Frequent Alterations of Driver Genes in Chromosome X and Their Clinical Relevance in Extranodal NK/T-Cell Lymphoma

Extranodal NK/T-cell lymphoma (ENKTCL) is a rare and aggressive mature T-cell malignancy which is invariably related to Epstein-Barr virus (EBV) infection with male predominance. Although many genetic studies of ENKTCL have revealed frequent alterations in several genes (e.g., DDX3X and TP53), they have mainly focused on mutational changes. Therefore, the landscape of somatic alterations, including copy number alterations (CNAs) and structural variations (SVs), in ENKTCL remains elusive.

Here we performed whole-exome sequencing (WES) of 66 ENKTCL patients, followed by targeted-capture sequencing (targeted-seq) of 147 lymphoma-associated genes in 168 patients from Japanese (n=64) and French (n=104) institutions. Combined analysis of WES and targeted-seq identified 25 significantly mutated genes using dNdScv. These genes included six novel genes, namely HLA-B (8%), HLA-C (6%) ROBO1 (5%), CD58 (4%), POT1 (3%), and MAP2K1 (2%). Besides them, recurrent mutations (in ≥ 3 cases) were observed in seven genes previously reported to be drivers in ENKTCL. Thus, 32 genes were recurrently mutated in our cohort. Consistent with our previous report, we detected SVs truncating the 3′-untranslated region of CD274 in 14% of cases. We also identified significant focal amplification (n = 1) and deletions (n = 10) and arm-level gains (n = 12) and losses (n = 7) by GISTIC2.0. Focal CNAs included 15 driver genes, including CD274 amplifications (18%) and CDKN2A (17%) and ARID1A (11%) deletions.

When mutations, CNAs, and SVs were combined, a total of 34 genes were recurrently altered in our cohort. CD274 (24%) was the most frequently altered, followed by TP53 (20%), CDKN2A (19%), ARID1A (15%), HLA-A (15%), BCOR (15%), and MSN (15%). The frequently altered pathways were immune surveillance (46%), epigenetic regulation (43%), and tumor suppressors (41%). Notably, several driver genes (MSN, BCOR, DDX3X, and KDM6A) were present on chromosome (chr) X. Among arm-level CNAs, chrX losses were only observed in females (39% and 37% in Xp and Xq loss, respectively). In addition, the proportion of mutations affecting these drivers on chrX was higher in males and females with chrX loss than in females without chrX loss (43% vs 23%, p = 0.02). These results suggest that driver genes on chrX are coordinately involved in ENKTCL pathogenesis and may explain its male predominance (58% in our cohort).

Among driver genes on chrX, the most frequently altered gene was MSN (15%), which was affected by loss-of-function mutations and/or deletions and specifically altered in ENKTCL among lymphoma subtypes. To assess the functional consequence of MSN alterations, we investigated the effect of MSN knockout (KO) or overexpression (OE) in NK/T cell lines. MSN OE diminished the proliferation of a MSN-null cell line (YT1), while cell growth was enhanced by CRISPR-mediated MSN KO in MSN wild-type cell lines. In line with this, transcriptomic analysis with RNA-seq demonstrated the enrichment of proliferation-related signatures in MSN-null cells. Next, we analyzed the hematopoietic system in Msn KO mice and found that Msn disruption caused an increase of a subset (CD11b^-CD27⁺) of NK cells but a decrease in a more mature subset (CD11b⁺CD27⁺) in the bone marrow. In addition, cell-cycle progression was increased in these Msn KO NK subsets. Transcriptomic analysis showed the significant enrichment of proliferation-related signatures and several signaling pathway signatures.

Finally, we integrated 34 drivers and 19 significant arm-level CNAs using consensus clustering based on non-negative matrix factorization and identified two molecular groups. In group 1 (n=33), tumor suppressors (CDKN2A and TP53), epigenetic regulators (KDM6A and BCOR), and JAK-STAT pathway genes (JAK3 and STAT5B) were more frequently altered, whereas group 2 (n=30) was enriched with MSN and CD274 alterations. Clinically, group 1 was significantly associated with a more advanced stage and worse prognosis than group 2, irrespective of the PINK score.

In summary, we have illustrated the genetic landscape of ENKTCL, which was shaped by frequent disruptions of driver genes on chrX. Particularly, we characterized the role of loss-of-function MSN alterations in NK cell function in vitro and in vivo. In addition, we proposed a novel molecular classification with biological and clinical relevance. Our findings can help improve diagnostic and therapeutic strategies in ENKTCL.