Loss of cerebrovascular Shaker-type K⁺ channels: a shared vasodilator defect of genetic and renal hypertensive rats

Abstract

The cerebral arteries of hypertensive rats are depolarized and highly myogenic, suggesting a loss of K⁺ channels in the vascular smooth muscle cells (VSMCs). The present study evaluated whether the dilator function of the prominent Shaker-type voltage-gated K⁺ (K_V1) channels is attenuated in middle cerebral arteries from two rat models of hypertension. Block of K_V1 channels by correolide (1 μmol/l) or psora-4 (100 nmol/l) reduced the resting diameter of pressurized (80 mmHg) cerebral arteries from normotensive rats by an average of 28 ± 3% or 26 ± 3%, respectively. In contrast, arteries from spontaneously hypertensive rats (SHR) and aortic-banded (Ao-B) rats with chronic hypertension showed enhanced Ca²⁺-dependent tone and failed to significantly constrict to correolide or psora-4, implying a loss of K_V1 channel-mediated vasodilation. Patch-clamp studies in the VSMCs of SHR confirmed that the peak K⁺ current density attributed to K_V1 channels averaged only 5.47 ± 1.03 pA/pF, compared with 9.58 ± 0.82 pA/pF in VSMCs of control Wistar-Kyoto rats. Subsequently, Western blots revealed a 49 ± 7% to 66 ± 7% loss of the pore-forming α_1.2- and α_1.5-subunits that compose K_V1 channels in cerebral arteries of SHR and Ao-B rats compared with control animals. In each case, the deficiency of K_V1 channels was associated with reduced mRNA levels encoding either or both α-subunits. Collectively, these findings demonstrate that a deficit of α_1.2- and α_1.5-subunits results in a reduced contribution of K_V1 channels to the resting diameters of cerebral arteries from two rat models of hypertension that originate from different etiologies.

during the pathogenesis of chronic hypertension, the cerebral circulation develops an elevated myogenicity that limits the transmission of high systemic pressure to the blood-brain barrier (6, 8, 19, 20, 30, 34). The increased myogenicity is associated with a reduced K⁺ permeability of the vascular smooth muscle cells (VSMCs), resulting in membrane depolarization, opening of voltage-gated Ca²⁺ channels, and vasoconstriction (27, 28, 41, 46). Although a reduced K⁺ permeability during hypertension infers a loss of functional K⁺ channels in the VSMCs, the identity of the K⁺ channel type(s) that loses function remains highly controversial. At least three different gene families of K⁺ channels regulate cerebrovascular tone under normal conditions, including the inwardly rectifying K⁺ (K_IR) channels, the large-conductance Ca²⁺-activated K⁺ (BK_Ca) channels, and the voltage-gated K⁺ (K_V) channels of the Shaker (K_V1) and Shab-type (K_V2) subtypes (2–4, 9–11, 14, 32, 35, 37, 38, 40, 53). The stimulus of high blood pressure could potentially reduce the activation or expression level of one or more of these K⁺ channel families in cerebral resistance vessels to compromise vasodilation.

In this context, we previously reported (2) that the Shaker-type voltage-gated K⁺ (K_V1) channels are major contributors to the resting diameter of rat cerebral arteries. Thus it would be logical to assume that the depolarization and the increased myogenicity of cerebral arteries exposed to hypertension may relate to this predominant channel. Hence, the present study was designed to determine whether K_V1 channels show attenuated function leading to a reduced contribution to the resting diameter of cerebral arteries in hypertensive rats and to investigate the molecular alterations. It has been shown that underlying biochemical disturbances, blood pressure profiles, end-organ damage, and drug sensitivities differ between rat models of hypertension (44), suggesting that the ion channel profile in the vasculature also may be different. Thus we searched for possible abnormalities of K_V1 channel expression and function in two different rat models of hypertension, the spontaneously hypertensive rat (SHR) and the aortic-banded (Ao-B) rat.

METHODS

Animals and removal of arteries.

Fourteen- to eighteen-week-old male Wistar-Kyoto (WKY) rats and SHR were purchased from Taconic Farms (Germantown, NY). Ten-week-old male Sprague-Dawley rats purchased from Harlan Laboratories (Madison, WI) were used for interrenal aortic banding to produce Ao-B and sham-operated (Sham) rats as described previously (43). The ligature in the Ao-B rats was secured to allow only partial return of blood flow through the banded aorta, while in Sham rats the aortic ligature was not tightened. After surgery, the animals were kept for 6–8 wk to permit hypertension to develop in the Ao-B rats. Mean systolic blood pressure was measured in each animal by cannulation of the carotid artery under anesthesia induced by ketamine (100 mg/kg im) and acepromazine (1 mg/kg im). The average systolic blood pressures in WKY rats and SHR were 117 ± 2 (n = 51) and 188 ± 2 (n = 48) mmHg, respectively. The average systolic blood pressures in Sham and Ao-B rats were 116 ± 2 (n = 50) and 189 ± 2 (n = 33) mmHg, respectively. After the blood pressure measurement, the rats were decapitated for removal of the brain. The brain was placed in physiological salt solution (PSS) of the following composition (mmol/l): 145 NaCl, 4 KCl, 1 MgCl₂, 10 HEPES, 1.8 CaCl₂, and 10 glucose (pH = 7.4). The middle cerebral arteries were removed for vessel perfusion and patch-clamp studies as previously described (2). The entire circle of Willis and its branches were removed to evaluate channel transcript and protein expression by real-time reverse transcription-polymerase chain reaction (RT-PCR) and Western blotting, respectively. Procedures involving animals were approved by the Institutional Animal Care and Use Committee of the University of Arkansas for Medical Sciences and the Medical College of Wisconsin.

Isolated vessel perfusion.

Middle cerebral arteries from WKY rats, SHR, and Sham and Ao-B rats were cannulated and pressurized (80 mmHg) on glass micropipettes with a standard perfusion system and solutions as previously described (2). Internal diameter was measured by videomicroscopy calibrated for micrometer resolution, and membrane potential (E_m) was measured with glass microelectrodes (2). The normal function of VSMCs and endothelial cells was confirmed by a vasoconstrictor response to 1 μmol/l 5-hydroxytryptamine (5-HT) and subsequent dilation to 1 μmol/l acetylcholine (ACh), respectively. In some experiments, normal function of VSMCs was confirmed by observing a vasoconstrictor response to 60 mmol/l KCl. Correolide (Cor) at 1 μmol/l concentration and/or 5-(4-phenylalkoxypsoralen) (psora-4) (49) at cumulative concentrations of 30 and 100 nmol/l were added to the bath solution to preferentially block K_V1 channels, and steady-state diameter responses were measured after 30 min of drug exposure.

Measurement of K⁺ current.

Single VSMCs from WKY and SHR middle cerebral arteries were enzymatically isolated as described previously (33) for analysis on a patch-clamp station with a List EPC-7 amplifier (List Instruments, Darmstadt, Germany) and a microcomputer running pCLAMP 6.0 software (Axon Instruments, Foster City, CA). Whole cell K_V current was elicited in freshly dissociated VSMCs by progressive 8-mV depolarizing steps (400 ms, 5-s intervals) from −70 mV to +58 mV. As detailed previously (2), contaminating current through BK_Ca channels was minimized by using a low Ca²⁺ concentration (10 nmol/l) in the recording solutions and including 100 nmol/l iberiotoxin in the bath solution. Standard single-pulse protocols were used to assess the voltage dependence of activation of Cor-sensitive K_V current (2). Activation curves were constructed by applying 500-ms voltage steps from −70 mV to +58 mV in 8-mV increments to VSMCs exposed to symmetrical KCl (145 mmol/l). The peak amplitudes of inward tail currents were recorded upon return to a holding potential of −70 mV and normalized against maximal tail current amplitude. The Cor-sensitive component of voltage-gated K⁺ current that represented current attributed to K_V1 channels was obtained by subtracting tail currents obtained in the presence of Cor from control tail currents obtained in drug-free solution.

RT-PCR.

Cerebral arteries were pooled from two or three rats, and total RNA was isolated with the RNeasy Protect Mini Kit (Qiagen, Germantown, MD) according to the manufacturer's instructions. Total RNA was subjected to DNase treatment using a DNA-free Kit (Ambion, Austin, TX) to remove contaminating DNA. Five hundred nanograms of the RNA isolate was reverse transcribed (iScript cDNA Synthesis Kit; Bio-Rad, Hercules, CA) to generate cDNA per the manufacturer's instructions. The cDNA sequences for the pore-forming α_1.2- and α_1.5-subunits, which coassemble as tetramers to form K_V1 channels in rat cerebral VSMCs (2), were amplified by real-time RT-PCR (iQ SYBR Green Supermix Kit, Bio-Rad) using subunit-specific primers on an iCycler iQ Multicolor Real-Time PCR Detection System (Bio-Rad). The primer target sequences, the expected product sizes, and relevant GenBank entries were 1779-1905, 127 bp, and NM_012970 (rat α_1.2) and 1533-1638, 106 bp, and NM_012972 (rat α_1.5), respectively. Smooth muscle-specific α-actin (169-280, 112 bp, X06801) was used as an internal standard. Each amplification reaction contained 2.5–5 ng of the cDNA product. The melting curve analysis of the real-time PCR generated single peaks, and only single bands of the predicted size were detected on a 2% agarose gel (Supplemental Fig. S1, A–C). 1 Resulting bands were subsequently excised and confirmed by direct DNA sequencing to be fragments matching the GenBank sequences. Standard curves for α_1.2- and α_1.5-subunits generated by twofold serial dilutions of cDNA from 25 ng to 1.56 ng per amplification reaction resulted in one-cycle (1.0 ± 0.1 and 1.1 ± 0.1, respectively) increases in the threshold cycle (C_t), demonstrating sensitive detection of changes in transcript level in this range (Supplemental Fig. S1, D and E). Control reactions contained all other components except the cDNA template. The transcript level of the sodium channel α-subunit (Na_V1.1, NM_030875), which is abundantly expressed in neurons but not in VSMCs, was 340-fold less abundant in the cerebral arteries compared with whole brain preparation (Supplemental Fig. S2A) when quantified by RT-PCR using a previously reported primer pair (17). On the other hand, α_1.2 transcript was 2-fold less in cerebral arteries (Supplemental Fig. S2B) and α_1.5 transcript was 49-fold more in cerebral arteries than in the whole brain preparation (Supplemental Fig. S2C). The latter finding agrees with earlier reports that brain tissue does not express α_1.5-subunits (7, 50). Hence, the contribution from brain tissues to our real-time PCR products evaluating the relative abundance of α_1.2 and α_1.5 transcripts in cerebral arteries was deemed to be insignificant. The relative abundances of the α_1.2 and α_1.5 transcripts in the cerebral arteries of WKY, SHR, Sham, and Ao-B rats were estimated by the ΔΔC_t method (39) and reported as percentage of WKY or Sham rat expression.

Western immunoblot.

The protein expression of the α_1.2- and α_1.5-subunits in cerebral arteries was detected with monoclonal antibodies obtained commercially (NeuroMab Facility, Davis, CA). Each sample of cerebral arteries pooled from three or four rats was homogenized on ice in Triton lysis buffer of the following composition (mmol/l): 150 NaCl, 1 EGTA, 1.5 MgCl₂, and 50 HEPES, with 1% Triton X-100 and 10% glycerol (pH = 7.3). The lysis buffer also contained protease inhibitors of the following concentrations (mg/ml) to prevent protein degradation: 0.5 leupeptin, 0.5 aprotinin A, 0.5 antipain, and 1.74 phenylmethylsulfonyl fluoride. Large tissue debris and nuclear fragments from the homogenate were removed by two low centrifuge spins (1,000 g for 10 min; 14,000 g for 10 min) at 4°C, and the membrane fractions were obtained after a subsequent centrifugation at 100,000 g for 1 h to perform Western blotting. Smooth muscle-specific α-actin (catalog no. A2547, Sigma, St. Louis, MO) was used to ensure equal lane loading. Immunoreactive bands were detected by enhanced chemiluminescence (Amersham Pharmacia Biotech, Little Chalfont, UK), and band intensity was quantitated by densitometry using Scion Image software and expressed as a percentage of the WKY or Sham signal. The purity of the cerebral artery protein preparation from the underlying brain tissue was confirmed by the absence of signal from antibodies directed against a sodium (Na_V1.1) channel α-subunit (NeuroMab Facility) that is densely expressed in brain but not in vascular tissue (Supplemental Fig. S3).

Drugs.

Cor, a kind gift from Drs. Maria Garcia and Gregory Kaczorowski (Merck Research Laboratories, Rahway, NJ), was prepared as a 20 mmol/l stock in dimethyl sulfoxide (DMSO) and stored at 4°C. Psora-4 (Sigma) was prepared as a 3 mmol/l stock in DMSO and stored at −20°C. The DMSO solvent did not alter the resting diameter of cerebral arteries or amplitudes of K_V1 current (2). Iberiotoxin (Sigma) was reconstituted as a 100 μmol/l stock in PSS and stored in aliquots at −20°C. Drug concentrations in studies refer to final bath concentrations.

Statistical analysis.

Data are expressed as means ± SE. Data were analyzed with ANOVA with repeated measures and paired t-tests, as appropriate. A probability value of <0.05 was considered statistically significant.

RESULTS

Reduced contribution of K_V1 channels in SHR arteries.

Before any experiments, cannulated middle cerebral arteries from WKY rats and SHR were tested for 5-HT (1 μmol/l)-mediated constriction and ACh (1 μmol/l)-induced dilation to confirm their viability. The arteries from WKY rats robustly constricted in response to 5-HT by 41 ± 5%, and subsequent application of ACh relaxed the arteries by 33 ± 7% (n = 8 each) (Fig. 1). Similarly, the arteries from SHR constricted by 42 ± 5% in response to 5-HT and dilated by 41 ± 9% (n = 11 each) after the subsequent addition of ACh (Fig. 1). We had previously reported (2) that Shaker-type K_V1 channels contribute significantly to the resting diameter of cerebral arteries of Sprague-Dawley rats. In the present study, the contribution of K_V1 channels to the resting diameter of WKY and SHR cerebral arteries was compared by analyzing the diameter response to Cor (1 μmol/l), a previously characterized pharmacological blocker of these channels (2, 21, 26, 36). A representative example of cerebral arterial diameters from WKY rats and SHR before and after treatment with Cor is shown in Fig. 2A, with average values displayed in Fig. 2B. The internal diameter of WKY arteries was 154 ± 10 μm after a 1-h equilibration period at a perfusion pressure of 80 mmHg. The WKY vessels constricted to 112 ± 9 μm, a 28% decrease in diameter, in response to pharmacological block of K_V1 channels by 1 μmol/l Cor (n = 8). In contrast, SHR cerebral arteries showed a smaller diameter of 90 ± 5 μm under the same conditions and were insensitive to 1 μmol/l Cor, suggesting a reduced contribution of K_V1 channels to the resting diameter of these vessels (n = 11).

Fig. 1.

Effect of 5-hydroxytryptamine (5-HT) and acetylcholine (ACh) on the resting diameter of Wistar-Kyoto (WKY) rat and spontaneously hypertensive rat (SHR) cerebral arteries. Isolated, cannulated middle cerebral arteries were perfused at an intraluminal pressure of 80 mmHg. Average diameters of cerebral arteries from WKY (n = 8) and SHR (n = 11) in drug-free solution (Con), after addition of 1 μmol/l 5-HT, and after subsequent addition of 1 μmol/l ACh are shown. *Significant difference between diameters before and after 5-HT treatment; +significant difference between diameters before and after subsequent addition of ACh.

Fig. 2.

Effect of correolide (Cor) on the resting diameter of WKY and SHR cerebral arteries. Isolated, cannulated cerebral arteries were perfused at an intraluminal pressure of 80 mmHg. A: images captured by videomicroscopy from a representative experiment showing the diameter response of WKY and SHR arteries to 1 μmol/l Cor. B: average diameters of WKY arteries (n = 8) and SHR arteries (n = 11) before (Con) and after Cor treatment. *Significant difference between WKY arterial diameters in control and Cor-containing solution; +significant difference between WKY and SHR under the same conditions (P < 0.05).

To further evaluate whether the contribution of K_V1 channels to the resting diameter of cerebral arteries is reduced in SHR, another set of experiments was performed with psora-4 (49), another putative preferential blocker of K_V1 channels. The average diameters from five experiments are displayed in Fig. 3. The internal diameter of WKY arteries was 148 ± 6 μm (n = 5) after a 1-h equilibration period at a perfusion pressure of 80 mmHg. The cumulative application of 30 and 100 nmol/l psora-4 reduced the arterial diameter to 115 ± 5 μm (22 ± 3% decrease) and 110 ± 6 μm (26 ± 3% decrease), respectively. Addition of 1 μmol/l Cor to arteries pretreated with 100 nmol/l psora-4 did not elicit further constriction (110 ± 7 μm). The diameter response to psora-4 (−39 ± 4 μm) was almost identical to the diameter response of similar WKY arteries to 1 μmol/l Cor (−42 ± 9 μm, Fig. 2B). In contrast, SHR cerebral arteries showed a smaller resting diameter of 89 ± 3 μm (n = 5), did not significantly constrict in response to 30 (87 ± 3 μm) or 100 (85 ± 4 μm) nmol/l psora-4, and also did not respond to further addition of 1 μmol/l Cor (86 ± 4 μm). In all arteries, 60 mmol/l KCl induced a strong diameter reduction that verified functional VSMCs (Fig. 3). Thus the lack of response of SHR cerebral arteries to both 1 μmol/l Cor and 100 nmol/l psora-4 strongly suggests that the dilator influence of K_V1 channels is absent in these arteries.

Fig. 3.

Effect of 5-(4-phenylalkoxypsoralen) (psora-4) on resting diameter of WKY and SHR cerebral arteries. Isolated, cannulated cerebral arteries were perfused at an intraluminal pressure of 80 mmHg. Average diameters of cerebral arteries from WKY (n = 5) and SHR (n = 5) in drug-free solution (Con), after cumulative addition of 30 and 100 nmol/l psora-4, and after the subsequent addition of 1 μmol/l Cor are shown. At the end of each experiment, arteries were treated with 60 mmol/l KCl to elicit maximal constriction and confirm functional vascular smooth muscle cells (VSMCs). *Significant difference between WKY arterial diameters in control and drug-containing solutions; +significant difference between WKY and SHR under the same conditions (P < 0.05).

Hyperpolarizing influence of K_V1 channels is absent in SHR arteries.

Microelectrode studies revealed a blunted contribution of Cor-sensitive K_V1 channels to the resting E_m of SHR cerebral arteries. In agreement with earlier findings (27, 28, 46), the resting E_m of SHR cerebral arteries was more depolarized than that of WKY arteries, averaging −43 ± 3 (n = 9) and −55 ± 1 (n = 8) mV, respectively (Fig. 4). Adding 1 μmol/l Cor to preferentially block K_V1 channels profoundly reduced the E_m of WKY cerebral arteries from −55 ± 1 mV to −38 ± 1 mV, verifying a strong contribution of these channels to resting E_m (Fig. 4, A and C). In contrast, SHR cerebral arteries failed to depolarize further in response to Cor (Fig. 4, B and C), suggesting that the hyperpolarizing influence of the K_V1 channels was absent.

Fig. 4.

Effect of Cor on membrane potential in WKY and SHR cerebral arteries. Isolated, cannulated middle cerebral arteries were perfused at 80 mmHg. Microelectrode recordings of membrane potential (E_m) were obtained in WKY (A) and SHR (B) arteries under control (Con) conditions and after 1 μmol/l Cor was added to the superfusate. C: average E_m values for WKY and SHR arteries (n = 8, 9 animals). *Significant difference between E_m in WKY arteries in control and Cor-containing solutions; +significant difference between WKY and SHR under the same conditions (P < 0.05).

Cerebral VSMCs of SHR show less K_V1 current.

Whole cell patch-clamp studies revealed a reduction in total K⁺ current and a loss of functional K_V1 channels in the VSMCs of SHR. Smaller amplitudes of total K⁺ current were elicited in SHR than WKY cells (Fig. 5, A and B, left). The peak membrane densities of total current at +58 mV averaged 9.44 ± 1.24 (SHR; n = 10) and 12.88 ± 0.84 (WKY; n = 9) pA/pF in control solution. Subsequently, pharmacological block of K_V1 channels with 1 μmol/l Cor revealed an underlying component of Cor-sensitive K_V1 current (I_Cor) that was diminished in SHR cells (Fig. 5, A and B, right). Digital subtraction of residual current (after Cor) from total current (before Cor) revealed a loss of I_Cor in the VSMCs of SHR (Fig. 5C). Correspondingly, the current-voltage (I-V) relationship for I_Cor was significantly suppressed in the VSMCs of SHR compared with WKY rats (Fig. 5D). For example, peak K_V1 current density was 43 ± 12% less in SHR (n = 10) compared with WKY (n = 9) cells, averaging 5.47 ± 1.03 and 9.58 ± 0.82 pA/pF, respectively. The density of Cor-insensitive K⁺ currents in SHR cells was not statistically different from that in WKY cells.

Fig. 5.

Effect of Cor on whole-cell K⁺ currents in WKY and SHR cerebral VSMCs. Currents were recorded in control solution and after 1 μmol/l Cor. A and B: total and Cor-sensitive K⁺ currents were lower in the VSMCs of SHR than WKY. Cell capacitances were 20.7 (WKY) and 20.2 (SHR) pF. C: Cor-sensitive K⁺ current (I_Cor) corresponding to Shaker-type voltage-gated K⁺ (K_V1) channel current in the same WKY and SHR cells. D: I_Cor density plotted as a function of membrane potential. A reduced current density was evident in VSMCs of SHR (n = 10 cells; n = 7 animals) compared with WKY (n = 9 cells; n = 7 animals). *Significant difference between WKY and SHR at the same voltage (P < 0.05).

Theoretically, a reduced level of K_V1 current may result from a reduced number of channel proteins in the plasma membrane and/or a shift to the right in voltage-dependent channel activation due to alterations such as phosphorylation state (1) or α-subunit composition (31). To determine whether the reduced K_V1 current observed in VSMCs of SHR resulted from a lower activation state of the channel resulting from such alterations, we analyzed the voltage dependence of activation of the Cor-sensitive K_V1 current. Tail currents were recorded in control and Cor-containing solutions in WKY and SHR cerebral VSMCs (Fig. 6A). The Cor-sensitive component of K_V current was analyzed by subtracting tail currents obtained in the presence of Cor from tail currents obtained in drug-free solution, and the voltage sensitivity for activation was evaluated by comparing half-maximal activation potential (V_0.5) values (Fig. 6B) (2). A V_0.5 value of −3.3 ± 2.3 mV (n = 4) was calculated for Cor-sensitive K_V1 current in VSMCs of WKY rats, consistent with the V_0.5 value of −1 mV reported earlier in cerebral VSMCs of Sprague-Dawley rats (Fig. 6B) (2). A similar V_0.5 value of −4.5 ± 2.5 mV (n = 4) was calculated for Cor-sensitive K_V1 current in VSMCs of SHR. These comparable V_0.5 values for Cor-sensitive K_V1 current inferred that the diminished K_V1 current density in the cerebral VSMCs of SHR was not related to a reduced voltage sensitivity of the channel. These numbers are also consistent with the V_0.5 value of −5 mV reported for α_1.2/α_1.5 heterotetramers expressed in mouse L cells, which represents a hybrid voltage reflecting the V_0.5 values of +22 mV and −13 mV for α_1.2 and α_1.5 homotetramers, respectively, in the same preparation (31).

Fig. 6.

Analysis of the voltage dependence of Cor-sensitive K_V1 current in WKY and SHR cerebral VSMCs. Tail currents were evoked in symmetrical K⁺ (145 mmol/l) solutions by hyperpolarizing pulses to −70 mV, which followed 8-mV prepulse steps from −70 mV to +58 mV. A: effect of 1 μmol/l Cor on steady-state activation in VSMCs of WKY rats and SHR. B: voltage sensitivity for activation of the Cor-sensitive K_V1 current was not significantly different between WKY and SHR cerebral VSMCs (n = 4 each). I/I_max, current normalized to maximal current.

Reduced α_1.2 transcript in SHR arteries.

Further studies determined whether the loss of functional K_V1 channels in the VSMCs of SHR correlated with a reduced expression of mRNAs encoding the α_1.2 and α_1.5 subunits that form the channel pore (2). Representative real-time PCR curves comparing amplification products of α_1.2, α_1.5, and α-actin between WKY and SHR arteries are shown in Fig. 7, A, B, and C, respectively. The average C_t for detection of the α_1.2 amplification product was 24.45 ± 0.16 for WKY arteries (n = 5, pooled samples) and 25.51 ± 0.29 for SHR arteries (n = 5), implying that the expression of the α_1.2 transcript averaged 51 ± 5% less in SHR arteries as estimated by the ΔΔC_t method (Fig. 7D). With the same RT products, the average C_t for detection of the α_1.5 amplification product was 26.22 ± 0.26 (n = 5) and 26.20 ± 0.35 (n = 5) for WKY and SHR arteries, respectively, inferring no difference in the abundance of α_1.5 mRNA (Fig. 7D). The average C_t for detection of amplification product of the internal standard α-actin was 21.87 ± 1.06 (n = 5) and 21.89 ± 1.05 (n = 5) for WKY and SHR arteries, respectively, inferring no difference in the abundance of α-actin mRNA.

Fig. 7.

RT-PCR amplification products corresponding to α_1.2- and α_1.5-subunit and α-actin mRNAs from cerebral arteries of WKY rats and SHR. For each comparison, total RNA was pooled from arteries of 2 or 3 rats. Representative RT-PCR curves comparing amplification products of α_1.2 (A), α_1.5 (B), and α-actin (C) between WKY and SHR arteries show reduced abundance of the α_1.2 product in cerebral arteries of SHR. RFU, relative fluorescence unit. D: the average relative abundance of α_1.2 transcript was 51% less in arteries from SHR compared with WKY, as estimated by the ΔΔC_t method (where C_t is threshold cycle). However, the relative abundance of α_1.5 was not significantly different between arteries of WKY and SHR (n = 5 each). *Significant difference between WKY and SHR (P < 0.05).

Reduced expression of α_1.2 and α_1.5 proteins in SHR arteries.

Western blotting was used to determine whether the reduced level of α_1.2 transcript in SHR cerebral arteries was associated with a loss of the α_1.2 protein that coassembles with α_1.5 to form K_V1 channel tetramers. Membrane proteins isolated from cerebral arteries of four WKY rats or SHR were pooled to obtain a single preparation that was subjected to SDS-PAGE and immunodetection. A lower intensity of an 80-kDa immunoreactive band corresponding to the mature, glycosylated form of α_1.2 was observed in SHR compared with WKY membranes (Fig. 8A) (2, 56). The SHR membranes also showed a reduced intensity of the 75-kDa band corresponding to the α_1.5 protein (Fig. 8B), in contrast to our earlier finding of similar α_1.5 mRNA expression between WKY and SHR cerebral arteries (Fig. 8, B and D). Expression of the internal standard, α-actin, was similar between lanes of a single Western blot (Fig. 8, A and B). Densitometric analysis of four Western blots using proteins pooled from different animals indicated that the immunoreactivity corresponding to the α_1.2 and α_1.5 subunits was 40 ± 14% and 64 ± 9% less, respectively, in SHR compared with WKY arteries (Fig. 8, C and D).

Fig. 8.

Western blots comparing the expression of α_1.2 and α_1.5 proteins between WKY and SHR cerebral arteries. A and B: reduced band densities were detected for α_1.2 and α_1.5 in arteries from SHR compared with WKY. In the same blots, the band density for α-actin, the internal standard, was similar between lanes. Each lane was loaded with 15 μg of protein, and duplicates of the pooled samples from 3 rats were used. The aspect ratio for α-actin images was modified to 1:2 (height:width). C and D: average densitometric values from 4 Western blots indicated that the abundances of the α_1.2 and α_1.5 proteins, respectively, were less in SHR compared with WKY cerebral arteries. *Significant difference between WKY and SHR (P < 0.05).

Loss of K_V1 channels in cerebral arteries of aortic-banded rats.

To determine whether the loss of K_V1 channels in cerebral arteries exposed to high blood pressure is an abnormality shared by other rat models of hypertension, we also screened for K_V1 channel abnormalities in Ao-B rats. In particular, we determined whether a reduced expression of at least one K_V1 α transcript also was associated with a loss of K_V1 channel proteins and a reduced dilator influence of K_V1 channels in these animals. The internal diameter of arteries from Sham rats was 132 ± 13 μm after a 1-h equilibration period (n = 6) (Fig. 9B). In contrast, arteries from Ao-B rats showed a 19% smaller diameter of 107 ± 11 μm under the same conditions (n = 6) (Fig. 9B), although this difference was not significant because of the variable resting diameters in the Ao-B rats. Subsequently, arteries were treated with 5-HT (1 μmol/l) and ACh (1 μmol/l) to ensure arterial responsiveness to a physiological vasoconstrictor and vasodilator (Fig. 9A). Treatment with 1 μmol/l Cor reduced the resting diameter of cerebral arteries from Sham rats by 38 ± 5% (Fig. 9B). In contrast, the cerebral arteries from Ao-B rats failed to constrict to Cor, suggesting a loss of K_V1-mediated dilation (Fig. 9B). Next, the expression levels of α_1.2 and α_1.5 mRNA were compared between arteries of Sham and Ao-B rats. Representative real-time PCR curves comparing amplification products of α_1.2, α_1.5, and α-actin between Sham and Ao-B arteries are shown in Fig. 10, A, B, and C, respectively. Using the same ΔΔC_t analysis of RT-PCR reactions shown in Fig. 7, we determined that the expression levels of both K_V1 α transcripts were significantly reduced in the cerebral arteries of Ao-B rats (Fig. 10D). The abundance of α_1.2 mRNA was 34 ± 4% less in arteries of Ao-B rats than Sham rats (n = 5 each). Similarly, the expression level of α_1.5 mRNA was depressed 40 ± 5% in the arteries of Ao-B rats, an abnormality that we did not detect in the cerebral arteries of SHR.

Fig. 9.

Effect of 5-HT, ACh, and Cor on the resting diameter of cerebral arteries from sham-operated (Sham) and aortic-banded (Ao-B) rats. Arteries were perfused at an intraluminal pressure of 80 mmHg. A: average diameters of cerebral arteries from Sham (n = 6) and Ao-B rats (n = 6) in drug-free solution (Con), after addition of 1 μmol/l 5-HT, and after the subsequent addition of 1 μmol/l ACh. *Significant difference between diameters before and after 5-HT treatment; +significant difference between diameters before and after subsequent addition of ACh. B: average diameters of arteries from Sham rats (n = 6) and Ao-B rats (n = 6) before (Con) and after Cor treatment. *Significant difference between Sham diameters in control and Cor-containing solutions.

Fig. 10.

RT-PCR amplification products corresponding to α_1.2, α_1.5, and α-actin mRNAs from cerebral arteries of Sham and Ao-B rats. For each comparison, total RNA was pooled from arteries of 2 or 3 rats. Representative RT-PCR curves comparing amplification products of α_1.2 (A), α_1.5 (B), and α-actin (C) between arteries of Sham and Ao-B rats show reduced abundance of α_1.2 and α_1.5 amplification products in cerebral arteries of Ao-B rats. D: average relative abundances of α_1.2 and α_1.5 transcripts were 34% and 40% less respectively in arteries from Ao-B compared with Sham rats, as estimated by the ΔΔC_t method (n = 5 each). *Significant difference between Sham and Ao-B rats (P < 0.05).

Finally, Western blotting was used to determine whether the reduced levels of α_1.2 and α_1.5 mRNAs in cerebral arteries of Ao-B rats resulted in a loss of the respective proteins. In these studies, the immunodensities of the bands corresponding to the α_1.2 and α_1.5 subunits were less in cerebral arteries of Ao-B rats compared with Sham rats (Figs. 11, A and B). Expression of the internal standard, α-actin, was similar between lanes. In eight Western blots using different arterial samples, the immunoreactive signals corresponding to α_1.2 and α_1.5 were 53 ± 8% and 67 ± 10% less, respectively, in cerebral arteries of Ao-B rats compared with Sham animals (Fig. 11C).

Fig. 11.

Western blots comparing the expression of α_1.2 and α_1.5 proteins between arteries from Sham and Ao-B rats. A and B: a reduced band density was detected for α_1.2 (A) and α_1.5 (B), respectively, in arteries from Ao-B rats compared with Sham rats. In the same blots, the band density for α-actin, the internal standard, was similar between lanes. Each lane was loaded with 15 μg of protein, and duplicates of pooled samples from 3 rats were used. The aspect ratio for α-actin images was modified to 1:2 (height:width). C: average densitometric values from 8 Western blots indicated that the abundances of the α_1.2 and α_1.5 proteins were less in arteries of Ao-B compared with Sham rats. *Significant difference between Sham and Ao-B rats (P < 0.05).

DISCUSSION

The cerebral arteries of SHR are abnormally depolarized and constricted (14, 27, 28, 41, 46), a finding that extends to the small cerebral arteries of Dahl salt-sensitive hypertensive rats (52). However, the identity of the K⁺ channels that mediate the enhanced excitability of cerebral arteries during chronic hypertension remains highly controversial. At least three different K⁺ channel gene products contribute to the resting diameter of small cerebral arteries under normal conditions, including the K_IR channel, BK_Ca channel, and K_V channel families (2–4, 9–11, 14, 32, 35, 37, 38, 40, 53). Various studies have alternatively predicted functional abnormalities of K_IR, BK_Ca, or K_V channels in the cerebral circulation of SHR. For example, an enhanced function of K_IR channels in SHR cerebral arteries was suggested by Nakahata et al. (40). These authors observed accentuated K⁺-induced dilations blocked by BaCl₂ in cerebral intraparenchymal arterioles of SHR compared with WKY rats. Abnormalities of BK_Ca channels also have drawn recent attention. Amberg and Santana (3) reported a reduced dilator influence of BK_Ca channels in isolated, pressurized middle cerebral arteries of SHR, which was attributed to a loss of activating β₁-subunits. In contrast, an increased expression of BKα subunits and enhanced contribution of BK_Ca channels to the resting diameter of SHR cerebral arterioles in situ was described earlier by Liu et al. (38). The basilar arteries of SHR also purportedly rely more on BK_Ca channels for resting quiescence, an observation attributed to a higher density of BK_Ca current in freshly isolated VSMCs (37, 53). Finally, a lower density of whole cell K_V1 current was reported in VSMCs from SHR compared with WKY basilar arteries by Xie et al. (53), and a reduced K_V2 channel function was reported in angiotensin (ANG) II-induced hypertensive rats by Amberg and Santana (4).

Considering the different findings of these reports and new evidence that Shaker-type K_V1 channels are major contributors to the resting diameter of rat cerebral arteries (2, 9), the present study investigated whether a loss of K_V1 channel-mediated dilation is a shared abnormality between different rat models of hypertension. We first focused on the SHR as a genetic model of hypertension that shares many features with human essential hypertension, including an elevated vascular tone and reliance on interactions between complex, polygenic mechanisms for elevation of blood pressure (18, 23–25, 42–44, 48). However, it is recognized that the genetic differences between SHR and WKY rats exceed those related to susceptibility genes for hypertension; this is a limitation of the interstrain comparison (44). For this reason, we further compared K_V1 channel-mediated dilation and expression between cerebral arteries of Ao-B rats and normotensive Sham rats. In contrast to SHR, the Ao-B rat provides a high-ANG II “renal” model of hypertension that shares the same genetic background (Sprague-Dawley) as its normotensive Sham control (43, 44). Thus abnormalities of K_V1 channels in cerebral arteries of Ao-B rats cannot be attributed to differences in genetic background but rather rely on factors related to the development of hypertension.

Relying on shared evidence from the SHR and Ao-B rat models, our results strongly suggest that a downregulation of K_V1 channels primarily accounts for the “hypertensive phenotype” of depolarization and constriction in small cerebral arteries. First, vessel perfusion studies revealed a pronounced contribution of K_V1 channels to the resting E_m and diameter in WKY arteries, consistent with our previous observations (2). However, the contribution of K_V1 channels to the E_m and diameter of SHR vessels was absent or significantly blunted. Second, the VSMCs isolated from SHR cerebral arteries showed a reduced density of voltage-gated, Cor-sensitive K⁺ current. Third, SHR arteries exhibited a reduced expression of α_1.2 mRNA and the expression levels of the pore-forming α_1.2 and α_1.5 proteins also were reduced. Fourth, cerebral arteries of Ao-B rats exhibited nearly the same pattern of K_V1 channel deficits as those of SHR, including loss of K_V1 channel-mediated dilation, reduced expression of α_1.2 mRNA, and loss of α_1.2 and α_1.5 proteins. Additionally, expression of mRNA coding for the α_1.5 subunit was lower in arteries of Ao-B rats compared with Sham animals, suggesting an additional defect in the Ao-B rat model that shows high circulating levels of ANG II. In this regard, ANG II has been reported to directly inhibit K_V channel activity by activation of protein kinase C (PKC)ε and inhibition of protein kinase A (PKA) (29). Additionally, gene expression of α_1.5 is regulated by the Sp1 transcription factor that is sensitive to oxidative stress (12, 22), a feature of hypertension that has been partially attributed to ANG II (45, 47). Although the signaling molecules in VSMCs that link elevated ANG II to K_V1 channel suppression are unclear, our findings suggest that the reduced expression of even a single mRNA coding for either α_1.2 or α_1.5 may limit their optimal coassembly into K_V1 channel heterotetramers, resulting in a sparse expression of functional K_V1 channels on the VSMC surface. Indeed, our patch-clamp studies demonstrated a lower density of K_V1 current in VSMCs of SHR compared with WKY rats. Additionally, the V_0.5 values for activation of K_V1 channels in VSMCs of WKY and SHR were comparable and reflected the hybrid voltage profile of typical K_V α_1.2/α_1.5 heterotetrameric channels (2, 31). Thus the loss of K_V1 channels in SHR appears to occur without a change in α-subunit composition.

Recently, Amberg and Santana (4) reported that Shab-type voltage-gated K⁺ (K_V2) channels also contribute to the resting diameter of rat middle cerebral arteries. In their study, stromatoxin (ScTx)-induced block of the K_V2 channels reduced resting diameter by 13% in arteries of control rats. In contrast, cerebral arteries of ANG II-induced hypertensive rats (AHR) showed only a 6% constriction in response to ScTx-induced constriction and a lower density of ScTx-sensitive K⁺ current, suggesting attenuated K_V2 channel function. Considered with our results presented here in similar isolated cerebral arteries also pressurized at 80 mmHg, it seems that both K_V1 channels and K_V2 contribute to the resting diameter of these small vessels. In our study, Cor- and/or psora-4-induced block of K_V1 channels reduced the diameter of arteries from WKY and Sprague-Dawley (Sham) rats by 26–28%, whereas block of K_V2 channels by ScTx reduced diameter by only 13% as reported by the other authors (4). Collectively, these findings infer that a loss of two different gene families of K_V channels, K_V1 and K_V2, may contribute to the elevated cerebrovascular tone in hypertensive rats. It is difficult to evaluate whether shared mechanisms of hypertension coordinately downregulate the K_V1 and K_V2 channels, since expression levels of K_V2 channel transcript have not been reported for the AHR model to date (4).

Notably, the reduction in K_V1 channel expression in cerebral arteries is unlikely to contribute to hypertension since the cerebral circulation is not a site of blood pressure regulation. However, other reports indicate that the loss of K_V1 channels during hypertension may be an abnormality shared between vascular beds. For example, a diminished K_V current density was observed by Cox et al. (16) in patch-clamped VSMCs from small mesenteric arteries of the SHR. Similarly, Wang et al. (51) reported that K_V channel current was attenuated in a rat model of pulmonary arterial hypertension (PAH) induced by hypoxia. In the latter study, a reduced expression of α_1.2 and α_1.5 mRNA and protein was reported in small pulmonary arteries (51), similar to the findings of this study in the cerebral circulation. A downregulation of K_V1 channel function, mRNA, and protein also has been documented in small pulmonary arteries of humans with primary PAH (54, 55). Thus K_V1 channels may play a more global role in elevating vascular tone during different forms of hypertension than divulged by our findings, and there appears to be a link between the established phase of high blood pressure and the loss of vascular K_V1 channels.

It is important to acknowledge several limitations of our study that may have influenced the findings. First, several conclusions in this study rely on the specificity of the K_V1 channel antagonist correolide (Cor). Whereas the optimal concentration of Cor required to selectively block K_V1 channels in VSMCs is unclear, we chose the concentration of 1 μmol/l in this study on the basis of several findings. The ability of Cor to selectively inhibit K_V1 channels has been repeatedly demonstrated (21, 26, 36). For example, K_V3.2 or K_V4 channels are insensitive to 10 μmol/l Cor in heterologous expression systems because of a structural difference in the S5-S6 linker of the α-subunit compared with K_V1 channels (21, 26). K_V2.1 subunits, on the other hand, showed only ∼20-fold less affinity (K_d ∼200 nmol/l) than K_V1.3 subunits (K_d ∼11 nmol/l) for the tritium-labeled derivative of Cor (21). However, the estimated IC₅₀ of COR for α_1.2 and α_1.5 from ⁸⁶Rb⁺ efflux studies in human embryonic kidney cells is reported to be 700 and 1,150 nmol/l, respectively, but these values were obtained after drug incubation periods of only 10 min, whereas Cor binding has been shown to be time and use dependent (21, 26). Thus the steady-state potency of Cor may be higher than reported by this study. In this regard, Albarwani et al. (2) showed that adding the nonspecific K_V channel blocker 4-aminopyridine (1 mmol/l) did not further reduce the diameter of rat cerebral arteries preincubated for 30 min with 1 μmol/l Cor, suggesting maximal block of vascular K_V1 channels by this concentration of Cor. For these reasons, 1 μmol/l Cor was chosen in this study as the optimal concentration to block K_V1 channels without significant block of other K_V channel types. A similar concentration has been used by other authors to selectively block K_V1 channels in different vascular beds (2, 5, 9, 10).

Additionally, results from experiments using psora-4, a different K_V1 channel blocker, suggest that 1 μmol/l Cor maximally and preferentially blocks vascular K_V1 channels. Psora-4 is a 5-phenylalkoxypsoralen reported to selectively inhibit K_V1 channels with low nanomolar affinity (49). Although its blocking effect on K_V2 channels has not been studied to date, psora-4 inhibits K_V1 α_1.2 and α_1.5 homotetrameric channels with IC₅₀ values of 49 and 8 nmol/l, respectively, but fails to block other channels including K_V3 and BK_Ca channels and neuronal sodium channels (49). Hence, we regarded 30 and 100 nmol/l psora-4 as the most likely concentrations to preferentially inhibit the α_1.2/α_1.5 heterotetrameric channels in cerebral arteries. The higher concentration of psora-4 reduced the resting diameter of WKY arteries by 26 ± 3%, a response almost identical to the constriction induced by 1 μmol/l Cor. Furthermore, addition of 1 μmol/l Cor to WKY arteries pretreated with 100 nmol/l psora-4 did not further constrict the arteries, inferring a common blocking action of the two drugs.

A second limitation of our study is that RNA and protein isolates were derived from entire cerebral arteries and not limited to the smooth muscle cell layer. As a result, and although K_V1 channels are densely expressed in rat cerebral VSMCs (2), a minimal contribution of other cell types to α_1.2 and α_1.5 mRNA and protein could not be ruled out. Since our immunoreactive bands on Western blots corresponding to α_1.2 and α_1.5 in membrane preparations were derived from whole arterial lysates rather than purified plasma membranes, it was critical to use patch-clamp methods to verify that the loss of these subunits corresponded to a reduced density of K_V1 current, a correlation that we demonstrated in the cerebral VSMCs of SHR. It should also be acknowledged that mRNA and protein analyses were performed with the entire circle of Willis and its branches, whereas the patch-clamp and vessel perfusion studies used only the middle cerebral arteries. However, Albarwani et al. (2), who first detected α_1.2/α_1.5 heterotetramers in the rat circle of Willis and its branches by coimmunoprecipitation, also found evidence of α_1.2/α_1.5 heterotetramers in the VSMCs of middle cerebral arteries by analyzing activation voltages of Cor-sensitive currents in patch-clamp studies. Thus it appears that the expression of α_1.2/α_1.5 channels is not limited to a single segment of the rat cerebral circulation.

Finally, it seems incongruent that our vessel perfusion studies failed to detect any contribution of K_V1 channels to the resting diameter of SHR cerebral arteries, whereas K⁺ current attributed to K_V1 channels persisted in patch-clamped VSMCs of SHR, albeit it was significantly reduced. These apparently discrepant findings could be explained if the functional K_V1 channels still present in the VSMCs of SHR cerebral arteries in situ are inactivated by cytoplasmic factors, which are diluted during dialysis of the patch-clamped cell with pipette solution to unmask a residual population of functional K_V1 channels. Notably, the signaling molecules that inactivate K_V1 channels in VSMCs include intracellular calcium and PKC, and these molecules are purported to be elevated during the myogenic response to high blood pressure (1, 13, 15). Overall, the functional role of K_V1 channels is more easily resolved by evaluating their contribution to resting membrane potential and diameter in vessel perfusion studies, which reflects the collective impact of channel opening, rather than by K_V1 currents in patch-clamp recordings that are difficult to quantify at physiological E_m.

In summary, our results suggest a previously uncharacterized role for the K_V1 channels as mediators of the compensatory vasoconstriction in cerebral arteries of genetic and renal hypertensive rats. The cellular mechanisms that appear to restrict K_V1 channel activity in the VSMCs of arteries exposed to high blood pressure may include the transcriptional loss of at least one pore-forming α-subunit involved in the biogenesis of the channels.

GRANTS

This work was supported by grants from the American Heart Association [Predoc 0410009Z (A. A. Tobin) and Predoc 0715570Z (B. K. Joseph)], the University of Arkansas for Medical Sciences (Medical Research Endowment Fund, S. W. Rhee), the National Heart, Lung, and Blood Institute (R01-HL-59238, N. J. Rusch), and a research grant from the Sultanate of Oman (S. Albarwani).