Mechanosensitive ion channels are thought to initiate pressure-induced vasoconstriction, however, the molecular identity of these channels is unknown. Recent work from our laboratory suggests that members of the Degenerin/Epithelial Na+ Channel (DEG/ENaC) family may be components of the mechanosensitive ion channel complex in vascular smooth muscle (VSM); however, the specific DEG/ENaC proteins mediating myogenic constriction are unknown. The goal of this study is to determine if specific knockdown of β or γENaC, using dominant-negative (DN) or small-interference RNA (siRNA) molecules, inhibits pressure-induced vasoconstriction in mouse renal interlobar arteries. To address this goal, isolated arteries were transiently transfected with β or γENaC DN-cDNA or siRNA molecules. After 24 h, vessels were either 1) cannulated and pressurized for pressure-diameter response curves or 2) dissociated and immunolabeled to determine VSM cell endogenous ENaC protein expression. We found that transfection of βENaC DN-cDNA or siRNA suppresses β-, but not γENaC protein expression. Similarly, γENaC DN-cDNA or siRNA suppresses γ-, but not βENaC protein expression. In addition, transfection of β- or γENaC DN-cDNA or siRNA molecules inhibits pressure-induced vasoconstriction, but does not block agonist-induced vasoconstriction. Our results provide the first direct evidence that β and γENaC proteins are essential in mediating myogenic vasoconstriction in mouse renal interlobar arteries.
- transfected isolated renal vessel
- stretch-activated cation channel
- dominant negative
vascular smooth muscle cells (VSMC) from most small resistance arteries exhibit an innate ability to contract in response to elevations in intraluminal pressure. The active vasoconstriction is considered myogenic, or of muscular origin, typically occurring independently of neural, metabolic, or hormonal factors (7). Pressure-induced VSM membrane depolarization (6, 15, 23) and subsequent Ca2+ influx via voltage-gated Ca2+ channels largely mediate myogenic constriction (6, 22, 31). The response involves transduction of a mechanical stimulus (pressure-induced stretch) into a cellular event (depolarization/contraction). Although mechanosensitive nonselective cation channels are thought to initiate pressure-induced depolarization (6, 22, 31), the molecule(s) involved have not been fully characterized.
Members of the Degenerin/Epithelial Na+ Channel (DEG/ENaC) family of proteins are candidates for mechanosensitive ion channels in vascular smooth muscle. Evidence suggests DEG/ENaC proteins may act as mechanosensors. First, members of this evolutionarily conserved family form mechanosensors in the nematode, Caenorhabditis elegans (26). Second, ENaC channels can be activated by shear stress, a mechanical factor (5, 28). Third, ENaC proteins are required for normal mechanosensation in a variety of mammalian cell types (3, 10, 13, 21, 25, 32). Through selective pharmacological inhibition of DEG/ENaC channels, our laboratory has recently provided evidence that DEG/ENaC channels play an important role in myogenic constriction in isolated, pressurized rat middle cerebral arteries (9) and mouse renal interlobar arteries (18). These pharmacological inhibitors have provided a basic tool to screen for DEG/ENaC channel involvement because they block a broad spectrum of DEG/ENaC channels. While we detected β and γ, but not α, ENaC transcripts and protein in freshly dispersed VSMCs from rat cerebral and mouse renal vessels (9, 18), direct evidence supporting a role of β and/or γENaC in myogenic constriction is still lacking. The goal of this investigation was to determine whether the presence of β and/or γENaC is important in eliciting myogenic constriction in renal interlobar arteries. To achieve this goal, we used novel approaches, dominant-negative (DN) and small-interference RNA (siRNA), to silence endogenous ENaC expression. Our findings provide the first direct evidence that β- and γENaC are essential for myogenic vasoconstriction in mouse renal interlobar arteries.
MATERIALS AND METHODS
All protocols and procedures employed in this study were reviewed and approved by the Institutional Animal Care and Use Committee of the University of Mississippi Medical Center.
Preparation of dominant negative (DN) expression vectors and siRNA. The generation of our DN expression vectors has been previously described (8). Briefly, premature stop codons (X) were engineered into nucleotide bases encoding for amino acid I41 for βENaC (EGFP-βI41X) and L160 for γENaC (EGFP-γL160X) and ligated into Enhanced Green Fluorescent Protein C-terminal expression vector (pEGFP-C1). The EGFP expression vector was used as a negative control and EGFP fluorescence was used to identify transfection efficiency of the arteries (Fig. 1) and VSMC (Fig. 2).
Small interfering RNA (siRNA) molecules designed against mouse βENaC (SCNN_1b; ID # 69163) and γENaC (SCNN_1g; ID # 151576) were obtained from Ambion (Austin, TX). As a control, we used a non-targeting siRNA (Dharmacon, Chicago, IL) which activates the RNA-induced silencing complex (RISC).
Transfection of renal interlobar arteries.
To transfect renal interlobar arteries with DN-cDNA or siRNA, we used a modified protocol previously published by Kaide et al. (19). Briefly, male C57BL/6J mice (6–8 wks; Jackson Laboratory, Bar Harbor, ME) were anesthetized with halothane and decapitated. Kidneys were excised and placed in ice cold physiological saline solution [(PSS, pH adjusted to 7.4 with NaOH) containing (in mM) 130 NaCl, 4 KC1, 1.2 MgSO4, 4 NaHCO3, 1.8 CaC12, 10 HEPES, 1.18 KH2PO4, 6 glucose, 0.03 EDTA]. Interlobar arteries were dissected from hemi-sected kidneys and transferred to a 35 mm culture dish and transiently transfected with the DN-cDNA (10 μg of EGFP, EGFP-βI41X, or EGFP-γL160x) or siRNA (10 μg of non-targeting, βENaC, or γENaC) using Lipofectamine 2000 (Invitrogen, Carlsbad, CA) according to manufacturer’s instructions. Additional vessels were incubated with Lipofectamine (Lfx) alone as a vehicle control. After 4 hrs, a 1:1 ratio of Dulbecco’s Modified Eagle Medium (DMEM):F-12 supplement (GIBCO Laboratories) containing 100U/ml penicillin/streptomycin was added to the vessels. The vessels were maintained in organ culture for 20–24 hrs in a humidified incubator (95% air-5% CO2) at 37°C. At the end of the culture period, vessels were either used for vascular reactivity studies or were further dissociated for quantitative immunofluorescence studies.
ENaC immunofluorescence in dispersed VSMCs.
To determine endogenous expression of β- and γENaC following vessel transfection with β- and γENaC DN-cDNA or siRNA, VSMCs were enzymatically dissociated from the isolated vessels and fixed as described previously (18). Anti-ENaC antibodies were raised in rabbits [beta mouse ENaC (617–638): NH2-CNYDSLRLQPLDTMESDSEVEAI-COOH, gamma mouse ENaC (618–639): NH2-CPAPEAPVPGTPPPRYNTLRLD-COOH, Sigma Genosys, Woodlands TX]. The βENaC antigenic sequence has 100% identity with rat and 90% identity with human isoforms. The γENaC antigenic sequence has 84% identity with rat and 79% identity with human isoforms. Antibodies were affinity purified against the original antigenic sequence (Sigma Genosys) and screened by ELISA, immunolabeling in COS-7 cells transfected with β and γENaC cDNA and immunolabeling in VSMCs with antigen competition. Samples were labeled with rabbit anti- β- or γENaC antibodies (1:100) and mouse anti-smooth muscle α-actin (1:200) to identify VSMC and normalize immunofluorescence. Samples were examined using fluorescence confocal microscopy (TCS-SP2, Leica Microsystems, Exton, PA). To normalize data, β- and γENaC fluorescence was divided by α-actin fluorescence for each cell. All α-actin-labeled cells were used for analysis independent of EGFP fluorescence. Data were averaged from 3 animals in each group.
To determine vascular reactivity following transfection with β- and/or γENaC DN-cDNA or siRNA, mouse renal interlobar arteries were cannulated and studied in a vessel chamber (CH/1/SH, Living Systems, Burlington, VT) and analyzed using MetaMorph software (Universal Imaging, Dowingtown, PA) as described previously (18). Following an initial incubation period (30 min; 75 mmHg), a concentration-response curve to phenylephrine (PE; 10–9-10–5 M) was generated in each vessel to determine vessel viability. Vessels with less than 60% maximal response (calculated as percent of baseline inner diameter) were excluded. After washing and re-equilibrating the vessel, a pressure-diameter curve was generated by exposing the interlobar arteries to step-wise increases in intraluminal pressure from 25 to 150 mmHg (25 mmHg steps, 5 min each). Then, arteries were equilibrated for 30 min with Ca2+-free PSS (same as above PSS plus 2 mM EGTA and omit 1.8 mM CaCl2) to determine the passive pressure-diameter curve as described above. Change in diameter was calculated as the difference between the active (PSS) and passive (Ca2+-free PSS) inner diameter at each pressure. Data were averaged from 5 animals in each group. When more than one treated artery was used from a single animal, the results of the multiple experiments were averaged and included as a single value.
All data are expressed as means ± SE. A one-way ANOVA and two-way ANOVA with repeated measures were used to make comparisons where appropriate. Differences among groups were compared using the Student-Newman-Keuls post hoc test. Statistical significance was considered at P < 0.05.
ENaC DN-cDNA or siRNA suppresses endogenous β and γENaC protein expression in renal interlobar arteries.
We used quantitative immunofluorescence to determine if β or γENaC DN isoforms or siRNA suppresses ENaC protein. Representative images and quantitative group data are shown in Figs. 2 and 3. Compared with the Lfx and EGFP controls, transfection of EGFP-βI41X reduced endogenous βENaC expression ∼60% (Fig. 2, A and B); and EGFP-γL160X reduced endogenous γENaC expression ∼50% (Fig. 2, C and D). In addition, transfection with EGFP-βI41X did not suppress endogenous γENaC expression (Fig. 2, C and D) nor did EGFP-γL160X alter endogenous βENaC expression (Fig. 2, A and B). Similarly, compared with non-targeting siRNA controls, βENaC-siRNA suppressed ∼70% endogenous β- (Fig. 3, A and B), but not γENaC (Fig. 3, C and D). In addition, γENaC-siRNA suppressed ∼60% endogenous γ- (Fig. 3, C and D), but not βENaC (Fig. 3, A and B). These data show DN and siRNA molecules selectively silence β or γENaC expression in renal interlobar arteries.
ENaC DN-cDNA or siRNA inhibits renal interlobar myogenic constriction.
PE-induced vasoconstrictor responses are shown in Fig. 4 for DN (Fig. 4A) and siRNA (Fig. 4B) transfected arteries. PE elicited similar vasoconstrictor responses in all groups. This demonstrates suppression of endogenous ENaC does not alter vascular reactivity, per se. In contrast, transfection with β and γENaC DN-cDNA and siRNA molecules dramatically inhibited pressure-induced constriction (Fig. 5 and 6). Transfection with EGFP-γL160X maximally inhibited ∼70% the myogenic response, while transfection with EGFP-βI41X blocked ∼80% (Fig. 5, C-E). Compared with the Lipofectamine (Lfx) vehicle control, EGFP alone did not significantly alter myogenic tone (Fig. 5, B and E). As shown in Fig. 6, βENaC-siRNA and γENaC-siRNA significantly reduced myogenic constriction by 65 and 40%, respectively, in renal interlobar arteries compared with the non-targeting siRNA control.
Previous studies from our laboratory demonstrated that DEG/ENaC channel blockade abolished pressure induced vasoconstriction in rat middle cerebral arteries (9) and mouse interlobar arteries (18). Although these experiments provided evidence members of the DEG/ENaC family are required to mediate myogenic constriction, they did not identify which individual DEG/ENaC proteins are involved. The major finding of this investigation is that β and γENaC expression in mouse renal interlobar arteries is required for myogenic constriction.
Until recently, mammalian ENaC proteins have been described mainly in kidney epithelial cells where they form constitutively active channels that play a rate limiting role in sodium absorption, and therefore, an essential role in the control of sodium balance, blood volume, and blood pressure (14, 20). However, our findings suggest that ENaC proteins in VSMCs may contribute to local blood flow regulation via myogenic constriction. This is a novel finding and one that supports the unique function of this protein family as mechanosensors. The involvement of DEG/ENaC in mechanotransduction has been shown in many other species and cell types (2, 12, 14, 20, 24); including mammalian dorsal root ganglion, arterial baroreflex sensory neurons, osteoblasts, keratinocytes, pain and touch receptors, and renal tubular cells (3, 10, 13, 21, 25, 28). Here, we provide evidence β and γENaC proteins are also involved in mechanotransduction in VSMCs.
Similiar to previous findings from our lab, we found (9, 18), β and γENaC expression is concentrated at or near the membrane, as indicated by colocalization with α-actin. Previous work by ourselves and others show α-actin staining is quite different in VSMCs that have been freshly dissociated compared with VSMCs maintained in culture (9, 18, 33). In VSMCs maintained in culture, α-actin staining is filamentous and distributed through out the cell. However, in freshly dispersed VSMCs, α-actin is concentrated near the membrane, making it an excellent marker of near-membrane proteins. In contrast to our previous findings, in the current study we detected more cytoplasmic localization of β and γENaC protein, suggesting some redistribution of ENaC towards the cytoplasm. However, it is important to note that despite the presence of cytoplasmic ENaC expression, most ENaC expression is still localized to the membrane region. The factors accounting for the cytoplasmic localization of ENaC are unknown, but are most likely due to the overnight incubation period since this was the only methodological difference between the previous (18) and current studies. We speculate that substances present in the culturing media and/or the loss of transmural pressure in the arteries are contributing factors. In previous reports, we have shown a pronounced cytoplasmic staining pattern in renal VSMCs maintained in culture for a longer period (1–3 wk) (18). Taken together, these findings suggest that localization of ENaC proteins within the VSMC, i.e., cytoplasmic vs. membrane, is not “hard-wired”, but can be influenced by the cells’ environment. However, whether loss of transmural pressure and supplements in the culturing media are responsible for regulating ENaC localization remains to be determined.
Two approaches were used to determine if ENaC proteins are required for myogenic constriction; siRNA and dominant-negative ENaC isoforms. As expected, siRNA, a known posttranscriptional gene silencer, reduced protein levels of βENaC and γENaC by ∼50% 24 h after transfection. The effect of siRNA on ENaC channel activity per se cannot be determined from our investigation since electrophysiological measurements of channel activity were not made. However, our findings indicate the 50% reduction in immunoreactive protein levels of βENaC or γENaC resulted in an almost complete loss of the myogenic constrictor response. This indicates normal levels of both βENaC and γENaC protein are required for a normal myogenic constrictor response and further suggests they may interact or associate.
In addition to siRNA, we inhibited ENaC channel function using a DN approach. DN ENaC isoforms also inhibited β and γENaC protein levels to a similar extent as siRNA, however, most likely by a different mechanism. The NH2-terminal fragment of a related degenerin family member, Mec-4, has been shown to inhibit the response to touch in C. elegans (17). A similar NH2-terminal fragment of γENaC has been shown to inhibit αβγENaC channel activity, most likely by associating with other full-length subunits to form nonfunctional channels (1). The improperly formed channels are thought to be unstable and degraded, the basis for a loss of channel activity. Based on this finding, we anticipated either DN-cDNA would inhibit protein levels of β and γENaC. There are at least two possible explanations for these findings. First, β and γENaC may form homomeric channels. However, we think this is unlikely given our finding that siRNA towards either subunit abolished 70–80% of the response. A second possibility is that the DN (EGFP-βI41X) interacts with its full-length βENaC counterpart prior to assembly or association with γENaC. Further studies are required to determine the mechanisms underlying DN inhibition of ENaC channel function.
A critical issue is whether βENaC and γENaC can form an ion-conducting pore in the absence of αENaC. In heterologous expression systems, evidence suggests βENaC and γENaC are capable of associating in the absence of αENaC. When coexpressed without αENaC, βENaC immunoprecipitates with γENaC, suggesting a biochemical association between βENaC and γENaC that persists in the absence αENaC. Using the Xenopus oocyte expression system, a report by Bonny et al. (2a) demonstrates βγENaC can generate amiloride-sensitive macroscopic currents after 6 days of incubation. A delayed trafficking of βγENaC channels in the ooctye system may be the basis for difficulty in identifying currents generated by βγENaC channels by others. Bonny et al.’s finding is significant because it demonstrates βENaC and γENaC are sufficient to form a pore. Alternatively, it is possible that another degenerin protein, such as an ASIC protein, also associates with βENaC and γENaC to form the pore of a channel in VSMCs.
Presently, there is no electrophysiological evidence that ENaC proteins form a channel in VSMCs. However, the presence of a non-voltage-gated, epithelial-like Na+ channel in VSMCs was reported by Renterghem and Lazdunski 15 years ago (30). Similar to ENaC channels, the channel reported by Renterghem and Lazdunski exhibited high Na+ selectivity, a 10-pS conductance and was not voltaged gated. Unlike ENaC channels, the VSMC channel was insensitive to amiloride, butsensitive to higher concentrations of phenamil, an amiloride analog. Since that time, the presence and importance of these channels have received little attention. Electrophysiologic evidence confirming the presence of ENaC channels in VSMCs remains an important area of future investigation.
In comparison to responses in freshly isolated renal interlobar arteries (18), a small loss of the myogenic constrictor response was observed in the transfection control experiments (Lfx, EGFP, NT-siRNA), particularly at higher pressures (≥100 mmHg). However, this is not surprising given the 24-h incubation period required to induce gene silencing. Although it may be coincidental, we speculate the depression of the myogenic constrictor response following the 24-h incubation may be related to the redistribution of β and γENaC towards the cytoplasm noted previously. A redistribution of ENaC proteins towards the cytoplasm may reflect a reduction in the amount of active, membrane-associated ENaC channels contributing to signal transduction, which might be expected to lead to a reduction in myogenic tone development.
It is generally accepted that pressure-induced VSM membrane depolarization (6, 15, 23) and subsequent Ca2+ influx via voltage-gated Ca2+ channels largely mediates myogenic constriction (6, 22, 31). Many studies have followed this initial finding to determine a common signaling mechanism initiating the myogenic response. However, it has been difficult to ascribe myogenic constriction to a single mechanism because increased pressure activates multiple factors that influence VSM depolarization and their relative contribution varies immensely between, and even within, vascular beds (16). Potential mechanisms include pressure-induced activation of membrane-associated enzyme systems, ion transportors and mechanosensitive ion channels (7).
Although we do not know the precise function of ENaC proteins or whether their function is consistent across vascular beds, we speculate that βγENaC subunits are part of a large complex, previously modeled in C. elegans (24). ENaC subunits form the pore of the channel in VSMCs and interact with extracellular matrix and cytoskeletal proteins to form a larger mechanosensitive complex. This speculation is supported by findings that demonstrate cytoskeletal and extracellular matrix proteins interact with certain DEG/ENaC proteins (2, 21, 24, 27). Furthermore, at least one cytoskeletal protein may be required to mechanically gate a mechanosensitive channel formed by DEG proteins in C. elegans (26). We further speculate that upon mechanical activation, the channel opens and allows influx of Na+ and/or Ca2+ ions, presumably leading to depolarization. While our laboratory has shown that inhibition of DEG/ENaC abolishes pressure-induced increases in cytosolic Na+ and Ca2+ (18), we have not yet correlated this to pressure-induced depolarization.
In the current investigation, we provide evidence ENaC proteins are involved in the myogenic response in the renal circulation. However, ENaCs may not contribute to myogenic responses in every vascular bed. The involvement of other mechanosensitive proteins, such as transient receptor potential channels (11, 29) or acid-sensing ion channels, may play a role in myogenic constriction in other beds. Results from these experiments identify a new group of proteins involved in vascular function; however, many fundamental mechanistic questions remain. How is the channel gated by mechanical stimuli? Does the channel interact with extracellular matrix or other degenerin proteins? How is the channel regulated by hormonal and autocrine factors?
This work is supported by National Institutes of Health Grants HL-082425 (N. Jernigan), HL-071603 (H. Drummond), AHA 0655305B (H. Drummond), and HL-51971.
The authors thank A. Hoover for technical assistance and laboratory colleague S. Grifoni for assistance and discussion.
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
- Copyright © 2006 the American Physiological Society