Findings

Intragenic deletions of the dystrophin-encoding and muscular dystrophy-associated DMD gene have been recently described in many common human mesenchymal tumors, including gastrointestinal stromal tumor (GIST), rhabdomyosarcoma (RMS) and leiomyosarcoma (LMS) which show myogenic differentiation [1]. In particular, DMD deletions were found in 19 of 29 GIST tumors, 3 of 4 RMS tumors and 3 of 7 LMS tumors. Moreover, DMD deletions and their protein expression were not found in non-myogenic cancers, or in benign counterparts of GIST, RMS and LMS. In GIST, dystrophin is downregulated in metastatic GIST and primary high-risk GIST, but not in low-risk GIST, which implies this downregulation is a late event in progression of this disease. Finally, dystrophin inhibits myogenic sarcoma cell migration, invasion, anchorage independence and invadopodia formation; and when deregulated, dystrophin restoration inhibits invasiveness and migration in sarcoma cell lines. These data validate dystrophin as a tumor suppressor and likely anti-metastatic factor. In light of these findings, we evaluated DMD copy number and gene expression levels in our series of GIST patients who had already been studied with wide genome assays, to investigate more fully the correlation between dystrophin status and disease annotations.

Patient selection and tumor sample collection

Dystrophin status was evaluated using already-available data from wide genome assays done on tumor specimens collected during surgery and immediately frozen. We included samples from 29 patients with mutant KIT/PDGFRA GIST and 6 with wild-type (WT) KIT/PDGFRA GIST. Among the KIT/PDGFRA WT GIST group, 4 cases were SDH deficient and 2 cases were quadruple KIT/PDGFRA/SDH/BRAF-KRAS-NF1 WT. Out of all 35 patients, 19 had already been reported [2]. Table 1 shows patients’ clinical and molecular data. The genomic analysis study was approved by local Ethical Committee.

Table 1 Clinical and molecular data of the patients included in the study
Full size table

Copy number analysis

Genomic DNA was extracted with QiaAmp DNA mini kit (Qiagen, Milan, Italy), labelled and hybridized to SNP array Genome Wide SNP 6.0 (Affymetrix) following manufacturer’s instructions. Quality control was performed by Contrast QC and MAPD calculation. Copy number analysis was performed by Genotyping Console and visualized with Chromosome Analysis Suite Software (Affymetrix). Hidden Markov Model algorithm was used to detect amplified and deleted segments.

RNA sequencing

Whole-transcriptome sequencing was performed on RNA isolated from fresh-frozen tumor tissue with the RNeasy spin-column method (Qiagen). Whole-transcriptome RNA libraries were prepared in accordance with Illumina’s TruSeq RNA Sample Prep v2 protocol (Illumina, San Diego, California). Paired-end libraries were sequenced at 2 × 75bp read length using Illumina Sequencing by synthesis (SBS) technology. Averages of 85 million reads per sample were analyzed. Mapping on the human reference genome was done with TopHat/BowTie software, while expression level of the DMD gene was expressed as number of mapped reads.

Results and discussion

The genome-wide analysis of our series highlighted a recurrent intragenic deletion on chromosome X for the DMD gene, which codes for human dystrophin. Nine of the 29 KIT/PDGFRA mutant GIST (31%) showed DMD gene deletions (Figure 1A), which were focal and intragenic in 8 cases, and involved loss of a whole chromosome in one case (GIST_188; Figure 1B). None of the 6 KIT/PDGFRA WT GIST samples had DMD alterations.

Figure 1
figure 1

DMD deletions and gene expression in KIT/PDGFRA mutant IST samples. (A) Signal Log2 Ratio of copy number data from SNP 6.0 arrays on the X chromosome, showing focal losses and deletion of 1 entire chromosome arm. (B) Graphical representation of sizes of the genomic losses of DMD gene, located on the reverse X chromosome strand. (C) DMD expression showed as average number of mapped reads from RNA sequencing of the samples that carry DMD deletions. All cases retain the expression of the short isoform starting from exon 63.

Full size image

As DMD is an X-linked gene, deletions were nullizygous in males and heterozygous in females. All focal events involved the 5′ portion of the gene, with the region between exon 2 and exon 7 as the most recurrently involved, and an average deletion size of 770 Mb.

RNA-sequencing performed on 3 out of 9 cases with deletions showed that the genomic losses abrogated expression of the largest DMD transcript, while preserving expression of the short isoform, with a transcription start site at exon 63 (Figure 1C).

Patients with KIT/PDGFRA mutant GIST tended to have higher frequency of DMD deletions in more advanced cases, as DMD loss was seen only in the 5 patients with metastasis, whereas 18 out of the 20 patients with localized disease wild-type DMD (Table 1; P = 0.0004, Fisher exact test).

Our study confirms the presence of DMD deletions in KIT/PDGFRA mutant GIST. In particular we also observed that this molecular event is associated with more advanced clinical disease such as metastatic tumors. Although these findings are quite preliminary they suggest potential therapeutic strategies that target and restore DMD function in treating metastatic GIST. Larger studies are necessary to correlate DMD status and specific mutations of KIT/PDGFRA receptors to explore novel therapies for GIST that present primary resistance or that initially responds to imatinib but later progresses [3]. As is well known, molecular mechanisms of secondary resistance are various and heterogeneous, the most common being the acquisition of secondary KIT/PDGFRA mutations or selection of sub-clones with resistant mutations [4]. In our series, of 7 patients who presented primary mutations in KIT exon 11, 2 also had secondary mutations, in KIT exon 17 or KIT exon 18. The 2 patients with mutations in PDGFRA had them in exon 18 D842V or exon 12. The small series does not permit any conclusive considerations on the correlation of DMD involvement with KIT/PDGFRA kinase genotype. But as the clinical use of sunitinib and regorafenib after imatinib failure does not completely cover the molecular landscape of the progressing GIST, a novel approach that targets dystrophin deregulation may have relevance in GIST treatment. Moreover, compared with KIT/PDGFRA mutant GIST, in our series all patients with KIT/PDGFRA WT GIST—including 4 who were SDH deficient and 2 who had quadruple KIT/PDGFRA/SDH/BRAF-KRAS-NF1 WT did not present dystrophin deregulation, even those with metastasis, which confirms once again that KIT/PDGFRA WT GIST should be considered a distinct disease in molecular background and clinical presentation [5].

Conclusion

In conclusion, deregulation of dystrophin seems to be associated with GIST progression and more data should be accumulated in order to define it as a therapeutic target.