The kidneys play a major role in the chronic regulation of blood pressure, primarily via modulation of sodium and water excretion. Abnormalities in the intrinsic properties of the kidney, such as reduced renal phosphate (1) and insulin clearances (2), altered renal hemodynamics (3), and increased sodium reabsorption in the proximal tubules (4, 5), are often associated with essential hypertension and may contribute to its existence. However, whether reduced kidney function is a consequence of hypertension (renal adaptation) or whether hypertension develops as a consequence of kidney dysfunction is currently unclear. On one hand, it seems that in spontaneously hypertensive rats (SHR) a high arterial pressure is required to excrete a given amount of salt and water (6). As a consequence, a blunting of the renal pressure-diuresis response may help to prevent exaggerated urine excretion, as previously described in SHR (7). This would agree with a more recent finding describing an alteration in SHR of an adrenoceptor subtype gene that stimulates solute excretion, a defective modulation that is probably not secondary to the elevated blood pressure (8). On the other hand, normalization of systolic blood pressure using acute (9, 10) or long-term (11, 12) treatments with antihypertensive drugs has long been known to improve renal function. Meanwhile, there is evidence that the glomerular capillary pressure and the local renal blood flow both contribute to glomerulosclerosis in SHR (13). Nevertheless, a possible worsening of kidney function under sustained hypertension has not yet been described. Our hypothesis is that chronic hypertension remains a major pathogenic factor in the progression of glomerular degeneration.
The plasma clearance of a physiologically inert substance is a measure of the glomerular filtration rate (GFR) as long as this substance is not synthesized, metabolized by the kidney or any other organ, or undergoes tubular reabsorption or secretion. In this respect, the determination of inulin clearance is superior to other methods for describing glomerular filtration; hence, it is considered to be the “gold standard.” However, this method suffers from being time-consuming, and it is invasive in such a way that the experiment is usually terminal in rats, and therefore does not allow longitudinal studies. Moreover, local renal tissue adjustments to a pathological situation cannot be addressed by this method. A recent study by Baumann and Rudin (14) demonstrated that dynamic MRI can noninvasively estimate kidney function in the rat with high temporal and spatial resolution. In this study, the renal clearance was assessed by measuring the uptake and release of the gadolinium complex Gd(DOTA) within selected areas of the kidney based on first-order kinetics, and by assuming that the paramagnetic agent is cleared only through the kidney.
Using dynamic MRI in anesthetized rats, the present study demonstrated renal clearance of gadopentetate dimeglumine Gd(DTPA), an extracellular contrast agent with similar relaxivity (15) and pharmacokinetic (16) properties to Gd(DOTA). We then tested whether kidney function assessed this way correlates with the glomerular filtration rate as measured with the inulin method. Finally, we investigated the hypothesis that chronic hypertension leads to progressive renal dysfunction. The role of high blood pressure was differentiated from intrinsic changes in kidney function by observing the Gd(DTPA) renal clearance response to the acute administration of a calcium antagonist drug.
Experiments were carried out with a group of 12–13-week-old normotensive Wistar-Kyoto (WKY) rats (mean arterial pressure (MAP) = 93 ± 8 mm Hg, body weight (BW) = 295 ± 7 g, N = 10) and two groups of spontaneously hypertensive rats (SHR). The first group of SHR was age-matched with the WKY group (young-SHR, MAP 135 ± 7 mm Hg, 256 ± 4 g, N = 10). The second group was composed of 22–25-week-old rats (old-SHR, MAP 140 ± 8 mm Hg, 360 ± 9 g, N = 12). The antihypertensive drug experiment was carried out using a separate group of SHR (14–15-weeks old, MAP 142 ± 10 mm Hg, 308 ± 6 g, N = 8). All animals received standard rodent chow and had free access to drinking water. Experimental procedures were carried out in compliance with the guidelines of the Novartis Institutional Animal Care and Use Committee.
Images were obtained using a 3.0-T/60-cm superconductive magnet (Bruker Medical Inc., Billerica, MA) equipped with a 12-cm i.d. actively shielded gradient system (maximum strength 200 mT/m) and using a 72-mm 1H volume resonator. All experiments were performed in rats that were anaesthetized with 2% isoflurane in a 70:30 nitrous oxide : oxygen mixture administered via a facemask. Prior to the MRI measurement, the tail vein was cannulated for the administration of the contrast agent Gd(DTPA) (Magnevist, Berlex Laboratories, Wayne, NJ). The animal was then positioned in the center of the probe, and body temperature was kept at 37°C ± 1°C by heating the air inside the magnet. A snapshot fast low-angle shot (FLASH) pulse sequence (17) was used to acquire images, as previously described (14). Radiofrequency spoiling was not used. However, it can be assumed that sufficient spoiling was provided by the imaging gradients. After slice positioning using orthogonal scout scans, 128 consecutive images (repetition time (TR) = 6 ms, echo time (TE) = 2.2 ms, spectral width = 100 kHz, field of view (FOV) = 5 × 5 cm, matrix = 128 × 64, slice thickness = 1.2 mm, flip angle = 25°) were collected in the coronal plane of the animal. The flip angle was set to 25° to enhance the sensitivity of the sequence to T1 shortening. The zero-filling technique was applied for image reconstruction (spatial resolution in plane 390 μm2). Motion artifact in images was minimized with sufficient signal averaging (16 averages), resulting in a 10-s acquisition time per image. Immediately after acquiring the 20th image of the series, 25 μmoles/rat of Gd(DTPA) (or 50 μl/rat of Magnevist) were injected in the tail vein and then flushed with 0.5 ml of saline within 2 s.
Image and Data Analysis
Time-course changes in signal intensities were translated into local Gd(DTPA) concentrations assuming a linear dependence with signal enhancement, as described previously (14). In brief, as a result of Gd(DTPA) administration, the effective longitudinal relaxation time T1 and the effective transverse relaxation time T2 can be expressed as:
where T10 and T20 are baseline relaxation times, α1 and α2 are the relaxivities of Gd(DTPA), and Ct is the Gd(DTPA) concentration at a time t after the injection of the contrast agent. The signal enhancement can then be computed according to Ref. 18:
with the pulse angle β = 0.44 (25°), and assuming that
In this study, changes in the relaxation rates were calculated using T10 = 1.4, 1.59, and 1.99 s, and T20 = 44, 48, and 86 ms as longitudinal and transverse relaxation times in the absence of Gd(DTPA) in the cortex, outer medulla, and inner medulla, respectively (19). Gd(DTPA) relaxivities used in the procedure were α1 = 1.2, 1.5, and 1.3 mM–1.s–1, and α2 = 7.9, 10, and 10 mM–1.s–1 for the cortex, outer medulla, and inner medulla, respectively (19). Ultimately, T10, T20, α1, and α2 were used as fitting parameters to describe the time-course of changes in Gd(DTPA) concentration in selected areas of the kidney. Time-to-peak (TTP) values were determined from the γ-variate fitting of the variation in Gd(DTPA) concentration during the initial phase after the tracer administration. The half-time (t1/2) of tracer disappearance was determined from the exponential fit within selected kidney regions as follows:
where k = ln2/t1/2. Finally, the kidney clearance index was derived from the first-order rate constant kcl according to:
assuming the tracer transport from the cortex (c) to the outer medulla (m) to be dominant during the initial phase. The rate constant kcl was determined by fitting the observed Gd(DTPA) concentration in the medulla to that calculated assuming a first-order transfer of Gd(DTPA) from the cortex to the medulla over the first minute after injection. The selection of regions-of-interest (ROIs) of usually 2 × 4 pixels was made from the third or fourth image of the series recorded after the injection of Gd(DTPA) that displayed good contrast between the cortex, outer medulla (usually about 2–3 mm from the surface), and inner medulla (between the outer medulla and the pelvis).
Measurement of the GFR
To correlate kcl with GFR over a wide range of renal function, renal clearance was measured in both the 12–13-week-old WKY (N = 10) and age-matched SHR (N = 10) groups. For this purpose, rats were placed on a temperature-regulated table and anesthetized under conditions similar to the MRI setup. GFR measurements were performed 2–4 days after the MRI assessment. Blood loss was minimal during the surgical preparation. A polyethylene (PE-10) catheter was inserted into the right carotid artery for blood sampling. Meanwhile, heparinized saline (0.9% NaCl) containing 0.13 mCi/mg [3H]inulin (Amersham Pharmacia Biotech Ltd., Buckinghamshire, UK) was infused through an additional PE-10 catheter in the left jugular vein (a bolus of 2 μCi diluted in 0.5% of BW saline followed by an infusion of 3.2 μCi diluted in 8 ml of saline at a rate of 0.05 ml/min for 150 min). Finally, urine was continuously collected via a PE-50 catheter attached to the bladder. The bladder was emptied of urine before the [3H]inulin infusion was initiated. After 30 min equilibration, urine samples were collected every 30 min in preweighed centrifuge tubes along with blood samples. Endogenous inulin clearance was computed after analyzing plasma and urine for [3H]inulin. Twenty and 50 μl of urine and plasma from respective samples were placed in a scintillation vial to measure 3H radioactivity with a liquid scintillation analyzer (model LS 6000IC, Beckman Coulter Inc., Fullerton, CA). GFR (ml/min/kg) was expressed as:
where UI is the urine concentration of inulin (mg/ml), V is the urine flow rate (ml/min), PI is the plasma concentration of inulin (mg/ml), and BW the rat body weight (kg).
Renal Response to the Acute Reduction of Blood Pressure
Each SHR served as its own control, and all experiments were performed in anesthetized rats as described above. The renal Gd(DTPA) was first measured after the administration of saline, then reassessed in the same animal one week later, 30–60 min after an acute 1 mg/kg administration of verapamil (Abbott Laboratories, North Chicago, IL) in the tail vein. Changes in the MAP were monitored by the tail-cuff method (IITC Model 31, IITC Inc., Life Science Instruments, Woodland Hills, CA) during the anesthesia period.
Results are expressed as mean ± SEM. Intergroup comparisons were made using two-way analysis of variance (ANOVA) and simultaneous multiple comparison procedures (with Bonferroni correction). The correlation between GFR and kcl data was assessed by linear regression analysis. P < 0.05 was considered statistically significant.
Dynamic Contrast-Enhanced MRI
Changes in signal intensity in a normal rat kidney following Gd(DTPA) administration, from a differential image series, are shown in Fig. 1. Significant tracer uptake could clearly be monitored from signal enhancement in the venous blood and specific regions of the kidney with a 10-s time resolution. Figure 2 illustrates the differences observed between 12-week-old WKY and age-matched SHR. Time-course changes in Gd(DTPA) concentrations are shown for the cortex, inner medulla, and outer medulla of both kidneys. The initial change in signal intensity first occurred in the vena cava within 10 s after the i.v. injection of Gd(DTPA), then in the renal cortex, quickly followed by change in the outer medulla and finally the inner medulla. In all of the regions except the inner medulla, the maximal signal intensity was followed by a slow decrease, reflecting the tracer transport from one compartment to the other. Interestingly, blood levels of Gd(DTPA) remained higher after peak values were reached in young-SHR as compared with age-matched WKY rats. This poor tracer elimination translated into low Gd(DTPA) concentrations in all kidney areas, including peak values. Similar trends were observed in old-SHR. Figure 3 illustrates differences in the kinetic variables related to these time-course variations as observed in the three groups of rats. While TTP values remained essentially the same in the venous blood and the cortex, a delayed accumulation of Gd(DTPA) in the medullas could already be measured in young-SHR. Cortical dysfunction became visible only at a more advanced stage of hypertension, i.e., in old-SHR. In those animals, a slow tracer uptake by the cortex and a slower elimination process, as reflected by greater TTP and t1/2 values, respectively, were observed. These changes obviously contributed to a greater accumulation of Gd(DTPA) in the blood pool, especially in old-SHR. In summary, the kidney clearance index as derived from the first-order rate constant kcl was 34% lower in young-SHR (P < 0.001) and 48% lower in old-SHR (P < 0.001), as compared with 12-week-old WKY rats (Fig. 3).
Correlation With GFR
The Gd(DTPA) clearance as estimated from the MRI-determined time constant kcl correlated well with GFR values as measured by the inulin method in the same animals (r = 0.75, N = 17, P < 0.0005; Fig. 4). Linear regression analysis indicated that the slope is different from 1 in such a way that a 66% decrease in GFR (e.g., SHR: 7.92 ± 0.96 ml/min/kg vs. WKY: 23.54 ± 3.35 ml/min/kg, P < 0.001) corresponds to a 29% decrease in kcl (e.g., SHR: 1.09 ± 0.06 min–1 vs. WKY: 1.53 ± 0.09 min–1). A ∼40% lower urine flow in SHR may have accounted for such a difference (SHR: 3.3 ± 0.4 μl/min vs. WKY: 5.9 ± 0.7 μl/min, P < 0.01) as well as for higher plasma inulin concentrations by ∼70% during the infusion period (SHR: 5.1 ± 0.3 mg/ml vs. WKY: 1.6 ± 0.2 mg/ml, P < 0.001).
Acute Effects of Verapamil of Kidney Function
The acute administration of verapamil resulted in a greater and more sustained reduction (P < 0.05, two-way ANOVA) of the MAP in anesthetized SHR over a 2-hr period as compared with control measurements (Fig. 5). This was accompanied by a 47% increase in the apparent first-order rate constant kcl (1.02 ± 0.15 vs. 1.50 ± 0.18 min–1, P < 0.05) (Fig. 6). Greater filtration was also evidenced by a rapid elimination of the tracer from the renal cortex (t1/2: 94 ± 35 vs. 44 ± 10 s–1) while maximal Gd(DTPA) concentrations were increased (P < 0.05, two-way ANOVA) in each compartment (blood: 1.8 ± 0.3 vs. 2.6 ± 11.3 mM; cortex: 1.4 ± 0.2 vs. 1.9 ± 0.3 mM; outer medulla: 1.0 ± 0.2 vs. 1.7 ± 0.3 mM; inner medulla: 3.8 ± 0.8 vs. 5.7 ± 0.7 mM). Interestingly, no sign of diuresis was shown throughout the 20-min course of measurement, since Gd(DTPA) concentrations in the inner medulla (mostly urine) were almost twofold higher than in controls.
An impaired kidney function most likely resulting from local renal adjustments to the high systemic blood pressure was characterized noninvasively in anaesthetized 3-month-old SHR by measuring Gd(DTPA) uptake and elimination. An even lower rate constant for the tracer transport from the cortex to the medulla was measured in 6-month-old SHR, suggesting that chronic hypertension contributes to a worsening of kidney function in aging SHR. An important result of this study is that the apparent rate constant kcl, describing the transport of tracer from the cortex to the outer medulla, can be considered a good estimate of the renal clearance as indicated by a significant linear correlation between kcl and GFR values. Finally, the acute administration of the calcium antagonist verapamil at an antihypertensive dose almost normalized SHR renal function, as reflected by an increased filtration of Gd(DTPA).
Our assessment of kidney function is in part based on a first-order kinetic model that describes the transport of Gd(DTPA) from the cortex (mostly blood, since the signal enhancement follows a pattern similar to the one observed in the vena cava area) to the medulla (mostly urine) (14). Similarly to Baumann and Rudin (14), we assumed that the tracer distribution follows a unidirectional pattern, i.e., blood → cortex → medulla → pelvis. Our TTP values support this assumption. In their studies, Baumann and Rudin also demonstrated a linear relationship between the total tracer amount injected and the initial rate of tracer uptake in all compartments analyzed. This led them to propose a first-order kinetic model of tracer elimination. During the initial phase the medullar Gd(DTPA) uptake rate is determined by the concentration in the cortex multiplied by a rate constant kcl. In other words, the increase in signal intensity in the cortex begins within seconds after administering Gd(DTPA), with a maximum reached at approximately 30 s. A slow, constant fall in cortical levels of gadolinium then follows concomitantly with an increase in the inner medullary levels of the tracer, reflecting the onset of glomerular filtration. At this point, it should be noted that the rapid disappearance, short halflife, size, and hydrophilicity support the postulate that free glomerular filtration remains the predominant route of Gd(DTPA) urinary elimination (16, 20). Also, the absence of tubular secretion and reabsorption (21) reinforces its utilization as a relatively specific GFR agent.
Experimentally, the measurement of the changes in the Gd(GDTPA) concentration in renal cortex and medulla as a function of time is straightforward. The determination of absolute concentrations in the inner medulla is prone to uncertainties. At relatively low Gd concentrations, contrast enhancement is governed by T1 effects. This is, however, not the case for high Gd tissue levels. In this case, T2 effects become dominant, resulting in signal attenuation due to T2 shortening. In our study, this latter effect was often observed near the pelvic area (i.e., the inner medulla) within 10 min after the contrast agent was injected, reflecting the efficient clearance of the Gd(DTPA) in the urine.
Essential hypertension has long been associated with low natriuresis due to an increased reabsorption by the proximal tubule (4, 5) while fluid excretion by the proximal and distal convoluted tubules appears to be undisturbed (22). As to glomerular filtration, the literature data are more controversial. On one hand, our GFR and kcl data disagree with a pressure-induced hyperfiltration phenomenon (23) or preservation of normal GFR in hypertension (24, 25). On the other hand, they support the concept that an increase in arterial pressure is associated with reduced GFR (13, 26, 27), and that a weak autoregulatory capacity of the glomerulus contributes to the maintenance of hypertension (e.g., through an excessive retention of salt (28)). It is true that preglomerular vascular resistance in young SHR allows a normal glomerular capillary pressure to be maintained (29, 30). However, at a later stage, glomerulosclerosis in the juxtamedullary cortex can develop in response to an increase in glomerular hydrostatic pressure, which then leads to a decreased renal filtration (13). It is also possible that the mechanism for a decreased Gd(DTPA) or inulin transport in SHR involves hemodynamic changes in the kidney itself. Thus, at the glomerulus level, a defect in the transmission of the renal artery pressure to the interstitium could cause the rightward shift of the pressure-natriuresis relation (11). Therefore, in light of a phenomenon that may occur prior to the development of hypertension, the question arises as to how reducing blood pressure helps restore normal kidney function.
It has been postulated that abnormalities of the cellular calcium levels contribute to an increase in intraglomerular pressure (31), which is frequently associated with glomerular sclerosis in SHR. When a calcium antagonist is given at a dose great enough to abolish autoregulation, the glomerular pressure becomes normal (32) or reduced (33), which in turn may preserve renal function (32). There is also good evidence that renal vessels in SHR are more sensitive to calcium antagonists than are systemic vessels (34). As a result, calcium channel blockers have been shown to exert hemodynamic, natriuretic, and diuretic effects in the kidney. The rise in GFR can account for the increase in solute excretion. However, studies undertaken to localize the action site of calcium channel antagonists have suggested that a decrease in fractional distal reabsorption is also involved in their diuretic and natriuretic properties (35, 36). In this study, we examined the responsiveness of Gd(DTPA) renal intake and elimination from SHR to an antihypertensive dose of verapamil. The results indicated that a substantial pressure reduction during calcium channel blockade resulted in an increased filtration of Gd(DTPA) and a more rapid elimination from the blood, cortex, and inner medulla. We assume that an inhibition of water resorption is less likely to account for the observed changes in kcl than the rise in glomerular filtration for the increase in solute excretion when treating with verapamil (as opposed, for example, to the diuretic furosemide). It is true that the MRI approach measures Gd(DTPA) clearance based on changes in concentration of the contrast agent in a single voxel, instead of the cumulative amount. However, the fact that we observed an increase in kcl, peak concentrations, and elimination of Gd(DTPA) excludes the possibility of a major tracer dilution in the urine resulting from increased water excretion, as hypothetized with furosemide (14).
Finally, our data show that the relative GFR difference measured between SHR and WKY rats was about double that observed from respective kcl values. Except for the infusion of the fluid containing inulin at a rate that may have induced up to ∼50% more diuresis in WKY rats, experimental conditions were kept rigorously similar for both the MRI and the inulin assessments. In this respect, we consider the MRI approach superior to the inulin method, as kidney function can be assessed under nondiuresis conditions, thereby eliminating a variable that has an important impact on final data.
In summary, we have demonstrated that the Gd(DTPA) clearance provides a good estimate of the glomerular filtration as long as water reabsorption remains unchanged, or its role on tracer dilution is negligible. With that assumption in mind, essential hypertension was associated with an impairment of kidney function in 3-month-old SHR. The mechanism for the long-term effects of high blood pressure on kidney function cannot be determined from this study. However, it is reasonable to suggest that structural changes such as glomerulosclerosis may accompany the worsening of kidney function, as observed in 6-month-old SHR. Finally, we have shown that acute verapamil treatment of young-SHR acted to enhance the Gd(DTPA) clearance concomitantly with a reduction of systemic blood pressure. However, one can assume that acute treatment with verapamil may not be a cure for secondary renal damages due to chronic hypertension, and thus may not fully restore kidney function as observed in young-SHR. We conclude that chronic hypertension remains a major pathogenic factor in the progression of glomerular degeneration, as opposed to a primary glomerular defect independent of systemic pressure.