Whole Genome Survey of Copy Number Variation in the Spontaneously Hypertensive Rat
Relationship to Quantitative Trait Loci, Gene Expression, and Blood Pressure
Copy number variation has emerged recently as an important genetic mechanism leading to phenotypic heterogeneity. The aim of our study was to determine whether copy number variants (CNVs) exist between the spontaneously hypertensive rat (SHR) and its control strain, the Wistar-Kyoto rat, whether these map to quantitative trait loci in the rat and whether CNVs associate with gene expression or blood pressure differences between the 2 strains. We performed a comparative genomic hybridization assay between SHR and Wistar-Kyoto strains using a whole-genome array. In total, 16 CNVs were identified and validated (6 because of a relative loss of copy number in the SHR and 10 because of a relative gain). CNVs were present on rat autosomes 1, 3, 4, 6, 7, 10, 14, and 17 and varied in size from 10.0 kb to 1.6 Mb. Most of these CNVs mapped to chromosomal regions within previously identified quantitative trait loci, including those for blood pressure in the SHR. Transcriptomic experiments confirmed differences in the renal expression of several genes (including Ms4a6a, Ndrg3, Egln1, Cd36, Sema3a, Ugt2b, and Idi21) located in some of the CNVs between SHR and Wistar-Kyoto rats. In F2 animals derived from an SHR×Wistar-Kyoto cross, we also found a significant increase in blood pressure associated with an increase in copy number in the Egln1 gene. Our findings suggest that CNVs may play a role in the susceptibility to hypertension and related traits in the SHR.
- DNA copy number variations
- inbred SHR
- gene expression
- microarray analysis
- blood pressure
High blood pressure (BP; hypertension) is a major risk factor for coronary, cerebrovascular, and renal disease. Most cases of hypertension have unknown etiology and are, thus, classified as essential hypertension. Hypertension has a significant genetic contribution. Despite the progress made toward the understanding of rare monogenic forms of hypertension in humans, the genetic background of essential hypertension remains poorly understood.1 The spontaneously hypertensive rat (SHR) is one of the most widely used genetic models for hypertension. The SHR model is characterized by hypertension, insulin resistance, hypertriglyceridemia, and hypercholesterolemia. Genetically, the SHR was derived in 1963 from inbreeding Wistar rats with the highest BPs.2 Using linkage analysis, there have been multiple efforts to map genes influencing BP and related phenotypes in the SHR. These efforts have resulted in the successful identification of several chromosome regions containing quantitative trait loci (QTLs) regulating BP or related cardiovascular and metabolic phenotypes in the SHR.3 Yet, despite many experiments, very few genes that underlie these QTLs have been unambiguously identified.4 Copy number variations (CNVs), defined as gains and losses of DNA typically >1 kb and up to several megabases, are being increasingly recognized as a source of differences in genomic sequence5–8 and have been proposed as possible mechanisms for phenotypic variation in humans,8,9 the mouse, and, recently, the rat.10–12
The aim of this study was to first identify differences in copy numbers between the genome of the SHR and its most commonly used control strain, the Wistar-Kyoto rat (WKY). We then determined whether CNVs map to known QTLs and whether genes located within these CNV regions are differentially expressed between the SHR and the WKY. Finally, we investigated the relationship of one of the identified CNVs to BP in a genetic cross between the SHR and the WKY.
Genomic DNA was prepared using the DNeasy kit (Qiagen) from livers and kidneys of 3 male SHRs and 3 WKYs aged 8 to 10 weeks obtained from the research colony maintained at the University of Leicester (all of the animals were initially obtained from the breeding stock of Charles River Laboratories, Margate, United Kingdom). Total RNA was extracted from rat kidneys, spleens, and livers using Qiagen RNeasy kits (Qiagen) according to the manufacturer’s instructions. Total RNA yield was quantified by UV spectrophotometry, and RNA integrity was verified using an Agilent 2100 Bioanalyzer (Agilent Technologies).
Oligonucleotide Array Comparative Genomic Hybridization
Unamplified genomic DNA (1 μg) was labeled with Cy3 (SHR) or Cy5 (WKY). We performed comparative genomic hybridization using long oligonucleotide arrays containing 385 000 isothermal probes, 45-mer to 75-mer, spanning the rat reference genome with a median spacing of ≈5 kb (NimbleGen Systems). The oligonucleotide design, array fabrication, DNA labeling, array comparative genomic hybridization (aCGH) experiments, data normalization, and calculations of copy number ratio (using log2 of the ratio of signal from the fluorophores Cy3 and Cy5) were performed at NimbleGen according to recommended and published procedures. CNVs were identified on the basis of the relative intensity of log2 ratio profiles of each experiment, using the circular binary segmentation algorithm from Olshen et al.13 Our criteria of ≥3 probes in a segment and mean amplitude of a log2 shift across segment ±0.5 were used to define the final set of high-confidence CNV calls. The BLAT (BLAST-like alignment tool) hit count was defined as the number of matches in which the probe sequence identity×length of matching sequence/length of the probe was ≥0.9. Gene annotation and overlap were determined using Ensembl GeneBuild 3.4.
Validation of Array Data Using Quantitative Real-Time PCR
To validate CNVs detected by oligo-aCGH, quantitative real-time PCR (qPCR) assays were developed to measure copy number in implicated chromosomal regions relative to a control region of invariant copy number in both strains (selected based on aCGH profile) in DNA extracted from the kidney.14 Relative copy numbers were determined by qPCR using Power SYBR Green chemistry and the ABI Prism 7900HT Sequence Detection System (Applied Biosystems). Predesigned QuantiTect primers (Qiagen) were used when available or primers were designed using ABI Primer Express Software (version 2.0) and the Ensembl Rat GenomeBuild 3.4 (Table S1, available in the online Supplementary Data at http://hyper.ahajournals.org). Each assay was performed in triplicate using 20-μL reactions containing 10 μL of Power SYBR Green PCR Master Mix (2×; Applied Biosystems), 200 nM concentration of forward and reverse primer, and 20 ng of genomic DNA extracted from kidneys to confirm the presence of the same CNVs in DNA from different tissues. Melting curve analysis and sequence analysis were performed to check PCR product specificity. Amplification was performed according to the following conditions: 1 cycle at 95°C for 10 minutes, 50 cycles at 95°C for 15 seconds, and 60°C for 1 minute. Experiments were performed on the test and control primers to verify comparable efficiency in amplification before analysis of copy number in the strains. Mean DNA starting quantities and SDs were estimated on the basis of threshold cycle differences between the control and test loci.14 Primer sequences are available in Table S1.
Gene Expression Analysis
To evaluate whether mRNA levels of genes that lie within CNV regions are differentially expressed between SHRs and WKYs, we used qPCR as described previously.15 To avoid interstrain variation of primer binding efficiency, we performed melting curve analysis and only used pairs of primers with similar binding efficiencies determined from a dilution curve. mRNA levels of test genes were compared with those for β-actin of the same sample, and the results displayed are the relative level of renal expression in SHRs compared with the WKY control strain. We initially measured mRNA levels in the kidney because of the known role of this organ in the regulation of BP.16 We also measured mRNA levels in the heart and spleen for a broader assessment of CNV-related changes in mRNA expression. We used QuantiTect predesigned and validated primer sets for each gene (Qiagen).
Association Analysis of Egln1 Gene Copy Number and BP in an SHR×WKY F2 Population
We measured the copy number of Egln1 (CNV1b) in 229 F2 rats derived from a cross of SHRs and WKYs. The generation, characterization, and BP measurements of this F2 cross, which were derived from the same SHR and WKY colonies as animals analyzed for CNVs in this study, have been described in detail previously.17 Briefly, indirect systolic BP in the tail artery was measured by tail plethysmography at 20 weeks of age in rats on a normal diet. To determine relative copy number in F2 rats, we performed qPCR in triplicate and determined the normalized relative copy number as stated above.
Statistical analysis of comparisons between SHR and WKY genes and copy numbers was undertaken using paired and unpaired t tests where appropriate; otherwise, the Mann-Whitney test was used. To determine whether there is any difference in systolic BPs between F2 rats that carry different Egln1 or Ugt2b copy numbers, we used 1-way ANOVA and Kruskal-Wallis test with the Dunn multiple comparison test.
High-Resolution Comparative Genomic Hybridization Analysis
The 3 comparative genomic hybridizations consistently revealed 16 CNVs on 8 SHR chromosomes (1, 3, 4, 6, 7, 10, 14, and 17; Figure 1 and Table 1). Six of these regions were attributed to a relative loss of DNA event in the SHR, and 10 regions exhibited an increase in copy number in the SHR. The segments varied in size (range: 10 kb to 1.6 Mb; mean: 250 kb). Changes on the Y chromosomes were not analyzed because of lower probe density and greater mapping uncertainty for these regions in the current assembly.
Validation of CNVs
To validate CNVs at each locus, we developed a real-time PCR assay to quantitatively determine whether these regions show CNV. Figure 2 shows the real-time PCR results, which confirmed the CNVs in all of the regions. In those regions where there was a gain of DNA event in the SHR compared with the WKY, the increase in copy number was between 2 and 4 copies. We also determined which of our CNVs overlapped with CNVs detected by Guryev et al.12 and found that 8 of our CNVs (1b, 3b, 4a, 6a, 7b, 7c, 10a, and 17a) were detected in both studies.
Genes Located Within CNV Regions and Their Expression
Eleven of the 16 identified CNVs contain annotated or putative genes (Table 2). We compared expression of these genes in the kidney, heart, and spleen between SHRs and WKYs. Some of the genes did not have detectable expression in these tissues. However, others showed significant differential expression between the 2 strains (Table 2). Notably, Ms4a6a located in CNV1a (first CNV on chromosome 1) had markedly decreased expression in all 3 of the tissues in the SHR compared with the WKY, consistent with the loss of DNA in the SHR (Figure 3). CNV1b, which causes a gain of the Egln1 gene in the SHR, was associated with significantly increased mRNA levels of this gene in the kidney and heart of the SHR compared with the WKY (Figure 3). We did not detect any expression of Egln1 in the spleen. Within CNV3a (chromosome 3) lies the Ndrg3 (N-myc downstream regulated gene 3) involved in cell differentiation. The expression of the Ndrg3 was significantly decreased in the SHR kidney, heart, and spleen. CNV4a harbors the Cd36 predicted gene. We observed decreased expression of this gene in the SHR kidney and spleen, as described previously by other groups (data not shown).18 CNV4b contains a putative gene similar to NEDD4-binding protein 1, which was expressed at very low levels in both the SHR and WKY but showed a trend toward increased expression in the SHR (data not shown). Also in the same region is the Sema3a gene, which shows increased expression in the SHR kidney and heart. CNV14a contains the Ugt2b gene, thought to be a glucosyl transferase. mRNA levels of this gene were significantly higher in the kidney of the SHR compared with the WKY. CNV17a contains 2 genes, Idi21 (isopentenyl diphosphate δ-isomerase) involved in cholesterol synthesis and similar to nucleolar GTP-binding protein 1 involved in chronic renal failure. Expression of Idi21 was significantly upregulated in the SHR kidney (Figure 3).
CNV Location Within Rat QTLs and Orthologous Regions in Humans
Of the 16 confirmed CNVs, several (1a, 1b, 3b, 4a, 4b, 7b, 7c, and 10a) are located within BP QTLs mapped in the SHR (Table 3). Furthermore, 1b, 3b, and 4b CNVs map to QTLs in crosses derived from SHR and WKY strains.19–22 Both 1a and 1b CNVs on chromosome 1 map to the BP QTL in our previously constructed WKY.SHR-Sa (D1Wox19-D1Mit2) strain (Figure 4). CNV1b is located 4,707,195 bp from the boundary of our minimal SISA congenic strain (Figure 4).19,21 Other CNVs (3a, 7a, 10a, and 14a) map to QTLs for related traits, such as vascular growth, urinary albumin excretion, and cholesterol level in the SHR. CNVs 3a, 10b, and 17a also map to known BP QTLs in strains other than the SHR.
When we searched the Copy Number Variation Discovery Project for orthologous regions of these stretches of DNA in humans, we found that CNVs on chromosome 1, 6, and 14 have orthologous regions in humans with putative structural variations, as described in the database of genomic variants (http://www.sanger.ac.uk/humgen/cnv/42mio/ and http://projects.tcag.ca/variation/; Table 3).
Egln1 CNV and BP
We studied the association between copy number in CNV1b (Egln1 gene) and BP, in a large F2 cross from SHRs and WKYs (Figure 5). There was a significant difference in systolic BP in the F2 animals that carry different copy numbers of the Egln1 gene (ANOVA P=0.0083). The mean systolic BP of animals with 1 copy of the Egln1 gene was 155.5±17.2 mm Hg, whereas animals that have >1 copy on only 1 chromosome had a mean systolic BP of 157.0±13.7 mm Hg and animals with 2 duplications had a mean systolic BP of 163.5±13.9 mm Hg.
In this comparative genome hybridization analysis of the SHR and the WKY strains, we have identified new and validated existing CNVs in the rat. Of the 16 CNVs detected, 8 were revealed previously in the study by Guryev et al12 that compared the SHR and the Brown Norway strain. The 7 additional CNVs detected by our experiments may have arisen in the SHR since the divergence from the WKY strain.
We also demonstrate that, in many instances, CNV is associated with changes in expression of the genes located in the CNV regions. Our results highlight protein-coding genes within the CNVs that may contribute to the pathogenesis of disease in the SHR. Kidney, liver, and spleen mRNA analysis of genes revealed changes in 8 genes that overlap with the detected CNVs. However, some of the genes did not show expression change, which paralleled the copy number status of the genomic segment. Such observations are in line with findings by Guryev et al,12 and others that show that increased copy number can be positively,23,24 negatively,25 or neutrally correlated with gene expression. Changes in expression may be caused by a loss of DNA that can delete a transcriptional repressor or duplication in DNA in a gene enhancer. Other possible regulatory elements that may be lost or duplicated are microRNAs or microRNA targets, which regulate gene expression. Lack of SHR-WKY differences in tissue expression of genes showing a distinct pattern of CNV may be explained by one of the following: (1) regulation of these genes is not affected by the rearrangement; (2) expression was measured in an inappropriate tissue; or (3) feedback mechanisms exist to correct for any changes in copy number.25 Our data, therefore, support previous findings that some but not all CNVs influence the expression profile of genes in a way that may affect pathogenesis of disease.
Potential Relevance to BP QTLs in SHRs
Several CNVs (1a, 1b, 3b, 4a, 4b, 7b, 7c, and 10a) that we identified are located within BP QTL regions identified in crosses of the SHR with other rat strains. For example, the CNV on chromosome 4 includes a predicted Cd36 gene; Cd36 is a duplicated gene that has been identified previously as a cause of defect in insulin action, fatty acid metabolism, and hypertension in the SHR.18,26 The QTLs containing CNVs on chromosomes 1b, 3b, and 4b have been directly detected in crosses between the SHR and WKY. To examine the involvement of one of these CNVs in BP in the SHR, we took forward CNV1b, which contains the Egln1 gene. Egln1, also known as the hypoxia-induced factor (HIF)-prolyl hydroxylase gene, showed increased expression in the kidney of the SHR (Table 2), and in previous studies, using renal transplantation, we have shown that the BP QTL located in this region on chromosome 1b at least partly mediates its effect via the kidney.16 The protein product of Egln1 is a hydroxylase enzyme, which promotes degradation of the HIF subunits,27 and a study by Li et al28 demonstrated that blunting this protein was associated with an increase of BP. The observed increase in HIF-prolyl hydroxylase gene expression in the SHR could, therefore, lead to an augmented degradation of HIF 1α, the key molecule that regulates adaptation to hypoxia, protects renal medulla from ischemia, and, thus, controls renal sodium excretion. Consistent with this, we observed a significant association of Egln1 copy number with BP in the expected direction in an F2 cross derived from a cross of SHRs and WKYs, providing strong support for a possible causal involvement of this CNV with BP in the SHR.
CNVs in the rat can be successfully exploited to model CNVs apparent in the human genome, as shown for the CNV in the Fcgr3 gene.29 The human orthologs of the Ms4a6a and HIF-prolyl hydroxylase 2 gene contained within the rat CNVs identified in this study lie within regions with reported germ-line copy number polymorphisms in humans. Human orthologs of the olfactory receptor genes are also known to be duplicated in humans and other mammalian species.30 Conservation of CNVs across species suggests that selective pressure may drive acquisition or retention of specific gene dosage alterations. In fact, the study by Guryev et al,12 which surveyed >10 rat models, showed that human and rat CNV regions share more in common than the mouse, reinforcing the importance of the rat as a model organism for studying phenotypic effects of structural variations relevant to complex disease in humans.
A limitation of our study is that the number of CNVs detected is an underestimate of the real number of structural variants in SHRs, because many genomic regions were not effectively surveyed because of technical limitations (eg, probe density or nonrandom distribution) of the technology platform that was used. Newer arrays with higher densities and next-generation sequencing platforms using paired end-tag sequencing may resolve this limitation.31 Another limitation of the study is that we could not accurately size the CNV regions detected, because probes are not equally distributed along the rat genome because of various factors, such as repeat elements. Probes may, therefore, fall far from the border of each CNV region. The borders may be determined by newer arrays with higher probe density or custom arrays. It is also possible that somatic mutations that arise in specific tissues may contribute to cardiovascular disease; however, this was not a focus for our study, which was designed to investigate germ-line transmitted CNVs. Finally, although we show that several CNVs map to QTL regions in the SHR and some have a significant effect on expression of coincident genes, additional studies are required to support their phenotypic relevance. Even where a direct association with BP (or another trait) is shown, as we have for the CNV containing Egln1, further investigation, including, ultimately, genetic manipulation of the copy number, is required to provide definitive proof of its causal involvement.
Human genomic variation exists in many forms; CNVs are a form of variation in which large amounts of DNA are either deleted or duplicated. Such CNVs have not yet been systematically studied with regard to their relevance to hypertension. In the present study we examined the DNA of the SHR model for CNVs. We demonstrated that there are CNVs that exist in this model compared with its control strain, the WKY, which coincide with BP QTLs and may contribute to the pathogenesis of its hypertension. This provides a hypothesis for human genetic research on the involvement of CNVs in hypertension, which will require large-scale studies that are composed of well-phenotyped cohorts.
This study makes use of data generated by the Genome Structural Variation Consortium, whom we thank for prepublication access to their CNV discovery (and/or) genotyping data, made available through the Web sites http://www.sanger.ac.uk/humgen/cnv/42mio/ and http://projects.tcag.ca/variation/ as a resource to the community.
Sources of Funding
This work was supported by a Wellcome Trust Functional Genomics Programme Grant in Cardiovascular Disease and the University of Ballarat Research Grant. F.J.C. was supported by the British Heart Foundation. N.J.S. holds a British Heart Foundation chair.
- Received September 1, 2009.
- Revision received September 20, 2009.
- Accepted February 15, 2010.
Sebat J, Lakshmi B, Troge J, Alexander J, Young J, Lundin P, Maner S, Massa H, Walker M, Chi M, Navin N, Lucito N, Healy J, Hicks J, Ye K, Reiner A, Gilliam CC, Trask B, Patterson N, Zetterberg A, Wigler M. Large-scale copy number polymorphism in the human genome. Sci. 2004; 305: 525–528.
Redon R, Ishikawa S, Fitch KR, Feuk L, Perry GH, Andrews TD, Fiegler H, Shapero MH, Carson AR, Chen W, Cho EK, Dallaire S, Freeman JL, GonzÃ¡lez JR, GratacÃ2s M, Huang J, Kalaitzopoulos D, Komura D, MacDonald JR, Marshall CR, Mei R, Montgomery L, Nishimura K, Okamura K, Shen F, Somerville MJ, Tchinda J, Valsesia A, Woodwark C, Yang F, Zhang J, Zerjal T, Zhang J, Armengol L, Conrad DF, Estivill X, Tyler-Smith C, Carter NP, Aburatani H, Lee C, Jones KW, Scherer SW, Hurles ME. Global variation in copy number in the human genome. Nature. 2006; 444: 444–454.
Bailey JA, Gu Z, Clark RA, Reinert K, Samonte RV, Schwartz S, Adams MD, Myers EW, Li PW, Eichler EE. Recent segmental duplications in the human genome. Science. 2002; 297: 1003–1007.
Pollex RL, Hegele RA. Copy number variation in the human genome and its implications for cardiovascular disease. Circulation. 2007; 115: 3130–3138.
Cutler G, Marshall LA, Chin N, Baribault H, Kassner PD. Significant gene content variation characterizes the genomes of inbred mouse strains. Genome Res. 2007; 17: 1743–1754.
Olshen AB, Venkatraman ES, Lucito R, Wigler M. Circular binary segmentation for the analysis of array-based DNA copy number data. Biostatistics. 2004; 5: 557–572.
Lee AS, Gutierrez-Arcelus M, Perry GH, Vallender EJ, Johnson WE, Miller GM, Korbel JO, Lee C. Analysis of copy number variation in the rhesus macaque genome identifies candidate loci for evolutionary and human disease studies. Hum Mol Genet. 2008; 17: 1127–1136.
Tomaszewski M, Charchar FJ, Lynch MD, Padmanabhan S, Wang WY, Miller WH, Grzeszczak W, Maric C, Zukowska-Szczechowska E, Dominiczak AF. Fibroblast growth factor 1 gene and hypertension: from the quantitative trait locus to positional analysis. Circulation. 2007; 116: 1915–1924.
Clemitson JR, Pratt JR, Frantz S, Sacks S, Samani NJ. Kidney specificity of rat chromosome 1 blood pressure quantitative trait locus region. Hypertension. 2002; 40: 292–297.
Clemitson J, Dixon RJ, Haines S, Bingham AJ, Patel BR, Hall L, Lo M, Sassard J, Charchar FJ, Samani NJ. Genetic dissection of a blood pressure quantitative trait locus on rat chromosome 1 and gene expression analysis identifies SPON1 as a novel candidate hypertension gene. Circ Res. 2007; 100: 992–999.
Frantz SA, Kaiser M, Gardiner SM, Gauguier D, Vincent M, Thompson JR, Bennett T, Samani NJ. Successful isolation of a rat chromosome 1 blood pressure quantitative trait locus in reciprocal congenic strains. Hypertension. 1998; 32: 639–646.
Stoll M, Kwitek-Black AE, Cowley AW Jr, Harris EL, Harrap SB, Krieger JE, Printz MP, Provoost AP, Sassard J, Jacob HJ. New target regions for human hypertension via comparative genomics. Genome Res. 2000; 10: 473–482.
Pravenec M, Churchill PC, Churchill MC, Viklicky O, Kazdova L, Aitman TJ, Petretto E, Hubner N, Wallace CA, Zimdahl H, Zidek V, Landa V, Dunbar J, Bidani A, Griffin K, Qi N, Maxova M, Kren V, Mlejnek P, Wang J, Kurtz TW. Identification of renal Cd36 as a determinant of blood pressure and risk for hypertension. Nat Genet. 2008; 40: 952–954.
Marx J. Cell biology: how cells endure low oxygen. Science. 2004; 303: 1454–1456.
Li N, Chen L, Yi F, Xia M, Li PL. Salt-sensitive hypertension induced by decoy of transcription factor hypoxia-inducible factor-1α in the renal medulla. Circ Res. 2008; 102: 1101–1108.
Aitman TJ, Dong R, Vyse TJ, Norsworthy PJ, Johnson MD, Smith J, Mangion J, Roberton-Lowe C, Marshall AJ, Petretto E, Hodges MD, Bhangal G, Patel SG, Sheehan-Rooney K, Duda M, Cook PR, Evans DJ, Domin J, Flint J, Boyle JJ, Pusey CD, Cook HT. Copy number polymorphism in Fcgr3 predisposes to glomerulonephritis in rats and humans. Nature. 2006; 439: 851–855.
Fullwood MJ, Wei CL, Liu ET, Ruan Y. Next-generation DNA sequencing of paired-end tags (PET) for transcriptome and genome analyses. Genome Res. 2009; 19: 521–532.