Oral Presentation ANZBMS-MEPSA-ANZORS 2022

Making personalization of cancer treatment a reality with Spatial transcriptomics (#99)

Jazmina L Gonzalez-Cruz 1 , Andrew Causer 2 , Debottam Sinha 1 , Min Teoh 2 , Jerry Tay 1 , Rahul Ladwa 3 4 , Christopher Perry 4 5 , Ian Frazer 1 , Benedict Panizza 4 5 , Quan Nguyen 2
  1. The University of Queensland Diamantina Institute, Brisbane, QLD, Australia
  2. Institute of Molecular Biology, The University of Queensland, Brisbane, QLD, Australia
  3. Department of Medical Oncology, Princess Alexandra Hospital, Brisbane, QLD, Australia
  4. Faculty of Medicine, , The University of Queensland, Brisbane, QLD, Australia
  5. Department of Otolaryngology, Princess Alexandra Hospital, Brisbane, QLD, Australia

  Human Papillomavirus (HPV) is responsible of >70% of all Oropharyngeal Squamous cell carcinoma (OPSCC). Immune checkpoint inhibitors (ICI) are available to treat recurrent metastatic OPSCC patients. Unfortunately, only a <20% of OPSCC patients benefits from this approach.  

 We postulate that the tumour microenvironment plays a critical role in determining ICI therapy outcome. Therefore, we performed spatial transcriptomics (ST) on tumour tissue isolated from a patient diagnosed with metastatic HPV+ OPSCC.

    The patient’s primary OPSCC tumour responded to chemo-radio therapy, followed by Nivolumab (anti-PD-1) therapy to target lung metastasis. After Nivolumab, new OPSCC soft pallet tumours resurged, and Pembrolizumab (a-PD-1) and Lenvatinib (VEGFR inhibitor) were given. This approach initially regressed the secondary OPSCC. However, Lenvatinib’s dose was reduced, and OPSCCs recurred.

  We performed ST on the OPSCC soft palate tumours and healthy tissue, after Nivolumab treatment. We observed unbiased clustering, based on differentially expressed genes (DEG) of ST array spots, that recapitulated the H&E tissue annotations (i.e., epithelium vs muscle). Interestingly, only one cluster, cluster 6, was identified as Oropharyngeal cancer cells, with most of the cells in S and G2M cell cycle phases, indicating that this cluster was the proliferative tumour. The profile of upregulated genes in cluster 6 (49 genes, log2>1.5, adj.p-value <0.001) included oncogenes (18.4%); poor prognosis biomarkers (24.5%); drug resistance genes (20.40%); known targets (18.4%)  including EGFR and VEGFA; and experimentally-druggable-targets (20.4%). We didn't find PD-1 or PD-L1 expression in the tumour, suggesting that Lenvatinib was mainly responsible for the OPSCC regression, likely due to the high VEGFA levels in cluster 6.

  ST data accurately predicted disease evolution in this patient and could be used to generate a list of clinically relevant therapeutic candidates. We foresee that this technology will allow oncologist to accurately personalize cancer management, and open new approaches to drug discovery.