HIF1A gene polymorphisms and human diseases: graphical review of 97 association studies

Abstract

Hypoxia-inducible factors (HIFs) belong to a family of transcription factors (TF) responsive to a low O₂ availability, which is often a characteristic feature of solid tumors. The alpha subunit of the HIF heterodimer is O₂-sensitive, and once stabilized in hypoxia, it functions as a master regulator of various genes involved in hypoxia pathway. Changes in the HIF1A (hypoxia inducible factor 1, alpha subunit) nucleotide sequence or expression has been shown to be associated with development of several diseases. Due to increasing research interest in HIF1A gene a review of association studies was needed. We here reviewed published data on single nucleotide polymorphisms (SNPs) in HIF1A in various diseases; in total, 34 SNPs were tested for an association with 49 phenotypes, and the results were visualized using the Cytoscape software. Among all collected polymorphisms 16 SNPs showed significant associations with 40 different phenotypes, including six SNPs associated with 14 cancer types. Missense SNPs (rs11549465 and rs11549467) within the oxygen-dependent degradation domain were most frequently studied. The study provides a comprehensive tool for researchers working in this area and may contribute to more accurate disease diagnosis and identification of therapeutic targets.

Introduction

Mammalian cells rapidly respond and adapt to low oxygen conditions (hypoxia). In hypoxic cells important responses are activated for metabolic, bioenergetics, and redox demands (reviewed in (Majmundar et al., 2010)). A primary transcriptional response to hypoxia is mediated by the hypoxia-inducible factor (HIF), which is known as a pivotal regulator under hypoxia stress. Besides its adaptive response in cellular stress, it carries important roles in physiological and pathological processes (Semenza 2003, Majmundar et al., 2010). The HIF protein complex is a heterodimer consisting of an oxygen sensitive (alpha, A) and an oxygen stable (beta, B) subunit (Wang and Semenza 1993). In mammals three different HIFA isoforms are present, among which HIF1A is expressed ubiquitously while HIF2A and HIF3A expression vary depending on type of tissue cells (Bertout et al., 2008). In normally oxidized conditions (normoxia) A-subunits are hydroxylated at conserved proline residues (p.Pro402 and p.Pro564) (Ivan et al. 2001) by oxygen regulated prolyl hydroxylase domain-containing enzymes (EGLN1, 2, 3 or PHD1, 2, 3) (Wang et al. 1995, Chan et al. 2005, Kaelin and Ratcliffe 2008). Marked HIFA-subunits are recognized by E3 ubiquitin ligase from the von Hippel-Lindau protein complex (pVHL) for proteasomal degradation (Wang et al. 1995, Majmundar et al., 2010). The pVHL can target the N-terminal transactivation domain (N-TAD) within the oxygen-dependent degradation domain (ODD domain), which controls HIF1A degradation by ubiquitin-proteasome pathway and consists of approximately 200 amino acids (Huang et al. 1998) (Supplementary Figure 1). The removal of the ODD domain renders HIF1A stability even under normoxic conditions, consequently resulting in autonomous HIF1A heterodimerisation, DNA binding and transactivation independently from hypoxic signaling (Huang et al. 1998).

Transcription factor HIF1A mediates transcriptional responses to hypoxia for a high number of genes to control cellular oxygen supply and maintain cell viability during periods of low oxygen concentration (Wang et al. 1995, Wenger et al., 2005, Vilela et al., 2008, Keith et al., 2012). Analysis of 98 HIF1A target genes collected from 51 published studies revealed 20 associated pathways (Slemc and Kunej 2016). HIF1A was implicated in: metabolism and redox homeostasis (glucose catabolism, regulation of lipid metabolism), vascular responses in hypoxia (ischemia-induced angiogenesis, endothelial cells), cancer (tumorigenesis, metastasis, tumor angiogenesis, cancer stem cells, regulation by cancer metabolism), inflammation (regulation in inflammatory cells, myeloid cell function, tumor-associated macrophages), and also as part of a systemic response to hypoxia reviewed in (Majmundar et al., 2010). Besides regulating expression of protein-coding genes, it has been shown that HIF1A also regulates noncoding RNA genes (ncRNA) including microRNAs (miRNAs) (Gee et al. 2014) and transcribed-ultraconserved regions (T-UCRs) (Ferdin et al. 2013).

Although HIF1A gene has been a topic of several studies it is not yet included in the Diseasome map, which visualizes known interactions between genes and diseases (Goh et al. 2007). The aim of the present study was to review published reports of associations between HIF1A gene polymorphisms and diseases or phenotypic traits in human and to graphically visualize associations as the gene-disease network.

A literature review of HIF1A polymorphisms, associated diseases and phenotypes was performed using PubMed (http://www.ncbi.nlm.nih.gov/pubmed) and Web of Science (WoS; http://apps.webofknowledge.com/WOS). The literature search was performed on publications published between 2003 and 2015 using keywords, including: HIF1A, polymorphisms, SNP, human, cancer, and association studies. The nomenclature for SNPs was unified to the reference SNP (rs) IDs according to the Human Genome Variation Society (HGVS) guidelines (http://www.hgvs.org/mutnomen/). Versions of dbSNP build 137 from NCBI and Ensembl (release 85) were used to obtain information about variations (biotype of the polymorphism and genomic location). Associations between the HIF1A gene and phenotypes were visualized using Cytoscape 3.2.1 software (Shannon et al. 2003). Associated phenotypes have been sorted into disease categories as in the Diseasome map, which includes 22 categories of diseases and disorders: bone, cancer, cardiovascular, connective tissue, dermatological, developmental, ear, nose and throat, endocrine, gastrointestinal, hematological, immunological, metabolic, muscular, neurological, nutritional, ophthamological, psychiatric, renal, respiratory, skeletal, multiple, and unclassified (Goh et al. 2007). Due to heterogeneity in phenotype nomenclature in publications, the terminology was unified according to the Disease Ontology (DO) database (http://disease-ontology.org/).

Extraction of literature data and editing according to the genomic databases

The workflow of the study and main results are presented in Fig. 1. Polymorphisms’ names were edited according the latest genomic browser releases. Usually, articles had polymorphisms defined with rs ID numbers, but in more than 20 cases it was necessary to identify corresponding rs ID number in genomic databases (Prior et al. 2003, Yamada et al. 2005, Hong et al. 2007, Konac et al. 2007, Lee et al. 2008). Two polymorphisms from the published literature (rs62639821 and rs1957577) were not included in the study, since they are now annotated to different genomic locations. Disease and phenotype names used in the studies were edited according to the nomenclature in the DO database, if available. Namely; some studies used different terms for the same disease, or only some aspect(s) of disease development, progression or consequence were recorded. Out of 49 tested phenotypes it was possible to translate 38 phenotypes according to the DO database, however for the rest of the tested diseases and phenotypes an appropriate term was not yet available in the database. Because some of the studied traits are phenotypes and not diseases, it will be necessary in the future studies to combine data from additional ontology databases like Human Phenotype Ontology (HPO) to verify the terminology of all traits.

Figure 1.

Workflow of the study and main results. We performed literature search in PubMed and Web of Science and extracted data related to polymorphisms and phenotype. The collected data were edited and complemented with additional information. Next, we sorted studies according to association type between polymorphism and phenotype (positive, protective, negative). Disease and phenotype terminology was curated according to Disease Ontology database and sorted into classes according to Diseasome map. Collected associations between SNPs and phenotypes were visualized as a network.

Visualization of the HIF1A gene - phenotype network

Literature review of 97 association studies in humans revealed that 34 HIF1A SNPs were tested for an association with 49 phenotypes (Figure 2, Supplementary Figure 2). Those 49 phenotypes belong to 11 disease categories: cancer, cardiovascular, bone, renal, hematological, immunological, connective tissue disorder, respiratory, gastrointestinal, muscular and unclassified. Among tested associations 16 SNPs showed positive associations with 40 phenotypes, including 14 cancer types (Figure 2A). Some studies also showed the protective role of polymorphisms; meaning a protective association for the specific disease phenotype (such as loss of articular cartilage, or advanced prostate cancer, and so on); three polymorphisms in nine associations were associated with lower risk for development of different diseases (blue lines in Fig. 2A). The studied HIF1A SNPs are located within various genomic regions: 5’-flanking, exons, introns, 3’-UTR, and 3’-flanking (Supplementary Figure 3). An integrated review revealed that among 34 SNPs, 10 showed variable associations (i.e., either present or absent; Table 1), 6 were always positively associated with the tested phenotypes, and 18 showed no association with the tested phenotypes (Table 2, Fig. 2B). Among all SNPs, missense polymorphisms within the ODD domain (rs11549465 and rs11549467) were most frequently studied and for which variable nomenclature was used in publications. Moreover, several studies that tested associations of these two SNPs with phenotypes showed conflicting results for the same cancer type. For example, opposing results were reported for rs11549465 in prostate cancer susceptibility studies; with association (Chau et al. 2005, Fu et al. 2005, Orr-Urtreger et al. 2007, Jacobs et al. 2008, Foley et al. 2009) and no association (Li et al. 2007, Li et al. 2012).

Figure 2.

Visualization of reported HIF1A genotype-phenotype associations studies in human. (A) Associations between HIF1A SNPs and studied phenotypes. Blue lines denote protective role of polymorphism. (B) Reported negative associations between studied SNPs and phenotypes.

Figure legend: CAD, coronary artery disease; COPD, chronic obstructive pulmonary disease; CRC, colorectal cancer; ESCC, esophageal squamous cell carcinoma; NSCLC, non-small cell lung cancer; OSCC, Oral squamous cell carcinoma; T1DM, diabetes mellitus type 1; T2DM, diabetes mellitus type 2; SLE, systemic lupus erythematosus. The term “performance” is used as a hypernym for: adaptation to living at high altitude, endurance status, maximal oxygen consumption, power-oriented athletes, symptom-limited exercise test duration over time, and response to acute hypoxia.

Table 1.

HIF1A SNPs showing both positive and negative association with disease phenotypes.

#	SNP ID/substitution	SNP name as in reference	Genomic location	Association with diseases and phenotypes
1	rs1957757 [T>C]		intron 6
			(+)	○ symptom-limited exercise test duration (Sarzynski et al. 2010)
			(−)	• prostate cancer risk (Jacobs et al. 2008)
				○ systemic sclerosis (Wipff et al. 2009)
				○ systemic lupus erythematosus (Feng et al. 2014)
2	rs12434438 [G>A]		intron 6
			(+)	○ systemic sclerosis (Wipff et al. 2009)
				○ T2DM (together with rs1319462) (Yamada et al. 2005)
			(−)	○ Systemic lupus erythematosus (Feng et al. 2014)
3	rs10873142 [C>T]		intron 8
			(+)	○ coronary artery disease with stable exertional angina (Hlatky et al. 2007)
				○ idiopathic osteonecrosis of the femoral head in men (Hong et al. 2007)
			(−)	○ acute myocardial infarction and frequent intradialytic hypotension (Hemodialysis patients) (Zheng et al. 2009)
				○ early- onset pre-eclampsia (Andraweera et al. 2014)
				○ lung cancer (Konac et al. 2009)
				○ COPD (Ding et al. 2015)
4	rs41508050 [C>T]		exon 10
			(+)	○ coronary artery disease with stable exertional angina (Hlatky et al. 2007)
		Thr418Ile	(−)	• lung cancer (Konac et al. 2009)
5	rs2301113 [C>A]		intron 10
			(+)	○ early-stage NSCLC (Liu et al. 2014)
			(−)	○ prostate cancer risk (Jacobs et al. 2008)
				○ acute myocardial infarction and frequent intradialytic hypotension (Hemodialysis patients) (Zheng et al. 2009)
				○ elite endurance status (Doring et al. 2010)
				○ COPD(Ding et al. 2015)
6	rs11549465 [C>T]		exon 12
		Pro582Ser	(+)	• cancer risk (He et al. 2013, Yang et al. 2013, Hu et al. 2014, Ye et al. 2014) and cancer metastasis (Zhang et al. 2013) cancer malignancy (Wu et al. 2014)
				• glioma (Xu et al. 2011)
				• oral cancer risk (together with rs11549467) (Chen et al. 2009)
				• esophageal squamous cell carcinoma (Ling et al. 2005)
				• head and neck squamous cell carcinoma (Tanimoto et al. 2003)
				• renal cell carcinoma (Ollerenshaw et al. 2004)
		C2028T		• lung cancer (NSCLC) (Koukourakis et al. 2006)
				• breast cancer risk (poor prognosis) (Naidu et al. 2009, Lee et al. 2008, Kim et al. 2008)
				• CRC risk and metastasis (Kuwai et al. 2004, Kang et al. 2011), ulcerative CRC (Fransen et al. 2006)
				• prostate cancer risk/susceptibility (Foley et al. 2009, Orr-Urtreger et al. 2007); androgen-independent prostate cancer (Chau et al. 2005, Fu et al. 2005)
				• pancreatic cancer (Ruiz-Tovar et al. 2012) (Wang et al. 2011)
		C1772T		• cervical and endometrial cancers (Konac et al. 2007)
				• urinary cancers (Li et al. 2013)
				• increased female specific cancer risk (Zhao et al. 2009)
				• gastrointestinal tract cancer risk (Xu et al.)
				○ CAD with stable exertional angina (Hlatky et al. 2007)
				○ ischemic heart disease (collaterals formation) (Resar et al. 2005);
				○ chronic obstructive pulmonary disease susceptibility (with rs11549467) (Putra et al. 2011a)
				○ risk of cellulite (Emanuele et al. 2010)
				○ abdominal aortic aneurysm (Strauss et al. 2012)
				○ acute kidney injury (Kolyada et al. 2009)
				○ endurance status (McPhee et al. 2011, Doring et al. 2010)
				○ maximal oxygen consumption (Prior et al. 2003)
				○ power-oriented athletes and muscle activity (Ahmetov et al. 2008, Cieszczyk et al. 2011, Gabbasov et al. 2012)
				✓ knee osteoarthritis (Fernandez-Torres et al. 2015)
				✓ diabetic nephropathy (Gu et al. 2012)
				✓ prostate cancer (Jacobs et al. 2008)
				✓ T2DM (Yamada et al. 2005, Nagy et al. 2009)
				✓ T1DM (Nagy et al. 2009)
				✓ early-onset pre-eclampsia (Andraweera et al. 2014)
			(−)	• gastric cancer (Li et al. 2009)
				• breast cancer risk (Zagouri et al. 2012, Vleugel et al. 2005, Meka et al. 2015); sporadic breast cancer (Apaydin et al. 2008)
				• ovarian cancer (Konac et al. 2007)
				• digestive cancer (Yang et al. 2014, Sun et al. 2015)
				• prostate cancer (Li et al. 2007, Li et al. 2012)
				• cervical cancer (Fu Sl Fau - Miao et al. 2014)
				• bladder cancer occurrence (Nadaoka et al. 2008)
				• oral squamous cell carcinoma (Munoz-Guerra et al. 2009, Shieh et al. 2010)
				• lung cancer (Konac et al. 2009, Kuo et al. 2012) (associated with lung cancer with TP53 LOH) (Putra et al. 2011b), NSCLC (Kim et al. 2010)
				• renal cell carcinoma (Morris et al. 2009) (Qin et al. 2011)
				• hepatocellular carcinoma risk (Hsiao et al. 2010)
				• endometrial cancer risk (potential association with tumorigenesis and increased tumor vasculature) (Horree et al. 2008)
				• survival of patients with colorectal cancer (Lee et al. 2011)
				○ systemic sclerosis (Wipff et al. 2009)
				○ acute myocardial infarction and frequent intradialytic hypotension (Zheng et al. 2009)
				○ pre-eclampsia (Heino et al. 2008, Kim et al. 2012, Nava-Salazar et al. 2011)
				○ erythrocytosis (Percy et al. 2003)
				○ giant cell arteritis (Torres et al. 2010)
				○ peripheral artery disease (Bahadori et al. 2010)
				○ osteonecrosis (Chachami et al. 2013)
				○ premature coronary artery disease (Lopez-Reyes et al. 2014)
				○ ischemic heart disease (coronary artery collaterals) (Alidoosti et al. 2011)
				○ lumbar disc degeneration (Lin et al. 2013)
				○ systemic lupus erythematosus (Feng et al. 2014)
				○ power-oriented athletes (Eynon et al. 2010)
				○ response to acute hypoxia (Hennis et al. 2010)
				○ idiopathic osteonecrosis of the femoral head in men (Hong et al. 2007)
				○ acute mountain sickness (Droma et al. 2008)
7	rs11549467 [G>A]		exon 12
		Ala588Thr	(+)	• cancer risk (Zhou et al. 2014, Yang et al. 2013, Liu and Zhang 2013, Hu et al. 2014)
				• urinary cancers (Li et al. 2013)
				• digestive cancers (Yang et al. 2014, Sun et al. 2015)
				• oral squamous cell carcinoma (Chen et al. 2009, Munoz-Guerra et al. 2009)
				• head and neck squamous cell carcinoma (Tanimoto et al. 2003)
				• gastric cancer (Li et al. 2009)
				• ulcerative CRC (together with rs11549467) (Fransen et al. 2006)
				• hepatocellular carcinoma risk (Hsiao et al. 2010)
				• renal cell carcinoma (Ollerenshaw et al. 2004)
				• pancreatic cancer (Ruiz-Tovar et al. 2012, Wang et al. 2011)
				• prostate cancer (Li et al. 2012)
				○ chronic obstructive pulmonary disease susceptibility (with rs11549465) (Putra et al. 2011a)
				○ abdominal aortic aneurysm (Strauss et al. 2012)
				○ lumbar disc degeneration (Lin et al. 2013)
				✓ breast cancer risk (Zhao et al. 2009)
				✓ T1DM (Nagy et al. 2009)
		G1790A; G2046A	(−)	• oral squamous cell carcinoma (Shieh et al. 2010)
				• breast cancer risk (Naidu et al. 2009, Apaydin et al. 2008, Kim et al. 2008)
				• lung cancer (Konac et al. 2009), (associated with adenocarcinoma with 1p34 LOH) (Putra et al. 2011b)) and lung carcinoma (NSCLC) (Koukourakis et al. 2006, Kuo et al. 2012, Kim et al. 2010)
				• CRC risk, progression and metastasis (Knechtel et al. 2010, Szkandera et al. 2010)
				• prostate cancer (Li et al. 2007, Orr-Urtreger et al. 2007, Chau et al. 2005)
				• ovarian, cervical and endometrial cancers (Konac et al. 2007, Fu Sl Fau - Miao et al. 2014)
				• renal cell carcinoma (Qin et al. 2011, Morris et al. 2009)
				• bladder cancer (Nadaoka et al. 2008)
				• survival of patients with colorectal cancer(Lee et al. 2011)
				○ pre-eclampsia (Heino et al. 2008, Kim et al. 2012, Nava-Salazar et al. 2011)
				○ peripheral artery disease (Bahadori et al. 2010)
				○ giant cell arteritis (Torres et al. 2010)
				○ coronary artery disease with stable exertional angina (Hlatky et al. 2007)
				○ osteonecrosis (Chachami et al. 2013)
				○ premature coronary artery disease (Lopez-Reyes et al. 2014)
				○ endurance status (Doring et al. 2010)
				○ knee osteoarthritis (Fernandez-Torres et al. 2015)
8	rs199775054 [G>C]		exon 12
		Ala593Pro	(+)	• hepatocellular carcinomas (Park et al. 2009)
		Ala593Pro	(−)	• colon, gastric, breast, lung cancer and acute leukemia (Park et al. 2009)
9	rs113182457 rs60361955 [insGT]		intron 13
		GT dinucleotide repeat	(+)	○ lung carcinoma (NSCLC) (Koukourakis et al. 2006)
		GT14 allele	possible (+)	○ adaptation to living at high altitude (Suzuki et al. 2003)
		rs10645014	(−)	• lung cancer (Konac et al. 2009)
10	rs2057482 [T>C]		3’-UTR
			(+)	• cervical cancer (Fu Sl Fau - Miao et al. 2014)
				• rectal cancer risk (Frank et al. 2010)
				○ idiopathic osteonecrosis of the femoral head in men (Hong et al. 2007)
				○ early-stage NSCLC (Liu et al. 2014)
				○ premature coronary artery disease (Lopez-Reyes et al. 2014)
				✓ hepatocellular carcinoma (Guo et al. 2015)
			(−)	• cancer risk (Hu et al. 2014)
		Ex15+197C>T		• breast cancer risk (Lee et al. 2008)
				• renal cell carcinoma (Qin et al. 2011)
				• prostate cancer (Li et al. 2012)
				○ acute myocardial infarction and frequent intradialytic hypotension (Zheng et al. 2009)
				○ knee osteoarthritis (Fernandez-Torres et al. 2015)
		C191T		○ coronary artery disease with stable exertional angina (Hlatky et al. 2007)

Legend:

(+) observed association, (−) no association.

• Cancer

○ Other diseases and phenotypes

✓ Protective role of polymorphism

Table 2.

HIF1A SNPs showing both positive and negative association with disease phenotypes.

#	SNP ID [substitution]	Synonym, as named in the reference	Genomic location	Association with diseases and phenotypes
ASSOCIATION
1	rs2783778 [C>T]	-	5’-flanking	○ acute myocardial infarction and frequent intradialytic hypotension (Hemodialysis patients) (Zheng et al. 2009)
2	rs7148720 [T>C]	-	5’-flanking	○ acute myocardial infarction and frequent intradialytic hypotension (Hemodialysis patients)(Zheng et al. 2009)
3	rs1535679 [A>C]	−2755C>A	5’-flanking	○ idiopathic osteonecrosis of the femoral head in men (Hong et al. 2007)
4	rs28708675 [A>T]	A-2500T	5’-flanking	○ maximal oxygen consumption (Prior et al. 2003)
5	rs1319462 [G>A]	-	3’-flanking	○ T2DM (together with rs12434438) (Yamada et al. 2005)
6	rs1957755 [G>A]	-	intron 4	○ symptom-limited exercise test duration (Sarzynski et al. 2010)
NO ASSOCIATION
1	rs41362550 [T>C]	-	5’-flanking	○ CAD with stable exertional angina (Hlatky et al. 2007)
2	rs7143164 [G>C]	-	intron 1	○ acute myocardial infarction and frequent intradialytic hypotension (Hemodialysis patients) (Zheng et al. 2009)
				○ systemic lupus erythematosus (Feng et al. 2014)
				○ COPD (Ding et al. 2015)
				• prostate cancer (Jacobs et al. 2008)
3	rs1951795 [A>C]	-	intron 1	○ elite endurance status (Doring et al. 2010)
				○ systemic lupus erythematosus (Feng et al. 2014)
4	rs12435848 [A>G]	-	intron 1	• prostate cancer risk (Jacobs et al. 2008)
5	rs2301104 [G>C]	-	intron 1	○ COPD (Ding et al. 2015)
6	rs10129270 [G>A]	-	intron 1	○ COPD (Ding et al. 2015)
7	rs8005745 [T>A]	-	intron 1	○ COPD (Ding et al. 2015)
8	rs779897997[C>A]	S28Y	exon 2	○ T2DM (Yamada et al. 2005)
		C111A		• ovarian, cervical and endometrial cancers (Konac et al. 2007)
				• breast cancer (Apaydin et al. 2008, Naidu et al. 2009)
9	rs4899056 [T>C]	-	intron 4	• prostate cancer (Jacobs et al. 2008)
				○ COPD (Ding et al. 2015)
10	rs11158358 [G>C]	-	intron 6	• prostate cancer (Jacobs et al. 2008)
				○ elite endurance status (elite endurance athletes (EEA)) (Doring et al. 2010)
11	rs2301111 [G>C]	-	intron 7	• prostate cancer (Jacobs et al. 2008)
12	rs966824 [T>C]	-	intron 7	○ COPD (Ding et al. 2015)
13	rs41492849 [C>T]	C1720T	exon 12	• OSCC (Shieh et al. 2010)
				• lung cancer (Konac et al. 2009)
14	rs34005929 [G>A]	1740G>A	exon 12	○ pre-eclampsia (Heino et al. 2008)
				• OSCC (Shieh et al. 2010)
				• lung cancer (Konac et al. 2009)
15	rs61755645 [A>T]	A1828T	exon 12	○ pre-eclampsia (Heino et al. 2008)
				• OSCC (Shieh et al. 2010)
16	rs4902080 [T>C]	-	intron 12	○ CAD with stable exertional angina (Hlatky et al. 2007)
				○ COPD (Ding et al. 2015)
17	rs4902082 [C>T]	T+140 C	intron 14	○ maximal oxygen consumption (Prior et al. 2003)
18	rs17099207 [G>A]	-	3’-flanking HIF1A and	○ elite endurance status (elite endurance athletes (EEA)) (Doring et al. 2010)

Legend:

• Cancer

○ Other diseases and phenotypes

Abbreviations: OSCC, oral squamous cell carcinoma; CAD, coronary artery disease; RCC, renal cell carcinoma; CRC, colorectal cancer; T2DM, diabetes mellitus (type 2); T1DM, diabetes mellitus (type 1); NSCLC, non-small-cell lung cancer.

HIF1A and cancer risk

Among 16 HIF1A SNPs associated with 40 different phenotypes six SNPs have been associated with increased risk for 14 cancer types: rs113182457 (Koukourakis et al. 2006), rs11549465 (Tanimoto et al. 2003, Ollerenshaw et al. 2004, Chau et al. 2005, Fu et al. 2005, Ling et al. 2005, Fransen et al. 2006, Koukourakis et al. 2006, Hong et al. 2007, Konac et al. 2007, Orr-Urtreger et al. 2007, Jacobs et al. 2008, Kim et al. 2008, Lee et al. 2008, Chen et al. 2009, Foley et al. 2009, Naidu et al. 2009, Zhao et al. 2009, Kang et al. 2011, Wang et al. 2011, Xu et al. 2011, Ruiz-Tovar et al. 2012, He et al. 2013, Li et al. 2013, Xu et al. 2013,Yang et al. 2013, Zhang et al. 2013, Hu et al. 2014, Wu et al. 2014, Ye et al. 2014, Fernandez-Torres et al. 2015), rs11549467 (Tanimoto et al. 2003, Ollerenshaw et al. 2004, Fransen et al. 2006, Chen et al. 2009, Li et al. 2009, Munoz-Guerra et al. 2009, Zhao et al. 2009, Hsiao et al. 2010, Wang et al. 2011, Li et al. 2012, Ruiz-Tovar et al. 2012, Li et al. 2013, Liu and Zhang 2013, Yang et al. 2013, Hu et al. 2014, Yang et al. 2014, Zhou et al. 2014, Sun et al. 2015), rs199775054 (Park et al. 2009), rs2057482 (Frank et al. 2010, Fu Sl Fau - Miao et al. 2014, Liu et al. 2014, Guo et al. 2015) and rs2301113 (Liu et al. 2014). These six SNPs were most frequently associated with breast, lung, colorectal (CRC), gastric, prostate, oral cancer and renal cell carcinoma (RCC). It is known that solid tumors frequently have low levels of O₂, which can be a result of cancer cells growing more rapidly than their supporting vascular network. Hypoxic stress might also be caused by a perfusion defect as a result of abnormal tumor blood vessel structure and function (Majmundar et al., 2010). Consequently, these events cause HIF1A levels to increase in solid tumors (Bertout et al., 2008), but its levels can additionally be increased by HIF-independent pathways (Majmundar et al., 2010). In addition, HIF1A has a capability to directly reprogram the metabolic state in cells, which is important in hypoxic settings such as vascular disease and cancer (Majmundar et al., 2010). Thus, HIF1A in hypoxic cells regulates the transcription of many genes involved in key aspects of cancer biology, including immortalization, maintenance of stem cell pools, cellular differentiation, genetic instability, vascularization, metabolic reprogramming, autocrine growth factor signaling, invasion/metastasis, and treatment failure (Semenza 2007). Increased expression of HIF1A also often associates with poor clinical prognosis of many cancer types (Semenza 2007).

HIF1A and association with other phenotypes and diseases

Genetic variability of HIF1A was also found to be associated with cardiovascular system diseases like: ischemic heart disease, coronary artery disease (CAD) with stable exertional angina, premature coronary artery disease, pre-eclampsia, acute myocardial infarction and frequent intradialytic hypotension. Polymorphisms associated with cardiovascular system diseases were: rs10873142 (Hlatky et al. 2007), rs11549465 (Resar et al. 2005, Hlatky et al. 2007, Strauss et al. 2012, Andraweera et al. 2014), rs11549467 (Strauss et al. 2012), rs2057482 (Lopez-Reyes et al. 2014), rs2783778 (Zheng et al. 2009), rs41508050 (Hlatky et al. 2007) and rs7148720 (Zheng et al. 2009). Cardiovascular diseases like atherosclerosis in the heart, brain, and limb muscle, are known to be susceptible to ischemic injury (Beckman et al., 2002, Kett-White et al. 2002). Besides, myocardial ischemia is the most common cause of cardiac hypoxia in clinical medicine and it occurs when O₂ delivery cannot meet myocardial metabolic requirements in the heart (Shohet and Garcia 2007). Expression of HIF1A is essential and sufficient for promoting reperfusion in ischemic skeletal muscle (Majmundar et al., 2010). Moreover, reduced HIF1A activation was also found in hypoxic skin wounds of aged diabetic mice (Liu et al. 2008), emphasizing the role of age in ischemic response. This is important since peripheral arterial disease is associated with age (Beckman et al., 2002). Pro-angiogenic roles of HIF1A were also associated with hypertrophic cardiomyopathy, myocardial infarction, skin wound healing, and retinal neovascularization (reviewed in (Majmundar et al., 2010)). Many studies suggested that the occurrence of local hypoxia in the muscle causes a drop in O₂ pressure within the myocyte during exercise (Richardson et al. 1995, Richardson et al., 2001).

Because the cardiovascular system subsequently influences human body performance, HIF1A variations have been associated with the following performance related phenotypes: maximal oxygen consumption, adaptation to living at high altitude, lumbar disc degeneration, idiopathic osteonecrosis of the femoral head, power-oriented athlete performance and muscle activity, endurance status, and symptom-limited exercise test duration over time (rs113182457 (Suzuki et al. 2003), rs10873142 (Hong et al. 2007), rs11549465 (Prior et al. 2003, Hong et al. 2007, Ahmetov et al. 2008, Doring et al. 2010, Cieszczyk et al. 2011, McPhee et al. 2011, Gabbasov et al. 2012), rs11549467 (Lin et al. 2013), rs1535679 (Hong et al. 2007), rs1957757 and rs1957755 (Sarzynski et al. 2010), rs2057482 (Hong et al. 2007), and rs28708675 (Prior et al. 2003)). Some polymorphisms were also associated with other phenotypes such as systemic sclerosis (rs12434438 (Wipff et al. 2009)), acute kidney injury (rs11549465 (Kolyada et al. 2009)), cellulite (rs11549465 (Emanuele et al., 2010)) and chronic obstructive pulmonary disease (rs11549465 (Putra et al. 2011a, Putra et al., 2011b)). Moreover, HIF1A SNPs were shown to be involved in metabolic disorders such as: diabetic nephropathy (rs11549465 (Gu et al. 2013)), which is a result of longstanding type 1 diabetes mellitus (T1DM) (rs11549465 (Nagy et al. 2009)) and type 2 diabetes mellitus (T2DM) (rs11549465 (Yamada et al. 2005, Nagy et al. 2009), rs12434438 (Yamada et al. 2005), and rs1319462 (Yamada et al. 2005)).

Most frequently studied missense SNPs within ODD domain

The review of genotype-phenotype studies revealed that HIF1A association studies were most often focused on two missense SNPs located within the ODD domain: rs11549465 (p.Pro582Ser; C1772T), and rs11549467 (p.Ala588Thr; A1790G). For these two SNPs opposing associations with phenotype were found. In addition, both rs11549465 and rs11549467 are germline SNPs within the ODD/pVHL interaction domain (Clifford et al. 2001). However, rs11549465 is also known to have an ability to enhance transactivation (Huang et al. 1998, Tanimoto et al. 2003), but was not identified as a site for HIF1A hydroxylation and is not mediating pVHL binding (Yamada et al. 2005). Moreover, these results are in agreement with a previous study by Percy et al. (Percy et al. 2003), who showed that the substitution p.Pro582Ser in vitro does not prevent VHL binding to a fragment of HIF1A after hydroxylation at p.Pro564. In the present study we have also collected published negative associations between polymorphisms and phenotypes, since they could be tested also in other populations or in higher number of subjects of the same populations to re-evaluate significance.

As in most association studies, also HIF1A association studies focused on missense SNPs and much less attention was given to synonymous (sSNP) and other non-coding SNPs. This could reflect a long-term assumption that sSNPs are inconsequential, since the primary amino-acid sequence of the protein stays unchanged. Although HIF1A association studies mostly focused on missense SNPs within the ODD domain affecting oxygen-dependent proteolysis, polymorphisms within other HIF1A domains have also been shown to affect HIF1A activity. Possible scenarios that may disrupt the role of HIF1A functional domains and therefore affect HIF1A stability and its role as the main TF in hypoxia are: (1) variations within the bHLH domain may prevent binding of HIF1A to HRE recognition sites within the promoter region of target genes, and likewise variations within HRE may create or destruct HRE binding sites for HIF1A to influence downstream targets; (2) variations within the PAS domain could affect dimerization with ARNT (HIF1B), which would result in HIF1A inability to function as a transcriptional regulator; (3) variations within the ODD domain could affect stability of a protein in normoxia, since conserved proline residues (p.Pro402 and p.Pro564) are usually targeted for VHL/proteasome degradation; (4) variations within the nuclear localization signal (NLS) may have an effect on translocation of HIF1A into the cell nucleus by nuclear transport; and (5) variations within N-TAD and C-TAD could influence transcriptional activation of HIF1A and interaction with its co-activators. Our literature review showed that HIF1A variations outside the ODD domain also have functional effects and were associated with diseases and phenotypes.

HIF1A and clinical trials

There are several clinical studies recorded in the database ClinicalTrials (https://clinicaltrials.gov/) that collates all publicly and privately supported clinical studies of human participants conducted around the world. Although 31 ongoing or already completed clinical trials included HIF1A in various cancers we found only one that specifically explores the effect of a HIF1A polymorphism in breast cancer (clinical trial identifier: NCT01935102; (Allegrini et al. 2014)). A missense SNP rs11549465 within the oxygen-dependent degradation domain that has also been extensively reviewed in our study, was tested in the aforementioned clinical trial. Specifically, the trial was aimed to identify interactions between the vascular endothelial growth factor A (VEGFA) SNPs with SNPs in several other genes: KDR (VEGFR2), CXCL8 (IL-8), HIF1A, HIF2A, and THBS1 (TSP-1) for possible differential bevacizumab response in a population of metastatic breast cancer. The multifactor dimensionality reduction (MDR) analyses found no interactions involving HIF1A genotypes. However, we need to add, that the sample size, especially for interaction analyses, was quite small, possibly leading to low power of statistical detection. For example, for HIF1A SNP rs11549465 the study involved only 8 individuals with the CC genotype, 100 heterozygotes and only 4 with the TT genotype. Nevertheless, the study was able to detect a significant interaction between KDR (VEGFR2) rs11133360 and CXCL8 (IL-8) rs4073 genotypes that identified a favorable genetic profile predicting a better therapeutic outcome. We can expect more such studies exploring HIF1A SNPs in clinical trials especially in terms of predicting their effect on favorable or unfavorable therapy outcomes in several other cancers mentioned also in our study. Apart from such pharmacogenetics clinical trials, we can anticipate also gene therapy trials to correct the most damaging germ line mutations in HIF1A as was recently shown for editing the sickle cell anemia mutation by the CRISPR/Cas9 approach (DeWitt et al. 2016).

Conclusions

Hypoxia is a characteristic feature for many pathological settings; hence, a stable HIF1A protein is critical for cells to adapt, thrive and survive in a hypoxic environment. The present HIF1A genotype-phenotype integrated review and manually curated graphical presentation could help researchers in better planning of future experiments as well as identifying traits and pathologies likely to be affected by HIF1A variability. Potential novel discoveries about functional HIF1A variations might be important to improve our understanding of HIF1A regulation. In addition, altered HIF1A regulation and its direct influence on various cellular pathways could better explain the role of HIF1A on a cellular level and its contribution to human complex diseases like cancer, diabetes, and obesity, or its contribution to the function of immune response and cardiovascular systems.