Vascular smooth muscle cell culture in microﬂuidic devices

This paper presents a microﬂuidic device enabling culture of vascular smooth muscle cells (VSMCs) where extracellular matrix coating, VSMC seeding, culture, and immunostaining are demonstrated in a tubing-free manner. By optimizing droplet volume differences between inlets and outlets of micro channels, VSMCs were evenly seeded into microﬂuidic devices. Furthermore, the effects of extracellular matrix (e.g., collagen, poly- L -Lysine (PLL), and ﬁbronectin) on VSMC proliferation and phenotype expression were explored. As a platform technology, this microﬂuidic device may function as a new VSMC culture model enabling VSMC studies. V C 2014 Author(s). All article content, except where otherwise noted, is licensed under a Creative Commons Attribution 3.0 Unported License .


I. INTRODUCTION
Vascular smooth muscle cells (VSMCs) are essential blood vessel components responsible for the regulation of vessel tones and their dysfunctions may lead to diverse vascular diseases 1,2 (e.g., abdominal aortic aneurysm, 3,4 atherosclerosis, [5][6][7] and restenosis 8,9 ). Current studies of VSMCs are based on conventional cell culture approaches (e.g., culture dish and flask), 10,11 which allow no control over the spatial/temporal distribution of the cells and biomolecules and thus cannot recapitulate local in vivo microenvironments.
Microfluidics is the science and technology of manipulating and detecting fluids in the microscale. 12,13 Due to its dimensional comparison with biological cells and capabilities of defining local biophysical, biochemical, and physiological cues, microfluidics has been used to construct more in vivo like cell culture models, [14][15][16] enabling tumor, 17,18 neuron, 19 and vascular 20,21 studies.
As to applications of microfluidics to vascular smooth muscle cell studies, preliminary studies were confined within three areas.  As the first demonstration, soft lithography was used to form micro and nano patterned extracellular matrix proteins and geometrical topographies enabling the regulation of VSMC morphologies and phenotypes. 22,23,[27][28][29][30][31]36,37,39 Furthermore, microfluidics based gradient-compliance substrates were proposed to investigate durotaxis (migration from less stiff to more stiff substrates) of VSMCs. 26,35 Meanwhile, threedimensional culture of VSMCs based on layer-by-layer assemblies 32,38 or circular micro channels 42 were demonstrated for tissue engineering studies. However, the combination of microfluidics with vascular smooth muscle cells is still at an early stage and no systematic studies of on-chip VSMC culture were previously demonstrated. a) Y. C. Wei, F. Chen, and T. Zhang contributed equally to this work. To address this issue, this paper proposed a microfluidic platform for VSMC culture where extracellular matrix coating (Figures 1(a) and 1(b)), VSMC seeding, culture and observation (Figures 1(c)-1(e)), and immunostaining (Figures 1(f)-1(h)) are demonstrated in a tubing-free manner. More specifically, the VSMC seeding evenness in microfluidic channels was investigated and the effects of extracellular matrix on VSMC phenotypes were explored. In addition, as gravity was used as the driving force for cell loading and immunostaining, this platform does not need external pumps and tubes, which can be operated in biological labs with low access requirements.

A. Materials
Dulbecco Modified Eagle Medium (DMEM) was purchased from Hyclon and the other cell-culture reagents were purchased from Life Technologies. SU-8 photoresist (MicroChem) was used for mold master fabrication while collagen (Sigma), poly-L-Lysine (PLL) (Sigma), and fibronectin (Sigma) were used for glass surface coating. Materials for alpha actin immunostaining included 4% paraformaldehyde (Sigma), primary antibody (Bioss), GTVision Detection System (Gene Technology), and goat serum (GIBCO).

B. Device design and fabrication
Polydimethylsiloxane (PDMS) based microchannels with channel dimensions of 10.00 L Â 0.80 W Â 0.10 H (mm) were proposed in this study, where each channel was divided into eight observation windows (1.00 L Â 0.80 W Â 0.10 H (mm)) and labeled as 1 to 8 for cell distribution eveness evaluation. Two microchannel ports (channel inlet and outlet) with a diameter of 4 mm and a PDMS thickness of 2 mm were used in this study to facilitate liquid droplet manipulation. Four channels were included in one mask to characterize operation repeatability (see Figure 2(a)).
The PDMS device was replicated from a single-layer SU-8 mold based on the conventional soft lithography (see Figure 2(b)). Briefly, SU-8 5 was spin coated on a glass slide, flood exposed, post exposure baked, and hard baked (without development) to form an adhesive layer, followed by SU-8 2100 spin coating, exposure and development, forming the mold master with a height of 100 lm. PDMS prepolymers and curing agents were mixed, degassed, poured on channel masters, and baked in an oven. Cured PDMS channels were then peeled from the SU-8 masters with ports punched through and bonded to glass slides.

C. Cell culture
VSMCs of CRL-1999 (ATCC) were cultured with DMEM media supplemented with 10% fetal bovine serum and 1% penicillin and streptomycin. Immediately prior to an experiment, cells were trypsinized, centrifuged, and resuspended in supplemented culture medium with a concentration of 1 Â 10 6 cells per ml. Cell passage generations between p6 and P12 were used in the experiments.

D. Device operation and data analysis
The fabricated microfluidic devices were sterilized in a hood under ultraviolet (UV) overnight, followed by surface coating (PLL of 0.1 mg per ml, or fibronectin of 0.1 mg per ml or collagen of 1 mg per ml). As the first step, phosphate-buffered saline (PBS) solutions were flushed into the microfluidic devices by using micro pipets. Then, PBS solutions in two ports were replaced with coating solutions containing PLL, fibronectin, or collagen at a volume of 20 ll and 15 ll, respectively. Due to the gravity difference, the coating solutions were loaded into microfluidic channels for one-hour channel soaking. In the end, the coating solutions were removed by aspiration and the coated surfaces were rinsed with DMEM thoroughly.
The cell loading and observation protocol is as follows. DMEM solutions in two ports were removed and replaced with cell suspension solutions at 20 vs. 18 ll, 20 vs. 15 ll, 20 vs. 10 ll or 20 vs. 0 ll, respectively (1 Â 10 6 cells per ml). After five minutes of cell loading, microscopic images were taken to evaluate the cellular seeding evenness. Then each microfluidic device was placed in a petri dish containing PBS to limit evaporation and osmolality shifts and then incubated in a cell incubator. DMEM was replaced every 12 h where solutions in two ports were removed and replaced with 20 ll fresh culture medium.
The on-chip immunostaining procedure is as follows. Cells within microfluidic devices were washed twice with PBS, fixed by 4% paraformaldehyde (15 min), permeabilized by 0.3% Triton X-100 (15 min), blocked with 10% goat serum (1 h). Then cells within microfluidic devices were incubated with primary antibodies (1:100) at 37 C (1 h), secondary antibodies at 37 C (1 h), and the Diaminobenzidine (DAB) solution (2 min), sequentially. Note that in all the steps requiring liquid handling, a setup of 20 vs. 10 ll was used where gravity was the driving force for liquid manipulation inside microfluidic devices.
In order to evaluate the cell loading evenness and the cellular proliferation status inside microfluidic channels, cell number analysis was conducted based on manual processing of phase-contrast images taken along the length of the micro channels at 0 h, 12 h, 24 h, 48 h, and 72 h (Â10 magnification and 592 images in total).

III. RESULTS AND DISCUSSION
Microfluidic technologies hold great promise for the creation of advanced cell culture models. [14][15][16] However, the majority of currently available microfluidic cell culture approaches are not compatible with existing laboratories since syringe pumps and plastic tubes are used for cell loading and liquid dispersion. In addition, these physical connections with microfluidic devices lead to cell seeding on tubing walls and medium instabilities due to tubing disturbances. 44,45 To address this issue, Beebe et al. [44][45][46] and Takayama et al. [47][48][49] proposed the concept of tubing-free microfluidic cell culture where surface tension or gravity was used for cell loading and liquid manipulation. Based on surface tension, difference in menisci of unequal volumes of two drops is used to drive liquid flow, which was extensively optimized and demonstrated for cellular micro environment reconstructions. [44][45][46] Meanwhile, although gravity based cell loading (relying on the height rather than menisci difference of drops for the channel inlet and outlet) was utilized for cell loading and characterization, [47][48][49] corresponding operation optimization and condition investigation were not reported. This is the optimal choice since the averaged cell density (96.0 6 16.3/mm 2 ) approaches the idea cell loading density of 100/mm 2 and the ratio of standard deviation to average was the lowest among all the four setups (Figures 3(b) and 3(e)). In this setup, cell suspension solutions were assumed to replace the culture medium solutions inside microfluidic channels thoroughly as the first step, followed by cell seeding on the substrates.

A. Cell loading distribution evaluation in microfluidic channels
Cell loading densities for the setup of 20 ll vs. 10 ll were quantified as 220.  (Figures 3(c) and 3(e)). In addition, the quantified averaged cell density number was much higher than the ideal cell loading density (167.1 6 37.2/mm 2 vs. 100/mm 2 ). The 10 ll volume difference may lead to a prolonged cell flushing process. As the cell suspension solution was A trend in cell density increase from the port loaded with 20 ll to the port loaded with 0 ll was located (57.5 6 35.5/mm 2 (position 1) vs. 124.7 6 30.1/mm 2 (position 8)) (Figures 3(d)  and 3(e)). It was speculated that the 20 ll volume difference lead to a prolonged cell flushing process, together with cellular seeding on micro channels. Due to the existence of the 20 ll volume difference, a higher flow rate compared to the setup of 20 ll vs. 10 ll was produced, which lead to cell settlement in the downstream rather than in the upstream of microfluidic channels. Thus a higher cell seeding density in position 8 than position 1 was recorded.

B. Effect of extracellular matrix on VSMC phenotypes
VSMCs have been demonstrated to have contractile or synthetic phenotypes. 11,50 Contractile VSMCs are featured with limited proliferation capabilities while synthetic VSMCs are characterized by high proliferation and extracellular matrix secretion. In the local microenvironments of VSMCs, extracellular matrix of fibronectin and collagen were located and their effects on VSMC phenotypes were investigated in conventional cell culture models. 51,52 This study was designed to investigate the effects of extracellular matrix on the phenotypes of VSMCs in the microfluidic environment. Figure 4(a) shows the microscopic images of VSMCs cultured on the glass substrate without modifications in a time sequence, followed by an endpoint immunostaining of alpha actin filaments. Based on the manual cell counting, the cell numbers were quantified as 88.5 6 17.9/mm 2 (12 h), 87.7 6 21.6/mm 2 (24 h), 94.5 6 24.5/mm 2 (48 h), and 102.4 6 28.5/mm 2 (72 h) (see Figure 4(b)). Within the first 24 h of cell seeding, VSMCs were noticed to have limited surface spreading areas and no cellular proliferations (88.5 6 17.9/mm 2 (12 h) vs. 87.7 6 21.6/mm 2 (24 h)) due to the bare glass surface without extracellular matrix coating. Starting from 24 h after cell seeding, VSMCs were noticed to increase in surface spreading areas with noticeable cellular proliferations from 87.7 6 21.6/mm 2 (24 h) to 102.4 6 28.5/mm 2 (72 h).
It was speculated that within the first 24 h, VSMCs secreted extracellular matrix macromolecules on glass substrates as extracellular modifications, which lead to cellular proliferations from 24 h to 72 h. These results indicate that VSMCs maintained the synthetic phenotype during the on-chip culture process, featured with extracellular molecule secretions and cellular proliferations. Figure 5(a) shows the microscopic images of VSMCs cultured on collagen coated glass substrates in a time sequence, followed by an endpoint immunostaining of alpha actin filaments. Based on the manual cell counting, the cell numbers were quantified as 93.4 6 13.2/mm 2 (12 h), 106.7 6 12.9/mm 2 (24 h), 112.7 6 13.9/mm 2 (48 h), and 111.3 6 13.4/mm 2 (72 h) (see Figure 5(b)).
In summary, for VSMCs cultured in microfluidic channels coated with extracellular matrix (e.g., collagen, PLL, and fibronectin), cellular proliferations were initially noticed, followed by stable cell numbers without cell doubling, indicating that VSMCs changed from the synthetic phenotype to the contractile phenotype. Since cells in the middle areas of the microchannels with limited access to fresh medium (positions 4 and 5) and cells in the areas close to channel ports with easy access to fresh medium (positions 1 and 8) demonstrated similar trends ( Figures 5(b), 6(b), and 7(b)), these proliferation rate variations and phenotype transitions have no relationship with nutrition transportation. Thus it was speculated that the height of microfluidic channels (100 lm) may provide geometry limitations 53 or the thickness of PDMS (2 mm) may limit oxygen transportation, 54 which was responsible for cellular proliferation and phenotype variations.

IV. CONCLUSION
This paper presented a tubing-free microfluidic platform for VSMC culture where extracellular matrix coating, VSMC seeding, culture and observation, and immunostaining were demonstrated. Based on operation optimization, even distributions of cells within microfluidic channels (96.0 6 16.3/mm 2 ) was realized. Compared to bare glass surface, extracellular matrix (e.g., collagen, PLL, and fibronectin) coated substrates lead to VSMC proliferations and phenotype variations in a time sequence. As a platform technology, this microfluidic device was confirmed to be capable of functioning as a new VSMC culture model for VSMC studies.