Functional and genetic screening of acute myeloid leukemia associated with mediastinal germ cell tumor identifies MEK inhibitor as an active clinical agent

Mediastinal germ cell tumors (GCT) are capable of malignant transformation into tumors of endodermal, mesodermal, and ectodermal origin. With regard to hematologic malignancies, GCT are most frequently associated with acute myeloid leukemia (often with megakaryoblastic differentiation); but myelodysplastic syndromes, malignant histiocytosis, myeloproliferative neoplasms, and acute lymphoblastic leukemia have all been described [1‐7]. Identification of shared cytogenetic abnormalities in the GCT and hematologic malignancy (most commonly isochromosome 12p) indicate a clonal relationship and common precursor [8].

The prognosis for GCT with a concurrent hematologic malignancy is poor, as the disease is usually refractory to conventional chemotherapy and patients are unable to attain a clinical remission [9]. For those that do attain a remission, relapse without consolidative allogeneic stem cell transplant is inevitable, and successful outcomes are limited to two cases in which patients underwent allogeneic hematopoietic stem cell transplant early in the first complete remission [1, 10]. Given these dismal outcomes, novel therapeutic approaches are needed.

A 23-year-old male presented with fevers, hemoptysis, and an 18-cm anterior mediastinal mass (Fig. 1a). A computed tomography-guided core biopsy of the mass contained mostly necrotic tissue with rare atypical cells that were partially positive for alpha-fetoprotein (AFP), glypican, and pancytokeratin by immunohistochemistry. Serum AFP was elevated at >60,000 ng/mL, leading to a presumptive diagnosis of a mediastinal germ cell tumor. The patient was treated with four cycles of VIP (vinblastine, ifosfamide, etoposide) with concomitant decrease in serum AFP to 300 ng/mL. Prior to resection of the residual mediastinal mass (4 months after the original diagnosis), a complete blood count (CBC) demonstrated elevation of the WBC to 55 k/μL with 28 % circulating myeloid blasts. His lactate dehydrogenase (LDH) was markedly elevated at 17,481 U/L, although his AFP decreased to 211 ng/mL. A bone marrow biopsy demonstrated acute erythroid leukemia with 70–80 % markedly left-shifted erythroid precursors with dysplastic and megaloblastoid changes and myeloid blasts accounting for more than 20 % of non-erythroid cell population (Fig. 1b). Cytogenetics showed a complex clone (Fig. 2d). Targeted massively parallel sequencing analysis also identified TP53 p.G245S and NRAS p.G12C point mutations at 87 and 86 % allele frequency, respectively (see Additional file 1 and Additional file 2: Table S1 for details). As the patient’s mediastinal mass had not yet been resected, a regimen that would treat both his AML and his residual germ cell tumor (containing platinum-based therapy) was desired. For this reason, he was treated with HAEP (high-dose cytarabine, etoposide, and cisplatin), previously reported as an effective salvage therapy for AML [11]. Following induction, the patient achieved morphologic complete remission (CR) of his AML; however, low-level residual disease was identified by cytogenetics (3 of 20 cells with the prior abnormal clone) and massively parallel sequencing analysis (1 % TP53 and 2 % NRAS mutant allele frequencies).

He was treated with a single cycle of HAEP consolidation therapy, after which he underwent resection of the mediastinal mass. Histopathologic examination demonstrated mature teratoma with foci of rhabdomyosarcoma and angiosarcoma (Fig. 1c, d). Targeted Sanger-based DNA sequencing analysis identified the same TP53 and NRAS point mutations present in the acute erythroid leukemia, confirming a shared clonal origin.

Six weeks after resection of the mediastinal mass, a restaging bone marrow biopsy demonstrated relapsed disease with a morphologically abnormal erythroid series, 5–10 % myeloid blasts, and persistent TP53 and NRAS mutations at 87 and 86 % allele frequency (Fig. 2e). Cytogenetics identified the original abnormal clone and the emergence of two additional clones (Fig. 2d). Since the patient was not fit for re-induction chemotherapy due to recent surgery, his leukemic cells were isolated and tested against a panel of small molecule kinase inhibitors to identify in vitro drug sensitivity. Trametinib was identified as one of the most potent in vitro inhibitors of leukemic cell viability (Fig. 2a). In vitro small inhibitory RNA (siRNA) studies further identified that silencing of NRAS had a profound inhibitory effect on blast viability (Fig. 2b), consistent with his known activating NRAS mutation (for full list of silenced genes please see Additional file 3: Table S2). The patient began treatment with single-agent trametinib therapy obtained on a compassionate use basis and had an immediate reduction in his peripheral blasts after a week of therapy (Fig. 2c).

A bone marrow biopsy after 1 month of trametinib revealed a slight reduction in the blast count (4–5 %) but persistent erythroid dysplasia. Although the AML remained stable, multiple cytogenetically complex clones were identified and the TP53 and NRAS mutant allele frequency increased to 96 and 97 %, respectively (Fig. 2e). After 3 months of trametinib monotherapy, the patient developed pneumonia and Staphylococcus aureus bacteremia. A bone marrow biopsy at that time demonstrated 8 % blasts and erythroid dysplasia. Trametinib was held twice due to infections, and during these intervals, there was an increase in peripheral blasts, which resolved following resumption of trametinib (Fig. 2c). A bone marrow biopsy after 5 months of trametinib therapy demonstrated stability of his AML with 6 % blasts and erythroid dysplasia. Karyotyping and repeat sequencing analysis demonstrated continued clonal evolution, with multiple cytogenetically complex subclones (Fig. 2d). There was a persistent high-level NRAS and TP53 mutant allele frequency but two additional WT1 point mutations were also identified at allele frequencies of 20 and 15 %, further indicating clonal evolution (Fig. 2e). Repeat testing confirmed that the leukemic clone remained sensitive to trametinib in vitro (data not shown). However, imaging studies performed in preparation for allogeneic stem cell transplantation demonstrated a 3-cm pulmonary nodule and serum AFP also rose to 12 ng/dL, consistent with relapse of metastatic GCT. Given his poor prognosis, the patient opted for palliative care and died a month later with frank relapse of his AML after trametinib discontinuation.

In conclusion, we report a case of metasynchronous mediastinal GCT and AML in a young male patient with identical TP53 and NRAS mutations in both malignancies. In vitro functional assays were used to confirm the importance of the NRAS mutation to leukemic blast proliferation and the ability of trametinib to inhibit growth. Trametinib therapy led to sustained disease control in vivo over 6 months following failure of conventional cytotoxic chemotherapy. This case highlights the power of in vitro functional assays to identify novel targeted agents to treat rare and aggressive malignancies and supports the exploratory use of small molecule kinase inhibitors in the treatment of AML, particularly in rare presentations of AML where therapeutic options are limited and outcomes are poor.