Predictive value of PD-L1 and MHC-II expression for cHL clinical outcomes

On 2 February 2018, Margaretha G.M. Roemer, from the Dana-Farber Cancer Institute, Boston, MA, and colleagues, published online ahead of print in the Journal of Clinical Oncology a study assessing the predictive value of genetic and protein-expression alterations in 9p24.1, programmed death ligand-1 (PD-L1), major histocompatibility complex class I (MHC-I), MHC-II and β2-microglobulin (β2M), for clinical outcomes in relapsed/refractory (R/R) classical Hodgkin lymphoma (cHL) patients.

To assess these parameters the authors extracted Hodgkin and Reed–Sternberg (HRS) tumor cells from biopsies of certain R/R cHL patient cohorts, from the phase II CheckMate 205 trial (NCT02181738), only after receipt of nivolumab (anti-PD-1) treatment. Genetic abnormalities and protein expression were evaluated with fluorescence in-situ hybridization and immunohistochemistry, respectively. The primary endpoints of this study included progression-free survival (PFS) and clinical responses after nivolumab therapy.

Patient characteristics
  • N = 180 patients with R/R cHL from cohorts B and C from CheckMate 205
    • Cohort B (n =80) received single-agent brentuximab vedotin (BV) after ASCT relapse (ASCT→BV)
    • Cohort C (n = 100) received either:
      • BV before ASCT (BV→ASCT) (n = 33)
      • BV after ASCT relapse (ASCT→BV) (n = 58) or
      • BV before and after ASCT relapse (BV→ASCT→BV) (n = 9)
  • Patients were then treated with nivolumab (3mg/kg) for 2 weeks until disease progression or unacceptable toxicity
  • Time from ASCT to nivolumab treatment (median; range):
  • Cohort B (40, 2–228 months)
  • Cohort C (21, 3–204 months)
Key findings
9p24.1/PD-L1 alterations:
  • All evaluable tumor specimens (n = 99) had 9p24.1 gene copy number changes:
    • Copy gain: 60% (n = 59)
    • Amplification: 27% (n = 27)
    • Polysomy: 10% (n = 10)
  • From those specimens, n = 97 were further analyzed for PD-L1 protein expression:
  • There was significant association between PD-L1 protein expression and magnitude of 9p24.1 genetic alterations (P = 0.001)
  • Following nivolumab: R/R patients were more likely to have lower-level 9p24.1 alterations and less PD-L1 expression on HRS cells, than responders (CR, PR or stable disease) (P = 0.006; P = 0.047)
  • Shorter PFS was associated with nivolumab R/R patients with lower-level 9p24.1 alterations and less PD-L1 protein on HRS cells (P < 0.001; P = 0.026)
  • 1 genetic alterations were associated with PFS in patients with either ≤ 12 months or > 12 months between ASCT and nivolumab
  • PD-L1 protein levels were associated with PFS only in patients with > 12 months between ASCT and nivolumab and not in those with a ≤ 12-month interval
MHC-I, MHC-II and β2M expression alterations:
  • N = 72 available specimens
  • β2M expression in HRS cells:
    • Negative, 71% (n = 51)
    • Decreased, 22% (n = 16)
    • Positive, 7% (n = 5)
  • MCH-I expression in HRS cells:
    • Negative, 64% (n = 46)
    • Decreased, 29% (n = 21)
    • Positive, 7% (n = 5)
  • MCH-II expression in HRS cells:
    • Negative, 29% (n = 21)
    • Decreased, 32% (n = 23)
    • Positive, 39% (n = 28)
  • Highly significant association between MCH-I and β2M expression (P < 0.001)
  • No association between MCH-II and MCH-I or β2M (P = 0.645; P = 0.5764)
MHC-I and MHC-II and clinical outcome:
  • N = 11/12 (92%) of patients achieving CR following nivolumab, had tumors negative for β2M and MCH-I
  • No association between PFS and β2M or MCH-I expression on HRS cells
  • N = 11/12 (92%) of patients achieving CR following nivolumab, had tumors with MCH-II expression (67% positive and 25% decreased)
  • MCH-II expression was not predictive for PFS in patients receiving nivolumab ≤ 12 months after ASCT
  • MCH-II expression was predictive for PFS in patients receiving nivolumab > 12 months after ASCT

The main findings from this study demonstrated that predictive factors for superior PFS for R/R cHL patients are: (a) higher level of 9p24.1 gene copy gain, (b) increased PD-L1 protein expression on HRS cell and (c) MHC-II expression on HRS cells, in patients with nivolumab > 12 months after ASCT. MCH-II expression was also predictive for CR. According to the authors, the main limitation of this study was the limited availability of tumor biopsy specimens from the CheckMate 205 trial. These findings indicate the ‘biomarker value’ of PD-L1 and MCH-II expression for clinical outcomes in R/R cHL and present evidence in support of an alternative MHC-II-associated molecular mechanism of action for nivolumab in cHL.

  1. Roemer et al. Major Histocompatibility Complex Class II and Programmed Death Ligand 1 Expression Predict Outcome After Programmed Death 1 Blockade in Classic Hodgkin Lymphoma. Journal of Clinical Oncology. 2018 Feb 2. doi: 10.1200/JCO.2017.77.3994. [Epub ahead of print]