The domestic dog represents the only model which allows all of these facets to be studied within the same species, and in respect to glioma TICs and adult NSCs, syngeneic samples from the same individual. As the ability to interrogate the canine genome increases, the clearly defined breed predispositions in regard to canine gliomagenesis may elucidate biologically relevant genomic alterations found in human glioma, enhancing our ability to classify this heterogeneous neoplasm by molecular biology, which may greatly improve our understanding over traditional pathologic grading schema. In order to more completely define minimal regions of alteration within large chromosomal amplifications or deletions often found in human glioblastoma, we identified regions of shared CNAs (between human GBMs and canine GSCs) and mapped these regions to the corresponding canine chromosomes. PTEN (located on HSA 10q.23.3) is usually deleted along with large segments of HSA 10q in human glioblastom. A long-standing question in glioma genetics is whether the target of HSA 10q deletion is PTEN alone or whether there are additional target genes that are co-deleted with PTEN that contribute to the tumorigenic phenotype.In the dog, PTEN is located at the telomeric end of CFA 26, which was deleted in the GSCs derived from the secondary and tertiary xenografts. Genes co-deleted in our canine GSCs alongside PTEN in CFA 26 included Dkk1, which functions to inhibit Wnt signaling and may function to inhibit differentiation potential or clonogenic growth potential. Additionally, cyclic guanosine monophosphate (cGMP)-dependent kinase (PRKG1) was also co-deleted with PTEN in the canine tertiary tumors and, alongside PRKG2, has been suggested to be potential tumor suppressor genes in colon carcinoma and glioma. PRKG1 also functions to inhibit angiogenesis through vasodilator-stimulated phosphoprotein (VASP). Thus, the deletion PRKG1 in both canine and human GSCs may have a mechanistic role in the aberrant cell cycle control and angiogenic phenotype of malignant gliomas.In addition to the deletions in CFA 26 containing PTEN (HSA 10q 89.2-90.9Mb), the canine tertiary xenograft tumors also exhibit an expanded deletion in CFA 4 that is syntenic to areas of HSA 10 that flank the PTEN loci and are often deleted in human GBMs (HSA 10q, 59.6-88.7 and 91-91.16Mb). Therefore, while the secondary and tertiary canine GSCs exhibit loss at CFA 26 corresponding to PTEN, progression of xenograft malignancy in our canine GSC population is associated with an additional, separate chromosomal deletion of CFA 4 that contains a syntenic region in immediate proximity to PTEN on the human genome. We examined genes present within this co-deleted segment of CFA 4 corresponding to HSA 10 and identified several genes within this region that have suspected roles in gliomagenesis. ANXA7 (located on HSA 10q21.1-q21.2), which is frequently deleted in human GBMs, has been previously suggested as a tumor suppressor gene independent of its human chromosomal proximity to PTEN. It has been hypothesized that loss of ANXA7 stabilizes and thus augments EGFR signaling and is negatively associated with patient survival. The loss of ANXA7 secondary to deletion of CFA 4, and separate from that containing PTEN (CFA 26), in canine gliomas supports an independent and important role for both genes as potential tumor suppressors in GBMs and demonstrates the power of the dog as a predictive model for interrogating the glioma genome. Additional genes co-deleted within CFA 4 and corresponding to the syntenic regions of HSA 10q involve other tumor suppressor gene candidates including BMPR1A, and CCAR1. We, and others have shown that disruptions in BMP signaling may impact the tumorigenic potential and capacity for differentiation in GSCs, and BMPR1A deletions are known to predispose individuals to colon cancer formation. CCAR1, or cell cycle and apoptosis regulator 1 (also referred to as cell cycle and apoptosis regulatory protein 1 or CARP-1) is located on HSA 10q21-q22 and has been reported to suppress the clonogenic growth, tumorigenicity, and invasion of human breast cancer cells and is integral to the induction of apoptosis following EGFR inhibition. Loss of CCAR1 may explain in part the failure of some patients to respond to EGFR inhibition therapeutically or enable tumor cells to survive in conditions of low growth factor concentration.We performed similar analyses on the other commonly altered genomic foci shared between our canine GSCs and human GBMs. The canine GSC genomic deletion containing CDK2NA is evolutionarily related to a very small region of HSA 9p21-p22, highlighting the importance of CDKN2A in glioma biology and diminishing the potential importance of a large number of other genes (passenger genes) that are co-deleted in the large CNAs found in HSA 9. Indeed, the canine chromosomal regions flanking the small locus containing CDKN2A are amplified, suggesting a specific role of CDKN2A deletion in glioma biology while potentially excluding a number of linked co-deleted genes in HSA 9. Within that minimally deleted region containing the p16/Ink4a locus in our canine GSCs are several interferon genes (IFNB1, IFNA5, IFNA13, and IFNA7) that are also deleted. This may be of interest for the IFNAR1 gene was recently identified by Cerami et al. as a linker gene in the PIK3R1 module through an automated network analysis (Human Interaction Network). This module, linking several IFNA genes and IFNB1 to IFNAR1, was found altered in 25% of GBMs in the TCGA database, but as the authors express, was of unknown significance due to the close proximity of IFN genes to CDKN2A. The co-deletion of CDKN2A and the IFNAR1 gene in both canine and human gliomas strengthens the argument that IFNAR1 is a potentially significant gene in the biology of GBMs.An additional chromosomal amplification associated with the more malignant xenograft gliomas is seen in CFA 6, which shares evolutionarily conserved synteny with HSA 7. This is of particular interest since trisomy HSA 7 is a common finding in GBMs. This region contains the gene, Glioblastoma Amplified Sequence (GBAS), which has been reported to be amplified in up to 40% of human GBMs. Other commonly altered foci in human GBMs show striking conservation in our canine GSCs including those surrounding amplification of MDM4, MDM2, and in the genomic regions commonly deleted in HSA 6 and 13. A number of these CNAs that are conserved across both species contain well-known potential tumor suppressor or oncogenes. One such example is Gli1, a transcription factor involved in the transduction of sonic hedgehog signaling and that may have a role in the promotion of tumor cell invasion. The Gli1 gene (HSA 12q13.2-13.3) is amplified both in a subset of human GBMs as well as in our canine glioma (CFA 10). By contrast, it is of interest that amplification of the epidermal growth factor receptor (EGFR) seen in primary human GBMs was not seen in our canine tertiary tumors, consistent with its lack of amplification in secondary human GBMs. Thus, these foci of shared genomic alterations allow us to identify a series of genomic events responsible for driving both human and canine gliomagenesis with more certainty than would be possible through the genomic study of gliomas from only one species.

Agency
National Institute of Health (NIH)
Institute
National Cancer Institute (NCI)
Type
Investigator-Initiated Intramural Research Projects (ZIA)
Project #
1ZIABC011099-04
Application #
8349324
Study Section
Project Start
Project End
Budget Start
Budget End
Support Year
4
Fiscal Year
2011
Total Cost
$634,619
Indirect Cost
Name
National Cancer Institute Division of Basic Sciences
Department
Type
DUNS #
City
State
Country
Zip Code
Riddick, Gregory; Song, Hua; Ahn, Susie et al. (2011) Predicting in vitro drug sensitivity using Random Forests. Bioinformatics 27:220-4
Edwards, Lincoln A; Woolard, Kevin; Son, Myung Jin et al. (2011) Effect of brain- and tumor-derived connective tissue growth factor on glioma invasion. J Natl Cancer Inst 103:1162-78
Riddick, Gregory; Fine, Howard A (2011) Integration and analysis of genome-scale data from gliomas. Nat Rev Neurol 7:439-50
Wuchty, Stefan; Arjona, Dolores; Li, Aiguo et al. (2011) Prediction of Associations between microRNAs and Gene Expression in Glioma Biology. PLoS One 6:e14681
Wuchty, Stefan; Zhang, Alice; Walling, Jennifer et al. (2010) Gene pathways and subnetworks distinguish between major glioma subtypes and elucidate potential underlying biology. J Biomed Inform 43:945-52
Bozdag, Serdar; Li, Aiguo; Wuchty, Stefan et al. (2010) FastMEDUSA: a parallelized tool to infer gene regulatory networks. Bioinformatics 26:1792-3
Edwards, Lincoln A; Fine, Howard A (2010) The Ids have it. Cancer Cell 18:543-5
Li, Aiguo; Bozdag, Serdar; Kotliarov, Yuri et al. (2010) GliomaPredict: a clinically useful tool for assigning glioma patients to specific molecular subtypes. BMC Med Inform Decis Mak 10:38
Kotliarov, Yuri; Bozdag, Serdar; Cheng, Hangjiong et al. (2010) CNAReporter: a GenePattern pipeline for the generation of clinical reports of genomic alterations. BMC Med Genomics 3:11
Son, Myung Jin; Woolard, Kevin; Nam, Do-Hyun et al. (2009) SSEA-1 is an enrichment marker for tumor-initiating cells in human glioblastoma. Cell Stem Cell 4:440-52

Showing the most recent 10 out of 11 publications