7q21.3

EDSARTH2,EDSCV,OI4

Tumorigenesis

Up- or down-regulation of COL1A2 has been reported in certain cancers. Prior to widespread use of microdissection methods, it was impossible to distinguish changes in tumor cell versus stromal cell expression of COL1A2. In vitro studies have generally shown inhibitory
Roles of tumor cell COL1A2 expression and, as discussed below, promoting roles of stromal cell expression. Thus some of the conflicting data concerning tumor expression may be resolved with better localization of the cells responsible for COL1A2 expression in specific tumors.

Tbx2 is a member of T-box family of transcription factors whose expression is de-regulated in some melanoma, breast and pancreatic cancers. It has been reported by Teng et al (2007) that endogenousTbx2 expression correlates with Col1A2 in several fibroblast cell lines and that overexpression of Tbx2 represses the human COL1A2 in transformed WI-38 fibroblasts and HT1080 human fibrosarcoma cells. Tbx2 appears to act as a co-repressor at a site -107 to +50 on the human COL1A2 promoter. In studies overexpressing murine Tbx2 in NIH3T3 fibroblasts and a rat osteoblastic cell line, Col1α gene was upregulated and downregulated respectively .
In primary infiltrating ductal carcinomas, COL1A2 expression in fibroblasts adjacent to breast tumor cells was increased with stage I tumors compared to nearby normal tissue while in stage II and III tumors a decrease in COL1A2 was observed in adjacent stromal fibroblasts. Additionally, co-culture of normal fibroblasts with breast tumor cell lines resulted in down-regulation of collagen mRNA and protein in fibroblasts.
Microarray analysis of medulloblastoma and primitive neuroectodermal tumor (PNET) specimens reveal overexpression of COL1A2 compared to normal brain tissue. Increased collagen type I protein was also found in medulloblastoma .
Expression profiling of microarray data showed an increased COL1A2 expression in gastric adenocarcinomas compared to normal tissue.
Gene silencing as a result of epigenetic modifications such as histone deacetylation and CpG methylation is increasingly being recognized as an important player in cancer development. Treatment with agents that reverse these processes is an emerging area for cancer therapy. In several human hepatoma cell lines, treatment with the histone deacetylase inhibitor, trichostatin A (TSA), upregulated COL1A2 as demonstrated by microarray and qRT-PCR analysis. Several reports have demonstrated the silencing of COL1A2 due to aberrant methylation within the promoter region of COL1A2 and at CpG islands in primary cancer tissues and from several cancer lines including breast cancer, fibrosarcoma, hepatoma and colorectal carcinoma cells. COL1A2 gene transcription was reestablished upon application of the demethylating agent, 5Aza-dC. Studies such as these that correlate the anticancer effects of TSA and 5Aza-dC with the upregulation of COL1A2 point to a role for COL1A2 as a candidate tumor suppressor gene.
Upregulation of both COL1A2 gene and protein expression has been shown in several studies by cDNA array, tumor tissue microarray and qRT-PCR analysis in subtypes of malignant pleural mesothelioma tumors as compared to normal mesothelial cell lines and pleural mesothelium.
Serial analysis of gene expression from five samples of gastric cancer by Yasui et al (2004) demonstrated that COL1A1 and COL1A2 were upregulated in these tissues compared to normal epithelium. Also, differential expression of COL1A2 occurred with tumor stage and therefore may be marker for invasion and metastasis.
Expression of COL1A2 was analyzed in different stages of large B-cell lymphomas with cDNA microarrays. Downregulation of COL1A2 was observed in more advanced stages of the disease.
Screening of esophageal squamous cancer patients versus normal controls revealed a loss of heterozygosity in one or two of the polymorphic loci located within the promoter or first intron region of the COL1A2 gene in a total of three patients.
The association between the development of oral submucous fibrosis (OSF), a precancerous condition of the oral cavity, and polymorphisms of collagen genes was examined in patients with a history of betal quid chewing, a habit that is a risk factor for this collagen related disorder. Polymorphisms for both COL1A1 and COL1A2 were noted and correlated with an individual propensity for development of OSF depending on the level of exposure to betal quid.
Increased Ets expression is implicated in ECM remodeling, especially within the context of tumor invasion and metastasis. Overexpression of Ets in human dermal fibroblast cultures suppressed the TGFβ-induced activation of the COL1A2 promoter toward a phenotype of increased matrix breakdown and decreased matrix deposition indicative of several diseases including cancer.
Rearrangements in chromosome band 8q12 are characteristic of lipoblastomas and drive the promoter swapping events in the PLAG1 oncogene. In four lipoblastomas that were examined by Hibbard et al (2000), fusion genes between COL1A2-PLAG1 were identified in each case. Fusion of COL1A2 occurred along the entire coding sequence of PLAG1 and results in a full-length PLAG1 protein and truncated COL1A2 protein product with undetermined functionality. Collagen fibers surround the nodular arrangement of lipoblastomas, and it is possible that the lipoblastoma cells are responsible for the production of this capsule. The COL1A2 promoter, therefore, may play a significant role in lipoblastoma.
A finding in RAS - and EGF -transformed cells is that a variety of genes associated with ECM molecules are repressed including COL1A2. Farnesyltransferase inhibitors were found to upregulate the COL1A2 gene and reverse the phenotype of RAS- transformed NIH3T3 fibroblasts. Consistent with this are studies demonstrating that overexpression of COL1A2 suppresses tumorigenesis in RAS-transformed NIH3T3 cells. It appears then that COL1A2 functions as an EGF/RAS regulated growth repressor.

The mechanism by which COL1A2 alters tumor growth may involve interactions with collagen receptors on both tumor cells and tumor stromal cells. Studies of the collagen receptor Endo180 have identified roles in tumor growth mediated by its expression on both stromal fibroblasts and breast carcinoma cells. In the latter case, expression of Endo180 enhanced tumor growth. An additional role of stromal collagen is in regulation of tumor angiogenesis as discussed in more detail below.

Tumor angiogenesis

Three serial analysis of gene expression (SAGE) tags for COL1A2 were significantly more abundant in an analysis of tumor endothelium isolated from human colon carcinomas versus normal endothelium. COL1A1 was also strongly upregulated in tumor endothelium, identifying type I collagen as a potential tumor-endothelium marker. The functional significance of this increased expression was unclear, although early in vitro studies had found that type I collagen is induced during sprouting of post-confluent endothelial cell monolayers. COL1A2 (and COL1A1) was subsequently identified as an important target of the angiogenesis inhibitor thrombospondin-1 (TSP1).
Type I collagen was identified as up-regulated in metabolically labeled proteins secreted by vascular outgrowths from TSP1 null muscle tissue explanted into a 3D collagen gel compared to equivalent explants from wild type mice. Using quantitative RT-PCR, endogenous TSP1 was confirmed to specifically decrease mRNA levels for COL1A1 and COL1A2. Thus, increased angiogenic responses in tissues lacking the angiogenesis inhibitor TSP1 were associated with increased COL1A2 expression. A functional role for this gene expression in the angiogenic switch was then shown using antisense morpholino oligonucleotides to suppress the increased COL1A2 expression in TSP1 null explants. Suppressing either COL1A1 or COL1A2 using antisense morpholino oligonucleotides decreased vascular outgrowth. Therefore, type I collagen gene expression appears to be necessary for angiogenesis, and inhibitors of their expression may inhibit pathological angiogenesis. This is consistent with the known roles of collagen-binding integrins in angiogenesis , up-regulated type I collagen during tube formation in post-confluent 2D endothelial cell cultures, and the hypothesis of Folkman and colleagues that collagen is used as a "mandrill" to organize new blood vessels. Regulation of COL1A1 and COL1A2 via TSP1 may explain the reciprocal regulation of these genes in leiomyomas.

Cancer Metastasis

COL1A2 was independently identified as one of 17 genes highly correlated with metastatic potential from gene expression profiling applied to a set of 279 tumors of diverse types. COL1A2 was up-regulated in metastatic tumors. Subsequent studies showed that the significance of COL1A2 to this signature is limited to certain cancer sites, but the significance of a microenvironment gene signature that includes COL1A2 for predicting disease-free survival has been confirmed in breast cancers. The COL1A1 gene, which encodes the other subunit of type I collagen, was also found in these metastasis signatures, implying that increased type I collagen protein expression would be associated with increased metastasis. This has been confirmed in human and porcine cutaneous melanomas, where increased expression was found in fibroblasts associated with invasive melanomas. Furthermore, pharmacological inhibition of collagen gene expression in the porcine model limited invasive growth and vascularization of the tumors. Increased COL1A2 and COL1A1 expression were also found in gastric carcinomas. Increased COL1A2 expression was significantly associated with tumor stage in this study. Increased COL1A2 expression was also associated with invasion and metastasis in gastric carcinoma.

As in regulating tumor growth, the mechanism by which increased COL1A2 expression increases metastasis may involve interactions with collagen receptors on both tumor cells and tumor stromal cells. LNCaP prostate carcinoma cells expressing elevated alpha-2 integrin / beta-1 integrin showed enhanced motility to type I collagen in vitro and developed more frequent bone metastases in vivo.