Interleukin-6 Knockout Prevents Angiotensin II Hypertension: Role of Renal Vasoconstriction and JAK2/STAT3 Activation

Abstract

Chronic angiotensin II (AngII) infusion stimulates IL-6 release, and we and others have shown that preventing the increase in IL-6 significantly attenuates AngII hypertension. This study measured renal blood flow (RBF) chronically, using Transonic flow probes in wildtype (WT) and IL-6 knockout (KO) mice, to determine the role of renal blood flow regulation in that response. AngII infusion at 200, 800, and 3600 ng/kg/min caused a dose-dependent decrease in renal blood flow in WT mice, and the response at 800 ng/kg/min was compared between WT and IL-6 KO mice. AngII infusion increased plasma IL-6 concentration in WT mice and increased MAP (19 hrs/day; DSI telemetry) from 113±4 to 149±4 mmHg (Δ 36 mmHg) over the 7-day infusion period, and that effect was blocked in IL-6 KO mice (119±7 to 126±7 mmHg). RBF decreased to an average of 61±8% of control over the 7-day period (control = 0.86±0.02 ml/min) in the WT mice; however, the average decrease to 72±6% of control (control = 0.88±0.02 ml/min) in the KO mice was not significantly different. There also was no difference in afferent arteriolar constriction by AngII in blood-perfused juxtamedullary nephrons in WT vs. KO mice. Phosphorylation of JAK2 and STAT3 in renal cortex homogenates increased significantly in AngII-infused WT mice, and that effect was prevented completely in AngII-infused IL-6 KO mice. These data suggest that IL-6-dependent activation of the renal JAK2/STAT3 pathway plays a role in AngII hypertension, but not by mediating the effect of AngII to decrease total renal blood flow.

INTRODUCTION

Evidence linking inflammatory mechanisms to cardiovascular diseases such as atherosclerosis ^1–4 and hypertension ^5–9 continues to build. However, the identity of the precise inflammatory mediators and the mechanisms that underlie their cardiovascular actions remain unclear. The cytokine interleukin-6 (IL-6) is released from vascular tissue in response to angiotensin II (AngII), ^10–13 and our laboratory showed that acute, stress-induced hypertension, which is AngII-dependent, ¹⁴ is attenuated in IL-6 knockout (KO) mice. ¹⁵ Therefore, we tested whether chronic AngII hypertension was dependent on IL-6, and reported significant attenuation of AngII hypertension in IL-6 KO mice. ¹⁶ That was confirmed by Coles et al. in IL-6 knockout mice, and in wildtype mice with blockade of circulating IL-6. ¹⁷ In addition, Harrison’s laboratory ¹⁸ showed that generalized loss of T cell-mediated inflammatory processes significantly attenuated AngII hypertension. These data strongly implicated a role for IL-6 in AngII hypertension.

However, Harrison’s study also showed that DOCA-salt hypertension was blunted by loss of T cells, ¹⁸ and Luther et al. reported that acute stimulation of IL-6 by AngII infusion in human subjects was blocked by spironolactone. ¹⁹ Those reports suggested the effect of IL-6 knockout to attenuate AngII hypertension could be due to an effect of IL-6 on mineralocortocoid-mediated actions. We tested that, and reported that mineralocorticoid hypertension was not attenuated in IL-6 knockout mice, ²⁰ thus supporting a specific link between IL-6 and AngII hypertension.

The goal of this study was to determine whether an effect of IL-6 on AngII-mediated renal vasoconstriction could be a potential mechanism for the dependence of AngII hypertension on IL-6. Interleukin-6 is required for AngII–induced increases in vascular superoxide levels, ²¹ and renal vascular actions in particular are supported by a report from Granger’s laboratory that IL-6 infusion increased plasma renin activity and decreased renal plasma flow and GFR in pregnant rats. ²² Therefore, we used chronically-implanted renal artery flow probes to test the hypothesis that chronic AngII-induced renal vasoconstriction would be attenuated in IL-6 KO mice. In addition, we measured the afferent arteriolar vasoconstrictor response to AngII in anesthetized WT and IL-6 KO mice, using the blood-perfused juxtamedullary nephron preparation, ^{23, 24} and measured renal JAK2 and STAT3 phosphorylation.

METHODS

Procedures involving animals were approved by the Animal Care and Use committee of the Medical College of Georgia. The experiments were conducted in twelve- to fourtheen-week-old (23–28 g) male mice from Jackson Laboratories. The IL-6 KO mice (B6.129S2-Il6^tm1Kopf/J) are on a C57BL/6J background, and those were the WT control mice we used. Mice were allocated to 3 experiments.

Experiment 1: Blood pressure

Mice were assigned randomly to 4 groups: AngII infusion in IL-6 KO (n=6) and WT (n=8) mice, and untreated control IL-6 KO (n=5) and WT (n=5) mice. Biotelemetry transmitter devices (Data Sciences, PA-C10) were implanted in the left carotid artery under isoflurane anesthesia using aseptic technique. Mice were transferred to a light-and temperature-controlled room in the animal facilities, and were housed individually in standard mouse cages with tap water and standard rodent chow available ad libitum. They were given 5–7 days to recover from surgery before Control measurements were made.

Experimental Protocol

After 4 days of stable blood pressure, AngII mice were implanted with micro-osmotic minipumps (Alzet 1007D) that delivered AngII at 800 ng/kg/min, and WT and KO mice were run simultaneously with minipumps filled with the same AngII solution. On the 7th day of infusion, mice were killed to collect kidney tissue and blood samples.

Experiment 2: Renal Blood Flow

There were 2 groups: AngII infusion at 800 ng/kg/min in IL-6 KO (n=6) and WT (n=5) mice. Under isoflurane anesthesia and using aseptic technique, a 0.5 mm Transonic flow probe (0.5PSL) was implanted on the right renal artery via flank incision, with the mouse in a ventral recumbent position, and with the probe cable running subcutaneously from the scapular region where the connector was anchored. To prevent the probe from twisting, we attached a small piece of brass shim stock (~ 3 × 6 mm square) to the underside of the probe using superglue, prior to gas sterilization. This provided a stabilizing plate that was tucked under the kidney and surrounding fat, thus providing much more resistance to twisting. Recovery and housing were the same as described for Experiment 1.

Additional WT mice were run at 200 (n=3) and 3600 (n=8) ng/kg/min to provide comparative information relative to the lowest AngII dose reported by Welch and Wilcox (200 ng/kg/min) ^{25, 26} and the high dose we used in our previous study (3600 ng/kg/min).¹⁶

Experimental Protocol

Renal blood flow (RBF) was measured 4–6 hours each day by placing the mice individually in customized Plexiglas cages with a Dragonfly (model SL-88-10) electrical swivel mounted on top. After 4 days of stable blood flow, all mice were implanted with micro-osmotic minipumps (Alzet 1007D) that delivered AngII at 200, 800, or 3600 ng/kg/min, and the WT and KO mice infused with 800 ng/kg/min were run simultaneously using the same AngII solution. After 7 days of infusion, the minipumps were removed for recovery period measurements. Recovery periods were used as the time control in this experiment to eliminate the need for running additional mice as control groups, and no tissues or plasma samples were taken from these mice.

Experiment 3: Afferent Arteriolar Diameter

Videomicroscopy experiments were conducted in vitro on IL-6 or WT mouse kidneys using the blood-perfused juxtamedullary nephron technique, as previously described. ^{23, 24} For each experiment, a mouse and rat were anesthetized with sodium pentobarbital (50 mg/kg; i.p.). Perfusate blood was collected from a rat because it is not possible to obtain sufficient blood from blood donor mice for kidney perfusion. Arteriolar diameters were recorded at 12-second intervals. Sustained afferent arteriolar diameter was calculated from the average of measurements made during the final 2 minutes of each treatment period. Following a 15-min equilibration period, baseline afferent arteriolar diameters were measured. Each kidney was treated with increasing concentrations of AngII (10⁻⁷ to 10⁻⁵ mol/L) and diameter was monitored for 5 minutes/concentration.

Analytical Methods

Blood pressure measurement

Mouse cages were placed individually on Data Sciences receivers, and pulsatile arterial pressure was recorded from 1500 hours to 1000 hours (i.e. 19 hours) each day. Analog signals from the transmitters were sampled for 5 seconds every 1–2 minutes at 500 Hz, and the average of those measurements was recorded as the daily mean arterial pressure (MAP) for each animal.

Renal blood flow measurement

Mice were connected 4–6 hours per day to the Dragonfly swivel, which in turn was connected to a Transonic TS402 series flowmeter, and the signal was sampled continuously at 100 Hz using PowerLab and a Macintosh computer. The average of the entire collection period was used as that day’s RBF, and mice were excluded from the study if the flow signal was not pulsatile.

Plasma IL-6

Plasma IL-6 concentrations were measured by enzyme immunoassay (R&D Systems) from blood samples obtained by ventricular puncture under isoflurane anesthesia.

JAK2/STAT3: Tissue homogenization for protein work

Tissues were quick frozen with liquid nitrogen, pulverized in a liquid nitrogen-cooled mortar and pestle and solubilized in a 255 mM sucrose/10mM Tris buffer (pH 7.4) with protease inhibitors (0.5 mM PMSF, 2mM EGTA, 10 μg/ul aprotinin and 10 μg/ml leupeptin) and tyrosine phosphatase (1 mM sodium orthovanadate) inhibitors. Homogenates were centrifuged (14,000g for 10 minutes, 4°C) and supernatant total protein (Bio-Rad) was measured.

JAK2/STAT3: Western Blotting

Supernatant were separated on SDS-polyacrylamide gels (7.5 % SDS-PAGE) and transferred to Immobilon–P membrane. Membranes were blocked and probed overnight (4°C) with primary antibody (1:1000, phospho-specific JAK2, and phospho-specific STAT3, Cell Signaling). Blots were washed, and an anti-rabbit horseradish peroxidase –linked secondary antibody (1:7500, Amersham Labs) was added for one hour and incubated with the blots at 4°C. Blots were washed and enhanced chemiluminescence (Super Signals Ultra, Pierce) was used to visualize labeled bands. Blots were stripped and reprobed with the total antibody (1:1000, JAK2 BioSource, STAT3 BD Transduction Labs). Beta Actin was used to ensure equal total protein loading between lanes. Band density was quantified using the program NIH Image.

Statistical Analysis

Time- and treatment-dependent changes in MAP and RBF were analyzed with repeated measures ANOVA. Significant F-values were followed by Dunnett’s test to determine specific within-group differences, and completely randomized ANOVA plus Fisher’s PLSD test on each day to determine specific between-group differences. JAK2/STAT3 data were analyzed with completely randomized ANOVA. Within group analysis of AngII effects on afferent arteriolar diameter were made by one-way analysis of variance for repeated measurements combined with the Dunnett’s multiple range test. Across group comparisons were made using Student’s t-test for unpaired data. Significance was p < 0.05, and data are expressed as mean ± SEM.

RESULTS

Experiment 1: Blood pressure

Mean arterial pressure (MAP) averaged 113±4 mmHg in WT mice during the control period and increased gradually during infusion of AngII at 800 ng/kg/min, plateauing after 3–4 days and averaging 149±4 mmHg (Δ36 mmHg) by day 7 (Figure 1). Control period MAP in the KO mice averaged 119±7 mmHg, and there was no significant change in MAP during AngII infusion (average on day 7 was 126±7 mmHg). There also was no change in MAP in the WT or KO control groups during the experiment (Figure 1).

FIGURE 1.

Mean arterial pressure (mean±SEM) in WT and IL-6 KO mice infused with angiotensin II (AngII) at 800 ng/kg/min and in control WT and KO mice during the control (C) period and AngII treatment (T) period. * = p < 0.05 vs baseline (within-group).

Experiment 2: Renal Blood Flow

In the mice infused with AngII at 800 ng/kg/min, renal blood flow (RBF) averaged 0.86±0.02 and 0.88±0.02 ml/min in WT and KO, respectively, during the control period, and decreased significantly in both groups after starting AngII infusion, averaging 61±8% and 72±6% of control over the 7-day period in WT and KO, respectively (Figure 2). However, despite that slightly lesser decrease in the KO mice and slight separation between KO and WT mice at days 1, 5, and 6, there was no statistically significant between-group difference between the WT and KO mice in the repeated measures ANOVA for those 2 groups only, or when their areas-over-the-curve were compared by t-test. When all 4 groups were tested together by ANOVA, there was a significant between-group F-value, but still no difference on any day between the WT and KO mice at 800 ng/kg/min. Their areas-over-the-curve also were not different when all 4 groups were compared by completely randomized ANOVA. The ANOVA for all 3 AngII doses in WT mice, on the other hand, revealed a significant effect of AngII dose on RBF (p<0.001). There was no decrease in RBF in the WT Ang200 mice, i.e. no within-group difference, consistent with previous reports, ^{25, 26} but RBF decreased significantly beginning on day 1 of AngII in all other groups.

FIGURE 2.

Renal blood flow expressed as percent control (mean±SEM) in WT and IL-6 KO mice during the control (C) period, AngII treatment (T) period, and recovery (R) period. WT mice were infused with AngII at 200, 800, or 3600 ng/kg/min, and the IL-6 KO mice were infused with AngII at 800 ng/kg/min. * = p < 0.05 vs baseline (within-group); § = p < 0.05 vs. 800 dose (between-group); # = p < 0.05 vs. 3600 dose (between-group).

Experiment 3: Afferent Arteriolar Diameter

AngII significantly decreased afferent arteriolar diameter in WT and IL-6 KO mice beginning with the lowest concentration (0.1 nM; Figure 3). There was a linear decrease in diameter with increasing AngII concentration in both groups, reaching a nadir of approximately 80% of control diameter at an AngII concentration of 10 nM, and there were no differences between WT and KO mice at any concentration. Diameter returned to baseline in both groups after AngII was removed.

FIGURE 3.

Afferent arteriole diameter in blood-perfused juxtamedullary nephrons from WT and IL-6 KO mice in response to AngII. * = p < 0.05 vs control (within-group).

Plasma IL-6 and JAK2/STAT3

Plasma IL-6 was below detection in all IL-6 KO mice and in the control WT mice, and averaged 27.7±5.1 pg/ml in the WT mice infused with AngII at 800 ng/kg/min. Seven WT mice infused with AngII at 3600 ng/kg/min were added to the assay to provide comparison with our previous report, ¹⁶ and indeed there was a significantly greater increase in plasma IL-6 concentration in those mice (48±9 pg/ml). Figure 4 shows significant activation (phosphorylation) of JAK2 and STAT3 in renal cortex of WT mice infused with AngII at 800 ng/kg/min, and that effect was completely absent in the IL-6 KO mice.

FIGURE 4.

JAK2 and STAT3 phosphorylation in control WT mice, AngII infused WT mice (800 ng/kg/min), control IL-6 KO mice, and AngII infused IL-6 KO mice (800 ng/kg/min). * = p < 0.05 vs control.

DISCUSSION

The main findings from this study are that chronic AngII infusion in WT mice on normal salt intake caused dose-dependent stimulation of IL-6, significant hypertension, renal cortical JAK2/STAT3 phosphorylation, and dose-dependent decreases in RBF and afferent arteriole diameter. Interleukin-6 knockout completely prevented the hypertension and JAK2/STAT3 activation without causing significant attenuation of the renal vasoconstrictor response. These data show that AngII-induced renal vasoconstriction does not require IL-6, and suggest that other actions likely contribute to the effect of IL-6 knockout to prevent AngII hypertension.

In the first study that showed attenuation of AngII hypertension by knockout of IL-6, we used a high AngII infusion dose of 90 ng/min (~ 3600 ng/kg/min) in mice on 4% salt intake. Therefore one goal of this study was to measure renal blood flow, and also blood pressure, under less severe conditions. However, it still was necessary to choose an AngII dose that would decrease renal blood flow over the time course (< 1 week) ^{16, 17} where we knew IL-6 knockout began to attenuate AngII hypertension. Welch and Wilcox showed that AngII infusion at 400 ng/kg/min in mice did not cause hypertension or decrease renal blood flow by day 6. ²⁵ Therefore, we used 800 ng/kg/min, which also is similar to the dose used by Coles et al.. ¹⁷

We found that AngII infusion at that dose decreased RBF by approximately 40% in WT mice over the 7-day infusion. That was surprising in itself, because even though AngII is a well-known renal vasoconstrictor, we are not aware of any previous chronic RBF measurements in AngII hypertension using a renal flow probe, and the magnitude and sustained nature of the decrease were remarkable. That result was one reason for adding WT mice at the 200 and 3600 ng/kg/min AngII doses, and indeed the chronic renal vasoconstrictor effect of AngII was dose-dependent. However, the principal finding in this study is that IL-6 knockout did not significantly attenuate the decrease in renal blood flow at the 800 ng/kg/min dose. On the other hand, it also is apparent that the RBF plots in the KO versus WT mice are not superimposable, and that tends to focus attention on a comparison between those two RBF responses. The more remarkable observation, however, is that AngII hypertension was prevented completely in the IL-6 KO mice, yet they still had a significant decrease in renal blood flow.

This lack of effect of IL-6 KO on AngII-induced renal vasoconstriction was surprising in light of the reported effect of IL-6 infusion to increase blood pressure and cause renal vasoconstriction in pregnant rats. ²² In addition, IL-6 is linked to pulmonary hypertension, ^{27, 28} it is required for AngII-mediated increases in vascular superoxide, ²¹ and we showed that acute-stress hypertension in mice, which is AngII dependent, ¹⁴ is blunted in IL-6 KO mice. ¹⁵ Crosswhite and Sun also have just reported that knockdown of IL-6 decreased blood pressure and vascular superoxide production in rats with cold-induced hypertension, ²⁹ which is an AngII-dependent model. ³⁰ Moreover, for IL-6 KO not to attenuate the decrease in RBF, yet completely block the hypertension, means that some decrease in tubular sodium reabsorption must have occurred during AngII infusion in those mice, and we do not have an explanation for that. Yet, despite these arguments, and even if one speculates that additional mice in the IL-6 KO group would have yielded a statistically significant difference between the WT and KO mice, there still would remain a significant decrease in RBF caused by AngII infusion in the IL-6 KO mice. The use of chronically-implanted flow probes, the recovery data, and the lack of effect of AngII infusion at 200 ng/kg/min, assuage concerns about the validity of that finding.

However, if the blood pressure and blood flow data from Experiments 1 and 2 are used to estimate renal vascular resistance during AngII infusion, it would be lower in the IL-6 KO mice (lower MAP with no difference in RBF). Therefore, even though AngII significantly decreased RBF in the IL-6 KO mice, the lesser increase in renal vascular resistance could indicate impaired vasoconstrictor responses to AngII in IL-6 KO mice. The data from Dr. Inscho’s laboratory (Figure 3) argue against that, by showing no difference in the afferent arteriolar constrictor effect of AngII in WT vs. IL-6 KO mice. In addition, Coles et al. ¹⁷ showed there was no difference in the contractile response to AngII in aortic rings isolated from WT vs. IL-6 KO mice. These data are consistent with our data that showed no significant difference in the AngII-induced decrease in RBF in WT vs. IL-6 KO mice. This suggests that other factors, such as autoregulation, may have contributed to greater estimated renal vascular resistance in the hypertensive WT mice.

These results do not completely rule out potential vascular actions of IL-6, such as mediating effects of AngII on filtration fraction or medullary blood flow. The fact that the RBF responses in WT vs KO mice at 800 ng/kg/min were not superimposable suggests there even may have been mild attenuation of more global AngII-induced vasoconstriction that simply was not statistically significant at this infusion dose. In addition, comparing the RBF and MAP responses to AngII in this study and previous studies ^{16, 17, 25, 26} suggests there could be different thresholds for IL-6 action on AngII hypertension versus renal vasoconstriction. Therefore, it is possible that there is a level of AngII hypertension that is associated with, and perhaps dependent upon, IL-6-mediated vasoconstriction.

The effect of IL-6 KO to prevent AngII hypertension without preventing the decrease in RBF nonetheless is strong evidence that renal vasoconstriction likely is not the primary explanation for the blood pressure effect. However, the inhibition of renal JAK2/STAT3 phosphorylation by IL-6 KO still implicates a role for the kidneys. JAK2/STAT3 signaling has been shown to be an early upstream component of the renal and glomerular actions of AngII in diabetes, ^31–33 and Guilluy et al. recently reported that AngII hypertension was attenuated by chronic blockade of JAK2 activation. ³⁴ Satao et al. showed that AngII alone did not cause STAT3 phosphorylation in a cultured human proximal tubular cell line, but did potentiate the stimulatory action of IL-6. ³⁵ This suggests IL-6 could play a role in the effect of AngII on tubular sodium reabsorption, and it is consistent with evidence that IL-6 activates collecting duct ENaC ³⁶ and increases sodium-glucose cotransporter expression and STAT3 phosphorylation in primary cultured renal proximal tubular cells. ³⁷ There is no direct evidence that IL-6 contributes to AngII-mediated regulation of tubular reabsorption, but it is intriguing to consider based on those studies, our in vivo evidence that renal JAK2 and STAT3 are activated in AngII hypertension and require IL-6, and our evidence that IL-6 knockout can attenuate AngII hypertension without blocking renal vasoconstriction.

PERSPECTIVES

The effect of T cell elimination to attenuate AngII and mineralocorticoid hypertension, together with the effect of IL-6 elimination to attenuate only AngII hypertension, suggests that branching down from more global levels of the inflammatory response reveals more specific relationships between inflammatory mediators and certain types of hypertension. Likewise, our findings that IL-6 knockout can completely block AngII-induced JAK2/STAT3 activation and hypertension without blocking AngII-induced decreases in RBF suggest that inflammatory mediators, such as IL-6, may have dose- and tissue-specific mechanisms for controlling the response to hypertensive stimuli. That also is supported by the effect of IL-6 knockout to cause ~ 50% inhibition of AngII hypertension at the 3600 ng/kg/min dose, ¹⁶ but completely block it at the 800 ng/kg/min dose. It will be important now to determine whether the in vitro evidence supporting renal tubular actions of IL-6 translate to effects on AngII-induced increases renal sodium reabsorption in vivo, and determine whether such effects are amplified at lower AngII infusion doses.