Background: PD-1 blockade has demonstrated antitumor activity in diffuse large B-cell lymphoma (DLBCL). In order to select patients who are more likely to respond to PD-1 blockade, predictive biomarkers are being intensively pursued. PD-1 and PD-L1 expression in DLBCL treated with R-CHOP as standard-of-care has been examined in several studies. However, only overall PD-1+ cells but not T-cell specific PD-1 expression have been quantitated, and results of the prognostic significance of PD-L1 expression have been inconsistent in DLBCL patients. Whether ongoing PD-1/PD-L1 interaction in tumors contributes to the adverse prognostic effect of PD-1 and PD-L1 expression is unknown.

Patients and Methods: Tissue microarray (TMA) sections ofdiagnostic samples from 414 patients with de novo DLBCL treated with R-CHOP were fluorescently stained with CD3 (T cell marker), PD-1, and PD-L1. Positive cells and PD-1/PD-L1 interaction were quantitatively assessed by the Automated Quantitative Analysis (AQUA) algorithms. To characterize PD-1/PD-L1 expression in each sample, CD3+, PD-1+, PD-L1+, and PD-1+CD3+ (double positive, i.e., PD-1+ T) cell numbers were divided by total nucleated cells (DAPI+) in the representative tumor areas. Ongoing PD-1/PD-L1 interactions were evaluated by the proximity of PD-1+ T cells to the PD-L1+ compartments and proportion of PD-1+ T cells co-localized with PD-L1, generating an interaction score for each sample. Biomarker expression and interaction scores were correlated with patient survival.

Results: CD3+, PD-1+, PD-L1+, and PD-1+CD3+ cells were identified in 90%, 30%, 53%, and 18% of the cohort, respectively. The mean percentages of CD3+, PD-1+, PD-L1+, and PD-1+CD3+ cells in DLBCL tissues were 12.1%, 0.44%, 3.6%, and 0.19%, respectively. Activated B-cell-like DLBCL had a significantly higher mean percentage of PD-L1+ cells, but not PD-1+CD3+ cells, compared to germinal center B-cell-like DLBCL (p=0.0012). Three prognostic tiers were identified. Firstly, absence to very low percentage (0-3.5%) of tumor-infiltrating CD3+ T cells in the tissues was associated with poorer overall survival (OS) of DLBCL patients (p=0.028). CD3high DLBCL (>3.5% of all cells were CD3+) was associated with significantly higher percentages of PD-L1+ and PD-1+ cells (p<0.0001 and p=0.005, respectively). Secondly, in CD3high DLBCL, presence of PD-1+ T cells correlated with significantly poorer OS and progression-free survival (PFS) of patients (p=0.028 and p=0.0099, respectively). In contrast, presence/absence of PD-1+ non-T cells was not prognostic. Presence of PD-1+ T cells in CD3high DLBCL was associated with a higher mean percentage of PD-L1+ cells (p=0.019). However, PD-1/PD-L1 interaction scores did not add to the prognostic effect. This may suggest that phenotype of PD-1/PD-L1 interaction is not a biomarker to specify the immunosuppressive role of the PD-1/PD-L1 axis. Lastly, in CD3high DLBCL cases with PD-1−negative T cells, PD-L1 positivity correlated with significantly poorer OS and PFS of patients (p=0.0034 and p=0.0059, respectively), although PD-L1+ was not associated with non-T PD-1 expression either, suggesting PD-1−independent function of PD-L1. In the overall cohort, presence of PD-1+ T cells was associated with a trend of poorer PFS (p=0.079), whereas PD-L1 expression was not prognostic.

Conclusions: PD-1 expression in T cells and lack of tumor-infiltrating CD3+ T cells are associated with significantly poorer survival of DLBCL patients. In addition, our data suggest that PD-1 can be a prognostic marker independent of PD-1/PD-L1 co-localization, and that PD-L1 has other PD-1-independent inhibitory functions. These results demonstrate the importance of both T-cell influx and antitumor function in prolonging the survival of DLBCL patients, and provide rationales for PD-1 blockade and immunotherapies activating T cell-mediated immune responses in DLBCL.

Disclosures

Tran: Genoptix, Inc: Employment. Adams: Navigate BioPhama Services, Inc: Employment. Roscoe: Navigate BioPhama Services, Inc: Employment. Hsi: Seattle Genetics: Consultancy, Honoraria, Speakers Bureau; Abbvie: Research Funding; Eli Lilly and Co.: Research Funding; Cellerant Therapeutics: Research Funding. Piris: Kura Oncology: Research Funding. Winter: Glaxo-Smith-Kline: Research Funding; Merck: Research Funding. Vaupel: Genoptix Laboratory: Employment. Dakappagari: Genoptix, Inc: Employment.

Author notes

*

Asterisk with author names denotes non-ASH members.

Sign in via your Institution