Abstract 4956: Discovering novel targetable pathways by combining functional and multi-omic data from primary ovarian cancer samples

Background: There is a critical necessity to reveal novel and tractable targets for anti-cancer treatments in indications with high unmet medical need, such as high grade serous ovarian cancer (HGSOC). However, standard process for target discovery using models such as outgrown cell lines and well-averaged readouts has yielded a less than 5% approval rate for drugs entering trials (Thomas et al. 2016 Bio.org).

Here, we describe patient-centric target discovery through the use of disease relevant primary OC samples and single cell functional characterization using a platform with proven hemonc translatability (Kornauth et al. 2021, Snijder et al. 2017). We integrate data from our functional drug testing platform under multiple drug perturbations with matching genomic and transcriptomic data to reveal associations with novel downstream regulators of sensitivity.

Methods: Sensitivity of the cancer cell compartment in primary malignant ascites samples (n = 20; 75% HGSOC) to 85 small molecule drugs, was evaluated using a proprietary and translatable deep learning-driven single cell imaging platform (Vladimer et al. 2017). Cancer cell sensitivity from the drugs was combined with WES, bulk-RNAseq and drug induced changes in phosphoproteome, and single cell RNAseq transcriptome to identify perturbed targets and pathways.

Results: Here we describe a family of TKIs including ALKi that induce cytotoxicity of cancer cells in primary samples, not previously captured in publicly available cell line drug sensitivity screening data (Iorio et al. 2016). We report novel sensitivity of OC driven by non-canonical targets of ceritinib such as FAK1 or IGF1R, mediated by the downstream signaling hub YBX1 (Kuenzi et al. 2017), involved in NFB pathway regulation (Motolani et al. 2021). Indeed, transcriptomic scRNA analysis upon ceritinib treatment of primary OC cells revealed rapid perturbation of numerous NFB pathway members, alongside YBX1 inactivation.

Conclusions: Combining functional endpoints and single cell-based differential expression analysis of primary OC samples, we have identified the NFB pathway and the regulator YBX1 as a promising novel sensitivity for HGSOC treatment development. These and several other important targetable nodes identified, sit outside the recently suggested JAK/STAT pathway (Izar et al. 2020), thereby demonstrating a pipeline towards novel drug target and pathway discovery driven by patient-centric, disease relevant models of high-need indications.

Citation Format: Irene Gutierrez-Perez, Bekir Ergüner, Pisanu Buphamalai, Joost Van Ham, Paul Heinz, Valentin Aranha, Rin Okumura, Elisabeth Waltenberger, Isabella Alt, Claudia Baumgaertler, Maja Stulic, Edgar Petru, Christoph Minichsdorfer, Judith Lafleur, Lukas Hefler, Laudia Hadjari, Lucia Dzurillova, Jozef Sufliarsky, Nikolaus Krall, Thorsten Füreder, Gregory Ian Vladimer, Bojan Vilagos, Robert Sehlke. Discovering novel targetable pathways by combining functional and multi-omic data from primary ovarian cancer samples. [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2023; Part 1 (Regular and Invited Abstracts); 2023 Apr 14-19; Orlando, FL. Philadelphia (PA): AACR; Cancer Res 2023;83(7_Suppl):Abstract nr 4956.

Abstract 1893: Deep learning supported high content analysis of primary patient samples identifies ALK inhibition as a novel mechanism of action in a subset of ovarian cancers

Background: High unmet need of ovarian cancer (OC) suggests the discovery of new targeted therapeutics is crucial to improve patient prognosis. Unlike artificial model systems such as cell lines, primary cancer samples recapitulate the complexity of the original microenvironment consisting of cancer cells as well as stromal and immune cells; this is especially important when evaluating IO targets and signalling pathways. Supported by our previous success predicting therapy for late stage haematological cancer patients in the EXALT-I trial using AI-supported functional single cell quantification of drug action (Kornauth et. al. 2021) we set out to systematically reveal novel targets and pathways in OC using small molecule drugs (SMDs) as tools.

Single cell phenotypic screening of OC MPAs (malignant pleural effusion and ascites) was enabled by the quantification of drug effects using an end-to-end scalable deep learning driven image analysis tool chain. This custom state-of-the-art AI software is critical to enable robust primary cell screening given the diversity of cells within each sample. This revealed anaplastic lymphoma kinase (ALK), as well as structurally related targets and pathway associated proteins, as being potential novel targets in a subset of OC patient samples. There is sparse literature evidence for therapeutic utilisation of the ALK pathway in OC, and the diversity of responses indicates a further novel patient selection method.

Methods: MPAs from OC patients (n = 20) were collected and the sensitivity of the cancer cells to 85 SMDs was evaluated using high content microscopy. Individual cells were segmented and classified using convolutional neural networks and drug responses were estimated from the resulting cell counts. The integration of these results with whole exome and RNA sequencing guided target and pathway prioritisation.

Results: Screening for novel sensitivities using SMDs as tools uncovered inhibitors of ALK and related targets as having strong cancer cell cytotoxic effects, recapitulated in solid tumour biopsies. Transcriptomic profiling revealed pathway correlations to ALK inhibitor sensitivity, however non-annotated polypharmacological effects of each drug cannot yet be excluded.

Conclusions: Quantifying SMD sensitivity in a disease relevant model system identified ALK as a promising and overlooked target in OC, providing an upstream and potentially more specific target to the recently suggested MEK, PI3K and STAT3 (Papp et. al. 2018, Izar et al. 2020). While further work to confirm the target is required, this study supports a notion of patient-centric drug development using disease relevant models and deep learning. Our work introduces a novel patient-centric tool to advance understanding of the OC target landscape and provides a resource for the development of novel therapeutic approaches.

Citation Format: Irene Gutierrez-Perez, Joost Van Ham, Valentin Aranha, Rin Okumura, Elisabeth Waltenberger, Isabella Alt, Claudia Baumgaertler, Maja Stulic, Edgar Petru, Christoph Minichsdorfer, Lukas Hefler, Judith Lafleur, Nikolaus Krall, Thorsten Füreder, Gregory Ian Vladimer, Robert Sehlke, Bojan Vilagos. Deep learning supported high content analysis of primary patient samples identifies ALK inhibition as a novel mechanism of action in a subset of ovarian cancers [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2022; 2022 Apr 8-13. Philadelphia (PA): AACR; Cancer Res 2022;82(12_Suppl):Abstract nr 1893.

Abstract 4150: Enriching for adenosine antagonist patient responses through deep learning

Background: Adenosine signaling is a key metabolic pathway regulating tumor immunity. The conversion of inflammatory extracellular ATP into immunosuppressive adenosine invokes signaling of the A2a receptor (A2AR) in the tumor microenvironment. This dampens immune responses and creates a pro-tumor niche. A number of novel IO drugs targeting the adenosine pathway through inhibition of the ectonucleotidases CD39 and CD73 or the A2AR/BR are in clinical trials, including our non-CNS penetrant A2AR-selective antagonist, EXS21546 (NCT04727138). While early clinical results from other adenosine receptor-inhibitors (A2ARi) have shown modest monotherapy activity in nonspecific patient populations, we believe greater success can be achieved by leveraging methods that enable the evaluation of single-cell effects in patient samples preclinically. Using our deep learning driven image analysis platform, we define an adenosine-induced, tumor protective immunosuppression biomarker to augment A2AR antagonist responder identification.

Here we describe efforts to transcriptionally and functionally map the adenosine suppressed immune potential and activation by inhibition of A2AR with EXS21546 in primary material. The goal is to reveal gene signatures indicative of adenosine immunosuppression deployable in clinical studies, increasing the likelihood of trial success by identifying patients that have the highest efficiency potential for A2AR targeted therapy.

Methods: Leveraging patient material as disease relevant model systems, and collecting baseline and treatment condition transcriptomics, we model the patient specific anti-cancer immune repertoire and validate patient selection methods functionally with a translatable high content imaging platform amenable to primary human material supported by end-to-end deep learning driven image analysis.

Results: Combining single cell transcriptomic and functional response data, we demonstrate preclinical mechanistic studies of A2AR antagonism on infiltrating immune cells, with the ultimate aim of discovering predictive algorithms to enrich patients more likely respond to adenosine pathway inhibitors. Patient selection gene signatures are functionally validated using a high content imaging platform with proven translational capabilities (Kornauth et al, Cancer Disc., 2021), demonstrating the association of anti-cancer immune activity with inhibition of adenosine signaling by EXS21546. Signatures and patient selection algorithms are cross-validated with publicly available data.

Discussion: Gathering multiple layers of data from primary tumor tissues, we reveal and map the association of immune response potential to A2AR inhibition in cancer.

Patient stratification gene signatures identified have the aim to be implemented in future studies of our candidate A2ARi EXS21546, to deliver the right drug at the right time in the right patients.

Citation Format: Isabella Alt, Robert Shelke, Anna Lobley, Claudia Baumgaertler, Maja Stulic, Pierre Fons, Mark Whittaker, Klaus Hackner, Lucia Dzurillova, Edga Petru, Laudia Hadjari, Judith Lafleur, Josef Singer, Nikolaus Krall, Lukas Hefler, Thorsten Füreder, Christina Taubert, Christophe Boudesco, Andrew Payne, Gregory Ian Vladimer. Enriching for adenosine antagonist patient responses through deep learning [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2022; 2022 Apr 8-13. Philadelphia (PA): AACR; Cancer Res 2022;82(12_Suppl):Abstract nr 4150.

Clinical scores for the prediction of esophageal varices in patients with liver cirrhosis

INTRODUCTION

Esophageal bleeding is one of the most important and dramatic complications of liver cirrhosis in our everyday practice. Considering the costs of repeated upper endoscopy (UE) there is an increasing number of studies focusing on noninvasive para-meters for the assessment of esophageal varices (EV).

PATIENTS AND METHODS

Retrospective study included 74 patients with alcoholic and viral liver cirrhosis treated at Clinic of Gastroenterology and Hepatology, Clinical Center of Serbia. The data were obtained from patients medical records including history, biochemical, ultrasonography and UE findings.

RESULTS

The average value of the RLLD/INR for patients who showed evidence of EV during UE and in those who didn't was 10.46 ± 3.09 and 12.24 ± 3.43, respectively (p = 0.019, p < 0.05). Cutoff value (11.5) of RLLD/INR showed a sensitivity of 64.15% and specificity of 66.67% (1.92LR+, and 0.54 LR-, AUROC 0.639) for the detection of EV. The average value of PC/SBD for patients who showed evidence of EV during UE and in those who didn't was 619.79 ± 492.96 and 1423.1 ± 908.2, respectively (p = 0.0, p < 0.05). The average value of RLLD/SA was 5.5 ± 0.17 and 4.57 ± 0.17 (p = 0.015, p < 0.05) for patients who showed evidence of EV -during UE and in those who didn't, respectively.

CONCLUSION

Noninvasive assessment of EV using scores based on ultrasonography and laboratory is simple, inexpensive, and could be a useful tool in limiting the number of repeated UE.