Final Report Summary - TUMADOR (Development of very promising humanized therapeutic mAbs efficiently neutralizing a novel immunosuppressive pathway involved in a wide range of cancers)
Executive Summary:
Based on the broad immuno-suppression and strong tumor promotion (proliferation and dissemination) exerted by adenosine, therapeutic strategies aimed at targeting adenosine pathway are highly promising. However, given involvement of the four known specific adenosine receptors, the use of mAbs neutralizing the enzymatic function of CD73, the sole 5’ ecto-nucleotidase expressed at surface membrane, thus preventing dephosphorylation of AMP into adenosine, represents an attractive, potent and safe strategy for the development of innovative immunotherapy.
The TumAdoR project aims at paving the way to clinical trials of potent neutralizing anti-CD73 mAb candidates in cancer patients. It should validate a novel cancer immunotherapy approach applicable to a wide range of cancers preventing dissemination and relapse through restoration of anti-tumour immune memory.
The program has progressed very well with the development and selection of the lead candidate for a neutralizing anti huCD73 mAb. The final lead candidate anti-huCD73 Ab has been humanized, and produced in large batches for safety evaluation. The lead candidate has been further characterized and compared with the Abs from the competitors, and back-up anti-huCD73 Abs have been generated.
Many novel data on huCD73 biology on CD4+ T cells and A2A functionality on CD8+ T cells and NK cells have been generated, increasing the relevance of CD73 therapeutic targeting. Current data also indicate lack of safety issues.
Furthermore, standardized tools for IHC (automated protocol and quantification grid) have been developed for the analysis of CD73 target expression on large cohorts of human solid tumors. Several partners of the project have generated knowledge on CD73 distribution in human selected tumors and its impact on tumor progression and patient survival allowing a comprehensive huCD73 expression that will be critical for the future orientation of the clinical development. A new mAb has been selected, for use in future IHC companion diagnostic for anti-huCD73 treatment.
A major investment has been made by the partners to develop mouse models for in vivo evaluation of anti-CD73 mAbs, despites technical difficulties the consortium select relevant models to assess CD73 biology in vivo and the impact of the drugs on immune parameters. Finally, tools for the immune-monitoring of the future clinical trial with the drug candidate are now all in place.
The results obtained in the course of TumAdoR have been presented at several scientific events (31 dissemination activities, 1 PhD, 3 publications, 8 presentations, 14 posters, others 5).
Project Context and Objectives:
Cancer is a clinical and societal burden that justify important Research and Development effort to better understand the biology of these disease and their spread, prerequisite for the identification of new therapeutic pathway and potent clinical tools. Among them monoclonal antibodies (mAbs) for targeted cancer therapies has generated in the last decade a lot of enthusiasm and hope for patients and physicians in the frame of more efficient, safer, personalized medicine. The mAbs oncology market is constantly growing, reaching nearly 20 billion Euros in 2012. One of the aims of the TumAdoR project is to boost the EU competitiveness of health-related industries and businesses, as well as address global health issues.
The major challenge for success that has to be addressed in cancer immunotherapy is that tumors fool the host immune system, favoring tumor spread and shortening patient survival. Some tumor-driven immunomodulating mechanisms have been identified: secretion of immunosuppressive cytokines, recruitment of suppressive immune cells, local amino-acid starvation, altered activation of immune receptors and engagement of inhibitory receptors. In addition to being involved in the natural progression of cancer, this immunosuppressive environment affects the efficacy of anticancer treatments. At the beginning of TumAdoR project in 2013, few targets contributing to tumor immune-suppression have been clinically evaluated using neutralizing mAbs and the successes in clinical trials of anti-PD1 and PD-L1 mAbs opened the door to a highly promising novel class of therapeutics targeting tumor-driven immunosuppressive pathways. However, as severe autoimmune syndromes related to disruption of immune tolerance have been reported after treatment with anti-CTLA-4 mAbs, development of strategies targeting other tumor-driven immunosuppressive pathways with increased efficacy and lower toxicity was urgently needed.
Extracellular Ado has multiple activities that contribute to tumor progression (Figure 1). Ado has direct effects on tumor cells contributing to their proliferation and favoring tumor cell motility and metastases (for review Allard 2012). Furthermore Ado has direct effect on endothelial cells and increase angiogenesis and protects cancer cells from hypoxia (Ahmad 2009). In addition to these direct tumor promoting activities, Ado has critical immunomodulatory properties, suppresses both innate and adaptive immune responses and inhibits pro-inflammatory cytokine production. Finally, Ado inhibits immune cell recruitement by inhibiting expression of adhesion molecules on endothelial cells and T cells.
The conversion of ATP into Ado is tightly regulated by cell membrane ectoenzymes of the NTPDase family and tissue non-specific alkaline phosphatases (Figure 2). The conversion of ATP into AMP is predominantly catalyzed by CD39. CD73, a glycosylphosphatidylinositol (GPI)-anchored nucleotidase present in cell membrane lipid rafts, active as a disulfide-linked (for review Yegutkin 2008). Importantly, whereas the activity of CD39 could be reverted by the action of NDP kinase and adenylate kinase, CD73 activity is virtually irreversible thus representing a crucial checkpoint in conversion of pro-inflammatory ATP into immunosuppressive Ado. Indeed, ATP is a danger signal released by damaged and dying cells that primes immune responses through the ligation of P2X and P2Y purino-receptors. By contrast, through the activation of G-protein-coupled seven transmembrane spanning receptors (A2A, A2B), Ado suppresses the immune responses of not only CD4+ and CD8+ T cells, but also of the antigen presenting cells, NK and NKT cells (for review Stagg 2010).
Thus, based on the broad immuno-suppression and strong tumor promotion (proliferation and dissemination) exerted by adenosine (Ado), therapeutic strategies aimed at inhibiting adenosine appeared highly promising. For one research team at the Centre Anticancéreux Léon Bérard (CLB, coordinator of TumAdoR project) the use of mAbs neutralizing the enzymatic function of CD73 receptor expressed at surface membrane, thus preventing the generation of Ado, represented an attractive strategy for the development of innovative immunotherapy. In addition, CD73 neutralization appears to reduce tumorigenesis and metastasis. Combining the expertise of 6 top-level scientific partners in Europe and the US, the TumAdoR project aims at paving the way to clinical trials of potent neutralizing anti-huCD73 mAb candidates in cancer patients. Neutralizing CD73 activity is expected to be efficient in a wide range of cancers, and to have limited adverse side effects.
TumAdoR project should in the coming years validate a novel cancer immunotherapy approach. This novel approach will enable the development of specific anti-tumor immune response inhibiting cancer progression and dissemination and preventing relapse through restoration of anti-tumor immune memory.
Specifically, the objectives of the TumAdoR program were to 1- generate a comprehensive knowledge of CD73 distribution in tissues and major type of tumors, 2- understand CD73 biology as an immune checkpoint in human, 3- develop therapeutic antibodies neutralizing CD73 enzymatic activity, 4- set up new preclinical models for relevant in vivo therapeutic evaluation of anti-CD73 therapeutic candidate, 5- develop specific assays for toxicity evaluation of anti-CD73 therapeutic candidate, 6- set new Immuno-monitoring tools to monitor anti-CD73 mAb clinical trial.
Project Results:
Progress in Target and associated biomarkers expression, and clinical correlation on survival/relapse.
Information on CD73 prevalence in human tumor types and its impact on tumor progression and patient survival was limited, so the consortium undertook the development of specific tools to analyze CD73 expression and markers associated with CD73 biology on well clinically annotated cohorts of selected tumors types.
A commercial antibody was selected that provided specific and unambiguous staining (rabbit mAb D7F9A, Cell Signalling Technology). In addition, a short-list of biomarkers associated with CD73 and its biological function has been generated through bioinformatics analysis. The evaluation, harmonization and automatization of the IHC protocols for CD73 detection as well as the quantification grid have been cross-validated between the partners (D7.5 D2.8). A major investment was performed by CLB, UTU and CHUV partners to analyze CD73 expression on multiple tumor types using either Tissue multi array (TMA) or full sections according to the available tumor material and the size of the cohort to be analyzed. Retrospective tumor collections with clinical annotations have been defined in different partners’ hospital institutions and IHC studies have been done on different tumor types including Melanoma, Lung, Breast, Ovarian, Bladder, Prostate, Head and Neck (H&N), hepatobiliopancreatic system and thyroid carcinoma as well as different subtypes of Lymphoma, all with unmet medical needs.
Overall, we observe variable expression on tumor cells according to pathologies ranging from few percentages of cases (breast tumors, lymphoma) to 100% of cases being positive (some hepatobiliopancreatic tumors). Cancers showing the highest rate of CD73 expression are represented by the hepatocellular and pancreatic ductal malignancies , lung, bladder, H&N carcinoma and melanoma. In most instances, the expression on tumor cells is focal and appears in several pathologies to be associated with the invasive front (lung, some thyroid carcinomas, melanoma) (D2.8). Beside expression on tumor cells, blood vessels express CD73 in most tumors, as well as mesenchymal stromal cells. CD73 expression on immune cells is also observed in most tumors at variable levels from few percentages up to >50% depending on tumor type (D2.10).
Clinical correlations have been performed when the percentage of expression was significant and clinical follow-up sufficient. In certain tumor types CD73 expression on tumor cells is of bad prognosis, while in others CD73 expression is of good prognosis. The expression on stromal cell and immune cells is in some instance associated with good prognosis, but this might be linked to the stromal cells and immune cell types themselves as CD73 expression varies according to different infiltrating cell types.
Thus, based on tumor cell expression and outcome, some tumor types are considered as potential primary tumor targets. Of note CD73 IHC might be of help in the diagnosis of some cases of pancreas and thyroid tumors difficult to interpret by anatomo-pathologists. In parallel, a proprietary anti-huCD73 has been generated and selected for potential future development of an IHC companion diagnostic (D2.9).
Furthermore, UTU partner established a standardized test format for subsequent large-scale analysis to quantify in biological fluids soluble CD73 (sCD73) by ELISA and demonstrated the strong correlation with the enzymatic activity (D2.4). This tool has been used to measure sera concentrations of sCD73 in the FinRisk 2002 cohort (1369 men, 1426 women). Prevalent or incident cancers were diagnosed in 358 individuals (12,8%) but no difference in sCD73 serum concentrations was observed between individuals with and without cancer. However, among patients with prevalent cancer, those who died had statistically higher sCD73 values than those who survived. (D2.4). Based on these results we believe that sCD73 alone may not have additive value in finding the right individuals for clinical trials or to follow up their treatment response. However, the feasibility of combining sCD73 measurements with other parameters have also been investigated and analyses are ongoing (D7.4).
Relevance of CD73 biology as an immune checkpoint in human
Focusing on understanding CD73 biology within T cell populations, CLB and LICR partners demonstrated that, in contrast to murine cells, CD39 and CD73 are poorly co-expressed within human T cell subsets. CD39 is selectively expressed on Tregs (suppressive T cells) and activated T cells whereas CD73 is expressed on a discrete CD4+ memory subset (helper T cells) and on all naïve CD8+ T cells (cytotoxic T cells). On CD8+ T cells this expression decrease during the differentiation process (D3.3 D7.1). On CD4+ T cells, CD73 characterizes a memory/effector T cell population (Teffs) of Th1.17 profile producing both IFNγ and IL-17 at high levels (potent effector). CD73 renders this CD4 population highly sensitive to suppression by AMP and by ATP in presence of CD39+ Tregs (D3.3 D3.1 D7.1). Of interest, Ado suppresses most cytokine production by T cells except IL-17 and IL-22. Our current model is that CD73 expression on this potent CD4+ Teffs limits its reactivity in an inflammatory environment rich in ATP and that CD39+ Tregs transform Th1.17 potent anti-tumor effector into “Th17 only” with pro-tumoral activity. Moreover, their expression of functional MDR1, the multidrug resistance protein 1, render them resistant to chemotherapy in vitro (D3.1) suggesting their capacity to resist to chemotherapy regimen in vivo that could contribute to the anti-tumor immune response. Finally, their lack of other inhibitory immune check-points (ICPi, PD1, CTLA-4, TIGIT, TIM3) (D3.3 D7.1) suggest that CD73 play a dominant role as an ICPi able to regulate the function of these unique Teffs.
In summary these results highlight a unique biology of CD73 regulating a potent anti-tumor effector, and suggest CD73 as a potential resistance mechanisms to anti-PD1/CTLA4 treatment.
Figure 3: Autocrine Adenosine regulates tumor polyfunctional CD73+CD4+ effector T cells devoid of immune checkpoints: 1- ATP through inflammasome activation favors Th1.17; 2- Inflammation leads to CCL22 secretion and Treg recruitment ; 3- CD39high activated Tregs will target CD73+ Teffs through Ado production transforming potent Th1.17 anti-tumor effectors into protumoral Th17 T cells
LICR partner also evaluated the role of Ado receptors (ADORAs) in mediating T and NK cells immunosuppression. Overall, the receptor mainly expressed by immune cells on peripheral PBMC is A2A. We show that Ado immunosuppressive effects on human T cells are primarily exerted through A2A and that CD8+ TCM cells are more sensitive than the other CD8+ T cell memory subsets since they express higher A2A levels (D3.5). In addition, TILs expanded from distinct tumor types also preferentially express A2A and have a similar sensitivity to Ado compared to peripheral T cells. Triggering of A2A in T cells increased PKA activation resulting in impaired mTORC1 (but not mTORC2) pathway, cytokine production and metabolic fitness. Our findings unveil the A2A/PKA/mTORC1 pathway as the main axis for Ado to impair T cell function and metabolic fitness (D3.5). Finally, NK cell function was also impaired by Ado. However, both A2A and A2B but also receptor-independent mechanisms may play a role in mediating these effects on this population (D3.8).
In summary these data identify A2a as a dominant receptor mediating Ado suppressive activity on cytotoxic effectors (CD8 T cells and NK) through inhibition of the mTOR pathway involved in metabolic adaptation of the T cells in the unfavorable tumor environment.
In this WP, the role of CD73 on tumor cells was also investigated in various assays by UTU and CLB partners. The most consistent observation relates to a role of CD73 in tumor invasion in 3D culture models (D3.6). Thus targeting CD73 may have additional benefit by preventing tumor cell invasion.
Furthermore, CLB partner also analyzed the regulation of CD73 expression (D3.2) and its implication during the migration of cells associated with tumor invasiveness (epithelial to mesenchymal transition, or EMT process (D3.7). Clear association between EMT and CD73 expression was made on TNBC human tumors and extended in carcinosarcoma of the breast (D3.7). This observation was also confirmed in a plastic mouse breast model undergoing EMT in vivo under immune pressure (D3.7). Together with the IHC observation that CD73 is focally expressed at the invasive front of certain tumors (D2.8) these data suggest a role of CD73 in the invasive properties of CD73+ tumor cells, and that the immunosuppressive production of Ado may contribute in some tumor-types to increase tumor dissemination properties.
Altogether these biological observations strengthen the importance of CD73 biology as an immune checkpoint in human, and validate its enzymatic neutralization as a therapeutic objective.
Development of neutralizing anti-huCD73 mAbs.
The generation and characterization of innovative and potent humanized CD73 neutralizing mAbs has represented a major collective effort. Monoclonal Abs were produced, purified and evaluated in various in vitro efficacy assays (affinity, specificity, inhibition of enzymatic activity, inhibition of Adenosine-mediated effect on immune cells, through degradation of AMP into Adenosine), using different cellular models (D4.1 D4.2). A group of anti-huCD73 Abs initially selected were further characterized (epitope mapping and mechanism of action (MOA)) (D4.3). One lead candidate was selected and has been successfully humanized (D4.4 D4.6).
In the course of the project, two pharmaceutical companies have developed their own blocking anti-CD73 Abs and they have initiated phase I clinical trials, Medimmune in 2015 and BMS in 2016. Based on patent information, the two competitors Abs were produced and compared in vitro to the selected lead anti-huCD73 Ab candidate. The latter has distinct properties and epitope from the ones of the competitors (MedImmune, BMS), and better efficacy in several in vitro assays.
In parallel, surrogate anti-moCD73 blocking Abs were generated for use in syngeneic murine tumor models (an immunocompetent model for immunotherapy assessment) (D4.5 D4.7). In vivo efficacy experiments are underway.
New preclinical models for relevant in vivo therapeutic evaluation.
A major investment has been made by the partners to develop mouse models for in vivo evaluation of anti-CD73 mAbs (Table 1): i) syngeneic mouse tumor models expressing moCD73, ii) syngeneic mouse tumor models deleted for moCD73 (moCD73KO ID8 cell line) (D5.1) iii) transplantable syngeneic mouse tumor models expressing huCD73 (huCD73 transfection in moCD73KO ID8 and B16F10 cells) (D5.2) iv) generation of hCD73 KI mouse strain (D5.3); v) transplantable xenogeneic human tumor cell lines (huCD73WT and huCD73KO) (D5.4) vi) transplantable xenogeneic huCD73 human tumor cell lines in immuno-deficient mice reconstituted with human immune cells (D5.6).
Type of models Tumor cell lines Main interest/objective Limitations
syngeneic mouse tumor models expressing moCD73 B16F10 Melanoma, ID8-agg Ovarian, LLC-1 & SV-2*Lung, 4T1 Mammary, CT26 Colon carcinomas Analyse immune response in immuno competent mice Will not allow evaluation of anti-hCD73 not cross-reacting with mCD73
syngeneic mouse tumor models deleted for moCD73 ID8-agg Ovarian Evaluate the importance of mCD73 expressed on tumor cells
transplantable syngeneic mouse tumor models expressing huCD73 (huCD73 transfection in moCD73KO ID8 and B16F10 cells) B16F10 Melanoma, ID8-agg Ovarian The replacement of mCD73 by hCD73 allows to evaluate the therapeutic anti-hCD73 drug candidate, even in wt mice Issue with immunogenicity of hCD73 in wt mice
generation of hCD73 KI mouse strain Allow to evaluate the targeting of hCD73 expressed on host cells (the tumor cell being hCD73 pos or neg) If ki functional, this is a perfect model to use with hCD73 expressing mouse tumor cell lines
transplantable xenogeneic human tumor cell lines (huCD73WT and huCD73KO) HCT116 & HT29 & Isreco-1 Colorectal, NCI-H292 (huCD73WT and huCD73KO) lung, MDA-MB-231 Breast, MIA PaCa-2 Pancreatic carcinoma Allow to target hCD73 expressed on human xenogenic tumor cell line in immunodeficeint mice The absence of adaptive immune system prevent to evaluate the immunomodulatory effect of hCD73 neutralisation
transplantable xenogeneic huCD73 human tumor cell lines in immuno-deficient mice reconstituted with human immune cells EBV LCL B cell lymphoma cell line CD73+ The model allow to target hCD73 expressed on tumor cells or human immune cells The models are complex, heavy and difficult to standardize, and with a short window of experiments due to to xeno-GVH
Table 1: Mouse models for in vivo evaluation of anti-CD73 mAbs
The TumAdoR consortium had to counteract numerous technical difficulties to assess the efficacy of the lead candidate due to 1) immunogenicity of huCD73 that led to the loss of huCD73 on transfected tumor cells (ID8 and B16 models) (D5.2) and 2) the absence of in vivo validation of the huCD73 Knock-In mouse model generated by Polygene partner (D5.3) that prevents for the moment the evaluation of the in vivo efficacy of the drug candidate. Indeed, for unknown reason the KI construct was expressed at the mRNA level but the human CD73 protein could not be detected in vivo in the KI mice (western blot and Immunofluorescence).
The use by CLB partner of transplantable xenogeneic human tumor cell lines in SCID mice did not demonstrate clear activity of anti-huCD73 blocking Ab, probably due to lack of efficient immune system in these models (D5.4). As therapeutic anti-huCD73 have been developed based on the immune regulatory role of CD73, we focused on models with human immune system. The last series of models developed by UTHSCSA partner, thus, consist of human tumors growing in immunodeficient mice reconstituted with autologous human peripheral blood cells. These models being complex to set-up and not standardized, their use within EU-funded project lifetime has been challenging. For several reasons (early GVHD, unstandardized tumor growth) the experiments performed in the humanized NSG mice implanted with the human tumor models did not allow yet to reach a conclusion regarding the activity of the anti-huCD73 mAb candidates (D5.6).
An alternative plan is ongoing by Polygene Partner to generate huCD73 transgenic (huCD73tg) mice under the control of other promotors to have expression of huCD73 respectively in T/B lymphocytes and in endothelium (cells normally expressing CD73 in human and mice) to evaluate activity of a blocking anti-CD73 Ab. In these mice, tumor cell lines expressing huCD73 will not be rejected as huCD73tg mice will be tolerant to huCD73 Although the transgenic mice are not available at the end of the project, they will be used as soon as possible to investigate in vivo the impact of the anti-huCD73 blocking Ab
Finally, a massive collective effort was done between IPH, LICR and CLB partners to explore, in therapeutic setting, the impact of moCD73 neutralization in multiple purely syngeneic models (Table 2) using several surrogate mAbs. Combination with anti-PD1/L1 mAbs were also performed in several of these models. Experiments are still ongoing.
Cell line Tumor model Injection site Mouse Strain moCD73 expression Immunotherapy response to PD-1/PD-L1 blockade
B16F10 Melanoma Sub-cutaneous C57BL/6 Negative None
ID8-agg Ovarian carcinoma Intra-peritoneal C57BL/6 Positive None
LLC-1 Lung carcinoma Sub-cutaneous C57BL/6 Negative Partial
SV-2* Lung carcinoma Sub-cutaneous C57BL/6 Positive None
4T1 Mammary carcinoma Sub-cutaneous BALB/c High Partial
CT26 Colon carcinoma Sub-cutaneous BALB/c Negative Partial
Table 2 : Different syngeneic tumor models developed to investigate the impact of neutralizing anti CD73 mAbs.
Development of specific assays for toxicity evaluation.
For the validation of the interest of CD73-based therapeutic approach to treat cancer, it is key to verify the absence of safety issues related to CD73 blockade. Assays to evaluate toxicity on blood vessel integrity, vascular leakage, and immune response to test antigens have been established and performed in CD73KO mice by UTU partner. huCD73 expression on frozen normal human tissues using the lead anti-huCD73 mAb has been performed by IPH partner. No safety issues have been observed (Table 3).
Assays CD73KO CD73 small molecule inhibitors in wt mice Anti-huCD73 mAb in wt mice Anti-huCD73 candidate in Cynomolgus Anti-hCD73 in in vitro assay Anti-huCD73 candidate on healthy human tissues
Blood vessel integrity No leakage No leakage ND Ongoing No leakage ND
Immune response to test antigens Normal response Normal response ND ND ND ND
Pathological evaluation No abnormality ND ND Ongoing ND ND
Cross-Reactivity on healthy human tissues ND ND ND ND ND No aberrant cross-reactivity
Table 3: In vitro and in vivo assays to evaluate side effects and off target effects of anti-hCD73.
Due to absence of huCD73 KI mice, the evaluation of the safety/toxicity of the lead candidate has not been performed in mice. As an alternative approach, a novel and potent inhibitor of CD73 enzymatic activity was used in CD73WT mice by UTU partner. These experiments did not reveal any increased vascular leakiness, suggesting that inhibition of the enzymatic activity of CD73 with antibodies should not lead to serious side effects i.e suggesting that the new therapy targeting CD73 will be safe.
IPH partner further evaluated the lead anti-huCD73 mAb candidate antibody on frozen healthy tissue TMA containing 15 types of normal organs each. A membranous/cytoplasmic expression on endothelial cells and lymphocytes in lymphoid organs and tissues, as well as on cells from immune privileged sites was mainly observed, in agreement with the literature on CD73 expression.
As the lead candidate cross-reacted with cynomolgus CD73, a preliminary evaluation of safety was done in cynomolgus. The experimental part of the study has been performed, complete analysis will be available after the project end.
From the current TumAdoR consortium current data no toxicity issues are expected following CD73 enzymatic activity neutralisation in clinic.
In the frame of TumAdoR project a series of imaging tools allowing the longitudinal monitoring of tumor growth in vivo has been developed by the partners (luciferase/bioluminescence/NightOWL by UTHSCSA and CLB partners, Quantum FX Low Dose Computed Tomograph by CLB and LICR partners). These tools will allow in the future to reduce drastically the number of animals used (by 50 to 80% in spontaneous tumor models), through early identification of tumor development and constitution of homogenous experimental group of mice. Furthermore, these imaging tools will allow a longitudinal follow-up of mice under treatment, that would require 10 time many mice in conventional tumor growth evaluation (mice sacrifice at each time point), in particular for deep tumors (Lung and ovarian/pancreas models).
Thus the imaging tools develop during the TumAdoR consortium will allow a major reduction in the number of mice needed for longitudinal follow-up of mice under treatment.
Development of new Immuno-monitoring tools.
Specific tools to assess target expression and to monitor relevant immune parameters in patients during the treatment have been developed by LICR and CLB partners (Table 4). Multi-parametric flow cytometry tools have been setup to analyse impact on differentiation/ polarization of T cells subsets and activation status of B cells, NK cells, DCs, monocytes in combination with CD73 and other biomarkers expression (D7.1).
Parameters Assays Vadidation in heathy donors SOP transfer in the consortium Evalua-tion in patients Compatibility with anti-CD73 drug candidate
Multi-parametric flow cytometry tools for CD73 and CD39 on immune cell subsets and activation status -differentiation/ polarization of T cells
- immune subsets and activation status
- B cells, NK cells, DCs, monocytes done
done
done done
done
done done
done
done Yes
Yes
Yes
TAA-specific immune responses - fluorescent bare-coding MHC/antigen peptide multimers multiparametric
- flow cytometry analyses by intracytoplasmic cytokines staining done
done done
done done
done Yes
Yes
sCD73 in plasma - Elisa
- enzymatic activity of sCD73 done
done done
done done
done Yes
Yes
CD73 expression in situ IHC done done done Yes
CD73 and immune cells in situ Multi-IF ND ND done Yes
Table 4: Immunomonitoring tools to evaluate immune parameters during anti-hCD73 clinical trial.
To measure TAA-specific immune responses, two experimental approaches have been pursued: (i) fluorescent bare-coding MHC/antigen peptide multimers for antigen specific T cell monitoring and (D7.2) (ii) multiparametric flow cytometry analyses by intracytoplasmic cytokines staining following TCR (T cell receptor)-mediated T cell activation. Flow cytometry strategies designed to follow specific T cell responses can be applied to any antigen of choice depending on the tumor targeted. This work was also completed by the development of tools to assess the expression of CD73 during drug treatment in case the drug candidate competes with the commercial antibodies.
An assay to monitor the enzymatic activity of sCD73 in plasma has been setup. Furthermore, a multiplex assay allowing the concomitant analysis of 48 analytes (including cytokine/chemokine/growth factors) have been successfully evaluated in plasma from cancer patients. These potentially will serve as new biomarkers for the patients selected for and/or treated with anti-CD73 antibody therapy in the future.
Overall, the work performed during TumAdoR project allowed to generate comprehensive knowledge of CD73 and CD39 expression in the main peripheral blood immune cell subsets and of sCD73 enzymatic activity in cell-free human serum.
The generated tools used could be determinant for future immunomonitoring in clinical trial and the information generated could represent important normal baseline value for patient comparison and follow up.
Overall, efforts put in this consortium to develop and harmonize immunomonitoring tools and standard operating procedures (flow cytometry panels, samples collection/handling, CD73 IHC protocols and reading grids) could be very relevant to speed up validation of the strategy to apply in clinical samples obtained during the future clinical trial.
Final TumAdoR results
Although we have experienced delay in WP2 and WP5 (technical difficulties, absence of functional huCD73 KI mice), the program has progressed very well with the development and selection of the lead candidate for a neutralizing anti huCD73 mAb (WP4). IPH partner has selected the final lead candidate of humanized anti-huCD73 Ab, and produced large batches of different Abs (WP4), for in vivo efficacy (WP5) and safety evaluation (WP6) by IPH, LICR, CLB and UTHSCSA partners. IPH partner has further characterized the lead candidate, and generated back-up anti-huCD73 Abs. Many novel data on huCD73 biology on CD4+ T cells and A2A functionality on CD8+ T cells and NK cells have been generated, increasing the relevance of CD73 therapeutic targeting (WP3). Current data also indicate lack of safety issues (WP6). Furthermore, standardized tools for IHC (automated protocol and quantification grid) have been developed for the analysis of CD73 target expression on large cohorts of human solid tumors (WP2). CHUV, CLB and UTU partners have generated knowledge on CD73 distribution in human selected tumors and its impact on tumor progression and patient survival allowing a comprehensive knowledge of CD73 expression that will be critical for the future orientation of the clinical development. A mAb has been generated, for use in future IHC companion diagnostic for anti-huCD73 treatment (WP2). A major investment has been made by the partners to develop mouse models for in vivo evaluation of anti-CD73 mAbs, despites numerous technical difficulties (immunogenicity of huCD73 in wt mice, absence of in vivo validation of the huCD73 Knock-In mouse model), the consortium select relevant models to assess CD73 biology in vivo and the impact of the drugs on immune parameters (WP5). Finally, tools for the immune-monitoring of the future clinical trial with the drug candidate are now all in place (WP7).
Overall, with the exception of the in vivo proof of activity, the program has reached most of its objective albeit with an overall 6-months delay.
In conclusion, the TumAdoR project, thanks to the complementary expertise and objectives of the partners, has generated:
- New knowledge on CD73 as an ICP blocker on CD4+ Teffs,
- New knowledge on A2A and A2B expression, function and signaling on CD8+ T cells and NK cells
- New in vitro assays to evaluate drug candidates blocking CD73 pathway
- New preclinical models to assess efficacy of this strategy, ,
- Humanized efficient neutralizing anti-huCD73 mAbs, with higher efficacy in vitro than competitors and differentiated MOA
- Tools for future companion diagnostic IHC to assess CD73 expression
- Comprehensive knowledge of huCD73 expression in human solid tumors that will be critical for the future orientation of the clinical development and first clinical trial
- Immuno-monitoring tools to follow relevant immune parameters in patients under anti-CD73 therapy to assess biological efficacy and investigate on and off targets.
Potential Impact:
Main dissemination activities and exploitation of results
The results obtained in the course of TumAdoR have been presented at several scientific events (14 posters, 8 oral presentations) and described in peer-reviewed publications (to date in 3 published papers and 1 Thesis; 8 papers in preparation including 1 in revision).
Information regarding flow cytometry panels, SOPs for samples collection/handling and screening of huCD73 expression on FFPE-embedded tumors by IHC have been generated and distributed to all partners for harmonization purposes. They will be disseminated to the scientific community with the several peer-review publications arising from TumAdoR project.
For the moment, 4 patents arose from the project and will allow exploitation of the results of the consortium. In addition, one PhD thesis was defended in direct linked with the project, and one more is planned for September 2018. Communication to larger public was also made via project website, press releases and partner’s websites. Finally, member of the TumAdoR consortium contributed to disseminate knowledge on immunotherapy and immune checkpoint role/function and treatment through multiple seminars in institution or meetings and through courses for master and MD students (total audience represent more than 1000 students).
List of Websites:
Coordinator of the TumAdoR project :
Christophe Caux,
Centre Anticancéreux Léon Bérard
28, rue Laennec
F-69373 LYON Cedex 08
Tel: +33 4 78 78 51 39
E-mail: christophe.caux@lyon.unicancer.fr
Main contacts for the other partners of the TumAdoR project :
Innate Pharma: Carine Paturel, Carine.PATUREL@innate-pharma.fr
Turun Yliopisto (University of Turku): Sirpa Jalkanen, sirjal@utu.fi
Université de Lausanne (LICR laboratory): Pedro Romero, Pedro.Romero@hospvd.ch
UT Health Science Center at San Antonio: Tyler Curiel, curielt@uthscsa.edu
PolyGene AG: Doron Schmerling, doron.shmerling@polygene.ch
AYMING: Aurore Niemiec, aniemiec@ayming.com
http://www.tumador.eu
Acknowledgements :
This project has received funding from the European Union’s Seventh Programme for research, technological development and demonstration under grant agreement No 602200.
Based on the broad immuno-suppression and strong tumor promotion (proliferation and dissemination) exerted by adenosine, therapeutic strategies aimed at targeting adenosine pathway are highly promising. However, given involvement of the four known specific adenosine receptors, the use of mAbs neutralizing the enzymatic function of CD73, the sole 5’ ecto-nucleotidase expressed at surface membrane, thus preventing dephosphorylation of AMP into adenosine, represents an attractive, potent and safe strategy for the development of innovative immunotherapy.
The TumAdoR project aims at paving the way to clinical trials of potent neutralizing anti-CD73 mAb candidates in cancer patients. It should validate a novel cancer immunotherapy approach applicable to a wide range of cancers preventing dissemination and relapse through restoration of anti-tumour immune memory.
The program has progressed very well with the development and selection of the lead candidate for a neutralizing anti huCD73 mAb. The final lead candidate anti-huCD73 Ab has been humanized, and produced in large batches for safety evaluation. The lead candidate has been further characterized and compared with the Abs from the competitors, and back-up anti-huCD73 Abs have been generated.
Many novel data on huCD73 biology on CD4+ T cells and A2A functionality on CD8+ T cells and NK cells have been generated, increasing the relevance of CD73 therapeutic targeting. Current data also indicate lack of safety issues.
Furthermore, standardized tools for IHC (automated protocol and quantification grid) have been developed for the analysis of CD73 target expression on large cohorts of human solid tumors. Several partners of the project have generated knowledge on CD73 distribution in human selected tumors and its impact on tumor progression and patient survival allowing a comprehensive huCD73 expression that will be critical for the future orientation of the clinical development. A new mAb has been selected, for use in future IHC companion diagnostic for anti-huCD73 treatment.
A major investment has been made by the partners to develop mouse models for in vivo evaluation of anti-CD73 mAbs, despites technical difficulties the consortium select relevant models to assess CD73 biology in vivo and the impact of the drugs on immune parameters. Finally, tools for the immune-monitoring of the future clinical trial with the drug candidate are now all in place.
The results obtained in the course of TumAdoR have been presented at several scientific events (31 dissemination activities, 1 PhD, 3 publications, 8 presentations, 14 posters, others 5).
Project Context and Objectives:
Cancer is a clinical and societal burden that justify important Research and Development effort to better understand the biology of these disease and their spread, prerequisite for the identification of new therapeutic pathway and potent clinical tools. Among them monoclonal antibodies (mAbs) for targeted cancer therapies has generated in the last decade a lot of enthusiasm and hope for patients and physicians in the frame of more efficient, safer, personalized medicine. The mAbs oncology market is constantly growing, reaching nearly 20 billion Euros in 2012. One of the aims of the TumAdoR project is to boost the EU competitiveness of health-related industries and businesses, as well as address global health issues.
The major challenge for success that has to be addressed in cancer immunotherapy is that tumors fool the host immune system, favoring tumor spread and shortening patient survival. Some tumor-driven immunomodulating mechanisms have been identified: secretion of immunosuppressive cytokines, recruitment of suppressive immune cells, local amino-acid starvation, altered activation of immune receptors and engagement of inhibitory receptors. In addition to being involved in the natural progression of cancer, this immunosuppressive environment affects the efficacy of anticancer treatments. At the beginning of TumAdoR project in 2013, few targets contributing to tumor immune-suppression have been clinically evaluated using neutralizing mAbs and the successes in clinical trials of anti-PD1 and PD-L1 mAbs opened the door to a highly promising novel class of therapeutics targeting tumor-driven immunosuppressive pathways. However, as severe autoimmune syndromes related to disruption of immune tolerance have been reported after treatment with anti-CTLA-4 mAbs, development of strategies targeting other tumor-driven immunosuppressive pathways with increased efficacy and lower toxicity was urgently needed.
Extracellular Ado has multiple activities that contribute to tumor progression (Figure 1). Ado has direct effects on tumor cells contributing to their proliferation and favoring tumor cell motility and metastases (for review Allard 2012). Furthermore Ado has direct effect on endothelial cells and increase angiogenesis and protects cancer cells from hypoxia (Ahmad 2009). In addition to these direct tumor promoting activities, Ado has critical immunomodulatory properties, suppresses both innate and adaptive immune responses and inhibits pro-inflammatory cytokine production. Finally, Ado inhibits immune cell recruitement by inhibiting expression of adhesion molecules on endothelial cells and T cells.
The conversion of ATP into Ado is tightly regulated by cell membrane ectoenzymes of the NTPDase family and tissue non-specific alkaline phosphatases (Figure 2). The conversion of ATP into AMP is predominantly catalyzed by CD39. CD73, a glycosylphosphatidylinositol (GPI)-anchored nucleotidase present in cell membrane lipid rafts, active as a disulfide-linked (for review Yegutkin 2008). Importantly, whereas the activity of CD39 could be reverted by the action of NDP kinase and adenylate kinase, CD73 activity is virtually irreversible thus representing a crucial checkpoint in conversion of pro-inflammatory ATP into immunosuppressive Ado. Indeed, ATP is a danger signal released by damaged and dying cells that primes immune responses through the ligation of P2X and P2Y purino-receptors. By contrast, through the activation of G-protein-coupled seven transmembrane spanning receptors (A2A, A2B), Ado suppresses the immune responses of not only CD4+ and CD8+ T cells, but also of the antigen presenting cells, NK and NKT cells (for review Stagg 2010).
Thus, based on the broad immuno-suppression and strong tumor promotion (proliferation and dissemination) exerted by adenosine (Ado), therapeutic strategies aimed at inhibiting adenosine appeared highly promising. For one research team at the Centre Anticancéreux Léon Bérard (CLB, coordinator of TumAdoR project) the use of mAbs neutralizing the enzymatic function of CD73 receptor expressed at surface membrane, thus preventing the generation of Ado, represented an attractive strategy for the development of innovative immunotherapy. In addition, CD73 neutralization appears to reduce tumorigenesis and metastasis. Combining the expertise of 6 top-level scientific partners in Europe and the US, the TumAdoR project aims at paving the way to clinical trials of potent neutralizing anti-huCD73 mAb candidates in cancer patients. Neutralizing CD73 activity is expected to be efficient in a wide range of cancers, and to have limited adverse side effects.
TumAdoR project should in the coming years validate a novel cancer immunotherapy approach. This novel approach will enable the development of specific anti-tumor immune response inhibiting cancer progression and dissemination and preventing relapse through restoration of anti-tumor immune memory.
Specifically, the objectives of the TumAdoR program were to 1- generate a comprehensive knowledge of CD73 distribution in tissues and major type of tumors, 2- understand CD73 biology as an immune checkpoint in human, 3- develop therapeutic antibodies neutralizing CD73 enzymatic activity, 4- set up new preclinical models for relevant in vivo therapeutic evaluation of anti-CD73 therapeutic candidate, 5- develop specific assays for toxicity evaluation of anti-CD73 therapeutic candidate, 6- set new Immuno-monitoring tools to monitor anti-CD73 mAb clinical trial.
Project Results:
Progress in Target and associated biomarkers expression, and clinical correlation on survival/relapse.
Information on CD73 prevalence in human tumor types and its impact on tumor progression and patient survival was limited, so the consortium undertook the development of specific tools to analyze CD73 expression and markers associated with CD73 biology on well clinically annotated cohorts of selected tumors types.
A commercial antibody was selected that provided specific and unambiguous staining (rabbit mAb D7F9A, Cell Signalling Technology). In addition, a short-list of biomarkers associated with CD73 and its biological function has been generated through bioinformatics analysis. The evaluation, harmonization and automatization of the IHC protocols for CD73 detection as well as the quantification grid have been cross-validated between the partners (D7.5 D2.8). A major investment was performed by CLB, UTU and CHUV partners to analyze CD73 expression on multiple tumor types using either Tissue multi array (TMA) or full sections according to the available tumor material and the size of the cohort to be analyzed. Retrospective tumor collections with clinical annotations have been defined in different partners’ hospital institutions and IHC studies have been done on different tumor types including Melanoma, Lung, Breast, Ovarian, Bladder, Prostate, Head and Neck (H&N), hepatobiliopancreatic system and thyroid carcinoma as well as different subtypes of Lymphoma, all with unmet medical needs.
Overall, we observe variable expression on tumor cells according to pathologies ranging from few percentages of cases (breast tumors, lymphoma) to 100% of cases being positive (some hepatobiliopancreatic tumors). Cancers showing the highest rate of CD73 expression are represented by the hepatocellular and pancreatic ductal malignancies , lung, bladder, H&N carcinoma and melanoma. In most instances, the expression on tumor cells is focal and appears in several pathologies to be associated with the invasive front (lung, some thyroid carcinomas, melanoma) (D2.8). Beside expression on tumor cells, blood vessels express CD73 in most tumors, as well as mesenchymal stromal cells. CD73 expression on immune cells is also observed in most tumors at variable levels from few percentages up to >50% depending on tumor type (D2.10).
Clinical correlations have been performed when the percentage of expression was significant and clinical follow-up sufficient. In certain tumor types CD73 expression on tumor cells is of bad prognosis, while in others CD73 expression is of good prognosis. The expression on stromal cell and immune cells is in some instance associated with good prognosis, but this might be linked to the stromal cells and immune cell types themselves as CD73 expression varies according to different infiltrating cell types.
Thus, based on tumor cell expression and outcome, some tumor types are considered as potential primary tumor targets. Of note CD73 IHC might be of help in the diagnosis of some cases of pancreas and thyroid tumors difficult to interpret by anatomo-pathologists. In parallel, a proprietary anti-huCD73 has been generated and selected for potential future development of an IHC companion diagnostic (D2.9).
Furthermore, UTU partner established a standardized test format for subsequent large-scale analysis to quantify in biological fluids soluble CD73 (sCD73) by ELISA and demonstrated the strong correlation with the enzymatic activity (D2.4). This tool has been used to measure sera concentrations of sCD73 in the FinRisk 2002 cohort (1369 men, 1426 women). Prevalent or incident cancers were diagnosed in 358 individuals (12,8%) but no difference in sCD73 serum concentrations was observed between individuals with and without cancer. However, among patients with prevalent cancer, those who died had statistically higher sCD73 values than those who survived. (D2.4). Based on these results we believe that sCD73 alone may not have additive value in finding the right individuals for clinical trials or to follow up their treatment response. However, the feasibility of combining sCD73 measurements with other parameters have also been investigated and analyses are ongoing (D7.4).
Relevance of CD73 biology as an immune checkpoint in human
Focusing on understanding CD73 biology within T cell populations, CLB and LICR partners demonstrated that, in contrast to murine cells, CD39 and CD73 are poorly co-expressed within human T cell subsets. CD39 is selectively expressed on Tregs (suppressive T cells) and activated T cells whereas CD73 is expressed on a discrete CD4+ memory subset (helper T cells) and on all naïve CD8+ T cells (cytotoxic T cells). On CD8+ T cells this expression decrease during the differentiation process (D3.3 D7.1). On CD4+ T cells, CD73 characterizes a memory/effector T cell population (Teffs) of Th1.17 profile producing both IFNγ and IL-17 at high levels (potent effector). CD73 renders this CD4 population highly sensitive to suppression by AMP and by ATP in presence of CD39+ Tregs (D3.3 D3.1 D7.1). Of interest, Ado suppresses most cytokine production by T cells except IL-17 and IL-22. Our current model is that CD73 expression on this potent CD4+ Teffs limits its reactivity in an inflammatory environment rich in ATP and that CD39+ Tregs transform Th1.17 potent anti-tumor effector into “Th17 only” with pro-tumoral activity. Moreover, their expression of functional MDR1, the multidrug resistance protein 1, render them resistant to chemotherapy in vitro (D3.1) suggesting their capacity to resist to chemotherapy regimen in vivo that could contribute to the anti-tumor immune response. Finally, their lack of other inhibitory immune check-points (ICPi, PD1, CTLA-4, TIGIT, TIM3) (D3.3 D7.1) suggest that CD73 play a dominant role as an ICPi able to regulate the function of these unique Teffs.
In summary these results highlight a unique biology of CD73 regulating a potent anti-tumor effector, and suggest CD73 as a potential resistance mechanisms to anti-PD1/CTLA4 treatment.
Figure 3: Autocrine Adenosine regulates tumor polyfunctional CD73+CD4+ effector T cells devoid of immune checkpoints: 1- ATP through inflammasome activation favors Th1.17; 2- Inflammation leads to CCL22 secretion and Treg recruitment ; 3- CD39high activated Tregs will target CD73+ Teffs through Ado production transforming potent Th1.17 anti-tumor effectors into protumoral Th17 T cells
LICR partner also evaluated the role of Ado receptors (ADORAs) in mediating T and NK cells immunosuppression. Overall, the receptor mainly expressed by immune cells on peripheral PBMC is A2A. We show that Ado immunosuppressive effects on human T cells are primarily exerted through A2A and that CD8+ TCM cells are more sensitive than the other CD8+ T cell memory subsets since they express higher A2A levels (D3.5). In addition, TILs expanded from distinct tumor types also preferentially express A2A and have a similar sensitivity to Ado compared to peripheral T cells. Triggering of A2A in T cells increased PKA activation resulting in impaired mTORC1 (but not mTORC2) pathway, cytokine production and metabolic fitness. Our findings unveil the A2A/PKA/mTORC1 pathway as the main axis for Ado to impair T cell function and metabolic fitness (D3.5). Finally, NK cell function was also impaired by Ado. However, both A2A and A2B but also receptor-independent mechanisms may play a role in mediating these effects on this population (D3.8).
In summary these data identify A2a as a dominant receptor mediating Ado suppressive activity on cytotoxic effectors (CD8 T cells and NK) through inhibition of the mTOR pathway involved in metabolic adaptation of the T cells in the unfavorable tumor environment.
In this WP, the role of CD73 on tumor cells was also investigated in various assays by UTU and CLB partners. The most consistent observation relates to a role of CD73 in tumor invasion in 3D culture models (D3.6). Thus targeting CD73 may have additional benefit by preventing tumor cell invasion.
Furthermore, CLB partner also analyzed the regulation of CD73 expression (D3.2) and its implication during the migration of cells associated with tumor invasiveness (epithelial to mesenchymal transition, or EMT process (D3.7). Clear association between EMT and CD73 expression was made on TNBC human tumors and extended in carcinosarcoma of the breast (D3.7). This observation was also confirmed in a plastic mouse breast model undergoing EMT in vivo under immune pressure (D3.7). Together with the IHC observation that CD73 is focally expressed at the invasive front of certain tumors (D2.8) these data suggest a role of CD73 in the invasive properties of CD73+ tumor cells, and that the immunosuppressive production of Ado may contribute in some tumor-types to increase tumor dissemination properties.
Altogether these biological observations strengthen the importance of CD73 biology as an immune checkpoint in human, and validate its enzymatic neutralization as a therapeutic objective.
Development of neutralizing anti-huCD73 mAbs.
The generation and characterization of innovative and potent humanized CD73 neutralizing mAbs has represented a major collective effort. Monoclonal Abs were produced, purified and evaluated in various in vitro efficacy assays (affinity, specificity, inhibition of enzymatic activity, inhibition of Adenosine-mediated effect on immune cells, through degradation of AMP into Adenosine), using different cellular models (D4.1 D4.2). A group of anti-huCD73 Abs initially selected were further characterized (epitope mapping and mechanism of action (MOA)) (D4.3). One lead candidate was selected and has been successfully humanized (D4.4 D4.6).
In the course of the project, two pharmaceutical companies have developed their own blocking anti-CD73 Abs and they have initiated phase I clinical trials, Medimmune in 2015 and BMS in 2016. Based on patent information, the two competitors Abs were produced and compared in vitro to the selected lead anti-huCD73 Ab candidate. The latter has distinct properties and epitope from the ones of the competitors (MedImmune, BMS), and better efficacy in several in vitro assays.
In parallel, surrogate anti-moCD73 blocking Abs were generated for use in syngeneic murine tumor models (an immunocompetent model for immunotherapy assessment) (D4.5 D4.7). In vivo efficacy experiments are underway.
New preclinical models for relevant in vivo therapeutic evaluation.
A major investment has been made by the partners to develop mouse models for in vivo evaluation of anti-CD73 mAbs (Table 1): i) syngeneic mouse tumor models expressing moCD73, ii) syngeneic mouse tumor models deleted for moCD73 (moCD73KO ID8 cell line) (D5.1) iii) transplantable syngeneic mouse tumor models expressing huCD73 (huCD73 transfection in moCD73KO ID8 and B16F10 cells) (D5.2) iv) generation of hCD73 KI mouse strain (D5.3); v) transplantable xenogeneic human tumor cell lines (huCD73WT and huCD73KO) (D5.4) vi) transplantable xenogeneic huCD73 human tumor cell lines in immuno-deficient mice reconstituted with human immune cells (D5.6).
Type of models Tumor cell lines Main interest/objective Limitations
syngeneic mouse tumor models expressing moCD73 B16F10 Melanoma, ID8-agg Ovarian, LLC-1 & SV-2*Lung, 4T1 Mammary, CT26 Colon carcinomas Analyse immune response in immuno competent mice Will not allow evaluation of anti-hCD73 not cross-reacting with mCD73
syngeneic mouse tumor models deleted for moCD73 ID8-agg Ovarian Evaluate the importance of mCD73 expressed on tumor cells
transplantable syngeneic mouse tumor models expressing huCD73 (huCD73 transfection in moCD73KO ID8 and B16F10 cells) B16F10 Melanoma, ID8-agg Ovarian The replacement of mCD73 by hCD73 allows to evaluate the therapeutic anti-hCD73 drug candidate, even in wt mice Issue with immunogenicity of hCD73 in wt mice
generation of hCD73 KI mouse strain Allow to evaluate the targeting of hCD73 expressed on host cells (the tumor cell being hCD73 pos or neg) If ki functional, this is a perfect model to use with hCD73 expressing mouse tumor cell lines
transplantable xenogeneic human tumor cell lines (huCD73WT and huCD73KO) HCT116 & HT29 & Isreco-1 Colorectal, NCI-H292 (huCD73WT and huCD73KO) lung, MDA-MB-231 Breast, MIA PaCa-2 Pancreatic carcinoma Allow to target hCD73 expressed on human xenogenic tumor cell line in immunodeficeint mice The absence of adaptive immune system prevent to evaluate the immunomodulatory effect of hCD73 neutralisation
transplantable xenogeneic huCD73 human tumor cell lines in immuno-deficient mice reconstituted with human immune cells EBV LCL B cell lymphoma cell line CD73+ The model allow to target hCD73 expressed on tumor cells or human immune cells The models are complex, heavy and difficult to standardize, and with a short window of experiments due to to xeno-GVH
Table 1: Mouse models for in vivo evaluation of anti-CD73 mAbs
The TumAdoR consortium had to counteract numerous technical difficulties to assess the efficacy of the lead candidate due to 1) immunogenicity of huCD73 that led to the loss of huCD73 on transfected tumor cells (ID8 and B16 models) (D5.2) and 2) the absence of in vivo validation of the huCD73 Knock-In mouse model generated by Polygene partner (D5.3) that prevents for the moment the evaluation of the in vivo efficacy of the drug candidate. Indeed, for unknown reason the KI construct was expressed at the mRNA level but the human CD73 protein could not be detected in vivo in the KI mice (western blot and Immunofluorescence).
The use by CLB partner of transplantable xenogeneic human tumor cell lines in SCID mice did not demonstrate clear activity of anti-huCD73 blocking Ab, probably due to lack of efficient immune system in these models (D5.4). As therapeutic anti-huCD73 have been developed based on the immune regulatory role of CD73, we focused on models with human immune system. The last series of models developed by UTHSCSA partner, thus, consist of human tumors growing in immunodeficient mice reconstituted with autologous human peripheral blood cells. These models being complex to set-up and not standardized, their use within EU-funded project lifetime has been challenging. For several reasons (early GVHD, unstandardized tumor growth) the experiments performed in the humanized NSG mice implanted with the human tumor models did not allow yet to reach a conclusion regarding the activity of the anti-huCD73 mAb candidates (D5.6).
An alternative plan is ongoing by Polygene Partner to generate huCD73 transgenic (huCD73tg) mice under the control of other promotors to have expression of huCD73 respectively in T/B lymphocytes and in endothelium (cells normally expressing CD73 in human and mice) to evaluate activity of a blocking anti-CD73 Ab. In these mice, tumor cell lines expressing huCD73 will not be rejected as huCD73tg mice will be tolerant to huCD73 Although the transgenic mice are not available at the end of the project, they will be used as soon as possible to investigate in vivo the impact of the anti-huCD73 blocking Ab
Finally, a massive collective effort was done between IPH, LICR and CLB partners to explore, in therapeutic setting, the impact of moCD73 neutralization in multiple purely syngeneic models (Table 2) using several surrogate mAbs. Combination with anti-PD1/L1 mAbs were also performed in several of these models. Experiments are still ongoing.
Cell line Tumor model Injection site Mouse Strain moCD73 expression Immunotherapy response to PD-1/PD-L1 blockade
B16F10 Melanoma Sub-cutaneous C57BL/6 Negative None
ID8-agg Ovarian carcinoma Intra-peritoneal C57BL/6 Positive None
LLC-1 Lung carcinoma Sub-cutaneous C57BL/6 Negative Partial
SV-2* Lung carcinoma Sub-cutaneous C57BL/6 Positive None
4T1 Mammary carcinoma Sub-cutaneous BALB/c High Partial
CT26 Colon carcinoma Sub-cutaneous BALB/c Negative Partial
Table 2 : Different syngeneic tumor models developed to investigate the impact of neutralizing anti CD73 mAbs.
Development of specific assays for toxicity evaluation.
For the validation of the interest of CD73-based therapeutic approach to treat cancer, it is key to verify the absence of safety issues related to CD73 blockade. Assays to evaluate toxicity on blood vessel integrity, vascular leakage, and immune response to test antigens have been established and performed in CD73KO mice by UTU partner. huCD73 expression on frozen normal human tissues using the lead anti-huCD73 mAb has been performed by IPH partner. No safety issues have been observed (Table 3).
Assays CD73KO CD73 small molecule inhibitors in wt mice Anti-huCD73 mAb in wt mice Anti-huCD73 candidate in Cynomolgus Anti-hCD73 in in vitro assay Anti-huCD73 candidate on healthy human tissues
Blood vessel integrity No leakage No leakage ND Ongoing No leakage ND
Immune response to test antigens Normal response Normal response ND ND ND ND
Pathological evaluation No abnormality ND ND Ongoing ND ND
Cross-Reactivity on healthy human tissues ND ND ND ND ND No aberrant cross-reactivity
Table 3: In vitro and in vivo assays to evaluate side effects and off target effects of anti-hCD73.
Due to absence of huCD73 KI mice, the evaluation of the safety/toxicity of the lead candidate has not been performed in mice. As an alternative approach, a novel and potent inhibitor of CD73 enzymatic activity was used in CD73WT mice by UTU partner. These experiments did not reveal any increased vascular leakiness, suggesting that inhibition of the enzymatic activity of CD73 with antibodies should not lead to serious side effects i.e suggesting that the new therapy targeting CD73 will be safe.
IPH partner further evaluated the lead anti-huCD73 mAb candidate antibody on frozen healthy tissue TMA containing 15 types of normal organs each. A membranous/cytoplasmic expression on endothelial cells and lymphocytes in lymphoid organs and tissues, as well as on cells from immune privileged sites was mainly observed, in agreement with the literature on CD73 expression.
As the lead candidate cross-reacted with cynomolgus CD73, a preliminary evaluation of safety was done in cynomolgus. The experimental part of the study has been performed, complete analysis will be available after the project end.
From the current TumAdoR consortium current data no toxicity issues are expected following CD73 enzymatic activity neutralisation in clinic.
In the frame of TumAdoR project a series of imaging tools allowing the longitudinal monitoring of tumor growth in vivo has been developed by the partners (luciferase/bioluminescence/NightOWL by UTHSCSA and CLB partners, Quantum FX Low Dose Computed Tomograph by CLB and LICR partners). These tools will allow in the future to reduce drastically the number of animals used (by 50 to 80% in spontaneous tumor models), through early identification of tumor development and constitution of homogenous experimental group of mice. Furthermore, these imaging tools will allow a longitudinal follow-up of mice under treatment, that would require 10 time many mice in conventional tumor growth evaluation (mice sacrifice at each time point), in particular for deep tumors (Lung and ovarian/pancreas models).
Thus the imaging tools develop during the TumAdoR consortium will allow a major reduction in the number of mice needed for longitudinal follow-up of mice under treatment.
Development of new Immuno-monitoring tools.
Specific tools to assess target expression and to monitor relevant immune parameters in patients during the treatment have been developed by LICR and CLB partners (Table 4). Multi-parametric flow cytometry tools have been setup to analyse impact on differentiation/ polarization of T cells subsets and activation status of B cells, NK cells, DCs, monocytes in combination with CD73 and other biomarkers expression (D7.1).
Parameters Assays Vadidation in heathy donors SOP transfer in the consortium Evalua-tion in patients Compatibility with anti-CD73 drug candidate
Multi-parametric flow cytometry tools for CD73 and CD39 on immune cell subsets and activation status -differentiation/ polarization of T cells
- immune subsets and activation status
- B cells, NK cells, DCs, monocytes done
done
done done
done
done done
done
done Yes
Yes
Yes
TAA-specific immune responses - fluorescent bare-coding MHC/antigen peptide multimers multiparametric
- flow cytometry analyses by intracytoplasmic cytokines staining done
done done
done done
done Yes
Yes
sCD73 in plasma - Elisa
- enzymatic activity of sCD73 done
done done
done done
done Yes
Yes
CD73 expression in situ IHC done done done Yes
CD73 and immune cells in situ Multi-IF ND ND done Yes
Table 4: Immunomonitoring tools to evaluate immune parameters during anti-hCD73 clinical trial.
To measure TAA-specific immune responses, two experimental approaches have been pursued: (i) fluorescent bare-coding MHC/antigen peptide multimers for antigen specific T cell monitoring and (D7.2) (ii) multiparametric flow cytometry analyses by intracytoplasmic cytokines staining following TCR (T cell receptor)-mediated T cell activation. Flow cytometry strategies designed to follow specific T cell responses can be applied to any antigen of choice depending on the tumor targeted. This work was also completed by the development of tools to assess the expression of CD73 during drug treatment in case the drug candidate competes with the commercial antibodies.
An assay to monitor the enzymatic activity of sCD73 in plasma has been setup. Furthermore, a multiplex assay allowing the concomitant analysis of 48 analytes (including cytokine/chemokine/growth factors) have been successfully evaluated in plasma from cancer patients. These potentially will serve as new biomarkers for the patients selected for and/or treated with anti-CD73 antibody therapy in the future.
Overall, the work performed during TumAdoR project allowed to generate comprehensive knowledge of CD73 and CD39 expression in the main peripheral blood immune cell subsets and of sCD73 enzymatic activity in cell-free human serum.
The generated tools used could be determinant for future immunomonitoring in clinical trial and the information generated could represent important normal baseline value for patient comparison and follow up.
Overall, efforts put in this consortium to develop and harmonize immunomonitoring tools and standard operating procedures (flow cytometry panels, samples collection/handling, CD73 IHC protocols and reading grids) could be very relevant to speed up validation of the strategy to apply in clinical samples obtained during the future clinical trial.
Final TumAdoR results
Although we have experienced delay in WP2 and WP5 (technical difficulties, absence of functional huCD73 KI mice), the program has progressed very well with the development and selection of the lead candidate for a neutralizing anti huCD73 mAb (WP4). IPH partner has selected the final lead candidate of humanized anti-huCD73 Ab, and produced large batches of different Abs (WP4), for in vivo efficacy (WP5) and safety evaluation (WP6) by IPH, LICR, CLB and UTHSCSA partners. IPH partner has further characterized the lead candidate, and generated back-up anti-huCD73 Abs. Many novel data on huCD73 biology on CD4+ T cells and A2A functionality on CD8+ T cells and NK cells have been generated, increasing the relevance of CD73 therapeutic targeting (WP3). Current data also indicate lack of safety issues (WP6). Furthermore, standardized tools for IHC (automated protocol and quantification grid) have been developed for the analysis of CD73 target expression on large cohorts of human solid tumors (WP2). CHUV, CLB and UTU partners have generated knowledge on CD73 distribution in human selected tumors and its impact on tumor progression and patient survival allowing a comprehensive knowledge of CD73 expression that will be critical for the future orientation of the clinical development. A mAb has been generated, for use in future IHC companion diagnostic for anti-huCD73 treatment (WP2). A major investment has been made by the partners to develop mouse models for in vivo evaluation of anti-CD73 mAbs, despites numerous technical difficulties (immunogenicity of huCD73 in wt mice, absence of in vivo validation of the huCD73 Knock-In mouse model), the consortium select relevant models to assess CD73 biology in vivo and the impact of the drugs on immune parameters (WP5). Finally, tools for the immune-monitoring of the future clinical trial with the drug candidate are now all in place (WP7).
Overall, with the exception of the in vivo proof of activity, the program has reached most of its objective albeit with an overall 6-months delay.
In conclusion, the TumAdoR project, thanks to the complementary expertise and objectives of the partners, has generated:
- New knowledge on CD73 as an ICP blocker on CD4+ Teffs,
- New knowledge on A2A and A2B expression, function and signaling on CD8+ T cells and NK cells
- New in vitro assays to evaluate drug candidates blocking CD73 pathway
- New preclinical models to assess efficacy of this strategy, ,
- Humanized efficient neutralizing anti-huCD73 mAbs, with higher efficacy in vitro than competitors and differentiated MOA
- Tools for future companion diagnostic IHC to assess CD73 expression
- Comprehensive knowledge of huCD73 expression in human solid tumors that will be critical for the future orientation of the clinical development and first clinical trial
- Immuno-monitoring tools to follow relevant immune parameters in patients under anti-CD73 therapy to assess biological efficacy and investigate on and off targets.
Potential Impact:
Main dissemination activities and exploitation of results
The results obtained in the course of TumAdoR have been presented at several scientific events (14 posters, 8 oral presentations) and described in peer-reviewed publications (to date in 3 published papers and 1 Thesis; 8 papers in preparation including 1 in revision).
Information regarding flow cytometry panels, SOPs for samples collection/handling and screening of huCD73 expression on FFPE-embedded tumors by IHC have been generated and distributed to all partners for harmonization purposes. They will be disseminated to the scientific community with the several peer-review publications arising from TumAdoR project.
For the moment, 4 patents arose from the project and will allow exploitation of the results of the consortium. In addition, one PhD thesis was defended in direct linked with the project, and one more is planned for September 2018. Communication to larger public was also made via project website, press releases and partner’s websites. Finally, member of the TumAdoR consortium contributed to disseminate knowledge on immunotherapy and immune checkpoint role/function and treatment through multiple seminars in institution or meetings and through courses for master and MD students (total audience represent more than 1000 students).
List of Websites:
Coordinator of the TumAdoR project :
Christophe Caux,
Centre Anticancéreux Léon Bérard
28, rue Laennec
F-69373 LYON Cedex 08
Tel: +33 4 78 78 51 39
E-mail: christophe.caux@lyon.unicancer.fr
Main contacts for the other partners of the TumAdoR project :
Innate Pharma: Carine Paturel, Carine.PATUREL@innate-pharma.fr
Turun Yliopisto (University of Turku): Sirpa Jalkanen, sirjal@utu.fi
Université de Lausanne (LICR laboratory): Pedro Romero, Pedro.Romero@hospvd.ch
UT Health Science Center at San Antonio: Tyler Curiel, curielt@uthscsa.edu
PolyGene AG: Doron Schmerling, doron.shmerling@polygene.ch
AYMING: Aurore Niemiec, aniemiec@ayming.com
http://www.tumador.eu
Acknowledgements :
This project has received funding from the European Union’s Seventh Programme for research, technological development and demonstration under grant agreement No 602200.
