Genomic landscape of mycosis fungoides

Abstract Introduction: Cutaneous T-cell lymphomas (CTCLs) are a rare, clinically heterogenous group of extranodal lymphomas that arise from mature skin-resident T cells. Mycosis fungoides (MF) is the most common primary CTCL, accounting for 50% of all cases. MF occurs on a spectrum of indolent to more advanced and aggressive forms, encompassing large-cell transformation (LCT) and the leukemic variant, Sezary syndrome (SS). Currently, there are no defining genomic and molecular features for the various subtypes of MF. We proposed to investigate disease biology of MF on a genomic and transcriptome level with the goal to shed light on possible genomic markers of prognosis and treatment targets for the various subtypes of MF. Methods: All samples were collected and underwent expert pathology review as part of the Atlas of Blood Cancer Genomes consortium. All nucleic acid extractions and sequencing library preparations were performed with Duoseq (Research Kit EPXv3, Data Driven Bioscience, Durham, NC) following the manufacturer's protocol. All libraries were sequenced on the Illumina platform and analyzed according to manufacturer recommendations. Whole exome, targeted panel, and transcriptome sequencing was performed on a cohort of 67 patients with MF from 15 different institutions. Gene set enrichment analysis (GSEA) were performed on biological pathways from the Molecular Signatures Database (MSigDB). For determination of ancestry, we used iAdmix, a tool that analyzes a panel of single nucleotide polymorphisms (SNPs) in the DNA sequencing data and uses a maximum likelihood estimation algorithm accounting for the relative presence of these SNPs in the Hapmap3 populations to determine most likely ancestry profiles. The genomic determination of ancestry, correlated very highly with self-reported race. Results: Median age at diagnosis was 61 years. Majority of the patients were male (56%) and self-reported as white race (57%), while 34% self-reported as black race. Sixty-four percent of patients had Stage III-IV disease consistent with advanced CTLC. Further, 31% of the patients had LCT during the course of their disease, and 52% were alive at the time of last follow-up. Copy number variants in regions containing CDKN2A, TP53, RB1 and ZEB1 were the most common genomic alterations, concordant with findings from a previous large SS study. LCT patients had worse progression-free survival compared to their non-LCT counterparts (p 0.01). Gene enrichment studies showed that cell cycle pathways (p=1x 10-10) and genes involved in the G2M checkpoint (p=1x 10-10) were more upregulated in LCT samples compared to those without LCT. Conversely, inflammatory response pathways and NFkB signaling pathways were more upregulated in those without LCT (p = 1 x 10-10and 1.21 x 10-5, respectively). Black patients also showed worse overall survival compared to white patients, with the known driver gene ARID1A more frequently mutated in black compared to white patients (p 0.05). Similar statistically significant results were noted when classification was made by SNP based ancestry. No statistically significant association of LCT with CD30 status was observed in this study. Conclusions: Our study is the largest multi-center genomic study of MF. We show that MF with LCT was associated with poor progression free survival and upregulation of cell cycle pathways and G2M checkpoint. Furthermore, black patients had poor overall survival compared to their white counterparts, with increased incidence of ARID1A mutations. These results enhance our understanding of the heterogeneity of MF and highlight potential targetable pathways for patients with aggressive forms of the disease.