Diabetes downregulates BKβ1 potassium channel subunit in retinal arteriolar smooth muscle

Abstract

Retinal vasoconstriction and reduced retinal blood flow precede the onset of diabetic retinopathy. The pathophysiological mechanisms that underlie increased retinal arteriolar tone during diabetes remain unclear. Normally, local Ca²⁺ release events (Ca²⁺-sparks), trigger the activation of large-conductance Ca²⁺-activated K⁺(BK)-channels which hyperpolarize and relax vascular smooth muscle cells (VSMCs), thereby causing vasodilatation. In the present study, we examined BK channel function in retinal VSMCs from streptozotocin(STZ)-induced diabetic rats. The BK channel inhibitor, Penitrem A, constricted non-diabetic retinal arterioles (pressurized to 70mmHg) by 28%. The BK current evoked by caffeine was dramatically reduced in retinal arterioles from diabetic animals even though caffeine-evoked [Ca²⁺]_i release was unaffected. Spontaneous BK currents were smaller in diabetic cells, but the amplitude of Ca²⁺-sparks was larger. The amplitudes of BK currents elicited by depolarizing voltage steps were similar in control and diabetic arterioles and mRNA expression of the pore-forming BKα subunit was unchanged. The Ca²⁺-sensitivity of single BK channels from diabetic retinal VSMCs was markedly reduced. The BKβ1 subunit confers Ca²⁺-sensitivity to BK channel complexes and both transcript and protein levels for BKβ1 were appreciably lower in diabetic retinal arterioles. The mean open times and the sensitivity of BK channels to tamoxifen were decreased in diabetic cells, consistent with a downregulation of BKβ1 subunits. The potency of blockade by Pen A was lower for BK channels from diabetic animals. Thus, changes in the molecular composition of BK channels could account for retinal hypoperfusion in early diabetes, an idea having wider implications for the pathogenesis of diabetic hypertension.

Introduction

Diabetes causes changes to the structure and function of blood vessels in the retina leading to visual impairment and blindness¹. The cellular and molecular basis of diabetic retinopathy is not wholly understood although large prospective clinical trials have established the importance of hyperglycaemia in precipitating this disease in both type 1 and type 2 diabetic patients^2,3. A major pathway through which hyperglycaemia is believed to contribute to retinal microangiopathy is the disruption of retinal blood flow. Patient-based studies have shown that retinal haemodynamic abnormalities occur prior to the onset of clinical diabetic retinopathy⁴ and that the development and progression of diabetic retinopathy correlates with the extent of the blood flow changes observed^5-8. In diabetic patients without retinopathy, retinal arteriolar vasoconstriction^{9, 10} and decreased total retinal blood flow has been reported^{4, 11}. However, as the disease progresses, the arterioles begin to dilate¹² and bulk retinal blood flow increases in proportion to the severity of retinopathy, and thus the degree of retinal ischemia¹¹. Presently, the molecular mechanisms underlying retinal vasoconstriction during early diabetes are unknown, yet an improved understanding of the pathophysiology involved could be crucial to the development of better therapies for the treatment of diabetic retinopathy.

A major factor controlling the contractile state of arterioles is the activity of ion channels on the plasma membranes of the VSMCs¹³. BK channels are known to play a crucial role in the regulation of arterial smooth muscle tone since blockade of these channels using the specific inhibitor iberiotoxin, causes membrane depolarisation and vasoconstriction in pressurized isolated vessels¹⁴. BK channels are activated by local Ca²⁺ release events, termed Ca²⁺ sparks, resulting from opening of ryanodine receptor (RyR) channels in the sarcoplasmic reticulum (SR;¹⁵). The activation of BK channels results in outward K⁺ current that opposes VSMC contraction by causing membrane hyperpolarisation. This reduces Ca²⁺ influx by reducing activation of voltage-dependent Ca²⁺ channels¹⁵. BK channels are composed of α-subunits and accessory β-subunits¹⁶. The α-subunit forms the K⁺ selective pore, while the β subunits influence the kinetics, pharmacology and Ca²⁺ sensitivity of BK currents^{17, 18}. Four members of the BK β-subunit family have been identified to date (β1-β4) and the β1-subunit is expressed predominantly in VSMCs¹⁹. Targeted deletion of the β1-subunit gene reduces the Ca²⁺ sensitivity of BK channels and the coupling of Ca²⁺ sparks to BK channel activity in VSMCs from cerebral arteries^{19, 20}. The functional significance of the β1-subunit of VSMC BK channels is underlined by the observation that knockout mice are hypertensive and display enhanced vascular reactivity to application of vasoconstrictors^{19, 20}.

In the present study we have examined the effects of STZ-induced diabetes on the properties of BK channels in VSMCs of the rat retinal microcirculation. A clear advantage of using STZ-diabetic rats is that it is well documented that retinal arteriolar vasoconstriction and decreased retinal blood flow occur in this animal model following several weeks of diabetes²¹. These animals subsequently exhibit many of the vasodegenerative changes associated with human diabetic retinopathy²². We hypothesised that diabetes causes a downregulation of the α- and/or β1-subunit, thereby reducing the capacity of the BK channels to hyperpolarise retinal VSMCs and resist vasoconstriction. We show that diabetes reduces coupling between Ca²⁺ release from RyR sensitive Ca²⁺ stores and BK channel activation. No alteration in the expression of the pore-forming α-subunit was evident, but β1-subunit expression was reduced at both the mRNA and protein level. Consistent with this, we found that BK channels in retinal VSMCs from STZ-diabetic animals exhibit a diminished sensitivity to Ca²⁺ and we provide pharmacological evidence supporting the idea that the expression of functional α+β1 subunit complexes is reduced in diabetes. These results suggest that changes in the molecular composition of BK channels in retinal VSMCs during diabetes might contribute to the onset and early progression of diabetic retinopathy.

Methods

All procedures with animals were performed in accordance with the Guide for the Care and Use of Laboratory Animals published by the US National Institutes of Health (NIH Publication No. 85-23, revised 1996) and the United Kingdom Animals (Scientific Procedures) Act, 1986. Full details of the methods and materials used are in the online data supplement available at http://circres.ahajournals.org.

Results

Characteristics of experimental animals

Diabetic animals had higher mean plasma glucose levels (27.8 ± 1.3mmol/L; n=41) than non-diabetic animals (7.6 ± 0.3mmol/L; n=37 p<0.001). Mean glycosylated haemoglobin values were 5.9 ± 0.2% and 17.6 ± 0.8% in the non-diabetic and diabetic groups, respectively (p<0.001). Both sets of animals gained weight during the 3-month experimental period, but the increase was greater in non-diabetic than diabetic rats (304 ± 17g vs 141 ± 9.7g, respectively; p<0.001).

Functional Significance of BK channels in retinal VSMCs

We examined the contribution of BK channels to the regulation of retinal arteriolar tone by pressurizing freshly isolated retinal arterioles from non-diabetic animals to 70mmHg and measuring the change in internal diameter induced by 100nmol/L Penitrem A (Pen A), a potent inhibitor of BK channels²³. Pen A caused a 28% decrease in the diameter of pressurized retinal vessels (Fig 1). These results suggest that BK activity plays an important vasodilatory role in retinal blood vessels.

Figure 1. Physiological significance of BK channels in retinal arterioles.

Top panel, representative photomicrographs of a pressurized retinal arteriole (70mmHg) before and 10-min after the application of 100nmol/L Pen A. Bottom panel, mean ± SEM of the Pen A-induced constriction in non-diabetic retinal arterioles (n=8 vessels, p<0.05; paired t-test).

Caffeine-induced BK currents, but not Ca²⁺ activated Cl^- currents, are smaller in retinal VSMCs from diabetic animals

To investigate BK channel activity during diabetes we tested activation following Ca²⁺ release from caffeine-sensitive Ca²⁺ stores. Retinal arterioles were bathed in low Cl^- Hanks’ solution containing 1mmol/L 9-anthracene carboxylic acid (9AC) to block Ca²⁺ activated Cl^- channels (Cl_Ca channels). Application of 10mmol/L caffeine for 5 seconds evoked large, noisy transient outward currents at positive membrane potentials (Fig 2A,C). These currents were fully abolished by addition of 100nmol/L Pen A (Fig 2A; n=6). Fig 2B shows typical caffeine-induced BK currents in retinal VSMCs from non-diabetic and diabetic animals. Fig 2C shows the average peak current density plotted as a function of voltage for the caffeine-induced BK currents. BK currents evoked by Ca²⁺ release from caffeine-sensitive Ca²⁺ stores were dramatically reduced in VSMCs from diabetic animals.

Figure 2. Caffeine-induced BK currents recorded from retinal VSMCs of non-diabetic and diabetic animals.

(A) Left, whole-cell records showing caffeine-induced outward currents in a non-diabetic vessel held at a range of potentials. The vessel was bathed in low Cl^- Hanks’ solution with 1mmol/L 9AC to block Cl_Ca currents. Right, the BK channel inhibitor, Pen A (100nmol/L), completely abolished the caffeine-induced currents. (B) Typical traces showing caffeine-induced BK currents at -80, 0, +40 and +80 mV in non-diabetic and diabetic retinal arterioles. (C) Summary current-voltage relationships for the caffeine-induced BK currents in non-diabetic (for each point, n=5-13) and diabetic (n=7-19) vessels.

The attenuated BK currents in arterioles from diabetic rats could be explained by decreased Ca²⁺ release from the caffeine-sensitive stores. To test this, global [Ca²⁺]_i responses were measured in retinal VSMCs from non-diabetic and diabetic animals using fura 2 based Ca²⁺ microfluorimetry (Fig 3A). No differences in the peak amplitude of [Ca²⁺]_i transients evoked by 10mmol/L caffeine were observed. We also compared the size of caffeine-evoked Cl_Ca currents at a range of voltages in myocytes from non-diabetic and diabetic arterioles bathed in normal Hanks’ solution containing 100nmol/L Pen A to block BK channels. Application of 10mmol/L caffeine evoked currents that reversed close to E_Cl (0mV; Fig 3B,D). These were completely abolished in low Cl^- Hanks’ solution containing the Cl^- channel inhibitor, 9AC (Fig 3B; n=6). Isolation of the caffeine-evoked Cl^- currents in diabetic arterioles required longer pre-incubation times with Pen A (∼20 min as compared to ∼5-min for non-diabetic vessels). Fig 3C,D show representative traces and summary data for the caffeine-evoked Cl^- currents in non-diabetic and diabetic retinal VSMCs. The mean peak current densities were similar for normal and diabetic cells at all voltages tested. Thus, it appears that the smaller caffeine-induced BK currents in diabetic retinal VSMCs reflect reduced coupling efficiency between Ca²⁺ release and BK channel activation rather than defective Ca²⁺ release.

Figure 3. Caffeine-evoked global [Ca²⁺]_i transients and Cl_Ca currents in non-diabetic and diabetic retinal VSMCs.

(A) Left, time-course record showing the effects of 10mmol/L caffeine on global [Ca²⁺]_i in a non-diabetic retinal arteriole segment. Right, mean data showing that caffeine-induced [Ca²⁺]_i transients were unaffected by diabetes. (B) Left, whole-cell currents elicited by 10 mmol/L caffeine in a non-diabetic vessel held at a range of test potentials and bathed in normal Hanks’ solution containing the BK channel antagonist, Pen A (100nmol/L). Right, the Cl_Ca channel inhibitor, 9AC (1mmol/L), completely blocked the caffeine-evoked transient inward and outward currents. (C) Original records showing caffeine-induced Cl_Ca currents at -80, 0 and +80mV in non-diabetic and diabetic vessels. (D) Plot showing the mean peak current density against voltage for the caffeine-evoked Cl_Ca currents in non-diabetic (n=12) and diabetic (n=10) retinal VSMCs.

Ca²⁺ sparks are larger, but STOCs are smaller in diabetic retinal VSMCs

From a physiological perspective it is well established that Ca²⁺ sparks activate BK channels generating spontaneous transient outward K⁺ currents (STOCs), modulating vascular tone^{15, 24}. STOCs were recorded from retinal VSMCs of non-diabetic and diabetic animals at test potentials between -80mV to +80mV (increased in 40mV increments) using the perforated patch clamp technique. STOCs were evident at membrane potentials positive to -40mV and were completely abolished in the presence of 100nmol/L Pen A (n=6, non-diabetic VSMCs). Individual STOC events were superimposed (Fig 4A) precluding an accurate assessment of STOC amplitudes and frequencies, so we calculated the integral of the STOC densities for recordings lasting 5-10 minutes. Visual inspection of the original traces suggested that STOC amplitudes and frequencies were greatly reduced in diabetic retinal VSMCs, and integrated current densities were considerably smaller (Fig 4A). These data confirm that retinal VSMCs from diabetic animals demonstrate less spontaneous BK current activity than cells from non-diabetic animals.

Figure 4. STOC activity is reduced but Ca²⁺ sparks are greater in retinal VSMCs from diabetic animals.

(A) Top panel, whole-cell recordings of STOC activity in a non-diabetic and diabetic vessel at a holding potential of +40mV. Bottom panel, graph showing the mean integrated current density versus voltage for STOCs from non-diabetic (n=8) and diabetic (n=8) retinal arterioles. (B) Top panel, line-scan image recorded from a non-diabetic retinal VSMC showing two consecutive Ca²⁺ sparks originating from the same Ca²⁺ spark site. The graph below plots the fractional fluorescence change (F/F₀) for this panel. Bottom panel, line scan image and graph on slower time scales from another non-diabetic cell in which Ca²⁺ sparks amalgamate to produce a cell-wide global Ca²⁺ oscillation. (C) Representative line-scan images of basal Ca²⁺ sparks in non-diabetic and diabetic retinal VSMCs. Bottom, average temporal profile for each spark has been plotted and the traces superimposed.

Using confocal imaging techniques, we have recently described the presence of two distinct populations of spontaneous Ca²⁺ sparks in retinal VSMCs: “basal” sparks that arise from resting fluorescence levels (i.e. from F/F_o = 0.95-1.05) and “sparks on oscillations” that overlay global Ca²⁺ transients²⁵. Fig 4B shows representative images of basal Ca²⁺ sparks, Ca²⁺ sparks on oscillations and global Ca²⁺ oscillations in non-diabetic arteriolar myocytes. We considered the possibility that the lower STOC activity in diabetic retinal VSMCs may reflect reduced spontaneous sub-cellular [Ca²⁺]_i signaling in these cells. Fig 4C shows typical images of basal Ca²⁺ sparks in non-diabetic and diabetic arteriolar myocytes. Below the image panels, the time course of the normalized fluorescence for each event has been plotted and the traces superimposed. It is evident that the peak amplitude of the Ca²⁺ spark in the diabetic cell is around twice that of the non-diabetic. Quantitative data for basal Ca²⁺ sparks, Ca²⁺ sparks on oscillations and global oscillations in non-diabetic and diabetic VSMCs are summarized in Table 1 of the online data supplement. The peak amplitude of both populations of Ca²⁺ sparks was substantially larger in diabetic than in non-diabetic VSMCs (ΔF/F_o basal sparks, 0.92 ± 0.06 and 0.42 ± 0.03, respectively; ΔF/F_o sparks on oscillations, 1.56 ± 0.21 and 0.36 ± 0.04, respectively; p<0.001 in both cases), while the frequency and duration (FDHM) of these events remained unchanged. No differences were observed in amplitude, frequency or duration of global Ca²⁺ oscillations between non-diabetic and diabetic VSMCs. These results show that decreased sub-cellular Ca²⁺ signaling activity cannot explain the decreased STOC activity observed in diabetic VSMCs.

Diabetes reduces the Ca²⁺ sensitivity of BK channels in diabetic retinal VSMCs

Both the comparisons of caffeine-evoked currents and Ca²⁺ transients, and of STOC and spark activity, suggest decreased coupling between Ca²⁺ release and BK channel activation in retinal VSMCs after short-term diabetes. This might be explained by a reduced number of functional BK channels in diabetic myocytes. Quantitative RT-PCR, however, failed to reveal any differences in BKα transcript expression in non-diabetic and diabetic retinal arterioles (Fig 5A). To assess BK channel density, we also compared depolarization dependent whole-cell BK currents (Fig 5Bi,ii). Vessels were bathed in low Cl^- Hanks’ solution containing 1mmol/L 9AC and 10mmol/L 4-aminopyridine (4AP) to block Cl_Ca and K_v channels²⁶, respectively. 10mmol/L 4AP has no effect on BK channels in retinal arterioles; caffeine-evoked BK currents at +40mV in the absence and presence of 10mmol/L 4AP were 115 ± 29.8 and 114 ± 26pA/pF, respectively (n=4; p=0.84; paired t-test). Voltage-activated BK currents in control VSMCs were unaffected by the removal of extracellular Ca²⁺, inhibition of voltage-dependent Ca²⁺ channels with 10μmol/L nifedipine or blockade of RyR receptors with 100μmol/L tetracaine (Fig 5Biii-v). This demonstrates that BK currents evoked by depolarisation are independent of Ca²⁺ influx and Ca²⁺ store release in these cells i.e. they appear to be entirely dependent on voltage gating of the BK channels. As such, the peak current density should be proportional to the number of functional BK channels, assuming that the single-channel conductance and the voltage dependence of activation remain constant. Single channel conductance’s of BK channels were similar for non-diabetic and diabetic VSMCs (160 ± 2pS and 160 ± 4pS, respectively; n=21 and 15 patches, p>0.05). Fig 5C shows the average peak current density as a function of voltage for the voltage-activated BK currents in retinal VSMCs from non-diabetic and diabetic animals. No differences were observed, suggesting that BK channel density is unaltered in retinal VSMCs during diabetes. Consistent with this, in single channel recordings we found no difference in the number of BK channels per membrane patch in non-diabetic (2.4 ± 0.2 channels) and diabetic (2.1 ± 0.2 channels) retinal VSMCs (p>0.05, n=40 patches).

Figure 5. Diabetes does not affect BK channel density.

(A) Histogram showing relative BKα transcript expression in non-diabetic and diabetic retinal arterioles as determined by quantitative PCR. Amplifications were performed in triplicate. 9 non-diabetic and 9 diabetic rats were used in total and RNA was isolated from 15-25 retinal arterioles collected from 3 non-diabetic and 3 diabetic animals per replicate. BKα transcript expression was normalized to β-actin. (Bi) Family of whole-cell voltage clamp currents evoked in a non-diabetic retinal VSMC by voltage steps ranging between -100 mV to +100mV from an initial holding potential of -80mV in the presence of 9AC (1 mmol/L) and 4AP (10mmol/L). (Bii) The voltage-dependent current was abolished by the BK channel inhibitor, Pen A (100nmol/L; % change, -99.3 ± 1.1%; p<0.05; n=4), but was unaffected by (Biii) the removal of extracellular Ca²⁺ (2.5± 8.5%; p>0.05; n=4), (Biv) the L-type Ca²⁺ channel inhibitor, nifedipine (10μmol/L; -1.6 ± 17.2%; p>0.05 n=5) and (Bv) the RyR antagonist tetracaine (100μmol/L; 3.4 ± 15.2%; p>0.05; n=5). Responses were constant across the full voltage range. For clarity current records are presented for single steps between -80 to +80mV. Dashed lines: zero current (C) Average peak current density as a function of voltage for the voltage-activated BK current in non-diabetic (n=13) and diabetic (n=13) vessels.

Another possible explanation for the reduced coupling between Ca²⁺ release and BK channel activation is a decrease in the sensitivity of the BK channels to activation by Ca²⁺. We examined the Ca²⁺-sensitivity of BK channels using inside-out membrane patches from retinal VSMCs (Fig 6). The open probability (P_o) at +80mV was determined for a range of Ca²⁺ concentrations between 0.01-100μmol/L (Fig 6A,B). The activity of BK channels from both non-diabetic and diabetic animals increased with increasing Ca²⁺ concentrations, but the P_o vs Ca²⁺ curve was shifted to the right and the Hill slope was reduced for diabetics (Fig 6B). No differences in P_o’s were observed at Ca²⁺ concentrations of 0.01 and 0.1μmol/L, and this may explain the similarity in voltage-activated, whole cell BK-currents described above. P_o vs voltage relations were also determined at a single Ca²⁺ concentration, 10μmol/L (Fig 6C). There was a strong rightward shift along the voltage axis (>100mV) for BK channels from diabetic animals.

Figure 6. Ca²⁺ sensitivity of single BK channels is reduced in diabetes.

(A) Representative single BK channel records in inside-out patches (holding potential +80mV) from non-diabetic and diabetic VSMCs exposed to increasing [Ca²⁺]. (B) Summary data of the mean ± SEM P_o at the five Ca²⁺ concentrations tested. Non-diabetic, n=8-11; diabetics, n=8-11. Curves are fitted with the Hill equation as described in the supplementary methods. Fit parameters are as follows: non-diabetic (Kd = 0.86 μmol/L, Hill slope 1.1), diabetic (Kd = 1.9 μmol/L, Hill slope 0.97). (C) P_o-V relations determined at 10μmol/L Ca²⁺. Non-diabetic, n=2-11; diabetics, n=1-11. Curves are Boltzmann fits with the following parameters: non-diabetic (V_1/2 = -59.7 mV); diabetic (V_1/2 = 57.9 mV).

β1 expression and function is lower in diabetic retinal VSMCs

The results above suggest that BK channels in diabetic VSMCs have a reduced Ca²⁺ sensitivity. Since the Ca²⁺ sensitivity of BK channels is known to be dependent on the presence of β1 accessory subunits^{17, 18}, a downregulation of the β1 subunit could explain the changes observed in diabetes. Expression of the BKβ1 subunit in retinal arterioles from non-diabetic and diabetic animals was investigated at the mRNA level. β1 transcripts were approximately 60% less abundant in diabetic than in non-diabetic arterioles (Fig 7A). We also estimated changes in expression of the β1 subunit by immunohistochemistry. A punctuate distribution of BKβ1-associated fluorescence was apparent throughout the VSMC layer of retinal arterioles from non-diabetic animals (Fig 7B). Consistent with the RT-PCR analysis, BKβ1 immunostaining decreased dramatically in retinal arterioles of diabetic rats (Fig 7B) i.e. BKβ1 expression is reduced in retinal VSMC in diabetes.

Figure 7. BKβ1 subunit expression and function.

(A) Downregulation of BKβ1 mRNA in retinal VSMC cells from diabetic arterioles. BKβ1 expression in diabetic arterioles is presented relative to non-diabetic vessels. Amplifications were performed in triplicate (same samples as for Fig 5A) and normalized as described for BKα transcripts (B) Left, confocal images of non-diabetic and diabetic retinal arterioles embedded within retinal flatmount preparations and labeled with anti-BKβ1 Ab (green) and propidium iodide (red: nuclear label). Labeling of the circular smooth muscle is reduced in the tissue from the diabetic animal. Right, summary data showing statistically significant reduction in anti-BKβ1 fluorescence for diabetic samples (n=6 retinas, 30 vessels) relative to non-diabetics (n=6 retinas, 25 vessels). (C) Sensitivity of single BK channels in inside out patches to 1μmol/L tamoxifen (holding potential +80mV; 1μmol/L free [Ca²⁺]) from non-diabetic and diabetic retinal VSMCs. Right, summary data showing the differential effects of tamoxifen on the P_o of single BK channels from non-diabetic (n=7) and diabetic (n=8) vessels. (D) Pharmacology of single BK channels from non-diabetic (n=5) and diabetic (n=9) retinal VSMCs exposed to Pen A. Mean data is expressed as the % inhibition of P_o after 5-min of exposure to 100nmol/L Pen A.

Besides increasing the Ca²⁺ sensitivity of the BKα-subunit, the BKβ1-subunit also modifies the kinetics and pharmacological properties of BK channels¹⁸. The β1 subunit increases the stability of BK channel open states²⁷. If there is a decrease in the coupling ratio of α/β1 subunits in diabetes then BK channel open times should be reduced. We compared the open times of BK channels from control and diabetic retinal VSMCs by constructing open time histograms at +80 mV with 10 μmol//L Ca²⁺. Histograms were fitted with a single exponential function. BK channels from diabetic myocytes (τ_open = 2.6 ± 1.5ms) had shorter open times than those from control VSMCs (τ_open = 6.36 ± 0.54ms; p<0.05), supporting the view that β1 subunit function is decreased during diabetes. Recently it has been shown that the xenoestrogen, tamoxifen, markedly increases the P_o of BK channels but only when they are associated with a β1 subunit²⁸. Exposure of BK channels from non-diabetic cells to 1μmol/L tamoxifen increased the P_o by approximately 1.5-fold but had no significant effect on BK channels from diabetic retinal VSMCs (Fig 7C). From these data, we conclude that the reduced coupling efficiency between Ca²⁺ release and BK channel activation in diabetic retinal VSMCs results from reduced Ca²⁺ sensitivity of the BK channels, reflecting a decrease in the functional expression of the BKβ1 accessory subunit.

It has also previously been reported that BKα+β channels are markedly less sensitive to blockade by iberiotoxin when compared to BKα channels alone¹⁷. To determine whether the BKβ1 subunit might affect the sensitivity of BK channels to Pen A, we compared the effects of this inhibitor on the P_o of BK channels from non-diabetic and diabetic retinal VSMCs (Fig 7D). Application of 100nmol/L Pen A to BK channels from non-diabetic cells caused close to 100% block of channel activity. By contrast, the P_o of BK channels from diabetic cells was reduced by only 27%. These results are consistent with the possibility that BKβ1 increases the sensitivity of BK channels to Pen A, although further studies utilizing heterologous expression systems are required to confirm this.

Discussion

It has been recognised for the past 25 years that abnormal blood flow to the retina occurs during early diabetes and that this may contribute to the pathogenesis of diabetic retinopathy¹². Despite this, surprisingly little is known about the precise mechanisms linking chronic hyperglycaemia to retinal arteriolar vasoconstriction and reduced retinal blood flow prior to the onset of overt retinopathy. In the present study, we have identified a major pathophysiological mechanism that may play a specific role in the development of retinal perfusion abnormalities in diabetes. We have obtained molecular and functional data to suggest that during diabetes, the BKβ1 subunit is downregulated in retinal VSMCs, while the expression of the pore-forming α-subunit remains unaltered. Consistent with this, we observed a marked reduction in the Ca²⁺ sensitivity of BK channels and an uncoupling of BK channel activation from Ca²⁺ release in diabetic retinal VSMCs. Of particular note, the decreased expression of the β1 subunit drastically reduced the ability of spontaneous Ca²⁺ sparks to activate BK channel-mediated STOCs. Because STOCs act to hyperpolarise and relax VSMCs^{15, 29}, loss of STOC activity could well underlie the observed arteriolar vasoconstriction seen in the development of diabetic retinopathy.

Information is generally lacking regarding the effects of diabetes on vascular BK channel function. Coronary microvessels from diabetic dyslipidemic swine exhibit an uncoupling in the relationship between Ca²⁺ sparks and STOC activation³⁰ and whole-cell BK current density is reduced in microvascular smooth muscle cells of mesenteric arteries from Zucker diabetic fatty (ZDF) rats³¹ and fructose-fed, insulin resistant rats³². The precise mechanisms underlying the changes in vascular BK channel activity were not resolved in these studies, although it is interesting that an alteration in BKβ1 expression was not apparent in ZDF rats³¹. This discrepancy with the present study could possibly be attributed to the different animal models used and the origins of the microvessels studied. In general terms, however, the findings from the current work do strengthen the view that BK channel function is impaired in VSMCs of the microcirculation during diabetes. Intriguingly, impaired BK channel function may be limited to the microvasculature since the P_o of BK channels is increased in thoracic aortic VSMCs of STZ-induced diabetic mice³³.

Hypertension is approximately twice as frequent in patients with diabetes compared to patients without the disease³⁴. Furthermore work linked to the United Kingdom Prospective Diabetes Study (UKPDS) showed that the incidence of micro- and macrovascular diabetic complications was strongly associated with elevated blood pressure (BP)³⁵. Presently the etiology of hypertension in diabetic patients is not fully understood, although loss of the normal vascular relaxation to insulin in both type 1 (insulin-deficiency) and type 2 (insulin-resistance) diabetes may contribute³⁶. The present study is the first to raise the possibility that there is a selective downregulation of BKβ1 resulting in an impairment of RyR-BK channel signaling which could be an important component in the pathogenesis of hypertension in diabetic patients. However, detailed profiling of BKβ1 expression in a range of vascular tissues during diabetes, particularly in small arteries and arterioles that are central to BP regulation³⁷, is required before any firm conclusions can be made. It is likely that a widespread downregulation of BKβ1 in diabetes could be sufficient to induce hypertension since BKβ1 knockout mice are indeed hypertensive^{19, 20}. Furthermore, downregulation of BKβ1 has been reported in animal models of both acquired and genetic hypertension^{38, 39}. Also, a recent population-based epidemiological study has reported that a gain-of-function polymorphism (E65K) in the BKβ1 subunit, which increased the apparent Ca²⁺- and voltage-sensitivity of the pore-forming BKα subunit, is associated with a low prevalence of diastolic hypertension in this population⁴⁰. Future work should be directed towards a clearer understanding of the role BKβ1 downregulation in the development of hypertension in diabetic patients as well as elucidating the nature of the signaling mechanisms that regulate BKβ1 expression during diabetes.

Another novel observation in the current study was that Ca²⁺ spark amplitudes were greater in retinal VSMCs from diabetic animals, even though the frequency and duration of these events were unaltered. It seems unlikely that this represents a compensatory response to the downregulation of BKβ1 and loss of Ca²⁺ spark/BK channel coupling since the spatiotemporal properties of Ca²⁺ sparks were unaltered in VSMCs of BKβ1 knockout animals^{19, 20}. Ca²⁺ spark amplitude should depend on the number and activity (open times) of RyR channels in each spark unit and the driving force for Ca²⁺ efflux from the SR. It seems unlikely that the larger Ca²⁺ spark amplitudes can be attributed to changes in the SR Ca²⁺ load, since caffeine-evoked [Ca²⁺]_i transients and caffeine-activated Cl_Ca currents were similar in retinal VSMCs from non-diabetic and diabetic animals. Diabetic retinal vessels are known to accumulate increased levels of advanced glycation end-products⁴¹ and these adducts are known to accumulate heavily on RyR channels during diabetes⁴². The possible contribution of AGE crosslinking of RyRs to alterations in Ca²⁺ spark activity in diabetic retinal VSMCs deserves further investigation.

In summary, we have presented data that strongly supports the hypothesis that diabetes downregulates the expression of the BKβ1 subunit and consequently decreases Ca²⁺ dependent activity of BK channels in retinal VSMCs. Our findings may have important implications with respect to the early pathogenesis of diabetic retinopathy. It is also fascinating to speculate that this mechanism might contribute to the development of hypertension in diabetic patients, although more extensive studies are needed to fully evaluate this possibility.