MicroRNA-371-373 cluster and methylome analysis suggests that a subset of “somatic-type” malignancies arising in germ cell tumors may originate in yolk sac tumor components

Background: Somatic-type malignancies (SM) arising in germ cell tumors (GCTs) represent aggressive and chemoresistant neoplasms frequently occurring as metastases after chemotherapy. Historically, SM have been interpreted as originating in teratoma; however, recent observations suggest that a subset may derive from yolk sac tumor. In this study, we evaluate the relationship between conventional histologic types of GCT and SM of germ cell origin by assessing expression of miR-371~373 and genomic methylation. Methods: A cohort of 97 samples (41 SM and 38 conventional GCTs, including paired conventional GCT and SMs from individual tumors) were assessed for miR-371~373 expression (RT-qPCR) and genomic DNA methylation (clinically validated assay). Results: Expression of miR-371~373 was higher in conventional non-teratomatous GCT than in SM (considered as a single category containing all histologic subtypes). However, miR-371~373 expression was heterogeneous among the latter, with significantly higher levels in sarcomatoid yolk sac tumor (SYST) and glandular neoplasms than in other SMs. Genomic DNA methylation analysis showed that SMs (considered as a single category) did not form a discrete cluster. Instead, they grouped into subclusters that intercalated with conventional GCTs. Analysis of paired conventional GCT and SM or “SM-like” components (i.e., components with SM histology that did not meet size criteria) from individual tumors demonstrated a clear separation according to histology, suggesting that epigenetic processes play a significant role in the transformation of GCT to SM. Discussion: A subset of SMs shows molecular similarities to (and may derive from) yolk sac tumor, with SYST and glandular yolk sac tumor possibly representing intermediate phenotypes. The transformation of conventional GCT into SM seems to be driven largely by epigenetic mechanisms, suggesting that trials of targeted treatment with epidrugs may be beneficial in selected patients. FFPE tissues were manually dissected, and DNA was extracted. The DNA was then processed for hybridization and fluorescence staining using the Infinium MethylationEPIC (850k) BeadChip array (Illumina, San Diego, USA) according to the manufacturer’s instructions. The arrays were scanned with the Illumina iScan microarray scanner. Raw idat files were analyzed using R version 4.3.2, and raw intensities were processed with the Bioconductor R package Minfi (23). Each sample’s quality was assessed by mean detection p-value (p < 0.05). Samples were normalized using quantile normalization, and batch effects were corrected using the removeBatchEffect function in the R package limma (24). Probes that failed in one or more samples (detection p < 0.01) and probes with single-nucleotide polymorphisms were removed. Beta-values were calculated using the default offset value of 100 recommended for Illumina assays. Top variable probes were identified by calculating the variance of beta values across all samples, and the top 1000 probes were selected to plot the unsupervised heatmap using the R package ComplexHeatmap (25). PCA was performed based on the top 1000 variable probes using the prcomp function in R.