class differences in the effects of calcium channel ... · class differences in the effects of...

Kidney International, Vol. 55 (1999), pp. 1849–1860

VASCULAR BIOLOGY – HEMODYNAMICS – HYPERTENSION

Class differences in the effects of calcium channel blockersin the rat remnant kidney model

KAREN A. GRIFFIN, MARIA M. PICKEN, GEORGE L. BAKRIS, and ANIL K. BIDANI

Departments of Medicine and Pathology, Loyola University Medical Center and Hines Veterans’ Administration Hospital,Maywood, and the Department of Preventive Medicine, Rush-Presbyterian St. Luke’s Medical Center, Chicago, Illinois, USA

Conclusions. These class differences between CCBs in theirClass differences in the effects of calcium channel blockers inrelative impact on systemic BP profiles, renal autoregulation,the rat remnant kidney model.and glomerular pressure transmission may have clinically sig-Background. Controversy persists as to the existence of classnificant implications and may account for the variable glomeru-differences between calcium channel blockers (CCBs) in theirloprotection that has been observed with these agents in bothability to provide renoprotection and as to potential mecha-experimental models and in humans.nisms involved.

Methods. Rats with 5/6 renal ablation were left untreatedor received diltiazem, verapamil, or felodipine after the firstweek, and the relationship between continuous radiotelemet-

The rat remnant kidney (RK) model, produced byrically measured blood pressure (BP) and glomerulosclerosisright nephrectomy and segmental infarction of two thirds(GS) was assessed at seven weeks. Additionally, the effects of

these CCBs on renal autoregulation and hypertrophy were of the left kidney, has been extensively used to investi-examined at three weeks after renal ablation. gate the progressive nature of chronic human renal dis-

Results. Although an excellent linear correlation was ob- ease [1–8]. The initially normal remnant nephrons un-served between the average BP levels and GS in all groupsdergo compensatory increases in size and function, but(r 5 0.75 to 0.84, P , 0.01), significant protection was notwith time, these rats develop a syndrome of hyperten-achieved with any of the CCBs, but for different reasons. The

antihypertensive effects of diltiazem were not sustained beyond sion, proteinuria, and progressive glomerulosclerosisthe second week. Verapamil significantly reduced the average (GS), the pathogenesis of which remains controversialBP (144 6 4 mm Hg vs. 181 6 8 in untreated rats) but shifted [1–8]. Substantial evidence supports the concept that thethe slope of the relationship between BP and GS (increase in

compensatory vasodilation of the preglomerular vascula-percentage GS/mm Hg increase in average systolic BP) to theture and the associated impairment in renal autoregula-left (x intercept 121 vs. 144 mm Hg for untreated rats, P ,

0.01) so that GS was not reduced. Felodipine also significantly tion after an approximate 5/6 renal ablation permit anreduced the average BP (144 6 3 mm Hg) and shifted the exaggerated transmission of systemic pressure to the glo-slope to the left (x intercept 123 mm Hg), but additionally

merular capillaries, resulting in glomerular capillary hy-made the slope steeper (2.3 6 0.5 vs. 0.82 6 0.2 in untreatedpertension and injury [2, 3, 5, 8, 9]. Essentially identicalrats). Because of these differing effects on the relationship

between BP and GS, the rank order of GS for any given BP mechanisms have also been postulated to play a majorelevation was as follows: felodipine . verapamil . diltia- role in the progression of diabetic renal disease [9–15].zem 5 untreated. Felodipine, but not verapamil or diltiazem, The well-documented adverse effects of systemic hyper-caused additional impairment of the already impaired renal

tension in accelerating the progression of chronic renalautoregulation in untreated rats, thereby explaining its adversedisease regardless of etiology is consistent with such aeffects on GS. By contrast, the adverse effects of verapamil

on GS were attributable to the greater amplitude of BP fluctu- concept [16–20]. As a logical extension of this concept,ations that was observed in the verapamil-treated rats such the use of antihypertensive agents has become a primarythat for any given average BP, these rats were exposed to

therapeutic strategy to slow the rate of nephron loss.greater peak pressures as compared with the other groups.However, antihypertensives are not uniformly protec-None of the CCBs had a significant effect on glomerular hyper-

trophy. tive. Substantial controversy persists as to the success ofindividual antihypertensive agents other than angioten-sin-converting enzyme inhibitors in achieving such reno-Key words: glomerulosclerosis, nephrectomy, hypertrophy, blood pres-

sure, hypertension, telemetry, renoprotection. protection [21–25]. Evidence is particularly controversialwith calcium channel blockers (CCBs), with beneficial,Received for publication July 20, 1998neutral, and even adverse effects having been reportedand in revised form November 23, 1998

Accepted for publication December 29, 1998 [26–34]. Additionally, class differences between CCBshave been noted in some but not other studies [31–34]. 1999 by the International Society of Nephrology

1849

Griffin et al: CCBs and glomerulosclerosis1850

The reasons for such conflicting results with CCBs are and dark (18:00 to 6:00 hr) cycle. All rats received foodand water ad libitum throughout the study.not clear, but several possibilities exist.

First, an accurate assessment of the ambient blood pres-Radiotelemetry studiessure (BP) profiles is a critical prerequisite for a valid com-

parison of the relationship between the relative antihyper- The rats were anesthetized with sodium pentobarbital(45 mg/kg intraperitoneally), subjected to approximatelytensive and glomeruloprotective efficacies of different

regimens. Rapid, spontaneous, and large BP fluctuations 5/6 renal ablation (right nephrectomy and ligation of allbut one posterior extrarenal branch of the left renalhave been documented by chronic radio-telemetry in

conscious, unrestrained rats (and most other species), artery), and prepared for telemetric monitoring of BP(Data Sciences International, St. Paul, MN, USA) at theparticularly when hypertension is present [35–39]. Such

BP lability makes it extremely unlikely that the conven- time of the renal ablation surgery as previously described[30, 36, 37]. Each rat had a BP sensor (model TA11PA-tional tail-cuff measurements that have been used in

most of these investigations can either assess the overall C40; Data Sciences International) inserted intraperito-neally. The sensor’s catheter was inserted into the aortapressure burden with sufficient accuracy or can ensure

equivalent ambient systemic BP profiles to allow valid below the level of the renal arteries, and the radio fre-quency transmitter was fixed to the peritoneum. Thecomparisons between different regimens. Therefore,

some of the reported differences in glomeruloprotection rats were housed individually in plastic cages that wereplaced on top of the receiver. The signals from the pres-achieved by CCBs despite “equivalent” BP control may,

in fact, be due to BP differences that are not detected sure sensor were converted, temperature compensated,and sent via the radio frequency transmitter to the telem-by the conventional tail-cuff method [30, 37].

Second, CCBs can have adverse effects on renal auto- etry receiver. The receiver was connected to a BCM-100consolidation matrix, which transmitted the informationregulation, allowing a greater transmission of systemic

BP and its fluctuations to the renal microvasculature to the Dataquest IV acquisition system (Data SciencesInternational). Systolic BP (SBP) in each animal was[40–42]. Consistent with such an interpretation, we havecontinuously recorded at 10-minute intervals for approx-recently reported that the dihydropyridine CCB nifedi-imately seven weeks. Tail vein blood samples were ob-pine further impairs the already compromised autoregu-tained at three days for measurement of serum creatininelatory ability of RKs [30] and alters the relationship(SCr) as an index of the degree of renal mass reductionbetween BP and GS such that a greater degree of GS is[30]. At approximately seven days, the rats were ran-observed for any given BP in nifedipine-treated rats asdomly assigned to the untreated group or received diltia-compared with untreated animals. It is possible that classzem (750 to 1500 mg/liter of drinking water), verapamildifferences exist between CCBs in the severity of their(750 mg/liter of drinking water), or felodipine (0.1% byadverse effects on renal autoregulation, which may resultweight in standard rat chow). The felodipine chow wasin differences in glomeruloprotection. Finally, CCBskept covered with aluminum foil to prevent photodegra-have been reported to have significant growth effectsdation. In the case of diltiazem, the rats were initiallyin the RK model, and a reduction in the glomerularstarted on 750 mg/liter of drinking water. The amounthypertrophy response has been postulated to contributeof diltiazem was increased to 1000 to 1500 mg/liter into the glomeruloprotection provided by these agents in8 of 10 rats over the first few days in this group becausesome studies [29, 34, 43]. It is possible that the differentialof a poor initial antihypertensive response. After seveneffects of CCBs on GS in the RK model may reflectweeks, tail vein SCr and 24-hour urine collections fordifferences in their respective “antigrowth” effects onprotein excretion were obtained. The rats were thenglomerular hypertrophy.anesthetized with intravenous sodium pentobarbital (40These studies were undertaken to examine if classmg/kg). A tracheostomy was performed using polyethyl-differences exist between CCBs in their ability to protectene (PE-200) tubing, and the rats were surgically pre-against GS in the RK model and whether BP-dependentpared for clearance studies as described previously [30,or BP-independent mechanisms account for such differ-44]. In brief, a carotid artery was cannulated with PE-ences. The following classes of CCBs were investigated:50 tubing and connected to a Windograf (model 40-8474;the benzothiazepine (diltiazem), the phenylalkylamineGould Inc., Glen Burnie, MD, USA) for continuous re-(verapamil), and the dihydropyridine (felodipine).cording of mean arterial pressure (AP). A femoral veinwas cannulated with PE-50 tubing, and a priming dose

METHODS of inulin in 150 mm NaCl was administered, followedTwo sets of investigations were performed on male by a continuous maintenance infusion of 150 mm NaCl

Sprague-Dawley rats (body wt 200 to 300 g) that were containing inulin at 0.055 ml/min to maintain the plasmafed a standard protein diet (Purina, St. Louis, MO, USA) concentration of inulin at approximately 50 mg/dl and

for replacement of surgical and ongoing fluid losses. Theand synchronized to a 12:12 hour light (6:00 to 18:00 hr)

Griffin et al: CCBs and glomerulosclerosis 1851

left ureter was then cannulated with polyethylene tubing one kidney section for each animal was measured usinga digitizing pad, as described previously [30, 45]. Thefor collection of urine samples. A 1.0 mm R series flow

probe (Transonic Systems Inc., Ithaca, NY, USA) was mean glomerular volume (VG) was then calculated fromthe respective mean AG as VG 5 b/k (AG

3/2), where b 5placed around the left renal artery for measurement ofrenal blood flow (RBF) by a flowmeter (Transonic Sys- 1.38 is the size distribution coefficient and k 5 1.1 is the

shape coefficient for glomeruli idealized as spheres [46].tems Inc.), as described previously [30, 44]. At the con-clusion of the surgery, a 150 mm NaCl bolus equal to

Analyses, calculations, and statistics1% of the body weight was administered. Two 20-minuteclearances of inulin were obtained. Blood samples were Urinary protein was measured by the quantitative sul-

fosalicylic acid method, with human serum albumin serv-obtained at the midpoint of each urine collection. At theconclusion of these studies, the rats were killed, and the ing as the standard. Serum creatinine was measured using

a creatinine analyzer (Beckman Instruments Inc., Fuller-kidneys were harvested for morphological studies.ton, CA, USA) [30, 36, 37]. Inulin in urine and plasma

Autoregulatory studies filtrates was determined spectrophotometrically by thediphenylamine method, as described previously [3, 36,Rats assigned to the autoregulatory studies underwent

renal ablation without implantation of radiotransmitters 44]. GFR was calculated using standard formulae. RPPwas calculated as the mean AP 2 renal venous pressureand were followed for three weeks without radioteleme-

try. At approximately seven days, the rats were randomly (assigned a value of 5 mm Hg). Renal vascular resistancewas calculated as RPP (mm Hg)/RBF (ml/min/kg). Auto-assigned to receive either no treatment (U) or diltiazem,

verapamil, or felodipine as in the radiotelemetry studies. regulatory index (AI) was calculated by the method ofSemple and de Wardener [47] as follows: AI 5 [(RBF2 2The dose of diltiazem used was 1000 mg/liter of drinking

water for these studies. After three weeks, the rats were RBF1)/RBF1]/[RPP2 2 RPP1)/RPP1]. An AI of zero indi-cates perfect autoregulation, whereas an AI of one indi-anesthetized with intravenous sodium pentobarbital (40

mg/kg) and surgically prepared as described earlier here. cates that the vessels act as passive conduits for bloodflow [47]. All results are expressed as mean 6 se. Statisti-After the measurement of RBF and glomerular filtration

rate (GFR) at the ambient mean arterial pressures cal analysis was performed using analysis of variance,followed by Student–Newman–Keuls test or by Kriskall–(APs), RBF autoregulatory studies were performed us-

ing aortic miniclamps positioned above and below the Wallis nonparametric analysis of variance, followed byDunn’s multiple comparison test, as appropriate [48].left renal artery to raise or lower renal perfusion pressure

(RPP) as previously described [30, 44]. The RBF was Linear regression analysis was used to calculate theslopes and intercepts of the relationship between BP andallowed to stabilize for one to two minutes at each pres-

sure before RBF measurements were made. At the con- GS in each group. Analysis of covariance was used tocompare the slopes and intercepts between the groupsclusion of these studies, the rats were killed, and the

kidneys were harvested for morphological and morpho- using the Minitab Software package (Minitab Inc., StateCollege, PA, USA) [48]. A P value of less than 0.05 wasmetric studies.considered statistically significant.

Morphologic methods

Transverse sections of the kidney through the papillaRESULTS

were fixed in situ by perfusion for five minutes at theRadiotelemetry studiesmeasured BP with 1.25% glutaraldehyde in 0.1 m caco-

dylate buffer. Sections were cut at a thickness of 2 mm Table 1 shows that the initial body weight, SCr, and24-hour urine protein excretion for the four groups wereand were stained with hematoxylin and eosin and peri-

odic acid-Schiff. Sections were evaluated systematically not significantly different from each other. Similarly, theSCr at three days was not significantly different betweenin each kidney for glomerular injury (segmental sclero-

sis) in a blinded fashion by standard morphologic meth- the groups, indicating comparable renal mass reductionin all groups. Although the diltiazem- and felodipine-ods [3, 30, 36, 37]. At least 100 glomeruli in each animal

were evaluated, and the severity of glomerulosclerosis treated rats tended to gain less weight as a group, becauseof the large interanimal variability in each group, neitherwas expressed as the percentage of glomeruli exhibiting

such injury. the difference in weight gain nor the final body weightwere statistically significant. Similarly, the kidney weights

Morphometric methods at seven weeks were also not significantly different be-tween the four groups. Measurements of GFR, RBF,Glomerular volume was measured by area perimeter

analysis (Bioquant System IV software; R&M Biomet- and RVR during the clearance studies also failed toreveal any significant differences.rics Inc., Nashville, TN, USA). The cross-sectional area

(AG) of 75 consecutive glomerular profiles contained in Figure 1 illustrates the course of the 24-hour averages


Table 1. Functional data for the radiotelemetry studies

Initial Final z7 weeks

Body wt D wt GFR RBFBody wt SCr Proteinuria 3 days SCr AP RVR

N g mg/dl mg/24 hrs mg/dl g mm Hg mm Hg/(ml /min /kg) ml /min /kg

Untreated 11 25665 0.3 60.04 3 60.3 0.960.05 427623 171 622 153 610 6.961.3 3.060.4 25 63Diltiazem 10 26166 0.3 60.03 3 60.3 0.860.08 384631 123 626 168 610 7.160.8 2.560.5 25 63Verapamil 12 25164 0.3 60.03 4 60.9 0.760.04 430615 179 616 127 69a 4.860.6 2.660.4 28 62Felodipine 11 25066 0.3 60.02 3 60.5 0.760.08 391616 141 616 116 68a 5.761.3 2.360.5 24 63

a P , 0.05 compared to untreated or diltiazem-treated rats

Fig. 2. Overall average blood pressure (BP) during the first seven daysand during the subsequent approximate six weeks in 5/6 renal ablatedrats that after the seventh day were left untreated (h; N 5 11), or

Fig. 1. Course of systolic blood pressure (BP; 24-hr averages) over received diltiazem (750 to 1500 mg/liter of drinking water; ; N 5 10),approximately seven weeks in the four groups. All rats underwent a verapamil (750 mg/liter of drinking water; ; N 5 12), or felodipine5/6 renal ablation (right nephrectomy 1 infarction of approximately (0.1% by weight in chow; j; N 5 11). BP was radiotelemetricallytwo thirds of the left kidney). After 7 days, the rats were left untreated recorded continuously at 10-minute intervals, starting at 6 p.m. on the(h; N 5 11) or received diltiazem (750 to 1500 mg/liter of drinking day of renal ablation surgery to sacrifice approximately seven weekswater; j; N 5 10), verapamil (750 mg/liter of drinking water; .; N 5 later. *P , 0.01 vs. first week, dP , 0.01 vs. untreated and diltiazem.12), felodipine mixed in diet (0.1%; ★; N 5 11) by weight in chow.Blood pressure (BP) was radiotelemetrically recorded continuously at10-minute intervals, starting at 6 p.m. on the day of renal ablationsurgery to sacrifice approximately seven weeks later.

of the average SBP during the first week before theinitiation of treatment and during the following approxi-mately six weeks (Fig. 2). The average SBP during thefirst week was not significantly different between theof SBP after approximate 5/6 renal ablation in untreated

animals and in the three groups treated with CCBs. For groups. There was no significant change in the averageSBP in the untreated and diltiazem-treated rats duringthe first seven days after renal ablation and before the

initiation of antihypertensive therapy, the BP was similar the subsequent six weeks. By contrast, both verapamiland felodipine caused a significant reduction in the aver-in the four groups. There was an initial reduction in BP

with all three treatment regimens. However, the antihy- age SBP during the final approximate six weeks as com-pared with their respective BP during the first week.pertensive effect of diltiazem was modest and transient.

By the end of the third week after renal ablation (the Additionally, the average SBP of verapamil- and felodip-ine-treated animals was significantly lower than that ofsecond week after the initiation of therapy), no differ-

ences could be observed in the systolic BP of the diltia- the untreated and diltiazem-treated animals during thefinal six weeks, but was not significantly different fromzem-treated group from that of the untreated animals.

By contrast, BP was significantly lower throughout the each other. Figure 3 A and D illustrate the course oftelemetric SBP recordings at 10-minute intervals in a ratcourse in the verapamil- and felodipine-treated animals,

although some decrease in the antihypertensive effec- from each of the four groups. Figure 3A graphicallydisplays the spontaneous fluctuations of SBP in an un-tiveness of these agents was also observed over the time

course of the studies. The differences between the four treated rat. The BP fluctuation pattern was not alteredby diltiazem or felodipine (Fig. 3 B and D); however,groups are summarized and illustrated by a comparison


Fig. 3. Illustrations of the course of systolicblood pressure (BP) recorded every 10 min-utes for approximately seven weeks in a ratwith an approximate 5/6 renal ablation fromeach of the four groups: (A) untreated, (B)diltiazem, (C) verapamil, and (D) felodipine.

verapamil-treated rats exhibited an exaggerated ampli- This difference is quantitatively shown in Figure 4C, inwhich the standard deviations are plotted as a percentagetude of BP fluctuations, as illustrated in Figure 3C. This

difference in the pattern of BP fluctuations is more of the average SBP. As can be noted, this parameterwas significantly increased by verapamil treatment.clearly illustrated by the decompressed 24-hour segments

of BP recordings of the rats treated with verapamil (Fig. The protein excretion rate (mg/24 hr) and the percent-age of glomeruli exhibiting GS in the RKs of these four4A) and felodipine (Fig. 4B). As can be noted, the vera-

pamil-treated rats exhibited greater peak increases in groups after approximately seven weeks of radioteleme-try are presented in Figure 5. There were no statisticallySBP occurring cyclically at one- to three-hour intervals.


Fig. 3. Continued.

significant differences in proteinuria and GS between the average SBP during the final approximate six weeks wascompared between the four groups, significant differ-four groups, although both proteinuria and GS tended to

be the greatest in felodipine-treated rats. A direct and ences were observed. The slope of this relationship be-tween GS and SBP in untreated rats, that is, the increasestrong relationship between the average SBP and GS

was observed in each of the individual four groups, with in percentage of GS/mm Hg increase in average SBP(0.82 6 0.2), was not significantly altered by diltiazemcorrelation coefficients ranging between 0.75 and 0.84

(Fig. 6). However, when the relationship of GS to the (0.81 6 0.2). Verapamil treatment did not significantly


Fig. 4. (A) A decompressed 24-hour segment from the blood pressure(BP) recording shown in Figure 3C in a rat treated with verapamil.The mean SBP 6 sd for the 24 hours was 124.4 6 14.2 mm Hg, with16% of the 144 readings being more than 140 mm Hg. (B) A similardecompressed segment from the BP recording shown in Figure 3D ina rat treated with felodipine. The mean SBP 6 sd for the 24 hours was124.3 6 5.2 mm Hg, with none of the BP readings being more than 140mm Hg. (C) The standard deviations expressed as a percentage of theoverall average SBP, an index of the amplitude of BP fluctuations areshown for each of the four groups during the first week (before therapy)and during the subsequent six weeks (after therapy). Symbols are: (h)N 5 11 untreated; ( ) N 5 10 diltiazem; ( ) N 5 12 verapamil; (j)N 5 11 felodipine.

alter the slope per se (1.1 6 0.3), but did result in asignificant shift to the left as indicated by the significantlydifferent x intercept for the verapamil group (121 mmHg) as compared with the untreated rats (143 mm Hg,P , 0.01). By contrast, felodipine not only caused asignificant shift to the left of the slope (x intercept 123mm Hg, P , 0.01 versus untreated) but additionallycaused it to be much steeper (2.3 6 0.5 increase in per-centage GS/mm Hg increase in average SBP, P , 0.01vs. each of the other three groups). These changes in theslopes of the relationship between BP and GS resulted inGS, which was greatest in the felodipine-treated rats, andlowest in the diltiazem-treated rats, with the verapamil-treated rats exhibiting an intermediate degree of GS forany given BP elevation.

Fig. 5. Proteinuria and percentage of glomerular injury at the end ofapproximately 7 weeks in the four groups of rats with approximate 5/6 Autoregulatory studiesrenal ablation who were left untreated (h; N 5 11), received diltiazem

No significant differences were present for the initial( ; N 5 10), verapamil ( ; N 5 12), or felodipine (j; N 5 11) afterthe first week. body weight, SCr, and 24-hour protein excretion rates in


Fig. 6. (A) Correlations of the percentage of glomeruli with sclerosis at approximately 7 weeks in individual rats with ablation from all fourgroups with their average systolic BP during the final six weeks (the mean of all approximately 6000 BP readings obtained radiotelemetricallyevery 10 minutes after the first week). After the first seven days, the rats had received no treatment, diltiazem, verapamil, or felodipine. (B) Linearregression analysis of the slopes of the relationship between the average BP during the final six weeks and percentage GS for each of the fourgroups of rats with 5/6 renal ablation who received no treatment, diltiazem, verapamil or felodipine after approximately seven days. Symbols are:(h) untreated (N 5 11; r 5 0.82, slope 0.82 6 0.2, x intercept 143 mm Hg); (j) diltiazem (N 5 10; r 5 0.81 slope 0.86 6 0.2, x intercept 137 mmHg); (.) verapamil (N 5 12; r 5 0.75, slope 1.08 6 0.3, x intercept 121 mm Hg); (★) felodipine (N 5 11; r 5 0.84, slope 2.3 6 0.5, x intercept123 mm Hg). P for the correlation coefficient for each group , 0.002. The slope for the felodipine group was significantly different from the slopesof all other groups (P , 0.01). The intercepts for the verapamil-treated and felodipine-treated groups were significantly different than that of theuntreated or diltiazem-treated groups (P , 0.01). There were no significant differences between the untreated and diltiazem-treated groups.

Table 2. Functional and structural data for the autoregulatory studies

Baseline Final z3 weeks

RBF GFRBody wt SCr 3 days SCr Body wt AP RVR Kidney wt Glom vol

N g mg/dl mg/dl g mm Hg mm Hg/(ml /min /kg) g lm23 31026ml /min /kg

Untreated 12 24867 0.4 60.03 0.8 60.06 315615 168 68 9.361.7 23.263.3 2.6 60.2 2.4960.14 2.060.16Diltiazem 11 24565 0.4 60.03 0.8 60.06 290611 157 65 5.660.7a 30.963.3 2.6 60.3 2.3160.15 1.8460.16Verapamil 11 24464 0.3 60.03 0.8 60.05 265615 135 67a 5.360.5a 26.061.9 3.1 60.3 2.2660.21 1.8360.12Felodipine 12 24864 0.3 60.02 0.9 60.07 273611 143 67a 6.360.8a 24.662.1 2.5 60.4 2.3260.08 1.9760.14

a P , 0.05 compared to untreated or diltiazem treated rats

the four groups or rats that underwent autoregulatory shown in Figure 7. All four groups demonstrated signifi-cant impairment in RBF autoregulation (autoregulatorystudies (Table 2). Serum creatinine at three days was

also not significantly different between the four groups, indices of more than 0.4). The autoregulatory ability wasimpaired to a similar degree in control, diltiazem-, andindicating comparable renal mass reduction as in the

radiotelemetry studies. Table 2 also shows that the final verapamil-treated rats, as indicated by the lack of sig-nificant differences in the autoregulatory indices forbody weight, RK weight, and glomerular volume at three

weeks were not significantly different between the four these groups. By contrast, felodipine administration re-sulted in significant additional impairment in RBF auto-groups. Differences were also not observed in the RBF

and GFR measurements performed at ambient BP be- regulatory responses, as indicated by the significantlyhigher autoregulatory indices in this group as comparedfore autoregulatory studies. However, in contrast to the

RVR values at seven weeks, RVR was significantly lower with the other three groups (P , 0.01). In fact, theautoregulatory indices of approximately one observedat three weeks in all of the CCB-treated groups as com-

pared with untreated rats. The difference in RVR be- in the felodipine group indicate a complete loss of auto-regulatory ability and a resultant passive behavior oftween individual CCB-treated groups was not statisti-

cally significant. resistance vessels in this group. Calculation of RVRchanges during the autoregulatory studies confirmed thisThe results of the RBF autoregulation studies are


on the systemic BP and its relationship to GS. We havepreviously demonstrated that GS in untreated individualrats with 5/6 renal ablation exhibits an exceedingly closelinear relationship to their radiotelemetrically measuredaverage SBP [36, 37]. We have therefore suggested thatthe impact of an antihypertensive agent on the slope ofthis relationship between BP and GS may provide amore valid method to separate the agent’s BP-dependentand BP-independent effects on GS [30, 37]. If a givenantihypertensive agent reduces both BP and GS but doesnot significantly alter the slope of the relationship be-tween BP and GS from that observed in untreated ani-mals, the glomeruloprotection can be ascribed primarilyto its BP-dependent (antihypertensive) effects [37]. BP-independent protection by an antihypertensive agentwould be expected to reduce the slope and/or shift it to

Fig. 7. Renal blood flow (RBF) autoregulatory studies in the four the right, while BP-independent deleterious effects ongroups of rats that underwent an approximate 5/6 renal ablation three GS would be manifested by a shift of the slope to theweeks before. After day 7, the rats were left untreated (h), received

left or the slope becoming more steep [30].diltiazem (j), verapamil (.), or felodipine (★). Graded changes inrenal perfusion pressure (RPP) were produced by aortic miniclamp and Our studies provide further support for this concept.resulted in significant changes in RBF with every change in RPP in all There was an excellent correlation between BP and GSgroups (P , 0.01, repeated means analysis of variance). The bottom

in each of the experimental groups (r 5 0.75 to 0.84),of the figure shows the calculated autoregulatory indices (AI) for thechanges in RBF for perfusion pressure changes between 150 and 100 indicating the marked BP dependence of glomerular in-mm Hg. An AI of 0 indicates perfect autoregulation, whereas an AI of jury in this model, even in treated rats. However, signifi-approximately 1.0 indicates resistance vessels acting as passive conduits

cant class differences were observed between the threewithout autoregulatory responses. The AI for RPP changes for the twochanges in RPP were not significantly different from each other; therefore, CCBs in the dosages used in their effect on the systemica single AI for the RBF change between (150 to 100 mm Hg) is pre- BP, as well as on the relationship between BP and GS.sented. The AI was significantly greater in felodipine-treated rats com-

Because of the very transient antihypertensive effects ofpared with each of the other three groups (* P , 0.05 maximum).diltiazem, the overall pressure burden was not signifi-cantly reduced, despite dosages that were twofold tofourfold greater than those reported to be antihyperten-

finding, with the felodipine-treated rats exhibiting no sive in other rat models [49, 50] and tended to reducegrowth. This relative antihypertensive ineffectivenesssignificant change in RVR during the change in pressuremay reflect differences in the pathogenesis of systemicfrom 140 to 100 mm Hg (the mean RVR at approximatelyhypertension between these models. Additionally, diltia-100 mm Hg was 94% of the value at 140 mm Hg). Byzem also failed to alter the relationship between BPcontrast, a significant fall in RVR was observed in theand GS, indicating a lack of significant BP-independentother three groups, with the mean RVR values at approx-effects of this agent in this model. By contrast, in theimately 100 mm Hg, averaging 81, 78, and 82% in thedosages used, both verapamil and felodipine resulted inuntreated, diltiazem-treated and verapamil-treated rats,significant and comparable reduction of the daily averagerespectively.SBP. Despite this antihypertensive effectiveness, bothverapamil and felodipine failed to provide significant

DISCUSSION glomeruloprotection because of their concurrent adverseA large body of experimental and clinical evidence effects on the slope of the relationship between average

supports the importance of BP control in slowing the BP and GS. In the case of verapamil, the slope per seseemingly inexorable progression of both diabetic and was not changed, but it was shifted to the left, indicatingnondiabetic renal disease [4, 9, 16–20]. Despite their not only that a greater degree of GS was observed atantihypertensive effectiveness and wide clinical usage, any given average SBP, but also that BP would need tothe renoprotective ability of individual CCBs remains be reduced to lower levels to provide complete glomeru-controversial [23–34]. These studies show that none of loprotection in verapamil-treated animals. By contrast,the three major classes of CCBs were able to provide felodipine not only caused a similar shift to the left ofglomeruloprotection in the 5/6 renal ablation model of the slope of the relationship between BP and GS, butprogressive GS. However, the mechanisms responsible additionally made it significantly steeper. Consequently,for this failure were different for each of the three CCBs a given increase in BP resulted in a disproportionally

greater increase in GS in felodipine-treated animals asand stemmed from the class differences in their effects


compared with all other groups, including verapamil. effects of felodipine on renal autoregulation [57–60].Given the spontaneous, large, and frequent BP fluctua-These similar, yet different, adverse effects of verapamil

and felodipine on the relationship between BP and GS tions that are observed in these animals in the unanesthe-tized state, the efficiency of the renal autoregulatorysuggest that both CCBs had BP-independent deleterious

effects on GS that were likely mediated through separate responses is likely to be the primary determinant of thedegree to which such fluctuations are transmitted to glo-mechanisms.

Among the potential BP-independent mechanisms merular capillaries [30, 38, 39]. Therefore, a completeloss of this protective mechanism, as observed in thethat have been postulated to be pathogenetically signifi-

cant in the RK model, the glomerular hypertrophy felodipine-treated animals, is expected to greatly exag-gerate the glomerular capillary exposure to systemic hy-(growth) response has been suggested by several investi-

gators to be of major importance [6, 7, 29, 51, 52]. The pertension with resultant increased GS, as was observed.In contrast to felodipine, the adverse effects of vera-precise cellular mechanisms have not yet been defini-

tively established, but both an increase in glomerular pamil on the relationship between BP and GS are notexplained by additional deleterious effects on renal auto-capillary wall tension through the Laplace Law effects

(tension 5 pressure 3 radius) and an increase in the regulation. It is possible that the shift in the slope to theleft of the relationship between average BP and GS isactivity of pathogenetic growth factors associated with

the glomerular hypertrophy response have been postu- a consequence of the fact that animals treated with vera-pamil exhibit a greater amplitude of BP fluctuations aslated [6, 7, 51–53]. The present data do not show that

differences in glomerular hypertrophy play a significant compared with the untreated and felodipine-treated rats.Therefore, for any given average BP, the glomerularrole in CCB-treated animals, as none of the CCBs had

a significant effect on glomerular hypertrophy. capillaries in verapamil-treated rats may have been ex-posed to higher peak pressures. If valid, such an interpre-Differential effects on the relationship between BP

and GS may also result from differential effects of antihy- tation would suggest that differences in BP fluctuationpatterns in addition to the average SBP may need topertensive agents on renal autoregulation. Normally, the

autoregulatory vasoconstrictor responses of the preglo- be taken into account when comparing the effects ofantihypertensive agents on glomerular injury. To ourmerular vasculature provide the primary protection

against increases in systemic pressure (episodic or sus- knowledge, such a change in BP fluctuation patterns byverapamil or other antihypertensive agents has not beentained) from being transmitted to glomerular capillaries

[3, 8, 16, 38, 39]. These autoregulatory responses have described previously. The mechanism(s) responsible re-mains to be identified. Verapamil is known to have addi-been demonstrated to be impaired in RKs, although

the precise mechanisms remain undefined [3, 8]. Normal tional effects possibly unrelated to calcium channelblockade, such as an alteration of sympathetic activity,autoregulatory vasoconstriction in response to pressure

increases is believed to be mediated by calcium entry which may have contributed to the change in the BPfluctuation patterns [58, 61]. Whether such an effect oc-through voltage-gated, “L”-type calcium channels as a

result of stretch (pressure)-induced depolarization [54– curs in other species (humans) is not known.Although the verapamil preparation used in these56]. Consistent with such a mechanism, acute administra-

tion of all three classes of CCBs has been shown to studies was not a long-acting formulation, the short-act-ing nature of the preparation is unlikely to account forimpair renal autoregulation in intact kidneys of both

normotensive and hypertensive rats [40–42]. The reason the observed frequency pattern of BP fluctuations (one-to three-hr cycles throughout the day). In this context,that felodipine, but not diltiazem or verapamil during

chronic administration, had an additional deleterious im- it is of note that the effects of felodipine, a long-actingdihydropyridine CCB, on BP fluctuation patterns, renalpact on the already impaired renal autoregulation in

rats with RKs remains to be defined. Although all three autoregulation, and GS are very similar to that pre-viously observed by us with nifedipine, a short-actingclasses of CCBs exert their effects through blocking the

“L”-type, voltage-gated channels, differences in binding dihydropyridine CCB [30]. These data, therefore, do notsupport the suggestion that the potentially deleteriousaffinity have been demonstrated, depending on the rela-

tive state of activity of the calcium channel [57–60]. renal effects of CCB may only be particular to the short-acting CCB preparations, as has been suggested for ad-Moreover, although all three classes of CCBs bind to

the a1 subunit of the “L”-type calcium channel, they bind verse cardiovascular events [62–65]. Although it is alsonot yet clear if the potentially adverse renal and cardiacto different and discrete sites of the channel protein,

with probably specific structure–function relationships effects are both mediated by similar mechanisms, it is ofinterest that the recent concerns about possible adverse[57–60]. The dihydropyridine CCBs bind to a site that

is very proximate to the putative transmembrane voltage cardiovascular effects of CCBs have also primarily beendirected at the dihydropyridine CCBs.sensor segment (S4) of the a1 subunit of the calcium

channel, and this may account for the greater adverse In summary, our data demonstrate that none of three


capillary hydrostatic pressure in the rat remnant nephron. J Clinmajor classes of CCBs are successful in providing glom-Invest 88:101–105, 1991

eruloprotection in the 5/6 renal ablation model. How- 9. Neuringer JR, Brenner BM: Hemodynamic theory of progressiveever, the reasons for the lack of renoprotection differed renal disease: A 10-year update in brief review. Am J Kidney Dis

22:98–104, 1993between the three groups and may reflect significant class10. Hostetter TH, Rennke HG, Brenner BM: The case for intrarenaldifferences between the CCBs. The mechanisms ranged hypertension in the initiation and progression of diabetic and other

from a lack of antihypertensive effectiveness (benzothia- glomerulopathies. Am J Med 72:375–380, 198211. Zatz R, Meyer TW, Rennke HG, Brenner BM: Predominancezepine, diltiazem), a change in BP fluctuation patterns

of hemodynamic rather than metabolic factors in the pathogenesis(phenylalkylamine, verapamil), to a complete abolition of diabetic glomerulopathy. Proc Natl Acad Sci USA 82:5963–5967,of protective renal autoregulation (dihydropyridine, fel- 1985

12. Anderson S, Rennke HG, Garcia DL, Brenner BM: Short andodipine). The nature of these differences, however, sug-long term effects of antihypertensive therapy in the diabetic rat.gests that it is possible that in other models or disease Kidney Int 36:526–536, 1989

states or with different dosages, different effects on GS 13. Hayashi K, Epstein M, Routzenhiser R, Forster H: Impairedmyogenic responsiveness of the afferent arteriole in streptozotocin-may be seen and thus account for the contradictory re-induced diabetic rats: Role of eicosanoid derangements. J Am Socsults that have been reported with the different CCBs.Nephrol 2:1578–1586, 1992

For instance, these results do not exclude the possibility 14. Ohishi K, Okwueze MI, Vari RC, Carmines PK: Juxtamedullarymicrovascular dysfunction during the hyperfiltration stage of diabe-that if BP is reduced to the low normotensive range withtes mellitus. Am J Physiol 267:F99–F105, 1994any of the three classes of CCBs, GS may be ameliorated.

15. Vallon V, Blantz RC, Thomson S: Homeostatic efficiency ofHowever, with comparable but suboptimal BP reduction tubuloglomerular feedback is reduced in established diabetes mel-

litus in rats. Am J Physiol 269:F876–F883, 1995into an intermediate zone, the rank order of GS is ex-16. Baldwin DS, Neugarten J: Hypertension and renal disease. Ampected to be dihydropyridines . phenylalkylamines .

J Kidney Dis 10:186–191, 1987benzothiazepine. These data additionally emphasize the 17. Walker WG: Hypertension-related renal injury: A major contribu-

tor to end-stage renal disease. Am J Kidney Dis 22:164–173, 1993complex and dynamic nature of the relationship between18. Rossing P, Hommel E, Smidt UM, Parving HH: Impact of arterialsystemic BP and glomerular injury, and the inherent limi-

blood pressure and albuminuria on the progression of diabetictations of conventional BP measurements in investigating nephropathy in IDDM patients. Diabetes 42:715–719, 1993the role of hypertensive mechanisms in the pathogenesis 19. Klahr S, Levey AS, Beck GJ, Caggiula AW, Hunsicker, Kusek

JW, Striker G, for the Modification of Diet in Renal Diseaseof GS or in the evaluation of the relative effectivenessStudy Group: The effects of dietary protein restriction and blood-of antihypertensive agents in providing renoprotection. pressure control on the progression of chronic renal disease. N EnglJ Med 330:877–888, 1994

20. Klag MJ, Whelton PK, Randall BL, Neaton JD, Brancati FL,ACKNOWLEDGMENTSFord CE, Shulman NB, Stamler J: Blood pressure and end-stage

This research was supported by the National Institutes of Health renal disease in men. N Engl J Med 334:13–18, 1996grant DK-40426. This work was presented in part at the 30th Annual 21. Anderson S, Rennke HG, Brenner BM: Therapeutic advantageMeeting of the American Society of Nephrology in New Orleans, LA, of converting enzyme inhibitors in arresting progressive renal dis-USA, on 5 November, 1996, and was published in abstract form (J Am ease associated with systemic hypertension in the rat. J Clin InvestSoc Nephrol 7:1586, 1996). The authors thank Ms. Jennifer Quinn, 77:1993–2000, 1986Ms. Shilpa Gude, and Ms. Wanda Plott for technical assistance, and 22. Yoshida Y, Kawamaura T, Ikoma M, Fogo A, Ichikawa I: EffectsMs. Martha Prado for secretarial assistance. of antihypertensive drugs on glomerular morphology. Kidney Int

36:626–635, 1989Reprint requests to Karen A. Griffin, M.D., Loyola University Medi- 23. Anderson S, Rennke HG, Brenner BM with the technical assis-

cal Center, Section of Renal Disease and Hypertension, 2160 South First tance of Zayas MA, Lafferty HM, Troy JL, Sandstrom DJ:Avenue, Maywood, Illinois 60153, USA. Nifedipine versus fosinopril in uninephrectomized diabetic rats.

Kidney Int 41:891–897, 199224. Kasiske BL, Kalil RSN, Ma JZ, Liao M, Keane WF: Effect ofREFERENCES antihypertensive therapy on the kidney in patients with diabetes:

A meta-regression analysis. Ann Intern Med 118:129–138, 19931. Purkerson ML, Hoffsten PE, Klahr S: Pathogenesis of the glom-25. Lewis EJ, Hunsicker LG, Bain RP, Rohde RR: The effect oferulopathy associated with renal infarction in rats. Kidney Int

angiotensin-converting-enzyme inhibition on diabetic nephropa-9:407–417, 1976thy. N Engl J Med 329:1456–1462, 19932. Brenner BM: Nephron adaptation to renal injury or ablation. Am

26. Harris DC, Hammond WS, Burke TJ, Schrier RW: VerapamilJ Physiol 249:F324–F337, 1985protects against progression of experimental chronic renal disease.3. Bidani AK, Schwartz MM, Lewis EJ: Renal autoregulation andKidney Int 31:41–46, 1987vulnerability to hypertensive injury in remnant kidney. Am J Phys-

27. Jackson B, Johnson CI: The contribution of systemic hypertensioniol 252:F1003–F1010, 1987to progression of chronic renal failure in the rat remnant kidneys:4. Klahr S, Schreiner G, Ichikawa I: The progression of renalEffect of treatment with an angiotensin converting enzyme inhibi-disease. N Engl J Med 318:1657–1666, 1988tor or a calcium inhibitor. J Hypertens 6:495–501, 19885. Olson JL, Heptinstal RH: Non-immunologic mechanisms of glo-

28. Brunner FP, Thiel G, Hermle M, Bock A, Mihatsch MJ: Long-merular injury. Lab Invest 59:564–578, 1988term enalapril and verapamil in rats with reduced renal mass.6. Yoshida Y, Fogo A, Ichikawa I: Glomerular hemodynamicKidney Int 36:969–977, 1989changes vs. hypertrophy in experimental glomerular sclerosis. Kid-

29. Kloke HJ, Branten AJ, Hyysmans FT, Wetzels JF: Antihyper-ney Int 35:654–660, 1989tensive treatment of patients with proteinuric renal diseases: Risks7. Fogo A, Ichikawa I: Evidence of central role of glomerular growthor benefits of calcium channel blockers? Kidney Int 53:1559–1573,promoters in the development of sclerosis. Semin Nephrol 9:329–1998342, 1990

8. Pelayo JC, Westoo JY: Impaired autoregulation of glomerular 30. Griffin KA, Picken MM, Bidani AK: Deleterious effects on cal-


cium channel blockade on pressure transmission and glomerular Practical Approach Series, edited by Rickwood D, Hames BD,New York, OIRL Press at Oxford University Press Inc., 1993injury in rat remnant kidneys. J Clin Invest 96:798–800, 1995

31. Epstein M: Calcium antagonists and diabetic nephropathy. Arch 49. Sato T, Nara Y, Note S, Yamori Y: Effect of calcium antagonistson hypertension and diabetes in new hypertensive diabetic models.Intern Med 151:2361–2364, 1991

32. Demaric BK, Bakris GL: Effects of different classes of calcium J Cardiovasc Pharmacol 10(Suppl 10):S192–S194, 198750. Grellet JP, Bonoron-Adele SM, Tariosse LJ, Besse PJ: Diltia-antagonists on proteinuria in diabetic subjects. Ann Intern Med

113:987–988, 1990 zem and left ventricular hypertrophy in renovascular hypertensiverats. Hypertension 11:495–501, 198833. Bakris GL: Hypertension in diabetic patients, an overview of

interventional studies to preserve renal function. Am J Hypertens 51. Daniels BS, Hostetter TH: Adverse effects of growth in theglomerular microcirculation. Am J Physiol 258:F1409–F1416, 19906:1405–1475, 1993

34. ter Wee PM, DeMicheli AG, Epstein M: Effects of calcium antag- 52. Striker GE, He CJ, Liu ZH, Zalups RK, Esposito C, StrikerLJ: Biology of disease: Pathogenesis of nonimmune glomeruloscle-onists on renal hemodynamics and progression of nondiabetic

chronic renal disease. Arch Intern Med 154:1185–1202, 1994 rosis. Studies in animals and potential applications to humans. LabInvest 73:596–605, 199535. Parati G, Omboni S, Dirienzo M, Frattola A, Albini F, Mancia

G: Twenty-four hour blood pressure variability: Clinical implica- 53. Wolf AV: Demonstration concerning pressure-tension relationsin various organs. Science 115:243–244, 1952tions. Kidney Int 41:S24–S28, 1992

36. Bidani AK, Griffin KA, Picken M, Lansky DM: Continuous 54. Cohen AJ, Fray JCS: Calcium ion dependence of myogenic renalplasma flow autoregulation: Evidence from the isolated perfusedtelemetric blood pressure monitoring and glomerular injury in the

rat remnant kidney model. Am J Physiol 265:F391–F398, 1993 rat kidney. J Physiol (Lond) 330:449–460, 198255. Hayashi K, Epstein M, Loutzenhiser R: Pressure-induced vaso-37. Griffin K, Picken M, Bidani A: Radiotelemetric BP monitoring,

antihypertensives and glomeruloprotection in remnant kidney constriction of renal microvessels in normotensive rats. Circ Res65:1475–1484, 1989model. Kidney Int 46:1010–1018, 1994

38. Holstein-Rathlou NH, Marsh DJ: Renal blood flow regulation 56. Kirber MT, Walsh JV Jr, Singer JJ: Stretch-activated ion channelsin smooth muscle: A mechanism for the initiation of stretch-and arterial pressure fluctuations: A case study in nonlinear dynam-

ics. Physiol Rev 74:637–681, 1994 induced contraction. Pflugers Arch 412:339–345, 198857. McDonald TF, Pelzer S, Trautwein W, Pelzer DJ: Regulation39. Holstein-Rathlou NH, He J, Wagner AJ, Marsh DJ: Patterns

of blood pressure variability in normotensive and hypertensive and modulation of calcium channels in cardiac, skeletal, andsmooth muscle cells. Physiol Rev 74:365–507, 1994rats. Am J Physiol 269:R1230–R1239, 1995

40. Navar LG, Champion WJ, Thomas CE: Effects of calcium channel 58. Triggle DJ: Biochemical and pharmacologic differences amongcalcium channel antagonists: Clinical implications in calcium antag-blockade on renal vascular resistance responses to changes in per-

fusion pressure and angiotensin-converting enzyme inhibition in onists, in Clinical Medicine, edited by Epstein M, Philadelphia,Hanely & Belfus, 1992, pp 1–27dogs. Circ Res 58:874–881, 1986

41. Loutzenhiser R, Epstein M: Effects of calcium antagonists on 59. Campbell KP, Leung AT, Sharp AH: The biochemistry and mo-lecular biology of the dihydropyridine-sensitive calcium channel.renal hemodynamics. Am J Physiol 249:F619–F629, 1985

42. Carmines PK, Mitchell KD, Navar LG: Effects of calcium antag- Trends Neurosci 11:425–430, 198860. Catterall WA: Structure and function of voltage-gated ion chan-onists on renal hemodynamic and glomerular function. Kidney Int

41(Suppl 36):S43–S48, 1992 nels. Trends Neurosci 16:500–506, 199361. Motulski HJ, Snavely MD, Hughes RJ, Insel PA: Interaction43. Haller H: Calcium antagonists and cellular mechanisms of glo-

merulosclerosis and atherosclerosis. Am J Kidney Dis 21(Suppl of verapamil and other calcium channel blockers with a1- and a2-adrenergic receptors. Circ Res 52:226–231, 19833):26–31, 1993

44. Griffin KA, Bidani AK, Ouyang J, Ellis V, Churchill M, 62. Psaty BM, Heckbert SR, Koepsell TD, Siscovick DS, Raghuna-than TE, Weiss NS, Rosendaal FR, Lemaitre RN, Smith NL,Churchill PC: Role of endothelium-derived nitric oxide in hemo-

dynamic adaptations after graded renal mass reduction. Am J Phys- Wahl PW, Wagner EH, Furberg CD: The risk of myocardialinfarction associated with antihypertensive drug therapies. JAMAiol 264:R1254–R1259, 1993

45. Griffin KA, Picken M, Bidani AK: Method of renal mass reduc- 274:620–625, 199563. Furberg CD, Psaty BM, Meyer JV: Nifedipine: Dose-related in-tion is a critical modulator of subsequent hypertension and glomer-

ular injury. J Am Soc Nephrol 4:2023–2031, 1994 crease in mortality in patients with coronary heart disease. Circula-tion 92:1326–1331, 199546. Weibel ER: Stereological Methods: Practical Methods for Biologi-

cal Morphometry. London, Academia, 1979, pp 51–57 64. Alderman MH, Cohen H, Roque R, Madhavan S: Effect oflong-acting and short-acting calcium antagonists on cardiovascular47. Semple SJG, de Wardener HE: Effect of increased renal venous

pressure on circulatory “autoregulation” of isolated dog kidneys. outcomes in hypertensive patients. Lancet 349:594–598, 199765. Jeffrey AC: Calcium-channel blockers for hypertension: Uncer-Circ Res 7:643–648, 1959

48. Fry JC: Biological data analysis: A practical approach, in The tainty continues. N Engl J Med 338:679–680, 1998

class differences in the effects of calcium channel ... · class differences in the effects of...

Documents