# A Phase 2 study of a Checkpoint Inhibitor in Men with Progressive Metastatic Castrate Resistant Prostate Cancer Characterized by a Mismatch Repair Deficiency or Biallelic CDK12 Inactivation

> **NIH VA I01** · VA GREATER LOS ANGELES HEALTHCARE SYSTEM · 2022 · —

## Abstract

Alternative treatment approaches are urgently needed to improve quantity and quality of life of patients
with metastatic castration resistant prostate cancer (mCRPC). Immunotherapy in the form of checkpoint inhibition
has yielded auspicious results in many cancers, including lung cancer, melanoma, renal cell carcinoma, and
lymphoma amongst several others. The frequency of tumor-specific somatic mutations and hence neoantigen
formation strongly predicts for objective response to checkpoint inhibitors (CPIs). Prostate cancer is typified by
a relatively low mutational burden, so it is not surprising CPIs for mCRPC are largely ineffective, although
occasional responses have been observed. Over the last few years, mismatch repair deficiency (dMMR) and
biallelic inactivation of CDK12 (CDK12-/-) have been observed in a small subset of mCRPC patients. Importantly,
these genetic lesions have been associated with increased mutational burden due to increased point mutations
in the case of dMMR and heightened formation of focal tandem duplications in the case of CDK12-/-. Accordingly,
dMMR or CDK12-/- tumors are expected to be sensitive to CPIs.
 The frequency of dMMR has been underestimated due to poor sensitivity of detection assays, which can
fail to detect allelic inactivation of MMR genes. Moreover, neither the frequency of dMMR or CDK12-/- nor the
response to checkpoint inhibition or dMMR or CDK12-/- tumors has been studied amongst Veterans. The
relevance of this knowledge gap amongst Veterans is highlighted by key demographic differences in Veterans
compared to the US population at large, which could in principle affect the frequencies of dMMR and CDK12-/-
as well as response to checkpoint inhibition. Using a sensitive next generation sequencing platform and
bioanalytic tool to detect microsatellite instability, a surrogate for dMMR, we predict that we can detect dMMR or
CDK12-/- in at least 15% of Veterans. Furthermore, we hypothesize that Veterans with dMMR or CDK12-/- will
exhibit a high response rate to checkpoint inhibition. We propose three aims:
1. Identify the frequency of dMMR and CDK12-/- as determined by NGS and a sensitive analytic tool for MSI
 detection amongst Veterans with mCRPC. The MSI bioanalysis, known as mSINGS (microsatellite instability
 by next generation sequencing), as well as targeted gene sequencing inclusive of CDK12 are built-in
 components of OncoPlex, one of the CLIA-certified NGS platforms being implemented within the VA
 Precision Oncology Program Cancer of the Prostate (POPCAP) network.
2. Perform an open label phase 2 clinical trial of pembrolizumab, an anti-PD1 CPI, to determine the efficacy of
 pembrolizumab amongst dMMR and CDK12-/- mCRPC Veterans who have received prior AR signaling
 ≥
 inhibitors. The primary endpoint will be response rate, defined as a composite of: objective response rate by
 iRECIST 1.1, PSA50 at 12 weeks, radiographic progression free survival at 6 months.
3. Perform exploratory analyses to...

## Key facts

- **NIH application ID:** 10417047
- **Project number:** 5I01CX002006-03
- **Recipient organization:** VA GREATER LOS ANGELES HEALTHCARE SYSTEM
- **Principal Investigator:** MATTHEW B RETTIG
- **Activity code:** I01 (R01, R21, SBIR, etc.)
- **Funding institute:** VA
- **Fiscal year:** 2022
- **Award amount:** —
- **Award type:** 5
- **Project period:** 2020-04-01 → 2024-03-31

## Primary source

NIH RePORTER: https://reporter.nih.gov/project-details/10417047

## Citation

> US National Institutes of Health, RePORTER application 10417047, A Phase 2 study of a Checkpoint Inhibitor in Men with Progressive Metastatic Castrate Resistant Prostate Cancer Characterized by a Mismatch Repair Deficiency or Biallelic CDK12 Inactivation (5I01CX002006-03). Retrieved via AI Analytics 2026-05-22 from https://api.ai-analytics.org/grant/nih/10417047. Licensed CC0.

---

*[NIH grants dataset](/datasets/nih-grants) · CC0 1.0*