Figures
Abstract
Left ventricular hypertrophy (LVH) is an independent risk factor for cardiovascular morbidity and mortality, and a powerful predictor of adverse cardiovascular outcomes in the hypertensive patients. It has complex multifactorial and polygenic basis for its pathogenesis. We hypothesized that rare copy number variants (CNVs) contribute to the LVH pathogenesis in hypertensive patients. Copy number variants (CNV) were identified in 258 hypertensive patients, 95 of whom had LVH, after genotyping with a high resolution SNP array. Following stringent filtering criteria, we identified 208 rare, or private CNVs that were only present in our patients with hypertension related LVH. Preliminary findings from Gene Ontology and pathway analysis of this study confirmed the involvement of the genes known to be functionally involved in cardiac development and phenotypes, in line with previously reported transcriptomic studies. Network enrichment analyses suggested that the gene-set was, directly or indirectly, involved in the transcription factors regulating the “foetal cardiac gene programme” which triggered the hypertrophic cascade, confirming previous reports. These findings suggest that multiple, individually rare copy number variants altering genes may contribute to the pathogenesis of hypertension-related LVH. In summary, we have provided further supporting evidence that rare CNV could potentially impact this common and complex disease susceptibility with lower heritability.
Citation: Boon-Peng H, Mat Jusoh JA, Marshall CR, Majid F, Danuri N, Basir F, et al. (2016) Rare Copy Number Variants Identified Suggest the Regulating Pathways in Hypertension-Related Left Ventricular Hypertrophy. PLoS ONE 11(3): e0148755. https://doi.org/10.1371/journal.pone.0148755
Editor: Ryuichi Morishita, Osaka University Graduate School of Medicine, JAPAN
Received: June 23, 2015; Accepted: December 20, 2015; Published: March 1, 2016
Copyright: © 2016 Boon-Peng et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Data Availability: We are in the midst of depositing the dataset to dbGaP. The data shall be made accessible upon request. Such clause will be made clear to the dbGaP repository, thus data are available upon request to the corresponding author of this study: Professor Khalid Yusoff (VC@ucsiuniversity.edu.my).
Funding: This study is supported by the Fundamental Research Grant Scheme (FRGS) of Ministry of Higher Education Malaysia in 2007 (211501080005 [ST FRGS 1342]) and 2010 (600-RMI/ST/FRGS 5/3 Fst [61/2010]), 100-IRDC/BIOTEK 16/6/2 (13/2007) and 600-RMI/LRGS 5/3 (2/2011). BP Hoh was awarded with Dr Ranjeet Bhagwan Singh International Fellowship (Academy of Science Malaysia). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Competing interests: ACL holds a NeuroDevNet doctoral fellowship. SWS holds the GlaxoSmithKline-CIHR Chair in Genome Sciences at the University of Toronto and The Hospital for Sick Children. This does not alter the authors' adherence to PLOS ONE policies on sharing data and materials.
Introduction
Cardiovascular diseases remain the most significant cause of mortality globally, in high- and middle- or low-income countries [1]. Hypertension is the main driver for this epidemiologic reality [2,3]. Left ventricular hypertrophy (LVH) is a common outcome of hypertension, especially when uncontrolled whereby the LV wall thickens and/or LV mass increases in response to the biomechanical stress rendered by the elevated blood pressure, which is initially compensatory to the wall stress [4–6]. Hypertension related LVH is a complex, multifactorial and polygenic pathophysiologic condition. Almost a third of hypertensive patients develop LVH, [7], despite controlled with anti-hypertensive medications in more than half of them [8]. Thus while elevated blood pressure may explain the development of LVH, there may be other contributory factors. Hereditary may contribute up to 60% of this risk for developing LVH [9,10]
Although LVH can be reversed by pharmacological control of blood pressure such as losartan, identifying those at risk of developing LVH may have significant impact on the prognosis of patients with hypertension by providing a means of prevention of its development. Hence identifying the causative genes and / or the core biological pathway(s) leading to pathogenesis of LVH in hypertension is crucial in addressing such question.
At present, the fundamental hypothesis for genetic influence on complex diseases predominantly lies on the “common disease–common variant (CD-CV)” model [11,12] in which a disease trait is caused by a combination of common alleles (defined as ≥ 5% in a population), each contributing modest additive effects. Although several Genome Wide Association (GWA) studies have been conducted, only a handful of SNPs associated with left ventricular hypertrophy (LVH) have been identified [13,14] (https://www.genome.gov/). A number of genes known to play a role in the susceptibility of hypertension related-LVH such as angiotensin converting enzyme have not been able to be detected [15–17]. While another major candidate gene calcineurin has recently been reported in animal model [18], but it was not found in GWA studies for human. Recently an alternative hypothesis has been proposed to explain the failure to detect associations, namely the “missing heritability”, [11] where rarer variants are believed to carry a relatively larger effect on complex disease susceptibility. In an attempt to map the susceptible genes of hypertensive LVH, we adopted an alternative approach, with a postulation that some variants predisposing to hypertension related LVH are highly penetrant, individually rare or population specific, and of recent origin, even specific to single case [19,20].
Materials and Methods
Sample recruitment
A total of 116 blood samples of the hypertensive subjects were recruited from the PURE (Prospective Urban-Rural Epidemiologic) / REDISCOVER (REsponDing to IncreaSing CardiOVascular disEase pRevalence) Study from 2007 to 2010 carried out in Malaysia [21]. We defined the control group as those hypertensive patients without LVH; while cases were defined as those hypertensive patients with LVH. The following inclusion criteria were used for sample recruitment:
- 30–60 years of age
- Hypertension defined as systolic blood pressure ≥140 mmHg and/or diastolic blood pressure ≥90 mmHg
The exclusion criteria of the study were:
- Prescribed anti-hypertensive drugs at the time of enrolment
- Smoker
- Alcohol drinkers
- Diabetes
Echocardiography
Echocardiographic measurements were made using the Echo Pac in the Non Invasive Cardiac Lab, Universiti Teknologi MARA, Faculty of Medicine Selayang Campus. Doppler, two-dimensional (2D), and M-mode (2D-guided, in the parasternal short axis view) echocardiograms were performed using a standard protocol. Measurements were made using a computerized review station equipped with a digitizing tablet and monitor overlay for calibration and quantification.
Transthorasic echocardiogram criteria for LV mass index used the formula: LV mass index = (0.8 (1:04 ([LVIDD PWTD IVSTD] 3—[LVIDD] 3) (0.6g / Height2) (Devereux Criteria), where:
- LVIDD = Left ventricular internal dimension in diastole
- PWTD = Posterior wall thickness at end in diastole
- IVSTD = Interventricular septal thickness at end-diastole
Subjects were diagnosed as LVH when Left Ventricular Mass Index (LVMI) exceeded 110 g/m2 in women and 125 g/m2 in men.
All subjects provided written informed consent. Ethics approval was obtained from ethics committees of the Universiti Teknologi MARA (UiTM)[REC/UITM/2007(10)].
Microarray analysis
Genomic DNA was extracted either from whole, or clotted blood using commercially available kit. Genotyping was carried out with the Illumina Human 660W-Quad Beadchip (San Diego, SA, USA). Briefly, 500 ng of genomic DNA was denatured overnight, and enzymatically fragmented, precipitated with isopropanol, centrifuged at 4°C, and resuspended in hybridization buffer. All beadchips were prepared for hybridization in a capillary flow-through chamber. Samples were loaded to beadchips and incubated overnight in the Illumina Hybridization Oven. Unhybridized or non-specific products were washed, and beadchips were preceded with staining and extension. The beadchips were scanned on the Illumina Beadarray Reader using default settings, and intra-chip normalization was performed using Illumina Genome Studio with a GenCall cut-off point 0.1 and call rare cut-off of 99%. Built-in controls–both sample dependent and sample independent, were inspected to assess the quality of the experiment.
CNV detection, quality control and analysis of rare CNV
The Log R ratio (LRR) and B allele frequencies (BAF) were first exported from Genome Studio (Illumina). The Illumina cluster file comprising >120 HapMap samples was used as reference to generate intensities and genotypes. Stringent criteria of quality control were applied to the array [22–24]. Samples were excluded if: (i) genotype call rate of <99%; (ii) LRR values with a standard deviation above 0.35; (iii) standard deviation for B allele frequencies of >0.13; (iv) cross samples batch normalized ratio standard deviation >0.27.
Samples passing QC were carried out for further analysis. CNVs were called using three independent algorithms: CNV partition v2.3.4 (Genome Studio, Illumina), PennCNV (Wang et al., 2007) and iPattern (The Centre for Applied Genomics, Toronto). The application of multiple algorithms minimizes the number of potential false positive discoveries; and thus increases the chance of obtaining more high confidence calls [24].
CNV analyses were performed using the original array coordinates based on Human Genome Assembly NCBI (Build hg18). We applied stringent filtering criteria in CNV analysis by excluding CNVs calls with: (i) less than five consecutive probes; (ii) located in regions with high GC content (>70%); (iii) approximately 30 kb adjacent to the centromere or telomeres; (iv) size less than 1 Kb; (iv) sex chromosomes; and (v) called by only one out of the three algorithms.
In addition, manual visual inspection was used to exclude potential false positives (typically >1 Mb) due to unknown artefacts. We excluded samples outliers with respect to executive aggregate length of CNVs.
CNVs that passed all QCs were considered rare or novel if they: (i) did not overlap with any known copy number polymorphism (frequency >1%); (ii) had <50% reciprocal overlap (by length) with CNVs reported in Database for Genomic Variants (DGV), whereby the ‘case’ CNV overlapped with at least 50% of the control CNV and the conversely, the control CNV overlapped with at least 50% of the case CNV [23]; (iii) occurred as singleton in the 116 samples genotyped in this study. In other word, this means the rare CNVs in this study are uniquely identified for both length and locus, via the comparison to known variants with less than 50% overlap. Recurrent CNVs specific to case group in this study were also identified.
The rare and/or recurrent CNVs specific to the case group identified in this study were further assessed by comparing to control sample datasets from HapMap3 and subsequently from Singapore Genome Variation Project (SGVP, http://www.statgen.nus.edu.sg/~SGVP/) as the population matched controls. CNVs in cases with >50% reciprocal overlap to these control datasets were excluded. We limited our rare CNVs cut-off size of >1 kb, instead of >30 kb suggested by most authors [19,20,24].
Replication study
We replicated the study on additional 143 samples, consisting 51 case and 92 controls. CNV typing was carried out using Illumina OmniExpress (San Diego, SA, USA), comprising >750,000 SNV probes according to the manufacturer’s protocols. Criteria for CNV calls were as mentioned above. The CNVs were called using PennCNV, QuantiSNP [25] and iPattern.
Pathway analysis
Gene Ontology (GO) analysis was carried out using, the DAVID (Database for Annotation, Visualization and Integrated Discovery, version 6.7) (david.abcc.ncifcrf.gov), Ingenuity (http://www.ingenuity.com/) and GeneGO Metacore (https://portal.genego.com/). An interaction network was generated on the”case-specific” genes, using MetaCore (GeneGo).
qPCR Validation
Candidate genes of interest harbouring the rare CNVs were validated by quantitative Real-Time PCR (qPCR) SyBr Green assay. Primers were designed using Primer3 (http://frodo.wi.mit.edu/primer3/) and checked with UCSC Genome Browser. Detailed information of the primer designed is shown in S1 Table. Normalization to the control gene Forkhead Box P2 (FOXP2) (primers: 5'-TGACATGCCAGCTTATCTGTTT-3' and 5'-GAGAAAAGCAATTTTCACAGTCC-3') was used to give an estimate of copy number.
Results
Clinical demographic data
Table 1 shows the clinical and phenotypic data of the recruited subjects in the stage 1 study. Forty-four subjects were diagnosed as hypertension with LVH (denoted herein as case), while 72 were hypertension without LVH (denoted herein as control). There was no significant difference between the case and control groups with regards to ethnicity, age, systolic and diastolic blood pressure. As expected, significant differences were observed between case and control groups with regards to LV mass, IVSD and LVMI, indicating well characterized samples between cases and controls. BMI showed a significant difference between the case and controls, as expected. Males are significantly higher in case group than the females in our study cohort. This is in line with the previous studies, suggesting gender differences in the pathogenesis of LVH [26].
Characterization of CNVs
Fig 1 describes the experimental workflow for identification of CNV specific to subjects hypertension related LVH. Of the 116 hypertensive subjects studied, 29,472, 33,473 and 33,203 CNVs were called by CNVPartition, PennCNV, and iPattern, respectively (Table 2). A total of 22,337 CNVs were successfully merged with at least two algorithms (referred as “stringent calls”), corresponding to an average of 202.9 CNVs per genome, with a median size of 3,893 bp (average size 19,428 bp) (Table 2). The number of CNVs per genome was relatively higher than previous reports [22,27,28] but in line with Pinto et al. (2011) which reported an average of 240 calls genome for the Illumina 660W platform. This could be due to several reasons: (i) different platforms utilized for CNV detection and its resolution; ii) levels of QC stringency applied during CNV call; iii) the algorithms applied when performing CNV call; [29]. Of particular note, the Illumina 660W is a platform seeded with probes to allow detection for common CNVs, thus expected to have higher sensitivity. A total of 1,973 CNVs unique to our study subjects remained after applying the QC filtering steps by excluding known polymorphic CNVs (copy number polymorphism, CNP), calls adjacent to telomere and centromere regions, and those reported in DGV and SGVP. The pre-defined criterion of 50% reciprocal overlapped was applied, whereby a CNV that was at least 50% unique by length when compared to every CNV in the control datasets was taken as putative novel [23]. This included 851 CNVs in cases (74 gain, 777 loss) and 1,122 in controls (162 gain, 960 loss). The length distribution of the CNVs observed is shown in Fig 2.
qPCR was performed as an independent technical validation. FOXP2 was used as the reference gene. Results of the qPCR validation are shown in S1 Table. Seven out of 9 (77%) of the loci validated were positive calls. Candidate genes underlying these CNVs were excluded in the subsequent pathway analysis.
Case- and control-specific CNV
We further identified from the dataset, 208 CNVs specific to cases (35 gains; 173 losses) and 283 specific to controls (75 gains; 208 losses), which corresponded to an average of 4.72 and 3.93 CNV per genome in case and control groups, respectively. The overall CNV length distribution between case and control was similar between the cases and controls (average length 50,152.94 bp vs 53,194.88 bp) (P = 0.163) suggesting that CNV length may have minimal impact to LVH development. Recurrent CNVs for gene LOC348021, was observed in 3 cases; while CDH15 and KCNIP4 were observed in 2 cases, respectively (Fig 3, S2 Table). We performed targeted dosage analysis with qPCR for these candidate genes in additional 36 cases recruited from the same cohort, but no CNVs were detected.
Figures produced with custom tracks listing CNV calls and uploaded to http://genome.ucsc.edu. The hypertensive LVH cases are designated as black bars.
Gene Ontology and Pathway Analysis
Gene ontology and pathway analyses were then performed on the genes specific to case using 3 approaches.
DAVID pathway analysis
We first performed analysis with DAVID. It is a publicly available threshold-based ontology analysis, which provides extensive option for interrogation of approximately 40 databases including KEGG, Interpro etc [30]. Results presented by DAVID are based upon the use of functional annotation clustering and the top five single terms among clusters will be reported.
Forty two case-specific genes were found involve in “ion binding” (GO:0043167) (P = 0.029); whereas “translational initiation” (GO:006413) (P = 0.003) and “translational factor activity / nucleic acid biding” (GO:0008135) (P = 0.037) were presented as the most significant GO processes (S3 Table).
Ingenuity Pathway Analysis (IPA)
IPA is a JAVA based commercialized web-based system. Its annotations of genes and pathways are based on its own database built on the findings from various literatures and publicly available databases including GO and Entrez Gene. Enrichment analyses of IPA are carried out using the right-tailed Fisher’s exact test and Benjamini-Hochberg multiple testing corrections [30].
The “case-specific” genes identified in this study were enriched in diseases involved in inflammatory response (P = 3.43E-04), cancer (P = 2.87E-03), respiratory disease (P = 2.87E-03) connective tissue disorder (P = 5.56E-03) and skeletal and muscular disorders (P = 5.56E-03) (S4 Table). Functional enrichment analyses of IPA revealed top two most significant enriched genes in cell-to-cell signalling interaction (P = 3.43E-04–4.65E-02), and tissue development (P = 1.04E-04–4.91E-02) (S3 Table). Network analysis revealed UBC signalling as the major interacting network (Fig 4).
Network was generated by Ingenuity Pathway Analysis by merging up the top canonical pathways with default settings. Network analysis revealed UBC signalling as the major interacting network.
GeneGO MetaCore analysis
GeneGO Metacore pathway analysis is a commercial web-delivered application based on unique, cur-rated database from Thompson Reuters containing approximately 20 validated functional ontologies that can be used for filtering and enrichment, as well as interactive canonical pathways capturing ~200,000 pathways.
Metacore analysis identified 69 genes most significantly involved in heart development (P = 1.47E-57); whereas another two most relevant tissue groups involved were skeletal muscle (67 genes; P = 1.64E-52), and smooth muscle (66 genes; P = 7.72E-52). The GO process were most significant with: “regulation of purine nucleotide metabolic process” (P = 1.28E-06), “regulation of neuron migration” (P = 5.00E-06), “intracellular signal transduction” (P = 1.22E-05), “negative regulation of renin secretion into blood stream” (P = 4.03E-05), and “hydrogen peroxide catabolic process” (P = 4.35E-05).
We then predicted the most probable networks that may involve in LVH pathogenesis by pooling the 3 most enriched pathways from Metacore canonical network analysis. It was found that the gene-set was, directly or indirectly related to the transcription factors Sp1, p53 and CREB1, and androgen receptor signalling cascades (Fig 5).
Network was generated by GeneGo Metacore pathway analysis programme by merging the top canonical pathways with default settings.
Replication
The subsequent replication with an independent group of samples was carried out using Illumina Omni Express SNP microarray, and the CNV was called using the stringent criteria pipeline is as described above, except that the CNVpartition was replaced by QuantiSNP, owing to its sensitivity in identifying CNVs from this platform (only a total of 829 CNVs were identified using CNVpartition from Genome Studio).
Rare CNVs were detected in 128 out of the 148 subjected included in the replication study. An average of 11 CNVs per genome were identified, ranging from 2 to 27 CNVs per genome (Table 3). The lower number of CNV detected in the replication stage is expected, since Omni Express was not designed for the detection of CNV, unlike the 660W which was a hybrid platform (i.e. seeded with probes for common CNV detection), thus lower power in CNV detection. A total of 303 CNVs unique to our study subjects remained after applying the QC filtering steps as mentioned in the methodology. These included 111 CNVs specific in cases (55 gain, 56 loss; 2.52 CNVs per genome) and 113 control specific CNVs (80 gain, 112 loss; 1.35 CNVs per genome) (P = 0.1895) (S6 Table). qPCR validation on 5 randomly selected CNVs identified in this replication stage revealed 100% positive calls (S1 Table).
Notably, SUMF1, F2R and IQGAP2 (found in the case group) and ACSF3 (found in the control group) identified in the initial study were replicated in the replication cohort, suggesting potential role of these genes in pathophysiology of hypertension related LVH.
Gene Ontology and pathway enrichment analyses were carried out using DAVID and IPA. This analysis was carried out in 2 stages: (i) we first assessed the gene list revealed from the replication cohort; (ii) subsequently the gene list was combined from the “Illumina 660W” cohort and the “Illumina Omni Express” replication cohort (S7 Table). DAVID pathway analysis revealed the “case-specific” genes significantly enriched in EGF signalling (IPR:013111) (P = 0.0179); and “activation of protein kinase activity (GO:0032147)” (P = 0.0064). When the two stages were combined, EGF-like domain (IPR:006210) (P = 1.36E-03) remained the most significant enrichment; whilst the category “ion-binding” (GO:0043167) comprised the most number of genes (P = 0.040; 53 genes).
IPA analyses revealed significant enrichment in diseases related to infectious diseases (P = 7.73E-07–4.77E-02) and respiratory diseases (P = 7.73E-02–4.10E-02), suggesting disease mechanism may be related to immune inflammatory response, in line with the finding during the earlier stage (S4 Table). Interestingly, the most significant enrichment on physiological system development and function was “cardiovascular system development and function” (P = 7.48E-04–3.77E-02) (S7 Table). When merging the top 3 networks from IPA, Ubiquitin C (UBC) related signalling appeared to be the major network (Fig 6), and remained as the major interacting network analysis when gene lists from both stages were combined (Fig 5).
Network was generated by Ingenuity Pathway Analysis by merging up the top canonical pathways with default settings. Network analysis revealed UBC signalling as the major interacting network.
We note that some of the genes observed might not have any direct functional involvement with LVH or hypertension. However, collective findings from this study was connected though via biological interactions with numerous genes, many of which are known to be functionally involved in cardiac phenotypes in “foetal gene programmes” [31–33].
To ensure that the pathways and ontologies identified are unique in case, we performed a pathway analysis on the candidate genes identified in the control groups using DAVID. Apparently none of the pathways or ontologies identified were similar to the case groups (S9 Table).
Discussion
In this study we applied an alternative approach for discovery of candidate genes involved in hypertension related-LVH. To our knowledge, this is the first report to evaluate the impact of rare CNVs on hypertension related-LVH conducted in the Southeast Asia population. We have provided further supporting evidence to show that rare CNV may have impact on common and complex disease susceptibility [22,27,34]. In addition our results also supported the previously reported signalling pathways for the development of LVH in the hypertensive subjects [5,31,35].
Several candidate genes were of interest and postulated to play a role in the LVH development. The coagulation factor II, F2R, is a G-protein coupled receptor family member (Fig 3, S2 Table). This gene was thought to function as a transmembrane receptor involved in the regulation of thrombotic response (http://www.ncbi.nlm.nih.gov/gene/2149), and may play a role in platelet activation and vascular development. Studies found that F2R interacts with IL-6 in the susceptibility of myocardial infarction [36,37], and coronary heart disease in hypertensive patients [38]. However its role in the inflammatory mechanism of cardiac hypertrophy is not well understood. IQGAP2, a neighbouring gene of F2R, binds with CALM1 [39] and it regulates cell morphology and motility via interactions with components of cytoskeleton, cell adhesion and several other signalling molecules (http://www.ncbi.nlm.nih.gov/gene/10788), hence believed to be involved in the process of cardiac development. A recent study on whole genome exome sequencing reported a mutation of IQGAP2 that caused LVH (Zhi et al., 2012). Both F2R and IQGAP2 were found recurrent in our replication cohort. KCNIP4, a member of the voltage-gated potassium (Kv) channel-interacting protein family, has been reported to be associated with ischemic heart disease [40]. It was found to be associated with LVMI amongst the isolated Amish cohort in a genome-wide association study, though the signal was not significant in subsequent replication study [41]. KCNIP4 is essential to the low repetitive firing and back propagation of action potential in neurons and shapes the action potential in the heart [42].
We observed several recurrent case specific CNVs, and attempted to further replicate these CNVs with additional 36 independent samples with qPCR on top of the 148 samples replicated with Illumina Omni Express. None of these samples however, were copy number variable, thus suggesting that these case-specific CNVs may be at a very low frequency. This postulation is indeed reasonable, considering that most rare / de novo CNVs occur less than 1% in a population [43–45].
Research of life sciences has shifted from gene identification to gene annotations including functions, interactions and the involvement of pathways [22,30,46–48]. Our exploratory data support the findings implicating the immune system and inflammation pathways in the aetiology of stressed induced cardiac hypertrophy [31,49]. At present, there is no single specific signalling pathway that could explain entirely the functional changes leading to cardiac hypertrophy. However, a number of transcription factor signalling pathways have been suggested [50]. Among these, the involvement of cAMP response element binding protein 1 (CREB1) transcription factor signalling as an important regulator in cardiac hypertrophy has been well acknowledged [14,51,52].
Specificity protein 1 (SP1) transcription factor and its signalling pathway is a major component during the foetal stage of cardiac development in human. It was identified as a transcriptional regulatory mechanism in the reduction of foetal metabolic programme during pressure overload induced cardiac hypertrophy [32].
Transcription factor p53 and its signalling are known to mediate apoptosis induced by multiple stresses [53], and crucially involved in cardiac hypertrophy [54].
The influence of androgen receptor signalling in cardiac hypertrophy has been proposed earlier [55–57], either by acting directly on the heart or by affecting the vascular system [58–61]. However, little is known about its role and the underlying molecular mechanism in cardiac hypertrophy as findings remain contradicting [28,55,57,62]
The Ubiquitin C (UBC) via its ubiquitination activities has been reported to play numerous physiological functions including that causing hypertrophic response [63–65]. The promotion of ubiquitination activities via expression of Atrogin-1 (FBXO32) and repression of calcineurin A apparently leads to inhibition of cardiac hypertrophy [66].
Numerous studies have provided evidence that induction of Ang II promotes the growth of cardiomyocytes via transactivation of EGF signalling subsequently activation of MAPK signalling [67]. It is interesting to note that transactivation EGF receptor signalling by AT1R activates CREB1 [68,69].
Several potential limitations have been identified in this study. First, relatively a small sample size, therefore statistical and functional analyses could not be carried out. Secondely, rare CNVs (defined as frequency <1%) could hardly reach a significant number eligible for statistical analyses [70]. Therefore it is difficult to replicate the findings in independent studies. Thirdly, false negative results probably due to stringent filtering criteria for CNV calls, thus potentially limiting discovery of novel and informative findings. On top of that, inclusion of smaller CNVs in this study. We included all CNVs sized >1 kb instead of >30 kb as practiced by most reports [20,22,27,28,34,71,72]. We believed that selection of larger CNVs might introduced potential bias to the impact of CNVs in disease pathogenesis especially in complex diseases with lower inheritability such as hypertension related LVH, and presuming that smaller CNV may well contribute equal impact with the larger ones especially when exons or splicing sites of a gene is being disrupted by CNV breakpoints. However, smaller CNVs causes higher false discovery rate due to poor signal to noise ratio, thus data should handle with caution. Lastly, The biological replication of the rare CNV was carried out using a different platfrom (Illumina OmniExpress, which was not meant for CNV detection), leading to a significant drop in CNVs calls per sample. (and decrease in unique CNVs), indicating that the replication was not as powerful as it was expected, therefore the lowered the power of CNV detection. However, this does not increase the false positive rate of this study, therefore the results reported are considerable accurate.
Despite this, our data was analysed with a most stringent QC criteria by comparing with several datasets including DGV, HapMap3 and the SGVP, involving more than 1,300 samples, and were subsequently selectively validated. As such, our CNV call is considered reliable. However, we acknowledge the potential constraint of these datasets as they were genotyped with different platforms therefore further interpretation of the finding should be taken cautiously.
It should be noted that this study however does not attempt to prove the involvement with the disease of any specific variants or even any specific genes or pathways. Rather, it provides insights that common diseases like hypertension related LVH can be affected by rare CNVs that influence or disrupt the underlying genes in the relevant pathways. In other word, rare CNVs identified in this study are not proven to be the causative variants, rather they may contribute as a portion of risk factor to the common disease susceptibility such as hypertension-related LVH. An independent cohort with larger number of samples is crucial to warrant the findings of this study. Functional studies characterizing the role in cardiac development of genes within these rare CNVs are a priority to illuminate the mechanisms of LVH development in hypertensive patients.
In summary, this study delivers two major messages. First, we show that rare variant plays a role in the susceptibility of common and complex diseases such as LVH. Indeed, using the rare CNV strategy, we have demonstrated further supporting evidence of the previously identified signalling pathways leading to cardiac hypertrophy. In particular, collective results of this study further support the activation of foetal cardiac gene programme during cardiac hypertrophy. Second, LVH is a complex event that depends on the activation of different signalling pathways. The finding of this work may eventually lead on to work concerning differences in response to drugs, which can prevent the development of LVH in patients with hypertension.
Supporting Information
S1 Table. Candidate genes primers sequences for SyBr Green qRT-PCR assay.
https://doi.org/10.1371/journal.pone.0148755.s001
(DOC)
S2 Table. Case- and control specific CNVs identified in the 116 hypertension related LVH subjects studied.
Chr, chromosome; hg18, human genome assembly 18 (March 2006). Dashes indicate that no gene is involved or disrupted by CNV breakpoints. Highlighted are genes with evidence for cardiovascular involvement.
https://doi.org/10.1371/journal.pone.0148755.s002
(DOC)
S3 Table. Gene ontology and pathway analyses identified in hypertension-related LVH.
https://doi.org/10.1371/journal.pone.0148755.s003
(DOC)
S4 Table. Top significant disease groups identified by Ingenuity (IPA) enrichment analysis.
https://doi.org/10.1371/journal.pone.0148755.s004
(DOC)
S5 Table. Top significant tissue groups identified by GO MetaCore enrichment analysis.
https://doi.org/10.1371/journal.pone.0148755.s005
(DOC)
S6 Table. Case- and control specific CNVs identified in the 116 hypertension related LVH subjects in the replication study.
Chr, chromosome; hg18, human genome assembly 18 (March 2006). Dashes indicate that no gene is involved or disrupted by CNV breakpoints. Highlighted are genes that are identified in the earlier stage of the study.
https://doi.org/10.1371/journal.pone.0148755.s006
(DOC)
S7 Table. Gene ontology and pathway analyses identified in the hypertension-related LVH from the replication study.
https://doi.org/10.1371/journal.pone.0148755.s007
(DOC)
S8 Table. Top 5 networks identified by Ingenuity (IPA) from the replication study.
https://doi.org/10.1371/journal.pone.0148755.s008
(DOC)
S9 Table.
Gene ontology and pathway analyses identified using DAVID in the hypertension patients without LVH (denoted as controls) from (i) Illumina 660W; (ii) Illumina Omni Express; (iii) combination from Illumina 660W and Illumina Omni Express datasets.
https://doi.org/10.1371/journal.pone.0148755.s009
(DOC)
Acknowledgments
This study is supported by the Fundamental Research Grant Scheme (FRGS) of Ministry of Higher Education Malaysia in 2007 (211501080005 [ST FRGS 1342]) and 2010 (600-RMI/ST/FRGS 5/3 Fst [61/2010]), 100-IRDC/BIOTEK 16/6/2 (13/2007) and 600-RMI/LRGS 5/3 (2/2011). BP Hoh was awarded with Dr Ranjeet Bhagwan Singh International Fellowship (Academy of Science Malaysia). S.W.S. holds the GlaxoSmithKline-CIHR Chair in Genome Sciences at the University of Toronto and The Hospital for Sick Children. This does not alter our adherence to PLOS ONE policies on sharing data and materials. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript. The authors wish to acknowledge the support from the University of Toronto McLaughlin Centre and The Hospital for Sick Children Foundation.
Author Contributions
Conceived and designed the experiments: HBP SWS KY. Performed the experiments: HBP JAMJ. Analyzed the data: HBP JAMJ CRM BT. Contributed reagents/materials/analysis tools: SWS KY. Wrote the paper: HBP CRM KY. Collected the samples and phenotypic data of the subjects: FM ND FB.
References
- 1. Reardon SF, Bischoff K (2011) Income inequality and income segregation. AJS 116: 1092–1153. pmid:21648248
- 2. Gersh BJ, Sliwa K, Mayosi BM, Yusuf S (2010) Novel therapeutic concepts: the epidemic of cardiovascular disease in the developing world: global implications. Eur Heart J 31: 642–648. pmid:20176800
- 3. Sliwa K, Stewart S, Gersh BJ (2011) Hypertension: a global perspective. Circulation 123: 2892–2896. pmid:21690504
- 4. Lorell BH, Carabello BA (2000) Left ventricular hypertrophy: pathogenesis, detection, and prognosis. Circulation 102: 470–479. pmid:10908222
- 5. Heineke J, Molkentin JD (2006) Regulation of cardiac hypertrophy by intracellular signalling pathways. Nat Rev Mol Cell Biol 7: 589–600. pmid:16936699
- 6. Meijs MF, de Windt LJ, de Jonge N, Cramer MJ, Bots ML, Mali WP, et al. (2007) Left ventricular hypertrophy: a shift in paradigm. Curr Med Chem 14: 157–171. pmid:17266575
- 7. Verdecchia P, Angeli F, Pittavini L, Gattobigio R, Benemio G, Porcellati C (2004) Regression of left ventricular hypertrophy and cardiovascular risk changes in hypertensive patients. Ital Heart J 5: 505–510. pmid:15487267
- 8. Ching S, Chia Y, Wan Azman W (2012) Prevalence and Determinants of Left Ventricular Hypertrophy in Hypertensive Patients at a Primary Care Clinic. Malaysian Family Physician 7: 3.
- 9. Chien KL, Hsu HC, Su TC, Chen MF, Lee YT (2006) Heritability and major gene effects on left ventricular mass in the Chinese population: a family study. BMC Cardiovasc Disord 6: 37. pmid:16945138
- 10. Deschepper CF, Boutin-Ganache I, Zahabi A, Jiang Z (2002) In search of cardiovascular candidate genes: interactions between phenotypes and genotypes. Hypertension 39: 332–336. pmid:11882568
- 11. Manolio TA (2009) Cohort studies and the genetics of complex disease. Nat Genet 41: 5–6. pmid:19112455
- 12. McCarthy MI (2009) Exploring the unknown: assumptions about allelic architecture and strategies for susceptibility variant discovery. Genome Med 1: 66. pmid:19591663
- 13. Loirand G, Pacaud P (2010) The role of Rho protein signaling in hypertension. Nat Rev Cardiol 7: 637–647. pmid:20808285
- 14. Muller FU, Boknik P, Knapp J, Linck B, Luss H, Neumann J, et al. (2001) Activation and inactivation of cAMP-response element-mediated gene transcription in cardiac myocytes. Cardiovasc Res 52: 95–102. pmid:11557237
- 15. Choudhary R, Sastry B, Subramanyam C (2005) Positive correlations between serum calcineurin activity and left ventricular hypertrophy. Int J Cardiol 105: 327–331. pmid:16274778
- 16. Molkentin JD (2004) Calcineurin–NFAT signaling regulates the cardiac hypertrophic response in coordination with the MAPKs. Cardiovasc Res 63: 467–475. pmid:15276472
- 17. Schunkert H, Hense H-W, Holmer SR, Stender M, Perz S, Keil U, et al. (1994) Association between a deletion polymorphism of the angiotensin-converting-enzyme gene and left ventricular hypertrophy. New England Journal of Medicine 330: 1634–1638. pmid:8177269
- 18. Rau CD, Wang J, Avetisyan R, Romay MC, Martin L, Ren S, et al. (2015) Mapping genetic contributions to cardiac pathology induced by Beta-adrenergic stimulation in mice. Circ Cardiovasc Genet 8: 40–49. pmid:25480693
- 19. Walsh T, McClellan JM, McCarthy SE, Addington AM, Pierce SB, Cooper GM, et al. (2008) Rare structural variants disrupt multiple genes in neurodevelopmental pathways in schizophrenia. Science 320: 539–543. pmid:18369103
- 20. Sebat F, Musthafa AA, Johnson D, Kramer AA, Shoffner D, Eliason M, et al. (2007) Effect of a rapid response system for patients in shock on time to treatment and mortality during 5 years. Crit Care Med 35: 2568–2575. pmid:17901831
- 21. Rasiah R, Yusoff K, Mohammadreza A, Manikam R, Tumin M, Chandrasekaran SK, et al. (2013) Cardiovascular disease risk factors and socioeconomic variables in a nation undergoing epidemiologic transition. BMC Public Health 13: 886. pmid:24066906
- 22. Silversides CK, Lionel AC, Costain G, Merico D, Migita O, Liu B, et al. (2012) Rare copy number variations in adults with tetralogy of Fallot implicate novel risk gene pathways. PLoS Genet 8: e1002843. pmid:22912587
- 23. Lionel AC, Crosbie J, Barbosa N, Goodale T, Thiruvahindrapuram B, Rickaby J, et al. (2011) Rare copy number variation discovery and cross-disorder comparisons identify risk genes for ADHD. Sci Transl Med 3: 95ra75. pmid:21832240
- 24. Pinto D, Darvishi K, Shi X, Rajan D, Rigler D, Fitzgerald T, et al. (2011) Comprehensive assessment of array-based platforms and calling algorithms for detection of copy number variants. Nat Biotechnol 29: 512–520. pmid:21552272
- 25. Colella S, Yau C, Taylor JM, Mirza G, Butler H, Clouston P, et al. (2007) QuantiSNP: an Objective Bayes Hidden-Markov Model to detect and accurately map copy number variation using SNP genotyping data. Nucleic Acids Res 35: 2013–2025. pmid:17341461
- 26. Schunkert H, Hense HW, Holmer SR, Stender M, Perz S, Keil U, et al. (1994) Association between a deletion polymorphism of the angiotensin-converting-enzyme gene and left ventricular hypertrophy. N Engl J Med 330: 1634–1638. pmid:8177269
- 27. Soemedi R, Wilson IJ, Bentham J, Darlay R, Topf A, Zelenika D, et al. (2012) Contribution of global rare copy-number variants to the risk of sporadic congenital heart disease. Am J Hum Genet 91: 489–501. pmid:22939634
- 28. Glessner JT, Reilly MP, Kim CE, Takahashi N, Albano A, Hou C, et al. (2010) Strong synaptic transmission impact by copy number variations in schizophrenia. Proc Natl Acad Sci U S A 107: 10584–10589. pmid:20489179
- 29. Mokhtar SS, Marshall CR, Phipps ME, Thiruvahindrapuram B, Lionel AC, Scherer SW, et al. (2014) Novel population specific autosomal copy number variation and its functional analysis amongst Negritos from Peninsular Malaysia. PLoS One 9: e100371. pmid:24956385
- 30. Hong MG, Pawitan Y, Magnusson PK, Prince JA (2009) Strategies and issues in the detection of pathway enrichment in genome-wide association studies. Hum Genet 126: 289–301. pmid:19408013
- 31. Galindo CL, Skinner MA, Errami M, Olson LD, Watson DA, Li J, et al. (2009) Transcriptional profile of isoproterenol-induced cardiomyopathy and comparison to exercise-induced cardiac hypertrophy and human cardiac failure. BMC Physiol 9: 23. pmid:20003209
- 32. Sack MN, Disch DL, Rockman HA, Kelly DP (1997) A role for Sp and nuclear receptor transcription factors in a cardiac hypertrophic growth program. Proc Natl Acad Sci U S A 94: 6438–6443. pmid:9177236
- 33. Kuwahara K, Nishikimi T, Nakao K (2012) Transcriptional regulation of the fetal cardiac gene program. J Pharmacol Sci 119: 198–203. pmid:22786561
- 34. Prakash SK, LeMaire SA, Guo DC, Russell L, Regalado ES, Golabbahsh H, et al. (2010) Rare copy number variants disrupt genes regulating vascular smooth muscle cell adhesion and contractility in sporadic thoracic aortic aneurysms and dissections. Am J Hum Genet 87: 743–756. pmid:21092924
- 35. Gallego-Delgado J, Connolly SB, Lazaro A, Sadlier D, Kieran NE, Sugrue DD, et al. (2009) Transcriptome of hypertension-induced left ventricular hypertrophy and its regression by antihypertensive therapies. Hypertens Res 32: 347–357. pmid:19325563
- 36. Gigante B, Bennet AM, Leander K, Vikstrom M, de Faire U (2010) The interaction between coagulation factor 2 receptor and interleukin 6 haplotypes increases the risk of myocardial infarction in men. PLoS One 5: e11300. pmid:20585578
- 37. Gigante B, Vikstrom M, Meuzelaar LS, Chernogubova E, Silveira A, Hooft FV, et al. (2009) Variants in the coagulation factor 2 receptor (F2R) gene influence the risk of myocardial infarction in men through an interaction with interleukin 6 serum levels. Thromb Haemost 101: 943–953. pmid:19404549
- 38. Gigante B, Bellis A, Visconti R, Marino M, Morisco C, Trimarco V, et al. (2007) Retrospective analysis of coagulation factor II receptor (F2R) sequence variation and coronary heart disease in hypertensive patients. Arterioscler Thromb Vasc Biol 27: 1213–1219. pmid:17347481
- 39. Brill S, Li S, Lyman CW, Church DM, Wasmuth JJ, Weissbach L, et al. (1996) The Ras GTPase-activating-protein-related human protein IQGAP2 harbors a potential actin binding domain and interacts with calmodulin and Rho family GTPases. Mol Cell Biol 16: 4869–4878. pmid:8756646
- 40. Matarin M, Brown WM, Scholz S, Simon-Sanchez J, Fung HC, Hermandez D, et al. (2007) A genome-wide genotyping study in patients with ischaemic stroke: initial analysis and data release. Lancet Neurol 6: 414–420. pmid:17434096
- 41. Parsa A, Chang YP, Kelly RJ, Corretti MC, Ryan KA, Robinson SW, et al. (2011) Hypertrophy-associated polymorphisms ascertained in a founder cohort applied to heart failure risk and mortality. Clin Transl Sci 4: 17–23. pmid:21348951
- 42. Holmqvist MH, Cao J, Hernandez-Pineda R, Jacobson MD, Carroll KI, Sung MA, et al. (2002) Elimination of fast inactivation in Kv4 A-type potassium channels by an auxiliary subunit domain. Proc Natl Acad Sci U S A 99: 1035–1040. pmid:11805342
- 43. Pinto D, Delaby E, Merico D, Barbosa M, Merikangas A, Klei L, et al. (2014) Convergence of genes and cellular pathways dysregulated in autism spectrum disorders. Am J Hum Genet 94: 677–694. pmid:24768552
- 44. Vaags AK, Lionel AC, Sato D, Goodenberger M, Stein QP, Curran S, et al. (2012) Rare deletions at the neurexin 3 locus in autism spectrum disorder. Am J Hum Genet 90: 133–141. pmid:22209245
- 45. Hitz MP, Lemieux-Perreault LP, Marshall C, Feroz-Zada Y, Davies R, Yang SW, et al. (2012) Rare copy number variants contribute to congenital left-sided heart disease. PLoS Genet 8: e1002903. pmid:22969434
- 46. Patnala R, Clements J, Batra J (2013) Candidate gene association studies: a comprehensive guide to useful in silico tools. BMC Genet 14: 39. pmid:23656885
- 47. Noh HJ, Ponting CP, Boulding HC, Meader S, Betancur C, Buxbaum JD, et al. (2013) Network topologies and convergent aetiologies arising from deletions and duplications observed in individuals with autism. PLoS Genet 9: e1003523. pmid:23754953
- 48. Lusis AJ, Weiss JN (2010) Cardiovascular networks: systems-based approaches to cardiovascular disease. Circulation 121: 157–170. pmid:20048233
- 49. Coggins M, Rosenzweig A (2012) The fire within: cardiac inflammatory signaling in health and disease. Circ Res 110: 116–125. pmid:22223209
- 50.
Müller F, Neumann J, Schmitz W (2000) Transcriptional regulation by cAMP in the heart. Control of Gene Expression by Catecholamines and the Renin-Angiotensin System: Springer. pp. 11–17.
- 51. Funakoshi Y, Ichiki T, Takeda K, Tokuno T, Iino N, Takeshita A (2002) Critical role of cAMP-response element-binding protein for angiotensin II-induced hypertrophy of vascular smooth muscle cells. J Biol Chem 277: 18710–18717. pmid:11907026
- 52. Vasan RS, Glazer NL, Felix JF, Lieb W, Wild PS, Felix SB, et al. (2009) Genetic variants associated with cardiac structure and function: a meta-analysis and replication of genome-wide association data. JAMA 302: 168–178. pmid:19584346
- 53. Fridman JS, Lowe SW (2003) Control of apoptosis by p53. Oncogene 22: 9030–9040. pmid:14663481
- 54. Sano M, Minamino T, Toko H, Miyauchi H, Orimo M, Qin Y, et al. (2007) p53-induced inhibition of Hif-1 causes cardiac dysfunction during pressure overload. Nature 446: 444–448. pmid:17334357
- 55. Marsh JD, Lehmann MH, Ritchie RH, Gwathmey JK, Green GE, Schiebinger RJ, et al. (1998) Androgen receptors mediate hypertrophy in cardiac myocytes. Circulation 98: 256–261. pmid:9697826
- 56. Lind JM, Chiu C, Ingles J, Yeates L, Humphries SE, Heather AK, et al. (2008) Sex hormone receptor gene variation associated with phenotype in male hypertrophic cardiomyopathy patients. J Mol Cell Cardiol 45: 217–222. pmid:18617186
- 57. Ikeda Y, Aihara K, Sato T, Akaike M, Yoshizumi M, Suzaki Y, et al. (2005) Androgen receptor gene knockout male mice exhibit impaired cardiac growth and exacerbation of angiotensin II-induced cardiac fibrosis. J Biol Chem 280: 29661–29666. pmid:15961403
- 58. English KM, Jones RD, Jones TH, Morice AH, Channer KS (2001) Gender differences in the vasomotor effects of different steroid hormones in rat pulmonary and coronary arteries. Horm Metab Res 33: 645–652. pmid:11733866
- 59. Hayward CS, Webb CM, Collins P (2001) Effect of sex hormones on cardiac mass. Lancet 357: 1354–1356. pmid:11343761
- 60. Jones RD, Hugh Jones T, Channer KS (2004) The influence of testosterone upon vascular reactivity. Eur J Endocrinol 151: 29–37. pmid:15248819
- 61. Kienitz T, Quinkler M (2008) Testosterone and blood pressure regulation. Kidney Blood Press Res 31: 71–79. pmid:18319594
- 62. Liu PY, Death AK, Handelsman DJ (2003) Androgens and cardiovascular disease. Endocr Rev 24: 313–340. pmid:12788802
- 63. van Empel VP, De Windt LJ (2004) Myocyte hypertrophy and apoptosis: a balancing act. Cardiovasc Res 63: 487–499. pmid:15276474
- 64. Zolk O, Schenke C, Sarikas A (2006) The ubiquitin-proteasome system: focus on the heart. Cardiovasc Res 70: 410–421. pmid:16497285
- 65. Usui S, Maejima Y, Pain J, Hong C, Cho J, Park JY, et al. (2011) Endogenous muscle atrophy F-box mediates pressure overload-induced cardiac hypertrophy through regulation of nuclear factor-kappaB. Circ Res 109: 161–171. pmid:21617130
- 66. Li HH, Kedar V, Zhang C, McDonough H, Arya R, Wang DZ, et al. (2004) Atrogin-1/muscle atrophy F-box inhibits calcineurin-dependent cardiac hypertrophy by participating in an SCF ubiquitin ligase complex. J Clin Invest 114: 1058–1071. pmid:15489953
- 67. Shah BH, Soh JW, Catt KJ (2003) Dependence of gonadotropin-releasing hormone-induced neuronal MAPK signaling on epidermal growth factor receptor transactivation. J Biol Chem 278: 2866–2875. pmid:12446705
- 68. Molnar P, Perrault R, Louis S, Zahradka P (2014) The cyclic AMP response element-binding protein (CREB) mediates smooth muscle cell proliferation in response to angiotensin II. J Cell Commun Signal 8: 29–37. pmid:24327051
- 69. Ichiki T (2006) Role of cAMP response element binding protein in cardiovascular remodeling: good, bad, or both? Arterioscler Thromb Vasc Biol 26: 449–455. pmid:16293792
- 70. McCarroll SA, Altshuler DM (2007) Copy-number variation and association studies of human disease. Nat Genet 39: S37–42. pmid:17597780
- 71. Williams NM, Zaharieva I, Martin A, Langley K, Mantripragada K, et al. (2010) Rare chromosomal deletions and duplications in attention-deficit hyperactivity disorder: a genome-wide analysis. Lancet 376: 1401–1408. pmid:20888040
- 72. Norton N, Li D, Rieder MJ, Siegfried JD, Rampersaud E, Zuchner S, et al. (2011) Genome-wide Studies of Copy Number Variation and Exome Sequencing Identify Rare Variants in 'BAG3' as a Cause of Dilated Cardiomyopathy. The American Journal of Human Genetics 88: 273–282. pmid:21353195