Does PD-1/PD-L1 Inhibition Really Work in Prostate Cancer?

Early attempts at checkpoint inhibition in prostate cancer seemed to offer little promise.  The initial trial with CTLA-4 inhibition with ipilumumab was performed in patients with metastatic castration-resistant prostate cancer (mCRPC) in the post-docetaxel setting and resulted in a HR 0.85, 95% CI 0.72-1.00; p=0.053.1 

Although close, this was considered a negative trial.  However, it was felt that the presence of patients with visceral metastases in that trial had a negative interaction with ipilumumab treatment effect.  As a result, the results of a pre-docetaxel randomized phase 3 trial with ipilumumab had high expectations, as patients with visceral metastasis were excluded.2   

Unfortunately, that trial also yielded no survival benefit.  PD-1 inhibition was also felt to have little activity in prostate cancer, since prostate cancer exhibits a rather low mutational spectrum.3   An early attempt with nivolumab, treated 17 men with mCRPC, and zero patients had a significant response to therapy.4  Many in the field felt the discouraging results were due to the “non-inflamed” phenotype of prostate cancer with low PD-L1 expression on both tumor cells and few tumor infiltrating lymphocytes.5

More recent work found PD-L1 and PD-L2 expression to be increased on circulating dendritic cells after patient exposure to enzalutamide, and higher expression correlated with lower PSA declines in response to enzalutamide.6   As a result, Dr. Julie Graff at Oregon Health and Science University designed a trial adding 4 cycles of pembrolizumab to patients progressing on enzalutamide, as a quick and simple test of activity in this setting.  Early data have shown 5 of 27 (19%) treated patients have a confirmed and sustained PSA response, and some patients have also had impressive RECIST soft tissue responses as well.7,8  Although, the patients in this trial have not all been molecularly characterized, it is doubtful that microsatellite instability (MSI) or hypermutation are present in all the pembrolizumab responding patients.  For patients who are MSI high or who harbor hypermutation, it is possible that PD-1 or PD-L1 inhibition alone may have significant antitumor effect,9 as evidenced by the recent FDA approval of pembrolizumab for patients with any tumor that is MSI high.  Yet, more work must be done to confirm early reports that MSI occurs in 3-12% of patients with mCRPC.10,11  It is possible that certain histologic subtypes, such as ductal prostate carcinoma, may harbor a higher frequency of MSI high tumors.12  The single agent efficacy of PD-L1 inhibition alone is being tested in mCRPC patients with MSI in a trial with durvalumab (Trial 1). 

This exciting early data has led to multiple combination therapeutic strategies currently being tested in clinical trials.  To further extend the theme of PD-1/PD-L1 inhibition combination with enzalutamide, there are two current trials starting pembrolizumab (Trial 2) or atezolizumab (Trial 3) up front at the same time enzalutamide is initiated.  The pembrolizumab trial, termed KEYNOTE 365, also has a cohort in combination with docetaxel chemotherapy (Trial 2).  Other approaches are evaluating the phenomenon of abscopal effect by combining either pembrolizumab (Trial 4) or atezolizumab (Trial 5) with radium-223 to facilitate neoantigen release.  There are also ongoing logical approaches to prime the immune system with vaccine-based therapies in combination with a PD-1 or PD-L1 antibody.  These approaches include combining sipuleucel-T with atezolizumab (Trial 6), pTVG-HP plasmid DNA vaccine with pembrolizumab (Trial 7) or PROSTVAC with nivolumab and/or ipilumumab (Trial 8).  The approach of combining multiple immune checkpoint inhibitors will likely facilitate higher response rates.  Tremilumumab and durvalumab are being combined in mCRPC trials (Trial 9), one restricted only to a chemotherapy naïve mCRPC population (Trial 10).   Ipilumumab is also being combined with nivolumab in multiple mCRPC disease states (Trial 11).  Finally, with the excitement surrounding DNA repair alterations10,13 and PARP inhibition14 in mCRPC, combinations of durvalumab with olaparib (Trial 12) as well as pembrolizumab with olaparib (Trial 2) are ongoing.

The future of PD-1 and PD-L1 inhibition in prostate cancer will depend heavily upon appropriate patient selection either by identification of predictive biomarkers or ideal clinical disease states.  Although there is some single-agent activity in mCRPC, current approaches to identify multiple rational biologic combinations is warranted to facilitate higher response rates and improve overall outcomes.  Our field should embrace and support this approach by directing patient accruals to these and other commonly-themed clinical trials.

Written by: Evan Yu, MD

Highlighted Trials:

  1. Durvalumab for MSI high mCRPC  
  2. Pembrolizumab combination with enzalutamide, docetaxel or olaparib 
  3. Atezolizumab combination with enzalutamide 
  4. Pembrolizumab combination with radium-223 
  5. Atezolizumab combination with radium-223 
  6. Atezolimab combination with sipuleucel-T 
  7. Pembrolizumab combination with TVG-HP plasmid DNA vaccine 
  8. Nivolumab and/or Ipilumumab combination with PROSTVAC 
  9. Durvalumab combination with tremelimumab 
  10. Durvalumab combination with tremelimumab for chemotherapy naïve mCRPC 
  11. Nivolumab combination with ipilumumab Nivolumab combination with ipilumumab 
  12. Durvalumab combination with olaparib and/or cediranib 
References
  1. Kwon ED et al.  Lancet Oncology 2014; 15:700-12.
  2. Beer TM et al.  J Clin Oncol 2017; 35:40-7.
  3. Lawrence MS et al.  Nature 2013; 499:214-8.
  4. Topalian SL et al.  N Engl J Med 2012; 366:2443-54.
  5. Martin AM et al.  Prostate Cancer Prostatic Dis 2015; 18:325-32.
  6. Bishop JL et al.  Oncotarget 2015; 6:234-42.
  7. Graff JN et al.  European Society of Medical Oncology Congress 2016; Abstract 719O.
  8. Graff JN et al.  Oncotarget 2016; 7:52810-7.
  9. Le T et al.  N Engl J Med 2015; 372:2509-20.
  10. Robinson D et al.  Cell 2015; 161:1215-28.
  11. Pritchard CC et al.  Nat Commun 2014; 5:4988.
  12. Schweizer MT et al.  Oncotarget 2016; 7:82504-10.
  13. Pritchard CC et al.  N Engl J Med 2016; 375:443-53.
  14. Mateo J et al.  N Engl J Med 2015; 373:1697-708