G protein-coupled receptor kinase 4 gene variants in human essential hypertension Robin A. Felder*, Hironobu Sanada*†, Jing Xu‡, Pei-Ying Yu‡, Zheng Wang‡, Hidetsuna Watanabe*, Laureano D. Asico*, Wei Wang‡, Shaopeng Zheng‡, Ikuyo Yamaguchi‡, Scott M. Williams§, James Gainer¶, Nancy J. Brown¶, Debra Hazen-Martin储, Lee-Jun C. Wong**, Jean E. Robillard††, Robert M. Carey‡‡, Gilbert M. Eisner‡, and Pedro A. Jose‡ *Department of Pathology, University of Virginia Health Sciences Center, Charlottesville, VA 22908; ‡Department of Pediatrics and Physiology and Biophysics, Georgetown University Medical Center, Washington, DC 20007; §Department of Microbiology, Meharry Medical College, Nashville, TN 37208; ¶Department of Medicine and Pharmacology, Vanderbilt University Medical Center, Nashville, TN 37232; 储Department of Pathology, Medical University of South Carolina, Charleston, SC 29403; **Institute for Molecular and Human Genetics, Georgetown University Medical Center, Washington, DC 20007; ††Department of Pediatrics, University of Michigan College of Medicine, Ann Arbor, MI 48109; and ‡‡Department of Medicine, University of Virginia Health Sciences Center, Charlottesville, VA 22908 Communicated by Maria Iandolo New, Weill Medical College of Cornell University, New York, NY, December 21, 2001 (received for review August 10, 2001)
Essential hypertension has a heritability as high as 30 –50%, but its genetic cause(s) has not been determined despite intensive investigation. The renal dopaminergic system exerts a pivotal role in maintaining fluid and electrolyte balance and participates in the pathogenesis of genetic hypertension. In genetic hypertension, the ability of dopamine and D1-like agonists to increase urinary sodium excretion is impaired. A defective coupling between the D1 dopamine receptor and the G protein兾effector enzyme complex in the proximal tubule of the kidney is the cause of the impaired renal dopaminergic action in genetic rodent and human essential hypertension. We now report that, in human essential hypertension, single nucleotide polymorphisms of a G protein-coupled receptor kinase, GRK4␥, increase G protein-coupled receptor kinase (GRK) activity and cause the serine phosphorylation and uncoupling of the D1 receptor from its G protein兾effector enzyme complex in the renal proximal tubule and in transfected Chinese hamster ovary cells. Moreover, expressing GRK4␥A142V but not the wild-type gene in transgenic mice produces hypertension and impairs the diuretic and natriuretic but not the hypotensive effects of D1-like agonist stimulation. These findings provide a mechanism for the D1 receptor coupling defect in the kidney and may explain the inability of the kidney to properly excrete sodium in genetic hypertension.
L
ong-term regulation of blood pressure is vested in the organ responsible for the control of body fluid volume, the kidney (1, 2). Dopamine facilitates the antihypertensive function of the kidney because it is both vasodilatory and natriuretic (3). Dopamine (produced by renal proximal tubules) via D1-like receptors is responsible for over 50% of incremental sodium excretion when sodium intake is increased (3–6). The paracrine兾 autocrine dopaminergic regulation of sodium excretion is mediated by tubular but not by hemodynamic mechanisms (6). The ability of dopamine and D1-like agonists to decrease renal proximal tubular sodium reabsorption is impaired in genetic rodent hypertension and human essential hypertension (3, 5, 7–15). Indeed, the aberrant D1-like receptor function in the kidney precedes and cosegregates with high blood pressure in spontaneously hypertensive rats. In addition, disruption of the D1 receptor in mice produces hypertension (12, 13). The pivotal role of dopamine in the excretion of sodium after increased sodium intake has led to the hypothesis that an aberrant renal dopaminergic system is important in the pathogenesis of some forms of genetic hypertension (3, 5, 7–17). Several mechanisms potentially responsible for the failure of endogenous renal dopamine to engender a natriuretic effect in genetic hypertension have been investigated and ruled out, including decreased renal dopamine production and receptor expression, aberrant nephron segment distribution of dopamine receptors, defective effector enzymes (adenylyl cyclase or phospholipase C), and 3872–3877 兩 PNAS 兩 March 19, 2002 兩 vol. 99 兩 no. 6
abnormal renal sodium transporters (3, 8, 13, 17). Also, the coding region of the D1 receptor is unchanged in hypertensive subjects (16), as well as in rodents with genetic hypertension (unpublished studies). In renal proximal tubules from humans with essential hypertension and from rodents with genetic hypertension, the D1-like receptor is uncoupled from its G protein兾effector enzyme complex (3, 7–10, 16). This uncoupling is thought to be the mechanism for the failure of dopamine to engender a natriuresis in genetic hypertension (3, 5, 11–15). This mechanism is similar to but distinct from homologous desensitization (18, 19) because the uncoupling in hypertension is independent of renal dopamine levels (3, 16, 20). Similarly, the uncoupling is not due to heterologous desensitization because the responsiveness of other G protein-coupled receptors (e.g., parathyroid hormone, adrenergic, and cholecystokinin receptors) remains intact in the prehypertensive spontaneously hypertensive rat (3, 8, 21–23). G protein-coupled receptor kinases (GRKs) have been implicated in genetic and acquired hypertension because they participate in the desensitization of G protein-coupled receptors, including D1 receptors. The GRK-mediated desensitization is caused, in part, by serine phosphorylation of the receptor (18, 19). We have reported that basal serine-phosphorylated D1 receptor is increased in renal proximal tubules from genetically hypertensive rodents as well as from humans with essential hypertension (3, 16). The seven members of the GRK family are divided into three subfamilies: GRK1 and GRK7 belong to the rhodopsin kinase subfamily; GRK2 and GRK3 belong to the -adrenergic receptor kinase subfamily; and GRK4, GRK5, and GRK6 belong to the GRK4 subfamily (24, 25). GRK2 expression and activity are increased in lymphocytes from patients with essential hypertension (26) and genetically hypertensive rats (27). However, the increase in GRK2 expression and activity in the spontaneously hypertensive rat followed the development of hypertension (27). Moreover, we found no difference in the sequence of the coding region of GRK2 between hypertensive and normotensive human subjects (unpublished data). It is possible that the increase in GRK2 activity in lymphocytes of hypertensive patients (26, 27) is secondary to the high blood pressure, as has been suggested for the increase in GRK5 activity and expression in rodents with genetic and induced hypertension (28). Moreover, the ubiquiAbbreviations: GRK, G protein-coupled receptor kinase; CHO, Chinese hamster ovary; SNP, single nucleotide polymorphism. †To
whom reprint requests should be addressed at: c兾o Robin A. Felder, Department of Pathology, Box 168, Charlottesville, VA 22908. E-mail:
[email protected].
The publication costs of this article were defrayed in part by page charge payment. This article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. §1734 solely to indicate this fact.
www.pnas.org兾cgi兾doi兾10.1073兾pnas.062694599
tous expression of GRK2 and GRK3 is at odds with the recognized preeminence of the kidney in the pathogenesis of both rodent and human essential hypertension (1, 2). The limited expression of GRK4 (24, 25) and the fact that the GRK4 gene locus is linked to hypertension (29) make GRK4 an attractive candidate for a pathogenetic mechanism in human hypertension. Therefore, we sought to determine whether GRK4 is expressed in renal proximal tubules and whether genetic variants of GRK4 affect renal function and blood pressure. Methods
Fig. 2. Phosphorylation of the D1 receptor in renal proximal tubule cells from hypertensive (HT) subjects. Lysates of renal proximal tubule cells were immunoprecipitated with a D1 receptor antibody and immunoblotted with an anti-phosphoserine antibody. The amount of basal phosphorylated D1 receptor was greater in hypertensive than in normotensive subjects. Fenoldopam (5 M) increased the phosphorylation of the D1 receptor in cells from normotensive but not in cells from hypertensive subjects. Sense or scrambled GRK4 oligonucleotides had no significant effect on the quantity of phosphorylated D1 receptor in hypertensive or normotensive subjects. In contrast, antisense GRK4 oligonucleotides inhibited the phosphorylation of the D1 receptor in both hypertensive or normotensive subjects (HT ⬎ NT). (Inset) Antiphosphoserine immunoblots of the anti-D1 receptor antibody immunoprecipitates, unless otherwise indicated. Lanes 1– 4, HT; lanes 6 –9, NT; lanes 1 and 6, basal phosphorylation; lanes 2 and 7, effect of fenoldopam; lanes 3 and 8, effect of GRK4 sense兾scrambled oligonucleotides; lanes 4 and 9, effect of GRK4 antisense oligonucleotides; lane 10, Western blot of human proximal tubule cells with D1 antibody preadsorbed with the immunizing peptide; and lane 11, Western blot of human proximal tubule cells with D1 antibody. Lane 5, molecular size marker, 80 kDa. The graph depicts the composite studies from five hypertensive and four normotensive subjects. Homozygous GRK4 gene variants were found in four of the five hypertensive subjects. The % area was normalized to 100% for either hypertensive or normotensive subjects. The inhibition of phosphorylation of the D1 receptor was associated with an enhancement of the fenoldopam-induced increase in cAMP accumulation in hypertensive subjects (see Fig. 3). a, P ⬍ 0.05 normotensive vs. hypertensive, t test; *, P ⬍ 0.05 vs. other hypertensive groups, ANOVA, Scheffe´’s test; #, P ⬍ 0.05 vs. normotensive basal, ANOVA, Scheffe´’s test; and ⫹, P ⬍ 0.05 vs. normotensive fenoldopam alone, t test. Data are mean ⫾ SE.
Tissue Culture. Human kidneys were obtained as fresh surgical
specimens from white patients who had unilateral nephrectomy because of renal carcinoma. The patient records of the subjects were reviewed and classified into those with either normal blood pressure (n ⫽ 9) or essential hypertension (n ⫽ 14). Subjects with systolic blood pressures less than 140 mm Hg and diastolic blood pressures less than 90 mm Hg were considered normotensive. Subjects with systolic blood pressures equal to or greater than 140 mm Hg and兾or diastolic blood pressures equal to or greater than 90 mm Hg and兾or on antihypertensive medications were considered hypertensive. Cultures of renal proximal tubule cells from histologically verified normal sections (5 ⫻ 105 cells per well in 24-well plastic plates coated with 0.075% Type I collagen) were incubated at 37°C in 95% air兾5% CO2 and grown in a serum-free medium consisting of a 1:1 mixture of DMEM and Ham’s F-12 medium supplemented with selenium (5 ng兾ml), insulin (5 g兾ml), transferrin (5 g兾ml), hydrocortisone (36 ng兾ml), triiodothyronine (4 pg兾ml), and epidermal growth factor (10 ng兾ml) (16). When subconfluent (90–95%), the cells were subcultured (passages 6–8) for use in experimental protocols by using trypsinEDTA (0.05%, 0.02%). The culture conditions are conducive for growth of human renal proximal tubules that retain characteristics of renal proximal tubule cells (16). Transfection and Cell Culture. The rat D1 (rD1) receptor cDNA was subcloned in the expression vector pPUR (CLONTECH). Calcium phosphate-mediated transfection of the resulting construct Felder et al.
was used to create stable Chinese hamster ovary (CHO) cell line expressing the pTet-Off regulator plasmid (CLONTECH; ref. 30). GRK4␥ and GRK4␦ cDNAs, obtained from reverse transcription–PCR of mRNA from human kidney cortex, were subcloned into a pTet-Off response plasmid (CLONTECH). Additionally, CHO cell lines stably transfected with rD1 were transiently transfected with GRK4␥ wild type or variants with TransIT-LT2 (Panvera, Madison, WI). Determination of cAMP Accumulation. The cells were washed twice
with D-PBS, after which 1 mM 3-isobutyl-1-methyl-xanthine was added to each well. The cells were incubated at 37°C for 30 min in the presence or absence of drugs: D1-like receptor agonist, fenoldopam, or the D1-like receptor antagonist SCH23390 (Research Biochemicals, Natick, MA), or forskolin (Sigma). Then, the cells were lysed with 0.1 M HCl and frozen at ⫺80°C. cAMP concentration was measured by RIA (13, 16). Protein concentration was measured with the BCA protein assay kit (Pierce). Light Microscopic Immunohistochemistry. Immunohistochemistry
of kidney tissues and cells in culture fixed in HISTOCHOICE (Amresco, Solon, OH) was performed as described (16). Affinity-column purified polyclonal human D1 receptor antibodies were raised against a synthetic peptide sequence GSGETQPFC (amino acids 299–307; ref. 16). Two commercially available GRK4 isoform antibodies were used (Santa Cruz BiotechnoloPNAS 兩 March 19, 2002 兩 vol. 99 兩 no. 6 兩 3873
MEDICAL SCIENCES
Fig. 1. D1-like agonist stimulation of GRK activity in renal proximal tubule cells from hypertensive (HT) subjects. The D1-like agonist fenoldopam (5 M) increased GRK activity (measured by the phosphorylation of rhodopsin) in HT but not in normotensives (NT) with time. (Inset) Fenoldopam stimulation of rhodopsin phosphorylation in HT (⬇40 kDa); addition of GRK4␥兾␦ antibody (GRK4; Inset) decreased the 10-min phosphorylation of rhodopsin (Inset). GRK activity was measured by the phosphorylation of rhodopsin (31). Of the four isoforms of GRK4 in humans, only GRK4␣ can phosphorylate rhodopsin (25). Because the D1 agonist did not stimulate GRK activity in cells from normotensive subjects, the effect of GRK ␣兾 antibody was not tested. Number of studies: n ⫽ 5兾group except at 1 min and 5 min where n ⫽ 4兾group. #, P ⬍ 0.05 vs. 0 time, t test; *, P ⬍ 0.05 vs. 0 time, ANOVA for repeated measures, Scheffe´ ’s test; a, P ⬍ 0.05 HT vs. NT, t test. Data are mean ⫾ SE.
phosphorothioate oligonucleotides (5 nM) for 4–16 h. The cells were then washed and reincubated for a total incubation time of 24 to 72 h. Two antisense oligonucleotides were used: 5⬘-CAC GAT GTT CTC GAG CTC CAT-3⬘, complementary to bases 255–275, and 5⬘-CTC CAT GTC CTG GCG CCG-3⬘, complementary to bases 243–260 (25). Sense (5⬘-ATG GAG CTC GAG AAC ATC GTG-3⬘) or scrambled antisense (5⬘-ACC CTT GCG TCC GCT GCG-3⬘) oligonucleotides were used as controls. These sequences are common to all GRK4 isoforms. GRK4 and GRK6 have the same sequence in the first 18 bases but not in the 5⬘ non-coding region (24, 25, 32). Immunoprecipitation. Proximal tubule cells were incubated with
Fig. 3. Normalization of D1-like agonist-induced stimulation of cAMP accumulation in renal proximal tubule cells from hypertensive (HT) subjects by GRK4 antisense oligonucleotides. GRK4 antisense propyne兾phosphorothioate oligonucleotides normalized the ability of fenoldopam to stimulate cAMP accumulation in HT; a slight increase was noted in normotensives (NT). GRK4 sense or scrambled oligonucleotides did not significantly affect the ability of fenoldopam to stimulate cAMP accumulation in either NT or HT. Basal cAMP production was similar in hypertensive (940 ⫾ 34 fmol兾mg protein兾30 min, n ⫽ 8) and normotensive subjects (1,015 ⫾ 36 fmol兾mg protein兾30 min, n ⫽ 6; P ⫽ 0.10, t test). GRK4␥兾␦ immmunoreactive levels (Inset) and GRK4␣兾 (not shown) were attenuated by antisense (5⬘-CAC GAT GTT CTC GAG CTC CAT-3⬘; lane 2, NT; lane 5, HT) but not by sense兾scrambled oligonucleotides (lane 3, NT; lane 6, HT) compared with vehicle-treated controls (lane 1, NT; lane 4, HT). Results were similar by using two different oligonucleotides as described in Methods. Four of the eight hypertensive subjects were homozygous in at least one of the polymorphic sites (nucleotide positions 448 and 679, n ⫽ 1; 679, n ⫽ 1; and 1711, n ⫽ 2). Uncoupling without GRK4 gene variants could be interpreted to indicate the presence of other meaningful GRK4 polymorphisms or there other causes of uncoupling besides GRK4 gene variants. Data are mean ⫾ SE. Number of experiments is in parentheses. Each n represents cells from one subject. a, P ⬍ 0.05 NT vs. HT, t test; *, P ⬍ 0.05 vs. other groups in HT, ANOVA for repeated measures, Scheffe´ ’s test.
gy); one GRK4 antibody recognized both the ␣ and  isoforms, whereas another recognized both the ␥ and ␦ isoforms. The specificity of these antibodies has been previously reported (24). Determination of GRK Activity. GRK activity was measured according to Benovic (31). Renal proximal tubular extracts were prepared by homogenization in ice-cold lysis buffer containing (in mM): 25 Tris䡠HCl (pH 7.5), 5 EDTA, 5 EGTA with leupeptin (10 g兾ml), aprotinin (20 g兾ml), and 1 PMSF. The crude homogenate was centrifuged at 30,000 ⫻ g for 30 min. The pellet was solubilized by 200 mM NaCl on ice for 30 min and centrifuged at 30,000 ⫻ g for 30 min. The supernatant was used for all GRK assays and immunoblotting. Twenty micrograms of protein extract was incubated with rhodopsin-enriched rod outer segments in assay buffer with 10 mM MgCl2 and 0.1 mM ATP (containing [␥-32P]ATP). After incubation in white light for 15 min at room temperature, the reaction was stopped with ice-cold lysis buffer and centrifuged at 13,000 ⫻ g for 15 min. The pellet was resuspended in Laemmli buffer and subjected to 12% SDS兾PAGE. The gels were subjected to autoradiography, and the phosphorylated rhodopsin was quantified by using both densitometry and radioactive counting of the excised bands at the appropriate size. GRK activity was also measured in the presence or absence of a GRK4 ␥兾␦ antibody.
GRK4 Oligonucleotides. The effects of inhibition of GRK4 activity
on cAMP accumulation and levels of immunoreactive GRK4 ␥兾␦ were determined by treating renal proximal tubule cells with vehicle, sense, scrambled, or antisense GRK4 propyne兾 3874 兩 www.pnas.org兾cgi兾doi兾10.1073兾pnas.062694599
vehicle, fenoldopam, sense, scrambled, or antisense propyne兾 phosphorothioate GRK4 oligonucleotides (5 nM) as described above. The membranes were lysed with ice-cold lysis buffer (PBS兾1% Nonidet P-40兾0.5% sodium deoxycholate兾0.1% SDS兾1 mM EDTA兾1 mM EGTA兾1 mM sodium vanadate兾1 mM NaF兾1 mM PMSF兾10 g/ml aprotinin兾10 g/ml leupeptin). The lysates were incubated with IgG-purified anti D1 receptor on ice for 1 h and protein-A agarose for 12 h with rocking at 4°C. The proteins separated by SDS兾PAGE were electrophoretically transferred onto nitrocellulose membranes. The transblot sheets were blocked with 5–10% nonfat dry milk in 10 mM Tris䡠HCl (pH 7.5), 150 mM NaCl, and 0.1% Tween-20 and were incubated with diluted affinity-purified polyclonal anti-phosphoserine antibody (Zymed; ref. 16). In some cases, the cells were labeled with 32P and immunoprecipitated with anti-D1 receptor antibody. The autoradiograms and immunoblots, visualized with the ECL system (Amersham Pharmacia), were quantified by densitometry (16). Transgenic Mice. Two constructs were used to generate transgenic mice. The full-length wild-type hGRK4␥ cDNA was obtained by PCR by using the GRK4␣ cDNA in pTRE plasmid as template. The A142V polymorphism was obtained by using site-directed mutagenesis. The two cDNAs were subcloned into pcDNA3.1. Expression of the cDNA insert was under the control of the cytomegalovirus promoter and bovine growth hormone (BGH) poly(A) signal. Full-length cDNA was verified by sequencing. The mice were produced by microinjecting the cDNA constructs into fertilized eggs obtained from the mating of a (C57BL兾6J ⫻ SJL兾J) F1 female mouse and a (C57BL兾6J ⫻ SJL兾J) F1 male mouse at the University of Michigan Transgenic Animal Model Core. The presence of the transgene in the transgenic mice was verified by PCR. The first set of PCR primers (pcDNA, sense 5⬘-CGACTCACTATAGGGAGAC-3⬘; hGRK4, antisense 5⬘-ATGGTTCCCCTCTTAGGTAG-3⬘) generated a 530-bp fragment. The second set PCR primers (pcDNA, sense 5⬘-CGACTCACTATAGGGAGAC-3⬘; hGRK4, antisense 5⬘-CTTGATTCTTTGATCGACCTCCTCCC-3⬘) generated a 1,260-bp fragment. Twenty-two mice carrying wild-type hGRK4␥ and 10 mice carrying hGRK4␥ A142V were identified. Both wild-type and A142V hGRK4␥ were expressed in the kidney as determined by reverse transcription–PCR, immunoblotting, and immunohistochemistry. Blood Pressure and Renal Functional Studies. The animal studies
were approved by the Georgetown University Animal Care and Use Committee. The mice were anesthetized with pentobarbital (50 mg兾kg i.p.), placed on a heated board to maintain body temperature at 37°C, and tracheotomized (33). Catheters were inserted into the femoral vessels for fluid administration, blood drawing, and blood pressure monitoring. Urine was collected via a suprapubic cystostomy. After a 60-min equilibration period, a baseline 30-min collection period was obtained. Thereafter, fenoldopam was infused intravenously at 2 g兾kg body weight兾 min for 30 min. Urine was collected during the drug infusion and Felder et al.
sion of the experiment, the mice were killed with an i.v. injection of pentobarbital (100 mg兾kg).
Fig. 4. (A) Effect of GRK4␥ variants on D1-like agonist stimulation of cAMP production in CHO cells stably transfected with D1 receptor and GRK4␥ in tet-off vector. The filled symbols represent CHO cells treated with tetracycline in which GRK4␥ was turned off and served, therefore, as controls. In the absence of GRK4␥ expression, fenoldopam (10⫺5 M) increased cAMP accumulation to a similar extent in wild-type (W) and single (R65L, A142V, or A486V) or double variant (R65L and A486V)-transfected cells. The open symbols represent CHO cells not treated with tetracycline and therefore, GRK4␥ was expressed. Expression of wild-type GRK4␥ (W, E) decreased the ability of fenoldopam to stimulate cAMP accumulation. The decrease was greater with GRK4␥ variants at R65L (CGT to CTT; ƒ), or A486V (GCG to GTG; 䊐) and even greater with A142V (GCC to GTC; ‚) and the double variant (R65L and A486V; ⽧). Studies were performed in CHO cells expressing similar amounts of GRK4␥ and D1 receptor. Fenoldopam had no effect on cAMP accumulation in untransfected CHO cells, CHO cells with tet-off regulator, or response plasmid alone, or CHO cells expressing only GRK4␥ without the D1 receptor (data not shown). Basal cAMP accumulation was similar among the groups in the presence or absence of tetracycline (527 ⫾ 5 fmol兾mg protein with tetracycline兾30 min and 522 ⫾ 5 fmol兾mg protein without tetracycline兾30 min). Data are mean ⫾ SE (error bars are absent if the symbols are bigger than the error bars). n, The number of cell lines studied per group, performed in triplicate. #, P ⬍ 0.05 wild-type vs. R65L ⫹ A486V, A142V; *, P ⬍ 0.05 wild-type vs. others, ANOVA, Scheffe´ ’s test (cells not treated with tetracycline). (B) Effect of GRK4␥ variants on serine phosphorylation of the D1 receptor in CHO cells stably (n ⫽ 3) or transiently (n ⫽ 5) transfected with rD1 receptor and GRK4␥ in tet-off vector. Lysates of CHO cells were immunoprecipitated with anti-D1 antibody and immunoblotted with an anti-phosphoserine antibody (in two studies, the cells were labeled with 32P). (Inset) Compared with untransfected CHO cells (lane 1) and CHO cells transfected with wild-type GRK4␥ (lane 2), phosphorylation of the D1 receptor was increased in CHO cells transfected with the GRK4␥ variants R65L (lane 3), A486V (lane 4), R65L兾A486V (lane 5), and A142V (lane 6). *, P ⬍ 0.05 vs. GRK4␥ wild-type or nontransfected CHO cell; #, P ⬍ 0.05 vs. A142V or R65L兾A486V, ANOVA, Duncan’s test.
for another two 30-min periods, thereafter. Blood (50 l) was obtained from the femoral artery before the drug infusion and at the end of the last urine collection period. In some mice, bolus injections of fenoldopam (1, 10, 100, and 1000 ng) and the effect on blood pressure were monitored for 10 min, and the time for blood pressure to recover to preinjection levels. At the concluFelder et al.
Results and Discussion GRK4 was previously thought to be expressed mainly in testes and the brain (24, 25). However, GRK4␥ has been reported to be expressed in human myometrium (34). We detected mRNA of all of the reported GRK4 isoforms in renal proximal tubules (not shown). There were no differences in the protein expression of the isoforms of GRK4 (␣兾, ␥兾␦) in kidneys or cultured renal proximal tubule cells between hypertensive and normotensive subjects (not shown). However, basal and D1 agonist-stimulated GRK activities were elevated in renal proximal tubular cells from hypertensive subjects (Fig. 1). The increased GRK activity in cells from hypertensive subjects was probably caused, in part, by GRK4␥兾␦ because the antibody that recognizes these two GRK4 isoforms blocked the stimulatory effect of the D1-like agonist, fenoldopam, on GRK activity. An initial step in the desensitization process is the phosphorylation of the G protein-coupled receptor by GRK (18, 19). In our studies, basal (ligand-independent) GRK activity and serine phosphorylation of the D1 receptor in renal proximal tubule cells was higher in hypertensive subjects than in normotensive subjects (Figs. 1 and 2). As expected, fenoldopam increased the phosphorylation of the D1 receptor and cAMP accumulation in cells from normotensive subjects. In contrast, fenoldopam failed to increase further the phosphorylation of the D1 receptor and minimally increased cAMP accumulation in cells from hypertensive subjects (Figs. 2 and 3). The increased GRK4 activity was related to the diminished responses of the D1 receptor in hypertension because antisense GRK4 oligonucleotides completely blocked the serine phosphorylation of the D1 receptor and restored the ability of fenoldopam to stimulate cAMP accumulation in cells from hypertensive subjects. In contrast, in cells from normotensive subjects, antisense GRK4 slightly increased the ability of fenoldopam to stimulate cAMP accumulation and also diminished the magnitude of phosphorylation of the D1 receptor (Figs. 2 and 3). Antisense GRK4 oligonucleotides did not affect basal or forskolin-stimulated cAMP production (data not shown). Compared with fenoldopam alone, neither sense nor scrambled GRK4 oligonucleotides affected cAMP accumulation or receptor serine phosphorylation in either group. The almost complete suppression of the phosphorylation of the D1 receptor and normalization of the cAMP response by GRK4 antisense oligonucleotides in renal proximal tubules from hypertensive subjects suggest that the major GRK involved in the phosphorylation and desensitization of the D1 receptor in hypertension is GRK4 rather than other GRKs that may be expressed in this nephron segment. Several single nucleotide polymorphisms (SNPs) have been noted in the coding region of the GRK4 gene, but their clinical relevance has not been investigated (24). However, the GRK4 locus is linked with essential hypertension (29). Sequencing of GRK4 cDNAs from normotensive and hypertensive subjects confirmed the presence of SNPs (24): nucleotide 448, CGT to CTT (amino acid R65L) and the two reported variants, nucleotide 679, GCC to GTC (amino acid A142V), and nucleotide 1711, GCG to GTG (amino acid A486V). The increased GRK activity in cells from hypertensive subjects was attenuated by antibodies to GRK4␥兾␦; therefore, we determined whether the variations in the GRK4␥ or GRK4␦ genes have any functional consequences. Because we did not find any PNAS 兩 March 19, 2002 兩 vol. 99 兩 no. 6 兩 3875
MEDICAL SCIENCES
Statistical Analysis. The data are expressed as means ⫾ SE. Significant differences among and within groups were determined by ANOVA for n ⬎ 2 and t test for n ⫽ 2. N refers to the number of treatment groups or experiments, cell lines, or mice as indicated.
Table 1. Baseline characteristics of hGRK4␥ transgenes in anesthetized mice Variables
Wild type hGRK4␥ (n ⫽ 22)
A142V hGRK4␥ (n ⫽ 7)
26 ⫾ 1 1.54 ⫾ 0.04 0.46 ⫾ 0.01 396 ⫾ 15 1.38 ⫾ 0.42 222 ⫾ 90 97 ⫾ 2 68 ⫾ 2 77 ⫾ 2
28 ⫾ 2 1.51 ⫾ 0.07 0.52 ⫾ 0.02* 431 ⫾ 16 1.63 ⫾ 0.50 209 ⫾ 103 129 ⫾ 2* 97 ⫾ 4* 108 ⫾ 3*
Body weight, g Kidney weight, % body weight Heart weight, % body weight Heart rate, beats兾min Urine flow, l兾min Sodium excretion, nEq兾min Systolic blood pressure, mm Hg Diastolic blood pressure, mm Hg Mean blood pressure, mm Hg
Data are mean ⫾ SE. *, P ⬍ 0.05 vs. wild-type transgene. The mice were 3– 4 mo of age.
effect of GRK4␦ on D1 receptor function (unpublished observations), we concentrated our studies on GRK4␥. The ability of fenoldopam to stimulate cAMP accumulation in CHO cells (in the absence of GRK4␥ expression) was similar to that noted in HEK-293 cells, a cell with low endogenous GRK activity (Fig. 4A; ref. 35). The expression of wild-type GRK4␥ slightly decreased the ability of the D1 agonist to stimulate cAMP production (Fig. 4A). However, the D1 receptor-mediated cAMP production was markedly impaired by GRK4␥ SNPs (R65L, A142V, A486V, and combined R65L plus A486V). The effect of wild-type or GRK4␥ SNPs was not due to differences in the quantity of the expression of either the D1 receptor or GRK4␥ (not shown). Wild-type GRK4␥ or its variants did not affect the ability of forskolin to stimulate cAMP accumulation (not shown), indicating specificity of the interaction of GRK4␥ with the D1 receptor. The action of fenoldopam was selective for the D1 receptor because the fenoldopam effect was blocked by the D1-like antagonist SCH23390 (not shown). Expression of GRK4␥ SNPs was also associated with increased basal phosphorylation of the D1 receptor (Fig. 4B). These studies suggest that the increased basal phosphorylation of the D1 receptor by GRK4␥, may explain, in part, the decreased responsiveness of the D1 receptor in hypertension. Of the GRK4 isoforms (␣, , ␥, and ␦), only GRK4␣ and GRK4␦ have been reported to be involved in the regulation of G protein-coupled receptors (e.g., -adrenergic, luteinizing hormone, metabotropic glutamate, and M2 muscarinic receptors; refs. 24 and 35–40). GRK4␥ may preferentially regulate D1 receptors because it has not been found to have much activity in regulating other G proteincoupled receptors (24). Of the three GRK4␥ SNPs, GRK4␥ A142V had the most drastic effect on D1 receptor function in the expression studies. Therefore, we compared the consequences of the expression of GRK4␥ wild type and GRK4␥ A142V in transgenic mice. The important role that GRK4␥ SNPs play in the regulation of blood pressure was buttressed by the demonstration that transgenic mice expressing GRK4␥ A142V were hypertensive whereas those expressing the wild-type GRK4␥ were normotensive (Table 1). Furthermore, the heart weights were greater in hypertensive GRK4␥ A142V than in normotensive wild-type GRK4␥ mice. The hypotensive effects of bolus i.v. administration of fenoldopam were comparable in wild-type and GRK4␥ A142V transgenes (Fig. 5A). However, the i.v. infusion of fenoldopam increased urine flow and sodium excretion in GRK4␥ wild-type mice but not in the hypertensive GRK4␥ A142V transgenes (Fig. 5B). These results are in agreement with previous studies showing that the desensitization of the D1 receptor in genetic hypertension is renal specific (3, 7, 8, 16). The renal vasodilatory effect of D1-like agonists may be preserved in genetic hypertension (3). Thus, the ability of fenoldopam to decrease renal vascular resistance is maintained in the spontaneously hypertensive rat (SHR) (41). Fenoldopam increases effective renal plasma flow 3876 兩 www.pnas.org兾cgi兾doi兾10.1073兾pnas.062694599
to a greater extent in salt-sensitive hypertensive than in normotensive subjects (15). The ability of dopamine to relax renal artery strips is also increased in stroke-prone SHRs (42). However, fenoldopam has been reported to fail to vasodilate the kidney of some patients with essential hypertension (43). The ability of fenoldopam to counteract the renal vasoconstrictor effect of exogenous angiotensin II and inhibit tubuloglomerular feedback is also impaired in the SHR (44, 45). It is possible that differences in published reports may be related to differences in classes of hypertension being studied (salt-sensitive vs. saltresistant) or different experimental conditions (basal tone vs.
Fig. 5. Effect of the D1-like agonist, fenoldopam, in anesthetized mice overexpressing wild-type (n ⫽ 10) or A142V (n ⫽ 4) GRK4␥. (A) Effect of bolus i.v. injections of fenoldopam (1–1000 g) on arterial blood pressure was observed for 10 min. (B) Effect of a 30-min i.v. infusion of fenoldopam (2 g兾kg兾min) on sodium excretion (UNaV) and urine flow (V). Percentage changes from a 30-min vehicle infusion period are depicted. *, P ⬍ 0.05 vs. basal, ANOVA on ranks, Dunnet’s test; ⫹, P ⬍ 0.05 vs. basal, t test; #, P ⬍ 0.05 GRK4␥ A142V vs. WT (wild-type). Blood pressures were not affected by fenoldopam infusion except for transient decreases in GRK4␥ A142V to the same level as the wild-type transgenes during recovery 1 (not shown). Differences in blood pressure and renal responses to D1-like agonist infusion cannot be explained by differences in GRK4␥ or D1 receptor expression (data not shown).
Felder et al.
reactivity to vasoconstrictors; ref. 3). Nevertheless, the preservation of the extrarenal vasodilatory, and the ‘‘distal’’ renal tubular responses to D1-like agonists explains the ability of fenoldopam, a D1-like agonist, to decrease blood pressure and produce a natriuresis in hypertensive subjects (15, 46). Genes regulating the renin angiotensin system have been implicated in the pathogenesis of essential hypertension (47, 48). In a cohort of Ghanian hypertensive subjects, we found that there was a significant interaction of GRK4 SNPs (termed FJ) with the SNPs of angiotensin converting enzyme, AT1 angiotensin receptor, and angiotensinogen (49). In a white population, hypertensives (n ⫽ 89) and normotensives (n ⫽ 90) differed significantly in disease risk as a function of R65L兾A486V haplotypes (unpublished data). Similar effects were also found in a Japanese population (hypertensives, n ⫽ 122, normotensives, n ⫽ 54), where normal subjects were deficient in 65L兾486V alleles, and interactions among A142V, A486V, and angiotensin converting enzyme alleles significantly associated with disease phenotypes (unpublished data). These GRK4 SNPs impaired the function of D1 receptors, whether endogenously (renal proximal tubule cells) or exogenously (CHO cells) expressed, and increased blood pressure and impaired the diuretic and natriuretic effects of D1-like agonist stimulation. The failure of the kidney to excrete excess sodium chloride is thought to be crucial in the
development of hypertension (1, 2). The desensitization of the D1 receptor in renal proximal tubules in hypertension may lead to a decreased ability of the kidney to eliminate a sodium chloride load (3, 6, 13, 15). Genes that regulate renal sodium transport are important in the regulation of blood pressure (50). In summary, we have found a ligand-independent increase in GRK activity and decrease in D1 receptor function in renal proximal tubule cells in human essential hypertension because of activating SNPs of GRK4␥. These studies show a candidate gene, whose locus is linked to (29), and whose SNPs are associated with, essential hypertension, which has been shown to have pathophysiologic consequences in renal proximal tubule and CHO cells, to impair renal sodium excretion, and to cause hypertension in transgenic mice. We describe here a transgenic mouse model in which the pathogenesis of hypertension appears to mimic that reported for human essential hypertension. This model may be useful in the study of new approaches to the treatment of hypertension.
1. Guyton, A. C. (1980) Circulatory Physiology III: Arterial Pressure and Hypertension (Saunders, Philadelphia). 2. Guidi, E., Menghetti, D., Milani, S. Montagnino, G., Plazzi, O. & Bianchi, G. (1996) J. Am. Soc. Nephrol. 7, 1131–1138. 3. Jose, P. A., Eisner, G. M. & Felder, R. A. (1998) Pharmacol. Ther. 80, 149–182. 4. Hegde, S. S., Jadhav, A. L. & Lokhandwala, M. F. (1989) Hypertension 13, 828–834. 5. Felder, R. A., Seikaly, M. G., Cody, P., Eisner, G. M. & Jose, P. A. (1990) Hypertension 15, 560–569. 6. Siragy, H. M., Felder, R. A., Howell, N. L., Chevalier, R. L, Peach, M. J. & Carey, R. M. (1989) Am. J. Physiol. 257, F469–F477. 7. Nishi, A., Eklo ¨f, A.-C., Bertorello, A. M. & Aperia, A. (1993) Hypertension 21, 767–771. 8. Hussain, T. & Lokhandwala, M. F. (1997) Am. J. Physiol. 272, F339–F346. 9. Debska-Slizien, A., Ho, P., Drangova, R. & Baines, A. D. (1994) J. Am. Soc. Nephrol. 5, 1125–1132. 10. Gesek, F. A. & Schoolwerth, A. C. (1991) Am. J. Physiol. 261, F526–F536. 11. Hansell, P. (1995) Exp. Nephrol. 3, 15–22. 12. Li, X. X., Xu, J., Zheng, S., Albrecht, F. E., Robillard, J. E., Eisner, G. M. & Jose, P. A. (2001) Am. J. Physiol. Regul. Integr. Comp. Physiol. 280, R1650– R1656. 13. Albrecht, F. E., Drago, J., Felder, R. A., Printz, M. P., Eisner, G. M., Robillard, J. E., Sibley, D. R., Westphal, H. J. & Jose, P. A. (1996) J. Clin. Invest. 97, 2283–2288. 14. Chen, C. & Lokhandwala, M. F. (1992) Clin. Exp. Hypertens. 14, 615–628. 15. O’Connell, D. P., Ragsdale, N. V., Boyd, D. G., Felder, R. A. & Carey, R. M. (1997) Hypertension 29, 115–122. 16. Sanada, H., Jose, P. A., Hazen-Martin, D., Yu, P.-Y., Xu, J., Bruns, J. E., Phipps, J., Carey, R. M. & Felder, R. M. (1999) Hypertension 33, 1036–1042. 17. Kinoshita, S., Sidhu, A. & Felder, R. A. (1989) J. Clin. Invest. 84, 1849–1856. 18. Ferguson, S.S. (2001) Pharmacol. Rev. 53, 1–24. 19. Carman, C. V. & Benovic, J. L. (1998) Curr. Opin. Neurobiol. 8, 335–344. 20. Racz, K., Kuchel, O., Buu, N. T. & Tenneson, S. (1986) Circ. Res. 57, 889–897. 21. Michel, M. C., Siepmann, F., Bu ¨scher, R. Philipp, T. & Brodde, O.-E. (1993) Hypertension 22, 169–177. 22. Onsgard-Meyer, M. J., Berndt, T. J., Khraibi, A. A. & Knox, F. G. (1994) Am. J. Physiol. 267, R78–R83. 23. Ladines, C. A., Zeng, C., Asico, L. C., Sun, X., Pocchiari, F., Semeraro, C., Pisegna, J., Wank, S., Yamaguchi, I., Eisner, G. M., et al. (2001) Am. J. Physiol. Regul. Integr. Comp. Physiol. 281, R1071–R1078. 24. Premont, R. T., Macrae, A. D., Stoffel, R. H., Chung, N., Pitcher, J. A., Ambrose, C., Inglese, J., MacDonald, M. E. & Lefkowitz, R. J. (1996) J. Biol. Chem. 271, 6403–6410. 25. Sallese, M., Mariggio, S., Collodel, G., Moretti, E., Piomboni, P., Baccetti, B. & De Blasi, A. (1997) J. Biol. Chem. 272, 10188–10195. 26. Gros, R., Benovic, J. L., Tan, C. M. & Feldman, R. D. (1997) J. Clin. Invest. 99, 2087–2093. 27. Gros, R., Chorazyczewski. J., Meek, M. D., Benovic, J. L., Ferguson, S. S. & Feldman, R. D. (2000) Hypertension 35, 38–42.
28. Ishizaka, N., Alexander, R. W., Laursen, J. B., Kai, H., Fukui, T., Oppermann, M., Lefkowitz, R. J., Lyons, P. R. & Griendling, K. K. (1997) J. Biol. Chem. 272, 32482–32488. 29. Casari, G., Barlassina, C., Cusi, D., Zagato, L., Muirhead, R., Righetti, M., Nembri, P., Amar, K., Gatti, M., Macciardi, F., et al. (1995) Hypertension 25, 320–326. 30. Yamaguchi, I., Walk, S. F., Jose, P. A. & Felder, R. A. (1996) Mol. Pharmacol. 49, 373–378. 31. Benovic, J. L. (1991) Methods Enzymol. 200, 351–362. 32. Benovic, J. L. & Gomez, J. (1993) J. Biol. Chem. 268, 19521–19527. 33. Asico, L. D., Ladines, C., Fuchs, S., Accili, D., Carey, R. M., Semeraro, C., Pocchiari, F., Felder, R. A., Eisner, G. M. & Jose, P. A. (1998) J. Clin. Invest. 102, 493–498. 34. Brenninkmeijer, C. B., Price, S. A., Lopez Bernal, A. & Phaneuf. S. (1999) J. Endocrinol. 162, 401–408. 35. Tiberi, M., Nash, S. R., Bertrand, L., Lefkowitz, R. J.,& Caron, M. G. (1996) J. Biol. Chem. 271, 3771–3778. 36. Menard, L., Ferguson, S. S., Barak, L. S., Bertrand, L., Premont, R. T., Colapietro, A. M., Lefkowitz, R. J. & Caron, M. G. (1996) Biochemistry 35, 4155–4160. 37. Munshi, U. M., Peegel, H. & Menon, K. M. (2001) Eur. J. Biochem. 268, 1631–1639. 38. Dale, L. B., Bhattacharya, M., Anborgh, P. H., Murdoch, B., Bhatia, M., Nakanishi, S. & Ferguson, S. S. (2000) J. Biol. Chem. 275, 38213–38220. 39. Sallese, M., Salvatore, L., D’Urbano, E., Sala, G., Storto, M., Launey, T., Nicoletti, F., Knopfel, T. & De Blasi, A. (2000) FASEB J. 14, 2569–2580. 40. Tsuga, H., Okuno, E., Kameyama, K. & Haga, T. (1998) J. Pharmacol. Exp. Ther. 284, 1218–1226. 41. Lappe, R. W., Todt, J. A. & Wendt, R. L. (1986) J. Pharmacol. Exp. Ther. 236, 187–191. 42. Gao, Y. J., Nishimura, Y. & Suzuki, A. (1995) Clin. Exp. Pharmacol. Physiol. 22, Suppl. 1, S96–S98. 43. Bughi, S., Horton, R., Antonipillai, I., Manoogian, C., Ehrlich, L. & Nadler, J. (1989) J. Clin. Endocrinol. Metab. 69, 1116–1121. 44. Chatziantoniou, C., Ruan, X. & Arendshorst, W. J. (1995) Proc. Natl. Acad. Sci. USA 92, 2924–2928. 45. Ha¨berle, D. A., Konigbauer, B., Kawabata, M. & Ushiogi, Y. (1991) Klin. Wochenschr. 69, 587–596. 46. Post, J. B., 4th, & Frishman, W. H. (1998) J. Clin. Pharmacol. 38, 2–13. 47. Jeunemaitre, X., Soubrier, F., Kotelevtsev, Y. V., Lifton, R. P., Williams, C. S., Charru, A., Hunt, S. C., Hopkins, P. N., Williams, R. R., Lalouel, J. M., et al. (1992) Cell 71, 169–180. 48. Sigmund, C.D. (2001) Am. J. Hypertens. 14, 33S–37S. 49. Williams, S. M., Addy, J. H., Phillips, J. A., 3rd, Dai, M., Kpodonu, J., Afful, J., Jackson, H., Joseph, K., Eason, F., Murray, M. M., et al. (2000) Hypertension 36, 2–6 50. Lifton, R. P., Gharavi, A. G. & Geller, D. S. (2001) Cell 104, 545–556.
Felder et al.
PNAS 兩 March 19, 2002 兩 vol. 99 兩 no. 6 兩 3877
MEDICAL SCIENCES
We wish to acknowledge Dr. Jeffrey L. Benovic for initially providing the rhodopsin used in the GRK assay as well as the monoclonal GRK2 antibody. These studies were supported in part by grants from the National Institutes of Health, the National Institute of Diabetes and Digestive and Kidney Diseases, the National Heart, Lung, and Blood Institute, and The National Center for Research Resources.
