Vascular α₁-adrenergic sensitivity is enhanced in chronic kidney disease

Abstract

Chronic kidney disease (CKD) is often complicated by difficult-to-control hypertension, in part due to chronic overactivation of the sympathetic nervous system (SNS). CKD patients also exhibit a greater increase in arterial blood pressure for a given increase in sympathetic nerve activation, suggesting an augmented vasoconstrictive response to SNS activation (i.e., neurovascular transduction). One potential mechanism of increased sympathetic neurovascular transduction is heightened sensitivity of the vascular α₁-adrenergic receptors (α₁ARs), the major effectors of vasoconstriction in response to norepinephrine release at the sympathetic nerve terminals. Therefore, we hypothesized that patients with CKD have increased vascular α₁AR sensitivity. We studied 32 patients with CKD stages III and IV (age 59.9 ± 1.3 yr) and 19 age-matched controls (CON, age 63.2 ± 1.6 yr). Using a linear variable differential transformer (LVDT), we measured change in venoconstriction in response to exponentially increasing doses of the selective α₁AR agonist phenylephrine (PE) administered sequentially into a dorsal hand vein. Individual semilogarithmic PE dose-response curves were constructed for each participant to determine the PE dose at which 50% of maximum venoconstriction occurred (ED₅₀), reflecting α₁AR sensitivity. In support of our hypothesis, CKD patients had a lower PE ED₅₀ than CON (CKD = 2.23 ± 0.11 vs. CON = 2.63 ± 0.20, P = 0.023), demonstrating increased vascular α₁AR sensitivity. Additionally, CKD patients had a greater venoconstrictive capacity to PE than CON (P = 0.015). Augmented α₁AR sensitivity may contribute mechanistically to enhanced neurovascular transduction in CKD and may explain, in part, the greater blood pressure reactivity exhibited in these patients.

INTRODUCTION

Chronic kidney disease (CKD) affects ~15% of the US population and is independently associated with an increased risk of cardiovascular disease (CVD) and mortality (10, 17, 20, 25). The majority of CKD patients have comorbid hypertension (27) that is often difficult to control and contributes to CVD risk. One major mechanism contributing to both increased blood pressure and CVD risk in CKD is chronic activation of the sympathetic nervous system (SNS). Prior studies have shown that CKD patients have higher plasma norepinephrine levels (6, 16) and elevated resting muscle sympathetic nerve activity (MSNA) that begins at early stages of CKD and are progressively more elevated with worsening renal function (13, 18, 23). Mechanisms underlying chronic SNS overactivity in the setting of reduced renal function include renal afferent nerve activation (12), chronic inflammation (5, 33), and activation of the renin-angiotensin system (24).

While it is clear that sympathetic nerve activity is chronically elevated in CKD (13, 18, 23), our understanding of the vasoconstriction that takes place in response to sympathetic nerve activation, i.e., sympathetic neurovascular transduction, in this population is currently limited. Our previous work showed that CKD patients have an exaggerated increase in blood pressure for the same degree of increase in MSNA during sympathoexcitation induced by handgrip exercise compared with controls (22). Furthermore, prior studies have shown that CKD patients have a greater systemic blood pressure response during the administration of phenylephrine (PE), a selective α₁AR agonist (4). Whereas chronic SNS overactivity might be expected to lead to a relative decrease in vasoconstriction in response to SNS activation in an effort to maintain physiological homeostasis, these observations suggest that neurovascular transduction may in fact be increased in CKD.

One mechanism that may underlie increased sympathetic neurovascular transduction is increased vascular α₁AR sensitivity. The α₁ARs are the major mediators of vasoconstriction in response to norepinephrine (NE) release from sympathetic nerve terminals, and therefore plays a major role in blood pressure regulation. Vascular α₁AR sensitivity can be examined in vivo in humans through use of a linear variable differential transformer (LVDT) that directly measures the degree of venoconstriction in response to exponentially increasing doses of the selective α₁AR agonist PE from a dorsal hand vein. This technique has previously been used to study vascular α₁AR sensitivity in various conditions including aging (15, 21), pregnancy (19), and hypertension (8), and has been shown to be reproducible in end-stage renal disease (ESRD) (1). Importantly, while the LVDT method quantifies venous α₁AR sensitivity, this has been shown to reflect arterial α₁AR sensitivity of the resistance arterioles (28). Furthermore, as opposed to intra-arterial infusions of PE, the LVDT method utilizes small doses of PE that act locally without affecting systemic blood pressure, avoiding confounding effects of baroreflex engagement, and allows for the generation of individual dose-response curves to assess α₁AR sensitivity. We hypothesized that patients with CKD stages III and IV have greater vascular α₁AR sensitivity compared with age-matched hypertensive controls.

METHODS

Ethical approval.

This study was approved by the Atlanta Veterans Affairs (VA) Health Care System Research and Development Committee and the Emory University Institutional Review Board. Written informed consent was obtained for all study participants and all study procedures conformed to the standards set forth by the Declaration of Helsinki.

Participants.

Participants reported to the human physiology laboratory at the Atlanta VA Health Care System Hospital after abstaining from food, alcohol, and caffeine for at least 12 h, and exercise for a minimum of 24 h. Participants treated with β-blockers and α-blockers held these medications for 24 h before experimentation; other medications were taken as prescribed. All participants emptied their bladders before experimentation to control for the potential confounding effects of SNS activation with bladder distension (9). Exclusion criteria for both groups included uncontrolled hypertension (BP>160/90 mmHg), vascular disease, use of clonidine, clinical evidence of heart failure or heart disease determined by electrocardiogram (ECG) or echocardiogram, ongoing drug or alcohol abuse within the past 12 mo, diabetic neuropathy, regular participation in exercise (>20 min twice per week) and pregnancy or plans to become pregnant.

Instrumentation.

Blood pressure measurements at rest were performed in triplicate using American College of Cardiology/American Heart Association (ACC/AHA) guidelines via an automated device while the subjects were seated quietly (Omron, Hem907XL, Hoffman Estates, IL) (32). A blood sample was collected from an antecubital vein for assessment of a basic metabolic panel. Subsequently, participants were moved to the semi-recumbent position on a reclining hospital bed. A 24-gauge intravenous catheter was placed into a suitable dorsal hand vein and normal saline (0.9% NaCl; NS) was continuously infused at a rate of 0.6 mL/min to ensure patency of the venous line. The hand was positioned above heart level to ensure complete emptying of superficial veins at rest. A blood pressure cuff was placed on the upper arm to induce intermittent venous distension. A small tripod holding the moveable central core of the LVDT (model MHR 100; Shaevitz) was mounted on the dorsal hand and secured with tape. The LVDT movable central core measures changes in vascular diameter during venous distension induced by inflating an upper arm cuff to 50 mmHg. The central core was placed directly over the dorsal hand vein, ~1.0 cm distal from the end of the intravenous catheter. The LVDT was connected to data acquisition hardware (PowerLab, AD Instruments, Bella Vista, NSW, Australia) to record changes in voltage as the central core moved vertically with venous distension.

Experimental protocol.

At least 30 min of rest was given after placement of the intravenous catheter to allow for the vein to recover from needle insertion. For baseline measurements, NS was infused at 0.6 mL/min and venous distension was induced by inflating the upper arm cuff for 3 min. This inflation caused an upward inflection on the LVDT signal relative to baseline and the resultant plateau was interpreted as maximal venous distension. Following three stable baseline measurements during NS infusion to determine baseline maximum vascular distension, exponentially increasing doses of the selective α₁AR agonist PE ranging from 15 to 12,000 ng/min were infused in a constant volume of 0.6 ml/min for 7 min per dose via a minisyringe infusion pump (Harvard Apparatus, Holliston, Massachusetts). The upper arm cuff was inflated to 50 mmHg to induce venous filling during the last 3 min of PE infusion at each dose. Similar to the baseline measurements, cuff inflation during PE infusion also resulted in an upward inflection in the LVDT signal; however, the magnitude of this change was decreased since the vein was unable to distend to the same extent that it did at baseline due to the presence of PE. A total of 10 doses of PE were delivered unless maximal venoconstriction was achieved earlier as evidenced by no venous distension with upper arm inflation, or plateauing of venoconstriction elicited during 2 or more consecutive doses of PE. All participants received a minimum of 5 doses.

Data analysis.

Vascular diameters were expressed as a percent reduction from baseline maximum venodilation, plotted against PE dose rates in individual semi-logarithmic dose-response graphs and analyzed using a 4-variable sigmoid dose-response model using Sigma Plot 13 (Systat Software). The log dose that produced a 50% of maximal venoconstriction (ED₅₀) reflects sensitivity to PE (i.e., α₁-AR sensitivity); lower ED₅₀ reflects higher sensitivity, whereas a higher ED₅₀ reflects lower sensitivity. To facilitate curve-fitting for all participants, the minimum venoconstriction was constrained to 0% and the maximum venoconstriction (E_max) was constrained to 100% for each individual participant’s curve (31), since it is not physiologically possible to venoconstrict more than 100% (i.e., empty vein). Since the absolute values for E_max cannot be accurately determined when curves are constrained, a separate analysis was performed to assess maximum venoconstrictive capacity between groups as a secondary outcome. For this analysis, a linear mixed model was performed to compare the percent change in venoconstriction between groups without the use of logistic modeling.

Statistics.

Demographic data between groups were compared via unpaired, two-tailed, t-tests for continuous variables or χ² analysis for categorical variables. For the primary analysis, unpaired, two-tailed t-tests were used to compare ED₅₀ values between groups. For the secondary analysis of maximum venoconstrictive capacity, a two-factor (group by dose) linear mixed model was used to compare responses over time and between groups as previously described (7). All data are expressed as means ± SE, and exact P values are reported.

RESULTS

Participants.

37 CKD participants and 23 age-matched controls were enrolled for study participation. We were unable to collect usable data from 3 CKD participants and 2 CON due to excessive hand movement during the experiment which precluded our ability to analyze the LVDT signal. Additionally, data collected from four participants were excluded (2 CKD, 2 CON) due to the inability to fit acceptable sigmoidal PE dose-response curves in these participants; therefore, results are reported for N = 32 for CKD and N = 19 for CON. Participant demographic data are presented in Table 1. There were no differences in age, sex, race, weight, blood pressure, smoking status, comorbidities, or medication use between groups (Table 1). Mean serum [creatinine] was 1.8 ± 0.09 mg/dl in the CKD group and 1.0 ± 0.04mg/dl in the CON group as expected. Thirty CKD patients had stage III CKD, while 2 of the CKD patients had stage IV CKD. There were no differences in serum sodium, potassium, chloride, or bicarbonate concentrations between groups, although plasma [glucose] was greater in the CKD group compared with CON (106.1 ± 5.4 vs. 86.3 ± 4.5 mg/dl, P = 0.02).

Table 1.

Participant demographic data

Characteristic	CON	CKD	P
n	19	32
Age, yr	59.9 ± 1.3	63.2 ± 1.6	0.15
Systolic blood pressure, mmHg	128.3 ± 3.6	130.2 ± 2.8	0.68
Diastolic blood pressure, mmHg	80.2 ± 2.1	78.4 ± 1.8	0.54
Mean arterial pressure, mmHg	96.2 ± 2.3	95.7 ± 1.9	0.86
Serum creatinine, mg/dL	1.0 ± 0.04	1.8 ± 0.09	<0.0001
Sex (M/F)	14/5	27/5	0.36
Race, n (%)			0.41
Black	18 (94.7%)	27 (84.4%)
White	1 (5.3%)	5 (15.6%)
Body mass index, kg/m2	29.0 ± 1.3	31.0 ± 1.1	0.27
Diabetes mellitus	3 (15.8%)	9 (28.1%)	0.37
Hypertension	13 (68.4%)	28 (87.5%)	0.18
Antihypertensive medications, n (%)
Calcium channel blockers	4 (21.1%)	16 (50.0%)	0.06
ACE inhibitors/ARBs	7 (36.8%)	20 (62.5%)	0.11
Diuretics	5 (26.3%)	9 (28.13%)	0.98
β-Blockers	4 (21.1%)	13 (40.1%)	0.20
α-Blockers	1 (5.26%)	2 (6.5%)	0.92
Hydralazine	0 (0%)	2 (6.3%)	0.48

Values are means ± SE or as otherwise indicated. ACE, angiotensin-converting enzyme, ARB, angiotensin receptor blocker.

Vascular α₁-adrenergic receptor sensitivity.

All participants in both groups exhibited the expected response of increasing venoconstriction with increasing doses of PE. Representative dose-response curves are displayed in Fig. 1. Individual dose-response curves revealed wide variability between participants in the log ED₅₀ values for both groups (CON range = 0.85–4.25, CKD range = 1.16–3.51, Fig. 2). Comparison of mean values revealed that CKD patients had a lower mean PE ED₅₀ compared with CON (Fig. 2, P = 0.023), demonstrating increased vascular α₁AR sensitivity.

Fig. 1.

Representative dose-response curves plotting percent constriction as a function of phenylephrine (PE) dose for control (CON; left) and chronic kidney disease (CKD; right) participants.

Fig. 2.

Mean log ED₅₀ values between groups. CON, control. CKD, chronic kidney disease; n = 19 CON (14 M/5 F) and 32 CKD (27 M/5 F). Mean log ED₅₀ values were compared via unpaired, 2-tailed t-tests. *P = 0.023.

Maximal venoconstrictive capacity.

Because E_max was constrained to 100%, absolute values of E_max are not reported. To further evaluate maximal venoconstrictive capacity, a linear mixed model analysis was performed for the cumulative data between groups and revealed a greater slope of rise in the degree of venoconstriction in response to increasing doses of PE, and a greater degree of venoconstriction in response to the same dose of PE at higher concentrations in CKD compared with CON (CKD = 26.3 ± 1.8, CON = 18.8 ± 1.6% constriction/log PE dose, P = 0.015).

DISCUSSION

In this investigation, we compared vascular α₁AR sensitivity between patients with CKD stages III and IV versus age- and comorbidity-matched controls. While it was previously established that patients with CKD have chronic overactivation of the sympathetic nerves (13, 18, 23), we now report the novel finding that venoconstriction in response to sympathetic nerve activation is also enhanced in CKD. Specifically, CKD patients exhibit greater vascular α₁AR sensitivity, as evidenced by a lower ED₅₀ in response to increasing doses of the selective α₁AR agonist PE. Additionally, we also showed that maximum venoconstrictive capacity is higher in CKD. This combination of heightened sympathetic nerve activity with an augmented AR sensitivity at the level of the vasculature likely has an additive effect on increasing blood pressure and CVD risk in patients with reduced renal function, and could explain, in part, why hypertension is difficult to control in this population.

The current findings suggest that increased neurovascular transduction of SNS activity in CKD is less likely due to greater release of NE from sympathetic nerves, but rather increased sensitivity at the receptor level. As early as 1982, Beretta-Piccoli et al. (4) observed a greater pressor response to intravenous NE infusion in CKD compared with healthy controls. Interestingly, there were no differences in pressor responses to angiotensin II between groups, suggesting that this reactivity was specific to an adrenergic mechanism. We have now expanded on these findings by implicating the α₁AR as the culprit in the enhanced pressor response selectively to NE. This observation may seem counterintuitive to the concept that receptors downregulate in response to chronic stimulation. For example, since CKD patients have chronically elevated resting MSNA, this would presumably lead to a compensatory decrease in the sensitivity of the α₁AR to maintain homeostasis. Although concomitant measurements of MSNA and α₁AR were not performed in this study, it appears that this expected decrease in α₁AR sensitivity is not present, since α₁AR sensitivity was in fact enhanced in the CKD group. Chronic elevation of sympathetic nervous activity, combined with enhanced (rather than desensitized) α₁AR sensitivity likely has an additive effect on cardiovascular risk in patients with reduced renal function. This finding of combined elevations in sympathetic nerve activity and enhanced sensitivity at the AR level is not unique to CKD, as increased resting MSNA and augmented vascular α₁AR sensitivity have also been observed in African-Americans (2, 26, 29). Potential mechanisms contributing to enhanced α₁AR sensitivity and the observed increase in venoconstrictive capacity include synergistic actions of other circulating hormones such as angiotensin II (3, 30), sensitization of the α₁AR by free radicals (11), or uremic metabolites that are known to accumulate in CKD. All of these scenarios are possible in the setting of CKD, and future work should seek to elucidate the precise mechanisms contributing to this enhanced α₁AR sensitivity, and the sensitivity of other subtypes of ARs (e.g., α₂ or β₂) in CKD. Additionally, it is currently unclear whether this enhanced sensitivity is due to alterations in adrenergic density, second messengers, or other molecular mechanisms within the vascular smooth muscle. Future work should seek to further explore this question. It is important to note that we measured venous rather than arterial α₁AR sensitivity. A prior study directly compared venous with arterial α₁AR sensitivity in vivo in healthy humans and reported a correlation (r = 0.70, P < 0.02) between the two vessels (28); therefore, sensitivity of venous α₁ARs is thought to reflect that of arterioles. In addition, venous adrenergic sensitivity itself is clinically relevant, given that venous capacitance is important for blood volume distribution, cardiac output, and blood pressure regulation (14).

An important point that remains to be clarified is whether CKD leads to enhanced α₁AR sensitivity or whether enhanced α₁AR sensitivity promotes the development of CKD. One prior human study showed that α₁AR sensitivity was enhanced in patients with end-stage renal disease on erythropoietin therapy compared with a group of much younger, healthy controls. However, it was unclear whether differences in medication use, demographics, or comorbidities accounted for differences in α₁AR sensitivity or when vascular sensitivity changes might occur in relation to disease onset or progression (1). The current study shows that patients with CKD stages III and IV, including patients with relatively minimal decrements in estimated glomerular filtration rate, have enhanced vascular α₁AR sensitivity compared with well-matched hypertensive controls, suggesting that renal dysfunction is independently associated with increased α₁AR sensitivity and that these vascular derangements begin early in the course of renal disease. Perturbations in vascular sensitivity and venoconstrictive capacity occurring early in disease development may exacerbate disease progression through accelerating development of hypertension or through direct renal effects. Additionally, although these findings provide mechanistic insights into the increased sympathetic neurovascular transduction in CKD, it is unclear whether enhanced vascular α₁AR sensitivity can be modulated. Eichler et al. (8) demonstrated an upward shift of the PE ED₅₀ in response to 2–4 wk of prazosin therapy in hypertensive males, suggesting that α₁AR sensitivity can be attenuated through α₁AR blockade. Additionally, previous studies have demonstrated that angiotensin II can sensitize the α₁AR (3, 30). Since the renin-angiotensin system is chronically elevated in CKD (24), it is also possible that angiotensin receptor blockers may attenuate α₁AR sensitivity. Future work should seek to determine the extent to which augmented vascular α₁AR sensitivity can be ameliorated in CKD, through both pharmacological (i.e., α₁ or angiotensin receptor blockade) as well as biobehavioral (i.e., diet and exercise) interventions.

There are several methodological considerations that should be mentioned, as they relate to interpretation of our findings. First, our control group was not free from disease and by design included participants with hypertension and diabetes mellitus. Although this may be viewed as a strength because these comorbidities are also present within the CKD group, allowing for assessment of the independent effects of renal disease on α₁AR sensitivity, the inclusion of an additional healthy control group would have further allowed for comparisons of α₁AR sensitivity to be made in relation to disease onset. Second, maximal venoconstrictive capacity was assessed via a linear mixed model rather than direct comparisons of E_max derived from individual dose-response curves. Constraining E_max to 100% reflects physiological reality (veins cannot be more constricted than 100%) and is an accepted strategy for modeling dose-response curves but will not accurately reflect absolute values of E_max (31). Third, we measured α₁AR sensitivity in the hand and interpreted these findings to be representative of systemic α₁AR sensitivity. It is possible that this measurement may not be reflective of α₁AR sensitivity in other limbs (i.e., the legs). Additionally, we observed considerable variability in the ED₅₀ within the control group. This large intersubject variability in the ED₅₀ has previously been reported in hypertensive patients (8) and may be related to the multiple comorbidities (e.g., diabetes mellitus and hypertension) as well as medication use that was included in this group. Interestingly, our observation that this variability seemed to decrease in our CKD group may suggest that the pathological processes within CKD that lead to enhanced α₁AR sensitivity are rather consistent across subjects despite the presence of comorbidities and medication use in this group. Finally, it is not clear whether CKD leads to increased α₁AR, or α₁AR sensitivity leads to CKD, or whether there is a shared mechanism contributing to both. However, this phenomenon does not appear to be due to hypertension, since groups were well matched for hypertension and antihypertensive medication. In addition, a prior study had demonstrated no difference in the PE ED₅₀ between hypertensive subjects and normotensive controls (8). This finding, combined with the fact that hypertension status and medication use were carefully balanced between the groups suggests that this enhanced sensitivity likely resulted from CKD, rather than being secondary to hypertension.

In conclusion, we have demonstrated that vascular α₁AR sensitivity is enhanced in CKD. We further report that maximal venoconstrictive capacity is also higher in CKD compared with CON. These findings provide mechanistic insights related to abnormal neurocirculatory control in CKD that contributes to difficult-to-control hypertension in this population.

Perspectives and Significance

Patients with CKD exhibit increased MSNA at rest and often present with comorbid hypertension that is difficult to control. We have extended these findings and have implicated the α₁AR as being an additional component that may contribute to hypertension and cardiovascular risk. Moreover, the combination of increased MSNA and sensitivity of α₁AR likely may have an additive effect on elevated CVD risk in CKD. Future work should delineate the mechanisms underlying augmented vascular α₁AR sensitivity and its role in CKD progression and whether it can be modulated through treatment.

GRANTS

Research was supported by the National Heart Lung and Blood Institute Grant R01 HL-135183 (PI: J. Park) and the National Institute of Diabetes and Digestive and Kidney Diseases (T32DK007656, PI: J Sands); and the Department of Veterans Affairs, Veterans Health Administration, Office of Research and Development, Clinical Studies Center, Decatur, Georgia.

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the author(s).

AUTHOR CONTRIBUTIONS

J.P. conceived and designed research; J.D.S., D.L.M., D.R.D., and J.P. performed experiments; J.D.S., D.L.M., and Y.L. analyzed data; J.D.S., C.M.S., and J.P. interpreted results of experiments; J.D.S. prepared figures; J.D.S. drafted manuscript; J.D.S., D.L.M., Y.L., S.Y.P., I.T.F., and J.P. edited and revised manuscript; J.D.S., D.L.M., C.M.S., Y.L., S.Y.P., I.T.F., D.R.D., and J.P. approved final version of manuscript.