|Home | About | Journals | Submit | Contact Us | Français|
Migration of vascular smooth muscle cells (VSMC) into neointima contributes to atherosclerosis and restenosis. This migration requires co-ordinated plasmalemmal fluxes of water and ions. Here, we show that aortic VSMC migration is dependent upon the regulation of transmembrane Cl− flux by ClC-3, a Cl− channel/transporter. The contribution of ClC-3 to plasmalemmal Cl− current was studied in VSMC by electrophysiological recordings. Cl− current was negligible in cells perfused with zero [Ca2+]. Raising intracellular [Ca2+] to 0.5 μM activated a Cl− current (ICl.Ca), approximately half of which was eliminated upon inhibition by KN-93 of calmodulin-dependent protein kinase II (CaMKII). ICl.Ca was also halved by inositol-3,4,5,6-tetrakisphosphate (IP4), a cellular signal with the biological function of specifically preventing CaMKII from activating ICl.Ca. Gene disruption of ClC-3 reduced ICl.Ca by 50%. Moreover, ICl.Ca in the ClC-3 null VSMC was not affected by either KN-93 or IP4. We conclude that ICl.Ca comprises two components: one is ClC-3 independent, while the other is ClC-3 dependent, activated by CaMKII, and inhibited by IP4. We also assayed VSMC migration in transwell assays. Migration was halved in ClC-3 null cells versus wild-type cells. Additionally, inhibition of ClC-3 by either niflumic acid, KN-93 or IP4, each reduced cell-migration in wild-type cells, but not in ClC-3 null cells. These cell-signaling roles of ClC-3 in VSMC migration suggest new therapeutic approaches to vascular remodeling diseases.
Proliferation and migration of vascular smooth muscle cells (VSMC) are critical steps in the process of vascular remodeling that is associated with hypertension, atherosclerosis, and restenosis1,2. The identification of proteins that participate in VSMC proliferation and migration, and characterization of their regulation, can assist therapeutic intervention in vascular remodeling1–3.
Cell migration is underpinned by changes in cell shape that are driven by temporally and spatially separated phases of localized cell swelling and cell shrinkage4,5. The latter are activities that require precise spatio-temporal control of ion channels and transporters so as to co-ordinate movements of ions and water across the plasma membrane. The regulation of plasmalemmal Cl− flux by the ClC-3 channel/transporter6 contributes to the migration of certain tumor cells7,8. However, it has not previously been studied if ClC-3 mediates the migration of VSMC in vitro or in vivo.
The manner in which ClC-3 activity is controlled underscores its potential importance as a regulator of cellular biology. In some cell types, ClC-3 dependent Cl− flux is stimulated by CaMKII8–10. Furthermore, the degree of activation of ClC-3 by CaMKII is physiologically constrained by inositol 3,4,5,6-tetrakisphosphate (Ins(3,4,5,6)P4), a member of the inositol phosphate signaling family11. In the current study we have investigated if ClC-3 and Ins(3,4,5,6)P4 regulate plasmalemmal Cl− flux and migration of VSMC. To recapitulate the pro-migratory phenotype of VSMC during vascular disease, we used primary cell cultures12; we obtained VSMC from wild-type and ClC-3 null mice.
Detailed methods are available in the online data supplement (please see http://hyper.ahajournals.org). ClC-3 null and wild-type aortic vascular smooth muscle cells were obtained as previously described13. Cells were grown in 5% fetal bovine serum, in Dulbecco’s Modified Eagle Medium (DMEM) containing 1% vitamins,1% glutamine, 1% penicillin/streptomycin, 1% non-essential amino acids and 25 mM HEPES (pH 7.4) and were maintained at 37°C in 5% CO2. Plasmalemmal Cl− currents were measured using the whole-cell patch clamp technique in conditions where Cl− was the only permeant anion. The migration assays were performed at 37 °C in 5% CO2 in 24-well hanging cell-culture inserts with 8 μM pore size (Millipore, Billerica, MA). To each insert, 40,000 cells were added in 200 μl DMEM. Each insert was placed in a well that contained 650 μl DMEM. Migration assays were begun 90 min after seeding with the addition of either 40 ng/ml PDGF-BB (R&D Systems, Minneapolis, MN) or 10% FBS, plus either a test reagent (added to both compartments of the transwell chamber) or a matched control containing an equivalent concentration of vehicle (DMSO). Cells were incubated for 17 hr (with FBS) or 5 hr (with PDGF). In control experiments, total cell number and viability were checked with a Countess Cell Counter (Life Technologies, New York). Migrated cells on the bottom surface of the insert were identified using a Diff-Quik Stain set (Siemens Healthcare Diagnostics Inc., Newark DE; 1 min in fixative, 2 min in solution I, and 4 min in solution II followed by rinsing in distilled water. Six 132 mm2 sections of the inserts were imaged by bright field microscopy (20X objective), and then the number of migrated cells was counted; values from the six different areas were averaged. All current/voltage data were statistically analyzed by ANOVA with a Bonferroni post-hoc test. Other data were analyzed with Student’s t-test. In the figure legends, ‘n’ = number of experiments. Values of P < 0.05 were considered significant.
We have investigated if ClC-3 influences the migration of VSMC in transwell assays. Cells were obtained from both wild-type and ClC-3 null mice; the success of the gene disruption is illustrated by Western analysis (Fig. 1A). Cell migration in response to PDGF, a pro-migratory stimulus14, was substantially reduced in ClC-3 null cells compared to WT cells (Fig. 1B, C, D). That same phenotype was also observed in an experimental paradigm (serum-stimulated migration15,16) that more closely mimics the complex growth factor environment in vivo (Fig. 1E, F, G). The reduced migration of ClC-3 null cells was not an indirect consequence of there being fewer cells (see legend to Fig. 1). Thus, the ClC-3 null cells did not proliferate more slowly. The latter observation is consistent with a previous2 conclusion that disruption of the ClC-3 gene did not affect VSMC proliferation per se, even though it specifically halved the proliferative response to the proinflammatory cytokine, TNF-α.
Plasmalemmal Cl− fluxes mediated by ClC-3 may be enhanced by either cell swelling (IClswell), or by elevated [Ca2+]IN (ICl.Ca) (see refs17,18). In VSMC, IClswell was unaffected by elimination of ClC-3, irrespective of whether Ca2+ was present19 or absent (Fig. S1). We next measured ICl.Ca in wild-type and ClC-3 null cells using media containing BAPTA to buffer [Ca2+]IN (Inset to Fig. 2A). Little Cl− current was observed at zero [Ca2+]IN, but with 0.5 μM [Ca2+]IN, whole-cell current was increased in both cell types (Fig. 2A,B). ICl.Ca was outwardly rectifying (Fig. 2A,B); kinetic analysis (Fig. S2) revealed time-dependent activation of Cl− currents at depolarizing potentials. These are the prototypical characteristics of ICl.Ca20–24. These kinetic properties were not significantly affected by disruption of the ClC-3 gene (Fig. S2).
ICl.Ca was approximately 50% smaller in ClC-3 null VSMC cells than in wild-type cells (Fig. 2). It has not previously been demonstrated that ICl.Ca is decreased following knock-out of the ClC-3 gene, in any cell type. Thus, we highlighted this novel result by depicting steady-state currents at −100 mV and +40 mV (Fig. 2C). We chose these voltages as they are equidistant from the ECl at − 31 mV; while opposite in polarity, the two resultant currents are elicited by electrical driving forces that are equal in magnitude. We conclude that ICl.Ca in the wild-type cells is comprised of approximately equal contributions from both ClC-3 dependent and ClC-3 independent components (Fig. S3).
A generic Cl− channel blocker, niflumic acid (NFA)25, inhibited ICl.Ca in both cell types (Fig. 2A,B). NFA also inhibited serum-stimulated migration in the wild-type cells (Fig. 2D); cell numbers and viability were not affected by NFA (Fig. 2 legend). NFA did not alter migration in ClC-3 null cells (Fig. 2D). Our data suggest the ClC-3 dependent component of ICl.Ca contributes to VSMC migration (Fig. S3).
How does Ca2+ activate ICl.Ca in a ClC-3 dependent manner in aortic VSMC? We are unaware of evidence that Ca2+ can stimulate ClC-3 directly, but Ca2+ can act indirectly by stimulating CaMKII, which in some cell types can activate ClC-38,9,26. However, other studies with smooth muscle cells from the pulmonary artery and the airways23,27–29 indicated that CaMKII does not stimulate ICl.Ca and instead inhibits it. We determined the impact of endogenous CaMKII upon ICl.Ca in aortic smooth muscle cells using KN-93, a specific, cell-permeant CaMKII inhibitor30. KN-93 reduced by about 40–50% the degree to which 0.5 μM [Ca2+]IN activated ICl.Ca in wild-type cells (Fig. 3A,S4). Significantly, KN-93 did not affect ICl.Ca in the ClC-3 null cells (Fig. 3B). Control experiments indicated there were similar degrees of CaMKII activity in the two cell types in (Fig. S5). Thus, in aortic VSMC, endogenous CaMKII mediates Ca2+-dependent activation of ICl.Ca through a ClC-3 dependent process (Fig. S3).
In certain cell types, the degree to which CaMKII activates ICl.Ca is supervised by Ins(3,4,5,6)P410,11,20; cellular levels of Ins(3,4,5,6)P4 are elevated downstream of the Ins(1,4,5)P3/Ca2+ cascade11. Ins(3,4,5,6)P4 does not inhibit CaMKII activity per se31 (as KN-93 does; see Fig. S3). Instead, Ins(3,4,5,6)P4 specifically prevents CaMKII from activating ICl.Ca11,20. This is a receptor-coupled mechanism by which Ins(3,4,5,6)P4 regulates salt and fluid secretion, insulin release and neuronal excitability11,20. A physiologically-relevant11 concentration of 10 μM Ins(3,4,5,6)P4 reduced the size of ICl.Ca by 40 – 50% (Fig 3A). The inhibitory effects of KN-93 and Ins(3,4,5,6)P4 upon ICl.Ca were not additive (Fig. 3A,S6), consistent with the end-point for both agents being inhibition of ClC-3 activity, albeit by different mechanisms (Fig. S3). Our data (Fig. 3A) represent the first demonstration that Ins(3,4,5,6)P4 regulates ICl.Ca in any VSMC.
Ins(3,4,5,6)P4 did not affect swelling-activated Cl− currents in VSMC (Fig. S1), which are not dependent upon ClC-3 (Fig. S1). Moreover, Ins(3,4,5,6)P4 did not alter ICl.Ca in the ClC-3 null cells (Fig 3B). These results provide important information concerning the specificity of Ins(3,4,5,6)P4 action; it has not previously been shown that ClC-3 is absolutely required for Ins(3,4,5,6)P4 to regulate ICl.Ca. Overall, our data demonstrate that ClC-3 provides VSMC with a specific signaling mechanism by which Ins(3,4,5,6)P4 and CaMKII can regulate one component of ICl.Ca (Fig. S3).
CaMKII is known to regulate VSMC motility3. We next investigated if the regulation of migration by CaMKII requires ClC-3. We again used serum as a pro-migratory stimulus; serum drives VSMC migration in vitro by elevating intracellular [Ca2+] and activating CaMKII3,14–16,32. When we added increasing concentrations of KN-93 (10–100 μM) to inhibit CaMKII, we found dose-dependent inhibition of migration of wild-type cells (maximal effect = 45%; Fig. 4A); KN-93 did not alter the number or the viability of wild-type cells at the highest dose used (Fig. 4 legend). Significantly, 10–50 μM KN-93 did not affect migration of ClC-3 null cells (Fig. 4B). These data substantiate our conclusion that CaMKII regulates cell migration through ClC-3 (Fig. S3).
At its maximal does of 100 μM, KN-93 slightly inhibited the migration of ClC-3 null cells (Fig. 4B), an effect that was possibly related to the 22% decrease in cell number at this concentration (Fig. 4 legend). We next added Ins(3,4,5,6)P4, which does not inhibit CaMKII per se, but instead selectively prevents the activation of ClC-3 by CaMKII10,20. 10 μM Ins(3,4,5,6)P4 inhibited migration of wild-type cells by 48% (Fig. 4C), without affecting cell number or viability (legend to Fig. 4). Ins(3,4,5,6)P4 did not inhibit migration in ClC-3 null cells (Fig. 4D), further supporting our conclusion that ClC-3 regulates VSMC migration.
Our study makes a number of new observations concerning the importance of ClC-3 and Ins(3,4,5,6)P4 for plasmalemmal Cl− flux (ICl.Ca) and cell motility in VSMC. Our results provide the first demonstration, in any cell type, that the size of ICl.Ca is substantially decreased following disruption of the ClC-3 gene (Fig. 2). It is also a novel finding that endogenous CaMKII acts through ClC-3 to stimulate ICl.Ca in a smooth muscle cell line (Fig. S3). This new information concerning ClC-3 is important because spatio-temporal control of ion channels and transporters is key to the process of cell migration4,5. Indeed, we report in this study that ClC-3 contributes to VSMC migration in a CaMKII-dependent manner (Figs. 3,,4).4). Our study also adds to the known repertoire of biological functions of Ins(3,4,5,6)P410 by showing that it regulates VSMC migration.
Previous work21 with parotid acinar cells indicated that ICl.Ca was not affected by elimination of ClC-3. That result has been frequently interpreted as an indication that ClC-3 does not contribute to ICl.Ca in any cell33–37. That assumption is now proved to be incorrect by our new data (Fig. 2). Clearly, the regulation of ICl.Ca in aortic smooth muscle cells differs from that in parotid acinar cells, and we believe that CaMKII is the distinguishing factor: in our aortic smooth muscle cells, ClC-3 dependent activation of ICl.Ca requires CaMKII (Fig. 3). In parotid acinar cells, CaMKII does not activate ICl.Ca38. These differences between parotid cells and VSMCs in the role of CaMKII can explain why ICl.Ca is differentially impacted by the elimination of ClC-3. Moreover, the activation of ICl.Ca by CaMKII in aortic smooth muscle cells also uncovers a new aspect of signaling specialization in vascular biology. Previous work23,27,29 has shown that CaMKII inhibits ICl.Ca in smooth muscle cells from the pulmonary artery.
There is a long-standing debate as to whether ClC-3 is either directly responsible for plasmalemmal ICl.Ca26 or instead indirectly regulates Cl− flux34. Elsewhere39 it has been explained why it is difficult to unequivocally prove whether a protein is either directly responsible for an ion current, or instead contributes to a process that regulates another channel. It is beyond the scope of our study to resolve this complex question. Nevertheless, our results indicate that the halving of ICl.Ca upon disruption of the ClC-3 gene (Fig. 2) reflects the loss of a specific signaling pathway, namely, CaMKII-dependent ICl.Ca. Consistent with that conclusion, ClC-3 dependent ICl.Ca was blocked by addition of a CaMKII inhibitor, KN-93 (Fig. 3) or by Ins(3,4,5,6)P4, an intracellular signal that specifically attenuates ClC-3 activation by CaMKII (Fig. 3,S3).
Our study shows that total ICl.Ca in VSMC includes a component that is CaMKII- and ClC-3 dependent, and another component that is independent of both CaMKII and ClC- 3 (Fig. S3). The latter may reflect that portion of ICl.Ca that is directly activated by Ca2+ 35,40,41. Unlike Ca2+, CaMKII is particularly suited for regulating longer-term responses, as the kinase can remain active (“autonomous”) even in a post-stimulation context, at a time when [Ca2+]IN has returned to resting levels42. That CaMKII “memory” could explain why the CaMKII-dependent portion of ICl.Ca, which is mediated by ClC-3, is important to a sustained biological activity such as cell migration. Nevertheless, there may be additional targets by which CaMKII regulates migration3.
Finally, our data are very relevant to an earlier study showing that ClC-3 null mice are less susceptible to neointima formation following carotid ligation2. That phenotype was attributed to a reduced rate of TNF-α dependent proliferation of VSMC2. Our new data suggest that reduced VSMC motility may be an important contributing factor. Both migration and proliferation of VSMC contributes to vascular remodeling1, so our study provides new insight into the mechanisms underlying this condition.
The regulation of ClC-3 influences migration (see above) of VSMC, which is a causal factor for diseases of vascular remodeling, such as the formation of neointima2 and atherosclerosis1. ClC-3 may also drive glioblastoma pathophysiology8 and cardiac hypertrophy17. Thus, genetic and/or pharmacological intervention in cell-signaling by ClC-3 may ultimately benefit multiple aspects of human health. To this end, Ins(3,4,5,6)P4 specifically prevents the activation of ClC-3 by CaMKII (see above), without altering CaMKII activity per se20.Moreover, the inositol phosphate did not affect either IClswell (Fig S1) or the ClC-3 independent portion of ICl.Ca (Fig. 2). We suggest that Ins(3,4,5,6)P4 is a potential lead compound for synthesizing a drug that might inhibit vascular remodeling by specifically targeting ClC-3.
We describe novel, cell-signaling roles of IP4 and ClC-3 in VSMC migration, thereby revealing new therapeutic approaches to vascular remodeling diseases.
Sources of Funding: This work was supported by the Intramural Research Program of the NIH/National Institute of Environmental Health Sciences and an NIH RO1 grant to F.S.L (HL62483). S.B.G., S.-G.W., A.Z., and S.B.S. performed experiments. F.S.L. supplied the cells. All authors contributed to writing the manuscript.
Conflict of Interest: None