Background Immune checkpoint inhibitors, such as anti-PD-1, have transformed cancer therapy in the last decade. Yet, questions persist regarding their efficacy and pose ongoing challenges. Identifying more robust biomarkers is crucial in immuno-oncology, and the iterative process between in silico discovery and experimental studies emphasizes the discovery of the most suitable therapies for patients. Our aim was to pinpoint druggable, predictive biomarkers of resistance to immune checkpoint inhibitors by leveraging publicly available transcriptomic and clinical data.

Methods We set up an online analysis platform (https://rocplot.com/immune) with 1,434 samples from distinct tumor types^{1 2} and selected best candidate genes overexpressed in non-responding patients for further investigations using in vitro and in vivo techniques.

Results Yes-Associated Protein 1 (YAP1) was the top druggable target overexpressed in non-responding patients (ROC AUC=0.699, FC=1.8, Mann-Whitney P=1.1E-08, P_ROC=7.5E-11) in the anti-PD-1 melanoma pre-treatment cohort (n=415). Patients with higher YAP1 expression experienced worse progression-free survival (HR=2.51, P=1.2E-06, log-rank test) and overall survival (HR=2.15, P=1.2E-05, log-rank test). We hypothesised that YAP1 inhibition with the FDA-approved drug Verteporfin (VP) could potentiate the effects of anti-PD-1 in melanoma. We found that VP significantly inhibited melanoma cell viability after 24 hours at 5 μM (P<0.0001) and 10 μM (P=0.0147) concentrations, while longer incubation periods (48 hours) showed efficacy with 0.1 μM (P=0.0014) and 1 μM (P=0.0004) concentrations. Next, we injected C57BL/6JRj mice with immunologically ‘cold’ B16-F10 and YUMM1.7 melanoma cell lines and treated them with VP, anti-PD1, the combination of both, or an IgG2a isotype control. In comparison to anti-PD-1 monotherapy (P=0.008), or control (P=0.021) groups, the combination of VP with anti-PD-1 demonstrated greater efficiency in reducing tumor size. We found increased macrophage and T cell infiltration in the combination therapy group based on elevated CD80, CD86, CD68, and CD45 expression levels. Neither Verteporfin alone nor anti-PD-1 alone exhibited significant inhibition of tumor growth compared to the control (P=0.425 and P=0.971, respectively). No notable difference was observed among the cohorts in B16-F10 mice.

Conclusions To sum up, a database was set up for the discovery of immunotherapy biomarkers, and a druggable predictive biomarker (YAP1) was chosen for in vitro, and in vivo validation. Verteporfin, an inhibitor of YAP1, exhibited higher efficacy in treating melanoma in mice, either as a standalone treatment or in conjunction with anti-PD-1, outperforming both anti-PD-1 monotherapy and the control group.

References

1. Kovács SA, Győrffy B. Transcriptomic datasets of cancer patients treated with immune-checkpoint inhibitors: a systematic review. J Transl Med. 2022;20(1):249.

2. Kovács SA, Fekete JT, Győrffy B. Predictive biomarkers of immunotherapy response with pharmacological applications in solid tumors. Acta Pharmacol Sin. 2023;44(9):1879–1889.

Ethics Approval The study was approved by the National Scientific Ethical Committee on Animal Experimentation in Hungary, approval number PE/EA/010176/2022.