Kidney International, Vol. 58 (2000), pp. 740–747

Renal nitric oxide production during the early phase ofexperimental diabetes mellitus

SHOSHANA KEYNAN, BOAZ HIRSHBERG, NOMY LEVIN-IAINA, ISAIAH D. WEXLER,RACHEL DAHAN, ETTY REINHARTZ, HAIM OVADIA, YORAM WOLLMAN,TAMARA CHERNIHOVSKEY, ADRIAN IAINA, and ITAMAR RAZ

Department of Internal Medicine and Department of Neurology, Hadassah University Hospital, Jerusalem; andDepartment of Nephrology, Ichilov Hospital, Tel Aviv Medical Center, Tel Aviv, Israel

Renal nitric oxide production during the early phase of experi- of the mediators implicated in the development of DNmental diabetes mellitus. is the nitric oxide (NO) system [4]. NO acts as a neuroneal

Background. Diabetic nephropathy (DN) is characterizedand vascular messenger that activates soluble guanylateby hyperfiltration and hypertrophy in experimental models ofcyclase, resulting in increased levels of cGMP. NO acti-diabetes mellitus (DM). Several studies have demonstrated

that the pathophysiologic and morphologic changes in DN are vates multiple intracellular pathways, including endothe-mediated by either an increase or decrease in renal nitric oxide lium-dependent relaxation, inhibition of platelet aggrega-(NO) production and/or activity. The goal of the present study

tion and adhesion, and tubuloglomerular feedback (TGF)was to determine the effects that the early diabetic state hasin the kidney [4]. NO synthase (NOS) catalyzes the con-on NO production in the kidney of rats with streptozotocin-

induced DM. version of arginine to citrulline and NO. NOS exists inMethods. Experimental DM was induced in rats with strep- three main isoforms: neuroneal NOS (nNOS), endothe-

tozotocin. Urinary NO production was measured, and levelslial NOS (eNOS), and inducible NOS (iNOS). nNOSand activity of the different NOS isoforms were determinedand eNOS are Ca11/calmodulin-dependent, whereas cal-by a combination of techniques, including immunoblotting,

immunohistochemistry, diaphorase staining, and reverse tran- cium-independent iNOS is regulated by a variety of cyto-scription-polymerase chain reaction. kines. Within the kidney, high levels of nNOS are located

Results. During the first week of DM, urinary NO metabo-in the macula densa. eNOS is found in glomerular affer-lites (uNO2 1 NO3) were reduced as compared with controls,ent and efferent arterioles, and iNOS has been localizedwhich were unrelated to changes in serum levels of NO. Total

NO synthase (NOS) activity was reduced in the renal cortex to the cells of the mesangium [5].beginning at 30 hours after the induction of DM. NADPH The exact role of NO in mediating DN has not beendiaphorase staining of renal cortical slices showed reduced NOS

fully elucidated, and the results of previous studies haveactivity in the macula densa in diabetic animals. By immunohis-not been consistent with each other. NO appears to medi-tochemical staining with antibodies to the different isoforms

of NOS, it was found that protein levels of the neuroneal NOS ate DN by its effects on both renal hemodynamics and(nNOS) isoform was diminished in the macula densa. No changes morphology. In experimentally-induced diabetes melli-were found in the levels of endothelial NOS (eNOS) activity

tus (DM), the elevation of the glomerular filtration rateand protein in the renal cortex in the early diabetic state.(GFR) was related to an increase in the synthesis andConclusions. This study provides strong evidence that renal

production of NO is reduced in early DM and that this reduc- action of NO [3]. Komers, Allen, and Cooper demon-tion is associated with decreased levels of nNOS activity and strated that in streptozotocin (STZ) induced DM, uri-protein in the macula densa.

nary NO metabolites (uNO2 1 NO3) were increased,and inhibition of NO production by N-nitro-l-argininemethyl ester (L-NAME), a NOS inhibitor, preventedRenal hypertrophy and hyperfiltration characterizethe increase in GFR that normally accompanies DMthe early stage of diabetic nephropathy (DN) [1–3]. Onein these animals [6]. On the other hand, NO-inducedvascular relaxation is impaired in DM, and this defectKey words: streptozotocin, renal hypertrophy, nephropathy, macula

densa, neuronal nitric oxide synthase. is repaired when l-arginine, a precursor of NO, is pro-vided in excess, suggesting that early experimental DM

Received for publication August 18, 1999is associated with a defect in NO production [7]. Long-and in revised form January 3, 2000

Accepted for publication March 16, 2000 term dietary supplementation of arginine, which increasesthe level of NO production, prevented the changes in 2000 by the International Society of Nephrology

740

Keynan et al: Renal NO production in DM 741

renal function and structure normally found in STZ- five, and seven days. In preliminary experiments, it wasdetermined that changes (body weight, glucose level, andinduced experimental DM [8, 9].

The aim of the present study was to characterize renal kidney growth) seen in experimental animals housed inmetabolic cages were similar to those in animals main-NO production and the expression of the various NOS

isoforms in early experimental DM. The relationship tained in regular cages. At the end of the study, theanimals were anesthetized with an intraperitoneal injec-between renal hypertrophy and elevated glomerular fil-

tration seen shortly after the onset of experimental DM tion of sodium pentobarbital, and lung, liver, heart,spleen, and both kidneys were rapidly removed. Afterand changes in the NO system was also investigated. Our

results indicate that the primary alteration in the renal weighing the kidneys, the left kidney of each experimen-tal animal was kept intact for preparation of histologicNO system in early DM involves a decrease in the level

of nNOS activity and protein expression in the macula slides, while the right cortex and medulla were separated.Tissues were stored at 2708C in preparation for thedensa that occurs simultaneously with the increase in

GFR and renal hypertrophy. analytic procedures described later in this article.

Urinary and plasma nitric oxide determinationMETHODS

Urine and plasma NO2 1 NO3, the stable metabolitesMaterials of NO, were determined by the Griess reaction, after

the reduction of NO3 to NO2 by nitrate-reductase pre-Antibodies to endothelial and nNOS were obtainedfrom Transduction Laboratories (Lexington, KY, USA). pared in our laboratory [10].[2,3-3H] l-arginine (43.5 Ci/mmol) was purchased from

Nitric oxide synthase activityNEN DuPont (Savyon, Israel). Aminoguanidine was pur-chased from Cal-Biochem (La Jolla, CA, USA). N-nitro- Nitric oxide synthase activity in tissue homogenates

was measured by monitoring the conversion of [3H] argi-l-arginine methyl ester (L-NAME), b-nicotinamidephosphate, NADPH, STZ, and all other materials were nine to [3H] citrulline [11]. Tissues were homogenized

in 50 mmol/L Tris-HCl buffer (pH 7.4) and 1 mmol/Lobtained from Sigma Israel (Rehovot, Israel).ethylenediaminetetraacetic acid (EDTA). The homoge-

Animals and experimental protocol nates were spun at 10,000 3 g for 10 minutes at 48C. Toremove the endogenous arginine, the supernatants wereThe experiments were conducted on Hebrew University

Sabra male rats weighing approximately 200 g. The ani- collected and incubated with Dowex 50W-X8 (Na1; Bio-Rad, Hercules, CA, USA) for five minutes at 48C. Themals were maintained and treated in accordance with the

regulations of the Hadassah Committee on Animal Care samples were spun again, and aliquots (500 mg of protein)were incubated with 10 mmol/L NADPH (reducedand Use and followed the Principles of Laboratory Ani-

mal Care (NIH publication no. 85-23, revised 1985). Rats form), 10 mmol/L CaCl2, 50 mmol/L l-valine, and [3H]arginine (0.5 mCi) with or without NOS inhibitors [10were housed in a temperature- and humidity-controlled

room with a 12-hour light/dark cycle. The animals were mmol/L L-NAME, egtazic acid (EGTA), or aminoguani-dine]. After 20 to 50 minutes of incubation at 378C, themaintained on standard laboratory animal chow and pro-

vided water ad libitum. Diabetes was induced by intrave- reaction was terminated by adding 3 mL of ice-cold HEPESbuffer (pH 5.5) containing 2 mmol/L EDTA, and thennous injection of STZ (diluted in Dextrose 5%, pH 4.5)

at a dose of 65 mg/kg body wt. Daily blood glucose passed on a column of Dowex (0.5 mL). A 0.5 mL aliquotfrom the eluted [3H] citrulline was quantitated by liquidlevels were monitored with a portable glucose-measuring

device (Glucometer Elite, Bayer, Elkhart, IN, USA) us- scintillation spectroscopy. The total protein concentra-tion of tissue homogenates was measured (standard Bio-ing blood obtained by tail prick of conscious rats.

The two groups, control and STZ rats, were all aged Rad protein assay).matched. For measurement of food consumption, uri-

Immunohistochemistrynary and plasma NO metabolites, and creatinine clear-ance, individual animals were placed in metabolic cages For immunohistochemistry staining, kidney tissue was

snap frozen in isopentane at 2208C, cut in 6 mm thickfor periods of time lasting for seven days. Body weight,food, and fluid intake were measured. Urine collections sections, and thaw mounted on slides. The slides were

incubated in 4% formaldehyde (pH 7.4) for fixationwere obtained, and blood samples were drawn from aninternal jugular vein. The rats were allowed 48 hours (30 min at room temperature), washed in phosphate-

buffered saline (PBS), pH 7.4, and incubated for blockingof adaptation in the metabolic cages. Plasma and urinecreatinine levels were determined, and the creatinine in PBS, pH 7.4, containing 5% bovine serum albumin

(BSA), and 0.1% Triton X-100 for 20 minutes at roomclearance was calculated by standard laboratory meth-ods. For the other experiments involving tissue analyses, temperature. This was followed by a 16-hour incubation

at 48C with polyclonal anti-nNOS (1:100 dilution). Subse-rats were housed two per cage for 30 hours and three,

Keynan et al: Renal NO production in DM742

quently, the slides were washed and incubated with anti- Reverse transcription-polymerase chain reactionrabbit Alexa-488 green fluorescence (1:200 dilution). Total RNA was isolated from kidney cortex of controlOther slides were incubated with eNOS monoclonal anti- and STZ-diabetic rats at day 5 using TRIzol reagentbody (1:100 dilution) in PBS containing 1% BSA and (GIBCO BRL, Grand Island, NY, USA) according to0.1% Triton X-100. The slides were washed and incu- manufacturer’s instructions.bated with the secondary biotinylated horse anti-mouse Reverse transcription followed by DNA amplificationantibody followed by peroxidase-avidin substrates ac- was performed using the access reverse transcription-cording to the manufacturer’s recommended protocol polymerase chain reaction (RT-PCR) system kit (Pro-(Vectastain elite ABC kit; Vector Laboratories, Burl- mega, Madison, WI, USA). Varying amounts of total RNAingame, CA, USA). were used in the experiments to assure that the assay was

semiquantitative. As a control, deoxynucleotide primersDiaphorase staining complementary to the rat glyceraldehyde-3-phosphate

Sections of renal tissue, 10 mm thick, prepared from dehydrogenase (GAPDH) primers (sense 59AATGCATfrozen kidney specimens sections were incubated for CCTGCACCACCAA39 and antisense 59ATACTGT

TACTTATACCGATG39) were utilized under the samefour hours in 4% buffered formaline, pH 7.0, and wereconditions. Specific rat nNOS primers were 59GAATwashed three times in PBS. The frozen sections wereACCAGCCTGATCCATGG-AAC39 (sense, positionincubated for two hours at 378C in the dark with a sub-2461 to 2484) and 59TCCTCCAGGAGGGTGTCCAC-strate mixture containing 0.1 mol/L Tris-HCl (pH 7.8),CGCA39 (antisense, position 3039 to 3062) [11].1 mmol/L b-NADPH, 0.1 mmol/L nitro blue tetrazolium,

The PCR was performed on a PTC-100 DNA thermal0.3% Triton X-100, and 0.1% formaldehyde. All sectionscycler (MJ Research, Watertown, MA, USA). The con-were rinsed with PBS and fixed in Kaiser-glycerol gelatinditions for DNA amplification were as follows: Aftersolution according to the method of Ovadia et al [12].initial denaturation at 948C for two minutes, 35 cyclesThe enzyme activity was evaluated by either countingof amplification (948C for 1 min, 608C for 1 min, and 728Cthe positive macula densa stained out of 100 glomerulifor 2 min) were performed, followed by final extension atper group or by counting the number of visible macula728C for five minutes. The PCR products were separateddensa cells per stained macula. The results are expressedon an ethidium bromide-stained 1.5% agarose gels.as the mean number of positive stained cells per glomer-

uli [13]. To determine the validity of using frozen tissueStatistical analysisversus perfused tissue, preliminary experiments were

All values are expressed as mean 6 SEM. Student’sdone with kidney sections from rats perfused with para-t-test was used to assess statistical significance betweenformaldehyde followed by fixation in formalin and incu-groups. Statistical significance was defined as P , 0.05.bation in a 30% sucrose solution for three days. Tissue

was otherwise prepared in similar manner and reactedwith NADPH diaphorase. No differences were detected RESULTSbetween frozen sections and sections from rats perfused Male Sabra rats weighing approximately 180 to 200 gwith paraformaldehyde. received STZ at a dose of 65 mg/kg body weight and

were killed at various time periods after the inductionImmunoblottingof DM. Table 1 summarizes the characteristics of these

Samples (50 mg protein) of tissue homogenates were animals and concurrent controls. As expected, animalsseparated by sodium dodecyl sulfate-polyacrylamide gel with DM failed to gain weight. In contrast, both kidneyelectrophoresis (SDS-PAGE; 7.5% acrylamide) and weight and the kidney:body weight ratio increased intransferred to Immobilon-P membrane (Millipore, Bed- diabetic animals (statistically significant after 3 days),ford, MA, USA). Blots were blocked for one hour at indicating renal growth. These changes occurred evenroom temperature in 5% nonfat milk and incubated with though food intakes was controlled for both experimen-monoclonal anti-eNOS antibody (1:1000 dilution) for 16 tal and control animals 16.3 6 1.6 and 17.1 6 1.3 (g),hours at 48C. After washing in TBS-Tween (0.1%), the respectively.membrane was incubated with anti-mouse IgG antibody The 24-hour urinary NO2 1 NO3 excretion (Fig. 1) wasconjugated with horseradish peroxidase (1:1000 dilution) lower in STZ-rats from day 3 to the end of experiment,for one hour at room temperature. The bands were visu- compared with nondiabetic rats despite concurrently oc-alized by enhanced chemiluminescence (ECL kit; Amer- curring increases in creatinine clearance (Fig. 2). Duringsham Pharmacia Biotech, Buckinghamshire, UK). Bands the same period (7 days), plasma NO metabolites werecorresponding to eNOS were quantitated by scanning unchanged (control 24.5 6 1.6 vs. diabetes 25.7 6 1.8densitometry (OD 3 mm2) of the Western blot autora- mmol/L). These findings suggest that reduced urine NO

metabolites are related to decreased renal production ofdiogram using a Fluro-S-multimager (Bio-Rad).

Keynan et al: Renal NO production in DM 743

Table 1. Physical and biochemical parameters of the control and diabetic rats

Control Diabetes Control Diabetes Control Diabetes Control Diabetes

30 hours 3 days 5 days 7 days

N 5 7 5 6 5 6 5 5Initial body weight g 176619.5 195617.2 18868 18667 18868 18868 19465 19163Final body weight g 179614.3 196613 220619 185619 224613 180614 22869 18264Kidney weight g (L1R)/2 0.6160.03 0.6260.06 0.6660.22 0.7860.09 0.7460.06 0.8160.13 0.7560.01 0.9460.02a

Kidney:body weight ratio 31023 3.460.00 3.260.00 3.060.0 4.260.1b 3.360.0 4.560.0b 3.360.0 5.260.0b

Blood glucose mmol/L 5.661.8 19.963.2b 7.860.8 16.561.3b 7.860.6 22.761.9b 7.760.6 18.361.6b

aP , 0.05, bP , 0.01

Fig. 1. Changes in urinary nitric oxide metabolites (uNO2 1 NO3) inFig. 2. Creatinine clearance in STZ rats during the first week of diabe-streptozotocin (STZ) rats during the first week of diabetes. Symbolstes. Symbols are: (d) control; (.) diabetic rats; *P , 0.01 diabetic ratsare: (d) control; (.) diabetic rats; *P , 0.01; **P , 0.05.vs. control.

NO due to decreased levels of NOS in the kidney. Inequals eNOS 1 nNOS). Additional experiments weresupport of this assumption, it was found that total NOSdone to assess whether the changes in cNOS activity areactivity in the kidney cortex was reduced beginning atdue to changes in protein levels of eNOS or nNOS. Using30 hours despite NOS activity increasing significantly ina monoclonal antibody specific to eNOS, we found thatthe liver (Table 2). The tissue NOS activity was almostalthough eNOS levels tends to increase in the liver ho-negligible in spleen, lung, heart, and kidney medulla evenmogenates of diabetic rats, it did not reach statisticalwhen large amounts (up to 1 mg) of total cellular proteinsignificance compared with control (2.43 6 0.23, 1.87 6were assayed.0.65 P 5 0.08, respectively; Fig. 3A). In kidney cortexSeveral studies were performed to determine whichhomogenates, five days after the induction of diabetes,renal isoforms of NOS were decreased in early DM. ToeNOS protein levels were the same (control 0.35 6 0.22determine whether iNOS activity was reduced in thevs. diabetes 0.36 6 0.12; Fig. 3B). We could not visualizecortex, the NOS assay was carried out in the presenceany positive bands to nNOS protein in cortical homoge-of aminoguanidine, as the it is a more selective inhibitornates using up to 100 mg of protein from control orof iNOS. Total NOS activity in the cortical homogenatediabetic animals using a polyclonal antibody to nNOS.did not change either in the experimental or controlHowever, there are no changes in the levels of nNOSanimals, suggesting that the contribution of iNOS to themRNA in the kidney cortex as assessed by RT-PCRoverall NOS activity is negligible. Further evidence that(Fig. 4).most of the NOS activity is derived from the calcium-

Since no changes in the NOS isoforms could be de-dependent forms was demonstrated after the additiontected in cortical homogenates, immunohistochemicalof 10 mmol/L EGTA, a calcium chelator, which reducedstaining was utilized to determine whether the diabeticNOS activity to background levels.state induced localized changes in NOS activity and pro-Thus, the changes in kidney cortex and liver NOS

activity (Table 2) are those of cNOS (constitutive NOS tein. Kidney slices from diabetic rats killed at three, five,

Keynan et al: Renal NO production in DM744

Table 2. Nitric oxide synthase activity in liver and kidney cortex

Control Diabetes Control Diabetes Control Diabetes Control Diabetes

30 hours 3 days 5 days 7 days

N 5 7 5 7 5 6 5 5Liver 100615 115619 100613 126613a 100623 270630b 100629 15069Kidney cortex 100616 57618b 100612 53625b 100615 36627b 100618 50617a

Thirty hours to 7 days after induction of diabetes nitric oxide synthase activity was assayed in the tissue homogenates by following the conversion of L-[3H]arginine to [3H]-citrulline (pmol/min/mg protein) with L-valine (50 mmol/L) present in the assay system. The activity was defined as citrulline formation totallyinhibited by 10 mmol/L L-NAME.

The activity of control rats was determined as 100% in liver and kidney cortex homogenates, 0.045 and 0.0178 pmol/min/mg protein, respectively.aP , 0.05, bP , 0.01, diabetic rats vs. control

Fig. 3. Western blot analysis of liver and kidney cortex homogenate.Fifty micrograms of liver (A) or kidney cortex (B) homogenate wasseparated on 7.5% SDS-PAGE. After blotting, the membrane wasreacted with monoclonal anti-endothelial nitric oxide synthase (anti-eNOS) antibody (1:1000). PC, human endothelial cell lysate.

Fig. 5. NADPH-diaphorase staining of sections from normal and dia-betic rat kidney. (A) Normal rat kidney was prepared and stained asdescribed under the Methods section (magnification 3400). (B) Popula-Fig. 4. Effect of five days diabetes on expression of neuronal nitriction of nNOS positive macula densa of the experimental rats. Twelveoxide synthase (nNOS) mRNA in kidney cortex. A representative ethid-sections from diabetic (N 5 5) and control (N 5 5) animals were stainedium bromide-stained agarose gel of kidney cortex nNOS (601 bp) andwith NADPH-diaphorase. The number of positive macula densa perGAPDH (515 bp) RT-PCR products. Lanes 1 through 3, control rats;100 counted glomeruli is expressed in percentage. *P , 0.01.lanes 4 through 6, STZ-diabetic rats.

ing, both in terms of the number of NADPH-diaphorase–and seven days after the induction of DM reacted withpositive macula densa per 100 glomeruli (Fig. 5) and thenitroblue tetrazolium (NADPH diaphorase), an indica-number of visibly stained cells per positive macula densator of NOS activity, did not demonstrate any changes inin the diabetic rats as compared with the controls (3.30 6staining intensity within the glomerulus. However, there

was a significant reduction in nitroblue tetrazolium stain- 0.97 vs. 6.42 6 1.23 on day 5, respectively, P , 0.01).

Keynan et al: Renal NO production in DM 745

is no difference in cellular staining when antibodies toeNOS are used for immunostaining. In addition, the onlychange in NADPH diaphorase staining, a marker forNOS activity, occurred in the macula densa—an areaknown to have high nNOS activity [23]. It is also unlikelythat the source of decreased urinary NO metabolites indiabetic rats was secondary to decreased systemic NOproduction as the plasma NO levels were the same inthe diabetic state as in the control and liver NOS activitywas significantly increased.

Choi et al showed by Western blot that protein levelsFig. 6. Confocal images of the normal kidney sections immunostainedfor nNOS. (a) Differential interference contrast (DIC) images were of all three NOS isoforms increased one week after in-collected simultaneously using a transmitted light detector. (b) A Zeiss duction of diabetes [24]. It should be noted, however,LSM 410 confocal laser scanning system attached to the Zeiss Axiovert

that the decrease in nNOS activity in the macula densa135M microscope was utilized Green fluorescence of Alexa-488 (sec-ondary antibody was used at 1:200; Molecular Probes, Eugene, OR, is demonstrated in our study by histologic analysis usingUSA) was excited with an argon laser. both an anti-nNOS antibody and diaphorase staining.

There were also no significant alterations in eNOS stain-ing throughout the cortex, and the only change inNADPH diaphorase occurred in the macula densa.The decline in NOS activity in the macula densa was

It also appears that the expression of NOS isoformsparalleled by a decrease in the number of stained maculais an evolving process. In a study by Sugimoto et al,densa (16% stained macula densa in control vs. 2% init was demonstrated that eNOS protein and NADPHdiabetic glomeruli on day 5 after the induction of DM), asdiaphorase staining intensity began increasing in the af-detected by immunohistochemistry using the polyclonalferent arterioles and glomeruli beginning at one weekanti-nNOS antibody (Fig. 6). No significant changes inof DM [20]. Yagihashi et al showed that nNOS activitystaining patterns between the control and diabetic stateand NADPH diaphorase staining remain low in the mac-could be detected when the eNOS antibody was utilizedula densa in prolonged DM [25]. Taken together, the(data not shown).data suggest that nNOS activity in the kidney declinesvery early in DM and remain low, whereas eNOS activity

DISCUSSION is initially stable, as shown by our data, and increases asthe diabetic state is prolonged.The NO system has been implicated in the pathogene-

sis of DN [4]. In pathological conditions, such as in DM, An increase in NO metabolites in the urine and plasmaof diabetic rats was shown before, and the apparenthyperglycemia can induce endothelial dysfunction, which

may result from decreased production of NO or other differences may be due to the different time periodsstudied, for example, during a later phase of DM, at afactors that interfere with NO activity [14, 15]. However,

the mechanisms by which changes in the level of NO are time when renal hypertrophy begins to plateau [3, 6, 26].Moreover, the increase in both plasma and urine NOrelated to renal hemodynamics and morphology occurring

in DM have not been fully elucidated [16–22]. In the metabolites reported by other groups does not necessar-ily imply increased renal production of NO. It is possiblepresent study, it has been shown that during the early

stage of DM, there is a decrease in urinary NO metabo- that increased NO metabolites were due to increaseddietary intake of the diabetic rats [27] (food consumptionlites and cortical renal NOS activity associated with a

decrease in nNOS activity and immunoreactivity in the was not controlled in most of those studies) or increasedorgan production of NO metabolites because of highermacula densa. These changes in both urinary metabolites

and nNOS activity were observed during the induction levels of tissue free radicals as the course of DM pro-gressed. The recent study by Sugimoto et al discussedof early experimental diabetes at a stage when renal

hyperfiltration and kidney growth is most evident [1–3]. previously in this article supports the idea that urinaryNO metabolites increase only after the diabetic state isThe results presented previously in this article suggest

that there is a link between the decrease in urinary NO prolonged [20]. These workers did not see an increasein NO metabolites until the second week of DM, whichmetabolites and the reduction in nNOS activity in the

macula densa, as we did not find any other changes in paralleled the increase in cortical eNOS activity and wasblocked by the NOS inhibitor L-NAME [20]. It is possi-other NOS isoforms induced by the early diabetic state.

Renal medulla NOS activity is negligible in either the ble that there is a biphasic pattern in renal NO produc-tion in experimental DM. As was shown recently bynormal or diabetic state. In the cortex, there is minimal

iNOS activity, and the levels of eNOS protein and nNOS Pieper, the disease duration is an important factor inthe NO-mediated endothelium relaxation of large arterymRNA did not change in the cortical homogenates; there

Keynan et al: Renal NO production in DM746

[28]. In preliminary experiments, the urinary excretion the macula densa [34]. It is possible that reduced local-ized NO production in the macula densa may enhanceof NO metabolites decreases in the first week after the

induction of diabetes compared with control rats. At two the effect of angiotensin II on the EA, thereby causingvasoconstriction of the EA that is greater than the angio-weeks, we noted an increase that was very significant at

five weeks. The same phenomena is seen in Wistar rats, tensin II-mediated vasoconstriction of the AA. This, inturn, would cause elevated intraglomerular pressure andwith the only difference being that the increase in urinary

NO metabolites begins earlier, at two weeks after the increased GFR in the early diabetic state. The later in-crease in eNOS, which is primarily located in the afferentinduction of diabetes (A. Iaina, unpublished observation).

Experiments are now being conducted in our laboratory arterioles as shown by Sugimoto et al, would result invasodilation of the AA, thereby sustaining continuedto determine the mechanism for this biphasic response

in urinary NO metabolites in response to diabetes. It is renal hyperfiltration [20].In conclusion, we have shown that in early experimen-possible that the early decline in urinary NO metabolites

reflects a decrease in nNOS activity and protein levels tal DM, there is a reduction in macula densa nNOS,which is associated with hyperfiltration, renal hypertro-in the macula densa followed by increased renal excre-

tion of NO metabolites beginning after the first week phy, and a decrease in urinary NO metabolites. Furtherstudies are necessary to define the causal relationshipthat results from an increase in eNOS activity.

Our study would suggest that the decline in nNOS between the decrease in nNOS in the macula densa, andthe increased hyperfiltration and hypertrophy seen inactivity is an early change that is temporally related to

renal hyperfiltration and renal growth seen in the first the first week of experimental DM.week of experimental DM. As was suggested by Craven,Studer, and DeRubertis and others, the NO generation ACKNOWLEDGMENTSor action is impaired in the diabetic glomeruli, which in This study was partly supported by a grant from the Chief Scientist,

Ministry of Health, Israel. Dr. Keynan is partly supported by a grantturn might contribute to glomerular capillary hyperten-from the Center for Absorption in Science, Ministry of Immigrantsion [29–31]. The relevance of decreased NO activityAbsorption.

during diabetes to the renal hemodynamics changes wasalso discussed by Wang, Brooks, and Edwards [32]. Hyper- Reprint requests to Itamar Raz, M.D., Department of Medicine, Ha-

dassah University Hospital, Kiryat Hadassah, P.O.B. 12000, Jerusalem,filtration results from an increase in intraglomerular pres-Israel 91120.

sure that, in turn, is dependent on the balance betweenthe vascular tone of the afferent arteriole (AA) andthe efferent arteriole (EA). Several studies, based on APPENDIXexperimental inhibition of NOS, have suggested that in- Abbreviations used in this article are: AA, afferent arteriole; BSA,creased production of NO is responsible for the hyperfil- bovine serum albumin; DM, diabetes mellitus; DN, diabetic nephropa-

thy; EA, efferent arteriole; eNOS, endothelial NOS; GAPDH, glyceral-tration seen in DM [3, 27, 33]. NO antagonizes the vaso-dehyde-3-phosphate dehydrogenase; GFR, glomerular filtration rate;constrictive effects of angiotensin II on the AA, thereby iNOS, inducible NOS; L-NAME, N-nitro-l-arginine methyl ester; NO,

increasing the GFR [34]. The increased levels of NO nitric oxide; nNOS, neuroneal NOS; NOS, nitric oxide synthase; PBS,phosphate-buffered saline; PCR, polymerase chain reaction; RT, re-also counteract the effects that increased renal bloodverse transcription; TGF, tubuloglomerular feedback, STZ, streptozo-flow has on tubuloglomerular feedback (TGF)-mediated tocin; and uNO2 1 NO3, urinary nitric oxide metabolites.

constriction of the AA [35].It might be that the hyperfiltration seen in early DM REFERENCES

is related to other factors besides NO. Altered renal1. Seyer-Hansen K: Renal hypertrophy in experimental diabetes:hemodynamics increased extracellular fluid expansion, Relation to of diabetes. Diabetologia 13:141–143, 1977

and glomerular hypertrophy is a causative factor in the 2. Ross J, Goldman JK: Effect of streptozotocin-induced diabeteson kidney weight and compensatory hypertrophy in the rat. Endo-increased renal blood flow that is not counteracted bycrinology 88:1079–1082, 1971TGF [36]. Other factors such as prostaglandins and atrial 3. Bank N, Aynedjian HS: Role of EDRF (nitric oxide) in diabetic

naturietic peptide may mediate increased vascular relax- renal hyperfiltration. Kidney Int 43:1306–1312, 19934. Craven PA, DeRubertis FR, Melhem M: Nitric oxide in diabetication of the AA despite lower levels of NO.

nephropathy. Kidney Int 52(Suppl 60):S46–S53, 1997Decreased macula densa NO production may also5. Star RA: Intrarenal localization of nitric oxide synthase isoforms

have a disproportionate effect on the EA in early DM. and soluble guanylyl cyclase. Clin Exp Pharmacol Physiol 24:607–610, 1997Recent studies have provided evidence that NO has a

6. Komers R, Allen TJ, Cooper ME: Role of endothelium-derivedprominent role in modulating the resistance of the EAnitric oxide in the pathogenesis of the renal hemodynamic changes

[34, 37]. The EA is very sensitive to the vasoconstrictive of experimental diabetes. Diabetes 43:1190–1197, 19947. Pieper GM, Siebeneich W, Moore-Hilton G, Roza AM: Reversaleffects of angiotensin II [38] that is counteracted by the

by l-arginine of a dysfunctional arginine/nitric oxide pathway invasodilatory effects of NO. The effect of NO may bethe endothelium of the genetic diabetic BB rat. Diabetologia

most important in the early segment of the EA, which 40:910–915, 19978. Reyes AA, Karl IE, Kissane J, Klahr S: l-arginine administrationis located in the glomerulus, and in close proximity to

Keynan et al: Renal NO production in DM 747

prevents glomerular hyperfiltration and decreases proteinuria in duction in diabetes: Protective role of arginine on endothelial dys-function. Hypertension 31:1047–1060, 1998diabetic rats. J Am Soc Nephrol 4:1039–1045, 1993

23. Mundel P, Bachmann S, Bader M, Fischer A, Kummer W, Mayer9. Reyes AA, Klahr S: Dietary supplementation of l-arginine ame-B, Kriz W: Expression of nitric oxide synthase in kidney maculaliorates renal hypertrophy in rats fed a high-protein diet. Proc Socdensa cells. Kidney Int 42:1017–1019, 1992Exp Biol Med 206:157–161, 1994

24. Choi KC, Kim NH, An MR, Kang DG, Kim SW, Lee J: Alterations10. Schwartz D, Blum M, Peer G, Wollman Y, Maree A, Serbanof intrarenal renin-angiotensin and nitric oxide systems in strepto-I, Grosskopf I, Cabili S, Levo Y, Iaina A: Role of nitric oxidezotocin-induced diabetic rats. Kidney Int 51(Suppl 60):S23–S27,in radiocontrast acute renal failure in rats. Am J Physiol 267:F374–1997F379, 1994

25. Yagihashi N, Nishida N, Seo HG, Taniguchi N, Yagihashi S:11. Bredt DS, Hwang PM, Glatt CE, Lowenstein C, Reed RR,Expression of nitric oxide synthase in macula densa in streptozo-Snyder SH: Cloned and expressed nitric oxide synthase structurallytocin diabetic rats. Diabetologia 39:793–799, 1996resembles cytochrome P-450 reductase. Nature 351:714–718, 1991

26. Maree A, Peer G, Iaina A, Blum M, Wollman Y, Csernihovsky12. Ovadia H, Rosenmann H, Shezer E, Halimi M, Ofran I, GabizonT, Silverberg DS, Cabili S: Nitric oxide in streptozotocin-inducedR: Effect of scrapie infection on the activity of neuronal nitric-diabetes mellitus in rats. Clin Sci (Colch) 90:379–384, 1996oxide synthase in brain and neuroblastoma cells. J Biol Chem

27. Tolins JP, Shultz PJ, Raij L, Brown DM, Mauer SM: Abnormal271:16856–16861, 1996 renal hemodynamic response to reduced renal perfusion pressure13. Kihara M, Umemura S, Sugaya T, Toya Y, Yabana M, Kobayashi in diabetic rats: Role of NO. Am J Physiol 265:F886–F889, 1993

S, Tamura K, Kadota T, Kishida R, Murakami K, Fukamizu A, 28. Pieper GM: Enhanced, unaltered and impaired nitric oxide-medi-Ishii M: Expression of neuronal type nitric oxide synthase and ated endothelium-dependent relaxation in experimental diabetesrenin in the juxtaglomerular apparatus of angiotensin type-1a re- mellitus: Importance of disease duration. Diabetologia 42:204–213,ceptor gene-knockout mice. Kidney Int 53:1585–1593, 1998 1999

14. Cosentino F, Luscher TF: Endothelial dysfunction in diabetes 29. Craven PA, Studer RK, DeRubertis FR: Impaired nitric oxide-mellitus. J Cardiovasc Pharmacol 32(Suppl 3):S54–S61, 1998 dependent cyclic guanosine monophosphate generation in glomer-

15. Honing ML, Morrison PJ, Banga JD, Stroes ES, Rabelink TJ: uli from diabetic rats: Evidence for protein kinase C-mediatedNitric oxide availability in diabetes mellitus. Diabetes Metab Rev suppression of the cholinergic response. J Clin Invest 93:311–320,14:241–249, 1998 1994

30. Craven PA, Studer RK, DeRubertis FR: Impaired nitric oxide16. De Nicola L, Minutolo R, Bellizzi V, Andreucci M, La Verderelease by glomeruli from diabetic rats. Metabolism 44:695–698,A, Cianciaruso B: Enhancement of nitric oxide synthesis by1995l-arginine supplementation in renal disease: Is it good or bad?

31. Baylis C, Mitruka B, Deng A: Chronic blockade of nitric oxideMiner Electrolyte Metab 23:144–150, 1997synthesis in the rat produces systemic hypertension and glomerular17. Craven PA, Studer RK, Felder J, Phillips S, DeRubertis FR:damage. J Clin Invest 90:278–281, 1992Nitric oxide inhibition of transforming growth factor-beta and col-

32. Wang YX, Brooks DP, Edwards RM: Attenuated glomerularlagen synthesis in mesangial cells. Diabetes 46:671–681, 1997cGMP production and renal vasodilation in streptozotocin-induced18. Persson AEG, Ollerstam A, Salomonsson M, Thorup C: Hyper-diabetic rats. Am J Physiol 264(5 Pt 2):R952–R956, 1993tension and impaired renal angiotensin II response in rats after

33. Mattar A, Fujihara C, Ribeiro M, Denucci G, Zatz R: Renalchronic neuronal nitric oxide synthase inhibition. Kidney Int 54effects of acute and chronic nitric oxide inhibition in experimental(Suppl 67):S216–S217, 1998diabetes. Nephron 74:136–143, 199619. Arima S, Ren Y, Juncos LA, Ito S: Platelet-activating factor dilates 34. Ito S, Abe K: Contractile properties of afferent and efferent arteri-

efferent arterioles through glomerulus-derived nitric oxide. J Am oles. Clin Exp Pharmacol Physiol 24:532–535, 1997Soc Nephrol 7:90–96, 1996 35. Arendshorst WJ, Brannstrom K, Ruan X: Actions of angiotensin

20. Sugimoto H, Shikata K, Matsuda M, Kushiro M, Hayashi Y, II on the renal microvasculature. J Am Soc Nephrol 10:S149–S161,Hiragushi K, Wada J, Makino H: Increased expression of endo- 1999thelial cell nitric oxide synthase (ecNOS) in afferent and glomeru- 36. Wardle EN: How does hyperglycaemia predispose to diabeticlar endothelial cells is involved in glomerular hyperfiltration of nephropathy? Q J Med 89:943–951, 1996diabetic nephropathy. Diabetologia 41:1426–1434, 1998 37. Elger M, Sakai T, Kriz W: The vascular pole of the renal glomeru-

21. Wilcox CS, Welch WJ: Macula densa nitric oxide synthase: Ex- lus of rat. Adv Anat Embryol Cell Biol 139:1–98, 1998pression, regulation and function. Kidney Int 54(Suppl 67):S53– 38. Brands MW, Granger JP: Control of renal function and bloodS57, 1998 pressure by angiotensin II: Implication for diabetic glomerular

injury. Miner Electrolyte Metab 24:371–380, 199822. Pieper GM: Review of alterations in endothelial nitric oxide pro-