The immune landscape of prostate cancer and nomination of PD-L2 as a potential therapeutic target

Shuang G. Zhao, Jonathan Lehrer, S. Laura Chang, Rajdeep Das, Nicholas Erho, Yang Liu, Martin Sjöström, Robert B. Den, Stephen J. Freedland, Eric A. Klein, R. Jeffrey Karnes, Edward M. Schaeffer, Melody Xu, Corey Speers, Paul L. Nguyen, Ashley E. Ross, June M. Chan, Matthew R. Cooperberg, Peter R. Carroll, Elai DavicioniLawrence Fong, Daniel E. Spratt, Felix Y. Feng

Research output: Contribution to journalArticlepeer-review

48 Scopus citations


Background: Immunotherapy has been less successful in treating prostate cancer than other solid tumors. We sought to better understand the immune landscape in prostate cancer and identify immune-related biomarkers and potential therapeutic targets. Methods: We analyzed gene expression data from 7826 prospectively collected prostatectomy samples (2013-2016), and 1567 retrospective samples with long-term clinical outcomes, for a total of 9393 samples, all profiled on the same commercial clinical platform in a CLIA-certified lab. The primary outcome was distant metastasis-free survival (DMFS). Secondary outcomes included biochemical recurrence-free survival (bRFS), prostate cancer-specific survival (PCSS), and overall survival (OS). All statistical tests were two-sided. Results: Unsupervised hierarchical clustering of hallmark pathways demonstrated an immune-related tumor cluster. Increased estimated immune content scores based on immune-specific genes from the literature were associated with worse bRFS (hazard ratio [HR]=1.26 [95% confidence interval [CI]=1.12 to 1.42]; P<.001), DMFS (HR=1.34 [95% CI=1.13 to 1.58]; P<.001), PCSS (HR=1.53 [95% CI=1.21 to 1.92]; P<.001), and OS (HR=1.27 [95% CI=1.07 to 1.50]; P=.006). Deconvolution using Cibersort revealed that mast cells, natural killer cells, and dendritic cells conferred improved DMFS, whereas macrophages and T-cells conferred worse DMFS. Interestingly, while PD-L1 was not prognostic, consistent with its low expression in prostate cancer, PD-L2 was expressed at statistically significantly higher levels (P<.001) and was associated with worse bRFS (HR=1.17 [95% CI=1.03 to 1.33]; P=.01), DMFS (HR=1.25 [95% CI=1.05 to 1.49]; P=.01), and PCSS (HR=1.45 [95% CI=1.13 to 1.86]; P=.003). PD-L2 was strongly associated with immune-related pathways on gene set enrichment analysis suggesting that it is playing an important role in immune modulation in clinical prostate cancer samples. Furthermore, PD-L2 was correlated with radiation response pathways, and also predicted response to postoperative radiation therapy (PORT) on multivariable interaction analysis (P = .03). Conclusion: In the largest study of its kind to date, these results illustrate the complex relationship between the tumorimmune interaction, prognosis, and response to radiotherapy, and nominate PD-L2 as a potential novel therapeutic target in prostate cancer, potentially in combination with radiotherapy.

Original languageEnglish (US)
Pages (from-to)301-310
Number of pages10
JournalJournal of the National Cancer Institute
Issue number3
StatePublished - Mar 1 2019

ASJC Scopus subject areas

  • Oncology
  • Cancer Research


Dive into the research topics of 'The immune landscape of prostate cancer and nomination of PD-L2 as a potential therapeutic target'. Together they form a unique fingerprint.

Cite this