ABSTRACT
Gastrointestinal stromal tumor (GIST) is the most commonly occurring mesenchymal neoplasm of the gastrointestinal tract, accounting for 80 percent of these tumors. GIST is highly unresponsive to standard chemotherapy, particularly in patients with advanced or metastatic disease. Recent molecular studies have shown that activating c-kit (KIT) mutations are detectable in a large proportion (>75%) of tumors, between 78% (1) and 89% (2). Approximately 30% of tumors without an identifiable KIT mutation exhibit PDGFRA mutations (3). Furthermore, the KIT mutations are heterogeneous, some being known to confer a relatively better prognosis than others (1). Gross cytogenetic abnormalities associated with GIST appear to be similar regardless of whether a KIT mutation is identified. The molecular genetic alterations associated with multistep GIST tumorigenesis, particularly those which confer intrinsic or acquired resistance to both standard as well as targeted therapeutic approaches, however, are not fully recognized. As an initial approach to identify chromosomal sites of candidate gene(s), which may predict overall clinical and biologic behavior of GISTs, as they relate to response to the specific therapeutic drug Gleevec, we analyzed six GIST samples using both conventional as well as array-based Comparative Genomic Hybridization (CGH). The common abnormalities detected by CGH in low and high grade tumors included loss of all or part of chromosome 14; an entire chromosome 14 was lost in four tumor samples, with the remaining two samples exhibiting loss of the 14q22-q32.3 region. Other recurrent abnormalities included loss of the 1p chromosomal region (four tumor samples), loss of part or entire chromosome 9 (only in metastatic tumors), loss of chromosomes 15 and 22, and a gain of the chromosome 3q region in three samples each. Array- based CGH performed using human BAC arrays (1400V11 Spectral Genomics Chip™) on the other hand, not only detected recurrent abnormalities of the chromosomes and chromosomal sites mentioned above, but also identified losses of additional chromosomal sites on chromosomes 6q and 13q. Array-based CGH further delineated the regions of loss or gain to specific chromosome bands or sub bands based on location of chromosomal site-specific BACs. The specific regions of losses are located at 1p36.2-36.3,6q12,9p13-ter,13q33.33-q34, and gain or amplification of chromosomal DNA at bands 3q26-27. It is interesting to note that some of the chromosomal sites of losses and gains also harbor gene(s) such as AFAR, RIZ1, p73(1p35-36), Akt-1(14q32), p14 and p16(9p21), NF2(22q12), ZASCI, p63 and PIK3CA oncogenes (3q26-27), all of which are known to play a major role in solid tumor pathogenesis. The role of these genes in GIST, however, remains to be seen. Thus, the results of our current study demonstrate that such a combined approach to detect global genomic alterations in GIST is significant in providing information related to chromosomal sites of potential tumor suppressor genes associated with multistep tumorigenesis; some of which also may be predictive of prognosis and clinical outcome following specific targeted therapy, which is the focus of future studies.
