Molecular Characterization of Testicular Mesothelioma and the Role of Asbestos as a Causative Factor

Cases of malignant mesothelioma of the TVT diagnosed between 1992 and 2021 were identified from the surgical pathology archives of South Australian Pathology, South Australia following ethics approval from the Central Adelaide Human Research Ethics Committee, Adelaide, South Australia, Australia (approval number 13440). Only cases with a definite diagnosis by current diagnostic guidelines and no previous or concurrent pleural or peritoneal mesothelioma were included in this study, and no cases presented here have been previously published.¹⁵  Formalin-fixed, paraffin-embedded (FFPE) tumor tissues from 3 cases were retrieved for molecular analysis. Areas of tumor tissue were identified on hematoxylin-eosin–stained sections by a pathologist. Genomic DNA fractions were extracted from 10-μm-thick slices of tumor tissue after macrodissection of the tumor tissue using a QIAamp DNA FFPE tissue extraction kit as per manufacturer’s instructions (QIAGEN, Hilden, Germany). DNA quality was determined using the Qubit 2.0 fluorometer (Thermo Fisher Scientific, Waltham, Massachusetts).

Immunohistochemistry for mouse monoclonal BAP1 C-4 antibody (sc-28383, 1:50 dilution; Santa Cruz Biotechnology, Dallas, Texas) and rabbit monoclonal MTAP antibody (EPR22570-76, ab254265; 1:200 dilution, Abcam) was performed on paraffin sections cut at 4-μm thickness. Labeling was carried out using Ventana BenchMark ULTRA, Automated IHC/ISH platform (Ventana Medical Systems, Orlo Valley, Arizona), as per the manufacturer’s instruction using validated clinical procedures approved by The National Association of Testing Authorities. Antigen retrieval was performed using Ventana reagents Cell Conditioning Solution 1 and Ventana Amplification Kit (760-080). Positive control tissue was included on every slide.

Whole-exome sequencing was performed at The Australian Cancer Research Foundation Cancer Genomics Facility using established quality-controlled platforms and workflows. The sequencing was performed on an Illumina Nextseq 500 on a high-output flow cell (Illumina, San Diego, California). The reads were dual-indexed: index1 was 18 bp, including an 8-bp sample barcode and 10-bp unique molecular identifier tag, and index2 was 8 bp, with a sample barcode only and no unique molecular identifier.

The resulting exome sequencing reads (FASTQ data files) were mapped to the human reference genome (hg19) using the Burrors Wheeler Aligner-MEM and the genome analysis toolkit. Somatic variant calling was done using a cohort of 6 callers: MuTect (v1.5), Seurat, SomaticSniper, Strelka, Varscan, and Shimmer. Single-nucleotide polymorphisms (SNPs) and indels were evaluated if they were identified by at least 3 callers.

We obtained pathology details and case histories for 11 cases of mesothelioma of the TVT diagnosed at Flinders Medical Centre between 1992 and 2021. However, 1 case of testicular mesothelioma with a prior history of pleural mesothelioma and 1 case of concurrent peritoneal and testicular mesothelioma were excluded because in those cases primary origin in the TVT could not be definitively established. The distribution of histologic subtypes was epithelioid (7 of 9 cases) and biphasic (2 of 9 cases) (Figure 1, A through D; Table). The median age at diagnosis was 74 years, and the mean age at diagnosis was 63 years. Of the cases available for molecular analysis, 2 were biphasic mesothelioma and 1 was epithelioid. The morphology of mesothelioma was indistinguishable from that at other sites. Definite asbestos exposure was identified in 7 individuals, with latency periods ranging between 24 and 72 years from first reported exposure to diagnosis of mesothelioma, and there was no information regarding presence or absence of asbestos exposure in 2 patients. Data on exposure history and pleural plaques was available for 7 patients, with only 1 patient having plaques, and with 4 patients having had occupational exposures. Survival data were available for 8 individuals, 5 of whom were still alive at the time of writing this paper. These patients were censored on March 1, 2022, the date of last follow-up. The other 3 individuals died 9 months, 5 years, and 5 years and 9 months following diagnosis.

All cases were labeled with mesothelial markers such as calretinin, CK5/6, AE1/3, WT1, or D2-40 and were negative for carcinoma markers such as CEA, TTF-1, BerEp4, PAX8, CDX2, PSA, or BG8, following the guidelines for mesothelioma diagnosis.¹⁵ 

A detailed breakdown of immunohistochemistry results is provided in Supplemental Table 1 (see supplemental digital content containing 2 tables at https://meridian.allenpress.com/aplm in the December 2023 table of contents). Where BAP1 and MTAP results were available, 4 out of 5 tumors were labeled positive for BAP; only 1 was BAP1-negative. MTAP was retained in all the 5 tumors.

Whole-exome sequencing was performed on normal and tumor FFPE tissue from the original diagnostic biopsy specimens of 3 patients: RM0001, RM0002, and RM0003 (Supplemental Table 2). The tumor cellularity of the testicular mesothelioma tissues was between 90% and 95% for each specimen. We achieved a mean sequencing read coverage of 86 738 906 reads for the normal tissue and 335 587 528 reads for the tumor tissue. Notably, RM0003 displayed a somatic inframe deletion and splice-region variant in NF2, with a variant allele frequency (VAF) of 0.362, which was predicted to have a moderate impact on protein function. RM0003 also showed 2 somatic missense variants in TP53 with VAFs of 0.282 and 0.364. Both missense variants in TP53 were predicted to have a moderate impact on protein function and have been previously reported in a wide range of cancers. RM0001 had a somatic SNP causing a premature stop codon in the NF2 gene with a VAF of 0.097, which was predicted to have a high impact. This variant has been previously reported in meningioma and schwannoma.^16,17  RM0002 showed a missense mutation in TRAF7, predicted to have a moderate impact; TRAF7 mutations are reported as novel tumor-suppressor genes in mesothelioma.¹⁸  A splice acceptor variant of CCDC137 was also seen, predictive of high impact. Descriptions of each mutation can be found in Supplemental Table 2.

The distribution of histologic subtypes with a predominance of the epithelioid subtype (7 cases) versus biphasic (2 cases) was similar to that of pleural or peritoneal mesothelioma and similar to that reported by others. The median and mean ages at diagnosis (74 and 64 years, respectively) were quite typical of mesothelioma, similar to the 72-year median age previously reported.¹⁹ 

Sequencing analysis has been previously exploited to uncover the genetic mutations of 2 cases of mesothelioma of the TVT. Zhang and colleagues²⁰  performed whole-genome sequencing in 4 successive tumor samples from a patient diagnosed with mesothelioma of the TVT with no reported history of asbestos exposure. Genomic analysis revealed a mutation in KIF25 in all samples, with mutations in OR2L8 and LRRFIP1 found in at least 3 samples. Later, Kowalik and colleagues²¹  performed targeted sequencing on 409 cancer-related genes in an 81-year-old man with mesothelioma of the TVT, revealing mutations in PARP1, RNF213, PAX8, KMT2C, MTRR, and KMT2C. The molecular abnormalities in these 2 cases differed from the profile of mutated genes described in pleural and peritoneal mesothelioma. For the first time we have identified cases of mesothelioma of the TVT that had somatic variants identified in genes also recurrently mutated in pleural and peritoneal mesothelioma, suggesting a shared etiology. We identified an NF2 insertion-deletion mutation in RM0003 and an SNP causing a premature stop codon in RM0001, which is frequently mutated in both pleural and peritoneal mesothelioma at frequencies of ∼19% and ∼23%, respectively.^14,22  The NF2 gene encodes NF2 (also known as Merlin), a protein involved in cell adhesion and proliferation. Inactivation of NF2, leading to Hippo pathway inactivation, has been described as a late clonal event in pleural mesothelioma.²³  The VAF for NF2 for RM0001 mutations was 0.097, which indicates that this event could be a subclonal event. The TP53 gene encodes the key tumor suppressor p53, which is involved in cell cycle regulation, among many other important roles. Two mutations in TP53 were identified in RM0003, which was observed in ∼8% of pleural mesotheliomas.²²  The BAP1 gene encodes a nuclear deubiquitinase, with functions in chromatin remodeling, DNA damage repair, cell cycle control, and differentiation; it is frequently deleted in both pleural and peritoneal mesotheliomas. RM0003 exhibited loss of BAP1 protein labeling as assessed by immunohistochemistry; however, the BAP1 gene sequence from this sample was wild-type, but copy number and potential loss of heterogeneity in the BAP1 gene could not be confirmed due to the quality of DNA extracted from FFPE tissues. TRAF7 mutations have been reported in pleural mesotheliomas²²  but not in diffuse peritoneal mesotheliomas, but are common in well-differentiated papillary mesothelioma.²⁴ 

With regard to the CCDC137 mutation, CCDC137, as the target gene of ABCA5, has been implicated in drug resistance in mesothelioma.²⁵ 

There is currently no mutational profile in mesothelioma that is linked to a certain causative agent: that is, no mutation that can confirm or exclude a causative role for asbestos. Other causes of TVT mesothelioma have been suggested, including a preexisting hydrocele, and 5 of the 9 cases presented here had a hydrocele, but a hydrocele is now regarded as an (early) symptom, rather than a cause of TVT mesotheliomas.²⁶ 

Regardless of other possible causes, the role of asbestos in the pathogenesis of mesothelioma has been unequivocally established and recognized as a causative factor. To consider mesothelioma of the TVT, it is imperative to understand the relationship between the TVT and the peritoneum. The testes descend to their final position around the ninth month of life in approximately 97% of males. At that point, the peritoneum forms a fold, resulting in the parietal and visceral layers of the TVT. The TVT represents an anatomic extension of the peritoneal membrane into the scrotum. The tunica vaginalis is part of the peritoneum, and the same causal effects contributing to causation of mesothelioma in the peritoneum apply to the tunica vaginalis. Asbestos exposure has been variably recorded as an associated risk factor in 30% to 67% of cases of mesothelioma of the TVT reported in the literature. For mesothelioma of the TVT, a preoperative diagnosis is rarely made and the true incidence of asbestos exposure in mesothelioma of the TVT is likely underestimated due to the absence of any mention of a history of asbestos exposure, especially for mesotheliomas at a very unusual primary site. This does not equate to no exposure. Recently Vimercati and colleagues⁹  performed a systematic literature review (289 cases between 1942 and 2018) and found that only 58.2% of cases commented on the patient’s history of asbestos exposure; for the rest of the cases these data were not available. The comprehensive registry-based study by Marinaccio et al²⁷  records 70% of cases as having asbestos exposure, very close to that for pleural mesothelioma in their database, and a recent case series that included medicolegal referral cases recorded asbestos exposure in 83% of 18 cases.¹⁹  The reporting on exposure to asbestos is clearly dependent on the data available, but absence of history does not equal absence of exposure. We obtained a detailed history of asbestos exposure in 7 of the 9 cases reported here.

Interestingly, 5 of the 6 cases for which the information was available had a history of or a concurrent hydrocele. Pleural effusion caused by mesothelioma in situ as a precursor of diffuse pleural mesothelioma is now well recognized²⁸  and has recently been described in the testis, suggesting that a similar mechanism may be applicable.²⁶ 

The minimum latency interval between first exposure to asbestos and diagnosis of malignant mesothelioma is often quoted as 10 years, with common latencies between 20 and 40 years.^29–31  Some studies have suggested shorter latencies for peritoneal mesothelioma, but in a study of 2644 mesothelioma cases with asbestos exposure history collected by the Italian mesothelioma register in the period of diagnosis 1993–2001, Marinaccio and colleagues²⁷  found a median latency of 42 years for males, with a minimum 21-year latency. The cases presented here had a latency period of 24–72 years (median, 50 years), which is in keeping with reported latencies for pleural and peritoneal mesothelioma^32,33  and those described for mesothelioma of the TVT.^3,19 

For the first time, we performed whole-exome sequencing of 3 cases of mesothelioma of the TVT with ascertained histories of asbestos exposure identifying aberrations in NF2, BAP1, and TP53. Taken together these data suggest at least some mesothelioma of the TVT can harbor genetic mutations in genes commonly mutated in pleural and peritoneal mesotheliomas, where asbestos has been unequivocally established as a causative factor. Further work will be required to more completely characterize the molecular landscape of testicular mesothelioma to facilitate a better understanding of the underlying pathogenesis.