Perivascular adipose tissue-derived adiponectin activates BK_Ca channels to induce anticontractile responses

Abstract

This study aims to identify the potential mechanisms by which perivascular adipose tissue (PVAT) reduces tone in small arteries. Small mesenteric arteries from wild-type and large-conductance Ca²⁺-activated K⁺ (BK_Ca) channel knockout mice were mounted on a wire myograph in the presence and absence of PVAT, and contractile responses to norepinephrine were assessed. Electrophysiology studies were performed in isolated vessels to measure changes in membrane potential produced by adiponectin. Contractile responses from wild-type mouse small arteries were significantly reduced in the presence of PVAT. This was not observed in the presence of a BK_Ca channel inhibitor or with nitric oxide synthase (NOS) inhibition or in BK_Ca or adiponectin knockout mice. Solution transfer experiments demonstrated the presence of an anticontractile factor released from PVAT. Adiponectin-induced vasorelaxation and hyperpolarization in wild-type arteries were not evident in the absence of or after inhibition of BK_Ca channels. PVAT from BK_Ca or adiponectin knockout mice failed to elicit an anticontractile response in wild-type arteries. PVAT releases adiponectin, which is an anticontractile factor. Its effect on vascular tone is mediated by activation of BK_Ca channels on vascular smooth muscle cells and adipocytes and by endothelial mechanisms.

the release of adipokines from perivascular fat [perivascular adipose tissue (PVAT)] has been linked to reduced tone in a variety of vascular beds from a number of species including humans (11–13, 15). The possible health implications of PVAT are highlighted by the fact that such anticontractile activity is lost in obesity and disease states such as diabetes and hypertension (15).

A major factor secreted by adipocytes is adiponectin, a 30-kDa protein belonging to the collagen superfamily. Serum concentrations of adiponectin are lower in hypertension, obesity, and diabetes, and high blood pressure develops in high-fat diet-fed mice with adiponectin gene deletion (9, 21, 23, 41, 43).

The mechanisms by which PVAT produces its anticontractile effect have yet to be defined but could include activation of the vascular smooth muscle large-conductance Ca²⁺-activated K⁺ (BK_Ca) channel. This channel is known to be involved in the development of hypertension (44), downregulated in diabetes (35), and impaired in metabolic syndrome (2) and insulin resistance (6) and therefore may be involved in mediating PVAT anticontractility. It is a transmembrane channel composed of four α- and four β-subunits. The α-subunits form the pore, and the β-subunits have a regulatory role. The channel is activated by Ca²⁺ sparks and is located close to the sarcoplasmic reticulum. Therefore it is exposed to relatively high concentrations of Ca²⁺ ions, which bind to the α-subunit. This process, facilitated by the β-subunit, activates the channel (3, 45). In turn, this leads to K⁺ efflux and membrane hyperpolarization. Consequently, voltage-gated Ca²⁺ channels are closed, intracellular Ca²⁺ concentration is reduced, and vasodilation ensues.

The aim of the present investigation was to identify adipokines responsible for PVAT-induced anticontractility and to determine possible mechanisms of anticontractile transduction in a murine model. Using wire myography and electrophysiology to study mouse knockout models for both the α- and β-subunits of the BK_Ca channel, we provide evidence that suggests that PVAT-derived adiponectin is responsible for reducing vascular tone and acts by opening myocyte BK_Ca channels, thereby identifying a critical role for this channel in this important physiological process. We also demonstrate that adipocyte BK_Ca channels are involved in the response.

METHODS

Animal Models

Procedures were performed in accord with Institutional Guidelines and the United Kingdom Animals (Scientific) Procedures Act of 1986, and the work was conducted under licenses 40/03409 and S1 40/03409, issued by the British Home Office. Kcnma1^−/− mice lacking the pore-forming slo1 or α-subunit of the BK_Ca channel (and wild-type Kcnma1^+/+ counterparts expressing the α-subunit) and BKβ1^−/− mice lacking the regulatory β₁-subunit of the BK_Ca channel were bred in house at the University of Manchester. Targeted disruption of the Slo gene and the BKβ1 gene has been described previously (3, 36). Genotyping confirmed successful generation of Kcnma1 mice. B6.129-Adipoq^tm1Chan/J adiponectin-deficient (adipo^−/−) mice were purchased from The Jackson Laboratory and used at 8–10 wk of age. C57BL/6J mice were purchased from Harlan UK and used as wild-type/control animals for BKβ1^−/− mice and adipo^−/− mice. Male and female Kcnma1^+/+, Kcnma1^−/−, BKβ1^−/−, and C57BL/6J mice (12–18 wk old, ∼25 g body wt) and adipo^−/− mice were killed by stunning and cervical dislocation.

Wire Myography

The mesenteric beds from Kcnma1^−/−, Kcnma1^+/+ counterpart, BKβ1^−/−, C57BL/6J, and adipo^−/− mice were dissected, and small arteries were isolated. One vascular segment was cleaned of PVAT, while an adjacent section was left with PVAT intact. Both arteries were mounted on 40-μm tungsten wires in a myograph (Danish MyoTech, Aarhus, Denmark) containing physiological saline solution (PSS, in mmol/l: 119 NaCl, 4.7 KCl, 25 NaHCO₃, 1.17 KH₂PO₄, 1.17 MgSO₄, 0.026 EDTA, 1.6 CaCl₂, and 5.5 glucose). Solutions were equilibrated with 95% air-5% CO₂ gas. After an initial 30-min equilibration, vessel wall tension and diameter were normalized in a standardized procedure and stabilized for 1 h. Wire myography was used to assess the norepinephrine (1 × 10⁻⁹–3 × 10⁻⁵ mol/l)-induced contractile responses of arteries with and without PVAT under a number of conditions: control, in the presence and absence of endothelium, N^G-nitro-l-arginine (l-NNA, 10⁻⁵ mol/l; Sigma Aldrich), paxilline (10⁻⁶ mol/l; Sigma Aldrich), indomethacin (10⁻⁶ mol/l; Sigma Aldrich), and adiponectin blocking peptide for PAb to adiponectin receptor 1 (AdipoR1) (5 × 10⁻³ g/ml; Enzo Life Sciences) (15). Modulators were preincubated for 30 min. Responses were normalized to an initial constriction to 60 mmol/l high-potassium solution (KPSS). The response to exogenously applied globular adiponectin (3 × 10⁻³ g/ml; Enzo Life Sciences) was also assessed in arteries preconstricted with 3 × 10⁻⁵ mol/l norepinephrine.

Solution Transfer Experiments

Arteries with and without PVAT were preconstricted with norepinephrine (3 × 10⁻⁵ mol/l). This maximal concentration was used to produce a uniform constriction for each artery and to avoid alteration of norepinephrine concentration in each bath upon transfer. Once a stable constriction had developed, the organ bath solution was taken from a donor wild-type (Kcnma1^+/+ or C57BL/6J) small artery with PVAT and used to replace the solution from a recipient artery without PVAT.

PVAT Transfer Experiments

Mesenteric arteries were dissected as described above. PVAT was dissected from the mesenteric bed and stored in ice-cold PSS. Care was taken to ensure that no residual blood vessels remained in the PVAT sample. After the initial constriction to KPSS and cumulative concentration-response curves (CCRC) to norepinephrine, PVAT from a wild-type (C57BL/6J) mouse was positioned in close proximity to a wild-type or BK_Ca knockout (BKβ1^−/−) artery lacking PVAT with a tungsten wire. A CCRC to norepinephrine was performed. After washout, PVAT from a knockout (BKβ1^−/−) artery was then placed close to a wild-type (C57BL/6J) or knockout (BKβ1^−/−) artery lacking PVAT and the CCRC was repeated. In a separate series of experiments, PVAT from an adipo^−/− artery was placed beside a wild-type (C57BL/6J) artery lacking PVAT and a CCRC to norepinephrine was performed.

Electrophysiology

Small segments of mesenteric artery (1st or 2nd order) without PVAT but with intact endothelium, from Kcnma1^+/+ and Kcnma1^−/− mice, were pinned to the Sylgard base of a thermostatically controlled bath and superfused (10 ml/min) with PSS (and bubbled with 95% air-5% CO₂). The myocytes were impaled via the adventitial surface with microelectrodes filled with 3 mol/l KCl (resistance 40–80 MΩ) as described previously (7) and in Table 1. Vessels were exposed to acetylcholine (10⁻⁵ mol/l), NS-1619 (3 × 10⁻⁵ mol/l), iberiotoxin (10⁻⁵ mol/l) and levcromakalim (10⁻⁵ mol/l) (see Fig. 5). Because of the high cost of globular adiponectin, flow was stopped and the peptide was added directly to the recording chamber to give a final concentration of 3 μg/ml.

Table 1.

EC₅₀ and log EC₅₀ data calculated from concentration-response curves to norepinephrine where values were normalized to maximum response from within each curve

	No PVAT		PVAT
Group	EC50	Log EC50	EC50	Log EC50	No PVAT n	PVAT n
Kcnma1+/+ control	3.3 × 10−6	−6.0 ± 0.1	2.3 × 10−5	−5.5 ± 0.1*	11	11
C57BL/6J	1.1 × 10−6	−6.1 ± 0.1	3.2 × 10−5	−6.0 ± 0.1	10	10
C57BL/6J + blocking peptide	1.1 × 10−6	−6.1 ± 0.1	9.4 × 10−7	−6.1 ± 0.1	5	8
adipo−/−	7.1 × 10−6	−6.3 ± 0.2	7.2 × 10−6	−5.6 ± 0.8	12	12
Kcnma1+/+ no endothelium	5.3 × 10−7	−6.5 ± 0.2	4.1 × 10−5	−5.7 ± 0.5	7	6
C57BL/6J no endothelium	1.1 × 10−6	−6.1 ± 0.1	1.7 × 10−6	−5.8 ± 0.1	6	5
Kcnma1+/+ l-NNA	4.2 × 10−7	−6.5 ± 0.1	8.0 × 10−7	−5.7 ± 0.5	6	6
C57BL/6J l-NNA	1.2 × 10−6	−6.0 ± 0.1	1.5 × 10−6	−5.9 ± 0.1	5	6
Kcnma1+/+ indomethacin	2.4 × 10−6	−5.0 ± 0.7	4.8 × 10−6	−5.6 ± 0.5	6	6
Kcnma1+/+ l-NNA + indomethacin	4.4 × 10−6	−5.3 ± 0.8	1.0 × 10−6	−6.1 ± 0.2	6	6
Kcnma1−/− control	9.7 × 10−6	−6.1 ± 0.1	4.8 × 10−6	−5.8 ± 0.2	10	10
BKβ1−/− control	1.7 × 10−4	−5.2 ± 0.1	4.5 × 10−5	−5.5 ± 0.3	16	13
Kcnma1+/+ paxilline	1.0 × 10−6	−6.2 ± 0.2	1.4 × 10−6	−6.0 ± 0.2	5	5
C57BL/6J paxilline	6.5 × 10−7	−6.3 ± 0.1	3.0 × 10−6	−5.9 ± 0.1	11	11
PVAT transfer experiments
C57 control			1.4 × 10−6	−6.0 ± 0.2	10	10
+C57 transferred PVAT	9.5 × 10−7	−6.2 ± 0.1			10
+BKβ1−/− PVAT	5.3 × 10−7†	−7.7 ± 0.7†			10
BKβ1−/− + C57 transferred PVAT	2.3 × 10−6	−6.0 ± 0.2			10
adipo−/− + C57 transferred PVAT	2.1 × 10−6	−5.7 ± 0.1			6

PVAT, perivascular adipose tissue; l-NNA, N^G-nitro-l-arginine.

^*P < 0.05, t-test, PVAT vs. no PVAT;

^†P < 0.05, t-test, vs. PVAT transfer experiments C57 control.

Fig. 5.

Typical traces showing sharp electrode recordings from isolated mesenteric artery without PVAT from Kcnma1^+/+ (A) and Kcnma1^−/− (B) mice. The presence of BK_Ca channels was confirmed by hyperpolarization in response to NS-1619 (BK_Ca channel opener). Acetylcholine (ACh) confirms endothelium presence. Levcromakalim (LK) hyperpolarizes via ATP-sensitive K⁺ (K_ATP) channels. Iberiotoxin (IbTx) reverses adiponectin-induced hyperpolarization. m.p., Membrane potential.

Statistical Analyses

Data are expressed as means ± SE. For cumulative response curves, differences between dose-tension relations were analyzed by repeated-measures ANOVA and Bonferroni post hoc tests were used to calculate significance at different concentrations. EC₅₀ and log EC₅₀ analysis was calculated as follows. The maximum response to KPSS was calculated as the difference between baseline and the maximum tension elicited by exposure to KPSS. Constrictions to norepinephrine were normalized to the maximum response within each curve, and the EC₅₀ and log EC₅₀ were calculated from this curve. These data were analyzed with Student's t-test and are presented in Table 1. Responses to exogenous adiponectin and electrophysiology data were analyzed with Student's t-test. A probability value of P < 0.05 was considered statistically significant. Analyses were performed with GraphPad Prism version 3.00 for Windows (GraphPad Software).

RESULTS

Wire Myography: PVAT Has an Anticontractile Effect in Murine Small Arteries

The response to a KPSS constriction performed at the beginning of the protocol was used to normalize the data presented. The maximum response to 60 mmol/l KPSS (ΔmN/mm ± SE) produced by arteries from Kcnma1^+/+ mice (n = 57) with PVAT (0.31 ± 0.03) was not significantly different from that produced by arteries from Kcnma1^+/+ mice (n = 57) without PVAT (0.38 ± 0.05). Responses to KPSS produced by arteries from Kcnma1^−/− mice (n = 36) with PVAT (0.37 ± 0.1) were not significantly different from those from arteries (n = 37) lacking PVAT (0.42 ± 0.04). Comparison of responses to KPSS from arteries from BKβ1^−/− mice (n = 34) surrounded by PVAT (0.4 ± 0.07) revealed no significant differences from their PVAT-lacking (n = 34) counterparts (0.34 ± 0.05). Similarly, no differences were detected between arteries from C57BL/6J mice (n = 56) with PVAT (0.33 ± 0.04) and those without PVAT (0.4 ± 0.04) in response to KPSS.

Cumulative concentrations of norepinephrine produced constriction in all groups. Wild-type [Kcnma1^+/+ (n = 11) and C57BL/6J (n = 10)] vessels with PVAT constricted significantly less than those without PVAT (Fig. 1, A and B, respectively). A significant shift (P < 0.05, test) was detected in log EC₅₀ values calculated in Kcnma1^+/+ arteries (−6.0 ± 0.1 no PVAT vs. −5.5 ± 0.4 PVAT). There was no difference in EC₅₀ values from C57BL/6J arteries with or without PVAT (Table 1).

Fig. 1.

A and B: contractile responses of mesenteric Kcnma1^+/+ (A) and C57BL/6J (B) arteries to cumulative concentrations of norepinephrine (NE). PVAT, perivascular adipose tissue; KPSS, high-potassium physiological saline solution. C: solution transfer experiments. D: C57BL/6J arteries in the presence of adiponectin blocking peptide. E: response to exogenous adiponectin in the presence (+E) and absence (−E) of endothelium. F: adipo^−/− arteries. *P < 0.05 ANOVA.

Organ Bath Solution Transfer Studies Demonstrate That Adipocytes Release a Vasorelaxing Factor

To ensure that a similar degree of constriction was achieved in each artery (maximum) and that the concentration of norepinephrine remained constant, arteries with and without PVAT were preconstricted with norepinephrine (3 × 10⁻⁵ mol/l). Once a stable constriction had developed, the organ bath solution was taken from a donor C57BL/6J small artery with PVAT and used to replace the solution from a recipient artery without PVAT. Control experiments transferring solution to and from wild-type arteries without PVAT (n = 4) were performed to determine whether the transfer maneuver would alter tension. Such a transfer did not significantly alter tension (from 5.5 ± 0.1 to 5.5 ± 0.1 mN/mm). Solutions from wild-type vessels (C57BL/6J) with PVAT (n = 6) caused a significant reduction in tension (P < 0.05, t-test) when transferred onto wild-type arteries (from C57BL/6J mice) without PVAT (Fig. 1C). The presence of adiponectin blocking peptide for AdipoR1 (n = 3) inhibited this effect (Fig. 1C). AdipoR1 peptide blockade (n = 8) also inhibited the PVAT-induced anticontractile response during norepinephrine concentration-response curves in C57BL/6J arteries (Fig. 1D).

Application of Exogenous Adiponectin Induces Vascular Relaxation

Applying globular adiponectin to C57BL/6J arteries preconstricted with norepinephrine (3 × 10⁻⁵ mol/l) caused dilations in both the presence and the absence of PVAT and in the presence and absence of endothelium that were significantly different from zero (P < 0.05, paired t-test) (Fig. 1E). Interestingly, arteries from Kcnma1^+/+ mice without PVAT (n = 4) also dilated significantly (33.4 ± 10.6%) to adiponectin, but there was no response from arteries from Kcnma1^+/+ mice with PVAT.

Anticontractile Effect Is Absent in Adiponectin Knockout Mice

The contractile responses to norepinephrine were assessed in arteries from mice lacking adiponectin. There were no significant differences in responses in the presence or absence of PVAT in arteries from adipo^−/− animals (n = 6) (Fig. 1F).

Endothelial Contribution to PVAT Activity

Endothelium removal abolished the anticontractile effects of PVAT response in wild-type (Kcnma1^+/+ and C57BL/6J) arteries (Fig. 2, A and B). The anticontractile effect was absent in the presence of the nitric oxide synthase (NOS) inhibitor l-NNA in intact wild-type (Kcnma1^+/+ and C57BL/6J) arteries (Fig. 2, C and D). Inhibition of prostacyclins with indomethacin also inhibited the anticontractile response in Kcnma1^+/+ arteries (Fig. 2E). Combination of l-NNA and indomethacin confirmed these data (Fig. 2F). The presence of indomethacin and l-NNA abolished the anticontractile effect but also significantly increased tension (P < 0.05, ANOVA) compared with arteries lacking PVAT. There were no significant differences in EC₅₀ values between any of the groups (Table 1).

Fig. 2.

A and B: contractile responses of mesenteric Kcnma1^+/+ (A) and C57BL/6J (B) arteries to cumulative concentrations of norepinephrine in the absence of endothelium. C and D: contractile responses of Kcnma1^+/+ (C) and C57BL/6J (D) arteries + N^G-nitro-l-arginine (l-NNA). E and F: responses of Kcnma1^+/+ arteries + indomethacin (E) and + indomethacin and l-NNA (F). *P < 0.05 ANOVA.

Vascular BK_Ca Activity Is Crucial to PVAT Relaxation

Deletion of the α (Kcnma1^−/− mice)- or β (BKβ1^−/− mice)-subunit of the BK_Ca channel resulted in a reduction of the anticontractile response (Fig. 3, A and B). The result remained the same irrespective of whether the endothelium was present or not (data not shown).

Fig. 3.

Role of the large-conductance Ca²⁺-activated K⁺ (BK_Ca) channel in mediating the PVAT response. A–D: constrictions produced by Kcnma1^−/− (A) and BKβ1^−/− (B) arteries to norepinephrine and Kcnma1^+/+ (C) and C57BL/6J (D) arteries in the presence of paxilline. E: solution transfer experiments showing the tension produced in Kcnma1^−/− and BKβ1^−/− arteries constricted with norepinephrine before and after transfer.

Pharmacological inhibition of the BK_Ca channel with paxilline also abolished the response in both Kcnma1^+/+ and C57BL/6J mice (Fig. 3, C and D). There was no significant difference in tension developed regardless of the presence of PVAT. There were no significant differences in EC₅₀ values between any of the groups (Table 1).

Anticontractile Factor Released from PVAT Requires BK_Ca Channels to Elicit Effects

Transfer of solution from Kcnma1^+/+ (n = 5) and C57BL/6J (n = 6) arteries with PVAT to arteries lacking PVAT from Kcnma1^−/− and BKβ1^−/− mice failed to elicit a change in tension (Fig. 3E).

BKβ1^−/− PVAT Does Not Elicit an Anticontractile Effect

In this series of experiments PVAT was taken from a mouse and transferred to arteries lacking PVAT to investigate whether the type of PVAT (wild type or a knockout variety) itself can influence the vascular response. The data are labeled as follows: an artery lacking PVAT in the presence of transferred wild-type or knockout PVAT is labeled “no PVAT (WT)” or “no PVAT (KO),” respectively.

PVAT from a C57BL/6J mouse (n = 10) was placed next to a C57BL/6J artery lacking PVAT [Fig. 4A, no PVAT (WT)]. This significantly reduced tension (P < 0.05, ANOVA) compared with response elicited before PVAT transfer (Fig. 4A, no PVAT). The response in the presence of transferred PVAT was not significantly different from that of a C57BL/6J artery surrounded by its own PVAT (Fig. 4A, PVAT).

Fig. 4.

A: effects of transferred C57BL/6J PVAT to arteries lacking PVAT [no PVAT (WT)] (*P < 0.05 ANOVA vs. no PVAT). B: tension elicited by C57BL/6J arteries lacking PVAT in the presence of transferred BKβ1^−/− PVAT [no PVAT (KO)] (*P < 0.05 ANOVA vs. PVAT). C: effects of transferred PVAT from C57BL/6J mice to BKβ1^−/− arteries lacking PVAT [no PVAT (WT)]. D: constrictions elicited by C57BL/6J arteries lacking PVAT in the presence of transferred PVAT from adipo^−/− mice [no PVAT (Adipo KO)] (*P < 0.05 ANOVA vs. PVAT).

When PVAT from a BKβ1^−/− mouse (n = 10) was incubated with a C57BL/6J artery lacking PVAT [Fig. 4B, no PVAT (KO)], the tension elicited was significantly greater (P < 0.05, ANOVA) compared with the response of a C57BL/6J artery with PVAT (Fig. 4B, PVAT). EC₅₀ and log EC₅₀ were significantly shifted under these conditions (Table 1).

The presence of C57BL/6J PVAT (n = 10) next to a BKβ1^−/− artery lacking PVAT [Fig. 4C, no PVAT (WT)] did not produce a significantly different response compared with a BKβ1^−/− artery surrounded by PVAT (Fig. 4C, PVAT).

Finally, incubation of PVAT (n = 6) taken from an adipo^−/− mouse failed to induce anticontractility in a C57BL/6J artery lacking PVAT [Fig. 4D, no PVAT (Adipo KO)]. The constrictions under these conditions were significantly greater (P < 0.05, ANOVA) compared with a C57BL/6J artery surrounded by native PVAT (Fig. 4D, PVAT).

Adiponectin Has No Effects in the Absence of the BK_Ca Channel

Myography.

Addition of exogenous globular adiponectin to preconstricted wild-type arteries (Kcnma1^+/+ and C57BL/6J) resulted in a vasodilation (Fig. 1D). Application of adiponectin to arteries with or without PVAT from knockout mice (both Kcnma1^−/− and BKβ1^−/−) did not elicit a dilation [Kcnma1^−/− − PVAT = 8.4 ± 7.4% (n = 4); Kcnma1^−/− + PVAT = −7.5 ± 5.4% (n = 4); BKβ1^−/− − PVAT = −0.8 ± 1.6% (n = 5); BKβ1^−/− + PVAT = −7.1 ± 2.9%; (n = 3)].

Electrophysiology.

The resting membrane potential of smooth muscle cells in Kcnma1^+/+ arteries was −53.6 ± 0.4 mV (n = 6). This was not significantly different from that of Kcnma1^−/− arteries (−51.4 ± 0.8 mV) (n = 7). The presence of a functional endothelium was confirmed by a hyperpolarization in response to acetylcholine. There was no significant difference in change in membrane potential between Kcnma1^+/+ (13.9 ± 1.0 mV) and Kcnma1^−/− knockout (16.4 ± 1.4 mV) arteries. Levcromakalim confirmed the presence of ATP-sensitive K⁺ (K_ATP) channels in the arteries. There were no differences in the magnitude of the levcromakalim-induced hyperpolarization between Kcnma1^+/+ (−23.6 ± 1.7 mV) and Kcnma1^−/− (−26.2 ± 1.8 mV) arteries. Furthermore, a hyperpolarization to levcromakalim occurred in Kcnma1^−/− arteries, which were unresponsive to adiponectin (Fig. 5B). There were significant differences (P < 0.01, t-test) in the responses of Kcnma1^+/+ arteries (−15.2 ± 0.7 mV) and Kcnma1^−/− arteries (0.0 ± 0.13 mV) to the BK_Ca channel opener NS-1619 (3.3 × 10⁻⁵ mol/l), confirming the absence of the channel in the Kcnma1^−/− mouse. Adiponectin produced a change in membrane potential of −10.5 ± 0.7 mV in Kcnma1^+/+ arteries, which was attenuated (−0.1 ± 0.2 mV) by the BK_Ca channel inhibitor iberiotoxin. The failure of adiponectin to generate myocyte hyperpolarization (−0.4 ± 0.3 mV) in Kcnma1^−/− vessels indicated that this channel is essential in mediating the electrical effects of this adipokine (Fig. 5B). The lack of hyperpolarization in responses to adiponectin in the absence of the BK_Ca channel indicates that this channel is essential in mediating the effect.

DISCUSSION

These studies have been carried out in small arteries from both wild-type mice (Kcnma1^+/+ and C57BL/6J) and strains without functional BK_Ca channels (Kcnma1^−/− and BKβ1^−/−). The results demonstrate that the white PVAT that surrounds such vessels releases a substance that can reduce tension in preconstricted arteries and that this appears to be mediated by the activation of BK_Ca channels. In mice without BK_Ca channels, the relaxant response is significantly reduced when either the functional pore or the regulatory domain is absent. Our hypothesis was further tested pharmacologically in arteries from wild-type mice, in which the effect was abolished with the BK_Ca channel blocker paxilline. Solution transfer studies confirm that the relaxing factor is released from PVAT. Contractility studies using arteries from adiponectin knockout animals and adiponectin blocking peptide protocols here and in our previously published work (15) indicate that these data are consistent with the involvement of adiponectin. Therefore exogenous globular adiponectin was applied to arteries from wild-type mice and shown to cause relaxation. Average circulating levels of total adiponectin in the human circulation of up to 30 μg/ml have been reported (1, 50), while murine levels are lower, >10 μg/ml (38). Similarly, the application of globular adiponectin caused hyperpolarization that could be abolished with the BK_Ca channel blocker iberiotoxin, again indicating activation of BK_Ca, and this effect was not observed in vessels without BK_Ca. PVAT-induced relaxation also involves endothelial release of nitric oxide (NO) consistent with reported adiponectin effects (4, 5, 47); this appears to be dependent on BK_Ca activation because no such effect was seen in arterial tissue from BK_Ca knockout mice. PVAT from BK_Ca knockout mice was unable to produce an anticontractile effect in wild-type arteries, suggesting that adipocyte BK_Ca channels are involved in adiponectin signaling.

Our results support the ubiquitous nature of PVAT-induced anticontractility. The demonstration of PVAT-induced anticontractility in murine small arteries adds to the abundance of evidence in other vascular beds including human subcutaneous vessels, internal mammary artery, rat mesenteric, and rat aorta (14, 15). The possible clinical relevance of the effect is highlighted by its absence or reduced efficacy in cardiovascular diseases (15): hypertension is associated with a reduced PVAT response in rats (10), and recently Lu and coworkers (32) reported a functional impairment of the anticontractile response in aortic rings from spontaneously hypertensive rats, which they attributed to angiotensin release. Furthermore, anticontractility is impaired in obesity in both animal and human models (15, 25), and data from diabetic models suggest that anticontractility is also decreased (28).

The most important finding in our study is the pivotal role the BK_Ca channel appears to play in mediating PVAT-induced anticontractility. Kcnma1^−/− mice lacking a functional pore or BKβ1^−/− mice lacking the regulatory β-subunit demonstrate a significant reduction in anticontractility, either during concentration-response curves to norepinephrine or via solution transfer studies. It has been reported that the BK_Ca channel retains some functionality in the absence of the regulatory β₁-subunit (3), and because the PVAT effect is reduced in β₁-subunit knockout arteries, this would imply that a Ca²⁺ signal is important in the signaling pathway from adipocyte to myocyte BK_Ca channel.

The use of an organ bath solution transfer protocol demonstrates that a factor is released from PVAT that influences both endothelial and smooth muscle function. The solution from a preconstricted wild-type artery surrounded with PVAT causes significant relaxation of preconstricted vessels devoid of PVAT. Several studies have used this technique to identify the presence of a transferable adipose-derived relaxing factor (ADRF) (15, 28). A recent review has compared the blossoming field of adipocyte research to that of the endothelial field in the 1980s (46). This may well be the case, the adipocyte being a complex signaling cell with interconnected molecular pathways, each being activated by different stimuli. Such a situation could account for the diverse candidates being proposed as PVAT-derived relaxing factor.

Current literature points to a number of possible candidates for the PVAT-derived relaxing factor, including Ang-(1–7) (27), H₂S₂ (42), and palmitic acid methyl ester (29). The target channels of PVAT modulators are also the subject of much discussion and debate. Potassium channels such as delayed-rectifier (K_V) channels (49), K_ATP channels (31), as well as small- and intermediate-conductance Ca²⁺-sensitive potassium channels (12) appear to play a central role in various PVAT pathways.

Previously, we successfully inhibited the PVAT anticontractile effect in human gluteal arteries by using a peptide blocking the effects of adiponectin, thereby suggesting that adiponectin is responsible (15), and achieved similar results in our preparation. Furthermore, contractile experiments performed on adiponectin-deficient mice demonstrated a lack of PVAT anticontractile effect. These findings contrast with those of an earlier study (9) in which an anticontractile effect of PVAT was still observed in mesenteric arteries from adiponectin-deficient mice. However, the superior mesenteric artery and not smaller mesenteric arteries were used in the aforementioned study and arteries were stimulated with serotonin, which suggests that both vessel caliber and stimulators are important when considering experimental design. PVAT from adiponectin-deficient mice failed to reduce tone in wild-type arteries lacking PVAT, suggesting that PVAT-derived adiponectin is required to elicit an anticontractile response. We confirmed the ability of our arteries to respond to adiponectin by observing a relaxation in response to exogenous globular adiponectin. We used globular adiponectin, which is known to act via AdipoR1 and has a low affinity for AdipoR2 (23, 55). In our study there appeared to be a greater effect of adiponectin in the absence of the endothelium, which could potentially be attributed to adiponectin having direct access to the smooth muscle layer (8). However, in vivo one would expect adiponectin to encounter the adventitia first and then the smooth muscle layer. Our results would therefore suggest that a vasocrine signaling system where factors are first released into the bloodstream is at work. Vasocrine signaling mechanisms have been previously reported where factors released from PVAT influence vascular behavior (57). Additionally, we see no effect of adiponectin in arteries from Kcnma^+/+ arteries with PVAT, which, again, may reflect an access problem or may be attributable to saturation of native adiponectin receptors with endogenous adiponectin. In our electrophysiology experiments direct application of adiponectin failed to generate myocyte hyperpolarization in vessels derived from Kcnma1^−/− mice and also failed to elicit relaxation in bath solution transfer experiments in similar vessels. Our electrophysiology experiments also showed that adiponectin hyperpolarized wild-type arteries and that this was reversible by inhibition of BK_Ca channels with iberiotoxin. Such results, together with our other findings, demonstrate that adiponectin is likely to be the key anticontractile component liberated from PVAT.

Adiponectin acts via at least three receptors, AdipoR1, AdipoR2, and T-cadherin (5, 23, 47, 55). AdipoR1 has been identified on both smooth muscle and endothelial cells. Therefore inhibition of AdipoR1 may affect not only receptors on the smooth muscle but also those present on the endothelium. Our study demonstrates a considerable role for endothelium-derived factors in mediating the anticontractile effect; both inhibition of NOS and prostacyclin abolish the anticontractile effect. There is a residual (although not significant) effect of PVAT in endothelium-denuded arteries from Kcnma1^+/+ mice, which leads us to believe that the effect is not fully mediated via the endothelium but also has a smooth muscle component. Adiponectin has well-documented endothelial protective mechanisms and can increase production of NO (4, 5, 47). This may explain why inhibition of NOS inhibits the PVAT effect in our preparation. In rat aorta, Gao and colleagues (12) have demonstrated a role for endothelium-derived NO in the response and Greenstein et al. (15) reported a blunting of the anticontractility in healthy human arteries in the presence of endothelial NOS inhibitors. We also provide data to show that combination of l-NNA and indomethacin significantly increases tension in arteries surrounded by PVAT. This demonstrates the importance of NO and prostaglandins as mediators of the adiponectin- and PVAT-induced anticontractility. It has been previously reported that salt-fed adiponectin-deficient mice that develop hypertension have low levels of NOS and prostaglandins that can be reversed with adiponectin treatment (37). Additionally, there is also evidence of an interlinked NO pathway cyclooxygenase in endothelial cells where one will compensate for a reduction in levels of the other (48). It may be that a similar pathway exists in our PVAT preparation and that inhibiting both leads to an increase in tension.

Transfer of PVAT from arteries lacking the BK_Ca channel does not induce an anticontractile effect on arteries carrying the channel, which suggests that adipocyte BK_Ca channels are involved in the response. Potassium currents (39, 40) and BK_Ca channel currents have been identified on human preadipocytes and play a role in differentiation (20). It is possible that BK_Ca channels may influence adipocyte membrane potential and in turn influence the release of adiponectin. It has been suggested by others that adipocyte membrane potential may play a role in the basal metabolic activity of adipocytes (39, 40). Recently, release of PVAT relaxing factor from rat aortic rings has been shown to be dependent on the presence of adipocyte calcium (29). As the BK_Ca channel requires Ca²⁺ for activation, this is suggestive of a role for the channel in the release of adiponectin. However, it must be pointed out that in this group's experiments iberiotoxin did not alter responses.

Therefore we propose that adiponectin binds to AdipoR1 on smooth muscle cells, which activates BK_Ca channels, and the resultant hyperpolarization leads to vasodilation. How adiponectin receptor activation in turn activates BK_Ca channels remains to be elucidated. A strong candidate for communicating the effects of adiponectin to the BK_Ca channel is the heterotrimeric serine/threonine protein kinase AMP-activated protein kinase (AMPK) (16). AMPK dysregulation has been linked with obesity and the vascular consequences of metabolic syndrome (17). Also, there is evidence to suggest that this pathway will be altered in obesity because AMPK expression is significantly decreased in obese rats (33). Adiponectin is known to activate AMPK in a number of cell types (4, 26, 54, 56). This results in stimulation of NO production in endothelial cells, which may contribute to PVAT anticontractility, and AMPK-induced release of Ca²⁺ from smooth muscle cell stores (4, 18, 22, 24, 34, 53, 58). The Ca²⁺ release may trigger BK_Ca activation. Indeed, this may explain why the presence of the Ca²⁺-sensing β-subunit is important in the PVAT anticontractile response. Further studies are in progress to determine whether adiponectin activation of AMPK activates in turn a plasmalemmal channel to allow Ca²⁺ entry, thereby provoking BK_Ca opening. NO can also stimulate PKG (19, 30), which in turn is known to activate BK_Ca channels (51). Data recently published from our lab (52) demonstrated that loss of PKG function inhibits the anticontractile effect.

In summary, the present data demonstrate that the BK_Ca channel is crucial in mediating the anticontractile effect of PVAT, at least in murine mesenteric arteries. Adiponectin released from PVAT activates AdipoR1 on endothelial and myocytes; the subsequent activation of BK_Ca leading to myocyte hyperpolarization, along with an endothelial contribution, decreases vascular tone.

GRANTS

This work was funded by the Wellcome Trust and The British Heart Foundation.

DISCLOSURES

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

AUTHOR CONTRIBUTIONS

Author contributions: F.M.L., S.B.W., and A.M.H. conception and design of research; F.M.L., S.B.W., Z.Y., M.E.W., G.E., and A.H.W. performed experiments; F.M.L., Z.Y., and A.H.W. analyzed data; F.M.L. and A.M.H. interpreted results of experiments; F.M.L. prepared figures; F.M.L. drafted manuscript; F.M.L., S.B.W., M.E.W., G.E., A.H.W., and A.M.H. edited and revised manuscript; F.M.L., S.B.W., Z.Y., M.E.W., G.E., A.H.W., and A.M.H. approved final version of manuscript.