Best Of Best Poster Oral Clinical Oncology Society of Australia Annual Scientific Meeting 2018

A bioinformatics approach to biomarker validation using publically available RNAseq data from immunotherapy trials (#304)

Wee Loong Chin 1 2 3 , Rachael Zemek 2 3 , Richard A Lake 2 3 , Anna K Nowak 1 2 3 , Timo Lassmann 4 , Willem Lesterhuis 2 3
  1. Sir Charles Gairdner Hospital, Perth, WA, Australia
  2. School of Medicine, University of Western Australia, Perth, WA, Australia
  3. National Centre for Asbestos Related Diseases, Perth, WA, Australia
  4. Telethon Kids Institute, Perth, WA, Australia

 

Category: Basic and Translational Research

 

1.    AIMS

Better predictive biomarkers of response to immunotherapy are urgently needed. Recently, we showed in two mouse cancer models that higher STAT1 activation is an important biomarker of immunotherapy response (Zemek R. et al., manuscript submitted). We hypothesised that STAT1 would also be a predictive biomarker in patients treated with immunotherapy. We therefore used a publically available RNAseq dataset from patients treated with immunotherapy to validate our findings in humans.

2.    METHODS

We used gene count data from the IMvigor 210 dataset comprising RNAseq data from pre-treatment tumours of 348 bladder cancer patients treated with Atezolizumab. Correlative clinical data (response, survival) was available. Statistical inference on gene counts and differential expression analysis was performed with the R statistical software package. Upstream Regulator Analysis was used to analyse differential analysis output. Gene Set Enrichment analysis (GSEA) was performed with a STAT1 gene signature derived from Care et al. (2015). Using the methodology defined in Reyneir et al. (2011), patients were separated into two cohorts (high vs low STAT1 pathway activation) for Kaplan-Meier survival analysis.

3.    RESULTS

STAT1 was expressed at statistically higher levels in responding patients (mean counts 1.3x times higher in responders, p value = 0.019). Differential expression analysis and Upstream Regulator Analysis demonstrated STAT1 as a significant regulator in responders (p-value of overlap 4.25E-10). GSEA analysis shows that a STAT1 gene signature is enriched in patients with treatment response. Finally, patients with higher STAT1 pathway activation demonstrated longer overall survival than those with lower STAT1 activation (median OS 9.5 months vs. 7.6 months, p=0.0023).

4.    CONCLUSIONS

The importance of STAT1 in the human response to immunotherapy is supported by four analyses on the IMvigor 210 dataset. This corroborates work in animal models and demonstrates the value of exploratory data analysis on publically available human RNAseq data.

  1. Care, M. A., Westhead, D. R. & Tooze, R. M. Gene expression meta-analysis reveals immune response convergence on the IFNgamma-STAT1-IRF1 axis and adaptive immune resistance mechanisms in lymphoma. Genome Med 7, 96, doi:10.1186/s13073-015-0218-3 (2015).
  2. Reynier, F. et al. Importance of correlation between gene expression levels: application to the type I interferon signature in rheumatoid arthritis. PLoS One 6, e24828, doi:10.1371/journal.pone.0024828 (2011).