A curated database of genetic markers from the angiogenesis/VEGF pathway and their relation to clinical outcome in human cancers

Abstract

Introduction. Angiogenesis causes local growth, aggressiveness and metastasis in solid tumors, and thus, is almost always associated with poor prognosis and survival in cancer patients. Because of this clinical importance, several chemotherapeutic agents targeting angiogenesis have also been developed. Genes and genetic variations in angiogenesis/VEGF pathway thus may be correlated with clinical outcome in cancer patients. Material and methods. Here, we describe a manually curated public database, dbANGIO, which posts the results of studies testing the possible correlation of genetic variations (polymorphisms and mutations) from the angiogenesis/VEGF pathway with demographic features, clinicopathological features, treatment response and toxicity, and prognosis and survival-related endpoints in human cancers. The scientific findings are retrieved from PUBMED and posted in the dbANGIO website in a summarized form. Results and conclusion. As of September 2011, dbANGIO includes 362 entries from 83 research articles encompassing 154 unique genetic variations from 39 genes investigated in several solid and hematological cancers. By curating the literature findings and making them freely available to researchers, dbANGIO will expedite the research on genetic factors from the angiogenesis pathway and will assist in their utility in clinical management of cancer patients. dbANGIO is freely available for non-profit institutions at http://www.med.mun.ca/angio

Cancer is a global burden with millions of new cases diagnosed each year [1]. Clinical management of cancer has become more successful in the last few decades, thanks to success in screening strategies, early detection and advances in treatment. However, approximately 12–14% of all deaths can be attributed to this disease each year, which makes it the third leading cause of death in the world [2].

The array of clinical outcome observed in cancer patients may be unique to each patient. For example, patients with similar disease characteristics may have different response to the same treatment. In addition, toxic side effects of treatment experienced by patients are variable in nature and degree. This inter-patient variability may result from interaction of many different factors. One of these factors is the genetic make-up of the individual which can determine/correlate with response to treatment and the toxic side effects (i.e. genetic predictive factors). In addition, while successful treatment is an important determinant of favorable prognosis and prolonged survival, prognosis and survival in cancer patients also vary considerably regardless of the treatment. A portion of this inter-patient variability can be explained by the genetic prognostic factors (i.e. factors that predict prognosis and survival in patients irrespective of treatment) [3–5].

The transition from early stage cancer (i.e. usually curable) to late stage cancer (i.e. almost incurable) is a critical step in tumor progression and requires changes in the tumor DNA and microenvironment. One of the physiological features contributing to this transition is angiogenesis (i.e. growth of new blood vessels in and around tumors). Abnormal angiogenesis during carcinogenesis facilitates local tumor growth, invasion and disease progression [6–8]. Generation of new lymph vessels (lymphangiogenesis) may also contribute to these disease characteristics. Following angio/lymphangiogenesis, metastasis is likely to occur [9–13]. In contrast to localized disease, metastatic cancer is more difficult to manage, almost always associated with poor prognosis, and responsible for the majority of the cancer-related deaths [14,15].

The well-known drivers of angio/lymphangiogenesis are the members of the Vascular Endothelial Growth Factor (VEGF) family [16]. In humans, there are five VEGF ligands (VEGFs) and five VEGF receptors (VEGFRs) [17]. Among them, four VEGFs (VEGFA, VEGFB, VEGFC, and FIGF; also called VEGFA-D) and three VEGFRs (FLT1, KDR, and FLT4; also called VEGFR1-3) are well-characterized. Binding of VEGFA to FLT1 or to KDR results in angiogenic and anti-angiogenic signaling, respectively. However, binding of FLT4 to either VEGFC or FIGF stimulates lymphangiogenesis [13,18]. Due to this clinical importance, angiogenesis/VEGF pathway is targeted in the treatment of various solid tumors [13,19–22] using monoclonal antibodies, such as bevacizumab against VEGFA, and small tyrosine kinase inhibitors against VEGFRs [23,24].

Because of the fact that the angiogenesis/VEGF pathway has a direct role in tumor biology and disease progression and is targeted by chemotherapeutic agents, this pathway is interesting from both predictive and prognostic points of view. This interest has so far resulted in many studies looking at the relation of genetic variations from this pathway and the clinical outcome in cancer patients. Here we describe dbANGIO, a database created in order to keep pace with this increasing number of publications and to help the scientific and medical community with easy and fast access to their findings.

Material and methods

dbANGIO is similar to one of our previous database projects [25] in structure and development, with minor modifications. The methodological details are presented at the dbANGIO website (http://www.med.mun.ca/angio/further.aspx). In short, the literature reports are retrieved from the PUBMED database (http://www.ncbi.nlm.nih.gov/sites/entrez?db = PubMed&itool = toolbar) using specific search terms. Inclusion criteria were applied to ensure the reports included were relevant to the dbANGIO scope. From each eligible report, select information (such as the ethnicity and the sample size of the cohort, the statistical results, outcomes investigated and information related to the polymorphisms) were retrieved. Additional information about the polymorphisms and the genes (such as SNP IDs, functional consequences, and the gene symbols) were also obtained using other sources. Once the data is compiled, an SQL database that contains the data was created and published on the Internet. The dbANGIO database is periodically updated using the data from new articles published in PUBMED.

Results and discussion

dbANGIO objectives and scope

dbANGIO is freely available at http://www.med.mun.ca/angio. It aims to aid researchers with access to succinct summaries of genetic predictive and prognostic studies that are published in PUBMED and that investigate the potential correlation of genetic variations (i.e. polymorphisms, germline and somatic mutations) of the genes functioning in the angiogenesis/VEGF pathway with the patient demographic information (e.g. sex, age, ethnicity), the disease clinical features (e.g. type of intervention/treatment, performance status) and pathological features at the time of diagnosis (e.g. stage, grade, histology), response to treatment and the treatment-related toxicities experienced by the patients (e.g. response rates, hematological and neurological toxicities), and clinically-relevant survival-related endpoints (e.g. overall, disease-specific, and progression-free survivals, time-to-recurrence). A list of particular features and clinical outcome information included in this database can be found at the dbANGIO search help page (http://www.med.mun.ca/angio/searchhelp.aspx).

Data compiled and posted in dbANGIO

As of September 2011, dbANGIO posts 362 entries in its website. These entries were collected from 83 research articles and are related to 154 unique genetic variations (inherited polymorphisms, somatic mutations, and inherited mutations) from 39 genes from the angiogenesis/VEGF pathway. dbANGIO does not focus on one type of cancer, thus contains entries from a variety of solid and hematological cancers. Cancer sites vary from commonly investigated cancers, such as breast, prostate and lung cancers to other cancers, such as vulvar cancer.

Genetic variations included in the dbANGIO encompasses a variety of variations, such as gene deletions (GSTM1 and GSTT1), chromosomal alterations in the form of deletions and imprinting, small insertion/deletions, microsatellite repeats, and single base substitutions. Almost all of the current entries in dbANGIO are related to inherited genetic polymorphisms (Table IA) identified using the non-tumor DNA (e.g. extracted from blood and tumor surrounding normal tissue). There was only one germline mutation studied so far: the VHL gene mutations in von Hippel-Lindau disease patients, which were found to be correlated with metastatic disease in one study [26]. In addition, only two current entries in dbANGIO are related to somatic mutations (mutations that occur in the tumor DNA). In one study, chromosomal alterations (i.e. deletions and imprinting) in chromosome 7p were investigated and found not to be correlated with several clinicopathological features and overall survival of Wilm's tumor patients from Japan [27]. In another study, small base pair deletions in the KRAS gene were found to be not correlated with response to treatment with bevacizumab and erlotinib, overall survival, and time-to-progression in cholangiocarcinoma and gallbladder carcinoma patients [28].

Table I A. Distribution of the dbANGIO entries with different types of genetic variations.

Type of genetic variation	n (%)
Genetic polymorphisms	359 (99.2%)
Germline mutations	1 (0.3%)
Somatic mutations	2 (0.5%)
Total	362 (100%)

VEGF data in dbANGIO

As of September 2011, in dbANGIO there were a total of 207/362 (57.2%) entries for the VEGF ligands and receptors: specifically two entries for FLT1, 21 entries for KDR, 173 entries for VEGFA, and 10 entries for VEGFC, one entry for FLT4, and no entries for VEGFB and VEGFD (Table IB). Overall, there were more results reported with negative findings (i.e. statistical correlations were not found) than positive findings (i.e. statistical correlations were found), which may represent either true negative findings, or may indicate that studies were underpowered to identify statistical correlations. The most studied VEGFA polymorphisms were the -1154G/A substitution in promoter (rs1570360; 15 entries; correlated with changes in VEGFA expression [29]), T-1498C T/C substitution in promoter (rs833061; 24 entries; correlated with changes in VEGFA expression [30]), + 936C/T substitution in 3′-UTR (rs3025039; 43 entries; correlated with changes in VEGFA expression [31]), -634G/C substitution in 5′-UTR (rs2010963; 38 entries; correlated with changes in VEGFA protein levels [32]), and the -2578C/A substitution (rs699947; 24 entries; correlated with changes in VEGFA expression [29]).

Table I B. The number of dbANGIO entries for the genes in the VEGF pathway.

Gene symbol	n (%)
VEGFA	173 (83.6%)
VEGFB	0 (0%)
VEGFC	10 (4.8%)
VEGFD (FIGF)	0 (0%)
FLT1	2 (1%)
KDR	21 (10.1%)
FLT4	1 (0.5%)
Total	207

Conclusion

In conclusion, genetic predictive and prognostic studies are relatively new, yet critical fields in oncology that can help improve the clinical outcome in cancer patients. An increasing number of such studies are focused on the genetic variations from the angiogenesis/VEGF pathway genes and their relation to clinical outcome in cancer patients. Here we describe dbANGIO, a manually curated, comprehensive, web-based, searchable, and free-to-access public database that summarizes the results of literature reports investigating the possible correlation of genetic variations from the angiogenesis/VEGF pathway with demographic, clinical, and pathological features, treatment response and toxicities experienced, and prognosis and survival characteristics in patients affected by a variety of human cancers. With its unique and comprehensive nature, we sincerely believe that dbANGIO will provide researchers with a better understanding, analysis, and interpretation of the genetic predictive and prognostic studies in cancer involving angiogenesis genes and designing future studies in this field.

Acknowledgements

This work was supported by the Memorial University of Newfoundland and the Research Development Award by The Medical Research Foundation of the Faculty of Medicine, Memorial University of Newfoundland. The author is grateful to Edwin Davis, Terry Upshall, Robert Ryan and Sean O'Neil from the Memorial University-HSIMS for their valuable help with construction and design of the dbANGIO database and website. The author also acknowledges the support of Dr. Ban Younghusband and thanks Michelle Simms for her assistance with the literature data and quality control process.

Declaration of interest: The authors report no conflicts of interest. The authors alone are responsible for the content and writing of the paper.