Hematologic Malignancies Arising in Patients with Germ Cell Tumors: Secondary Somatic Differentiation of Hematopoietic Malignancies from Germ Cell Precursors

Genomic analyses have recently illuminated our understanding of therapy-associated myeloid neoplasms in patients receiving therapy for other cancers. One of the most intriguing relationships between solid tumors and myeloid neoplasms involves a unique clinical entity of patients with germ cell tumors (GCT) and myeloid neoplasms. One in 17 patients with primary mediastinal germ cell tumor (PMGCT) develops a hematologic malignancy (most commonly AML, MDS, or histiocytosis) and the median survival in such patients is poor at only 5 months. Intriguingly, the presence of isochromosome 12p [i(12p)], a clonal marker common to GCTs but absent in de novo myeloid neoplasms, has been demonstrated as shared across GCTs and myeloid neoplasms in such individuals. While these data suggest a clonal relationship between the two, the exact nature of the shared origin is unknown. There are two competing hypotheses to explain this: (1) an embryonic progenitor with the capacity to differentiate into germ cell and hematopoietic lineages harbors the initiating genetic alterations and drives development of both malignancies or (2) the leukemia is derived directly from GCTs with the capacity to differentiate into hematopoietic lineages. To trace the clonal evolution of these seemingly unrelated cancer types and identify recurrent genomic lesions in leukemias present in GCT patients, we applied whole exome sequencing (WES), targeted genomic analyses, and/or RNA-seq to leukemias, GCTs, and germline DNA in a series of patients with myeloid neoplasms and concurrent GCTs.

We collected 12 patients with GCT and synchronously or metachronously occurring myeloid neoplasms (8 AML, 5 MDS/CMML, 2 histiocytic sarcoma (some had >1 hematologic malignancy)) with an average of 4 months between the two diagnoses. Consistent with prior reports, all were young men (median age 26) with PMGCT and nonseminomatous histology and a 3-month median survival from leukemia diagnosis (Fig. A). In each case, at least one copy number alteration or coding mutation was shared across the GCT and hematopoietic neoplasm, demonstrating the shared origin of both lesions. For example, half of the patients (6/12) carried i(12p) in both the GCT and hematopoietic neoplasm. In the i(12p) negative cases, somatic genetic alterations identified in the GCT were also found in the leukemia. The most common genomic alterations in leukemias beyond i(12p) included mutations activating RAS-PI3K-AKT signaling (including PTEN, RAS and PI3K isoform mutations) or inactivating TP53 (Fig. B). The only exception was a testicular-only GCT patient who developed clonally distinct acute promyelocytic leukemia; however, further analysis identified this as a chemotherapy-induced neoplasm with the PML-RARa breakpoint mapped to an etoposide sensitive area and this patient was not counted amongst the 12 cases.

We next traced the evolutionary history of clonally related GCTs and leukemias based on cancer cell fraction of all coding mutations and copy number alterations using WES of DNA from each tumor type and finger nails. In each instance, we identified clonal evolution of the hematopoietic malignancies from antecedent precursors within the GCT. To illustrate this, a 19-year-old male developed successive diagnoses of histiocytic sarcoma, CMML, and AML within 18 months of GCT diagnosis. Lineage tracing by WES of each of these four individual cancers revealed that all four were clonally related, and the histiocytic sarcoma, CMML, and AML were all derived from the GCT with a common precursor giving rise to the three hematopoietic malignancies (Fig. C-D). Moreover, the histiocytic sarcoma evolved separately from CMML/AML in this patient, where the AML represented leukemic transformation from the CMML.

These data conclusively demonstrate that myeloid neoplasms developing in patients with PMGCT represent secondary somatic differentiation of a hematologic progenitor from totipotent aberrant cells that are present in the GCT. Thus, GCT-associated leukemias have a unique ontogeny from de novo and/or secondary myeloid neoplasms, which originate from progenitors within the bone marrow. As such, GCT-associated leukemias have characteristic genomic alterations hallmarked by frequent i(12p) in combination with mutations activating RAS-PI3K-AKT signaling and inactivating TP53, and these patients do poorly even when treated with aggressive contemporary chemotherapy.

Figure

View largeDownload slide

Figure

View largeDownload slide

Close modal