METHODS

Study population

All subjects were ethnic Han, who account for about 96% of the total population on the mainland of China. A total of 148 hypertensive families came from Beijing suburbs (Fangshan and Shijingshan districts), Jiangsu province (Changshu, Taixing, and Zhangjiagang districts), and Shanxi Province (Hanzhong city). Genotypes of pedigree members in 148 families were verified for Mendelian segregation. Table 1 lists the structure of sibs in families.

To be eligible for our study, one member of sib pairs in each nuclear hypertensive family had to meet the following criteria: (1) age older than 15 years; (2) either parent with hypertension; (3) two or more sibs with hypertension; (4) no clinical or biochemical indices of secondary hypertension: detailed clinical differentiation including clinical laboratory tests and ultrasound or computed tomography (CT) investigations on certain patients were implemented to exclude patients with other known diseases, such as acute or chronic glomerulonephritis and pyelonephritis, glucorticoid remediable aldosteronism (GRA), Cushing’s syndrome, or phaeochromocytoma; (5) not taking cortisone based or symphaticotonic drugs (oestroprogestogenic hormones by women) and absence of nephrolithiasis; and (6) resting-sitting SBP⩾140 mm Hg or DBP⩾90 mm Hg on three different occasions or use of antihypertensive drugs. Hypertensive families were initially screened by local doctors who had received rigorous training and passed the proficiency test. Once families were identified, epidemiologists and cardiologists from Fu Wai Hospital went to the local site to confirm the family eligibility. After positive confirmation, a formal study was conducted at an office in the local hospital. Trained interviewers using standard questionnaires collected information on demographic characteristics, history of cardiovascular disease, lifestyle factors, and family history of hypertension. Anthropometric measurements, including height, weight, and waist and hip circumferences, as well as an electrocardiogram were obtained by standard protocols. Measurements of BP were obtained by experienced doctors or nurses using a standard mercury sphygmomanometer with appropriately sized cuffs on the right arm after subjects had rested for 10 minutes. Three measurements were taken with an interval of at least 30 seconds between readings. The mean values of three consecutive measurements were used in these data. The study protocol was approved by the local research ethics committee of the Cardiovascular Institute and Fu Wai Hospital, Chinese Academy of Medical Sciences and Peking Union of Medical College. Informed consent was obtained from all study subjects.

Genotyping

Genomic DNA was obtained from peripheral blood leucocytes by phenol/chloroform extraction. Table 2 lists seven highly polymorphic microsatellite markers. The primer sequences were acquired from the genome database http://gdbwww.gdb.org/ and forward primer was labelled with fluorescence at the 5′ end.

Touch down polymerase chain reaction

Touch down polymerase chain reaction (PCR) was performed in thermocyclers of a Perkim Geneamp 9700 (Applied Biosystems, USA) in a 5 μl reaction volume. The amplifications involved initial denaturation for 10 minutes at 95°C, followed by 42 cycles each of denaturation for 30 seconds at 94°C, annealing for 40 seconds at 56°C, and extension for 60 seconds at 72°C, with the exception of the first 17 cycles, in which the annealing temperatures were decreased from 63°C to 56°C by 0.5°C per cycle. At the end of the amplifications, an extension reaction was performed at 72°C for 10 minutes.

The products from PCR were mixed with the DNA size standard (Genescan Rox-HD400, Applied Biosystems, USA), dextran blue dye (loading buffer), and formamide. The electrophoresis was carried out in 1 × TBE at 51°C, 3 KV for 2.5 hours using an ABI 377 sequencer (Applied Biosystems, USA). Genotyping was performed by Genescan 3.1 and Genotype 2.1 software (Applied Biosystems, USA). The entire allelotyping procedure was verified by genomic DNA from CEPH families 1347–02.

Statistical analysis

Values for SBP and DBP were standardised according to data from the third national BP sampling survey of China conducted in 1991, which involved about 940 000 subjects selected randomly and stratified according to the city and rural area from 30 provinces. Population mean SBP and DBP were 119.1 (SD 20.6) mm Hg and 73.0 (SD 11.6) mm Hg based on the data from that survey. Standardised BP values were obtained with the ratio of the difference between the subjects’ BP values and the population mean BP values of 1991 to standard deviation. A transformation to the square root for SBP was done. A normal distribution test was conducted for SBP and DBP by SPSS 10.0.

The SAGE/SIBPAL2 4.0 program was used to perform qualitative trait and quantitative trait linkage analyses of D8S1145, D8S261, D8S282, D4S2366, D4S43, D12S398, and the CA repeat polymorphism in the 3′ flanking region of the AT1 gene. This weakly parametric method, based on the Haseman-Elston algorithm, regressed the mean corrected cross product of sib pairs on the mean proportion of alleles’ shared identity by descent (IBD).¹²

In addition, for the quantitative trait of BP, two point and multipoint linkage analyses were conducted by the variance component method in SOLAR 1.7.4.¹³ The covariate effects evaluated were age, sex, and the product of age and sex.

Besides conventional TDT, an extended TDT (S-TDT and combined Z score) was also performed. Because EH is a disease with a late onset, sometimes the data from parents are difficult to obtain. In this case, a bias can arise in the TDT if the genotype for one parent is missing. Whereas S-TDT compares the marker genotypes in affected and unaffected offspring, there is no corresponding bias in the S-TDT. Therefore, a combined Z score was obtained by combining information from TDT and S-TDT and making Bonferroni’s correction.¹⁴

RESULTS

Study sample

Table 3 shows the clinical characteristics of all subjects by parents and offspring.

The DBP followed a normal distribution (p=0.229) as did the SBP after a transformation to the square root (p=0.115).

SAGE/SIBPAL2 analysis

Qualitative trait linkage analysis showed that there was linkage evidence between the marker D8S1145 and EH (p=0.029). Some positive signals were seen only in chromosome 8 by quantitative trait linkage analysis (table 4). We obtained significant evidence for linkage of the D8S261 locus to SBP and DBP (p=0.002 and p=0.04, respectively). Allele sharing IBD analysis was performed separately for sib pairs with no, one, or two affected members. We detected a significant excess of two alleles sharing IBD above an expected sharing of 0.25 at the marker D8S261 (p=0.029 for affected and p=0.039 for unaffected sib pairs), as well as at the marker D8S1145 (p=0.013 for affected and p=0.038 for discordant sib pairs).

Two point and multipoint linkage analysis in SOLAR

Table 5 lists two point linkage lod scores of all markers selected and BP phenotypes. Two point linkage analysis for the LPL gene showed that there was evidence for linkage of SBP to the marker D8S261 (lod=2.52). However, there was no linkage (lod<1.0) for DBP and EH with either of those markers and no evidence of two point linkage was found in or near another three candidate genes (ADD1, AT1, and VAPR1A). With multipoint linkage analysis for the LPL gene, the markers D8S1145, D8S261, and D8S282 with SBP showed a suggestive linkage finding, with a maximum lod score of 2.03 near D8S261, but no such evidence of linkage of those three markers with DBP on chromosome 8 was identified, with a maximum lod score of 0.40. By means of a permutation, we obtained a significant adjusted p value from the marker D8S261 (p=0.0001) for SBP. However, no significant p values (p=0.5088 for DBP and p=0.5000 for EH) were found for DBP or EH or (p=0.5000) in or near the other three candidate genes (ADD1, AT1, and VAPR1A).

TDT/S-TDT analysis

A TDT/S-TDT analysis was performed after we had detected some positive linkage signals at markers D8S1145 and D8S261. We did not detect significant excess transmission of any allele for D8S1145 with EH. A significant association between D8S261 in the region of 8p22 and hypertension was found based on TDT/S-TDT analysis (table 6). Allele 3 of D8S261 showed significant linkage disequilibrium with EH in traditional TDT, S-TDT tests (χ²=8.643, p<0.01; Z′=2.408, p<0.05) and performance after Bonferroni’s correction (Z=3.517, p<0.01).

DISCUSSION

We have used Haseman-Elston and variance component based linkage methods to estimate the contribution of variation in the regions of the four candidate genes to interindividual variation in BP. Our results support the hypothesis that markers near the LPL gene are genetically linked to hypertension or BP in Chinese hypertensive families. Large samples of subjects residing in rural areas in China were used, where drugs for treating hypertension are uncommon. These areas are appropriate for collecting hypertensive families for the study of quantitative traits in BP. This rural population is relatively homogeneous. The rate of awareness and treatment of hypertension was low. Most hypertensive patients did not take antihypertensive drugs, and some patients took them irregularly. The patients were excluded from this study if they had taken medication in the past two weeks and could not be re-examined without withdrawing treatment for two weeks. This enabled us to examine the complete distribution of alleles underlying interindividual variation in the quantitative BP phenotype. Additionally, we had sufficient power to detect the previously found positive linkage signals in the Taiwanese sample if the LPL gene or markers near the genomic regions were indeed linked to hypertension in China. One genetic linkage study of young onset hypertension in Taiwan suggested a positive linkage between D8S1145 near the LPL gene and BP.³ Another study of non-diabetic members of Taiwanese families also found significant linkage of an LPL gene marker (D8S261) with SBP.⁴

Our results are consistent with their finding. In our study, a linkage of the marker D8S261 with SBP was found, with a lod score of 2.52. Furthermore, multipoint linkage analysis disclosed a suggestive linkage with a maximum lod score of 2.03 with SBP near D8S261. Moreover, a linkage of the marker D8S261 with DBP was also found in our study. A possible difference between our study and the study by Wu et al⁴ could be the phenotype used for the linkage analysis. It is also possible that a gene other than LPL in this region is responsible for linkage with DBP. Also, environmental variables often significantly affect phenotype and obscure genetic effects.

Although the way in which LPL is involved in the development of hypertension is not apparent, hypertension has been found to occur more often than expected in families with familial combined hyperlipidaemia. Membrane ion transport, which has been related to hypertension, can be altered by lipid abnormalities and could have some involvement in the mechanisms that link high triglyceride concentrations and hypertension.³ A linkage analysis of EH with candidate genes involved in the regulation of glucose and lipid metabolism in 95 Chinese families by Chu et al¹⁵ also provided a finding that the LPL gene could be linked to EH. However, a study by Hunt et al¹⁶ involving a white population in the United States resulted in a negative finding.

The present study cannot clearly distinguish between contributions of the LPL gene itself and that of other genes near to D8S261, although the LPL gene is considered as a prime candidate gene responsible for BP variability in our study. For the LPL gene, the three markers selected span a region that contains this gene. In this region, defined by markers D8S1145, D8S261, and D8S282 on chromosome 8, other potential candidate genes are located, including the α-1C adrenergic receptor gene (ADRA1C) and a platelet derived growth factor receptor-like (PDGFRL) gene. The ADRA1C gene is a member of the adrenergic receptor superfamily, which influences heart rate, cardiac contraction, and vascular tone.¹⁷ The α-1B adrenergic receptor gene (ADRA1B) was reported to show linkage with SBP.¹⁸ Bray et al¹⁹ reported a significant association between the polymorphism Arg16Gly within the β₂-adrenergic receptor gene (ADRB2) and SBP among discordant sibs. Another study reported a significantly higher frequency of the Gly16 allele at this locus in African-Caribbean hypertensive patients compared with normotensive controls.²⁰ The results from these linkage and association studies support the need for functional testing of 1B, β₂, and perhaps other adrenergic receptor genes, such as ADRA1C. It was postulated that the ADRA1C gene might be an important candidate in the regulation of peripheral blood flow and arterial pressure.

PDGFRL is a cytokine related receptor gene with a significant sequence similarity to the ligand binding domain of the platelet derived growth factor receptor β (PDGFRβ) gene. A study showed that proliferation of mesangial cells requires a PDGFRβ mediated signal transduction.²¹ The impairment of kidney function relevant to proliferation of mesangial cells is involved in the generation of hypertension; therefore, PDGFRL may be a logical candidate gene contributing to the development of EH. Consequently, we cannot exclude the possibility that the marker D8S261 suggested linkage or association with other potential candidate genes in this region.

Recent studies have emphasised the importance of association studies as a means of localisation of genes for complex human disease, as the genetic distance over which useful linkage disequilibrium is thought to be present is much smaller than that over which linkage can be detected.¹⁴ A TDT test that relies on linkage disequilibrium (the simultaneous presence of linkage and association) is often more powerful than other tests that rely on linkage alone for identification of markers closely linked to genes that contribute to disease susceptibility. Allele 3 of the marker D8S261 had a significant linkage disequilibrium with EH; the association with EH remained significant even after Bonferroni’s correction. It was deduced that allele 3 carriers might have a higher incidence of risk of EH than allele 3 non-carriers. In addition, it was probable that the allele of D8S261 was in linkage disequilibrium with other closer genes than LPL in this region. We are looking forward to further studies to confirm this result.

The ADD1 gene is regarded as a candidate for a “salt sensitivity gene”. As envisioned by Cusi et al,⁷ its variant has been shown to enhance Na⁺−K⁺ pump activity and increase renal tubular sodium reabsorption. However, the results of linkage or association studies on the ADD1 gene are controversial among different populations.^6,22–24 Our results did not support linkage of the ADD1 gene with EH. The ADD1 polymorphism might account for only a portion of genetic variation of BP and be associated with a particular form of hypertension, characterised by alterations in sodium handling and response of plasma renin activity.

The renin-angiotensin system plays an important part in the function of the cardiovascular system and the regulation of BP. No linkage was shown between the CA repeat polymorphism in the 3′ flanking region of the AT1 gene and interperson BP variation in our study. However, linkage of the CA repeat polymorphism with EH has been found in the Chinese Han, Tibetan, and Finnish populations.^8,10,25 In the Chinese population, subjects had moderate or severe hypertension (SBP⩾160 mm Hg or DBP⩾95 mm Hg); therefore, we cannot exclude the possibility that the AT1 gene may have an impact on interperson variation in BP, especially among those with clinically manifest hypertension.

Essential hypertension is considered to be a multifactorial disease resulting from a combination of several predisposing genes interacting with each other and environmental factors. Linkage studies of EH pose several problems, including delayed onset of phenotypic expression, varying penetrance, and lack of unequivocal diagnostic criteria.²⁶ When multiple genetic alleles contribute to a complex trait, linkage analysis has somewhat limited power for finding genes of modest effect. Also, environmental variables often significantly affect phenotype and obscure genetic effects. These factors may explain inconsistent linkage results.

In conclusion, we used two different weakly parametric linkage tests, sib pair linkage (SAGE/SIBPAL2) and pedigree based (SOLAR) approaches, to estimate the quantitative or qualitative contribution of variation at the four candidate genes or their near genomic regions to BP and EH. Positive linkage evidence of the LPL gene to hypertension and BP was found and TDT further supported this result. The LPL gene or other genes near to the D8S261 genomic region may influence individual BP variation in the Chinese population.