MLH1 Loss in Primary Prostate Cancer

PURPOSE

Among mismatch repair-deficient (MMRd) prostate cancers, loss of MLH1 is relatively uncommon and few cases have been reported in detail.

METHODS

Here, we describe the molecular features of two cases of primary prostate cancer with MLH1 loss detected by immunohistochemistry, and in one case, confirmed via transcriptomic profiling.

RESULTS

Both cases were microsatellite stable on standard polymerase chain reaction (PCR)–based microsatellite instability (MSI) testing, but showed evidence of MSI on a newer PCR-based long mononucleotide repeat (LMR) assay and by next-generation sequencing. Germline testing was negative for Lynch syndrome–associated mutations in both cases. Targeted or whole-exome tumor sequencing using multiple commercial/academic platforms (Foundation, Tempus, JHU, and UW-OncoPlex) showed modestly elevated, though variable, tumor mutation burden estimates (2.3-10 mutations/Mb) consistent with MMRd, but without identifiable pathogenic single-nucleotide or indel mutations in MLH1. Copy-number analysis confirmed biallelic MLH1 loss in one case and monoallelic MLH1 loss in the second case, without evidence of MLH1 promoter hypermethylation in either. The second patient was treated with single-agent pembrolizumab and demonstrated a short-lived prostate-specific antigen response.

CONCLUSION

These cases highlight the challenges in identifying MLH1-deficient prostate cancers using standard MSI testing and commercial sequencing panels, and support the utility of immunohistochemical assays and LMR- or sequencing-based MSI testing for detection of MMRd prostate cancers.

INTRODUCTION

Despite the rarity of mismatch repair deficiency (MMRd) in prostate cancer,^1-5 there is great interest in biomarkers that accurately identify such tumors because of the importance of germline screening in these patients, and their potential to respond favorably to immunotherapy. Although the standard polymerase chain reaction (PCR)–based microsatellite instability (MSI) detection assay^6,7 lacks sensitivity for noncolorectal MMRd cancers,^3,8,9 a newer panel of long mononucleotide repeat (LMR) MSI markers shows improved performance, but is not yet widely used.^9,10 Next-generation sequencing (NGS) is the current gold standard for MMRd confirmation in prostate cancer, but is costly and may lack sensitivity if tumor purity is low or intronic MMR gene sequence is not captured.² Finally, immunohistochemistry (IHC) is a sensitive, cheap, and widely available assay, frequently used for initial screening of MMRd, but may fail to detect missense mutations and even certain protein-truncating mutations.¹¹

CONTEXT

• Key Objective

• To our knowledge, for the first time, we describe in detail the performance of mismatch repair (MMR) biomarkers for prostate tumors with MLH1 alterations.

• Knowledge Generated

• We demonstrate low sensitivity and interplatform variability among many standard microsatellite instability (MSI) testing and commercial sequencing panels for identification of MMR deficient (MMRd) prostate cancers. Our data support using immunohistochemistry assays for initial MMRd screening in prostate cancer, paired with—or followed by—long mononucleotide repeat MSI testing and targeted tumor sequencing panels.

• Relevance

• Rapid and accurate diagnosis of MMRd prostate tumors will help identify patients likely to respond to the US Food and Drug Administration–approved PD-1 inhibitor pembrolizumab.

Herein, we present a comprehensive molecular evaluation of two prostate tumors with MLH1/PMS2 protein loss detected by IHC screening. These cases demonstrate the low sensitivity and interplatform variability of many current commercial molecular assays in the detection of MMRd prostate cancers with an intermediate phenotype and highlight the overall difficulty of identifying pathogenic MMRd alterations in this setting.

METHODS

Patient Samples

Written informed consent was obtained from both patients for publication of this case report. Patients were consented using the Johns Hopkins Authorization for Release of Health Information in a Case Report.

Mismatch Repair Protein Immunohistochemistry and Interpretation

We performed the MMR protein IHC staining as previously described.³ See the Data Supplement.

Targeted NGS

Targeted NGS data were collected in our cases across four commercial/academic platforms: (1) FoundationOne CDx (Cambridge, MA)¹²; (2) UW-OncoPlex¹³; (3) Tempus Labs (Chicago, IL) whole-exome sequencing (WES)¹⁴; and (4) the Johns Hopkins Solid Tumor Panel version 6.⁹

Estimation of TMB, FSB, and FSP

This was performed as previously described.^15-17 See the Data Supplement.

MSI Analysis System Version 1.2

MSI Analysis System Version 1.2 (Promega, Madison, WI) was performed on neoplastic and normal tissues on the basis of the manufacturer's directions.⁶ See the Data Supplement.

LMR MSI Panel

The LMR MSI Panel (Promega) was performed on neoplastic and normal tissues as previously described.⁹ See the Data Supplement.

NanoString IO360 RNA Expression Profiling

Formalin-fixed, paraffin-embedded prostate tissue RNA expression was profiled using the NanoString IO360 Immune Panel on the Nanostring GeoMx (Nanostring Technologies, Seattle, WA). See the Data Supplement.

Single-Nucleotide Polymorphism Microarray

Tumor DNA samples were run on the Illumina Infinium II single-nucleotide polymorphism (SNP) array with 300K markers and analyzed as described previously.¹⁸

RESULTS

Case Presentation 1

A 69-year-old man presented with a prostate-specific antigen (PSA) level of 6.9 ng/mL and a palpable nodule on digital rectal examination. Magnetic resonance imaging showed a large Prostate Imaging Reporting & Data System (PI-RADS) 5 lesion at the right apex with compression of the prostatic urethra. A targeted transrectal biopsy revealed low-risk Grade Group 1 prostatic adenocarcinoma (three cores). Soon after the biopsy, the patient went into urinary retention and was referred to Johns Hopkins with a PSA level of 7.9 ng/mL. A perineal biopsy demonstrated Grade Group 5 (Gleason score 4 + 5 = 9) prostatic adenocarcinoma with extensive lymphocytic inflammation, prompting MMR IHC analyses, which revealed loss of MLH1 and its binding partner PMS2, with intact MSH2 and MSH6 (Fig 1A). A family history of prostate cancer in the patient's father and maternal grandfather was noted; however, germline genetic testing via a commercial panel (Invitae, San Francisco, CA) revealed no pathogenic mutations.

FIG 1.

Molecular phenotype of case 1. (A) H&E and immunohistochemical staining of case 1 for MMR proteins. Tumor cells (arrow) show loss of MLH1 and PMS2 nuclear expression, while adjacent immune cells (arrowhead) provide an internal positive control. Scale bar = 100 μm. (B) LMR MSI analysis for case 1. (C) Spatial transcriptomic analysis of MMR gene expression in case 1. (D) Copy-number analysis for case 1 via UW-OncoPlex sequencing of SNPs. Each chromosome is depicted in a different color. H&E, hematoxylin and eosin; LMR, long mononucleotide repeat; MMR, mismatch repair; MSI, microsatellite instability; SMR, short mononucleotide repeat; SNP, single-nucleotide polymorphism.

Given the abnormal MMR IHC, the biopsy was also sent for an NGS panel (FoundationOne CDx), which revealed an elevated tumor mutation burden (TMB) of 10 Muts/Mb and additional pathogenic alterations (Table 1), without alterations in any of the four main MMR genes. The specimen was designated as microsatellite stable (MSS) on the basis of NGS assessment. When the reported pathogenic alterations were examined for frameshift mutations indicative of MSI, the tumor frameshift mutation burden (FSB) was 2.5 mutations/Mb and the frameshift mutation proportion (FSP) was 0.11.

TABLE 1.

Genomic Characteristics of Case 1

The patient elected to pursue a radical prostatectomy (RP) and was enrolled in a phase II clinical trial evaluating the safety, antitumor effect, and immunogenicity of neoadjuvant enoblituzumab (MGA271), an anti–B7-H3 (CD276) antibody in men with intermediate- and high-risk localized prostate cancer before RP (ClinicalTrials.gov identifier: NCT02923180). The RP specimen showed Grade Group 5 prostatic adenocarcinoma (Gleason 4 + 5 = 9) and pathologic stage pT3a Nx R0. Further evaluation of MMR deficiency using a standard five-marker MSI-PCR assay (MSI Analysis System v1.2; Promega) was interpreted as MSS. We next tested the RP sample using the LMR MSI Analysis System (Promega), consisting of four short mononucleotide repeat markers from the MSI V1.2 panel and four LMR markers.^9,10 Using this assay, the RP tumor was designated as MSI-H (5/8) with only one shift in the small microsatellite repeats (NR-21), but clear shifts in all four of the long microsatellite repeats (BAT60, BAT59, BAT56, and BAT52; Fig 1B).

As part of the clinical trial, the RP tumor sample further underwent WES (Tempus Labs) and spatial transcriptomic analysis. WES revealed a TMB of 2.3 mutations/Mb and multiple additional pathogenic alterations, although none involving the MMR genes (Table 1). The tumor's FSB was 1 mutation/Mb; the FSP was 0.29. By targeted RNA transcriptome analysis using 760 RNA target probes (Nanostring GeoMx DSP gene expression assay), the RP sample showed clear loss of MLH1 gene expression, with intact MSH2, MSH6, and PMS2, consistent with an underlying alteration resulting in reduced MLH1 mRNA abundance (Fig 1C).

As a final attempt to discern the underlying mechanism for loss of MLH1 and PMS2, we leveraged the University of Washington-OncoPlex NGS assay,¹³ which includes capture of nonrepetitive, noncoding sequence of the MMR genes in its panel to better assess intronic mutations, copy-number alterations, and genomic rearrangements. This assay showed a TMB of eight mutations/Mb with a FSB of three mutations/Mb, and a FSP of 0.4, and confirmed MSI-H status using the NGS MSI assay (mSINGS, 22.5% unstable microsatellites; cutoff for MSI-H designation is 20% on this assay¹⁹; Table 1)^8,13; moreover, clear evidence of focal homozygous MLH1 deletion was found. Homozygous loss was further confirmed by SNP b-allele frequency analysis (Fig 1D). MLH1 promotor methylation testing was negative.

Case Presentation 2

A 50-year-old man with no known family history of prostate cancer initially presented with a PSA level of 4.1 ng/mL. Systematic 12-core biopsy showed Grade Group 4 (Gleason score 4 + 4 = 8) prostatic adenocarcinoma (eight cores involved). The patient underwent a RP in 2013, which showed Grade Group 5 prostate adenocarcinoma (Gleason 4 + 5 = 9) and a pathologic stage of pT3a N0 R0. One month postoperatively, his PSA level was 0.1 ng/mL, but rose to 0.4 ng/mL after 12 months of follow-up. After a negative bone scan, the patient underwent salvage radiotherapy at a total dose of 70.2 Gy in 39 fractions with concurrent androgen-deprivation therapy (ADT) with 6 months of leuprolide. Fifteen months after radiation was completed, he was found to have biochemical recurrence, with a PSA level of 2.8 ng/mL and a PSA doubling time of 3 months. He was enrolled in a biomarker-unselected phase II clinical trial evaluating PSA response rates to olaparib, a poly (ADP-ribose) polymerase inhibitor, in men with biochemically recurrent prostate cancer after RP with a PSA doubling time of <6 months (ClinicalTrials.gov identifier: NCT03047135).

As part of the trial, the RP tumor sample was sent for an NGS panel (FoundationOne CDx; Table 2), which showed a TMB of six mutations/Mb, an FSB of 2.4 mutations/Mb, and an FSP of 0.13. The specimen was designated MSS on the basis of NGS assessment. No pathogenic alterations were reported in any MMR genes. However, a hotspot RNF43 frameshift mutation in a homopolymeric run was reported²⁰ (Table 2), which suggested the possibility of MMRd. MMR IHC analysis confirmed MMRd and demonstrated MLH1 and PMS2 loss (Fig 2A). Germline genetic testing via a commercial panel (ColorGenomics, Burlingame, CA) revealed no evidence of pathogenic germline mutations.

TABLE 2.

Genomic Characteristics of Case 2

FIG 2.

Molecular phenotype of case 2. (A) H&E and immunohistochemical staining of case 2 for MMR proteins. Tumor cells (arrow) show loss of MLH1 and PMS2 nuclear expression, while adjacent stromal cells (arrowhead) provide an internal positive control. Scale bar = 100 μm. (B) Serum PSA levels for case 1 before and after initiation of pembrolizumab monotherapy and ADT. (C) LMR MSI analysis for case 2. (D) SNP microarray analysis of chromosome 3 region containing MLH1 in case 2. (E) Copy-number analysis for case 2 via UW-OncoPlex sequencing of SNPs. Each chromosome is depicted in a different color. ADT, androgen deprivation therapy; H&E, hematoxylin and eosin; LMR, long mononucleotide repeat; MMR, mismatch repair; MSI, microsatellite instability; PSA, prostate-specific antigen; SNP, single-nucleotide polymorphism.

After 5 months of olaparib treatment without PSA response, the patient was placed on pembrolizumab without concurrent ADT, because of the MMRd IHC results. His PSA level decreased from 10.8 to 6.7 ng/mL on pembrolizumab, but the response was short-lived (Fig 2B). After 5 months of pembrolizumab monotherapy, he was started on ADT concurrent with pembrolizumab. After 2 years of pembrolizumab and ADT, he developed PSA progression and was treated with ADT and darolutamide, an androgen receptor inhibitor, during which he experienced a >90% decline in PSA.

Given the discordance between the sequencing and the IHC screening results, further evaluation to confirm MMRd was sought. A standard five-marker MSI-PCR assay (MSI Analysis System v1.2; Promega) found the patient's tumor sample to be MSS, while the LMR MSI Analysis System (Promega) designated it MSI-L (1/8), showing only a subtle shoulder-pattern shift in one microsatellite repeat (BAT-52; Fig 2C).

To further evaluate for the etiology of MMRd, the RP specimen was sent for panel-based NGS in the Johns Hopkins Molecular Pathology laboratory. On this assay, the TMB was 3.5 mutations/Mb and multiple pathogenic alterations were identified, although none involving the MMR genes. The tumor's FSB was 7.0 mutations/Mb; the FSP was 0.32, with a high number of frameshift alterations identified in homopolymeric repeats (Table 2). To evaluate MLH1 copy-number alteration status, we used SNP microarray analysis, which identified MLH1 loss of heterozygosity implying a monoallelic deletion (Fig 2D).

Since no second hits to MLH1 were uncovered by either sequencing assay, and the standard five-marker panel MSI testing proved ambiguous, we turned to the UW-OncoPlex NGS assay, which showed a TMB of 10 mutations/Mb, with an FSB of three mutations/Mb and an FSP of 0.25. This case was found to be MSI-indeterminate via the NGS MSI assay (mSINGS, 17% unstable microsatellites; cutoff for MSI-H is 20%), as well as to harbor probable shallow MLH1 copy-number loss (Fig 2E). Although numerous somatic mutations were detected, none involved MMR genes (Table 2). MLH1 promotor methylation testing was negative.

DISCUSSION

Herein, we present two highly characterized MMRd prostate cancers with MLH1 protein loss, a phenotype that is relatively rare, comprising around 10% of MMRd prostate cancers.^2,4,5,21 Because IHC is sensitive to underlying copy-number alterations and gene rearrangements even when tumor content is low, contemporary studies demonstrate retained MMR protein expression in <10% of MSI-H tumors.¹¹ The present cases—exhibiting MLH1 and PMS2 loss—confirm the utility of the MMR IHC assay for screening in prostate cancer, but highlight the challenges of confirmatory molecular assays in this setting.

Low sensitivity of conventional MSI testing in prostate cancer may be due to numerous factors, including tumor-specific patterns of microsatellite mutation,^22,23 low tumor neoplastic cellularity in the tested sample, the choice of metastatic or primary sample,⁵ or lower relative net proliferation rate in this tumor type, resulting in fewer cell divisions over which to accrue instability. This challenge is illustrated by the present cases, both of which were designated as MSS using the five-marker MSI PCR assay. Notably, however, we confirm that the LMR PCR-based MSI assay showed increased sensitivity in our MLH1-deficient tumors,⁹ where case 1 was found to be MSI-H (4/4 LMR markers shifted) and case 2 was MSI-L on the LMR panel.

Even by sequencing, MMRd prostate cancer often has a mild mutational phenotype, with a relatively lower TMB compared with MMRd colorectal or endometrial tumors.^2,4,5,21 Similarly, among gastrointestinal MMRd cancers, tumors with loss of MLH1/PMS2 coexpression by IHC are associated with a lower TMB compared with tumors with loss of MSH2/MSH6 co-expression.²⁴ These findings are consistent with our MLH1-deficient prostate cancer cases, both of which showed relatively low TMB (estimated at 10 or less mutations/Mb, varying by assay). This intermediate phenotype may also explain the short-lived response to pembrolizumab in case 2, since higher TMB correlates with improved clinical responses to immune checkpoint blockade across and within tumor types.^25-27

Importantly, our cases also highlight other challenges associated with using TMB estimates to screen for MMRd cancers. Although WES is the gold standard for TMB estimation, most commercially available sequencing platforms use panel-based sequencing, which may overestimate absolute TMB because of over-representation of cancer driver genes. Panel size, depth of sequencing, and variant filtering methodology can also dramatically affect TMB estimation.²⁸ This is illustrated by these cases, which were assessed across four commercial/academic platforms and showed a wide range of TMB estimates (3.5-10 and 2.3-10 mutations/Mb, respectively). Consistent with this variability, recent studies indicate that TMB alone may not be an optimal predictor of response to pembrolizumab.^29,30 Moreover, an elevated TMB may be due to environmental^31,32 or gene defects other than MMRd, such as POLD1, POLE, and MUTYH.³³

Because MMRd is associated with a 100- to 1,000-fold increase in base substitutions and frameshifts in small homopolymeric repeats,¹⁴ FSB or FSP may be better predictors of immunotherapy response in MMRd prostate cancer than TMB.³⁴ This raises the question of whether these indicators may also be better biomarkers of underlying MMRd than TMB in this tumor type. FSB ranged widely from one to seven mutations among our two MLH1-deficient cases, while FSP ranged from 0.11 to 0.4. These values trend higher than—but still overlap with—the distribution of FSB (median: 0, range: 0-3.75) and FSP (median: 0, range: 0-0.22) in a small recent series of non-MMRd prostate cancers that underwent commercial sequencing at our institution (LAS, personal communication, June, 2022). Taken together, these data suggest that FSB standardization and FSP standardization are also critical before these biomarkers can be compared across sequencing platforms and between patient groups.

It is notable that sequencing the same samples from the same patients across multiple commercial platforms failed to identify copy-number alterations in MLH1 that were detected by the UW-OncoPlex NGS assay (which assesses intronic sequence around the MMR genes) or by SNP microarray analysis. In cases where either tumor purity is low or intronic sequence is not captured, panel-based sequencing may be insensitive for detecting copy-number alterations and gene rearrangements. These shortcomings may be particularly relevant for MMRd screening in prostate cancer since prostate tumors may have relatively more stromal contamination compared with colorectal tumor samples and MMRd alterations in the prostate frequently involve copy-number alterations. Although case 2 showed only monoallelic MLH1 loss, there was no evidence of a second genomic hit in this tumor to explain loss of MLH1 protein and MSI by LMR and sequencing assays. Although MLH1 promoter hypermethylation is a common cause of MLH1 loss in colorectal and endometrial carcinomas, it has not been reported to occur in prostate cancer using contemporary clinical-grade assays and we did not find evidence of MLH1 hypermethylation in case 2. Importantly, we cannot exclude a germline or somatic translocation disrupting MLH1, as similar findings have been previously described in Lynch syndrome³⁵ but require whole-genome sequencing or fluorescence in situ hybridization assays to confirm.

In conclusion, these cases highlight the challenges associated with many of the current commercial molecular assays for identifying MMRd prostate cancers. Consistent with recent guidelines from the College of American Pathologists,³⁶ our data support the utility of IHC assays for initial MMRd screening in prostate cancer, paired with (or followed by) LMR MSI testing and carefully designed targeted NGS tumor sequencing panels, such as UW-OncoPlex.

AUTHOR CONTRIBUTIONS

Conception and design: Sanaz Nourmohammadi Abadchi, Emmanuel S. Antonarakis, Eugene Shenderov, Tamara L. Lotan

Financial support: Colin C. Pritchard

Provision of study materials or patients: Emmanuel S. Antonarakis

Collection and assembly of data: Sanaz Nourmohammadi Abadchi, Emmanuel S. Antonarakis, Colin C. Pritchard, Eric Q. Konnick, Eugene Shenderov, Tamara L. Lotan

Data analysis and interpretation: Sanaz Nourmohammadi Abadchi, Laura A. Sena, Colin C. Pritchard, James R. Eshleman, Eric Q. Konnick, Stephen J. Salipante, Eugene Shenderov, Tamara L. Lotan

Manuscript writing: All authors

Final approval of manuscript: All authors

Accountable for all aspects of the work: All authors

AUTHORS' DISCLOSURES OF POTENTIAL CONFLICTS OF INTEREST

The following represents disclosure information provided by authors of this manuscript. All relationships are considered compensated unless otherwise noted. Relationships are self-held unless noted. I = Immediate Family Member, Inst = My Institution. Relationships may not relate to the subject matter of this manuscript. For more information about ASCO's conflict of interest policy, please refer to www.asco.org/rwc or ascopubs.org/po/author-center.

Open Payments is a public database containing information reported by companies about payments made to US-licensed physicians (Open Payments).