PD-L1 may serve as a prognostic biomarker in HPV-associated head and neck cancer patients

Presentation: AHNS014
Topic: Immunotherapy and Other Adjuvant Treatments
Type: Oral
Date: Wednesday, April 18, 2018
Session: 11:30 AM - 12:20 PM Immunotherapy and Adjuvent Therapies
Authors: Austin K Mattox1, Francesco Sabbatino2, Vincenzo Villani2, Soldano Ferrone2, William C Faquin3, Hang Lee4, Jill Brooks5, Armida Lefranc-Torres6, Derrick T Lin6, Lori J Wirth7, Christopher C McConkey8, Hisham Mehanna5, Sara I Pai2
Institution(s): 1Department of Oncology, Johns Hopkins University School of Medicine, Baltimore, MD, 2Department of Surgery, Massachusetts General Hospital, Harvard Medical School, Boston, MA, 3Department of Pathology, Massachusetts General Hospital, Harvard Medical School, Boston, MA, 4Biostatistics Center, Massachusetts General Hospital, Harvard Medical School, Boston, MA, 5Institute of Head and Neck Studies and Education (InHANSE), University of Birmingham, UK, 6Department of Otology and Laryngology, Massachusetts Eye and Ear Infirmary, Boston, MA, 7Department of Medical Oncology, Dana-Farber Cancer Institute, Harvard Medical School, Boston, MA, 8Warwick Clinical Trials Unit, University of Warwick, UK

Several mechanisms of immune evasion have been reported in head and neck squamous cell carcinomas (HNSCCs), including the down-regulation of the HLA Class I antigen processing machinery (APM) components and activation of the PD-1:PD-L1 axis. Programmed death-ligand 1 (PD-L1) expression can be induced by interferon-gamma secretion by T cells and may serve as a marker for the activation of tumor antigen-specific immune responses. Since defects in HLA Class I/II expression can impact tumor antigen-specific immune responses, we explored PD-L1 expression in the setting of HLA Class I/II mutations in HNSCCs and correlated the expression of these markers with clinical outcome to standard of care treatment.

We analyzed the combined published and provisional dataset for HNSCC for genomic alterations in HLA Class I heavy chains (HLA-A/B/C), β2 microglobulin (β2M), and HLA Class II α and β chain (HLA-DR/DQ/DP) genes in 406 HNSCCs and report that mutation frequencies in APM components are infrequent. When alterations in APM were present, non-HPV-HNSCC tended to have more mutations in HLA Class I alleles, while HPV-HNSCC tended to have more mutations in β2M, which were truncating mutations or homozygous deletions (Figure 1). We found that HLA Class I mutations were associated with decreased OS in both HPV-HNSCCs (specifically HLA-A, log rank p = 0.00252) and non-HPV-HNSCCs (specifically HLA-C, log rank p = 0.0328), as shown in Figure 2.

We then evaluated whether the APM gene mutation frequency correlated with the frequency of cell surface protein expression of HLA Class I and II in HNSCCs. HLA Class I protein expression was present in 85.7% (66 of 77) of HNSCCs, whereas HLA Class II protein expression was present in 41.3% (31 of 75) of HNSCCs. PD-L1 was expressed within the tumor microenvironment of 84.4% (54 of 64) of HNSCCs and was associated with HLA Class I protein expression (Fisher’s exact p = 0.035), which was most significant in HPV-HNSCCs (Fisher’s Exact p = 0.008). PD-L1 expression did not correlate with prognosis to standard of care treatment in HNSCC. However, in HPV-HNSCC patients, PD-L1 positivity demonstrated a trend towards improved overall survival (Log rank p = 0.0674). Furthermore, a trend emerged where the combination of HPV status and HLA Class II expression can predict PD-L1 status within the tumor microenvironment (p = 0.08). 

Figure 1 Spectrum of mutations in antigen presenting machinery observed in non-HPV-HNSCC (A) and HPV-HNSCC (B). The mutation frequency of β2M was markedly higher in HPV-HNSCC.

Figure 2 Kaplan-Meier curves of overall survival by presence of mutations in HLA-A in HPV-HNSCC (A) and in HLA-C in non-HPV-HNSCC (B). Mutations in HLA-A in HPV-HNSCC and mutations in HLA-C in non-HPV-HNSCC are associated with decreased OS. The red line indicates tumors with mutations present, while the blue line indicates tumors with no mutations present.