The Genomic and Transcriptomic Landscape of Systemic Mastocytosis

Background and Aim

Systemic mastocytosis (SM) is characterized by a heterogeneous spectrum of disease variants with different clinical features and prognosis. In indolent SM (ISM), the clinical course is usually stable and patients (pts) have a normal life expectancy. In contrast, aggressive SM (ASM) and mast cell leukemia (MCL) are characterized by severe organ damage and are fatal forms with short survival times. All forms of SM share the presence of activating mutations in the KIT tyrosine kinase (most frequently, the D816V mutation), which suggests that cooperating events might be responsible for the profoundly different clinical presentation and outcome of advanced SM (ASM and MCL) as compared to ISM. However, no robust studies using genome-wide approaches to scan for additional molecular aberrations in ASM and MCL have been published to date. This prompted us to use exome and RNA sequencing and single nucleotide polymorphism (SNP)-arrays to search for novel (and potentially druggable) molecular lesions.

Methods

Seven pts with MCL, 6 pts with ASM and 2 pts with ISM were studied using whole exome sequencing (WES) and single nucleotide polymorphism (SNP)-arrays. WES (80x) was performed on a Hiseq 2500 (Illumina). SNP-arrays were done using Cytoscan HD Arrays (Affymetrix). Matched normal/mast cell DNA was analyzed in all but 2 archival MCL cases for whom no germline DNA source was available. RNA sequencing was also performed in 3/7 MCL and 1/6 ASM cases for whom high quality RNA was available. Data analysis and integration was performed according to in-house optimized pipelines.

Results

SNP-arrays found arm-level copy number (CN) or copy-neutral loss-of-heterozygosity (CN-LOH) events in 5/7 MCL pts (loss of 7p, 8p, 9q, 12p, 13q and CN-LOH at 4q, 18q and 21q). In contrast, ASM and ISM showed only focal CN and CN-LOH events. Overall, a median of 28 (range, 20-78) and 77 (range, 55-132) submicroscopic CN and CN-LOH events, respectively, were detected. The number of events did not significant differ between ISM, ASM and MCL. A total of 54, 102 and 1835 genes were significantly found to be recurrently involved in loss, gain and CN-LOH events, respectively. The average number of nonsynonymous mutations identified by WES was 15 in ISM and 35 in ASM and MCL. Signature analysis showed that the majority of sequence changes were C>T transitions and C>A transversions. Among genes recurrently mutated in ≥2 patients or concurrently identified by WES and SNP-arrays or known to be associated with cancer, some had already been reported to be mutated in SM: ASXL1, TET2, CBL, RUNX1, NRAS, IDH1, SRSF2, SF3B1. Alterations were also found in genes not previously implicated in SM, including RUNX3, MLL2, MLL3, CDC27, TP53BP1, CCND3, NCOR2, BCORL1, ATM, WRN, ARID1B, ARID3B, KDM1B, ARID4A, SETD1A, SETD1B, PRDM1. Pathway analysis showed enrichment of genetic lesions affecting the following cellular processes: PI3K/Akt and MAPK pathways, calcium signaling, chromatin modification, DNA methylation and DNA damage repair. RNA-seq results, filtered against those of a set of normal bone marrow samples, revealed a high number of RNA chimeras involving two adjacent genes in the same transcription orientation and invariably joining the pre-last exon of the upstream gene to the second exon of the downstream gene - a fusion pattern suggestive of intergenic splicing following transcription read-through events. Deletions at the intergenic region was excluded by SNP-arrays for all the predicted transcription-induced chimeras. In addition, in all the 4 pts analyzed, retention of some or all introns in mature transcripts was observed for some genes, including known tumor suppressors (TS) and other key genes like TP53, BAX, ERCC2, WRN, FANCD2, ATR; U2AF2, U2AF1, SRSF6.

Conclusions

• SNP-array and WES analysis of SM revealed a heterogeneous landscape of molecular alterations, although some key cellular pathways were found to be recurrently altered.

• More interestingly, RNA-seq uncovered perturbation of transcript elongation and splicing in all the advanced SM pts analyzed. Intron retention, in particular, seems to be a novel recurrent mechanism of TS gene inactivation at the posttranslational level in advanced SM. The mechanisms leading to intron retention and their cellular consequences deserve further investigation.

Supported by AIRC (project code 16996), progetto Regione-Università 2010-12 (L. Bolondi), FP7 NGS-PTL project.