Association of a Functional Cytochrome P450 4F2 Haplotype with Urinary 20-HETE and Hypertension

Abstract

Cytochrome P450 4F2 (CYP4F2) catalyzes the ω-hydroxylation of arachidonic acid to 20-hydroxyeicosatetraenoic acid (20-HETE), a natriuretic and vasoactive eicosanoid that participates in the development of hypertension. The relationship among CYP4F2 genetic variants in the regulatory region, formation of renal 20-HETE, and hypertension is unknown. Here are reported seven genetic variants around the CYP4F2 intronic regulatory region. Four of these variants made up two common haplotypes, Hap I (c.−91T/c.−48G/c.−13T/c.+34T) and Hap II (c.−91C/c.−48C/c.−13C/c.+34G). Hap I included a major functional variant, c.−91T→C, which was identified by reporter assay and electrophoretic mobility shift assay. Transfected into HEK293 cells, the Hap I construct showed a trend toward higher basal transcriptional activity and exhibited significantly greater LPS-stimulated activity than Hap II; these findings were the result of different NF-κB binding affinity between the two constructs. In vivo, a case-control study demonstrated that homozygosity for Hap I doubled the risk for hypertension in a Chinese population, even after adjustment for risk factors including age, gender, and body mass index. This association was confirmed in a family-based association study. In addition, Hap I was associated with elevated urinary 20-HETE. These results indicate that a functional variant of the CYP4F2 regulatory region, which increases the binding affinity of NF-κB, increases the risk for hypertension, likely by modulating the production of 20-HETE.

The cytochrome P450 4F2 (CYP4F2) gene, prominently expressed in human kidney and liver, encodes an ω-hydroxylase that catalyzes the metabolism of arachidonic acid,¹ leukotriene B4,² and tocopherol.³ The 20-hydroxyeicosatetraenoic acid (20-HETE), derived from arachidonic acid by CYP4F2 in the kidney, acts as a natriuretic and vasoactive eicosanoid and plays an important role in the control of renal function and systemic BP.^4,5 Considerable evidence showed that altered renal 20-HETE content and CYP genes were related to hypertension in animal models and in humans^6–10; however, the contribution of haplotypes in the CYP4F2 regulatory region to renal 20-HETE excretion as well as human hypertension remains unclear, because recent investigations indicated that the primary target variants for disease–gene association studies would be located in regulatory regions.¹¹

Hypertension is a common disease and an independent risk factor for stroke, heart failure, and ESRD. It is widely known that hypertension is a multifactorial disease whereby genetic determinants interact with environmental factors¹²; however, the molecular mechanisms are still not well understood. Previously, we focused on a single-nucleotide polymorphism, 421G→C (c.−48G→C, rs3093100), in the regulatory region of the CYP4F2 gene and found that the 421G allele was associated with increased susceptibility to hypertension via a negative Myb responsive element¹⁰; however, we could not conclude from these data that high risk for hypertension is attributed to downregulated CYP4F2 expression and subsequently decreased renal 20-HETE formation, because we did not measure urinary 20-HETE, which has both the pro- and antihypertensive action. Other variants or haplotypes in the CYP4F2 regulatory region might contribute to an altered gene transcription and renal 20-HETE formation and exert a major impact on the development of hypertension; therefore, in this study, we aimed to identify risk haplotype of the CYP4F2 regulatory region for hypertension and estimate its impact on gene transcription as well as the formation of renal 20-HETE.

Results

Regulatory Haplotypes of the CYP4F2 Gene

We showed previously that a single variant 421G→C (c.−48G→C) with altered Myb responsive element in the CYP4F2 intron 1 was associated with hypertension.¹⁰ To discover other variants in the regulatory region of CYP4F2, we amplified and sequenced 130 genomic DNA in this study. Six new genetic variants, c.−91T→C (rs3093098), c.−77T→C (rs3093099), c.−43C→T (rs3093101), c.−23G→A (rs3093102), c.−13T→C (rs3093103), and c.+34T→G (Trp12Gly, rs3093105), were found in intron 1 and exon 2 (Figure 1). Four of these variants, c.−91T→C, c.−48G→C, c.−13T→C, and c.+34T→G, had minor allelic frequencies of approximately 14%, and the remaining three variants were rare (minor allelic frequency 0.38%). Haplotype analysis showed three haplotypes: Hap I, Hap II, and Hap III (Table 1). Hap III was rare and therefore excluded from further analysis. Linkage disequilibrium (LD) analysis demonstrated that four common variants were in perfect LD (all of the D′ and r² were 1.00); therefore, Hap I could be represented by c.−13T, and Hap II could be represented by c.−13C.

Figure 1.

Map of the CYP4F2 gene around the regulatory region. Exons are indicated by bold letters, and the coding amino acid sequence is shown below the DNA sequence. The DNA sequencing region was from c.−209 to c.+419. Nucleotides were numbered according to the Genebank reference sequence AF467894, where the A of the ATG translation initiation codon is +1. The discovered genetic variants are denoted by arrows. Boxed sequences represent the basal cis-acting elements and putative binding sites of NF-κB and Myb.

Table 1.

Regulatory haplotypes of the CYP4F2 gene^a

Variants	Ref SNP_ID	Location	Hap I	Hap II	Hap III	MAF (%)
IVS1–91T→C	rs3093098	Intron 1	T	C	C	14.23
IVS1–77T→C	rs3093099	Intron 1	T	T	C	0.38
IVS1–48G→C	rs3093100	Intron 1	G	C	C	14.23
IVS1–43C→T	rs3093101	Intron 1	C	C	T	0.38
IVS1–23G→A	rs3093102	Intron 1	G	G	A	0.38
IVS1–13T→C	rs3093103	Intron 1	T	C	T	13.85
34T→G	rs3093105	Exon 2	T	G	G	14.23
n (%)			223 (85.77)	36 (13.85)	1 (0.38)

^aMAF, minor allele frequency; SNP, single nucleotide polymorphism.

Transcriptional Activities of Hap I and Hap II

To estimate the effect of different regulatory haplotypes on gene regulation, we generated reporter constructs of Hap I and Hap II (Figure 2A) and compared their transcriptional activities after transient transfection into HEK293. All constructs had basal transcriptional activities, and the series of deletion constructs resulted in the decrease of transcriptional activity (Figure 2B). The transcriptional activity of Hap I was higher than that of Hap II in pCAT219, with a trend in the same direction for pCAT564, whereas the Hap I in pCAT67 with only a Myb binding site showed decreased activity compared with Hap II. The most significantly increased activity of Hap I in the pCAT219 (27.23%; P < 0.05) was in accordance with a putative binding site of NF-κB at the c.−91 predicted by MATCH program¹³ and indicated that the putative NF-κB responsive element at c.−91T might play a more vital role than the Myb element at c.−48G. To evaluate the effect of c.−91T→C and c.−48G→C, we compared transcriptional activities of the pCAT219 constructs in different combinations (Figure 2C). The pCAT219 constructs with c.−91T (−91T/−48G and −91T/−48C) showed significantly higher activity than the corresponding constructs with c.−91C (−91C/−48G and −91C/−48C), whereas the differences between c.−48G constructs and c.−48C constructs were NS.

Figure 2.

Transcriptional activities of reporter constructs in the transfected HEK293 cells. (A) The series of deletion constructs with the c.−91T→C, c.−48G→C, and c.−13T→C variants. (B) Transcriptional activities of Hap I and Hap II in pCAT564, pCAT219, and pCAT67. (C) Transcriptional activities of pCAT219 constructs with different combination of the c.−91T→C and c.−48G→C variants. All experiments were performed three times independently. *P < 0.05.

NF-κB Binding Pattern of the c.−91T→C Variant

MATCH analysis revealed that the c.−91C allele disrupted the binding site of NF-κB (Figure 3A). Electrophoretic mobility shift assays (EMSA) were subsequently performed using radiolabeled oligonucleotides with the c.−91T allele and the c.−91C allele, as well as radiolabeled NF-κB consensus as a positive control, in the presence of HEK293 nuclear extract. As seen in Figure 3B, NF-κB consensus formed two complexes (lane 1) that were removed by self-competition (lanes 2 to 3) and unlabeled c.−91T competition (lanes 4 to 5). The radiolabeled c.−91T formed two similar complexes (lane 6) that were competed by 100-fold and 500-fold excess of unlabeled c.−91T (lanes 7 to 8), and the upper complex was abolished by unlabeled c.−91C (lane 9). The radiolabeled c.−91C formed only an upper complex that was removed by self-competition (lanes 10 to 12). Further studies showed that formation of these DNA–protein complexes was blocked by p50 antibody (lanes 13 to 16), indicating that the upper complex contained p50/p65 heterodimers and the lower complex contained p50/p50 homodimers. These results demonstrated that an NF-κB binding site existed at position c.−91 of the CYP4F2 gene, and the c.−91T→C variant altered its binding pattern. The c.−91T allele had the binding affinity to p50/p65 and p50/p50, whereas the c.−91C allele had the binding affinity only to p50/p65 heterodimers.

Figure 3.

EMSA with HEK293 nuclear extract. (A) Sense strands of putative NF-κB binding site. (B) Radiolabeled oligonucleotides NF-κB consensus (lanes 1 to 5), c.−91T (lanes 6 to 9), and c.−91C (lanes 10 to 12) were bound with nuclear extract. Competition assay was performed by adding 100- and 500-fold excess of unlabeled oligonucleotides. Radiolabeled c.−91T and c.−91C were bound with nuclear extract (lanes 13 and 15) and in the presence of p50 antibody (lanes 14 and 16). *Radiolabeled oligonucleotides. con., consensus oligonucleotide; anti-p50, p50 antibody; ns, nonspecific.

Effect of LPS Treatment on the Transcriptional Activity of Hap I

For confirmation that the raised CYP4F2 transcriptional activity of Hap I was attributed to the NF-κB mediated transactivation, LPS was added to the transient transfected HEK293 cells with various reporter constructs. Both pCAT564 and pCAT219 constructs showed elevated activities after LPS treatment. The magnitude of the increase was significantly greater in Hap I than in Hap II (25.28 versus 10.48% for pCAT564, and 22.64 versus 9.39% for pCAT219; P < 0.05; Figure 4); however, LPS treatment showed no effect on the pCAT67 constructs without NF-κB responsive element. These results suggested that LPS-stimulated NF-κB activation pathway contributed to the function of the CYP4F2 Hap I regulatory region.

Figure 4.

The basal and LPS-stimulated transcriptional activities of Hap I and Hap II in various reporter constructs. The final concentration of LPS is 1 μg/ml. *P < 0.05. All experiments were performed three times independently.

Impact of CYP4F2 Regulatory Haplotypes on Urinary 20-HETE

Because CYP4F2 is the major 20-HETE synthase in human kidney,¹ we measured the urinary 20-HETE levels in 132 participants to evaluate the functional consequence of the CYP4F2 regulatory haplotypes in vivo. No significant difference in gender, age, and body mass index (BMI) was observed among haplotypes. As urinary 20-HETE was not normally distributed; it underwent logarithmic transformation in statistic analysis and was presented as geometric mean (95% confidence intervals [CI]). The geometric mean of urinary 20-HETE levels in all patients was 12.59 ng/ml (95% CI 9.88 to 16.03). Significant difference among haplotypes was found in urinary 20-HETE, and the mean difference was significant for Hap I/Hap I versus Hap II/Hap II and for Hap I/Hap II versus Hap II/Hap II (Table 2). The upregulated effect of Hap I on 20-HETE production was observed in all models. In addition, the urinary 20-HETE in men was slightly higher than that in women. These results indicated that the CYP4F2 Hap I might contribute to the elevated renal formation of 20-HETE.

Table 2.

Urinary 20-HETE levels and the CYP4F2 regulatory haplotypes^a

Parameter	Hap I/Hap I (n = 75)	Hap I/Hap II (n = 53)	Hap II/Hap II (n = 4)
20-HETEb in total (ng/ml)	15.70 (11.51 to 21.39)	10.45 (7.06 to 15.46)	2.36 (0.49 to 11.42)
minimum (ng/ml)	1.20	0.60	0.79
maximum (ng/ml)	320.00	190.00	8.50
20-HETE in men (n = 66)		16.15 (11.41 to 22.88)
20-HETE in women (n = 66)		9.81 (7.02 to 13.70)

^aIn the ANOVA, P = 0.013, the mean difference was significant for Hap I/Hap I versus Hap II/Hap II and for Hap I/Hap II versus Hap II/Hap II. In a dominant model for Hap I, P = 0.015; in a recessive model for Hap I, P = 0.038. P = 0.041 for 20-HETE in men versus women.

^bAs urinary 20-HETE was not normally distributed, it was undergone logarithmic transformation in statistic analysis and presented as geometric mean (95% confidence intervals, CI).

Association of CYP4F2 Regulatory Haplotypes with Hypertension

We further estimated the association of CYP4F2 regulatory haplotypes with hypertension in a population-based case-control study and in a family-based study. The genotype frequencies in both studies were in Hardy-Weinberg equilibrium. As seen in Table 3, BMI, serum triglyceride, total cholesterol, and LDL cholesterol were risk factors for hypertension in the case-control population, and total cholesterol and LDL cholesterol were risk factors in hypertensive families. In the case-control study, the Hap I allele and homozygous Hap I genotype were modestly associated with hypertension (P = 0.035 and 0.027; Table 4). After adjustment for age, gender, BMI, triglyceride, total cholesterol, and LDL cholesterol, homozygous Hap I was an independent predictor of hypertension in this population (odds ratio [OR] 1.91; 95% CI 1.22 to 2.99; P = 0.005), particularly in women (OR 1.95; 95% CI 1.06 to 3.59; P = 0.033) and in age of 30 to 40 (OR 2.46; 95% CI 1.30 to 4.65; P = 0.006) after stratification by gender and age. For determination of the effect of Hap I on BP variation, each genotype was assessed in all models (Table 5). The homozygous Hap I carriers displayed higher diastolic BP (DBP) than the heterozygous and homozygous Hap II carriers in the recessive model.

Table 3.

Baseline characteristics of the participants in the case-control and family-based study^a

Characteristic	Case-Control Study			Family-Based Study
	Hypertensive (n = 269)	Control (n = 278)	P	Hypertensive (n = 189)	Control (n = 86)	P
Male/female (n)	127/142	128/150	0.784	103/86	42/44	0.383
Age (yr)	42.17 ± 12.89	40.66 ± 12.81	0.169	52.61 ± 13.30	41.75 ± 15.67	<0.001
BMI (kg/m2)	24.36 ± 4.77	22.02 ± 3.62	<0.001	25.03 ± 3.57	24.25 ± 3.85	0.166
SBP (mmHg)	146.80 ± 21.71	115.27 ± 10.58	<0.001	165.24 ± 21.95	125.63 ± 8.68	<0.001
DBP (mmHg)	93.08 ± 11.87	74.62 ± 7.32	<0.001	100.38 ± 24.85	80.39 ± 6.48	<0.001
Total cholesterol (mmol/L)	5.00 ± 1.07	4.23 ± 0.92	<0.001	5.40 ± 1.05	4.79 ± 0.97	<0.001
Triglycerides (mmol/L)	1.56 ± 1.35	1.05 ± 0.89	<0.001	1.89 ± 1.41	1.63 ± 1.44	0.235
HDL cholesterol (mmol/L)	1.58 ± 0.39	1.57 ± 0.35	0.791	1.50 ± 0.48	1.40 ± 0.31	0.166
LDL cholesterol (mmol/L)	3.00 ± 0.83	2.45 ± 0.72	<0.001	2.86 ± 0.72	2.47 ± 0.68	<0.001
Glucose (mmol/L)	4.50 ± 0.04	4.43 ± 0.05	0.328	4.75 ± 1.57	4.84 ± 2.57	0.730

^aData are means ± SD. SBP, systolic BP.

Table 4.

Frequencies of the CYP4F2 regulatory haplotypes in the case-control study

Parameter	Hypertensive (n = 269)	Control (n = 278)	P
Allele
Hap I (n [%])	484 (90.0)	477 (85.8)
Hap II (n [%])	54 (10.0)	79 (14.2)
OR (95% CI)	1.45 (1.03 to 2.15)			0.035
Genotype
Hap I/Hap I (n [%])	217 (80.7)	202 (72.7)
Hap I/Hap II (n [%])	50 (18.6)	73 (26.2)
Hap II/Hap II (n [%])	2 (0.7)	3 (1.1)
OR (95% CI)a	1.57 (1.05 to 2.35)			0.027
Adjusted OR (95% CI)a,b	1.91 (1.22 to 2.99)			0.005

^aIn a recessive model for Hap I. P ≥ 0.05 in a dominant and an additive model for Hap I.

^bAdjustment for age, gender, BMI, serum total cholesterol, serum triglyceride, and serum LDL.

Table 5.

Effect of the CYP4F2 regulatory haplotypes on BP and risk factors for hypertension^a

Parameter	Hap I/Hap I (n = 419)	Hap I/Hap II (n = 123)	Hap II/Hap II (n = 5)	Pb	Pc	Pd
SBP (mmHg)	131.64 ± 23.00	128.02 ± 23.77	126.00 ± 20.43	0.282	0.644	0.114
DBP (mmHg)	84.35 ± 13.56	81.52 ± 13.18	82.00 ± 10.37	0.118	0.778	0.039
MAP (mmHg)	108.00 ± 17.10	104.77 ± 17.04	104.00 ± 15.27	0.168	0.671	0.059
BMI (kg/m2)	23.02 ± 4.26	23.61 ± 4.64	25.19 ± 7.32	0.246	0.301	0.164
Total cholesterol (mmol/L)	4.60 ± 1.06	4.65 ± 1.11	4.68 ± 0.92	0.912	0.889	0.671
Triglycerides (mmol/L)	1.32 ± 1.24	1.23 ± 0.90	1.27 ± 0.83	0.766	0.952	0.467
LDL cholesterol (mmol/L)	2.72 ± 0.81	2.70 ± 0.88	2.77 ± 0.61	0.952	0.882	0.805

^aData are means ± SD. MAP, mean arterial pressure.

^bIn an additive model.

^cIn a dominant model for Hap I.

^dIn a recessive model for Hap I.

Transmission-disequilibrium test (TDT) analysis in the family-based study showed that the Hap I allele transmitted to affected offspring from heterozygous parents more frequently than expected, revealing that Hap I was significantly overtransmitted in the hypertensive offspring (z′ = 2.70, P < 0.05; Table 6), particularly in women (z′ = 3.16, P < 0.05) after stratification by gender. This was in line with the result from the population-based study that Hap I of CYP4F2 gene was associated with increased susceptibility to hypertension.

Table 6.

TDT and sibling TDT^a

Allele	TDT			Sibling TDT			Combined
	Transmitted	Not Transmitted	χ2	Y	Mean	z′	W	Mean	z′
Total
Hap I	40	22	5.23b	7	4.63	2.16b	47	35.63	2.70b
Hap II	22	40	5.23b	9	11.38	2.16b	31	42.38	2.70b
Male
Hap I	17	14	0.29	0	0	N/A	17	15.50	0.36
Hap II	14	17	0.29	0	0	N/A	14	15.50	0.36
Female
Hap I	23	8	7.26b	5	2.79	1.99b	28	18.29	3.16b
Hap II	8	23	7.26b	5	7.21	1.99b	13	22.71	3.16b

^aStatistical analysis was performed by the TDT-STDT Program 1.1. W, number of alleles combined from TDT and sibling TDT; Y, observed number of alleles among affected sibs.

^bP < 0.05.

Discussion

In this work, we reported for the first time that two common haplotypes, Hap I (c.−91T/c.−48G/c.−13T/c.+34T) and Hap II (c.−91C/c.−48C/c.−13C/c.+34G), existed in the regulatory region of the CYP4F2 gene. Hap I increased transcriptional activity via an NF-κB responsive element at c.−91T, resulting in elevated urinary 20-HETE level, and was associated with hypertension in both the case-control and family-based studies.

The CYP4F2 gene consists of 13 exons, and the ATG translation initiation codon was located in exon 2. Zhang et al.¹⁴ showed that the 5′-regulatory region of CYP4F2 was characterized by a proximal promoter, exon 1 and intron 1. The proximal promoter contained a typical TATA box, three consecutive TFIID binding sites, and two CCAAT boxes. The untranslated exon 1 had a repressive effect on gene transcription, whereas the intron 1 had a remarkable higher transcriptional activity than the promoter. We showed previously that a single variant, 421G→C (c.−48G→C), in the intron 1 was associated with hypertension,¹⁰ whereas constitution and function of the CYP4F2 risk haplotype for hypertension remained unknown. In this work, two common regulatory haplotypes were defined in the CYP4F2 intron 1, and each variant of haplotype could be used as a tag in association study. A recent report¹⁵ on CYP4F2 variants in the coding region demonstrated that the W12G (c.+34T→G) variant exhibited perfect LD with surrounding variants, which was in accordance with our result.

EMSA and reporter assay showed that the c.−91T→C variant altered the NF-κB binding pattern, and different NF-κB transactivation resulted in a trend toward higher basal transcriptional activity and greater LPS-stimulated transcriptional activity of the Hap I constructs than the corresponding Hap II constructs. The greater magnitude of increase in pCAT219 than in pCAT564 might be explained by the complex regulatory region in the pCAT564 construct, including the proximal promoter and the repressive exon 1, whereas the pCAT219 construct contained only intron 1. The lower activity of Hap I than Hap II in pCAT67 with only c.−48G→C was due to the previously identified negative Myb responsive element at c.−48G,¹⁰ and this result further indicated a potent upregulatory effect of the NF-κB transactivation pathway on CYP4F2 transcription.

NF-κB is a pivotal transcription factor that participates in the regulation of various target genes in different cells to exert its biologic functions. Our result from EMSA showed that p50/p65 and p50/p50 subunits of NF-κB bound to the c.−91T allele with increased transcriptional activity, whereas the c.−91C allele lost the binding affinity to p50/p50. The p50/p50 homodimer was described as a transactivator or a repressor in gene transcription,^16,17 and herein it seemed to enhance the transcription of CYP4F2. LPS was reported to activate a cell-signaling pathway that culminated in the activation of NF-κB and led to transcriptional regulation of genes¹⁸; therefore, we compared the basal or LPS-stimulated transcriptional activity of Hap I with Hap II. The increase of transcriptional activities by LPS treatment was greater in Hap I than in Hap II, indicating that the Hap I regulatory region with increased NF-κB binding affinity at c.−91T could increase the transcription of CYP4F2 through LPS-stimulated NF-κB transactivation pathway. On the basis of these findings, we hypothesized that NF-κB activation induced by environmental factors might upregulate the CYP4F2 expression and renal 20-HETE formation, resulting in an increased risk for hypertension in the Hap I homozygotes through a gain-of-function mechanism.

The effect of CYP4F2 gene plays a dual role in the regulation of BP via its metabolite from arachidonic acid in human kidney, 20-HETE. To elucidate the impact of CYP4F2 regulatory haplotypes on renal 20-HETE formation, we measured urinary 20-HETE levels. Our result showed that an increased urinary 20-HETE was significantly associated the Hap I, indicating that CYP4F2 protein or activity was increased in the Hap I carriers; however, it should be considered that variants in the coding region of CYP4F2, such as V433M, or other ω-hydroxylases, such as CYP4A11, might also participate in this metabolism. In addition, men had slightly higher levels of 20-HETE than women, which was in accordance with the report by Ward et al.⁸

In the renal arterioles, 20-HETE acts as a potent vasoconstrictor that contributes to the autoregulation of renal blood flow and renal vascular tone and elevates BP; at the level of the renal tubule, 20-HETE inhibits sodium reabsorption, increases sodium excretion, and opposes the development of hypertension.^4,5 In addition, renal CYP4F2 expression and 20-HETE formation would occur in the nephron, where they entered systemic circulation and exerted prohypertensive action.¹ Our case-control association study showed that homozygous Hap I was an independent predictor for hypertension, and Hap I elevated DBP in a recessive model. This relevance of Hap I to hypertension was confirmed by a TDT analysis in the family-based study and by population stratification. Considerable evidence from animal models supported that renal CYP activity and 20-HETE content were increased in hypertension. In the SHR, the CYP4A2 gene was preferentially overexpressed in the kidney,¹⁹ the renal 20-HETE was elevated, and agents that inhibit 20-HETE formation attenuated the development of hypertension.⁷ The delivery of CYP4A1 cDNA increased BP in Sprague-Dawley rat, whereas transferring antisense CYP4A1 cDNA into SHR decreased BP.²⁰ Our findings that Hap I of CYP4F2 regulatory region was associated with elevated urinary 20-HETE levels as well as increased risk for hypertension supported a prohypertensive effect of renal 20-HETE and CYP4F2 expression; however, Gainer et al.⁹ reported that a functional variant of CYP4A11 with reduced 20-HETE synthase activity was associated with hypertension, suggesting an antihypertensive action of 20-HETE. In vivo investigations of the synergistic or antagonistic effect of functional polymorphisms as well as haplotypes in the CYP4F2 and CYP4A11 genes would be the focus of further studies.

In summary, our findings indicate that a functional haplotype in the regulatory region of the CYP4F2 gene with increased NF-κB binding affinity is related to elevated urinary 20-HETE and increased susceptibility to hypertension in a Chinese population.

Concise Methods

Study Population

All patients were of Han Chinese origin and recruited from a relatively isolated Northeastern Chinese population. BP was measured and averaged three times from the right arm of seated patients after at least 5 min of rest. Hypertension was defined as an average systolic BP ≥140 mmHg or an average DBP ≥90 mmHg or current use of antihypertensive medication, excluding diabetes, renal disease, or secondary hypertension by history inquiry and physical examination. In the case-control study, 269 unrelated hypertensive patients and 278 unrelated normotensive individuals without history of cardiovascular diseases were recruited and matched for age and gender. The family-based study consisted of 66 families with 275 members who were not included in the case-control population. Each family had at least one affected parent and two affected offspring. Informed consent was obtained from all participants, and their blood and urine samples were allowed to be used in this study. The local ethics committee approved the study protocol.

Polymorphism Discovery and Genotyping

Genomic DNA was extracted and amplified from c.−209 to c.+419 of the CYP4F2 gene by PCR using the primers 5′-AGTGCTTACTAGGGAACTGGAG-3′ and 5′-AAGGATTCAATGCAGGCCTGGA-3′. Genetic variants were screened by direct sequencing of 130 amplicons. The c.−13T→C variant (rs3093103) was genotyped in the case-control population and hypertensive families by PCR-restriction fragment length polymorphism method with the primers 5′-CCTCTCAGCCCCTTTTTACC-3′ and 5′-ACTCCCTAAGCCTCGTACCC-3′, and Cac8I. Nucleotides were numbered according to the GeneBank reference sequence AF467894, where the A of the ATG translation initiation codon was +1.

Plasmid Construction

The regulatory regions of the CYP4F2 gene from c.−554 to c.+10 (564 bp) and from c.−209 to c.+10 (219 bp) were amplified by PCR with the forward primers 5′-GGCTGCAGGGATTGGTTGGC-3′ and 5′-AGTGCTTACTAGGGAACTGGAG-3′ and a common reverse primer 5′-GCTGGGACATCCTGCAGGGC-3′ using genomic DNA of different haplotypes as templates. These fragments cloned into the HindIII/XbaI sites of pCAT-Basic (Promega, Madison, WI) to generate pCAT564 and pCAT219 reporter constructs. The pCAT219 constructs with −91C/−48G and with −91T/−48C were constructed by the reverse megaprimers 5′-GCTGGGACATCCTGCAGGGCAGACGGGATGGACGGTGAGATCCTGAGGCCCAGAGAACGGCCCAG-3′ and 5′-GCTGGGACATCCTGCAGGGCAGACGGGATGGACGGTGAGATCCTGAGGCCCAGAGAAGGGCCCAG-3′. The pCAT67 constructs were generated by annealing synthesized oligonucleotides (from c.−57 to c.+10) with HindIII/XbaI sites. All constructs were confirmed by sequencing.

Transient Transfection and Reporter Assays

For transient transfection, the HEK293 cells were transfected with 4.0 μg/well plasmid DNA and 10 μl/well lipofectamine (Invitrogen, Carlsbad, CA). The cells were co-transfected with 0.2 μg/well pSV-β-Galactosidase vector (Promega) as a control. The CAT assay was performed using a CAT-ELISA kit (Roche, Mannheim, Germany) according to the manufacturer's instructions.

Preparation of Nuclear Extracts and EMSA

Nuclear extracts were prepared from HEK293 cells following the method of Andrews and Faller.²¹ The EMSA was performed as described previously.²² The sense strands of the c.−91T and c.−91C oligonucleotides were 5′-ACCTCCGGCACTGCCCGTCCCT-3′ and 5′-ACCTCCGGCACCGCCCGTCCCT-3′. The NF-κB consensus oligonucleotide (Promega) was used as a positive binding control. The supershift assay was performed by preincubation of nuclear extract with p50 antibody (Santa Cruz Biotechnology, Santa Cruz, CA) for 15 min at room temperature before the radiolabeled oligonucleotides were added.

Measurement of Urinary 20-HETE

Urine samples were collected from 132 genotyped participants with a typical diet in the case-control population. The urinary 20-HETE levels were assessed using 20-HETE ELISA kits (R&D, Detroit, MI) according to the manufacturer's instructions.

Statistical Analyses

Hardy-Weinberg equilibrium was tested by χ² test. Haplotype reconstruction was performed using PHASE software (version 2.0) (http://www.stat.washington.edu/stephens/software.html), and the pairwaise LD was calculated with HaploXT 1.1 (http://archimedes.well.ox.ac.uk/pise/haploxt.html). TDT was performed by TDT-STDT Program 1.1 (http://genomics.med.upenn.edu/spielman/WinEd.html). Statistical analyses were done with SPSS 10.0 for windows (SPSS, Chicago, IL).

Disclosures

None.