Cyclic stretch enhances reorientation and differentiation of 3-D culture model of human airway smooth muscle

Activation of airway smooth muscle (ASM) cells plays a central role in the pathophysiology of asthma. Because ASM is an important therapeutic target in asthma, it is beneficial to develop bioengineered ASM models available for assessing physiological and biophysical properties of ASM cells. In the physiological condition in vivo, ASM cells are surrounded by extracellular matrix (ECM) and exposed to mechanical stresses such as cyclic stretch. We utilized a 3-D culture model of human ASM cells embedded in type-I collagen gel. We further examined the effects of cyclic mechanical stretch, which mimics tidal breathing, on cell orientation and expression of contractile proteins of ASM cells within the 3-D gel. ASM cells in type-I collagen exhibited a tissue-like structure with actin stress fiber formation and intracellular Ca2+ mobilization in response to methacholine. Uniaxial cyclic stretching enhanced alignment of nuclei and actin stress fibers of ASM cells. Moreover, expression of mRNAs for contractile proteins such as α-smooth muscle actin, calponin, myosin heavy chain 11, and transgelin of stretched ASM cells was significantly higher than that under the static condition. Our findings suggest that mechanical force and interaction with ECM affects development of the ASM tissue-like construct and differentiation to the contractile phenotype in a 3-D culture model.


Introduction
Contraction of airway smooth muscle (ASM) plays a central role in airway narrowing in asthma. Increased ASM mass due to cell proliferation, hypertrophy, and migration is involved in the mechanism of pathophysiology of airway remodeling. Therefore, ASM is an important therapeutic target for airway diseases, specifically asthma and chronic obstructive pulmonary disease [1]. In order to uncover mechanisms underlying activation of ASM cells, two-dimensional (2-D) in vitro cultures of ASM cells have widely been used due to difficult availability of human ASM tissue samples [2][3][4]. However, ASM cells in vivo exist as a part of complex three-dimensional (3-D) structures with the extracellular matrix (ECM). Within the airway wall, ASM exists as an aligned population that wraps around the bronchiole in a helical fashion in vivo [5,6]. Due to this unique arrangement, the angle of orientation and cell alignment are major factors that determine the phenotypes and properties of ASM cells [7]. Therefore, development of bioengineered 3-D models of ASM tissues is warranted to assess functional properties for pharmacological and biophysical studies [8][9][10].
The lungs and airways are continually exposed to mechanical forces such as shear stress, compression, and stretch during tidal breathing and pulmonary circulation in vivo. These mechanical stresses are involved in the mechanisms underlying the normal physiology and development of the respiratory system and pathogenesis of asthma [11]. In 2-D culture models of ASM cells, cyclic stretch induces cell alignment perpendicular to the stretch axis with reorganization of the cytoskeleton [12,13]. However, the roles of mechanical stresses in the development of bioengineered 3-D models of ASM tissues are not known.
This study was designed to develop a 3-D model of ASM. For this purpose, human ASM cells were embedded in a collagen gel [9,10]. We further examined the effects of cyclic mechanical stretch, which mimics tidal breathing, on the regulation of cell orientation, formation of stress fibers, and phenotype. We postulated that when cultured three-dimensionally within collagen gel with cyclic stretch, ASM cells develop tissue-like behavior by upregulating expression of genes for contractile proteins.

Cells
Primary cultures of normal human bronchial smooth muscle cells from three different donors were obtained from Lonza (Walkersville, MD) and maintained in SmGM-2 culture medium (Lonza) containing 5% fetal bovine serum (FBS) in an atmosphere of 5% CO 2 and 95% air at 37°C [2,13,14]. Cells of passages 4-8 were used.

Fabrication of 3-D constructs
ASM (5 ×10 5 /ml) cells were suspended in a solution of 2 mg/ml of type I collagen (Cellmatrix; Nitta Gelatin, Osaka, Japan) in SmGM-2 cell culture medium (Invitrogen, Carlsbad, CA) containing 5% FBS at room temperature. The solution was transferred into the well (10 mm in length x 5 mm in width x 5 mm in depth) of a silicone chamber with sponge anchors on both sides (STB-CH-3.5GS; Strex, Osaka, Japan) ( Fig. 1A and B). The solution was able to infiltrate into the sponge. The gel was allowed to polymerize and attached to the sponge anchors by incubation at 37°C for 15 min. After the gels were polymerized, 2 ml of SmGM-2 cell culture medium containing 5% FBS was added to the gel (Fig. 1A), then the medium was changed every other day.

Application of cyclic and static mechanical stretch
After the gel constructs were incubated for 24 h and polymerized, a uniaxial sinusoidal stretch of 12% strain at 30 cycle/min was applied for 48 h using a stretching apparatus driven by a computer-controlled stepping motor (ST-140; Strex) [2,13,15]. Briefly, one end of the chamber was attached to a fixed frame, while the other end was attached to a movable frame (Fig. 1C). The other two sides were free to move. The movable frame was connected to a motor driven shaft whose amplitude and frequency of stretch was controlled by a programmable microcomputer. Strain was calculated from the displacement of the silicone chamber before and after the stretch. Cells incubated under a static condition in the silicone chamber were used as a time-matched control.

Measurement of cell orientation
Images of the nuclei stained with DAPI were obtained using a confocal microscope with at least three arbitrarily selected visual fields. Optical volumes 517.6 µm x 517.6 µm x 200 µm were flattened into a single plane image. The orientation of each nucleus of the cell was measured as an angle (θ) of the long axis between 0°and 90°with respect to the stretch axis ( Supplementary Fig. S1) using NIH ImageJ v1.33 software [13,15].

Statistical analysis
Data are expressed as means ± standard deviation (SD). An unpaired t-test or analysis of variance (ANOVA) followed by Bonferroni's or Games-Howell's post hoc test was used to evaluate the statistical significance. P < 0.05 was considered statistically significant. Statistical analyses were performed using SPSS ver. 24 (SPSS Inc., Chicago, IL).

Collagen gel and cells
Representative images of collagen gel and gel containing ASM cells after 72 h incubation in the silicone chamber are shown in Fig. 2A. Both edges of the gel were fixed to anchors. After 24 h incubation, the gel containing ASM cells was cyclically stretched or kept under the static condition for a further 48 h ( Fig. 2A). The shape of collagen gel without cells was not changed by incubation for 72 h ( Fig. 2A). In contrast, the width of the center of collagen gel constructs containing ASM cells was gradually shortened, indicating tension development due to cell contraction and collagen degradation. The width of the gel containing ASM cells became approximately 70% of the initial width 24 h after incubation ( Fig. 2A and B). There was no significant difference between the width of the gels of the static and stretched tissues at 72 h (Fig. 2B).  (Fig. 3B). Then, the orientation of each nucleus of the cell was measured. Histograms of cell orientation under the static and stretched conditions are shown in Fig. 3C. The average angles of the stretched cells were significantly smaller than those of static cells (Fig. 3D). The SD values, a measure of heterogeneity of cell direction [13], were not significantly different between the groups (Fig. 3E).

Effects of cyclic uniaxial stretch on cell and F-actin alignment
Supplementary material related to this article can be found online at doi:10.1016/j.bbrep.2018.09.003.

Effects of cyclic stretch on expression of contractile proteins
We investigated whether cyclic stretch induces differentiation of ASM cells to the contractile phenotype. Expression of α-SMA protein was used as an indicator of a contractile ASM phenotype. Fig. 4A shows immunofluorescent images of α-SMA-positive stress fibers, fluorescent F-actin, and merged images of ASM cells cultured within collagen gels with or without cyclic stretch. α-SMA-positive stress fibers were increased in the stretched cells. Real-time quantitative PCR data show that mRNA levels of contractile protein genes, α-SMA (ACTA2), calponin (CNN1), myosin heavy chain 11 (MYH11), and transgelin (TAGLN), in the stretched cells were significantly higher than those in the static cells (Fig. 4B). In our preliminary experiments, relative mRNA levels of ACTA2, CNN1, MYH11, and TAGLN in ASM cells cultured on the 2-D plastic dish to those in the 3-D gel under the static condition were 1.68, 4.03, 0.97, and 7.67, respectively.

Elevation of intracellular Ca 2+ concentration by methacholine
Next, we examined whether the ASM cells cultured within the

Discussion
The main findings of the present study are that in a 3-D culture model of ASM: (1) culture of human ASM cells within type-I collagen exhibited a tissue-like structure with F-actin formation, (2) uniaxial cyclic stretch enhanced alignment of nuclei and actin stress fibers, and (3) expression of mRNAs for contractile proteins such as α-SMA, calponin, myosin heavy chain 11, and transgelin of stretched ASM cells was significantly higher than that within the static gel. To our knowledge, we demonstrated for the first time that cyclic stretching enhanced cell reorientation with α-SMA expression in the 3-D model of ASM.
It is well-known that the cells in 2-D cultures distribute randomly under the static condition without stretching [12,13,15,20,21]. Interestingly, actin stress fibers of ASM cells slowly tend to align parallel to the stretch direction without applying cyclic stretch in our 3-D models. The average angle of the cells becomes closer to 45°when the distribution of cell orientations is random [13], but the average nucleus angle of static cells was approximately 35°under the static condition for 72 h (Fig. 3B). In our method, both ends of the 3-D gel were attached to sponge anchors, and the time-dependent shortening was observed in the gel containing ASM cells but not in that without cells, indicating tension development due to cell contraction and cell-ECM interaction (Fig. 2). Similar to our results, West et al. reported that when both ends of 3-D tissues, in which ASM cells together with NIH3T3 fibroblasts are cultured within collagen gels, are fixed, ASM cells tend to align along with the long axis with tension formation [9]. Therefore, the difference in cell orientation under the static condition between 2-D and 3-D models possibly derives from the directed intrinsic tension formation of the cells within the gels and mechanical interaction with surrounding ECM [21].
Cell reorientation toward the stretch direction was enhanced by cyclic stretch in the present 3-D culture of ASM cells (Fig. 3). Using a 2-D flat cell culture system, we and other groups demonstrated that cyclic stretch induces cell alignment perpendicular to the stretch direction in various cell types [12,13,15,20,21]. The difference in direction of alignment is in good agreement with the findings in vascular smooth muscle cells [21,22]. It is considered that the cells change their orientation to minimize the intracellular stress and cellular damage in 2-D cell cultures [21]. In contrast, when smooth muscle cells are embedded within the 3-D gel, the cells mechanically interact with the surrounding cells and ECM, leading to alignment along the stretch direction. However, due to the limitation of our imaging system, a distribution on cell density and orientation could not be acquired across the whole gel. Therefore, involvement of heterogeneity in cell density and orientation within a gel cannot be excluded. Another important issue is that cell migration within the gel might contribute to cell alignment and morphological change. Future studies and improvement of the present system are necessary.
We demonstrated that the expression of α-SMA protein and mRNAs for contractile proteins increased with the cyclic stretching (Fig. 4). Similar results were reported in a 3-D model of vascular smooth muscle [21]. Acquisition or increased expression of α-SMA, one of the six known eukaryotic actin isoforms, characterizes differentiation from a proliferative-to-contractile phenotype and increased contractile ability of ASM cells [23][24][25]. It is known that concentrations of contractile proteins decrease when ASM cells are cultured and passaged under the static 2-D condition [4,25]. Moreover, substrate stiffness regulates cellular properties and expression of contractile proteins in various cell types including ASM cells [17,26,27]. Our findings suggest that activation of cellular mechanotransduction is involved in the mechanisms of differentiation and expression of genes for contractile proteins induced by cyclic stretch in 3-D culture of ASM cells embedded in the collagen gel.
In the present study, a 12% uniaxial cyclic strain at 30 cycle/min was applied to the gels in accordance with methods described in our previous reports using 2-D culture of ASM cells [2]. The expected physiological range of tidal muscle stretch during breathing is from approximately 4% of muscle length during spontaneous breathing at rest to 12% during a sigh [28]. We previously demonstrated that human ASM cells release ATP in response to 12% uniaxial stretch (30 cycle/ min for 15 min) [2]. Moreover, a single 10% stretch induces [Ca 2+ ] i elevation via activating stretch-activated channels in 2-D culture of ASM cells [18]. Taken together, the sinusoidal stretch protocol (12% strain at 30 cycle/min) is within a physiological range and mimics the physical and biological properties of the airway wall.
High-throughput screening for drugs modulating contractile forces of ASM is beneficial to find novel therapeutic strategy and to understand the pathogenesis of asthma. Therefore, one of the goals of developing bioengineered ASM cells and tissues is to measure physiological and biophysical properties [3,8,9,24,29]. In our model, ASM cells within the collagen gel exhibited intracellular Ca 2+ mobilization in response to methacholine (Supplementary Fig. S2) as seen in intact tissues [30]. The increase of [Ca 2+ ] i plays a pivotal role in activation specifically of contraction of ASM cells [31]. Nesmith et al. designed and built an in vitro model of human ASM tissue and measured contraction [29]. Park et al., 2-D cultured ASM cells on polyacrylamide-based gel substrates and measured the contractile force of each ASM cell using Fourier-transform traction microscopy [3]. Future studies are necessary to extend our 3-D model and develop the system to assess its contractile force as well as [Ca 2+ ] i .
In summary, we developed an engineered ASM tissue-like construct that exhibits characteristics of ASM such as F-actin alignment, α-SMA expression, and elevation of [Ca 2+ ] i in response to methacholine using a 3-D culture of ASM cells within a collagen gel. Furthermore, cyclic mechanical stretch enhanced differentiation to the contractile phenotype. Our findings suggest that mechanical forces, both intrinsic tension formation and externally applied cyclic stretch mimicking tidal breathing, play an important role in development of ASM tissue-like behavior in a 3-D culture model.