Atherosclerosis-related molecular alteration of the human Ca_V1.2 calcium channel α_1C subunit

Abstract

Atherosclerosis is an inflammatory process characterized by proliferation and dedifferentiation of vascular smooth muscle cells (VSMC). Ca_v1.2 calcium channels may have a role in atherosclerosis because they are essential for Ca²⁺-signal transduction in VSMC. The pore-forming Ca_v1.2α1 subunit of the channel is subject to alternative splicing. Here, we investigated whether the Ca_v1.2α1 splice variants are affected by atherosclerosis. VSMC were isolated by laser-capture microdissection from frozen sections of adjacent regions of arteries affected and not affected by atherosclerosis. In VSMC from nonatherosclerotic regions, RT-PCR analysis revealed an extended repertoire of Ca_v1.2α1 transcripts characterized by the presence of exons 21 and 41A. In VSMC affected by atherosclerosis, expression of the Ca_v1.2α1 transcript was reduced and the Ca_v1.2α1 splice variants were replaced with the unique exon-22 isoform lacking exon 41A. Molecular remodeling of the Ca_v1.2α1 subunits associated with atherosclerosis caused changes in electrophysiological properties of the channels, including the kinetics and voltage-dependence of inactivation, recovery from inactivation, and rundown of the Ca²⁺ current. Consistent with the pathophysiological state of VSMC in atherosclerosis, cell culture data pointed to a potentially important association of the exon-22 isoform of Ca_v1.2α1 with proliferation of VSMC. Our findings are consistent with a hypothesis that localized changes in cytokine expression generated by inflammation in atherosclerosis affect alternative splicing of the Ca_v1.2α1 gene in the human artery that causes molecular and electrophysiological remodeling of Ca_v1.2 calcium channels and possibly affects VSMC proliferation.

Atherosclerosis is considered an inflammatory process that causes endothelial perturbation; local release of cytokines; and dedifferentiation, proliferation, and migration of vascular smooth muscle cells (VSMC) (1). Arterial VSMC constitute the media of the artery and play a crucial role in its elasticity and contractility. Contraction of VSMC is triggered by Ca²⁺ current through the voltage-gated Ca_v1.2 channels that are targets of Ca²⁺-channel-blocking drugs (2, 3). The vasodilating effect of these drugs is associated with high affinity binding to the pore-forming α_1C subunit of the Ca_v1.2 channel (4) that in the case of dihydropyridines depends on membrane potential (5–8). The expression of Ca_v1.2 changes during cellular differentiation and proliferation, and is strongly affected by hormones and cytokines (9–11). The Ca_v1.2α1 subunit gene is subject to complex alternative splicing (12–20) that may change both pharmacological (21–23) and physiological properties (14, 21, 24, 25) of the channel.

Although splice variations in segments of the vascular Ca_v1.2α1 transcripts have been recently established (13, 14, 20), the relationship between distinct Ca_v1.2α1 splice isoforms and vascular disease has not yet been investigated. Our study of Ca_v1.2α1 splice variants in VSMC is the first attempt to identify changes in the human Ca²⁺ channel associated with atherosclerosis. Our findings revealed an extended repertoire of the exon-21 Ca_v1.2α1 splice isoforms in nonatherosclerotic VSMC and established a potentially important switch to a unique exon-22 isoform as a molecular signature of the electrophysiologically remodeled proliferating pathophysiological state of VSMC in atherosclerosis.

Results

Reduced Expression of Ca_v1.2α1 in Atherosclerotic Regions of Human Artery.

The Ca²⁺ current through the l-type Ca_v1.2 channels triggers contraction of VSMC (26). We sought to characterize the Ca_v1.2α1 transcripts in VSMC affected (VSMC^D, diseased) and not affected (VSMC^N, nondiseased) by atherosclerosis. VSMC were identified and isolated from the tissue by laser-capture microdissection (LCM). Arterial tissue obtained during vascular surgery procedures (three femoral and three carotid arteries; see Table 2, which is published as supporting information on the PNAS web site) was prepared in 5- to 7-μm, frozen sections from the regions of atherosclerotic plaques and adjacent control areas that had no evidence of atherosclerosis. Fig. 1 shows representative immunohistochemical patterns of the tissue sections used for LCM. VSMC were identified (Fig. 1 A and F) in frozen sections by immunostaining with antibody against SM α-actin (27), used as a marker for VSMC. The SM α-actin staining correlated with immunostaining by anti-α_1C antibody in serial sections (Fig. 1 B and G). Consistent with dedifferentiation of VSMC^D (28, 29) immunostaining against both the SM α-actin and α_1C was visually reduced in atherosclerotic regions (Fig. 1G). To quantify the Ca_v1.2 transcript, we identified 200–300 VSMC in atherosclerotic and unaffected regions of the artery by rapid SM α-actin immunostaining and then isolated them by LCM. RNA extracted from the cells was then analyzed by quantitative real-time PCR with SYBR green. By studying six different preparations, we determined that the relative α_1C mRNA level in VSMC (normalized to 18S RNA) was reduced 3.7 ± 0.9 fold (mean ± SEM) in the atherosclerotic region as compared with the adjacent nondiseased tissue (P < 0.02, paired t test). This result confirms that atherosclerosis causes reduction in expression of the Ca_v1.2 channels in VSMC.

Fig. 1.

Representative immunohistochemical patterns of the vascular preparations used for LCM of VSMC and isolation of RNA from atherosclerotic (D) and adjacent nonatherosclerotic (N) regions of artery. Shown are photomicrographs of immunohistochemical staining of VSMC in serial sections of the same biopsy of arteries with antibodies against smooth muscle (SM) α-actin (A and F) (for the individual patient data, see Fig. 6), Ca_v1.2α1 (B and G), ubiquitous human nuclear protein Ki-67 (C and H), PDGF-BB (D and I), and PDGF-β receptor (E and J). (Scale bar: 50 μm.)

Reduced Ca_v1.2 expression was previously observed in human fibroblasts in response to mitogenic stimulation (9). The reduction in expression of the vascular Ca_v1.2 channels may also be due to mitogenic factors of local inflammation in the atherosclerotic plaque region that induce migration and dedifferentiation of VSMC (1, 28, 29). The nonatherosclerotic artery contains very few proliferating cells (Fig. 1C) and is characterized by very limited presence of cytokines such as PDGF-BB (Fig. 1D) and its receptor PDGFR-β (Fig. 1E). By contrast, in atherosclerotic regions there were much larger numbers of proliferating cells (Fig. 1H) that occur on conjunction with elevated PDGF-BB (Fig. 1I) and PDGFR-β (Fig. 1J). Consistent with effects of atherogenesis (29), the shape and arrangement of VSMC were changed (compare Fig. 1 A and F). Taken together, these data characterize quantitative differences between the unaffected and diseased regions of the arterial biopsies used in this work.

Molecular Remodeling of Ca_v1.2α1 Associated with Atherosclerosis.

To determine whether the altered expression of Ca_v1.2α1 in VSMC in atherosclerosis involves specific splice isoform(s), identity of the alternative exons of the Ca_v1.2α1 subunit (Fig. 2A) was determined by RT-PCR. Arterial preparations from the six patients (Table 2) were investigated to substantiate the statistical significance of the results. Each of them (see examples in Fig. 7, which is published as supporting information on the PNAS web site) showed the same pattern of alternative splicing of Ca_v1.2α1 in VSMC^N and its change in VSMC^D is summarized in Fig. 2 B–G. With a primer complementary to conserved exon 3, sense primers P283 and 22≫38 (see Fig. 2 legend) generate PCR products of 541 and 419 bp only when exon 1a and exon 1, respectively, are present. By using this assay, exon 1a was identified in the control human cardiac mRNA (mRNA^Card) (Fig. 2B, lane 1), but it was absent from the VSMC α_1C transcripts (Fig. 2B, lanes 3 and 5). However, exon 1 was found in both mRNA^Card (lane 2) and RNA isolated from VSMC^N (lane 4) and VSMC^D (Fig. 2B, lane 6,). Thus, the N terminus of the human vascular Ca_v1.2α1 is encoded by exon 1 and does not change in atherosclerosis.

Fig. 2.

Identification of the Ca_v1.2α1 splice variants. (A) Hypothetical transmembrane topology of Ca_v1.2α1 (17, 18). Outlined are four internal repeats, I–IV, each composed of six transmembrane segments, S1–S6. Protein segments encoded by numbered exons are marked by bold lines. Arrows point to alternative exons (8/8A, 21/22, and 31/32) that are subject to mutually exclusive splicing. Constitutively spliced exons 1/1A, 9A, 33, 34A, 41A, and 45 are shown by white boxes. (B–G) Identification of alternative exons of the Ca_v1.2α1 transcript. (B) Exon 1a (lanes 1, 3, and 5) and exon 1 (lanes 2, 4, and 6). P283 is the exon 1a-specific primer 5′-tggatccgccaATGCTTCGAGCCTTTGTTCAGC-3′. (C) Exons 8A and 8. (D) Exon 9a and 9. (E) Exons 21 and 22. (F) Exons 31–34. Shown are RT-PCR products (lanes 1 and 2) and their analytical digestion with NsiI (lanes 3 and 4) and PvuII (lanes 5 and 6). Splice variants identified by numbers on the left side of the gel photograph correspond to α_1C,127 (1), α_1C,73 (2), α_1C,125 (3), α_1C,126 (4), α_1C,71 (5) and α_1C,77 (6) (for details, see Fig. 10). (G) Differential utilization of exon 41A and lack of alternative exons 40B, 43A, 44A, and 45. Schematic diagrams illustrate the arrangement of alternative exons (black boxes) in RT-PCR products amplified from human mRNA^Card (lanes C) and RNA extracted from VSMC^N (lanes N) and VSMC^D (lanes D), Exons (boxes) are numbered as in A. The missing exons are shown as gray boxes. Numerals separated by ≫ or ≪ indicate the sense and antisense amplification primers, respectively, defined by nucleotide positions relative to the ORF of pHLCC71. To the right of schematics are RT-PCR products identified on agarose gels and their size in base pairs (arrows).

To discriminate between exons 8A and 8, RT-PCR products obtained with indicated primers (Fig. 2C) were digested with BamHI. Exon 8A has a unique BamHI restriction site that yields the 478- and 207-bp fragments from the 685-bp RT-PCR product. Both the exon 8 and exon 8a species of the Ca_v1.2α1 transcripts were positively identified in mRNA^Card (Fig. 2C, lane C), whereas only the BamHI-resistant exon 8-isoform of the α_1C transcripts was observed in VSMC (Fig. 2C, lanes N and D). Previously it was reported that vascular Ca_v1.2α1 incorporates the 75-bp combinatorial exon 9a between exons 9 and 10 (13, 14, 30, 31). We identified exon 9a in at least three VSMC^N Ca_v1.2α1 subunit transcripts (Fig. 8, which is published as supporting information on the PNAS web site) corresponding to the 1,170- and 1,095-bp PCR products (Fig. 2D, lane N). However, only the 1,095-bp DNA was amplified from VSMC^D, indicating that exon 9a was absent from the VSMC^D Ca_v1.2α1 (Fig. 2D, lane D).

The identity of alternative exons 21/22 was established by analytical AvrII digestion (Fig. 2E). Exon 22 has an AvrII restriction site that yields the 459- and 142-bp fragments from the 601-bp PCR product. By using AvrII digestion, the exon 22-isoform of α_1C was positively identified only in VSMC^D (Fig. 2E, lane D right). In contrast, VSMC^N (Fig. 2E, lane N) predominantly express the exon 21-isoform of the Ca_v1.2α1 transcript that is resistant to AvrII. Direct DNA sequencing of the crude PCR amplification products independently confirmed that result (Fig. 9, which is published as supporting information on the PNAS web site). Indeed, no distortion of the nucleotide peaks in the region of exon 21/22 was seen when compared with the exon-20 invariant region, suggesting that a switch to the exon-22 isoform of the vascular Ca_v1.2α1 subunit was almost complete in atherosclerosis.

RT-PCR applied to the region of exons 30–34 generated several amplification products identified by analytical restriction digestion and direct sequencing (Fig. 2F). Analytical digestion by NsiI (Fig. 2F, lanes 3 and 4) and PvuII (Fig. 2F, lanes 5 and 6) confirmed the structural assessment summarized in Fig. 3(boxed sequences). Only the exon-32 isoform of Ca_v1.2α1, digested by NsiI into the 353- and 223-bp fragments, was identified in VSMC^D (α_1C,77; Fig. 3). In contrast, four exon-32 isoforms of Ca_v1.2α1 (α_1C,71, α_1C,73, α_1C,125, and α_1C,126), and the exon-31 isoform resistant to NsiI (α_1C,127) were identified in VSMC^N. Analytical digestion by PvuII (Fig. 2F, lane 5) revealed heterogeneity of Ca_v1.2α1 in VSMC^N due to exon 33. Exon 33 was found to be missing in two of five Ca_v1.2α1 transcripts in VSMC^N (α_1C,73 and α_1C,125 in Fig. 2F). Additional heterogeneity in this regions of the structure was due to a 6-nt deletion (14, 20) in exon 32 (α_1C,125) and upstream extension of exon 34 (α_1C,126) in vascular Ca_v1.2α1 splice isoforms.

Fig. 3.

Distribution of alternative exons in transcripts of the Ca_v1.2α1 splice isoforms identified in VSMC^N and VSMC^D. Amino acid sequences encoded in alternative exons 31–34 are shown (boxes) beneath the chart. The α_1C subunit isoforms indicated on the left correspond to electrophysiologically characterized variants lacking exon 9a.

Finally, RT-PCR from the region of exons 40–46 revealed single 858- and 801-bp products of amplification in VSMC^N and VSMC^D RNA, respectively (Fig. 2G). DNA sequencing showed that the difference is due to the 57-bp combinatorial exon 41A that was absent from the VSMC^D RNA. However, alternative exons 40B, 43A, and 45 (18) and the cardiac 213-bp exon 44A (16) are absent from the vascular Ca_v1.2α1 transcripts. No additional heterogeneity was observed from the distal 3′-terminal region (data not shown).

Fig. 3 summarizes the results of the assessment of alternative splicing of the Ca_v1.2α1 transcripts. In VSMC^N, we identified an extended repertoire of Ca_v1.2α1 isoforms, all characterized by the presence of exons 21 and 41A. This selective heterogeneity of the Ca_v1.2α1 subunits was replaced in VSMC^D by the single exon-22 variant lacking exon 41A.

Electrophysiological Remodeling of Ca_v1.2 Calcium Channels in Atherosclerosis.

To determine whether molecular remodeling of Ca_v1.2α1 in atherosclerosis causes changes in electrophysiological properties of the VSMC Ca_v1.2 calcium channel, α_1C,77 of VSMC^D was compared with its splice variants in VSMC^N. Our goal here was to determine relative effects of the identified splicing variations on kinetics of inactivation, the I–V relationship, steady-state inactivation, recovery from inactivation and run-down of the channels with Ba²⁺ and Ca²⁺ as charge carriers. Because we have not established yet whether splice variation of β subunits is affected by atherosclerosis, for comparative analytical measurements all channels were expressed with the β_1a accessory subunit that provides for intermediate kinetics of inactivation (24). In a number of previous studies (14, 30, 31), effects of exon 9a incorporation on the channel properties have been characterized as marginal. Therefore, to simplify the interpretation of the results, we compared α_1C,77 with the exon 9a-deficient Ca_v1.2α1 splice variants α_1C,127, α_1C,73, α_1C,125, α_1C,126, and α_1C,71 detected in VSMC^N (Fig. 3). Major results are summarized in Table 1; see also Figs. 10 and 11, which are published as supporting information on the PNAS web site. All tested channels generated a μA maximum I_Ba and I_Ca that was sufficiently large to disregard contribution from endogenous channels (50–80 nA). Analysis of I_Ba revealed that kinetics of voltage-dependent inactivation is changed only slightly between the α_1C,77 channel and splice variants of VSMC^N. The I–V relationships for all isoforms (Table 3, which is published as supporting information on the PNAS web site) show peak I_Ba at +10 mV. However, voltage-dependence of activation and inactivation are both significantly (P < 0.05) different in α_1C,77, causing a notable change in the slope of the activation curve and a shift of the steady-state inactivation curve to more positive voltages. Confirming previous observations (24), these data indicate that sensitivity of the α_1C,77 channel to voltage gating is altered as compared with the tested VSMC^N channels.

Table 1.

Comparison of electrophysiological properties of the Ca_v1.2α1 Ca²⁺ channel splice variants identified in VSMC^N (α_1C,127, α_1C,73, α_1C,125, α_1C,126, and α_1C,71) and VSMC^D (α_1C,77) in dependence of β subunits and concentration of charge carriers

α1C isoform	Kinetics of inactivation				Activation		Steady-state inactivation
	Imax, μA (n)	Io, %	If, %	τf, ms	Va,0.5, mV	ka (n)	a, %	Vi,0.5, mV	ki (n)
α1C/β1a/α2δ-1, 40 mM Ba2+
α1C,127	−2.70 ± 0.35 (21)	20.0 ± 1.3†	43.6 ± 2.5	89.7 ± 7.3†	0.3 ± 1.8	9.2 ± 0.5* (9)	20.9 ± 3.8*	−24.4 ± 2.7*	12.4 ± 1.0* (8)
α1C,73	−4.36 ± 0.59 (19)	16.6 ± 1.5	49.0 ± 3.7	114.9 ± 10.7	1.8 ± 3.2	8.8 ± 0.6* (8)	12.4 ± 5.8	−17.2 ± 4.4	8.1 ± 1.1 (7)
α1C,125	−3.34 ± 0.31 (23)	11.6 ± 0.9*	51.8 ± 3.6	134.2 ± 10.9	−5.3 ± 1.4	9.5 ± 1.0* (8)	5.7 ± 1.4§	−21.4 ± 1.6*	12.1 ± 1.2* (8)
α1C,126	−3.21 ± 0.52 (18)	14.9 ± 0.9§	38.5 ± 3.3	133.6 ± 9.1	−1.2 ± 3.0	8.2 ± 0.9* (8)	15.1 ± 1.3*	−11.7 ± 2.4	8.8 ± 1.7 (5)
α1C,71	−4.07 ± 0.43 (26)	16.1 ± 1.5	43.5 ± 3.7	121.1 ± 8.6	0.5 ± 1.7	7.5 ± 0.6* (9)	9.1 ± 4.5	−22.9 ± 1.7*	8.2 ± 1.0 (6)
α1C,77	−4.04 ± 0.68 (38)	18.6 ± 1.1	44.0 ± 2.8	105.7 ± 8.4	0.6 ± 1.0	4.6 ± 0.3 (17)	7.6 ± 1.6	−8.8 ± 0.7	7.8 ± 1.0 (16)
α1C/β1a/α2δ-1, 40 mM Ca2+
α1C,127	−1.87 ± 0.20 (26)	4.7 ± 0.9	89.2 ± 0.5*	30.1 ± 0.9*	11.4 ± 1.7	12.2 ± 0.7* (6)	1.6 ± 1.1	−24.3 ± 1.3	14.1 ± 0.8* (5)
α1C,73	−2.84 ± 0.52 (15)	0.5 ± 1.8*†	97.0 ± 0.4†‡	24.4 ± 1.2†‡	7.3 ± 3.7	9.7 ± 0.4 (5)	1.2 ± 0.7	−23.9 ± 3.6	10.8 ± 0.9 (5)
α1C,125	−3.47 ± 0.41 (17)	5.4 ± 0.6	88.7 ± 0.5*	36.6 ± 2.4*	3.9 ± 1.7	8.4 ± 0.4 (13)	0.3 ± 0.3	−21.8 ± 3.1	10.9 ± 0.9 (7)
α1C,126	−3.22 ± 0.59 (13)	3.2 ± 0.5†	89.1 ± 0.9*	34.1 ± 1.9*	3.5 ± 1.7	9.0 ± 0.2 (5)	1.0 ± 0.5	−18.4 ± 1.4	8.7 ± 0.5 (5)
α1C,71	−3.69 ± 0.40 (20)	1.6 ± 1.0†	93.5 ± 1.0‡	31.4 ± 2.3*	3.3 ± 1.7	8.6 ± 0.5 (9)	1.8 ± 1.3	−17.7 ± 2.8	8.7 ± 0.2 (5)
α1C,77	−3.52 ± 0.80 (28)	4.3 ± 0.6	93.5 ± 0.5	24.6 ± 1.2	9.6 ± 1.7	10.0 ± 0.7 (10)	1.3 ± 0.8	−16.3 ± 2.4	9.0 ± 1.3 (8)
α1C/β2a/α2δ-1, 2.5 mM Ca2+
α1C,127	−0.46 ± 0.01 (6)	9.6 ± 3.2	80.5 ± 1.2	38.4 ± 1.0	24.2 ± 4.9*	10.9 ± 0.4* (3)	0.1 ± 0.2	−15.4 ± 4.2	14.0 ± 1.7* (3)
α1C,77	−1.42 ± 0.38 (10)	13.3 ± 0.2	73.8 ± 2.4	36.0 ± 3.0	0.9 ± 1.8	8.6 ± 0.2 (4)	0.7 ± 0.3	−14.9 ± 5.6	11.8 ± 1.8 (3)

Values are reported as means ± SEM. Differences were tested for by ANOVA. I_max, maximum amplitude of the current; I and I, sustained and fast components of the current; V_a,0.5, midpoint potential of activation; V_i,0.5, voltage at half-maximum of inactivation; a, fraction of noninactivating component of the current; k_a and k_i, slope factors; n, number of tested oocytes.

*, P < 0.05 by Dunnett's test using α_1C,77 as control. For all pairs comparisons, Tukey's test was used:

^†, P < 0.05 vs. α_1C,125;

^‡, P < 0.05 vs. α_1C,126;

^§, P < 0.05 vs. α_1C,127.

Replacement of Ca²⁺ for Ba²⁺ as the charge carrier evoked Ca²⁺-dependent inactivation that accelerated the I_Ca decay in all tested channels (Table 1). Unlike I_Ba, Ca²⁺ currents reached maximum at approximately +20 mV and exhibited U-shaped dependence of τ_f on membrane potential (Fig. 11B) characteristic for Ca²⁺-dependent inactivation (24, 32) that becomes faster with larger current. Although with the β_1a subunit (40 mM Ca²⁺) kinetics of inactivation of I_Ca varies between the different channel isoforms expressed in VSMC^N and compared with the α_1C,77 channel, no significant difference was observed between the α_1C,77 and α_1C,127 channels with the primary cardiac β_2a subunit (2.5 mM Ca²⁺; Fig. 4A). Another interesting finding is that two of the tested VSMC^N channels, α_1C,125 and α_1C,127 with the β_1a subunit, showed a significant (P < 0.05) sustained component of I_Ca that comprised 5.4% and 4.7%, respectively, of the total I_Ca by the end of a 1-s test pulse (Table 1). Contrary to the Ba²⁺ current data, significant changes of the voltage-dependence of the I_Ca activation and inactivation of the α_1C,77 channel were found only with α_1C,127 and observed with both β_1a and β_2a subunits. With 2.5 mM Ca²⁺ in the bath solution, this difference was even greater, and the channel assembled of the α_1C,77 and β_2a subunits showed a 15-mV shift of the I–V relation (Fig. 4B) and the activation curve (Fig. 4C) by ≈15 mV in the hyperpolarizing direction.

Fig. 4.

Comparison of electrophysiological properties of I_Ca through the α_1C,77 and α_1C,127 channels coexpressed with the primary cardiac β_2a subunit and measured with 2.5 mM Ca²⁺ in the bath solution. (A) Representative traces of I_Ca evoked by 1-s step depolarizations to +20 mV from V_h = −90 mV and normalized to the same amplitude. (B) Averaged I–V curves (filled circles) and voltage-dependences of the time constant of fast inactivation, τ_f, (open circles) for I_Ca. A 1-s test pulse in the range of −40 to +50 mV (10-mV increments) was applied from V_h = −90 mV with 30-s intervals. (C and D) Ensembles of activation (G/G_max − V) curves (C) and steady-state inactivation curves (D) fit by Boltzmann function. (E) Fractional recovery of I_Ca from inactivation. (F) Run-down of I_Ca. Step depolarizations of 250 ms to +20 mV were applied from V_h = −90 mV every 30 s, and the maximum amplitude of the current was normalized to the initial value. ∗, P < 0.05; SD with α_1C,77 by ANOVA with Tukey's test.

Recovery of the current from inactivation was measured with +10 mV (Ba²⁺) or +20 mV (Ca²⁺) prepulses of a 1-s duration followed by increasing time intervals (25 ms to 1 s) at −90 mV before 0.25-s test pulses to +10 mV (Ba²⁺) or +20 mV (Ca²⁺) were applied. The ratio of the maximum current, evoked in response to a prepulse, to that of the test pulse was calculated as a fraction of the current. All tested channels showed slower recovery from inactivation as compared with Ca_v1.2α1 mutants deprived of Ca²⁺-dependent inactivation (24). The Ca²⁺-conducting α_1C,77 channel recovered from inactivation significantly (P < 0.05) faster than α_1C,127 (Fig. 4E), but no significant difference was observed between the tested Ba²⁺-conducting channels (Fig. 10E). Finally, all tested channels showed typical run-down (10% of I_max in 4 min) except for α_1C,127 [Figs. 4F (Ca²⁺) and 10F (Ba²⁺)]. Taken together, these data revealed that a number of important electrophysiological properties of the α_1C,77 channel in VSMC^D are changed from those of the VSMC^N channel isoforms.

Possible Association of the Exon-22 Isoform of Ca_v1.2α1 and Proliferation of VSMC.

It is known that in response to locally elaborated cytokines, VSMC^D assume a proliferative phenotype and migrate in the atherosclerotic plaque area. To determine whether VSMC proliferation may be associated with reduced expression and the isoform switch of the Ca_v1.2α1 subunit, we examined effects of serum deprivation of human coronary artery SM cells in culture on alternative splicing of exons 21/22 and level of α_1C transcripts. Identity of the cells was established by the manufacturer (Clonetics, San Diego, CA) by positive staining for smooth muscle α-actin and negative staining for von Willebrand's factor VII. To halt cell proliferation, VSMC were grown to a confluent monolayer, and then the confluent cell culture was subjected to serum deprivation for 4 days. Under these conditions, DNA biosynthesis (assessed as [³H]thymidine incorporation) was decreased by 92%, the level of the Ca_v1.2α1 transcript increased 2.8 ± 0.12 fold (n = 3, 27- and 32-year-old male donors), and the AvrII-sensitive exon-22 isoform of the Ca_v1.2α1 transcript was not detected (Fig. 5, lane SF). However, when proliferation of nonconfluent VSMC was stimulated by the presence of 5% serum, DNA biosynthesis was restored and the presence of the exon-22 isoform of Ca_v1.2α1 was detected (Fig. 5, lane S). Although the isoform switch was not complete as in atherosclerosis, the cell culture results suggest that in a very different experimental system there is an association between proliferation of VSMC, decreased expression of the Ca_v1.2α1 transcript and appearance of the exon-22 isoform.

Fig. 5.

Evidence that the exon-21 isoform of Ca_v1.2α1 is not expressed in proliferating human arterial smooth muscle cells. Primary smooth muscle cells were grown in sparse culture in 5% serum (S) before serum-deprivation for 48 h (SF). Total RNA was isolated and analyzed by RT-PCR and subsequent AvrII restriction analysis as described in Fig. 2E. Shown are gels before (Left) and after (Right) AvrII overdigestion.

Discussion

This comprehensive study characterizes naturally occurring Ca_v1.2α1 splice variants in human arterial VSMC and their remodeling in atherosclerosis. The conclusions are that VSMC in the regions of atherosclerotic plaques as compared with cells from control nonaffected portions of the same artery have reduced expression of the Ca_v1.2α1 transcript. This is accompanied by replacement of multiple exon-21 isoforms of the Ca_v1.2α1 subunit with the single exon-22 isoform. That isoform exhibits altered electrophysiological properties and shows a possible association with VSMC proliferation.

Interestingly, two of the five investigated Ca_v1.2α1 isoforms of VSMC^N exhibited important functional differences from other ones: the α_1C,125 channel showed a significant residual component of I_Ca by the end of the 1-s depolarization (Table 1), whereas α_1C,127 exhibited slower recovery of I_Ca from inactivation and lack of run-down of both I_Ba and I_Ca (Figs. 10 and 11). These findings raise an interesting possibility that some Ca_v1.2 channel isoforms are less sensitive to Ca²⁺-induced inactivation that controls both the slow inactivation (33) and run-down of the channel (34). Because α_1C,125 and α_1C,127 are able to maintain a more sustained Ca²⁺ flux, they may have a specific role in VSMC functions that require prolonged Ca²⁺ influx, such as the maintenance of vascular tone and elasticity of arterial walls (35).

Atherosclerosis causes electrophysiological remodeling of VSMC through a replacement of multiple Ca_v1.2α1 variants situated in VSMC^N with the structurally different isoform α_1C,77 characterized by the presence of exon 22 in place of exon 21 coding for a portion of the transmembrane segment IIIS2. Careful electrophysiological analysis revealed a number of differences in the properties of the α_1C,77 channel as compared with the VSMC^N isoforms (Table 1). Our finding that I_Ca through the α_1C,77 channel recovers from inactivation significantly faster than α_1C,127 (Fig. 4E) and several other Ca_v1.2α1 isoforms in VSMC^N (Fig. 10F) suggests that alternative splicing in atherosclerosis may increase the Ca²⁺ current density in VSMC^D and affect regulation of the contractile vascular tone. A 15-mV shift of the α_1C,77/β_2a channel activation curve to more negative potentials (Fig. 4C) may also contribute to the increase of Ca²⁺ entry in VSMC^D, but it is not clear whether these electrophysiological properties would compensate and outmatch the reduced Ca_v1.2α1 expression.

Atherosclerosis is characterized by proliferation and migration of VSMC. Our results obtained with smooth muscle cells in culture (Fig. 5) indicated that inhibition of cell proliferation by serum deprivation in confluent monolayer completely eliminated the exon-22 isoform of the Ca_v1.2α1 transcript, which was reversible on stimulation of proliferation by addition of serum to nonconfluent cells. Cellular mechanisms leading to these changes may be very complex, but the association with cell proliferation is obvious. This association is consistent with and supported by earlier observations that in normal human fibroblasts, serum deprivation induced an increase, whereas mitogens (basic fibroblast growth factor, EGF, and insulin), second messengers (cAMP and Ca²⁺) or cell–cell contact inhibition caused strong reduction in the expression of the l-type Ca²⁺ channels (9). PDGF-BB is a mitogenic factor known to be a stimulus for neointimal proliferation and migration of VSMC in atherosclerosis (36, 37). Here we have shown increased expression of both PDGF-BB and its receptor in VSMC^D. Taken together, these data are consistent with the hypothesis that PDGF-BB, and perhaps other locally elaborated mitogens, are involved in the Ca_v1.2 α_1C isoform switch in atherosclerotic VSMC, resulting in the expression of the proliferation-specific exon-22 α_1C,77 channel.

Earlier studies have shown that exons 21 and 22 have different impact on voltage-dependent inhibition of the Ca_v1.2 channel by dihydropyridine Ca²⁺ channel blockers (21, 23) and the exon-22 α_1C,77 channel is more sensitive to dihydropyridines at negative potentials than the respective exon-21-isoforms. Thus, atherosclerosis-induced replacement of the exon-21 Ca_v1.2α1 isoforms for α_1C,77 should change response of VSMC^D to dihydropyridines very locally in the regions of the disease. This raises an intriguing possibility that reduced expression of the α_1C transcript and the isoform switch in the region of the atherosclerotic plaque both may lead to local heterogeneity in the VSMC response to Ca²⁺ channel blockers.

Our findings and other recent data (13–15, 20) based on the analysis of transcripts have a number of potential limitations. First, the physiological role of Ca_v1.2 variability in the maintenance of VSMC is unknown. It remains to be studied whether the orderly diversity of Ca_v1.2α1 in VSMC^N that utilizes 8 of 16 potential alternative exons (Fig. 3) is related to coordinated association with other subunits, specific subcellular distribution of the channels, and/or segregation of channels into large clusters. Investigation at the protein level may be helpful here. Second, electrophysiological experiments with β_1a (Fig. 10) and β_2a subunits (Fig. 4) showed that variation of β subunits may be another important determinant of the Ca_v1.2 remodeling in VSMC^D. Third, our results clearly showed that in VSMC^D the array of Ca_v1.2α1 isoforms is replaced by the single exon-22 variant, but we do not know yet whether this isoform is pathogenic and whether the small nuclear RNA-mediated skipping of exon 22 would rescue VSMC from atherosclerosis. These issues require further investigation.

In conclusion, our study revealed the exon-22 isoform α_1C,77 as a molecular signature of the electrophysiologically remodeled pathophysiological state of VSMC in atherosclerosis. Significantly reduced expression of the α_1C transcript in VSMC^D is another characteristic feature. Given that Ca_v1.2 channels are involved in Ca²⁺ signal transduction and transcription regulation (38), the isoform switch in Ca_v1.2 may occur as a transcriptional response to specific changes in local milieu, cytokine expression, and other factors of inflammation associated with atherosclerosis.

Materials and Methods

Arterial Tissue.

Arterial biopsy samples (three carotid and three femoral) were obtained at surgery from six patients of a mean age of 65 years (Table 2). After having received preoperative written informed consent for the protocol approved by the Western Institutional Review Board (Olympia, WA), vascular tissue was obtained intraoperatively during surgical procedures for atherosclerotic vascular disease. Small areas not visually affected by atherosclerosis and those occluded with heavy plaque burden were identified and dissected. Tissue was washed in 4°C saline, immersed in OCT compound (10.24% polyvinyl alcohol/4.26% polyethylene glycol) (Electron Microscopy Sciences, Hatfield, PA) and frozen in liquid nitrogen.

LCM and Isolation of mRNA.

Serial cryostat sections (5–7 μm thick) were cut from the frozen tissue samples. Before LCM, VSMC in the sections were fixed in acetone and quickly immunostained with anti-smooth muscle α-actin monoclonal IgG_2a (N1584, DAKO, Glostrup, Denmark). LCM was performed with PixCell II system (Arcturus, Sunnyvale, CA) by using a 7.5-μm laser spot. The excised VSMC were captured on LCM Caps. RNA was extracted from the collected cells by using PicoPure RNA isolation kit (Arcturus) and treated with RNase-free DNase (Qiagen, Valencia, CA).

PCR.

RT-PCR was carried out with a RETROscript kit (Ambion, Austin, TX) and RNA isolated from 200–300 microdissected cells. To increase specificity of PCR, a second round with nested primers was used. The identified Ca_v1.2α1 variants were subcloned into the Melton's vector. All nucleotide sequences were verified by DNA sequencing.

Real-time PCR was carried out in a GeneAmp 5700 Sequence Detection System (Applied Biosystems, Foster City, CA) by using the comparative threshold cycle method to determine the Ca_v1.2α1 expression in VSMC^N and VSMC^D of the same patient (Supporting Materials and Methods, which is published as supporting information on the PNAS web site).

Electrophysiology.

Ba²⁺ and Ca²⁺ currents were recorded in 40 mM Ba²⁺ or Ca²⁺ at 20–22°C by a two-electrode voltage clamp in Xenopus oocytes 3 days after microinjection with a mixture of mRNAs coding for a Ca_v1.2α1 variant and auxiliary β_1a and α₂δ-1 subunits (1:1:1, mol/mol). Additional experiments with selected channel isoforms (α_1C,77 and α_1C,127) that showed statistically significant differences in electrophysiological properties were carried out by using 2.5 mM Ca²⁺ close to physiological calcium concentration and the primary cardiac β_2a subunit (X64297). Inactivation time constants (τ) were determined from the double exponential fitting of the current decay. The activation and steady-state inactivation curves were fitted with Boltzmann equations. All fits were obtained with individual measurements and then averaged.

Statistical Analysis.

Results are presented as mean ± SEM. Differences of the measured electrophysiological parameters were tested for by ANOVA by using Tukey's test for all pair comparisons and by using Dunnett's test when data were compared with α_1C,77 as control. Probability values of P < 0.05 were considered to be statistically significant.

Further methodological details can be found in Supporting Materials and Methods.

Abbreviations

VSMC

arterial vascular smooth muscle cells

VSMC^D

diseased (atherosclerotic) VSMC

VSMC^N

nondiseased (nonatherosclerotic) VSMC

LCM

laser-capture microdissection

mRNA^Card

human cardiac mRNA.