Vasiliki Pelekanou

Quantitative Assessment of the Heterogeneity of PD-L1 Expression in Non–Small-Cell Lung Cancer

IMPORTANCE Early-phase trials with monoclonal antibodies targeting PD-1 (programmed cell death protein 1) and PD-L1 (programmed cell death 1 ligand 1) have demonstrated durable clinical responses in patients with non-small-cell lung cancer (NSCLC). However, current assays for the prognostic and/or predictive role of tumor PD-L1 expression are not standardized with respect to either quantity or distribution of expression. OBJECTIVE To demonstrate PD-L1 protein distribution in NSCLC tumors using both conventional immunohistochemistry (IHC) and quantitative immunofluorescence (QIF) and compare results obtained using 2 different PD-L1 antibodies. DESIGN, SETTING, AND PARTICIPANTS PD-L1 was measured using E1L3N and SP142, 2 rabbit monoclonal antibodies, in 49 NSCLC whole-tissue sections and a corresponding tissue microarray with the same 49 cases. Non-small-cell lung cancer biopsy specimens from 2011 to 2012 were collected retrospectively from the Yale Thoracic Oncology Program Tissue Bank. Human melanoma Mel 624 cells stably transfected with PD-L1 as well as Mel 624 parental cells, and human term placenta whole tissue sections were used as controls and for antibody validation. PD-L1 protein expression in tumor and stroma was assessed using chromogenic IHC and the AQUA (Automated Quantitative Analysis) method of QIF. Tumor-infiltrating lymphocytes (TILs) were scored in hematoxylin-eosin slides using current consensus guidelines. The association between PD-L1 protein expression, TILs, and clinicopathological features were determined. MAIN OUTCOMES AND MEASURES PD-L1 expression discordance or heterogeneity using the diaminobenzidine chromogen and QIF was the main outcome measure selected prior to performing the study. RESULTS Using chromogenic IHC, both antibodies showed fair to poor concordance. The PD-L1 antibodies showed poor concordance (Cohen κ range, 0.124-0.340) using conventional chromogenic IHC and showed intra-assay heterogeneity (E1L3N coefficient of variation [CV], 6.75%-75.24%; SP142 CV, 12.17%-109.61%) and significant interassay discordance using QIF (26.6%). Quantitative immunofluorescence showed that PD-L1 expression using both PD-L1 antibodies was heterogeneous. Using QIF, the scores obtained with E1L3N and SP142 for each tumor were significantly different according to nonparametric paired test (P < .001). Assessment of 588 serial section fields of view from whole tissue showed discordant expression at a frequency of 25%. Expression of PD-L1 was correlated with high TILs using both E1L3N (P = .007) and SP142 (P = .02). CONCLUSIONS AND RELEVANCE Objective determination of PD-L1 protein levels in NSCLC reveals heterogeneity within tumors and prominent interassay variability or discordance. This could be due to different antibody affinities, limited specificity, or distinct target epitopes. Efforts to determine the clinical value of these observations are under way.

Quantitative and pathologist-read comparison of the heterogeneity of programmed death-ligand 1 (PD-L1) expression in non-small cell lung cancer

PD-L1 is expressed in a percentage of lung cancer patients and those patients show increased likelihood of response to PD-1 axis therapies. However, the methods and assays for the assessment of PD-L1 using immunohistochemistry are variable and PD-L1 expression appears to be highly heterogeneous. Here, we examine assay heterogeneity parameters toward the goal of determining variability of sampling and the variability due to pathologist-based reading of the immunohistochemistry slide. SP142, a rabbit monoclonal antibody, was used to detect PD-L1 by both chromogenic immunohistochemistry and quantitative immunofluorescence using a laboratory-derived test. Five pathologists scored the percentage of PD-L1 positivity in tumor- and stromal-immune cells of 35 resected non-small cell lung cancer cases, each represented on three separate blocks. An intraclass correlation coefficient of 94% agreement was seen among the pathologists for the assessment of PD-L1 in tumor cells, but only 27% agreement was seen in stromal/immune cell PD-L1 expression. The block-to-block reproducibility of each pathologist’s score was 94% for tumor cells and 75% among stromal/immune cells. Lin’s concordance correlation coefficient between pathologists’ readings and the mean immunofluorescence score among blocks was 94% in tumor and 68% in stroma. Pathologists were highly concordant for PD-L1 tumor scoring, but not for stromal/immune cell scoring. Pathologist scores and immunofluorescence scores were concordant for tumor tissue, but not for stromal/immune cells. PD-L1 expression was similar among all the three blocks from each tumor, indicating that staining of one block is enough to represent the entire tumor and that the spatial distribution of heterogeneity of expression of PD-L1 is within the area represented in a single block. Future studies are needed to determine the minimum representative tumor area for PD-L1 assessment for response to therapy.

B7-H3 expression in NSCLC and its association with B7-H4, PD-L1 and Tumor Infiltrating Lymphocytes

Purpose: The immune checkpoint PD-1 and its receptor B7-H1 (PD-L1) are successful therapeutic targets in cancer but less is known about other B7 family members. Here, we determined the expression level of B7-H3 protein in non–small cell lung cancer (NSCLC) and evaluated its association with tumor-infiltrating lymphocytes (TIL), PD-L1, B7-H4, and major clinicopathologic characteristics is in 3 NSCLC cohorts. Experimental design: We used multiplexed automated quantitative immunofluorescence (QIF) to assess the levels of B7-H3, PD-L1, B7-H4, and TILs in 634 NSCLC cases with validated antibodies. Associations between the marker levels, major clinicopathologic variables and survival were analyzed. Results: Expression of B7-H3 protein was found in 80.4% (510/634) of the cases. High B7-H3 protein level (top 10 percentile) was associated with poor overall survival (P < 0.05). Elevated B7-H3 was consistently associated with smoking history across the 3 cohorts, but not with sex, age, clinical stage, and histology. Coexpression of B7-H3 and PD-L1 was found in 17.6% of the cases (112/634) and with B7-H4 in 10% (63/634). B7-H4 and PD-L1 were simultaneously detected only in 1.8% of NSCLCs (12/634). The expression of B7-H3 was not associated with the levels of CD3-, CD8-, and CD20-positive TILs. Conclusions: B7-H3 protein is expressed in the majority of NSCLCs and is associated with smoking history. High levels of B7-H3 protein have a negative prognostic impact in lung carcinomas. Coexpression of B7-H3 with PD-L1 and B7-H4 is relatively low, suggesting a nonredundant biological role of these targets. Clin Cancer Res; 23(17); 5202–9. ©2017 AACR.

Erratum to: Effect of neoadjuvant chemotherapy on tumor-infiltrating lymphocytes and PD-L1 expression in breast cancer and its clinical significance

BackgroundThe effects of neoadjuvant chemotherapy on immune markers remain largely unknown. The specific aim of this study was to assess stromal tumor-infiltrating lymphocytes (TILs) and programmed death ligand 1 (PD-L1) protein expression in a cohort of breast cancer patients treated with neoadjuvant chemotherapy.MethodsUsing quantitative immunofluorescence, we investigated stromal TILs and PD-L1 protein expression in pre-treatment and residual breast cancer tissue from a Yale Cancer Center patient cohort of 58 patients diagnosed with breast cancer from 2003 to 2009 and treated with neoadjuvant chemotherapy. We compared the TIL count and PD-L1 status in paired pre-treatment and residual cancer tissues and correlated changes and baseline levels with survival.ResultsOf the 58 patients, 46 (79.3%) had hormone-positive and 34 (58.6%) had node-positive breast cancer. Eighty-six percent of residual cancer tissues had TIL infiltration and 17% had PD-L1 expression. There was a trend for higher TIL counts in postchemotherapy compared to prechemotherapy samples (p = 0.09). Increase in TIL count was associated with longer 5-year recurrence-free survival (p = 0.02, HR = 3.9, 95% CI = 1.179–15.39). PD-L1 expression (both stromal and tumor cells) was significantly lower in post-treatment samples (p = 0.001). Change in PD-L1 expression after therapy or TILs and PD-L1 expression in the posttreatment samples did not correlate with survival.ConclusionsIncrease in stromal TILs in residual cancer compared to pretreatment tissue is associated with improved recurrence-free survival. Despite a trend for increasing TIL counts, PD-L1 expression decreased in residual disease compared to pretreatment samples.

Non-malignant respiratory epithelial cells preferentially proliferate from resected non-small cell lung cancer specimens cultured under conditionally reprogrammed conditions.

The “conditionally reprogrammed cells” (CRC) method, using a Rho kinase inhibitor and irradiated mouse fibroblast cells has been described for the efficient growth of cells from malignant and non-malignant samples from primary tumor and non-malignant sites. Using the CRC method, four institutions independently cultured tumor tissues from 48 non-small cell lung cancers (NSCLC, mostly from primary resected tumors) and 22 non-malignant lungs. We found that epithelial cells could be cultured from tumor and non-malignant lung. However, epithelial cells cultured from tumors had features of non-malignant respiratory epithelial cells which include: 1) among 22 mutations found in the original tumors only two mutations were found in the CRC cultures with reduced frequency (31% to 13% and 92% to 15% from original tumor and CRC culture respectively); 2) copy number variation was analyzed in 9 tumor and their CRC cultures and only diploid patterns were found in CRC cultures; 3) mRNA expression profiles were similar to those of normal respiratory epithelial cells; and 4) co-culture of tumor and non-malignant lung epithelial cells resulted in mostly non-malignant cells. We conclude that CRC method is a highly selective and useful method for the growth of non-malignant respiratory epithelial cells from tumor specimens and only occasionally do such CRC cultures contain a small subpopulation of cancer cells marked by oncogenic mutations. While our findings are restricted to resected primary NSCLC, they indicated the necessity to fully characterize all CRC cultures and the need to develop culture technology that facilitates the growth of primary lung cancers.

Nuclear IRF-1 expression as a mechanism to assess “Capability” to express PD-L1 and response to PD-1 therapy in metastatic melanoma

BackgroundPredictive biomarkers for antibodies against programmed death 1 (PD-1) remain a major unmet need in metastatic melanoma. Specifically, response is seen in tumors that do not express programmed death ligand 1 (PD-L1), highlighting the need for a more sensitive biomarker. We hypothesize that capacity to express PD-L1, as assessed by an assay for a PD-L1 transcription factor, interferon regulatory factor 1 (IRF-1), may better distinguish patients likely to benefit from anti-PD-1 immunotherapy.MethodsSamples from 47 melanoma patients that received nivolumab, pembrolizumab, or combination ipilimumab/nivolumab at Yale New Haven Hospital from May 2013 to March 2016 were collected. Expression of IRF-1 and PD-L1 in archival pre-treatment formalin-fixed, paraffin-embedded tumor samples were assessed by the AQUA method of quantitative immunofluorescence. Objective radiographic response (ORR) and progression-free survival (PFS) were assessed using modified RECIST v1.1 criteria.ResultsNuclear IRF-1 expression was higher in patients with partial or complete response (PR/CR) than in patients with stable or progressive disease (SD/PD) (p = 0.044). There was an insignificant trend toward higher PD-L1 expression in patients with PR/CR (p = 0.085). PFS was higher in the IRF-1-high group than the IRF-1-low group (p = 0.017), while PD-L1 expression had no effect on PFS (p = 0.83). In a subset analysis, a strong association with PFS is seen in patients treated with combination ipilimumab and nivolumab (p = 0.0051).ConclusionsAs a measure of PD-L1 expression capability, IRF-1 expression may be a more valuable predictive biomarker for anti-PD-1 therapy than PD-L1 itself.

Abstract 1310: Programmed death ligand-1 (PD-L1) heterogeneity in non-small cell lung cancer (NSCLC)

Early phase I trials with monoclonal antibodies targeting PD-1/PD-L1 have demonstrated durable clinical responses in patients with NSCLC. However, the prognostic/predictive role of tumor PD-L1 expression has not been determined. Current assays for evaluation of PD-L1 protein expression are not standardized. Here, we demonstrate PD-L1 protein distribution in NSCLC tumors using both conventional immunohistochemistry (IHC) and quantitative immunofluorescence (QIF), and compare results obtained using three PD-L1 antibodies targeting the intracellular (IC) and extracellular (EC) domains. We measured PD-L1 protein expression using two antibodies against the IC domain (E1L3N [Cell Signaling Technology] and SP142 [Spring Biosciences Inc.]) and one antibody binding the EC domain (E1J2J [Cell Signaling Technology]) in 49 NSCLC whole tissue sections and a TMA with the same 49 cases. Mel624 cells stably transfected with PD-L1, as well as Mel624 parental cells and human term placenta were used as controls and for antibody validation. PD-L1 protein expression in tumor and stroma was assessed using chromogenic IHC and the AQUA® method of QIF. For IHC, PD-L1 positivity was determined using previously reported cut-points for tumor (1%, 5%, and 50%) and stroma (5%). IHC inter-assay concordance was evaluated using kappa coefficient. AQUA scores were measured in an average of 28.6 fields of view per case or by TMA at two-fold redundancy. Linear regression coefficients (R2) were used to compare antibody QIF scores. Tumor-infiltrating lymphocytes were scored in hematoxylin/eosin stained slides using current consensus guidelines. Chromogenic IHC assays showed fair to poor concordance (kappa 0.124 - 0.552) for all cut-points. QIF showed PD-L1 immunopositivity was heterogeneous for all three PD-L1 antibodies. R2 values were lower when E1J2J (EC) was compared to E1L3N and SP142 (IC) (R2 = 0.090 and 0.079), while coefficients were higher when the IC antibodies were compared (R2 = 0.658). E1L3N and SP142 significantly correlated with high TILs (p = 0.007 and p = 0.021) but E1J2J did not (p = 0.281). E1J2J significantly correlated with older age (p = 0.038), but no other clinicopathological feature. SP142 significantly correlated with lymph node positive status (p = 0.030) Objective determination of PD-L1 protein levels in NSCLC reveals significant heterogeneity within tumors. There is significant inter-assay variability, even between assays detecting the same protein domain. This could be due to different antibody affinities, cross reactivity, or distinct target epitopes. Efforts to determine the clinical value of these observations are underway. Citation Format: Joseph McLaughlin, Kurt A. Schalper, Daniel E. Carvajal-Hausdorf, Vasiliki Pelekanou, Vamsidhar Velcheti, Herbert Haack, Matthew R. Silver, Roy Herbst, Patricia LoRusso, David L. Rimm. Programmed death ligand-1 (PD-L1) heterogeneity in non-small cell lung cancer (NSCLC). [abstract]. In: Proceedings of the 106th Annual Meeting of the American Association for Cancer Research; 2015 Apr 18-22; Philadelphia, PA. Philadelphia (PA): AACR; Cancer Res 2015;75(15 Suppl):Abstract nr 1310. doi:10.1158/1538-7445.AM2015-1310

Tumor-Infiltrating Lymphocytes and PD-L1 Expression in Pre- and Posttreatment Breast Cancers in the SWOG S0800 Phase II Neoadjuvant Chemotherapy Trial

Our aim was to examine the association of pretreatment tumor-infiltrating lymphocyte (TIL) count and PD-L1 levels with pathologic complete response (pCR) and assess immune marker changes following treatment in tumor specimens from the S0800 clinical trial, which randomized patients to bevacizumab + nab-paclitaxel, followed by doxorubicin/cyclophosphamide (AC) versus two control arms without bevacizumab (varying sequence of AC and nab-paclitaxel). TILs were assessed in 124 pre- and 62 posttreatment tissues (including 59 pairs). PD-L1 was assessed in 120 pre- and 43 posttreatment tissues (including 39 pairs) using the 22C3 antibody. Baseline and treatment-induced immune changes were correlated with pCR and survival using estrogen receptor (ER) and treatment-adjusted logistic and Cox regressions, respectively. At baseline, the mean TIL count was 17.4% (17% had zero TILs, 9% had ≥50% TILs). Posttreatment, mean TIL count decreased to 11% (5% had no TILs, 2% had >50% TILs). In paired samples, the mean TIL change was 15% decrease. Baseline PD-L1 was detected in 43% of cases (n = 5 in tumor cells, n = 29 stroma, n = 18 tumor + stroma). Posttreatment, PD-L1 expression was not significantly lower (33%). Higher baseline TIL count and PD-L1 positivity rate were associated with higher pCR rate even after adjustment for treatment and ER status (P = 0.018). There was no association between TIL counts, PD-L1 expression, and survival due to few events. In conclusion, TIL counts, but not PD-L1 expression, decreased significantly after treatment. Continued PD-L1 expression in some residual cancers raises the possibility that adjuvant immune checkpoint inhibitor therapy could improve survival in this patient population. Mol Cancer Ther; 17(6); 1324–31. ©2018 AACR.

Ki67 reproducibility using digital image analysis: an inter-platform and inter-operator study

Ki67 expression has been a valuable prognostic variable in breast cancer, but has not seen broad adoption due to lack of standardization between institutions. Automation could represent a solution. Here we investigate the reproducibility of Ki67 measurement between three image analysis platforms with supervised classifiers performed by the same operator, by multiple operators, and finally we compare their accuracy in prognostic potential. Two breast cancer patient cohorts were used for this study. The standardization was done with the 30 cases of ER+ breast cancer that were used in phase 3 of International Ki67 in Breast Cancer Working Group initiatives where blocks were centrally cut and stained for Ki67. The outcome cohort was from 149 breast cancer cases from the Yale Pathology archives. A tissue microarray was built from representative tissue blocks with median follow-up of 120 months. The Mib-1 antibody (Dako) was used to detect Ki67 (dilution 1:100). HALO (IndicaLab), QuantCenter (3DHistech), and QuPath (open source software) digital image analysis (DIA) platforms were used to evaluate Ki67 expression. Intraclass correlation coefficient (ICC) was used to measure reproducibility. Between-DIA platform reproducibility was excellent (ICC: 0.933, CI: 0.879–0.966). Excellent reproducibility was found between all DIA platforms and the reference standard Ki67 values of Spectrum Webscope (QuPath-Spectrum Webscope ICC: 0.970, CI: 0.936–0.986; HALO-Spectrum Webscope ICC: 0.968, CI: 0.933–0.985; QuantCenter-Spectrum Webscope ICC: 0.964, CI: 0.919–0.983). All platforms showed excellent intra-DIA reproducibility (QuPath ICC: 0.992, CI: 0.986–0.996; HALO ICC: 0.972, CI: 0.924–0.988; QuantCenter ICC: 0.978, CI: 0.932–0.991). Comparing each DIA against outcome, the hazard ratios were similar. The inter-operator reproducibility was particularly high (ICC: 0.962–0.995). Our results showed outstanding reproducibility both within and between-DIA platforms, including one freely available DIA platform (QuPath). We also found the platforms essentially indistinguishable with respect to prediction of breast cancer patient outcome. Results justify multi-institutional DIA studies to assess clinical utility.This study shows excellent reproducibility among 3 different digital image (DIA) analysis platforms in Ki67 scoring. An outstanding concordance was found among four operators using the same calibrated DIA platform suggesting a highly reproducible Ki67 scoring method. Finally, the authors also found that all three DIA platforms were essentially indistinguishable with respect to prediction of breast cancer patient outcome.

Immunological differences between primary and metastatic breast cancer.

Background Little is known about how the immune microenvironment of breast cancer evolves during disease progression. Patients and methods We compared tumor infiltrating lymphocyte (TIL) count, programmed death-ligand 1 (PD-L1) protein expression by immunohistochemistry and mRNA levels of 730 immune-related genes using Nanostring technology in primary and metastatic cancer samples. Results TIL counts and PD-L1 positivity were significantly lower in metastases. Immune cell metagenes corresponding to CD8, T-helper, T-reg, Cytotoxic T, Dendritic and Mastoid cells, and expression of 13 of 29 immuno-oncology therapeutic targets in clinical development including PD1, PD-L1, and CTLA4 were significantly lower in metastases. There was also coordinated down regulation of chemoattractant ligand/receptor pairs (CCL19/CCR7, CXCL9/CXCR3, IL15/IL15R), interferon regulated genes (STAT1, IRF-1,-4,-7, IFI-27,-35), granzyme/granulysin, MHC class I and immune proteasome (PSMB-8,-9,-10) expression in metastases. Immunotherapy response predictive signatures were also lower. The expression of macrophage markers (CD163, CCL2/CCR2, CSF1/CSFR1, CXCR4/CXCL12), protumorigenic toll-like receptor pathway genes (CD14/TLR-1,-2,-4,-5,-6/MyD88), HLA-E, ecto-nuclease CD73/NT5E and inhibitory complement receptors (CD-59,-55,-46) remained high in metastases and represent potential therapeutic targets. Conclusions Metastatic breast cancers are immunologically more inert than the corresponding primary tumors but some immune-oncology targets and macrophage and angiogenesis signatures show preserved expression and suggest therapeutic combinations for clinical testing.