Potent and conditional redirected T cell killing of tumor cells using Half DVD-Ig
- 1.6k Downloads
Novel biologics that redirect cytotoxic T lymphocytes (CTLs) to kill tumor cells bearing a tumor associated antigen hold great promise in the clinic. However, the ability to safely and potently target CD3 on CTL toward tumor associated antigens (TAA) expressed on tumor cells remains a challenge of both technology and biology. Herein we describe the use of a Half DVD-Ig format that can redirect CTL to kill tumor cells. Notably, Half DVD-Ig molecules that are monovalent for each specificity demonstrated reduced non-specific CTL activation and conditional CTL activation upon binding to TAA compared to intact tetravalent DVD-Ig molecules that are bivalent for each specificity, while maintaining good drug like properties and appropriate PK properties.
KeywordsDVD-Ig Half DVD-Ig Halfbody epidermal growth factor receptor redirected T-cell cytotoxicity rCTL
One of the most promising clinical uses of bispecific antibody technologies is the ability of a molecule with two antigen binding specificities to redirect immune effector cells to kill tumor cells. It was first described over 25 years ago that a bispecific antibody simultaneously targeting T cells and a tumor cell surface antigen can kill target tumor cells (Staerz and Bevan, 1989). The most successful clinical programs have been the targeting of different tumor antigens (e.g., CD19 or EpCAM) by anti-CD3 redirected T cells (Baeuerle and Reinhardt, 2009).
However, an agonistic antibody can carry the danger of unwanted side-effects, with a particularly stark example of the anti-CD28 agonistic mAb TGN-1412 (Horvath et al., 2012). This antibody was designed to stimulate T cell immunity for the treatment of cancer, but severe adverse events due to activation of T cells and induction of a cytokine storm were seen in healthy volunteers. Similar safety concerns may delay or hinder development of bispecific antibodies where an anti-CD3 binding domain is paired with antibody domains against a tumor-associated antigen for redirected tumor killing by cytotoxic T cells. The TGN-1412 incident highlighted the potential risk of developing agonist antibodies to the T-cell receptor or members of the co-stimulatory receptor family. Therefore, identifying therapeutic mAbs that safely target these cell surface receptors with desired features remains a challenging task.
Blinatumomab (Blincyto®), an FDA-approved bispecific molecule composed of an anti-CD3 single chain variable fragment (scFv) and an anti-CD19 scFv connected by a polypeptide linker, also known as bispecific T cell engager (BiTE), showed efficacy in clinical trials for B-cell non-Hodgkin’s lymphoma and B-precursor acute lymphocytic leukemia (ALL) with manageable side effects (Nagorsen and Baeuerle, 2011). Blinatumomab is monovalent for both CD3 and CD19, which may help to avoid unwanted CD3 signaling induced by receptor dimerization or clustering, or allow for conditional CD3 clustering upon binding to the target antigen, CD19. Blinatumomab represents the first-in-class BiTE antibody approved for clinical use and provides a novel therapeutic option for patients with refractory or relapsed B cell ALL (Benjamin and Stein, 2016). One concern often being raised for BiTE molecules is the requirement for continuous intravenous (IV) infusion dosing because of the small molecular weight and rapid clearance of the molecule from circulation. Recently, a tetravalent bispecific tandem diabody (TandAb) format was constructed to address these issues. The lead molecule of TandAb format, AFM11, has two binding sites for CD3 and two for CD19, and is currently in Phase I clinical development (Reusch et al., 2015). AFM11 elicited more potent in vitro B cell lysis than a BiTE molecule targeting the same antigens (Molhoj et al., 2007) . In addition, TandAb with the molecular weight at approximately 100 kDa, was reported to have a half-life ranging from 18 to 23 h after IV administration in mice (Reusch et al., 2015). AFM11 therefore can be administered weekly or twice weekly in humans. However, the safety and efficacy profile of AFM11, which is bivalent for CD3 binding, is still to be determined in clinical trials.
Recently we explored the construction of a series of ‘Halfbodies’, where full-length IgG molecules are split into two equal half molecules, by structural modeling assisted mutagenesis at the antibody CH3-CH3 interface. The amino acid residues that are important for antibody CH3 dimerization were first described by Carter and colleagues (Dall’Acqua et al., 1998). This structure-activity relationship study of antibody CH3 dimers revealed that certain residues, such as T366, L368, P395, F405, Y407 and K409, played an important role in maintaining the stability of the CH3 dimers. Two separate groups have previously reported that an IgG could be converted to a monomeric IgG by introducing mutations at residues 351, 366, 368, 395, 405, 407, 409 (Ying et al., 2012), or by introducing two Asn-linked glycans at positions 364 and 407 (Ishino et al., 2013). Although previously reported formats of mAb-based monovalent-binding molecules did improve the efficacy in targeting specific cell surface targets, further development of these formats have been hindered likely due to poor manufacturing and physiochemical properties of the antibodies (Cheadle, 2006; Filpula, 2007).
We discovered that a single point mutation in the CH3 domain and two mutations at cysteine residues within the IgG hinge region could result in Halfbodies as well, similar to the ones generated with multiple mutations at the CH3 domain (Table S1 and Fig. S1). With the Halfbody format, we demonstrated the conversion of agonistic or partial antagonist molecules to pure antagonists against the cell surface target (Table S2). In addition, we extended the Halfbody technology to DVD-Ig format to generate Half DVD-Ig molecules for monovalent CD3 binding. The monovalent CD3 containing Half DVD-Ig maintained the ability to bind CD3 but was conditional with regard to their ability to initiate immune synapse formation and mediate T cell activation upon binding to tumor-associated antigen which greatly reduced non-specific cytokine release for CD3-mediated T cell redirected cytotoxicity in vitro. Additionally, the half DVD-Ig molecule has good expression in mammalian cells, can be purified to homogeneity, and has improved half-life compared to other monovalent antibody variants (e.g., scFv, BiTE, and Fab). Therefore, we believe that Half DVD-Ig molecular format may provide an additional option for exploring cancer treatment using redirected T cell cytotoxicity with more practical dosing regimens.
Bispecific DVD-Ig molecules support T cell redirected killing
Construction, expression, and biochemical characterization of Half DVD-Ig molecules
Previously we demonstrated the generation of stable Half Ig molecules by introducing P395A, F405R, Y407R, and K409D mutations, at the CH3 domain to disrupt CH3 homodimerization and C226S and C229S mutations, at hinge region, to eliminate the two inter-chain disulfide bonds of the immunoglobulin molecules (Table S1). We further demonstrated that a single point mutation in the CH3 domain and two mutations at cysteine residues within the IgG hinge region could result in Halfbodies as stable as the ones with multiple mutations at CH3 domain (Table S1 and Fig. S1). With the Halfbody format, we achieved conversion of a cMet agonistic (or partial antagonist) antibody, to a pure antagonist. This was demonstrated in a cMet phosphorylation assay, where unlike the parental antibody, the Halfbodies neutralized HGF-induced cMet phosphorylation. Furthermore the halfbodies, but not the parental anti-cMet antibody, inhibited tumor cell proliferation (Table S2).
Anti-EGFR/Anti-CD3 Half DVD-Ig molecules for T cell redirected cytotoxicity
One potential issue with utilizing full length antibody constructs with Fc domains is the interaction of anti-(tumor associated antigen) molecules/anti-CD3 with Fc-receptors and the potential for cytokine release by non-specific T cell activation. In addition, bivalent CD3 binding by a full-length antibody might cross-link CD3 and have T cell mitogenic activity without the engagement of tumor antigens. Together, these T cell activation mechanisms can result in cytokine release syndrome (CRS), sometimes with potentially serious adverse events (Sathish et al., 2013).
Pharmacokinetic properties of Halfbodies
Pharmacokinetic profile of Half DVD-Ig compared to DVD-Ig molecules in Sprague Dawley rats
Anti-EGFR/anti-CD3 DVD-Ig protein
Anti-EGFR/anti-CD3 Half-DVD-Ig protein
One of the most promising clinical applications of bispecific technologies is the redirection of immune effector cells to kill tumor cells. A good review of the clinical history of immune cell rCTL mediated by various bispecific formats has been published by Lum and Thakur (2011). The first successful clinical program using redirected T cells was accomplished by Fresenius Biotech and Trion Pharma’s Catumaxomab (Removab®), which was approved in the EU in 2009 for malignant ascites. Catumaxomab was made of a rat/mouse quadroma which required devising a purification scheme to isolate pure heteroconjugates with specificity to CD3 on one arm and EpCAM on the other. Notably, catumaxomab was reported to have intact Fc effector function (Chelius et al., 2010). Since this molecule contains rat and mouse proteins, developing anti-drug antibody response in the clinic limited broader applicability of the approach. A major breakthrough in the field occurred when Micromet was able to demonstrate the efficacy and safety of blinatumomab as a B cell lymphoma therapy (Baeuerle and Reinhardt, 2009). Blinatumomab is a scFv-based bispecific T-cell engager (BiTE) molecule containing anti-CD3 and anti-CD19 domains. BiTE molecules have short half-lives, so patients must undergo continuous infusion by micropumps. BiTE molecules have monovalent binding domains and do not engage Fc receptors since they do not have an Fc domain. Therefore, BiTE molecules are thought to minimize the risk of CRS (Loffler et al., 2000).
The Halfbody format also has the potential to minimize the risk of a cytokine storm by not triggering non-specific T cell activation. It is well understood that crosslinking of the T cell receptor with anti-CD3 antibodies causes rapid T cell activation and can lead to serious clinical adverse events (Sgro, 1995). Bluestone and colleagues attempted to mitigate CRS for the treatment of type-I diabetes by making a chimeric antibody with an OKT3 variable domain and a huIgG1-mutant (L234A, L235A) constant region to prevent FcγR binding (Chatenoud and Bluestone, 2007). Although the in vitro lack of cytokine release looked promising, this result did not translate clinically because CRS was observed in patients, even though it was less severe than the parental wildtype OKT3 (Kaufman and Herold, 2009). We have shown that a half anti-EGFR/anti-CD3 DVD-Ig protein can support sub-nmol/L rCTL activation while minimizing the risk of cytokine release in our in vitro assay. It is still unclear whether high levels of cytokine production may be necessary for efficacy of an rCTL agent and what the clinically tolerable level of cytokine release might be (Lum and Thakur, 2011). Further studies are needed to understand if our results would translate clinically. The balance of safety and potency must be managed, and the Halfbody format provides a promising option that could be used for appropriate cancer indications.
Although the mean serum half-life of Halfbodies, estimated to be approximately 31 h, is much shorter than that of full-length antibodies, it is still a significant improvement over half-life of BiTE, Fab or scFv molecules (Chapman et al., 1999; Cheadle, 2006). This could be due to the fact that the molecular size of Halfbodies is larger than the threshold of renal clearance of proteins (McAleese and Eser, 2012). TandAb, a tetra-valent bispecific scFv antibody format with molecular size of ~110 kDa, is reported to have a serum half-life similar to Halfbodies reported here. With the estimated 12–24 h half-life in cynomolgus monkey, AFM13, an anti-CD16A X anti-CD30 TandAb, was enabled for a Phase I trial with weekly infusion dosing and the clinical PK results were reported to be promising (Rothe et al., 2015). If our observed rat PK correlates with human PK, we hypothesize that our Halfbody format would show a similar improvement in PK properties over scFv molecules. With advantages of monovalent-targeting, reduced Fc function, ease of expression and purification, preferred drug-like properties, and superior profile in serum half-life over other monovalent antibody fragments, Halfbody molecules may offer a promising novel therapeutic option.
Materials and Methods
Generation and characterization of Half DVD-Ig molecules
The cDNA sequences of anti-cMet variable domains were cloned from mouse hybridoma HB-11895 (5D5.11.6) (ATCC, Manassas, VA) with a mouse Ig primer set (Novagen, Cat # 69831-3). For cMet Halfbody molecules, mutations such as C226S, C229S, T366F, L368F, P395A, F405R, Y407R, K409D were introduced into the heavy chain hinge region using standard mutagenesis techniques. The variable domain sequences of the anti-CD3 antibody were derived from the OKT3 antibody and those of the anti-EGFR antibody were derived from cetuximab. Inner and outer domains were connected using a long-long linker (L-L): VH1-VH2 Linker, ASTKGPSVFPLAP (13a.a); VL1-VL2 Linker, TVAAPSVFIFPP (12a.a.). AbbVie-proprietary pHybE vectors with corresponding huIgG1 heavy chain and kappa light chain were used to transiently transfect 293-6E cells in Freestyle 293 medium (Invitrogen, Carlsbad, CA). Media containing the secreted protein was harvested 6–7 days post transfection and purified using protein A chromatography (Invitrogen, Carlsbad, CA) according to the manufacturer’s instructions. The proteins were then dialyzed into PBS buffer, pH 7.2 (Invitrogen, Carlsbad, CA). The monomer percentage of Half DVD-Ig proteins was analyzed by SDS-PAGE and size exclusion chromatography. Binding of anti-EGFR/anti-CD3 DVD-Ig and half DVD-Ig molecules to CD3 and EGFR was analyzed by FACS.
In vivo redirected cytotoxicity tumor models
Immune compromised female SCID mice were purchased from Charles Rivers (Wilmington, MA). Animals were acclimated to the animal facilities for a period of at least one week prior to commencement of experiments. Animals were kept in the light phase of a 12 h light:12 h dark schedule (lights on 0600 h). All experiments were conducted in compliance with AbbVie’s Institutional Animal Care and Use Committee and the National Institutes of Health Guide for Care and Use of Laboratory Animals guidelines in a facility accredited by the Association for the Assessment and Accreditation of Laboratory Animal Care. A431 cells were obtained from the Ludwigshafen Institute, cells were grown between passage 3–6 in vitro in RPMI 1640 media with 10% FBS. Donor T-cells were purchased from Astarte Biologicals, Bothell, WA. T-cells were positive selected and counted using a flow cytometer by Astarte Biologicals. T-cells were brought up from frozen in RPMI 1640, 10% FBS, 2 mmol/L L-glutamine, 55 mmol/L 2-BME and 30 IU/mL recombinant human IL-2, and maintained in the incubator for 24 h before use. Tumor and T-cell viability was assessed by trypan blue exclusion and later mixed in a 50 mL conical tube prior to inoculation, achieving a 1:1 effector:tumor ratio (E:T). A total of 1 × 106 viable A431 cells and/or T-cells were inoculated subcutaneously into the right flank of female SCID mice on day 0. The total injection volume was 0.1 mL and was composed of a 1:1 mixture of S-MEM and matrigel (Corning, MA). The A431 tumor model provided 100% tumor take and robust tumor growth, making it feasible for evaluation as an early start tumor model. In the early start tumor model, dosing of reagents started 1 h post tumor inoculation. Tumor measurements were collected twice a week and animals were removed from the study once tumors reached ≤1,500 mm3.
In vitro redirected cytotoxicity assay
Raji, A431 or N87 (ATCC, Manassas, VA) target (T) cells were allowed to adhere to ACEA xCELLigence 96-well plates (ACEA Bio, San Diego, CA) overnight and cell indexes are periodically monitored. Cell index is a measurement of electrical impedance and reflects the viability of adherent cells. Human PBMC effector cells (E) were then plated at an E:T ratio of 10:1. The testing samples were appropriately diluted to obtain concentration-dependent titration curves. The cell indexes of targets in the DVD-Ig protein treated samples were divided by the cell indexes of control targets (no treatment) to calculate percent specific lysis. The data was graphed and EC50s were calculated in Prism (Graphpad Software, San Diego, CA). T cell activation was monitored by harvesting cells from 96 well plates and subjected to FACS analysis (FACSCanto II, BD Biosciences, San Jose, CA) after staining with anti-CD4, CD8, CD25, and CD69 fluorescently-labeled antibodies (eBioscience, San Diego, CA).
Cytokine release huPBMC assay
A similar assay has been previously described (Finco et al., 2014). Briefly, test antibodies were coated onto a 96-well polypropylene plate at 1 μg/mL for 90 min at 37°C. The plate was rinsed and human PBMC were added at 105 cells/well. After 48 h of culture, supernatants were harvested and assayed for cytokines using commercially available kits (Meso Scale Discovery, Rockville, MD).
Visualization of immune synapse formation
Purified CD3-positive T cells were stimulated with anti-CD3 plus anti-CD28 for three days to generate T cell blasts that express stored granzyme-B. T cell blasts were subsequently co-cultured with A431 cells in the presence or absence of Half DVD-Ig for 4–6 h to allow for immune synapse to form. Cells were fixed, washed, permeabilized, and stained with fluorescent antibodies against EGFR and granzyme-B. DAPI was used to visualize the nucleus. Conjugates were visualized on Amnis ImageStream by gating CD3 and EGFR doublets. Representative image overlays were cropped together to illustrate the various stages of synapse formation.
Jurkat reporter assays
Jurkat-NFAT luciferase reporter cells (Promega, Madison, WI) that express luciferase under the control of an NFAT transcription factor binding site were stimulated with 5 nmol/L Half DVD-Ig or 5 nmol/L DVD-Ig in the presence or absence of A431 tumor cells that express an abundance of the tumor-associated antigen, EGFR. After 4–6 h, the co-cultures were lysed with SteadyGlo luciferase substrate (Promega) and relative luciferase activity was read using a luminometer.
Pharmacokinetics in rats
Pharmacokinetic (PK) properties of Halfbodies were determined in Sprague-Dawley rats purchased from Charles River Laboratories (Wilmington, MA). Male rats were dosed intravenously with a single dose of 4 mg/kg of test Halfbody molecules. Blood samples for serum concentration analysis were collected via the tail vein in serum separator tubes at various time points. The serum samples were analyzed on a MSD Sector Imager 6000 (Meso Scale Discovery, Gaithersburg, MD) in a chemiluminescent MSD method using biotinylated cMet for capture and sulfo-tagged anti-human IgG Kappa for detection. Sample concentrations were calculated by XLfit4 software version 2.2.1 Build 16 (Microsoft Corporation, Redmond, WA). PK parameters were calculated using Winonlin software version 5.0.1 (Pharsight Corporation, Mountain View, CA). All rodents were obtained from vendors and were handled under procedures approved by the AbbVie Institutional Animal Care and Use Committee Association, which is accredited by the Association for Assessment and Accreditation of Laboratory Animal Care.
We thank Mark Anderson for c-Met ELISA binding, Qian Zhang for c-Met phosphorylation and Lora Tucker for proliferation assays.
ALL, acute lymphocytic leukemia; BiTE, bispecific T cell engager; CRS, cytokine release syndrome; CTLs, cytotoxic T lymphocytes; IV, intravenous; PBMCs, peripheral blood mononuclear cells; scFv, single chain variable fragment; TAA, tumor associated antigens; TandAb, tandem diabody
COMPETING FINANCIAL INTERESTS
The authors are employees or former employees (J. Liu, Jieyi Wang and P. Bardwell) of North Chicago, IL-based AbbVie Inc. and may own AbbVie stocks or stock options. The authors have no other relevant affiliations or financial involvement with any other organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript apart from those disclosed. The design, study conduct, and financial support for this research were provided by AbbVie. AbbVie participated in the interpretation of data, review, and approval of the publication.
COMPLIANCE WITH ETHICS GUIDELINES
Matthew M. Staron, Qingfeng Tao, Susanne Scesney, Gail Bukofzer, Luis E Rodriguez, Chee-Ho Choi, Jennifer Wang, Qing Chang, Feng Dong, Cherrie Donawho, Christine M. Grinnell, Edit Tarcsa, Charles Hutchins, Tariq Ghayur, Jijie Gu are employees of AbbVie Inc. and may own AbbVie stocks or stock options. Phillip D. Bardwell, Junjian Liu, and Jieyi Wang were employees of AbbVie at the time of the study. The authors have no other relevant affiliations or financial involvement with any other organization or entity with a financial interest or conflict with the subject matter or materials discussed in the current manuscript. The design, study conduct, and financial support for this research were provided by AbbVie. All institutional and national guidelines for the care and use of laboratory animals were followed. AbbVie participated in the collection, analysis and interpretation of the data, review, and approval of the publication.
- Lutterbuese R, Raum T, Kischel R, Hoffmann P, Mangold S, Rattel B, Friedrich M, Thomas O, Lorenczewski G, Rau D et al (2010) T cell-engaging BiTE antibodies specific for EGFR potently eliminate KRAS- and BRAF-mutated colorectal cancer cells. Proc Natl Acad Sci 107:12605–12610CrossRefPubMedPubMedCentralGoogle Scholar
- Reusch U, Duell J, Ellwanger K, Herbrecht C, Knackmuss SH, Fucek I, Eser M, McAleese F, Molkenthin V, Gall FL et al (2015) A tetravalent bispecific TandAb (CD19/CD3), AFM11, efficiently recruits T cells for the potent lysis of CD19(+) tumor cells. MAbs 7:584–604CrossRefPubMedPubMedCentralGoogle Scholar
- Rothe A, Sasse S, Topp MS, Eichenauer DA, Hummel H, Reiners KS, Dietlein M, Kuhnert G, Kessler J, Buerkle C et al (2015) A phase 1 study of the bispecific anti-CD30/CD16A antibody construct AFM13 in patients with relapsed or refractory Hodgkin lymphoma. Blood 125:4024–4031CrossRefPubMedPubMedCentralGoogle Scholar
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.