12. Okt. 2017

Interferon-Gamma a more accurate biomarker than PD-L1?

Karachaliou N et al. (abstract P1.07-015) showed that Interferon-Gamma (IFNG) production by T-cells plays a critical role in anti-cancer immune responses by augmentation of MHC Class I expression, growth arrest, post-proteasomal trimming of antigen epitopes, recruitment of effector cells, induction of T-regs fragility and PD-L1 expression. Further research is warranted in order to validate whether INFG is more accurate than PD-L1.

Background

PD-L1 is induced by oncogenic signals or via INFG. STAT3, through DNMT1, epigenetically silences STAT1 and RIG-I and opposes INFG signalling. TET1 is a DNA demethylase. I kappa B kinase epsilon (IKBKE), a noncanonical I-kappa-B kinase, is essential for INFG induction, but can also promote NFATc1 phosphorylation and T cell response inhibition. Eomesodermin (Eomes) regulates T cell exhaustion. CCL5 (or Rantes), dependent on STAT3, causes myeloid-derived suppressor cell (MDSC) recruitment. YAP1 can also drive MDSC recruitment via CXCL5 signalling. This study explored whether the expression of genes related to INFG signalling, T cell exhaustion and MDSC recruitment is associated with response to ICB.

Method

Total RNA from pre-treatment tissue samples of 17 NSCLC and 21 melanoma patients treated with nivolumab and pembrolizumab respectively, was analyzed by qRT-PCR. INFG, STAT3, IKBKE, STAT1, RIG-I and PD-L1 mRNA were examined. CCL5, YAP1, CXCL5, NFATC1, EOMES and TET1 expression was additionally assessed. Gene expression was categorized with respect to tertiles and patients were divided into two risk groups (low and intermediate/high). CD8[+ ]tumor-infiltrating lymphocytes (TILs) and PD-L1 protein expression in tumor and CD8[+ ]TILs were examined by immunohistochemistry (SP57 and SP142 assay, respectively). Progression free survival (PFS), overall survival (OS) and Disease Control Rate (DCR) were estimated.

Result

Seventeen NSCLC patients, previously treated with one or more prior systemic therapies, received nivolumab. IKBKE was positively correlated with INFG (r=0.65, p=0.0124) and PD-L1 (r=0.58, p=0.0225) expression. RIG-I was loosely anticorrelated with NFATc1 (r=-0.55, p=0.0518). Among all biomarkers explored, only INFG was associated with PFS, OS and DCR. Specifically, PFS was significantly longer for nivolumab-treated patients with intermediate/high versus low INFG expression (5.1 versus 2.0 months, p=0.0124). OS was longer (though not statistically significant) for patients with intermediate/high versus low INFG expression (10.2 versus 4.9 months, p=0.0687). DCR to nivolumab was 71.43% for patients with intermediate/high INFG versus 0% for patients with low INFG expression. Neither PD-L1 immunohistochemistry expression nor CD8[+ ]TILs were related to nivolumab outcome. The same results were observed for 21 melanoma patients treated with pembrolizumab.

Reference:
Karachaliou N et al., abstract P1.07-015: Interferon-Gamma (INFG) as a Biomarker to Guide Immune Checkpoint Blockade (ICB) in Cancer Therapy