Introduction:

A subset of patients with germ cell tumors (GCTs) develop platinum resistance disease resulting in inferior survival outomes. The mechanisms driving resistance are complex and not yet fully understood. We sought to evaluate the genomic and transcriptomic landscapes in primary and metastatic GCTs to uncover genetic factors that drive cisplatin resistance in GCTs.

Methods:

GCTs (N = 138) were analyzed by next-generation sequencing of DNA (592-gene or whole exome) and RNA (whole transcriptome). Prevalence was calculated for pathogenic mutations and high copy number amplifications (CNA ³ 6 copies). Primary (N = 65, primary), lymph node metastasis (N=14, lymph), and non-lymph metastatic lesions (N = 59, metastasis) were defined based on the tissue site relative to known primary site. An independent genitourinary pathologist reviewed H&E-stained slides and designated tumors as chemo-naïve (PreC, N = 66) or post-chemotherapy (PostC, N = 17) based on absence or presence of morphologic evidence of treatment related changes. Platinum Resistant Alterations (PRAs, defined from prior literature) included KRAS, TP53, and KIT mutations, and MDM2 amplification. A transcriptomic signature associated with platinum sensitivity (Platinum sensitivity score: PSS, high score suggests increased platinum sensitivity) was applied. Mann-Whitney U and tests were applied as appropriate, with P-values adjusted for multiple comparisons. 

Results:

Sixty-five primary tumors were sequenced, including 7 intracranial primary tumors, 5 mediastinal primary tumors, 16 ovarian primary tumors, and 37 testicular primary tumors.  Seventy-three samples were obtained from metastatic sites (including lymph nodes) based on clinician annotation.  This included two from bone, 11 from the brain, six from the liver, five from the lung, 14 from lymph nodes, four from the mediastinum, 12 from the peritoneum, one from the spermatic cord, and 18 from non-bone/liver/brain visceral sites (ie. kidney, small intestine, connective tissue, etc). Compared to non-lymph node and lymph node metastases, patients with biopsy from primary tumors had a significantly lower median age at the time of biopsy (24 vs 34 and 41 years, respectively, p < 0.001), were more frequently female (29.2% vs 13.6% and 7.1%, respectively, p < 0.42), and were more frequently chemo-naïve (92.1% vs 73.5% and 54.5%, respectively, p < 0.01) (Table 1). The genomic variation landscape of GCTs was sparse, and predominantly made up of recurrent genetic variants previously associated with driver mutation or chemotherapy resistance (KIT-Mt, KRAS-Mt, TP53-Mt, PTEN-Mt, KRAS-Amp and MDM2-Amp). As a result, our investigation focused on these genes (Fig 2). KIT mutations were observed in 14.5% of primary versus 1.8% of met and 0% of lymph cases.  TP53 mutations were identified in 10% of primary GCTs versus 17% of met and 16.7% of lymph cases. MDM2 CNAs were similar between primary, met and lymph cohorts (3.7% vs 1.9% vs 8.3%) (all p > 0.05).  PRA positive PreC GCTs had a significantly lower average PSS score compared to PRA negative tumors (-5.44 vs 2.17 arbitrary units, p = 0.02). KRAS-Amp alterations were often concurrent with another PRA and were generally associated with a positive PSS, which may suggest a potential link between MAPK signaling and platinum sensitivity. Although rare, both MDM2-Amp tumors had very low PSS.

Conclusion:

The prevalence of PRAs does not significantly vary by site. Lower PSS scores in chemo-naïve tumors were associated with PRAs, suggesting a potential mechanism for platinum-based chemotherapy resistance.

Funding: N/A