Evidence against a crucial role of renal medullary perfusion in blood pressure control of hypertensive rats

Abstract

Key points

• The development of new effective methods of treating arterial hypertension is hindered by uncertainty regarding its causes.

• According to one widespread concept hypertension is caused by abnormal blood circulation in the kidney, specifically by reduction of blood flow through the kidney medulla; however, this causal relationship has never been rigorously verified.

• We investigated whether in rats with three different forms of experimental hypertension prolonged selective elevation of renal medullary blood flow using local infusion of the vasodilator bradykinin would lower arterial pressure.

• We found that increasing medullary blood flow by almost 50% did not result in alleviation of hypertension, which argues against a causal role of such changes in the control of arterial pressure and suggests that attempts at improving renal medullary circulation are not likely to be a promising approach to combating hypertension.

Abstract

The crucial role of renal medullary blood flow (MBF) in the control of arterial pressure (MAP) has been widely accepted but not rigorously verified. We examined the effects of experimental selective MBF elevation on MAP, medullary tissue hypertonicity and renal excretion in hypertensive rats. We used three hypertensive rat models: (1) rats with hypertension induced by chronic angiotensin II infusions (AngII model), (2) rats with hypertension induced by unilateral nephrectomy followed by high salt diet (HS/UNX), and (3) spontaneously hypertensive rats (SHR). In acute experiments, MBF (laser‐Doppler measurement) was selectively increased with an intramedullary infusion of bradykinin (Bk) at 0.27 mg h⁻¹ kg⁻¹ BW over 4 h. MAP, renal artery blood flow (Transonic probe) and renal excretion parameters were measured simultaneously. In chronic studies with AngII and HS/UNX rats, Bk was infused over 2 weeks and MAP (telemetry probe) and renal excretion were repeatedly determined. In acute studies, with AngII, SHR and HS/UNX groups, Bk infusion caused a 47% increase in MBF (P < 0.01–0.001), whereas solvent infusion was without effect. During the experiments MAP decreased slightly and to the same extent with Bk and solvent infusion. Medullary tissue osmolality and [Na⁺] were lower in Bk‐ than in solvent‐infused AngII rats and in SHR. Two weeks of intramedullary Bk infusion tested in AngII and HS/UNX rats did not alter MAP or renal excretion; though in the latter group a significant MBF increase and medullary hypertonicity decrease was observed. Since no decrease in MAP in hypertensive rats was seen with Bk‐induced major renal medullary hyperperfusion or with a wash‐out of medullary solutes, our data argue against a crucial role of MBF in the pathogenesis of arterial hypertension.

Key points

• The development of new effective methods of treating arterial hypertension is hindered by uncertainty regarding its causes.

• According to one widespread concept hypertension is caused by abnormal blood circulation in the kidney, specifically by reduction of blood flow through the kidney medulla; however, this causal relationship has never been rigorously verified.

• We investigated whether in rats with three different forms of experimental hypertension prolonged selective elevation of renal medullary blood flow using local infusion of the vasodilator bradykinin would lower arterial pressure.

• We found that increasing medullary blood flow by almost 50% did not result in alleviation of hypertension, which argues against a causal role of such changes in the control of arterial pressure and suggests that attempts at improving renal medullary circulation are not likely to be a promising approach to combating hypertension.

Introduction

Despite almost a century of focused research and striking progress in treatment, the pathogenesis of arterial hypertension and related disorders has not been satisfactorily clarified and multiple aspects of the leading relevant theories remain a matter of debate. This points to a need for rigorous testing of key components of these concepts (Beard, 2013). Perhaps the most outstanding fundamental controversy is that between exponents of a primarily neurogenic origin of essential hypertension (Schlaich et al. 2004; Osborn, 2005; Guyenet, 2006; Esler et al. 2010) and the protagonists of the ‘nephrocentric’ concept rooted in the classical work and views centred on the role of the kidney and the phenomenon of ‘pressure natriuresis’ (Guyton, 1961; Cowley, 1992). Within kidney‐oriented research, attention has more recently focused on the importance of the intrarenal renin‐angiotensin system activity (Navar, 2010; Navar et al. 2011). At the same time, the long‐standing and widely shared opinion on the key role of changes in renal medullary blood flow (MBF) in the control of sodium excretion and blood pressure (Cowley, 1997; Mattson, 2003; Bergstroem & Evans, 2004) is maintained and incorporated into new concepts explaining the mechanisms of salt‐dependent hypertension (Cowley et al. 2015).

The quintessence of the proposed role of MBF in control of blood pressure (BP) is that the pressure elevation leads to increased medullary perfusion. However, the very existence of such sequence and the underlying mechanism are uncertain: the original view that it simply reflects poor local blood flow autoregulation was not confirmed (Mattson et al. 1993; Majid et al. 1997).

Increased blood flow and hydrostatic pressure in the medullary vasa recta was originally proposed to start a chain of events beginning with a decrease in medullary tissue hypertonicity (due to a wash‐out of medullary solutes) and leading, via a sequence of physical changes in the renal interstitium, to a reduction in tubular water and solute transport and increased excretion. If this is sustained over hours or days, the body fluid depletion which ensues results in restoration of normal blood pressure (Mattson, 2003). Alternatively, increased medullary perfusion could trigger, by an unknown mechanism, release of a humoral vasodilator, such as the still‐illusory medullipin, which would reduce systemic peripheral vascular resistence and lower blood pressure (Muirhead, 1990); some evidence was presented suggesting that along this pathway BP could be normalized very rapidly (Zou et al. 1995). Both explanations ascribe much importance to the observed changes in renal medullary perfusion, but its causal role has not been rigorously tested, even though selective reduction in medullary blood flow was indeed reported in spontaneously hypertensive rats (SHR) and Dahl salt‐sensitive rats (Cowley et al. 1995; Miyata & Cowley, 1999). In our earlier study we used 30‐min renal intramedullary infusions of bradykinin (Bk) to examine if the resulting medullary vasodilatation and hyperperfusion would induce an acute decrease in BP in anaesthetized rats with various forms of experimental hypertension and found that this was not the case (Bądzyńska & Sadowski, 2012). Thus, sound evidence was provided against the proposal that an increase in MBF can cause a rapid decrease of elevated BP, allegedly mediated by release of a humoral vasodilator (Muirhead, 1990; Zou et al. 1995).

In the present study we pursued this line of research, using again intramedullary infusion of bradykinin, in an attempt to resolve three issues. First, we examined if renal medullary hyperperfusion extending over hours or chronic hyperperfusion during 2 weeks would induce a decrease in BP. Second, we examined if such hyperperfusion is indeed followed by a decrease in medullary tissue hypertonicity, purportedly a key event in the development of pressure natriuresis; this issue has not been adequately studied in whole‐animal experiments. Third, we thought it of interest to find out how prolonged or chronic medullary vasodilatation would affect renal excretion and how these changes are related to the medullary tissue hypertonicity level and to the systemic and renal haemodynamics.

Methods

Ethical approval

The procedures used in experiments performed with male Sprague‐Dawley (S‐D) or in male spontaneously hypertensive rats (SHR) were approved by the extramural IV Local Ethical Committee for Animal Experimentation, Warsaw, and conform with UK and EU legislation regarding ethical aspects of animal experimentation. All the authors understand and their work complies with the ethical principles listed in the guidelines of The Journal of Physiology. All rats derived from the animal house of the Mossakowski Medical Research Centre, Polish Academy of Sciences.

The aim was to examine if prolonged (4 h) or chronic (14 days) selective experimental elevation of the renal medullary perfusion (blood flow) in hypertensive rats lowered arterial blood pressure and/or medullary tissue osmolality and sodium concentration, and if it affected renal excretion.

Experiments were performed with three rat models of hypertension:

(1) AngII model

S‐D rats (n = 16), aged 12 weeks, weighing 280–340 g, were maintained on a standard sodium diet (0.25% Na⁺). They received continuous infusion of AngII at a dose that was directly subpressor, using Alzet osmotic minipumps (model 2002, Durect Co, Cupertino, CA, USA). Metacam (0.4 mg kg⁻¹ BW, Boehringer, Ingelheim, Germany) and baytril (10 mg kg⁻¹ BW, Bayer, Leverkusen, Germany) were first given subcutaneously for analgesia and to prevent infection, respectively. Subsequently, under inhalatory anaesthesia using an isoflurane (4%) and oxygen mixture, Alzet minipumps were implanted subcutaneously in the nap area to deliver angiotensin II (Sigma‐Aldrich, Basel, Switzerland) at 35μg kg⁻¹ day⁻¹ for 14 days, a procedure which induced mild hypertension (angiotensin‐induced hypertension) (Chin et al. 1998). The animals were kept in standard cages, with free access to food and drinking water, and their general condition was monitored daily.

(2) SHR model

Spontaneously hypertensive rats (n = 19), aged 10–12 weeks, weighing 280–300 g, with established hypertension, were maintained on a standard sodium diet.

(3) HS/UNX model

S‐D rats (n = 20, of these three rats were excluded because they failed to develop hypertension) weighing 320–360 g, aged 12 weeks, were subjected to right‐side nephrectomy (UNX) under analgesia and isoflurane/oxygen anaesthesia as described for Alzet minipump implantation in the AngII model, and with the usual post‐operative monitoring. Thereafter the rats were placed on high‐salt (HS) diet (4% Na⁺ w/w, Labofeed, Kcynia, Poland) for 14 days. This procedure (HS/UNX) was previously shown to result in the development of hypertension (Carlström et al. 2007; Bądzyńska & Sadowski, 2012).

Acute and chronic studies were performed, using different experimental protocols.

Acute studies

The same procedures and basic protocol were used with each of the three hypertension models. The rats were anaesthetized with thiopental (Sandoz GmbH, Kundl, Austria, 100 mg kg⁻¹ i.p.) which provided stable anaesthesia for about 5 h, with an additional small subcutaneous dose when needed. With the AngII and SHR models the right‐side nephrectomy was first performed from a right flank incision; the wound was then closed with sutures. In the HS/UNX rats, uninephrectomy had been done 2 weeks earlier, as required by the model. Right‐side nephrectomies were performed to avoid any neural or humoral effects of the contralateral kidney on the experimental (left) kidney function.

The details of experimental procedures were as described earlier (Bądzyńska & Sadowski, 2009, 2012). Briefly, mean arterial blood pressure (MAP) and heart rate (HR) were measured via a right femoral artery catheter connected with a Stoelting blood pressure meter (Stoelting, Wood Dale, IL, USA). The left kidney was exposed from a subcostal flank incision and immobilized in a plastic holder, and the ureter was cannulated for timed urine collection to measure urine flow (V), sodium excretion (U _Na V) and total solute excretion (U _osm V). In one of the two angiotensin‐induced hypertension groups, plasma and urine inulin levels were determined for calculation of inulin clearance, a measure of the glomerular filtration rate (GFR). In these rats, inulin (Sigma‐Aldrich, Poznan, Poland) was added to intravenously infused isotonic saline to obtain a 1.5% solution, which was given throughout the experiment. To measure total renal blood flow (RBF), a Transonic cuff probe was placed on the renal artery and connected with a Transonic flowmeter (Type T106, Transonic System Inc., Ithaca, NY, USA). Perfusion of the renal medulla (MBF) was measured throughout experiments as laser‐Doppler flux, using Periflux 4001 system (Perimed, Jarfalla, Sweden).

Intramedullary infusions and experimental protocol

A 32‐gauge stainless steel cannula, connected with an infusion pump, was inserted into the kidney, with the tip located at the outer‐inner medullary border, and an infusion of 0.9% NaCl, i.e. bradykinin (Bk) solvent was started at a volume rate of 0.5 ml h⁻¹. About one hour was then allowed for equilibration of the haemodynamic parameters measured and urine excretion, followed by a 30‐min control measurement period and urine collection. Subsequently, saline solvent infusion was replaced by Bk (Sigma‐Aldrich, Basel, Switzerland), which was given at 0.27 mg h⁻¹ kg⁻¹ BW over 4 h (Bądzyńska & Sadowski, 2012). Ten minutes later MAP, HR, RBF and MBF recording was started and continued throughout the experiment, along with 30‐min urine collections. In parallel time‐control experiments, Bk solvent was infused while measurements were being made as usual. At the end of experiments, the kidney was removed for studies of the medullary tissue, and the inner renal medulla was excised for measurement of tissue osmolality and sodium concentration [Na⁺]. The rats were killed with an overdose of thiopenthal.

Determination of osmolality and sodium concentration ([Na⁺]) in renal inner medulla

After the infusion experiments, the kidney was excised, and sectioned along the sagittal paracentral plane so as to preserve the inner (white) medulla intact. To obtain a standard medullary piece from the excised white medulla, the tissue was punched out using a stainless steel tube slightly oval at the cross‐section. The pieces included most of the white medulla but not the tip region of the papilla. They were weighed and subjected to 1 h extraction in a 1 ml volume of boiling distilled water, and osmolality and sodium concentration of the extraction fluid (equilibrated with tissue fluid) was determined. The final results were expressed as milliosmoles and millimoles of Na⁺ per kilogram of wet medullary tissue (Dobrowolski et al. 1992).

Chronic studies

Experimental measurements were performed during 14 days of infusions of Bk or solvent into the renal medulla of conscious rats in two models of hypertension (AngII and HS/UNX).

In the AngII model (n = 13), a right‐side nephrectomy was first performed, together with implantation of the minipump for AngII infusion, under analgesia and anaesthesia, followed by post‐operative monitoring as described in the Acute studies section. Seven days later aortic telemetry probes were implanted in all the rats, for chronic measurement of arterial pressure and heart rate (HR). The probes consisting of a battery‐operated radiotransmitter and a catheter (type TA11PA‐C40, Data Science International, St Paul, MN, USA) were implanted under aseptic conditions and sodium pentobarbital anaesthesia (50 mg kg⁻¹ i.p.). The abdominal aorta was exposed and the catheter was inserted by direct puncturing of the vessel. The transmitter was placed inside the peritoneal cavity and fixed to the abdominal muscle wall. Metacam (0.4 mg kg⁻¹ BW, Boehringer, Ingelheim, Germany) and baytril (10 mg kg⁻¹ BW, Bayer, Leverkusen, Germany) were used as post‐operative analgesia and to prevent infection, respectively. Five days of post‐operative monitoring were allowed for midline wound healing and full general recovery. Then, on day 0, 3‐h control telemetric measurements of arterial pressure were performed and the rats were placed in metabolic cages, and 24‐h urine collections were made for determination of renal excretion parameters (V, U _Na V and U _osm V).

On the following day the rats were anaesthetized using the usual inhalation of an isoflurane/oxygen mixture, with Metacam as analgesic, and the left kidney was exposed from the left subcostal incision. For chronic intramedullary Bk or solvent infusions, a minipump + microcannula set (Alzet, Durect Co., Cupertino, USA), originally designed for intracerebral infusions in rodents, was applied. The microcannula was inserted to position the tip at the depth of the outer‐inner medullary border, and was fixed to the renal capsule using tissue glue. It was connected with an osmotic minipump that was implanted in the abdominal subcutaneous tissue to deliver bradykinin or its solvent into the left renal medulla over 14 days. In this model another osmotic minipump was implanted to continue AngII infusion for another 14‐day period, till the end of the experiment.

For the HS/UNX model (n = 18, of these three rats were excluded because they failed to develop hypertension) the procedures were similar except that right‐side nephrectomy was performed 14 days earlier, at the onset of HS diet administration, and chronic infusions other than bradykinin or solvent were not needed.

On days 7 and 14 of Bk or solvent infusion, MBP and HR were measured telemetrically and the renal excretion parameters (V, U _Na V and U _osm V) were determined in metabolic cages. In addition, after 14 days of infusions and completing final measurements, the HS/UNX rats were anaesthetized with intraperitoneal thiopental as described for acute studies, and 1‐h measurements were conducted of MBP, HR, RBF, MBF, V, U _Na V and U _osm V, followed, after the experiment, by measurement of medullary tissue osmolality and [Na⁺] in the excised kidney.

Analytical procedures

Urine volume was determined gravimetrically. Urine, plasma and tissue sodium concentrations were measured with the flame photometer (Jenway PFP7, Essex, UK), and urine and tissue osmolality using a cryoscopic osmometer (Osmomat 030, Gonotec GmbH, Berlin, Germany). Plasma and urine inulin levels were determined using a modified method described previously (Nolin et al. 2002) and the inulin clearance, a measure of GFR, was calculated from the standard formula. Fractional sodium excretion (FE_Na) was calculated as the percentage ratio of U _Na V to filtered sodium load (equivalent to GFR × P _Na).

Statistics

Data are expressed as means ± SEM. Significance of changes within one group over time was first evaluated by repeated measures analysis of variance (ANOVA), followed by modified Student's t test for dependent variables, using Bonferroni correction for multiple comparisons (Wallenstein et al. 1980). Differences between profiles for bradykinin and solvent infusion were first analysed by repeated measures multivariate ANOVA, followed by Duncan's test (STATISTICA 10.0, StatSoft Polska Inc.). P ≤ 0.05 was taken to indicate significance of differences.

Results

Acute experiments

Baseline values of the mean arterial pressure (MAP) and renal haemodynamic and excretion parameters in three models of experimental hypertension: angiotensin‐induced (AngII), genetically determined (SHR), and in the HS/UNX model of salt‐dependent hypertension are collected in Table 1. The renal data present the values for the left kidney since all the rats had undergone contralateral nephrectomy: the HS/UNX group at the onset of exposure to high salt diet and the two other groups directly before acute study. Since there were substantial differences in kidney weight between the individual hypertension models, renal blood flow (RBF), V, U _Na V and U _osm V are expressed per gram kidney weight (KW). As expected, the highest MAP values were found in SHR, and the excretion rates were distinctly higher in the HS/UNX rats than in the other models.

Table 1.

Baseline values of MAP, renal haemodynamics and excretion parameters, and kidney weight (KW) in three models of hypertension

Group	n	MAP (mmHg)	RBF (ml min−1 g−1 tissue)	MBF (PU)	V (μl min−1 g−1 tissue)	U Na V (μmol min−1 g−1 tissue)	U osm V (μosm min−1 g−1 tissue)	KW (g)
AngII	16	154 ± 4	7.4 ± 0.4	115 ± 10	7.8 ± 0.8	0.7 ± 0.12	6.8 ± 0.5	2.01 ± 0.06
SHR	19	178 ± 3	7.5 ± 0.5	187 ± 12	4.3 ± 0.5	0.5 ± 0.1	4.4 ± 0.3	1.24 ± 0.04
HS/UNX	17	144 ± 3	6.5 ± 0.5	117 ± 12	11.8 ± 1.5	2.8 ± 0.4	10.1 ± 1.0	2.54 ± 0.06

MAP, mean arterial pressure; RBF, renal blood flow; MBF, renal medullary perfusion; V, excretion of water U _osm V, total solute excretion; U _Na V, total sodium excretion. Pooled control data from Bk and solvent infusion studies. AngII, angiotensin‐induced hypertension; SHR, spontaneously hypertensive rats; HS/UNX, hypertension induced by high‐salt diet + uninephrectomy.

The results of acute experiments (4‐h infusions of Bk or solvent) in the AngII, SHR and HS/UNX groups are presented in Figs 1, 2, 3, and the main data for an additional AngII group are given in Table 2. During intramedullary infusion of both Bk and solvent, MAP decreased slightly in each of the three main groups but not in the additional AngII group.

Figure 1. Effects of 4 h of bradykinin or solvent infusion on renal haemodynamics and excretion parameters in rats with angiotensin (AngII)‐induced hypertension.

MAP, mean arterial pressure; RBF, renal blood flow; MBF, renal medullary perfusion; V, excretion of water; U _osm V, total solute excretion; U _Na V, total sodium excretion. Means ± SEM: averaged 30‐min values measured in the control period and at the end of 80, 160 and 240 min of bradykinin infusion. Bradykinin, continuous line (n = 9); solvent, interrupted line (n = 7). ^#Significant difference between profiles at P < 0.01 (RBF) and P < 0.001 (MBF) (repeated measurement ANOVA); ^*significantly different from pre‐infusion control at P < 0.01 or less (Duncan post hoc test).

Figure 2. Effects of 4 h of bradykinin or solvent infusion on renal haemodynamics and excretion parameters in SHR.

MAP, mean arterial pressure; RBF, renal blood flow; MBF, renal medullary perfusion; V, excretion of water; U _osm V, total solute excretion; U _Na V, total sodium excretion. Means ± SEM: averaged 30‐min values measured in the control period and at the end of 80, 160 and 240 min of bradykinin infusion. Bradykinin, continuous line (n = 12); solvent, interrupted line (n = 7). ^#Significant difference between profiles at P < 0.001 (MBF) and P < 0.007 (U _Na V) (repeated measures ANOVA); ^*significantly different from pre‐infusion control at P < 0.05 or less (Duncan's post hoc test).

Figure 3. Effects of 4 h of bradykinin or solvent infusion on renal haemodynamics and excretion parameters in rats with salt‐induced hypertension (HS/UNX).

MAP, mean arterial pressure; RBF, renal blood flow; MBF, renal medullary perfusion; V, excretion of water; U _osm V, total solute excretion; U _Na V, total sodium excretion. Means ± SEM: averaged 30‐min values measured in the control period and at the end of 80, 160 and 240 min of bradykinin infusion. Bradykinin, continuous line (n = 9); solvent, interrupted line (n = 8); ^#Significant difference between profiles at P < 0.0001 (repeated measures ANOVA); ^*significantly different from pre‐infusion control at P < 0.0001; ^†significant difference between the values for Bk and solvent at the same time point at P < 0.02 (Duncan's post hoc test).

Table 2.

Renal haemodynamics and excretion parameters as affected by intramedullary bradykinin (Bk) infusion in an additional group of AngII‐induced hypertensive rats

		Bk
	Control	80 min	160 min	240 min
MAP (mmHg)	152 ± 4	155 ± 4	147 ± 4	155 ± 4
RBF (ml min−1 g−1 KW)	9.9 ± 0.6	10.0 ± 0.4	9.9 ± 0.5	8.8 ± 0.5
MBF* (PU)	42 ± 3	68 ± 9	69 ± 15	73 ± 19
GFR (ml−1 min−1 g KW)	1.0 ± 0.1	0.8 ± 0.1	–	0.9 ± 0.1
U Na V * (μmol min−1 g−1 KW)	1.6 ± 0.3	1.8 ± 0.3	2.4 ± 0.7	2.8 ± 0.5
FENa * (%)	1.21 ± 0.22	1.98 ± 0.58	–	2.53 ± 0.52

Means ± SEM; n = 5: averaged 30‐min values measured in the control period and at the end of 80, 160 and 240 min of Bk infusion. GFR, glomerular filtration rate (inulin clearance); FE_Na, fractional sodium excretion. Other denotations as in Table 1. ^*Significant increase during experiments at P < 0.03 or less (repeated measures ANOVA).

To assess first if Bk effectively increased medullary perfusion (MBF), as intended, the profile of the induced change over time was compared to that in time‐control experiments (solvent infusion) using multivariate repeated measurement ANOVA (Table 3). It is seen that highly significant MBF increases (relative to the profile for the solvent) were observed for each of the three hypertension models.

Table 3.

Statistical comparison of the change in medullary perfusion (MBF) during 4 h of bradykinin versus solvent infusion into the renal medulla in three models of rat hypertension

Group	SS	MS	F	P<
AngII (9, 9)	6747	2249	10.0454	0.0001
SHR (12, 8)	4409	1470	4.4751	0.0071
HS/UNX (9, 8)	12843	4281	10.3097	0.0001

Comparison made using repeated measures multivariate ANOVA. AngII, angiotensin‐induced hypertension; SHR, spontaneously hypertensive rats; HS/UNX, hypertension induced by high‐salt diet + uninephrectomy (n values for bradykinin, solvent). SS, sum of squares, MS, mean square, F value. The degree of freedom for interaction of time(t) × treatment(k) factor equals (n − t)(n − k), i.e. (4 − 1)(2 − 1) = 3 for each group.

Figure 1 shows that in the AngII group MBF increased about 46% during Bk and slightly decreased during solvent infusion. Notably, a similar differentiated response was also seen for RBF: bradykinin appeared to prevent a significant decrease in RBF observed in the time‐control study. There was no significant difference in the slightly decreasing profiles of mean arterial pressure (MAP) between the Bk and solvent group (repeated measures ANOVA, F(3,1) = 2.22; P < 0.2).

The renal excretion parameters (V, U _Na V, U _osm V) showed a transient significant increase in the first hour of Bk infusion, the change paralleled that in MBF but not in RBF. However, the changes in U _Na V or V did not correlate significantly with MBF (r = 0.62 and 0.57, n = 9, P < 0.2). At the end of the experiment the excretion rates were at the control level. In order to double‐check for the puzzling fluctuations in excretion and examine the possible role of GFR, another series of Bk infusions was later performed (Table 2). In this series, MAP, RBF and GFR were stable and not altered by intramedullary Bk infusion. There was the usual almost 70% increase in MBF and a progressive increase in both U _Na V and FE_Na (for the three parameters the increase was significant at P < 0.03 or less, as determined by repeated measures ANOVA), which suggested inhibition of the tubular sodium transport. Thus, the transiency of the excretion increase (Fig. 1) was not confirmed.

As shown in Fig. 2, in the SHR group MBF increased about 47% during Bk and 15% during solvent infusion, a difference that was distinct and significant (Table 3). The respective increases in RBF were 22 and 18%; the two profiles did not differ significantly. Apparently, both RBF and MBF were progressively increasing in the solvent‐infused rats and infusion of Bk caused an additional major increase in MBF only. MAP decreased slowly over 4 h, similarly in the Bk and time‐control groups. Notably, in bradykinin‐infused rats there was no correlation between the final decrease in MAP and the increase in MBF (r = −0.43, n = 12, P < 0.2

The renal excretion parameters increased progressively and significantly over time. For U _Na V the increase was significantly steeper in the Bk than in the solvent study (repeated measures ANOVA, F = 4.48; P < 0.007); however, it was not significantly correlated with MBF (r = 0.42, n = 10, P < 0.2). Since the natriuresis may have been related to increasing GFR, two additional Bk infusion experiments were performed in which GFR was also determined. It was found to increase from control levels of 0.61‐0.62 ml min⁻¹ g⁻¹ KW to 0.91‐1.07 and 0.74‐0.86 ml min⁻¹ g⁻¹ KW in the two last experimental periods; GFR did not change in solvent‐infused rats.

Figure 3 shows that in the HS/UNX group Bk distinctly increased MBF (47%) while no change was seen during solvent infusion. The MBF change occurred with stable RBF, which means that the effect of Bk was selective for the medulla. The renal excretion parameters (V, U _osm V, U _Na V) tended to increase transiently during the first hour of Bk (but not solvent) infusion, but the increments in excretion were not correlated with MBF.

The data for total solute (osmolality) and sodium concentration in the renal medulla at the end of the experiments are presented in Table 4. It is seen that in the AngII and SHR models both tissue osmolality and sodium were distinctly and consistently lower after Bk compared to solvent infusion. Notably, SHR was the only group in which increasing MBF was associated both with lower medullary tissue osmolality and [Na⁺], and increasing sodium excretion. In the HS/UNX model the medullary osmolality and sodium only tended to be lower in Bk infusion than in solvent infusion experiments.

Table 4.

A comparison of tissue osmolality and sodium concentration [Na⁺] in the inner medulla, measured at the end of BK and solvent infusion studies in three rat hypertension models

	Medullary osmolality (mosm kg−1 wet medullary tissue)		Medullary [Na+] (mmol kg−1 wet medullary tissue)
	Bk	Solvent	Bk	Solvent
AngII	717 ± 85	788 ± 57	128 ± 12*	167 ± 10
	(9)	(7)	(9)	(6)
SHR	822 ± 32*	972 ± 63	131 ± 22	180 ± 15
	(10)	(9)	(11)	(6)
HS/UNX	795 ± 60	845 ± 47	154 ± 10	168 ± 14
	(6)	(7)	(5)	(7)

AngII, angiotensin‐induced hypertension; SHR, spontaneously hypertensive rats; HS/UNX, hypertension induced by high‐salt diet + uninephrectomy. Values are means ± SEM (n). ^*Significantly lower than in the corresponding solvent infusion study (P < 0.05, unpaired t test).

Chronic studies

The results of follow‐up studies of MAP and the renal excretion parameters during 2 weeks of renal intramedullary infusion of Bk (0.27 mg h⁻¹ kg⁻¹ BW) or solvent in conscious rats with AngII‐ and HS/UNX‐induced hypertension are shown in Figs 4 and 5, respectively. We thought it of interest to compare these two models in chronic studies because of their different functional characteristics: at comparable baseline MAP levels, the excretion of water, sodium and total solutes was far greater (about 7‐fold, 22‐fold and 6‐fold, respectively) in the HS/UNX than in AngII model rats.

Figure 4. A follow‐up study of renal haemodynamics and excretion parameters during 2 weeks of intramedullary infusion of bradykinin and solvent in AngII‐induced hypertension.

MAP, mean arterial pressure; RBF, renal blood flow; V, excretion of water; U _osm V, total solute excretion; U _Na V, total sodium excretion. Bradykinin, filled bars (n = 8); solvent, open bars (n = 5).

Figure 5. A follow‐up study of renal haemodynamics and excretion parameters during 2 weeks of intramedullary infusion of bradykinin and solvent in HS/UNX hypertension model.

MAP, mean arterial pressure; RBF, renal blood flow; V, excretion of water; U _osm V, total solute excretion; U _Na V, total sodium excretion. Bradykinin, filled bars (n = 8); solvent, open bars (n = 7). ^*Significantly different from day 0 value at P < 0.05 (paired Student's t test).

In both models MAP showed only minor changes over the 2‐week follow‐up period. In neither model did Bk or solvent infusion alter MAP significantly. In both models diuresis (V) tended to increase with Bk but not with solvent infusion over 2 weeks; otherwise no consistent changes in renal excretion or differences in the excretion pattern between Bk and solvent studies were observed.

In the HS/UNX model, at the end of the chronic studies, while osmotic minipump infusions were still being continued, the rats were anaesthetized and the final acute measurements were done. Table 5 shows very similar MAP and RBF values for Bk‐ and solvent‐infused rats. On the other hand, MBF was 2.3‐fold higher in the Bk group (difference significant at P < 0.03) while medullary [Na⁺] was about 18% lower. There were no differences in the renal excretion parameters between the two groups. Similar acute experiments were also performed with AngII rats but because of the high variability of the parameters measured the data were found unsuitable for analysis.

Table 5.

Mean arterial pressure (MAP), renal haemodynamic parameters (RBF, MBF), sodium and osmolality of the renal medulla, and renal excretion parameters (V, U _Na V, U _osm V) in HS/UNX rats after 2 weeks of bradykinin

	Bk (n = 8)	Solvent (n = 8)
MAP (mmHg)	151 ± 11	145 ± 5
RBF (ml min−1 g−1 KW)	5.3 ± 0.3	5.2 ± 0.7
MBF (PU)	132 ± 21*	57 ± 8
[Na+] (mmol kg−1 MW)	186 ± 14	228 ± 58
[osm] (mosm kg−1 MW)	571 ± 34	593 ± 110
V (μl min−1 g−1 KW)	17.2 ± 5.3	14.5 ± 3.1
U Na V (μmol min−1 g−1 KW)	2.8 ± 0.8	2.8 ± 0.5
U osm V (μosm min−1 g−1 KW)	8.5 ± 2.0	8.0 ± 1.7
KW (g)	2.26 ± 0.06	2.83 ± 0.17

Data from acute experiments performed after 2 weeks of bradykinin (Bk, 0.27 mg h⁻¹ kg⁻¹ BW) or solvent infusion into the medulla of the left kidney. [Na⁺] and [osm], sodium and total solute concentration, respectively, of the renal medulla; KW, kidney weight; MW, medulla weight. ^*Significantly different from the corresponding value during solvent infusion (P < 0.003, Student's t test for unpaired samples).

Discussion

It was not certain if prolonged administration of bradykinin (Bk) would induce sustained elevation of MBF, even though in our earlier study the short‐term effect was quite distinct (Bądzyńska & Sadowski, 2012). However, in the present acute studies with each of the three rat hypertension models prolonged intramedullary Bk infusion effectively increased renal medullary perfusion. This experimental approach made it possible to achieve local vasodilatation over hours, which provided a necessary basis for a thorough examination of the functional consequences of medullary hyperaemia. There is evidence that intramedullary bradykinin increases MBF without affecting MAP in two normotensive rat strains (Mattson & Cowley, 1993; Bądzyńska & Sadowski, 2009). Extending bradykinin‐induced medullary vasodilatation over 2 weeks was even more challenging: we found that this can indeed be accomplished by chronic intramedullary infusion of Bk using implanted osmotic minipumps.

In each hypertension model MAP decreased slowly throughout experiments, to the same extent with bradykinin and solvent infusion. The reason for the increase was probably slow waning of the initial systemic vasoconstrictor response to anaesthesia and experimental surgery, and a progressive decrease in total peripheral and renal vascular resistance. We checked carefully and found that the pressure decrease was unrelated to bradykinin infusion or the consequent increase in MBF.

Increase in medullary perfusion does not cause a decrease in arterial pressure

The crucial issue examined here is whether, as claimed in numerous reports and promoted in an array of reviews (Cowley, 1992, 1997; Mattson, 2003; Bergstroem & Evans, 2004; Cowley et al. 2015), increased perfusion of the renal medulla (MBF) would trigger a chain of events which eventually lead to a decrease in systemic arterial blood pressure (MAP), especially in hypertensive animals.

When our Bk‐infusion experiments were analysed in detail, it was seen that: (1) MBF increases were not associated with MAP decreases beyond the usual tendency to a slight decrease (observed both in Bk‐ and solvent‐infused rats); (2) within any one group there was no correlation between the changes in the two variables; (3) scrutiny of the course of MAP and MBF changes over hours disclosed no short‐time associations in time. On the whole, there was no evidence of correlation or even a clear association between distinctly increasing MBF (dependent on Bk infusion) and the MAP level. Notably, this was true for three different hypertension models, characterized by different baseline values of arterial blood pressure, different total renal and medullary blood perfusion rates and different rates of renal excretion (Table 1). It will be noticed that these functional differences were related to the different pathogenetic backgrounds of the models: increased systemic and intrarenal activity of the renin‐angiotensin system (RAS) in the AngII rats, genetically determined hypertension in SHR, and salt‐dependent hypertension in the HS/UNX rats characterized by inhibition of the RAS.

It should be noted that a correlation between MBF and MAP, if present, would not per se indicate a causal relationship; however, in the absence of such correlation, as seen in the present study, causality is unlikely. A limitation of our interpretation is that medullary vasodilatation is not the only effect of Bk. A probable simultaneous inhibition by Bk of the tubular reabsorption (Mamenko et al. 2014) might be a confounding factor; however, the similarity of renal sodium excretion in Bk‐ and solvent‐infused rats argues against a prominent role for such an effect. Nevertheless, to confirm our interpretation it will be important to test the effects of medullary vasodilatation using some approach other than bradykinin infusion and to examine the effects of selective medullary hypoperfusion.

We found no indication of a decrease in MAP over 2 weeks of intramedullary Bk administration in the AngII and HS/UNX rat models. It is noteworthy that in the latter group (studied in a final acute experiment) the MBF value after chronic Bk infusion was a double that after solvent infusion, whereas arterial pressure even tended to be lower in the latter (Table 5). While the chronic studies should be extended to involve other hypertension models, they strongly suggest the effectiveness of the use of bradykinin for chronic medullary vasodilatation and argue against any effect on MAP. To confirm this conclusion, chronic measurement of MBF is required, but the simultaneous intramedullary infusion and local MBF measurement present a technical challenge.

Effect of increase in MBF on medullary tissue hypertonicity

In the proposed chain of events from an increase in MBF to the eventual decrease in MAP (Cowley, 1997; Mattson, 2003), the intermediate stage is a decrease in medullary tissue hypertonicity, explained as a result of ‘wash‐out’ of medullary solutes. The linking of this change to medullary perfusion changes requires whole‐kidney studies. The electrolyte component of the medullary hypertonicity can be followed in the in situ kidney by continuous measurement of tissue electrical admittance (reciprocal impedance) (Sadowski & Dobrowolski, 2003). We found in physiological experiments that spontaneous increases in MBF were almost immediately followed by decreases in medullary tissue ion concentrations and MBF decreases were followed by opposite changes (Sadowski et al. 1997).

In the present study we found that in the AngII and SHR models an increase in MBF was associated with a distinct decrease in medullary tissue osmolality and sodium concentration, while in the HS/UNX model only a slight tendency to decreased MBF was seen (Table 4). It is emphasized that in each of these models at the end of experiments, i.e. about 10 min before harvesting the medullas for tissue studies, MBF was significantly higher in the Bk‐ than in solvent‐infused rats (Figs. 1, 2, 3). An association between increased MBF (with stable RBF) and lowering of medullary [Na⁺] was also seen in the HS/UNX rats after 2 weeks of Bk infusion (Table 5).

Actual medullary tissue ionic hypertonicity reflects a balance of two processes: a delivery to the interstitium of sodium and attendant anions reabsorbed from the medullary ascending limb of Henle's loop, and evacuation of the solutes by the medullary vasa recta (a wash‐out): the latter effect would be attenuated or offset by the first one (Sadowski & Dobrowolski, 2003). It is difficult to avoid interference by changes in tubular reabsorption in attempts to demonstrate the ‘wash‐out’ effect of MBF on medullary hypertonicity since the experiments usually involve graded reduction of renal perfusion pressure (RPP). In a focused study in normotensive rats we found that while MBF clearly decreased during RPP reduction within the RBF autoregulation range (evidence of poor autoregulation in this setting), the medullary tissue osmolality and ion concentration showed no increase until RPP reached a level of about 80 mmHg, probably owing to falling distal delivery of the tubular fluid to the deep nephrons and decreasing distal tubular reabsorption and solute delivery to the interstitium (Dobrowolski et al. 1998).

On the whole, our data show that hyperperfusion of the medulla can modestly lower the medullary hypertonicity in some models of rat hypertension. In theory, this change could limit the reabsorption in medullary tubule segments and increase the renal excretion, leading to fluid depletion and normalization of elevated blood pressure. Therefore the next crucial event to follow a decrease in medullary hypertonicity would be increased renal excretion.

Increase in MBF: any relation to renal excretion?

Our data indicate that the relationships between Bk‐induced MBF increases and changes in RBF, GFR, medullary hypertonicity levels and renal excretion were different in individual hypertension models.

In the SHRs and in one of the two AngII groups, U _Na V increased consistently during Bk‐induced MBF elevation. The change was associated with, and possibly dependent on, an increase in GFR; however, in the AngII group an increase in FE_Na also suggested an inhibition of tubular sodium transport. In the SHRs the increase was associated with a decrease in medullary hypertonicity. A possible suggestion that the change in excretion was the consequence of increasing MBF (and a medullary wash‐out phenomenon) is weakened by the lack of correlation between U _Na V and MBF changes. On the whole, the relative contribution to increasing sodium excretion of an increase in GFR and filtered sodium load versus wash‐out of medullary solutes by increasing vasa recta blood flow would be difficult to assess. Moreover, a role of bradykinin's direct inhibition of sodium transport in the distal tubule cannot be excluded (Mamenko et al. 2014). Irrespective of the mechanism underlying the observed increase in natriuresis, it did not influence the MAP profile, which was similar with or without bradykinin's vasodilator action.

In the main AngII group and the HS/UNX rats (Figs 1 and 3) a transient increase in renal excretion was observed after 1 h of Bk infusion, followed by a prompt recovery to control levels. The increasing excretion parameters did not parallel MAP, MBF or RBF, which argues against an interrelationship. In the AngII group the significantly elevated MBF at the end of Bk infusion (Fig. 1) coincided with lowered medullary tissue sodium concentration (Table 4), in accordance with the wash‐out concept. However, these changes were not associated with increased excretion: at the end of Bk infusion U _Na V was even lower than the control pre‐infusion level. The mechanism of the transient increase in renal excretion observed in AngII and less evident in the HS/UNX rats is unclear. Since it was not correlated with renal haemodynamic changes, one can speculate that Bk directly inhibits tubular transport.

There was also no clear indication that Bk administration over 2 weeks increased renal excretion, at least not in the the AngII and the HS/UNX rat models (Figs 4 and 5). On the whole, we found no sound evidence that increasing medullary blood flow over 4 h or even 2 weeks was an important determinant of renal excretion. The similarity in the excretion profiles in Bk‐ and solvent‐infused rats, observed both in acute and chronic studies, argues against an important role of medullary perfusion in the control of tubular reabsorption and excretion.

Concluding remarks

We found that experimental prolonged or chronic renal medullary hyperperfusion in three different rat models of hypertension did not result in reduction of arterial pressure. This was evident even when the MBF increase appeared to reduce the medullary tissue hypertonicity and sodium concentration. Nor did medullary hyperperfusion consistently affect renal output of sodium, water or total solutes. These findings argue against the concept that medullary blood flow is an important controller of renal excretion and arterial blood pressure. Since the results were similar in three models of hypertension with different baseline activities of the renin‐angiotensin system, it is not likely that RAS status was crucial for the absence of a hypotensive response to MBF elevation. On the other hand, the results shift the balance of experimental evidence towards the alternative major theory which ascribes the development of hypertension to hyperactivity of the sympathetic nervous system (SNS). Such hyperactivity is firmly documented for the SHR (Schlaich et al. 2004; Osborn et al. 2005; Esler et al. 2010; Pinterova et al. 2011) and salt‐induced hypertension (Fujita et al. 2007) and is very probable for angiotensin‐induced hypertension (Lohmeier, 2012; Osborn et al. 2007). Thus, our results contest a role of renal medullary perfusion rate in hypertension models with a neurogenic background but not necessarily in models in which the pathogenetic role of the nervous system has not been confirmed.

Additional information

Competing interests

All authors declare that there are no competing financial, personal or professional interests.

Author contributions

All experiments and measurements were performed in the laboratories of the M. Mossakowski Medical Research Centre, Polish Academy of Sciences, Warsaw. B.B. and J.S. conceived and designed the work. B.B. performed acute and chronic experiments and acquired the data. I.B. participated in some of the experiments and performed tissue studies. B.B. analysed and interpreted the data and drafted the manuscript. O.G. performed some of the chemical analytical procedures, and performed and interpreted the statistical analysis. J.S. participated in data interpretation and prepared the final manuscript version. All authors have approved the final version of the manuscript and agree to be accountable for all aspects of the work. All persons designated as authors qualify for authorship, and all those who qualify for authorship are listed.

Funding

This research was financed by National Science Centre, Poland (Grant: 2014/15/B/NZ4/01100).

Biography

Bożena Bądzyńska graduated from the Biological Faculty, University of Warsaw, Poland. She started her career as a research assistant in the Department of Renal and Body Fluid Physiology, Medical Research Centre of the Polish Academy of Sciences, Warsaw, where she continues to work as an associate professor. Her interest has focused on integrative renal physiology, in particular on the control of intrarenal circulation by vasoactive humoral agents such as atrial natriuretic peptide, angiotensin II, prostaglandins or, more recently, bradykinin. The challenge emerging from this research is to separate facts from myths regarding the role of renal medullary perfusion in cardiovascular and body fluid regulation.