Degenerin/Epithelial Na⁺ Channel Proteins

The myogenic response is an important regulatory mechanism for blood flow autoregulation to renal, mesenteric, and cerebral circulation.¹ In these vascular beds, blood flow remains relatively constant despite changes in perfusion pressure via parallel changes in vascular resistance; when arterial pressure rises, resistance vessels constrict. Mechanosensitive channels in vascular smooth muscle cells (VSMCs) may be the first step in signal transduction of a mechanical stimulus (stretch) to a cellular response (vasoconstriction). However, the identity of mechanosensitive channels in VSMCs is unknown.¹

The degenerin/epithelial Na⁺ channel (DEG/ENaC) cation channel family is a class of proteins that may function as mechanosensors in a diverse range of species and cell types.² First identified in the nematode Caenorhabditis elegans, DEG/ENaC channels are required for mechanosensation.² The degenerins share a common topology and sequence homology with ENaC and acid-sensing ion channel (ASIC) proteins, 2 groups of mammalian DEG/ENaC proteins.^2,3 Expression of ASIC proteins may be limited to neuronal cells and taste buds,^2,4 whereas ENaC proteins are expressed in a variety of cell types, including epithelial cells, sensory neurons, keratinocytes, taste buds, and osteoblasts.^2,5,6

Although initial reports regarding the role of ENaC in mechanosensation are equivocal,^5,7–12 recent evidence supports the role of DEG/ENaC proteins as mechanosensors. DEG/ENaC transcripts and proteins are expressed in sensory ganglia containing rich populations of mechanoreceptive neurons, and proteins are localized at the site of mechanotransduction in nerve terminals innervating arterial baroreceptors, touch receptors in hairless skin, hair follicles, and vibrissae.^13–19 Furthermore, ASIC null mice have defects in specific populations of tactile mechanoreceptors.^17,18 Together, these data suggest that DEG/ENaC proteins may be components of mechanosensitive ion channel complexes. However, the role of DEG/ENaC proteins as mechanosensors in mammalian vascular smooth muscle has not been addressed. We hypothesized that DEG/ENaC proteins may also function as mechanosensors in vascular smooth muscle. To test this hypothesis, we used: (1) RT-PCR, (2) immunoblotting and immunolocalization to determine whether specific ENaC transcripts and proteins are expressed in isolated vessels and VSMCs, and (3) amiloride and benzamil, specific DEG/ENaC inhibitors at low micromolar and submicromolar doses, to examine the role of ENaC proteins as mechanosensors during the pressure-induced myogenic constriction in isolated rat cerebral arteries.

Experimental Procedures

Vessel Isolation

To isolate cerebral blood vessels, adult Sprague-Dawley rats (250 to 300 g) were deeply anesthetized with halothane and flushed with Hanks’ balanced salt solution via transcardial injection. The posterior communicating artery, middle cerebral artery, and posterior cerebral artery and their branches were dissected away under a microscope and immediately placed in RNA STAT-60 (Tel-Test “B”) for RNA isolation, fixative for immunostaining, lysis buffer for immunoblotting, or dissociation solution to liberate VSMCs. All procedures followed were in accordance with University of Mississippi Medical Center Institutional Animal Care and Use Committee guidelines.

Cerebral VSMC Dissociation

Cerebral vessels were isolated as described above and muscle cells dissociated according to Davis et al.²⁰

Reverse Transcription–Polymerase Chain Reaction

Total RNA was isolated from rat cerebral vessels and dissociated cerebral VSMCs, DNase treated, and reverse transcribed using random primers and avian myeloblastosis virus reverse transcriptase (Promega). Amplification conditions are indicated in supplemental Table I (available online at http://www.hypertensionaha.org). PCR products were visualized with ethidium bromide and sequenced to confirm identity.

Western Blotting

Standard protocols were used to isolate the Triton X-100 soluble proteins from isolated cerebral vessels. A detailed protocol is available in an online supplement (http://www.hypertensionaha.org).

Immunofluorescence

To determine localization of ENaC subunits in cerebral vessels, we evaluated immunofluorescence labeling for ENaC proteins in isolated cerebral vessels and dissociated cerebral VSMCs. Specific methods for immunostaining are included in the online supplement.

Role of ENaC in Myogenic Response: Studies With Amiloride and Benzamil

Myogenic responsiveness in rat middle cerebral artery segments were analyzed as detailed in the online supplement.

Results

RT-PCR Detection

Figure 1 A shows the detection of βrENaC and γrENaC by RT-PCR and αrENaC by nested PCR in isolated cerebral arteries in a representative experiment. In freshly dissociated rat cerebral VSMCs (Figure 1B), we detected γrENaC in the first round of PCR and βrENaC in the second round. We were unable to detect αrENaC in the second round of PCR. These data suggest that at least β and γ ENaC subunits are expressed in VSMCs.

• Download figure
• Download PowerPoint

Western Blot Detection

Western blots, used to detect ENaC expression in isolated rat cerebral vessels, are shown in Figure 2. In isolated cerebral vessels, the β and γ but not α antibodies labeled a major band just <78 kDa, near the predicted molecular weight of ENaC proteins (≈70 to 78 kDa). Experiments with 2 other anti-αENaC antibodies, kindly provided by Drs Douglas Eaton (Emory University School of Medicine, Atlanta, Ga) and Dale Benos (University of Alabama at Birmingtham), provided similar results (data not shown). These data suggest that β and γ subunits are the predominant subunits expressed in cerebral vessels.

• Download figure
• Download PowerPoint

Immunofluorescence

To determine whether ENaC subunits localized to VSMCs, we used immunofluorescence in freshly dissociated rat cerebral artery VSMCs (Figure 3 and supplemental Figure III [available online at http://www.hypertensionaha.org]) and isolated vessels (Figure 4). We detected expression of β and γENaC but not αENaC in VSMCs. Images demonstrating close association with α-actin and absence of αENaC immunostaining are available in supplemental Figure III. Typical of freshly dissociated muscle cells, α-actin staining is localized below the membrane. β and γENaC staining patterns were similar but not identical to α-actin, suggesting β and γENaC may be expressed at or near the cell surface (supplemental Figure III). To determine whether β and γENaC are expressed within the same muscle cell, some dissociated VSMC samples and isolated vessel segments were costained with rabbit anti-βENaC (βENaC [R]) and sheep anti-γENaC (γENaC [S]). As shown in Figure 3 (bottom row), β and γENaC expression appears to be clustered in similar locations within a smooth muscle cell when examined at higher magnification. Middle cerebral artery segments, the segments used to examine myogenic tone, also colabel for βENaC and γENaC (Figure 4).

• Download figure
• Download PowerPoint

• Download figure
• Download PowerPoint

Amiloride/Benzamil Block of Myogenic Constriction

Extraluminal amiloride and benzamil, which inhibit ENaC,²¹ blocked the myogenic response at very low micromolar and submicromolar concentrations (Figure 5). The data shown are absolute changes in inner diameter to stepwise 20 mm Hg increases in transmural pressure. Each data point represents the average of 3 readings in n=5 to 6 animals. Under control conditions, vessel diameters decreased in response to increasing pressure. At 20 mm Hg, absolute vessel diameters were not different among the groups (Figure 5 legend). The data at 50 μmol/L amiloride are not shown, because the data are not different from Ca²⁺ free bathing solution, indicating a total block of the myogenic response at this dose. At 60 mm Hg and higher, the change in diameter for all treatments was different from control. These data demonstrate that low micromolar and submicromolar doses of amiloride and benzamil inhibit myogenic vasoconstriction in response to increasing perfusion pressure by blocking activity of ENaC proteins.

• Download figure
• Download PowerPoint

Discussion

Mechanosensitive cation channels have been found in nearly all cell types, yet the molecular identities of these channels remain unknown.²² Many mechanisms have been proposed to transduce mechanical stimuli into cellular responses in VSMCs, including: (1) membrane-bound enzyme systems, such as phospholipase A2 or C; (2) activation of ion transporters and exchangers, such as Na⁺-Ca²⁺ and Na⁺-H⁺ exchangers; and (3) mechanosensitive ion channels, including voltage-gated Ca²⁺ channels, Ca²⁺-activated K⁺ channel, transient receptor potential (TRP) proteins, Cl⁻ channels and cation selective channels.²⁰ We propose that DEG/ENaC subunits form the latter, cation selective channels. The major finding of this investigation is that ENaC subunits are expressed in cerebral VSMCs, and pharmacological blockade of ENaC abolishes the vasoconstrictor response to increases in perfusion pressure. These data suggest that DEG/ENaC are required for pressure-induced vessel responses.

Role of DEG/ENaC Proteins as Mechanosensors

A growing body of evidence suggests DEG/ENaC proteins form mechanosensors in many tissues and species. DEG/ENaC proteins are expressed in tissues rich in mechanoreceptors and are required for normal responses to mechanical stimulation.^{13–15,17,18,23} However, the role of DEG/ENaC proteins as mechanosensors in vascular tissue has not been evaluated previously.

Detection of ENaC Subunits in VSMCs

To test the hypothesis that ENaC proteins may be components of a mechanosensitive channel in VSMCs, we first determined whether ENaC subunits are expressed in vascular tissue. We used RT-PCR, Western blotting, and immunostaining to determine whether ENaC message and protein are expressed in isolated rat cerebral vessels and specifically in VSMCs. Message for all 3 subunits was detected in isolated vessels, although a second round of amplification was necessary to detect αENaC. Interestingly, we were unable to detect αENaC expression in vessels or freshly dissociated cerebral VSMCs by immunoassay or RT-PCR, respectively. There are at least 2 possible explanations for these results. First, αENaC may be expressed below detection sensitivity for immunolabeling. Second, αENaC may not be expressed in VSMCs but expressed in another cell type, such as endothelial cells.²⁴

To determine the localization of ENaC subunits in VSMCs, we immunostained freshly dissociated VSMCs cells for ENaC subunits and α-actin, a membrane-associated cytoskeletal protein specific to smooth muscle. In freshly dissociated VSMCs, α-actin stains just below the sarcolemmal membrane.²⁵ The close association between β and γENaC and α-actin immunolabeling indicates the ENaC subunits are localized at or near the membrane. Furthermore, clustering of β and γENaC staining at or near the membrane suggests the 2 subunits may associate. This expression pattern places the channel at an ideal site, where it can be gated by mechanical stress at the membrane.

Mechanosensitive Ion Channel Complex

Many investigators speculate mechanosensitive ion channels interact with cytoskeletal and extracellular matrix proteins that tether the channel and permit mechanical gating. Pore-forming DEG/ENaC subunits are thought to interact with intracellular, membrane-associated, stomatin-related proteins, which may associate directly or indirectly with the cytoskeleton.² Although stomatin-related transcripts are expressed in VSMCs (H.A. Drummond, 2002, unpublished data), it is unknown whether they associate with ENaC subunits. On the extracellular side, DEG/ENaC proteins may be linked to extracellular matrix proteins such as collagen.²

It is likely that the mechanosensor channel formed by DEG/ENaC subunits in VSMCs is not the same channel found in epithelial cells. Despite similar single-channel conductances (ENaC 5 to 40 pS; mechanosensitive cation channels 30 to 40 pS), ion selectivity and gating are different in epithelial ENaC and VSMC mechanosensitive channels.^1,8 A lack of αENaC in VSMCs could explain gating differences with epithelial ENaC channels. Because the αENaC subunit confers constitutive activity to the ENaC channel, loss of αENaC would suggest the channel formed by β and γENaC is quiescent,^26,27 which is consistent with the nature of mechanosensitive ion channels.^20,28,29 However, because amiloride binding sites are also present in β and γENaC, it is likely that a mechanically gated channel formed by βγENaC could still be blocked by amiloride.³⁰

Role of ENaC Proteins in Myogenic Response of Isolated Cerebral Vessels

To determine whether ENaC proteins play a role in the myogenic response, we evaluated stretch-induced vasoconstriction in isolated rat middle cerebral arteries. Our data indicate that >50% of myogenic response was blocked with 1 μmol/L amiloride. Using benzamil, a more potent and selective ENaC inhibitor, we blocked ≈40% and 75% of the myogenic constrictor response with 30 nmol/L and 1 μmol/L benzamil, respectively.^21,27 At these doses, the selectivity of amiloride and benzamil is a critical issue regarding interpretation of the data. There is substantial evidence that amiloride and benzamil, at the low doses used in the present study, inhibit ENaC but do not inhibit other molecules such as the l-type Ca²⁺ channel, Na⁺-H⁺ exchanger, Na⁺-Ca²⁺ exchanger, and TRP6, a transient receptor potential channel family member implicated recently as a potential mediator of the myogenic response.^21,31,32 These channels and transporters are blocked by doses 100 to 10 000× greater than used in the present study. Amiloride analogs, such as hexamethyl amiloride, dimethyl amiloride, and methyl isobutyl amiloride, are potent inhibitors of the Na⁺/H⁺ exchanger. These analogs are inadequate controls because disruption of these channels and transporters can alter ion gradients and impact myogenic tone independently.^1,33–40

Other investigators have suggested that changes in extracellular sodium may alter myogenic responsiveness by altering smooth muscle mechanoreceptor sensitivity to stretch.^37,38 In isolated vessels, increases in extracellular sodium decrease myogenic tone and vice versa.³⁷ Interestingly, the αβγENaC channel displays “self-inhibition,” a phenomenon in which amiloride-sensitive current slowly decreases over seconds to minutes after increases in extracellular sodium.⁴¹ Feedback inhibition of ENaC channels in VSMCs may account for myogenic tone sensitivity to extracellular sodium.

Although the doses of amiloride and benzamil used in the present study are highly specific for ENaC, the results do not directly prove that ENaC proteins are functioning as mechanosensors. It is possible that ENaC proteins are required for a downstream step in the signal transduction cascade leading to vasoconstriction. However, we consider this unlikely because amiloride does not block vascular responses to other vasoactive agents.^37,42 In cerebral arteries, 100 μmol/L amiloride, a dose 100× greater than used here, does not suppress KCl or prostaglandin F2-induced vasoconstriction or alter cerebral artery intracellular Ca²⁺ or pH.⁴² Additionally, in facial veins, 10 μmol/L amiloride blocks myogenic tone but does not block histamine-induced tone.³⁷ It is it also unlikely that the effect of amiloride on the myogenic response is attributable to its action on endothelial cells because the vessels are de-endothelialized when an air bolus is passed through the vessel.⁴³

We speculate β and γENaC subunits, perhaps in conjunction with other degenerin subunits, form the pore of a mechanosensitive, nonselective cation channel. Increases in pressure or vessel wall strain activate the mechanosensor via protein–protein interactions between pore-forming subunits and intracellular and extracellular proteins anchoring the channel. The mechanically gated cation influx (most likely Na⁺ and Ca²⁺) triggers secondary signal transduction pathways (phospholipases, ion transporters, and channels) that lead to the elevations in intracellular Ca²⁺ and muscle contraction.

Perspectives

The role of DEG/ENaC proteins as mechanosensors in vascular tissue has not been evaluated previously. The results of this study suggest ENaC proteins may mediate pressure-induced vasoconstriction and blood flow regulation in cerebral vessels. It is likely that DEG/ENaC proteins act as mechanosensors in smooth muscle cells in other circulation with strong myogenic control of blood flow. In the kidney, end organ damage associated with hypertension has been attributed to blood pressure rather than the hormonal milieu. Therefore, understanding how vascular cells sense pressure, as well as other mechanical stimuli, is a key step in understanding mechanisms leading to end organ damage with hypertension.

Footnotes