MTE24.01 Immunohistochemical Assessment of Biomarkers for Immune Checkpoint Inhibitors

Abstract

Programmed death receptor-1 (PD-1) is a type 1 membrane protein of the immunoglobulin superfamily that has an important role in restricting immune-mediated tissue danage secondary to inflammation and/or infection.1 The clinical advantage of antibodies that target either PD-1 or PD-L1 to block this ligand-receptor interface, allowing cancer killing by T cells became clear when CTLA4, an antagonist against the T-cell, such as ipilimumab, and afterward PD-1, showed an increase survival in patients with metastatic melanoma.2 Clinical investigations in lung cancer have demonstrated the benefit of PD-1 inhibitors pembrolizumab in advanced non–small cell lung cancer (NSCLC) and nivolumab in advanced squamous and nonsquamous NSCLC; both approved as second-line therapies by the US Food and Drug Administration (FDA).3-5 Others PD-L1 inhibitors such as atezolizumab and durvalumab have demonstrated effectiveness in several tumor types6-7 but they were not approved for clinical use until now. PD-1 inhibitors induce around of 20% of complete response frequency in patients with NSCLC, and persistent response in a subgroup of patients treated by immune checkpoint inhibitors. Garon et al.3 showed that tumors with PD-L1 expression ≥ 50% by immunohistochemistry (IHC) were significantly more expected to respond to pembrolizumab than those with less than 50% malignant cell expression. In contrast, response rates to nivolumab are significantly greater in patients with nonsquamous NSCLC, showing ≥ 1% tumor cell positivity.5 These differences are related to the combination of antibody clone and detection system as a companion diagnostic for selecting lung cancer patients for pembrolizumab therapy. Previous investigations reported response taxes in PD-L1–positive tumors of 31% to 52%, but particularly more than 16% of PD-L1–negative tumors also showed treatment response.1 This finding indicates that PD-L1 expression improves for responders but the absence of expression is not a complete indicator of advantage. PD-L1 expression did not predict differential response to nivolumab in lung squamous cell carcinoma as compared with docetaxel.4 PD-L1 in NSCLC is expressed on the membrane of tumor cells, and/or on immune infiltrating cells dendritic cells, antigen-presenting cells and T lymphocyte. PD-1, the PDL1 receptor, is expressed on tumor infiltrating lymphocytes, mainly CD4 T cells, T and B regulatory, NK, monocytes and DC. Concerning PD-L1 binding, PD-1 inhibits kinases involved in T cell activation. Two potential mechanisms are involved in expression of immune checkpoints on tumor cells and their immune stromal component: oncogenic signaling, and response to inflammatory signals.8 Tumor cells express multiple ligands and receptors and antitumor immune response can be enhanced by multi-level blockade of immune checkpoints. PD-1/PD-L1 commitment leads to HSP-2 phosphatase activity which dephosphorylates Pi3K and thus downregulate AKT.8 The positive score on tumor cells has not been evaluated nor enhanced or standardized.3,8 Brambilla and Ming8 assessed a score of positivity for prognosis analysis using E1L3N Cell Signaling antibody commercially available. They found that 20% of lung tumors cell expressed PD-L1 (≥ 20% intensity 2+3+), and 29% the immune stromal cells (T, macrophages, DC) ≥ 10% intensity 2+3+. PD-L1 positivity in both tumor and immune cells were seen in only 9% of NSCLC, 20,7% were both negative. There was no prognostic relevance of PD-L1 (tumor cells or stroma) whatever cut off by 10% increment or linear scoring was used. Only immune PD-L1 expression was correlated with a highly intense immune infiltrations. Previous published evaluations of prognostic value were discordant likely because immune checkpoints modulators play both positive and negative roles in the immune inhibitory pathways with some redundancy, and patients series and assays were not comparable. The two meta-analyzes with different antibodies, cutoffs, patient series, ethnicities and contribution of oncogene driven cancers, initial resection sample or contemporary biopsy rendered their interpretation extremely problematic. Global result was supporting a poor prognosis of “PD-L1 positivity” on tumor cells. Most of phase I trials works with four antibodies targeting PD-1 or its primary ligand PD-L1, response taxes appear higher in patients with increased tumor PD-L1 membrane expression by IHC. However, different antibody assays, absence of standardization, different score to determine PD-L1 positivity, companion test type, and a short number of specimens available for testing, accopled to the variability of the intervals between biopsy and test, has certainly disadvantaged the conclusion and prevent consensus to be reached.10 The best threshold was provided by Garon et al, with ≥ 50% of tumor cells PD-L1 positive to allow the highest response rate of 45% to pembrolizumab.3 In most trial series, biopsies or resected specimen were used and considerable difference between these samples occurs due to tumor heterogeneity. The reliability of small biopsy samples is questioned.10 Indeed lung tumor heterogeneity is characteristic and PD-L1 is typically heterogeneous in its distribution in the tumor majority as is PD-L1 positive immune cells. Multiple questions are still addressed before PD-L1 is considered as a definitive molecular predictor of effectiveness. As for prognostic evaluations, thresholds of ≥ 1%, ≥ 5%, ≥ 10%, ≥ 50% or continuous H score have been used. In addition, in a few trials, PD-L1 expression in TILs was predictive more than PD-L1 on tumor cells but the best cut off was not revealed. PDL1 expression predicts response to immune checkpoint inhibitors. Concordant results showing a better response if PDL1 + in several trials, using drug specific test and for Nivolumab also histology specific. We should evaluate membranous staining in tumor sample with at least 100 tumors cells and immune cells. Perspective for upgrading includes: 1) heterogeneity of the expression of PDL1 within tumor, primitive vs metastases number and size of samples; 2) surgical tissue versus biopsy and 3) archival versus new biopsy and 4) standardize the assays. Published abstracts showed high rates of concordance between primary and metastases (81%). Obtaining multiple biopsies from different areas of the tumor would enhance the validity of the results of IHC evaluation (160 patients=48% discordance).