The yin and yang of K_V channels in cerebral small vessel pathologies

Abstract

Cerebral small vessel diseases (SVDs) encompass a group of genetic and sporadic pathological processes leading to brain lesions, cognitive decline, and stroke. There is no specific treatment for SVDs, which progress silently for years before becoming clinically symptomatic. Here, we examine parallels in the functional defects of parenchymal arterioles in CADASIL, a monogenic form of SVD, and in response to subarachnoid hemorrhage, a common type of hemorrhagic stroke that also targets the brain microvasculature. Both animal models exhibit dysregulation of the voltage-gated potassium channel, K_V1, in arteriolar myocytes, an impairment that compromises responses to vasoactive stimuli and impacts cerebral blood flow autoregulation and local dilatory responses to neuronal activity (neurovascular coupling). However, the extent to which this channelopathy-like defect ultimately contributes to these pathologies is unknown. Combining experimental data with computational modeling, we describe the role of K_V1 channels in the regulation of myocyte membrane potential at rest and during the modest increase in extracellular potassium associated with neurovascular coupling. We conclude that parenchymal arteriole resting membrane potential and myogenic tone depend strongly on K_V1.2/1.5 channel density, and that reciprocal changes in K_V channel density in CADASIL and subarachnoid hemorrhage produce opposite effects on extracellular potassium-mediated dilation during neurovascular coupling.

INTRODUCTION

Dementia and stroke rank among the most pressing health issues worldwide (1,2). Cerebral small vessel diseases (SVDs), termed for various pathologies associated with small vessel dysfunction in the brain, have emerged as a central link between these two major co-morbidities. SVDs account for at least 40 % of dementia cases and more than 30 % of strokes (1,3). They encompass multiple distinct diseases that can be separated based on their underlying genetic defects, risk factors, and clinical presentations. Despite the severe nature of these diseases, there are no treatments with proven efficacy against cerebral SVDs.

Major progress has been made in identifying overlapping mechanisms involved in the functional defects of small (parenchymal) arterioles within the brain in response to subarachnoid hemorrhage (SAH) (4) and cerebral autosomal dominant arteriopathy with subcortical infarcts and leukoencephalopathy (CADASIL), a monogenic form of SVD (5). Although, strictly speaking SAH is not a cerebral SVD, an increasing body of evidence suggests that SAH-induced perfusion deficits originate from the parenchymal microvasculature (6–13). These deficits appear to be driven by changes in the balance of matrix metalloproteinase (MMP) activity which supports the novel concept that perturbations of the extracellular matrix can be a point of convergence between these cerebral small vessel pathologies (14). These changes, in turn, alter the number of voltage-gated K⁺ (K_V) channels at the arteriolar smooth muscle cells (SMCs) plasma membrane (15–18). The K_V channel superfamily, comprising 12 subfamilies (K_V1–K_V12), is one of the most diverse families of K⁺ channels identified to date (19). Molecular cloning of K⁺ channels from different cell types has revealed that K_V subfamilies share a common primary structure, reflecting hetero- or homo-multimeric assembly of four pore-forming α subunits and auxiliary β subunits, with the identity of the α subunit being the primary determinant of biophysical and pharmacological properties of individual channels (20–22). In vascular SMCs, K_V channels are active under physiological conditions, and the K⁺ efflux that they mediate serves as a key negative feedback control mechanism that opposes moment-to-moment pressure-induced constriction (23). Thus, any increase or decrease in functional K_V channel density on the plasma membrane of arteriolar SMCs is predicted to impact arteriolar diameter.

The potential dire consequences of disrupting the number of functional SMC K_V channels are evident in the two divergent cerebral small vessel pathologies presented here—SAH and CADASIL. In the hemorrhagic stroke (i.e. subarachnoid hemorrhage) model, MMP activation is understood to cause K_V channel endocytosis through activation of the epidermal growth factor receptor (EGFR), resulting in increased vasoconstriction and decreased cerebral blood flow (CBF) (15,16). On the other hand, in a mouse model of CADASIL, an increase in tissue inhibitor of metalloproteinase-3 (TIMP-3) inhibits MMP (ADAM17) activity, leading to an increased number of functional K_V channels in the plasma membrane of arteriolar SMCs and attenuation of vasoconstriction (17,18). These findings support the concept that the number of K_V channels are tuned to provide the appropriate membrane potential control in response to changes in intravascular pressure, and that any change in channel number will adversely affect cerebral arteriolar function.

Here, we weave experimental data and computational modeling to describe small vessel dysfunction after SAH and in CADASIL from a K_V channelopathy perspective, revealing how the “Yin and Yang” of SMC K_V channels—the balance between upregulation and downregulation—is crucial for maintaining parenchymal arteriole (PA) function, CBF and, ultimately, brain health.

1. K_V CHANNELS IN MYOCYTES PROFOUNDLY REGULATE THE MYOGENIC TONE OF BRAIN PARENCHYMAL ARTERIOLES

In vivo, small diameter arteries and arterioles exist in a partially constricted state, largely owing to the vasoconstrictor influence of intravascular pressure. These blood vessels can constrict further as pressure increases and, conversely, dilate when the pressure lessens (24). This response, referred to as myogenic tone, supports CBF autoregulation and protects capillaries from the disruptive effects of high blood pressure (25). Within the brain, myogenic tone also provides the vasodilatory reserve necessary for PAs to locally dilate and increase blood delivery in response to neuronal activity. This use-dependent increase in blood flow, termed functional hyperemia, is supported by a variety of mechanisms collectively referred to as neurovascular coupling (NVC) (1,26–29). Thus, the myogenic response within the cerebral vasculature, specifically in small diameter arterioles, is crucial for CBF control.

This myogenic response can be reproduced ex vivo using an arteriography system by increasing intravascular pressure through a cannula inserted into the arteriolar lumen, enabling the study of the molecular players that contribute to this phenomenon. Intravascular pressure causes a graded membrane potential (V_m) depolarization of SMCs that leads to an increase in the open-state probability of voltage-dependent Ca²⁺ channels (VDCCs), thereby enhancing Ca²⁺ influx and ultimately causing myogenic constriction (23,30). In mouse PAs, elevation of luminal pressure from 10 mmHg to 40 mmHg typically causes an 18-mV depolarization of the SMC membrane from −53 mV to −35 mV; the associated doubling in the level of myogenic tone manifests as a constriction that represents approximately a 35–40% decrease in arterial diameter (17).

Like most biological processes, myogenic tone is modulated by negative feedback elements (31). Among them, K⁺ channels in SMCs and endothelial cells (ECs) of arterioles can serve as a brake on pressure-induced depolarization and constriction (23). At a physiological extracellular K⁺ concentration ([K⁺]_o) of 3 mM and an intracellular K⁺ concentration of 140 mM, the equilibrium potential for K⁺ (E_K) is −102.7 mV. At 40 mm Hg, the estimated pressure experienced in vivo by cerebral arterioles of this size (32), the SMC membrane potential is about −35 mV, creating a 68-mV driving force for K⁺ efflux. Consequently, opening of K⁺ channels exerts a rapid and strong hyperpolarizing effect that acts to oppose pressure-induced constriction and, more broadly, provides a vasodilatory influence.

Arteriolar SMCs predominantly express four types of K⁺ channels: ATP-sensitive (K_ATP), large conductance Ca²⁺-activated (BK), inward rectifier (K_IR), and voltage-gated (K_V) (23,25,33,34). Interestingly, it appears that only K_IR and K_V channels are active under physiological conditions in PA SMCs examined in vitro. K_ATP channels are expressed in large cerebral pial arteries on the surface of the brain (35), but the absence of PA dilation in response to the K_ATP agonist cromakalim suggests a minimal role for this channel type within the brain cortex (36). The apparent lack of functional K_ATP channels in PAs is consistent with previous reports that small cerebral arterioles fail to dilate in response to cromakalim (37–40). In many vascular beds, including cortical surface arteries, BK channels are activated by local Ca²⁺-release events (Ca²⁺ sparks) through ryanodine receptors (RyRs), and exert tonic negative feedback on myogenic tone (41,42). Consistent with this tonic negative feedback element opposing myogenic tone, treatment with BK channel blockers such as paxilline causes significant and sustained constriction of pial cerebral arteries (43). In contrast, BK channel blockade does not alter PA diameter, suggesting that BK channel-mediated K⁺ efflux does not constitute a significant feedback mechanism against myogenic tone in PAs under physiological conditions (17,44–46), possibly because Ca²⁺ spark activity is very low in these arterioles (at least in rat and mouse) (47–50). However, considering that BK currents are detectable in isolated PA SMCs (9,46), and can be activated in certain conditions such as acidosis (36,50), a role for BK (and K_ATP) channels under all physiological and pathological conditions or in all species is possible.

K_IR channels, predominantly K_IR2.1 channels, are present in both PA SMCs (51–53) and ECs (54). K_IR2.1 channels conduct strong inward current at V_m negative to E_K and a small outward current that peaks and then decreases as V_m becomes increasingly positive to E_K (23,54). The role of K_IR channels in regulating smooth muscle membrane potential has been described in previous reviews (23,34,40). Inhibition of this channel with 30 µM Ba²⁺ or genetic ablation in ECs does not affect PA resting diameter (55). This is likely because channel activity at −35 mV, the resting V_m of PAs at 40 mmHg, is very low, with a conductance, gK_IR, about a 10,000-time less than the maximal conductance of this channel (54,56,57). In addition to its aforementioned dependence on V_m, K_IR is also activated by external K⁺. This enables SMCs and ECs to sense elevations in extracellular K⁺ in response to neuronal activity and mediate NVC (7,45,51,55). Therefore, the physiological role of K_IR in the brain microcirculation during NVC has been likened to that of a “vascular K⁺ electrode”, as defined by Longden and Nelson (54). The voltage and K⁺ dependence of K_IR synergize in response to small increases in external K⁺, leading to dramatic hyperpolarization that is capable of bringing V_m close to E_K and causing near maximal dilation (45,51,55).

We have established that K_V channels play a tonic and profound role in opposing vasoconstrictor influences (e.g., myogenic tone) in PAs from rodent brains (17,58,59). SMC membrane depolarization in response to pressure or vasoconstrictors increases K_V channel open probability, creating a hyperpolarizing K⁺ efflux that counterbalances depolarizing current through VDCCs and other Na⁺/Ca²⁺ permeable channels to act as a break on vasoconstriction. K_V channels in vascular SMCs show fast activation kinetics in response to membrane potential depolarization, with activation time constants (τ_act) on the order of tens of milliseconds (Figure 1A). In addition, the steady-state activation and inactivation properties of K_V allow significant K⁺ currents between −40 and +10 mV, implicating K_V channels in the regulation of PA SMC membrane potential at physiological intravascular pressures (Figure 1B). The contribution of K_V channels to the regulation of resting PA diameter is illustrated by the significant constriction caused by inhibition of K_V channels with 1 mM 4-aminopyridine (4-AP) at physiological pressure (40 mmHg) (Figure 1C) (17,23,58,60,61).

Figure 1. K_V channels exert a tonic dilatory influence on the diameter of intracerebral arterioles.

(A) Families of K_V currents from an isolated arteriolar smooth muscle cell elicited by voltage pulses from −70 mV to +50 mV in the presence of 100 nM iberiotoxin to inhibit large conductance (BK) currents. (B) steady-state activation (circles) and inactivation (triangles) properties of K_V currents measured from isolated arteriolar smooth muscle cells. Solid lines, Boltzmann fits to the data. (C) Typical recording of the internal diameter of a pressurized parenchymal arteriole (40 mm Hg) showing the constriction caused by the perfusion of the K_V blocker 4-AP, 1 and 5 mM. A and B are from (58) and C is from (17).

The voltage dependence of K_V channels has been described using a Boltzmann-type steady-state activation term (Figure 2). The characteristics of this voltage dependence—voltage at half-maximal activation (V_0.5) and slope factor (k_Kv)—appear to vary significantly among channel subtypes (17,22,23). The involvement of K_V1 channel family members, predominantly K_V1.2 and K_V1.5 subtypes, in the regulation of arteriolar tone has been documented in the cerebral vasculature of multiple species (17,58,60,62,63). Studies have also reported that K_V2 channels, primarily those of the K_V2.1 subfamily, contribute to counteracting myogenic constriction in rat pial cerebral arteries (64–67). Other groups have also provided evidence supporting the contribution of K_V7 channels (K_V7.1, K_V7.4, and K_V7.5 subtypes) to the regulation of pial cerebral arterial diameter (61,68). Regarding intracerebral PAs, a number of lines of experimental evidence support the view that K_V1.2 and K_V1.5 subtypes are predominantly expressed and primarily regulate PA SMC V_m and arteriolar diameter. First, V_0.5 values measured in PA SMCs were −3.2 mV in rat (58) and +6 mV in mouse PA SMCs (17), similar to values reported for heteromultimers of K_V1.2/1.5 channels (69). In contrast, the V_0.5 value of K_V7 channels are ~ −30 mV (70). Second, RT-PCR revealed mRNA expression of K_V1.2 and K_V1.5, but not K_V1.1, K_V1.3, K_V1.4, K_V1.6 or K_V2.1 in rat PAs (58). Third, stromatoxin, an inhibitor of K_V2 channels (66,71), does not significantly constrict mouse parenchymal arterioles (17). Furthermore, it has been shown that the 4-AP sensitivity of K⁺ currents, V_m, and tone are dependent on the expression of K_V1.2 and K_V1.5 subtypes (60,69,72–74).

Figure 2. Steady state K_V current density in PA SMCs from normal and diseased animal models.

Steady state K_V current densities for CTL, CADASIL (TgNotch3^R169C), and SAH models are fitted using a linear equation $I_{\mathit{\text{Kv}}}=\frac{1}{c_m}{\mathit{\text{mG}}}_{\mathit{\text{Kv}}}(V_m-E_K)$ with a Boltzmann-type activation term $m=\frac{1}{1+\text{exp}(-(\frac{V_m-V_{\mathit{\text{Kv}},0.5}}{k_{\mathit{\text{kv}}}}))},$, from experimental data (17) and Koide & Wellman unpublished data. C_m is the membrane capacitance; V_m is the membrane potential; G_Kv is the whole-cell conductance of K_V channels; E_K is the reversal potential for K⁺. At physiological membrane potentials pA differences in K_V currents are predicted (Figure inset). Model parameters: G_kv = 1.6 [nS]; V_Kv,0.5 = 6 [mV]; k_kv = 14 [mV] for control; G_kv = 0.8 [nS]; V_Kv,0.5 = 6 [mV]; k_kv = 14 [mV] for SAH; and G_kv = 3.2 [nS]; V_Kv,0.5 = 2.6 [mV]; k_kv = 15.8 [mV] for CADASIL model; C_m = 12.8 [pF]; [K⁺]_i = 150 [mM]; [K⁺]_o = 3 [mM].

2. IMPACT OF PATHOLOGICAL INCREASES (CADASIL) OR DECREASES (SAH) IN PA K_V CHANNEL NUMBERS

The structural organization of the cerebral circulation may explain the formation of regional pathological lesions, which are associated with small vessel dysfunction in SVDs. The two-dimensional network of surface (pial) arteries branches out and dives down into the brain parenchyma as PAs, which ultimately transition into capillary networks of interconnected vessels with a considerable capacity for redirecting blood flow (75). However, PAs upstream of the interconnected capillary networks control the blood supply of a specific volume of brain territory; as such, they represent bottlenecks to blood flow and thus are the element most vulnerable to cerebral SVDs (76,77). PAs also exhibit unique features, including the lack of extrinsic innervation and the presence of astrocytic processes that enwrap almost their entire basolateral surface (78–80). PAs are the last smooth muscle-containing vessels upstream of the capillary bed, and their active luminal diameter in mouse brain is typically < 20 µm—about one-third the width of a human hair. These anatomical features make ex vivo experimental approaches for studying their function extremely challenging, which has limited the investigation of this pathophysiologically important vascular bed.

In both SAH and CADASIL, dysregulation of brain PA reactivity precedes the onset of neurological deficits. Using well-established rabbit (15,43,81), rat (6,7) and mouse (12,13) models of SAH, and a transgenic mouse model of CADASIL (TgNotch3^R169C) in which a human NOTCH3 receptor mutation—the molecular cause of CADASIL—is overexpressed (17,82), we have discovered that K_V channel activity in PA SMCs is abnormal in both pathological conditions (17,18,59). Although abnormal K_V activity could conceivably reflect changes in channel gating properties or recruitment of new K_V channel family members, our experimental data demonstrated unchanged τ_act, V_0.5, or k in both disease models. Instead, the observed changes in K_V channel activity better correlates with a change in the number of functional channels at the SMC plasma membrane (Figure 2). Specifically, SAH causes a decrease in functional K_V channels and enhanced vasoconstriction. Conversely, CADASIL model mice exhibit an increase in SMC membrane K_V channel numbers and a decrease in cerebral arteriolar tone. One possible mechanism underlying both of these vascular pathologies is altered trafficking of K_V channels. In this context, altered shedding of the epidermal growth factor receptor ligand, HB-EGF, caused by aberrant MMP and/or ADAM (a disintegrin and metalloproteinase) activity, leads to enhanced (SAH) or decreased (CADASIL) EGFR-mediated endocytosis of K_V channels (14–18).

Based on current density data obtained experimentally using the perforated-patch configuration of the patch-clamp technique (17) (Koide & Wellman unpublished data), we calculated the number of K_V channels in control (CTL), CADASIL and SAH conditions at a given voltage (−40 mV) using the Goldman–Hodgkin–Katz constant field equation ((83); see (17) for calculation details). We found that the average number of functional K_V channels per myocyte in CTL mice was 3,060 ± 479 (6 cells from 6 animals), a number that increased to 4,970 ± 655 (6 cells from 6 animals) in CADASIL (TgNotch3^R169C) mice. A comparison of CTL and SAH rats revealed the opposite change, with 3,465 ± 346 channels per myocyte observed in CTL rats (6 cells from 5 animals) compared with 1809 ± 423 in SAH rats (6 cells from 6 animals) (Koide & Wellman unpublished data). In each case, the cell capacitances of myocytes were similar, indicating a similar membrane surface area in all three conditions. Therefore, the CADASIL-causing mutation results in a ~57 % increase in K_V channel number, whereas SAH reduces K_V channel number by ~48 %. The differences in K_V current densities between disease and control animals, at different membrane potentials, are depicted in Figure 2. Steady-state K_V current data for CTL and CADASIL mice are fitted to a Boltzmann-type Vm-dependent activation, as previously done (17). The voltage dependency of K_V1 in SAH is predicted based on a 48% reduction in K_V1 conductance relative to CTL. SMC outward K_V1 current increases dramatically at depolarized potentials, reflecting K_V channel opening with depolarization, and current density correlated with the K_V channel number in all conditions.

Given the role of K⁺ efflux through K_V channels in opposing pressure-induced membrane depolarization and vasoconstriction, the prediction is that increased (CADASIL) (17) or decreased (SAH) (6,84) K_V channel numbers would alter the degree of negative-feedback control, resulting in hyperpolarization (CADASIL) or depolarization (SAH) of PA SMCs. Consistent with this prediction, experimentally measured V_m was hyperpolarized (−45 mV) (17) and depolarized (−28 mV) (6) in pressurized (40 mmHg) PAs obtained from CADASIL mice and SAH model animals, respectively, compared to CTL groups (−35 mV). The relationship between myogenic tone and membrane potential in CTL, CADASIL and SAH animals is summarized in Figure 3A. At lower intravascular pressure (5 and 10 mmHg), tone and membrane potential did not differ among the three groups. However, as pressure is increased to a more physiological level (40 mmHg), V_m in PAs from CADASIL mice remained almost 10 mV more hyperpolarized than that in CTL PAs (−45 mV vs. −35 mV), whereas V_m was pulled to −28 mV in SAH (6,17).

Figure 3. Relationship between myogenic tone and membrane potential.

(A) Values of membrane potential and myogenic tone at different intravascular pressure (mm Hg) from CADASIL (TgNotch3^R169C, blue triangles) and SAH (red circles) animals are consistent with the linear regression obtained from CTL animals (black triangles represent CTL mice from (17) and black circles represent CTL rats from (6)) showing a similar relationship between tone and membrane potential. (B) A detailed model of SMC membrane potential and Ca²⁺ dynamics was adapted from (102) and modified by incorporating the K_V1 current of PA SMCs from CTL animals (Figure 2), while adjusting other transmembrane currents (K_IR, NSC, VDCC, NaK, PMCA) to produce resting V_m and Ca²⁺ concentration in agreement with experimental data (17,103). The effect of altered K_V1 channel density in CADASIL (blue triangles) and SAH (red triangles) was examined assuming all other model parameters remain the same as in CTL (black circles). (C) The effect of increasing pressure was simulated by depolarizing SMC membrane through increasing Na⁺ permeability (P_Na). Model simulations, in agreement with the corresponding experiments in (A), show differences between CADASIL and SAH animals in V_m (bottom) and Ca²⁺ (top) as pressure increases and highlight the inhibitory role of K_V channels and the effect K_V channel density on myogenic tone. Parameters as in reference (102) except: P_VDCC=6.3×10⁻⁵ cm/s; P_NaNSC=1.23×10⁻⁶ cm/s; I_PMCA=8.58 pA; I_NaK=7.76 pA/pF; G_KIR = 0.5 nS/(mM)^0.5; G_Na,leak=0.12 nS; C_m = 12.8 pF.

We also used a mathematical model of SMC membrane permeability to examine whether the reported changes in V_m could be explained by measured differences in K_V density between CADASIL and SAH animals and CTL groups (6,17) (Figure 3B). This model accounts for the activity of important SMC components, including K_V1, K_IR, BK, K_ATP, VDCC, and non-selective cation (NSC) channels, as well as Na⁺/K⁺ ATPases (NaK) and plasma membrane Ca²⁺ pumps (PMCA). Figure 3C demonstrates the effect of increased (CADASIL) and decreased (SAH) K_V channel density under the assumption that the presence of all these other components remains unchanged. Simulations show the effect of increasing transmembrane Na⁺ permeability (P_Na) accounts for the pressure-induced smooth muscle depolarization in myogenic response, presumably through the opening of stretch activated NSC channels. In agreement with the experimental data (Figure 3A; 5–10 mmHg), the difference in membrane potential between control and disease states is negligible at low P_Na current (i.e., low intraluminal pressure) as the contribution of K_V currents to the total transmembrane current diminishes (Figure 3B and 3C). The model predicts that high P_Na current (i.e., high intraluminal pressure) leads to depolarization and opening of K_V channels, highlighting the inhibitory role of K_V channels on the myogenic response. A larger number of K_V channels (CADASIL) provide greater negative feedback, resulting in hyperpolarized V_m, less VDCC activity, lower intracellular Ca²⁺, and less myogenic tone. Conversely, a decrease in the number of K_V channels (SAH) provides less negative-feedback, resulting in more depolarized V_m, increased VDCC activity, higher intracellular Ca²⁺, and increased myogenic tone. Thus, the model demonstrates that changes in SMC K_V channel density can account for changes in V_m, intracellular Ca²⁺ concentration and myogenic tone, consistent with observations from CADASIL and SAH model animals. In both pathological conditions, a change in the gain of the K_V-mediated negative feedback loop is expected to have profound effects on the responses of arterioles to pressure fluctuations (CBF autoregulation), but also to impact the effect of external K⁺ variations during NVC.

3. IMPACT OF SMALL VESSEL PATHOLOGIES ON THE INTERPLAY BETWEEN K_V AND K_IR CHANNELS DURING NVC

NVC is the process that links localized neuronal activity in the brain to vasodilation of proximate PAs so as to ensure adequate delivering of oxygen and nutrients to metabolically active regions of the central nervous system. A key component of NVC is the K_IR-dependent SMC hyperpolarization and vasodilation that results from modest increases in [K⁺]_o within the perivascular space (51,52,85). In isolated PAs, raising [K⁺]_o from 3 mM to 8 mM hyperpolarizes the SM V_m to near the new E_K of −76 mV, causing near maximum vasodilation by decreasing Ca²⁺ influx (51). This dramatic V_m hyperpolarization is achieved by K_IR activation. Raising [K⁺]_o from 3 mM to 8 mM activates K_IR channels and results in a 50-fold increase in membrane K⁺ permeability (86). In addition to a modest rise in [K⁺]_o (e.g., from 3 mM to 8 mM), membrane hyperpolarization itself increases K_IR channel activity, causing a rightward shift of the K_IR current-voltage relationship and increased outward current amplitude (Figure 4A, left inset) as a result of the positive shift in E_K and an increase in K_IR conductance (87). In contrast, a modest rise in [K⁺]_o monotonically decreases outward K_V current due to a reduction in the K⁺ chemical gradient (Figure 4A, right inset). Additionally, modest increases in [K⁺]_o and the associated K_IR channel activation and V_m hyperpolarization would be expected to decrease K_V (and possibly BK) channel open probability. Thus, the interaction between K_IR and K_V/BK channels during modest increases in [K⁺]_o creates a “tug-of-war” dynamic that would be expected to play an important role in determining SMC K⁺ permeability and PA diameter during NVC. Sufficient K_IR activation is required to overcome the concurrent decrease in K_V/BK channel activity to produce V_m hyperpolarization and vasodilation. Under physiological conditions, this is the case, with modest increases in [K⁺]_o associated with neuronal activation leading to SMC V_m hyperpolarization, vasodilation, and increased local CBF during NVC.

Figure 4. Effect of PA SMC K_V current density and the interplay with K_IR current on V_m dynamics at rest and during potassium challenge.

(A) Combined contribution of K_IR and K_v currents in healthy and diseased models during rest and [K⁺]_o stimulus. Activation of K_IR current by [K⁺]_o and hyperpolarization is accounted: $I_{\mathit{\text{Kir}}}=\frac{G_{\mathit{\text{Kir}},\mathit{\text{max}}}(V_m-E_K)}{1+\text{exp}(\frac{(V_m-V_{K_{\mathit{\text{IR}}},0.5})}{k_{\mathit{\text{Kir}}}})};G_{\mathit{\text{Kir}},\mathit{\text{max}}}=G_{\mathit{\text{Kir}}}{[K^{+}]}_o^{0.5}$ where G_kir,max is the maximal K_IR conductance. Solid lines show the sum of the two currents at rest, and dashed lines are during elevation of [K⁺]_o from 3 to 8 mM. The shaded regions show the range of voltages within which K_IR current increases more than K_V current decreases during the K⁺ stimulus, i.e. the resting V_m window where the K⁺ challenge will result in hyperpolarization. As K_V current density increases (from SAH, red lines; to CTL black lines; to CADASIL, blue lines) the window shrinks in size and shifts to more hyperpolarized potentials. (B) Representative simulation using the model of PA SMCs from Figure 3. SMCs from CTL (black line), CADASIL (blue line) and SAH (red line) conditions hyperpolarize following an increase in extracellular K⁺, [K⁺]_o, from 3 mM to 8 mM. The change in membrane potential is less for CADASIL as a result of the more hyperpolarized resting V_m prior to the K⁺ challenge. G_KIR = 0.76 [nS/(mM)^0.5]; k_KIR = 7 [mV]; V_KIR,0.5= E_K+12 [mV].

However, evidence indicates impaired NVC in CADASIL (14,18) and SAH (7,13) animal models. Since K_V channel number in PA SMCs is significantly altered in these two pathologies, we used computational modeling to investigate the effect of different K_V channel densities on SMC membrane potential when [K⁺]_o is elevated from 3 mM to 8 mM. The current-voltage curves of K_V currents obtained from Figures 1 and 2 are summed with representative PA SMC K_IR currents in Figure 4A. Solid lines represent combined K_V and K_IR currents at rest ([K⁺]_o = 3 mM), and dashed lines represent currents when [K⁺]_o is elevated to 8 mM. Shaded areas show the increase in net hyperpolarizing current predicted by shifting [K⁺]_o from 3 mM to 8 mM and highlight the V_m window where the K_IR influence is dominant (i.e., overcomes the opposing influence of decreased K_V channel activity), producing a net increase in membrane K⁺ permeability. Within this V_m range, or “K_IR window”, hyperpolarization can be achieved during increases in [K⁺]_o to 8 mM, whereas outside of this window, the K⁺ stimulus will result in unstable membrane potential or even depolarization. Therefore, K⁺-induced hyperpolarization is possible when the resting V_m is within the K_IR window (Figure 4A shaded areas). Despite differences in K_V current density, our modeling indicates that conditions exist where increasing [K⁺]_o from 3 mM to 8 mM would be predicted to hyperpolarize PA SMCs of CTL, CADASIL, and SAH animals (Figure 4B).

In SAH, the prediction is that decreased K_V channel density extends the V_m window where the K_IR influence is dominant (Figure 4A red region). This is consistent with our results showing that parenchymal arterioles isolated from the brains of SAH animals and pressurized ex vivo dilate in response to modest increases in extracellular K⁺ (7). However, SAH is associated with an inversion of NVC in animal models; that is, instead of causing vasodilation, neuronal activation causes vasoconstriction in these animals, both in brain slices and in vivo (7,13,88). We have proposed that this inversion of NVC is the result of a pathological increase in basal [K⁺]_o reflecting enhanced K⁺ efflux by astrocytic endfeet, rather than impaired K_IR function. When summed with neurally evoked K⁺ efflux, the net elevation in basal [K⁺]_o leads to a more depolarized SMC E_K that lies outside of the influence of the K_IR window. Thus, the polarity of the vascular response is switched from dilation to constriction due to V_m depolarization and enhanced Ca²⁺ entry through VDCCs (7,13,88).

In CADASIL, the representative model predicts that an increase in K_V channel density reduces the size of the K_IR window, but also shifts the window to more hyperpolarized potentials (Figure 4A, blue region). However, increased K_V channel density in CADASIL model mice also brings resting V_m to more hyperpolarized values (17). According to the simulation, this limits the impact of the increased K_V density on a 8-mM [K⁺]_o challenge, and allow K⁺-induced dilations to occur. Consistent with this, PAs from CADASIL model mice respond to 8, 15 and 20 mM [K⁺]_o with near maximum dilation ex vivo, comparable to that in CTL animals (17). Nevertheless, an altered hyperpolarization window may compromise K-induced dilation under some conditions, for example when V_m is more depolarized by the presence of an additional vasoconstrictor. Also, a higher K_V channel density in CADASIL decreases the vasodilatory reserve, because PAs are in a less constricted state compared with controls. This could lead to deficits in vascular autoregulation and the ability of K⁺-induced signaling to efficiently redirect blood flow in the vascular network during NVC, as previously reported (14,18).

CONCLUSIONS

Our understanding of cerebral SVDs has advanced greatly over the past decade. Recent work has established the contribution of the yin and yang of the K_V channel balance to the pathological progression of cerebral small vessel dysfunction. Here, we combine data from different animal models with detailed computational modeling to further understand elements of this pathology.

K_V channels, estimated to number 3,000–3,500 per SMC in physiological conditions, play a profound role in regulating PA SMC resting V_m and PA myogenic tone. These channels are sensitive to inhibition by 4-AP, and the current footprint obtained in native PA myocytes aligns with the properties of K_V1.2 and K_V1.5 subunits, results supported by the expression of mRNAs for these subunits (17,58,59). Changing the number of channels per SMC results in abnormal SMC V_m and myogenic responses—a yin and yang dynamic that helps to account for cerebral microvascular defects in both CADASIL and SAH. Indeed, in addition to being dramatically altered by CADASIL and SAH, K_V channel expression and function in the vascular wall can also be disrupted in the context of other major causes of cerebral SVDs, such as diabetes (58), aging (89), and hypertension (34,90). Thus, targeting K_V1 channels in the vascular wall with the aim of restoring normal hemodynamic function may be a future therapeutic option for such disorders (91).

Secondly, an in-depth characterization of the SMC membrane K⁺ permeability control by the K_V−K_IR interplay during NVC will enable further advances in our understanding of the impact of SVDs on functional hyperemia. Here, we establish an inroad to this analysis, introducing the new concept that changes in the number of K_V channels impact the tug-of-war between activation of SMC K_IR channels and deactivation of K_V channels, which can impair sensitivity to vasoreactive signaling.

Overall, increased K_V channel density in CADASIL narrows the V_m range over which K_IR channel can induce dilation, but also brings resting V_m to more hyperpolarized values, partially counteracting the impact of the tug-of-war between K_V and K_IR channels at the expense of the vasodilatory reserve. In SAH, decreased K_V channel density leads to depolarization of resting V_m and increases myogenic tone. Compromised NVC in this disease model may be attributable to high extracellular K⁺ levels resulting from excessive K⁺ efflux from astrocytes that depolarizes rather than hyperpolarizes SMCs.

FUTURE DIRECTIONS

Recent studies by our group and others have identified a previously unanticipated role for the endothelium, specifically K_IR channels in capillary ECs, in sensing neural activity at the capillary level and translating it into a propagating hyperpolarizing electrical signal that dilates upstream arterioles (55,92,93). These findings highlight the importance of conducted hyperpolarization along PAs and hold the promise of resolving controversies regarding SVD-induced neurovascular dysfunction, potentially providing a paradigm-shifting concept. The recent development of cell-type–specific, genetically encoded fluorescent voltage sensors brings the possibility of in vivo and ex vivo optical electrophysiology within reach (94–97). In the context of electrical signaling between capillaries and arterioles, this may allow us to image how changes in K_V channels alter the regenerative hyperpolarization of the endothelium during NVC. Importantly, CADASIL is caused by mutations in the NOTCH3 receptor, which is expressed not only in vascular SMCs, but also in pericytes (5). The role of pericytes, and their possible dysfunction in NVC, has recently been brought to light by several groups (93,98–101). Although consensus on this point has remained elusive, the increasing number of studies suggests that pericytes provide an additional layer of CBF regulation and therefore could play a key role in facilitating or dampening capillary-to-arteriole signaling. However, whether K_V channels are expressed in pericytes, and whether they are up-regulated in CADASIL and potentially down-regulated in SAH, remains unknown. Resolving these questions may prove critical to a full understanding of the regulation of CBF by K_V channels in the intracerebral microcirculation.

PERSPECTIVE

K_V channels located in SMCs of brain parenchymal arterioles oppose pressure-induced depolarization. Changes in K_V channel density occur in pathological processes that target the brain microcirculation, impairing intracerebral arteriole constriction in response to changes in intravascular pressure. This blunting of a fundamental vascular function is expected to impact cerebral blood flow autoregulation and local dilation in response to neuronal activity (functional hyperemia). Interventions aimed at restoring K_V channel function to improve brain perfusion may thus be a future therapeutic direction in the treatment of cerebral small vessel pathologies.

Abbreviations used

4-AP

4-aminopyridine

[K⁺]_o

external K⁺ concentration

ADAM

a disintegrin and metalloproteinase

BK

large conductance Ca²⁺- activated K⁺ channel

CADASIL

cerebral autosomal dominant arteriopathy with subcortical infarcts and leukoencephalopathy

CBF

cerebral blood flow

CTL

control

EC

endothelial cell

EGFR

epidermal growth factor receptor

E_K

reversal potential for K⁺

HB-EGF

heparin-binding epidermal growth factor-like growth factor

K_ATP

ATP-sensitive K⁺ channel

K_IR

inward rectifier K⁺ channel

K_V

voltage-gated K⁺ channel

MMP

matrix metalloproteinase

NaK

Na⁺/K⁺ ATPase

NSC

non-selective cation channels

NVC

neurovascular coupling

PA

parenchymal arteriole

PMCA

plasma membrane Ca²⁺ ATPase

RyR

ryanodine receptor

SAH

subarachnoid hemorrhage

SMC

smooth muscle cell

SVD

small vessel disease

TIMP-3

tissue inhibitor of metalloproteinase-3

VDCC

voltage-dependent Ca²⁺ channel

V_m

membrane potential