J Am Soc Nephrol 13: 2888–2897, 2002

A Role for Uric Acid in the Progression of Renal Disease

DUK-HEE KANG,*†‡ TAKAHIKO NAKAGAWA,* LILI FENG,*
SUSUMU WATANABE,* LIN HAN,* MARILDA MAZZALI,* LUAN TRUONG,§
RAYMOND HARRIS,¶ and RICHARD J JOHNSON*†
*Division of Nephrology, Baylor College of Medicine, Houston, Texas; †Division of Nephrology, University of
Washington, Seattle, Washington; ‡Division of Nephrology, Ewha Women’s University College of Medicine,
Ewha Medical Research Center, Seoul, Korea; §Division of Pathology, Baylor College of Medicine, Houston,
Texas; and ¶Division of Nephrology, Vanderbilt School of Medicine, Nashville, Tennessee.

Abstract. Hyperuricemia is associated with renal disease, but it fibrosis (1.89 " 0.45 versus 1.52 " 0.47; P # 0.05). Hyper-
is usually considered a marker of renal dysfunction rather than uricemic rats developed vascular disease consisting of thick-
a risk factor for progression. Recent studies have reported that ening of the preglomerular arteries with smooth muscle cell
mild hyperuricemia in normal rats induced by the uricase proliferation; these changes were significantly more severe
inhibitor, oxonic acid (OA), results in hypertension, intrarenal than a historical RK group with similar BP. Allopurinol sig-
vascular disease, and renal injury. This led to the hypothesis nificantly reduced uric acid levels and blocked the renal func-
that uric acid may contribute to progressive renal disease. To tional and histologic changes. Benziodarone reduced uric acid
examine the effect of hyperuricemia on renal disease progres- levels less effectively and only partially improved BP and renal
sion, rats were fed 2% OA for 6 wk after 5/6 remnant kidney function, with minimal effect on the vascular changes. To
(RK) surgery with or without the xanthine oxidase inhibitor, better understand the mechanism for the vascular disease, the
allopurinol, or the uricosuric agent, benziodarone. Renal func- expression of COX-2 and renin were examined. Hyperuricemic
tion and histologic studies were performed at 6 wk. Given rats showed increased renal renin and COX-2 expression, the
observations that uric acid induces vascular disease, the effect latter especially in preglomerular arterial vessels. In in vitro
of uric acid on vascular smooth muscle cells in culture was also studies, cultured vascular smooth muscle cells incubated with
examined. RK rats developed transient hyperuricemia (2.7 uric acid also generated COX-2 with time-dependent prolifer-
mg/dl at week 2), but then levels returned to baseline by week ation, which was prevented by either a COX-2 or TXA-2
6 (1.4 mg/dl). In contrast, RK!OA rats developed higher and receptor inhihbitor. Hyperuricemia accelerates renal progres-
more persistent hyperuricemia (6 wk, 3.2 mg/dl). Hyperurice- sion in the RK model via a mechanism linked to high systemic
mic rats demonstrated higher BP, greater proteinuria, and BP and COX-2–mediated, thromboxane-induced vascular dis-
higher serum creatinine than RK rats. Hyperuricemic RK rats ease. These studies provide direct evidence that uric acid may
had more renal hypertrophy and greater glomerulosclerosis be a true mediator of renal disease and progression.
(24.2 " 2.5 versus 17.5 " 3.4%; P # 0.05) and interstitial

Hyperuricemia has long been associated with renal disease. tologic findings have been observed in autopsies of 79 to 99%
Approximately 20 to 60% of patients with gout have mild or of patients with gout (3).
moderate renal dysfunction (1); before the availability of uric Despite the association of gout with renal disease, contro-
acid lowering agents, as many as 10 to 25% of patients with versy exists as to whether uric acid has an etiologic role (4 – 6).
gout developed end-stage renal disease (2). The histologic First, it has been difficult to ascribe the generalized renal injury
lesion termed “gouty nephropathy” consists of glomeruloscle- in gout to the deposition of urate crystals, for they are often
rosis, interstitial fibrosis, and renal arteriolosclerosis, often only focally present. Second, many patients with gout have
with focal interstitial urate crystal deposition (2,3). These his- hypertension or are elderly, and the renal lesions might simply
reflect hypertensive or aging-associated renal damage (1).
Third, results of the studies are mixed as to whether lowering
uric acid will slow renal progression in patients with gout (7,8).
Received May 31, 2002. Accepted August 1, 2002. The inability to resolve this issue has emphasized the need for
Dr. Jared Grantham served as guest editor and supervised the review and final additional studies (6).
disposition of this manuscript.
To investigate the role of uric acid in renal disease, we
Correspondence to Dr. Duk-Hee Kang, Division of Nephrology, Ewha Wom-
en’s University Hospital, 70 Chongno 6-ka Chongno-ku Seoul 110-126, Korea. recently developed a model of hyperuricemia in rats (9). Most
Phone: 82-2-760-5121; Fax: 82-2-760-5008; E-mail: dhkang@ewha.ac.kr mammals have a low serum uric acid due to the presence of
1046-6673/1312-2888 uricase; in humans and the Great Apes, the uricase gene was
Journal of the American Society of Nephrology mutated and rendered nonfunctional. We therefore induced
Copyright © 2002 by the American Society of Nephrology hyperuricemia in rats by providing low doses of oxonic acid,
DOI: 10.1097/01.ASN.0000034910.58454.FD which is a uricase inhibitor. Unlike previous models of uricase
J Am Soc Nephrol 13: 2888–2897, 2002 Uric Acid and Renal Disease Progression 2889

inhibition, which result in massive uricosuria with intrarenal mg/dl in the drinking water; Sanofi, Barcelona, Spain)
crystal deposition and obstructive renal disease, this model (RK!OA!BZ). 2% OA could be given without apparent toxicity to
resulted in mild hyperuricemia without intrarenal crystal dep- rats according to our previous studies (9 –11), and there was no
osition. Nevertheless, subtle interstitial renal injury developed, mortality during the study period. The doses of AP and BZ were
determined on the basis of the uric acid-lowering effect in previous
and this was associated with activation of the renin angiotensin
studies (9,10). The mean daily intakes of allopurinol and benzioda-
system (RAS) and the development of hypertension (9). The rone were 20.4 " 5.3 mg/kg per d and 21.3 " 6.2 mg/kg per d in
hyperuricemic animals also developed an afferent arteriolopa- RK!OA!AP and RK!OA!BZ groups, respectively. At 6 wk, rats
thy that occurred independently of changes in BP (10). The were sacrificed for histologic evaluation.
vascular injury was mediated in part by direct effects of uric
acid to induce vascular smooth muscle cell (VSMC) prolifer- Systolic BP, Uric Acid, Proteinuria, and Renal
ation and also by activation of the RAS (10). Whereas hyper- Function
uricemia induced both renal injury and vascular disease in Systolic arterial BP was monitored by a tail cuff sphygmomanom-
normal rats, hyperuricemia was also shown to accelerate cy- eter using an automated system with a photoelectric sensor (IITC; Life
closporine-induced vascular injury and interstitial renal disease Sciences, Woodland Hills, CA) that has been shown to closely cor-
(11). The exacerbation of cyclosporine nephrotoxicity by uric relate with intraarterial measurement of BP (15). All rats were pre-
acid was also associated with increased activation of the RAS conditioned for this machine at least two times before the actual
and blockade of specific nitric oxide pathways (11). measurement of BP. Serum uric acid concentration was determined by
The observation that mild hyperuricemia activates the RAS a carbonate phosphotungstate method and uric acid standard (Sigma,
and causes renal disease via a crystal independent pathway St. Louis, MO) (16). Urinary protein excretion (24 h) were measured
raised the hypothesis that hyperuricemia may be a general risk using the sulfosalicylic acid method, and blood urea nitrogen and
creatinine were determined colorimetrically utilizing a commercial kit
factor for renal progression. Interestingly, hyperuricemia is a
(Sigma Diagnostics, St. Louis, MO).
common feature of renal disease of all etiologies; as GFR falls
the serum uric acid increases due to reduced renal excretion.
The hyperuricemia in renal failure patients is generally mild,
Renal Morphology and Immunohistochemistry
Tissue for light microscopy and immunoperoxidase staining was
due to a compensatory increase in fractional excretion, reduced
fixed in Methyl Carnoy’s solution and embedded in paraffin. Four-
production, and increased excretion of uric acid via nonrenal micrometer sections were stained with the periodic acid-Schiff (PAS)
(gastrointestinal) routes (12). Because the hyperuricemia is reagent and counterstained with hematoxylin. Indirect immunoperox-
often mild, gout is rare in patients with end-stage renal disease idase staining of 4-!m sections was performed as described previ-
and the increase in uric acid is often considered innocuous. ously (17), with specific monoclonal and polyclonal antibodies di-
However, the possibility that the hyperuricemia could be con- rected to the following antigens: "-smooth muscle actin ("-SMA)
tributing to renal progression has not been tested. To test this with mouse monoclonal ("-SM-1; Sigma, St. Louis, MO), renin with
hypothesis, we examined the role of uric acid in the classic monocloncal mouse anti-human antibody (Sanofi Recherche, Motpel-
model of renal progression induced by 5/6 nephrectomy (rem- lier, France); collagen type III with goat anti-human antibody (South-
nant kidney model). Additionally, given the fact that renal ern Biotechnology, Birmingham, AL); COX-2 with rabbit anti-mouse
cortical COX-2 expression is increased after subtotal renal polyclonal antibody (Cayman, Ann Arbor, MI); monocyte-macroph-
ages with mouse monoclonal antibody ED-1 (Serotec, Indianapolis,
ablation (13,14) and COX-2 is a major regulator of renin
IN). Controls included omitting the primary antibody and substitution
production, we also investigated changes in COX-2 by of the primary antibody with preimmune rabbit or mouse serum.
hyperuricemia. To examine whether there is any evidence of endothelial or smooth
muscle proliferation, double immunostaining was performed with
Materials and Methods "-smooth muscle cell actin and an antibody to the proliferating cell
Experimental Protocol nuclear antigen (PCNA) (PC 10; Cappel, Aurora, OH). Tissue sec-
All animal procedures were approved by the Animal Care Com- tions were first incubated with PCNA antibody overnight at 4°C
mittees of the University of Washington and Baylor College of followed sequentially by biotinylated horse anti-mouse IgG serum,
Medicine. Male Sprague-Dawley rats (190 to 200 g; Simonsen Lab- peroxidase-conjugated avidin D, and color development with diami-
oratories, Gilroy, CA) underwent a remnant kidney (RK) operation nobenzidine (DAB) with nickel chloride. After incubation in 3%
after baseline measurement of BP and renal function. The RK oper- H2O2 for 8 min to eliminate any remaining peroxidase activity,
ation was performed by a right subcapsular nephrectomy and surgical sections were incubated with primary antibody for "-smooth muscle
resection of the upper and lower thirds of the left kidney in total of 17 cell actin for 3 h at room temperature followed by biotinylated horse
rats. After randomization by the percent remnant kidney weight re- anti-mouse IgG for 30 min at room temperature. After incubation to
moved [(right kidney weight $ weight of upper and lower poles of alkaline phosphatase Streptavidin (Vector, Burlingame, CA), color
left kidney)/right kidney weight % 100], the animals were divided into was developed using AP-RED substrate kit (Zymed, San Francisco,
four groups. Group 1 (n & 4) received normal-salt (NS) diet (0.27% CA).
NaCl) (RK); group 2 (n & 5) received NS diet containing 2% oxonic
acid (OA, hepatic uricase inhibitor; Sigma, St. Louis, MO) Quantification of Morphologic Data
(RK!OA); group 3 (n & 4) received 2% OA NS diet and allopurinol All analyses were performed blinded. The numbers of endothelial
(xanthine oxidase inhibitor, 13 mg/dl in the drinking water; Schein and smooth muscle cells were counted in the afferent arteriole, inter-
Pharmaceutical, Florham Park, NJ) (RK!OA!AP); and group 4 (n & lobular artery, and arcuate artery. All cross-sectioned arterioles and
4) received 2% OA NS diet and benziodarone (uricosuric agent, 13 arteries available in each section were evaluated. Vessels that were not
2890 Journal of the American Society of Nephrology J Am Soc Nephrol 13: 2888–2897, 2002

sectioned transversally, providing an asymmetrical wall, were ex- measuring 3H-thymidine uptake (22). The effect of a selective COX-2
cluded from the present analysis. To assess smooth muscle cell inhibitor (NS398 [10 !M/L]; Cayman Chemical, Ann Arbor, MI) and
hypertrophy and/or hyperplasia, the cross-sectional wall area and thromboxane A2 (TXA-2) receptor antagonist (SQ-29548 [2.5 !M/
thickness of each artery was measured by computer image analysis ml]; Biomol, Plymouth Meeting, PA) on uric acid-induced SMC
(Optimas 6.2; Media Cybernetics, Silver Springs, MD). Arteries and proliferation was also evaluated. At the concentrations of all reagents
arterioles were identified by their anatomic location and branching used there was no evidence for cytotoxicity as documented by trypan
pattern from neighboring vessels and shape of endothelial and smooth blue staining and LDH release.
muscle cells described elsewhere (18). Considering the continuous
tapering course of interlobular artery, we evaluated the interlobular Statistical Analysis
artery at the same level (proximal 1/3 from corticomedullary junction) All data are presented as mean " SD. Differences in the various
of individual kidneys. Afferent arterioles were carefully distinguished parameters between groups were evaluated by two-way ANOVA
from efferent arteriole by their general characteristics, including the followed by correction for multiple comparison. Differences in pa-
presence of thin and continuous endothelial cells with a thicker rameters at each time point after RK surgery and drug administration
smooth muscle wall than the efferent arteriole. were compared by paired t test. The relation between variables was
Renin expression was quantified by the number of glomeruli with assessed by Pearson correlation analysis. Significance was defined as
positive staining for juxtaglomerular renin using a minimum 40 glo- P # 0.05.
meruli in each biopsy, a method which has previously been shown to
correlate with changes in tissue renin content (19). The expression of
cortical COX-2 was measured as the percent positive area of COX-2
Results
immunostaining in renal cortex using computer image analyzer. The
Hyperuricemia Induces Hypertension and Renal
percent glomerulosclerosis and tubulointerstitial fibrosis score (0 to 5) Hypertrophy in the Remnant Kidney Model
were evaluated on the basis of PAS staining as described previously Baseline uric acid levels averaged 1.1 mg/dl in all four
(17). groups (Figure 1). In RK rats, a transient increase in uric acid
was observed (2 wk, 2.7 " 0.6 mg/dl; P # 0.05), but levels
Generation of Rat COX-1 and COX-2 Riboprobe for
RNase Protection Assay (RPA)
A fragment of COX-1 cDNA produced by BstX1 (from bp 1350 to
bp 1690), 340 bp, was blunted and subcloned into the SmadI site of
pGEM4Z. A 241-bp fragment of rat COX-2 cDNA produced by
BamHI and EcoRI (from bp 409 to bp 650) was subcloned in
pGEM4Z (20).

RPA for COX-1 and COX-2 mRNA

Following incubation of VSMC with uric acid for 4, 8, and 24 h,
total RNA was prepared from the VSMC monolayers using the
RNeasy96 total RNA isolation protocol manufactured by Qiagen
(Valencia, CA). After determination of RNA purity and concentration
by spectrophotometry, 2 !g of RNA samples were hybridized for 30
min at 90°C with a mixture of 32P-UTP–labeled riboprobe and the
housekeeping gene (L32) (1 % 105 cpm for each probe), and RPA was
performed as described previously using a RPA kit (21) (Torrey Pines
Biolabs, Houston, TX) according to the manufacturer’s instruction.
The protected hybridized RNA was denatured at 85°C and electro-
phoresed on 10% polyacrylamide gels. The gels were transferred to
3M Whatman filter paper, dried, and exposed to Kodak X-O mat film
overnight at $70°C.

In Vitro Effect of Uric Acid and Selective COX-2

Inhibitor on Smooth Muscle Cell Proliferation
Rat vascular smooth muscle cells (RVSMC) were purchased from
American Type Culture Collection (CRL-2018) and cultured in
DMEM supplemented with 10% FBS, gentamicin (G418, 50 mg/ml),
and L-glutamine. After cells were grown to 70% confluence in mul-
tiwell plates (Beckton Dickinson, Franklin Lakes, NJ), the culture
medium was changed into serum-free media for 24 to 48 h before each
experiment. After synchronization of cell growth, cells were washed
three times with HBSS and exposed to uric acid (3 mg/dl) (Sigma, St. Figure 1. Changes in serum uric acid (A) and BP (B) in the remnant
Louis, MO) for 6, 24, 48, and 72 h. At each time point, the number of kidney (RK; !!!"!!!), remnant kidney ! oxonic acid (RK!OA; ©),
cells was counted by hemocytometer and Coulter Counter. Alterna- remnant kidney ! oxonic acid ! allopurinol (RK!OA!AP; –%–),
tively, cell number was determined using CellTiter Proliferation As- and remnant kidney ! oxonic acid ! benziodarone (RK!OA!BZ;
say (Promega, Madison, WI). Cell proliferation was also assessed by !!–"–!!) rats. *P # 0.05 versus RK and RK!OA!AP.
J Am Soc Nephrol 13: 2888–2897, 2002 Uric Acid and Renal Disease Progression 2891

returned to baseline by week 6 (1.4 " 0.5 mg/dl). In contrast,

RK!OA rats developed more severe and persistent hyperuri-
cemia (2 wk, 4.0 " 0.6 mg/dl; 6 wk, 3.2 " 0.6 mg/dl; P # 0.01
versus RK alone at 2 and 6 wk).
This rise in serum uric acid in RK!OA rats was followed by
an elevation of BP at 4 wk compared with RK rats alone
(Figure 1B). Allopurinol, a xanthine oxidase inhibitor, signif-
icantly decreased the level of serum uric acid and prevented the
development of hypertension. Benziodarone (a uricosuric
agent) only partially reduced uric acid and had a lesser effect
on BP (Figure 1).
During the study period, the mean daily intake of food was
similar among groups (RK versus RK!OA versus
RK!OA!AP versus RK!OA!BZ, 24 " 0.4 versus 21 " 0.5 Figure 2. Changes in proteinuria in the RK (!!!"!!!), RK!OA (©),
versus 22 " 0.6 versus 22 " 0.5 g/kg). Despite comparable RK!OA!AP (–%–), and RK!OA!BZ (!!–"–!!) rats. *P # 0.05
body weight gain, remnant kidney weight (RKW) and RKW/ versus RK and RK!OA!AP; #P # 0.05 versus RK, RK!OA!AP
body weight (BW) at 6 wk were significantly higher in and RK!OA!BZ.
RK!OA rats, suggesting more severe renal hypertrophy (Ta-
ble 1). The percent increase in RKW/BW in RK!OA rats was
1.6-fold compared with RK rats (RK!OA versus RK, 380.4 " prevented the increase in proteinuria induced by OA (Figure
37.3 versus 241.9 " 20.5%; P # 0.05). The renal hypertrophy 2). Allopurinol was especially effective and reduced protein-
induced by oxonic acid was significantly prevented by the uria to the same level as that observed in the RK model alone.
administration of allopurinol or benziodarone (Table 1). RK rats displayed classic evidence of progressive renal
disease with focal glomerulosclerosis and interstitial fibrosis.
Hyperuricemia Accelerates Renal Progression in the RK!OA rats showed similar histologic features; in particular,
Remnant Kidney Model there was no evidence for tubular obstruction from crystals.
Baseline levels of serum creatinine were comparable in four Glomerulosclerosis was more common in RK!OA rats com-
groups (RK versus RK!OA versus RK!OA!AP versus pared with RK rats (24.2 " 2.5 versus 17.5 " 3.4%; P # 0.05)
RK!OA!BZ, 0.72 " 0.01 versus 0.69 " 0.04 versus 0.75 " (Figure 3). Interstitial fibrosis was also greater in RK!OA rats
0.05 versus 0.70 " 0.03 mg/dl). At 6 wk of OA administration, compared with RK alone (1.89 " 0.45 versus 1.52 " 0.47; P
RK!OA rats displayed worse renal function with high serum # 0.05) (Figure 3). Allopurinol prevented both the increase in
creatinine compared with the RK alone group (1.44 " 0.08 glomerulosclerosis and interstitial fibrosis. In contrast, the uri-
versus 1.23 " 0.13 mg/dl, RK!OA versus RK; P # 0.05). cosuric agent, benziodarone, was able to prevent the increase in
Administration of allopurinol or benziodarone prevented the glomerulosclerosis but not the increase in interstitial fibrosis in
increase in serum creatinine (RK!OA!AP, 1.12 " 0.15 mg/ RK!OA rats (Figure 3).
dl; RK!OA!BZ, 1.23 " 0.15 mg/dl; P # 0.05 versus
RK!OA, respectively). Hyperuricemia Induces Severe Preglomerular Vascular
After renal ablation, RK rats showed a progressive increase Disease in RK Rats
in urinary protein excretion. The administration of OA in RK The preglomerular arterial vessels (arcuate artery, interlob-
rats resulted in greater proteinuria at 4 and 6 wk (Figure 2) ular artery, and afferent arteriole) of RK!OA rats showed
compared with the RK rats. Both allopurinol and benziodarone dramatic differences compared with control RK rats (Figure 4,

Table 1. Changes in body weight (BW) and remnant kidney weight (RKW) at 6 wka
RK RK!OA RK!OA!AP RK!OA!BZ
(n & 4) (n & 5) (n & 4) (n & 4)

Baseline BW (g) 202.3 " 13.0 206.1 " 1.3 192 " 13.1 195.6 " 6.0
Final BWb (g) 358.4 " 35.0 336.6 " 23.2 338.1 " 18.4 345.9 " 23.2
Final RKWb (g) 1.31 " 0.05 1.47 " 0.11e 1.29 " 0.28 1.30 " 0.28
Final RKW/BWb (g/kg) 3.7 " 0.2 4.2 " 0.4e 3.8 " 0.6 3.7 " 0.6
Increase in RKW (%)c 241.9 " 20.5 380.4 " 37.3e 256.8 " 35.7 236.2 " 28.4
a
OA, oxonic acid; AP, allopurinol; BZ, benziodarone. Data are expressed as mean " SD.
b
BW and RKW after 6 wk of RK operation.
c
(RKW/BW at 6 wk)/(Baseline RKWc/BW) % 100.
d
Baseline RKW & Baseline right kidney weight $ (weight of upper and lower poles of left kidney).
e
P # 0.05 versus RK, RK!OA!AP, and RK!OA!BZ.
2892 Journal of the American Society of Nephrology J Am Soc Nephrol 13: 2888–2897, 2002

Figure 4. Morphology of preglomerular vessels in the RK and

RK!OA rats. Compared with RK rats (A, %400, PAS; D, %400,
PCNA!"-SMA double-staining), there was marked wall thickening
with smooth muscle cell (SMC) proliferation in RK!OA rats (B,
%400, PAS; E, %400, PCNA!"-SMA double-staining). Occasional
arteries in RK!OA rats showed an increase in SMC number, migra-
tion of SMC into intima consistent with an obliterative arteriopathy
(C, %400, PAS). Perivascular adventitial fibrosis was also prominent
in RK!OA rats (F, %400, Collagen III immunostaining).

Role of BP in the Vascular Changes Induced by

Hyperuricemia
Figure 3. Comparison of renal scarring. (A) glomerulosclerosis; (B) Because the RK!OA rats had higher systolic BP compared
tubulointerstitial fibrosis in the RK, RK!OA, RK!OA!AP, and with the control RK rats (Figure 1), we hypothesized that the
RK!OA!BZ rats. *P # 0.05 versus RK, RK!OA!AP, and vascular injury observed in the RK!OA rats might be attrib-
RK!OA!BZ; #P # 0.05 versus RK and RK!OA!AP.
uted to vascular injury induced by the higher pressure. To
differentiate the effect of hyperuricemia-induced elevation in
A through C). RK!OA rats showed a marked increase in BP from hyperuricemia itself on SMC proliferation, we eval-
vessel wall thickness as well as an increase in smooth muscle uated the changes in the preglomerular vessels in a historic RK
cell number (Figure 4B). In some RK!OA rats, the arterial control group with comparable BP to RK!OA rats (n & 6;
vessels showed changes consistent with an obliterative arteri- RK!OA versus historic RK with high BP, 177.8 " 20.1
opathy (Figure 4C). Double staining of PCNA and "-SMA versus 175.7 " 45.8 mmHg; P & NS) and body weights (336.6
documented the presence of numerous proliferating SMC in " 23.2 versus 362.5 " 55.4 g; P & NS) (17,23). As shown in
the RK!OA group compared with control RK rats (Figure 4, Figure 5, the number of SMC of preglomerular arteries in
D and E). RK!OA rats was higher compared with the RK rats with
Periarterial adventitial fibrosis was also prominent in comparable BP. This suggests that the increase in BP observed
RK!OA rats (Figure 4F) in association with more cell infil- in RK!OA rats does not account for the differences in vascu-
tration. Quantification of the proliferating smooth muscle cell lar injury observed.
number (PCNA/"-SMC!cells) and medial wall thickness of
the various preglomerular vessels is shown in Table 2. Another Hyperuricemia Activates the RAS and COX-2 Systems
characteristic finding in preglomerular vessels was the infiltra- in the RK Model
tion of inflammatory macrophages assessed by ED-1! cells, We have previously reported that hyperuricemia increases
especially in subendothelial and adventitial areas. There were juxtaglomerular renin expression in normal rats and in rats
also some ED-1! cells in media. treated with cyclosporine (9 –11). We have also shown that
Allopurinol significantly prevented the SMC proliferation preglomerular arteriopathy in hyperuricemic rats can be pre-
and wall thickening of afferent arteriole and interlobular artery vented by treatment with either ACE inhibitors or angiotensin
in RK-OA rats (Table 2). In contrast, benziodarone did not II type 1 receptor (AT1) blockers (10). We therefore investi-
show any effect on SMC proliferation. Interestingly, the gated whether renin was increased by hyperuricemia in the RK
number of endothelial cells in the interlobular and arcuate model. Consistent with the other studies, we found increased
arteries was also increased with OA administration (Table 2) renin expression in RK!OA rats versus RK alone (RK!OA
and was not affected by the administration of allopurinol or versus RK, 39.2 " 7.39 versus 31.2 " 4.48% of glomeruli with
benziodarone. juxtaglomerular renin staining; P # 0.05). Both allopurinol
J Am Soc Nephrol 13: 2888–2897, 2002 Uric Acid and Renal Disease Progression 2893

Table 2. Changes in macrovasculature in RK and OA-treated RK ratsa

RK RK!OA RK!OA!AP RK!OA!BZ

Afferent Arteriole
number of SMC (per cross-section) 4.0 " 0.4 6.7 " 1.6b 4.0 " 0.5 6.2 " 0.8
number of EC (per cross-section) 2.8 " 0.7 3.1 " 0.6 3.1 " 0.2 3.3 " 0.4
wall area (!m2) 231.3 " 21.8 269.5 " 13.8b 215.4 " 10.5 257.3 " 31.9
Interlobular Artery
number of SMC (per cross-section) 8.1 " 1.0 13.2 " 2.2b 9.8 " 1.1 12.6 " 4.2
number of EC (per cross-section) 6.7 " 0.6 7.9 " 0.7c 8.3 " 0.4 8.4 " 1.9
wall area (!m2) 659.3 " 203.2 1090.0 " 224.5b 670.4 " 253.0 983.7 " 310.8
Arcuate Artery
number of SMC (per cross-section) 17.8 " 1.7 30.8 " 7.7c 25.4 " 1.0 27.8 " 7.8
number of EC (per cross-section) 9.5 " 1.1 14.5 " 2.4c 13.9 " 1.2 15.5 " 2.6
wall area (!m2) 3058 " 847.9 5737.6 " 3206.9b 3110.2 " 1532.9 5935.1 " 2916.6
a
Data are expressed as mean " SD. SMC, smooth muscle cell, EC, endothelial cell.
b
P # 0.05 versus RK and RK!OA!AP.
c
P # 0.05 versus RK.

Figure 6. Correlation of serum uric acid with renin expression in renal

Figure 5. BP-independent effect of hyperuricemia on SMC prolifer- cortex.
ation. The number of SMC of interlobular arteries in RK!OA rats
was significantly higher compared with the RK rats with comparable
BP. *P # 0.05 versus normal, RK, and RK with high BP.
The increase in COX-2 expression was quantified by com-
puter image analysis of the cortex. RK!OA rats showed
(20.2 " 7.93%) and benziodarone (27.8 " 8.35%) signifi- significantly greater COX-2 expression than RK rats alone
cantly reduced renal renin expression induced by OA (P # (0.24 " 0.05 versus 0.70 " 0.36%; P # 0.05), and both
0.005). There was a strong correlation between renin content of allopurinol (0.27 " 0.18%) and benziodarone (0.48 " 0.26%)
individual rats with the serum uric acid (r2 & 0.52; P # 0.05) could prevent the increase in COX-2 expression. Within indi-
(Figure 6). vidual rats, the uric acid levels correlated directly with the
A major regulator of renin expression is cyclooxygenase 2 degree of COX-2 expression (Figure 8A). In turn, the COX-2
(COX-2) (13). We therefore examined the expression of expression also correlated with the increase in juxtaglomerular
COX-2 in this model. In normal rat cortex, COX-2 immuno- renin (Figure 8B). These studies demonstrate that both the
reactivity is present almost exclusively in macula densa cells COX-2 and RAS are induced with hyperuricemia in the RK
(14). Consistent with the studies by Wang et al. (14) in the RK model.
model, COX-2 expression was increased in RK rats compared
with normal rats (0.10 " 0.01 versus 0.24 " 0.05%; P # 0.05), Uric Acid Stimulates VSMC Proliferation via the COX-
especially in the tubular cells of the cortical thick ascending 2 Pathway
limb. RK!OA rats displayed even more COX-2 expression The observation that uric acid induced vascular disease
that involved not only the cortical thick ascending limb but also independently of BP suggested a potential direct effect of uric
smooth muscle cells in afferent arterioles and interlobular acid on smooth muscle cells. Consistent with in vivo finding,
arteries (Figure 7). stimulation with uric acid (3 mg/dl; Sigma, St. Louis, MO)
2894 Journal of the American Society of Nephrology J Am Soc Nephrol 13: 2888–2897, 2002

Figure 7. Expression of COX-2 in the RK and RK!OA rats. In

normal rats and RK rats, COX-2 immunoreactivity is present almost
exclusively in macula densa cells (A, RK rats, %200), whereas
RK!OA rats displayed more COX-2 expression that involved not
only the tubular cells of cortical thick ascending limb but also smooth
muscle cells in afferent arterioles (C, %400, arrow) and interlobular
arteries (D, %400).
Figure 8. Correlation of COX-2 expression with serum uric acid (A)
and renin expression (B).

resulted in time-dependent increase in rat aortic smooth muscle

cell (VSMC) number (Figure 9A). Discussion
We observed de novo expression of COX-2 in the vascular Hyperuricemia is common in patients with renal disease, but
smooth muscle wall of the RK!OA rats; we therefore hypoth- it has almost never been considered a risk factor for progres-
esized that COX-2 expression induced by uric acid may play an sion (25). However, two studies have found that hyperuricemia
important role in the development of arteriopathy observed in is an independent risk factor for progression in IgA nephrop-
the RK!OA rats. Consistent with this observation, we found athy (26,27). Furthermore, in a recent study of 6400 subjects
that uric acid could induce a marked increase in COX-2 mRNA with normal renal function, a serum uric acid of '8.0 mg/dl,
in cultured VSMC (Figure 9B), whereas comparable COX-1 when compared with a serum uric acid level of #5.0 mg/dl,
mRNA expression was noted. To determine the role of COX-2 was associated with a 2.9-fold increased risk for developing
renal insufficiency within 2 yr in men and a 10.0-fold increased
in the SMC proliferation, we examined the effect of a selective
risk in women (28). This increased relative risk was indepen-
COX-2 inhibitor (NS398 [10 !M/L]: Cayman Chemical, Ann
dent of age, body mass index, systolic BP, total cholesterol,
Arbor, MI) on SMC number. NS398 significantly inhibited
serum albumin, glucose, smoking, alcohol use, exercise habits,
uric acid–induced SMC proliferation (Figure 9C), suggesting
proteinuria, and hematuria (28). Indeed, an elevated uric acid
the possible role of prostaglandins generated by COX-2 in uric was more predictive for the development of renal insufficiency
acid–mediated SMC proliferation. Importantly, NS398 did not than proteinuria (28). We therefore investigated the effects and
affect cell number in the absence of uric acid stimulation. the action mechanism of hyperuricemia on the progression of
Previous studies have reported that angiotensin II-induced renal disease in the classic RK model in rats.
VSMC proliferation is mediated by COX-2– generated Serum uric acid increased in RK rats but was greater in rats
thromboxane (TX-A2) (24). TX-A2 was therefore hypothe- administered 2% OA (the uricase inhibitor). The serum uric
sized to be a likely mediator of the proliferative response of acid level peaked at 2 wk in both RK and RK!OA rats and
VSMC to uric acid. VSMC were incubated with uric acid (3 then gradually decreased over the ensuing 4 wk. The fall in uric
mg/dl) in the absence or presence of the selective TXA2 acid in the RK and RK!OA rats may reflect enhanced extra-
antagonist (SQ-29548 [2.5 !M/ml]; Biomol, Plymouth renal (mainly enteric) excretion and depressed production of
Meeting, PA). SQ-29548 significantly reduced uric acid- uric acid (reduced xanthine oxidase activity) in chronic renal
induced SMC proliferation. failure as reported by Vaziri et al. (12).
J Am Soc Nephrol 13: 2888–2897, 2002 Uric Acid and Renal Disease Progression 2895

There were many new findings in this study. The major

novel finding was that modest hyperuricemia markedly exac-
erbated renal progression. Specifically, hyperuricemic
RK!OA rats showed more renal hypertrophy, hypertension,
proteinuria, impaired renal function, and greater glomerulo-
sclerosis and interstitial fibrosis. The renal injury did not
appear to be mediated by urate crystal deposition, as there were
no features of tubular obstruction consistent with intrarenal
urate crystal deposition.
One of the mechanisms by which uric acid may have exac-
erbated the renal injury may be activation of the RAS, which
has been shown to be an important mediator of progression in
renal disease, not only by its hemodynamic effects to increase
systemic and glomerular pressure, but also by its direct fibro-
genic effect on renal and vascular cells. We have previously
reported that the administration of OA to normal rats increases
juxtaglomerular renin expression and that treatment with ena-
lapril can control BP, improve the arteriolopathy, and prevent
the renal injury in this model (9,10). We also found that OA
treatment in a rat model of cyclosporine nephropathy results in
hyperuricemia, increased renin expression, and accelerated
progression (11). Saito et al. (29) have also reported that serum
uric acid levels in humans also correlate with plasma renin
activity. Although we did not measure plasma renin activity,
we did document an increase in renal renin expression in
hyperuricemic rats that correlated with serum uric acid. Fur-
thermore, prevention of the hyperuricemia in OA-treated rats
with allopurinol, and to a lesser extent with benziodarone,
resulted in lower renin levels in oxonic acid-treated rats in
concert with less renal injury.
A second new finding in the study was the observation that
hyperuricemic RK rats developed severe preglomerular vascu-
lopathy. Although thickening of the afferent arterioles had
been noted in previous studies of hyperuricemic rats with
normal renal function (10), hyperuricemia was associated with
a severe vasculopathy in the RK model. The preglomerular
vessels showed thickening with an increase in SMC number
and macrophage infiltration in subendothelial, medial, and
adventitial areas. These vascular changes, which occasionally
resulted in an obliterative arteriopathy, could potentiate the
renal injury by causing ischemia to the postglomerular circu-
lation (30). The reduction in lumen could also provide a
stimulus for the increased renin expression observed (31) and
might also contribute to the development of the marked hy-
pertension in these rats (32).
The mechanism for the development of the vascular disease
was studied. We have previously reported (10), as has Rao et
Figure 9. Uric acid–induced VSMC proliferation via COX-2 pathway. al. (33), that uric acid may directly stimulate VSMC prolifer-
Stimulation with uric acid (3 mg/dl, ©) resulted in time-dependent ation. In this study, we identified a new mechanism involving
increase in rat aortic smooth muscle cell (VSMC) number (A, n & 6) COX-2. Specifically, we found that rat aortic VSMC showed
compared with control media without uric acid (!!!"!!!). A marked
de novo expression of COX-2 mRNA after incubation with uric
increase in COX-2 mRNA expression was observed in cultured
VSMC at 4 and 8 h after stimulation with uric acid (B, n & 3) in
acid. Incubation of the SMC with either a COX-2 inhibitor or
contrast to no change in COX-1 mRNA expression. Selective COX-2 with a TX-A2 receptor inhibitor could prevent the proliferation
inhibitor (NS398, 10 !M/L) significantly inhibited uric acid-induced in response to uric acid. COX-2 was also shown to be ex-
SMC proliferation, whereas NS398 did not affect 3H-thymidine up- pressed de novo in the preglomerular vessels of RK!OA rats,
take in the absence of UA stimulation (C, n & 6 per group). *P # 0.01 and its expression correlated both with the uric acid levels and
versus other groups. with the degree of SMC proliferation. These studies thus
2896 Journal of the American Society of Nephrology J Am Soc Nephrol 13: 2888–2897, 2002

suggest a critical role for uric acid-mediated, COX-2– gener- In summary, we have identified uric acid as a new and
ated thromboxane in VSMC proliferation in this model. Inter- potentially important mediator of renal progression in the rem-
estingly, COX-2 has also been recently shown to be important nant kidney model of progressive renal disease. Hyperuricemia
in the mechanism by which angiotensin II stimulates VSMC increased systemic BP, proteinuria, renal dysfunction, and pro-
proliferation (24). gressive renal scarring. Hyperuricemia also induces vascular
In addition to COX-2, it is likely that angiotensin II is also disease via a COX-2– dependent pathway. Although one must
contributing to the vasculopathy. As mentioned, renin was also be cautious in the interpretation of animal models to human
increased in this model and correlated with the COX-2 content. disease, these studies provide a mechanism to explain recent
It is unclear if the increased renin reflects direct effects of uric epidemiologic data showing that uric acid is an independent
acid on the juxtaglomerular cells or an indirect effect via risk factor for renal progression (26 –28). Furthermore, our
stimulation of COX-2 in the macula densa and arterioles (13) studies suggest hyperuricemia may be one of the key and
or an indirect effect related to the development of vascular previously unknown mechanisms for the activation of the renin
disease with reduced renal perfusion. Regardless, the genera- angiotensin and COX-2 systems in progressive renal disease.
tion of angiotensin II may result in VSMC proliferation and Studies are needed to determine if early treatment of hyperuri-
hypertrophy (24) as well as inflammatory cell infiltration (34). cemia can slow progressive renal disease in humans.
We have previously reported that the vasculopathy in normal
rats with OA-induced hyperuricemia can be largely prevented Acknowledgments
by blocking the RAS, and we have also found that uric acid– This study was supported by a grant of the Korea Health 21 R&D
mediated VSMC proliferation can be partially inhibited by Project, Ministry of Health & Welfare, Republic of Korea (02-PJ1-
blocking the AT1 receptor (10). It is thus likely that both PG3-20599-0014; to D-H Kang) and by NIH RO1 DK52121 and
angiotensin II and COX-2 are involved in the vascular prolif- HL-68607 (to R.J. Johnson).
eration and inflammation observed in our model. In vivo stud-
ies evaluating the effect of inhibition of the RAS or COX-2 in References
this model of progressive renal disease with hyperuricemia 1. Berger L, Yü T: Renal function in gout: IV: An analysis of 524
would provide a more direct proof of the importance of these gouty subjects including long-term follow-up studies. Am J Med
systems. 59: 605– 613, 1975
Importantly, the changes in preglomerular vessels observed 2. Talbott JH, Terplan KL: The kidney in gout. Medicine 39:
in our RK!OA rats were not solely a consequence of an 405– 467, 1960
3. Gonick HC, Rubini MD, Gleason IO, Sommers SC: The renal
elevation in systemic BP (Figure 5), as evidenced by more
lesion in gout. Ann Int Med 62: 667– 674, 1965
prominent vascular changes in preglomerular vessels of hype- 4. Beck L: Requiem for gouty nephropathy. Kidney Int 30: 280 –
ruricemic RK rats when compared with historic RK rats with 287, 1986
comparable systemic BP but significantly lower uric acid level. 5. Nickeleit V, Mihatsch MJ: Uric acid nephropathy and end stage
This finding suggests uric acid–induced vasculopathy is inde- renal disease: Review of a non-disease. Nephrol Dial Transpl 12:
pendent of elevated systemic BP. Additionally, although in- 1832–1838, 1997
creased COX-2 expression was reported in arteries from hy- 6. Johnson RJ, Kivlighn SD, Kim Y-G, Suga S, Fogo A: Reap-
pertensive rats (35), increased cortical expression of COX-2 praisal of the pathogenesis and consequences of hyperuricemia in
induced by renal ablation or angiotensin II infusion has been hypertension, cardiovascular disease, and renal disease. Am J Kid
shown to be BP-independent (36). Dis 33: 225–234, 1999
An interesting observation was that allopurinol was mark- 7. Rosenfeld JB: Effect of long-term allopurinol administration on
serial GFR in normotensive and hypertensive hyperuricemic
edly effective at lowering uric acid and preventing OA-induced
subjects. Adv Exp Med Biol 41B: 581–596, 1974
renal dysfunction, proteinuria, hypertension, vascular disease, 8. Perez-Ruiz F, Calabozo M, Fernandez-Lopez MJ, Herrero-Beites
renal hypertrophy, and renal scarring. The benefit of allopuri- A, Ruiz-Lucea E, Garcia-Erasukin G, Duruelo J, Alonso-Ruiz A:
nol correlated with its ability to lower uric acid. However, it is Treatment of chronic gout in patients with renal function impair-
possible that some of the benefit may also relate to the known ment. An open, randomized actively controlled study. J Clin
ability of allopurinol to prevent oxidant formation. Benzioda- Rheumatol 5: 49 –55, 1999
rone, a uricosuric agent, was less effective at lowering uric acid 9. Mazzali M, Hughes J, Kim YG, Jefferson JA, Kang D-H, Gordon
and only partially reduced the increase in renin expression, and KL, Lan HY, Kivlighn S, Johnson RJ: Elevated uric acid in-
this may account for its partial benefit on the disease process. creases blood pressure in the rat by a novel crystal-independent
Interestingly, it was more effective at blocking the glomerular mechanism. Hypertension 38: 1101–1106, 2001
changes (proteinuria and glomerulosclerosis) than the vascular 10. Mazzali M, Kanellis J, Han L, Feng L, Chen Q, Kang DH,
Watanabe S, Nakagawa T, Lan HY, Johnson RJ: Hyperuricemia
or interstitial changes. This may reflect that the glomerular
induces a primary arteriolopathy in rats by a blood pressure-
changes may be mediated more by a threshold phenomenon independent mechanism. Am J Physiol Renal Physiol 282: F991–
related to the uric acid level. Another possibility is that the F997, 2002
interstitial injury was not blocked as effectively due to the 11. Mazzali M, Kim Y-G, Suga S, Gordon KL, Kang D-H, Jeferson
effects of increased urinary uric acid on tubular cells that JA, Hughes J, Kivlighn SD, Lan HY, Johnson RJ: Hyperuricemia
occurred as a consequence of the uricosuric action of exacerbates chronic cyclosporine nephropathy. Transplantation
benziodarone. 71: 900 –905, 2001
J Am Soc Nephrol 13: 2888–2897, 2002 Uric Acid and Renal Disease Progression 2897

12. Vaziri ND, Freel RK, Hatch M: Effect of chronic experimental tensin II-mediated proliferation of vascular smooth muscle cells.
renal insufficiency on urate metabolism. J Am Soc Nephrol 6: Circ Res 86: 906 –914, 2000
1313–7, 1995 25. Hebert LA, Wilmer WA, Falkenhain ME, Ladson-Wofford ST,
13. Harris RC, Breyer MD: Physiological regulation of cyclooxy- Nahman NS, Rovin BH: Renoprotection: One or many therapies?
genase-2 in the kidney. Am J Physiol Renal Physiol 281: F1–F11, Kidney Int 59: 1211–1226, 2001
2001 26. Syrjänen J, Mustonen J, Pasternak A: Hypertriglyceridemia and
14. Wang J-L, Cheng H-F, Shappell S, Harris RC: A selective hyperuricemia are risk factors for progression of IgA nephropa-
cycloxygenase-2 inhibitor decreases proteinuria and retards pro- thy. Nephrol Dial Transpl 15: 34 – 42, 2000
gressive renal injury in rats. Kidney Int 57: 2334 –2342, 2000 27. Ohno I, Hosoya T, Gomi H, Ichida K, Okabe H, Hikita M: Serum
15. Bunag RD, Buttefield J: Tail-cuff blood pressure measurement uric acid and renal prognosis in IgA nephropathy. Nephron 87:
without external preheating in awake rats. Hypertension 4: 898 – 333–339, 2001
903, 1982 28. Iseki K, Oshiro S, Tozawa M, Iseki C, Ikemiya Y, Takishita S:
16. Henry RJ, Sobel C, Kim J: A modified carbonate phosphotung- Significance of hyperuricemia on the early detection of renal
state method for the determination of uric acid and comparison failure in a cohort of screened subjects. Hypertens Res 24:
with the spectrophotometric uricase method. Am J Clin Pathol 691– 697, 2001
28: 152–160, 1957 29. Saito I, Saruta T, Kondo K, Nakamura R, Oguro T, Yamagami
17. Kang DH, Joly AH, Oh SW, Hugo C, Kerjaschki D, Gordon KL, K, Ozawa Y, Kato E: Serum uric acid and the renin-angiotensin
Mazzali M, Jefferson JA, Hughes J, Madsen KM, Schreiner GF,
system in hypertension. J Am Geriatrics Soc 26: 241–247, 1978
Johnson RJ: Impaired angiogenesis in the remnant kidney model
30. Fine LG, Orphanides C, Norman JT: Progressive renal disease:
(I): Potential role of vascular endothelial growth factor and
The chronic hypoxia hypothesis. Kidney Int 65: S74 –S78, 1998
thrombospondin-1. J Am Soc Nephrol 12: 1434 –1447, 2001
31. Goldblatt H: The renal origin of hypertension. Physiol Rev 27:
18. Elger M, Sakai T, Kritz W: The vascular pole of the renal
120 –165, 1947
glomerulus of rat. Adv Ana Embryol Cell Biol 139: 1–98, 1998
32. Johnson RJ, Herrera-Acosta J, Schreiner GF, Rodriguez-Iturbe
19. Eng E, Veniant M, Floege J, Fingerle J, Alpers CE, Menard J,
B: Subtle acquired renal injury as a mechanism of salt-sensitive
Clozel JP, Johnson RJ: Renal proliferation and phenotypic
changes in rats with two-kidney, one-clip Goldblatt hyperten- hypertension. N Engl J Med, 346: 913–923, 2002
sion. Am J Hypertens 7: 177–185, 1994 33. Rao GN, Corson MA, Berk BC: Uric acid stimulates vascular
20. Feng L, Sun W, Xia Y, Tang W, Chanmugam P, Soyoola E, smooth muscle cell proliferation by increasing platelet derived
Wilson CB, Hwang D: Cloning two isoforms of rat cyclooxy- growth factor A-chain expression. J Biol Chem 266: 8604 – 8608,
genase: Differential regulation of their expression. Arch Biochem 1991
Biophys 307: 361–368, 1993 34. Theuer J, Dechend R, Muller DN, Park JK, Fiebeler A, Barta P,
21. Xia Y, Pauza ME, Feng L, Lo D: RelB regulation of chemokine Ganten D, Haller H, Dietz R, Luft FC: Angiotensin II induced
expression modulates local inflammation. Am J Pathol 151: inflammation in the kidney and in the heart of double transgenic
375–387, 1997 rats. BMC Cardiovasc Disord 2: 3–14, 2002
22. Couser WG, Pippin JW, Shankland SJ: Complement (C5b-9) 35. Henrion D, Dechaux E, Dowell FJ, Maclour J, Samuel JL, Levy
induces DNA synthesis in rat mesangial cells in vitro. Kidney Int BI, Michel JB: Alteration of flow-induced dilatation in mesen-
59: 905–912, 2001 teric resistance arteries of L-NAME treated rats and its partial
23. Kang DH, Hughes J, Mazzali M, Schreiner GF, Johnson RJ: association with induction of cyclo-oxygenase-2. Br J Pharma-
Impaired angiogenesis in the remnant kidney model (II): VEGF col 121: 83–90, 1997
administration reduces renal fibrosis and stabilizes renal func- 36. Hernandez J, Astudillo H, Escalante B: Angiotensin II stimulates
tion. J Am Soc Nephrol 12: 1448 –1457, 2001 cyclooxygenase-2 mRNA expression in renal tissue from rats
24. Young W, Mahboubi K, Haider A, Li I, Ferreri NR: Cyclooxy- with renal failure. Am J Physiol Renal Physiol 282: F592–F598,
genase-2 is required for tumor necrosis factor-alpha- and angio- 2002

Access to UpToDate on-line is available for additional clinical information

at http://www.jasn.org/