Reactive Hyperemia Occurs via Activation of Inwardly-Rectifying Potassium Channels and Na⁺/K⁺-ATPase in Humans

Subjects

With Institutional Review Board approval and after written informed consent, a total of 24 young healthy adults [18 men, 6 women; age=23±1 years (range:18-34 years); weight=73.1±1.5 kg; height=175±1 cm; body mass index=23.9±0.5 kg/m²; forearm volume (FAV)=945±39 ml; means±s.e.m.] participated in the present study. All studies were performed according to the Declaration of Helsinki.

Arterial catheterization, arterial blood pressure and heart rate

A 20 gauge, 7.6 cm catheter was placed in the brachial artery of the non-dominant arm under aseptic conditions after local anesthesia (2% lidocaine) for local administration of study drugs, blood sampling, and mean arterial pressure (MAP) measurement. Heart rate (HR) was determined using a 3-lead electrocardiogram (Cardiocap/5, Datex-Ohmeda Louisville, CO).

Forearm blood flow and vascular conductance

Forearm blood flow (FBF) was measured via venous occlusion plethysmography using mercury-in-silastic strain gauges and techniques as previously described³⁰. FBF was expressed as milliliters per deciliter of tissue per minute (ml/dl FAV/min). As an index of forearm vasodilation and to account for individual differences in baseline vascular tone, forearm vascular conductance (FVC) was calculated as (FBF/MAP) × 100 expressed as ml/dl FAV/min/100mmHg. Immediately following the release of the occlusion cuff for the RH (see below), the same cuffcycled between inflation at ∼50 mmHg (4 seconds) and deflation (3 seconds) to cause venous occlusion and this yielded one blood flow measurement every 7 seconds for the first 56 seconds (8 flow measures). After 8 flow measures, the inflation:deflation cycle was changed back to 7:8 seconds, as was used at rest.

RH protocol

After measurement of baseline FBF, the cuff on the upper arm was rapidly inflated to 200 mmHg for 5 minutes of ischemia. This location and duration of ischemia was chosen to mimic the RH protocol utilized in investigations of the contributions of various endothelial-derived vasodilator pathways to the RH response^{18, 19, 22, 23} and importantly, has recently demonstrated peak RH flow to be more strongly associated with cardiovascular disease risk than measures of flow-mediated vasodilation¹². After 5 minutes, the cuff was rapidly deflated and flow measures commenced for 2.5 minutes (150 seconds). To determine the effect of repeated bouts of RH, 8 additional young subjects were instrumented non-invasively and underwent four successive bouts of RH separated by 20 minutes of rest (see Online Supplemental Materials for details).

Vasoactive drug infusion

All drug infusions were through the brachial artery catheter to create a local effect in the forearm, we recompleted during baseline measures prior to the arterial occlusion, and saline was utilized as a control infusate. Specific timing and duration of infusions is provided below in the Experimental Protocols section.

To inhibit K_IR channels, barium chloride (BaCl₂; K_IR channel inhibitor; 10% w/v BDH3238, EMD Chemicals, Gibbstown, NJ) was infused at 0.9 μmol/dl FAV/min within an absolute range of 8 μmol/min to 10 μmol/min for five minutes prior to each arterial occlusion. To inhibit Na⁺/K⁺-ATPase, ouabain octahydrate (Na⁺/K⁺-ATPase inhibitor; Sigma 03125, St. Louis, MO) was infused at 2.7 nmol/min for 15 minutes prior to arterial occlusion^{30, 34}. On subsequent RH trials, ouabain was reinfused for 5 minutes prior to arterial occlusion to provide continuous inhibition. This approach of using BaCl₂ and ouabain to inhibit K_IR channels and Na⁺/K⁺-ATPase, respectively, has been used previously by our group^{30, 35} and others^{36, 37}. We administered N^G-monomethyl-l-arginine (l-NMMA; NO synthase inhibitor; Clinalfa/Bachem, Weil am Rhein, Germany) to inhibit the production of NO in combination with ketorolac (non-selective cyclooxygenase inhibitor; Hospira, Lake Forest, IL) to inhibit the synthesis of PGs. The doses of l-NMMA and ketorolac were 5 mg/min and 1200 μg/min respectively and given for 5 minutes³⁰ prior to arterial occlusion.

Experimental protocols

In all experimental protocols, subjects rested quietly for 30 minutes after insertion of the catheter before the first experimental trial and for 20 minutes between each RH trial.

Data acquisition and analysis

Data were collected and stored on a computer at 250 Hz and were analyzed off-line with signal-processing software (WinDaq, DATAQ Instruments, Akron, OH). MAP was determined from the arterial pressure waveform. FBF was determined from the derivative of the forearm plethysmography signal. For resting hemodynamic measures, the average of the last minute of baseline was used. To quantify the RH response, we averaged and plotted values from each subject at all FBF time points (7, 14, 21, 28, 35, 42, 49, 56, 60, 75, 90, 105, 120, 135, 150 seconds post-cessation of arterial occlusion) and the total reactive hyperemic FBF [area under the curve (AUC)] was determined as the sum of FBF above baseline at each time point. The peak RH FBF and vasodilation (FVC) was determined for each subject individually and these values were also averaged. In all subjects, these individual peaks occurred at either the first, second, or third flow measurements. When FBF/FVC measurements for all subjects were averaged at each time point, the peak nearly always occurred at the first flow measurement (see Results). To quantify the impact of the vasoactive inhibitors, the magnitude of inhibition (%Δ) was calculated as: (FBF_peak/total inhibition – FBF_peak/total control)/(FBF_peak/total control)×100 and always quantified from the control condition. For the SNP control trials, FBF was averaged across the last minute of baseline and SNP infusion.

Protocol 1: Independent and combined effects of K_IR channel and Na⁺/K⁺-ATPase inhibition

In Group 1 of Protocol 1, subjects received BaCl₂ alone following the control trial in order to assess the independent role of K_IR channels in RH. A representative tracing of one subject who participated in this protocol is provided in Figure 1 in control conditions (Panel A) and following BaCl₂ infusion (Panel B). Baseline FBF and FVC are presented in Tables 1 and ​and2.2. During RH, BaCl₂ significantly reduced the peak response (-50±6%; Figure 2A and B) and impaired FBF for the first 75 seconds (Figure 2A). Taken together, the total RH FBF was also significantly reduced from control levels (-62±3%; Figure 2C). The addition of ouabain did not further impact peak RH FBF (-60±7%; BaCl₂ vs BaCl₂+ouabain; P=0.25) but there was a strong trend towards an additional effect on total RH FBF (-82±4%; P=0.07). The addition of L-NMMA+ketorolac did not have a further impact (Peak: -68±7%; Total: -88±3%). Changes in peak vasodilation (FVC) paralleled those of FBF (Table 2).

Open in a separate window

Figure 1

Representative Tracing of Baseline and Reactive Hyperemia

Representative tracing (n=1) of the last 30 seconds of rest and the first minute of the reactive hyperemia response in control (saline; A) conditions and with inhibition of K_IR channels via BaCl₂ (B). Tracings are shown for the electrocardiogram (ECG), intra-arterial pressure (I.A. Press.), and venous occlusion plethysmography (VOP) output from which heart rate, mean arterial pressure, and forearm blood flow, respectively, are calculated and/or derived. Notes: The vertical scale for VOP is 4 times greater during rest (pre-occlusion) than during reactive hyperemia (post-occlusion). Vertical deflections indicate balancing of the plethysmography signal to maintain a consistent baseline.

Open in a separate window

Figure 2

Protocol 1: Independent effects of K_IRchannel inhibition (Group 1)

A. Forearm blood flow (FBF) response following 5 minutes of arterial occlusion in the following conditions: control (black circles), independent K_IR channel inhibition (BaCl₂; dark grey triangles), combined K_IR channel and Na⁺/K⁺-ATPase inhibition (BaCl₂+ouabain; light grey squares), and combined inhibition of K_IR channels, Na⁺/K⁺-ATPase, NO and PGs (BaCl₂+ouabain+L-NMMA+ketorolac; white diamonds) conditions. BaCl₂ significantly inhibited the response for 75 seconds and there was little additional effect of ouabain, or L-NMMA+ketorolac. *P<0.05 vs BaCl₂; †P<0.05 vs BaCl₂+ouabain; ‡ P<0.05 vs BaCl₂+ouabain+L-NMMA+ketorolac. B. Peak reactive hyperemic FBF was significantly attenuated from control by BaCl_2, and ouabain had no additional effect whereas there was a slightly greater reduction with the addition of L-NMMA+ketorolac. *P<0.05 vs Control; †P<0.05 vs BaCl₂. C. Similarly, total reactive hyperemic FBF (area under curve) was significantly reduced from control by BaCl_2, and ouabain had no additional effect whereas L-NMMA+ketorolac further reduced this response. *P<0.05 vs Control; †P<0.05 vs BaCl₂.

In Group 2 of Protocol 1, subjects received ouabain alone following the control trial in order to assess the independent role of Na⁺/K⁺-ATPase in RH (Figure 3). Ouabain had no effect on peak RH FBF (2±6%; Figure 3A and B) but did significantly reduce FBF during 14-90 seconds of hyperemia, resulting in a significant attenuation of the total RH FBF (-44 ±12%; Figure 3C). The addition of BaCl₂ significantly reduced peak RH FBF (-62±8%) as well as further reduced total RH FBF (-92±8%) whereas there was no additional effect of L-NMMA+ketorolac on either peak (-63±7%) or total RH FBF (-94±8%). Changes in peak vasodilation (FVC) paralleled those for FBF (Table 2).

Open in a separate window

Figure 3

Protocol 1: Independent effects of Na⁺/K⁺-ATPase inhibition (Group 2)

A. Forearm blood flow (FBF) response following 5 minutes of arterial occlusion in control (black circles), independent Na⁺/K⁺-ATPase inhibition (Ouabain; dark grey triangles), combined Na⁺/K⁺-ATPase and K_IR channel inhibition (Ouabain+BaCl₂; light grey squares), and combined inhibition of Na⁺/K⁺-ATPase, K_IR channels, NO and PGs (Ouabain+BaCl₂+L-NMMA+ketorolac; white diamonds) conditions. Ouabain did not affect initial FBF, but thereafter reduced FBF from control until 90 seconds post-cuff deflation. The addition of BaCl₂ further attenuated FBF for 30 seconds, whereas addition of L-NMMA+ketorolac had no further effect. *P<0.05 vs Ouabain; †P<0.05 vs Ouabain+BaCl₂; ‡P<0.05 vs Ouabain+BaCl₂+L-NMMA+ketorolac. B. Peak reactive hyperemic FBF was not affected by ouabain. Infusion of BaCl₂ significantly reduced peak FBF from control, and L-NMMA+ketorolac had no further impact. *P<0.05 vs Control; †P<0.05 vs Ouabain. C. Total reactive hyperemic FBF (area under curve) was significantly reduced from control by ouabain, and BaCl₂ had an additional effect whereas L-NMMA+ketorolac did not. *P<0.05 vs Control; †P<0.05 vs Ouabain.

Protocol 2: Effects of combined inhibition of NO and PGs

In Protocol 2, we assessed the combined contribution of NO and PGs to RH and subsequently inhibited K_IR channels and Na⁺/K⁺-ATPase (Figure 4). As would be expected with effective inhibition, L-NMMA+ketorolac significantly reduced baseline FBF and FVC (Tables 1-​-3).3). The mean of the first FBF measures was augmented with L-NMMA+ketorolac (Figure 4A); however, when each individual subjects' peak response was averaged, this comparison only approached being significant (+18±8%; P=0.07; Figure 4B). FBF was attenuated with L-NMMA+ketorolac 30-60 seconds following the end of arterial occlusion (Figure 4A), yet the total RH FBF remained similar to control (-10±12%; P=0.24; Figure 4C). The additional inhibition of K_IR channels and Na⁺/K⁺-ATPase via BaCl₂ and ouabain, respectively, significantly attenuated both peak (-61±8%; Figures 4A and B) and total (-69±6%; Figure 4C) RH FBF.

Open in a separate window

Figure 4

Protocol 2: Effects of combined inhibition of nitric oxide and prostaglandins

A. Forearm blood flow (FBF) response following 5 minutes of arterial occlusion in control (black circles), combined inhibition of NO and PG synthesis (L-NMMA+ketorolac; light grey squares), and combined inhibition of NO, PGs, K_IR channels and Na⁺/K⁺-ATPase (L-NMMA+ketorolac +BaCl₂+ouabain; white diamonds) conditions. L-NMMA+ketorolac attenuated the response from control only from 30-60 seconds post-cuff deflation. The addition of BaCl₂+ouabain significantly reduced FBF for 30 seconds and thereafter had no further effect. *P<0.05 vs L-NMMA+ketorolac; †P<0.05 vs L-NMMA+ketorolac +BaCl₂+ouabain. B. Peak reactive hyperemic FBF was not affected by L-NMMA+ketorolac and was significantly attenuated by L-NMMA+ketorolac +BaCl₂+ouabain. *P<0.05 vs Control; †P<0.05 vs L-NMMA+ketorolac. C. Similar to peak, total reactive hyperemic FBF (area under curve) was not affected by L-NMMA+ketorolac and was significantly attenuated by L-NMMA+ketorolac +BaCl₂+ouabain. *P<0.05 vs Control; †P<0.05 vs L-NMMA+ketorolac.

Comparison of RH protocols

A summary of the relative (%Δ) effects of independent and combined roles of K_IR channels and Na⁺/K⁺-ATPase, as well as combined NO and PGs as compared to control conditions is presented in Figure 5. In these pooled comparisons (BaCl₂: n=8; ouabain: n=8; BaCl₂+ouabain: n=16; L-NMMA+ketorolac: n=8; BaCl₂+ouabain+L-NMMA+ketorolac: n=24) the reduction of peak FBF from BaCl₂ alone (-50±6%) to combined BaCl₂+ouabain (-61±6%) was similar (P=0.15). Total RH FBF was attenuated by BaCl₂ alone (-62±3%) and ouabain alone (-44±12%) and this effect was enhanced when these inhibitors were combined (-87±4%), nearly eliminating the hyperemic response to ischemia. L-NMMA+ketorolachad no independent effects on peak (P=0.68) or total RH FBF (P=0.69) nor did they enhance any inhibition beyond that which occurred with BaCl₂+ouabain infusion (Peak: -64±4%; P=0.60; AUC:-84±4%; P=0.62).

Open in a separate window

Figure 5

Summary: Effects of inhibition of K_IRchannels, Na⁺/K⁺-ATPase, nitric oxide and prostaglandins on peak and total reactive hyperemia

Combined results from the three experimental protocols are presented for relative impact (%Δ) on both peak (A) and total (B) reactive hyperemic forearm blood flow (FBF) in each experimental condition (BaCl₂: n=8; Ouabain: n=8; BaCl₂+ouabain: n=16; L-NMMA+ketorolac: n=8; BaCl₂+ouabain+L-NMMA+ketorolac: n=24). BaCl₂ reduced peak FBF and this attenuation was unchanged with the addition of ouabain or L-NMMA+ketorolac. Neither ouabain alone nor L-NMMA+ketorolac attenuated peak FBF. BaCl₂ and ouabain both independently reduced total FBF and in combination (BaCl₂+ouabain), the reduction was enhanced. There was no additional reduction by L-NMMA+ketorolac, nor did L-NMMA+ketorolac independently reduce total FBF. *P<0.05 vs zero; †P<0.05 vs BaCl₂; ‡P<0.05 vs Ouabain

Historical overview of RH in humans

Beginning with the initial observation nearly 70 years ago of a rapid and profound hyperemia in response to a period of ischemia, there was interest in determining the underlying signals for this response^{1, 2, 7, 20, 23}. Early experiments determined that an intact nervous system was not requisite to observe this response^{2, 10} and subsequent studies pursued investigating local mechanisms of vascular control that might be involved^{8, 17-20, 22, 24}. Alongside these studies aimed to determine the physiological basis of RH, the test itself began to be used as a measure of vascular health in a variety of at-risk populations^{4-7, 9, 10}. Different groups of subjects that demonstrated “endothelial dysfunction” as commonly assessed by intra-arterial infusion of endothelium-dependent vasodilators (e.g. acetylcholine) or flow-mediated vasodilation of the brachial artery were shown to have attenuated RH responses^{4-7, 9, 10}. These associations among cardiovascular disease, endothelial health, and impaired RH have further stimulated an interest in the potential underlying mechanisms of this response. Importantly, recent evidence indicates that peak RH flow in response to five minutes of ischemia (via upper arm cuff inflation) may in fact be a better predictor of cardiovascular events than the more commonly-assessed brachial flow-mediated vasodilation¹². Thus, elucidating the mechanisms underlying this physiological response has clear significant clinical implications.

Several potential mediators and downstream targets underlying RH have been postulated and experimentally tested in both humans and experimental animals. Due to the ischemic nature of the stimulus, classic metabolic candidates for regulating RH include adenosine or K_ATP channel activation. Augmenting adenosine signaling through caffeine (adenosine receptor antagonist) withdrawal or dipyridamole (inhibitor of cellular uptake of adenosine) does improve RH; however, direct inhibition of adenosine receptors (via theophylline or caffeine) does not impair peak RH FBF^{22, 26} and has a minimal effect on total FBF²². Similarly, results from inhibition of K_ATP channels have been equally as unsuccessful in explaining RH. Inhibition of K_ATP channels via sulfonylureas such as tolbutamide or glibenclamide modestly reduces total RH FBF but has no impact on the peak response in some studies^{17, 24}, whereas other investigators have demonstrated no effect of K_ATP channel inhibition on either peak or total RH ³⁹.

Endothelium-derived NO and PGs contribute little to RH in humans

It is well known that endothelial-derived NO contributes to cardiovascular health in humans due to its multifaceted cardioprotective properties²⁷. Whether NO mediates RH was a logical proposition and has been investigated in a variety of existing studies^{8, 17-20}. There is discrepancy within the literature, and our present finding that NO (in combination with PGs) does not contribute to peak RH FBF fits with the results of most^{5, 20, 36, 40} but not all^{8, 41} of these studies. Some of the previous work has shown a modest role for NO in the total hyperemic response^{17, 18, 20}. Additionally, previous studies demonstrated only a minimal contribution of endothelial-derived PGs to peak and/or total RH^{18, 21, 22}. However, it is important to note that significant cross-talk occurs between these two endothelial pathways, such that inhibition of one pathway often does not impact vascular responses to a variety of stimuli, whereas combined inhibition reveals a significant role^{32, 33}. Only one study to date inhibited NO and PGs in combination, and this was done with intra-arterial L-NMMA and oral ibuprofen¹⁸. In this prior study, there was no impact on the peak change in FBF in response to ischemia, whereas there was some reduction (∼35%) in the prolonged hyperemic response¹⁸. Our current findings agree with these observations that even in combination, NO and PGs do not contribute to peak RH FBF, and while we observed a reduction in absolute FBF in the latter portion of RH, this was not of sufficient magnitude to impair the total RH FBF (Figure 4). An interesting observation in the present study was that peak RH was somewhat augmented after combined NO and PG inhibition, and this could also reflect a critical role for vascular hyperpolarization in the response as NO is capable of suppressing EDH-mediated vasodilation⁴². Although presently unclear, another explanation for this augmented response is our inhibition of cyclooxygenase may have shifted arachadonic acid to the cytochrome p450 pathway, thus increasing the production of EETs to cause additional vasodilation²⁷.

Critical role for K_IR channel and Na⁺/K⁺-ATPase activation in RH in humans

Based on our assessment in the present study, there is a prominent role for K_IR channels and Na⁺/K⁺-ATPase in RH in humans (Figure 5). Interestingly, only K_IR channel activation, but not Na⁺/K⁺-ATPase contributed to the peak RH FBF. Selective inhibition of K_IR channels reduced the peak hyperemic response ∼50%, and a total of ∼60% was observed when inhibition of Na⁺/K⁺-ATPase was performed simultaneously. Inhibition of K_IR channels and Na⁺/K⁺-ATPase independently reduced the total RH FBF by ∼60% and ∼40% respectively, and this effect is enhanced when these are inhibited in combination. In this context, there is a remarkable reduction in the total response (-87±4%) from control, nearly abolishing RH. Collectively, the magnitude of the observed attenuation due to K_IR channel inhibition on peak hyperemia, and combined K_IR and Na⁺/K⁺-ATPase inhibition on the total hyperemic response, is by far the greatest in the known studies to date on this topic.

Endothelium-dependent vasodilation that occurs beyond NO and PGs causes hyperpolarization of endothelial cells and vascular smooth muscle cells²⁷. “Classic' EDH is sensitive to inhibition of K_Ca channels that when activated cause hyperpolarization of endothelial and smooth muscle cells through direct electrical communication or stimulation of K_IR channels and Na⁺/K⁺-ATPase²⁹. Our findings of a significant role for K_IR channel and Na⁺/K⁺-ATPase activation in RH in humans is consistent with the classic proposed mechanism of EDH. However, we must also recognize that the RH response may be endothelium-independent, yet still occur through vascular smooth muscle cell hyperpolarization²⁸. Recently, K_IR channels have been shown to be particularly important for the amplification of hyperpolarizing stimuli as they are directly responsive to changes in membrane potential⁴³ and thus, this unique property may explain the profound impact of BaCl₂ administration we observe on both peak and total reactive hyperemia. We are not able to address cell-specific issues related to the K_IR and Na⁺/K⁺-ATPase activation that we observe in the present study due to the limitations of our human in vivo model.

Experimental considerations

All of the inhibitors utilized were administered prior to arterial occlusion and RH. This may lead one to question the efficacy of our inhibitors after the 5 minutes of occlusion and subsequent large increases in blood flow. While not directly assessed, we used doses previously established in our laboratory³⁰ and given the large magnitude of the effects on peak and total RH FBF we observed, do not think this consideration affects our primary conclusions. If anything, we maybe potentially underestimating a role for K_IR channels and Na⁺/K⁺-ATPase in RH. Further, we show that RH responses are largely repeatable over time, with only a slight decline (-9±5%) in peak RH in the 4^th trial as compared to the 1^st trial (see Supplemental Figure I) and no change in AUC. Thus, the marked impact of BaCl₂ and ouabain we observe cannot be attributed to reduced responses with repeated trials.

BaCl₂ has been demonstrated to be primarily selective for K_IR channels up to a concentration of 100 μmol/L⁴⁴. Dawes and colleagues demonstrated that a dose at half of what we used increased antecubital venous plasma concentrations in the infused forearm to 50 μmol/L³⁶ and thus, it can be assumed that our dose would result in concentrations within the selective range for K_IR channels. A direct assessment of the selectivity of BaCl₂ for K_IR channels is difficult in humans as it is not possible to make membrane potential measurements or isolate selective stimulation of this channel. Along these lines, at greater concentrations, BaCl₂ has been shown to inhibit other potassium channels, most prominently K_ATP channels. While we believe that BaCl₂ in the dose we administered is selective for K_IR channels, if we are in fact inhibiting K_ATP channels, this likely does not provide an alternate explanation for our findings as the majority of existing data shows little-to-no impact of inhibiting K_ATP channels on peak and/or total RH ^{17, 24, 39}. Further, our group^{30, 31, 35} and others^{36, 37} have demonstrated that each of the pharmacological inhibitors we use in this investigation do not impact overall vascular responsiveness nor have systemic effects in the doses utilized.

Presently, the exact stimulus for vascular hyperpolarization in response to local ischemia is unknown. Potential candidates include substances that have been shown to cause vasodilation through K_IR channels and/or Na⁺/K⁺-ATPase such as K^{+30, 45}, ATP³⁰, bradykinin³⁷, H₂O₂⁴⁶, and EETs⁴⁷ and it is possible that concentrations of these substances may rise during ischemia as has been observed in animal models, particularly in the coronary circulation⁴⁸. However, to the best of our knowledge, limited studies in humans have made interstitial measures of the candidate substances during ischemia in skeletal muscle, and to date, no significant increases have been observed⁴⁹. Alternatively, evidence suggests that mechanosensitive mechanisms such as the myogenic response and stretch of endothelial cells contribute to the earliest portion of RH⁴⁰. In this context, recent data indicate that low intravascular pressure can stimulate transient receptor potential channels that elicit changes in endothelial cell calcium which can stimulate vasodilation via hyperpolarization⁵⁰ and this might also serve as a stimulus for K_IR channel and/or Na⁺/K⁺-ATPase activation. Identifying the stimulus for K_IR channel and Na⁺/K⁺-ATPase activation that occurs during RH represents an intriguing future area of research and potentially would provide valuable insight into explaining impaired RH responses in clinical populations.

We thank the subjects who volunteered to participate and Leora J. Garcia and Devin V. Dinenno for their assistance.

Sources of Funding: NIH awards HL102720, HL095573 (F.A.D.).