Kinomic profiling to predict sunitinib response of patients with metastasized clear cell Renal Cell Carcinoma

Abstract

Introduction

Treatment with Sunitinib, a potent multitargeted receptor tyrosine kinase inhibitor (TKI) has increased the progression-free survival (PFS) and overall-survival (OS) of patients with metastasized renal cell carcinoma (mRCC). With modest OS improvement and variable response and toxicity predictive and/or prognostic biomarkers are needed to personalize patient management: Prediction of individual TKI therapy response and resistance will increase successful treatment outcome while reducing unnecessary drug use and expense. The aim of this study was to investigate whether kinase activity analysis can predict sunitinib response and/or toxicity using tissue samples obtained from primary clear cell RCC (ccRCC) from a cohort of clinically annotated patients with mRCC receiving sunitinib as first-line treatment.

Materials and Methods

EuroTARGET partners collected ccRCC and matched normal kidney tissue samples immediately after surgery, snap-frozen and stored at -80°C until use. Phosphotyrosine-activity profiling was performed using PamChip® peptide microarrays (144 peptides derived from known phosphorylation sites in Protein Tyrosine Kinase substrates) of lysed tissue samples (5 µg protein input) of 163 mRCC patients. Evolve software Was used to analyze kinome profiles and Bionavigator was used for unsupervised and supervised clustering. The kinexus kinase predictor (www.phosphonet.ca) was used to analyze the peptide lists within the clusters.

Results

Kinome data was available from 94 patients who received sunitinib as 1st-line treatment and had complete follow-up of their clinical data (PFS, OS and toxicity) for at least 6 months. Matched normal tissue was available from 14 mRCC patients. Supervised clustering of basal kinome activity could correctly classify mRCC patients with PFS >9 months versus PFS<9 months with an accuracy of 61 %. Unsupervised hierarchical clustering revealed 3 major clusters related to immune signaling, VEGF pathway, and immune signaling/cell adhesion. Basal kinase activity levels of patients with short PFS were substantially higher compared to patients who experienced extended PFS.

Discussion/Conclusion

Based on kinase levels ccRCC tumors can be subdivided into 3 clusters which may reflect the aggressiveness of these tumors. The accuracy of response prediction of 61 % based on basal kinase levels is too low to justify implementation. STK assays may help to predict sunitinib toxicity and guide clinical management. Additionally, it is possible that mRCC patients with an immune kinase signature are better checkpoint inhibitor candidates, but this needs to be studied.