Coordinated responses to individual tumor antigens by IgG antibody and CD8+ T cells following cancer vaccination
- 1.7k Downloads
One of today’s greatest hurdles for cancer immunotherapy is the absence of information regarding which tumor antigens are already recognized by patients receiving immunotherapies, and whether those therapies then boost or generate an immune response against tumor proteins. For CD8+ T cells in particular, patient-specific immune recognition and responses at the level of individual tumor antigens are rarely characterized. Because of this, some immunologists have turned to serum antibodies as an alternative measure of antigen-specific anti-tumor immunity. In this work, we sought to simultaneously interrogate serum IgG and CD8+ T cell recognition of individual tumor antigens to determine whether antigen-specific serum IgG antibodies provide a window into the behavior of antigen-specific CD8+ T cell responses. Using antibody-based assays to evaluate immune response repertoires and focus T cell antigen exploration could afford substantial advantages for discovering and monitoring the anti-cancer immune responses of patients enrolled on clinical trials.
We vaccinated female BALB/c mice with a novel combination of an autophagosome-enriched vaccine derived from 4T1 mammary carcinoma along with poly-I:C adjuvant, then screened serum for IgG binding to arrays of 15mer peptides containing known mutation sites in 4T1. Simultaneously, we primed CD8+ T cell cultures from these same animals with 8-11mer peptides derived from these antigens. These primed T cells were then stimulated to measure recognition of the peptides or live 4T1 cells by IFNγ release.
Vaccinated animals demonstrate increases in antigen-specific CD8+ T cell recognition of 4T1 tumor cells and peptides. For proteins confirmed in 4T1 cells and vaccine by mass spectrometry, there is a correlation between this increased CD8+ T cell IFNγ release and serum IgG binding to individual peptide antigens.
These results suggest it is possible to observe some features of a patient’s antigen-specific T cell repertoire via an antibody surrogate, which has implications for tumor antigen discovery and clinical monitoring of antigen-specific anti-tumor immunity.
KeywordsImmunological Monitoring Antigen CD8 + T cell IgG Antibody Vaccine Autophagosome Poly-I:C 4T1
Liquid chromatography tandem mass spectrometry
Single nucleotide variant
Quantitative tandem mass tag
A large background of autoantibody signals to thousands of normal human proteins is frequently observed in IgG biomarker surveys [1, 2, 3, 4, 5]. On average, over 20% of the entire surveyed human proteome is targeted by a unique landscape of these autoantibodies in healthy individuals . Such preexisting or “natural” antibody landscapes are thought to be the result of prior adaptive immunity to similar peptide mimics found in commensal microbes, foods, environmental exposures, infections, and autologous proteins. In spite of the frequency of autoantibodies observed in humans, and the similarity between many types of tumor antigens and autologous targets, it is not known whether these serum antibodies or changes in their abundance might also hint at the antigen-specific behavior of an individual’s T cell repertoire. Others have used antibody as a surrogate measure of antigen-specific anti-tumor immunity [4, 5, 6], and we hypothesized that IgG antibody signals would be more likely to overlap with features of antigen-specific CD8+ T cell recognition than expected by chance. Potential mechanisms for such a relationship could occur via overlap with the underlying CD4+ T cell repertoire necessary for activating both CD8+ T cells and B cells, or from antibody-aided T cell activation via Fc receptor targeting of antigens to antigen presenting cells. Improved understanding of the antigen-specific relationships between antibody and T cell responses to tumor antigens could lead to improved immune monitoring for cancer patients and a deeper understanding of what features define clinically-relevant tumor antigens.
Although the overall benefit of B cell responses to cancer remains controversial [7, 8, 9], there is a long history of surveys for antigen-specific anti-tumor antibodies [10, 11]. One type of anti-tumor immunity increasingly recognized as important to improved outcomes for patients with cancer are responses to tumor-specific single nucleotide variant (SNV) neoantigens [12, 13, 14, 15, 16], which differ by a single amino acid from their wild-type (WT) counterpart autoantigens. We sought to screen for IgG antibodies to peptides centered at previously reported mutation sites in the 4T1 tumor model in both SNV neoantigen and their WT autoantigen counterpart versions. We hypothesized that patterns in these antibody profiles would relate to the 4T1 vaccine-induced T cell recognition of those same mutation sites. In viral immunity, there are documented examples of IgG antibody responses mirroring CD4+ responses at the level of individual antigens [17, 18, 19]. Similar to the viral literature, potential links have been observed between anti-tumor antibodies and T cell responses to specific tumor antigens [20, 21], and increased antigen-specific antibody responses have been observed in association with improved outcomes following immunotherapy treatments typically understood to depend on T cells [4, 6, 22].
For this work, we chose 4T1, a metastatic murine mammary carcinoma model in BALB/c mice with a limited number of previously described neoantigens, and a vaccine that is known to both work in 4T1 therapeutically and generate cross-reactive immunity to diverse unrelated tumors [23, 24]. This model provided an opportunity to repeatedly interrogate antigen-specific immune responses to specific components of our vaccine in a well-controlled system. The 4T1 tumor cell line was a gift of Emmanuel Akporiaye (Earle A. Chiles Research Institute, Portland, OR), from stocks received from Suzanne Ostrand-Rosenberg (UMBC, Baltimore, MD). Cell line identity was confirmed identical to ATCC 4T1 and free from Mycoplasma and other common eukaryotic contaminants via microsatellite profiling (IDEXX RADIL). Tumor cells were thawed directly from the confirmed bank and passaged less than 4 times before use. Cells were cultured in complete media consisting of RPMI-1640 (Lonza) with 1% L-Glutamine (Lonza), 1% Sodium Pyruvate (Lonza), 1% Non-essential Amino Acids (Lonza), 0.1% Beta Mercaptoethanol, 50 mg/L Gentimicine Sulfate, and 10% fetal bovine serum (Atlas Biologicals Lot # 1070612). Production of three 4T1 autophagosome-enriched vaccine lots was performed as previously described [23, 25]. In brief, tumor cells were seeded into T225 flasks, grown to ~ 70% confluence, and treated with 20 mM ammonium chloride and 100 nM Bortezomib (Velcade) to induce autophagosome formation. Treated 4T1 cells were harvested and sonicated to release autophagasomes. Suspended autophagasomes were harvested with centrifugation at 12,000 G. Protein content was measured by a BCA assay using bovine serum albumin as a standard, and harvested 4T1 autophagosome-enriched vaccine was diluted to a protein concentration of 1 mg/mL in hetastarch vehicle and frozen at − 80 °C until use. Age-matched 14–20 week old female BALB/c mice (Jackson Laboratories) were vaccinated in both inguinal nodes with a total of 10 μg 4T1 autophagosome-enriched vaccine plus 3 μg of Vaccigrade poly-I:C (Sigma-Aldrich) in 20 μL hetastarch carrier, vaccine and carrier alone, poly-I:C adjuvant and carrier alone, or left untreated. Animals were boosted after two weeks with a single subcutaneous injection of the same total dose in the left flank. After another two weeks serum was harvested for analysis or mice were challenged with 5000 live 4T1 cells in the left mammary fat pad. Tumor growth in challenged mice was measured thrice weekly for 30 days until immunohistochemistry and tumor bearing serum experiments, or until a maximal area of 150 mm2, which was the determinant for death in overall survival experiments.
Day 30 4T1 tumors were pretreated for 24 h in a zinc solution, placed in 70% ethanol, and then paraffin embedded until staining as previously described . Five μm sections were cut and fluorescently stained with DAPI and specific antibodies to CD8a (53–6.7, BD Pharmingen), F4/80 (Cl:A3–1, Bio Rad), CD3 (SP7, Spring Bioscience), FOXP3 (FJK-16 s, eBioscience), and CD4 (RM4–5, BD Biosciences) via tyramide signal amplification. Multispectral fields were imaged with a multispectral microscope (PerkinElmer, Vectra) and 15 representative 20× fields per sample were quantified with vendor software (PerkinElmer, Inform).
4T1 whole exome sequencing and variant detection
DNA was then isolated from our 4T1 cell line bank using a Qiagen DNeasy kit and sent to a contractor for whole-exome sequencing (Otogenetics) at a target 50× coverage depth. Using CLC Genomics Workbench v7.04, the resulting Illumina FASTQ files were aligned to the mm10 reference genome using CLC NGS core tools, a BWS algorithm, to preserve annotations. Known SNVs and indels in BALB/cJ versus mm10 were subtracted using a variant file downloaded from the Sanger mouse genome project (www.sanger.ac.uk/science/data/mouse-genomes-project). Heterozygous non-synonymous protein-coding variants detected > 10 times were determined to be 4T1-specific SNV mutation candidates.
4T1 15mer mutation site peptide arrays
Mutation site candidates identified from our sequencing were compared to a list of heterozygous nonsynonymous protein coding 4T1 SNVs identified in prior publications [27, 28]. One of these studies reported immunologic response data to 17 4T1 neoantigens , and we included all of these previously reported immunogenic 4T1 mutation sites on the arrays. As space allowed in the array design, we additionally included 66 of the 81 total mutation sites identified by both our independent sequencing and confirmed by at least one of the other reports. The Mouse ENSMBL protein database was downloaded from BioMart (www.ensembl.org/biomart) , and 15mer wild-type peptide sequences were extracted centered at the 75 selected coordinates. The 15mer wild-type sequences were then altered to the identified SNV versions for a total of 150 WT and SNV peptides. These 150 peptides were printed in triplicate on 60 replicate custom peptide arrays along with the known 4T1 retroviral antigen AH1  and anti-mouse IgG control spots by JPT Peptides (Berlin, Germany). Twenty of these arrays were used in preliminary experiments, and forty in follow-up experiments paired with T cell data. Whole mouse sera were pooled from 2 to 3 animals per experimental group, diluted 1:200, and incubated on the peptide arrays for one hour at 30 °C. IgG signals were detected with a fluorescent anti-mouse IgG secondary. All samples reacted to anti-mouse IgG control spots. Each array spot was imaged with a high resolution fluorescence scanner and its intensity quantified with GenePix spot-recognition software (Molecular Devices). Resulting IgG fluorescence intensity values were averaged across each of the three replicate spots for further analysis. In preliminary studies, the average intensity values from all initial 20 arrays were normalized simultaneously using an interquartile range transformation performed using BRB-ArrayTools v4.5.0 developed by Dr. Richard Simon and the BRB-Array Tools Development Team (brb.nci.nih.gov/BRB-ArrayTools). These preliminary arrays are presented in Additional file 1: Data file S1, and were used for selecting antigens to additionally investigate in T cell assays. The 40 follow-up arrays paired with T cell data were analyzed as the raw average of the three replicate spots without such normalization, are plotted in figures, and additionally included in Additional file 1: Data file S1.
In vitro T cell IFN γ release peptide assays
Antigens selected for additional profiling via IFNγ T cell assays were selected based on a profile of the preliminary peptide array data. We selected thirty-one antigens that spanned a range of properties: sites with a strong preexisting IgG background signal, sites with a post-vaccine IgG signal increase across multiple experiments, sites with high and low predicted MHCI affinity, and mutation sites without any of these distinctions but previously reported as immunogenic . All experiments were performed using pooled splenocytes from 2 to 3 individual female BALB/c mice harvested two weeks after their second vaccination. In the case of CD8+ enriched experiments, CD4+ cells were depleted in vivo three days prior to spleen harvest using 200 μg of GK1.5 anti-CD4 antibody administered IP. CD4 depletion was confirmed by flow cytometry. After ACK lysis of red blood cells, 1 × 106 splenocytes were plated into each well of 96 well round-bottom tissue culture plates and given primary stimulation in complete media with 10% FBS and 5 μM of either SNV or WT versions of mutation site peptides manufactured by A&A Labs (San Diego, California). 15mer peptides matched the IgG arrays, and 8-11mer minimal peptides designs were based on predicted ability to bind MHCI. Both WT and SNV 8-11mer peptides were based on the length and frame of the top predicted MHCI binding minimal 8-11mer SNV peptide identified using NetMHCpan v2.8 Server. NetMHCpan, was used to calculate predicted H2-Kd, H2-Dd, and H2-Ld MHCI binding scores for all possible WT and SNV 8mers, 9mers, 10mers, and 11mers that include the mutation site . After 48 h of primary peptide stimulation, IL2 was added at 10 Cetus units/mL. After an additional 96 h, contents of each well were washed and split onto either 2° peptide restimulation with 5 × 105 irradiated splenocytes, irradiated splenocytes alone, media only, or plated onto 1 × 105 live 4T1 cells. Supernatants were harvested after an additional 20 h, frozen at − 80 °C, and later analyzed for IFNγ by ELISA.
TMT LC-MS/MS of 4T1 cells and autophagosome-enriched vaccine
Quantitative tandem mass tag (TMT) liquid chromatography tandem mass spectrometry (LC-MS/MS) was performed by the Proteomics Shared Resource at Oregon Health & Science University on three 4T1 autophagosome-enriched vaccine lots and three paired samples of untreated whole 4T1 cells. Samples were lysed using a probe sonicator and protein concentration was estimated using BCA assay. Forty μg of protein per sample was trypsin digested in solution. In brief, samples were dried, dissolved in 10 μL of 4X buffer (8 M urea,1 M Tris (pH 8.5), 8 mM CaCl2, 0.2 M methylamine), reduced, alkylated, diluted to a final 2 M urea concentration and digested by addition of 1.6 μg of sequencing grade trypsin overnight (ProMega) Completion of the digestion was confirmed by 1-D gel analysis. Twenty-five μg of each digested sample was then solid phase extracted using Oasis HLB 1cm3 cartridges (Waters Corporation), and peptides dried by vacuum centrifugation. Samples were labeled with 10-plex TMT reagents (Thermo Scientific), pooled together, and on-line two dimensional reverse phase/reverse phase (RP-RP) liquid chromatography used to separate into 9 fractions at high pH, and each fraction further separated at low pH. Peptides were analyzed using an Orbitrap Fusion Mass Spectrometer (Thermo Scientific) with a synchronous precursor selection MS3 TMT method . Twenty μL samples (32.9 μg) were injected onto a NanoEase 5 μM XBridge BEH130 C18 300 μM × 50 mm column (Waters) at 3 μL/min in a mobile phase containing 10 mM ammonium formate (pH 10). Peptides were eluted by sequential injection of 20 μL volumes of 14, 20, 22, 24, 26, 28, 30, 40, and 90% acetonitrile (ACN) in 10 mM ammonium formate (pH 10) at a 3 μL/min flow rate. Eluted peptides were diluted with mobile phase containing 0.1% formic acid at a 24 ul/min flow rate and delivered to an Acclaim PepMap 100 μM × 2 cm NanoViper C18, 5 μM trap (Thermo Scientific) on a switching valve. After 10 min of loading, the trap column was switched on-line to a PepMap RSLC C18, 2 μM, 75 μM × 25 cm EasySpray column (Thermo Scientific). Peptides were then separated at low pH in the 2nd dimension using a 7.5–30% ACN gradient in mobile phase containing 0.1% formic acid at 300 nL/min flow rate. Each 2nd dimension LC run required 2 h for separation and re-equilibration, so the entire LC/MS method required 18 h for completion. Survey scans were performed in the Orbitrap mass analyzer (resolution = 120,000), and data-dependent MS2 scans performed in the linear ion trap using collision-induced dissociation (normalized collision energy = 35) following isolation with the instrument’s quadrupole. Reporter ion detection was performed in the Orbitrap mass analyzer (resolution = 60,000) using MS3 scans following synchronous precursor isolation of the 10 most intense ions in the linear ion trap, and higher-energy collisional dissociation in the ion-routing multipole (normalized collision energy = 65).
Mass spectrometry data was processed against the UniProt Swiss-Prot canonical mouse protein database (v. 2014_05, 16,669 sequences) with SEQUEST HT in Proteome Discoverer v1.4 (Thermo Scientific). Search settings were: monoisotopic parent ion mass tolerance of 1.25 Da, monoisotopic fragment ion tolerance of 1.0 Da, tryptic cleavage with up to 2 missed cleavages, variable modification of oxidized methionine, and static modifications for TMT reagents (peptide N-term and lysines) and alkylated cysteines. Peptide sequence assignments were validated using Percolator  q-values (less than 0.05) and 20 ppm delta mass agreement between measured and theoretical peptide masses. TMT reporter ion intensities of individual peptides were exported as text files and processed with in-house scripts. A median reporter ion intensity cutoff of 1500 was used to reject low quality peptides, and all reporter ion intensities for unique peptides matched to each respective protein were summed to create total protein intensities. A minimum of 2 peptides contributing to the protein total was required for each identification to improve data quality. Protein identification, quantitative information, and additional UniProt annotations were tabulated for all proteins and are listed in Additional file 1: Data file S1. A total of 4416 proteins were identified and quantification was done on 4196 proteins (excluding contaminants). Only this discovery confirmation, and not quantitative abundance, was used to separate experimental groups in Fig. 6, and Additional file 2: Figures S5–7.
Analyses were performed on either summary data or individualized experiments, and this information is placed alongside the specific type of test performed and p-value (P) within the figure legends. All statistical tests were considered significant at the P < 0.05 level and were performed with Prism 7 (GraphPad). In general, parametric comparisons were either two sample t-tests or paired t-tests, and non-parametric tests were Wilcoxon matched-pairs signed rank tests. Significance of all correlations was determined by linear regression and Pearson correlation coefficient. Graphs and statistics in Figs. 2, 3, 4, 5, and 6 and Additional file 2: Figures S3–7 can be recreated from values presented in Additional file 1: Data file S1.
We next sought to determine whether IgG antibodies could be used to indicate immunity to known mutation sites in 4T1 after treatment with our vaccine model. We were led to this by others whose observations have suggested links between IgG antibody and T cell responses to human tumor antigens [6, 21]. We isolated DNA and performed whole-exome sequencing on our 4T1 cell bank, and used the sequencing data to identify heterozygous single nucleotide variants by comparison to a female BALB/cJ reference sequence; subsequently referred to as SNVs. We used SNVs that were both identified in previous reports [27, 28, 35] and confirmed by our sequencing to design a custom 15mer peptide array for 75 SNV neoantigens and 75 alternate allele wild type (WT) autoantigens centered at 4T1 mutation sites, as well as the known retroviral antigen AH1  (Additional file 2: Figure S2). A number of these mutation sites have been previously reported as immunogenic to murine CD4+ or CD8+ T cells . Each array contained all peptides printed in triplicate along with anti-mouse IgG controls. We ran preliminary IgG arrays with sera harvested from animals used in the tumor challenge and in vitro experiments already presented (Fig. 1a-e), and these preliminary arrays were used to identify outlier antigens of interest for final IgG array experiments paired with IFNγ T cell assays. Data from both preliminary IgG arrays and final IgG arrays paired with T cell IFNγ release data are available (Additional file 1: Data file S1).
Prior to the final IgG array experiments presented in Fig. 2, we had used preliminary IgG arrays to identify a number of 4T1 antigens which had strong increases in IgG signals after vaccination or strong preexisting IgG signals in naïve animals. We used these results to select a smaller set of antigens for investigation in parallel T cell assays, and ran these T cell assays with splenocytes from the same animals whose serum IgG was analyzed in Fig. 2. These T cells were stimulated with either 15mer peptides matching the IgG array or predicted MHCI binding 8-11mer peptide designed with MHCI prediction software . The chosen mutation sites are listed in relation to IgG array data in Additional file 1: Data file S1, and were tested as both WT autoantigen and SNV neoantigen peptides.
Our IgG array data demonstrated that many of the strongest vaccine-induced antibody responses to 4T1 peptides could be detected, albeit at lower levels, in naive animals (Additional File 2: Fig. S4).
We observed IgG antibodies to 4T1 peptides in naïve female BALB/c mice, and this population of antibodies demonstrates stronger recognition of neoantigen peptides than autoantigen counterpart peptides which differ by a single amino acid. This result suggests a bias for the recognition of certain antigens prior to tumor exposure, perhaps caused by an individual’s unique history of tolerance to autoantigens or prior exposure to cross-reactive foreign antigens. Recent work has demonstrated a dramatic dependence of checkpoint blockade immunotherapies on specific microflora [38, 39], an effect typically ascribed to global cytokine changes, but perhaps also due in part to bacterial antigen cross-reactivity with tumor antigens. Germ free animals lacking such immunologic history do not respond well to cancer therapy . The correlations between vaccine-induced T cell responses and antigen-specific IgG antibody signals observed in this study suggest that IgG may be one biomarker to observe such relationships between past immunologic history and future or ongoing anti-tumor immunity. And beyond the potential role of antibody as a biomarker, antibody may be directly involved in transferring prior immunologic knowledge to help prime or boost T cell populations.
Antigen-specific antibody can increase T cell activation through improved antigen uptake and cross presentation by antigen presenting cells in mice , and a similar Fc receptor dependent effect has also been observed in humans: some patients receiving monoclonal antibodies to EGFR (Cetuximab) generate elevated circulating EGFR853–861–specific CD8+ T cells . Therefore, antibodies that happen to bind a tumor peptide or protein – such as those observed in this work – could provide a mechanism for improved CD8+ responses to those same antigens via increased cross-presentation efficiency. In addition to the overall trends we observed between IgG and CD8+ responses, we were struck that minimal peptides from Wdr33:H13Y – a mutation site with strong preexisting IgG signals in naïve animals – were recognized regularly in vaccinated animals by IgG and produced large increases in CD8+ T cell recognition of both 4T1 tumor and Wdr33:H13Y peptides. This exemplary case builds on our overall hypothesis that some antigen-specific CD8+ T cell responses to our vaccine are additionally favored by B cell and/or CD4+ recognition of similar peptides. While the mechanism is still uncertain, the observations reported here warrant further investigation into the roles of IgG and other antibody isotypes on future anti-tumor responses. Future work should focus on isotype-specific relationships and include data from clinical samples; this will be essential to determine these nuances of antibody and T cell interrelationships.
In the viral literature, the dependence of future responses on past ones to similar antigens is well documented [43, 44, 45, 46], and there are recent reports suggesting that preexisting immunity not only occurs, but is common; many healthy adults have memory T cells reactive to peptides from viruses they have never encountered . Perhaps such T cell repertoires relate to the observed universe of preexisting IgG autoantibodies common in humans . If true, these IgG antibodies may help point to the types of antigens targeted by – or that evade – cancer immune surveillance. We may find that some patients require not just a release of tolerance or boosting of the natural priming environment by checkpoint blockade, but perhaps vaccines and therapies directed toward those antigens inadvertently avoided by their own unique history of antigen exposure. In the future, it may be possible to determine this antigenic history and correct gaps in immune surveillance with either personalized cancer vaccines or generalized complex vaccines like the autophagosome-enriched vaccine studied here.
This study presents a simultaneous characterization of humoral and cellular immunity to tumor antigens at the level of individual peptides from a vaccine model whose components are currently undergoing human clinical trials [48, 49], and provides rationale for further investigation into the role of preexisting and post-treatment anti-tumor antibodies on anti-tumor CD8+ T cell immunity. Though this work focuses on neoantigens and autoantigens centered at mutation sites, these results may translate to other antigen populations, including non-mutated tumor antigens with ectopic expression patterns and antigens related to non-cancer diseases.
Thank you to Dr. Walter J. Urba, Dr. Eric Tran, Dr. David J. Messenheimer, Dr. Traci Hilton, Dr. Andrew Gunderson, Dr. Hong-Ming Hu, and the current and former members of the Earle A. Chiles Research Institute for their extensive support discussing and revising this manuscript.
Support for these studies was provided by the Chiles Foundation, Robert W. and Elise Franz, Lynn and Jack Loacker, Wes and Nancy Lematta, M.J. Murdock Charitable Trust, the Harder Family, and the Providence Medical Foundation. Support for Tyler W. Hulett additionally provided by OCTRI-OSLER TL1 and the ARCS Foundation – Portland. Mass spectrometric analysis by Larry L. David, Ashok P. Reddy, and Phillip A. Wilmarth performed by the OHSU Proteomics Shared Resource with partial support from NIH core grants P30EY010572, P30CA069533 and shared instrument grant S10OD012246.
Immunologic assays, integrated analysis, text, and figure preparation by TWH. Senior oversight by SMJ and BAF. Support with immunologic assays provided by MEA, SMJ and BAF. Support with multispectral immunohistochemistry by CBM and SMJ. All TMT mass spectrometry and analysis by LLD, APR, and PAW. 4T1 exome analysis and variant calls by CJ. Peptide array normalization and support by CJ. All authors contributed to editing this text. All authors read and approved the final manuscript.
Ethics approval and consent to participate
Recognized principles of laboratory animal care were followed (Guide for the Care and Use of Laboratory Animals, National Research Council, 2011), and all animal protocols were approved by the EACRI Animal Care and Use Committee. All mice were maintained in a specific pathogen–free environment. No humans were treated in this study and therefore ethical approval per Helsinki Declaration was not required.
Consent for publication
All authors provide their consent for publication of this manuscript.
BAF is a co-founder of UbiVac, which has licensed the autophagosome-enriched vaccine intellectual property. The other authors declare no potential conflicts of interest.
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
- 7.Messina JL, Fenstermacher DA, Eschrich S, Qu X, Berglund AE, Lloyd MC, et al. 12-chemokine gene signature identifies lymph node-like structures in melanoma: potential for patient selection for immunotherapy? Sci Rep. 2012;2 https://doi.org/10.1038/srep00765.
- 10.Chen Y-T. The journey from autologous typing to SEREX, NY-ESO-1, and cancer/testis antigens. Cancer Immun Arch. 2012;12:8.Google Scholar
- 11.Pfreundschuh M. The genealogy of SEREX. Cancer Immun Arch. 2012;12:7.Google Scholar
- 20.Jäger E, Chen Y-T, Drijfhout JW, Karbach J, Ringhoffer M, Jäger D, et al. Simultaneous humoral and cellular immune response against cancer–testis antigen NY-ESO-1: definition of human histocompatibility leukocyte antigen (HLA)-A2–binding peptide epitopes. J Exp Med. 1998;187:265–70.CrossRefPubMedPubMedCentralGoogle Scholar
- 24.Yu G, Li Y, Cui Z, Morris NP, Weinberg AD, Fox BA, et al. Combinational immunotherapy with Allo-DRibble vaccines and anti-OX40 co-stimulation leads to generation of cross-reactive effector T cells and tumor regression. Sci Rep. 2016;6 https://doi.org/10.1038/srep37558.
- 25.Page DB, Hulett TW, Hilton TL, Hu H-M, Urba WJ, Fox BA. Glimpse into the future: harnessing autophagy to promote anti-tumor immunity with the DRibbles vaccine. J Immunother Cancer. 2016;4 https://doi.org/10.1186/s40425-016-0130-4.
- 28.Castle JC, Loewer M, Boegel S, Tadmor AD, Boisguerin V, de Graaf J, et al. Mutated tumor alleles are expressed according to their DNA frequency. Sci Rep. 2014;4 https://doi.org/10.1038/srep04743.
- 36.Jensen S. Polarization of a type 1 anti-tumor immune response in the context of a strong tumor antigen results in the regression of the mammary adenocarcinoma, 4T1. Scholar Archive. 2003;3143 http://digitalcommons.ohsu.edu/etd/3143.
- 38.Routy B, Le Chatelier E, Derosa L, Duong CPM, Alou MT, Daillère R, et al. Gut microbiome influences efficacy of PD-1–based immunotherapy against epithelial tumors. Science. 2017; https://doi.org/10.1126/science.aan3706.
- 39.Gopalakrishnan V, Spencer CN, Nezi L, Reuben A, Andrews MC, Karpinets TV, et al. Gut microbiome modulates response to anti–PD-1 immunotherapy in melanoma patients. Science. 2017; https://doi.org/10.1126/science.aan4236.
- 42.Srivastava RM, Lee SC, Andrade Filho PA, Lord CA, Jie H-B, Davidson HC, et al. Cetuximab-activated natural killer and dendritic cells collaborate to trigger tumor antigen-specific T-cell immunity in head and neck cancer patients. Clin Cancer Res. 2013;19:1858–72.CrossRefPubMedPubMedCentralGoogle Scholar
- 48.Search of: Dribbles - List Results - ClinicalTrials.gov. https://clinicaltrials.gov/ct2/results?term=Dribbles&Search=Search. Accessed 19 May 2017.
- 49.Search of: poly I:C - List Results - ClinicalTrials.gov. https://clinicaltrials.gov/ct2/results?term=poly+I%3AC&Search=Search. Accessed 19 May 2017.
Open AccessThis article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.