Maurer20, Xuemin Gu20, Jaclyn Neely20, Satyendra Suryawanshi20, Ronald Levy17, Nikhil Khushalani21

1University of Texas MD Anderson Cancer Center, Houston, TX, USA; 2Memorial Sloan Kettering Cancer Center, New York, NY, USA; 3Institut Gustave Roussy, Villejuif, Ile-de-France, France; 4Center for Applied Medical Research (CIMA), University of Navarra, Pamplona, Navarra, Spain; 5University of Pennsylvania, Philadelphia, PA, USA; 6Earle A. Chiles Research Institute, Providence Cancer Center, Portland, OR, USA; 7Universitätsklinikum Essen, Essen, Nordrhein-Westfalen, Germany; 8University of Pittsburgh, Pittsburgh, PA, USA; 9CHU Rennes, Service Hématologie Clinique and INSERM 0203, Unité d'Investigation Clinique, Rennes, Bretagne, France; 10Centre Hospitalier Régional Universitaire de Lille, Lille, Nord-Pas-de-Calais, France; 11Simon Cancer Center, Indiana University, Indianapolis, IN, USA; 12University of Chicago School of Medicine, Chicago, IL, USA; 13Johns Hopkins University School of Medicine, Lutherville, MD, USA; 14Multidisciplinary Oncology and Therapeutic Innovations, Hôpital Nord, Marseille, Provence-Alpes-Cote d'Azur, France; 15Dana-Farber Cancer Institute, Boston, MA, USA; 16Centre Hospitalier Régional Universitaire Hôpital Jean Minjoz, Besançon, Franche-Comte, France; 17Stanford University School of Medicine, Stanford, CA, USA; 18Hospices Civils de Lyon-Université de Lyon, Pierre Benite, Auvergne, France; 19Hospital Universitari Vall d'Hebron, Universitat Autònoma de Barcelona, Barcelona, Catalonia, Spain; 20Bristol-Myers Squibb, Princeton, NJ, USA; 21H. Lee Moffitt Cancer Center, Tampa, FL, USA
Correspondence: Erminia Massarelli (emassarelli@coh.org)

Background

Urelumab is a fully human CD137 agonistic monoclonal antibody (mAb) that enhances T cell and natural killer (NK) cell antitumor activity in preclinical models. Nivolumab, a fully human programmed death-1 (PD-1) mAb that blocks the inhibitory function of the PD-1 receptor on T cells, has shown single-agent activity in many advanced malignancies. We hypothesized that the distinct, non-redundant mechanisms of these two mAbs could enhance antitumor activity without compromising safety. Here we report safety/tolerability, pharmacodynamics, and preliminary efficacy of urelumab and urelumab plus nivolumab combination therapy in patients with advanced malignancies.

Methods

The monotherapy study evaluated urelumab in patients with advanced malignancies (0.1 or 0.3 mg/kg Q3W) or advanced non-Hodgkin lymphoma (8 mg Q3W or Q6W). The combination study evaluated urelumab (3 or 8 mg Q4W) plus nivolumab (3 mg/kg or 240 mg Q2W) in patients with advanced solid tumors or B cell lymphoma (dose escalation) or patients with diffuse large B cell lymphoma (DLBCL), melanoma, non-small cell lung cancer (NSCLC), or squamous cell carcinoma of the head and neck (SCCHN; cohort expansion). Based on preliminary safety/tolerability/pharmacokinetic assessments of urelumab, cohort expansion focused on flat doses of 8 mg.

Results

Overall, patients who received urelumab monotherapy (N=123) experienced infrequent treatment-related serious AEs (7%) and treatment-related AEs (TRAEs) leading to discontinuation (5%; Table 2). In 104 patients treated with urelumab plus nivolumab (melanoma, n=40; NSCLC, n=20; SCCHN, n=22; DLBCL, n=22), the most frequent TRAE was fatigue (26%); grade 3/4 ALT/AST elevations (3%/3%) and TRAEs leading to discontinuation (7%) were infrequent. No treatment-related deaths were reported. Urelumab stimulated peripheral IFN-γ–induced cytokine production; induction was greater with urelumab plus nivolumab. In most melanoma tumors evaluated, a trend toward increased T and NK cell number and expression of IFN-γ and CXCL9 was observed upon treatment with the combination. Six patients with lymphoma treated with urelumab monotherapy had a partial (n=3) or complete (n=3) remission. Nine of 86 evaluable patients treated with the combination had partial responses (melanoma, n=8; SCCHN, n=1); no patients with NSCLC or DLBCL had confirmed responses at the interim analysis. Of 71 patients treated with urelumab plus nivolumab with RECIST/IWG assessments, 33 had reductions in tumor burden (Fig. 4).
Table 2

(Abstract O7). Treatment-related safety events

Patients, n (%)

Urelumab monotherapy N=23

Urelumab + nivolumab N=104

Treatment-related AEs

65 (53)

65 (53)

Most frequent treatment-related AEsa

  

Fatigue

18 (15)

27 (26)

AST increased

16 (13)

9 (9)

ALT increased

12 (10)

13 (13)

Treatment-related grade ¾ AST elevation

4 (3)

3 (3)

Treatment-related ¾ ALT elevation

3 (2)

3 (3)

Treatment-related serious AEs

9 (7)

10 (10)

Treatment-related AEs leading to discontinuation

6 (5)

7 (7)

Treatment-related deaths

0

0

AE, adverse event; AKT, alanine aminotransferase; AST, aspartate aminotransferase.

aTreatment-related AEs occurring in ≥10% of all patients

Fig. 4

(Abstract O7). Best percent reduction in target lesion tumor burden with urelumab plus nivolumab

Conclusions

Urelumab with or without nivolumab is safe/tolerable at flat and weight-based doses of 8 mg and 0.1 mg/kg. Although urelumab has demonstrated single-agent pharmacodynamic and antitumor activity in lymphoma, combination with nivolumab did not appear to provide significant additive/synergistic clinical benefit at the doses evaluated.

Trial Registration

ClinicalTrials.gov identifier NCT01471210 and NCT02253992.

O8 Beyond immune checkpoint: first-in-class antibody targeting soluble NKG2D ligand sMIC for cancer immunotherapy

Jennifer Wu, Jinyu Zhang, Fahmin Basher, Mark Rubinstein

Medical University of South Carolina, Charleston, SC, USA
Correspondence: Jennifer Wu (wujjd@musc.edu)

Background

In response to oncogenic insult, human cells were induced to express a family of MHC I-chain related molecules A and B (MICA and MICB, generally termed MIC) on the surface which serve as the ligands for the activating immune receptor NKG2D expressed by all human NK, CD8 T, NKT, and subsets of gamma-delta T cells. Theoretically, engagement of NKG2D by tumor cell surface MIC is thought to signal and provoke the immune system to eliminate transformed cells. Clinically, almost all advanced tumors in cancer patients produce soluble MIC through proteolytic shedding mediated by metalloproteases, or by release in exosomes derived from the cell membrane. Tumor-derived sMIC is known to be highly immune suppressive and profoundly insults the immune system by downregulating receptor NKG2D expression on effector NK and T cells, driving the expansion of tumor-favoring myeloid suppressor cells, skewing macrophages into alternatively activated phenotypes, and perturbing NK cell peripheral maintenance. High levels of serum sMIC significantly correlate with advanced diseases of many types of cancer. These observations clearly endorse sMIC to be a cancer immune therapeutic target. However, due to mice lacking homologues to human MIC, this concept was not proven until our recent studies.

Methods

Using a “humanized” MIC-transgenic spontaneous mouse model which recapitulates the NKG2D-mediated onco-immune dynamics of human cancer patients, we addressed whether sMIC is a cancer immunotherapeutic target and whether antibody targeting sMIC synergizes with immune checkpoint blockade or adoptive T or NK cell therapy.

Results

We show that therapy with a first-in-field non-blocking antibody B10 that does not block the interaction of MIC with NKG2D revamps endogenous innate and antigen-specific CD8+ T cell responses and remodels immune reactive tumor microenvironment, by restoring NK cell hemostatic maintenance and function, enhancing CD8+ T cell infiltration to tumors and TCR clonality/diversity, modulating CD8+ T cells metabolic preferences, eliminating MDSCs and TAMS. Anti-sMIC stand-alone therapy resulted in effective debulking of primary tumors and eliminated metastasis. Using multiple pre-clinical animal models, we further demonstrate that antibody B10 neutralizing sMIC synergizes with CTLA-4 and PD-1/PD-L1 checkpoint blockade therapy and adoptive cell based therapy with no observed toxicity.

Conclusions

Our study has launched a new avenue of cancer immunotherapy which can be readily translated into clinical trials.
Fig. 5

(Abstract O8). These are the examples in human cancer patients, prostate cancer, Oral cancer, and HBV-induced Liver cancer, where high levels of circulating sMIC correlates with advanced disease stages and poor survival.

Fig. 6

(Abstract O8). a. Therapy of the clinically relevant spontaneous prostate tumor TRAMP/MICB model (Liu et al, 2013, JCI 123 (10) 4410 ) with CuraB10 (also called B10G5) or control IgG (placebo) at advanced stage via I.P. injection at the dose of 3.8 mg/KG body weight twice weekly for 8 weeks (b). Mice with CuraB-1o therapy all enjoyed longtime survived whereas mice in placebo group are succumbed to cancer (c). Prostate weight. Comparisons made between Placebo group and CuraB-10 group and between before and after treatment of CuraB-10. (d). Representative images of the prostate. Top showing large tumor burden. Bottom showing normal prostate size. (e). All mice in the control group developed metastasis whereas no metastasis was detected in animals received CuraB-10 therapy. In summary, the data demonstrate that CuraB-10 stand-alone therapy can effectively induce regression of primary tumors and eliminate metastasis

Fig. 7

(Abstract O8). We further addressed the synergistic effect of CuraB-10 therapy with FDA approved checkpoint blockade therapy using the clinically relevant TRAMP/MIC spontaneous prostate tumor mouse model. Two points: 1) a percentage of TRAMP/MIC mice do not respond to checkpoint (CTLA4 or PD-1) blockade therapy, whereas all TRAMP/MIC mice are responsive to CuraB-10 therapy; also, a population of TRAMP/MICB animals died at 3-4 weeks of CTLA4 Rx alone. 2) CuraB-10 synergizes with checkpoint blockade (CTLA4 or PD-1) therapy when used in combination

Fig. 8

(Abstract O8). Because rodents do not express MIC, we engineered mouse melanoma B16 tumor cells to express sMIC (B16-sMIC). We implanted B16-sMIC into syngeneic host. When tumors grew to 50-100mm3 in size, treatment starts. Four treatments were given: Adoptive transfer of melanoma antigen-specific Pmel CD8 T cells once, B10G5 (CuraB-10) twice 2 week Adoptive transfer of melanoma antigen-specific Pmel CD8 T cells once, control IgG (cIgG) twice 2 week B10G5 alone 4) cIgG alone Tumor growth curve demonstrating that treatment with B10G5 (CuraB-10) effectuates the effect of Pmel CD8 T therapy Survival curve. Tumor volume of 1000mm3 was defined as survival end point. In one experiment, 2/7 animals received B10G5 and Pmel therapy had complete tumor regression. Note: currently Adoptive T cell transfer (ACT) requires prior-depletion of patient’s immune cells with chemotherapy to be effective. With B10G5 therapy, not only lymph depletion is not required prior to ACT, but also ACT is more effective

Diet, Exercise and/or Stress and Impact on the Immune System

O9 β-adrenergic signaling induced by cool housing temperatures mediates immune suppression and impairs the efficacy of anti-PD-1 checkpoint blockade immunotherapy in laboratory mice

Mark Bucsek, Guanxi Qiao, Cameron MacDonald, Bonnie Hylander, Elizabeth Repasky

Roswell Park Cancer Institute, Buffalo, NY, USA
Correspondence: Mark Bucsek (mark.bucsek@roswellpark.org)

Background

Recent work from our laboratory has shown that anti-tumor immunity is suppressed in mice housed at standard temperatures (ST; 22°C) which could be reversed by housing mice at warmer, thermoneutral temperatures (TT; 30°C) [1]. However, the mechanisms causing this impairment at ST remain unclear. Cold stress is mediated specifically by activation of the sympathetic nervous system and the release of norepinephrine (NE), which is highly suppressive when signaling through β-adrenergic receptors (β-ARs) on immune cells. We found that NE levels are significantly elevated in tumor-bearing mice housed at ST compared to TT, which led us to hypothesize that chronic stress induced by cool housing temperatures increases β-AR signaling that dampens the anti-tumor immune response and the efficacy of immune modulating therapies.

Methods

We used both physiologic (housing temperature; ST and TT) and pharmacologic blockade (β-blockers) to modulate β-AR signaling levels in immune-competent and SCID mice bearing 4T1 or B16-OVA tumors. Flow cytometry was used for immune cell analysis. Anti-PD-1 checkpoint blockade was given in 6, 200μg doses (Days 0, 2, 4, 6, 9, and 12) starting the day after tumors became detectable.

Results

We found that the addition of β-blockade significantly delayed 4T1 and B16-OVA tumor growth in mice housed at ST, recapitulating the slower tumor growth observed in mice housed at TT. However, β-blockade had no impact on tumor growth in SCID mice at ST or TT indicating dependence on the adaptive immune system. Analysis of 4T1 and B16-OVA tumors from immune-competent mice showed increased IFN-γ expression in both CD4+ and CD8+ T cells in mice treated with β-blockade indicating a more robust anti-tumor immune response. Lastly, we investigated the impact of β-AR signaling on anti-PD-1 checkpoint blockade efficacy and found that reducing β-AR signaling by both physiologic (TT) and pharmacologic (β-blockade) strategies improved responses in both tumor models. Further analysis of 4T1 tumors from mice treated with β-blockade and anti-PD-1 showed an increase in IFN-γ, producing CD8+ T cells compared to either β-blockade or anti-PD-1 alone.

Conclusions

Taken together, these data indicate that elevated β-AR stress signaling caused by cool housing temperatures impairs anti-tumor immunity and the response of tumors to anti-PD-1 checkpoint blockade.

Acknowledgements

Supported by: The Peter T. Rowley Breast Cancer Research Grant, The Harry J. Lloyd Charitable Trust, the Roswell Park Alliance Foundation, and 5T32CA085183-12.

Reference

1. Kokolus K, et al: Baseline tumor growth and immune control in laboratory mice are significantly influenced by subthermoneutral housing temperature. PNAS 2013, 110:20176-20181.

Immune Metabolism

O10 NAD-Sirt1 axis is central to the unique immuno-metabolic phenotype of Th1/17 hybrid cells in regulating its enhanced anti-tumor potential

Shilpak Chatterjee1, Anusara Daenthanasanmak1, Paramita Chakraborty1, Kyle Toth1, Megan Meek1, Elizabeth Garrett-Mayer1, Michael Nishimura2, Chrystal Paulos1, Craig Beeson1, Xuezhong Yu1, Shikhar Mehrotra1

1MUSC, Charleston, SC, USA; 2Loyola Cancer Center, Maywood, IL, USA
Correspondence: Shilpak Chatterjee (chatherj@musc.edu)

Background

Th17 cells hold promise for immunotherapy of cancer [1]. While the anti-tumor potential of Th17 cells primarily depends upon IFN-γ secretion and persistence [1], a long-term tumor control has still remained elusive. Given that both the “effector” and “stemness like” features are prerequisites for T cells to mount durable anti-tumor responses, we hypothesized that combining the culture conditions of Th1 (effector) and Th17 (stemness like) cells could generate hybrid Th1/17 cells with improved anti-tumor properties.

Methods

Melanoma epitope tyrosinase reactive CD4+ T cells obtained from h3T TCR transgenic mice were differentiated ex vivo to Th1, Th17, and Th1/17 cells before adoptive transfer (0.25×106 cells/animal i.v.) to C57BL/6 recipient animals with subcutaneously established B16 melanoma. Quantitative PCR (q-PCR), flow cytometry, and metabolomic analyses were used to evaluate the expression of various metabolism and stemness associated genes as well as protein expression in the T cells. To compare the metabolic commitment between different subsets (Th1, Th17 and Th1/17), real time metabolic flux analyzer (Seahorse Biosciences, USA) and radioactive tracer studies were used.

Results

The combined culture conditions of Th1 and Th17 generates hybrid Th1/17 cells with a IFN-γhi, IL17hi, GM-CSFhi, CD107ahi, T-bethi, Granzyme Bhi, IL23Rhi, IL22hi, Bcl6hi, Tcf7hi signature. These hybrid Th1/17 cells exhibit enhanced tumor control in subcutaneous and lung metastasis models of murine melanoma. A hypothesis generating transcriptional, metabolic, and proteomic profiling, followed by confirmatory experiments established that the enhanced anti-tumor properties were attributed to increased NAD+ mediated activity of histone deacetylase Sirt1 in hybrid Th1/17 cells. Inhibition of NAD+ and Sirt1 activity either pharmacologically or by genetic ablation (Sirt1-KO T cells) led to loss of stable anti-tumor control. Importantly, anti-tumor T cells or tumor infiltrating lymphocytes programmed in the presence of exogenous NAD+ also led to the similar metabolic phenotype and improved anti-tumor control.

Conclusions

The present study discloses that metabolic status plays an important role in dictating the anti-tumor response of the T cells. Combining the culture conditions of Th1 and Th17 cells renders hybrid Th1/17 cells with a unique immune-metabolic feature that enables them to orchestrate distinct transcriptional programs leading to highly effector and stem-like T cells.

Reference

1. Muranski P, Boni A, Antony PA, et al: Tumor-specific Th17-polarized cells eradicate large established melanoma.Blood 2008, 112:362-373.

O11 The Wnt5a-beta-catenin pathway triggers a metabolic switch that drives indoleamine 2,3-dioxygenase activity and dendritic cell tolerization in the melanoma microenvironment: optimizing checkpoint inhibitor immunotherapy

Fei Zhao1, Kathy Evans1, Christine Xiao1, Alisha Holtzhausen2, Brent A. Hanks1

1Duke University Medical Center, Durham, NC, USA; 2Lineberger Comprehensive Cancer Center, University of North Carolina, Chapel Hill, NC, USA
Correspondence: Brent A. Hanks (hanks004@mc.duke.edu)

Background

Despite recent advances, many cancers remain refractory to available immunotherapies by developing various strategies to evade the immune system. Emerging evidence indicates that the tolerization of local dendritic cells (DCs) within the tumor microenvironment plays a critical role in immune evasion. The role of metabolic re-programming in DC tolerization remains poorly characterized and the mechanisms by which cancers may utilize these pathways to promote the establishment of an immunotolerant microenvironment have not been described.

Methods

We investigated the role of the Wnt-beta-catenin pathway in the metabolic reprogramming of melanoma-derived DCs using real-time metabolic flux analysis. The impact of DC metabolic re-programming on the enzymatic activity of indoleamine 2,3-dioxygenase (IDO) was analyzed by HPLC while protoporphyrin IX(PpIX) levels were quantified by flow cytometry. The role of DC fatty acid oxidation (FAO) on regulatory T cell (Treg) generation was investigated using pharmacologic and genetic approaches. The impact of FAO inhibition on anti-tumor immune responses to anti-PD-1 antibody therapy were investigated in a transgenic melanoma model.

Results

We show that the Wnt5a-beta-catenin-PPARg pathway shifts DCs from glycolysis to FAO in the melanoma microenvironment in a manner dependent upon induction of the mitochondrial fatty acid transporter, CPT1A (Fig. 9). This metabolic shift promotes DC tolerization by 1) elevating DC levels of the PpIX prosthetic group of IDO, resulting in the enhanced activity of this enzyme (Fig. 10) and 2) potently suppressing DC-expression of IL-6 and IL-12, both culminating in the generation of Tregs both in vitro and in vivo (Fig. 11). Genetic silencing and the pharmacologic inhibition of CPT1A potently enhances the ability of DCs to stimulate effector T cell responses. Indeed, genetic silencing of melanoma-expressed Wnt5a significantly promotes T cell tumor infiltration and augments PD-L1 expression in this melanoma model. Consistent with these findings, we further show FAO inhibition to enhance the efficacy of anti-PD-1 therapy while augmenting melanoma antigen-specific T cell responses (Fig. 12).
Fig. 9

(Abstract O11). Wnt5a Promotes DC FAO in the Melanoma Microenvironment. A. Schematic of tumor-infiltrating DC (TIDC) metabolic analysis. B. Melanoma-derived Wnt5a promotes TIDC OXPHOS. C. Wnt5a promotes DC FAO

Fig. 10

(Abstract O11). Wnt5a-induced FAO Promotes DC Synthesis of PpIX and Enhances IDO Enzyme Activity. A. Wnt5a stimulates DC PpIX synthesis. B,C. Wnt5a promotes DC IDO activity in a FAO-dependent manner both in vitro and in vivo

Fig. 11

(Abstract O11). Wnt5a-induced DC OXPHOS Promotes Treg Generation in the Melanoma Microenvironment. Melanoma-derived Wnt5a conditions DCs to promote Treg generation in vivo

Fig. 12

(Abstract O11). Inhibition Wnt5a-DC FAO Enhances Melanoma PD-L1 Expression and Augments anti-PD-1 antibody Efficacy. A. Schematic of Wnt5a paracrine signaling pathway. B,C. Genetic silencing of Wnt5a in melanoma promotes T cell infiltration and PD-L1 upregulation. D. Inhibition of CPT1A/FAO synergizes with anti-PD-1 antibody therapy in melanoma

Conclusions

Our findings implicate the Wnt5a-beta-catenin-PPARg-CPT1A paracrine signaling axis as a driver of DC FAO and functional DC tolerization in the melanoma microenvironment and connect this pathway with the promotion of a “non-inflamed” phenotype in melanoma. This work describes a novel association between DC metabolism and the regulation of IDO enzymatic activity and suggests that this pathway may be a potent pharmacological target for increasing the responsiveness of “non-inflamed” tumors to anti-PD-1 antibody immunotherapy.

O12 Mitochondrial biogenesis is repressed in tumor-infiltrating CD8+ T cells resulting in metabolic insufficiency and T cell dysfunction

Nicole Scharping1, Ashley V Menk2, Rebecca Moreci2, Ryan Whetstone1, Rebekah Dadey1, Simon Watkins1, Robert Ferris1, Greg M Delgoffe1

1University of Pittsburgh, Pittsburgh, PA, USA; 2University of Pittsburgh Cancer Institute, Pittsburgh, PA, USA
Correspondence: Nicole Scharping (nes63@pitt.edu)

Background

CD8+ tumor-infiltrating T lymphocytes (CD8+ TIL) in the tumor microenvironment (TME) are unable to effectively control their tumor targets due to a variety of immunosuppressive mechanisms, including direct tumor cell-T cell inhibition and soluble immunosuppressive factors. This allows cancer to progress unchecked as T cells are rendered functionally inert. Recently, poor metabolite availability in the TME has been identified as an additional suppressive mechanism exploited by bioenergetically-dysregulated tumors. Because T cell activation also has robust metabolic demands, we hypothesized that CD8+ TIL dysfunction was a result of metabolic insufficiency.

Methods

Metabolic capacity was measured at the single cell level by 2NBDG and MitoTracker FM. Metabolic output was measured by Seahorse extracellular flux analysis. T cell reprogramming was performed by retroviral transduction on OVA-specific transgenic T cells in vitro before adoptive transfer into B16OVA bearing mice.

Results

We found CD8+ TIL are characterized by dramatic loss of mitochondrial mass in B16, MC38, and LLC implantable mouse tumors and human CD8+ TIL, which correlates with upregulation of co-inhibitory checkpoint molecules PD-1 and Tim-3. CD8+ TIL mitochondrial mass loss is caused by decreased mitochondrial biogenesis, due in part to repression of the transcriptional co-activator PGC1α resulting from chronic Akt signaling. Surprisingly, anti-PD-1 therapy had no effect on increasing PGC1α or mitochondrial mass in CD8+ TIL. We then asked whether improving CD8+ TIL metabolism genetically might result in enhanced effector function, so we reprogrammed tumor-specific CD8+ T cells to upregulate mitochondrial biogenesis prior to adoptive cell therapy. We found increased mitochondrial mass, restored cytotoxic functionality, and dramatically improved tumor regression in mice with reprogrammed CD8+ TIL. To better understand why mitochondrial loss causes T cell dysfunction, we are exploring the importance of mitochondria for T cell functionality, including ATP and nucleotide production, calcium buffering, and ROS production.

Conclusions

Our data support a model in which chronically-activated CD8+ TIL are unable to metabolically support their effector functions. By understanding these metabolic insufficiencies, we can both better understand T cell dysfunction and design metabolic modulation strategies to improve cancer immunotherapy.

Inflammation, Innate Immunity, and the Microbiome

O13 Intestinal microbiota and relapse after hematopoietic-cell transplantation

Jonathan Peled, Sean Devlin, Anna Staffas, Melissa Lumish, Kori Porosnicu Rodriguez, Katya Ahr, Miguel Perales, Sergio Giralt, Ying Taur, Eric Pamer, Marcel R. M. van den Brink, Robert Jenq

Memorial Sloan Kettering Cancer Center, New York, NY, USA
Correspondence: Jonathan Peled (peledj@mskcc.org)

Background

The major causes of mortality after allogeneic hematopoietic-cell transplantation (allo-HCT) are relapse, graft-versus-host disease (GVHD), and infection. We have previously reported that alterations in the intestinal flora are associated with GVHD, bacteremia, and reduced overall survival after allo-HCT. As intestinal bacteria are potent modulators of systemic immune responses including antitumor effects triggered by checkpoint blockade, we hypothesized that components of the intestinal flora could be associated with relapse after allo-HCT.

Methods

The intestinal microbiota of 541 patients admitted for allo-HCT was profiled by means of 16S ribosomal sequencing of prospectively collected stool samples. We hypothesized that evolutionarily related species exhibit functional similarities, and we therefore defined clusters of related operational taxonomic units (crOTUs) to evaluate for associations with clinical outcomes. To group OTUs by evolutionary distances, a phylogenetic tree was empirically constructed from a sequence alignment of all OTUs identified in the whole cohort (Fig. 13). We examined the relationship between abundance of microbiota species or groups of related species and relapse/progression of disease during two years of follow-up after allo-HCT using cause-specific Cox proportional hazards in a retrospective discovery-validation study (Fig. 14).
Fig. 13

(Abstract O13). Phylogenetic tree of OTUs and clusters of related operational taxonomic units (crOTUs). Each black point is a crOTU. Phylum is color coded along the circumference. Members of the same phyla were largely grouped together, indicating that the tree was broadly concordant with standard taxonomy

Fig. 14

(Abstract O13). Multivariate screening of microbial features for association with relapse. Volcano plot of multivariate p values of crOTUs against the multivariate hazard ratios for relapse/progression of disease in the discovery set. crOTUs are color coded by p value. Multivariate adjustment was performed for Disease Risk Index score, graft source, and conditioning intensity. The most abundant species in each of the labeled crOTUs are 1614: Eubacterium limosum. 2022-3: Streptococcus sinensis. 1638: Eubacterium limosum. 1630-1: Eubacterium limosum. 1790: Parvimonas micra. 0951-3: Leptotrichia hongkongensis 2986: Flavonifractor plautii. 1439: Actinomyces odontolyticus

Results

The intestinal presence of a group comprised mostly of Eubacterium limosum in the validation set was associated with less relapse/progression of disease (HR 0.52, CI 0.31–0.87, p = 0.01, Fig. 15). The two-year cumulative incidence of relapse/progression among patients with and without this group of bacteria was 33.8% and 19.8%, respectively. The relative abundance of this group was also associated with less relapse/progression of disease (HR 0.82, CI 0.71–0.95, p = 0.009). These associations remained significant in multivariate models and were strongest among recipients of T cell-replete allografts.
Fig. 15

(Abstract O13). crOTU 1614, which includes members of family Eubacteriaceae is associated with decreased relapse after allo-HCT.Cumulative incidence of relapse/POD in the discovery (n = 271) and validation (n = 270) sets stratified by presence or absence of crOTU 1614

Conclusions

We found associations between the abundance of a group of bacteria in the intestinal flora and relapse/progression of disease after allo-HCT. These might serve as potential biomarkers or therapeutic targets to prevent relapse and improve survival after allo-HCT.

Oncolytic Viruses

O14 Phase I/II CANON study: oncolytic immunotherapy for the treatment of non-muscle invasive bladder (NMIBC) cancer using intravesical Coxsackievirus A21

Nicola Annels1, Hardev Pandha1, Guy Simpson1, Hugh Mostafid2, Kevin Harrington3, Alan Melcher4, Mark Grose5, Bronwyn Davies5, Gough Au5, Roberta Karpathy5, Darren Shafren5

1University of Surrey, Guildford, England, UK; 2Royal Surrey County Hospital, Guildford, England, UK; 3Institute for Cancer Research, London, England, UK; 4The Institute for Cancer Research, London, England, UK; 5Viralytics, Inc., Sydney, New South Wales, Australia
Correspondence: Nicola Annels (n.annels@surrey.ac.uk)

Background

As a clinical setting in which local live biological therapy is already well established, non-muscle invasive bladder cancer (NMIBC) presents intriguing opportunities for oncolytic virotherapy. Coxsackievirus A21 (CVA21, CAVATAKTM) is a novel intercellular adhesion molecule-1 (ICAM-1)-targeted immunotherapeutic virus which exerts potent oncolytic activity against NMIBC cell lines and ex-vivo human bladder tumour. CVA21 in combination with low doses of Mitomycin C enhances CVA21 viral replication and oncolysis by increasing surface expression levels of ICAM-1.

Methods

A two stage Phase I/II study (CANON) was initiated to study the tolerance of escalating intravesicular (IV) doses of CVA21 administered alone or in combination with MitomycinC (10mg) in 16 first-line NMIBC cancer patients prior to TURBT surgery. Cystoscopy photography was performed before and after treatment. Tissues were analysed for CVA21 replication, apoptosis, changes in immune cell infiltrates (multi-spectral imaging) and immune-checkpoint molecules.

Results

IV administration of CAVATAK was well tolerated with no adverse events. Anti-cancer activity including viral induced tumour inflammation was demonstrated by serial cystoscopy including a complete response observed in one of 3 patients in the highest dose monotherapy cohort. Tumour targeting by CVA21 was shown by detection of secondary viral load peaks in the urine and by immunohistochemical analysis of TURBT tissue displaying tumour-specific viral replication and apoptotic cell death. Nanostring analysis revealed an upregulation of interferon-response and immune checkpoint inhibitory genes in CVA21-treated tissues compared to untreated historical controls. Notable changes in immune cell infiltrates and expression of PD-L1 within the CVA21-treated NMIBC tissue were also observed. Increased urinary levels of the chemokine, HMGB1, was observed in six of eleven patients following exposure to CVA21.

Conclusions

The utility of CVA21 as a potent immunotherapeutic agent has been demonstrated by the observed tumour targeting and viral replication. Upregulation of checkpoint molecules following CVA21 exposure may also allow potential sequential combination therapies with checkpoint targeting.

Trial Registration

ClinicalTrials.gov identifier NCT02316171.

O15 Pre-existing immunity to oncolytic virus potentiates its therapeutic efficacy.

Jacob Ricca1, Taha Merghoub2, Jedd D Wolchok3, Dmitriy Zamarin1

1Memorial Sloan Kettering Cancer Center, New York, NY, USA; 2Ludwig Collaborative Laboratory, Memorial Sloan Kettering Cancer Center, New York, NY, USA; 3Department of Medicine, Memorial Sloan Kettering Cancer Center, New York, NY, USA
Correspondence: Jacob Ricca (riccaj92@gmail.com)

Background

Despite the significant promise of oncolytic viral (OV) therapy in preclinical models, clinical efficacy of systemically-administered viruses has proven to be modest. One major limitation of the systemic OV therapy is neutralization of the virus by pre-existing immunity, or development of neutralizing antibodies shortly after therapy initiation, which limit viral delivery to tumor sites. Recently, we and others have demonstrated that intratumoral therapy with OV can lead to systemic anti-tumor immunity and abscopal effects, and several clinical trials are currently exploring intratumorally administered OVs in patients. The effect of pre-existing anti-viral immunity or the development of new anti-viral immunity on the anti-tumor efficacy, however, is not well defined.

Methods

Using oncolytic Newcastle Disease Virus (NDV) as a model, we explored the effect of pre-existing immunity to the virus on its therapeutic efficacy using syngeneic B16-F10 melanoma and MB49 bladder carcinoma models.

Results

BL6 mice were immunized with NDV and subsequently implanted with B16 or MB49 murine cancer cells. Immunized and naïve tumor-bearing mice were treated intratumorally with NDV. As expected, pre-immunized animals demonstrated decreased levels of NDV replication. Surprisingly, pre-existing immunity to the virus did not decrease the antitumor efficacy and led to superior tumor clearance and long-term animal survival. Analysis of tumor-infiltrating lymphocytes from the treated animals demonstrated marked increase in infiltration with CD8+ and CD4+FOXP3- cells, and significant decrease in CD4+FOXP3+ cells, an effect that was significantly more pronounced in the pre-immunized animals. This was observed in both virus-injected and contralateral flank tumors, in absence of viral spread to distant tumor sites. Concurrent adoptive transfer of luciferase-tagged tumor-specific Trp-1 lymphocytes demonstrated increased intratumoral accumulation of Trp-1 cells in pre-immunized mice. Furthermore, lymphocytes isolated from tumors of NDV-treated pre-immunized mice produced more IFNg than those of NDV-treated naïve mice when cultured with tumor cells in vitro, suggestive of antigen spreading. Finally, in an animal model of recurrent cancer after “cure” with NDV, re-treatment with NDV resulted in regression of tumors and long-term animal survival, an effect accompanied by significant increase in tumor-infiltrating immune cells.

Conclusions

Our findings demonstrate that pre-existing immunity to OVs might not deter, and even augment the efficacy of intratumoral OV therapy, which is likely mediated by enhanced tumor-specific immune response. This is a clinically-relevant question, which suggests that prior anti-viral immunity should not be a deterrent to OV therapy with locoregional administration, though it remains to be demonstrated whether such findings would translate to other oncolytic viruses.

Promoting and Measuring Anti-Tumor Immunity

O16 Immunoscore® Colon analytical performance

Luciana Batista1, Florence Marliot2, Angela Vasaturo3, Sabrina Carpentier4, Cécile Poggionovo1, Véronique Frayssinet1, Jacques Fieschi1, Marc Van den Eynde5, Franck Pagès6, Jérôme Galon3, Fabienne Hermitte1

1HalioDx, Marseille, Provence-Alpes-Cote d'Azur, France; 2Université Paris Descartes, APHP, Paris, Ile-de-France, France; 3INSERM, Paris, Ile-de-France, France; 4MI-mAbs, Marseille, Provence-Alpes-Cote d'Azur, France; 5Université Catholique de Louvain, Brussels, Brussels Hoofdstedelijk Gewest, Belgium; 6APHP, Paris, Ile-de-France, France
Correspondence: Fabienne Hermitte (fabienne.hermitte@haliodx.com)

Background

The Immunoscore® was validated as a powerful prognostic marker in colon cancer in a study conducted by the Immunoscore® worldwide consortium, led by the Society for Immunotherapy of Cancer (SITC) involving 23 pathology centers from 17 countries, and including more than 3800 patients. HalioDx has developed a standardized version of the test that was used in this study. Here we show the concordance with the research version and present the main analytical performances of the system.

Methods

For each colon tumor block, 2 slides are stained using an automated IHC staining instrument: one with CD3 and one with CD8. Digital images of stained slides are obtained using a whole slide scanner, and analyzed by a software program (Immunoscore® Analyzer, HalioDx). The Immunoscore® Analyzer automatically processes images for tissue detection (core of the tumor, CT and invasive margin, IM). Densities of positive lymphocytes in the CT and IM are reported. For each marker and each zone, densities distributions have been established in the SITC study training set. The Immunoscore® is reported in 5 categories from 0 to 4, or as IS High (IS3 and IS4), Low (IS0 and IS1) & intermediate (IS2). Precision of HalioDx Immunoscore® Colon assay in terms of repeatability and reproducibility was evaluated using commercial FFPE colon cancer blocks, with 152 independent stainings from 4 samples, corresponding to 62 CD3 and CD8 pairs. Intra-block and inter-block variability were assessed from 8 additional blocks. Accuracy based on inter-laboratory concordance was determined using 119 samples. The European Hospital Georges Pompidou (HEGP - center of reference for the SITC study) workflow was used as reference.

Results

The inter-instrument, inter-lot and inter-operator/-reader precision in terms of cell density (cells/mm2) CV were below 12%, 22% and 18%, respectively. Only 1 change in Immunoscore® category (out of 62 IS assessments) was observed, from IS1 to IS0. The equivalency between HalioDx and HEGP workflows was assessed in terms of cell densities. Deeming regression slopes were not significantly different from 1 for both CD3 and CD8 antibodies. Pearson correlation coefficients were above 0.89. The concordance table is provided in Fig. 16, corresponding to a weighted Cohen’s kappa coefficient of 0.88.
Fig. 16

(Abstract O16).

Conclusions

The Immunoscore® Colon is a robust, easy-to-use and accurate assay. It is the first standardized immune-based assay for the classification of cancer.

References

1. Galon J, et al: J Pathol 2014, 232(2):199-209.

2. Galon J, et al: JCO 2016, 34(15_suppl):3500.

O17 Mechanisms of chitosan/IL-12 immunotherapy for the treatment of bladder cancer

Sean G. Smith1, Khue Nguyen2, Sruthi Ravindranathan3, Bhanu Koppolu1, David Zaharoff1

1Joint Department of Biomedical Engineering, North Carolina State University and the University of North Carolina, Raleigh, NC, Cary, NC, USA; 2Cell and Molecular Biology, University of Arkansas, Fayetteville, AR, USA; 3Biomedical Engineering, University of Arkansas, Fayetteville, AR, USA
Correspondence: Sean G. Smith (sgseangr@ncsu.edu)

Background

Bladder cancer afflicts 430,000 people every year globally and is plagued by recurrence rates as high as 50%. One way to mitigate the risk of recurrence is by engaging adaptive immunity. Our group has been able to direct adaptive immunity via intravesical treatment with CS/IL-12, a coformulation of interleukin(IL)-12 and the biopolymer chitosan. Four twice-weekly administrations of CS/IL-12 routinely eliminate more than 90% of orthotopic bladder tumors in mice while providing systemic protection from recurrence and rechallenge for the duration of the lifespan of treated mice. The purpose of this study is to gain insights into the mechanisms underlying both the initial elimination and later rejection of bladder tumors by exploring the importance of the number of administrations, lymphocyte subtypes, and the immune cell infiltration throughout and following treatment.

Methods

Female C57BL/6J mice were implanted orthotopically with 75,000 MB49 bladder cancer cells. Beginning 7 days after implantation, mice were treated intravesically 2x/week for two weeks with CS/IL-12 (1 μg). The importance of the number of treatments was investigated by monitoring survival while varying the treatment number. The role of lymphocyte subtypes was investigated by monitoring survival after depleting CD4+, CD8+, or NK1.1+ cells prior to and throughout treatment or rechallenge. Cellular responses 24 hours after each treatment were measured in the bladder, bladder draining lymph nodes (BDLNs), and the spleen via flow cytometry.

Results

Varying the number of treatments revealed that a single administration significantly extended survival beyond saline with 4/10, 2/8, 6/9, and 7/8 mice surviving long term after 1, 2, 3, or 4 applications respectively. Depletion studies showed a dependence on CD8+ T cells for tumor elimination (Fig. 17a) and on CD4+ T cells for rejection of subsequent tumor rechallenge (Fig. 17b). Flow cytometry revealed fluctuations in the immune-cell populations over the course of treatment (Fig. 18). The first treatment was characterized by a 54% increase of macrophages in the bladder and a 56% increase in the CD8:Regulatory T cell ratio in the BDLNs. By the third treatment there was an influx of CD4+ and CD8+ T cells in the bladder as well as increased CD8+ T cells in the BDLNs.
Fig. 17

(Abstract O17). Depletion studies reveal role of T cell subtypes. Mice were depleted of NK1.1+, CD4+, or CD8+ cells either throughout treatment with CS/IL-12 (a) or rechallenge at a distant site (b). Significant differences (P<0.05) in median survival by the Log-Rank test are indicated by *

Fig. 18

(Abstract O17). Immune infiltrates after each treatment. Bladder tumor tissues and bladder draining lymph nodes (BDLNs) were disassociated and analyzed via flow cytometry 24 hours after intravesical treatment with either PBS or CS/IL-12. Regulatory T cells were defined as CD3+CD4+CD25+FoxP3+. Each symbol represents an individual mouse. Significant differences (P<0.05) by T-tests are indicated by *

Conclusions

Even a single administration of CS/IL-12 eliminates established tumors, though higher rates of survival were possible with 3 or 4 treatments. The initial response is inflammatory and driven by macrophage infiltration and CD8+ T cells while the memory responses is directed by CD4+ T cells.

Survivorship Issues Related to Immunotherapy

O18 Incidence and outcomes of central nervous system metastasis in metastatic melanoma patients treated with anti-PD-1 therapy

Gustavo Schvartsman, Roland Bassett, Jennifer L McQuade, Lauren E Haydu, Michael A Davies, Hussein Tawbi, Isabella Glitza

University of Texas MD Anderson Cancer Center, Houston, TX, USA
Correspondence: Gustavo Schvartsman (gschvartsman@mdanderson.org)

Background

Central nervous system (CNS) metastasis are common in patients with metastatic melanoma (MM) and represent a frequent site of treatment failure with current therapies. However, little is known about the incidence, characteristics and outcomes of CNS metastasis in MM patients treated with anti-programmed death-1 (PD-1) and in conjunction with more intensive local CNS treatment strategies.

Methods

Under an IRB-approved protocol, outcomes of MM patients treated with anti-PD-1 at The University of Texas MD Anderson Cancer Center from January 2012 to February 2016 were reviewed. The association between development of CNS metastasis and overall survival (OS) was assessed using Cox regression analysis with time to CNS metastasis treated as a time-varying covariate.

Results

We identified 264 MM patients who received anti-PD-1 treatment, including 74 (28%) who had CNS metastasis prior to the first dose of anti-PD-1. With a median follow-up of 10.4 months (range 0-51.6) from the start of this therapy, 37 (19% of patients without prior CNS metastasis) developed CNS metastasis after the initiation of anti-PD-1. Of those, 27 patients were diagnosed with CNS metastasis during anti-PD-1 or within 90 days of treatment discontinuation, and 10 patients were diagnosed with CNS mets >90 days after last anti-PD-1 dose. The majority of these patients were male (62%), their mean age at new CNS metastasis was 62 years (range 31-86), and most patients had a history of cutaneous primary (59%). Of the 26 patients who were tested for mutations, BRAF was identified in 8 (22%, V600E in 6 patients, V600K in 2 patients), NRAS in 5 (14%) and KIT in 6 (16%). 86% received at least one CNS directed treatment approach. 62% were treated with stereotactic radiosurgery, 11% received whole-brain radiation and 30% underwent surgery. Median OS from start of anti-PD-1 was 34 months (range 0-51.6 months) for the whole anti-PD-1 treatment cohort. Development of CNS metastasis while on anti-PD-1 therapy was strongly significantly associated with death (HR 3.39, 95% CI 2.06, 5.59, p < .0001).

Conclusions

To our knowledge, this is the first report describing the incidence of CNS metastasis as an initial site of disease progression in MM patients treated with anti-PD-1 and associated with worse OS, despite additional CNS directed therapy.

Tumor Microenvironment

O19 CD8α+ dendritic cells regulate leukemia antigen-specific CD8+ T cell tolerance

Douglas Kline1, Xiufen Chen2, Dominick Fosco2, Justin Kline3

1Committee on Immunology, University of Chicago, Chicago, IL, USA; 2Department of Medicine, University of Chicago, Chicago, IL, USA; 3Committee on Immunology and Department of Medicine, University of Chicago, Chicago, IL, USA
Correspondence: Justin Kline (jkline@medicine.bsd.uchicago.edu)

Background

Batf3-lineage CD8α+ and CD103+ dendritic cells (DCs) are required for the spontaneous priming of CD8+ T cells against solid tumors. In contrast, the APCs that regulate immune responses against hematological malignancies have not been characterized. Syngeneic transplantable and genetically-engineered acute myeloid leukemia (AML) models associated with a dense CD8+ T cell tolerant state were employed to identify the APCs responsible for inducing T cell tolerance in vivo.

Methods

Transplantable C1498 and genetically-engineered Mx1-Cre x LSLAML1-ETO/+ x FLT3ITD/ITDx R26-LSLSIY/+ (MAFFS) AML models were employed to characterize APCs involved in generating leukemia-specific CD8+ T cell tolerance.

Results

Following systemic introduction of viable, CellTrace violet-labeled AML cells, leukemia cell-derived fluorescence was observed exclusively within splenic CD8α+ DCs, whereas uptake of proteins from dead AML cells was mediated by CD11b+ macrophages. CD8α+ DCs were also uniquely capable of cross-presenting leukemia antigens to CD8+ T cells directly ex vivo. Interestingly, antigen encounter by leukemia-specific CD8+ T cells was severly reduced in Batf3-/- mice, indicating that CD8α+ DCs mediate T cell recognition of leukemia antigens, and that their absence is associated with immunological ignorance of AML antigens. Moreover, leukemia-specific CD8+ T cells in wildtype AML-bearing mice failed to respond following vaccination with the tolerizing antigen, while those in leukemia-bearing Batf3-/- mice expanded vigorously, demonstrating that CD8α+ DCs induce leukemia-specific tolerance in vivo. Activation of CD8α+ DCs with the TLR-3 agonist, poly(I:C) restored functional anti-leukemia T cell responses and protected mice from disease progression in a Batf3-dependent manner. RNA-seq analysis of "tolerogenic" versus "naive" CD8α+ DCs from leukemia-bearing mice revealed ~200 differentially expressed genes in the former, suggesting that tolerance induction by CD8α+ DCs is an active process.

Conclusions

Our data support a growing body of evidence that has defined a prominent role for Batf3-dependent DCs in regulating anti-cancer immune responses. Batf3-lineage DCs generate functional CD8+ T cell responses against solid tumors, but actively and exclusively induce CD8+ T cell tolerance to systemic leukemia, indicating that the same DC lineage can imprint disparate T cell fates in mice with solid verses hematopoietic malignancies, and suggesting that environmental cues perceived by CD8α+ DCs may dictate their ability to activate or tolerize cancer-specific CD8+ T cells. These results highlight stark differences in the regulation of anti-cancer immunity in hosts with solid versus hematological malignancies.

O20 Neuropilin-1-deficient regulatory T cell-derived interferon-γ drives infectious instability and tumor clearance

Abigail Overacre1, Maria Chikina1, Erin Brunazzi1, Gulidanna Shayan2, William Horne1, Jay Kolls1, Robert L Ferris1, Greg M. Delgoffe1, Tullia C Bruno3, Creg Workman1, Dario Vignali1

1University of Pittsburgh, Pittsburgh, PA, USA; 2Tsinghua University, Pittsburgh, PA, USA; 3University of Pittsburgh/Department of Immunology, Pittsburgh, PA, USA
Correspondence: Abigail Overacre (overacre@pitt.edu)

Background

Regulatory T cells (Tregs) play an integral role in maintaining immune homeostasis; however, they are detrimental in cancer through suppression of the anti-tumor immune response. Therefore, identifying Treg targets that are specifically required in the tumor microenvironment is warranted. We have previously shown that the Neuropilin-1 (Nrp1) pathway is required for functional stability of intratumoral Tregs, but remains disposable in maintaining peripheral immune homeostasis. However, 1) the mechanisms that drive Treg functional instability, 2) the fate and impact of functionally unstable Nrp1-deficient (Nrp1–/–) Tregs on the tumor microenvironment and 3) how NRP1 affects function in human Tregs remain unknown.

Methods

In order to further understand the role of Nrp1–/– Tregs in cancer, we injected B16.F10 melanoma into Nrp1L/LFoxp3Cre-YFP/DTR-GFP cellular heterozygous mice comprised of 50% WT Tregs and 50% Nrp1–/– Tregs. Using these mice, we performed whole transcriptome sequencing to determine global transcriptomic changes. Once identified, differentially regulated pathways were tested both ex vivo through functional assays and in vivo with Treg transfers into Foxp3–/– mice. Lastly, we obtained head and neck squamous cell carcinoma and metastatic melanoma samples to determine the abundance and function of NRP1 on human Tregs.

Results

Using Nrp1L/LFoxp3Cre-YFP/DTR-GFP mice, we found that intratumoral Nrp1–/– Tregs produce interferon-γ (IFNγ), driving the functional destabilization of surrounding WT Tregs, which in turn boosts antitumor immunity and facilitates tumor clearance. Furthermore, we have shown that NRP1 is expressed on a proportion of TIL Tregs in head and neck cancer as well as metastatic melanoma and that the IFNγ pathway is likely conserved in human Tregs. In addition, human TIL Tregs pre-treated with IFNγ show significantly reduced suppressive function in comparison to those without pre-treatment.

Conclusions

Overall, we have shown that Nrp1 is required for functional stability of intratumoral Tregs, and in its absence, there is an alteration in the tumor microenvironment, leading to an enhanced anti-tumor immune response. These studies uncover a novel potential target for cancer immunotherapies that preserves peripheral immune health. This is of clinical interest, given that NRP1 is expressed on select Tregs in human melanoma and head and neck cancer and that NRP1+ Tregs show a suppressive advantage over NRP1 Tregs.

Adoptive Cellular Therapy

P1 Converting tumor-mediated PD-L1 inhibition into CAR T cell costimulation to potentiate adoptive T cell therapy

Prasad S. Adusumilli

Memorial Sloan Kettering Cancer Center, New York, NY, USA
Correspondence: Prasad S Adusumilli (adusumip@mskcc.org)

Background

To overcome tumor-mediated inhibition of chimeric antigen receptor (CAR) T cells, we herein investigated the impact of tumor PD-L1 upregulation on CAR T cell exhaustion and anti-tumor efficacy, and further developed clinically translatable T cell-extrinsic as well as -intrinsic strategies to overcome PD-L1 inhibition in models of lung cancer (LC) and malignant pleural mesothelioma (MPM).

Methods

Human T cells were transduced with MSLN-specific CAR with CD28 and CD3zeta domains (M28z) were tested in vitro and in clinically-relevant LC and MPM mouse models by bioluminescence imaging (BLI) of tumor burden progression. To counteract PD-1/PD-L1 inhibition in vivo, we evaluated the efficacy of PD-1 blocking antibody or cell-intrinsic genetic-engineering strategies by cotransducing M28z CAR T cells with a PD-1 dominant negative receptor (PD1-DNR) or with PD-1/4-1BB fusion protein.

Results

A single, low-dose ofM28z CAR T cells is able to resist the progression of established tumor for 40 days, but mice eventually died with progressing tumor. Tumor harvest analysis demonstrated the PD-1 and PD-L1 upregulation on CAR T cells and tumor cells (Figure Panel A). We then confirmed in vitro that PD-L1 inhibits M28z T cell effector functions (proliferation, cytotoxicity and cytokine secretion). The addition of PD-1 blocking potentiates CAR T cell therapy in vivo but its efficacy requires multiple injections (Panel B). In contrast, a single dose of M28z T cells coexpressing PD-1-DNR restore effector functions, enhance tumor burden control (Panel C) and prolong median survival (56 vs 82 days, p=0.001). Converting PD-L1 inhibition into a positive costimulatory signal by PD-1/4-1BB construct cotransduction into M28z CAR T cells enhanced cytokine secretion and T cell accumulation (Panel D).

Conclusions

Our results demonstrate the therapeutic benefit of providing optimal costimulation and coinhibitory blockade to counteract PD-L1/PD-1 immunosuppression, thus potentiate CAR T cell therapy for lung cancer and mesothelioma.

References

1. Cherkassky L, Morello A, Villena-Vargas J, Feng Y, Dimitrov DS, Jones DR, et al: CAR T cells with cell-intrinsic checkpoint blockade resist tumor-mediated inhibition. J Clin Invest 2016, PMID: 27454297.
Fig. 19

(Abstract P1).

P2 Lack of moesin improves adoptive T cell therapy by potentiating anti-tumor functions

Ephraim A Ansa-Addo, Zihai Li

Medical University of South Carolina, Charleston, SC, USA
Correspondence: Ephraim A Ansa-Addo (ansaaddo@musc.edu)

Background

Moesin is a member of the ezrin-radixin-moesin (ERM) protein family that are crucial for organizing membrane domains [1]. However, the role of ERM proteins in regulating signal transduction activities is still less clear and identifying new target proteins regulated by the ERMs for drug targeting remains an important area within the field due to their increased levels in multiple cancers. Whether ERM proteins play any role during the differentiation of naïve CD4+ T cells to TGF-β-induced Tregs is completely unknown.

Methods

We utilized a combination of knockout (MsnKO) mice, polyribosome profiling, RT-PCR and immunoblotting to demonstrate that a lack of moesin promotes efficient adoptive T cell therapy in mice by controling translational upregulation of moesin by TGF-β in T cells.

Results

The lack of moesin led to poorer development and function of both peripherally-inducible Tregs and in vitro-induced Treg cells (Fig. 20). We found that the loss of moesin significantly delayed tumor recurrence in a mouse model of melanoma and supported the rapid expansion of adoptively transferred CD8 + T cells against cancer-associated antigens (Fig. 21). Of note, moesin knockout CD4+ T cells exhibited no defects in T cell receptor activation, proliferation or cytokine production, suggesting no alternations in T cell activation in these mice. Instead, our data indicate that moesin interacts with TGF-β receptor II and controls its surface abundance and stability (Fig. 22). Indeed, the lack of moesin significantly impaired optimal TGF-β signaling (Fig. 23) and improved adoptive T cell therapy under cancer setting (Fig. 24).
Fig. 20

(Abstract P2). a Immunoblot of moesin knockout (KO or MsnKO) in multiple organs of mice. b Representative enlarged spleen size and increased cell number as observed in MsnKO mice. Data are reported as the mean ± SEM; ***P <0.001 by Student’s t-test. n = 6 per group. c and d Flow cytometry of inflammatory cytokines, TNF-α and IL-4 (c), and IFN-γ and IL-17A (d) produced by CD4+ T cells from spleens of WT and MsnKO mice after PMA-Ionomycin stimulation for 4 h. Data are reported as the mean ± SEM; *P <0.05 and ***P <0.001 by Student’s t-test. WT n = 3, MsnKO n = 4. e and f Flow cytometry analysis and absolute number of pTregs (Foxp3+Heliolo) in the small intestine lamina propria (sLP) (e) and colon lamina propria (cLP) (f) of 10-12 weeks old mice. Data are reported as the mean ± SEM; *P <0.05, **P<0.01 and ***P <0.001 by Student’s t-test. n = 4 per group, MsnKO sLP n = 3. g and h Flow cytometry analysis and absolute number of pTregs (Foxp3+Helioslo) in the spleen (g) and peripheral blood (h) of 10-12 weeks old mice. Data are reported as the mean ± SEM; ns, not significant; *P <0.05 and **P <0.01 by Student’s t-test. n = 4 per group

Fig. 21

(Abstract P2). a Treatment scheme for B6 mice injected subcutaneously with B16-F1 melanoma tumor cells (2.5 x105) 7 days prior to lympho- 31 depletion with 6Gy total body irradiation and adoptive transfer (ACT) of 2 x106 Pmel-1 T cells (i.v.) at day 8. b Tumor growth kinetics in individual mice treated as indicated in A, (each line represents one mouse); WT (n=6), MsnKO (n=4). c Frequency of donor Pmel-1 CD8+ T cells circulating in the blood of WT and MsnKO mice from B at 8 weeks. d Survival analysis of WT and MsnKO mice upon B16 melanoma tumor injection and adoptive cell transfer. *P = 0.05 by Log-rank test (B and D), *P < 0.05 by Student’s t-test (C). Data are reported as the mean ± SEM

Fig. 22

(Abstract P2). a Immunoprecipitation (IP), with anti-TβRII and control immunoglobulin G antibodies of EL4 cell lysates, followed by immunoblot of pulldown and input samples with the indicated antibodies. b Confocal microscopy images of EL4 cells stimulated with TGF-β (5 ng/ml for 1 h) and stained with anti-TβRII and anti-moesin antibodies. c and d HEK293FT cells co-transfected with plasmids encoding wild-type moesin-tagged with CFP at the carboxy terminus (Msn-WT-CFP) and TβRII-tagged with haemagglutinin at the carboxy terminus (TβRII-HA). Immunoprecipitation of solubilised proteins using anti-GFP and anti-HA antibodies and immunoblot of the pull-down samples. Input - whole cell lysate immunoblotting (throughout). e-g Immunoprecipitation and immunoblot (as in c) of HEK293FT cells co-transfected with CFP-tagged wild-type or phosphomimetic moesin mutants and TβRII-HA constructs. Data are representative of at least three (a-c) or four (d-f) independent experiments. h and i Primary CD4+ T cells (h) and B220+ B cells (i) from the spleen of WT and MsnKO mice were treated with cyclohexamide at the indicated times and surface TβRII analyzed by flow cytometry. Data represents the mean ± SD of at least three independent experiments. ***P <0.001 by two-way analysis of variance (ANOVA). j Flow cytometry analysis of primary CD4+ T cells isolated from the spleen of WT and MsnKO mice, and treated with brefeldin A (BFA), 20 μg/ml for up to 5 h and then washed. Cell surface TβRII was left to recover for up to 12 h prior to analysis. Data represents the mean ± SD of at least three independent experiments. ***P <0.001 by two-way analysis of variance (ANOVA)

Fig. 23

(Abstract P2). a-c EL4 LAF (EL4) cells transduced with lentiviral scrambled vector (Scram) or moesin shRNA (MsnKD), stimulated with TGF-β (5 ng/ml) and analyzed for intracellular pSmad2/3 by flow cytometry at the times indicated. Data are representative of the mean ± SD of at least three independent experiments. *P <0.05 and **P <0.01 by Student’s t-test. (d and e) Analyses of phospho-Smad2/3 (pSmad2/3) by flow cytometry after 3-day cultures (d) and pSmad3 and total Smad3 by immunoblotting (e) stimulated at the indicated times in WT and MsnKO iTregs. Data are representative of at least two independent experiments.

Fig. 24

(Abstract P2). (i, ii) Under pathological conditions such as cancer, TGF-β production by tumor cells binds to cognate receptors on T cell surfaces and triggers signaling events that lead to Foxp3 expression and induced Treg cell development. (iii-iv) Signaling via the TGF-β non-canonical Akt-hnRNP E1-axis leads to post-transcriptional moesin expression. The TGF-β signaling pathway, Rho kinase and/or Phosphatidylinositol 4,5-bisphosphate[PtdIns(4,5)P2] pathways, lead to phosphorylation/activation of moesin and aids moesin binding to the F-actin. (v-viii) Moesin-F-acting binding may promote internalization of TGF-β receptors possibly via early endosomes which feeds forward to promote optimal TGF-β signaling leading to efficient Foxp3 induction and iTreg cell differentiation. (ix) Moesin may also promote efficient recycling of TGF-β receptors to maintain the abundance of TβRII on the cell surface. (x) Differentiated induced Treg cells then suppress the proliferation of other immune cells such as cytotoxic T lymphocytes and effector T (Teff) cells to limit anti-tumor responses and promote tumor progression

Conclusions

This finding is important and suggests that modulation of moesin (via inhibitors or agonists), such as developed recently for ezrin [2], could serve as a potential candidates for use in immunotherapy combinations for the treatment of cancer as well as advance our knowledge.

Acknowledgements

This work was supported by the US National Institutes of Health (R01DK098819 to D.C.R.; P01CA186866, R01CA188419 and R01AI070603 to Z.L.).

References

1. Hirata T, Nomachi A, Tohya K, Miyasaka M, Tsukita S, Watanabe T, Narumiya S: Moesin-deficient mice reveal a non-redundant role for moesin in lymphocyte homeostasis. Int Immunol 2012, 24(11):705-717.

2. Celik H, Bulut G, Han J, Graham GT, Minas TZ, Conn EJ, Hong SH, Pauly GT, Hayran M, Li X, Ozdemirli M, Ayhan A, Rudek MA, Toretsky JA, Uren A: Ezrin inhibition up-regulates stress response gene expression. J Biol Chem 2016, 291(25):13257-13270.

P3 Preclinical evaluation of an optimal-affinity MAGE-A4 T cell receptor for adoptive T cell therapy

Andrew Gerry, Joseph P Sanderson, Karen Howe, Roslin Docta, Qian Gao, Eleanor A L Bagg, Nicholas Tribble, Miguel Maroto

Adaptimmune, Oxfordshire, England, UK
Correspondence: Andrew Gerry (andrew.gerry@adaptimmune.com)

Background

Adoptive immunotherapy employing optimal affinity T cell receptor (TCR) engineered T cells is a highly attractive treatment modality for multiple cancer indications. In order to ensure the safety of novel T cell receptor therapies, it is important both that expression of the target antigen is tightly restricted to tumor, and that the TCR does not display off-target activity. Here we describe development of an optimal-affinity MAGE-A4 TCR for adoptive T cell therapy.

Methods

Expression profiling of the cancer-germline antigen MAGE-A4 was performed in tumor and normal tissues, determined by analyzing public RNAseq datasets and by in-house qPCR. We then generated an enhanced affinity TCR that recognizes a validated MAGE-A4 HLA-A*02 peptide, selected based on potency and specificity in in vitro testing from panels of engineered TCRs originating from multiple parental TCRs. The selected TCR was subject to full preclinical characterization using Adaptimmune’s extensive preclinical testing process. This process involves potency testing against both tumor cell lines and primary tumor tissue in 2D and 3D, and safety testing consisting of extensive screening of TCR-transduced T cell responses to a wide range of tumor lines, normal human primary cells and induced pluripotent stem cell-derived cells. In addition, the fine specificity of the TCR was characterized to allow the generation of a binding motif and the identification of putative mimotype peptides within the human proteome.

Results

The MAGE-A4 antigen was found to be highly over-expressed in several clinically important solid tumor indications, such as lung squamous cell cancer (60%), head and neck cancer (42%), bladder cancer (34%) and esophageal cancer (33%), while expression in non-tumor material was limited to expression in the testes and placenta, both immune-privileged tissues. We generated an enhanced-affinity TCR that demonstrated enhanced potency against MAGE-A4-expressing tumor cell lines and fresh tumor tissue, whilst retaining absence of relevant response against MAGE-A4-negative cells and non-MAGE peptide mimotypes.

Conclusions

MAGE-A4 is an attractive target antigen for adoptive T cell therapy using enhanced affinity TCRs. We have generated and characterized an optimal enhanced-affinity TCR, which shows enhanced potency against MAGE-A4-positive tumor targets whilst maintaining specificity. These data will be used to support an IND for the use of this TCR for investigatory clinical trials.

P4 Case report: specific peptide enhanced affinity receptor T cells (SPEAR® T cells) demonstrate long-term persistence and both in vivo and ex vivo tumoricidal activity

Gareth Betts1, Natalie Bath1, Luca Melchiori1, Daniel E Lowther1, Indu Ramachandran1, Gabor Kari1, Samik Basu1, Gwendolyn Binder-Scholl1, Karen Chagin1, Lini Pandite1, Tom Holdich1, Rafael Amado1, Hua Zhang2, John Glod2, Donna Bernstein2, Bent Jakobsen3, Crystal Mackall4

1Adaptimmune, Philadelphia, PA, USA; 2National Cancer Institute, Bethesda, MD, USA; 3Adaptimmune, Oxfordshire, England, UK; 4Stanford University School of Medicine, Stanford, CA, USA
Correspondence: Samik Basu (samik.basu@adaptimmune.com)

Background

SPEAR® T cells reactive against the NY-ESO-1 specific HLA-A02:01 restricted peptide (SLLMWITQC) have demonstrated clinical activity (ORR 50%) in patients (n=12) with advanced synovial sarcoma (SS). The mechanisms underlying tumor relapse in the presence of persisting SPEAR® T cells remain unclear. Here, we report on phenotypic and functional studies on both engineered T cells and tumor biopsies from a patient with a NY-ESO-1+ SS treated with NY-ESO-1C259 SPEAR® T cells.

Methods

Engineered T cell persistence was determined by qPCR for the vector backbone and flow cytometry for HLA-A2:01-SLLMWITQC reactive pentamer+ T cells in post-infusion PBMC samples. Multi-parameter flow cytometric analyses were performed on pre-infusion manufactured product and post-infusion PBMCs to assess memory subsets using CD45RA and CCR7, exhaustion using CD28 and PD-1, and functionality by IFN-ɣ and Gzmb. Tumor and NY-ESO-1C259 T cells from patient PBMCs were isolated at 28 months post-infusion to determine their ex vivo killing capacity against a NY-ESO-1+ cell line, A375 . Antigen expression and immunomodulatory milieu (e.g. PD-L1) in baseline and post-treatment biopsies were assessed by immunohistochemistry. Serum cytokines were measured by a Luminex based immunoassay. Tumor response was determined by RECIST v1.1.

Results

The patient achieved a partial response to NY-ESO-1C259 SPEAR® T cells with progression at 9 months post-infusion. Persistence at 28 months with NY-ESO-1C259 T cells was observed by qPCR and flow cytometry. Over the course of treatment, the phenotype of the engineered cells changed from a mix of TEMRA (CD45RA+CCR7-), TEM (CD45RA-CCR7-), and TSCM (CD45RA+CCR7+) populations at the time of infusion to a predominately TSCM (~98.7%) within five months. PBMC derived NY-ESO-1C259 SPEAR® T cells 28 months post-infusion exhibited substantial ex vivo killing of NY-ESO-1+ A375 cells without additional ex vivo re-stimulation. Pre- and post-infusion biopsies showed NY-ESO-1 expression and exhibited minimal to moderate leukocytic (CD45+) infiltration accompanied by minimal lymphocytic infiltration post-infusion. Of note, PD-L1 expression was exclusive to CD45+ cells.

Conclusions

Despite an initial response to NY-ESO-1C259 SPEAR® T cells, this patient eventually relapsed despite the persistence of functional SPEAR® T cells and antigen positive tumor . The basis for tumor progression following response remains unclear, but does not appear to result from T cell exhaustion. Other possibilities include loss of antigen expression and/or diminished tumor infiltration, which could result from the large peripheral TSCM population, known to traffic to lymphoid tissue rather than tumor.

Trial Registration

ClinicalTrials.gov identifier NCT01343043.

Consent

Written informed consent was obtained from patient for publication. A copy is available for editor review.

P5 Engineering 2nd generation SPEAR® T cells to overcome TGF-β-mediated immunosuppression for adoptive cell therapy

Ryan Wong1, Jonathan D Silk1, Katherine Adams1, Garth Hamilton1, Alan D Bennett1, Sara Brett2, Junping Jing2, Adriano Quattrini1, Manoj Saini1, Guy Wiedermann1, Andrew Gerry1, Bent Jakobsen1, Gwendolyn Binder-Scholl3, Joanna Brewer1

1Adaptimmune, Oxfordshire, England, UK; 2GSK, Stevenage, England, UK; 3Adaptimmune, Philadelphia, PA, USA
Correspondence: Andrew Gerry (andrew.gerry@adaptimmune.com)

Background

Adoptive cell therapy (ACT) with NY-ESO SPEAR® T cells, is showing promising initial clinical responses in phase I/II trials for both solid and liquid tumors including synovial sarcoma and multiple myeloma. However, the depth and durability of response may be affected by the inhibitory tumor microenvironment. Tumors utilize many different methods to inhibit anti-tumor immunity including secretion of inhibitory cytokines, such as transforming growth factor-β (TGF-β) and induction/recruitment of other inhibitory cells including regulatory T cells and myeloid-derived suppressor cells. These inhibitory cells also secrete cytokines such as IL-10 and TGF-β that potentially reduce the efficacy of T cells. TGF-β is expressed at high levels in a range of cancer indications.

Methods

We investigated whether SPEAR® T cells can be engineered to express additional proteins, allowing them to overcome such immune resistance mechanisms, potentially improving clinical responses. TGF-β inhibits T cells by binding to a dimer of TGFβRII, which then recruits a dimer of TGFβRI forming a heterotetrameric complex that activates inhibitory intracellular SMAD signaling pathways. Truncating the intracellular signaling domain produces a dominant negative TGF-β receptor (dnTGFβRII) that, although capable of binding TGF-β, is unable to signal. We therefore generated SPEAR® T cells co-expressing enhanced affinity T cell receptors (TCR) that recognize a peptide from NY-ESO/LAGE-1A in the context of HLA-A2, together with dnTGFβRII and tested their function in vitro.

Results

Firstly we showed that the function of NY-ESO SPEAR® T cells is inhibited with physiologically-relevant concentrations of TGF-β. We further show that dnTGFbRII can be co-expressed with enhanced affinity NY-ESO TCR in SPEAR® T cells. T cells expressing dnTGFβRII had reduced SMAD phosphorylation in response to TGF-β compared with cells expressing TCR alone, indicating that inhibitory signaling in response to TGF-β was reduced. Subsequently we showed that T cells expressing dnTGFβRII were partially or completely resistant to the effects of TGF-β, using assays for T cell proliferation, cytotoxicity (in 2D and with 3D microtissue models) and Th1 cytokine release (IFN-γ and IL-2) in response to antigen positive tumor cells.

Conclusions

Together these data indicate that co-expression of dnTGFβRII may be a viable approach to improve the efficacy of SPEAR® T cells in treating cancer.

P6 Inducible MyD88/CD40 (iMC) costimulation drives ligand-dependent tumor eradication by CD123-specific chimeric antigen receptor T cells

MyLinh Duong1, An Lu1, Peter Chang1, Aruna Mahendravada1, Nicholas Shinners1, Kevin Slawin1, David M Spencer2, Aaron E Foster1, J Henri Bayle1

1Bellicum Pharmaceuticals, Houston, TX, USA; 2Bellicum Pharmaceuticals and Baylor College of Medicine, Houston, TX, USA
Correspondence: J. Henri Bayle (jhbayle@bellicum.com)

Background

CD123/IL-3Rα is a promising chimeric antigen receptor (CAR)-T cell target due to its high expression on both acute myeloid leukemia (AML) blasts and leukemic stem cells (AML-LSCs). However, the antigen is also expressed at lower levels on normal stem cell progenitors, presenting a major toxicity concern should CD123-specific CAR-T cells show long-term persistence. Here, we describe a CAR platform, “GoCAR-T”, that uses a proliferation-deficient, first generation, CD123-specific CAR together with a ligand-dependent costimulatory switch (inducible MyD88/CD40 (iMC)) to provide physician-controlled eradication of CD123 + tumor cells and regulate long-term CAR-T cell engraftment.

Methods

T cells were activated and transduced with a bicistronic retrovirus encoding iMC (MyD88 and CD40 cytoplasmic signaling domains fused with tandem copies of FKBPv36 (binding domain for the dimerizing ligand rimiducid (Rim)) and a first generation CAR targeting CD123 (SFG--iMC-CD123.ζ). Ligand dependence for costimulation with iMC was assessed in coculture assays with CD123+ AML cell lines (KG1, THP-1 and MOLM-13) by examination of cytokine production and observation by IncuCyte-based live cell imaging. In vivo efficacy was assessed by i.v. injection of 106 EGFPluc-expressing CD123+ THP-1 tumor cells into immunodeficient NSG mice. After seven days 2.5x106 non-transduced or iMC-CD123.ζ-modified T cells were injected and Rim (1 mg/kg) or vehicle only administered i.p. on days 0 and 15 post-T cell injection. Animals were evaluated for tumor burden using IVIS bioluminescent imaging (BLI) and for T cell persistence by flow cytometry and qPCR at day 30 post-T cell injection.

Results

In coculture assays, both CD123 antigen recognition and Rim-dependent iMC costimulation were required for IL-2 production (285±41 versus 2,541±255 pg/ml for control and 1 nM Rim, respectively), robust CAR-T cell proliferation (87-fold increase with Rim stimulation) and enhanced KG1 cell killing. In NSG mice engrafted with CD123+ THP-1-EGFPluc tumor cells, only animals treated with iMC-CD123.ζ-modified T cells and Rim controlled tumor growth, showing a 2-log reduction in tumor burden with Rim treatment. Two weeks after the second Rim injection, CAR-T cells were infrequent (<1.0%) in the spleen and bone marrow of both CAR groups, suggesting that active costimulation is required for CAR-T persistence.

Conclusions

GoCAR-T, a platform comprising a ligand-dependent activation switch and a proliferation-deficient first generation CAR, efficiently eradicates CD123+ leukemic cells when costimulation is provided by systemic rimiducid administration. Deprivation of iMC costimulation results in reduction of CAR-T levels, providing a user-controlled system for managing persistence and safety of CD123-specific CAR-T cells.
Fig. 25

(Abstract P6). Rimiducid-dependent MyD88/CD40 costimulation enhances antitumor activity of a first-generation CD123-specific CAR

P7 Heterodimeric IL-15 treatment enhances tumor infiltration, persistence and effector functions of adoptively transferred tumor-specific T cells in the absence of lymphodepletion

Cristina Bergamaschi1, Sinnie Sin Man Ng1, Bethany Nagy1, Shawn Jensen2, Xintao Hu1, Candido Alicea1, Bernard Fox2, Barbara Felber1, George Pavlakis1

1National Cancer Institute at Frederick, Frederick, MD, USA; 2Providence Cancer Center, Portland, OR, USA
Correspondence: Cristina Bergamaschi (cristina.bergamaschi@nih.gov)

Background

Adoptive cell transfer (ACT) is a promising immunotherapeutic approach for cancer. Host lymphodepletion is associated with favorable ACT therapy outcomes, but it may cause detrimental effects in humans. Among the benefits provided by lymphodepletion, ablation of cells forming a cytokine “sink” results in high levels of homeostatic cytokines that support proliferation and survival of the transferred lymphocytes. Interleukin-15 (IL-15) is a lymphocyte growth and activation factor presently in clinical trials for immunotherapy of metastatic cancers. We previously showed that bioactive IL-15 in vivo comprises a stable complex of the IL-15 chain with the IL-15 receptor alpha chain (IL-15Rα), termed heterodimeric IL-15 (hetIL-15). In this study, we tested the hypothesis that hetIL-15 administration enhances ACT in the absence of lymphodepletion.

Methods

We evaluated the effects of the combination regimen ACT+hetIL-15 in the absence of lymphodepletion by transferring melanoma-specific Pmel-1 T cells into B16 melanoma-bearing mice. Tumors were analyzed by both flow cytometry and multi-parameter immunohistochemistry. Tumor-infiltrating transferred Pmel-1 were analyzed for their persistence, proliferation and effector functions.

Results

hetIL-15 treatment delayed tumor growth by promoting infiltration and persistence of both adoptively transferred Pmel-1 cells and endogenous CD8+ T cells into the tumor. In contrast, persistence of Pmel-1 cells was severely reduced following irradiation in comparison to hetIL-15 treatment. Importantly, we found that hetIL-15 led to the preferential enrichment of Pmel-1 cells in B16 tumor sites in an antigen-dependent manner. Upon hetIL-15 administration, tumor-infiltrating Pmel-1 cells showed a “non-exhausted” effector phenotype, characterized by increased IFN-g secretion, proliferation and cytotoxic potential and low level of PD-1. hetIL-15 treatment also resulted in an improved Pmel-1 to Treg ratio in the tumor.

Conclusions

This study shows that hetIL-15 administration improves the outcome of ACT in immunocompetent hosts and is able to replace the need for lymphodepletion prior ACT for cancer therapy. Applications of heterodimeric IL-15 to ACT will provide new tools and techniques for cancer immunotherapy protocols. Elimination of the need for lymphodepletion will make more patients eligible for cell transfer protocols. In addition, IL-15 could be a general method to place T cells into tumors, increasing the success rate of other immunotherapy interventions.

P8 Withdrawn

P9 Partially differentiated polyfunctional T cells dominate the periphery after tumor-infiltrating lymphocytes therapy for cancer

Marco Donia1, Julie Westerlin Kjeldsen2, Rikke Andersen1, Marie Christine Wulff Westergaard1, Valentina Bianchi3, Mateusz Legut3, Meriem Attaf3, Garry Dolton3, Barbara Szomolay4, Sascha Ott5, Rikke Lyngaa6, Sine Reker Hadrup6, Andrew Kelvin Sewell3, Inge Marie Svane1

1Department of Oncology, Center for Cancer Immune Therapy, Herlev Hospital, Herlev, Hovedstaden, Denmark; 2Center for Cancer Immune Therapy, Herlev Hospital, Herlev, Hovedstaden, Denmark; 3Cardiff University School of Medicine, Cardiff, Wales, UK; 4Systems Immunology Institute, Cardiff University, Cardiff, Wales, UK; 5Warwick Systems Biology Centre, University of Warwick, Coventry, England, UK; 6Section for Immunology and Vaccinology, National Veterinary Institute, Technical University of Denmark, Frederiksberg, Hovedstaden, Denmark
Correspondence: Marco Donia (doniamarco@gmail.com)

Background

Infusion of highly heterogeneous populations of autologous tumor-infiltrating lymphocytes (TILs) can result in tumor regression of exceptional duration. Initial tumor regression has been associated with persistence of tumor-specific TILs one month after infusion, but mechanisms leading to long-lived memory responses are currently unknown. Here, we studied the dynamics of bulk tumor-reactive CD8+ T cell populations in patients with metastatic melanoma following treatment with TILs.

Methods

We analyzed the function and phenotype of tumor-reactive CD8+ T cells contained in serial blood samples of sixteen patients treated with TILs.

Results

Polyfunctional tumor-reactive CD8+ T cells accumulated over time in the peripheral lymphocyte pool. Combinatorial analysis of multiple surface markers (CD57, CD27, CD45RO, PD-1 and LAG-3) showed a unique differentiation pattern of polyfunctional tumor-reactive CD8+ T cells. This subpopulation acquired simultaneously expression of the early differentiation marker CD27, alongside typical features of late effector cells such as loss of CD45RO and up-regulation of PD-1 and CD57. The differentiation and functional status appeared very similar from 1 month to 1 year after infusion. Despite some degree of clonal diversification occuring in vivo within the bulk tumor-reactive CD8+ T cells, further analyses showed that CD8+ T cells specific for defined tumor-antigens had similar differentiation status.

Conclusions

We demonstrated that tumor-reactive CD8+ T cell subsets that persist after TIL therapy are mostly polyfunctional,and display a stable partially differentiated phenotype. These atypical CD27+ incompletely differentiated polyfunctional TILs may have a high capacity for persistence and be crucial in keeping patients tumor free.

Trial Registration

ClinicalTrials.gov identifier NCT00937625.

P10 Peptide vaccine enhances antitumor effect of adoptive cell transfer using genetically engineered T cells

Aaron Fan1, Takumi Kumai2, Esteban Celis2

1Augusta University, Augusta, GA, USA; 2Georgia Cancer Center, Augusta University, Augusta, GA, USA
Correspondence: Aaron Fan (afan@augusta.edu)

Background

Adoptive cell therapy (ACT) has shown promise in tumor eradication in cancer, but current methods require harmful and toxic adjunct procedures. Our laboratory has developed a potent peptide vaccination strategy called TriVax that bypasses the necessity of these adjunct procedures. Previous studies show that retrovirally (RV)-transduced T cells are effective in ACT against various cancers. The present study aimed to determine whether RV-transduced T cells could respond to TriVax specific for melanosomal tumor antigen gp100, and to see if the responses could be enhanced when transduced with a constitutively active form of STAT5 (CA-STAT5), which has been shown to increase survival of CD8 effector/effector memory T cell populations.

Methods

CD8 T cells were purified from B6 mouse splenocytes and RV-transduced with a gene encoding the clonal T cell receptor (TCRs) for mouse gp100. In some experiments, cells were also co-transduced with a gene encoding for CA-STAT5 (co-expressing Thy1.1). Transduction efficiency was assessed using flow cytometry with gp100 tetramer and Thy1.1 antibody. Functional activity was measured using flow cytometry and EliSpot (IFNγ) assays. These cells were then adoptively transferred (1.0x105 tetramer+ cells) into naïve and tumor bearing congenic CD45.1 B6 mice, which were then vaccinated with TriVax. Tumor growth was assessed 3 times per week, and immune status was assessed in blood using flow cytometry every 7 days.

Results

TriVax administration selectively expanded the ACT cell population expressing gp100-TCR. Cell numbers in spleen indicate a 14-fold expansion 25 days after vaccination from what was initially transferred. When co-transduced with CA-STAT5, an even higher fold expansion (55-fold) was observed. CA-STAT5+ cells also expanded more robustly than CA-STAT5- cells when stimulated with a subsequent vaccine boost, demonstrating a 5000-fold increase in cell numbers with 85% of CD8+ T cells also being positive for gp100 tetramer. ACT of these cell populations into tumor bearing mice also yielded a dramatic increase in cell numbers, which greatly enhanced the survival of mice in treatment groups.

Conclusions

CD8 T cells RV-transduced to express a gp100 TCR and CA-STAT5 are capable of expansion in response to TriVax, bypassing the necessity of adjunct procedures. Co-expression of CA-STAT5 greatly enhances the boost effect of TriVax, leading to a dramatic anti-tumor effect.

P11 Stable tumor-infiltrating lymphocytes (TIL) phenotype following cryopreservation

Ian Frank, Amanda Stramer, Michelle A Blaskovich, Seth Wardell, Maria Fardis, James Bender, Michael T Lotze

Lion Biotechnologies, Inc., Tampa, FL, USA
Correspondence: Ian Frank (Ian.frank@lionbio.com)

Background

Lion Biotechnologies focuses on the development and commercialization of cancer immunotherapies based on tumor-infiltrating lymphocytes (TIL). Cryopreservation is a beneficial process which allows the final cell product to be shipped in a safe manner with less time constraints [1]. Clinical studies using cryopreserved TIL have not been conducted so far. Freezing and thawing of the cells may cause phenotypic changes such as loss of cell surface receptors [2]. Here, we tested fresh versus frozen/thawed TIL samples and evaluated the expression of phenotypic markers.

Methods

Briefly, PreREP TILs were obtained by culturing melanoma tumor fragments in IL-2 (6000 IU/ml). Rapid expansion protocol (REP) cells were initiated using allogeneic PBMC feeder cells with OKT3 and IL-2 in a Grex-100 flask. When the desired confluency was reached, the cells were cryopreserved in a 5% DMSO solution. We used flow cytometry to phenotype the fresh and thaw/rested TIL at 0h, 24h and 7d following reREP TIL. Flow cytometric analysis was performed using fluorescent antibodies specific for TCRa/b, CD4, CD8, CD27, CD28, CD56, CCR7, CD45RA, CD95, PD-1, and CD25.

Results

No significant differences in CD4, CD8 and TCRa/b expression or memory markers comparing fresh TIL versus thaw/rested TIL at 24h was noted. CD27 expression on TIL was reduced by 50% on thawed TIL, however after a 24h resting period the expression recovered. All other surface antigens that we tested were within 10% difference in their expression levels as compared to baseline.

Conclusions

Cryopreservation did not affect the measured phenotypic characteristics of TIL. We are further investigating the possibility of using cryopreserved TIL in a clinical setting.

References

1. Axelsson S, Faresjo M, Hedman M, Ludvigsson J, Casas R: Cryopreserved peripheral blood mononuclear cells are suitable for the assessment of immunological markers in type 1 diabetic children.Cryobiology 2008, 57:201–208.

2. Sadeghi A, Ullenhag G, Wagenius G, Tötterman TH, Eriksson F: Rapid expansion of T cells: Effects of culture and cryopreservation and importance of short-term cell recovery. Acta Oncol 2013, 52:978-986.

P12 Recognition of autologous neoantigens by tumor infiltrating lymphocytes derived from human breast cancer metastases

Stephanie L. Goff1, Nikolaos Zacharakis1, Yasmine Assadipour1, Todd D Prickett1, Jared J Gartner1, Robert Somerville2, Mary Black1, Hui Xu1, Harshini Chinnasamy1, Isaac Kriley1, Lily Lu1, John Wunderlich1, Paul F Robbins1, Steven Rosenberg1, Steven A Feldman1

1Surgery Branch, National Cancer Institute, National Institutes of Health, Bethesda, MD, USA; 2National Cancer Institute, Bethesda, MD, USA
Correspondence: Stephanie L. Goff (stephanie.goff@nih.gov)

Background

Adoptive transfer of tumor infiltrating lymphocytes (TIL) can effect long-term durable regression in patients with metastatic melanoma but has not been widely tested in common epithelial cancers. When examining the TIL of successfully treated patients with melanoma, a heterogeneous T cell population can be identified with reactivity against melanoma differentiation antigens, cancer germline antigens, and personalized non-synonymous somatic mutations. Common epithelial cancers, including breast cancer, express far fewer somatic mutations than melanoma, however, in a study of metastatic gastrointestinal cancer, lymphocytes targeting neoantigens were identified in the majority of specimens. This pilot study investigates the ability to grow TIL from breast cancer metastases, to identify personalized non-synonymous somatic mutations and potential neoantigens, and to adoptively transfer TIL into patients with breast cancer.

Methods

Eligible patients were evaluated and treated under IRB-approved protocols for tissue procurement, genomic testing, and adoptive cell transfer. Portions of resected tumors were placed in culture under standard TIL conditions. DNA was extracted from tumor and matched normal peripheral blood samples for whole exome sequencing (WES). Non-synonymous somatic mutations were identified and tested for potential immunogenicity using previously described tandem mini-gene and long (25mer) peptide techniques. Recognition was assessed by IFNγ release on ELISPOT and/or CD137 (4-1BB) upregulation with appropriate controls.

Results

Nine patients underwent surgical resection in this ongoing pilot study, and TIL were successfully grown from the tumors of all patients. All were primarily CD3+ (median 79%) with a small population of natural killer cells. Of the CD3+ cells, 7 of 9 patients had a predominantly CD4+ population (median CD4:CD8 ratio 2.2, range 0.4-5.8). For eight tumors, WES was performed, and non-synonymous somatic mutations were identified as potential neoantigens (median count 96.5, range 71-148). Autologous T cell reactivity has been identified against tumor-specific mutations in 6 of 8 patients.

Conclusions

Tumor-infiltrating lymphocytes derived from metastatic breast cancer can react to tumor-specific non-synonymous somatic mutations in vitro. TIL grown from breast cancers are predominantly CD4+ and can form the basis of an adoptive cell transfer experimental approach to patients with metastatic breast cancer.

Trial Registration

ClinicalTrials.gov identifier NCT01174121.

P13 Comparison of RECIST 1.0 to RECIST 1.1 in the evaluation of adoptive cell transfer

Kasia Trebska-McGowan1, Isaac Kriley1, Parisa Malekzadeh1, Eden Payabyab1, Richard Sherry2, Steven Rosenberg1, Stephanie L. Goff1

1Surgery Branch, National Cancer Institute, National Institutes of Health, Bethesda, MD, USA; 2National Cancer Institute, Bethesda, MD, USA
Correspondence: Stephanie L. Goff (stephanie.goff@nih.gov)

Background

Adoptive transfer of tumor infiltrating lymphocytes (TIL) can effect long-term durable regression in patients with metastatic melanoma. The earliest studies utilized WHO criteria for evaluation of response, but more recent studies replaced the bidimensional analysis with simplified unidimensional criteria of RECIST 1.0. With improved cross-sectional imaging, there became concern that the pathophysiology of lymph node response required distinct evaluation criteria, as tumor clearance may not completely eradicate normal structures. RECIST 1.1 was developed and validated in a variety of histologies with various treatments, but has not been evaluated for adoptive cell transfer.

Methods

Eligible patients were enrolled on an IRB-approved protocol of adoptive cell transfer, randomizing patients to receive one of two lymphodepleting regimens prior to transfer of TIL. This study was reported using RECIST 1.0 criteria, with 24 complete responders and 30 partial responders among 99 treated patients [1]. The official tumor measurements of target lesions and notations of non-target lesions were re-evaluated using RECIST 1.1 criteria, which limits the total number of target lesions, the number of target lesions/organ, and uses short-axis measurements for lymph node disease.

Results

By design, the number of target lesions/patient was decreased, from 4 (range 1-10) to 3 (range 0-5). Thirty-eight lymph nodes did not meet short-diameter criteria for target lesions (10-14mm), and an additional 12 measured <10mm were reclassified as “non-pathologic”. With retrospective application of RECIST 1.1, three patients would not have met eligibility criteria for lack of evaluable disease. In assessing overall response to treatment, 25 patients met CR criteria, with an additional 27 with PR. While there were five patients who achieved CR at an earlier time point, overall time to response was not significantly different (median 16.0 v 19.8 months p=0.19). One patient with lymph node disease did not achieve CR by original criteria, was an early CR in this analysis, but recurred three months later.

Conclusions

Adoption of RECIST 1.1 demonstrated comparable response rates for this trial. A hallmark of our modern studies is the durability of complete responses with RECIST 1.0, and only by further application of the new criteria will we be able to confirm the validity of lymph node response criteria in adoptive cell transfer.

Trial Registration

ClinicalTrials.gov identifier NCT01319565.

References

1. Goff SL, Dudley ME, Citrin DE, Somerville RP, Wunderlich JR, Danforth DN, et al: Randomized, Prospective Evaluation Comparing Intensity of Lymphodepletion Before Adoptive Transfer of Tumor-Infiltrating Lymphocytes for Patients with Metastatic Melanoma.J Clin Oncol 2016, 34(20):2389-2397.

P14 Bioluminescent redirected lysis assay (BRLA) as an efficient potency assay to assess tumor-infiltrating lymphocytes (TILs) for immunotherapy

Aishwarya Gokuldass, Michelle A Blaskovich, Charlene Kopits, Brian Rabinovich, Michael T. Lotze

Lion Biotechnologies, Inc., Tampa, FL, USA
Correspondence: Aishwarya Gokuldass (aishwarya.gokuldass@lionbio.com)

Background

Administration of TILs is a promising treatment for patients with melanoma and other solid tumors. TIL therapy involves culturing and expanding T cells isolated from a patient’s tumor in vitro and then reinfusing them back into the patient. TIL antitumor activity is commonly measured using tumor cells from the patient’s tumor, when available. However, autologous tumor cell lines are difficult to grow consistently. In order to test the potency of TILs that are infused into patients, we developed a surrogate target cell line that can be used to assess the lytic potential of TILs in a BRLA.

Methods

Mouse mastocytoma P815 cells expressing endogenous CD16, transduced with eGFP and firefly luciferase were selected as a surrogate for autologous tumor cells. The CD16 Fc receptor binds anti-CD3 antibodies providing a potent TCR activation signal. The P815 cells were sorted and cloned and then co-cultured with TILs +/- anti-CD3 antibodies to assess tumor reactivity through TCR activation and basal non-specific killing. Following 4 hours of incubation, luciferin was added to the wells for 5 minutes. After the incubation, bioluminescence intensity was determined. The % cytotoxicity and survival were calculated thus: % Survival = (experimental luminescence-background)/ (maximum luminescence-background)*100; % Cytotoxicity = 100-(% Survival); and Interferon-γ release in the media of the co-cultured TILs and P815 cells was analyzed by ELISA and LAMP1 (CD107a) expression analyzed by flow cytometry to evaluate the cytotoxicity of TILs.

Results

TCR-mediated TIL cytotoxicity can be measured in a highly sensitive dose-dependent manner. As shown below, this assay is highly sensitive. The following data will be presented: Highly Sensitive Potency Assay for TILs. A) % cytotoxicity of TIL (from patient M1033T-1) when co-cultured with P815 Clone G6 (with and without anti-CD3 antibody) at different ratios of effector to target cells. B) ELISA data showing amount of IFN-γ released at different ratios of effector to target cells. C) % LAMP1 marker expressed by the M1033T-1 when co-cultured with P815 Clone G6 and anti-CD3 at 1:1 effector to target cells for 4 and 24hr co-culture.

Conclusions

Our ‘Bioluminescent Redirected Lysis Assay’ (BRLA) using an engineered P815 cell line can be used as an assay to measure TIL potency. It requires no radionuclides and is more efficient than traditional cytotoxicity assays.
Fig. 26

(Abstract P14). a % cytotoxicity of TIL (from patient M1033T-1) when co-cultured with P815 Clone G6 (with and without anti-CD3 antibody) at different ratios of effector to target cells. b ELISA data showing amount of IFN-γ released at different ratios of effector to target cells. c % LAMP1 marker expressed by the M1033T-1 when co-cultured with P815 Clone G6 and anti-CD3 at 1:1 effector to target cells for 4 and 24hr co-culture

P15 Culturing and live cell confocal imaging of ovarian cancer spheroids with monocytes and interferons

Daniel S. Green1, Olena Kamenyeva2, Kathryn C Zoon3, Christina M Annunziata1

1Translational Genomics Section, Women's Malignancy Branch, National Cancer Institute, Bethesda, MD, USA; 2Biological Imaging Section, Research Technologies Branch, National Institute of Allergy and Infectious Disease, National Institutes of Health, Bethesda, MD, USA; 3Laboratory of Infectious Diseases, National Institute of Allergy and Infectious Disease, National Institutes of Health, Bethesda, MD, USA
Correspondence: Daniel S. Green (daniel.green2@nih.gov)

Background

Ovarian cancer is the number one cause of death due to gynecological malignancies and the fifth leading cause of death in cancer in women. 75% of patients will have recurrent disease. Recurrent disease is chemotherapy refractive that progresses to chemotherapy resistant disease. Currently there are no definitive second-line treatments for patients. Patients with ovarian cancer have a 5-year survival rate of 25-30%, making it one of the most lethal malignancies.

Methods

One of the limitations of in vitro studies is the use of immortalized cell lines grown in two-dimensional flasks. To address this issue, many laboratories have been growing spheroids. We have employed 3-dimensional culturing techniques to create complex spheroids from ovarian cancer cell lines that have phenotypic structure similar to metastatic ovarian cancer. To further characterize spheroids we have created a system for sorting spheroids based on size. The spheroids can then be cultured in 2-dimensions for the study of chemotherapy sensitivities, and the efficacy of monocyte and interferon therapies for the treatment of ovarian cancer. Further, we have developed a single-photon confocal microscopy protocol for the multi-parameter imaging of live spheroids and monocytes with and without interferons. To create experimental robustness, we have employed a technique to image multiple conditions (4) over long periods of time (14-16 hours), allowing for the simultaneous imaging of both control and experimental conditions. Post-acquisition analysis of the images can be used to study migration of the individual cells within the spheroids, loss of cell dye viability, and migration of monocytes into the spheroids.

Results

We have found that the size of the spheroid defines, in part its sensitivity to standard chemotherapy agents. Post-acquisition analysis of the images have been used to study migration of the individual cells within the spheroids, loss of cell dye viability, and migration of monocytes into the spheroids. The addition of interferons with or without monocytes significantly reduces the movement of the individual cells within the spheroids. Furthermore, we have found that the addition of interferons slows monocyte migration and initiates monocyte attachment and entry into the spheroids.

Conclusions

The combination of novel cell culturing techniques with modern imaging and post-acquisition data analysis will increase our understanding of ovarian cancer response to both standard chemotherapy and novel cell based therapies.

Acknowledgements

This work was supported by the Intramural Research Program of the National Institutes of Health (NIH), National Institute of Allergy and Infectious Diseases (NIAID), and National Cancer Institute (NCI).

P16 A novel xenograft model of chimeric antigen receptor-mediated toxicity sheds light on the influence of T cell source on the severity of the toxic sequellae

Joanne Hammill, Christopher Helsen, Craig Aarts, Jonathan Bramson

McMaster University, Hamilton, ON, Canada
Correspondence: Joanne Hammill (hammilja@mcmaster.ca)

Background

Chimeric antigen receptor (CAR)-engineered T lymphocytes are demonstrating striking clinical success, yet these treatments can be accompanied by severe on- and off-tumor toxicities. The availability of murine models to study these toxic phenomenon are currently lacking.

Methods

In our model, human T lymphocytes are engineered with a second-generation CD28-based CAR, targeted with a designed ankyrin repeat protein (DARPin) with picomolar affinity against HER2 (anti-HER2 DARPin CAR-T cells), and adoptively transferred into NRG mice at doses ranging from 6 x 105 – 1.2 x 107 CAR-T cells/mouse.

Results

Toxicity manifested in a drop in core body temperature and weight and, in some cases, lethality. The onset and severity of the toxicity varied with the source of the T lymphocytes (i.e., donor) used to generate CAR-T cells. We evaluated seven different donors and the toxicities associated with each donor’s cell product was reproducible across multiple experiments and multiple manufacturing runs. Anti-HER2 DARPin CAR-T cells were toxic in tumor-free mice, but toxicities were most severe in the presence of HER-2+ tumors demonstrating both on-tumor and off-tumor effects. The CAR-T cells did not respond to murine HER-2 indicating that the toxicity was both off-tumor and off-target. The toxicity was not due to the DARPin itself, as CAR-T cells bearing DARPins with other specificities were not toxic in this model. Further characterization of the model indicated that severity of toxicity was dose-dependent and could be exacerbated by the presence of a HER2+ tumor. Moribund mice were found to have aggregates of T cells in their lungs, liver, and heart and displayed a cytokine storm in the blood. The toxicity was triggered by CD4+ T cells in the anti-HER2 DARPin CAR-T cell product. While anti-HER2 DARPin CAR-T cells generated from donors demonstrating reduced toxicity were able to mediate anti-tumor efficacy in vivo in a xenograft model of ovarian carcinoma at low doses, they exhibited a narrow therapeutic window consistent with data emerging from the clinic where CAR-T therapy is effective.

Conclusions

This model offers a promising avenue to test strategies for the prevention or mitigation of toxicities associated with adoptive CAR-T cell transfer as well as insights into the contribution of T cell source to toxicities. Investigations are ongoing.

Acknowledgements

This work was supported by the Samuel Family Foundation, the Terry Fox Foundation, the Canadian Breast Cancer Foundation, and Triumvira Immunologics.

P17 Ex vivo generation of highly purified and activated natural killer cells from human peripheral blood in accordance with GMP/GCTP for clinical studies

Yui Harada, Yoshikazu Yonemitsu

Kyushu University, Fukuoka, Fukuoka, Japan
Correspondence: Yui Harada (rkfraile@med.kyushu-u.ac.jp)

Background

Natural killer (NK) cells play a crucial role during the innate immune responses, and as such form part of the body of immunological defense against various diseases, including infectious diseases and malignancies. Therefore, adoptive immunotherapy using NK cell is emerging as promising treatments for intractable malignancies; however, there has been still developing because of difficulties in culture, shortage of overall effector numbers, contamination of considerable numbers of T cells, and their limited anticancer potencies. We here established the simple feeder-free method to generate purified (>90%) and highly activated NK cells from human peripheral blood-derived mononuclear cells (PBMCs) in accordance with GMP/GCTP for clinical studies.

Methods

Under approval of the institutional ethical committee, PBMCs were collected from healthy volunteers by using CliniMACS Prodigy® (automatic/closed system). CD3+ cells were depleted by CliniMACS CD3 beads, and CD3-depleted PBMCs were cultured at a concentration of 5 x 105 cells/ml with high concentration of hIL-2 and 5% human AB serum for 14 days. Fresh medium wad added every 4-5 days throughout the culture period. Then, we confirmed the expression of surface markers, CD107a mobilization and cell-mediated cytotoxicity against various tumor cells and normal cells with or without monoclonal antibody drugs in vitro and antitumor effects against K562 in vivo.

Results

Among the several parameters, we found that simply 1) only CD3-depletion, 2) high dose IL-2, and 3) use of specific culture medium were sufficient to obtain the highly purified, expanded (~200-fold) and activated CD3-/CD56+ NK cells from PBMCs. Almost all activated NK cells expressed lymphocyte-activated marker CD69, and showed dramatically high expression of NK activation receptors (i.e. NKG2D, NKp30, NKp46, etc.), interferon-g, perforin and granzyme B. Importantly, only 2 hours reaction at effector/target ratio=1:1 was sufficient to kill almost all K562 cells, and antitumor activity was also representative on tumor bearing mice in vivo. Cytolysis was specific for various tumor cells, but not for normal cells, irrespective of MHC class I expression.

Conclusions

These findings strongly support that NK cells activated by our method is purified, expanded, and near fully activated. The cells were currently under investigation in clinical trial (phase I/IIa).

P18 T cell antigen couplers (TAC) demonstrate strong effectivity against solid tumors with no measurable toxicities, demonstrating an enhanced therapeutic index

Christopher Helsen1, Joanne Hammill1, Kenneth Mwawasi1, Galina Denisova1, Jonathan Bramson1, Rajanish Giri2

1McMaster University, Hamilton, ON, Canada; 2Indian Institute of Technology Mandi, Mandi, India
Correspondence: Christopher Helsen (helsenc@mcmaster.ca)

Background

Engineering T cells with chimeric antigen receptors (CARs) is proving to be an effective method for directing T cells to attack tumors in an MHC-independent manner [1, 2]. Current generation CARs aim to recapitulate T cell signaling by incorporating modular functional components of the TCR and co-stimulatory molecules. We sought to develop an alternate method to re-direct the T cell receptor which employs the native TCR. To this end, we developed the T cell Antigen Coupler (TAC) technology, a membrane-anchored receptor redirects the TCR and co-receptor in the presence of tumor antigen.

Methods

The utility of the TAC receptor has been assessed using in vitro and in vivo assays. In vitro assays are based on receptor surface staining, cytokine release and cytotoxicity assays. In vivo studies examined the anti-tumor effect of TAC-engineered T cells against established xenografts.

Results

In vitro testing has demonstrated robust and specific cytokine production and cytotoxicity by TAC-engineered human T cells directed against either HER-2 or CD19. In vivo TAC-engineered T cells revealed strong activity against HER-2 expressing solid xenograft tumor models such as MDA-MB-231 and OVCAR-3. Curiously, the TAC-engineered T cells outperformed a matched CD28-based second generation CAR in these models, demonstrating both increased anti-tumor efficacy and reduced toxicity. Whereas, mice treated with CAR engineered T cells showed serious toxicities that were both donor- and dose-dependent, we did not observe any toxicities arising from the TAC-engineered T cell, even at doses that produced complete tumor regression.

Conclusions

These differences in toxicities and efficacy highlight the biological differences of TAC and CAR receptors and indicated the potential for a superior therapeutic index for TAC engineered T cells. Current TAC’s in development target lymphatic malignancies and have shown great promise in early in vitro and in vivo assays.

Acknowledgements

Samuel Family Foundation, the Terry Fox Foundation, the Canadian Breast Cancer Foundation and Triumvira Immunologics.

References

1. Barrett DM, Singh N, Porter DL, Grupp S, June CH: Chimeric antigen receptor therapy for cancer.Annu Rev Med 2014, 65:333–347.

2. Gill S, June CH: Going viral: chimeric antigen receptor T-cell therapy for hematological malignancies.Immunol Rev 2015, 263 :68–89.

3. Zhao Z, et al: Structural Design of Engineered Costimulation Determines Tumor Rejection Kinetics and Persistence of CAR T Cells.Cancer Cell 2015, 28:415–428.

4. Sadelain M, Brentjens R, Rivière I: The basic principles of chimeric antigen receptor design.Cancer Discov 2013, 3:388–398.

P19 T cell receptor gene engineered T cells targeting human papillomavirus (HPV)-16 E7 induce regression of HPV-16+ human tumors in a murine model

Benjamin Jin1, Tracy Campbell1, Lindsey M Draper2, Sanja Stevanovic1, Zhiya Yu3, Bianca Weissbrich4, Nicholas P Restifo3, Cornelia L Trimble5, Steven Rosenberg3, Christian S Hinrichs3

1Experimental Transplantation and Immunology Branch, National Cancer Institute, Bethesda, MD, USA; 2National Cancer Institute, Bethesda, MD, USA; 3Surgery Branch, National Cancer Institute, Bethesda, MD, USA; 4Kite Pharma EU, Amsterdam, Noord-Holland, Netherlands; 5Johns Hopkins University, Baltimore, MD, USA
Correspondence: Christian S Hinrichs (hinrichs@nih.gov)

Background

The human papillomavirus (HPV)-16 E7 oncoprotein is constitutively expressed by HPV-16-associated cancers and absent from healthy tissues, and it is therefore an attractive therapeutic target for T cell receptor (TCR) gene engineered T cell therapy. However, E7 displays manifold mechanisms of immune evasion, and its potential as a target for TCR-T cell therapy is unknown.

Methods

The nucleotide sequence of a TCR targeting HPV-16 E711-19 was determined by study of cervix-infiltrating T cells from a subject with HPV-16+ cervical intraepithelial neoplasia. Expression of the TCR was optimized by hydrophobic substitutions in the alpha chain transmembrane region and introduction of an additional inter-chain TCR constant region disulfide bond. Second-party T cells genetically engineered to express the optimized TCR (E7 TCR T cells) were evaluated in in vitro immunologic assays and in an in vivo murine model of human cervical cancer.

Results

E7 TCR T cells displayed binding to HLA-A*02:01-E711-19 tetramers. HLA-A*02:01 restriction was confirmed with major histocompatibility complex (MHC) blocking and with testing for recognition of target cells either expressing or lacking HLA-A*02:01. In functional assays, E7 TCR T cells recognized and killed a panel of HLA-A*02:01+ HPV-16+ cervical and oropharyngeal cancer cells lines but not cell lines that lacked either the restriction element or target antigen. Tumor recognition by CD8 and CD4 T cells that expressed the E7 TCR was observed. They displayed minimal if any cross-reactivity against peptides from human proteins that were identified by BLAST search based on protein sequence similarity or the presence of key shared residues mapped by alanine scanning of E711-19. E7 TCR T cells recognized E711-19 peptide-pulsed T2 cells at concentrations as low as 10-11 M indicating relatively high functional avidity. Direct comparison to a previously described E6-specific TCR revealed greater functional avidity, a slower TCR-pMHC Koff-rate, and superior tumor cell recognition for the E7 TCR. In a murine xenograft model, a single intravenous injection of E7 TCR T cells induced complete regression of tumors from one human cervical cancer line and controlled tumors from another human cervical cancer line.

Conclusions

E7 TCR T cells specifically recognized and killed HPV-16+ cancer cells in vitro and mediated regression of HPV-16+ tumors in vivo. These findings provide the preclinical basis for a new personalized TCR-T cell therapy for metastatic HPV-16+ cancers including many cervical, oropharyngeal, and anal malignancies. A clinical trial for E7 TCR T cells is now active (NCI-16-C-0138).

P20 haNK, a cytotoxic human high affinity natural killer cell line, exerts enhanced ADCC mediated by avelumab (an anti-PD-L1 antibody) against multiple human tumor cell lines

Kwong Tsang1, Massimo Fantini1, James W Hodge2, Rika Fujii2, Ingrid Fernando1, Caroline Jochems1, Christopher Heery1, James Gulley1, Patrick Soon-Shiong3, Jeffrey Schlom4

1Laboratory of Tumor Immunology and Biology, National Cancer Institute, National Institutes of Health, Bethesda, MD, USA; 2National Cancer Institute, National Institutes of Health, Bethesda, MD, USA; 3NantKwest, Inc., Culver City, CA, USA; 4Center for Cancer Research, National Cancer Institute, National Institutes of Health, Bethesda, MD, USA
Correspondence: James W Hodge (jh241d@nih.gov)

Background

Immune checkpoints have been implicated in the down-regulation of antitumor immunity. Anti-PD-1/PD-L1 checkpoint inhibitory monoclonal antibodies (mAbs) can restore immune function in the tumor microenvironment, and have demonstrated clinical benefit in patients with melanoma, Hodgkin’s lymphoma, lung and bladder carcinomas, and other tumor types. In addition to its checkpoint inhibitory function, avelumab, a fully human IgG1 anti-PD-L1 mAb, can mediate antibody-dependent cellular cytotoxicity (ADCC) to lyse human tumor cells in the presence of natural killer (NK) cells [1]. NK cells can be used for cancer therapy. However, obtaining sufficient numbers of functionally active NK cells from patients is technically challenging since only about 10% of the population expresses on NK cells the high-affinity FcR that provides maximum ADCC. One alternative is to use established NK cell lines that have antitumor activity. High affinity NK cells (haNK), provided by NantKwest, derived from the human NK cell line NK-92, are genetically engineered to express high-affinity human CD16 (FcgRIIIA-V158) and transduced with the human IL-2 gene. In addition, haNK cells have little inhibitory killer-cell immunoglobulin-like receptor (KIR) expression, a unique feature that may be a factor in their highly cytotoxic activity against a broad range of malignancies. We report here our investigation of 1) whether haNK cells used in combination with avelumab can lyse human tumor cells via the ADCC mechanism, and 2) the factors that may influence this cytotoxic activity.

Methods

Cell lines used in our experiments included human carcinomas of the head and neck, cervix, bladder, and colon, as well as prostate and pancreatic cancers. haNK cells irradiated with 10 Gy were used as effector cells at various effector-to-target-cell ratios in all experiments. Four- and 18-hour 111In release assays were performed to evaluate ADCC activity.

Results

Our results show that 1) haNK cells can lyse a range of human carcinoma cells when avelumab is used to target PD-L1 expression; 2) the addition of anti-CD16 neutralizing antibody significantly inhibits ADCC lysis, implicating CD16 ligation as a major mechanism of action for ADCC lysis mediated by haNK and avelumab; and 3) in combination with avelumab, haNK cells mediate significantly higher levels of ADCC lysis than do NK cells isolated from healthy donor peripheral blood mononuclear cells (PBMC).

Conclusions

These results provide a rationale for using haNK cells in combination with avelumab or other ADCC-mediating cytotoxic mAbs to treat human malignancies.

P21 Adoptive transfer of ex vivo-expanded PD-1+ CD8+ and CD4+ T cells eliminates myeloma in mice

Weiqing Jing, Jill Gershan, Grace Blitzer, James Weber, Laura McOlash, Bryon D Johnson

Medical College of Wisconsin, Milwaukee, WI, USA
Correspondence: Bryon D Johnson (bjohnson@mcw.edu)

Background

Adoptive T cell therapy (ACT) has emerged as a potential curative therapy for patients with advanced solid tumors. However, for hematologic cancers, identifying and enriching the cancer antigen-reactive T lymphocytes for ACT remains a challenge. Our lab previously demonstrated that blockade of the PD-1/PD-L1 pathway (in the context of non-myeloablative, lymphodepleting whole body irradiation) was capable of eliminating established myeloma in mice [1, 2]. In the current study, we tested whether PD-1 is a marker for myeloma-reactive T cells.

Methods

C57BL/KaLwRij (KaLwRij) mice were inoculated with 5T33 myeloma cells intravenously, and 28 days later, splenic PD-1+ and PD-1- T cells (CD8+ and CD4+) were isolated by flow cytometric sorting. The purified T cells were expanded in culture for 7 days on plate-bound CD3 and soluble CD28 antibodies plus IL-2, IL-7 and IL-15. Some of the expanded T cells were assayed in vitro for myeloma reactivity. Equal mixtures of expanded CD4+ and CD8+ T cells, or each subset alone, were infused to myeloma-bearing Rag1-deficient recipients, and the mice were followed for myeloma progression.

Results

We found that numbers of cancer antigen-reactive T lymphocytes in myeloma-bearing mice were enriched in PD-1+ CD8+ and CD4+ T cell subsets. PD-1+ T cells could be efficiently expanded ex vivo for adoptive transfer, and the expanded cells maintained their anti-myeloma reactivity. Adoptive transfer of the ex vivo-expanded PD-1+ T cells effectively eliminated established myeloma in Rag1-deficient recipients. In contrast, adoptive transfer of expanded PD-1- T cells failed to demonstrate anti-myeloma efficacy in vivo. Notably, both CD8+ and CD4+ PD-1+ T cell subsets played a role in eradicating myeloma, but combined administration of the ex vivo-expanded subsets was most efficacious.

Conclusions

Our results show that PD-1 is a biomarker for myeloma-specific CD8+ and CD4+ T cells in mice. Furthermore, these PD-1-expressing T cells can be expanded in culture for effective adoptive cell immunotherapy of myeloma-bearing recipients.

Acknowledgements

Support from the Midwest Athletes Against Childhood Cancer (MACC) Fund, Children's Research Institute, and a former grant from the Multiple Myeoma Research Foundation.

References

1. Jing W, Gershan JA, Weber J, Tlomak D, McOlash L, Sabatos-Peyton C, Johnson BD: Combined immune checkpoint protein blockade and low dose whole body irradiation as immunotherapy for myeloma. J Immunother Cancer 2015, 3:2.

2. Kearl TJ, Jing W, Gershan JA, Johnson BD: Programmed death receptor-1/programmed death receptor ligand-1 blockade after transient lymphodepletion to treat myeloma. J Immunol 2013, 190:5620-5628.

P22 Entinostat sensitized osteosarcoma cells for cytotoxic effect of natural killer cells

Simin Kiany, Huang Gangxiong, Eugenie S Kleinerman

University of Texas MD Anderson Cancer Center, Houston, TX, USA
Correspondence: Simin Kiany (skiany@mdanderson.org)

Background

The goal of this study is to find an alternate therapy for osteosarcoma (OS) lung metastasis. We previously showed that NK cell therapy significantly decreased, but did not cure, OS lung metastasis in a mouse model. Other studies have shown that histone deacetylase inhibitors (HDACi) sensitize tumor cells to NK cell cytotoxicity, primarily by increasing expression of ligands for NK cells on tumor cells; thus, to augment NK cell therapy, we combined it with the HDACi, entinostat.

Methods

Flow cytometry, western blot analysis, and Q-PCR were used to determine whether entinostat increased expression of NK cell ligands on OS cells. Effects of entinostat on NK cell viability, receptor expression, and cytotoxicity were explored using a viability test, flow cytometry, and calcein release assay, respectively. NK cells were isolated from blood buffy coats and were expanded ex vivo using genetically engineered K562 cells and human IL-2. Q-PCR was used to measure microRNAs expression in OS cells and a CHIP assay was used to determine increased histone acetylation at the MICA/B gene promoters. For the in vivo study, mice were given 10, 5, or 2.5 mg/kg of entinostat orally to determine the subtherapeutic dose of the drug.

Results

We demonstrated that 2 μM entinostat for 48 h upregulated expression of NK cell ligands on OS cell lines. Increased expression of ligands on OS cells resulted in increased susceptibility of OS cell lines to NK cell cytotoxicity in vitro. NK cell treatment with up to 2 μM entinostat did not affect NK cell viability or NK cell receptor (NKG2D, NKp30, NKp44, NKp46, and DNAM-1) expression. NK cells pretreated with entinostat for 24 h did not decrease cytotoxicity of NK cells to OS cell lines. We also showed two mechanisms by which entinostat controls MICA/B expression on OS cells: 1) by increasing H4 acetylation at the MICA/B genes promoters and 2) by down-regulating mir-20a, mir-93, and mir-106b expression. We demonstrated that the sub-therapeutic dose of entinostat that significantly increased MICA/B on OS lung metastasis was 5 mg/kg three times a week for 5 weeks. Combining 5mg/kg entinostat and NK cell therapy is our ongoing in vivo experiment.

Conclusions

We demonstrated that entinostat immunosensitized OS cells to NK cell lysis by inducing upregulation of ligands for NK cells on OS cells. Our results suggest that NK cell therapy combined with entinostat provides an innovative approach to enhance the immunotherapeutic effect of NK cells against OS pulmonary metastases.

P23 Chimeric antigen receptor macrophages (CARMA) for adoptive cellular immunotherapy of solid tumors

Michael Klichinsky, Marco Ruella, Olga Shestova, Saad Kenderian, Miriam Kim, John Scholler, Carl H June, Saar Gill

Center for Cellular Immunotherapies, University of Pennsylvania, Philadelphia, PA, USA
Correspondence: Michael Klichinsky (mklich@mail.med.upenn.edu)

Background

Anti-CD19 chimeric antigen receptor (CAR) redirected T cells have demonstrated profound efficacy in relapsed/refractory B cell malignancies. However, CAR T cells have thus far failed to recapitulate these results in solid tumors. Accumulating evidence suggests that macrophages naturally traffic to and persist in solid tumors. The goal of this study was to evaluate the anti-tumor function of genetically engineered CAR macrophages (CARMA) and assess their potential for translation as a novel immunotherapeutic platform for solid tumors.

Methods

To examine the function of CARs in macrophages, first generation CARs with a CD3ζ intracellular domain were introduced into the THP-1 macrophage model. In vitro function was assessed via quantitative phagocytosis and luciferase-based specific killing assays. To assess the function of the CD3ζ-domain, CD3ζ-null CAR constructs were compared. Ad5f35 was used to transduce primary human monocyte derived macrophages with an anti-HER2 CAR construct, and anti-tumor function was tested in vitro and in vivo. The impact of inhibiting the CD47/SIRPα “do-not-eat-me” signal was tested with blocking antibodies and CRISPR-Cas9 mediated SIRPα knockout. Macrophage M1/M2 phenotype was determined by flow cytometry. Immunodeficient mouse xenograft models of a human HER2(+) ovarian cancer cell-line (SKOV3) were used for in vivo efficacy studies.

Results

CAR19, but not untransduced macrophages, phagocytosed CD19+ (but not CD19-) K562 cells. Deletion of CD3ζ in CAR19 macrophages abrogated phagocytosis and significantly reduced specific killing. Phagocytosis was inhibited by pharmacologic blockade of phagocytic signaling - suggesting that CAR activation drives productive cell-signaling in macrophages. Phagocytosis was also demonstrated in HER2 and mesothelin CARMA models. Blockade of the inhibitory CD47/SIRPα “do-not-eat-me” signal enhanced CARMA phagocytosis of antigen-bearing target cells in a dose-dependent manner (Fig. 27). We identified Ad5f35 as an efficient transduction approach for engineering primary human macrophages, resulting in >70% CAR transduction efficiency. Primary human anti-HER2 CARMA demonstrated targeted phagocytosis and specific killing. Ad5f35 transduction polarized human macrophages to the pro-inflammatory M1 phenotype, and rendered them resistant to downstream M2 subversion by immunosuppressive cytokines and cell lines. Anti-HER2 CARMA was evaluated in vivo in an ovarian cancer xenograft model. Mice that received CARMA had a decrease in tumor burden of approximately two orders of magnitude and had a 30-day survival benefit relative to untreated or control macrophage treated mice (p=0.018, Fig. 28).
Fig. 27

(Abstract P23). Timelapse of a CAR19 macrophage (mRFP+) phagocytosing a CD19+ K562 cells (GFP+) cell (A). CARMA but not Wt macrophages phagocytosed CD19+ but not CD19- tumor cells (B). Phagocytosis was confirmed by Imagestream cytometry, gating on mRFP+GFP+ events (C). Blockade of CD47 and SIRPa led to a dose dependent increase in CARMA phagocytosis. Non-blocking anti-CD47 clone 2D3 (opsonization control) did not enhance phagocytosis (D)

Fig. 28

(Abstract P23). Anti-HER2 primary human CARMA were tested in immunodeficient mouse models of human HER2+ ovarian cancer (schema, a). CARMA but not control untransduced macrophages reduced tumor burden (b, c) and prolonged survival by 30 days (p=0.018, d)

Conclusions

Here, we introduce for the first time that human macrophages engineered with a CAR exert antigen-specific tumor phagocytosis and killing, and propose a novel immunotherapeutic platform for the treatment of diverse solid tumors.

P24 Regulation of T cell sensitivity by TCR-proximal signaling components during anti-melanoma responses

Duane Moogk1, Shi Zhong2, Zhiya Yu3, Ivan Liadi4, William Rittase5, Victoria Fang6, Janna Dougherty1, Arianne Perez-Garcia7, Iman Osman8, Cheng Zhu5, Navin Varadarajan4, Nicholas P Restifo3, Alan Frey9, Michelle Krogsgaard10

1Perlmutter Cancer Center, NYU School of Medicine, New York, NY, USA; 2Xiangxue Pharmaceutical Co., Ltd., Guangdong, Guizhou, People’s Republic of China; 3Surgery Branch, National Cancer Institute, Bethesda, MD, USA; 4University of Houston, Houston, TX, USA; 5The George W. Woodruff School of Mechanical Engineering, Georgia Institute of Technology, Atlanta, GA, USA; 6Skirball Institute of Biomolecular Medicine, NYU School of Medicine, New York, NY, USA; 7Kite Pharma, Inc., Santa Monica, CA, USA; 8Ronald O. Perelman Department of Dermatology and Department of Medicine and Urology, Perlmutter Cancer Center at NYU School of Medicine, New York, NY, USA; 9New York University Langone School of Medicine, New York, NY, USA; 10Department of Pathology and Perlmutter Cancer Center at NYU School of Medicine, New York, NY, USA
Correspondence: Michelle Krogsgaard (krogsm01@nyumc.org)

Background

Immunotherapies for cancers have made great strides in recent years, yet new and improved approaches are required to achieve more durable responses in a greater number of patients. The in vitro expansion phase of adoptive T cell therapy prior to reinfusion into the patient present opportunities to genetically enhance T cell subsets to improve in vivo performance. While the most common genetic modification is the incorporation of engineered antigen-specific TCRs or chimeric antigen receptors, modification to signaling pathways in T cell memory subsets in order to enhance T cell sensitivity is an underexplored strategy. This is mainly because contributions of TCR signaling components that confer differences in activation sensitivity and functional outcomes between CD8+ Tcm and Tem are unclear.

Methods

To understand how TCR-proximal signaling differs significantly between T cell memory subsets, we derived TCM and TEM from the humanized TCR-transgenic melanoma mouse model (JR209). We quantified differences in TCR activation and feedback regulation by novel live-cell imaging technologies, phospho-specific protein assays and used modeling of early TCR signaling to reveal the physiological significance of these differences.

Results

One of the critical steps of T cell triggering is the coordinated phosphorylation and binding of CD3 and Zap-70 by Lck following TCR ligation by pMHC. Here, we show that Tcm and Tem possess differential constitutive Lck activities. Immediately proximal to Lck signaling, we observed enhanced Zap-70 phosphorylation in TEM following TCR ligation compared with TCM. Further, we observed increased intracellular calcium influx and cytotoxic effector function in TEM compared with TCM, and provide evidence that this results from a lower probability of TCM reaching threshold activation signaling due to the decreased magnitude of TCR-proximal signaling. We show that the differences in Lck constitutive activity between CD8+ Tcm and Tem are driven in part by differential regulation by SH2 domain-containing phosphatase-1 (Shp-1) and C-terminal Src kinase (Csk). We demonstrate that inhibition of Shp-1 results in increased constitutive Lck and cytotoxic activity in TCM to levels similar to that of TEM.

Conclusions

Together, this work demonstrates that differential activities of TCR-proximal signaling components may contribute to establishing the divergent effector properties of TCM and TEM. Inhibition of negative regulatory molecules, for example Shp-1 or Csk, or generalized augmentation of T cell sensitivity with miRNA offer potential therapeutic approaches in T cell immunotherapy but must be considered in the context of specificity and optimal targeting.

P25 Co-expression of synthetic PD-1 fusion proteins augments HER2 CAR T cell functionality against glioblastoma

Daniel Landi, Kristen Fousek, Malini Mukherjee, Ankita Shree, Sujith Joseph, Kevin Bielamowicz, Tiara Byrd, Nabil Ahmed, Meenakshi Hegde

Baylor College of Medicine, Houston, TX, USA
Correspondence: Daniel Landi (landi@bcm.edu)

Background

HER2-CAR T cells home to the central nervous system (CNS) and induce tumor regression in patients with glioblastoma (GBM) [1]. However, most responses are transient, as CAR T cells fail to expand and have limited persistence. Ex vivo analyses of tumor infiltrating lymphocytes demonstrate high levels of inhibitory receptors, including PD-1. Monoclonal antibodies blocking PD-1/PD-L1 induce responses in some patients with solid tumors and potentiate anti-tumor T cell activity. However, because antibodies exhibit erratic CNS pharmacokinetics, combining this approach with CAR T cells is not optimal for CNS tumors. We hypothesize that co-expression of a synthetic PD-1 fusion protein will convert PD-L1 into a costimulatory T cell signal, improving expansion, persistence, and anti-tumor activity of adoptively transferred HER2-CAR T cells.

Methods

We generated an array of bicistronic vectors encoding our clinically utilized 2nd generation HER2-CAR (CD28ζ-endodomain) and a PD-1 fusion protein. All PD-1 fusion proteins contained the native PD-1 ectodomain fused to either the CD28 transmembrane and endodomain (PD-1/CD28) or CD8α-transmembrane and 4-1BB-endodomain (PD-1/4-1BB). T cell expansion, persistence, and exhaustion (LAG3, TIM3, PD-1) were measured using flow cytometry following coculture with autologous HER2+/PD-L1+ GBM cells for 2-4 weeks. Cytokine release at 24 hours was measured using standard ELISA, and the xCELLigence impedance-based system was used to evaluate cytolytic activity. Using high-resolution immunofluorescence imaging, we interrogated differences at the CAR T cell/GBM contact point, also referred to as the immunologic synapse (IS).

Results

Compared to conventional HER2-CAR T cells, PD-1/CD28 T cells expanded more quickly with significantly higher IL2/IFNγ-release at 24 hours (Fig. 29), whereas PD-1/4-1BB T cells demonstrated enhanced cytolytic ability against autologous GBM cells in prolonged killing assays (Fig. 30) and better long-term persistence. Inhibitory receptor expression following coculture was comparable among T cell products, but PD-1/4-1BB T cells maintained a greater percentage of central memory cells (CCR7+/CD45RA-). HER2-CAR T cells bearing PD-1 fusion proteins demonstrated increased levels of activated kinases in CD28/4-1BB signaling pathways. Three-dimensional reconstitution and quantification of confocal images of the CAR T cell/tumor interface revealed increased stability of the IS between the HER2-CAR and the HER2 molecule on GBM.
Fig. 29

(Abstract P25). 2nd generation HER2 CAR T cells with and without PD-1/CD28 fusion proteins were incubated with autologous GBM cells at a 1:1 ratio. Cytokine concentrations from culture supernatant collected at 24 hours were measured using standard ELISA

Fig. 30

(Abstract P25). The xCELLigence platform uses impedance across resistor-coated plates to measure adherant tumor cell viability. Once 10,000 tumor cells became adherant and confluent, 1,000 T cells were added that expressed either a HER2 CAR only (H2) or a HER2 CAR with a truncated PD-1 protein (X), PD-1/CD28 protein (28), or PD-1/4-1BB protein (BB). Non-transduced T cells and tumor only wells served as controls

Conclusions

We conclude that PD-L1 can be converted into a costimulatory signal using synthetic PD-1 fusion molecules to drive key T cell activation pathways and enhance stability of the CAR-target antigen interface, leading to improved HER2-CAR T cell functionality against GBM.

References

1. Ahmed N: Autologous HER2 CMV bispecific CAR T cells for progressive glioblastoma: Results from a phase I clinical trial. J Clin Oncol 2015, 33:3008.

P26 Adoptive therapy with tumor-infiltrating lymphocytes for melanoma: interim results of a phase II single-institution study

Sylvia Lee1, David Byrd2, John Thompson3, Shailender Bhatia4, Scott Tykodi4, Judy Delismon5, Liz Chu5, Siddiq Abdul-Alim5, Arpy Ohanian5, Anna Marie DeVito6, Stanley Riddell5, Kim Margolin7

1Seattle Cancer Care Alliance, Fred Hutchinson Cancer Research Center, University of Washington, Seattle, WA, USA; 2University of Washington, Seattle, WA, USA; 3Clinical Research Division at Fred Hutch Cancer Center, University of Washington, Seattle, WA, USA; 4Fred Hutchinson Cancer Research Center, University of Washington, Seattle, WA, USA; 5Fred Hutchinson Cancer Research Center, Seattle, WA, USA; 6Seattle Cancer Care Alliance, Seattle, WA, USA; 7Department of Medical Oncology, City Of Hope, Duarte, CA, USA
Correspondence: Sylvia Lee (smlee@fredhutch.org)

Background

Adoptive cell therapy using tumor-infiltrating lymphocytes (TIL) has been established as an effective treatment option for metastatic melanoma but is available at only a small number of institutions. We have developed a TIL program at the Fred Hutchinson Cancer Research Center and present our initial results in melanoma patients.

Methods

Patients with metastatic melanoma were enrolled in a two-step fashion to a nonrandomized, phase II TIL trial: Step 1 for TIL generation and Step 2 for TIL treatment. TIL are cultured from tumor fragments, using standard methodologies developed at the NCI. At a time of future disease progression, patients are given cyclophosphamide 60mg/kg/day x 2 days, fludarabine 25mg/m2/day x 5 days, then TIL up to 1.5 x 10^11, followed by high-dose IL-2, 600,000 units/kg every 8 hours, up to 14 doses as tolerated. TIL populations are selected for infusion based on growth, phenotype, and MHC-class-I-dependent autologous tumor recognition, as assessed by interferon-gamma release.

Results

Thirty-two patients have been enrolled on Step 1 for TIL generation. Of these, 23(72%) achieved adequate TIL number of >40 x10^6 by 5 weeks of culture. TIL in 17(53%) patients also demonstrated positive autologous tumor reactivity. To date, 7 patients have received TIL therapy on Step 2; all had progressed on prior immunotherapy (Table 3). In these 7 patients, the RECIST 1.1 responses are: 2 CR, 1 PR, 3 SD, 1PD. All patients experienced tumor regression of some or all baseline target lesions; the patient with PD progressed in the brain, but experienced a partial response of extracranial sites. Two patients were on concurrent BRAF therapy and had reached a plateau in their response to BRAF inhibitors prior to start of TIL treatment, and both experienced further tumor reduction after TIL. Responses were associated with a higher percent CD8+TIL, a lower percent Treg(CD4+CD25+CD127-), and TCR oligoclonality, while functional markers PD-1 and TIM3 did not associate with response.
Table 3

(Abstract P26). Characteristics and responses of treated patients

 

Age/Gender

Prior Treatments

M Classification and Metastatic Sites

Response

Response Duration (months)

1

54 F

Anti-CLA-4, interferon (IFN)

M1c: Mesentery, small bowel, lymph nodes (LN)

PR

8

2

34 F

Anti-PD1, anti-CLA-4, IFN, IL-2 + radiation (XRT), temozolomide

M1c: Brain, kidneys, adrenals, bone, liver, LN, subcutaneous (SC)

SD

3

3

61 M

IL-2, BRAF inhibitor

M1c: Brain, liver, lungs, SC, mesentery, chest wall

PD (Progression in brain, but PR in extracranial sites)

8, for extracranial response

4

27 F

IL-2, anti-CTLA-4+IL-21, anti-CTLA-4 + XRT, recombinant IL-15

M1c: Brain, SC, LN, pleura, kidneys, peritoneum

SD

3

5

63 M

Anti-CTLA-4

M1a: Intramuscular, LN

CR

17, ongoing

6

52 M

Anti-CTLA-4

M1a: SC, LN

CR

20, ongoing

7

31 F

Anti-PD1, anti-CTLA-4, decarbazine, BRAF/MEK inhibitor

M1c: Brain, lungs, liver, SC

SD

5

Conclusions

Our single-institution study validates the utility of TIL therapy in patients with advanced melanoma, refractory to other immunotherapy. TIL can induce durable CRs and can also mediate additional tumor regression in patients on active BRAF inhibition. The replication of TIL methodology across different institutions, with reproducible clinical efficacy, supports the feasibility of its widespread application, as well as further investigation into optimizing elements of this treatment modality.

Acknowledgements

We wish to thank the NCI Surgery Branch and MD Anderson for their generous guidance with the development of our TIL therapy program, and Prometheus for their generous IL-2 support.

P27 Functional characterization of CD4+ and CD8+ CD19 chimeric antigen receptor T cells

Isabelle Magalhaes, Jonas Mattsson, Michael Uhlin

Karolinska Institutet, Stockholm, Stockholms Lan, Sweden
Correspondence: Isabelle Magalhaes (isabelle.magalhaes@ki.se)

Background

Chimeric antigen receptor (CAR) T cells targeting CD19 have shown dramatic results in patients with refractory B cell malignancies but the precise mechanisms of how CD19 CAR T cells kill tumor cells are not all well understood.

Methods

Second generation CD19 CAR T cells were produced from peripheral blood mononuclear cells (PBMCs) obtained from pediatric patients with acute lymphoblastic leukemia (ALL) and adult patients with chronic lymphocytic leukemia (CLL). Here we present the phenotype, cytokine and cytotoxic profile of CD4+ and CD8+ CD19 CAR T cells generated from patients and healthy donors (HDs).

Results

T cell frequency in PBMCs from patients (ALL and CLL) was low (<4% of lymphocytes). In patients with CLL, 90% of lymphocytes were CD19+ B cells. In patients with CLL, as compared to HDs, the majority of CD8+ T cells displayed a terminally differentiated phenotype, while CD4+ T cells were mostly effector memory cells. Retroviral transduction was performed on PBMCs without prior B cell depletion or T cell enrichment step. CAR T cell cytotoxicity towards CD19+ target cells was mediated via granzymes, but not perforin, Fas-L or TRAIL. In patients and HDs, CD4+ and CD8+ CD19 CAR T cells showed a comparable cytokine (IL-2, IFN-g, TNF) production, CD107a expression in response to stimulation with K562 CD19+ cells. The majority of CAR T cells were polyfunctional (≥2 functions). However, patients with CLL, as compared to HDs and patients with ALL, displayed a higher frequency of IFN-g producing and polyfunctional CD4+ and CD8+ CD19 CAR T cells. When stimulated with autologous CD19+ B cells, as compared to K562 CD19+, lower frequencies of CD4+ and CD8+ CD19 CAR T cells produced IL-2, IFN-g and TNF. The frequency of polyfunctional CD19 CAR T cells was lower when stimulated with autologous B cells as compared to K562 CD19+ cells. Stimulation with other CD19+ cell lines induced a different cytokine production profile of CD19 CAR T cells.

Conclusions

CD4+ and CD8+ CD19 CAR T cells display comparable differentiation phenotype, cytokine production and cytotoxic capacity. The presence of a high frequency of CD19+ B cells during the generation of CAR T cells did not have an impact on CAR T cell phenotype. However CAR T cells from patients with CLL produced more cytokines when stimulated with CD19+ target cells suggesting that activation of CD19 CAR T cells by B cells during cell expansion impacts the cytokine profile. Furthermore, our data show that the level of CD19 cell surface expression modulates CAR T cells cytokine production.

P28 Directing traffic: fostering chemokine receptor expression on tumor-infiltrating lymphocytes improves re-trafficking in vivo

Satoshi Nemoto1, Patricio Pérez Villarroel1, Ryosuke Nakagawa1, James J Mule2, Adam W. Mailloux1

1Translational Science, H. Lee Moffitt Cancer Center, Tampa, FL, USA; 2Immunology Program, Cutaneous Oncology Program, H. Lee Moffitt Cancer Center, Tampa, FL, USA
Correspondence: Adam W. Mailloux (adam.mailloux@moffitt.org)

Background

Previously, a critical role for chemokines was found in a unique immune-related gene expression signature (GES) for immune cell recruitment and tertiary lymphoid structure formation in metastatic melanoma [1]. Regarding adoptive cell therapy (ACT) of tumor-infiltrating lymphocytes (TIL), we hypothesized that expression of chemokine receptors (CCR)2, CCR5, and CCR7, which bind multiple GES chemokines, favor TIL re-trafficking to the tumor, thereby bolstering ACT efficacy.

Methods

We utilized a preclinical ACT model in which TIL from syngeneic MC-38 colon carcinomas grown in wild-type (WT) C57BL/6 mice, or gene knock-out (KO) C57BL/6 mice lacking CCR2, CCR5, or CCR7, were expanded ex vivo, and adoptively transferred into MC-38-bearing C57BL/6 recipient mice bearing congenic CD45.1. After seven days, tumor burden was assessed, and Spanning-tree Progression Analysis of Density-normalized Events (SPADE) of 18-parameter cytometry data was used to identify and quantify TIL subsets, and 51Cr-release assays quantified cytotoxicity among sorted TIL populations.

Results

ACT of WT TIL reduced tumor burden by 50% compared to untreated mice (p=0.0098). This benefit was lost when transferring CCR5KO or CCR7KO TIL, and reduced when transferring CCR2KO TIL (all p < 0.05). MC-38 tumors contained 90%, 85%, and 70% less transferred TIL with ACT of CCR2KO, CCR5KO, or CCR7KO TIL, respectively, compared to WT TIL (all p < 0.01) . SPADE identified eight novel subsets in re-trafficked TIL with unique patterns of activation, memory, and immune checkpoint markers (Fig. 31). Nearly all subsets displayed impaired re-trafficking with CCR2KO, CCR5KO, or CCR7KO TIL ACT (all p < 0.05). Cytotoxicity assays suggest a range of cytotoxic potential among identified subsets, with CD8+TIL3 and 4 subsets being most cytotoxic (all p < 0.05). During TIL expansion, IL-2 up-regulated CCR2, CCR5, and CCR7 in a cell density-dependent fashion (all p < 0.007). ACT of TIL with up-regulated CCRs increased tumor re-trafficking three-fold (p < 0.05), and decreased tumor burden by an additional 25% (p < 0.05) versus WT TIL without up-regulated CCRs.
Fig. 31

(Abstract P28).

Conclusions

CCR2, CCR5, and CCR7 are vital for TIL re-trafficking to MC-38 tumors following ACT, including cytotoxic subsets among eight novel TIL phenotypes. Fostering CCR expression during TIL expansion augments ACT efficacy in murine MC-38 colon carcinoma. Translational studies to human TIL ACT are currently underway.

Acknowledgements

Supported by NCI–NIH (CA148995-01; P30CA076292; P50CA168536), V Foundation, Dr. Miriam and Sheldon G. Adelson Medical Research Foundation, and the Chris Sullivan Foundation.

References

1. Messina JL, Fenstermacher DA, et al: 12-Chemokine gene signature identifies lymph node-like structures in melanoma: potential for patient selection for immunotherapy?Scientific Reports 2012, 2:765-770.

P29 Provision of inducible MyD88 and CD40 co-stimulation in CAR T cells results in potent antitumor activity in preclinical solid tumor models

Melinda Mata1, Phuong Nguyen1, Claudia Gerken1, Christopher DeRenzo1, David M Spencer2, Stephen Gottschalk1

1Baylor College of Medicine, Center for Cell and Gene Therapy, Houston, TX, USA; 2Bellicum Pharmaceuticals and Baylor College of Medicine, Houston, TX, USA
Correspondence: Melinda Mata (mxmata@texaschildrens.org)

Background

Although adoptive immunotherapy using T cells expressing chimeric antigen receptors (CAR) is successful in refractory hematological malignancies, limited clinical responses have been observed in solid tumors. We reasoned that introducing an inducible co-stimulation (iCO-STIM) gene into T cells would allow for improved activation of CAR T cells, resulting in enhanced antitumor activity. Due to the co-stimulatory properties of MyD88 and CD40 in T cells, we explored whether CAR T cells expressing an iCO-STIM molecule consisting of a myristoylation-targeting sequence, two FKBP dimerizing domains, MyD88ΔTIR, and the intracellular domain of CD40, have superior effector function relative to standard CAR T cells in vitro and in vivo.

Methods

T cells expressing a HER2-CAR and iCO-STIM (HER2-CAR/iCO-STIM T cells) were generated by transduction with a retroviral vector encoding FRP5.ζ (HER2-CAR), a 2A peptide, and iCO-STIM. HER2-CAR/iCO-STIM T cell effector function was then evaluated in vitro and in murine xenograft models.

Results

In the presence of CID (AP20187), HER2-CAR/iCO-STIM T cells produced significantly higher levels of IL2 (p < 0.05) compared to HER2-CAR/iCO-STIM T cells in the absence of CID or HER2-CAR.CD28.ζ T cells in co-culture assays with HER2+ tumor cells (MDA-HER2, LM7, A549). In contrast, HER2-negative tumor cells (MDA) did not induce IL2 production by HER2-CAR/iCO-STIM T cells +/- CID or HER2-CAR.CD28.ζ T cells. In repeat stimulation assays, HER2-CAR/iCO-STIM T cells showed robust antigen-dependent expansion in the presence of CID and were able to lyse HER2-positive LM7 cells after 4 re-stimulations compared to HER2-CAR.CD28.ζ T cells (p < 0.0001). In vivo, a low dose of HER2-CAR/iCO-STIM T cells (3x105) + CID had superior antitumor activity in the metastatic LM7 osteosarcoma murine xenograft model compared to HER2-CAR/iCO-STIM T cells without CID or HER2-CAR.CD28.ζ T cells, resulting in a significant survival advantage (p < 0.001). In 3 mice that developed late recurrences after HER2-CAR/iCO-STIM T cell + CID therapy, a second dose of CID, given 100 days post T cell injection, eliminated 2/3 tumors. Superior antitumor activity of HER2-CAR/iCO-STIM T cell + CID therapy was confirmed in the HER2+ A549 lung cancer murine xenograft model.

Conclusions

CID-mediated activation of MyD88 and CD40 co-stimulatory signals in HER2-CAR T cells results in superior effector function compared to standard HER2-CAR T cells. Thus, expressing iCO-STIM molecules in CAR T cells has the potential to improve the efficacy of CAR T cell therapy approaches for solid tumors. In addition, our results indicate that the CID/iCO-STIM system will enable the ‘remote control’ of infused T cells.

P30 In vivo administration of immune checkpoint inhibitors prior to tumor harvest enhances the function of tumor-infiltrating T lymphocytes expanded for adoptive T cell transfer

Mélissa Mathieu, Sandy Pelletier, John Stagg, Simon Turcotte

CRCHUM, Montréal, PQ, Canada
Correspondence: Mélissa Mathieu (melissamathieu911@gmail.com)

Background

T cell reactivity against mutated antigens, derived from cancer genomic alterations, is a key mediator of immunotherapy efficacy. Adoptive cell transfer (ACT) of mutation-reactive tumor-infiltrating T lymphocytes (TILs) however has been only effective in a minority of patients with metastatic gastrointestinal cancers. We hypothesize that the low frequency of mutation-reactive TILs and their exhaustion features may contribute to the lack of efficacy of TIL ACT, and that these factors can be overcome by in vivo administration of blocking antibodies against CTLA-4 or PD-1 prior to tumor harvest and TIL expansion.

Methods

We selected the mouse MC-38 colorectal cancer model to study CD8+ TIL reactive to 2 mutated, 2 self, and 1 retroviral antigens using co-cultures with peptide-pulsed splenocytes followed by IFN-γ intracellular staining. MC-38 cancer cells were inoculated subcutaneously and allowed to grow until 20mm2. Then anti-CTLA-4 (9H10) and/or anti-PD-1 (RPM1-14) antibodies were administered twice (4 days interval). Three days following the last antibody injection, tumors were harvested and analyzed for immune cell infiltration, phenotype and functionality. TILs were also expanded in vitro and characterized regarding the specificity and the functionality against the known antigens.

Results

The administration of anti-CTLA-4 prior to tumor harvest increased immune cell infiltration, decreased the proportion of myeloid derived suppressor cells and regulatory T cells in the tumors and enhanced CD8+ T cell IFN-γ and TNF-α production. Administration of both anti-CTLA-4 and anti-PD-1 was more effective to eliminate the tumor burden and recapitulated the effects observed following anti-CTLA-4 injection alone. Anti-CTLA-4 and/or anti-PD-1 alone did not appear to change the relative frequency of TIL reactive to mutated, self, or viral antigens but increased their polyfunctionality.

Conclusions

In vivo pre-conditioning of MC-38-bearing mice with immune checkpoint blocking antibodies may generate TILs more fit for ACT immunotherapy. Experiments are underway to compare the efficacy of this approach to conventional TIL ACT in this mouse model.

P31 Redirecting gene-engineered T cells through covalent attachment of targeting ligands to a universal immune receptor

Nicholas Minutolo1, Prannda Sharma1, Andrew Tsourkas2, Daniel J Powell1

1Department of Pathology and Laboratory Medicine, Perelman School of Medicine, University of Pennsylvania, Philadelphia, PA, USA; 2Department of Bioengineering, University of Pennsylvania, Philadelphia, PA, USA
Correspondence: Nicholas Minutolo (ngminutolo@gmail.com)

Background

Infusion of T cells genetically engineered to express a chimeric antigen receptor (CAR) is a promising approach for the treatment of certain cancers. Though CAR T cells are highly efficacious against CD19+ hematological malignancies, limitations exist in broadening their use. Conventional CAR T cells target a single tumor associated antigen (TAA), limiting their effectiveness against tumors with heterogeneous TAA expression as well as emerging antigen loss variants, as observed in CD19 CAR trials. Additionally, stably engineered CAR T cells may continually proliferate and activate in the presence of antigen, potentially causing fatal toxicity, without a method of elimination. To overcome these issues, we and others have developed a variety of universal immune receptors (UIRs) that allow for targeting of multiple TAAs with a single receptor-expressing T cell. Although these UIRs are a promising new technology, their reliance on noncovalent interactions between receptor and targeting ligand can lead to potential issues of affinity, specificity and activity.

Methods

Our UIR platform employs the use of the SpyCatcher and SpyTag proteins that, when combined, can form a covalent bond with high efficiency both in vitro and in vivo. Our SpyCatcher immune receptor is composed of an extracellular SpyCatcher domain attached to intracellular T cell signaling motifs in a lentiviral expression vector. Anti-TAA antibodies conjugated to SpyTag are used to confer redirected specificity to SpyCatcher expressing T cells. Measurements of T cell effector function include T cell cytokine secretion, activation, and targeted tumor cell lysis in vitro.

Results

In this study, we demonstrate the first universal immune receptor platform that allows for the endowment of function through post-translational covalent attachment of targeting ligands, securing their loading on the T cell surface while expanding recognition specificity. We demonstrate that the SpyCatcher immune receptor is expressed in primary human T cells and allows for specific T cell activation and cytokine secretion against plate bound SpyTag. Notably, in the presence of SpyTag-labeled targeting antibody, SpyCatcher T cells recognize and lyse antigen-expressing human tumor cells in a target-specific and dose-dependent fashion.

Conclusions

The SpyCatcher immune receptor is the first universal immune receptor designed for its capacity to covalently bind targeting ligands and redirect T cells against a diverse array of antigens, addressing current limitations of conventional CAR T cell therapy.

P32 Simple automated manufacturing of gene engineered T cells under serum free conditions for adoptive T cell therapy

Nadine Mockel-Tenbrinck, Daniela Mauer, Katharina Drechsel, Carola Barth, Katharina Freese, Ulrike Kolrep, Silke Schult, Mario Assenmacher, Andrew Kaiser

Miltenyi Biotec, Bergisch Gladbach, Nordrhein-Westfalen, Germany
Correspondence: Nadine Mockel-Tenbrinck (nadinet@miltenyibiotec.de)

Background

Adoptive immunotherapy using gene-modified T cells redirected against cancer has proven clinical efficacy and tremendous potential in several medical fields. However, such personalized medicine faces several challenges in the complexity associated with the current clinical manufacturing methods. The processes are mostly suboptimal requiring cell manufacturers to deal with open steps, liquid handlings between devices used, manual interventions, removal of activation beads and often the use of reagents for which commercial availability is not in line with the high and increasing demand. GMP compliant human AB serum is one of such reagents. Therefore, developing improved methods to generate gene-engineered T cells suitable for clinical use that do not require serum are essential for the commercial scalability of such therapies.

Methods

We have developed a robust and reproducible automated manufacturing process for the lentiviral gene-modification and expansion of selected T cells. The CliniMACS Prodigy TCT (for T Cell Transduction) process software allows automated purification and polyclonal T cell stimulation followed by gene-modification and expansion of T cells in a single-use closed tubing set.

Results

Here we show that the TCT process enables the manufacturing of gene-modified T cells without the need for serum supplementation (human AB serum) when using TexMACS GMP Medium. Furthermore, implementation of a humanized recombinant activation reagent TransAct allows for a simplification of the process whereby the “bead removal” step is unnecessary. Comparable results from healthy donor or patients starting cells are demonstrated.

Conclusions

Taken together the TCT process in combination with the CliniMACS Prodigy and minimal user interactions enables the preparation of gene-modified T cells in serum free conditions. Process risks, due to use of different devices, unnecessary manipulations, or potential shortage of human AB serum availability can be minimized by automation of the entire process as developed and shown here. Overall, these improvements are meant to fully support commercial treatment of a large number of patients.

P33 Tumor Infiltrating lymphocytes from soft tissue sarcoma have tumor-specific function

John Mullinax, MacLean Hall, Julie Le, Krithika Kodumudi, Erica Royster, Allison Richards, Ricardo Gonzalez, Amod Sarnaik, Shari Pilon-Thomas

H. Lee Moffitt Cancer Center, Tampa, FL, USA
Correspondence: John Mullinax (john.mullinax@moffitt.org)

Background

Adoptive Cell Transfer (ACT) using Tumor Infiltrating Lymphocytes (TIL) for unresectable metastatic melanoma results in a median overall survival of 52 months at our institution. This stands in contrast to the median overall survival for metastatic soft tissue sarcoma (mSTS) which is 12 months. The purpose of this report is to describe the phenotype and function of TIL from fresh surgical sarcoma specimens as a rationale for applying ACT to mSTS.

Methods

Fresh surgical sarcoma specimens were acquired under an IRB-approved protocol. Half of the specimen was digested and phenotyped by flow cytometry. The remaining half was plated as fragments for the isolation of TIL, which were expanded in vitro with conditions validated for melanoma-derived TIL. Cultured TIL were phenotyped by flow cytometry and propagated further with a rapid expansion protocol (REP). Tumor-specific reactivity was assessed by co-culture of TIL with autologous tumor or HLA-matched cell lines with measurement of IFN-gamma using ELISA.

Results

Sixteen patient-derived sarcoma specimens were collected. Histology of primary tumor included dedifferentiated liposarcoma (9), well-differentiated liposarcoma (2), undifferentiated pleomorphic sarcoma (2), and one each of gastrointestinal stromal tumor, myxofibrosarcoma and synovial sarcoma. Analysis of tumor digests indicated an average of 48% of cells from the lymphocyte gate were CD3+ (range 3.6%-76%). TIL were grown from all specimens, with TIL observed in 152 out of 192 (79%) fragments. The phenotype of the CD3+ subpopulations from TIL cultures included a median of 90.8% (range 2-99.9%) CD8+ and 2.3% (range 0-96.4%) CD4+ T cells. TIL were expanded in a REP to clinically meaningful numbers (mean 1385-fold) with no change in the CD8+ T cell proportion. Tumor-specific function of TIL generated from fragments was measured in two patients. There was tumor-specific IFN-gamma release (mean 148.8±13.5 pg/ml) in TIL co-cultured with HLA-matched cell lines and also in TIL co-cultured with autologous tumor digest (mean 259.9±14.7 pg/ml). The degree of IFN-gamma release was significantly greater when TIL were co-cultured with autologous digest compared to an HLA-mismatched cell line (p=0.0295).

Conclusions

CD3+CD8+ TIL can be isolated from human sarcoma tumors in vitro, expanded to meaningful numbers for therapeutic use, and demonstrate reactivity to the tumor from which they were cultured. These data form the basis for efforts to develop a clinical trial using ACT for patients with advanced sarcoma.

Acknowledgements

Work funded by a grant from the Chotiner Family Foundation.

P34 Preclinical development of tumor infiltrating lymphocyte (TIL) based adoptive cell transfer (ACT) immunotherapy for patients with sarcoma

Morten Nielsen1, Anders Krarup-Hansen2, Dorrit Hovgaard3, Michael Mørk Petersen3, Anand Chainsukh Loya4, Niels Junker5, Inge Marie Svane5

1Center for Cancer Immune Therapy and Department of Oncology, Herlev Hospital, Herlev, Hovedstaden, Denmark; 2Department of Oncology, Herlev University Hospital, Herlev, Hovedstaden, Denmark; 3Department of Orthopaedic Surgery, Copenhagen University Hospital, Copenhagen, Hovedstaden, Denmark; 4Department of Pathology, Copenhagen University Hospital, Copenhagen, Hovedstaden, Denmark; 5Department of Oncology, Center for Cancer Immune Therapy, Herlev University Hospital, Herlev, Hovedstaden, Denmark
Correspondence: Morten Nielsen (morten.nielsen.03@regionh.dk)

Background

ACT based on infusion of autologous TILs has the ability to induce complete and durable response in some patients with advanced malignant melanoma [1]. We believe that this approach could also be effective in sarcoma. In this preclinical study we are investigating feasibility of expanding TILs from sarcoma, as well as performing functional in vitro analyses on these TILs. This abstract is an update on our results so far.

Methods

A portion (>1 cm3) of the excised sarcoma tumor tissue is cut into fragments and placed in a growth medium containing IL-2 for initial TIL expansion. Afterwards, TIL cultures undergo a rapid expansion protocol (REP) expansion by adding OKT-3 and feeder cells. Phenotype and functional analyses are performed using flow cytometry and Elispot.

Results

To this date we were able to expand TILs from 18 of 20 tumor samples. TILs were harvested and frozen when an estimated number of 100x106 to 200x106 cells were reached. Mean expansion time was 32 days (16 - 61). 87.7% (36.4 – 99.1) of these cells were CD3+, and of these 66.7% (16.3 – 99.1) were CD4+, and 21.8 % (0.1 – 50.6) were CD8+. Most of the expanded TILs were effector memory subtype, while a smaller fraction was the more differentiated effector T cells. REP expansion rates ranged from 630 fold to 2.300 fold, and followed expansion pattern similar to TILs from malignant melanoma. TILs from 6 of 10 tested tumor samples with 4 different sarcoma subtypes (undifferentiated pleomorphic sarcoma, myxofibrosarcoma, myxoid liposarcoma and osteosarcoma) demonstrated reactivity against autologous tumor cells using Elispot. Further assessment is ongoing.

Conclusions

We were able to expand TILs from 90% of the included tumor samples to numbers needed for possible future clinical implementation. TILs were a mix of CD4+ and CD8+ with CD4+ being predominant. As of yet we have demonstrated TIL reactivity against autologous tumor cells from 6 of 10 tested patients. Thus, we conclude that it is feasible to translate TIL based ACT into clinical testing in sarcoma patients.

References

1. Rosenberg SA, Yang JC, Sherry RM, et al: Durable complete responses in heavily pretreated patients with metastatic melanoma using T-cell transfer immunotherapy. Clin Cancer Res 2011, 17(13):4550-4557.

P35 Human natural killer cells engineered to express a chimeric NK activating receptor have activity against highly suppressive cells of the solid tumor microenvironment

Charlotte Rivas1, Robin Parihar1, Stephen Gottschalk2, Cliona M Rooney1

1Baylor College of Medicine, Houston, TX, USA; 2Center for Cell and Gene Therapy, Baylor College of Medicine, Houston, TX, USA
Correspondence: Robin Parihar (rxpariha@txch.org)

Background

The suppressive microenvironment of solid tumors inhibits the anti-tumor activity of endogenous and chimeric antigen receptor (CAR)-bearing T cells, thereby limiting the efficacy of adoptive T cell therapies. Myeloid-derived suppressor cells (MDSCs; CD33+ CD11b+ HLA-DRlow) and regulatory T cells (Tregs; CD4+ CD25high FoxP3+ Helios+) contribute to the inhibitory tumor microenvironment (TME) through secretion of suppressive cytokines, expression of inhibitory ligands, and by promoting tumor neo-vascularization. NKG2D is an activating surface receptor expressed on natural killer (NK) cells, whose ligands are highly expressed by human Tregs and MDSCs. We genetically modified primary human NK cells to express a chimeric NKG2D molecule (ectodomain of endogenous NKG2D fused to the cytotoxic CD3-zeta chain; called NKG2D.z) in order to promote NK cell activation against Tregs and MDSCs. We hypothesized that NKG2D.z NK cells would exhibit enhanced cytolytic activity against suppressive autologous Tregs and MDSCs via the chimeric NKG2D, as well as secrete chemokines and cytokines that recruit and activate tumor-specific T cells. The objective of the current study was to compare the activity of NKG2D.z NK cells vs. non-transduced (NT)-NK cells on Tregs, MDSCs, or the combination.

Methods

We isolated MDSCs and Tregs from normal donors and patients with solid tumors, confirmed their phenotype by flow cytometry and their suppressive activity in T and NK cell proliferation assays, and used them as targets in NK cell cytotoxicity and co-culture assays.

Results

NT (endogenous) and mock-engineered (empty vector control) NK cells were unable to mediate cytotoxicity or release pro-inflammatory cytokines in response to autologous MDSCs or Tregs, either alone or in combination. In contrast, NKG2D.z NK cells exhibited enhanced cytolysis and secreted T cell-recruiting and -activating cytokines in response to both suppressive cell types alone. Further, when NKG2D.z NK cells were co-cultured in a highly suppressive environment containing both MDSCs and Tregs, their cytolytic and cytokine-secreting activities against either cell type were unimpaired. We have established an in vivo TME model comprising tumor cells with MDSCs and Tregs, and experiments testing the ability of NKG2D.z NK cells to eliminate suppressive cells and recruit endogenous or CAR-T cells in vivo are underway.

Conclusions

Our results suggest that our modified NK cells may reverse immune suppression by the TME and improve T cell-based immune therapies for solid tumors.

Acknowledgements

Authors thank Charles L. Sentman, PhD for initial use of the NKG2D.zeta sequence.

P36 Cytokines induced by pre-B leukemia progression mediates irreversible T cell dysfunction

Haiying Qin1, Sang Nguyen1, Paul Su1, Chad Burk1, Brynn Duncan1, Bong-Hyun Kim2, M. Eric Kohler1, Terry Fry1

1National Cancer Institute, NIH, Bethesda, MD, USA; 2Frederick National Laboratory for Cancer Research Leidos Biomedical Research, Inc., Frederick, MD, USA
Correspondence: Haiying Qin (qinh@mail.nih.gov)

Background

Acute lymphoblastic leukemia (ALL) is the most common childhood malignancy. The cure rate has reached up 90% with modern therapy, outcomes are still very poor for patients who relapse. Immunotherapy with donor lymphocyte infusions for ALL has not demonstrated the success seen in other hematologic malignancies, suggesting there are characteristics inherent to ALL that make it less amenable to detection and elimination by the immune system. T cell dysfunction in the setting of leukemia has been well described, but the mechanisms have not been fully elucidated. It is also unclear if this T cell dysfunction underlies some of the treatment failures seen in patients receiving adoptive therapy with chimeric antigen receptor (CAR) expressing T cells. Immune checkpoint blockade has led to advances in the treatment of many solid tumors, but it has yet to demonstrate similar success in ALL.

Methods

Using two preclinical models of pre-B cell ALL, we studied the negative impact of ALL progression on T cell function and the efficacy of CAR T cell therapy and immune checkpoint blockade. We also dissected the mechanism underlying the observed T cell dysfunction.

Results

Prophylactic vaccination protects mice against the murine pre-B ALL in a T cell-dependent manner. However, therapeutic vaccination is ineffective, even in the setting of low disease burden. Adoptive transfer of primed T cells from immunized donors can completely eradicate established leukemia; however, this efficacy is not seen with the adoptive transfer of T cells from leukemia-bearing mice. Expression of a CD19 CAR in T cells from leukemic mice fails to eradicate ALL, while the CAR T cells derived from naïve mice can. T cells from leukemic mice express markers of T cell dysfunction. The expression of these molecules was associated with elevated levels of IL6, TNFa and IL10 in the serum of leukemia-bearing mice. Incubation of naïve T cells in these cytokines ex vivo recapitulated the upregulation of exhaustion markers seen in vivo, suggesting these cytokines play a role in the observed T cell dysfunction. Blockade of PD-1, its ligand, PD-L1, or Tim3 were ineffective at reversing T cell dysfunction and preventing leukemia progressions in vivo, suggesting other mechanisms must be targeted to restore immune function in leukemia bearing hosts.

Conclusions

Cytokines induced by Pre-B Leukemia progression mediates irreversible T cell dysfunction. These findings underscore the need to elucidate the mechanisms of leukemia-induced immune suppression to fully optimize the use of CAR-expressing T cells in the treatment of ALL.

P37 Identification of a recurrent high-affinity MHC class I restricted neoepitope in neuroblastoma using ProTECT

Arjun A. Rao1, Noam Teyssier1, Jacob Pfeil1, Nikolaos Sgourakis1, Sofie Salama1, David Haussler2

1University of California, Santa Cruz, Santa Cruz, CA, USA; 2UC Santa Cruz Genomics Institute, University of California, Santa Cruz, Santa Cruz, CA, USA
Correspondence: Arjun A. Rao (aarao@ucsc.edu)

Background

T cells are trained to differentiate between cell-surface MHC-displayed peptide sequences from self- and non-self proteins and act on the latter. The numerous mutations often associated with cancers can occur in coding regions of the genome and modify the sequence of wild-type proteins, potentially creating targets for immunotherapies. We have developed an analysis pipeline ProTECT (Prediction of T cell Epitopes for Cancer Therapy) to identify and rank neo-epitopes in terms of immunogenicity. Running ProTECT on a set of recurrent neuroblastomas showed that recurrent tumors share neo-epitopes with their corresponding primary tumors suggesting that immunotherapies could provide long-term protection.

Methods

ProTECT accepts paired tumor and normal DNA sequencing fastq files, and tumor RNA sequencing fastqs. Mutations are called using a panel of callers, and are annotated to identify coding mutations. Prediction of self-MHC:mutated-peptides is carried out and the final binding predictions are ranked using an in-house algorithm. We support both MHCI and MHCII predictions.

Results

Running ProTECT on 6 recurrent neuroblastoma samples (NBL) from the TARGET (Therapeutically Applicable Research To Generate Effective Treatments) project revealed that the relapsed tumor inherits neo-epitopes predicted in the primary tumor. We also predicted neo-epitopes from 2 well-known hotspot mutations in NBL (NRAS Q61K and ALK R1275Q) that bind to common MHC alleles (HLA-A*01:01, HLA-A*03:01 and HLA-B*15:01). We carried out in vitro refolding and crystallization assays [1] for the five highest-ranking mutant NRAS and ALK predictions. Properly conformed MHC trimers were verified by a single, monodisperse peak after anion exchange chromatography and validated by SDS gel electrophoresis. Mass spec confirmed bound peptide for 4/5 tested predictions and 3 of these were used to set up hanging-drop crystallization trials in various conditions. Positive hits were obtained for one (MAQDIYRASY::HLA-B*15:01). ProTECT has also been run on a large subset of The Cancer Genome Atlas (TCGA). We aim to reveal clinically relevant hotspot-mutation:MHC pairs.

Conclusions

We have described a pipeline for identification and ranking of therapeutically relevant neo-epitopes. We have predicted potentially therapeutic targets for NBL that have been validated in vitro.

References

1. Garboczi DN, Hung DT, Wiley DC: HLA-A2-peptide complexes: refolding and crystallization of molecules expressed in Escherichia coli and complexed with single antigenic peptides . PNAS 1992, 89:8.

P38 A higher-affinity variant of a GD2-specific CAR significantly enhances potency in vivo and allows for a novel model of on-target off-tumor toxicity

Sarah A. Richman1, Selene Nunez-Cruz2, Zack Gershenson2, Zissimos Mourelatos3, David Barrett1, Stephan Grupp1, Michael Milone3

1Children's Hospital of Philadelphia Division of Oncology, Philadelphia, PA, USA; 2University of Pennsylvania, Philadelphia, PA, USA; 3Department of Pathology and Laboratory Medicine, Division of Neuropathology, University of Pennsylvania, Philadelphia, PA, USA
Correspondence: Sarah A. Richman (richmans@email.chop.edu)

Background

As many potential targets of chimeric antigen receptor (CAR) T cell immunotherapy are self-antigens that are over-expressed in tumors but also present at lower levels on some normal tissue, understanding the nature of on-target off-tumor toxicity and how to overcome it is important in the development of new CAR T cell therapies. Preclinical modeling of such toxicity is complicated by the fact that most antigens are not shared between humans and mice, and strategies have largely relied on co-injection of antigen-low tumors or introduction of human antigen into mouse tissue by viral gene delivery. The GD2 tumor antigen would provide an excellent model in which to study on-target off-tumor toxicity as the exact glycolipid antigen is naturally shared between mice and humans. However, as of yet, GD2-specific CAR T cells have yielded modest efficacy and little toxicity in preclinical studies. Here we have engineered a higher potency GD2 CAR by introducing an affinity-enhancing mutation (E101HK), previously described by Horwacik et al., that we show significantly enhances CAR T cell activity and provides a model for toxicity.

Methods

Primary human T lymphocytes were transduced with lentivirus encoding either wild-type 14G2a-based anti-GD2 CAR, E101K mutant GD2 CAR, or a negative control CAR. After standard stimulation and expansion, T cells were analyzed for CAR surface expression by flow cytometry and for in vitro effector function by chromium release and IFN gamma ELISA. To evaluate in vivo, NOD-SCID-IL2Rγc-/- (NSG) mice were injected with the luciferase-expressing GD2-high human neuroblastoma cell line SY5Y, and four days later 3,000,000 CAR+ T cells were injected via tail vein. Tumor burden was measured using in vivo bioluminescence, and tumor and normal tissues were evaluated histologically by H&E staining and immunohistochemistry.

Results

The higher affinity mutant displayed comparable surface expression and T cell expansion but significantly enhanced GD2-specific cytotoxicity and cytokine release in vitro and tumor control in vivo. However, this enhanced efficacy was associated with severe CNS toxicity causing neurologic symptoms and death, and post-mortem evaluation of tissues revealed extensive CAR T cell infiltration into certain brain structures, particularly cerebellum, known to contain GD2.

Conclusions

The introduction of an affinity-enhancing mutation into the GD2-specific CAR dramatically increases CAR T cell potency and permits off-tumor CAR T cell activity in areas of the brain containing GD2. This scenario provides a new opportunity to investigate the mechanism of this toxicity and test strategies to achieve a therapeutic window.

P39 Evaluating the potential of Müllerian inhibiting substance type II receptor (MISIIR) as a target for CAR T cell therapy against ovarian cancer

Alba Rodriguez-Garcia1, Matthew K Robinson2, Gregory P Adams2, Daniel J Powell3

1University of Pennsylvania, Philadelphia, PA, USA; 2Fox Chase Cancer Center, Philadelphia, PA, USA; 3Department of Pathology and Laboratory Medicine, Perelman School of Medicine, University of Pennsylvania, Philadelphia, PA, USA
Correspondence: Alba Rodriguez-Garcia (albarod@mail.med.upenn.edu)

Background

Ovarian cancer is responsible for 5% of cancer-related deaths among women, and the majority of the cases are diagnosed at a late stage, accounting for a 5-year survival rate of 27%. Therefore, there is a dire need for effective therapies. The recent success of adoptive cell therapy using T cells engineered to express anti-CD19 chimeric antigen receptors (CARs) for the treatment of hematologic malignancies, rationalizes the development of similar strategies for solid tumors such as ovarian cancer. The achievement of safe, effective therapy requires the selection of a target antigen that is overexpressed in malignant cells but present in few to no normal cells. The Müllerian inhibiting substance type 2 receptor (MISIIR) is a member of the TGF-β receptors family involved in the regression of the primordial female reproductive tract in male embryos. This action is exerted through its interaction with soluble Müllerian inhibiting substance (MIS), triggering a downstream signaling cascade that induces apoptosis. MIS signaling through MISIIR has been shown to cause growth inhibition in ovarian, breast, prostate and endometrial cancer cell lines in vitro. In humans, MISIIR is expressed at very low levels in a restricted set of healthy tissues but is overexpressed in gynecologic cancers, including 69% of epithelial ovarian cancers, making it a candidate target antigen.

Methods

Here, we evaluate for the first time the potential of MISIIR as a target for CAR T cell therapy. In this work, we generated and functionally tested 5 distinct CARs comprised of different human MISIIR-specific single-chain antibody variable fragments (scFv) isolated from a phage display library coupled to the T cell signaling domains from CD27 and CD3Z.

Results

All the CARs were efficiently expressed primary human T cells transduced using recombinant lentivirus technology and showed specific binding and reactivity against recombinant MISIIR protein. Interestingly, when co-cultured with target cells engineered to overexpress MISIIR, just one of the CARs, GM7-27Z, showed specific reactivity in terms of cytolytic function and proinflammatory cytokines secretion. The activity of this CAR was further evaluated in vitro and in vivo in a panel of tumor cells lines expressing different levels of the target antigen.

Conclusions

Although the assessment of CAR-mediated antitumor activity and on-target off-tumor toxicity potential in vivo is required, the results obtained so far support the further exploration of an anti-MISIIR CAR-based therapy for the effective treatment of ovarian cancer as well as other gynecologic malignancies.

P40 IL-2 in adoptive cell therapy–local production from an adenovirus vector instead of systemic administration results in safety and efficacy gains

João Santos1, Riikka Havunen2, Mikko Siurala2, Víctor Cervera-Carrascón1, Suvi Parviainen1, Marjukka Antilla3, Akseli Hemminki2

1TILT Biotherapeutics, Helsinki, Uusimaa, Finland; 2University of Helsinki, Helsinki, Uusimaa, Finland; 3Finnish Food Safety Authority, Helsinki, Uusimaa, Finland
Correspondence: João Santos (joao@tiltbio.com)

Background

The use of interleukin-2 (IL-2) has been a major asset to boost the therapeutic anti-tumor efficacy of adoptive cell therapy, including tumor infiltrating lymphocyte (TIL) therapy in the context of solid tumors. However, clinical assessments have revealed that its systemic administration results in poor accumulation at solid tumor sites. Additionally, the half-life of this recombinant molecule is short. High dose administration has therefore been used but this results in severe adverse events, including mortality. Hence, local production at the tumor is an attractive concept which might retain or even increase the useful aspects of IL-2, while reducing systemic side effects.

Methods

We aimed to evaluate the efficacy and safety of, intratumoral delivered IL-2-armed adenoviruses combined with T cell transfer in rodents. Experiments were set up using the syngeneic CB57BL/6 mouse B16.OVA melanoma tumor model infused with OVA-specific T cells, and the syngeneic Syrian hamsters Hapt1 pancreatic tumor model infused with TILs. Both therapeutic schedules involved once-a-week intratumoral administration of replication deficient serotype 5 (mice) or oncolytic serotype 5/3 chimeric (hamsters) IL-2-armed adenoviruses comparing with weekly-continuous systemic administration of recombinant IL-2.

Results

In both models, local production of IL-2 successfully replaced that need for systemic IL-2. In fact, efficacy was even higher than with systemic IL-2. Furthermore, the vectored delivery of IL-2 significantly potentiated the infiltration of CD8+ T cells and, significantly decreased the percentage of regulatory T cells. In animals that received systemic recombinant IL-2 therapy, significant histological changes were observed in the lungs, liver, heart, spleen and kidneys that should be considered as side-effects of the treatment.

Conclusions

In summary, local production of IL-2 seems appealing from the point of view of efficacy and safety in the context of adoptive cell therapy. This preclinical assessment provides the rational for clinical translation, which is ongoing by TILT Biotherapeutics Ltd.

P41 Successful expansion and characterization of tumor infiltrating lymphocytes (TILs) from non-melanoma tumors

Jyothi Sethuraman, Laurelis Santiago, Jie Qing Chen, Zhimin Dai, Seth Wardell, James Bender, Michael T Lotze

Lion Biotechnologies, Inc., Tampa, FL, USA
Correspondence: Jyothi Sethuraman (jyothi.sethuraman@lionbio.com)

Background

Adoptive cell therapy (ACT) has shown promise in comparison to other methods of cancer immunotherapy that rely on the active development of antitumor T cells in vivo to mediate cancer regression [1]. Administration of autologous TILs in melanoma patients has shown an overall response rate of 55% at NCI, 38% at Moffitt Cancer Center, 48% at MD Anderson Cancer Center and 40% in Sheba at the Ella Cancer Institute, Israel [1]. TILs have been found in a variety of solid tumors, and their presence has been shown to be a prognostic indicator of improved survival. Here, we demonstrate the feasibility of growing TILs and the potential to develop TIL therapies to treat other solid tumors, such as lung, breast, and bladder cancers.

Methods

Upon receiving surgically resected tumor specimens, samples were washed in HBSS and cut into small fragments prior to propagating in vitro in G-REX-10 cell culture flasks with interleukin-2 (commonly referred to as a pre-rapid expansion protocol [pre-REP]). After culture initiation, media was exchanged. The media was changed every 3 days subsequently for 2 weeks. TILs were harvested to assess cell count and viability, followed by immunophenotyping and cryopreservation.

Results

The average yield of TILs cultured and expanded from bladder, cervical, head & neck, lung, and triple-negative breast tumors is listed in Table 4. Phenotypic characterization of TILs from bladder, cervical and lung cancer were >60-70% CD8+ T cells whereas TILs from head and neck demonstrated variable distribution of CD8+ and CD4+ T cells. TILs propagated from TNBC were >80% CD4+ T cells. Regardless of the tumors, most cultures had < 20% CD56+ NK cells.
Table 4

(Abstract P41).

Tumor

Number of patient tumors

Average yield (and range) or TILs from pre-REP (106)

Bladder

2

290 (97-600)

Cervical

4

360 (147-800)

H&N

7

539 (132-738)

Lung

8

688 (50-915)

TNBC

13

429 (81-665)

Conclusions

We have been successful in culturing and expanding TILs from various non-melanoma solid tumors. We will initiate REP from pre-REP TILs from non-melanoma tumors to enable product development for subsequent possible clinical trials. Efforts are currently focused on culturing TILs from smaller tumor specimens/biopsies to assess utility in promoting expansion of TILs with central and effector memory phenotypes and selecting for mutanome reactive TILs.

References

1. Rosenberg SA, Restifo NP: Adoptive cell transfer as personalized immunotherapy for human cancer.Science 2015, 348:62-68.

P42 A tumor-penetrating recombinant protein anti-EGFR-iRGD enhances efficacy of antigen-specific CTL in gastric cancer in vivo

Huizi Sha, Shu Su, Naiqing Ding, Baorui Liu

The Comprehensive Cancer Center of Drum-Tower Hospital, Medical School of Nanjing University & Clinical Cancer Institute of Nanjing University, Nanjing, Jiangsu, People’s Republic of China
Correspondence: Huizi Sha (shahuizinju@126.com)

Background

Strategies that enhance the function of T cells are critical for immunotherapy. Targeted delivery of T cells through BiTE (bispecific T cell engager) platform to cancerous tissues shows potential in sparing unaffected tissues. However, it has been a major challenge for cells penetration in solid tumor tissues due to the complicated tumor microenvironment. Activated T cells expression integrin, which is the target of peptide RGD. Peptide iRGD (CRGDK/RGPD/EC) increased vascular and tissue permeability in a tumor-specific and neuropilin-1-dependent manner, allowing co-administered drugs to penetrate into extravascular tumor tissue. Recombinant protein anti-EGFR-iRGD was purified and examined.

Methods

Recombinant protein anti-EGFR-iRGD consisting of an anti-EGFR VHH (the variable domain from the heavy chain of the antibody) fused to iRGD, a tumor-specific binding peptide with high permeability were expressed in E. coli BL21 (DE3) and purified by nickel-nitrilotriacetic acid affinity chromatography. We use tumor cell lines and mice to analyze the targeting, penetrating and antitumor activity of antigen-secific T cells together with recombinant protein.

Results

We have successfully constructed a recombinant protein named anti-EGFR-iRGD, a dual targets of EGFR and integrin and high permeable protein. It could spread extensively throughout both the multicellular spheroids and the tumor mass. The recombinant protein anti-EGFR-iRGD could help more T cells infiltrating into tumor mass and also exhibited antitumor activity in tumor cell lines and mice.

Conclusions

Our results provide impetus for further studies for potentially using iRGD based fusion protein anti-EGFR-iRGD with immune therapy regimens for enhancing therapy of gastric cancer patients.
Fig. 32

(Abstract P42). Purification and verification of recombinant protein

Fig. 33

(Abstract P42). Penetration of CTL together with recombinant protein

P43 Immunodominance of cancer neoantigen and cancer-germline antigen T cell reactivities in successful immunotherapy of virally-induced epithelial cancer

Sanja Stevanovic1, Anna Pasetto2, Sarah R Helman1, Jared J Gartner2, Todd D Prickett2, Paul F Robbins2, Steven A Rosenberg2, Christian S Hinrichs1

1Experimental Transplantation and Immunology Branch, National Cancer Institute, National Institutes of Health, Bethesda, MD, USA; 2Surgery Branch, National Cancer Institute, National Institutes of Health, Bethesda, MD, USA
Correspondence: Sanja Stevanovic (sanja.stevanovic@nih.gov)

Background

Immunotherapy has clinical activity in human papillomavirus (HPV)-induced epithelial cancers, but the tumor antigens targeted by T cells resulting in cancer regression are poorly defined. The viral proteins expressed by these malignancies are generally considered the primary targets of T cell based immune attack. However, HPV+ cancers also harbor somatic mutations and express cancer-germline antigens that may be targets of tumor-specific T cells. Here, we aimed to elucidate the landscape of tumor antigens targeted by T cells in two patients with metastatic HPV+ cervical cancer who experienced durable complete tumor regression after adoptive transfer of tumor-infiltrating lymphocytes (TIL).

Methods

To this end, reactivity of therapeutic TIL was assessed in immunological assays against HPV-derived antigens (L1, L2, E1, E2, E4, E5, E6 and E7), mutated neoantigens and cancer-germline antigens identified by whole-exome and/or RNA sequencing of patient’s tumors. T cell receptor (TCR) clonotypes conferring specificity to tumor antigens were elucidated and quantified, and their in vivo persistence was profiled by TCR deep sequencing.

Results

T cell reactivity against the HPV-E6 and/or -E7 antigens was detected in both patient’s infused TIL, consistent with previously reported results. No T cell reactivity was detected against other six HPV-derived antigens. However, our data indicated that these patient’s infused TIL distinctly recognized mutated neoantigens or a cancer-germline antigen. Detailed TCR clonotype analysis showed that in one patient multiple CD8+ clonotypes (35% of infused TIL) recognized somatically mutated gene products (n=3) unique to patient’s tumor in addition to several CD8+ and/or CD4+ clonotypes (14% of TIL) targeting HPV-E6 and/or -E7. In the other patient, one CD8+ clonotype (67% of TIL) recognized the cancer-germline antigen Kita-kyushu lung cancer antigen 1 in addition to one CD4+ clonotype (14% of TIL) that targeted HPV-E7. Administered viral and non-viral tumor antigen-specific T cells in both patients remained functional and persisted at elevated levels in the circulation for months during ongoing remission.

Conclusions

Our data show that both patients who experienced complete tumor regression received TIL that contained low frequency of HPV-targeted T cells but a high frequency of mutated neoantigen- or cancer-germline antigen-targeted T cells. These results reveal a previously unappreciated role for T cells targeting non-viral antigens in HPV+ cervical cancer. By expanding the categories of potential tumor regression antigens for cervical cancer and possibly other HPV-induced malignancies, our findings provide new targets for personalized cancer vaccines and/or adoptive T cell therapies as well as for immune-monitoring of various cancer immunotherapies.

Consent

Written informed consent was obtained from patients.

P44 Adoptive cellular therapy (ACT) with allogeneic activated natural killer (aNK) cells in patients with advanced Merkel cell carcinoma (MCC): preliminary results of a phase II trial

Shailender Bhatia1, Melissa Burgess2, Hui Zhang3, Tien Lee4, Hans Klingemann4, Patrick Soon-Shiong4, Paul Nghiem1, John M Kirkwood5

1Fred Hutchinson Cancer Research Center, University of Washington, Seattle, WA, USA; 2University of Pittsburgh Cancer Institute, Pittsburgh, PA, USA; 3NantBioScience, Inc., Culver City, CA, USA; 4NantKwest, Inc., Culver City, CA, USA; 5UPMC Cancer Center, University of Pittsburgh Cancer Institute, Pittsburgh, PA, USA
Correspondence: Shailender Bhatia (sbhatia@uw.edu)

Background

MCC is an aggressive skin cancer associated with Merkel cell polyomavirus (MCPyV). Downregulation of class I major histocompatibility complex (MHC) expression in >80% of MCC tumors suggests potential susceptibility to adoptively transferred NK cells. aNK cells are derived from a NK cell line that is highly cytotoxic to a broad range of tumor cells. Phase I studies suggest that aNK cell therapy is well tolerated and has antitumor activity in patients with advanced hematologic or solid cancers. This study seeks to determine the efficacy of aNK cell therapy in patients with advanced MCC.

Methods

In this open-label phase II study, advanced MCC patients (planned N=24) receive aNK cells (2 × 109 cells/m2) intravenously on two consecutive days (1 cycle) every 2 weeks. Key eligibility criteria include age ≥18 years, unresectable stage III or IV MCC, and ECOG performance status ≤2. Up to two prior systemic chemotherapies and/or immunotherapies are allowed. The study uses a Simon optimal two-stage design. The primary efficacy endpoint is 4-month progression-free survival (PFS) rate. Secondary endpoints include objective response rate by RECIST 1.1, time to progression, overall survival, safety, and quality of life assessment (FACT-G). Planned correlative studies include genomic and proteomic tumor profiles, MCPyV status, and immunohistochemical assessment of MHC-1, correlated to treatment outcome.

Results

As of August 2016, 3 patients have been enrolled. Treatment-related adverse events have been grade 2 or milder with no serious adverse events. The efficacy criterion for the first stage of the study has been met, with a patient with advanced MCC refractory to multiple prior therapies including PD-1 blockade demonstrating an impressive partial response (PR) with >70% regression, ongoing at 20 weeks. Intriguing changes were observed clinically in another patient’s superficial tumors just a few hours after aNK infusions, although this patient developed progressive disease at 4 weeks. Correlative studies on tumor biopsies of the patient with PR are ongoing.

Conclusions

ACT with allogeneic aNK cells has been safe and well tolerated in the initial three patients with advanced MCC. Encouraging antitumor activity has been observed with an impressive PR in a patient with advanced MCC refractory to PD-1 blockade. The pre-specified efficacy criterion for the first stage of the trial has been met and enrollment continues on the trial. Updated results will be presented at the meeting.

Trial Registration

ClinicalTrials.gov identifier NCT02465957.

P45 Low dose conditioning chemotherapy and CD19-directed CAR T cells may elicit distinct immune programs associated with clinical responses

John M Rossi1, Marika Sherman1, Allen Xue1, Yueh-wei Shen1, Lynn Navale1, Steven A Rosenberg2, James N Kochenderfer3, Adrian Bot1

1Kite Pharma, Inc, Santa Monica, CA, USA; 2Surgery Branch, National Cancer Institute, Bethesda, MD, USA; 3Experimental Transplantation and Immunology Branch, National Cancer Institute, Bethesda, MD, USA
Correspondence: Adrian Bot (abot@kitepharma.com)

Background

Anti-CD19 CAR T cell therapy has shown promising clinical efficacy. Recent evidence points to a critical role for non-myeloablative conditioning chemotherapy, influencing the expansion, persistence and activity of CAR T cells. To diminish toxicities, Kochenderfer, et al. pioneered a conditioning chemotherapy regimen with low dose cyclophosphamide (300-500mg/m2) and fludarabine (30mg/m2) administered daily for 3 days [1]. This resulted in a response rate of 73% in patients with advanced Non-Hodgkin lymphoma, with lower rate of hematologic toxicities. Post CAR T cell-treatment peak levels of cytokines in blood were associated with T cell expansion, clinical efficacy or neurotoxicity [1].

Methods

We analyzed 41 blood biomarkers in 22 patients treated with anti-CD19 CAR T cells, preceded by low dose conditioning chemotherapy. We also measured cytokine levels produced by CAR T cells ex vivo, upon culture with CD19-expressing target cells. The expansion of CAR T cells in blood was measured by quantitative PCR.

Results

Conditioning chemotherapy enhanced IL-15 and decreased lymphocytes and perforin blood levels. Several cytokines peaked sequentially in blood post CAR T cell infusion: among those, IL-15 and GM-CSF at days 2-3, followed by IL-10 and Granzyme B, at day 6. CAR T cell expansion in blood occurred within 7-14 days. IL-15 blood levels associated with T cell expansion, clinical response and toxicities. In addition, early post-CAR T cell treatment levels GM-CSF and peak blood levels of IL-10 and Granzyme B, were associated with clinical efficacy or neurotoxicity. When stimulated ex vivo with CD19-expressing cells, CAR T cells produced a broad range of molecules including GM-CSF, IL-10 and Granzyme B, but not IL-15.

Conclusions

In conclusion, these preliminary findings suggest that three immune programs impact clinical outcome of CAR T cell treatment: a T cell proliferative program initiated by conditioning chemotherapy, together with an inflammatory and a cytotoxic program deployed by CAR T cells. In addition, this analysis highlights the need to carefully optimize the conditioning chemotherapy regimen.

Acknowledgements

This study was conducted under a CRADA between NCI and Kite Pharma.

Trial Registration

ClinicalTrial.gov identifier NCT00924326.

References

1. Kochenderfer J, Somerville R, Lu T, Shi V, Yang JC, Sherry R, et al: Anti-CD19 chimeric antigen receptor T cells preceded by low-dose chemotherapy to induce remissions of advanced lymphoma [abstract]. J Clin Oncol 2016, 34 Suppl:LBA3010.

P46 Artificial antigen presenting cells promote expansion of tumor infiltrating lymphocytes (TILs)

Anandaraman Veerapathran, Aishwarya Gokuldass, Amanda Stramer, Jyothi Sethuraman, Michelle A Blaskovich, Doris Wiener, Ian Frank, Laurelis Santiago, Brian Rabinovich, Maria Fardis, James Bender, Michael T Lotze

Lion Biotechnologies, Inc., Tampa, FL, USA
Correspondence: Anandaraman Veerapathran (anand.veerapathran@lionbio.com)

Background

For more than a decade, allogeneic peripheral blood mononuclear cells (PBMC) have been used as accessory feeder cells that provide “costimulatory signals” necessary for the expansion of tumor-infiltrating lymphocytes (TILs) in the presence of IL-2 and CD3 stimulation (Rapid Expansion Protocol [REP]) [1]. The intrinsic heterogeneity of allogeneic PBMC is an important variable when considering the expansion and resulting phenotype of post-REP TIL prepared for transplantation. The procurement of allo-PBMC in large numbers is also challenging and expensive. Our objective was to evaluate artificial antigen presenting cells (aAPC) as a potential substitute for PBMC. As such, we developed a novel aAPC from the CD64+ MOLM-14 human leukemia cell line, genetically engineered to express recombinant CD86 (B7-2) and CD137-L (41BBL) (MOLM14-86/137).

Methods

The MOLM-14-86/137 cell line was generated via transduction of wild type MOLM-14 with lentiviral virions encoding genomic RNA of CD86 or CD137 downstream of the U3 promoter from MSCV. MOLM-14-86/137 was ɣ-irradiated at 100Gy and co-cultured with TILs at a 1:100 ratio (TIL:aAPC) in media containing OKT3 (30 ng/ml) and IL-2 (3000 IU/ml) for 14 days. On Day 14, we calculated their expansion and examined their differentiation (flow cytometry), metabolic rate, and cytotoxicity.

Results

Compared to TIL co-cultured with PBMC or wild type MOLM-14, we obtained 95-100% TILs via co-culture with MOLM-14-86/137. This value was within the expected range using PBMC (Fig. 34). TIL differentiation, cellular respiration (OXPHOS) and redirected cytotoxicity were also within the range expected via co-culture with PBMC.
Fig. 34

(Abstract P46). Rapid expansion of TILs using irradiated engineered MOLM-14 or PBMC feeders. TIL were co-cultured with PBMC Feeders or MOLM14 (CD86/41BBL) at 1:100 ratios plus OKT3(30ng/ml) and IL-2 (3000 IU/ml). Cells were counted and split on Day 6 and 11. Each dot represents cell numbers determined on Day 14

Conclusions

According to our measurements, co-culture of TIL with MOLM-14-86/137 aAPC resulted in expansion, metabolic activity and cytotoxicity that were sufficiently similar to that obtained with PBMC. This data suggests that investigation of a MOLM-14-86/137 based REP protocol in a clinical setting is warranted.

References

1. Dudley ME, Wunderlich JR, Shelton TE, Even J, Rosenberg SA: Generation of tumor-infiltrating lymphocyte cultures for use in adoptive transfer therapy for melanoma patients. J Immunother 2003, 26:332–342.

P47 Blocking vasoactive intestinal peptide signaling modulates immune checkpoints and graft-versus-leukemia in allogeneic transplantation in mice

Edmund K Waller1, Jian-Ming Li1, Christopher Petersen1, Bruce R Blazar2, Jingxia Li1, Cynthia R Giver1

1Emory University, Atlanta, GA, USA; 2University of Minnesota, Minneapolis, MN, USA
Correspondence: Edmund K Waller (ewaller@emory.edu)

Background

The goal of allogeneic bone marrow transplantation (allo-BMT) is the elimination of leukemia cells through the graft-versus-leukemia (GvL) activity of donor cells, while limiting graft-versus-host disease (GvHD). Immune checkpoint pathways regulate GvL and GvHD activities, but blocking these pathways can cause lethal GVHD. Vasoactive intestinal peptide (VIP) is an immunosuppressive neuropeptide that regulates co-inhibitory pathways.

Methods

Murine models of MHC-mismatched allogeneic bone marrow transplantation were used to evaluate the effect of blocking VIP-signaling on the graft-versus-leukemia (GvL) and graft-versus host disease (GvHD) effect of donor T cells. Mice were transplanted with donor grafts from VIP-KO mice or recipients of wild-type grafts were treated with seven daily injections of a peptide antagonist to VIP (VIPhyb). Survival, GvHD and the growth of luciferase+ LBRM, a T cell lymphoblastic leukemia cell line, or C1498, a myeloid leukemia cell line, were monitored by bio-luminescent imaging. Expression of chemokine receptors, cytokines and check-point molecules were measured by flow cytometry. VIP expression on donor and host cells was visualized using a transgenic mouse in which GFP expression is driven by the VIP promoter. T cell repertoire from T cells in mice with GvHD or GvL was analyzed by deep sequencing.

Results

VIP is expressed transiently in donor NK, NKT, dendritic cells, and T cells after allo-transplant, as well as in host leukocytes. A peptide antagonist of VIP signaling (VIPhyb) increased T cell proliferation in vitro and reduced IL-10 expression in donor T cells. Treatment of allo-BMT recipients with VIPhyb, or transplanting donor grafts lacking VIP (VIP-KO), activated donor T cells in lymphoid organs, reduced T cell homing to GvHD target organs, and enhanced GvL without increasing GvHD in multiple allo-BMT models. Genetic or ex vivo depletion of donor NK cells or CD8+ T cells from allografts abrogated the VIPhyb-enhanced GvL activity (Fig. 35A). VIPhyb treatment led to downregulation of PD-1 and PD-L1 expression on donor immune cells (Fig. 35B), increased effector molecule expression, and expanded oligoclonal CD8+ T cells that protected secondary allo-transplant recipients from leukemia (Fig. 35C & D).
Fig. 35

(Abstract P47). Treatment of allo-BMT recipients with a VIP antagonist induces a CD8+ donor T cell-dependent GvL response associated with down-regulation of PD-L1 on donor pDC and expansion of oligoclonal donor CD8+ T cells. A: Survival curves and GVHD clinical scores for B10.BR-->B6 allo-BMT harboring C1498 leukemia cells. Donor T cell subsets were depleted of specific populations by MACS prior to transplantation as shown. B: Expression of PD-L1 on donor pDC in B6-->B10.BR transplants. C: TCR Vbeta and J gene segments present in donorCD8+ T cells (left), CD8+ T cells from mice with GVHD (middle) and Cd8+ T cells from mice with GvL (right). D: Lack of correspondence between TCR Vbeta and J genes sequenced in mice with GvHD and mice with GVL

Conclusions

VIP production by donor immune cells is dynamically regulated after allo-BMT, and transplanting VIP-KO cells, or daily treatment with VIPhyb, significantly enhanced survival of leukemia-bearing transplant recipients via a CD8+ T cell dependent GvL effect without increased GvHD in murine models of MHC mis-matched allo-BMT. Blocking VIP-signaling thus represents a novel pharmacological approach to separate GvL from GvHD and enhance adaptive T cell responses to leukemia-associated antigens in allo-BMT.

P48 Bortezomib sensitizes cancer stem cells from solid human tumors to natural killer cell-mediated killing

Ziming Wang1, Steven K Grossenbacher1, Ian Sturgill2, Robert J Canter1, William J Murphy1

1University of California, Davis, Sacramento, CA, USA; 2California State University, Sacramento, Sacramento, CA, USA
Correspondence: Ziming Wang (zswang@ucdavis.edu)

Background

Cancer stem cells (CSCs) from solid and hematopoietic tumors resist conventional cytotoxic therapies that target rapidly proliferating cells. Thus, residual CSCs can hide within the tumor niche and seed relapse and metastasis. Due to their relapse potential there is an urgent need to identify ways to therapeutically target CSCs. We previously found that cells expressing high amounts of the stem cell associated protein aldehyde dehydrogenase (ALDH) are effectively killed by activated natural killer (NK) lymphocytes. NK cells are known to kill malignant cells though apoptotic processes inherent to the target cell, such as TRAIL-DR5 or Fas-FasL binding, without prior immunization. We and others have also found that the FDA approved proteasome inhibitor, bortezomib, sensitizes tumor cells to NK cell killing by upregulating DR5 and intracellular machinery associated with apoptosis. Based on this previous work, we investigated the effects of bortezomib to promote NK cell killing of ALDHbright CSCs. We evaluated CSCs derived from solid tumors, in vitro and in vivo, for the induction of receptors associated with NK cell mediated killing, and for their susceptibility to NK killing after treatment.

Methods

In vitro sensitization and cytotoxicity assays were performed using cultured NK cells isolated from human peripheral blood. The glioblastoma and sarcoma cell lines, U87 and A673, respectively, were first treated with bortezomib, then co-cultured with activated NK cells at serial effector:target ratios. Target tumor cells were analyzed using flow cytometry for cell survival, and expression of Fas, DR4, DR5, and MICA/B on both ALDHbright and ALDHdim cells.

Results

Bortezomib led to a 3-fold increase in the percentage of viable ALDHbright glioblastoma and sarcoma cells, in vitro, compared to untreated controls. Bortezomib enhanced the expression of Fas and DR5 by 10% and 40%, respectively, in ALDHbright U87 cells. It increased the expression of DR4 by 20% in ALDHbright A673 cells. However, bortezomib had little effect on ALDHdim cells. Bortezomib pretreatment lead to a 98% decrease in viable ALDHbright cells following NK cytotoxicity assays in vitro. In vivo, bortezomib improved the efficacy of adoptive NK cell therapy in multiple mouse xenograft models.

Conclusions

ALDHbright CSCs are resistant to the cytotoxic effects of bortezomib. Bortezomib resistance is marked by increases in the expression of Fas, DR4, and DR5 and leads to increased susceptibility to lysis by activated NK cells. The combined use of bortezomib with activated natural killer cells could act as a potential anti-CSC strategy to improve outcomes for patients with solid tumors.

P49 Targeted NK cells display potent activity against glioblastoma and induce protective antitumor immunity

Congcong Zhang1, Michael C Burger2, Lukas Jennewein3, Anja Waldmann1, Michel Mittelbronn3, Torsten Tonn4, Joachim P Steinbach2, Winfried S Wels1

1Georg-Speyer-Haus, Institute for Tumor Biology and Experimental Therapy, Frankfurt, Germany; 2Institute for Neurooncology, Goethe University, Frankfurt, Germany; 3Edinger Institute, Goethe University, Frankfurt, Frankfurt, Germany; 4German Red Cross Blood Donation Service North-East, Dresden, Dresden, Germany
Correspondence: Winfried S Wels (wels@gsh.uni-frankfurt.de)

Background

Significant progress has been made over the last decade towards realizing the potential of natural killer (NK) cells for cancer immunotherapy. In addition to donor-derived primary NK cells, also continuously expanding cytotoxic cell lines such as NK-92 are being considered for adoptive cancer immunotherapy. High cytotoxicity of NK-92 has previously been shown against malignant cells of hematologic origin in preclinical studies, and general safety of infusion of NK-92 cells has been established in phase I clinical trials. To enhance their therapeutic utility, here we genetically modified NK-92 cells to express a chimeric antigen receptor (CAR), consisting of an ErbB2 (HER2)-specific scFv antibody fragment fused via a linker to a composite CD28-CD3ζ signaling domain. GMP-compliant protocols for vector production, lentiviral transduction and expansion of a genetically modified NK-92 single cell clone (NK-92/5.28.z) were established.

Methods

Functional analysis of NK-92/5.28.z cells revealed high and stable CAR expression, and selective cytotoxicity against ErbB2-expressing but otherwise NK-resistant tumor cells of different origins in vitro. Ongoing work focuses on the development of these cells for adoptive immunotherapy of ErbB2-positive glioblastoma (GBM). ErbB2 expression in primary tumors and cell cultures was assessed by immunohistochemistry and flow cytometry. Cell killing activity of NK-92/5.28.z cells was analyzed in in vitro cytotoxicity assays. In vivo antitumor activity was evaluated in NOD-SCID IL2Rγnull (NSG) mice carrying orthotopic human GBM xenografts and C57BL/6 mice carrying orthotopic ErbB2-expressing murine GBM tumors.

Results

We found elevated ErbB2 protein expression in >40% of primary GBM samples and in the majority of GBM cell lines investigated. In in vitro assays, NK-92/5.28.z in contrast to untargeted NK-92 cells lysed all ErbB2-positive established and primary GBM cells analyzed. Potent in vivo antitumor activity of NK-92/5.28.z was observed in orthotopic GBM xenograft models in NSG mice, leading to a marked extension of symptom-free survival upon repeated stereotactic injection of CAR NK cells into the tumor area. In immunocompetent mice, local therapy with NK-92/5.28.z cells resulted in cures of transplanted syngeneic GBM in the majority of animals, induction of endogenous antitumor immunity and long-term protection against tumor rechallenge at distant sites.

Conclusions

Our data demonstrate the potential of ErbB2-specific NK-92/5.28.z cells for adoptive immunotherapy of glioblastoma, justifying evaluation of this approach for the treatment of ErbB2-positive GBM in clinical studies.

P50 Shared T cell receptor sequences between HLA-A2+ patients vaccinated against a Melan-A epitope correlate with clinical benefit

Jason B Williams1, Yuanyuan Zha1, Thomas F Gajewski2

1University of Chicago, Chicago, IL, USA; 2University of Chicago Medical Center, Chicago, IL, USA
Correspondence: Jason B Williams (jaybwilliams1@uchicago.edu)

Background

Adoptive T Cell Therapy (ACT) of in vitro expanded T cell clones or transduced T cells redirected against defined tumor antigens has shown therapeutic efficacy in some patients. Ideas to improve upon this therapy are multifaceted, including combining ACT with checkpoint blockade, increasing the number of defined T cell receptors (TCR) against defined antigens, identifying new tumor-specific somatic mutations to target, and engineering TCRs to have increased avidity. However, even for well characterized antigens such as Melan-A, the optimal TCR is not known. Some engineered TCRs have shown off-tumor toxicity, and so selecting TCRs with maximal therapeutic efficacy but at the same time giving minimal side effects remains an important goal.

Methods

We reasoned that one strategy for selecting optimal TCRs might be to identify T cells expanded after active immunization against defined epitopes in patients who experienced clinical benefit but no apparent side effects. To this end, we performed deep TCR sequencing of HLA-2/Melan-A+ CD8+ T cells from 16 metastatic melanoma patients vaccinated against a Melan-A epitope.

Results

While changes in overall TCR clonality measured before and after vaccination did not correlate with clinical benefit, many TCRs showed a significant increase in representation of the total TCR repertoire after vaccination. Of the 6 patients that received a clinical benefit we found 122 public TCRβ and 124 public TCRα sequences. 105 of these sequences showed expansion after vaccination in 2 or more patients. Surprisingly, we did not observe the defined Melan-A-specific TCRs used previously in redirected ACT clinical trials, designated DMF4 and DMF5. Mapping of public sequences by frequency per patient and aligning TCRα/TCRβ sequences highlighted several potential TCRα/TCRβ pairings. One patient was of particular interest as he had participated in two vaccine trials, with a 32-month interim between trials and clinical benefit each time. By the end of the second treatment period, the patient’s TCR repertoire contained 55 public sequences. Interestingly, 7 of these sequences showed an initial contraction at the end of the first trial followed by a significant expansion by the end of the second trial, suggesting a strong clonotypic response to Melan-A.

Conclusions

Together, these data highlight multiple TCRα and TCRβ sequences correlating with clinical benefit in the setting of no treatment-related toxicities. Similar results have been observed in other trials utilizing CEA peptide or WT1 peptide immunization. These sequences should enable full-length cloning of TCRs to be used in redirect adoptive cell therapy.

Trial Registration

ClinicalTrials.gov identifier NCT00515528.

P51 T cells redirected to TEM8 have antitumor activity but induce ‘on target/off cancer toxicity’ in preclinical models

LaTerrica C. Williams1, Giedre Krenciute1, Mamta Kalra1, Chrystal Louis1, Stephen Gottschalk2

1Baylor College of Medicine, Houston, TX, USA; 2Center for Cell and Gene Therapy, Baylor College of Medicine, Houston, TX, USA
Correspondence: LaTerrica C. Williams (laterriw@bcm.edu)

Background

Targeting the tumor vasculature holds promise to improve the outcome for patients with refractory solid tumors. Tumor endothelial marker (TEM) 8 is an attractive target for T cell therapies since it is expressed at higher levels in malignant cells and the tumor endothelium as judged by studies using monoclonal antibodies (mAbs). However T cells do not require high expression of the targeted antigen for activation, because of the higher overall avidity of a multivalent T cell compared to bivalent mAbs. Thus, the aim of this project was to determine the safety and antitumor activity of T cells expressing TEM8/CD3-specifc T cell engagers (TEM8-ENG).

Methods

qPCR and FACS analysis was used to determine the expression of TEM8 in solid tumor and endothelial cells. TEM8-ENG T cells were generated by transducing T cells with a retroviral vector encoding a TEM8-ENG consisting of the TEM8-specific scFv L2 linked to a scFv recognizing CD3. TEM8-ENG T cell effector function was evaluated in vitro and in vivo. Appropriate controls were used including ENG T cells specific for an irrelevant antigen (CD19).

Results

To confirm the specificity of TEM8-ENG T cells we used targets that did not express TEM8 (BV173) or BV173 cells that were genetically modified to express human TEM8, murine TEM8, or murine TEM1. TEM8-ENG T cells recognized TEM8pos targets (BV173.hTEM8, BV173.mTEM8) as judged by their ability to secrete IFNγ in coculture assays and kill both targets in cytotoxicity assays; in contrast, TEM8neg cells (BV173, BV173.mTEM1) were not recognized. Specificity of TEM8-ENG T cells was further confirmed with TEM8pos U373 glioma cells and U373.k/oTEM8 cells. TEM8-ENG T cells recognized a panel of TEM8pos solid tumor cells (A431, A549, LM7, LAN1, U87), and primary endothelial cells (HHSEC, HPAEC) in contrast to TEM8neg tumor cells (KG1a, Daudi). In vivo, intratumoral administration of TEM8-ENG T cells induced regression of U373 gliomas in an orthoptic xenograft model. Intravenous administration of 1x107 TEM8-ENG T cells resulted in antigen-dependent expansion and death of 7/10 mice; no toxicity was observed after the injection of 1x106 TEM8-ENG T cells.

Conclusions

TEM8 is expressed in tumor endothelium, normal endothelial cells and solid tumor cells as judged by qPCR, FACS, and functional assays. TEM8-ENG T cells had antitumor activity in vivo, but displayed dose-dependent toxicity. Our studies highlight that mAbs and T cells may have different toxicity profiles, most likely due to differences in their avidity for the targeted antigen. TEM8-ENG T cell xenograft models represent an ideal model to study genetic approaches to prevent ‘on target/off cancer toxicities’ of cell therapies.

P52 A pathogen boosted adoptive cell transfer immunotherapy to treat solid tumors

Gang Xin1, David Schauder1, Aimin Jiang2, Nikhil Joshi3, Weiguo Cui1

1Blood Center of Wisconsin, Milwaukee, WI, USA; 2Department of Immunology, Roswell Park Cancer Institute, Buffalo, NY, USA; 3Koch Institute for Integrative Cancer Research and Department of Biology, Massachusetts Institute of Technology, Cambridge, MA, USA
Correspondence: Gang Xin (Gang.Xin@BCW.edu)

Background

Despite the remarkable success in treating hematological malignancies, adoptive cell transfer (ACT) still faces several challenges in treating solid tumors. The main stumbling blocks include insufficient quantity of tumor-specific T cells for transfer, impaired migration of transferred T cells into the tumor and the immunosuppressive microenvironment within the tumor.

Methods

To overcome these problems, we designed an innovative approach that not only overcomes immunosuppression, but also induces robust anti-tumor T cell responses in the tumor. We first genetically engineered dual-specific CD8 T cells that can recognize both a tumor associated antigen and a bacterial antigen in vitro. Then, we treated tumor-bearing mice with ACT using a small number of the dual-specific CD8 T cells. This was accompanied by intratumoral injection of a low dose of the bacteria, which was sufficient to break local immunosuppression.

Results

The dual-specific CD8 T cells expanded robustly and migrated to the tumor bed in response to the infection. At the same time, the second TCR of these effector CD8 T cells recognized tumor antigen and executed effector function, causing tumor regression. As a result of this enhanced anti-tumor effect, 60% of the treated mice successfully eradicated their solid tumor at the primary site.

Conclusions

Our approach not only overcomes immunosuppression, but also recruits robust anti-tumor T cell responses to the tumor. Overall, our study harnesses the power of multiple arms of the immune system with promising translational value, which can be used to target many types of solid tumors.

P53 Pharmacologic rejuvenation of exhausted T cells to improve adoptive TIL therapy

Xue Zeng1, Ashley V Menk1, Nicole Scharping2, Greg M Delgoffe2

1University of Pittsburgh Cancer Institute, Pittsburgh, PA, USA; 2University of Pittsburgh, Pittsburgh, PA, USA
Correspondence: Xue Zeng (xuz27@pitt.edu)

Background

Immunotherapy has emerged as a strategy for the treatment of cancer. One of these immunotherapies is adoptive tumor-infiltrating lymphocyte (TIL) therapy, in which T cells from resected tumors are expanded in vitro and then given to patients. However adoptive TIL therapy has little efficacy for many patients, because the tumor microenvironment creates an extreme environment for T cells. Our lab has revealed that T cells display metabolic defects, especially a loss of mitochondria, when they infiltrate the tumor microenvironment. This loss is related to T cell exhaustion. We hypothesize that these exhausted T cells were the most functional cells as they responded to tumor earliest and strongest. However their loss of mitochondria prevents them from further expansion when removed and cultured in vitro. Thus, we are utilizing what we have identified about their metabolic dysfunction to rejuvenate those T cells during ex vivo expansion. Our goal is to make exhausted T cells more metabolically active and provide a potent method for TIL therapy.

Methods

Tumor injection: mice were given either 250,000 B16 or MC38 tumor cells injected intradermally in the center of the back. T cell activation: TILs are activated with 3 ug/ml anti-CD3 (plate bound), 2 ug/ml anti-CD28, 50 units/ml IL-2. Adoptive TIL transfer: treated and non-treated TIL are given to the mice that bear tumors by intravascular injection.

Results

PD-1hi cells remain mitochondrially deficient and fail to proliferate ex vivo. Rosiglitazone can rescue mitochondrial mass and proliferation. PD-1 Tim-3hi cells are over proliferated by PD-1 Tim-3lo cells. Preliminary data has shown that glitazone compounds to long-term expansion protocols prevents loss of the previously-exhausted T cells during expansion.

Conclusions

Cells expressing high levels of PD-1 and Tim-3 have low mitochondrial mass and fail to proliferate effectively in vitro. Mixing congenically marked cells from the non-exhausted or exhausted compartment shows exhausted cells are quickly overtaken by the non-exhausted (less than 1 week). Adding glitazone compounds to stimulate mitochondrial biogenesis results in short-term improvement of T cell proliferation in vitro. Preliminary data has shown that glitazone compounds to long-term expansion protocols prevents loss of the previously-exhausted T cells during expansion.

Acknowledgements

UPCI Cytometry Core and Animal Facility (supported by NCI P30CA047904), University of Pittsburgh Department of Immunology, University of Pittsburgh Cancer Institute, Tumor Microenvironment Center, Chinese Scholar Council.

References

1. Delgoffe GM, Powell JD: Feeding an army: The metabolism of T cells in activation, anergy, and exhaustion.Mol Immunol 2015, 68(2 Pt C):492–496.

P54 A turbocharged chimeric antigen receptor against prostate cancer

Zeguo Zhao1, Mohamad Hamieh1, Justin Eyquem1, Gertrude Gunset1, Neil Bander2, Michel Sadelain1

1Center for Cell Engineering, Memorial Sloan Kettering Cancer Center, New York, NY, USA; 2Department of Urology, Weill Cornell Medical College, New York, NY, USA
Correspondence: Zeguo Zhao (zhaoz@mskcc.org)

Background

Both CD28- and 4-1BB-based second-generation CAR T cells elicit dramatic clinical responses in patients with refractory/relapsed CD19 positive malignancies, especially patients with acute lymphoblastic leukemia. We recently demonstrated that co-expressing the second-generation 19-28z CAR with 4-1BBL yields balanced tumoricidal function and T cell persistence, resulting in the greater therapeutic efficacy (Turbocharged CAR). However, due in part to their tumor microenvironment, solid tumors often resist CAR T cell therapy. We hypothesized that CD28-based second-generation CAR T cells coexpressing 4-1BBL would have better therapeutic efficacy against solid tumors than current second-generation CARs, owning to their unique intrinsic and immunomodulatory qualities.

Methods

Prostate-specific membrane antigen (PSMA) is a dimeric type II integral membrane glycoprotein, which is overexpressed in castrate-resistant, metastatic prostate cancer. We constructed a tricistronic PSMA-targeted CAR vector encoding the Pd28z CAR, 4-1BBL and a truncated, nonfunctional EGFR as a safety control (Pd28z–4-1BBL–EGFRt). Two second-generation CARs (Pd28z and PdBBz) served as controls.

Results

In a high tumor burden model of disseminated prostate cancer, we used the in vivo “stress test” in which the T cell dose is gradually lowered lowered to levels where CAR therapy begins to fail, in order to compare the relative functionality and persistence of these CAR T cells. CAR T cells coexpressing Pd28z with 4-1BBL exhibited higher tumor eradication and T cell persistence in NSG mice bearing diffuse metastatic prostate cancer, compared to both second-generation CARs Pd28z and PdBBz.

Conclusions

4-1BBL Turbocharged CAR T cells thus seem to possess striking therapeutic potential against solid tumors.

Biomarkers and Immune Monitoring

P55 Map of targets on dendritic cells (DC) in human tonsils and lymph nodes potentially facilitating antigen cross-presentation

David Askmyr1, Milad Abolhalaj2, Kristina Lundberg2, Lennart Greiff3, Malin Lindstedt2

1ENT Departement, Lund University Hospital, Lund, Skane Lan, Sweden; 2Department of Immunotechnology, Lund University, Lund, Skane Lan, Sweden; 3ENT Department, Skåne University Hospital, Lund, Lund, Skane Lan, Sweden
Correspondence: Milad Abolhalaj (milad.abolhalaj@immun.lth.se)

Background

Dendritic cells (DCs) orchestrate adaptive and innate immune responses and are therefore key targets for immunotherapy (IT). Of special interest for IT directed against cancer are DCs with a high capacity of antigen cross-presentation, potentially resulting in cell-mediated tumor antigen-specific effects. This study reports an update on the map of DC subtypes and phenotypical aspects that may be reached by adjuvant measures in human tonsils and lymph nodes, both potential sites for “vaccine” deposition.

Methods

From biopsies of tonsils (n=23) and neck lymph nodes (n=16), single cell suspensions were prepared by enzymatic digestion and DCs were identified by an 8-color flow cytometry Ab panel. DC subsets frequencies (CD1c+, CD123+, CD141+, CD1c- CD141-), maturity status, and surface receptor profiles, focusing on a variety of C-type lectin receptors. Toll-like receptor 2 (TLR2) and the chemokine receptor XCR1, were then described and investigated at protein level via flow cytometry. The results were then analyzed through two-group comparison tests.

Results

DCs with similar myeloid CD11c+/plasmacytoid CD123+ ratios and largely similar frequencies among CD45+ leukocytes were observed in tonsils as well as lymph nodes. However, among the DC subsets studied, CD141+ DCs showed a higher frequency in tonsils compared to lymph nodes. No maturity differences were found among the DC subsets in tonsils and lymph nodes based on expression of CD80 and CD86. DC subsets expressing XCR1, TLR2, CLECSF14, CD206, DEC205, and Dectin-1 were observed with similar frequencies in tonsils (n < 18) and lymph nodes (n < 11).

Conclusions

DCs in tonsils and lymph nodes largely share similar features in terms of frequency, maturation, and PRR expression. However, a higher frequency of CD141+ cells in tonsils may be of interest considering this subset’s capability in antigen cross-presentation. Our work suggests tonsils as well as lymph nodes as vaccine deposition sites in DC-mediated IT. Furthermore, specific adjuvant measures directed at C-type lectin receptors, TLR2, and XCR1 may be employed to achieve cross-presentation of antigen and cell-mediated tumor antigen-specific effects.

P56 PD-L1 and immune infiltrates are prognostic and differentially expressed in distinct subtypes of gastric cancer

Helen K Angell1, Kyoung-Mee Kim2, Seung-Tae Kim2, Sung Kim3, Alan D Sharpe1, Julia Ogden4, Anna Davenport5, Darren R Hodgson4, Carl Barrett6, Jeeyun Lee3, Elaine Kilgour4

1AstraZeneca, Cambridge, England, UK; 2Department of Pathology & Translational Genomics, Samsung Medical Center, Sungkyunkwan University School of Medicine, Seoul, South Korea; 3Division of Hematology-Oncology, Department of Medicine, Samsung Medical Center, Sungkyunkwan University School of Medicine, Seoul, South Korea; 4AstraZeneca, Macclesfield, England, UK; 5University Hospital of South Manchester, Manchester, England, UK; 6AstraZeneca, Waltham, MA, USA
Correspondence: Helen K Angell (helen.angell@astrazeneca.com)

Background

Gastric cancer (GC) is often diagnosed at an advanced stage, for which therapeutic options are largely limited to cytotoxic chemotherapy and five-year survival is less than 20%. Immune checkpoint blockade with anti-programmed cell death-1 (PD-1) or anti-programmed cell death ligand-1 (PD-L1) antibodies is emerging as a promising therapeutic approach for several cancer types. An important question is whether the clinical efficacy of PD-1/PD-L1 checkpoint blockade can be improved through combination with targeted agents, such as trastuzumab, for use in human epidermal growth factor receptor 2 (HER2)-positive disease and olaparib, a poly (ADP-ribose) polymerase (PARP) inhibitor. This study determines the association of PD-L1 expression and immune cell infiltrates with clinical outcome and investigates the overlap of these with microsatellite instability (MSI)-high, ATM low and HER2 high segments.

Methods

PD-L1 membrane expression on tumour cells (TC) and infiltrating immune cells (IC), CD3+ T lymphocytes, CD8+ cytotoxic T cells, ATM and HER2 were assessed by immunohistochemistry (IHC) in a cohort of 380 Korean gastric cancer patients. PD-L1 positivity was assessed by a pathologist (positive < 0%). CD3 and CD8 were quantified by HALO® image analysis (cells/mm2). EBV status was determined using in situ hybridization and MSI status was performed using PCR and MLH1 IHC.

Results

The ATM-low and HER2-high segments are mutually exclusive and differ markedly in their immune profile; the ATM-low segment being enriched for MSI (p < 0.01), PD-L1 TC positivity (p < 0.01) and CD8+ cytotoxic immune infiltrates (p=0.033), while the HER2 segment is enriched for MSS, with no enrichment for immune markers. The PD-L1 segment is associated with increased T cell infiltrates: CD3 (p < 0.01) and CD8 (p < 0.01), while the MSI-high segment is enriched for PD-L1 TC (p < 0.01), PD-L1 IC (p < 0.001), CD3 (p < 0.05) and CD8 (p < 0.01), and has significant overlap with the ATM-low but not HER2 segments. Multivariate analysis confirmed PD-L1 TC positivity (p < 0.01), high CD3 (overall survival [OS] P < 0.01; disease-free survival [DFS] p=0.021) and high CD8 (OS p < 0.01; DFS p=0.027) as independent prognostic factors for both DFS and OS. Patients with MSI-high tumours had better overall survival by both univariate (p < 0.01) and multivariate (p < 0.05) analysis.

Conclusions

Here we present evidence for segmentation of gastric cancers into four distinct molecular segments, namely ATM-low, HER2-high, PD-L1 positive and MSI-high. This illustrates the potential for subsets of GC patients to respond differently to immune therapy and the opportunity to employ different strategies for maximising the benefit from immune therapies in these segments.

P57 Four color T and B cell ELISPOT assays for simultaneous detection of analytes

Jodi Hanson, Richard Caspell, Alexey Karulin, Paul Lehmann

Cellular Technology Ltd, Shaker Hts, OH, USA
Correspondence: Jodi Hanson (jodi.hanson@immunospot.com)

Background

ELISPOT assays are a key research tool for enumerating antigen-specific T and B cells in PBMC. As both T and B cells occur in major classes, immune monitoring has to be concerned with identifying these as well. So far, ELISPOT assays have been primarily done single or double color. In this report, we demonstrate the development of four color T and B cell ELISPOT assays.

Methods

PBMC were cultured for 4 days with a peptide pool of CMV-, EBV- and Flu- viruses for T cell assays or polyclonal B cell activators for B cell assays. On day 4, cells were washed, counted and plated in a low autofluorescence PVDF plate. Plates were precoated with capture antibodies for detection of IFN-g, IL-2, GzB, or TNF-α (T cell assays) or Ig secretion (B cell assays). During a 4h culture, the cytokine or antibodies secreted by the individual T or B cells respectively was captured on the membrane. The plate-bound “spots” were visualized using cytokine-specific or IgG subclass- or Ig class-specific detection reagents, whereby each detection reagent is distinguished from the other 3 reagents through its unique fluorescence. The four-color assays were analyzed using an ImmunoSpot® S6 Analyzer.

Results

We show that the four color assay has the same sensitivity for detecting individual cells secreting analytes as the respective single color assays, and that the four fluorescent colors can be discerned unambiguously, without overlap. Cells secreting any of the four analytes can therefore be identified unambiguously in an automated fashion, without the need for compensation. Cells co-expressing analytes can be identified by superimposing the individual colors. Studying B cells and T cells experimentally has permitted us to verify the accuracy of co-expression measurements. Each B cell secretes only one type of Ig class/subclass. T cells, in contrast, frequently coexpress cytokines. Serial dilution experiments showed that for T cells the numbers of co-expressors linearly decreased with the numbers of cells plated. For B cells, no coexpressors were found.

Conclusions

The feasibility of four color T and B cell assays have been shown here. This is particularly important when conserving cell material thereby allowing researchers the opportunity for comprehensive immune monitoring spanning multiple cytokines.

P58 A positive control for the detection of functional CD4 T cells in human PBMC – CPI protein pool

Tameem Ansari, Annemarie Schiller, Srividya Sundararaman, Paul Lehmann

Cellular Technology Ltd, Shaker Heights, OH, USA
Correspondence: Tameem Ansari (tameem.asari@gimmunospot.com)

Background

Testing of PBMC for immune monitoring purposes requires verification of their functionality. This is of particular concern when testing cryopreserved PBMC or cells that have been shipped, stored for prolonged periods of time. The CEF peptide pool has been developed as a positive control for CD8 cell functionality. A positive control for detecting CD4 memory cell functionality so far is lacking.

Methods

Protein antigens from infectious/ environmental organisms have been selected that are ubiquitous. T cell reactivity to these antigens has been tested in an IFN-g Immunospot® assay from CTL. Cryopreserved PBMC from 100 Caucasian donors were selected from CTL’s ePBMC database for testing. Magnetic bead depletion experiments were performed to verify CD4 or CD8 response.

Results

Of an initial selection of 12 antigenic systems, (Varicella, Influenza, Parainfluenza, Mumps, Cytomegalovirus, Streptococcus, Mycoplasma, Lactobacillus, Neisseria, Candida, Rubella, and Measles) 3 were selected as a) eliciting CD4 cells exclusively and b) eliciting recall responses in the majority of donors. While individually none of the antigens triggered recall responses in all of the donors, the pool of these three antigens did. Only 2 of 100 donors did not respond to the CPI (Cytomegalo-, Epstein Barr-, and Influenza- virus) protein pool. These two however were impaired functionally non-viable cells with increased numbers of dead and apoptotic cells and showed increased apoptotic progression. Comparisons with CEF peptide pool, showed clear cut responses in ~50% of donors, borderline responses in xx and no responses in 30% of donors.

Conclusions

CPI protein pool is suited as a positive control in Caucasians. Studies are on the way to establish the suitability of this pool for functionality testing in Asians and Africans.

P59 Maximizing odds for detecting a positive T cell response by ELISPOT

Jodi Hanson, Diana Roen, Alexey Karulin, Paul Lehmann

Cellular Technology Ltd, Shaker Heights, OH, USA
Correspondence: Jodi Hanson (jodi.hanson@immunospot.com)

Background

It has been a matter of debate to determine the best cutoffs in ELISPOT assay analysis for the unambiguous identification of a positive T cell response. At present, the answers to the above question is largely based on empirical or mixed criteria. To come up with scientifically validated answers, parametric statistical analysis has to be used, which in turn requires knowledge about the distributional properties of ELISPOT counts in replicate wells.

Methods

PBMC of HLA-A2-0201 positive human subjects who had been infected with HCMV were plated with a HCMV peptide antigen (CEF-7 peptide, pp65 (495-503)) at 100,000 cells per well in IFH-g ELISPOT assays. However, we selected donors whose PBMC, when tested at 100,000 cells per well and challenged with the HLA-A2-0201-restricted HCMV peptide, pp65(495-503) did not display spot counts over medium background. We tested the PBMC for pp65 reactivity in 96 replicate wells to establish the distributional properties of these low frequency ELISPOTs. The distributional properties of the spot counts in the replicate wells were analyzed using diagnostic plots (QQ plots) and the Shapiro-Wilk normality tests.

Results

We observed that increasing the number of PBMC plated per well resulted in higher positive to negative count ratio, lower relative experimental error (CV), and higher power for detecting pp65-induced positive responses without causing false positive results from HCMV negative subjects. This decrease of CV and increase in the power of the test was directly proportional to the numbers PBMC plated. The distributional properties showed that the spot counts in replicate wells follow Normal distribution. We showed that parametric statistics, such as Student's t test can be used and provide higher statistical power detecting weak positive HCMV responses than nonparametric methods (Wilcoxon or DFR). We also show that compared to increasing cell numbers per well, the increase of replicate wells is a more efficacious means of establishing positivity when statistical analysis is at the limits of confidence.

Conclusions

The Normal distribution of ELISPOT counts permits us to make precise predictions regarding the numbers of replicate wells needed, and cut off values, especially when responses from donors are low.

P60 Association between microsatellite instability and clinical response across tumor types in the phase Ib KEYNOTE-012 and KEYNOTE-028 studies of pembrolizumab in PD-L1-expressing advanced solid tumors

Mark Ayers1, Diane Levitan2, Gladys Arreaza2, Fang Liu2, Robin Mogg2, Yung-Jue Bang3, Bert O'Neil4, Razvan Cristescu2

1Merck & Co., Inc., West Point, PA, USA; 2Merck & Co., Inc., Kenilworth, NJ, USA; 3Seoul National University College of Medicine, Seoul, Republic of Korea; 4Indiana University Health University Hospital, Indianapolis, IN, USA
Correspondence: Mark Ayers (mark.ayers@merck.com)

Background

High levels of microsatellite instability (MSI-H) can occur in some patients with colorectal cancer (CRC) due to defects in mismatch repair (MMR). In the phase II KEYNOTE-016 study, MSI-H CRC patients were more responsive to PD-1 inhibition with pembrolizumab, as were MSI-H non-CRC patients. Here, we evaluate loss of MLH1 gene expression across tumor types, and effect of MSI status on clinical response to pembrolizumab in the KEYNOTE-012 and KEYNOTE-028 studies in patients with recurrent, metastatic tumors including breast, gastric, urothelial, and CRC tumors.

Methods

Microarray analysis of loss of MLH1 gene expression was used as a surrogate for MSI-H status to assess prevalence of MSI-H in the proprietary Moffitt database of approximately 18,000 annotated, archived tumors. MSI status of tumors from KEYNOTE-012 and KEYNOTE-028 was evaluated using the Promega MSI Analysis system v1.2. Microsatellite markers were amplified from DNA isolated from tumors and separated and size analyzed by capillary electrophoresis. MSI-H status was identified by comparison of formalin-fixed, paraffin-embedded tumor samples to matched blood DNA. Samples were MSI-H if 2 or more markers changed or non-MSI-H if 1 or no markers changed relative to blood. Logistic regression analyses assessed the correlation between MSI status and overall response rate (ORR, centrally-assessed RECIST v1.1).

Results

In the Moffitt database, loss of MLH1 expression mirrored results from a clinical MSI immunohistochemistry assay in CRC patients. Loss of MLH1 expression also correlated with mutational load. Inferred MSI-H status was identified across multiple tumor types (aggregate prevalence 3.2%) in the Moffitt database (Fig. 36). In KEYNOTE-012 (n=96) and KEYNOTE-028 (n=265), MSI-H status was identified in 6% and 2% of patients, respectively, (3% overall), similar to the unselected population in the Moffitt database. Gastric cancer had the highest prevalence (n=4 patients), with estrogen receptor positive breast, biliary, esophageal, triple negative breast, endometrial, CRC, and bladder tumors also having MSI-H in at least one patient. For patients with MSI and response data (n=310), ORR was 70% versus 12% (1-sided p=0.0001) in those with MSI-H and non-MSI-H status, respectively.
Fig. 36

(Abstract P60).

Conclusions

In this retrospective analysis, identification of MSI-H status was associated with a statistically relevant increase in response to pembrolizumab. This suggests that determination of MSI-H status may be predictive of response to pembrolizumab regardless of histology. The association between MSI-H status and clinical benefit of anti-PD-1 therapy is being evaluated in ongoing studies (KEYNOTE-158 [NCT02628067], KEYNOTE-164 [NCT02460198], and KEYNOTE-177 [NCT02563002]).

Trial Registration

ClinicalTrials.gov identifier NCT01848834 and NCT02054806.

P61 Whole-blood RNA transcript-based signatures predict pre- and post-treatment response in two large independent clinical studies of patients with advanced melanoma treated with tremelimumab

Philip Friedlander1, Karl Wassman2, Chrisann Kyi1, William Oh1, Nina Bhardwaj3

1Icahn School of Medicine at Mount Sinai, New York, NY, USA; 2Check Point Sciences, Cambridge, MA, USA; 3Tish Cancer Institute, Icahn School of Medicine at Mount Sinai, New York, NY, USA
Correspondence: Nina Bhardwaj (nina.bhardwaj@mssm.edu)

Background

Tremelimumab is a cytotoxic T lymphocyte-associated antigen-4-blocking monoclonal antibody. An unmet clinical need exists for blood-based response-predictive gene signatures to facilitate use of cancer immunotherapy in the most clinically effective and cost-efficient manner.

Methods

Pre- and post-treatment (30 days) peripheral blood samples were taken from 210 treatment-naïve melanoma patients receiving tremelimumab in a worldwide, multicenter phase III study. Objective response was determined by an expert panel of radiologists using RECIST criteria. 169 mRNA transcripts were measured for the n=210 patients using reverse transcription polymerase chain reaction (RT-PCR) [1]. Pre- and post-treatment response-predictive signatures were identified in the n=210 training dataset. An independent population of n=150 refractory/relapsed melanoma patients receiving tremelimumab after chemotherapy enrolled in a worldwide, multicenter phase II study [2] was the test dataset.

Results

A 16-gene pre-treatment and 8-gene post-treatment mRNA signatures were identified in the n=210 training dataset. These pre- and post-treatment signatures were tested in the n=150 test dataset first, for objective response as determined by RECIST criteria, and second for one-year survival after treatment. The same genes, coefficients and constant from the training dataset were used in the test cases with the results in Table 5. Both the 16-gene pre- and 8-gene post-treatment response-prediction training gene signatures validated when compared to objective response in the test dataset. The one-year survival criteria also validated with even higher AUC’s for both pre- and post-treatment.
Table 5

(Abstract P61).

 

Pre-Treatment Blood-Based

Post-Treatment Blood-Based

 

16-Gene mRNA Signature

8-Gene mRNA Signature

Melanoma

Training

Test

Test

Training

Test

Test

Cases

N=210

N=150

N=150

N=210

N=150

N=150

Biomarker

Response

Response

Survival

Response

Response

Survival

Sensitivity

96.4%

65.0%

74.4%

85.7%

70.0%

69.8%

Specificity

66.5%

59.2%

55.1%

64.3%

60.8%

59.8%

NPV

99.2%

91.7%

84.3%

96.7%

92.9%

83.1%

AUC

0.8634

0.6376

0.6785

0.7853

0.6125

0.6512

Conclusions

This is the first large clinical study of tremelimumab, independently validated in a second large clinical study, to show both pre- and post-treatment response-predictive mRNA signatures in blood. The pre-treatment biological signature may represent expression levels of particular immune system genes that are needed for a robust immune response against cancer. They may identify patients whose immune systems are already primed to fight the cancer and are particularly amenable to a boost in effectiveness provided by immunotherapy. The 30-day post-treatment biological signature represents a timely way to determine whether the patient is responding positively to the immunotherapy.

Trial Registration

ClinicalTrials.gov identifier NCT00257205.

References

1. Ribas A, et al: Phase III randomized clinical trial comparing tremelimumab with standard-of-care chemotherapy in patients with advanced melanoma.J Clin Oncol 2013, 31(5):616-622.

2. Kirkwood JM, et al: Phase II trial of tremelimumab (CP-675,206) in patients with advanced refractory or relapsed melanoma.Clin Cancer Res 2010, 16(3):1042-1048.

P62 Extensive analysis of PD-1 and CTLA-4 in HVs and GBM patients: implications for monitoring patients on checkpoint inhibitors

Svetlana Bornschlegl, Michael P Gustafson, Dennis A Gastineau, Ian F Parney, Allan B Dietz

Mayo Clinic, Rochester, MN, USA
Correspondence: Svetlana Bornschlegl (bornschlegl.svetlana@mayo.edu)

Background

Checkpoint inhibitors are becoming widely used for immunotherapy but methods to monitor dosing and duration for each individual patient needs to be more fully understood. Immune monitoring by flow cytometry is a tool that can be utilized for measuring responses to immunotherapy in patients. In this study we assessed the expression of PD-1 and CTLA-4 on numerous cell types in healthy volunteers (HVs) and glioblastoma (GBM) patients enrolled in a dendritic cell clinical trial.

Methods

Peripheral blood was collected from 20 HV and 20 GBM patients receiving a DC vaccine in a clinical trial. Whole blood was stained using a previously established method for the identification of multiple cell populations by flow cytometry and novel analysis that captures data on over 120 phenotypes [1]. An extended analysis focused on T cell phenotypes was performed using markers for CD154, CD45RO, CD56, CD3, CD8, CD28, CD4, and CD45. T cell parent populations were characterized by SS, FS, CD45+, CD3+, CD4+, CD8+, CD4+/CD8+ sub populations. Non-T cell populations were assessed by various gating strategies. These populations were measured for PD-1+, CTLA4+, DP, and DN populations.

Results

We identified 15 parent populations, of which 11 expressed PD-1 and 9 expressed CTLA-4. Within subsets of the parent populations we found statistically significant differences (p <0.001) in PD-1 between CD8+ memory and CD8+ naïve cells, CD4+ memory and CD4+ naïve cells, CD8+ NKT and CD8+CD3+ cells, as well as NKT and NK cells. These statistical differences hold true for both HV and GBM patients. We also found HVs to have higher levels of CTLA-4 on CD4+CD8+ cells and B cells compared to GBM patients, and lower levels of PD-1 on CD8+ and naïve CD8+ cells.

Conclusions

This panel allows us to measure approximately 60 phenotypes related to checkpoint proteins. The data presented here identify PD-1 and CTLA-4 phenotypic differences within parent populations, within subsets of parent populations, and differences in healthy volunteers compared to GBM patients. These results may help optimize the targeting of checkpoint proteins as well as other immunotherapeutic approaches in clinical trials.

Acknowledgements

This study is funded in part by the Ivy Foundation.

Trial Registration

ClinicalTrials.gov identifier NCT01957956.

References

1. Gustafson MP, et al: A method for identification and analysis of non-overlapping myeloid immunophenotypes in humans.PloS ONE 2015, 10.3:e0121546.

P63 Objective measurement and clinical significance of IDO1 protein in hormone receptor-positive breast cancer

Daniel Carvajal-Hausdorf1, Nikita Mani1, Vamsidhar Velcheti2, Kurt Schalper1, David Rimm1

1Yale University School of Medicine, New Haven, CT, USA; 2Cleveland Clinic Main Campus, Cleveland, OH, USA
Correspondence: Daniel Carvajal-Hausdorf (daniel.carvajal@yale.edu)

Background

Immunostimulatory therapies targeting immune suppressive pathways produce durable clinical responses in advanced solid tumors. Indoleamine 2, 3-dioxygenase (IDO) is the rate-limiting oxidoreductase that catalyzes the degradation of tryptophan to kynurenine. IDO induces immune tolerance by downregulating CD8+ and effector CD4+ T cell responses. IDO1, the most active isoform, is expressed in diverse tumor types and can be targeted using small molecule inhibitors. Here, we used a validated quantitative in situ assay to measure the levels of IDO1 protein in a retrospective collection of hormone receptor-positive breast cancer (BC).

Methods

IDO1 was measured using quantitative immunofluorescence (QIF) in 362 stage I-III hormone-receptor-positive breast carcinomas represented in tissue microarray format. The levels of IDO1 were determined in the tumor compartment; and were stratified using the median as cut-point. The association between IDO1 levels, clinico-pathological features, CD3+, CD8+ and CD20+ tumor-infiltrating lymphocytes (TIL) and survival was studied.

Results

IDO1 protein was detected in 76.2% of hormone receptor-positive BC. There was no significant association between IDO1 levels and major clinico-pathological characteristics. Increased IDO1 correlated with decreased CD20+ infiltration (P=0.0004) but not with changes in CD3+ or CD8+ levels. Elevated IDO1 expression was associated with worse 20-year overall survival (log-rank P=0.02, HR=1.39, 95% C.I.: 1.05-1.82). IDO1 scores were independently associated with outcome in multivariable analysis.

Conclusions

IDO1 protein is expressed in the majority of hormone receptor-positive BC and is an independent negative prognostic marker. Additionally, IDO1 expression is negatively associated with tumor B cell infiltration. Measurement of IDO1 has the potential to identify a population that might derive benefit from IDO1 blockade.

References

1. Munn DH, Zhou M, Attwood JT, Bondarev I, Conway SJ, Marshall B, et al: Prevention of allogeneic fetal rejection by tryptophan catabolism. Science 1998, 281(5380):1191-1193.

2. Uyttenhove C, Pilotte L, Theate I, Stroobant V, Colau D, Parmentier N, et al: Evidence for a tumoral immune resistance mechanism based on tryptophan degradation by indoleamine 2,3-dioxygenase. Nat Med 2003, 9(10):1269-1274.

3. Ling W, Zhang J, Yuan Z, Ren G, Zhang L, Chen X, et al: Mesenchymal stem cells use IDO to regulate immunity in tumor microenvironment. Cancer Res 2014, 74(5):1576-1587.

4. Schalper KA, Carvajal-Hausdorf D, McLaughlin J, Altan M, Velcheti V, Gaule P, et al: Differential expression and significance of PD-L1, IDO-1 and B7-H4 in human lung cancer. Clin Cancer Res 2016.

P64 CD8+ T cell subsets may be associated with response to anti-CD137 agonist antibody treatment

Serena Chang1, Ronald Levy1, John Kurland2, Suba Krishnan2, Christoph Matthias Ahlers2, Maria Jure-Kunkel2, Lewis Cohen2, Holden Maecker1, Holbrook Kohrt1

1Stanford University School of Medicine, Stanford, CA, USA; 2Bristol-Myers Squibb, Princeton, NJ, USA
Correspondence: Serena Chang (ssc233@stanford.edu)

Background

Cancer immunotherapies historically, have low success rates. One way to increase these success rates involves investigating baseline or early response biomarkers that predict success. Selecting patients with early biomarkers for success and competent immune status are essential when choosing the best therapeutic strategy. CD137 agonists, activators of T and NK cells and a promising newer immunotherapy, have dramatically reduced tumor size and disease in murine tumor models [1] and are currently being examined in clinical trials [2].

Methods

This study investigated the immune response and potential early immune biomarkers in both solid tumor and hematologic cancer patients participating in the phase I BMS-663513 NCT01471210 clinical trial of urelumab monotherapy. We used peripheral blood mononuclear cells to determine the circulating immune competence of patients (n=8) before treatment (baseline/C1D1), 24 hours after the first treatment (C1D2), before the second treatment (C2D1), and up to 30 days after the third treatment (C3R). Analysis was performed using mass cytometry, surveying 38 different immune proteins simultaneously.

Results

At all time points, we observed a trend toward higher central memory and naïve CD8+ T cells in patients with stable disease (n=3) or partial response (n=1) vs. progressors (n=4), while the opposite was true in effector and effector memory RA cells. The most striking difference was seen when considering all CD8+CD27+ T cells, which were higher in those with stable disease or partial response, at all time points. CD8+FcεR1α+ T cells showed a similar trend, albeit to a lesser extent, while CD57+CD8+ T cells showed the opposite trend. CD8+ T cells in both groups were comparably responsive to PMA/ionomycin stimulation, producing multiple cytokines. These trends were not seen in CD4+ T cells or with head and neck solid tumor patients treated with cetuximab.

Conclusions

The aforementioned trends suggest that CD27+CD8+ T cells, and possibly other subsets of CD8+ T cells, should be further explored to determine whether they predict response to anti-CD137 agonist therapy. They also suggest that potential predictive measures of immune status prior to immunotherapy are detectable in peripheral blood.

References

1. Sabel MS, et al: Monoclonal antibodies directed against the T-cell activation molecule CD137 (interleukin-A or 4-1BB) block human lymphocyte-mediated suppression of tumor xenografts in severe combined immunodeficient mice. J Immunother 2000, 23(3):362-368.

2. Sznol M, et al: Phase I study of BMS-663513, a fully human anti-CD137 agonist monoclonal antibody, in patients (pts) with advanced cancer J. Clin Oncol 2008.

P65 Regulation of PD-L1 expression in melanoma and immune cells

Shuming Chen1, George Crabill1, Theresa Pritchard1, Tracee McMiller1, Drew Pardoll2, Fan Pan2, Suzanne Topalian1

1Department of Surgery, Johns Hopkins University School of Medicine, Sidney Kimmel Comprehensive Cancer Center, and Bloomberg-Kimmel Institute for Cancer Immunotherapy, Baltimore, MD, USA; 2Department of Oncology, Johns Hopkins University School of Medicine, Sidney Kimmel Comprehensive Cancer Center, and Bloomberg-Kimmel Institute for Cancer Immunotherapy, Baltimore, MD, USA
Correspondence: Shuming Chen (schen72@jhmi.edu)

Background

The therapeutic effects of PD-1/PD-L1 inhibition in multiple cancers indicates a critical role for this pathway in immunosuppression in the tumor microenvironment (TME), but factors regulating PD-L1 expression on tumor and immune cells are poorly understood. The dichotomous transcription factors STAT1 and STAT3 have both been reported to bind the PD-L1 promoter. The current study investigates the role of STAT1/3 and other signaling in cytokine-induced PD-L1 expression on human melanoma (MEL) cells and monocytes.

Methods

Seventeen cultured MELs or short-term monocyte cultures were treated with recombinant cytokines including IFN-g, IL-6, IL-10, IL-32g, or TNF-a. PD-L1 cell surface protein expression was detected by flow cytometry, and mRNA by quantitative real-time RT-PCR (qRT-PCR). STAT1 or/and STAT3 were knocked down by small interfering RNAs (siRNAs). Total and phosphorylated STAT1/3 proteins were quantified by Western blotting.

Results

While PD-L1 is expressed on 35-40% of MELs in situ, it was not expressed on 17 cultured MELs. IFN-g significantly enhanced PD-L1 protein expression on MELs (p=0.0003), increasing PD-L1 mRNA expression (qRT-PCR) in association with PD-L1 cell surface protein expression (FACS) in all MELs tested (p=0.0004). This suggests that IFN-g regulates PD-L1 expression primarily at the transcriptional level and not via translocation of intracellular protein stores. Enhanced PD-L1 expression in IFN-g-treated MELs correlated with increased STAT1 phosphorylation (p=0.05). Consistent with this, siRNA knockdown of STAT1 reduced PD-L1 expression by 40-70% in 2 MELs after 24-48hr IFN-g exposure. In contrast, STAT3 knockdown reduced IFN-g-induced PD-L1 expression by only 12-15%, on one of two MEL lines. In cultured monocytes from two donors, PD-L1 mRNA expression was induced by IFN-g, IL-10, IL-32-g and TNF-a, and significantly correlated with PD-L1 cell surface protein expression, suggesting that these cytokines acted at the transcriptional level. In monocytes, IFN-g was associated with markedly increased pSTAT1, and IL-10 with increased pSTAT3. These cytokines were not associated with increased pp65 or focal adhesion kinase (FAK) phosphorylation in monocytes.

Conclusions

In addition to IFN-g, other cytokines in the TME may play important, coordinate and selective roles in promoting PD-L1 expression on different cell types including tumor and stromal cells. pSTAT1 and pSTAT3 are associated with PD-L1 protein expression in response to different cytokine stimuli. Future studies will further characterize cytokine-triggered transcription factors and signaling pathways responsible for PD-L1 expression on tumor cells and immune cells. Understanding mechanisms regulating PD-L1 expression will help guide the development of more optimal predictive biomarkers and combinatorial therapies based on anti-PD-1.

P66 Gene expression markers of tumor infiltrating leukocytes

Patrick Danaher1, Sarah Warren1, Lucas Dennis1, Andrew M White1, Leonard D'Amico2, Melissa Geller3, Mary L Disis2, Joseph Beechem1, Kunle Odunsi4, Steven Fling2

1NanoString Technologies, Seattle, WA, USA; 2University of Washington, Seattle, WA, USA; 3University of Minnesota, Minneapolis, MN, USA; 4Roswell Park Cancer Institute, Buffalo, NY, USA
Correspondence: Patrick Danaher (pdanaher@nanostring.com)

Background

Assays of the abundance of immune cell populations in the tumor microenvironment promise to inform immune oncology research and the choice of immunotherapy for individual patients. We propose to measure the intratumoral abundance of various immune cells populations with gene expression. In contrast to IHC and flow cytometry, gene expression assays yield high information content from a clinically practical workflow. Previous studies of gene expression in purified immune cells have reported hundreds of genes showing enrichment in a single cell type, but the utility of these genes in tumor samples is unknown. We describe a novel statistical method for using co-expression patterns in large tumor gene expression datasets to validate previously reported candidate cell type marker genes.

Methods

We used co-expression patterns in 9986 samples from The Cancer Genome Atlas (TCGA) to validate previously reported cell type marker genes. We compared immune cell scores derived from these genes to measurements from flow cytometry and immunohistochemistry. We characterized the reproducibility of our cell scores in replicate runs of RNA extracted from FFPE tumor tissue.

Results

We identified a list of 60 marker genes whose expression levels quantify 14 immune cell populations. Cell type scores calculated from these genes are concordant with flow cytometry and IHC readings, show high reproducibility in replicate RNA samples from FFPE tissue and reveal an intricate picture of the immune infiltrate in TCGA. Most genes previously reported to be enriched in a single cell type have co-expression patterns inconsistent with cell type specificity.

Conclusions

Due to their concise gene set, computational simplicity and utility in tumor samples, these cell type gene signatures may be useful in future discovery research and clinical trials to understand how tumors and therapeutic intervention shape the immune response.
Fig. 37

(Abstract P66). Comparison of gene expression and flow cytometry cell type measurements in PBMCs

Fig. 38

(Abstract P66). Comparison of gene expression and IHC cell type measurements in FFPE

Fig. 39

(Abstract P66). Reproducibility of gene expression measurements of immune cell types

P67 Quantitative real-time PCR based diagnostic to assess NKT cell function in breast cancer patients

Roshanak Derakhshandeh, Tonya J Webb

University of Maryland, Baltimore, Baltimore, MD, USA
Correspondence: Roshanak Derakhshandeh (rderakhshandeh@som.umaryland.edu)

Background

Breast cancer is a leading cause of cancer-related death among women worldwide. Although, surgery, radiotherapy, and chemotherapy have improved the 5-year survival rate, new treatment methods are needed to combat this disease. To date, significant efforts have been invested into harnessing the therapeutic potential of the immune system for the treatment of cancer. However, tumor tolerance and immune suppression can severely limit its therapeutic efficacy. In fact, natural killer T (NKT) cells play an important role in cancer immune surveillance, but are reduced in cancer patients. In order to assess which patients will likely benefit from immune cell-based therapies, we have developed a quantitative method to rapidly assess the baseline function of NKT cells using stimulation with artificial antigen presenting cells followed by and quantitative real-time PCR (aAPC-qPCR).

Methods

In this study, we assessed NKT cell number and function in healthy donors and breast cancer patients by flow cytometry, ELISA, and qPCR. In addition, we assessed the percentage of tumor-infiltrating lymphocytes and PD-L1 expression within the tumor microenvironment by immunohistochemistry.

Results

Although % circulating NKT cell were significantly reduced in breast cancer patients (BCP), compared to healthy donors (HD), we detected NKT cell function in 82% HD (n=22) and 70% BCP (n=30). We compared high responders (high IFN-γ induction) to low responders and found that there was no significant difference in NKT cell number between these two groups. Following further characterization of these groups, it was found that low responders had a significant reduction in the induction of TNFα, LAG3, and LIGHT.

Conclusions

In summary, this data we have developed a novel diagnostic platform using aAPC-qPCR to determine NKT cell function in patients. This technology is important because NKT cell number did not correlate with function in our breast cancer patient cohort. Thus, our studies demonstrate that there is a critical need to assess baseline immune function prior to the initiation of immunotherapy. Future studies can focus identifying new breast cancer classifications according to immune gene expression patterns, and these tumor subtypes may provide a basis for new therapeutic opportunities.

Acknowledgements

Supported by NIH/NCI 1R21CA162273, R21CA184469, and R21CA199544 grants to Tonya Webb.

References

1. Siegel RL, Miller KD, Jemal A: Cancer Statistics 2015. CA Cancer J Clin 2015, 65(1): 5-29.

2. Hermans IF, et al: NKT cells enhance CD4+ and CD8+ T cell responses to soluble antigen in vivo through direct interaction with dendritic cells. J Immunol 2003, 171(10): 5140-5147.
Fig. 40

(Abstract P67). a Circulating % NKT cells are reduced in BC patients. Peripheral blood mononuclear cells (PBMC) were collected from healthy donors (HD) and breast cancer patients (BC), stained for NKT TCR and analyzed by flow cytometry. b NKT cell function correlates total T cell function in BC patients. PBMCs from BC patients were stimulated with CD1D-Ig aAPC loaded with a-GalCer to specifically activate NKT cells or anti-CD3/CD28 to activate T cells. Media serves as a negative control and PMA/ionomycin was the positive control. After 4 hours, RNA was extracted and qPCR was performed to assess IFN-g and 18S. Fold induction was calculated relative to the control

P68 Arming of CD56dim and CD56bright NK cells in IL-15-infused cancer patients

Sigrid Dubois, Kevin Conlon, Bonita Bryant, Jennifer Hsu, Nancy Beltran, Jürgen Müller, Thomas Waldmann

Lymphoid Malignancies Branch, Center for Cancer Research, National Cancer Institute, Bethesda, MD, USA
Correspondence: Sigrid Dubois (duboiss@mail.nih.gov)

Background

Survival and expansion of NK cells depends on interleukin (IL)-15. It has recently been shown that IL-15 infusions caused NK cell expansions in cancer patients. The objectives of our study were to assess the effects of IL-15 on functions of CD56dim and CD56bright subpopulations of NK cells in IL-15-treated cancer patients.

Methods

We monitored numbers, phenotypic changes, cytokine production and lytic activities of CD56dim and CD56bright NK cell subpopulations in the blood of cancer patients that had received daily infusions of 2 μg/kg IL-15 for 10 days.

Results

Ten-day continuous infusion of IL-15 led to expansions of both CD56dim and CD56bright subpopulations of NK cells. Phenotypic analyses revealed that IL-15 infusions caused the expression of CD56bright–typical surface proteins on CD56dim NK cells that included Trail, CD62L, NKG2D, CD94 and the IL-18 receptor, while CD56bright NK cells remained essentially unchanged. CD56bright NK cells retained their superior ability to produce IFNγ/TNFα in responses to IL-12/IL-18 when compared with CD56dim NK cells, and IL-15 infusions increased the percentage of cytokine-producing CD56bright NK cells. The cytotoxic capacity of CD56dim NK cells remained superior to CD56bright NK cells even after IL-15 infusions. However, cytotoxic competencies were increased for both subpopulations after IL-15 infusions that resulted in substantial lytic activities via natural cytotoxicity receptors, stress receptors, and antibody-dependent cytotoxicity even among CD56bright NK cells.

Conclusions

These data show that IL-15 infusions increase the functional abilities of both types of NK cells in cancer patients.

P69 Identification of a novel subset of tumor-resident human CD8+ T cells, marked by dual expression of CD103 and CD39

Rebekka Duhen1, Thomas Duhen2, Lucas Thompson3, Ryan Montler2, Andrew Weinberg1

1Earle A. Chiles Research Institute, Providence Cancer Center, Portland, OR, USA; 2AgonOx, Inc., Portland, OR, USA; 3Juno Therapeutics, Portland, OR, USA
Correspondence: Rebekka Duhen (rebekka.duhen@providence.org)

Background

Homing and retention of T cells in tissues is mediated by the interaction of adhesion molecules with their respective ligands. Among those, the integrin CD103 interacts with its ligand E-cadherin and allows T cells to home to epithelial tissues. T cells expressing high levels of CD103 have recently been identified as tissue-resident memory (TRM) cells and play a crucial role in protecting epithelial tissues against viral infections. Previous reports have shown that CD103+ CD8+ T cells were present in some but not all human solid malignancies.

Methods

Cell sorting, gene expression analysis, and TCR sequencing.

Results

Here, while confirming these data, we identify a subset of CD103+ CD8+ T cells that co-express the ectonucleotidase CD39. This subset is enriched in primary tumors and metastatic lymph nodes but absent in the blood and normal lymph nodes of cancer patients. We compared several tumor histologies and found highest frequencies in head and neck squamous cell carcinomas (HNSCC), ovarian, lung and rectal cancers (ranging from 20-80% of tumor-infiltrating CD8+ T cells), whereas those cells were absent or low in primary colon cancer and colorectal liver metastasis. Gene expression analysis of CD103/39 double positive CD8+ T cells revealed a gene signature reminiscent of TRM cells (CCR7lo, L-selectinlo, S1PR1lo and CD69hi), and their activated phenotype (HLA-DRhi, Ki67hi, Granzyme Bhi) implies strong tumor reactivity. Furthermore, TCR repertoire analysis shows high clonality and distinct CDR3 sequences in this subset compared to other CD8+ T cells present in the tumor. Based on this phenotype, gene signature, circulation pattern and clonality, we believe that CD103/39 double positive CD8+ T cells are being chronically stimulated within the tumor microenvironment, and may recognize neoantigens. In support of this finding, our in vitro data suggests that expression of CD39 is upregulated as a result of strong, sustained TCR stimulation in naïve CD8+ T cells. Finally, to better address the role of those cells in vivo we examined 9 different solid murine tumor models. Unexpectedly, CD103/39 double positive CD8+ T cells were only found in the transgenic MMTV-PyMT breast cancer model. Utilizing this model, we plan to study their differentiation and function in vivo and address their antigenic specificity and role in tumor development.

Conclusions

Taken together our findings suggest that targeting tumor-resident CD103/39 CD8+ T cells may be a promising approach to enhance immune-mediated tumor regression.

P70 Characterization of tumor infiltrating T cell receptor (TCR) repertoire in non-muscle invasive bladder cancer (NMIBC) patients treated with Bacillus Calmette-Guérin (BCG)

Max Kates1, Brandon Early2, Erik Yusko3, Taylor H Schreiber2, Trinity J Bivalacqua1

1Johns Hopkins University, Baltimore, MD, USA; 2Heat Biologics, Durham, NC, USA; 3Adaptive Biotechnologies, Seattle, WA, USA
Correspondence: Brandon Early (bearly@heatbio.com)

Background

Bladder cancer is the 5th most common malignancy in the US, and the majority of bladder cancer is diagnosed while confined to the most superficial layer (non-muscle invasive bladder cancer, NMIBC) [1]. Treatment goals in this minimal residual disease setting are ideal for immunotherapy: reduce local disease recurrence following surgical resection and prevent progression to muscle invasive bladder cancer (MIBC). Bacillus Calmette-Guérin (BCG), an attenuated form of Mycobacterium bovis, is an intravesical immunotherapy which remains the mainstay of NMIBC treatment since 1976. Despite the tenure in bladder cancer treatment, full characterization of the immunologic mechanism of action of BCG is still lacking [2]. In a pilot study, we sought to investigate the diversity of T cells infiltrating bladder tumors and compare the changes in T cell diversity among patients who were responders and unresponsive to BCG.

Methods

Six patients were selected from the IRB-approved Johns Hopkins bladder cancer tumor repository. All patients had T1 disease without concurrent carcinoma in situ (CIS) and received standard of care BCG induction and maintenance. Three patients were classic