Sinusoidal length oscillation- and receptor-mediated mRNA expression of myosin isoforms and α-SM actin in airway smooth muscle

Melissa Wahl, Thomas J. Eddinger, Chi-Ming Hai

Abstract

We tested the hypothesis that sinusoidal length oscillation and receptor activation interactively regulate the abundance of mRNA encoding α-smooth muscle (α-SM) actin and myosin isoforms in intact bovine tracheal smooth muscle. We found that sinusoidal length oscillation significantly downregulated abundance of mRNA encoding α-SM actin mRNA in unstimulated tissues but not in histamine- and carbachol-activated tissues. This observation suggests antagonistic interactions between mechanical stretch and receptor-mediated signal transduction in regulating the abundance of mRNA encoding α-SM actin in intact airway smooth muscle. This pattern of antagonistic interaction was also observed in cholinergic receptor activation experiments. Whereas carbachol significantly upregulated myosin heavy chain SMA isoform expression in muscle strips held at slack length, carbachol did not significantly alter SMA expression in muscle strips at sinusoidal length oscillation. Carbachol also significantly upregulated GAPDH expression in bovine tracheal smooth muscle. However, unlike SMA expression, upregulation of GAPDH expression mediated by cholinergic receptor activation appeared to be insensitive to the mechanical state of airway smooth muscle. Unlike carbachol, histamine did not significantly alter the expression of GAPDH, myosin heavy chain SMA and SMB, myosin light chain LC17a and LC17b, and α-SM actin in bovine tracheal smooth muscle. U0126 (10 μM) completely inhibited carbachol-induced ERK1/2 MAPK phosphorylation but did not significantly affect carbachol-induced upregulation of GAPDH and SMA expression, suggesting that the ERK1/2 MAPK pathway was not the underlying mechanism. A potential implication of these findings is that periodic stretching of airways during respiratory cycles may modulate mRNA expression by receptor agonists in airway smooth muscle cells in vivo.

  • asthma
  • carbachol
  • deep inspiration
  • gene expression
  • histamine

airway smooth muscle is the contractile component of airways. Abnormal contractility and proliferation of airway smooth muscle contribute to the elevated airway resistance in diseases such as asthma (1, 6, 9, 19). Airway smooth muscle cells function in an oscillatory mechanical environment during respiratory cycles. Mechanical deformation and receptor activation have been shown to induce proliferation and gene expression in cultured airway smooth muscle cells (34, 37, 38). However, it is not known to what extent mechanical deformation and receptor activation modulate gene expression in intact airway smooth muscle.

Actin and myosin are major contractile proteins in airway smooth muscle cells. α-Smooth muscle (α-SM) actin is the major actin isoform in differentiated smooth muscle cells in vivo (23). Smooth muscle myosin exists in multiple isoforms (11, 41). They include myosin heavy chain head isoforms (SMA, SMB), myosin heavy chain tail isoforms (SM1, SM2), and myosin light chain isoforms (LC17a, LC17b). Several laboratories have reported correlation between myosin heavy chain SMB isoform expression and myosin ATPase activity or shortening velocity in smooth muscle (3, 12, 22, 26, 40). Relative expression of the SM1 and SM2 myosin heavy chain isoforms has also been found to correlate with maximal shortening of single vascular smooth muscle cells (31). Similarly, several laboratories have reported correlation between LC17a and LC17b isoform expression and myosin ATPase activity or shortening velocity (29), but other laboratories did not observe such correlation (10). The regulation of α-SM actin and myosin mRNA transcription appears to involve complex combinations of multiple modules (23, 30). Such a complex system of gene transcriptional control may allow multiple stimulatory and inhibitory inputs into the regulation of α-SM actin and myosin mRNA transcription. In this study, we tested the hypothesis that sinusoidal length oscillation and receptor activation interactively regulate the abundance of mRNA encoding α-SM actin and myosin isoforms in intact bovine tracheal smooth muscle. ERK1/2 MAPK has been implicated in gene regulation and cell proliferation in airway smooth muscle (13, 33). U0126 inhibits MEK1/2, thereby inhibiting ERK1/2 phosphorylation and activation. Accordingly, we studied the effect of U0126 on receptor-mediated α-SM actin and myosin mRNA expression in bovine tracheal smooth muscle.

MATERIALS AND METHODS

Tissue preparation.

Bovine tracheas were collected from an abattoir and transported to the laboratory in ice-cold physiological salt solution (PSS) of the following composition (in mM): 140.1 NaCl, 4.7 KCl, 1.2 Na2HPO4, 2.0 MOPS (pH 7.4), 0.02 Na2EDTA, 1.2 MgSO4, 1.6 CaCl2, and 5.6 d-glucose. The smooth muscle layer together with adventitia and mucosa was excised from a tracheal segment and placed in a petri dish of cold PSS. The mucosal and adventitial layers were dissected away from the smooth muscle layer under a microscope. Smooth muscle strips, ∼4–5 mm in width, were prepared along the direction of muscle fibers. The ends of each muscle strip were tied with surgical silk. For length oscillation experiments, one end of each muscle strip was tied to a stainless steel hook that was connected to the arm of a lever (Aurora Scientific, Aurora, ON, Canada) for length oscillation and the other end was secured at a clamp connected to a length manipulator. A computer-controlled lever system was used to control muscle length and record force during length oscillation experiments.

Sinusoidal length oscillation.

The procedures for tissue equilibration and length manipulation were described previously (2, 18). Basically, a computer program controlled the sending of voltages to the length input port of the lever system at regular time intervals, thereby inducing sinusoidal oscillation of the lever arm and the muscle strip attached to the lever arm. For length oscillation experiments, muscle strips were first stretched to 12 g and then allowed to equilibrate for 1 h in PSS bubbled with air at 37°C. The muscle strips were then activated for 3 min with K-PSS, a solution similar to PSS in composition except that 104.95 mM NaCl was substituted by an equimolar concentration of KCl to check muscle viability. Muscle strips were then allowed to relax in PSS for 15 min and stretched to 12 g every 15 min during another hour of equilibration in PSS. Muscle strips were adjusted to optimal length (Lo) for contraction by releasing the muscle strips quickly to 2.5 g and then stimulated by K-PSS for 10 min. The active force developed in this contraction was recorded as Fo, which was used as an internal standard to normalize active force produced by carbachol and histamine in subsequent contractions. After the recording of Fo, muscle strips were allowed to relax in PSS for 15 min, and the muscle length was measured with a caliper (resolution = 0.1 mm). For slack length experiments, muscle strips were allowed to hang freely in a muscle chamber containing the solution to which the oscillating muscle strips were exposed.

Experimental protocols.

In PSS experiments, muscle strips were unstimulated and equilibrated in PSS at slack length or at sinusoidal length oscillation. In histamine experiments, muscle strips were stimulated by 100 μM histamine at slack length or at sinusoidal length oscillation. In carbachol experiments, muscle strips were stimulated by 1 μM carbachol at slack length or at sinusoidal length oscillation. For each protocol, muscle strips were allowed to equilibrate in the solution for 15 min before sinusoidal length oscillation. Using a computer-controlled Dual Mode Lever System (model 300 B; Aurora Scientific) and DMC/DMA Version 3.1 software, we oscillated muscle strips at a frequency of 1 Hz and an amplitude of 10% Lo. We recorded peak force during length oscillation for 20 cycles at the beginning and every hour of length oscillation. Solutions in muscle chambers were changed every hour.

RNA extraction and reverse transcription.

At the end of each experiment, muscle strips were weighed and then homogenized in 500 μl of lysis solution with 1% mercaptoethanol (GenElute Mammalian Total RNA Kit; Sigma). Tissue weights ranged from 30 to 50 mg. Tissue homogenate was then centrifuged at 16,000 rpm for 1 min at 20°C. The supernatant was collected for RNA extraction with the minicolumns according to the manufacturer's instructions. The reverse transcription (RT) procedure was similar to that described by Eddinger et al. (10). The extracted RNA was first mixed with 20 μg/ml oligo(dT) (Promega) and 500 μM dNTP (Promega) and heated at 68°C for 5 min. The mixture was rapidly cooled in wet ice for 20 s and then placed on a heating block set at 42°C. The following chemicals were then added to the tube for first-strand synthesis of cDNA: first-strand buffer (Invitrogen), 2 U/ml RNasin (Promega), 0.1 μg/μl acetylated BSA (Promega), 1.25 mM DTT (Promega), and 10 U/μl Superscript II RNase H-reverse transcriptase (Invitrogen). The RT reaction was carried out at the manufacturer's recommended temperature of 42°C for 2 h.

Polymerase chain reaction and DNA chip analysis.

Using aliquots of the RT product from each muscle sample, we set up four separate polymerase chain reaction (PCR) tubes to amplify the expression of GAPDH, myosin heavy chain isoforms SMA and SMB, myosin light chain isoforms LC17a and LC17b, and α-SM actin. The primer sequences for GAPDH and α-SM actin were designed with GeneFisher software (15) based on the published mRNA sequences for bovine GAPDH and α-SM actin in the National Center for Biotechnology Information Entrez Nucleotides database (accession numbers U85042 and BM431010, respectively). The forward primer sequence (5′ to 3′) for GAPDH was CTG GGG TCT TCA CTA CCA, corresponding to nucleotide positions 259–276 of the bovine GAPDH mRNA sequence. The reverse primer sequence (5′ to 3′) for GAPDH was TTG AGA GGG CCC TCT GA, corresponding to nucleotide positions 743–759 of the bovine GAPDH mRNA sequence. The expected size of the PCR product was 500 bp. The forward primer sequence (5′ to 3′) for α-SM actin was AGC ATC CAA CCC TTC TCA, corresponding to nucleotide positions 84–101 of the bovine α-SM actin mRNA sequence. The reverse primer sequence (5′ to 3′) for α-SM actin was TTC TCG AGG GAG GAG GA, corresponding to nucleotide positions 465–482 of the bovine α-SM actin mRNA sequence. The expected size of the PCR product was 398 bp.

The primer sequences for myosin heavy chain SMA and SMB isoforms were the same as described by Eddinger and Meer (12) and corresponded to nucleotide positions 610–629 and 806–787 as published by Nagai et al. (32). These primers flank the 21-nucleotide exon insert that is present in SMB but absent in SMA isoforms. The forward primer sequence (5′ to 3′) for SMA/SMB was CAG TCC ATT CTC TGC ACA GG. The reverse primer sequence (5′ to 3′) for SMA/SMB was TCA TTC TTG ACC GTC TTG GC. The expected sizes of the PCR products for SMA and SMB were 197 and 218 bp, respectively. As reviewed by Babu et al. (4), alternative splicing of the smooth muscle myosin gene near the 5′ end generates the myosin SMA and SMB isoforms. It is noteworthy that exon 5b for the seven-amino acid insert is separated from the flanking exon 5a and exon 6 by introns. Therefore, PCR products generated from amplification of the genomic DNA would have different sizes from the PCR products generated from RT-PCR of the mRNA. Because the primers for SMA/SMB were designed based on the sequence of rabbit smooth muscle myosin (32), we sequenced the SMA PCR product from bovine tracheal smooth muscle for comparison. DNA sequencing was done by automated DNA cycle sequencing at Brown University's DNA Sequencing Facility with the Big Dye Terminator kit 1.1 (Applied Biosystems, Foster City, CA). We have sequenced the SMA PCR product from both the forward and reverse directions with the appropriate primers. Results from the two directions of sequencing were consistent in producing the data in Table 1. We compared the bovine SMA PCR product sequence to the cDNA sequence for rabbit myosin heavy chain mRNA (GenBank accession no. M77812). As shown in Table 1, excluding the primer sequences, 146 of the 157 nucleotide bases (93%) of the bovine SMA PCR product sequence were identical to the rabbit SMA cDNA. The deduced amino acid sequences from the two sets of cDNA sequences indicated that 51 of the 52 amino acid residues (98%) were identical. We checked the bovine SMA PCR product sequence against the nucleotide sequences in GenBank with BLAST and found significant matches with smooth muscle myosin heavy chain sequences from human, mouse, and rat. The estimated probability that the matches happened by chance was extremely small (8 × 10−42–1 × 10−31). The sequence of the seven-amino acid insert in the myosin SMB isoform is highly conserved and identical in rat, rabbit, and mouse smooth muscles (4). The sequence data together with the BLAST results suggest that bovine tracheal smooth muscle expresses predominantly the myosin SMA isoform. It is noteworthy that human airway smooth muscle also expresses only the SMA isoform (28). We further confirmed our results by Western blotting with antibodies specific for smooth muscle myosin SMA and SMB isoforms (from T. J. Eddinger). We found that bovine tracheal smooth muscle expressed predominantly the SMA isoform, with the SMB signal indistinguishable from nonspecific binding (data not shown). Therefore, bovine and human airway smooth muscles are similar in expressing predominantly the SMA isoform (28).

Table 1.

Comparison of bovine SMA PCR product sequence with rabbit smooth muscle myosin SMA cDNA sequence

The primer sequences for myosin light chain LC17a and LC17b isoforms corresponded to nucleotide positions 346–365 and 534–553 as published by Lash et al. (25). These primers flank the 45-nucleotide exon insert that is present in the LC17b isoform but absent in the LC17a isoform. The forward primer sequence (5′ to 3′) for LC17a/LC17b was GCA CGT TCT CGT CAC ACT GG. The reverse primer sequence (5′ to 3′) for LC17a/LC17b was CAA GAG AAC CTA GAA GCG TC. The expected sizes of the PCR products for LC17a and LC17b were 208 and 253 bp, respectively. All primers were synthesized by Invitrogen.

For the PCR reaction, an aliquot of RT product was added to a PCR mix containing Taq buffer (Promega), 1.5 mM MgCl2 (Promega), each dNTP at 0.2 mM, 1 μM forward primer, and 1 μM reverse primer. The PCR reaction tubes were placed in a thermocycler (Gene Amp PCR System 2400; Applied Biosystems), and the solution was heated to 80°C. The cycle was paused, and 1 U of Taq polymerase (Promega) was added to each tube. After 2 min at 94°C, the PCR temperature protocol consisted of 30 cycles of 94°C denaturation for 90 s, 55°C annealing for 2 min, and 72°C primer extension for 3 min. The PCR products were then allowed to cool gradually to 4°C and stored at −20°C until DNA chip analysis.

PCR products were analyzed by using the DNA500 chip and a bioanalyzer (2100 Bioanalyzer; Agilent Technologies). A DNA500 chip contains interconnected fluid wells and microchannels. A DNA500 chip was set up by filling the microchannels with a sieving polymer and fluorescence dye. An internal standard with known sizes and concentrations of DNA fragments was loaded into the chip. Samples of PCR products were loaded into the sample wells of the chip. The DNA500 chip was then placed inside the Agilent 2100 Bioanalyzer, which consists of a 16-pin electrode cartridge, a laser detector, and a computer. The computer regulated the voltage at the 16-pin electrode to inject individual samples into the separation microchannel to separate the PCR products by size. The separated PCR products were detected by their fluorescence at the laser detector. Using the information from running the internal standard before running the samples, the Agilent 2100 Bioanalyzer estimates the size and concentration of individual PCR products. DNA500 chips were set up and loaded with PCR products according to the manufacturer's instructions. Of the 12 sample wells available for analysis, the first and last wells were loaded with a 100-bp DNA ladder (Invitrogen) to detect any drift in the system during DNA chip analysis. The remaining 10 wells were used for the analysis of PCR products. The bioanalyzer analyzed PCR products by size and concentration. The bioanalyzer accurately estimated the sizes of the PCR products as follows: GAPDH, 500 bp; α-SM actin, 398 bp; LC17a, 208 bp; LC17b, 253 bp; SMA, 197 bp; and SMB, 218 bp. Tissue wet weight of each muscle strip was measured before RNA extraction and used for standardization of PCR products.

Statistics.

Data are shown as means ± SE; n represents the number of tracheal rings. Student's t-test was used for the comparison of two means (P < 0.05 considered significant).

RESULTS

Effect of sinusoidal length oscillation on peak force development by unstimulated, histamine-stimulated, and carbachol-stimulated tissues.

In these experiments, equilibrated muscle strips held at Lo were treated by one of the following conditions: 1) PSS alone, 2) PSS containing 100 μM histamine, or 3) PSS containing 1 μM carbachol. After 15 min of treatment, muscle strips were either held at slack length or placed at sinusoidal length oscillation at Lo ± 0.1 Lo at 1 Hz. Figure 1A shows the peak forces developed by PSS-, histamine-, and carbachol-treated muscle strips at the beginning of sinusoidal length oscillation. The forces shown at cycles −1 and 0 represent isometric forces before sinusoidal length oscillation. The isometric forces were 0.18 ± 0.04, 0.85 ± 0.03, and 0.95 ± 0.07 Fo for PSS-, histamine-, and carbachol-treated muscle strips, respectively. During the first cycle of length oscillation (Fig. 1A), peak forces developed by PSS-, histamine-, and carbachol-treated muscle strips were 0.84 ± 0.07, 1.47 ± 0.09, and 1.68 ± 0.17 Fo, respectively. By the 20th cycle of length oscillation, the peak forces developed by PSS-, histamine-, and carbachol-treated muscle strips were 0.38 ± 0.04, 0.87 ± 0.09, and 1.05 ± 0.07 Fo, respectively. During these first 20 cycles of sinusoidal length oscillation, peak forces developed by histamine- and carbachol-stimulated muscle strips were similar and both significantly higher than the peak force developed by PSS-treated muscle strips (P < 0.05).

Fig. 1.

Peak forces during sinusoidal length oscillations in unstimulated, histamine-stimulated, and carbachol-stimulated tissues. A: peak forces at the beginning of length oscillation. B and C: peak forces after 2 and 4 h of sinusoidal length oscillation. The experiments were carried out in physiological salt solution (PSS), PSS + 100 μM histamine, or PSS + 1 μM carbachol. Symbols and vertical bars represent means ± SE (n = 8–11). SE bars are not visible when they are shorter than symbol size. Forces shown at cycles −1 and 0 represent isometric forces before sinusoidal length oscillation. Peak forces developed by histamine- and carbachol-treated tissues were significantly higher than peak forces developed by unstimulated tissues at all measured times (P < 0.05). Fo, active force of contraction at optimal length.

Figure 1B shows the peak forces developed by PSS-, histamine-, and carbachol-treated muscle strips after 2 h of sinusoidal length oscillation. The forces shown at cycles −1 and 0 represent isometric forces before sinusoidal length oscillation. The isometric forces developed by PSS-, histamine-, and carbachol-stimulated muscle strips were 0.036 ± 0.004, 0.23 ± 0.02, and 0.29 ± 0.02 Fo, respectively. During the 20 oscillatory cycles immediately after 2 h of length oscillation, the peak forces developed by histamine- and carbachol-treated muscle strips were both significantly higher than the peak force developed by PSS-treated muscle strips (P < 0.05). Furthermore, peak forces developed by histamine- and carbachol-stimulated muscle strips were also significantly different after 2 h of length oscillation.

Figure 1C shows the peak forces developed by PSS-, histamine-, and carbachol-treated muscle strips after 4 h of sinusoidal length oscillation. During the 20 oscillatory cycles immediately after 4 h of length oscillation, peak forces developed by histamine- and carbachol-treated muscle strips both remained significantly higher than the peak force developed by PSS-treated muscle strips (P < 0.05). Furthermore, peak forces developed by histamine- and carbachol-stimulated muscle strips were also significantly different after 4 h of length oscillation.

Control experiments for RT-PCR and DNA chip analysis.

Figure 2 shows the typical patterns of DNA chip analysis: lane a shows the bands of the 100-bp DNA ladder; lanes b and c represent negative RT and PCR controls, which do not show any bands at the expected positions of the PCR products; lane d shows the PCR product of GAPDH at the expected 500-bp position; and lane e shows the PCR products of the myosin heavy chain SMA and SMB splice variants, at the expected 197- and 218-bp positions. We found that SMB was typically low or not detectable in bovine tracheal smooth muscle. Figure 2, lane f, shows the PCR products of the myosin light chain LC17a and LC17b splice variants at the expected 208- and 253-bp positions; lane g shows the PCR product of α-SM actin at the expected 398-bp position. These results demonstrate the accuracy of PCR product analysis by the system of DNA chip analysis.

Fig. 2.

DNA chip analysis showing 100-bp DNA ladder (lane a), reverse transcription (RT) negative control (lane b), polymerase chain reaction (PCR) negative control (lane c), PCR product for GAPDH at 500 bp (lane d), PCR products for SMA and SMB at 197 and 218 bp (lane e), PCR products for LC17a and LC17b at 208 and 253 bp (lane f), and PCR product for α-smooth muscle (SM) actin at 398 bp (lane g). Thin lines at top and bottom of each lane represent fluorescent markers for sample separation.

We tested the linearity of the RT-PCR method by increasing and decreasing the RT product by 50%. Figure 3A compares PCR products generated from 1, 2, and 3 μl of RT product. The rationale of the experimental design was that if the relationship between RT and PCR products were linear, then a 50% increase in RT product from 2 to 3 μl should lead to a 50% increase in PCR products. Accordingly, the amount of PCR products generated from 3 μl of RT product should be 150% of the amount of PCR products generated from 2 μl of RT product. Similarly, a 50% decrease in RT product from 2 to 1 μl should lead to a 50% decrease in PCR product. As shown in Fig. 3A, the ratios of PCR products generated from 3-μl and 2-μl RT products clustered around the expected 150% for GAPDH, SMA, LC17a, LC17b, and α-SM actin. PCR product for the myosin heavy chain SMB splice variant was not detectable in bovine tracheal smooth muscle. The overall average was 155 ± 9%, which was not significantly different from the expected 150%. These results suggest linearity between RT and PCR products for detecting increases in mRNA abundance. As shown in Fig. 3A, the ratios of PCR products generated from 1-μl and 2-μl RT products were higher than the expected 50% for GAPDH, SMA, LC17a, LC17b, and α-SM actin. The overall average was 86 ± 2%, implying that a 50% decrease in RT product would result in only a 14% decrease in PCR product. These findings together imply that the RT-PCR method in this study accurately detects increases in mRNA abundance but potentially underestimates decreases in mRNA abundance.

Fig. 3.

Control studies for testing saturation of PCR reactions (A) and reproducibility of DNA chip analysis (B). A: comparison of PCR products generated by using 1, 2, and 3 μl of RT products for the PCR reaction. Dashed lines labeled “50%” and “150%” represent the expected ratios of PCR products generated from 1 and 3 μl of RT product relative to the PCR product generated from 2 μl of RT product if the relation between RT and PCR products is linear. B: comparison of the output readings from the 10 sample wells of 2 separate DNA chips loaded with the same PCR product in each well.

Next, we investigated the reproducibility of DNA chip analysis by loading 20 identical PCR samples into two separate DNA chips, each containing 10 sample wells. As shown in Fig. 3A, values from the two DNA chips distributed randomly about the two means. The average readings from DNA chip 1 (20.2 ± 0.4) and DNA chip 2 (19.5 ± 0.5) were not significantly different. These results indicate the reproducibility of DNA chip analysis of PCR products.

Effect of sinusoidal length oscillation on mRNA expression in unstimulated, histamine-stimulated, and carbachol-stimulated tissues.

In these experiments, PSS-treated muscle strips were either held at slack length or placed at sinusoidal length oscillation for 2 or 4 h. As shown in Fig. 4, sinusoidal length oscillation significantly downregulated α-SM actin expression by 44% in unstimulated muscle strips at 4 h (Fig. 4B) but showed no other significant changes. In contrast, when muscle strips were stimulated by 100 μM histamine and 1 μM carbachol, sinusoidal length oscillation did not significantly alter the abundance of mRNA encoding GAPDH, SMA, SMB, LC17a, and LC17b (Fig. 5). To determine the effect of muscle length per se on mRNA expression, we also compared unstimulated muscle strips held either at Lo or slack length in PSS but did not detect any significant differences in mRNA expression between muscle strips at the two muscle lengths (data not shown). Therefore, oscillatory strain appeared to be necessary to induce downregulation of α-SM actin expression in unstimulated muscle strips.

Fig. 4.

Effect of sinusoidal length oscillation on mRNA expression in unstimulated tissues. Paired muscle strips were either held at slack length or placed on sinusoidal length oscillation in PSS for 2 (A) or 4 (B) h. Bars and vertical lines represent means and SE (n = 5–8). *Significant difference in mRNA expression between tissues held at slack length and tissues placed on sinusoidal length oscillation (P < 0.05).

Fig. 5.

Effect of sinusoidal length oscillation on mRNA expression in histamine (A)- and carbachol (B)-stimulated tissues. Paired muscle strips were either held at slack length or placed at sinusoidal length oscillation for 4 h in PSS containing 100 μM histamine (A) or PSS containing 1 μM carbachol (B). Bars and vertical lines represent means and SE (n = 8–12).

Effect of histamine and carbachol on mRNA expression.

The results shown in Figs. 4 and 5 suggest that sinusoidal length oscillation and receptor activation interactively regulate mRNA expression in bovine tracheal smooth muscle. To address the possibility that receptor activation per se may modulate mRNA expression in bovine tracheal smooth muscle, we performed additional experiments to compare mRNA expression in unstimulated and receptor-activated tissues held at identical lengths, that is, either at slack length or at sinusoidal length oscillation. As shown in Fig. 6, relative to the unstimulated PSS control, 4-h stimulation of muscle strips by 100 μM histamine did not significantly alter the expression of GAPDH, SMA, SMB, LC17a, LC17b, or α-SM actin at either slack length or sinusoidal length oscillation. In contrast, 4-h stimulation by 1 μM carbachol significantly upregulated the expression of GAPDH and SMA by 34% and 65%, respectively, at slack length (Fig. 7A). However, when muscle strips were placed at length oscillation, 4-h stimulation by 1 μM carbachol significantly increased the expression of GAPDH by 45% but did not significantly alter SMA expression (Fig. 7B).

Fig. 6.

Effect of histaminergic receptor activation on mRNA expression in muscle strips held at slack length (A) or sinusoidal length oscillation (B). Paired muscle strips were either unstimulated in PSS alone or stimulated in PSS containing 100 μM histamine. In A, both unstimulated (PSS) and histamine-stimulated muscle strips were held at slack length for 4 h. In B, both unstimulated (PSS) and histamine-stimulated muscle strips were placed at sinusoidal length oscillation for 4 h. Bars and vertical lines represent means and SE (n = 6–9).

Fig. 7.

Effect of cholinergic receptor activation on mRNA expression in muscle strips held at slack length (A) or sinusoidal length oscillation (B). Paired muscle strips were either unstimulated in PSS alone or stimulated in PSS containing 1 μM carbachol. In A, both unstimulated (PSS) and carbachol-stimulated muscle strips were held at slack length for 4 h. In B, both unstimulated (PSS) and histamine-stimulated muscle strips were placed at sinusoidal length oscillation for 4 h. Bars and vertical lines represent means and SE (n = 7–9). *Significant difference between unstimulated (PSS) and carbachol-stimulated tissues (P < 0.05).

Effect of U0126 and actinomycin D on carbachol-induced contraction and mRNA expression.

ERK1/2 MAPK has been implicated in gene regulation and cell proliferation in airway smooth muscle (13, 33). U0126 inhibits MEK1/2, thereby inhibiting ERK1/2 phosphorylation and activation. If ERK1/2 MAPK activation is an important determinant of carbachol-stimulated upregulation of the abundance of mRNA encoding GAPDH and SMA, then U0126 should inhibit GAPDH and SMA expression in carbachol-stimulated muscle strips. In these experiments, muscle strips were stimulated with 1 μM carbachol in the absence or presence of 10 μM U0126 for 4 h at either slack length or length oscillation. As shown in Fig. 8, U0126 did not significantly affect peak forces developed by carbachol-stimulated tissues during sinusoidal length oscillations. These data are consistent with our previous finding that 10 μM U0126 had an insignificant effect on isometric contraction despite U0126's complete inhibition of carbachol-induced ERK1/2 phosphorylation (data not shown). As shown in Fig. 9, U0126 did not significantly alter GAPDH and SMA expression in carbachol-stimulated muscle strips held either at slack length or at length oscillation. U0126 appeared to upregulate GAPDH and SMA expression in carbachol-stimulated tissues, but the effect was statistically insignificant.

Fig. 8.

Effect of U0126 on peak forces developed by carbachol-stimulated tissues during sinusoidal length oscillation. Paired muscle strips were oscillated for 4 h in PSS containing 1 μM carbachol without U0126 or with 10 μM U0126. Symbols and vertical bars represent means ± SE (n = 8–10). Standard error bars are not visible when they are shorter than symbol size. Forces shown at cycles −1 and 0 represent forces at isometric muscle length before length oscillation.

Fig. 9.

Effect of U0126 on carbachol-stimulated mRNA expression in muscle strips held at slack length (A) or sinusoidal length oscillation (B). In A, paired muscle strips were held at slack length and incubated in PSS containing 1 μM carbachol either without U0126 or with 10 μM U0126 for 4 h. In B, paired muscle strips were placed at sinusoidal length oscillation and incubated in PSS containing 1 μM carbachol either without U0126 or with 10 μM U0126 for 4 h. Bars and vertical lines represent means and SE (n = 5–10).

Actinomycin D is an inhibitor of RNA polymerase and has been used to inhibit gene transcription in bovine tracheal smooth muscle (21). To determine whether carbachol-induced increases in SMA and GAPDH mRNA abundance were due to increases in gene transcription, we incubated bovine tracheal smooth muscle strips at slack length in PSS containing 5 μg/ml actinomycin D for 1 h and then stimulated the muscle strips with PSS containing 1 μM carbachol with or without 5 μg/ml actinomycin D for 4 h. As shown in Fig. 10, actinomycin D (5 μg/ml) did not significantly change SMA and GAPDH mRNA abundance in carbachol-stimulated tissues.

Fig. 10.

Effect of actinomycin D on carbachol-stimulated mRNA expression in muscle strips held at slack length. Paired muscle strips were held at slack length, pretreated with PSS containing 5 μg/ml actinomycin, and then stimulated in PSS containing 1 μM carbachol either without actinomycin or with 5 μg/ml actinomycin for 4 h. Bars and vertical lines represent means and SE (n = 5 or 6).

Figure 11 summarizes the major findings from this study that sinusoidal length oscillation and receptor activation interactively regulate the expression of α-SM actin, SMA, and GAPDH in bovine tracheal smooth muscle. As shown in Figs. 4 and 5 and summarized in Fig. 11, sinusoidal length oscillation induced downregulation of α-SM actin expression in unstimulated muscle strips, but the downregulation was abolished by histaminergic and cholinergic receptor activation. As shown in Fig. 7 and summarized in Fig. 11, cholinergic receptor activation induced upregulation of SMA and GAPDH expression, but the upregulation of SMA expression but not GAPDH expression was abolished by sinusoidal length oscillation.

Fig. 11.

Regulation of the abundance of mRNA encoding α-SM actin and myosin heavy chain by sinusoidal length oscillation and receptor activation in airway smooth muscle. This model summarizes the findings shown in this study. Solid lines represent direct effects of length oscillation and receptor agonists on mRNA expression. Dashed lines represent modulatory effects of sinusoidal length oscillation and receptor activation on mRNA expression.

DISCUSSION

A major finding of this study was the interactive regulation of gene regulation by sinusoidal length oscillation and receptor activation in intact airway smooth muscle. We found that sinusoidal length oscillation significantly downregulated the abundance of mRNA encoding α-SM actin in unstimulated bovine tracheal smooth muscle (Fig. 4) but not in histamine- and carbachol-activated tissues (Fig. 5). Conversely, cholinergic receptor activation significantly upregulates myosin SMA heavy chain expression at slack length but not at sinusoidal length oscillation (Fig. 7). The observed length oscillation-induced downregulation of the abundance of mRNA encoding α-SM actin (Fig. 4) was consistent with the finding of Lundberg et al. (27) that cyclic stretching at 24% amplitude and 2 Hz decreased α-SM actin expression in cultured aortic smooth muscle cells. However, Tock et al. (39) found that cyclic stretching at 5% amplitude and 1 Hz increased α-SM actin expression in cultured rat aortic smooth muscle cells. Yang et al. (42) investigated the effect of static stretching on undifferentiated lung mesenchymal cells and found that 10% stretching increased α-SM actin expression, whereas 15% or higher stretching decreased α-SM actin expression. Results from these studies suggest that mechanical stretch may upregulate or downregulate α-SM actin expression in cultured vascular smooth muscle and lung mesenchymal cells depending on the stretch amplitude. To our knowledge, this is the first report of mechanical oscillation-induced downregulation of α-SM actin expression in intact airway smooth muscle. This finding suggests the possibility that periodic stretching of airways during deep inspiration may modulate mRNA expression in airway smooth muscle cells in addition to bronchodilation and bronchoprotection (36).

Histamine- and carbachol-activated tissues did not respond to sinusoidal length oscillation with downregulation of α-SM actin expression (Fig. 5), suggesting antagonistic interactions between mechanical stretch and receptor activation in regulating α-SM actin expression in intact airway smooth muscle. This pattern of antagonistic interaction between mechanical stretch and receptor activation was also observed in carbachol-induced mRNA expression experiments. Whereas carbachol significantly upregulated myosin heavy chain SMA expression in muscle strips held at slack length, carbachol did not significantly alter SMA expression in muscle strips with sinusoidal length oscillation (Fig. 7). Recently, Gosens et al. (17) investigated the effect of long-term (8 days) cholinergic receptor activation on contractile protein expression in organ-cultured bovine tracheal smooth muscle. They found that 1 μM methacholine induced a small (∼20%) but statistically insignificant increase in myosin heavy chain protein expression, whereas 100 μM methacholine significantly downregulated myosin heavy chain protein expression. In this study, we observed a statistically significant 65% increase in myosin heavy chain SMA expression induced by 1 μM carbachol. These results together suggest that cholinergic receptor agonists may have a bimodal and concentration-dependent effect on myosin heavy chain expression in airway smooth muscle.

Carbachol also significantly upregulated GAPDH expression in bovine tracheal smooth muscle (Fig. 7). This effect was detectable in our study when we normalized PCR products by tissue wet weight. Recently, Glare et al. (16) reported variable GAPDH expression in bronchoalveolar lavage fluid cells and endobronchial biopsy tissues collected from control and asthmatic patients. Furthermore, recent findings suggest that GAPDH may be involved in the initiation of apoptosis in neurons and skeletal muscle cells (7, 20). These findings suggest that GAPDH is inadequate for the normalization of mRNA expression. The observed carbachol-induced upregulation of GAPDH expression in this study does not appear to represent a general increase in transcriptional activity, as the expression of α-SM actin and myosin light chain LC17a and LC17b did not change significantly in carbachol-stimulated tissues. Furthermore, we found that GAPDH expression was upregulated by cholinergic receptor activation regardless of whether muscle strips were held at slack length or at sinusoidal length oscillation (Fig. 7). Therefore, unlike SMA expression, upregulation of GAPDH expression by cholinergic receptor activation appeared to be insensitive to the mechanical state of airway smooth muscle.

ERK1/2 MAPK has been implicated in gene regulation and cell proliferation in airway smooth muscle (13, 33). Cholinergic receptor activation by 1 μM carbachol has been found to stimulate ERK1/2 MAPK activation in airway smooth muscle (14). We previously found that 10 μM U0126 completely inhibited ERK1/2 MAPK activation induced by 1 μM carbachol (data not shown). In this study, we found that the upregulation of GAPDH and SMA expression by 1 μM carbachol was not significantly affected by 10 μM U0126 (Fig. 9). Therefore, cholinergic receptor-mediated upregulation of SMA and GAPDH expression appears to be mediated by mechanisms other than the ERK1/2 MAPK pathway in bovine tracheal smooth muscle. Actinomycin D is an inhibitor of RNA polymerase and has been used to inhibit gene transcription in bovine tracheal smooth muscle (21). However, actinomycin D (5 μg/ml) did not significantly change SMA and GAPDH mRNA abundance in carbachol-stimulated tissues (Fig. 10). This finding suggested that cholinergic receptor-mediated increases in SMA and GAPDH mRNA abundance were not due to increases in gene transcription. Recently, p38 MAPK-mediated mRNA stabilization has been implicated in cytokine-induced increase in cyclooxygenase-2 mRNA abundance in airway smooth muscle (35). It has been recognized that mRNAs having AU-rich elements in their 3′-untranslated regions are stabilized by p38 MAPK (8). Bakheet et al. (5) compiled a database of AU-rich element mRNAs and found a wide repertoire of functionally diverse proteins. A search on their ARED 2.0 website revealed myosin, tropomyosin, and NADH dehydrogenase mRNAs as having AU-rich elements. Cholinergic receptor stimulation has been found to stimulate p38 MAPK phosphorylation in bovine tracheal smooth muscle (24). Therefore, we speculate that p38 MAPK-mediated mRNA stabilization could be an important mechanism in carbachol-induced increases in SMA and GAPDH mRNA abundance in bovine tracheal smooth muscle.

Histamine did not significantly alter the expression of GAPDH, myosin heavy chain SMA and SMB, myosin light chain LC17a and LC17b, or α-SM actin in bovine tracheal smooth muscle (Fig. 6). Panettieri et al. (34) found that histamine induced proliferation and c-fos expression in cultured airway smooth muscle cells but did not significantly alter the ratio of muscle-specific myosin heavy chain to total myosin heavy chain. Therefore, Panettieri et al. (34) concluded that histamine did not induce significant phenotypic modulations toward the production of more muscle-specific myosin heavy chains. Our observation that histamine did not significantly alter myosin heavy chain SMA expression confirms the findings of Panettieri et al. (34) in intact airway smooth muscle. We found that bovine tracheal smooth muscle expresses almost exclusively the myosin heavy chain SMA isoform, with near-zero SMB expression. Furthermore, we found that sinusoidal length oscillation did not significantly affect the relative ratios of myosin heavy chain SMA and SMB isoforms or LC17a and LC17b isoforms in bovine tracheal smooth muscle (Figs. 46). Interestingly, Ma et al. (28) also found no change in the expression of SMA in human asthmatic airway smooth muscle, together with the absence of SMB expression. These findings together suggest that phenotypic plasticity of myosin isoform expression may be limited in intact airway smooth muscle cells in vivo.

Results from this study are summarized in the model of interactive regulation of the abundance of mRNA encoding α-SM actin and myosin heavy chain by sinusoidal length oscillation and receptor activation shown in Fig. 11. In this model, the downregulation of α-SM actin expression by sinusoidal length oscillation is opposed by histaminergic and cholinergic receptor activation. Similarly, the upregulation of myosin heavy chain SMA expression by cholinergic receptor activation is opposed by sinusoidal length oscillation. To our knowledge, this is the first report of interactive regulation of the abundance of mRNA encoding α-SM actin and myosin heavy chain by sinusoidal length oscillation and receptor activation in intact airway smooth muscle. However, the upregulation of GAPDH by cholinergic receptor activation is not sensitive to sinusoidal length oscillation, suggesting the coexistence of mechanoinsensitive pathways of gene regulation in airway smooth muscle cells.

GRANTS

This study was supported by National Heart, Lung, and Blood Institute Grants HL-52714 (to C.-M. Hai) and HL-62237 (to T. J. Eddinger).

Acknowledgments

We thank Dr. Jeannette Krieger for performing the DNA sequencing experiments and Hak Rim Kim for performing the Western blotting experiments. Bovine trachea were generously provided by Rhode Island Beef for this study.

Footnotes

  • 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.

REFERENCES

View Abstract