Final Report Summary - IACT (Immunostimulatory Agonist antibodies for Cancer Therapy)
Executive Summary:
Cancer immunotherapy is beginning to realize its potential in the clinic, but immune checkpoint blockade (ICB) is able to elicit durable responses in only a fraction of patients with immunogenic cancers such as melanoma and lung cancer. Significant sub-groups of patients with these indications do not respond to ICB, and the same applies to patients with other cancer types. A major hurdle in this respect is the lower intrinsic immunogenicity of ICB-resistant tumors. In order to build on the first promising results of ICB, there is a clear need for the exploration of additional drugs and treatment regimens in clinical trials. Agonist immunostimulatory antibodies (IS-Abs) targeting activatory receptors on immune cells are a potential alternative for immune checkpoint inhibitors, in that these Abs have been demonstrated to turn ‘cold’ tumors into ‘hot’ tumors, with notable therapeutic results in pre-clinical models. The aim of the IACT project was to boost the further development of agonist IS-Abs, in particular IS-Abs targeting the activatory immune cell receptors CD40, OX40 and 4-1BB.
A main focus of the IACT project was the clinical development of the anti-human CD40 Ab ChiLob7/4 for which new GMP producer lines were generated that were used for manufacturing of a clinical batch, which is ready to be tested in two clinical indications: head and neck squamous cell carcinoma (HNSCC) and pancreatic ductal adenocarcinoma (PDA). Testing in HNSCC was initially planned in combination with a synthetic RNA vaccine against the E6 and E7 oncoproteins of human papillomavirus type 16. However, pre-clinical and subsequent clinical studies showed that the newly developed liposomal formulation of this vaccine (RNALIP) was already very potent and unlikely to be significantly complemented by anti-CD40 Ab. As a result, clinical testing of RNALIP-HPV16 and ChiLob7/4 is proceeding in parallel, with the aim of systematically comparing the pro-immunogenic impact of these drugs. ChiLob7/4 will furthermore be tested in the neo-adjuvant setting in patients with primary resectable PDA, in view of the fact that this offers a unique opportunity to measure drug impact on tumor microenvironment and intratumoral T-cell immune response. Pre-clinical experiments in newly developed mouse models for PDA demonstrated that anti-CD40 Ab treatment results in profound changes in tumor immune cell infiltrate and immune gene signatures. These findings serve as a template for the evaluation of pharmacodynamic biomarkers in the neo-adjuvant trial.
The second major objective of the IACT project was the development of optimized IS-Abs. A systematic evaluation of anti-human CD40 Abs showed that the immunostimulatory potency is not only dependent on the Fab domain, but also on the Fc-domain. Proper choice of Fc-domain can result in more potent agonist Abs. In the context of IACT, two enhanced variants of ChiLob7/4 were developed: a glyco-optimized®, hypo-fucosylated version, the enhancement of which involves increased secondary cross-linking through Fc-receptor positive bystander immune cells, and a human-IgG2-variant, the immuno-stimulatory action of which is independent of secondary cross-linking. The clinical development of a set of already available anti-human OX40 and anti-human 4-1BB Abs was abandoned in favor of the development of superior Abs that were newly isolated from synthetic human Ab libraries. From the same libraries, a unique set of dual-specific Abs was isolated, the Fab domains of which are capable of targeting any combination of two of the three immunostimulatory receptors concerned, thereby providing the first proof of concept for dual agonist Abs. Using the trimerbody® technology, an optimized, bi-specific anti-4-1BB Ab was generated, the agonist activity of which is targeted to the tumor-microenvironment through recognition of a clinically validated tumor antigen.
Project Context and Objectives:
Mission statement
Every year, more than 3 million Europeans are diagnosed with cancer, pointing to the continued importance of research into cancer prevention, diagnosis, treatment and patient care. Surgical resection still offers the best curative option, but in many cases cancer cells have already spread before the primary tumour is removed. Radiation and cytotoxic drugs offer treatments with the potential to ‘mop-up’ cells which escape surgery, but these are ‘blunt’ instruments of destruction which lack specificity and hence provide varying degrees of success. This leaves a desperate need for alternative, conceptually distinct, treatment approaches. Immunotherapy now offers such an alternative, with a realistic expectation of success in a wide range of aggressive cancers. In the last decade numerous approaches have shown that the immune system can be ‘trained’ and strengthened to both engage and destroy malignant cells. Furthermore, such responses are durable with the potential to provide life-long protection. Inevitably, these breakthroughs in understanding of anti-cancer immunity have resulted in the clinical testing of exciting new immunotherapy drugs, including some which are now approved for clinical use. However, despite pioneering work by European leading scientists, the USA currently dominates this field. Our IACT concerted action presents a unique opportunity for European academia and industry to gain the ascendency and be at the forefront of delivering effective immune-based cancer treatments. IACT will focus on the clinical translation of a drug concept shown to be highly promising in pre-clinical studies, for which the IACT consortium has achieved a highly competitive position: agonist immunostimulatory antibodies (IS-Abs).
Scientific background of therapeutic concept
a. Immunotherapy of cancer
The immune system not only protects us against infectious pathogens, but also against cancer. Unfortunately, this protection is not perfect and immune failure can lead to disease. Since the development of the small pox vaccine in the 18th century, we have learned to harness the immune defence against pathogens and thereby to control many infectious diseases. Until recently, the clinical efficacy of immune intervention strategies against cancer was low, mainly because the immunological difference between tumours and normal tissue is much more subtle than between pathogen-infected and healthy cells. Consequently, the development of effective immunotherapeutic strategies against cancer requires comprehensive insight into the fine-tuning of the immune system. Importantly, critical advances in this field have resulted in the FDA-approval of two cancer immunotherapy drugs: a cell-based vaccine for use in metastatic prostate cancer (Sipuleucel-T) and an immunostimulatory antibody (IS-Ab) for use in metastatic melanoma (Ipilimumab). Both drugs enhance the patient’s T-cell immune response against cancer. This success confirms the role of the immune system in controlling cancer and encourages further development of cancer immunotherapy drugs to improve on the results seen to date.
b. Antibody-based immunotherapeutics as compared to other approaches
Monoclonal antibodies (mAbs) constitute a highly attractive drug format, as illustrated by the successful clinical development of the cancer-targeting blockbuster drugs Rituximab, Trastuzumab/Herceptin and Bevacizumab/Avastin. Compared to other immunotherapeutic strategies, IS-Abs have the following major advantages:
• Manufacturing is well established, resulting in molecularly defined drugs that are stable, display excellent pharmacokinetics (PK) and can be modified by a wide array of technologies.
• IS-Abs have a broader application than cancer vaccines and adoptive cell therapy, the use of which is generally limited to patients with cancers displaying pre-identified antigens. Moreover, effective IS-Abs may be used in combination with these antigen-specific approaches to enhance their efficacy.
c. Immunostimulatory antibodies (IS-Abs) as cancer therapeutics
The immune response is regulated through cell surface receptors that stimulate or inhibit the activity of immune cells. While activating signals enable immune responses against harmful pathogens, inhibitory signals prevent uncontrolled immune reactivity against harmless substances (as in allergies) or normal somatic tissues (auto-immunity). Triggering the inhibitory receptors is a major mechanism by which cancer switches off the immune system. Pre-clinical experiments demonstrated that this inhibition can be reversed in two conceptually distinct ways. Antagonist IS-Abs such as Ipilimumab can be used to unleash immunity by blocking the inhibitory receptors (inhibiting the inhibitor). Alternatively, agonist IS-Abs jump-start immune cells by triggering activatory receptors (stimulating the stimulator). Both types of IS-Abs were shown to induce marked tumour regression in mice and to provide long lasting immunity against tumour recurrence [1-16]. While antagonist IS-Abs are extensively being tested in human patients, the clinical potential of agonist IS-Abs remains to be explored.
d. The importance of developing agonist IS-Abs in addition to antagonist IS-Abs
There is significant attraction to make agonist IS-Abs available for patient treatment in addition to the antagonist IS-Abs that are already in the clinic, first of all because there is ample evidence from pre-clinical tumour models that this approach can be at least as efficacious. In line with the notion that these IS-Ab classes represent two conceptually distinct ways of mobilizing immunity, the efficacy of agonist and antagonist IS-Abs differs for the mouse models studied. Furthermore, analyses of human tumours revealed that different mechanism of immune suppression underlie immune failure in cancer (sub-)types, pointing at the need for multiple treatment options/drugs to overcome immune failure in human cancer (personalized medicine). This need is supported by the notion that only a fraction of the melanoma patients respond to treatment with Ipilimumab (anti-CTLA-4), while clinical results for these antagonists, often referred to as ‘checkpoint blockers’ (also including anti-PD-L1), differ between individual patients and cancer types. Finally, combining agonist and antagonist IS-Abs in mice often work synergistically, pointing at the merit of exploring this approach in the clinic [5,12,16-21].
In spite of this strong rationale, clinical development of agonist IS-Abs has been stalled for several years. This is mainly due to an incident in the U.K. in 2006, in which administration of a highly potent agonist Ab against the immunostimulatory receptor CD28 (TGN1412) to healthy volunteers resulted in a life threatening cytokine-release syndrome, due to the uncontrolled activation of immune cells [22,23]. Importantly, it is now generally agreed that this clinical study was not conducted in an appropriate manner, while the underlying pre-clinical studies offered insufficient conceptual basis for the intended aim of suppressing – rather than enhancing – immunity by an agonist IS-Ab.
With resolution of these concerns, the clinical development of agonist IS-Abs as cancer therapeutics is currently regaining momentum, as illustrated by recent activity of USA industry and academia in this area, in particular with respect to anti-CD40 IS-Abs. Notably, pioneering pre-clinical work has been performed in Europe, including that by partners 1, 3, 4 and 6 in our consortium [2-6,9,11,13,24,25]. Moreover, partner 3 has recently concluded a successful phase I study with an agonist anti-CD40 IS-Ab (ChiLob7/4) [26] and, with support from Cancer Research UK, is also developing lead agonist Abs targeting the stimulatory receptors OX40 and 4-1BB. The availability of these clinical antibodies to the IACT consortium, together with the long-standing expertise of our team in this area of research, create a unique opportunity for European academia and industry to bring this promising class of drugs into the clinic and thereby develop new treatment options for cancer patients.
e. Selection of immunostimulatory pathways for targeting with agonist IS-Abs
The reactivity of immune cells is regulated by a great diversity of cell surface receptors [7,17,18]. We identified CD40, OX40 (CD134) and 4-1BB (CD137), all members of the tumour necrosis factor-receptor (TNFR) family, as the most interesting targets for agonist IS-Ab based on the following criteria:
• efficacy: striking pre-clinical evidence for therapeutic impact against tumours
• safety: the expectation that immune related adverse events will be manageable
• complementarity: enhancement of anti-cancer immunity through distinct mechanisms
The mechanism of action of anti-CD40 Abs primarily involves direct stimulation of innate myeloid cells and indirect enhancement of T-cell immunity through activation of dendritic cells [1,4,6,25,27], which makes this pathway complementary to direct T-cell stimulation by agonist IS-Abs against OX40 and 4-1BB. Anti-4-1BB Abs primarily stimulate CD8+ T-killer cells, whereas anti-OX40 Abs primarily potentiate the CD4+ T-helper response and were found to suppress CD4+ T-regulatory cells [3,8,9,11,13,15,16,28-30]. Importantly, T-cell stimulation through OX40 and 4-1BB is less potent than that mediated by the major co-stimulatory receptor CD28, and their triggering in the absence of T-cell receptor stimulation was not found to result in T-cell activation. Furthermore, OX40 and 4-1BB are – unlike CD28 - not constitutively expressed at the T-cell surface. This significantly reduces the risks of serious adverse events due to uncontrolled immune cell activation.
IACT research strategy and objectives
In conclusion, the recent progress in cancer immunotherapy indicates that this approach can offer new treatment options to meet the urgent unmet medical need of cancer patients. Compelling pre-clinical data indicates that agonist IS-Abs constitute a promising drug class in this respect, while lead/candidate agonist IS-Abs targeting three immune receptors are available to the IACT consortium. Moreover, our pre-clinical data point at concrete opportunities for further optimization of IS-Abs (§1.2.7). Taken together, this offers a unique starting point for testing the available drugs in the clinic, while working on the development of a 2nd generation of optimized drugs.
These considerations inspired the three-pronged IACT research strategy shown in Figure 2, which encompasses the following main objectives:
1. To drive further clinical development and testing of the available lead IS-Abs through:
o obtaining clinical proof of concept for the efficacy of anti-CD40 IS-Ab ChiLob7/4 in combination with chemotherapy in pancreatic cancer
o obtaining clinical proof of concept for the efficacy of anti-CD40 IS-Ab ChiLob7/4 with antigen-specific vaccination in human papillomavirus type 16 (HPV16+) head and neck cancer
o providing state-of-the-art anti-OX40 and 4-1BB IS-Abs to be used for first in human testing
2. To optimally support the successful development of IS-Ab drugs through
o the identification of pharmacodynamics (PD) biomarkers that are related to therapeutic efficacy, and the development of validated, harmonized assays for their measurement in the context of clinical trials
o the development of validated, harmonized in vitro and in vivo models to guide the pre-clinical development of IS-Ab drugs and their best possible application in clinical trials
3. To set up a pipeline for the development of 2nd-generation IS-Abs with improved therapeutic index along four innovative approaches that are based on the latest insights in IS-Ab function as well as on proprietary technologies of our SME partners.
References:
1. Beatty GL, Chiorean EG, Fishman MP, Saboury B, Teitelbaum UR, Sun W, Huhn RD, Song W, Li D, Sharp LL, et al.: CD40 agonists alter tumor stroma and show efficacy against pancreatic carcinoma in mice and humans. Science 2011, 331:1612-1616.
2. Diehl L, den Boer AT, Schoenberger SP, van der Voort EI, Schumacher TN, Melief CJ, Offringa R, Toes RE: CD40 activation in vivo overcomes peptide-induced peripheral cytotoxic T-lymphocyte tolerance and augments anti-tumor vaccine efficacy. Nat Med 1999, 5:774-779.
3. Diehl L, van Mierlo GJ, den Boer AT, van der Voort E, Fransen M, van Bostelen L, Krimpenfort P, Melief CJ, Mittler R, Toes RE, et al.: In vivo triggering through 4-1BB enables Th-independent priming of CTL in the presence of an intact CD28 costimulatory pathway. J Immunol 2002, 168:3755-3762.
4. French RR, Chan HT, Tutt AL, Glennie MJ: CD40 antibody evokes a cytotoxic T-cell response that eradicates lymphoma and bypasses T-cell help. Nat Med 1999, 5:548-553.
5. Gray JC, French RR, James S, Al-Shamkhani A, Johnson PW, Glennie MJ: Optimising anti-tumour CD8 T-cell responses using combinations of immunomodulatory antibodies. Eur J Immunol 2008, 38:2499-2511.
6. Honeychurch J, Glennie MJ, Johnson PW, Illidge TM: Anti-CD40 monoclonal antibody therapy in combination with irradiation results in a CD8 T-cell-dependent immunity to B-cell lymphoma. Blood 2003, 102:1449-1457.
7. Melero I, Hervas-Stubbs S, Glennie M, Pardoll DM, Chen L: Immunostimulatory monoclonal antibodies for cancer therapy. Nat Rev Cancer 2007, 7:95-106.
8. Melero I, Murillo O, Dubrot J, Hervas-Stubbs S, Perez-Gracia JL: Multi-layered action mechanisms of CD137 (4-1BB)-targeted immunotherapies. Trends Pharmacol Sci 2008, 29:383-390.
9. Melero I, Shuford WW, Newby SA, Aruffo A, Ledbetter JA, Hellstrom KE, Mittler RS, Chen L: Monoclonal antibodies against the 4-1BB T-cell activation molecule eradicate established tumors. Nat Med 1997, 3:682-685.
10. Melief CJ, Van Der Burg SH, Toes RE, Ossendorp F, Offringa R: Effective therapeutic anticancer vaccines based on precision guiding of cytolytic T lymphocytes. Immunol Rev 2002, 188:177-182.
11. Piconese S, Valzasina B, Colombo MP: OX40 triggering blocks suppression by regulatory T cells and facilitates tumor rejection. J Exp Med 2008, 205:825-839.
12. Takeda K, Okumura K, Smyth MJ: Combination antibody-based cancer immunotherapy. Cancer Sci 2007, 98:1297-1302.
13. Valzasina B, Guiducci C, Dislich H, Killeen N, Weinberg AD, Colombo MP: Triggering of OX40 (CD134) on CD4(+)CD25+ T cells blocks their inhibitory activity: a novel regulatory role for OX40 and its comparison with GITR. Blood 2005, 105:2845-2851.
14. Vonderheide RH: Prospect of targeting the CD40 pathway for cancer therapy. Clin Cancer Res 2007, 13:1083-1088.
15. Wang C, Lin GH, McPherson AJ, Watts TH: Immune regulation by 4-1BB and 4-1BBL: complexities and challenges. Immunol Rev 2009, 229:192-215.
16. Weinberg AD, Morris NP, Kovacsovics-Bankowski M, Urba WJ, Curti BD: Science gone translational: the OX40 agonist story. Immunol Rev 2011, 244:218-231.
17. Pardoll DM: The blockade of immune checkpoints in cancer immunotherapy. Nat Rev Cancer 2012, 12:252-264.
18. Peggs KS, Quezada SA, Allison JP: Cancer immunotherapy: co-stimulatory agonists and co-inhibitory antagonists. Clin Exp Immunol 2009, 157:9-19.
19. Postow MA, Callahan MK, Wolchok JD: The antitumor immunity of ipilimumab: (T-cell) memories to last a lifetime? Clin Cancer Res 2012, 18:1821-1823.
20. Robert C, Thomas L, Bondarenko I, O'Day S, M DJ, Garbe C, Lebbe C, Baurain JF, Testori A, Grob JJ, et al.: Ipilimumab plus dacarbazine for previously untreated metastatic melanoma. N Engl J Med 2011, 364:2517-2526.
21. Topalian SL, Hodi FS, Brahmer JR, Gettinger SN, Smith DC, McDermott DF, Powderly JD, Carvajal RD, Sosman JA, Atkins MB, et al.: Safety, activity, and immune correlates of anti-PD-1 antibody in cancer. N Engl J Med 2012, 366:2443-2454.
22. Horvath C, Andrews L, Baumann A, Black L, Blanset D, Cavagnaro J, Hastings KL, Hutto DL, MacLachlan TK, Milton M, et al.: Storm forecasting: additional lessons from the CD28 superagonist TGN1412 trial. Nat Rev Immunol 2012, 12:740; author reply 740.
23. Hunig T: The storm has cleared: lessons from the CD28 superagonist TGN1412 trial. Nat Rev Immunol 2012, 12:317-318.
24. Melero I, Johnston JV, Shufford WW, Mittler RS, Chen L: NK1.1 cells express 4-1BB (CDw137) costimulatory molecule and are required for tumor immunity elicited by anti-4-1BB monoclonal antibodies. Cell Immunol 1998, 190:167-172.
25. Schoenberger SP, Toes RE, van der Voort EI, Offringa R, Melief CJ: T-cell help for cytotoxic T lymphocytes is mediated by CD40-CD40L interactions. Nature 1998, 393:480-483.
26. Johnson PW, Glennie M: manuscript in preparation.
27. Diehl L, Den Boer AT, van der Voort EI, Melief CJ, Offringa R, Toes RE: The role of CD40 in peripheral T cell tolerance and immunity. J Mol Med (Berl) 2000, 78:363-371.
28. Ito T, Wang YH, Duramad O, Hanabuchi S, Perng OA, Gilliet M, Qin FX, Liu YJ: OX40 ligand shuts down IL-10-producing regulatory T cells. Proc Natl Acad Sci U S A 2006, 103:13138-13143.
29. So T, Croft M: Cutting edge: OX40 inhibits TGF-beta- and antigen-driven conversion of naive CD4 T cells into CD25+Foxp3+ T cells. J Immunol 2007, 179:1427-1430.
30. Vu MD, Xiao X, Gao W, Degauque N, Chen M, Kroemer A, Killeen N, Ishii N, Li XC: OX40 costimulation turns off Foxp3+ Tregs. Blood 2007, 110:2501-2510.
Project Results:
Below is a detailed description of the output of the experimental work packages WP2 – WP8.
WORK PACKAGE 2. COMBINATION OF ANTI-CD40 IS-AB WITH GEMCITABINE IN PATIENTS WITH PANCREATIC DUCTAL ADENOCARCINOMA.
Objectives
• Test the primary hypothesis that agonist anti-CD40 IS-Abs are capable of mobilizing strong systemic and intra-tumour anti-cancer immunity, in particular when combined with non-immunoablative chemotherapy
• Reproduce the synergy between gemcitabine and anti-CD40 Abs observed in mouse tumour models
Progress for each task
Task 2.1 Preparation and quality control of a clinical batch of ChiLob7/4 (Partner 10)
Generation of new Ab producer lines
BioNTech (partner 10) took on the task to generate new Ab producer lines, because the original, adherent producer line from Southampton (partner 3) could not be successfully adapted to suspension culture. It was decided to make new producer lines for both the original IgG1 version of this Ab, as well as an IgG2 version in view of greater agonist potency of the latter (see WP7, Task 7.2). After codon optimization of the gene constructs, excellent producer lines were obtained. Cell banking and QC were completed in a timely fashion.
Production of a GMP batch of CD40 antibody
GMP batch manufacturing of ChiLob7/4 could successfully be completed within the project runtime, despite unforeseen hurdles encountered. BioNTech, cooperating IACT partners and the CR-UK Biotherapeutics Development Unity (BDU) mitigated such hurdles by allocating additional resources, so to mitigate delays in the best possible manner. A comprehensive report on this process is provided below.
Manufacturing of ChiLob7/4 Drug Substance at BDU (CR-UK)
The aforementioned producer cell line was then transferred to a contract manufacturing organization (CMO) for process optimization and GMP drug substance manufacturing. The BDU of Cancer Research UK (CR-UK) was selected by BioNTech as subcontractor for the manufacturing services. However, several hurdles, including lengthy contract negotiations with the license giver of the expression vectors used for producer clone generation, were encountered while planning Ab production at the BDU. Despite the prompt and solution-oriented reaction by the consortium, and the commitment of BioNTech and CR-UK to considerable extra investments, the start of ChiLob7/4 manufacturing was thereby delayed by approximately one year. Although a manufacturing agreement between CR-UK and BioNTech could be signed in November 2014, the first available slot for manufacturing of the GMP batch at the BDU was then only in April 2016. Given that after manufacturing, additional time is required for quality control testing, GMP reporting and QP batch certification, the Deliverable 2.1/3.3 was then delayed by 13 months (expected: May 2016 as opposed to April 2015, as originally planned).
A 250 L pilot/engineering run was scheduled for Nov/Dec 2015. Cell expansion and all in-process controls looked very good, until a sudden drop in cell viability was detected. The CMC experts at the BDU together with BioNTech evaluated the impact of this finding and came to the conclusion that the run needed to be abrogated and a repetition had to be scheduled prior to the GMP run. Although the BDU at that time was fully booked, full commitment from all partners and the BDU in particular secured a slot for a repetition run directly in January 2016. The repetition run was successfully completed in March 2016, yielding approx. 500 g of pilot bulk drug substance. Ensured by this promising result, and enabled by the BDU re-scheduling other activities, the GMP run was successfully performed in April 2016 as planned. QC testing was completed in June 2016, and a GMP certificate for the Bulk Drug Substance (BDS) was issued in July 2016. In sum, the issues and hurdles described above resulted in a total delay of 15 months.
ChiLob7/4 Fill & Finish and Quality Control Testing
For ChiLob7/4, it was planned to directly proceed to fill & finish at the BDU without further processing steps, thereby producing the ChiLob7/4 drug product (DP) for clinical use. Just prior to GMP manufacturing of the BDS, however, revised calculations of the number of vials required for both clinical trials indicated that the intended batch size of 1,000 vials might be insufficient to ensure safe drug supply. This was in part due to a re-calculation of doses to be applied in the clinics, and in part due to a more extensive stability testing program set up for both ChiLob7/4 BDS and DP. The latter was initiated in response to the initial notion by the MHRA that ChiLob7/4 manufactured using a novel producer cell line may classify as a new IMP (from a manufacturing perspective) which would require more extensive quality and stability testing due to limited experience with the product. This view was confirmed by the Paul-Ehrlich Institute (PEI), the German national authority, in a scientific advice meeting in July 2016. As the established fill process at the BDU is only validated for fill campaigns of up to 1,000 vials, an alternative fill house had to be identified, qualified, and sub-contracted by BioNTech. BioNTech directly obtained quotes from two different fill houses. Fill & finish could earliest been performed in February 2017. This lack of capacity at the external fill houses alone therefore caused a delay of another 7 months, amounting to 22 months in total. Importantly, filling of >3,700 vials, each containing 2.6 mL of ChiLob7/4 DP formulated at 20 mg/mL, has successfully been performed in February 2017.
Quality control testing was then delayed to June 2017. This was due to the necessity for further method optimization for one release relevant assay (glycosylation analysis). Method optimization was promptly addressed and successfully completed, thereby allowing for complete QC testing of the clinical batch of ChiLob7/4 DP vials in a timely manner. According to the manufacturer, the product passed all release criteria. The necessary documents for GMP batch certification (i.e. the Certificate of Analysis (CoA) and the Certificate of Conformity (CoC) were issued by the manufacturer. With this accomplishment Tasks 2.1 was completed with an overall delay of 27 months.
ChiLob7/4 Labelling and Preparatory Activities for IMPD Submission
In order to start the clinical trials, a number of activities still need to be completed.
1) Completion of the CMC section for the IMPD
2) Labelling of ChiLob7/4 with labels approved by the national competent authorities and set-up of ChiLob7/4 distribution to clinical sites
3) Establishing an investigator-initiated trial (IIT) agreement and a quality assurance agreement (QAA) between BioNTech and Heidelberg (sponsor of IAPACA-NEO trial)
The status of each of these tasks is outlined in the following paragraphs:
1. Completion of the CMC section for the IMPD [still ongoing]
IMPD preparation was started in a timely manner already end of 2016. For the preparation of the IMPD, all activities to address comparability between the old Southampton batch and newly generated batch of ChiLob7/4 have been completed successfully, in close cooperation with the Southampton team and in line with requests by PEI as communicated during the Scientific Advice session. These experiments revealed only slight differences between the two batches, the only significant difference being lower activity of the new batch in an in vitro CDC (complement-dependent cytotoxicity) assay, which is likely to be related to differences in glycosylation as reported during last year’s IACT meeting. Based on these data, the safety profile of the new ChiLob7/4 is expected to be very similar to that of the old batch. In addition, these data support the Sponsors’ position that the new batch of ChiLob7/4 does not constitute a new IMP, thereby significantly decreasing the regulatory risks initially associated with the clinical testing of ChiLob7/4 in two clinical phase I trials (sponsored by partners Heidelberg and Southampton).Furthermore, the lack of binding of ChiLob7/4 to mouse CD40, expected on basis of the fact that this Ab was derived from a mouse monoclonal Ab raised in wild type mice that express mouse CD40 on all relevant tissues, was confirmed. This supported the notion that mice are not a relevant species for further toxicity studies. Although not explicitly requested by PEI, tissue cross-reactivity (TCR) analysis by means of IHC using tissue micro-arrays has been successfully performed at BioNTech, with the aim to provide further evidence for comparability between demonstrate comparability between the old and new ChiLob7/4 preparations. As the data that have been presented to the PEI at the Scientific Advice included such a (TCR) study with the old Southampton batch, it was decided to perform this check pro-actively with the newly generated batch as well to avoid further delays in approval. Despite the new ChiLob7/4 batch being expected not to be classified as a new IMP by regulatory authorities, partners xx and yy, together with the regulatory experts at BioNTech conluded that an ‘In-use compatibility study’ needs to be performed prior to submission (regulatory de-risking). This study is scheduled to be conducted by the BDU in March 2019. The main goal of this study is to test the antibody with the infusion system. According to the latest time lines, the full and comprehensive IMPD will be available for submission in April 2019.
2. Labelling of ChiLob7/4 with labels approved by the national competent authorities and set- up of ChiLob7/4 distribution to clinical sites
Labelling of the primary packaging is usually performed right after filling of the drug product into the vials, and prior to freezing of the DP. However, as the IMP is used for two independent trials which each run in two countries and therefore require labels in different languages, and as the respective clinical study protocols of these trials have not been finalized at the time of filling yet, it has been decided that the vials will be labelled only after approval of the clinical study protocol and therefore approval of the labels by the competent authorities. This minimizes the risk of re-labelling of the vials, which can be technically challenging if the product is stored frozen and may not thaw during re-labelling (as is the case for ChiLob7/4 DP), due to additional requests by the competent authorities, which would delay the provision of the study medication. Another advantage of this strategy is, that the vials can be labelled based on the actual recruitment rate of the trials (clinical sites) which minimizes ChiLob7/4 DP loss. As a consequence of this decision, labelling needs to be performed at the long-term storage temperature (-20±5°C) of ChiLob7/4 DP. To this end, every effort has been undertaken to identify material that is suitable for labelling at this temperature and to select a CMO that provides this service. The respective CMO could successfully be qualified in a GMP-audit by BioNTech in June 2017, and will perform storage, labelling and distribution of ChiLob7/4 DP in compliance with GMP.
3. Establishing an investigator-initiated trial (IIT) agreement and a quality assurance agreement (QAA) between BioNTech and Heidelberg (sponsor of IAPACA trial) [still ongoing]
Prior to start of the clinical trial IAPACA-NEO an IIT-agreement that clearly describes certain tasks, e.g. provision of study medication, which will be delegated to BioNTech by the sponsor, needs to be established. A similar IIT- agreement between BioNTech and Southampton (sponsor of the HARE-40 trial), has already been established for the HARE-40 trial and was used as a basis for the IIT-agreement between Heidelberg and BioNTech. Mature draft documents are available and will be finalized prior to trial start (IAPACA-NEO).
Impact of delays with clinical ChiLob7/4 on project time lines and clinical trial strategy:
In the course of 2016, it became evident that we could not adhere to the initial time schedule for the clinical studies with anti-CD40 Ab ChiLob7/4 (IAPACA trial) as shown in Fig. 1, because of the following 2 main reasons:
• Delay of the production and QC of the clinical batch of ChiLob7/4 Ab
• Information by the PEI that they may consider the new batch of ChiLob as a novel IMP due to subtle biochemical differences with the old batch of ChiLob7/4
In view of this, timing and design of the studies were evaluated and adjusted (see Fig 2), as extensively discussed during the 4th IACT consortium meeting in October 2016 and in the 2nd periodic report. The main changes as compared to the original trial schedule were:
• Start of the IAPACA-1 phase I dose escalation trial in Q1 of Year 5, instead of Q3 of year 2
• An increase of trial duration to 2 years, due to the need for subsequent enrollment of patients
• Start of the IAPACA-2 neo-adjuvant in Q1 of year 7, instead of Q3 of year 3.
Based on the adjusted time schedule, it was decided to apply for a 2-year no-cost extension of the IACT project in order to enable the IACT partners to complete at least the majority of clinical studies planned, including the IAPACA-1 study.
In the meantime, pre-clinical experiments in the lab of the IACT coordinator focused at the pharmacodynamics of agonist anti-CD40 Ab and cytostatic drugs (DoW Task 5.7) revealed that (i) transcriptional profiling of tumor tissue of treated animals constitutes a powerful strategy for studying drug-related PD biomarkers and (ii) that the PD biomarker signature induced by anti-CD40 Abs is much more clear if not combined with cytostatic drugs. These insights led to re-evaluation of the neo-adjuvant study design in the course of 2017, as shown in Fig 3. The main changes were:
• Planning of the IAPACA-2 neo-adjuvant study as a single agent study with ChiLob7/4, including a dose escalation phase to accommodate the fact that PEI will most likely regard this drug as novel IMP
• Execution of this study with the scope of the 2-year extension period applied for.
In essence, this amendment implies that also the neo-adjuvant study aimed at analyzing the pharmacodynamics of ChiLob7/4 in the tumor microenvironment would be concluded within the (extended) running time of the IACT project. These amendments are discussed in detail in the following pages.
In June 2018, it was decided to further amend trial design for the following reasons:
• Further unanticipated delays in completion of the IMPD.
• Additional pre-clinical data from the lab of the Coordinator, now obtained in three independent mouse tumor models including a clinically relevant pancreatic cancer model, supporting the notion that PD biomarker data on the intra-tumoral impact of agonist anti-CD40 Ab is best obtained by means of administration of anti-CD40 Ab as single agent.
• Very recent information from the ASCO meeting (Chicago; May 31-June 4, 2018) based on large 493-patient randomized trial (PRODIGE 24/CCTG PA.6 trial), indicating that FOLFIRINOX is superior and preferable to Gemcitabine as adjuvant treatment after surgery: http://www.ascopost.com/News/58907(opens in new window). The clinicians from partners 1 and 3 involved in the planning of the IAPACA trial have signaled that this new development will significantly complicate the accrual of patients for the IAPACA-1 study, because FOLFIRINOX and Gemcitabine/Abraxane are also gradually replacing Gemcitabine mono-agent treatment in patients with non-resectable disease, and the recently published study will further enhance this shift. In essence, this marks the conclusion of a discussion that has been going on over the past 3 months concerning the gradually decreasing feasibility of the IAPACA-1 study as a result of ongoing changes in clinical practice.
• Further information from the recent ASCO meeting that patients with primary resectable cancer benefit from neo-adjuvant treatment, based on a 246-patient randomized trial (PREOPANC-1): http://www.ascopost.com/News/58908(opens in new window). Based on this study, the interest in the field in neo-adjuvant studies like IAPACA-2A will significantly increase.
In view of these new developments, the IACT partners involved in the planning of the IAPACA trial agreed to further adapt the work plan as follows (see Fig. 4):
• Focus all efforts on the neo-adjuvant single agent study (renamed to IAPACA-NEO) with anti-CD40 Ab ChiLob7/4, as described under Task 2.3.
• Abandon IAPACA-1 as described under Task 2.2
Task 2.2 Dose escalation safety study in non-resectable metastatic PDA. (Partners 1, 2, 3)
As explained above, the plan to pursue this study was abandoned in June 2018
Task 2.3 Neo-adjuvant study: Mechanism of action neo-adjuvant study in resectable PDA (Partners 1, 2, 3)
As a result of cumulative delays in the preparation and QC of the clinical drug, chances in clinical practice and progressing pre-clinical insights, the design of the clinical study with anti-CD40 Ab ChiLob7/4 in PDA was amended multiple times (see above). We are currently aiming for a single agent study that focuses on recording the biological impact of the anti-CD40 Ab in the tumor, as shown in Fig.4.
Basis for these insights was the plan, as formulated in IACT project Task 5.7 to discover new PD biomarkers related to the action of agonist anti-CD40 Ab by performing unbiased mRNA gene expression profiling of tumor samples obtained from mice treated with anti-CD40 Ab alone or in combination with cytostatic drugs. While these experiments confirmed the therapeutic synergy between anti-CD40 Abs and cytostatic drugs, they also revealed that the immune gene expression signature in the tumor was much more pronounced in mice treated with anti-CD40 Ab as single agent as compared to mice receiving both CD40 Ab and cytostatic drug (see below). Even though the paradoxical aspects of these findings still need to be elucidated, it is evident that – at least based on our state of the art pre-clinical data – the likelihood of detecting a PD biomarker signature in the neo-adjuvant trial setting is much higher if the patients are dosed with anti-CD40 Ab ChiLob7/4 only.
Importantly, the detection of this PD biomarker signature is, and has always been, the primary aim of this trial, in that this would embody the first direct evidence that administration of anti-CD40 Ab does not only result in systemic exposure, but also in intra-tumoral drug exposure and delivery of its agonist signal to CD40 positive immune cells.
In addition to this conceptual advantage, performance of a single agent neo-adjuvant trial offers the advantage of reduced risk for adverse events that could delay of surgery due to the fact that ChiLob7/4 will not combined with GMZ. Notably, while GMZ is SOC treatment for patients with non-resectable disease (the patient group selected for IAPACA-1), it is considered an IMP for patients with primary resectable PDA, therefore necessitating the dose finding of ChiLob7/4 in conjunction with GMZ in patients with non-resectable disease.
Notably, while adaptation of the trial protocol to the ‘IAPACA-NEO’ study design is completed, the IMPD is not expected to be completed until April 2019. In view of the fact that completion of the CD40 Ab trial in PDA within the requested 2-year extension period (Nov 2017-Oct 2019) could not be considered as realistic anymore, the Project Coordinator was informed in December 2019 by EU Scientific Officer Jan van the Loo that it had become impossible for him to obtain approval for this extension. In view of this, the strategy of the partners towards performing this trial was amended as follows:
• ChiLob7/4, once released will first be tested for safety in HNSCC patients (See Figure 5 below; trial Arm 2A).
• The safety data of this study will form the basis for presenting the IAPACA-NEO study to the ethics authorities
• Alternative funding for completion of this trial will be requested from CRUK.
WORK PACKAGE 3. COMBINATION OF ANTI-HPV RNA VACCINATION WITH ANTI-CD40 IS-AB IN PATIENTS WITH RECURRENT HPV+ HEAD AND NECK SQUAMOUS CELL CARCINOMA
Objectives
• To test the primary hypothesis that agonist anti-CD40 IS-Abs are capable of mobilizing and boosting systemic and specific, intra-tumour anti-cancer immunity, induced by antigen specific RNA vaccination against HPV.
• To gather early clinical efficacy data on the effects of RNA vaccination when applied as single treatment or in combination with anti-CD40 IS-Ab.
Progress for each task
Task 3.1: Generation clinical grade anti-CD40 IS-Ab ChiLob7/4 (Partner 10)
See Task 2.1 above.
Task 3.2: Preclinical studies with HPV16 mRNA vaccine and anti-CD40 (Partner 3, 10)
Pre-clinical studies in mouse models by BioNTech (partner 10) demonstrated striking immunogenicity and therapeutic anti-tumor potency of the HPV16 E6/E7 RNA vaccine. Subsequent experiments focused on demonstrating synergy between RNA vaccination and CD40 Ab, as previously shown for other vaccine entities. An important aspect of these studies was to determine the most optimal dosing regimen in this respect, and to further elucidate the mechanism of action of CD40 Ab in the context of this combination treatment. Contrary to expectations, the pre-clinical studies in tumor models (TC1, CT26, 4T1) revealed no added value of CD40 Ab stimulation (FGK, rat IgG2) on top of RNA-based vaccination for tumor relevant antigens, in spite of clear anti-tumor, T-cell dependent efficacy of each of the single agents. The lack for synergy was also shown in extensive vaccination experiments with measurement of the T-cell immune response as readout. The reason for a lack in added value of CD40-ligation is thought to be primarily due the potency of the liposomally complexed E6/E7 RNA vaccine alone. Notably, this liposomally-complexed RNA vaccine was developed by partner 10 after planning of the IACT clinical studies.
Importantly, either drug formulation showed significant single agent activity. Based on these findings, the plan to combine these drugs in the context of the clinical studies under Task 3.5 was abandoned, and an alternative plan for testing ChiLob7/4 as single agent in HNSCC was developed (see below).
Task 3.3 Preparation and QC of a clinical batch of HPV RNA vaccine (Partner 10)
BioNTechs’ RNALIP Immunotherapy Pipeline
BioNTech has developed a nanoparticulate formulation which allows for intravenous administration of the RNA vaccine (RNALIP) and for selective expression of the encoded antigens in antigen-presenting cells in lymphoid tissues, thereby inducing potent antigen- specific immune responses already at low doses (Kranz et al., Nature, 2016). BioNTech takes the position that each technology improvement and each novel therapeutic concept should be introduced into clinical first-in-human testing in a sequential manner in order to allow for incorporating safety-related findings in any study in subsequent trial designs, an approach which efficiently mitigates potential risks inherently associated with each novel component. This view is shared by the German competent authority, the Paul-Ehrlich-Institut (PEI). The RNA vaccine tested in the HARE-40 trial is related to the RNA vaccines tested in BioNTech’s Lipo-MERIT and (TNBC-)MERIT trials, and although the HARE-40 trial is not sponsored by BioNTech, sequential trial initiation was deemed reasonable to ensure patient safety and hence was discussed and agreed with the sponsor/chief investigator of the HARE-40 trial. Two issues encountered during trial initiation and patient treatment in BioNTech’s Lipo- MERIT trial had a direct impact on the design and treatment start in BioNTech’s (TNBC-)MERIT trial, and hence on the design, submission and initiation of the HARE-40 trial and are described in detail below.
Preparation of RNALIP: Manufacturing vs. Reconstitution
In line with the EMA Guideline CPMP/BWP/3088/99 (Committee for Proprietary Medicinal Products (CPMP) ‘Note for Guidance on the Quality, Preclinical and Clinical Aspects of Gene Transfer Medicinal Products’, 24 April 2001) the preparation of RNA with liposomes was classified as reconstitution by BioNTech. German local regulatory authorities, however, unexpectedly classified the final step of RNALIP preparation as a manufacturing step which requires a dedicated manufacturing license. In addition, the regulatory agency asked for de-complexation of the IMP preparation protocol.
In response, BioNTech generated a simplified, robust two-step RNA(LIP) preparation procedure which is also applied to the RNA vaccine used in the HARE-40 trial. Furthermore, BioNTech undertook all necessary steps to verify, audit and contract a pharmacy which holds a manufacturing license for Advanced Therapeutic Medicinal Products (ATMPs), the drug class RNALIP falls into. BioNTech and the identified pharmacy prepared all relevant documents to obtain the manufacturing authorization for RNA(LIP) from the local competent authority. The manufacturing license was finally obtained in December 2014, and the pharmacy now supplies clinical sites in Germany and in Belgium with ready-to-use RNALIP (Lipo-MERIT and (TNBC-)MERIT trials). Of interest, the final step of RNALIP preparation was later classified as reconstitution in France, Sweden, and in the UK. Obtaining the manufacturing license implied a delay in starting the melanoma trial, and thereby in the generation of clinical safety data on RNALIP which needed to inform the design of (TNBC-)MERIT and HARE-40 trials, including the definition of a safe starting dose, thereby initially delaying HARE-40 submission by 12 months.
Clinical Trial Application and Study Start
After having resolved the regulatory issues regarding RNALIP manufacturing described above, the application to the British competent authorities was planned for summer 2015. However, results obtained from the very first patient treated with RNALIP in May 2015 (within the Lipo-MERIT trial) caused a further delay in submitting the ‘HARE-40’ clinical trial application: The treatment of the first melanoma patient revealed a high potency of RNALIP already at the very low doses initially used. The first patient experienced symptoms typically associated with immune system activation within hours after vaccination. These effects were dose-dependent but already observed at the second-lowest dose level. Additional adverse events were observed after escalation to the next higher dose level. The observed symptoms were readily manageable by co-medication but led to a dose reduction for the remaining vaccination cycles in this patient at the discretion of the treating physician. After review of the first patient’s clinical data an independent data safety monitoring board (DSMB) approved the inclusion of two further patients and recommended continuation of the melanoma trial as planned. Nevertheless, these findings mandated an adaptation of the intended dose-escalation scheme in all trials testing RNALIP (Lipo-MERIT, (TNBC-)MERIT, and HARE-40). To this extent and in agreement with the PEI (the responsible competent authority for the Lipo-MERIT trial), it was decided to first wait for and to evaluate clinical safety data generated from the first three patients in the Lipo-MERIT trial prior to finalizing the dose-escalation scheme for all RNALIP trials, including HARE-40. The data from these first three patients were available by November 2015 and revealed significant inter-individual variations both in cytokine secretion as well as in the intensity of adverse events typically associated with immune activation. These findings are in line with inter-individual variation of TLR function due to genetic polymorphisms and an ageing-associated decrease of TLR-mediated inflammatory reactions, and suggest that it is unlikely to identify one pharmacodynamically active and safe dose which fits all patients.
To better address the observed inter-individual variability, the dose identification strategy for all RNALIP trials at that time, including HARE-40, encompassed a dose-escalation in every single individual patient by starting treatment of each individual with a very low dose and stepwise increasing the individual dose until either the dose level of the respective cohort is reached or until a lower individual dose is identified which cannot be further safely increased due to inflammatory reactions. The individual dose-escalation process foresees that after every drug exposure and prior to the next drug application a decision has to be made by the responsible investigator whether the dose-escalation in the individual patient will be continued as planned or, in case of significant inflammatory adverse events, treatment has to be modified according to a predefined algorithm.
The novel treatment scheme had to be worked out and implemented into the HARE-40 trial design prior to the initial submission of the clinical trial protocol to the British authorities, which was therefore delayed to March 2016. The MHRA and ethics committee supported the clinical trial application and approved testing of the RNA vaccine in the adjuvant and advanced disease settings. The authorities, however, at that time did not approve testing of the RNA vaccine in neo-adjuvant patients as clinical safety data with that particular RNALIP product had not yet been available. Instead, the authorities suggested implementing that arm of the study only later via submission of a substantial amendment. The authorities further asked to remove all reference to the intended testing of ChiLob7/4 antibody from the protocol until the product had been manufactured. A substantial amendment should then be submitted together with the IMPD and IB describing ChiLob7/4 product characteristics. Consequently, an amended protocol version, in which the neo-adjuvant study arm as well as all reference to ChiLob7/4 had been removed, had to be prepared and was submitted to the MHRA and local ethics committee in May 2016. Approval by the British authorities was obtained in summer 2016.
The RNALIP products for the HARE-40 trial are reconstituted from individual components, i.e. the RNA drug products, sodium chloride as diluent, and liposomes as excipient. The components were to be packaged into one ‘HARE-40 kit’ to be shipped to and stored until use at the hospital pharmacy at 2-8 °C. By the time of protocol approval, new stability data obtained for the RNA drug product batches manufactured for HARE-40 revealed that the HARE-40 kit storage temperature of 2-8 °C was insufficient to safely ensure drug supply over the whole treatment cycle of 78 days for a given HARE-40 patient. As a consequence, the individual components for RNALIP reconstitution were separated to be supplied in two different kits (i.e. one kit containing the RNA drug products, to be shipped and stored at <-15 °C, and another kit containing sodium chloride and liposomes, to be shipped and stored at 2-8 °C). As packaging of the individual components into kits is a manufacturing step to be performed in full compliance with GMP, realization of the design change required adaptations to various controlled documents which are subject to controlled review and release processes.
In addition, the new approach now using two kits instead of one required another substantial amendment, which required changes not only to the protocol but also to IMPD and IB. This amendment was submitted in December 2016 and was approved by the authorities in spring 2017 (EudraCT number 2014-002061-30). Right upon approval, the last preparatory steps required to enable patient treatment, i.e. qualifying the hospital pharmacy staff for RNALIP reconstitution, shipment of sufficient HARE-40 kits for the initial patients to be treated, and site initiation by sponsor were conducted and the first HARE-40 patient received the first vaccination with RNALIP in May 2017.
Task 3.4 RNA vaccination dose escalation study in recurrent HPV16+ HNSCC (Partners 3, 5 and 10)
Task 3.5 Safety/immunogenicity study combining HPV mRNA vaccination with anti-CD40 IS-Ab in HPV16+ HNSCC (Partners 3, 5 and 10)
Amendment of trial design
The design of this study as described in the original IACT Description of Work (DOW) was as follows:
• Task 3.4: RNA vaccination dose escalation study in recurrent HPV+ HNSCC (Study A)
• Task 3.5: Safety/immunogenicity study combining HPV RNA vaccination with anti-CD40 IS-Ab in HPV16+ HNSCC (Study B)
Task 3.4 and 3.5 have been combined within the HARE-40 trial; Study A is represented by part I (Arms 1A-C) and Study B by part II (Arm 2A) of the trial (see Figure 5 for amended design).
The design of these tasks has been amended in view of the following insights that emerged in the course of the IACT project:
1. The ongoing studies in melanoma with the new liposomal BioNTech RNA vaccine formulation (RNA(LIP)) demonstrated impressive potency of the vaccine format with induction of antigen-specific T-cell immunity and early evidence of clinical responses.
2. Unexpectedly, substantial inter-patient variability was observed with respect to adverse events following administration of vaccine using the new vaccine platform (see Task 3.3 above). While no dose-limiting toxicity was identified, the decision had to be taken to keep patients in the hospital overnight for close monitoring. This has required considerable extra logistical effort as patients now have to be accommodated overnight for each visit until they have reached the target vaccine dose. This also meant that we undertook the ‘first in man’ element of the study in a single center and only recruited one patient at one time. Having gained experience with the adverse event profile (grade 2 fever and rigors) we are confident that we can relax the restrictions on parallel recruitment after the review of the trial by the safety board. We will however continue intra-patient dose escalation and careful monitoring of each new patient.
3. Pre-clinical studies indicated that combination of this new liposomal vaccine formulation with agonist anti-CD40 IS-Abs did not result in synergistic effects, most likely due to the increased potency of this vaccine. Nevertheless, these studies showed that each of these drugs had significant single agent activity against HP16-positive tumors.
Based on these insights, the design of the HARE-40 trial was amended as follows:
• In order to evaluate immunogenicity and to assess immune toxicity in patients in whom immune function is not affected by advanced disease, Arm 1A recruits patients who are currently disease free, but at significant risk of recurrence.
• The SRC review following vaccination of 6 patients at dose level 3 will trigger the recruitment of patients to the Arm 1A expansion cohort (intra-patient dose escalation to dose level 5) and the opening of the dose finding element of Arm 1B (N=15). The two study arms will allow direct comparison of immunogenicity in disease-bearing and disease-free patients. Two additional UK centers are in setup to support timely recruitment of patients with advanced disease.
• Following completion of production of the antibody (Task 2.1) we are also now in a position to test the anti-CD40 IS-Ab ChiLob7/4 in the clinic in HNSCC. Due to the mode-of action of the anti-CD40 IS-Ab, involving activation/modulation of CD40-positive immune cells in the tumor microenvironment, patients in the adjuvant setting will not benefit from this treatment. Hence, we propose to evaluate the antibody in parallel to vaccination in patients with advanced disease. In Arm 2A we will recruit patients who have an HPV16 negative cancer (exclusion criterion for Arm 1B) but are otherwise eligible for the trial. This will allow us to bring forward the testing of the anti-CD40 IS-Ab to early 2018, the protocol amendment to allow this is in preparation to be submitted during November 2017. The MHRA supported our proposal to add the anti-CD40 Ab arm, pending evaluation of the final batch of antibody, which we now have (Task 2.1) when we sought scientific advice prior to initial approval of the study.
• In view of the observed potency of our RNA vaccine formulation in patients with melanoma (see point 1 above), we will also add an expansion cohort to Arm 1B, as well as a study arm in which the vaccine will be administered in the neo-adjuvant setting (Arm 1C), enabling detailed evaluation of the impact of the vaccine on (the T-cell response in) the tumor micro-environment.
Trials status December 2018
Arm 1A (adjuvant setting) cohort 1 was opened for enrolment in May 2017, involving intra-patient dose escalation up to 29 ug of vaccine. The first 6 patients were enrolled and completed the intra-patient dose escalation scheduled without severe IMP-related AEs. Upon positive safety review of the results of this cohort in January 2018, Arm 1A (adjuvant setting) cohort 2 and Arm 1B (metastatic disease) have been opened, each of which involve intra-patient dose escalation up to 72.8 ug (9 patients each). In Arm 1A, we have now recruited all 15 of the planned patients. For Arm 1B, we have now recruited and treated 3 (of 29) patients, with 4 further patients on the waiting list. Upon positive evaluation of the results of these cohorts, Arm 1C (neo-adjuvant; 15 patients) can be opened.
The interest for our trial is intense with multiple referrals coming from the UK and other EU countries. Nevertheless, rapid recruitment of patients has been hampered by the fact that the 2-year extension of the IACT project was not approved, and was eventually rejected. In particular, this halted the inclusion of additional clinical centers (Poole, Liverpool, London, Pamplona) in the trial.
Preparation for Arm 2A, involving a dose escalation study with anti-CD40 Ab ChiLob7/4 in patients with advanced disease, will proceed independently of Arms 1A-C. The time-limiting factor for this arm is completion of the IMPD for ChiLob7/4 by partner BioNTech, as referred to above. The plan is to submit the HARE-40 Arm 2A protocol as soon as possible to the MHRA.
WORK PACKAGE 4: DEVELOPMENT OF LEAD ANTI-OX40 AND ANTI-4-1BB IS-ABS FOR CLINICAL USE
Objectives
• Systematic selection of lead anti-human OX40 and human 4-1BB IS-Abs based on multiple in vitro and in vivo predictive assays
• Engineer humanised mAb against OX40 and 4-1BB for clinical use.
Progress for each task
Task 4.1 In vivo testing in receptor knock-in mice: T-cell costimulation and anti-tumor efficacy (Partner 3)
Anti-human 4-1BB Abs:
Twenty candidate anti-4-1BB Abs were subjected to selection by measuring their capacity to costimulate (CD8+) T-cell proliferation in human PBMC-cultures. This resulted in identification of clone SAP3-28 as lead candidate plus several backups. Notably, these costimulation assays make use of PBMCs that have been subjected to high-density culture for 2 days; see WP6, Task 6.1.
Since candidates have been selected for agonist function, a continuation towards the human-IgG2 format seems the most logical step forward (for important information on this Ab format, see WP7, task 7.2). Even though depletion of Treg with huIgG1-formatted Abs may be of potential interest, the expected depletion of CD8+ effector T-cells seems to be an undesirable aspect of this concept. This also fits best with the recently identified merits of 4-1BB Abs in stimulating ADCC activity in conjunction with tumor-targeting Abs (Levy et al. Blood. 2011 117(8):2423-32; J Clin Invest. 2014 124(6):2668).
In vivo analysis of anti-human 4-1BB Abs was impeded by the failure of human 4-1BB-ki mouse to express the receptor of interest. Consequently, further testing was focused on the CD4+ T-cell assays (Task 4.3) and NK assays (Task 4.4).
Anti-human OX40 Abs:
Twenty anti-OX40 Ab candidates were screened for their ability to costimulate human T-cell proliferation in human PBMC-cultures in vitro, as well as hOX40 KI OT-I cells in vivo in mice, as a measure of agonistic function. For OX40, several good candidates were identified, including clones SAP15-3, 25-29 and 29-50. The most promising candidate with respect to costimulation of effector T-cell responses in human OX40 transgenic mice were found to be clones 25-29 and 29-50, with clone 29-50 consistently showing somewhat greater activity than 25-29.
Task 4.2 OX40: CD4+ T-regulatory cell suppression (Partner 6)
Optimized assays have been developed by partner 6 (Milan) to measure the capacity of anti-OX40 Abs to prevent the conversion of non-suppressive CD4+ T cells into Tr1 suppressor cells. CD4+CD25- T-responder cells were purified from total PBMC using MACS MicroBead Technology. As T-responder cells do not express OX40 constitutively, we induced its up-regulation stimulating T-responder lymphocytes with TNFα and IL-2 for 18 hours. The resulting OX40-expressing responder T cells were cultured for seven days in the presence of dexamethasone and vitamin-D3 to induce their conversion into suppressive Tr1-IL10 secreting cells. Compared to control cells, under dexamethasone and vitamin-D3 pressure, T-responder cells secret higher amount of IL-10. The addition of the anti-OX40 Abs, especially clones SAP15-3 and SAP25-29 (provided by partner 3), reduced IL-10 secretion by T-responder cells. Moreover SAP25-29 increases the production of IFNγ, suggesting a more inflammatory/non-suppressive phenotype of T-responder cells despite the presence of dexamethasone and vitamin-D3.
More extensive testing of the selected, anti-human OX40 Abs (SAP 15-3; 25-29; 29-50), as well as anti-human 4-1BB Ab 3-28 Abs was performed in the second project period. Two different assays were used to test the capacity of the selected Abs to suppress respectively the development of Tr1 cells (under the influence of Dex and VitD3) from fresh PBMC-derived CD4+ T-cells or the IL-10-secretion by already differentiated Tr1 cells. CD3 and OX40/4-1BB Abs were added as plate-coated Abs in these assays. In both assays, the fraction of IL-10-secreting CD4+ T-cells was assessed by FACS/ICS. The most significant outcomes were as follows:
• Most prominent differences were observed in Tr1 differentiation assays that started with either total or naïve CD4+ T-cells
• Most consistent inhibition of IL-10 secretion and IRF1 expression were observed with SAP29-50 (OX40) and SAP3-28 (4-1BB). Overall, impact of SAP25-29 was weaker, while SAP15-3 was clearly the weakest OX40 Ab
• SAP29-50 (OX40) also proved to be the most potent Ab in costimulating proliferation of CD4+ T-cells under regular (in absence of DEX or VitD3) conditions
• Results from other assays are less convincing with respect to differentiating between the Abs tested, which was mainly due to limited difference between positive and negative control (so window for accurate measurements too small).
In view of the above, OX40 Ab SAP29-50 appears to be most potent and will be considered as the lead candidate. Furthermore, these data support the biological activity of the (only) available humanized anti-4-1BB candidate SAP3-28.
Based on the data set shown, the Tr1 differentiation assay with IL-10 and IRF1 as readout is the test of choice for evaluation the dual-specific Abs generated by BioInvent (see WP8 below).
Task 4.3 4-1BB: CD4+ T-cell response suppression (Partner 6)
Partner 6 also tested anti-human 4-1BB Ab SAP3-28 for its biological activity in the in vitro assays described under Task 4.2. This confirmed that this antibody is capable to suppress Tr1 development.
Task 4.4 4-1BB: stimulation NK activity. (Partner 4)
These experiments focused on the recent finding that agonist anti-4-1BB Abs can potentiate the anti-tumor efficacy of tumor-targeting Abs that mediate their anti-tumor effect (at least in part) through ADCC (Levy et al. Blood. 2011 117(8):2423-32; J Clin Invest. 2014 124(6):2668). The EGFR-specific Ab Cetuximab and EGFR+ HT29 tumor cells were used a model for the initial experiments. In order to evaluate ability of the anti-human 4-1BB IS-Ab to stimulate NK cell activity in vitro, freshly isolated human peripheral blood NK (CD56+ CD3-) cells were incubated with Cetuximab and HT29 tumor cells. NK cells significantly upregulated 4-1BB, becoming receptive to 4-1BB costimulation. Several assays were performed to measure the stimulation of human NK cell-mediated ADCC towards HT29 cells in conjunction with Cetuximab. Of these, the in vivo assay focusing on the percentage of viable HT29 cells recovered upon peritoneal lavages and liver metastasis shows a prominent impact of the reference anti-4-1BB Ab (6B4). A pilot experiment with anti-4-1BB Ab SAP3-28 was ongoing.
An alternative in vitro assay tested the impact of 4-1BB Abs on CD25 and BCL-XL expression in CD8+ T-cells pre-estimulated with anti-CD3 (detected in flow cytometry). Reference 4-1BB Ab 6B4 results in a prominent induction of CD25 in approx. 50% of CD8+ T-cells, along with a shift towards expression of higher levels of BCL-XL. In addition, enhanced proliferation was detected by means of V450 fluorochrome dilution, albeit less reproducibly. Notably, the Southampton reference Ab SAP3-28 was not yet tested in these assays!
Task 4.5 Humanisation and testing of selected anti-OX40 and anti-4-1BB mAb (Partner 3)
Humanization of the selected Abs was successfully completed for all lead Abs. However, of the anti-OX40 Abs humanized, only Ab 25-29 retained full affinity for huOX40. Humanization of Ab 29-50 only succeeded in one of the humanized frameworks that were applied, and the resulting humanized Ab had a clearly diminished affinity as compared to the original muIgG1 Ab. The same applied to the humanization of Ab 3-28.
Changes in Work Plan due to strategic considerations
On June 1 2015, BioInvent (Partner 8) announced that it entered in a research collaboration with the University of Southampton (partner 3) concerning the development of T-regulatory cell checkpoint inhibitors. This collaboration involved licensing by CRUK/CRT to BioInvent of the lead anti-human OX40 and anti-human 4-1BB Abs that are subject of IACT work package 4. The major objective of this collaboration, the accelerated clinical development of this Abs, is highly complementary to the IACT work plan. Partners 8 and 3, as well as CRUK/CRT, have declared to the IACT coordinator (Offringa, partner 1) that this is a non-exclusive licensing agreement that does not affect the workings of the IACT project The Heidelberg Technology Transfer office has investigated this in more detail, and concluded that this is correct and, therefore, that no amendment of the Consortium Agreement needs to be made in this respect.
In the course of 2017, further work on Tasks 4.3 – 4.5 was suspended. The rationale for this decision was that the aforementioned collaboration between University of Southampton (partner 3) and BioInvent (partner 8) had resulted in the identification of fully human, phage-derived Abs with properties superior to the IACT lead Abs. In view of the competitiveness of the field and the need to focus on ‘best in class’ strategies (see above), it was decided to stop all pre-clinical development work on the original set of OX40 and 4-1BB IACT candidate Abs. Irrespective of this, these IACT Abs will still be used as a reference for other project goals.
While this means that the IACT goal for developing the pre-defined anti-OX40 and anti-4-1BB lead Abs towards clinical testing (Deliverable D4.5) will not be fulfilled, the IACT consortium did create a setting that facilitated the isolation of the new, superior Abs. Notably, these new Abs fall outside of the scope of the IACT project and IACT consortium agreement, but instead have been generated and tested in the context of the aforementioned collaborative agreement between University of Southampton and BioInvent.
WORK PACKAGE 5: BIOMARKER ANALYSIS ACCORDING TO STANDARD OPERATING PROCEDURES (SOPS)
Objectives
• Develop validated assays for effective measurement of drug pharmacokinetics (PK) and drug/treatment-related pharmacodynamic (PD)-biomarkers.
• Define and harmonized these assays by means of SOPs, that will be distributed among IACT partner
Progress for each task
Task 5.1 Pharmacokinetics (PK) assays (Partners 3, 7, 8, 9)
A validated sandwich ELISA, as developed by partner 3, is available for anti-CD40 Ab ChiLob (Chowdhury et al. J Immunol Methods 2010 363(1):1) and has already been applied successfully in the completed phase I trial (Johnson et al. 2015 Clin Cancer Res 21(6):1321). SOPs are available for sharing with IACT partners.
Task 5.2 Anti-drug antibody (ADA) response assays (Partners 3, 7, 8, 9)
A validated ELISA, as developed by partner 3, is available that uses anti-CD40 Ab ChiLob as capture antigen (Chowdhury et al. J Immunol Methods 2010 363(1):1-8) and has already been applied successfully in the completed phase I trial (Johnson et al. 2015 Clin Cancer Res 21(6):1321). SOPs are available for sharing with IACT partners.
Task 5.3 Serum cytokine/chemokine analyses (Partners: 1, 3, 4, 10)
SOPs for multiplex analysis of serum samples have been established by partner 3 in the context of the completed single agent trial (see WP2 Task 2.2 for more info). The completed analyses have revealed a sharp, >10-fold increase in MIP1b within the first day after dosing, most likely to reflect an infusion reaction rather than a result of engagement of the CD40 receptor. In addition, a minor (2-fold) increase in serum IL-12 could be observed in the course of the 1st week after dosing. This biomarker is more likely to be related to the biological activity of agonist anti-CD40 Ab. A limitation in detecting serum levels of IL-12 may be the limited half-life of this cytokine in the circulation. These results correspond to those obtained with another, very similar anti-CD40 Ab (SGN40; Moskowitz et al. 2013 Leuk Lymph 54: 277 and references therein).
In view of the limited usefulness of serum cytokines as mechanism of action-related biomarker, efforts to identify alternative biomarkers are being pursued (see Task 5.7).
Task 5.4 Analysis of blood cell subsets and activation markers (Partners: 1, 3, 4, 10)
Flow cytometric analysis of samples from the completed single agent study revealed a rapid (within 1 day of dosing) loss of peripheral B-cell counts, and a slower loss of NK cells. While the change in the number of CD40+ B-cells are clearly relevant to the impact of the anti-CD40 Ab, the mechanism by which the CD40-negative NK cells disappear is unclear. Measurements of dendritic cells (very low numbers) and T-cells (no significant effects) did not point at additional biomarkers that would be useful. These results again correspond to those obtained with anti-CD40 Ab SGN40 (Moskowitz et al. 2013 Leuk Lymph 54: 277 and references therein).
The disappearance of B-cells from the circulation can be used as a biomarker for systemic action of agonist CD40 Ab, but since this is known to happen in any human subject to which such Abs are administered, this biomarker will not be useful as a surrogate endpoint for clinical activity of anti-CD40 Ab.
Task 5.5 Analysis of HPV16 E6/E7-specific T-cell immunity (Partners 1, 3, 4, 10)
Validation of the following biomarker assays was completed by partner 3 (Ottensmeier):
• Serum ELISA for Ab titers
• ELISPOT-analysis using overlapping pools of synthetic peptides: 15-mers with an overlap of 11 residues, in total 59 peptide divided in 16 pools of 3-8 peptides
• IFNγ intracellular staining using the same peptide pools
• Staining with HLA-A2 tetramers for the following known epitopes. E6: 18-26 KLPQLCTEL, 29-38 TIHDIILECV, 52-60 FAFRDLCIV; E7: 11-19 YMLDLQPET(T), 82-90 LLMGTLGIV, 86-93 TLGIVCPI
In addition, impact of immunotherapeutic treatment on tumor microenvironment will focus on RNA-seq analysis of tumor samples. Baseline data for 40 HPV16+ HNSCC tumors have been compiled, as published in: Gene expression analysis of TIL rich HPV-driven head and neck tumors reveals a distinct B-cell signature when compared to HPV independent tumors. Wood O, Woo J, Seumois G, Savelyeva N, McCann KJ, Singh D, Jones T, Peel L, Breen MS, Ward M, Garrido Martin E, Sanchez-Elsner T, Thomas G, Vijayanand P, Woelk CH, King E, Ottensmeier C. Oncotarget. 2016 Jul 22. doi: 10.18632/oncotarget.10788.
Task 5.6 Evaluation of intra-tumour immunity by means of whole slide ‘Immunomap’ imaging (Partner 1)
Validated SOPs for the most important stainings (T, B-cell, DC, NK, macrophages) are available. Baseline data for T-cells and macrophages is currently available for 180 pathology-verified tumors (primary resectable pancreatic ductal adenocarcinoma; PDA). For an additional 40 tumors, data is available on a marker set that defines tertiary lymphoid structures (DC-markers, B-cells, T-cells high endothelial venules, ki67); this panel will be expanded to at least 80 tissue samples. Stainings of pancreatic tumors (>40) for the remaining markers (HLA-II, INOS, Arginase, CD86, CD40) are now available. Especially striking are the vast numbers of CD40-positive cells in the tumors. This clearly shows that this tumor is a splendid target for treatment with agonist anti-CD40 Ab. The detection of high numbers of CD40-positive cells in tumors was confirmed by stainings with two additional anti-human CD40 Abs, including our clinical Ab ChiLob7/4 (!), supporting the expectation that treatment of patients with agonist anti-CD40 Abs, when administered at dose levels resulting in systemic exposure, is likely to impact on the immune cell infiltrate in these tumors.
In order to enable the detection of cell-bound anti-CD40 Ab ChiLob7/4, partner 3 generated several anti-idiotype Abs. While these Abs detected cell-bound ChiLob7/4 under optimal in vitro conditions, the sensitivity of detection was unfortunately too low to allow detection of cell/tissue-bound ChiLob7/4 under physiological conditions, e.g. in tumor tissue of patients treated by ChiLob7/4. In view of this, this aim (as mentioned in the report on the 2nd project period) was deprioritized.
Task 5.7 Unbiased PD biomarker discovery by means of mRNA gene expression profiling of PBMC and tumour biopsies (partners 1, 3 and 10)
Identification of PD biomarkers on basis of the transcriptome of in vitro stimulated human PBMC rendered interesting findings. However, there is major variability in the signature obtained between different PBMC batches. Furthermore, it is evident that in vitro culture alone already triggers a time-dependent transcriptional profile. While this does not exclude meaningful findings using PBMC from ChiLob-treatment patients, it is unlikely that the in vitro experiments will provide significant guidance in this respect. Based on these findings, further experiments focused at the tumor transcriptome of mice treated with agonist CD40 Ab (see WP6/Offringa below). These latter experiments showed that transcriptome analysis of tumors from mice treated with anti-CD40 Abs (Tasks 6.5 and 6.6) proved to be highly informative in this respect.
While these experiments confirmed the therapeutic synergy between anti-CD40 Abs and cytostatic drugs, they also revealed that the immune gene expression signature in the tumor was much more pronounced in mice treated with anti-CD40 Ab as single agent as compared to mice receiving both CD40 Ab and cytostatic drug (for details, see under task 7.2 as well as report on Deliverable D5.4 as well as Annex 5 of project extension request). Even though the paradoxical aspects of these findings still need to be elucidated, it is evident that – at least based on our state of the art pre-clinical data – the likelihood of detecting a PD biomarker signature in the neo-adjuvant trial setting is much higher if the patients are dosed with anti-CD40 Ab ChiLob7/4 only (see WP2 for implementation). Importantly, the detection of this PD biomarker signature is, and has always been, the primary aim of this trial, in that this would embody the first direct evidence that administration of anti-CD40 Ab does not merely result in systemic exposure, but also in intra-tumoral drug exposure and delivery of its agonist signal to CD40 positive immune cells.
WORK PACKAGE 6. VALIDATED AND HARMONIZED PRE-CLINICAL IN VITRO AND IN VIVO MODELS (SOPS)
Objectives
Development, validation and harmonization of:
• in vitro models for assessing the immunostimulatory capacity as well as potential super-agonist activity of agonist IS-Abs
• in vivo immunization models for the evaluation of agonist IS Abs potency
• transplantable tumour models for initial evaluation of anti-tumour efficacy of agonist IS-Abs
• genetically engineered tumour models for in-depth evaluation of therapeutic index of selected therapeutic regimens involving agonist IS-Abs
Progress for each task
Task 6.1 In vitro assays for analysis of stimulatory activity agonist IS-Abs (Partners 3, 4, 6)
Experiments by partner 3 have shown that high density pre-culture, as initially determined in the context of assays with super-agonist anti-CD28 Ab TGN1412 (Kalinke et al. J Immunol 2014; 192:2091-2098), is an essential step in generating PBMC for use in in vitro assays to determine costimulatory activity agonist Abs in presence of low concentrations of anti-CD3 Ab OKT3. Furthermore, the optimal concentration of plate-bound OKT3 differs between PBMC donors. Consequently, a range of concentrations needs to be tested (0.05 – 1 ug/ml), at least initially (if multiple aliquots from same donor are available). Finally, responses by CD4+ and CD8+ T-cell subset differ between PBMC donors and Abs. Therefore, proliferation, as measured by CFSE-dilution, is determined by FACS for both T-cell subsets. SOPs for these assays have been made available to the IACT partners
Task 6.2 In vitro assays for detecting super-agonist activity IS-Abs(Partner 3)
High density pre-incubation as mentioned above similarly enables testing for superagonist activity (see above; SOPs available). In line with published data (Kalinke et al. J Immunol 2014; 192:2091-2098), induction of T-cell proliferation by anti-CD28 Ab TGN1412 in the absence of CD3 Abs (so super-agonist activity) was only seen for high density pre-cultured PBMCs. Notably, this super-agonist activity was found to be dependent on Fc-FcR-interaction, as F(ab)2 fragments lacked activity. Moreover, Ab-mediated blockade of FcRIIB, but not of FcRI or FcRIII, abolished super-agonist activity of TGN1412, demonstrating that the FcRIIB-dependency is the same as found for the agonist Abs targeting CD40, OX40, 4-1BB and CD27.
Both monocytes and B-cells were shown to provide this FcRIIB-dependent crosslinking capacity in in vitro cultures; cell titration experiments showed that approx. 10-fold less monocytes than B-cells were required for this accessory function. Furthermore, high-density pre-culture was found to strongly upregulate FcRIIB on monocytes, while not affecting levels on B-cells. Therefore, the increased responsiveness of high-density pre-cultured PBMC to agonist Abs is primarily due to upregulation of FcRIIB on the monocyte subset, rather than T-cell-intrinsic features as proposed in the aforementioned publication in J Immunol.
Task 6.3 In vitro assays for detection of CD4+ T-cell suppression (Partners 4, 6)
See description under Task 4.2 and 4.3.
Task 6.4 In vivo immunization models (Partners 1, 3, 4, 6)
In vivo assays for testing anti-mouse agonist Abs targeting CD40, OX40 and 4-1BB have been recently published by partner 3 (White & Glennie et al. J Immunol. 2011 187(4):1754-63; Cancer Immunol Immunother. 2013 62(5):941-8; J Immunol. 2014 193(4):1828-35; Cancer Cell. 2015 Jan 27(1):138-48) and protocols have been made available to the IACT partners. More recently, an in vivo assay for testing agonist anti-human OX40 Abs has been developed by making use of huOX40-Tg OT-1 T-cells. Unfortunately, the hu-4-1BB-Tg mice were found to lack functional expression of the receptor of interest, so an in vivo assay for testing anti-human-4-1BB Abs is not yet available and may not be achievable. In vivo immunization models using synthetic OVA peptides have been set up by partner 1 on basis of prior experience (Bijker, Offringa et al. Eur J Immunol. 2008 38(4):1033-42) and protocols are available to the IACT partners.
Task 6.5 Transplantable tumor models (Partners 1, 4, 6)
Standardized tumor models based on MC38-OVA, B16-OVA as well as PANC-02 have been set up by partner 1 and SOPs are available to other IACT partners. The PANC-02 model will be used to experiment with the relative timing of gemcitabine chemotherapy and anti-CD40 Ab dosing as discussed in WP2 Task 2.1.
Task 6.6 Genetically engineered mouse models (GEMMs) (Partner 1)
A melanoma GEMM is up and running in Heidelberg (partner 1). This model features conditional, melanocyte-specific expression of BrafV600E, overexpression of BCat and inactivation of PTEN. SMi targeting BRaf and MEK have profound anti-tumor activity, resulting in resp. tumor stasis and tumor regression. The anti-tumor impact of Temozolide is modest. Therapeutic impact of combinations of cytotoxic oncology drugs with agonist CD40 and OX40 Abs is currently under investigation (close collaboration between partners 1 and 3).
In order to improve the availability of chimeric anti-mouse Abs for these in vivo experiments, partner 10 has been involved in optimizing the production of these Abs (partner 10 has also enabled the production of clinical-grade anti-human CD40 Ab by generating new producer lines; see Task 2.1). In house production of surrogate agonist Ab obtained from Southampton (partner 3) (CD40, OX40, 4-1BB) has been set up in Heidelberg with help of Andreas Fregin and Oezlem Tureci of BioNTech (partner 10) sister-company Ganymed. Using chimeric anti-mouse CD40 3/23-muIgG1, multiple oncology SMi have been tested in in vivo assays for potential negative impact on DC/T-cell function (OVA immunization assays in OT-1 system), as well as in tumor models for potential synergy in anti-tumor impact with CD40 Ab. In particular, a MEK-inhibitor was shown to have no negative impact on DC/T-cell function while synergizing with CD40 Ab in anti-tumor efficacy (Offringa et al. manuscript in preparation). While these experiments have been performed in transplantable tumor models, genetically engineered models for skin melanoma and PDA are now available in the lab that feature traceable neo-antigen for follow up studies in autochthonous tumor systems.
WORK PACKAGE 7. DEVELOPMENT OF OPTIMIZED 2ND-GENERATION AGONIST IS-ABS THROUGH MODIFICATIONS IN FC-DOMAIN
Objectives
• Generation optimized IS-Abs through modification of the primary sequence of the Fc-domain, or through introduction of defined changes in the glycosylation of the Fc-domain.
• Selection of potential clinical candidates by means of functional assays that examine mAb function with respect to immune stimulation, induction of tumour-regression, as well as the absence of super-agonist activity
Progress for each task
Task 7.1 Introduction of mutations into the human IgG1 (h1) Fc-region to increase binding to human FcγRIIB (Partners 3, 4, 6)
In vivo experiments by partner 3 in mice have demonstrated that the mIgG1 is most potent Ab format for agonist Abs targeting CD40 and several T-cell costimulatory receptors (OX40, 4-1BB, CD27), and that the mechanism of action involves FcRIIb-dependent clustering of Abs and resulting immune-receptor multimerisation (White & Glennie et al. J Immunol. 2011 187(4):1754-63; Cancer Immunol Immunother. 2013 62(5):941-8; J Immunol. 2014 193(4):1828-35). Since all FcRs can in principle mediate this multimerisation-dependent agonist activity in vitro, bio-availability of FcR-mediated Ab-cross-linking seems to be a key feature in the in vivo setting.
A major hurdle in translating these findings towards the human system is that there are significant differences between the mouse and human Fc/FcR system, including the fact that in the human system there is no true counterpart of mIgG1. Nevertheless, it is known that certain mutations in the huIgG framework can increase the affinity to human FcgIIB, in particular the S267E and L328F mutations. Data from partner 3, as reported previously (see also report on Deliverable 7.3) demonstrated that the so-called huIgG SELF-variants of anti-mouse and anti-human CD40 IS-Ab display display greatly enhanced, FcRIIB-dependent agonist activity in in vitro and in vivo experiments. However, the in vivo tests also showed significant toxicity in mice, as reflected by weight loss, liver toxicity (increase in systemic ATS/ALT levels), ruffled skin and hunched posture. Similar studies by Ravetch and co-workers confirmed that the SELF huIgG framework enhanced both the agonist activity and in vivo toxicity of anti-CD40 IS-Abs, although in this case the toxicity was reported to primarily caused by thrombocytopenia (Dahan et al. 2016 Cancer Cell 26, 820). Based on these findings, and the poor predictive value of mouse models for the outcome of Fc/FcR-interactions, it was decided to de-prioritize the further development of IS-Abs with a huIgG SELF framework. This also applied to the targets 4-1BB and OX40 for the following 2 major reasons:
• The overall de-prioritization of work on 4-1BB and OX40-targeting IS-Ab in the IACT program (see WP4)
• Reports on early clinical trials with agonist anti-4-1BB Abs indicating that liver toxicity is a major hurdle with respect to the clinical application and development of anti-4-1BB Abs (Bartkowiak 2018, Clin Cancer Res 24 (4) and references therein)
Task 7.2 Investigation of human IgG2 (h2) as a potential format for generating potent and safe agonist IS-Ab (Partners 3, 4, 6)
Greatly enhanced, FcR-independent agonist activity can be achieved by converting Abs into the huIgG2-format, in particular the IgG2B-isoform, as described in the previous progress report. In view of these findings, the huIgG2-format represents a very interesting new design for agonist immunostimulatory Abs. The key question is whether the enhanced agonist potency of huIgG2-type stimulatory Abs results in an improved therapeutic index, or whether enhanced potency would simply be mirrored by increased toxicity. This has being investigated by partner 3 in a systematic fashion by means of a number of well-defined syngeneic tumor models in human CD40-transgenic mice. These experiments focused on three distinct anti-human CD40 Abs in huIgG2 format:
• Ab ChiLob7/4, originally formatted as huIgG1
• The super-agonist Ab CP-870,893, originally formatted as huIgG2
• Ab 341G2, an antagonist Ab in huIgG1 format that converts into a highly potent agonist in huIgG2 format (see below)
In essence, these experiments demonstrated that the latter two antibodies, although displaying highly potent immune stimulatory activity in in vitro and in vivo assays, were poorly tolerated in that they induced life threatening toxicity in the mice. The latter is in line with the clinical data for Ab CP-870,893, showing that the maximum tolerated dose of this Ab is 10-20-fold lower than that of ChiLob7/4 and SGN40.
In view of the above, the level of agonist activity displayed by huIgG2 ChiLob7/4 is still of great interest, in that it constitutes an intermediate between the ‘conventional’ huIgG1 agonist Ab (ChiLob7/4 & SGN40) and the huIgG2 super-agonist Ab CP-870,873. At patent application on this concept has been filed on Oct 2015, which is currently advanced towards the national phase: Modified antibodies containing modified IgG2 domains which elicit agonist or antagonistic properties and use thereof; WO 2015145360 A1; M. Glennie & A. White.
The limitation of these experiments referred to above is that clear-cut anti-tumor efficacy could only be observed when the agonist Abs were combined with tumor antigen-specific immunization (synthetic peptide vaccines). The conceptual limitation of such experiments is that in this setting one mainly looks at the capacity of the CD40 Abs to act as vaccine adjuvant, while the question of primary interest is whether these Abs can trigger anti-tumor immunity by inducing pro-immunogenic changes in the tumor micro-environment, in particular by activating dendritic cell and modulating myeloid cell function. Moreover, our experiments in Task 3.2 have shown that our mRNA-based vaccines are already very potent ‘as is’ and not significantly enhanced by co-administration of agonist anti-CD40 Abs.
Notably, based on technologies set up in the context of Tasks 5.7 6.5 and 6.6 partner 1 (Heidelberg) has tested chimeric anti-mouse CD40 3/23-muIgG1 in combination with multiple oncology small molecule inhibitors (SMi) in in vivo assays for potential negative impact on DC/T-cell function (OVA immunization assays in OT-1 system), as well as in tumor models for potential synergy in anti-tumor impact with CD40 Ab. In particular, MEK-inhibitors were shown to have no significant negative impact on DC/T-cell function while strongly synergizing with CD40 Ab in anti-tumor efficacy in three different tumor models. Detailed PD biomarker analysis by means of immunohistology, flow cytometry and whole transcriptome analysis showed the following:
• Successfully treated tumors with the combination of MEKi and CD40 Ab showed a shift from M2 macrophages to M1 macrophages, as well as a change in the CD8/Treg T-cell ratio in favor of CD8+ T-cells.
• This impact on the immune phenotype was a result of BOTH drugs together, and could not be achieved to similar extent by either single drug.
• The gene signature of treated tumors showed a strong suppression of cell proliferation/DNA synthesis markers, reflecting the anti-tumor impact, as well as a profound synergy by the drugs, in that neither alone rendered this effect
• Furthermore, a clear pro-immune gene signature was associated with the administration of the CD40 Ab. Interestingly, this effect was strongest in the mice treated with CD40 Ab only, and partly attenuated in mice that also received MEKi. Overall, this shows that the synergy between MEKi and CD40 Ab must be most at the level to tumor suppression, while MEKi partly counteracted the pro-immune impact of CD40 Ab in the tumor models, in spite of having little negative impact in the in vivo OT-1 system.
Further in-depth analysis of drug action is ongoing to gain further insights, with the aim of arriving at the most optimal treatment regimen.
Especially the transcriptome analysis strategy has resulted in very clear-cut gene expression profiles that can form the basis of validated, drug-specific PD biomarker profiles to be applied in clinical trials (see report on Deliverable 5.4).
In view of the above, partners 1 and 3 are proceeding by testing the anti-tumor impact of ChiLob7/4-huIgG2, as well as surrogate anti-mouse CD40 Ab 3/23-huIgG2, in combination with MEK-inhibitors.
In parallel to the development of anti-human CD40 Ab with an optimized IgG framework, partner 3 has deeply invested in obtaining more insight in the mechanism of action of these Abs by zooming in on the fine specificity of the Fab-domain of the Ab. The Southampton-group demonstrated that anti-human CD40 Abs can be divided into four different classes based on their behaviour in functional assays. For reasons of confidentiality related to the protection of intellectual property, the full details of these findings cannot be disclosed in this public report. However, most information can be found in the following publication: Yu, Glennie, White et al. Cancer Cell. 2018 Apr 9;33(4):664-675).
Exploration of the impact of the huIgG2 format on agonist activity of anti-4-1BB and anti-OX40 Abs showed a major increase for 4-1BB only. The reason for the minor increase for OX40 is not known at this time and is somewhat surprising since, also for the target CD28 the huIgG2 format resulted in greatly increased agonist activity. At this time, huIgG2 formatted anti-4-1BB Abs are, however, not considered for further development in view of the toxicity issues already reported concerning anti-4-1BB Abs (see under Task 7.1).
Task 7.3 Generation of glyco-optimized variants of anti-CD40 mAb ChiLob7/4 (Partner 7)
Glyco-optimized variants of ChiLob IgG1 and IgG2 (chimeric and humanized) generated through stable expression in a panel of Glycotope’s proprietary GlycoExpress cell lines display different glycosylation structures, i.e. slightly increased sialylation, levels of N-glycans bearing terminal mannose or hybrid type N-glycans as compared to Abs produced by standard CHO producer cells, but mainly low and high levels of fucosylation. Humanization and glyco-optimization were performed in parallel to differentiate between functional effects due to humanization and glyco-optimization. Glyco-optimized IgG1 and IgG2 variants (chimeric and humanized) of ChiLob7/4 were successfully produced for pre-clinical testing. For reasons of confidentiality related to the protection of intellectual property, the full details of these findings cannot be disclosed in this public report.
Task 7.4 Pre-clinical testing of glyco-optimized mAb in vitro (Partners 3, 4, 6, 7)
The aforementioned variants displayed highly comparable huCD40 binding (Biacore; flow cytometry) and passed QC with respect to purity, low content of aggregates (< 0.4%) and endotoxin.
A main goal of this project was to test whether glyco-optimization could result in increased FcyRIIB-dependent agonistic functions. However, no difference between the glyco-optimized variants was seen in the B-cell assay: all IgG1 Abs (chimeric and humanized) similarly stimulated B-cell proliferation in an X-linking dependent (and dose-dependent) manner, while all IgG2 Abs did so independent of secondary X-linking (X-linking was enabled using an X-linking antibody). However, in line with expectations, the low-fucosylated IgG1 and IgG2 Abs (chimeric and humanized) bound with higher affinity to activatory FcγRIIIA, with huIgG1 Abs having a much higher affinity than huIgG2. The in vitro ADCC assays using CD40 positive cancer cells corroborate these findings. For reasons of confidentiality related to the protection of intellectual property, some details of these findings were omitted from this public report.
In view of the above mentioned data, Glycotope GmbH proposes to focus on further functional assays in the context of different immune cells, instead of producing a larger batch of the Abs concerned.
The aim of these studies is the development of in vitro assays with human immune cells that reflect more closely the relevant tumor microenvironment and a wider repertoire of FcγRs than FcγRIIB. Most of the IACT research/optimization regarding the agonistic activity of ChiLob7/4, is based on the activation of FcγRIIB positive B cells. However, after the IACT project has been started, conflicting results have been published whether the agonistic activity of the human anti-CD40 antibodies in human patients solely depends on FcγRIIB receptors. Additionally, other immune cells, such as monocytes, macrophages and dendritic cells express the CD40 receptor and are known to play important roles in the tumor microenvironment. Therefore, we propose to develop assays with human monocytes, macrophages, NK cells, T cells and dendritic cells to further test the agonistic activity of the humanized and glyco-optimized ChiLob7/4 variants. These cell types express different FcγR types including FcγRIIIA, FcγRIIA and FcγRIIB. Especially activation and expansion of dendritic cells and reprogramming of tumor-associated macrophages are reported to inhibit cancer progression and metastasis. As DCs and macrophages express CD40 in addition to the FcγR, studies analyzing the agonist potential of the different anti-CD40 antibody glyco-variants in more complex assays, mimicking the tumor microenvironment, might reveal functional differences of the glyco-variants that cannot be detected in the classical B-cell assays. Furthermore, potential side effects of an ADCC enhanced huIgG1 should be analyzed. To this end, ADCC assays with primary CD40 positive macrophages and DC will be performed.
While this initial objective of improved FcgRIIB mediated effector cell activation by glyco-optimization was not achieved as described above, the low-fucosylated and humanized version of ChiLob (huIgG1 ChiLob7/4-FF) did show significantly enhanced agonist function as compared to reference ChiLob7/4 Ab, both in the B-cell activation and NF-κB signalling assay, when provided in the context of FcγRIIIA-expressing cells, such as NK-cells and DCs, in line with the propensity of this low-fucose Ab to bind more strongly to FcγRIIIA. In the functional assays concerned, including an allo-MLR T-cell activation assay, huIgG1 formatted ChiLob7/4-FF displays agonist activity that is comparable or even higher than that of huIgG2 formatted ChiLob7/4 while it is almost absent with the huIgG1 ChiLob7/4-SM (high-fucosylated and humanized version).
This finding is in line with data published by Ravetch et al. (Dahan et al. 2016 Cancer Cell 26, 820) and suggests that in the human system secondary cross-linking of agonist CD40 Ab by activatory FcγRIIIA may be as important as that mediated by inhibitory FcγRIIB in potentiating agonist function. Despite the enhanced agonistic activity, it is to mention that glyco-optimization of ChiLob7/4 IgG1 showed increased ADCC activity against CD40 positive tumor cells, but surprisingly not against CD40 positive immune cells (B cells, monocytes, DCs). The later may explain why FcγRIIIa mediated ADCC do not negatively interfere with agonist function. Therefore, the newly developed huIgG1 ChiLob7/4-FF (which has also been fully humanized by Glycotope GmbH) constitutes a very interesting new variant, in that it exhibits increased agonist activity at a level similar to that of huIgG2 ChiLob7/4, albeit in a fashion that – unlike for the huIgG2 version – does depend on secondary cross-linking by FcγR-expressing accessory cells. The latter difference is expected to result in differences in compartmentalization of in vivo agonist function. Notably, the fully humanized huIgG1 ChiLob7/4-FF Ab could be used to generate new IP.
In view of the above, the Glycotope GmbH team is performing further in-depth analysis of the FcγR-dependent agonist function of the low-fucosylated huIgG1-ChiLob7/4 Ab. While this work was included in the IACT extension request, the fact that this request was not granted implies that this work will continue outside of the IACT project.
Task 7.5 Pre-clinical testing of Glyco-optimized mAb in vivo (Partners 3, 4, 6, 7)
This task has been replaced by the follow up work described under task 7.4.
WORK PACKAGE 8. DEVELOPMENT OF OPTIMIZED 2ND-GENERATION AGONIST IS-ABS THROUGH MANIPULATION OF THE FAB-DOMAIN
Objectives
• Generation of dual-specific Abs that simultaneously trigger two complementary immune pathways
• Development of trimeric antibodies with enhanced agonist function
Progress for each task
Task 8.1 Isolation of dual-specific Abs (Partner 8)
The three target antigens CD40, OX40 and 4-1BB have been produced in HEK293EBNA cells as soluble, extra-cellular domains with 8xHis tag. As judged by SDS-PAGE all protein preparations showed high purity and according to size exclusion chromatography the monomeric protein form was dominating (above 80%). ELISA with commercial target-specific antibodies showed specific and dose-dependent binding. In addition, transfected CHO cells expressing the full length proteins have been generated. Flow cytometry analysis with commercial antibodies showed high cell surface expression of all three proteins.
Using the BioInvent n-CoDeR® human antibody library, mono-specific scFv-fragments have been successfully isolated against each of the three targets; 134, 112 and 100 unique scFv for the targets CD40, OX40 and 4-1BB, respectively. Binding of isolated scFv-fragments to soluble proteins was verified using ELISA and binding to cell-surface expressed proteins was verified using Fluorometric Microvolume Assay Technology (FMAT) and flow cytometry. Several of the scFv-fragments exhibit EC50 values in the low-nM range in ELISA.
Using the mono-specific scFv-fragments as starting material, various shuffling and combination procedures were performed. The resulting Ab-fragments were screened for binding to two out of the three antigens CD40, OX40 and 4-1BB
On the basis of these analyses, the following high-affinity, dual-specific Abs were generated and shipped to partners in Southampton, Pamplona and Milan:
• 3x CD40/OX40 dual-specific Abs: #2, #5 and #17
• 2x CD40/4-1BB dual-specific Abs: #4 and #19
• 2x OX40/4-1BB dual-specific Abs: #6 and #20
For reasons of confidentiality related to the protection of intellectual property, the full details of these findings cannot be disclosed in this public report.
Biochemical analysis of the Abs revealed that both expression and folding are comparable to that of conventional Abs. Most dual-specific Abs bind both ligands with EC50 in the sub- to low (single digit) nM range. This is true for binding to purified extracellular domains in ELISA as well as target-transfected and endogenously-expressing cells in flow cytometry.
Task 8.2 In vitro functional testing of dual-specific Abs (Partners 3, 4, 6, 8)
Functional testing of dual-specific Abs comprising these domains by the partners in Southampton, Milan and Pamplona indicated that one of the CD40 specific Ab and all three of the 4-1BB specific Abs tested displayed clear-cut agonist activity. The second CD40 Ab only displayed weak agonist activity, and the same applied to the only available OX40 Ab. In summary, the experiments performed by BioInvent in cooperation with other IACT partners have provided proof of concept for a novel format of dual-specific Abs that is distinct from previously reported designs.
For reasons of confidentiality related to the protection of intellectual property, the full details of these findings cannot be disclosed in this public report.
Task 8.3 Initial testing of a Trimerbody against mouse 4-1BB (Months 0-12; partner 9)
Four Trimerbody constructs against mouse 4-1BB were generated on basis of the scFv fragment derived from the well-defined anti-mouse 4-1BB conventional mAb (clone 1D8, IgG2a). Three 1D8 scFv-based were N-terminal Trimerbodies (1D8N) including different lengths of the flexible linker region connecting the scFv to the murine collagen XVIII trimerization domain: 25-, 5- and 0-mer (1D8N25, 1D8N5, and 1D8N0 respectively); and a short (0-mer 1D8 scFbased C-terminal trimerbody (1D8CO).
All four 1D8 scFv-based Trimerbodies were secreted as soluble proteins by transfected HEK-293 cells and specifically recognized in ELISA the immobilized murine m4-1BB:hFc protein. Trimerbody specificity was further confirmed by FACS on the surface of cells expressing the mouse 4-1BB receptor. Affinity studies (Biacore) against immobilized recombinant mouse 4-1BB Preliminary biosensor kinetics analysis indicates that 1D8 scFv-based N- terminal Trimerbodies have higher avidity and slower off-rates compared to native 1D8 IgG2a. Of interest, two Trimerbody configurations have a dissociation rates 10 times slower compared to 1D8 IgG2a. For reasons of confidentiality related to the protection of intellectual property, the full details of these findings cannot be disclosed in this public report. Size exclusion chromatography-multi-angle laser light scattering signal from oligomeric analysis and molar mass demonstrate that the 1D8 scFv-based Trimerbodies behave predominantly as trimers with none to minor higher molecular weight aggregation.
All Trimerbodies were very stable in human serum as demonstrated by the 80-100% preservation of the binding activity to m4-1BB:hFc even after 4 days incubation time. Additionally, Performing several freeze/thaw cycles of the different Trimerbody samples showed that three out of the four Trimerbodies were very stable when subjected to these conditions. Of interest, one particular configuration (1D8CO) was labile under these conditions.
In collaboration with partner 4 (Melero), In vitro functional assays for testing T cell co-stimulatory impact through mouse 4-1BB and subsequently used to test the signalling impact through the 4-1BB receptor by the mouse 1D8 IgG2a compared to 1D8 scFv-based Trimerbodies. We have now obtained solid and reproducible results supporting the fact that, though to a different extend, three Trimerbodies were superior to the parental 1D8 IgG2a conventional mAb. The T -cell co-stimulation in vitro readouts included cell proliferation/survival, Interferon-γ and IL-2 production. Based on these findings, two preferred 1D8 scFv-based Trimerbody configurations were selected for further studies. For reasons of confidentiality related to the protection of intellectual property, the full details of these findings cannot be disclosed in this public report.
Task 8.4 Generation and initial testing of a Trimerbody against human 4-1BB (Months 12-30; partner 9)
Based on the above findings (Task 8.3) generation of a human trimerbody, on basis of the SAP3.28 Ab from Southampton, focused on generation of an N18-trimer. A QC-ed preparation of trimerbody was produced and found to exhibit excellent binding to human 4-1BB. In view of the outcome of the functional tests with the anti-murine trimerbody (see below), a tumor antigen targeted hexavalent bispecific trimerbody was successfully generated.
Task 8.5 In vivo functional testing of anti-4-1BB Trimerbodies (Months 12-48; partners 1, 3, 4, 9)
Generation of the anti-mouse 4-1BB trimerbody based on the 1D8 monoclonal IgG type Ab was reported previously. More recent data show that the agonist activity of this trimerbody is, unfortunately, not significantly greater than that of the original, bivalent 1D8 Ab. In view of this finding, the trimerbody design was expanded to a hexavalent bispecific design that additionally includes three tumor-targeting domains against a clinically well-validated tumor target (Leadartis chooses not to disclose this target, which is expressed at the cell surface in certain human tumors). This new construct can effectively be produced and binding assays show that it efficiently binds both 4-1BB and the tumor antigen in a simultaneous fashion and slower dissociation rates compared to the 1D8 Ab. Moreover, the agonist activity of this bispecific construct is superior to that of both the parental anti 4-1BB 1D8 Ab and the original monospecific 1D8-based trimerbody when applied in the presence of plastic-immobilized or cell surface-expressed tumor target antigen.
The in vivo half-live of the new bispecific construct is approx. 20 hrs. This is better than that of the monospecific trimerbody, but shorter than that of conventional Abs. As such, it may offer a good compromise between in vivo exposure levels that require long half-life and safety considerations that are mitigated by shorter half-life. Moreover, targeting of 4-1BB agonist activity to the tumor microenvironment may offer an attractive ‘best in class’ strategy for improving the therapeutic index of agonist 4-1BB Abs, which is of interest in view of the fact that the conventional 4-1BB Abs tested thus far in clinic are not well-tolerated (see above). Further in vivo experiments with the bispecific, hexavalent construct are ongoing (beyond the scope of the terminated IACT project).
Potential Impact:
Achievements concerning preclinical research and drug development
The IACT project brought together a team of experts on the subject of agonist immunostimulatory antibodies with the aim of pushing this field, which is lagging behind that of the immune checkpoint blockers, forward. From a scientific point of view, the most prominent output was generated by the lab of Martin Glennie at Southampton University (partner 3) in the form of two high impact papers that were authored by him and Dr. Anne White, the post-doctoral scientist appointed on the IACT project for partner 3:
• Conformation of the human immunoglobulin G2 hinge imparts superagonistic properties to immunostimulatory anticancer antibodies. White AL, Chan HT, French RR, Willoughby J, Mockridge CI, Roghanian A, Penfold CA, Booth SG, Dodhy A, Polak ME, Potter EA, Ardern-Jones MR, Verbeek JS, Johnson PW, Al-Shamkhani A, Cragg MS, Beers SA, Glennie MJ. Cancer Cell. 2015 Jan 12;27(1):138-48.
• Complex Interplay between Epitope Specificity and Isotype Dictates the Biological Activity of Anti-human CD40 Antibodies. Yu X, Chan HTC, Orr CM, Dadas O, Booth SG, Dahal LN, Penfold CA, O'Brien L, Mockridge CI, French RR, Duriez P, Douglas LR, Pearson AR, Cragg MS, Tews I, Glennie MJ, White AL. Cancer Cell. 2018 Apr 9;33(4):664-675.e4
This work has greatly increased our insight in the structure function relationship of agonist immunostimulatory antibodies and is of great value for the design of the next-generation drugs.
IACT has also enabled the lab of Rienk Offringa at German Cancer Research Center (Partner 1) to pursue the search for optimal combinations of cytostatic drugs and agonist immunostimulatory antibodies, through the provision by the Glennie-lab of the gene constructs for mouse chimeric antibodies targeting mouse CD40, OX40 and 4-1BB. This work resulted in the identification of MEK-inhibitors as highly effective companion of agonist anti-CD40 antibodies in tumor treatment, as well as in the identification of whole tumor transcriptomics for PD biomarker analysis of drug action (Baumann, Offringa et al, Pro-immunogenic impact of MEK inhibition synergizes with agonist anti-CD40 immunostimulatory antibodies in T cell mediated tumor therapy, manuscript under consideration for publication in Immunity). Furthermore, the IACT-project supported the analysis of the natural anti-tumor T-cell response in human pancreatic cancer, which will serve as an important base line for the planned neo-adjuvant trial in this indication:
• Identification of a tumor-reactive T-cell repertoire in the immune infiltrate of patients with resectable pancreatic ductal adenocarcinoma. Poschke I, Faryna M, Bergmann F, Flossdorf M, Lauenstein C, Hermes J, Hinz U, Hank T, Ehrenberg R, Volkmar M, Loewer M, Glimm H, Hackert T, R. Sprick MR, Höfer T, Trumpp A, Halama N, Hassel JC, Strobel O, Büchler M, Sahin U & Offringa R. OncoImmunology 2016, DOI: 10.1080/2162402X.2016.1240859
Work by the lab of Ignacio Melero at CIMA, as supported in part by the IACT funding, resulted in several papers that provided further insight in the mechanism of action of agonist anti-4-1BB antibodies, including the following two:
• CD137 (4-1BB) Costimulation Modifies DNA Methylation in CD8+ T Cell-Relevant Genes. Aznar MA, Labiano S, Diaz-Lagares A, Molina C, Garasa S, Azpilikueta A, Etxeberria I, Sanchez-Paulete AR, Korman AJ, Esteller M, Sandoval J, Melero I. Cancer Immunol Res. 2018 Jan;6(1):69-78.
• Mitochondrial Morphological and Functional Reprogramming Following CD137 (4-1BB) Costimulation. Teijeira A, Labiano S, Garasa S, Etxeberria I, Santamaría E, Rouzaut A, Enamorado M, Azpilikueta A, Inoges S, Bolaños E, Aznar MA, Sánchez-Paulete AR, Sancho D, Melero I. Cancer Immunol Res. 2018 Jul;6(7):798-811.
In addition, IACT resulted in the involvement of two SMEs in the development of immunostimulatory agonist antibodies, in particular the vaccine company BioNTech/Ribological (partner 10) and the antibody company Glycotope (partner 7). The resulting engagement of BioNTech in this area of research is best illustrated by the fact that this SME emerged as the driver of the production of clinical grade anti-CD40 Ab ChiLob7/4, starting with the generation of new GMP-grade producer lines and evolving to the generation and QC of the clinical antibody. Moreover, BioNTech took a licence on the CRUK-owned ChiLob7/4 antibody with the aim of developing this further. Furthermore, glyco-optimization as performed by Glycotope resulted in a more powerful anti-CD40 Ab that has been filed in the following IP-application:
• WO2018178046: HUMANIZED ANTI-CD40 ANTIBODIES
For SME Leadartis (partner 9), IACT has been a very effective platform for exploiting their Trimerbody technology, of which in particular the bispecific, hexavalent constructs show great promise with respect to therapeutic index: the mitigation of toxicity by means of tumor targeting and an intermediate (16hr) half-life, as published in:
• A tumor-targeted trimeric 4-1BB-agonistic antibody induces potent anti-tumor immunity without systemic toxicity. Marta Compte, Seandean Lykke Harwood, Ines G. Muñoz, Rocio Navarro, Manuela Zonca, Gema Perez-Chacon, Ainhoa Erce-Llamazares, Nekane Merino, Antonio Tapia-Galisteo, Angel M. Cuesta, Kasper Mikkelsen, Eduardo Caleiras, Natalia Nuñez-Prado, M. Angela Aznar, Simon Lykkemark, Jorge Martínez-Torrecuadrada, Ignacio Melero, Francisco J. Blanco, Jorge Bernardino de la Serna, Juan M. Zapata, Laura Sanz & Luis Alvarez-Vallina. Nature Communications volume 9, Article number: 4809 (2018)
Furthermore, the following patient application was filed:
• EP18382401.0: TRIMERIC POLYPEPTIDE COMPLEXES AND USES THEREOF
Last but not least, SME partner BioInvent (partner 8) was successful in obtaining proof of concept for the use of their n-CoDer antibody libraries towards the generation of Dual (bispecific/bivalent) agonist antibodies. Impressive was the fact that this proof of concept was successfully obtained for all three targets (CD40, OX40, 4-1BB) in a single attempt. A patent application on these results is currently in preparation. Furthermore, BioInvent entered and exclusive collaboration with Glennie-Lab (partner 3) to accelerate the development of optimal anti-OX40 and anti-4-1BB clinical candidate antibodies, resulting in a series of antibodies with superior features as compared to the OX40 and 4-1BB antibodies that formed the basis of the IACT consortium.
Achievements concerning clinical testing of agonist immunostimulatory antibodies
Unfortunately, the primary clinical goal of IACT, the testing of agonist anti-CD40 antibody ChiLob7/4 in studies in patients with head and neck cancer (HNSCC) and pancreatic ductal adenocarcinoma (PDA) has not been achieved during the 4-year project period. The reason for this is the almost incredible number of unforeseen complications that were encountered during the production of the clinical-grade antibody, as described in detail in the progress report under work package 2, task 2.1. Nevertheless, the following two points need to be emphasized in this respect:
• Although the IACT partners involved were under enormous pressure due to these complications, both with respect to time and resources, they never cut corners that could have compromised the safety of the drug preparation
• In spite of the fact that these delays eventually resulted in loss of a significant part of the IACT-related EU funding, due to the fact that the 2-year no-cost project extension was not granted, the work on finalizing the QC trajectory for the clinical ChiLob7/4 antibody is continuing and the clinical studies will be conducted, albeit in modified form and through alternative project funding (see report on work packages 2 and 3). The main reason for the perseverance of the IACT partners in this respect is that the (modified) studies are still state of the art, with high potential significance for the development of this drug class.
Irrespective of the above, the IACT consortium inspired a second clinical trial that, although not focused on agonist immunostimulatory antibodies, develops very well: the HARE-40 trial involving the testing of the HPV16E6/E7-specific synthetic RNA vaccine that is based on the vaccine technology of SME-partner BioNTech (partner 10). Due to the pre-clinical finding that agonist anti-CD40 antibodies did not significantly enhanced the immunogenicity and anti-tumor impact of the HPV16 RNA vaccine, the IACT partners decided to perform a single agent vaccine study in patients with HPV16-positive HNSCC. Also for the preparation of this drug (vaccine), many unanticipated hurdles needed to be overcome, as detailed in the progress report under work package 3, Task 3.2. Although this resulted in a significant delay in the start of the trial, this trial was started before the end of the 4-year project period. The initial results from this study look very promising. In view of this, as well as the equally promising results of similar vaccination studies as performed by BioNTech in patients with metastatic melanoma, also this trial will be continued, in spite of the discontinued EU funding.
Socio-economic impact
Although it is difficult to measure - or even predict - the socio-economic impact of a 4-year research project, it is evident that the four SME-partners were intensely engaged in the IACT consortium and have used this opportunity, each in their own way, to (further) explore in the field of agonist immunostimulatory antibodies. Notably, it is commonly known that only a fraction of the drug development projects started results in a successful clinical drug, and that the path from bench to patient is very long and windy, as also witnessed by this final report. In this context, the IACT collaborative was, and still is, a very good shot at goal.
Dissemination activities
The dissemination activities of the IACT consortium over the three project periods were as follows:
First project period:
1.
During the first project year, we decided on connecting with other EU-consortia that focus on cancer immunotherapy, with the aim to identify potential synergies. To this end, a special workshop on cancer immunotherapy-focused EU-consortia was organized by the IACT coordinator (Offringa) and a representative of partner 10 (Dr. Cedrik Britten) during the 2014 edition of the yearly meeting of the Association for Cancer Immunotherapy (CIMT) in Mainz, Germany. EC Scientific Officer Dr. Jan van de Loo was included in the organization and consulted on selecting the EC projects that were featured during this session. During this session, representatives of each of the selected consortia gave a brief introduction to their project. The program was as follows:
• Introduction; the aims of this Session (Cedrik Britten, Mainz; Germany)
• EC funded consortia and cancer (Jan van de Loo, Brussels, Germany)
• Adoptive Engineered T-Cell Targeting To Activate Cancer Killing - ATTACK (R. Hawkins, Manchester, UK)
• Inflammation and Cancer Research in Europe - INFLA-CARE (A. Eliopoulos, Heraklion, Greece)
• Advanced T-cell Engineered for Cancer Therapy - ATECT (S.A. Quezada, London, UK)
• Glioma Actively Personalized Vaccine Consortium - GAPVAC (H. Singh, Tübingen, Germany)
• Mutanome Engineered RNA Immuno-Therapy - MERIT (U. Sahin, Mainz Germany)
• Immunostimulatory Agonist antibodies for Cancer Therapy - IACT (R. Offringa, Heidelberg, Germany)
• Panel discussion & Summary. How to reach more synergy across consortia? (Moderators: R. Offringa & U. Sahin)
This initial workshop has rendered a number of leads for interactions between these EU consortia. We will follow up on this during the 2015 IACT consortium meeting and during the 2016 edition of the CIMT conference. In view of the latter, The IACT project coordinator (Offringa) has recently been appointed as member of the CIMT board.
2.
IACT has been referred to by representatives of partners 3 (Glennie) and 1 (Offringa), during their talks at the CRI/CIMT/EATI/AACR cancer immunotherapy meeting in New York, Sept 2015.
3.
The 2015 IACT consortium meeting was organized in conjunction with an international conference, organized by IACT partner 5 (Melero) in Pamplona Spain, which featured an excellent line up of experts from the USA and Europe on immunostimulatory Abs. As such, this was an excellent platform for ‘marketing’ the IACT consortium. The program for this meeting looked as follows (IACT partners underlined):
International Symposium: ‘Immunostimulatory monoclonal antibodies and immunomodulation: harvesting the crop’.
Coordinators:
• Dr. Ignacio Melero, Centro de Investigación Médica Aplicada. Clínica Universidad de Navarra. Pamplona
• Dr. Juan José Lasarte, Centro de Investigación Médica Aplicada. Pamplona
Program October 18, 2015:
• Ronald Levy “Targeting the Tumor and Targeting the Host”. Stanford University Medical Center, Stanford, United States
• Drew Pardoll “Protumorogenic immunity: the dark side of the immune system”. Johns Hopkins Medical Institutions, Baltimore, United States
• Jim Allison ”Immune Checkpoint Blockade in Cancer Therapy: New Insights and Opportunities”. MD Anderson Cancer center. Houston. United States
• Antoni Ribas “Melanoma response and resistance to PD-1 blockade” UCLA Medical Center. Los Angeles, California, United States
Program October 19, 2015:
• George Coukos “Understanding tumor microenvironment to optimize immunotherapy. Lessons from ovarian cancer” University of Laussane Medical center, Switzerland
• Ignacio Melero “CD137 and friends in immunotherapy combinations”. Centro para la Investigación Médica Aplicada-Clínica Universidad de Navarra, Pamplona, Spain
• Sergio Quezada “Immune Regulation at the Tumour Site”. UCL Cancer Institute, London, United Kingdom
• Suzanne Topalian “PD-1 pathway blockade: a ‘common denominator’ treatment approach mediates a ‘personalized’
• response”. Johns Hopkins Medical Institutions, Baltimore, United States
• Mario Sznol “Immune Therapy of Advanced Melanoma based on Immune Checkpoint Inhibitors”. School of medicine of Yale. New Haven. United States
• Pam Sharma “Immunotherapy frontiers against prostate and bladder cancer”. MD Anderson Cancer Center. Houston. United States
• Tim Chan “Cancer genomes and immunotherapy efficacy”. Memorial Sloan Kettering Cancer Center, New York city, United States
• Sandra Demaria “Mechanisms of synergy of radiotherapy with antibodies targeting negative immune regulators in the tumor microenvironment”. New York University School of Medicine. New York city, United States
• Janice Taube “Tumor markers of response and resistance to PD-1/PD-L1 checkpoint blockade”. Johns Hopkins Hospital. Baltimore. United States
Program October 20, 2015
• Cornelis Melief “Therapeutic cancer vaccines and their application in combination treatment of cancer” Leiden University Medical Centre. ISA Pharmaceuticals. Leiden, The Netherlands
• Caroline Robert “Melanoma management : New Treatments and new Toxicities” Institute Gustave Roussy, Villejuif, France
• John Haanen ”Checkpoint inhibitors: who should we treat?” The Netherlands Cancer Research Institute. Amsterdam, The Netherlands
• Francisco Sánchez Madrid “Leukocyte activation receptor CD69 and ligands: Immunoregulatory role in inflammatory processes”. Hospital de la Princesa. Madrid, Spain
• Juan José Lasarte “FOXP3 as a target for Immunotherapy”. Centro para la Investigación Médica Aplicada. Pamplona, Spain
• María Luisa Toribio “Therapeutic antibody targeting of human T-cell acute lymphoblastic leukemia initiang cells”. Centro de Biología Molecular Severo Ochoa. Madrid, Spain
• Mario Colombo “Multiple layers of immunoregulation: the myeloid connection”. Instituto Nazionale Tumori Experimental Oncology. Milan, Italy
• Rienk Offringa "A rational approach towards implementation of immunotherapy for pancreatic ductal adenocarcinoma”. German Cancer Research Center, Heidelberg, Germany
• Christian Ottensmeier “Genomic assessment of tumour immune profiles in head and neck cancer”. University of Southampton School of Medicine, Southampton, United Kingdom
• Paolo Ascierto “Combining checkpoint inhibition maximizing the body’s immune response to cancers”. Fondazione Pascale Istituto Nazionale Tumori di Napoli, Italy
• Martin Glennie ”Designing immunostimulatory anti-TNFR mAb to promote anti-cancer immunity”. University of Southampton School of Medicine, Southampton, United Kingdom
Second project period:
4.
IACT coordinator Offringa has hosted a second workshop on EU-funded immunotherapy projects during the 2016 edition of the CIMT meeting. The program of this session was as follows (IACT speakers underlined):
• Rienk Offringa, German Cancer Research Center, Heidelberg, Germany:
• Fostering interaction and synergy between ongoing EU Cancer Immunotherapy Consortia
• Harpreet Singh, Immatics Biotechnologies, Tuebingen, Germany
• GAPVAC: First results from a personalized peptide vaccination study in glioblastoma
• Ann White, Southampton University, Southampton, UK
• IACT: The human IgG2 format imparts potent agonistic properties to immunostimulatory antibodies
• Fred Falkenburg, Leiden University Medical Center, Leiden, The Netherlands
• T-Control: Therapy of hematological malignancies using streptamer-selected cell products
• Jan van de Loo, Scientific and Policy Officer Cancer Research, European Commission, Brussels, Belgium
• EU Cancer (Immunotherapy) Consortia: what we are looking for in the next round.
• Zlato Trajanoski, Medical University of Insbruck, Innsbruck, Austria
APERIM: Advanced Bioinformatics platform for personalized cancer treatment
• Sebastian Kobold, Ludwig Maximilians University Munich, Munich, Germany
• IMMUTRAIN: Training Network for the Immunotherapy of Cancer
Third project period:
During the final project year, the IACT partners focused on consolidating the efforts to make the clinical trials going and applying for the no-cost project extension.
Exploitation of results
The IACT project was set up to allow the SME partners freedom to exploit the proceeds of their work. This applies in particular to the glyco-optimized forms of agonist immunostimulatory antibodies developed by Glycotope (partner 7), the novel dual antibodies and optimal anti-OX40 and 4-1BB antibodies generated by BioInvent (partner 8), the Trimerbodies and dual-specific variants generated by Leadartis (partner 9) and the huIgG1 and huIgG2-forms of the anti-CD40 Ab ChiLob7/4 as licensed by BioNTech (partner 10), as well as their proprietary HPV16E6/E7-specific synthetic RNA vaccine. In view of this, the further development of these drug formats is, at this stage, left at the discretion of these SME partners.
For the exploitation of the pre-clinical results from the academic labs of Glennie (partner 3) and Offringa (partner 1), completion of the delayed clinical studies with anti-CD40 antibody ChiLob7/4 is a critical step, because these studies constitute the best stepping stone towards follow up studies testing optimized antibodies or combination regimes. It is for this reason that the IACT partners involved keep pushing ahead with the ChiLob7/4 clinical studies, in spite of the fact that the request for a no-cost extension of the IACT project was not granted.
The bottom line is that the returns in life sciences and drug development do not come overnight. Maybe the initial 4-year time span chosen for the IACT project was too short; in fact it was proven to be too short. However, given the global competition and resources involved, drug development can only be successful if performed under time pressure. Notably, the ambitious time schedule and work plan was completely in line with the call by the EU for high-risk, high gain projects, as illustrated by the fact that the project was funded on the basis of high scores.
The manner in which the IACT project, as well as several other EC projects that include clinical trials, encounter delays with the execution of clinical studies is ‘food for thought’ for the EC committees designing calls and managing the funding. These committees may be formally correct in restricting the funding to the time frame allotted, allowing project extension under exceptional circumstances only. However, a more realistic and productive manner of judging these matters would be driven by the question whether the consortia concerned are pursuing the goals that have been set and are deploying the EU funding towards achieving these goals. In this respect, it should be emphasized that the extension request for the IACT project included the explicit reservation of the funds that were intended for the clinical trials. In view of this, the rejection of this request – which concerned no more than a no-cost extension involving already allocated EU-funding - could be considered rather non-constructive and bureaucratic.
In view of the latter point, a potential goal for the EC could involve the in-depth evaluation of IACT and other life science drug development projects with respect to the following questions:
• What is the ideal balance between an ambitious, high-risk, high-gain project proposal versus a realistic, down-to-earth proposal?
• Would a proposal of the latter type ever make it through the peer review process?
• Do the EC committees concerned have true experts on board with hands-on experience in the areas of clinical studies and drug development?
• Why is the EU lagging so far behind in drug development and clinical drug testing as compared to the USA and - to an increasing degree – China?
• What exactly are the European handicaps in this respect?
• Which concrete actions could be undertaken to overcome these hardwired handicaps?
List of Websites:
www.IACT-project.eu
Prof. Dr. Rienk Offringa
K.H. Bauer Foundation-endowed professor
Head Div. Pancreatic Cancer Research
Surgery Clinic, University Hospital Heidelberg, Im Neuenheimer Feld 110, 69120, Heidelberg
Head Dept. Molecular Oncology of Gastrointestinal Tumors
Head DKFZ-Bayer Joint Immunotherapeutics Lab.
German Cancer Research Center (DKFZ-D200); Im Neuenheimer Feld 280, 69120 Heidelberg
Phone (secr.): +49 (0)6221-42.3144
Phone (direct.): +49 (0)6221-42.3140
Conference phone:: +49 (0)6221-42.5989
FAX: +49 (0)6221-42.4625
E-mail: r.offringa@dkfz.de
Cancer immunotherapy is beginning to realize its potential in the clinic, but immune checkpoint blockade (ICB) is able to elicit durable responses in only a fraction of patients with immunogenic cancers such as melanoma and lung cancer. Significant sub-groups of patients with these indications do not respond to ICB, and the same applies to patients with other cancer types. A major hurdle in this respect is the lower intrinsic immunogenicity of ICB-resistant tumors. In order to build on the first promising results of ICB, there is a clear need for the exploration of additional drugs and treatment regimens in clinical trials. Agonist immunostimulatory antibodies (IS-Abs) targeting activatory receptors on immune cells are a potential alternative for immune checkpoint inhibitors, in that these Abs have been demonstrated to turn ‘cold’ tumors into ‘hot’ tumors, with notable therapeutic results in pre-clinical models. The aim of the IACT project was to boost the further development of agonist IS-Abs, in particular IS-Abs targeting the activatory immune cell receptors CD40, OX40 and 4-1BB.
A main focus of the IACT project was the clinical development of the anti-human CD40 Ab ChiLob7/4 for which new GMP producer lines were generated that were used for manufacturing of a clinical batch, which is ready to be tested in two clinical indications: head and neck squamous cell carcinoma (HNSCC) and pancreatic ductal adenocarcinoma (PDA). Testing in HNSCC was initially planned in combination with a synthetic RNA vaccine against the E6 and E7 oncoproteins of human papillomavirus type 16. However, pre-clinical and subsequent clinical studies showed that the newly developed liposomal formulation of this vaccine (RNALIP) was already very potent and unlikely to be significantly complemented by anti-CD40 Ab. As a result, clinical testing of RNALIP-HPV16 and ChiLob7/4 is proceeding in parallel, with the aim of systematically comparing the pro-immunogenic impact of these drugs. ChiLob7/4 will furthermore be tested in the neo-adjuvant setting in patients with primary resectable PDA, in view of the fact that this offers a unique opportunity to measure drug impact on tumor microenvironment and intratumoral T-cell immune response. Pre-clinical experiments in newly developed mouse models for PDA demonstrated that anti-CD40 Ab treatment results in profound changes in tumor immune cell infiltrate and immune gene signatures. These findings serve as a template for the evaluation of pharmacodynamic biomarkers in the neo-adjuvant trial.
The second major objective of the IACT project was the development of optimized IS-Abs. A systematic evaluation of anti-human CD40 Abs showed that the immunostimulatory potency is not only dependent on the Fab domain, but also on the Fc-domain. Proper choice of Fc-domain can result in more potent agonist Abs. In the context of IACT, two enhanced variants of ChiLob7/4 were developed: a glyco-optimized®, hypo-fucosylated version, the enhancement of which involves increased secondary cross-linking through Fc-receptor positive bystander immune cells, and a human-IgG2-variant, the immuno-stimulatory action of which is independent of secondary cross-linking. The clinical development of a set of already available anti-human OX40 and anti-human 4-1BB Abs was abandoned in favor of the development of superior Abs that were newly isolated from synthetic human Ab libraries. From the same libraries, a unique set of dual-specific Abs was isolated, the Fab domains of which are capable of targeting any combination of two of the three immunostimulatory receptors concerned, thereby providing the first proof of concept for dual agonist Abs. Using the trimerbody® technology, an optimized, bi-specific anti-4-1BB Ab was generated, the agonist activity of which is targeted to the tumor-microenvironment through recognition of a clinically validated tumor antigen.
Project Context and Objectives:
Mission statement
Every year, more than 3 million Europeans are diagnosed with cancer, pointing to the continued importance of research into cancer prevention, diagnosis, treatment and patient care. Surgical resection still offers the best curative option, but in many cases cancer cells have already spread before the primary tumour is removed. Radiation and cytotoxic drugs offer treatments with the potential to ‘mop-up’ cells which escape surgery, but these are ‘blunt’ instruments of destruction which lack specificity and hence provide varying degrees of success. This leaves a desperate need for alternative, conceptually distinct, treatment approaches. Immunotherapy now offers such an alternative, with a realistic expectation of success in a wide range of aggressive cancers. In the last decade numerous approaches have shown that the immune system can be ‘trained’ and strengthened to both engage and destroy malignant cells. Furthermore, such responses are durable with the potential to provide life-long protection. Inevitably, these breakthroughs in understanding of anti-cancer immunity have resulted in the clinical testing of exciting new immunotherapy drugs, including some which are now approved for clinical use. However, despite pioneering work by European leading scientists, the USA currently dominates this field. Our IACT concerted action presents a unique opportunity for European academia and industry to gain the ascendency and be at the forefront of delivering effective immune-based cancer treatments. IACT will focus on the clinical translation of a drug concept shown to be highly promising in pre-clinical studies, for which the IACT consortium has achieved a highly competitive position: agonist immunostimulatory antibodies (IS-Abs).
Scientific background of therapeutic concept
a. Immunotherapy of cancer
The immune system not only protects us against infectious pathogens, but also against cancer. Unfortunately, this protection is not perfect and immune failure can lead to disease. Since the development of the small pox vaccine in the 18th century, we have learned to harness the immune defence against pathogens and thereby to control many infectious diseases. Until recently, the clinical efficacy of immune intervention strategies against cancer was low, mainly because the immunological difference between tumours and normal tissue is much more subtle than between pathogen-infected and healthy cells. Consequently, the development of effective immunotherapeutic strategies against cancer requires comprehensive insight into the fine-tuning of the immune system. Importantly, critical advances in this field have resulted in the FDA-approval of two cancer immunotherapy drugs: a cell-based vaccine for use in metastatic prostate cancer (Sipuleucel-T) and an immunostimulatory antibody (IS-Ab) for use in metastatic melanoma (Ipilimumab). Both drugs enhance the patient’s T-cell immune response against cancer. This success confirms the role of the immune system in controlling cancer and encourages further development of cancer immunotherapy drugs to improve on the results seen to date.
b. Antibody-based immunotherapeutics as compared to other approaches
Monoclonal antibodies (mAbs) constitute a highly attractive drug format, as illustrated by the successful clinical development of the cancer-targeting blockbuster drugs Rituximab, Trastuzumab/Herceptin and Bevacizumab/Avastin. Compared to other immunotherapeutic strategies, IS-Abs have the following major advantages:
• Manufacturing is well established, resulting in molecularly defined drugs that are stable, display excellent pharmacokinetics (PK) and can be modified by a wide array of technologies.
• IS-Abs have a broader application than cancer vaccines and adoptive cell therapy, the use of which is generally limited to patients with cancers displaying pre-identified antigens. Moreover, effective IS-Abs may be used in combination with these antigen-specific approaches to enhance their efficacy.
c. Immunostimulatory antibodies (IS-Abs) as cancer therapeutics
The immune response is regulated through cell surface receptors that stimulate or inhibit the activity of immune cells. While activating signals enable immune responses against harmful pathogens, inhibitory signals prevent uncontrolled immune reactivity against harmless substances (as in allergies) or normal somatic tissues (auto-immunity). Triggering the inhibitory receptors is a major mechanism by which cancer switches off the immune system. Pre-clinical experiments demonstrated that this inhibition can be reversed in two conceptually distinct ways. Antagonist IS-Abs such as Ipilimumab can be used to unleash immunity by blocking the inhibitory receptors (inhibiting the inhibitor). Alternatively, agonist IS-Abs jump-start immune cells by triggering activatory receptors (stimulating the stimulator). Both types of IS-Abs were shown to induce marked tumour regression in mice and to provide long lasting immunity against tumour recurrence [1-16]. While antagonist IS-Abs are extensively being tested in human patients, the clinical potential of agonist IS-Abs remains to be explored.
d. The importance of developing agonist IS-Abs in addition to antagonist IS-Abs
There is significant attraction to make agonist IS-Abs available for patient treatment in addition to the antagonist IS-Abs that are already in the clinic, first of all because there is ample evidence from pre-clinical tumour models that this approach can be at least as efficacious. In line with the notion that these IS-Ab classes represent two conceptually distinct ways of mobilizing immunity, the efficacy of agonist and antagonist IS-Abs differs for the mouse models studied. Furthermore, analyses of human tumours revealed that different mechanism of immune suppression underlie immune failure in cancer (sub-)types, pointing at the need for multiple treatment options/drugs to overcome immune failure in human cancer (personalized medicine). This need is supported by the notion that only a fraction of the melanoma patients respond to treatment with Ipilimumab (anti-CTLA-4), while clinical results for these antagonists, often referred to as ‘checkpoint blockers’ (also including anti-PD-L1), differ between individual patients and cancer types. Finally, combining agonist and antagonist IS-Abs in mice often work synergistically, pointing at the merit of exploring this approach in the clinic [5,12,16-21].
In spite of this strong rationale, clinical development of agonist IS-Abs has been stalled for several years. This is mainly due to an incident in the U.K. in 2006, in which administration of a highly potent agonist Ab against the immunostimulatory receptor CD28 (TGN1412) to healthy volunteers resulted in a life threatening cytokine-release syndrome, due to the uncontrolled activation of immune cells [22,23]. Importantly, it is now generally agreed that this clinical study was not conducted in an appropriate manner, while the underlying pre-clinical studies offered insufficient conceptual basis for the intended aim of suppressing – rather than enhancing – immunity by an agonist IS-Ab.
With resolution of these concerns, the clinical development of agonist IS-Abs as cancer therapeutics is currently regaining momentum, as illustrated by recent activity of USA industry and academia in this area, in particular with respect to anti-CD40 IS-Abs. Notably, pioneering pre-clinical work has been performed in Europe, including that by partners 1, 3, 4 and 6 in our consortium [2-6,9,11,13,24,25]. Moreover, partner 3 has recently concluded a successful phase I study with an agonist anti-CD40 IS-Ab (ChiLob7/4) [26] and, with support from Cancer Research UK, is also developing lead agonist Abs targeting the stimulatory receptors OX40 and 4-1BB. The availability of these clinical antibodies to the IACT consortium, together with the long-standing expertise of our team in this area of research, create a unique opportunity for European academia and industry to bring this promising class of drugs into the clinic and thereby develop new treatment options for cancer patients.
e. Selection of immunostimulatory pathways for targeting with agonist IS-Abs
The reactivity of immune cells is regulated by a great diversity of cell surface receptors [7,17,18]. We identified CD40, OX40 (CD134) and 4-1BB (CD137), all members of the tumour necrosis factor-receptor (TNFR) family, as the most interesting targets for agonist IS-Ab based on the following criteria:
• efficacy: striking pre-clinical evidence for therapeutic impact against tumours
• safety: the expectation that immune related adverse events will be manageable
• complementarity: enhancement of anti-cancer immunity through distinct mechanisms
The mechanism of action of anti-CD40 Abs primarily involves direct stimulation of innate myeloid cells and indirect enhancement of T-cell immunity through activation of dendritic cells [1,4,6,25,27], which makes this pathway complementary to direct T-cell stimulation by agonist IS-Abs against OX40 and 4-1BB. Anti-4-1BB Abs primarily stimulate CD8+ T-killer cells, whereas anti-OX40 Abs primarily potentiate the CD4+ T-helper response and were found to suppress CD4+ T-regulatory cells [3,8,9,11,13,15,16,28-30]. Importantly, T-cell stimulation through OX40 and 4-1BB is less potent than that mediated by the major co-stimulatory receptor CD28, and their triggering in the absence of T-cell receptor stimulation was not found to result in T-cell activation. Furthermore, OX40 and 4-1BB are – unlike CD28 - not constitutively expressed at the T-cell surface. This significantly reduces the risks of serious adverse events due to uncontrolled immune cell activation.
IACT research strategy and objectives
In conclusion, the recent progress in cancer immunotherapy indicates that this approach can offer new treatment options to meet the urgent unmet medical need of cancer patients. Compelling pre-clinical data indicates that agonist IS-Abs constitute a promising drug class in this respect, while lead/candidate agonist IS-Abs targeting three immune receptors are available to the IACT consortium. Moreover, our pre-clinical data point at concrete opportunities for further optimization of IS-Abs (§1.2.7). Taken together, this offers a unique starting point for testing the available drugs in the clinic, while working on the development of a 2nd generation of optimized drugs.
These considerations inspired the three-pronged IACT research strategy shown in Figure 2, which encompasses the following main objectives:
1. To drive further clinical development and testing of the available lead IS-Abs through:
o obtaining clinical proof of concept for the efficacy of anti-CD40 IS-Ab ChiLob7/4 in combination with chemotherapy in pancreatic cancer
o obtaining clinical proof of concept for the efficacy of anti-CD40 IS-Ab ChiLob7/4 with antigen-specific vaccination in human papillomavirus type 16 (HPV16+) head and neck cancer
o providing state-of-the-art anti-OX40 and 4-1BB IS-Abs to be used for first in human testing
2. To optimally support the successful development of IS-Ab drugs through
o the identification of pharmacodynamics (PD) biomarkers that are related to therapeutic efficacy, and the development of validated, harmonized assays for their measurement in the context of clinical trials
o the development of validated, harmonized in vitro and in vivo models to guide the pre-clinical development of IS-Ab drugs and their best possible application in clinical trials
3. To set up a pipeline for the development of 2nd-generation IS-Abs with improved therapeutic index along four innovative approaches that are based on the latest insights in IS-Ab function as well as on proprietary technologies of our SME partners.
References:
1. Beatty GL, Chiorean EG, Fishman MP, Saboury B, Teitelbaum UR, Sun W, Huhn RD, Song W, Li D, Sharp LL, et al.: CD40 agonists alter tumor stroma and show efficacy against pancreatic carcinoma in mice and humans. Science 2011, 331:1612-1616.
2. Diehl L, den Boer AT, Schoenberger SP, van der Voort EI, Schumacher TN, Melief CJ, Offringa R, Toes RE: CD40 activation in vivo overcomes peptide-induced peripheral cytotoxic T-lymphocyte tolerance and augments anti-tumor vaccine efficacy. Nat Med 1999, 5:774-779.
3. Diehl L, van Mierlo GJ, den Boer AT, van der Voort E, Fransen M, van Bostelen L, Krimpenfort P, Melief CJ, Mittler R, Toes RE, et al.: In vivo triggering through 4-1BB enables Th-independent priming of CTL in the presence of an intact CD28 costimulatory pathway. J Immunol 2002, 168:3755-3762.
4. French RR, Chan HT, Tutt AL, Glennie MJ: CD40 antibody evokes a cytotoxic T-cell response that eradicates lymphoma and bypasses T-cell help. Nat Med 1999, 5:548-553.
5. Gray JC, French RR, James S, Al-Shamkhani A, Johnson PW, Glennie MJ: Optimising anti-tumour CD8 T-cell responses using combinations of immunomodulatory antibodies. Eur J Immunol 2008, 38:2499-2511.
6. Honeychurch J, Glennie MJ, Johnson PW, Illidge TM: Anti-CD40 monoclonal antibody therapy in combination with irradiation results in a CD8 T-cell-dependent immunity to B-cell lymphoma. Blood 2003, 102:1449-1457.
7. Melero I, Hervas-Stubbs S, Glennie M, Pardoll DM, Chen L: Immunostimulatory monoclonal antibodies for cancer therapy. Nat Rev Cancer 2007, 7:95-106.
8. Melero I, Murillo O, Dubrot J, Hervas-Stubbs S, Perez-Gracia JL: Multi-layered action mechanisms of CD137 (4-1BB)-targeted immunotherapies. Trends Pharmacol Sci 2008, 29:383-390.
9. Melero I, Shuford WW, Newby SA, Aruffo A, Ledbetter JA, Hellstrom KE, Mittler RS, Chen L: Monoclonal antibodies against the 4-1BB T-cell activation molecule eradicate established tumors. Nat Med 1997, 3:682-685.
10. Melief CJ, Van Der Burg SH, Toes RE, Ossendorp F, Offringa R: Effective therapeutic anticancer vaccines based on precision guiding of cytolytic T lymphocytes. Immunol Rev 2002, 188:177-182.
11. Piconese S, Valzasina B, Colombo MP: OX40 triggering blocks suppression by regulatory T cells and facilitates tumor rejection. J Exp Med 2008, 205:825-839.
12. Takeda K, Okumura K, Smyth MJ: Combination antibody-based cancer immunotherapy. Cancer Sci 2007, 98:1297-1302.
13. Valzasina B, Guiducci C, Dislich H, Killeen N, Weinberg AD, Colombo MP: Triggering of OX40 (CD134) on CD4(+)CD25+ T cells blocks their inhibitory activity: a novel regulatory role for OX40 and its comparison with GITR. Blood 2005, 105:2845-2851.
14. Vonderheide RH: Prospect of targeting the CD40 pathway for cancer therapy. Clin Cancer Res 2007, 13:1083-1088.
15. Wang C, Lin GH, McPherson AJ, Watts TH: Immune regulation by 4-1BB and 4-1BBL: complexities and challenges. Immunol Rev 2009, 229:192-215.
16. Weinberg AD, Morris NP, Kovacsovics-Bankowski M, Urba WJ, Curti BD: Science gone translational: the OX40 agonist story. Immunol Rev 2011, 244:218-231.
17. Pardoll DM: The blockade of immune checkpoints in cancer immunotherapy. Nat Rev Cancer 2012, 12:252-264.
18. Peggs KS, Quezada SA, Allison JP: Cancer immunotherapy: co-stimulatory agonists and co-inhibitory antagonists. Clin Exp Immunol 2009, 157:9-19.
19. Postow MA, Callahan MK, Wolchok JD: The antitumor immunity of ipilimumab: (T-cell) memories to last a lifetime? Clin Cancer Res 2012, 18:1821-1823.
20. Robert C, Thomas L, Bondarenko I, O'Day S, M DJ, Garbe C, Lebbe C, Baurain JF, Testori A, Grob JJ, et al.: Ipilimumab plus dacarbazine for previously untreated metastatic melanoma. N Engl J Med 2011, 364:2517-2526.
21. Topalian SL, Hodi FS, Brahmer JR, Gettinger SN, Smith DC, McDermott DF, Powderly JD, Carvajal RD, Sosman JA, Atkins MB, et al.: Safety, activity, and immune correlates of anti-PD-1 antibody in cancer. N Engl J Med 2012, 366:2443-2454.
22. Horvath C, Andrews L, Baumann A, Black L, Blanset D, Cavagnaro J, Hastings KL, Hutto DL, MacLachlan TK, Milton M, et al.: Storm forecasting: additional lessons from the CD28 superagonist TGN1412 trial. Nat Rev Immunol 2012, 12:740; author reply 740.
23. Hunig T: The storm has cleared: lessons from the CD28 superagonist TGN1412 trial. Nat Rev Immunol 2012, 12:317-318.
24. Melero I, Johnston JV, Shufford WW, Mittler RS, Chen L: NK1.1 cells express 4-1BB (CDw137) costimulatory molecule and are required for tumor immunity elicited by anti-4-1BB monoclonal antibodies. Cell Immunol 1998, 190:167-172.
25. Schoenberger SP, Toes RE, van der Voort EI, Offringa R, Melief CJ: T-cell help for cytotoxic T lymphocytes is mediated by CD40-CD40L interactions. Nature 1998, 393:480-483.
26. Johnson PW, Glennie M: manuscript in preparation.
27. Diehl L, Den Boer AT, van der Voort EI, Melief CJ, Offringa R, Toes RE: The role of CD40 in peripheral T cell tolerance and immunity. J Mol Med (Berl) 2000, 78:363-371.
28. Ito T, Wang YH, Duramad O, Hanabuchi S, Perng OA, Gilliet M, Qin FX, Liu YJ: OX40 ligand shuts down IL-10-producing regulatory T cells. Proc Natl Acad Sci U S A 2006, 103:13138-13143.
29. So T, Croft M: Cutting edge: OX40 inhibits TGF-beta- and antigen-driven conversion of naive CD4 T cells into CD25+Foxp3+ T cells. J Immunol 2007, 179:1427-1430.
30. Vu MD, Xiao X, Gao W, Degauque N, Chen M, Kroemer A, Killeen N, Ishii N, Li XC: OX40 costimulation turns off Foxp3+ Tregs. Blood 2007, 110:2501-2510.
Project Results:
Below is a detailed description of the output of the experimental work packages WP2 – WP8.
WORK PACKAGE 2. COMBINATION OF ANTI-CD40 IS-AB WITH GEMCITABINE IN PATIENTS WITH PANCREATIC DUCTAL ADENOCARCINOMA.
Objectives
• Test the primary hypothesis that agonist anti-CD40 IS-Abs are capable of mobilizing strong systemic and intra-tumour anti-cancer immunity, in particular when combined with non-immunoablative chemotherapy
• Reproduce the synergy between gemcitabine and anti-CD40 Abs observed in mouse tumour models
Progress for each task
Task 2.1 Preparation and quality control of a clinical batch of ChiLob7/4 (Partner 10)
Generation of new Ab producer lines
BioNTech (partner 10) took on the task to generate new Ab producer lines, because the original, adherent producer line from Southampton (partner 3) could not be successfully adapted to suspension culture. It was decided to make new producer lines for both the original IgG1 version of this Ab, as well as an IgG2 version in view of greater agonist potency of the latter (see WP7, Task 7.2). After codon optimization of the gene constructs, excellent producer lines were obtained. Cell banking and QC were completed in a timely fashion.
Production of a GMP batch of CD40 antibody
GMP batch manufacturing of ChiLob7/4 could successfully be completed within the project runtime, despite unforeseen hurdles encountered. BioNTech, cooperating IACT partners and the CR-UK Biotherapeutics Development Unity (BDU) mitigated such hurdles by allocating additional resources, so to mitigate delays in the best possible manner. A comprehensive report on this process is provided below.
Manufacturing of ChiLob7/4 Drug Substance at BDU (CR-UK)
The aforementioned producer cell line was then transferred to a contract manufacturing organization (CMO) for process optimization and GMP drug substance manufacturing. The BDU of Cancer Research UK (CR-UK) was selected by BioNTech as subcontractor for the manufacturing services. However, several hurdles, including lengthy contract negotiations with the license giver of the expression vectors used for producer clone generation, were encountered while planning Ab production at the BDU. Despite the prompt and solution-oriented reaction by the consortium, and the commitment of BioNTech and CR-UK to considerable extra investments, the start of ChiLob7/4 manufacturing was thereby delayed by approximately one year. Although a manufacturing agreement between CR-UK and BioNTech could be signed in November 2014, the first available slot for manufacturing of the GMP batch at the BDU was then only in April 2016. Given that after manufacturing, additional time is required for quality control testing, GMP reporting and QP batch certification, the Deliverable 2.1/3.3 was then delayed by 13 months (expected: May 2016 as opposed to April 2015, as originally planned).
A 250 L pilot/engineering run was scheduled for Nov/Dec 2015. Cell expansion and all in-process controls looked very good, until a sudden drop in cell viability was detected. The CMC experts at the BDU together with BioNTech evaluated the impact of this finding and came to the conclusion that the run needed to be abrogated and a repetition had to be scheduled prior to the GMP run. Although the BDU at that time was fully booked, full commitment from all partners and the BDU in particular secured a slot for a repetition run directly in January 2016. The repetition run was successfully completed in March 2016, yielding approx. 500 g of pilot bulk drug substance. Ensured by this promising result, and enabled by the BDU re-scheduling other activities, the GMP run was successfully performed in April 2016 as planned. QC testing was completed in June 2016, and a GMP certificate for the Bulk Drug Substance (BDS) was issued in July 2016. In sum, the issues and hurdles described above resulted in a total delay of 15 months.
ChiLob7/4 Fill & Finish and Quality Control Testing
For ChiLob7/4, it was planned to directly proceed to fill & finish at the BDU without further processing steps, thereby producing the ChiLob7/4 drug product (DP) for clinical use. Just prior to GMP manufacturing of the BDS, however, revised calculations of the number of vials required for both clinical trials indicated that the intended batch size of 1,000 vials might be insufficient to ensure safe drug supply. This was in part due to a re-calculation of doses to be applied in the clinics, and in part due to a more extensive stability testing program set up for both ChiLob7/4 BDS and DP. The latter was initiated in response to the initial notion by the MHRA that ChiLob7/4 manufactured using a novel producer cell line may classify as a new IMP (from a manufacturing perspective) which would require more extensive quality and stability testing due to limited experience with the product. This view was confirmed by the Paul-Ehrlich Institute (PEI), the German national authority, in a scientific advice meeting in July 2016. As the established fill process at the BDU is only validated for fill campaigns of up to 1,000 vials, an alternative fill house had to be identified, qualified, and sub-contracted by BioNTech. BioNTech directly obtained quotes from two different fill houses. Fill & finish could earliest been performed in February 2017. This lack of capacity at the external fill houses alone therefore caused a delay of another 7 months, amounting to 22 months in total. Importantly, filling of >3,700 vials, each containing 2.6 mL of ChiLob7/4 DP formulated at 20 mg/mL, has successfully been performed in February 2017.
Quality control testing was then delayed to June 2017. This was due to the necessity for further method optimization for one release relevant assay (glycosylation analysis). Method optimization was promptly addressed and successfully completed, thereby allowing for complete QC testing of the clinical batch of ChiLob7/4 DP vials in a timely manner. According to the manufacturer, the product passed all release criteria. The necessary documents for GMP batch certification (i.e. the Certificate of Analysis (CoA) and the Certificate of Conformity (CoC) were issued by the manufacturer. With this accomplishment Tasks 2.1 was completed with an overall delay of 27 months.
ChiLob7/4 Labelling and Preparatory Activities for IMPD Submission
In order to start the clinical trials, a number of activities still need to be completed.
1) Completion of the CMC section for the IMPD
2) Labelling of ChiLob7/4 with labels approved by the national competent authorities and set-up of ChiLob7/4 distribution to clinical sites
3) Establishing an investigator-initiated trial (IIT) agreement and a quality assurance agreement (QAA) between BioNTech and Heidelberg (sponsor of IAPACA-NEO trial)
The status of each of these tasks is outlined in the following paragraphs:
1. Completion of the CMC section for the IMPD [still ongoing]
IMPD preparation was started in a timely manner already end of 2016. For the preparation of the IMPD, all activities to address comparability between the old Southampton batch and newly generated batch of ChiLob7/4 have been completed successfully, in close cooperation with the Southampton team and in line with requests by PEI as communicated during the Scientific Advice session. These experiments revealed only slight differences between the two batches, the only significant difference being lower activity of the new batch in an in vitro CDC (complement-dependent cytotoxicity) assay, which is likely to be related to differences in glycosylation as reported during last year’s IACT meeting. Based on these data, the safety profile of the new ChiLob7/4 is expected to be very similar to that of the old batch. In addition, these data support the Sponsors’ position that the new batch of ChiLob7/4 does not constitute a new IMP, thereby significantly decreasing the regulatory risks initially associated with the clinical testing of ChiLob7/4 in two clinical phase I trials (sponsored by partners Heidelberg and Southampton).Furthermore, the lack of binding of ChiLob7/4 to mouse CD40, expected on basis of the fact that this Ab was derived from a mouse monoclonal Ab raised in wild type mice that express mouse CD40 on all relevant tissues, was confirmed. This supported the notion that mice are not a relevant species for further toxicity studies. Although not explicitly requested by PEI, tissue cross-reactivity (TCR) analysis by means of IHC using tissue micro-arrays has been successfully performed at BioNTech, with the aim to provide further evidence for comparability between demonstrate comparability between the old and new ChiLob7/4 preparations. As the data that have been presented to the PEI at the Scientific Advice included such a (TCR) study with the old Southampton batch, it was decided to perform this check pro-actively with the newly generated batch as well to avoid further delays in approval. Despite the new ChiLob7/4 batch being expected not to be classified as a new IMP by regulatory authorities, partners xx and yy, together with the regulatory experts at BioNTech conluded that an ‘In-use compatibility study’ needs to be performed prior to submission (regulatory de-risking). This study is scheduled to be conducted by the BDU in March 2019. The main goal of this study is to test the antibody with the infusion system. According to the latest time lines, the full and comprehensive IMPD will be available for submission in April 2019.
2. Labelling of ChiLob7/4 with labels approved by the national competent authorities and set- up of ChiLob7/4 distribution to clinical sites
Labelling of the primary packaging is usually performed right after filling of the drug product into the vials, and prior to freezing of the DP. However, as the IMP is used for two independent trials which each run in two countries and therefore require labels in different languages, and as the respective clinical study protocols of these trials have not been finalized at the time of filling yet, it has been decided that the vials will be labelled only after approval of the clinical study protocol and therefore approval of the labels by the competent authorities. This minimizes the risk of re-labelling of the vials, which can be technically challenging if the product is stored frozen and may not thaw during re-labelling (as is the case for ChiLob7/4 DP), due to additional requests by the competent authorities, which would delay the provision of the study medication. Another advantage of this strategy is, that the vials can be labelled based on the actual recruitment rate of the trials (clinical sites) which minimizes ChiLob7/4 DP loss. As a consequence of this decision, labelling needs to be performed at the long-term storage temperature (-20±5°C) of ChiLob7/4 DP. To this end, every effort has been undertaken to identify material that is suitable for labelling at this temperature and to select a CMO that provides this service. The respective CMO could successfully be qualified in a GMP-audit by BioNTech in June 2017, and will perform storage, labelling and distribution of ChiLob7/4 DP in compliance with GMP.
3. Establishing an investigator-initiated trial (IIT) agreement and a quality assurance agreement (QAA) between BioNTech and Heidelberg (sponsor of IAPACA trial) [still ongoing]
Prior to start of the clinical trial IAPACA-NEO an IIT-agreement that clearly describes certain tasks, e.g. provision of study medication, which will be delegated to BioNTech by the sponsor, needs to be established. A similar IIT- agreement between BioNTech and Southampton (sponsor of the HARE-40 trial), has already been established for the HARE-40 trial and was used as a basis for the IIT-agreement between Heidelberg and BioNTech. Mature draft documents are available and will be finalized prior to trial start (IAPACA-NEO).
Impact of delays with clinical ChiLob7/4 on project time lines and clinical trial strategy:
In the course of 2016, it became evident that we could not adhere to the initial time schedule for the clinical studies with anti-CD40 Ab ChiLob7/4 (IAPACA trial) as shown in Fig. 1, because of the following 2 main reasons:
• Delay of the production and QC of the clinical batch of ChiLob7/4 Ab
• Information by the PEI that they may consider the new batch of ChiLob as a novel IMP due to subtle biochemical differences with the old batch of ChiLob7/4
In view of this, timing and design of the studies were evaluated and adjusted (see Fig 2), as extensively discussed during the 4th IACT consortium meeting in October 2016 and in the 2nd periodic report. The main changes as compared to the original trial schedule were:
• Start of the IAPACA-1 phase I dose escalation trial in Q1 of Year 5, instead of Q3 of year 2
• An increase of trial duration to 2 years, due to the need for subsequent enrollment of patients
• Start of the IAPACA-2 neo-adjuvant in Q1 of year 7, instead of Q3 of year 3.
Based on the adjusted time schedule, it was decided to apply for a 2-year no-cost extension of the IACT project in order to enable the IACT partners to complete at least the majority of clinical studies planned, including the IAPACA-1 study.
In the meantime, pre-clinical experiments in the lab of the IACT coordinator focused at the pharmacodynamics of agonist anti-CD40 Ab and cytostatic drugs (DoW Task 5.7) revealed that (i) transcriptional profiling of tumor tissue of treated animals constitutes a powerful strategy for studying drug-related PD biomarkers and (ii) that the PD biomarker signature induced by anti-CD40 Abs is much more clear if not combined with cytostatic drugs. These insights led to re-evaluation of the neo-adjuvant study design in the course of 2017, as shown in Fig 3. The main changes were:
• Planning of the IAPACA-2 neo-adjuvant study as a single agent study with ChiLob7/4, including a dose escalation phase to accommodate the fact that PEI will most likely regard this drug as novel IMP
• Execution of this study with the scope of the 2-year extension period applied for.
In essence, this amendment implies that also the neo-adjuvant study aimed at analyzing the pharmacodynamics of ChiLob7/4 in the tumor microenvironment would be concluded within the (extended) running time of the IACT project. These amendments are discussed in detail in the following pages.
In June 2018, it was decided to further amend trial design for the following reasons:
• Further unanticipated delays in completion of the IMPD.
• Additional pre-clinical data from the lab of the Coordinator, now obtained in three independent mouse tumor models including a clinically relevant pancreatic cancer model, supporting the notion that PD biomarker data on the intra-tumoral impact of agonist anti-CD40 Ab is best obtained by means of administration of anti-CD40 Ab as single agent.
• Very recent information from the ASCO meeting (Chicago; May 31-June 4, 2018) based on large 493-patient randomized trial (PRODIGE 24/CCTG PA.6 trial), indicating that FOLFIRINOX is superior and preferable to Gemcitabine as adjuvant treatment after surgery: http://www.ascopost.com/News/58907(opens in new window). The clinicians from partners 1 and 3 involved in the planning of the IAPACA trial have signaled that this new development will significantly complicate the accrual of patients for the IAPACA-1 study, because FOLFIRINOX and Gemcitabine/Abraxane are also gradually replacing Gemcitabine mono-agent treatment in patients with non-resectable disease, and the recently published study will further enhance this shift. In essence, this marks the conclusion of a discussion that has been going on over the past 3 months concerning the gradually decreasing feasibility of the IAPACA-1 study as a result of ongoing changes in clinical practice.
• Further information from the recent ASCO meeting that patients with primary resectable cancer benefit from neo-adjuvant treatment, based on a 246-patient randomized trial (PREOPANC-1): http://www.ascopost.com/News/58908(opens in new window). Based on this study, the interest in the field in neo-adjuvant studies like IAPACA-2A will significantly increase.
In view of these new developments, the IACT partners involved in the planning of the IAPACA trial agreed to further adapt the work plan as follows (see Fig. 4):
• Focus all efforts on the neo-adjuvant single agent study (renamed to IAPACA-NEO) with anti-CD40 Ab ChiLob7/4, as described under Task 2.3.
• Abandon IAPACA-1 as described under Task 2.2
Task 2.2 Dose escalation safety study in non-resectable metastatic PDA. (Partners 1, 2, 3)
As explained above, the plan to pursue this study was abandoned in June 2018
Task 2.3 Neo-adjuvant study: Mechanism of action neo-adjuvant study in resectable PDA (Partners 1, 2, 3)
As a result of cumulative delays in the preparation and QC of the clinical drug, chances in clinical practice and progressing pre-clinical insights, the design of the clinical study with anti-CD40 Ab ChiLob7/4 in PDA was amended multiple times (see above). We are currently aiming for a single agent study that focuses on recording the biological impact of the anti-CD40 Ab in the tumor, as shown in Fig.4.
Basis for these insights was the plan, as formulated in IACT project Task 5.7 to discover new PD biomarkers related to the action of agonist anti-CD40 Ab by performing unbiased mRNA gene expression profiling of tumor samples obtained from mice treated with anti-CD40 Ab alone or in combination with cytostatic drugs. While these experiments confirmed the therapeutic synergy between anti-CD40 Abs and cytostatic drugs, they also revealed that the immune gene expression signature in the tumor was much more pronounced in mice treated with anti-CD40 Ab as single agent as compared to mice receiving both CD40 Ab and cytostatic drug (see below). Even though the paradoxical aspects of these findings still need to be elucidated, it is evident that – at least based on our state of the art pre-clinical data – the likelihood of detecting a PD biomarker signature in the neo-adjuvant trial setting is much higher if the patients are dosed with anti-CD40 Ab ChiLob7/4 only.
Importantly, the detection of this PD biomarker signature is, and has always been, the primary aim of this trial, in that this would embody the first direct evidence that administration of anti-CD40 Ab does not only result in systemic exposure, but also in intra-tumoral drug exposure and delivery of its agonist signal to CD40 positive immune cells.
In addition to this conceptual advantage, performance of a single agent neo-adjuvant trial offers the advantage of reduced risk for adverse events that could delay of surgery due to the fact that ChiLob7/4 will not combined with GMZ. Notably, while GMZ is SOC treatment for patients with non-resectable disease (the patient group selected for IAPACA-1), it is considered an IMP for patients with primary resectable PDA, therefore necessitating the dose finding of ChiLob7/4 in conjunction with GMZ in patients with non-resectable disease.
Notably, while adaptation of the trial protocol to the ‘IAPACA-NEO’ study design is completed, the IMPD is not expected to be completed until April 2019. In view of the fact that completion of the CD40 Ab trial in PDA within the requested 2-year extension period (Nov 2017-Oct 2019) could not be considered as realistic anymore, the Project Coordinator was informed in December 2019 by EU Scientific Officer Jan van the Loo that it had become impossible for him to obtain approval for this extension. In view of this, the strategy of the partners towards performing this trial was amended as follows:
• ChiLob7/4, once released will first be tested for safety in HNSCC patients (See Figure 5 below; trial Arm 2A).
• The safety data of this study will form the basis for presenting the IAPACA-NEO study to the ethics authorities
• Alternative funding for completion of this trial will be requested from CRUK.
WORK PACKAGE 3. COMBINATION OF ANTI-HPV RNA VACCINATION WITH ANTI-CD40 IS-AB IN PATIENTS WITH RECURRENT HPV+ HEAD AND NECK SQUAMOUS CELL CARCINOMA
Objectives
• To test the primary hypothesis that agonist anti-CD40 IS-Abs are capable of mobilizing and boosting systemic and specific, intra-tumour anti-cancer immunity, induced by antigen specific RNA vaccination against HPV.
• To gather early clinical efficacy data on the effects of RNA vaccination when applied as single treatment or in combination with anti-CD40 IS-Ab.
Progress for each task
Task 3.1: Generation clinical grade anti-CD40 IS-Ab ChiLob7/4 (Partner 10)
See Task 2.1 above.
Task 3.2: Preclinical studies with HPV16 mRNA vaccine and anti-CD40 (Partner 3, 10)
Pre-clinical studies in mouse models by BioNTech (partner 10) demonstrated striking immunogenicity and therapeutic anti-tumor potency of the HPV16 E6/E7 RNA vaccine. Subsequent experiments focused on demonstrating synergy between RNA vaccination and CD40 Ab, as previously shown for other vaccine entities. An important aspect of these studies was to determine the most optimal dosing regimen in this respect, and to further elucidate the mechanism of action of CD40 Ab in the context of this combination treatment. Contrary to expectations, the pre-clinical studies in tumor models (TC1, CT26, 4T1) revealed no added value of CD40 Ab stimulation (FGK, rat IgG2) on top of RNA-based vaccination for tumor relevant antigens, in spite of clear anti-tumor, T-cell dependent efficacy of each of the single agents. The lack for synergy was also shown in extensive vaccination experiments with measurement of the T-cell immune response as readout. The reason for a lack in added value of CD40-ligation is thought to be primarily due the potency of the liposomally complexed E6/E7 RNA vaccine alone. Notably, this liposomally-complexed RNA vaccine was developed by partner 10 after planning of the IACT clinical studies.
Importantly, either drug formulation showed significant single agent activity. Based on these findings, the plan to combine these drugs in the context of the clinical studies under Task 3.5 was abandoned, and an alternative plan for testing ChiLob7/4 as single agent in HNSCC was developed (see below).
Task 3.3 Preparation and QC of a clinical batch of HPV RNA vaccine (Partner 10)
BioNTechs’ RNALIP Immunotherapy Pipeline
BioNTech has developed a nanoparticulate formulation which allows for intravenous administration of the RNA vaccine (RNALIP) and for selective expression of the encoded antigens in antigen-presenting cells in lymphoid tissues, thereby inducing potent antigen- specific immune responses already at low doses (Kranz et al., Nature, 2016). BioNTech takes the position that each technology improvement and each novel therapeutic concept should be introduced into clinical first-in-human testing in a sequential manner in order to allow for incorporating safety-related findings in any study in subsequent trial designs, an approach which efficiently mitigates potential risks inherently associated with each novel component. This view is shared by the German competent authority, the Paul-Ehrlich-Institut (PEI). The RNA vaccine tested in the HARE-40 trial is related to the RNA vaccines tested in BioNTech’s Lipo-MERIT and (TNBC-)MERIT trials, and although the HARE-40 trial is not sponsored by BioNTech, sequential trial initiation was deemed reasonable to ensure patient safety and hence was discussed and agreed with the sponsor/chief investigator of the HARE-40 trial. Two issues encountered during trial initiation and patient treatment in BioNTech’s Lipo- MERIT trial had a direct impact on the design and treatment start in BioNTech’s (TNBC-)MERIT trial, and hence on the design, submission and initiation of the HARE-40 trial and are described in detail below.
Preparation of RNALIP: Manufacturing vs. Reconstitution
In line with the EMA Guideline CPMP/BWP/3088/99 (Committee for Proprietary Medicinal Products (CPMP) ‘Note for Guidance on the Quality, Preclinical and Clinical Aspects of Gene Transfer Medicinal Products’, 24 April 2001) the preparation of RNA with liposomes was classified as reconstitution by BioNTech. German local regulatory authorities, however, unexpectedly classified the final step of RNALIP preparation as a manufacturing step which requires a dedicated manufacturing license. In addition, the regulatory agency asked for de-complexation of the IMP preparation protocol.
In response, BioNTech generated a simplified, robust two-step RNA(LIP) preparation procedure which is also applied to the RNA vaccine used in the HARE-40 trial. Furthermore, BioNTech undertook all necessary steps to verify, audit and contract a pharmacy which holds a manufacturing license for Advanced Therapeutic Medicinal Products (ATMPs), the drug class RNALIP falls into. BioNTech and the identified pharmacy prepared all relevant documents to obtain the manufacturing authorization for RNA(LIP) from the local competent authority. The manufacturing license was finally obtained in December 2014, and the pharmacy now supplies clinical sites in Germany and in Belgium with ready-to-use RNALIP (Lipo-MERIT and (TNBC-)MERIT trials). Of interest, the final step of RNALIP preparation was later classified as reconstitution in France, Sweden, and in the UK. Obtaining the manufacturing license implied a delay in starting the melanoma trial, and thereby in the generation of clinical safety data on RNALIP which needed to inform the design of (TNBC-)MERIT and HARE-40 trials, including the definition of a safe starting dose, thereby initially delaying HARE-40 submission by 12 months.
Clinical Trial Application and Study Start
After having resolved the regulatory issues regarding RNALIP manufacturing described above, the application to the British competent authorities was planned for summer 2015. However, results obtained from the very first patient treated with RNALIP in May 2015 (within the Lipo-MERIT trial) caused a further delay in submitting the ‘HARE-40’ clinical trial application: The treatment of the first melanoma patient revealed a high potency of RNALIP already at the very low doses initially used. The first patient experienced symptoms typically associated with immune system activation within hours after vaccination. These effects were dose-dependent but already observed at the second-lowest dose level. Additional adverse events were observed after escalation to the next higher dose level. The observed symptoms were readily manageable by co-medication but led to a dose reduction for the remaining vaccination cycles in this patient at the discretion of the treating physician. After review of the first patient’s clinical data an independent data safety monitoring board (DSMB) approved the inclusion of two further patients and recommended continuation of the melanoma trial as planned. Nevertheless, these findings mandated an adaptation of the intended dose-escalation scheme in all trials testing RNALIP (Lipo-MERIT, (TNBC-)MERIT, and HARE-40). To this extent and in agreement with the PEI (the responsible competent authority for the Lipo-MERIT trial), it was decided to first wait for and to evaluate clinical safety data generated from the first three patients in the Lipo-MERIT trial prior to finalizing the dose-escalation scheme for all RNALIP trials, including HARE-40. The data from these first three patients were available by November 2015 and revealed significant inter-individual variations both in cytokine secretion as well as in the intensity of adverse events typically associated with immune activation. These findings are in line with inter-individual variation of TLR function due to genetic polymorphisms and an ageing-associated decrease of TLR-mediated inflammatory reactions, and suggest that it is unlikely to identify one pharmacodynamically active and safe dose which fits all patients.
To better address the observed inter-individual variability, the dose identification strategy for all RNALIP trials at that time, including HARE-40, encompassed a dose-escalation in every single individual patient by starting treatment of each individual with a very low dose and stepwise increasing the individual dose until either the dose level of the respective cohort is reached or until a lower individual dose is identified which cannot be further safely increased due to inflammatory reactions. The individual dose-escalation process foresees that after every drug exposure and prior to the next drug application a decision has to be made by the responsible investigator whether the dose-escalation in the individual patient will be continued as planned or, in case of significant inflammatory adverse events, treatment has to be modified according to a predefined algorithm.
The novel treatment scheme had to be worked out and implemented into the HARE-40 trial design prior to the initial submission of the clinical trial protocol to the British authorities, which was therefore delayed to March 2016. The MHRA and ethics committee supported the clinical trial application and approved testing of the RNA vaccine in the adjuvant and advanced disease settings. The authorities, however, at that time did not approve testing of the RNA vaccine in neo-adjuvant patients as clinical safety data with that particular RNALIP product had not yet been available. Instead, the authorities suggested implementing that arm of the study only later via submission of a substantial amendment. The authorities further asked to remove all reference to the intended testing of ChiLob7/4 antibody from the protocol until the product had been manufactured. A substantial amendment should then be submitted together with the IMPD and IB describing ChiLob7/4 product characteristics. Consequently, an amended protocol version, in which the neo-adjuvant study arm as well as all reference to ChiLob7/4 had been removed, had to be prepared and was submitted to the MHRA and local ethics committee in May 2016. Approval by the British authorities was obtained in summer 2016.
The RNALIP products for the HARE-40 trial are reconstituted from individual components, i.e. the RNA drug products, sodium chloride as diluent, and liposomes as excipient. The components were to be packaged into one ‘HARE-40 kit’ to be shipped to and stored until use at the hospital pharmacy at 2-8 °C. By the time of protocol approval, new stability data obtained for the RNA drug product batches manufactured for HARE-40 revealed that the HARE-40 kit storage temperature of 2-8 °C was insufficient to safely ensure drug supply over the whole treatment cycle of 78 days for a given HARE-40 patient. As a consequence, the individual components for RNALIP reconstitution were separated to be supplied in two different kits (i.e. one kit containing the RNA drug products, to be shipped and stored at <-15 °C, and another kit containing sodium chloride and liposomes, to be shipped and stored at 2-8 °C). As packaging of the individual components into kits is a manufacturing step to be performed in full compliance with GMP, realization of the design change required adaptations to various controlled documents which are subject to controlled review and release processes.
In addition, the new approach now using two kits instead of one required another substantial amendment, which required changes not only to the protocol but also to IMPD and IB. This amendment was submitted in December 2016 and was approved by the authorities in spring 2017 (EudraCT number 2014-002061-30). Right upon approval, the last preparatory steps required to enable patient treatment, i.e. qualifying the hospital pharmacy staff for RNALIP reconstitution, shipment of sufficient HARE-40 kits for the initial patients to be treated, and site initiation by sponsor were conducted and the first HARE-40 patient received the first vaccination with RNALIP in May 2017.
Task 3.4 RNA vaccination dose escalation study in recurrent HPV16+ HNSCC (Partners 3, 5 and 10)
Task 3.5 Safety/immunogenicity study combining HPV mRNA vaccination with anti-CD40 IS-Ab in HPV16+ HNSCC (Partners 3, 5 and 10)
Amendment of trial design
The design of this study as described in the original IACT Description of Work (DOW) was as follows:
• Task 3.4: RNA vaccination dose escalation study in recurrent HPV+ HNSCC (Study A)
• Task 3.5: Safety/immunogenicity study combining HPV RNA vaccination with anti-CD40 IS-Ab in HPV16+ HNSCC (Study B)
Task 3.4 and 3.5 have been combined within the HARE-40 trial; Study A is represented by part I (Arms 1A-C) and Study B by part II (Arm 2A) of the trial (see Figure 5 for amended design).
The design of these tasks has been amended in view of the following insights that emerged in the course of the IACT project:
1. The ongoing studies in melanoma with the new liposomal BioNTech RNA vaccine formulation (RNA(LIP)) demonstrated impressive potency of the vaccine format with induction of antigen-specific T-cell immunity and early evidence of clinical responses.
2. Unexpectedly, substantial inter-patient variability was observed with respect to adverse events following administration of vaccine using the new vaccine platform (see Task 3.3 above). While no dose-limiting toxicity was identified, the decision had to be taken to keep patients in the hospital overnight for close monitoring. This has required considerable extra logistical effort as patients now have to be accommodated overnight for each visit until they have reached the target vaccine dose. This also meant that we undertook the ‘first in man’ element of the study in a single center and only recruited one patient at one time. Having gained experience with the adverse event profile (grade 2 fever and rigors) we are confident that we can relax the restrictions on parallel recruitment after the review of the trial by the safety board. We will however continue intra-patient dose escalation and careful monitoring of each new patient.
3. Pre-clinical studies indicated that combination of this new liposomal vaccine formulation with agonist anti-CD40 IS-Abs did not result in synergistic effects, most likely due to the increased potency of this vaccine. Nevertheless, these studies showed that each of these drugs had significant single agent activity against HP16-positive tumors.
Based on these insights, the design of the HARE-40 trial was amended as follows:
• In order to evaluate immunogenicity and to assess immune toxicity in patients in whom immune function is not affected by advanced disease, Arm 1A recruits patients who are currently disease free, but at significant risk of recurrence.
• The SRC review following vaccination of 6 patients at dose level 3 will trigger the recruitment of patients to the Arm 1A expansion cohort (intra-patient dose escalation to dose level 5) and the opening of the dose finding element of Arm 1B (N=15). The two study arms will allow direct comparison of immunogenicity in disease-bearing and disease-free patients. Two additional UK centers are in setup to support timely recruitment of patients with advanced disease.
• Following completion of production of the antibody (Task 2.1) we are also now in a position to test the anti-CD40 IS-Ab ChiLob7/4 in the clinic in HNSCC. Due to the mode-of action of the anti-CD40 IS-Ab, involving activation/modulation of CD40-positive immune cells in the tumor microenvironment, patients in the adjuvant setting will not benefit from this treatment. Hence, we propose to evaluate the antibody in parallel to vaccination in patients with advanced disease. In Arm 2A we will recruit patients who have an HPV16 negative cancer (exclusion criterion for Arm 1B) but are otherwise eligible for the trial. This will allow us to bring forward the testing of the anti-CD40 IS-Ab to early 2018, the protocol amendment to allow this is in preparation to be submitted during November 2017. The MHRA supported our proposal to add the anti-CD40 Ab arm, pending evaluation of the final batch of antibody, which we now have (Task 2.1) when we sought scientific advice prior to initial approval of the study.
• In view of the observed potency of our RNA vaccine formulation in patients with melanoma (see point 1 above), we will also add an expansion cohort to Arm 1B, as well as a study arm in which the vaccine will be administered in the neo-adjuvant setting (Arm 1C), enabling detailed evaluation of the impact of the vaccine on (the T-cell response in) the tumor micro-environment.
Trials status December 2018
Arm 1A (adjuvant setting) cohort 1 was opened for enrolment in May 2017, involving intra-patient dose escalation up to 29 ug of vaccine. The first 6 patients were enrolled and completed the intra-patient dose escalation scheduled without severe IMP-related AEs. Upon positive safety review of the results of this cohort in January 2018, Arm 1A (adjuvant setting) cohort 2 and Arm 1B (metastatic disease) have been opened, each of which involve intra-patient dose escalation up to 72.8 ug (9 patients each). In Arm 1A, we have now recruited all 15 of the planned patients. For Arm 1B, we have now recruited and treated 3 (of 29) patients, with 4 further patients on the waiting list. Upon positive evaluation of the results of these cohorts, Arm 1C (neo-adjuvant; 15 patients) can be opened.
The interest for our trial is intense with multiple referrals coming from the UK and other EU countries. Nevertheless, rapid recruitment of patients has been hampered by the fact that the 2-year extension of the IACT project was not approved, and was eventually rejected. In particular, this halted the inclusion of additional clinical centers (Poole, Liverpool, London, Pamplona) in the trial.
Preparation for Arm 2A, involving a dose escalation study with anti-CD40 Ab ChiLob7/4 in patients with advanced disease, will proceed independently of Arms 1A-C. The time-limiting factor for this arm is completion of the IMPD for ChiLob7/4 by partner BioNTech, as referred to above. The plan is to submit the HARE-40 Arm 2A protocol as soon as possible to the MHRA.
WORK PACKAGE 4: DEVELOPMENT OF LEAD ANTI-OX40 AND ANTI-4-1BB IS-ABS FOR CLINICAL USE
Objectives
• Systematic selection of lead anti-human OX40 and human 4-1BB IS-Abs based on multiple in vitro and in vivo predictive assays
• Engineer humanised mAb against OX40 and 4-1BB for clinical use.
Progress for each task
Task 4.1 In vivo testing in receptor knock-in mice: T-cell costimulation and anti-tumor efficacy (Partner 3)
Anti-human 4-1BB Abs:
Twenty candidate anti-4-1BB Abs were subjected to selection by measuring their capacity to costimulate (CD8+) T-cell proliferation in human PBMC-cultures. This resulted in identification of clone SAP3-28 as lead candidate plus several backups. Notably, these costimulation assays make use of PBMCs that have been subjected to high-density culture for 2 days; see WP6, Task 6.1.
Since candidates have been selected for agonist function, a continuation towards the human-IgG2 format seems the most logical step forward (for important information on this Ab format, see WP7, task 7.2). Even though depletion of Treg with huIgG1-formatted Abs may be of potential interest, the expected depletion of CD8+ effector T-cells seems to be an undesirable aspect of this concept. This also fits best with the recently identified merits of 4-1BB Abs in stimulating ADCC activity in conjunction with tumor-targeting Abs (Levy et al. Blood. 2011 117(8):2423-32; J Clin Invest. 2014 124(6):2668).
In vivo analysis of anti-human 4-1BB Abs was impeded by the failure of human 4-1BB-ki mouse to express the receptor of interest. Consequently, further testing was focused on the CD4+ T-cell assays (Task 4.3) and NK assays (Task 4.4).
Anti-human OX40 Abs:
Twenty anti-OX40 Ab candidates were screened for their ability to costimulate human T-cell proliferation in human PBMC-cultures in vitro, as well as hOX40 KI OT-I cells in vivo in mice, as a measure of agonistic function. For OX40, several good candidates were identified, including clones SAP15-3, 25-29 and 29-50. The most promising candidate with respect to costimulation of effector T-cell responses in human OX40 transgenic mice were found to be clones 25-29 and 29-50, with clone 29-50 consistently showing somewhat greater activity than 25-29.
Task 4.2 OX40: CD4+ T-regulatory cell suppression (Partner 6)
Optimized assays have been developed by partner 6 (Milan) to measure the capacity of anti-OX40 Abs to prevent the conversion of non-suppressive CD4+ T cells into Tr1 suppressor cells. CD4+CD25- T-responder cells were purified from total PBMC using MACS MicroBead Technology. As T-responder cells do not express OX40 constitutively, we induced its up-regulation stimulating T-responder lymphocytes with TNFα and IL-2 for 18 hours. The resulting OX40-expressing responder T cells were cultured for seven days in the presence of dexamethasone and vitamin-D3 to induce their conversion into suppressive Tr1-IL10 secreting cells. Compared to control cells, under dexamethasone and vitamin-D3 pressure, T-responder cells secret higher amount of IL-10. The addition of the anti-OX40 Abs, especially clones SAP15-3 and SAP25-29 (provided by partner 3), reduced IL-10 secretion by T-responder cells. Moreover SAP25-29 increases the production of IFNγ, suggesting a more inflammatory/non-suppressive phenotype of T-responder cells despite the presence of dexamethasone and vitamin-D3.
More extensive testing of the selected, anti-human OX40 Abs (SAP 15-3; 25-29; 29-50), as well as anti-human 4-1BB Ab 3-28 Abs was performed in the second project period. Two different assays were used to test the capacity of the selected Abs to suppress respectively the development of Tr1 cells (under the influence of Dex and VitD3) from fresh PBMC-derived CD4+ T-cells or the IL-10-secretion by already differentiated Tr1 cells. CD3 and OX40/4-1BB Abs were added as plate-coated Abs in these assays. In both assays, the fraction of IL-10-secreting CD4+ T-cells was assessed by FACS/ICS. The most significant outcomes were as follows:
• Most prominent differences were observed in Tr1 differentiation assays that started with either total or naïve CD4+ T-cells
• Most consistent inhibition of IL-10 secretion and IRF1 expression were observed with SAP29-50 (OX40) and SAP3-28 (4-1BB). Overall, impact of SAP25-29 was weaker, while SAP15-3 was clearly the weakest OX40 Ab
• SAP29-50 (OX40) also proved to be the most potent Ab in costimulating proliferation of CD4+ T-cells under regular (in absence of DEX or VitD3) conditions
• Results from other assays are less convincing with respect to differentiating between the Abs tested, which was mainly due to limited difference between positive and negative control (so window for accurate measurements too small).
In view of the above, OX40 Ab SAP29-50 appears to be most potent and will be considered as the lead candidate. Furthermore, these data support the biological activity of the (only) available humanized anti-4-1BB candidate SAP3-28.
Based on the data set shown, the Tr1 differentiation assay with IL-10 and IRF1 as readout is the test of choice for evaluation the dual-specific Abs generated by BioInvent (see WP8 below).
Task 4.3 4-1BB: CD4+ T-cell response suppression (Partner 6)
Partner 6 also tested anti-human 4-1BB Ab SAP3-28 for its biological activity in the in vitro assays described under Task 4.2. This confirmed that this antibody is capable to suppress Tr1 development.
Task 4.4 4-1BB: stimulation NK activity. (Partner 4)
These experiments focused on the recent finding that agonist anti-4-1BB Abs can potentiate the anti-tumor efficacy of tumor-targeting Abs that mediate their anti-tumor effect (at least in part) through ADCC (Levy et al. Blood. 2011 117(8):2423-32; J Clin Invest. 2014 124(6):2668). The EGFR-specific Ab Cetuximab and EGFR+ HT29 tumor cells were used a model for the initial experiments. In order to evaluate ability of the anti-human 4-1BB IS-Ab to stimulate NK cell activity in vitro, freshly isolated human peripheral blood NK (CD56+ CD3-) cells were incubated with Cetuximab and HT29 tumor cells. NK cells significantly upregulated 4-1BB, becoming receptive to 4-1BB costimulation. Several assays were performed to measure the stimulation of human NK cell-mediated ADCC towards HT29 cells in conjunction with Cetuximab. Of these, the in vivo assay focusing on the percentage of viable HT29 cells recovered upon peritoneal lavages and liver metastasis shows a prominent impact of the reference anti-4-1BB Ab (6B4). A pilot experiment with anti-4-1BB Ab SAP3-28 was ongoing.
An alternative in vitro assay tested the impact of 4-1BB Abs on CD25 and BCL-XL expression in CD8+ T-cells pre-estimulated with anti-CD3 (detected in flow cytometry). Reference 4-1BB Ab 6B4 results in a prominent induction of CD25 in approx. 50% of CD8+ T-cells, along with a shift towards expression of higher levels of BCL-XL. In addition, enhanced proliferation was detected by means of V450 fluorochrome dilution, albeit less reproducibly. Notably, the Southampton reference Ab SAP3-28 was not yet tested in these assays!
Task 4.5 Humanisation and testing of selected anti-OX40 and anti-4-1BB mAb (Partner 3)
Humanization of the selected Abs was successfully completed for all lead Abs. However, of the anti-OX40 Abs humanized, only Ab 25-29 retained full affinity for huOX40. Humanization of Ab 29-50 only succeeded in one of the humanized frameworks that were applied, and the resulting humanized Ab had a clearly diminished affinity as compared to the original muIgG1 Ab. The same applied to the humanization of Ab 3-28.
Changes in Work Plan due to strategic considerations
On June 1 2015, BioInvent (Partner 8) announced that it entered in a research collaboration with the University of Southampton (partner 3) concerning the development of T-regulatory cell checkpoint inhibitors. This collaboration involved licensing by CRUK/CRT to BioInvent of the lead anti-human OX40 and anti-human 4-1BB Abs that are subject of IACT work package 4. The major objective of this collaboration, the accelerated clinical development of this Abs, is highly complementary to the IACT work plan. Partners 8 and 3, as well as CRUK/CRT, have declared to the IACT coordinator (Offringa, partner 1) that this is a non-exclusive licensing agreement that does not affect the workings of the IACT project The Heidelberg Technology Transfer office has investigated this in more detail, and concluded that this is correct and, therefore, that no amendment of the Consortium Agreement needs to be made in this respect.
In the course of 2017, further work on Tasks 4.3 – 4.5 was suspended. The rationale for this decision was that the aforementioned collaboration between University of Southampton (partner 3) and BioInvent (partner 8) had resulted in the identification of fully human, phage-derived Abs with properties superior to the IACT lead Abs. In view of the competitiveness of the field and the need to focus on ‘best in class’ strategies (see above), it was decided to stop all pre-clinical development work on the original set of OX40 and 4-1BB IACT candidate Abs. Irrespective of this, these IACT Abs will still be used as a reference for other project goals.
While this means that the IACT goal for developing the pre-defined anti-OX40 and anti-4-1BB lead Abs towards clinical testing (Deliverable D4.5) will not be fulfilled, the IACT consortium did create a setting that facilitated the isolation of the new, superior Abs. Notably, these new Abs fall outside of the scope of the IACT project and IACT consortium agreement, but instead have been generated and tested in the context of the aforementioned collaborative agreement between University of Southampton and BioInvent.
WORK PACKAGE 5: BIOMARKER ANALYSIS ACCORDING TO STANDARD OPERATING PROCEDURES (SOPS)
Objectives
• Develop validated assays for effective measurement of drug pharmacokinetics (PK) and drug/treatment-related pharmacodynamic (PD)-biomarkers.
• Define and harmonized these assays by means of SOPs, that will be distributed among IACT partner
Progress for each task
Task 5.1 Pharmacokinetics (PK) assays (Partners 3, 7, 8, 9)
A validated sandwich ELISA, as developed by partner 3, is available for anti-CD40 Ab ChiLob (Chowdhury et al. J Immunol Methods 2010 363(1):1) and has already been applied successfully in the completed phase I trial (Johnson et al. 2015 Clin Cancer Res 21(6):1321). SOPs are available for sharing with IACT partners.
Task 5.2 Anti-drug antibody (ADA) response assays (Partners 3, 7, 8, 9)
A validated ELISA, as developed by partner 3, is available that uses anti-CD40 Ab ChiLob as capture antigen (Chowdhury et al. J Immunol Methods 2010 363(1):1-8) and has already been applied successfully in the completed phase I trial (Johnson et al. 2015 Clin Cancer Res 21(6):1321). SOPs are available for sharing with IACT partners.
Task 5.3 Serum cytokine/chemokine analyses (Partners: 1, 3, 4, 10)
SOPs for multiplex analysis of serum samples have been established by partner 3 in the context of the completed single agent trial (see WP2 Task 2.2 for more info). The completed analyses have revealed a sharp, >10-fold increase in MIP1b within the first day after dosing, most likely to reflect an infusion reaction rather than a result of engagement of the CD40 receptor. In addition, a minor (2-fold) increase in serum IL-12 could be observed in the course of the 1st week after dosing. This biomarker is more likely to be related to the biological activity of agonist anti-CD40 Ab. A limitation in detecting serum levels of IL-12 may be the limited half-life of this cytokine in the circulation. These results correspond to those obtained with another, very similar anti-CD40 Ab (SGN40; Moskowitz et al. 2013 Leuk Lymph 54: 277 and references therein).
In view of the limited usefulness of serum cytokines as mechanism of action-related biomarker, efforts to identify alternative biomarkers are being pursued (see Task 5.7).
Task 5.4 Analysis of blood cell subsets and activation markers (Partners: 1, 3, 4, 10)
Flow cytometric analysis of samples from the completed single agent study revealed a rapid (within 1 day of dosing) loss of peripheral B-cell counts, and a slower loss of NK cells. While the change in the number of CD40+ B-cells are clearly relevant to the impact of the anti-CD40 Ab, the mechanism by which the CD40-negative NK cells disappear is unclear. Measurements of dendritic cells (very low numbers) and T-cells (no significant effects) did not point at additional biomarkers that would be useful. These results again correspond to those obtained with anti-CD40 Ab SGN40 (Moskowitz et al. 2013 Leuk Lymph 54: 277 and references therein).
The disappearance of B-cells from the circulation can be used as a biomarker for systemic action of agonist CD40 Ab, but since this is known to happen in any human subject to which such Abs are administered, this biomarker will not be useful as a surrogate endpoint for clinical activity of anti-CD40 Ab.
Task 5.5 Analysis of HPV16 E6/E7-specific T-cell immunity (Partners 1, 3, 4, 10)
Validation of the following biomarker assays was completed by partner 3 (Ottensmeier):
• Serum ELISA for Ab titers
• ELISPOT-analysis using overlapping pools of synthetic peptides: 15-mers with an overlap of 11 residues, in total 59 peptide divided in 16 pools of 3-8 peptides
• IFNγ intracellular staining using the same peptide pools
• Staining with HLA-A2 tetramers for the following known epitopes. E6: 18-26 KLPQLCTEL, 29-38 TIHDIILECV, 52-60 FAFRDLCIV; E7: 11-19 YMLDLQPET(T), 82-90 LLMGTLGIV, 86-93 TLGIVCPI
In addition, impact of immunotherapeutic treatment on tumor microenvironment will focus on RNA-seq analysis of tumor samples. Baseline data for 40 HPV16+ HNSCC tumors have been compiled, as published in: Gene expression analysis of TIL rich HPV-driven head and neck tumors reveals a distinct B-cell signature when compared to HPV independent tumors. Wood O, Woo J, Seumois G, Savelyeva N, McCann KJ, Singh D, Jones T, Peel L, Breen MS, Ward M, Garrido Martin E, Sanchez-Elsner T, Thomas G, Vijayanand P, Woelk CH, King E, Ottensmeier C. Oncotarget. 2016 Jul 22. doi: 10.18632/oncotarget.10788.
Task 5.6 Evaluation of intra-tumour immunity by means of whole slide ‘Immunomap’ imaging (Partner 1)
Validated SOPs for the most important stainings (T, B-cell, DC, NK, macrophages) are available. Baseline data for T-cells and macrophages is currently available for 180 pathology-verified tumors (primary resectable pancreatic ductal adenocarcinoma; PDA). For an additional 40 tumors, data is available on a marker set that defines tertiary lymphoid structures (DC-markers, B-cells, T-cells high endothelial venules, ki67); this panel will be expanded to at least 80 tissue samples. Stainings of pancreatic tumors (>40) for the remaining markers (HLA-II, INOS, Arginase, CD86, CD40) are now available. Especially striking are the vast numbers of CD40-positive cells in the tumors. This clearly shows that this tumor is a splendid target for treatment with agonist anti-CD40 Ab. The detection of high numbers of CD40-positive cells in tumors was confirmed by stainings with two additional anti-human CD40 Abs, including our clinical Ab ChiLob7/4 (!), supporting the expectation that treatment of patients with agonist anti-CD40 Abs, when administered at dose levels resulting in systemic exposure, is likely to impact on the immune cell infiltrate in these tumors.
In order to enable the detection of cell-bound anti-CD40 Ab ChiLob7/4, partner 3 generated several anti-idiotype Abs. While these Abs detected cell-bound ChiLob7/4 under optimal in vitro conditions, the sensitivity of detection was unfortunately too low to allow detection of cell/tissue-bound ChiLob7/4 under physiological conditions, e.g. in tumor tissue of patients treated by ChiLob7/4. In view of this, this aim (as mentioned in the report on the 2nd project period) was deprioritized.
Task 5.7 Unbiased PD biomarker discovery by means of mRNA gene expression profiling of PBMC and tumour biopsies (partners 1, 3 and 10)
Identification of PD biomarkers on basis of the transcriptome of in vitro stimulated human PBMC rendered interesting findings. However, there is major variability in the signature obtained between different PBMC batches. Furthermore, it is evident that in vitro culture alone already triggers a time-dependent transcriptional profile. While this does not exclude meaningful findings using PBMC from ChiLob-treatment patients, it is unlikely that the in vitro experiments will provide significant guidance in this respect. Based on these findings, further experiments focused at the tumor transcriptome of mice treated with agonist CD40 Ab (see WP6/Offringa below). These latter experiments showed that transcriptome analysis of tumors from mice treated with anti-CD40 Abs (Tasks 6.5 and 6.6) proved to be highly informative in this respect.
While these experiments confirmed the therapeutic synergy between anti-CD40 Abs and cytostatic drugs, they also revealed that the immune gene expression signature in the tumor was much more pronounced in mice treated with anti-CD40 Ab as single agent as compared to mice receiving both CD40 Ab and cytostatic drug (for details, see under task 7.2 as well as report on Deliverable D5.4 as well as Annex 5 of project extension request). Even though the paradoxical aspects of these findings still need to be elucidated, it is evident that – at least based on our state of the art pre-clinical data – the likelihood of detecting a PD biomarker signature in the neo-adjuvant trial setting is much higher if the patients are dosed with anti-CD40 Ab ChiLob7/4 only (see WP2 for implementation). Importantly, the detection of this PD biomarker signature is, and has always been, the primary aim of this trial, in that this would embody the first direct evidence that administration of anti-CD40 Ab does not merely result in systemic exposure, but also in intra-tumoral drug exposure and delivery of its agonist signal to CD40 positive immune cells.
WORK PACKAGE 6. VALIDATED AND HARMONIZED PRE-CLINICAL IN VITRO AND IN VIVO MODELS (SOPS)
Objectives
Development, validation and harmonization of:
• in vitro models for assessing the immunostimulatory capacity as well as potential super-agonist activity of agonist IS-Abs
• in vivo immunization models for the evaluation of agonist IS Abs potency
• transplantable tumour models for initial evaluation of anti-tumour efficacy of agonist IS-Abs
• genetically engineered tumour models for in-depth evaluation of therapeutic index of selected therapeutic regimens involving agonist IS-Abs
Progress for each task
Task 6.1 In vitro assays for analysis of stimulatory activity agonist IS-Abs (Partners 3, 4, 6)
Experiments by partner 3 have shown that high density pre-culture, as initially determined in the context of assays with super-agonist anti-CD28 Ab TGN1412 (Kalinke et al. J Immunol 2014; 192:2091-2098), is an essential step in generating PBMC for use in in vitro assays to determine costimulatory activity agonist Abs in presence of low concentrations of anti-CD3 Ab OKT3. Furthermore, the optimal concentration of plate-bound OKT3 differs between PBMC donors. Consequently, a range of concentrations needs to be tested (0.05 – 1 ug/ml), at least initially (if multiple aliquots from same donor are available). Finally, responses by CD4+ and CD8+ T-cell subset differ between PBMC donors and Abs. Therefore, proliferation, as measured by CFSE-dilution, is determined by FACS for both T-cell subsets. SOPs for these assays have been made available to the IACT partners
Task 6.2 In vitro assays for detecting super-agonist activity IS-Abs(Partner 3)
High density pre-incubation as mentioned above similarly enables testing for superagonist activity (see above; SOPs available). In line with published data (Kalinke et al. J Immunol 2014; 192:2091-2098), induction of T-cell proliferation by anti-CD28 Ab TGN1412 in the absence of CD3 Abs (so super-agonist activity) was only seen for high density pre-cultured PBMCs. Notably, this super-agonist activity was found to be dependent on Fc-FcR-interaction, as F(ab)2 fragments lacked activity. Moreover, Ab-mediated blockade of FcRIIB, but not of FcRI or FcRIII, abolished super-agonist activity of TGN1412, demonstrating that the FcRIIB-dependency is the same as found for the agonist Abs targeting CD40, OX40, 4-1BB and CD27.
Both monocytes and B-cells were shown to provide this FcRIIB-dependent crosslinking capacity in in vitro cultures; cell titration experiments showed that approx. 10-fold less monocytes than B-cells were required for this accessory function. Furthermore, high-density pre-culture was found to strongly upregulate FcRIIB on monocytes, while not affecting levels on B-cells. Therefore, the increased responsiveness of high-density pre-cultured PBMC to agonist Abs is primarily due to upregulation of FcRIIB on the monocyte subset, rather than T-cell-intrinsic features as proposed in the aforementioned publication in J Immunol.
Task 6.3 In vitro assays for detection of CD4+ T-cell suppression (Partners 4, 6)
See description under Task 4.2 and 4.3.
Task 6.4 In vivo immunization models (Partners 1, 3, 4, 6)
In vivo assays for testing anti-mouse agonist Abs targeting CD40, OX40 and 4-1BB have been recently published by partner 3 (White & Glennie et al. J Immunol. 2011 187(4):1754-63; Cancer Immunol Immunother. 2013 62(5):941-8; J Immunol. 2014 193(4):1828-35; Cancer Cell. 2015 Jan 27(1):138-48) and protocols have been made available to the IACT partners. More recently, an in vivo assay for testing agonist anti-human OX40 Abs has been developed by making use of huOX40-Tg OT-1 T-cells. Unfortunately, the hu-4-1BB-Tg mice were found to lack functional expression of the receptor of interest, so an in vivo assay for testing anti-human-4-1BB Abs is not yet available and may not be achievable. In vivo immunization models using synthetic OVA peptides have been set up by partner 1 on basis of prior experience (Bijker, Offringa et al. Eur J Immunol. 2008 38(4):1033-42) and protocols are available to the IACT partners.
Task 6.5 Transplantable tumor models (Partners 1, 4, 6)
Standardized tumor models based on MC38-OVA, B16-OVA as well as PANC-02 have been set up by partner 1 and SOPs are available to other IACT partners. The PANC-02 model will be used to experiment with the relative timing of gemcitabine chemotherapy and anti-CD40 Ab dosing as discussed in WP2 Task 2.1.
Task 6.6 Genetically engineered mouse models (GEMMs) (Partner 1)
A melanoma GEMM is up and running in Heidelberg (partner 1). This model features conditional, melanocyte-specific expression of BrafV600E, overexpression of BCat and inactivation of PTEN. SMi targeting BRaf and MEK have profound anti-tumor activity, resulting in resp. tumor stasis and tumor regression. The anti-tumor impact of Temozolide is modest. Therapeutic impact of combinations of cytotoxic oncology drugs with agonist CD40 and OX40 Abs is currently under investigation (close collaboration between partners 1 and 3).
In order to improve the availability of chimeric anti-mouse Abs for these in vivo experiments, partner 10 has been involved in optimizing the production of these Abs (partner 10 has also enabled the production of clinical-grade anti-human CD40 Ab by generating new producer lines; see Task 2.1). In house production of surrogate agonist Ab obtained from Southampton (partner 3) (CD40, OX40, 4-1BB) has been set up in Heidelberg with help of Andreas Fregin and Oezlem Tureci of BioNTech (partner 10) sister-company Ganymed. Using chimeric anti-mouse CD40 3/23-muIgG1, multiple oncology SMi have been tested in in vivo assays for potential negative impact on DC/T-cell function (OVA immunization assays in OT-1 system), as well as in tumor models for potential synergy in anti-tumor impact with CD40 Ab. In particular, a MEK-inhibitor was shown to have no negative impact on DC/T-cell function while synergizing with CD40 Ab in anti-tumor efficacy (Offringa et al. manuscript in preparation). While these experiments have been performed in transplantable tumor models, genetically engineered models for skin melanoma and PDA are now available in the lab that feature traceable neo-antigen for follow up studies in autochthonous tumor systems.
WORK PACKAGE 7. DEVELOPMENT OF OPTIMIZED 2ND-GENERATION AGONIST IS-ABS THROUGH MODIFICATIONS IN FC-DOMAIN
Objectives
• Generation optimized IS-Abs through modification of the primary sequence of the Fc-domain, or through introduction of defined changes in the glycosylation of the Fc-domain.
• Selection of potential clinical candidates by means of functional assays that examine mAb function with respect to immune stimulation, induction of tumour-regression, as well as the absence of super-agonist activity
Progress for each task
Task 7.1 Introduction of mutations into the human IgG1 (h1) Fc-region to increase binding to human FcγRIIB (Partners 3, 4, 6)
In vivo experiments by partner 3 in mice have demonstrated that the mIgG1 is most potent Ab format for agonist Abs targeting CD40 and several T-cell costimulatory receptors (OX40, 4-1BB, CD27), and that the mechanism of action involves FcRIIb-dependent clustering of Abs and resulting immune-receptor multimerisation (White & Glennie et al. J Immunol. 2011 187(4):1754-63; Cancer Immunol Immunother. 2013 62(5):941-8; J Immunol. 2014 193(4):1828-35). Since all FcRs can in principle mediate this multimerisation-dependent agonist activity in vitro, bio-availability of FcR-mediated Ab-cross-linking seems to be a key feature in the in vivo setting.
A major hurdle in translating these findings towards the human system is that there are significant differences between the mouse and human Fc/FcR system, including the fact that in the human system there is no true counterpart of mIgG1. Nevertheless, it is known that certain mutations in the huIgG framework can increase the affinity to human FcgIIB, in particular the S267E and L328F mutations. Data from partner 3, as reported previously (see also report on Deliverable 7.3) demonstrated that the so-called huIgG SELF-variants of anti-mouse and anti-human CD40 IS-Ab display display greatly enhanced, FcRIIB-dependent agonist activity in in vitro and in vivo experiments. However, the in vivo tests also showed significant toxicity in mice, as reflected by weight loss, liver toxicity (increase in systemic ATS/ALT levels), ruffled skin and hunched posture. Similar studies by Ravetch and co-workers confirmed that the SELF huIgG framework enhanced both the agonist activity and in vivo toxicity of anti-CD40 IS-Abs, although in this case the toxicity was reported to primarily caused by thrombocytopenia (Dahan et al. 2016 Cancer Cell 26, 820). Based on these findings, and the poor predictive value of mouse models for the outcome of Fc/FcR-interactions, it was decided to de-prioritize the further development of IS-Abs with a huIgG SELF framework. This also applied to the targets 4-1BB and OX40 for the following 2 major reasons:
• The overall de-prioritization of work on 4-1BB and OX40-targeting IS-Ab in the IACT program (see WP4)
• Reports on early clinical trials with agonist anti-4-1BB Abs indicating that liver toxicity is a major hurdle with respect to the clinical application and development of anti-4-1BB Abs (Bartkowiak 2018, Clin Cancer Res 24 (4) and references therein)
Task 7.2 Investigation of human IgG2 (h2) as a potential format for generating potent and safe agonist IS-Ab (Partners 3, 4, 6)
Greatly enhanced, FcR-independent agonist activity can be achieved by converting Abs into the huIgG2-format, in particular the IgG2B-isoform, as described in the previous progress report. In view of these findings, the huIgG2-format represents a very interesting new design for agonist immunostimulatory Abs. The key question is whether the enhanced agonist potency of huIgG2-type stimulatory Abs results in an improved therapeutic index, or whether enhanced potency would simply be mirrored by increased toxicity. This has being investigated by partner 3 in a systematic fashion by means of a number of well-defined syngeneic tumor models in human CD40-transgenic mice. These experiments focused on three distinct anti-human CD40 Abs in huIgG2 format:
• Ab ChiLob7/4, originally formatted as huIgG1
• The super-agonist Ab CP-870,893, originally formatted as huIgG2
• Ab 341G2, an antagonist Ab in huIgG1 format that converts into a highly potent agonist in huIgG2 format (see below)
In essence, these experiments demonstrated that the latter two antibodies, although displaying highly potent immune stimulatory activity in in vitro and in vivo assays, were poorly tolerated in that they induced life threatening toxicity in the mice. The latter is in line with the clinical data for Ab CP-870,893, showing that the maximum tolerated dose of this Ab is 10-20-fold lower than that of ChiLob7/4 and SGN40.
In view of the above, the level of agonist activity displayed by huIgG2 ChiLob7/4 is still of great interest, in that it constitutes an intermediate between the ‘conventional’ huIgG1 agonist Ab (ChiLob7/4 & SGN40) and the huIgG2 super-agonist Ab CP-870,873. At patent application on this concept has been filed on Oct 2015, which is currently advanced towards the national phase: Modified antibodies containing modified IgG2 domains which elicit agonist or antagonistic properties and use thereof; WO 2015145360 A1; M. Glennie & A. White.
The limitation of these experiments referred to above is that clear-cut anti-tumor efficacy could only be observed when the agonist Abs were combined with tumor antigen-specific immunization (synthetic peptide vaccines). The conceptual limitation of such experiments is that in this setting one mainly looks at the capacity of the CD40 Abs to act as vaccine adjuvant, while the question of primary interest is whether these Abs can trigger anti-tumor immunity by inducing pro-immunogenic changes in the tumor micro-environment, in particular by activating dendritic cell and modulating myeloid cell function. Moreover, our experiments in Task 3.2 have shown that our mRNA-based vaccines are already very potent ‘as is’ and not significantly enhanced by co-administration of agonist anti-CD40 Abs.
Notably, based on technologies set up in the context of Tasks 5.7 6.5 and 6.6 partner 1 (Heidelberg) has tested chimeric anti-mouse CD40 3/23-muIgG1 in combination with multiple oncology small molecule inhibitors (SMi) in in vivo assays for potential negative impact on DC/T-cell function (OVA immunization assays in OT-1 system), as well as in tumor models for potential synergy in anti-tumor impact with CD40 Ab. In particular, MEK-inhibitors were shown to have no significant negative impact on DC/T-cell function while strongly synergizing with CD40 Ab in anti-tumor efficacy in three different tumor models. Detailed PD biomarker analysis by means of immunohistology, flow cytometry and whole transcriptome analysis showed the following:
• Successfully treated tumors with the combination of MEKi and CD40 Ab showed a shift from M2 macrophages to M1 macrophages, as well as a change in the CD8/Treg T-cell ratio in favor of CD8+ T-cells.
• This impact on the immune phenotype was a result of BOTH drugs together, and could not be achieved to similar extent by either single drug.
• The gene signature of treated tumors showed a strong suppression of cell proliferation/DNA synthesis markers, reflecting the anti-tumor impact, as well as a profound synergy by the drugs, in that neither alone rendered this effect
• Furthermore, a clear pro-immune gene signature was associated with the administration of the CD40 Ab. Interestingly, this effect was strongest in the mice treated with CD40 Ab only, and partly attenuated in mice that also received MEKi. Overall, this shows that the synergy between MEKi and CD40 Ab must be most at the level to tumor suppression, while MEKi partly counteracted the pro-immune impact of CD40 Ab in the tumor models, in spite of having little negative impact in the in vivo OT-1 system.
Further in-depth analysis of drug action is ongoing to gain further insights, with the aim of arriving at the most optimal treatment regimen.
Especially the transcriptome analysis strategy has resulted in very clear-cut gene expression profiles that can form the basis of validated, drug-specific PD biomarker profiles to be applied in clinical trials (see report on Deliverable 5.4).
In view of the above, partners 1 and 3 are proceeding by testing the anti-tumor impact of ChiLob7/4-huIgG2, as well as surrogate anti-mouse CD40 Ab 3/23-huIgG2, in combination with MEK-inhibitors.
In parallel to the development of anti-human CD40 Ab with an optimized IgG framework, partner 3 has deeply invested in obtaining more insight in the mechanism of action of these Abs by zooming in on the fine specificity of the Fab-domain of the Ab. The Southampton-group demonstrated that anti-human CD40 Abs can be divided into four different classes based on their behaviour in functional assays. For reasons of confidentiality related to the protection of intellectual property, the full details of these findings cannot be disclosed in this public report. However, most information can be found in the following publication: Yu, Glennie, White et al. Cancer Cell. 2018 Apr 9;33(4):664-675).
Exploration of the impact of the huIgG2 format on agonist activity of anti-4-1BB and anti-OX40 Abs showed a major increase for 4-1BB only. The reason for the minor increase for OX40 is not known at this time and is somewhat surprising since, also for the target CD28 the huIgG2 format resulted in greatly increased agonist activity. At this time, huIgG2 formatted anti-4-1BB Abs are, however, not considered for further development in view of the toxicity issues already reported concerning anti-4-1BB Abs (see under Task 7.1).
Task 7.3 Generation of glyco-optimized variants of anti-CD40 mAb ChiLob7/4 (Partner 7)
Glyco-optimized variants of ChiLob IgG1 and IgG2 (chimeric and humanized) generated through stable expression in a panel of Glycotope’s proprietary GlycoExpress cell lines display different glycosylation structures, i.e. slightly increased sialylation, levels of N-glycans bearing terminal mannose or hybrid type N-glycans as compared to Abs produced by standard CHO producer cells, but mainly low and high levels of fucosylation. Humanization and glyco-optimization were performed in parallel to differentiate between functional effects due to humanization and glyco-optimization. Glyco-optimized IgG1 and IgG2 variants (chimeric and humanized) of ChiLob7/4 were successfully produced for pre-clinical testing. For reasons of confidentiality related to the protection of intellectual property, the full details of these findings cannot be disclosed in this public report.
Task 7.4 Pre-clinical testing of glyco-optimized mAb in vitro (Partners 3, 4, 6, 7)
The aforementioned variants displayed highly comparable huCD40 binding (Biacore; flow cytometry) and passed QC with respect to purity, low content of aggregates (< 0.4%) and endotoxin.
A main goal of this project was to test whether glyco-optimization could result in increased FcyRIIB-dependent agonistic functions. However, no difference between the glyco-optimized variants was seen in the B-cell assay: all IgG1 Abs (chimeric and humanized) similarly stimulated B-cell proliferation in an X-linking dependent (and dose-dependent) manner, while all IgG2 Abs did so independent of secondary X-linking (X-linking was enabled using an X-linking antibody). However, in line with expectations, the low-fucosylated IgG1 and IgG2 Abs (chimeric and humanized) bound with higher affinity to activatory FcγRIIIA, with huIgG1 Abs having a much higher affinity than huIgG2. The in vitro ADCC assays using CD40 positive cancer cells corroborate these findings. For reasons of confidentiality related to the protection of intellectual property, some details of these findings were omitted from this public report.
In view of the above mentioned data, Glycotope GmbH proposes to focus on further functional assays in the context of different immune cells, instead of producing a larger batch of the Abs concerned.
The aim of these studies is the development of in vitro assays with human immune cells that reflect more closely the relevant tumor microenvironment and a wider repertoire of FcγRs than FcγRIIB. Most of the IACT research/optimization regarding the agonistic activity of ChiLob7/4, is based on the activation of FcγRIIB positive B cells. However, after the IACT project has been started, conflicting results have been published whether the agonistic activity of the human anti-CD40 antibodies in human patients solely depends on FcγRIIB receptors. Additionally, other immune cells, such as monocytes, macrophages and dendritic cells express the CD40 receptor and are known to play important roles in the tumor microenvironment. Therefore, we propose to develop assays with human monocytes, macrophages, NK cells, T cells and dendritic cells to further test the agonistic activity of the humanized and glyco-optimized ChiLob7/4 variants. These cell types express different FcγR types including FcγRIIIA, FcγRIIA and FcγRIIB. Especially activation and expansion of dendritic cells and reprogramming of tumor-associated macrophages are reported to inhibit cancer progression and metastasis. As DCs and macrophages express CD40 in addition to the FcγR, studies analyzing the agonist potential of the different anti-CD40 antibody glyco-variants in more complex assays, mimicking the tumor microenvironment, might reveal functional differences of the glyco-variants that cannot be detected in the classical B-cell assays. Furthermore, potential side effects of an ADCC enhanced huIgG1 should be analyzed. To this end, ADCC assays with primary CD40 positive macrophages and DC will be performed.
While this initial objective of improved FcgRIIB mediated effector cell activation by glyco-optimization was not achieved as described above, the low-fucosylated and humanized version of ChiLob (huIgG1 ChiLob7/4-FF) did show significantly enhanced agonist function as compared to reference ChiLob7/4 Ab, both in the B-cell activation and NF-κB signalling assay, when provided in the context of FcγRIIIA-expressing cells, such as NK-cells and DCs, in line with the propensity of this low-fucose Ab to bind more strongly to FcγRIIIA. In the functional assays concerned, including an allo-MLR T-cell activation assay, huIgG1 formatted ChiLob7/4-FF displays agonist activity that is comparable or even higher than that of huIgG2 formatted ChiLob7/4 while it is almost absent with the huIgG1 ChiLob7/4-SM (high-fucosylated and humanized version).
This finding is in line with data published by Ravetch et al. (Dahan et al. 2016 Cancer Cell 26, 820) and suggests that in the human system secondary cross-linking of agonist CD40 Ab by activatory FcγRIIIA may be as important as that mediated by inhibitory FcγRIIB in potentiating agonist function. Despite the enhanced agonistic activity, it is to mention that glyco-optimization of ChiLob7/4 IgG1 showed increased ADCC activity against CD40 positive tumor cells, but surprisingly not against CD40 positive immune cells (B cells, monocytes, DCs). The later may explain why FcγRIIIa mediated ADCC do not negatively interfere with agonist function. Therefore, the newly developed huIgG1 ChiLob7/4-FF (which has also been fully humanized by Glycotope GmbH) constitutes a very interesting new variant, in that it exhibits increased agonist activity at a level similar to that of huIgG2 ChiLob7/4, albeit in a fashion that – unlike for the huIgG2 version – does depend on secondary cross-linking by FcγR-expressing accessory cells. The latter difference is expected to result in differences in compartmentalization of in vivo agonist function. Notably, the fully humanized huIgG1 ChiLob7/4-FF Ab could be used to generate new IP.
In view of the above, the Glycotope GmbH team is performing further in-depth analysis of the FcγR-dependent agonist function of the low-fucosylated huIgG1-ChiLob7/4 Ab. While this work was included in the IACT extension request, the fact that this request was not granted implies that this work will continue outside of the IACT project.
Task 7.5 Pre-clinical testing of Glyco-optimized mAb in vivo (Partners 3, 4, 6, 7)
This task has been replaced by the follow up work described under task 7.4.
WORK PACKAGE 8. DEVELOPMENT OF OPTIMIZED 2ND-GENERATION AGONIST IS-ABS THROUGH MANIPULATION OF THE FAB-DOMAIN
Objectives
• Generation of dual-specific Abs that simultaneously trigger two complementary immune pathways
• Development of trimeric antibodies with enhanced agonist function
Progress for each task
Task 8.1 Isolation of dual-specific Abs (Partner 8)
The three target antigens CD40, OX40 and 4-1BB have been produced in HEK293EBNA cells as soluble, extra-cellular domains with 8xHis tag. As judged by SDS-PAGE all protein preparations showed high purity and according to size exclusion chromatography the monomeric protein form was dominating (above 80%). ELISA with commercial target-specific antibodies showed specific and dose-dependent binding. In addition, transfected CHO cells expressing the full length proteins have been generated. Flow cytometry analysis with commercial antibodies showed high cell surface expression of all three proteins.
Using the BioInvent n-CoDeR® human antibody library, mono-specific scFv-fragments have been successfully isolated against each of the three targets; 134, 112 and 100 unique scFv for the targets CD40, OX40 and 4-1BB, respectively. Binding of isolated scFv-fragments to soluble proteins was verified using ELISA and binding to cell-surface expressed proteins was verified using Fluorometric Microvolume Assay Technology (FMAT) and flow cytometry. Several of the scFv-fragments exhibit EC50 values in the low-nM range in ELISA.
Using the mono-specific scFv-fragments as starting material, various shuffling and combination procedures were performed. The resulting Ab-fragments were screened for binding to two out of the three antigens CD40, OX40 and 4-1BB
On the basis of these analyses, the following high-affinity, dual-specific Abs were generated and shipped to partners in Southampton, Pamplona and Milan:
• 3x CD40/OX40 dual-specific Abs: #2, #5 and #17
• 2x CD40/4-1BB dual-specific Abs: #4 and #19
• 2x OX40/4-1BB dual-specific Abs: #6 and #20
For reasons of confidentiality related to the protection of intellectual property, the full details of these findings cannot be disclosed in this public report.
Biochemical analysis of the Abs revealed that both expression and folding are comparable to that of conventional Abs. Most dual-specific Abs bind both ligands with EC50 in the sub- to low (single digit) nM range. This is true for binding to purified extracellular domains in ELISA as well as target-transfected and endogenously-expressing cells in flow cytometry.
Task 8.2 In vitro functional testing of dual-specific Abs (Partners 3, 4, 6, 8)
Functional testing of dual-specific Abs comprising these domains by the partners in Southampton, Milan and Pamplona indicated that one of the CD40 specific Ab and all three of the 4-1BB specific Abs tested displayed clear-cut agonist activity. The second CD40 Ab only displayed weak agonist activity, and the same applied to the only available OX40 Ab. In summary, the experiments performed by BioInvent in cooperation with other IACT partners have provided proof of concept for a novel format of dual-specific Abs that is distinct from previously reported designs.
For reasons of confidentiality related to the protection of intellectual property, the full details of these findings cannot be disclosed in this public report.
Task 8.3 Initial testing of a Trimerbody against mouse 4-1BB (Months 0-12; partner 9)
Four Trimerbody constructs against mouse 4-1BB were generated on basis of the scFv fragment derived from the well-defined anti-mouse 4-1BB conventional mAb (clone 1D8, IgG2a). Three 1D8 scFv-based were N-terminal Trimerbodies (1D8N) including different lengths of the flexible linker region connecting the scFv to the murine collagen XVIII trimerization domain: 25-, 5- and 0-mer (1D8N25, 1D8N5, and 1D8N0 respectively); and a short (0-mer 1D8 scFbased C-terminal trimerbody (1D8CO).
All four 1D8 scFv-based Trimerbodies were secreted as soluble proteins by transfected HEK-293 cells and specifically recognized in ELISA the immobilized murine m4-1BB:hFc protein. Trimerbody specificity was further confirmed by FACS on the surface of cells expressing the mouse 4-1BB receptor. Affinity studies (Biacore) against immobilized recombinant mouse 4-1BB Preliminary biosensor kinetics analysis indicates that 1D8 scFv-based N- terminal Trimerbodies have higher avidity and slower off-rates compared to native 1D8 IgG2a. Of interest, two Trimerbody configurations have a dissociation rates 10 times slower compared to 1D8 IgG2a. For reasons of confidentiality related to the protection of intellectual property, the full details of these findings cannot be disclosed in this public report. Size exclusion chromatography-multi-angle laser light scattering signal from oligomeric analysis and molar mass demonstrate that the 1D8 scFv-based Trimerbodies behave predominantly as trimers with none to minor higher molecular weight aggregation.
All Trimerbodies were very stable in human serum as demonstrated by the 80-100% preservation of the binding activity to m4-1BB:hFc even after 4 days incubation time. Additionally, Performing several freeze/thaw cycles of the different Trimerbody samples showed that three out of the four Trimerbodies were very stable when subjected to these conditions. Of interest, one particular configuration (1D8CO) was labile under these conditions.
In collaboration with partner 4 (Melero), In vitro functional assays for testing T cell co-stimulatory impact through mouse 4-1BB and subsequently used to test the signalling impact through the 4-1BB receptor by the mouse 1D8 IgG2a compared to 1D8 scFv-based Trimerbodies. We have now obtained solid and reproducible results supporting the fact that, though to a different extend, three Trimerbodies were superior to the parental 1D8 IgG2a conventional mAb. The T -cell co-stimulation in vitro readouts included cell proliferation/survival, Interferon-γ and IL-2 production. Based on these findings, two preferred 1D8 scFv-based Trimerbody configurations were selected for further studies. For reasons of confidentiality related to the protection of intellectual property, the full details of these findings cannot be disclosed in this public report.
Task 8.4 Generation and initial testing of a Trimerbody against human 4-1BB (Months 12-30; partner 9)
Based on the above findings (Task 8.3) generation of a human trimerbody, on basis of the SAP3.28 Ab from Southampton, focused on generation of an N18-trimer. A QC-ed preparation of trimerbody was produced and found to exhibit excellent binding to human 4-1BB. In view of the outcome of the functional tests with the anti-murine trimerbody (see below), a tumor antigen targeted hexavalent bispecific trimerbody was successfully generated.
Task 8.5 In vivo functional testing of anti-4-1BB Trimerbodies (Months 12-48; partners 1, 3, 4, 9)
Generation of the anti-mouse 4-1BB trimerbody based on the 1D8 monoclonal IgG type Ab was reported previously. More recent data show that the agonist activity of this trimerbody is, unfortunately, not significantly greater than that of the original, bivalent 1D8 Ab. In view of this finding, the trimerbody design was expanded to a hexavalent bispecific design that additionally includes three tumor-targeting domains against a clinically well-validated tumor target (Leadartis chooses not to disclose this target, which is expressed at the cell surface in certain human tumors). This new construct can effectively be produced and binding assays show that it efficiently binds both 4-1BB and the tumor antigen in a simultaneous fashion and slower dissociation rates compared to the 1D8 Ab. Moreover, the agonist activity of this bispecific construct is superior to that of both the parental anti 4-1BB 1D8 Ab and the original monospecific 1D8-based trimerbody when applied in the presence of plastic-immobilized or cell surface-expressed tumor target antigen.
The in vivo half-live of the new bispecific construct is approx. 20 hrs. This is better than that of the monospecific trimerbody, but shorter than that of conventional Abs. As such, it may offer a good compromise between in vivo exposure levels that require long half-life and safety considerations that are mitigated by shorter half-life. Moreover, targeting of 4-1BB agonist activity to the tumor microenvironment may offer an attractive ‘best in class’ strategy for improving the therapeutic index of agonist 4-1BB Abs, which is of interest in view of the fact that the conventional 4-1BB Abs tested thus far in clinic are not well-tolerated (see above). Further in vivo experiments with the bispecific, hexavalent construct are ongoing (beyond the scope of the terminated IACT project).
Potential Impact:
Achievements concerning preclinical research and drug development
The IACT project brought together a team of experts on the subject of agonist immunostimulatory antibodies with the aim of pushing this field, which is lagging behind that of the immune checkpoint blockers, forward. From a scientific point of view, the most prominent output was generated by the lab of Martin Glennie at Southampton University (partner 3) in the form of two high impact papers that were authored by him and Dr. Anne White, the post-doctoral scientist appointed on the IACT project for partner 3:
• Conformation of the human immunoglobulin G2 hinge imparts superagonistic properties to immunostimulatory anticancer antibodies. White AL, Chan HT, French RR, Willoughby J, Mockridge CI, Roghanian A, Penfold CA, Booth SG, Dodhy A, Polak ME, Potter EA, Ardern-Jones MR, Verbeek JS, Johnson PW, Al-Shamkhani A, Cragg MS, Beers SA, Glennie MJ. Cancer Cell. 2015 Jan 12;27(1):138-48.
• Complex Interplay between Epitope Specificity and Isotype Dictates the Biological Activity of Anti-human CD40 Antibodies. Yu X, Chan HTC, Orr CM, Dadas O, Booth SG, Dahal LN, Penfold CA, O'Brien L, Mockridge CI, French RR, Duriez P, Douglas LR, Pearson AR, Cragg MS, Tews I, Glennie MJ, White AL. Cancer Cell. 2018 Apr 9;33(4):664-675.e4
This work has greatly increased our insight in the structure function relationship of agonist immunostimulatory antibodies and is of great value for the design of the next-generation drugs.
IACT has also enabled the lab of Rienk Offringa at German Cancer Research Center (Partner 1) to pursue the search for optimal combinations of cytostatic drugs and agonist immunostimulatory antibodies, through the provision by the Glennie-lab of the gene constructs for mouse chimeric antibodies targeting mouse CD40, OX40 and 4-1BB. This work resulted in the identification of MEK-inhibitors as highly effective companion of agonist anti-CD40 antibodies in tumor treatment, as well as in the identification of whole tumor transcriptomics for PD biomarker analysis of drug action (Baumann, Offringa et al, Pro-immunogenic impact of MEK inhibition synergizes with agonist anti-CD40 immunostimulatory antibodies in T cell mediated tumor therapy, manuscript under consideration for publication in Immunity). Furthermore, the IACT-project supported the analysis of the natural anti-tumor T-cell response in human pancreatic cancer, which will serve as an important base line for the planned neo-adjuvant trial in this indication:
• Identification of a tumor-reactive T-cell repertoire in the immune infiltrate of patients with resectable pancreatic ductal adenocarcinoma. Poschke I, Faryna M, Bergmann F, Flossdorf M, Lauenstein C, Hermes J, Hinz U, Hank T, Ehrenberg R, Volkmar M, Loewer M, Glimm H, Hackert T, R. Sprick MR, Höfer T, Trumpp A, Halama N, Hassel JC, Strobel O, Büchler M, Sahin U & Offringa R. OncoImmunology 2016, DOI: 10.1080/2162402X.2016.1240859
Work by the lab of Ignacio Melero at CIMA, as supported in part by the IACT funding, resulted in several papers that provided further insight in the mechanism of action of agonist anti-4-1BB antibodies, including the following two:
• CD137 (4-1BB) Costimulation Modifies DNA Methylation in CD8+ T Cell-Relevant Genes. Aznar MA, Labiano S, Diaz-Lagares A, Molina C, Garasa S, Azpilikueta A, Etxeberria I, Sanchez-Paulete AR, Korman AJ, Esteller M, Sandoval J, Melero I. Cancer Immunol Res. 2018 Jan;6(1):69-78.
• Mitochondrial Morphological and Functional Reprogramming Following CD137 (4-1BB) Costimulation. Teijeira A, Labiano S, Garasa S, Etxeberria I, Santamaría E, Rouzaut A, Enamorado M, Azpilikueta A, Inoges S, Bolaños E, Aznar MA, Sánchez-Paulete AR, Sancho D, Melero I. Cancer Immunol Res. 2018 Jul;6(7):798-811.
In addition, IACT resulted in the involvement of two SMEs in the development of immunostimulatory agonist antibodies, in particular the vaccine company BioNTech/Ribological (partner 10) and the antibody company Glycotope (partner 7). The resulting engagement of BioNTech in this area of research is best illustrated by the fact that this SME emerged as the driver of the production of clinical grade anti-CD40 Ab ChiLob7/4, starting with the generation of new GMP-grade producer lines and evolving to the generation and QC of the clinical antibody. Moreover, BioNTech took a licence on the CRUK-owned ChiLob7/4 antibody with the aim of developing this further. Furthermore, glyco-optimization as performed by Glycotope resulted in a more powerful anti-CD40 Ab that has been filed in the following IP-application:
• WO2018178046: HUMANIZED ANTI-CD40 ANTIBODIES
For SME Leadartis (partner 9), IACT has been a very effective platform for exploiting their Trimerbody technology, of which in particular the bispecific, hexavalent constructs show great promise with respect to therapeutic index: the mitigation of toxicity by means of tumor targeting and an intermediate (16hr) half-life, as published in:
• A tumor-targeted trimeric 4-1BB-agonistic antibody induces potent anti-tumor immunity without systemic toxicity. Marta Compte, Seandean Lykke Harwood, Ines G. Muñoz, Rocio Navarro, Manuela Zonca, Gema Perez-Chacon, Ainhoa Erce-Llamazares, Nekane Merino, Antonio Tapia-Galisteo, Angel M. Cuesta, Kasper Mikkelsen, Eduardo Caleiras, Natalia Nuñez-Prado, M. Angela Aznar, Simon Lykkemark, Jorge Martínez-Torrecuadrada, Ignacio Melero, Francisco J. Blanco, Jorge Bernardino de la Serna, Juan M. Zapata, Laura Sanz & Luis Alvarez-Vallina. Nature Communications volume 9, Article number: 4809 (2018)
Furthermore, the following patient application was filed:
• EP18382401.0: TRIMERIC POLYPEPTIDE COMPLEXES AND USES THEREOF
Last but not least, SME partner BioInvent (partner 8) was successful in obtaining proof of concept for the use of their n-CoDer antibody libraries towards the generation of Dual (bispecific/bivalent) agonist antibodies. Impressive was the fact that this proof of concept was successfully obtained for all three targets (CD40, OX40, 4-1BB) in a single attempt. A patent application on these results is currently in preparation. Furthermore, BioInvent entered and exclusive collaboration with Glennie-Lab (partner 3) to accelerate the development of optimal anti-OX40 and anti-4-1BB clinical candidate antibodies, resulting in a series of antibodies with superior features as compared to the OX40 and 4-1BB antibodies that formed the basis of the IACT consortium.
Achievements concerning clinical testing of agonist immunostimulatory antibodies
Unfortunately, the primary clinical goal of IACT, the testing of agonist anti-CD40 antibody ChiLob7/4 in studies in patients with head and neck cancer (HNSCC) and pancreatic ductal adenocarcinoma (PDA) has not been achieved during the 4-year project period. The reason for this is the almost incredible number of unforeseen complications that were encountered during the production of the clinical-grade antibody, as described in detail in the progress report under work package 2, task 2.1. Nevertheless, the following two points need to be emphasized in this respect:
• Although the IACT partners involved were under enormous pressure due to these complications, both with respect to time and resources, they never cut corners that could have compromised the safety of the drug preparation
• In spite of the fact that these delays eventually resulted in loss of a significant part of the IACT-related EU funding, due to the fact that the 2-year no-cost project extension was not granted, the work on finalizing the QC trajectory for the clinical ChiLob7/4 antibody is continuing and the clinical studies will be conducted, albeit in modified form and through alternative project funding (see report on work packages 2 and 3). The main reason for the perseverance of the IACT partners in this respect is that the (modified) studies are still state of the art, with high potential significance for the development of this drug class.
Irrespective of the above, the IACT consortium inspired a second clinical trial that, although not focused on agonist immunostimulatory antibodies, develops very well: the HARE-40 trial involving the testing of the HPV16E6/E7-specific synthetic RNA vaccine that is based on the vaccine technology of SME-partner BioNTech (partner 10). Due to the pre-clinical finding that agonist anti-CD40 antibodies did not significantly enhanced the immunogenicity and anti-tumor impact of the HPV16 RNA vaccine, the IACT partners decided to perform a single agent vaccine study in patients with HPV16-positive HNSCC. Also for the preparation of this drug (vaccine), many unanticipated hurdles needed to be overcome, as detailed in the progress report under work package 3, Task 3.2. Although this resulted in a significant delay in the start of the trial, this trial was started before the end of the 4-year project period. The initial results from this study look very promising. In view of this, as well as the equally promising results of similar vaccination studies as performed by BioNTech in patients with metastatic melanoma, also this trial will be continued, in spite of the discontinued EU funding.
Socio-economic impact
Although it is difficult to measure - or even predict - the socio-economic impact of a 4-year research project, it is evident that the four SME-partners were intensely engaged in the IACT consortium and have used this opportunity, each in their own way, to (further) explore in the field of agonist immunostimulatory antibodies. Notably, it is commonly known that only a fraction of the drug development projects started results in a successful clinical drug, and that the path from bench to patient is very long and windy, as also witnessed by this final report. In this context, the IACT collaborative was, and still is, a very good shot at goal.
Dissemination activities
The dissemination activities of the IACT consortium over the three project periods were as follows:
First project period:
1.
During the first project year, we decided on connecting with other EU-consortia that focus on cancer immunotherapy, with the aim to identify potential synergies. To this end, a special workshop on cancer immunotherapy-focused EU-consortia was organized by the IACT coordinator (Offringa) and a representative of partner 10 (Dr. Cedrik Britten) during the 2014 edition of the yearly meeting of the Association for Cancer Immunotherapy (CIMT) in Mainz, Germany. EC Scientific Officer Dr. Jan van de Loo was included in the organization and consulted on selecting the EC projects that were featured during this session. During this session, representatives of each of the selected consortia gave a brief introduction to their project. The program was as follows:
• Introduction; the aims of this Session (Cedrik Britten, Mainz; Germany)
• EC funded consortia and cancer (Jan van de Loo, Brussels, Germany)
• Adoptive Engineered T-Cell Targeting To Activate Cancer Killing - ATTACK (R. Hawkins, Manchester, UK)
• Inflammation and Cancer Research in Europe - INFLA-CARE (A. Eliopoulos, Heraklion, Greece)
• Advanced T-cell Engineered for Cancer Therapy - ATECT (S.A. Quezada, London, UK)
• Glioma Actively Personalized Vaccine Consortium - GAPVAC (H. Singh, Tübingen, Germany)
• Mutanome Engineered RNA Immuno-Therapy - MERIT (U. Sahin, Mainz Germany)
• Immunostimulatory Agonist antibodies for Cancer Therapy - IACT (R. Offringa, Heidelberg, Germany)
• Panel discussion & Summary. How to reach more synergy across consortia? (Moderators: R. Offringa & U. Sahin)
This initial workshop has rendered a number of leads for interactions between these EU consortia. We will follow up on this during the 2015 IACT consortium meeting and during the 2016 edition of the CIMT conference. In view of the latter, The IACT project coordinator (Offringa) has recently been appointed as member of the CIMT board.
2.
IACT has been referred to by representatives of partners 3 (Glennie) and 1 (Offringa), during their talks at the CRI/CIMT/EATI/AACR cancer immunotherapy meeting in New York, Sept 2015.
3.
The 2015 IACT consortium meeting was organized in conjunction with an international conference, organized by IACT partner 5 (Melero) in Pamplona Spain, which featured an excellent line up of experts from the USA and Europe on immunostimulatory Abs. As such, this was an excellent platform for ‘marketing’ the IACT consortium. The program for this meeting looked as follows (IACT partners underlined):
International Symposium: ‘Immunostimulatory monoclonal antibodies and immunomodulation: harvesting the crop’.
Coordinators:
• Dr. Ignacio Melero, Centro de Investigación Médica Aplicada. Clínica Universidad de Navarra. Pamplona
• Dr. Juan José Lasarte, Centro de Investigación Médica Aplicada. Pamplona
Program October 18, 2015:
• Ronald Levy “Targeting the Tumor and Targeting the Host”. Stanford University Medical Center, Stanford, United States
• Drew Pardoll “Protumorogenic immunity: the dark side of the immune system”. Johns Hopkins Medical Institutions, Baltimore, United States
• Jim Allison ”Immune Checkpoint Blockade in Cancer Therapy: New Insights and Opportunities”. MD Anderson Cancer center. Houston. United States
• Antoni Ribas “Melanoma response and resistance to PD-1 blockade” UCLA Medical Center. Los Angeles, California, United States
Program October 19, 2015:
• George Coukos “Understanding tumor microenvironment to optimize immunotherapy. Lessons from ovarian cancer” University of Laussane Medical center, Switzerland
• Ignacio Melero “CD137 and friends in immunotherapy combinations”. Centro para la Investigación Médica Aplicada-Clínica Universidad de Navarra, Pamplona, Spain
• Sergio Quezada “Immune Regulation at the Tumour Site”. UCL Cancer Institute, London, United Kingdom
• Suzanne Topalian “PD-1 pathway blockade: a ‘common denominator’ treatment approach mediates a ‘personalized’
• response”. Johns Hopkins Medical Institutions, Baltimore, United States
• Mario Sznol “Immune Therapy of Advanced Melanoma based on Immune Checkpoint Inhibitors”. School of medicine of Yale. New Haven. United States
• Pam Sharma “Immunotherapy frontiers against prostate and bladder cancer”. MD Anderson Cancer Center. Houston. United States
• Tim Chan “Cancer genomes and immunotherapy efficacy”. Memorial Sloan Kettering Cancer Center, New York city, United States
• Sandra Demaria “Mechanisms of synergy of radiotherapy with antibodies targeting negative immune regulators in the tumor microenvironment”. New York University School of Medicine. New York city, United States
• Janice Taube “Tumor markers of response and resistance to PD-1/PD-L1 checkpoint blockade”. Johns Hopkins Hospital. Baltimore. United States
Program October 20, 2015
• Cornelis Melief “Therapeutic cancer vaccines and their application in combination treatment of cancer” Leiden University Medical Centre. ISA Pharmaceuticals. Leiden, The Netherlands
• Caroline Robert “Melanoma management : New Treatments and new Toxicities” Institute Gustave Roussy, Villejuif, France
• John Haanen ”Checkpoint inhibitors: who should we treat?” The Netherlands Cancer Research Institute. Amsterdam, The Netherlands
• Francisco Sánchez Madrid “Leukocyte activation receptor CD69 and ligands: Immunoregulatory role in inflammatory processes”. Hospital de la Princesa. Madrid, Spain
• Juan José Lasarte “FOXP3 as a target for Immunotherapy”. Centro para la Investigación Médica Aplicada. Pamplona, Spain
• María Luisa Toribio “Therapeutic antibody targeting of human T-cell acute lymphoblastic leukemia initiang cells”. Centro de Biología Molecular Severo Ochoa. Madrid, Spain
• Mario Colombo “Multiple layers of immunoregulation: the myeloid connection”. Instituto Nazionale Tumori Experimental Oncology. Milan, Italy
• Rienk Offringa "A rational approach towards implementation of immunotherapy for pancreatic ductal adenocarcinoma”. German Cancer Research Center, Heidelberg, Germany
• Christian Ottensmeier “Genomic assessment of tumour immune profiles in head and neck cancer”. University of Southampton School of Medicine, Southampton, United Kingdom
• Paolo Ascierto “Combining checkpoint inhibition maximizing the body’s immune response to cancers”. Fondazione Pascale Istituto Nazionale Tumori di Napoli, Italy
• Martin Glennie ”Designing immunostimulatory anti-TNFR mAb to promote anti-cancer immunity”. University of Southampton School of Medicine, Southampton, United Kingdom
Second project period:
4.
IACT coordinator Offringa has hosted a second workshop on EU-funded immunotherapy projects during the 2016 edition of the CIMT meeting. The program of this session was as follows (IACT speakers underlined):
• Rienk Offringa, German Cancer Research Center, Heidelberg, Germany:
• Fostering interaction and synergy between ongoing EU Cancer Immunotherapy Consortia
• Harpreet Singh, Immatics Biotechnologies, Tuebingen, Germany
• GAPVAC: First results from a personalized peptide vaccination study in glioblastoma
• Ann White, Southampton University, Southampton, UK
• IACT: The human IgG2 format imparts potent agonistic properties to immunostimulatory antibodies
• Fred Falkenburg, Leiden University Medical Center, Leiden, The Netherlands
• T-Control: Therapy of hematological malignancies using streptamer-selected cell products
• Jan van de Loo, Scientific and Policy Officer Cancer Research, European Commission, Brussels, Belgium
• EU Cancer (Immunotherapy) Consortia: what we are looking for in the next round.
• Zlato Trajanoski, Medical University of Insbruck, Innsbruck, Austria
APERIM: Advanced Bioinformatics platform for personalized cancer treatment
• Sebastian Kobold, Ludwig Maximilians University Munich, Munich, Germany
• IMMUTRAIN: Training Network for the Immunotherapy of Cancer
Third project period:
During the final project year, the IACT partners focused on consolidating the efforts to make the clinical trials going and applying for the no-cost project extension.
Exploitation of results
The IACT project was set up to allow the SME partners freedom to exploit the proceeds of their work. This applies in particular to the glyco-optimized forms of agonist immunostimulatory antibodies developed by Glycotope (partner 7), the novel dual antibodies and optimal anti-OX40 and 4-1BB antibodies generated by BioInvent (partner 8), the Trimerbodies and dual-specific variants generated by Leadartis (partner 9) and the huIgG1 and huIgG2-forms of the anti-CD40 Ab ChiLob7/4 as licensed by BioNTech (partner 10), as well as their proprietary HPV16E6/E7-specific synthetic RNA vaccine. In view of this, the further development of these drug formats is, at this stage, left at the discretion of these SME partners.
For the exploitation of the pre-clinical results from the academic labs of Glennie (partner 3) and Offringa (partner 1), completion of the delayed clinical studies with anti-CD40 antibody ChiLob7/4 is a critical step, because these studies constitute the best stepping stone towards follow up studies testing optimized antibodies or combination regimes. It is for this reason that the IACT partners involved keep pushing ahead with the ChiLob7/4 clinical studies, in spite of the fact that the request for a no-cost extension of the IACT project was not granted.
The bottom line is that the returns in life sciences and drug development do not come overnight. Maybe the initial 4-year time span chosen for the IACT project was too short; in fact it was proven to be too short. However, given the global competition and resources involved, drug development can only be successful if performed under time pressure. Notably, the ambitious time schedule and work plan was completely in line with the call by the EU for high-risk, high gain projects, as illustrated by the fact that the project was funded on the basis of high scores.
The manner in which the IACT project, as well as several other EC projects that include clinical trials, encounter delays with the execution of clinical studies is ‘food for thought’ for the EC committees designing calls and managing the funding. These committees may be formally correct in restricting the funding to the time frame allotted, allowing project extension under exceptional circumstances only. However, a more realistic and productive manner of judging these matters would be driven by the question whether the consortia concerned are pursuing the goals that have been set and are deploying the EU funding towards achieving these goals. In this respect, it should be emphasized that the extension request for the IACT project included the explicit reservation of the funds that were intended for the clinical trials. In view of this, the rejection of this request – which concerned no more than a no-cost extension involving already allocated EU-funding - could be considered rather non-constructive and bureaucratic.
In view of the latter point, a potential goal for the EC could involve the in-depth evaluation of IACT and other life science drug development projects with respect to the following questions:
• What is the ideal balance between an ambitious, high-risk, high-gain project proposal versus a realistic, down-to-earth proposal?
• Would a proposal of the latter type ever make it through the peer review process?
• Do the EC committees concerned have true experts on board with hands-on experience in the areas of clinical studies and drug development?
• Why is the EU lagging so far behind in drug development and clinical drug testing as compared to the USA and - to an increasing degree – China?
• What exactly are the European handicaps in this respect?
• Which concrete actions could be undertaken to overcome these hardwired handicaps?
List of Websites:
www.IACT-project.eu
Prof. Dr. Rienk Offringa
K.H. Bauer Foundation-endowed professor
Head Div. Pancreatic Cancer Research
Surgery Clinic, University Hospital Heidelberg, Im Neuenheimer Feld 110, 69120, Heidelberg
Head Dept. Molecular Oncology of Gastrointestinal Tumors
Head DKFZ-Bayer Joint Immunotherapeutics Lab.
German Cancer Research Center (DKFZ-D200); Im Neuenheimer Feld 280, 69120 Heidelberg
Phone (secr.): +49 (0)6221-42.3144
Phone (direct.): +49 (0)6221-42.3140
Conference phone:: +49 (0)6221-42.5989
FAX: +49 (0)6221-42.4625
E-mail: r.offringa@dkfz.de