Combination of variations in inflammation- and endoplasmic reticulum-associated genes as putative biomarker for bevacizumab response in KRAS wild-type colorectal cancer

Chemotherapy combined with the angiogenesis inhibitor bevacizumab (BVZ) is approved as a first-line treatment in metastatic colorectal cancer (mCRC). Limited clinical benefit underpins the need for improved understanding of resistance mechanisms and the elucidation of novel predictive biomarkers. We assessed germline single-nucleotide polymorphisms (SNPs) in 180 mCRC patients (Angiopredict [APD] cohort) treated with combined BVZ + chemotherapy and investigated previously reported predictive SNPs. We further employed a machine learning approach to identify novel associations. In the APD cohort IL8 rs4073 any A carriers, compared to TT carriers, were associated with worse progression-free survival (PFS) (HR = 1.51, 95% CI:1.03–2.22, p-value = 0.037) and TBK1 rs7486100 TT carriers, compared to any A carriers, were associated with worse PFS in KRAS wild-type (wt) patients (HR = 1.94, 95% CI:1.04–3.61, p-value = 0.037), replicating previous findings. Machine learning identified novel associations in genes encoding the inflammasome protein NLRP1 and the ER protein Sarcalumenin (SRL). A negative association between PFS and carriers of any A at NLRP1 rs12150220 and AA for SRL rs13334970 in APD KRAS wild-type patients (HR = 4.44, 95% CI:1.23–16.13, p-value = 0.005), which validated in two independent clinical cohorts involving BVZ, MAVERICC and TRIBE. Our findings highlight a key role for inflammation and ER signalling underpinning BVZ + chemotherapy responsiveness.

missense variant on the NLR Family Pyrin Domain Containing 1 gene, in a) ANGIOPREDICT patients treated with BVZ + chemotherapy; b) patients from the MAVERICC cohort, treated with combination chemotherapy (FOLFIRI) and BVZ. c) patients from the TRIBE-A cohort, treated with combination chemotherapy (FOLFIRI) and BVZ. df ) KRAS wt patients in the same respective cohorts, any A versus TT. gi) KRAS mutant patients in the same respective cohorts, any A versus TT. X-axis: days from randomization, Y-axis: estimated probability of progression free survival (PFS). Log-rank test p-values p are given for each KM plot.
to an intronic region on the Sarcalumenin (SRL) gene in a) ANGIOPREDICT patients treated with BVZ + chemotherapy; b) patients from the MAVERICC cohort, treated with BVZ + chemotherapy; c) patients from the TRIBE cohort, treated with BVZ + chemotherapy. df) KRAS wt patients in the same respective cohorts, any G versus AA. KM plots not shown for KRAS mutants as low numbers for AA carriers in both Angiopredict and MAVERICC. X-axis: days from randomization, Y-axis: estimated probability of progression free survival (PFS). Logrank test p-values are p given for each KM plot. Table S1. 47 SNPs that have been selected to be associated with progression-free survival, using repeated penalized regression as outlined in Methods, in metastatic or locally advanced colorectal cancer patients after treatment with BVZ combination therapy (ANGIOPREDICT retrospective cohort). Hazard ratios given on the table were computed using Cox proportional hazards analysis for all patients treated with BVZ, adjusting for age, gender, grade of differentiation, chemotherapy line, chemotherapy type and status of somatic BRAF codon 600 V/E. HRhet and HR headings designate the hazard ratios for the heterozygous and homozygous on the minor allele genotypes respectively, compared to baseline (homozygous on the major allele genotypes), when a codominant model was most significant. When a dominant (d) or a recessive (r) model were most significant, the respective hazard ratios are given under the HR heading. The HRs for each included SNP represent the exponentiated coefficients computed in the respective multivariate Cox model and are interpretable as multiplicative effects on the hazard: the % by which the risk of relapse is increased with change to one alternative genotype (in keeping all other considered features to baseline). Thus, for example, a patient carrying at least one alternative allele C for rs7104785 on PIDD increases the risk of relapse by ~60% considering a dominant model for this SNP.

Supplementary
The order of magnitude of the probability p that the model coefficient for the indicated genotype could be zero (the hazard ratio could be equal to one) is indicated with .(dot) 0.05<p<0.1; * 0.01≤p≤0.05; ** p of order 0.001; *** p of order 0.0001 and less. If no order of magnitude for p is given then p ≥ 0.1 and is not statistically significant.

Provided as Excel File.
TRIBE cohorts, overall and KRAS wt respectively. The associations found in APD validate in MAVERICC. In TRIBE KRAS wt, the same sense HR is found for SRL rs13334970 AA carriers as in the other two cohorts, without reaching statistical significance. HRs, the respective 95% CI and the across-genotype Wald test p-values are presented under codominant, dominant and recessive models, using multivariable Cox proportional hazard regression models adjusted for available (see Methods) covariates. For each HR specifically, the order of significance of the HR being different from 1 is given by stars. The stars give the order of magnitude of the probability p that the model coefficient for the indicated genotype could be zero (the hazard ratio could be 1), with * p of order 0.01, ** p of order 0.001, *** p of order 0.0001. Cox proportional hazard regression models adjusted for available (see Methods) covariates. For each HR, the order of significance of the HR being different from 1 is given by stars (giving the order of magnitude of the probability p that the model coefficient for the indicated genotype could be zero) with . (dot) 0.05<p<0.1; * 0.01≤p≤0.05; ** p of order 0.001; *** p of order 0.0001 or less.

ANGIOPREDICT overall
Progression-free Survival

MAVERICC overall
Progression-free Survival

TRIBE overall
Progression-free Survival