Angiogenesis-associated signaling is a fundamental component in the remodeling of the extracellular matrix in response to loading. Genes encoding protein components within this signaling cascade are therefore suitable candidates for investigation into ACL injury susceptibility: namely, vascular endothelial growth factor A isoform (VEGFA), kinase insert-domain receptor (KDR), nerve growth factor (NGF), and hypoxia inducible factor-1α (HIF1A). A case-control genetic association study was conducted on 227 asymptomatic control participants and 227 participants with surgically diagnosed ACL ruptures of which 126 participants reported a non-contact mechanism of rupture. All participants were genotyped for seven polymorphisms within the four genes. The VEGFA rs699947 CC genotype (p = 0.010, OR: 1.92, 95% CI: 1.17–3.17) was significantly over-represented within participants with non-contact ACL ruptures. The VEGFA rs1570360 GA genotype was significantly over-represented in the CON group (p = 0.007, OR: 1.70, 95% CI: 1.16–2.50). Furthermore, the KDR rs2071559 GA genotype was significantly over-represented in the female controls (p = 0.023, OR: 2.16, 95% CI: 1.11–4.22). Inferred haplotype analyses also implicated genomic regions spanning the VEGFA and KDR genes. These novel findings suggest that regions within VEGFA and KDR may be implicated in the pathophysiology of ACL ruptures; highlighting the potential biological significance of angiogenesis-associated signaling in the aetiology of ACL ruptures. © 2014 Orthopaedic Research Society. Published by Wiley Periodicals, Inc. J Orthop Res 32:1612–1618, 2014.
The angiogenesis-associated signaling cascade, which is regulated by several growth factors, has been implicated in matrix remodeling following mechanical loading. Increased levels of angiogenic cytokines and growth factors have been observed in ruptured tendons and ligaments[1, 2] and after cyclic stretching of tendon fibroblasts. Increased vascular endothelial growth factor (VEGF) levels, for example, have been observed in injured ligament and tendon tissues. VEGF is an essential regulator of angiogenesis, with the A isoform, encoded by VEGFA, believed to have the highest angiogenic potency. Most of the biological effects of VEGFA are mediated via its receptor: Kinase insert-domain receptor (KDR or VEGFR-2), encoded by KDR. VEGF and its receptor form a crucial regulatory system for angiogenesis, as demonstrated by gene knockout studies wherein VEGF heterozygous (+/−) mice and VEGFR-2 null mutant mice were embryonically lethal. Several growth factors have been shown to up-regulate VEGF levels, including nerve growth factor (NGF); a pleiotropic molecule with angiogenic properties. NGF is encoded by the NGF gene and increased NGF mRNA levels were previously reported in injured ACL tissue.
Furthermore, injury to the ACL may result in hypoxia and studies revealed ACL and tendon fibroblasts are sensitive to the effects of hypoxia.[2, 7] Responses to hypoxia are mediated by hypoxia inducible factor-1α (HIF-1α); encoded by HIF1A, which activates transcription of several genes to maintain tissue homeostasis, including those involved in angiogenesis. HIF-1α can also be induced by cyclic stretching.
The angiogenesis-associated signaling cascade is an important component of matrix remodeling following mechanical loading. It is proposed that polymorphisms within genes encoding for this signaling pathway may be, at least in part, responsible for inter-individual variation in responses to mechanical load and consequently connective tissue integrity and function. In relation to the anterior cruciate ligament (ACL) of the knee, any biological variation in the regulation of this cascade may therefore alter ACL injury risk. Because several polymorphisms within genes encoding proteins involved in other remodeling processes of the ECM, namely the matrix metalloproteinases (MMPs), have been associated with ACL injury risk, further investigation into the key signaling molecules that are induced after mechanical loading are required.
The aim of this study was, therefore, to identify if genomic loci spanning the biologically significant candidate genes VEGFA, KDR, NGFB, and HIF1A, involved in the angiogenesis-associated signaling pathway, are associated with ACL injury susceptibility. For this reason several functional polymorphisms, which have previously been implicated in multifactorial phenotypes VEGFA rs699947, VEGFA rs1570360, VEGFA rs2010963, KDR rs2071559, KDR rs1870377, NGFB rs6678788, and HIF1A rs11459465 were investigated for an independent and collective association (inferred haplotype analyses) with risk of ACL rupture.
A total of 454 physically active South African Caucasian participants (of self-reported ancestry) were included in this study. Participants were recruited during the period of June 2006 to November 2012 and were described and investigated in a previous study. Of these, 227 were physically active asymptomatic controls (CON group) and 227 participants had surgically diagnosed ACL ruptures (ACL group). ACL ruptures were diagnosed using clinical criteria and confirmed by ultrasound, magnetic resonance imaging (MRI), arthroscopy, or during surgery. These participants were recruited from the Sports Science Orthopaedic Clinic in Cape Town, South Africa. The mechanism of injury was defined according to the American Orthopaedic Society for Sports Medicine classification Scheme and was recorded for all participants in the ACL group after detailed explanation of injury event was reported to the investigator. Of the 227 participants in the ACL group, 126 participants reported a non-contact mechanism of injury (NON subgroup) and were analysed as a separate subgroup. Control participants, with no history of ACL injury, were recruited from sporting clubs and gyms within the Cape Town area. Participants were included in the study based on the previously described inclusion and exclusion criteria and all were requested to give informed written consent. Furthermore, each participant completed questionnaires specifying personal details, injury details, and medical and sporting history. Sports participation (Table S1) was categorized into contact sports, non-contact jumping sports, non-contact non-jumping sports, and skiing sports, as previously defined, with slight modification. Male participants were matched for participation in non-contact jumping sports and skiing sports. However, significantly more males in the ACL group played contact sports (p = 0.002) compared to the CON group. In addition, significantly more male controls participated in non-contact non-jumping sports (p = 0.036) than the ACL group. Female participants were matched for all sporting activity (Table S1). This study was approved by the Research Ethics Committee of the Faculty of Health Sciences within the University of Cape Town, South Africa (HREC 164/2006). This case-control genetic association study is in accordance with the recommendations of the STREGA Statement for reporting the results of genetic association studies.
Approximately five mililiters of venous blood was collected from each patient by venipuncture of the forearm vein, into an EDTA Vacutainer® tube (Becton Dickinson, Franklin Lakes, NJ). Blood samples were stored at − 20°C prior to total DNA extraction using the standard protocol as described by Lahiri and Nurnberger (1991) with slight modifications.
SNP Selection and Genotyping
Biologically significant genes involved in the angiogenesis-associated signaling pathway were chosen for analysis. Seven single nucleotide polymorphisms (SNPs) within four genes (Fig. S1) were selected based on their functional significance, previous associations with multifactorial phenotypes and having a reported minor allele frequency > 5% in the Caucasian population (National Centre for Biotechnology Information [http://www.ncbi.nlm.nih.gov/]). All samples (n = 454) were genotyped for all seven polymorphisms.
Standard genotyping protocols were employed: polymerase chain reaction (PCR) and restriction fragment length polymorphism (RFLP) analysis for the VEGFA rs699947 (BglII) and KDR rs1870377 (AluI) SNPs or catalogued TaqMan™ Genotyping Assays (Applied Biosystems, Foster City, CA) for the VEGFA rs1570360, VEGFA rs2010963, KDR rs2071559, NGFB rs6678788, and HIF1A rs11549465 SNPs. Details of the primers sequences and PCR conditions are available on request. For the TaqMan™ Genotyping Assays, PCR reactions were conducted using the Applied Biosystems StepOnePlus™ Real-Time PCR system and the Applied Biosystems StepOnePlus™ Real-Time PCR software v2.2.2 (Applied Biosystems). The manufacturer's instructions were followed for all the reactions.
Negative controls (no DNA) and five repeat samples (known genotypes) were included on every PCR plate (RFLP and TaqMan™ analyses) as a quality control measure for reliable genotyping and detection of contamination. Genotypes were confirmed by two independent investigators, with an average of 98.7% call rate. All laboratory work was conducted at the UCT/MRC Research Unit for Exercise Science and Sports Medicine within the University of Cape Town.
Power analysis was performed using QUANTO v1.2.4 (http://hydra.usc.edu/gxe) to calculate sample size for the study, as previously described. Data was analysed using STATISTICA v11 (StatSoft Inc., Tulsa, OK) and GraphPad Prism v5.02 (GraphPad Software Inc., San Diego, CA). Basic descriptive statistics were compared using one-way analysis of variance to determine any significant differences between the characteristics of the CON and ACL groups. Pearson's chi-squared (χ2) analysis and Fisher's exact test were used to analyse any differences in the genotype and allele frequencies. Hardy–Weinberg equilibrium probabilities were calculated using Genepop v4.2 (http://genepop.curtin.edu.au). The HAPSTAT v3.0 case-control haplotype inference package was used to infer haplotypes. Statistically significant differences were accepted when p < 0.05.
All the participants' characteristics were previously described in an earlier study investigating the same study participants (Table S2). There were no significant genotype effects on age, sex, height, weight, BMI, and country of birth, except for the VEGFA rs1570360 SNP on country of birth (p = 0.012) (Table S3). Details of the participants' injury history are presented in Table S4.
Genotype and Allele Frequencies
The VEGFA rs699947 genotype frequency was significantly (p = 0.034) different between the CON group and the NON subgroup (Table 1). The CC genotype was significantly under-represented (p = 0.010, OR: 1.92, 95% CI: 1.17–3.17) in the CON group (19%), compared to the NON subgroup (32%). Furthermore, the VEGFA rs699947 A allele was significantly over-represented in the CON group (53%) compared to the NON subgroup (44%) (p = 0.032, OR: 1.40, 95% CI: 1.03–1.91). No significant differences (p = 0.055) were noted in the genotype distributions for VEGFA rs699947 between the CON and ACL groups when all males and females were compared although there was a trend for the CC genotype to be under-represented (p = 0.017) in the CON group (19%) compared to the ACL group (29%). The VEGFA rs1570360 genotype frequency was significantly (p = 0.023) different between the CON and ACL groups (Table 1). The GA genotype was significantly over-represented (p = 0.007, OR: 1.70, 95% CI: 1.16–2.50) in the CON group (48%) compared to the ACL group (35%). There was a trend towards significance (p = 0.087) in the genotype frequencies between the CON and NON groups, with the GG genotype significantly under-represented (p = 0.031) in the CON group. No other significant differences in genotype and allele frequency distributions were noted between the CON versus ACL groups and CON group versus. NON subgroups for any of the other SNPS investigated in this study (Table 1). The VEGFA rs1570360 and KDR rs1870377 polymorphisms deviated from HWE within the ACL group (Table 1) however; all other polymorphisms were in HWE for all the participant groups.
|CON||ACL||p valuea||NON||p valueb|
|CC||19 (44)||29 (65)||0.055||32 (40)||0.034|
|CA||55 (125)||48 (106)||48 (60)|
|AA||25 (57)||23 (52)||21 (26)|
|A allele||53 (239)||47 (210)||0.083||44 (112)||0.032|
|GG||42 (95)||51 (109)||0.023||55 (63)||0.087|
|GA||48 (108)||35 (75)||37 (42)|
|AA||9 (21)||13 (28)||9 (10)|
|A allele||33 (150)||31 (131)||0.414||27 (62)||0.083|
|GG||46 (104)||42 (96)||0.667||38 (48)||0.171|
|GC||44 (99)||45 (101)||45 (57)|
|CC||11 (24)||13 (29)||17 (21)|
|C allele||32 (147)||35 (159)||0.373||39 (99)||0.065|
|GG||25 (56)||30 (67)||0.465||32 (40)||0.339|
|GA||52 (117)||47 (106)||45 (57)|
|AA||24 (54)||23 (53)||23 (29)|
|A allele||50 (225)||47 (212)||0.424||46 (115)||0.317|
|TT||61 (136)||61 (134)||0.734||62 (77)||0.887|
|TA||33 (73)||31 (69)||31 (38)|
|AA||6 (14)||8 (18)||7 (9)|
|A allele||23 (101)||24 (105)||0.695||23 (56)||0.984|
|CC||54 (121)||50 (113)||0.717||49 (61)||0.712|
|CT||36 (82)||40 (90)||40 (50)|
|TT||10 (22)||9 (21)||10 (13)|
|T allele||28 (126)||29 (132)||0.628||31 (76)||0.461|
|CC||79 (179)||80 (182)||0.613||79 (100)||0.483|
|CT||20 (46)||18 (41)||18 (23)|
|TT||1 (2)||2 (4)||2 (3)|
|T allele||11 (50)||11 (49)||0.915||12 (29)||0.842|
When only the male participants were compared, no significant differences in genotype or allele frequency distributions were noted for any of the other polymorphisms analysed (Table S5). All the polymorphisms were in HWE in all three male groups, except for VEGFA rs1570360, which deviated from HWE within the ACL group. When only the female participants were analysed (Table S6), the KDR rs2071559 genotype frequency distribution was significantly different (p = 0.033). The GA genotype was significantly over-represented (p = 0.023, OR: 2.16, 95% CI: 1.11–4.22) in the female CON group (57%) compared to the female ACL group (38%). No other significant differences were observed in the female participants. All three female groups were in HWE, except the female NON subgroup at the KDR rs1870377 locus. The genotype and minor allele frequency distributions were similar to those reported for other European populations (http://www.ncbi.nlm.nih.gov/snp/).
Inferred haplotypes were constructed for the VEGFA gene using the genotype data (rs699947 C/A, rs1570360 G/A and rs2010963 G/C) for all groups. Of the eight possible combinations only four (C-G-G, C-G-C, A-G-G, A-A-G) were inferred at a frequency greater than 2%. The A-A-G haplotype was significantly over-represented (p = 0.017) in the CON group (33%) compared to the NON subgroup (26%) (Fig. 1A). In the female participants, the C-G-C haplotype was significantly under-represented (p = 0.030) in the CON group (30%) compared to the NON subgroup (41%) (Fig. 1C). No significant differences were observed in the frequency distribution in the male participants (Fig. 1B).
Four possible inferred haplotypes were constructed from the KDR polymorphisms (rs2071559 A/G and rs1870377 A/T). The G-A haplotype was noted to be significantly under-represented in the controls (7%) compared to the ACL group (12%) (p = 0.014), as well as the NON subgroup (12%) (p = 0.001) (Fig. 2A). Similar findings were observed in the female participants, the G-A haplotype was significantly under-represented in the controls (3%) compared to the ACL group (14%) (p = 0.036), as well as the NON subgroup (16%) (p = 0.003) (Fig. 2C). Similarly, the G-A haplotype was significantly under-represented (p = 0.035) in the male controls (9%) compared to the NON subgroup (11%) (Fig. 2B).
Angiogenesis-associated signaling is a fundamental component in remodeling of the extracellular matrix of the ligament after mechanical loading. The aim of this study was to explore seven functional polymorphisms within four genes encoding biologically significant signaling molecules involved in the angiogenesis-associated cascade (VEGFA rs699947, VEGFA rs1570360, VEGFA rs2010963, KDR rs2071559, KDR rs1870377, NGFB rs6678788, and HIF1A rs11459465) with risk of ACL injury. The main findings of this study were: (i) VEGFA rs1570360 was independently associated with ACL ruptures in all participants, (ii) VEGFA rs699947 was independently associated with risk of non-contact ACL ruptures, (iii) KDR rs2071559 was associated with risk of ACL rupture in female participants, and (iv) haplotype analyses have implicated specific genomic intervals spanning the VEGFA and KDR genes to be associated with risk of ACL injury.
Vascular endothelial growth factor is a key component of the angiogenesis-associated signaling cascade. Loss of a single copy of the VEGF gene results in embryonic lethality and VEGF has been implicated in several pathological conditions, including rheumatoid arthritis. In vitro studies have shown VEGF protein expression to be influenced by genetic variation in the VEGFA gene. The C allele of the functional rs699947 (-2578 C > A) promoter SNP was previously linked with enhanced VEGF protein expression. Moreover, the rs699947 SNP is in complete linkage with rs35569394, an 18bp insertion/deletion polymorphism at position -2549 of the VEGFA promoter. Individuals with the rs699947 C allele likely also have the 18bp deletion. The G allele of the functional rs1570360 (− 1154 G > A) promoter SNP is associated with increased VEGF production.
In the present study, the VEGFA rs699947 SNP was only independently associated with risk of non-contact ACL ruptures whereas the VEGFA rs1570360 SNP was independently associated with ACL ruptures in all participants. The VEGFA rs699947 CC genotype and C allele were significantly associated with a 1.9-fold and 1.4-fold increased risk of non-contact ACL ruptures, respectively. For VEGFA rs1570360, the GA genotype was associated with a 1.7-fold reduced risk of ACL ruptures. These associations are in agreement with reported biological functions of the C and G alleles of VEGFA rs699947 and rs1570360, respectively. An increase in VEGF expression potentially upregulates MMP expression levels which may ultimately compromise ECM homeostatic balance thereby altering the biomechanical properties of the ligament. Further research into these variants and the function of VEGF in the mechanism of ACL ruptures requires exploration. Nell et al. (2012) suggested a possible “Goldilocks effect” where either too much or too little expression of the protein may have detrimental consequences. The conservation and timing of the balance between stimulatory and inhibitory factors during matrix remodeling is critical to maintaining homeostasis.
The independent associations noted at the VEGFA loci were mirrored in the haplotype analysis conducted. A haplotype represents the co-segregation of more than one polymorphism; it has the ability to capture information from a larger genomic interval and is often more informative in the discovery of risk susceptibility regions. The C-G-C inferred haplotype of the VEGFA gene (rs699947 C/A, rs1570360 G/A and rs2010963 G/C) was significantly associated with risk of non-contact ACL ruptures in females. This haplotype contains all three alleles previously associated with an increase in the expression of VEGF.[18, 22] Conversely, the A-A-G inferred haplotype, associated with lower VEGF plasma levels, was significantly associated with decreased risk of non-contact ACL ruptures.
KDR is the receptor for VEGFA and the inability of VEGF to bind to its receptor may result in down-regulation of the angiogenesis signaling cascade. The GA genotype of KDR rs2071559 was associated with a 2.2-fold reduced risk of ACL rupture in females only. The G allele of KDR rs2071559 (− 604G > A) leads to structural alteration of a binding site in the promoter region reducing KDR transcription. The reason for this and other[11, 24] sex-specific associations remains unknown.
Haplotype analysis of the KDR gene (rs2071559 A/G and rs1870377 A/T) has further defined a region overlapping this gene to be associated with increased susceptibility; specifically the G-A inferred haplotype was significantly associated with increased risk of rupture in all participants, and in males and females separately. It has been reported that both SNPs have the potential to influence expression of KDR. As mentioned above, the G allele of rs2071559 was associated with reduced mRNA transcription while the A allele of the rs1870377 1719T > A SNP was associated with reduced VEGF-binding efficacy to KDR. One can, therefore, hypothesise that the G-A haplotype is defining a region within KDR to be involved in regulating the binding affinity of VEGF. Hence the functional effects of this genomic interval within KDR needs be explored.
NGF acts in concert with VEGF to regulate angiogenesis and hypoxia is a potent stimulator of VEGF expression in both normoxic and hypoxic conditions. Although, this study failed to identify independent associations between NGFB rs6678788 and HIF1A rs11549465 and ACL injury risk, these remain suitable genes requiring further investigation. Given the biological interactions between VEGF and KDR in regulating ECM homeostasis and the findings presented in this study, it would be feasible to investigate the cumulative biological effect of the genomic intervals defined in this study on ACL injury susceptibility. An understanding of how these key signaling molecules interact with each other within the angiogenesis cascade is essential for the effective design of appropriate therapeutic interventions. Further studies in independent populations are ultimately required to confirm the results presented here.
A limitation of this study is that the cases and controls were not matched for weight, which is a confounding variable and evidence shows human adipose tissue is a potent source of inflammatory interleukins and other cytokines. Moreover, interactions between inflammatory and angiogenic molecules have been recognized in the pathogenesis of many human diseases. Another limitation is that the genotyping data was not co-varied for the confounders identified in Table S2, namely age, sex, and weight. However, all data was stratified by sex. It is a limitation that sex was not matched between the CON and ACL groups when males and females were collectively analysed. Furthermore, sports participation data was self-reported and the male participants in the study were not matched for participation in contact and non-contact non-jumping sports. The mechanism of injury details were reported by the participants to the investigator. There is currently no technique to confirm the mechanism of injury and therefore this remains a limitation of the study.
In conclusion, this study provides novel evidence to implicate the genes encoding proteins involved in angiogenesis-associated signaling and matrix remodeling with risk of ACL rupture. The most interesting findings of this study are: (i) the independent associations of polymorphisms within the VEGFA and KDR genes and (ii) haplotype analyses identifying specific genomic intervals harbouring potential DNA motifs underlying susceptibility. These specific genomic regions should, therefore, be further explored to elucidate their role in the pathobiology of ACL ruptures.
The authors would like to thank Ms Sasha Mannion, Ms Melanie Hay and Dr Dion O'Cuinneagain for assistance with participant recruitment. MR was funded by the University of Cape Town and the National Research Foundation. MP was funded by the Thembakazi Trust. This research was funded in part by the University of Cape Town Research Council, the Medical Research Council and the National Research Foundation. Authors MC and AS have filed patents on the application of specific sequence variations (not included in this manuscript) related to risk assessment of ACL ruptures and Achilles tendinopathy.