A New Candidate Substrate for Cell-Matrix Adhesion Study: The Acellular Human Amniotic Matrix

In vivo adhesions between cells and the extracellular matrix play a crucial role in cell differentiation, proliferation, and migration as well as tissue remodeling. Natural three-dimensional (3D) matrices, such as self-assembling matrices and Matrigel, have limitations in terms of their biomechanical properties. Here, we present a simple method to produce an acellular human amniotic matrix (AHAM) with preserved biomechanical properties and a favorable adhesion potential. On the stromal side of the AHAM, human foreskin fibroblasts (HFFs) attached and extended with bipolar spindle-shaped morphology proliferated to multilayer networks, invaded into the AHAM, and migrated in a straight line. Moreover, αV integrin, paxillin, and fibronectin were observed to colocalize after 24 h of HFF culture on the stromal side of the AHAM. Our results indicate that the AHAM may be an ideal candidate as a cell-matrix adhesion substrate to study cell adhesion and invasion as well as other functions in vitro under a tensile force that mimics the in vivo environment.


Introduction
Initial studies have revealed that cell-matrix adhesions play a crucial role in cell morphology, migration, differentiation, and proliferation as well as tissue organization and matrix remodeling, all of which are essential for embryonic development and the remodeling and homeostasis of adult tissues [1][2][3][4]. Natural three-dimensional (3D) matrices have been adopted as physiological models to analyze cell-matrix interactions, rather than using traditional two-dimensional (2D) tissue culture [5][6][7][8].
Cell adhesion to a 3D matrix is referred to as in vivo native cell-matrix adhesion, which differs from focal and fibrillar adhesions [2,9]. Due to the difficulty of mimicking the 3D aspects of the matrix in vivo, 3D adhesion has been modeled in vitro using in self-assembling matrices and Matrigel [2,5]. The study of cells attached to a matrix by 3D adhesion is required to design or reproduce physiologically relevant conditions to decipher biological mechanisms, assay drug responses in vitro, and transplant cells in a native physiological state into the body [10].
Self-assembling matrices, Matrigel, and hydrogels are complex, costly to produce, and have limitations in terms of their geometric and biomechanical properties [11]. Thus, development of a tissue-derived matrix in vitro with favorable adhesion potential and biomechanical properties is anticipated. Great efforts have been made to decellularize a variety of tissues including heart valves, blood vessels, skin, nerves, skeletal muscle, tendons, ligaments, small intestinal submucosa, urinary bladder, and liver, [12][13][14][15][16][17][18]. However, it has been demonstrated that decellularization can alter the native 3D architecture of the extracellular matrix (ECM).
The human amniotic membrane (HAM) has many characteristics that are desirable for a biomaterial. It is inherently tough, yet devoid of blood vessels, lymphatics, and nerves. Additionally, the HAM is nutritionally permeable, inexpensive, easily obtained, and readily available [19,20]. Currently, decellularization of the HAM is focused on the removal of all cells on the epithelial side, and thus far no study has explored the cell-matrix adhesion potential of its stromal side. In this study, we introduce a method to thoroughly remove cells from both epithelial and stromal sides of the HAM, while preserving the favorable adhesion potential and biomechanical properties. Several types of cells seeded on the stromal side of the acellular human amniotic matrix (AHAM) exhibited bipolar spindle-shaped morphology and migrated in a straight line. Moreover, αV integrin, paxillin, and fibronectin were observed to colocalize on the stromal side of the AHAM at 24 h after seeding human foreskins fibroblasts (HFFs). Our results indicate that the AHAM is a promising potential candidate as a cell-matrix adhesion substrate to study cell adhesion and invasion as well as other functions in vitro under a tensile force that mimics the in vivo environment.

Preparation of the AHAM.
The study was approved by the General Hospital of Tianjin Medical University Ethical Committee for the usage of biological material for research purposes. All materials were used in compliance with ethical guidelines. Fresh HAM was obtained after caesarian deliveries. Maternal donors provided informed consent and were serologically negative for HIV, hepatitis B, hepatitis C, and syphilis. Briefly, blood clots were immediately cleaned off the placenta after surgery with sterile Ringer's solution containing antibiotics. The HAM was peeled from the chorion and rinsed extensively in Ringer's solution. Then, the HAM was cut into 10 × 10 cm 2 pieces and fixed to a stainless-steel ring supported by a tripod. The HAM was then submerged in Dulbecco's phosphate-buffered saline (DPBS) (Gibco BRL, Grand Island, NY, USA) containing 2% trypsin (Sigma Aldrich, St Louis, MO, USA) and 0.05 mg/mL EDTA (Sigma Aldrich) and incubated at 37 • C. A magnetic stirring bar (C350-21, BOLA) just beneath the HAM was used to stir the mixture at 100 rpm/min for the epithelial side for 20 min, and then another 20 min for the stromal side. After washing away debris with PBS, the AHAM was preserved in Dulbecco's Modified Eagle's Medium (DMEM) (Gibco BRL) containing antibiotics. The preparation procedures were carried out under aseptic conditions, and no particular sterilization process was performed.

Assessment of Biomechanical
Properties. All samples were tested under a uniaxial tensile low-strain rate and loading to failure using an electronic universal testing machine (CSS-44001; Changchun Research Institute for Testing Machines, China) according to the method by Wilshaw et al. [21]. Twelve paired samples of fresh HAM and AHAM were used. Before mounting the tissue strips (15 × 40 mm) on the holder, the thickness was measured at six points using a Mitutoyo thickness gauge with a resolution of 0.01 mm, and the average thickness was recorded. Strips were then clamped onto the holder with a gauge length of 15 mm. Zero extension was set at the point with a preload of 0.01 N. For each measurement, failure had to occur in the center of the sample, or the test was discarded. Data were interpreted using the software designed for the testing rig and further analyzed using SPSS 11.5 (IBM, USA).

Cell Culture.
HFFs were obtained from samples of healthy boys (12-14-year old) during circumcisions after receiving informed consent from the individual's guardian in compliance with ethical guidelines. After subcutaneous tissue removal, dermal specimens were fragmented with scissors into 5 mm 2 pieces. Samples were then rinsed eight times with PBS and vigorous agitation. These fragments were placed into Petri dishes and cultured in DMEM containing 20% (w/v) fetal bovine serum (FBS) (Gibco BRL), 100 U/mL penicillin, and 0.1 mg/mL streptomycin at 37 • C in a humidified atmosphere with 5% CO 2 .
Bone marrow was aspirated from four calf tibia bones in compliance with ethical guidelines. Cells were flushed out from the bone marrow with DMEM containing 50 U/mL penicillin, 0.05 mg/mL streptomycin, and 0.2 mM L-glutamine, collected, and centrifuged at 1509 g for 10 min. Cell pellets were resuspended in complete medium consisting of Minimum Essential Medium alpha supplemented with 15% FBS, 50 U/mL penicillin, and 0.05 mg/mL streptomycin, and then plated in T75 flasks. After 48 h of culture, nonadherent cells were removed, and fresh medium was added to the cells. At subconfluence, cells were harvested by trypsinization and replated at a 1 : 3 split ratio. Cells were cultured for no more than four passages. NIH 3T3 cells (ATCC Manassas, VA, USA) were cultured in DMEM containing 10% bovine calf serum, 100 U/mL penicillin, and 0.1 mg/mL streptomycin at 37 • C in a humidified atmosphere with 5% CO 2 . All reagents used for cell culture were purchased from Gibco BRL-Life Technologies.

Histology and
Immunohistochemistry. HAM and AHAM specimens were fixed in 10% neutral buffered formalin for 24 h, rinsed with PBS, and stored in 70% alcohol. After dehydration in an ascending ethanol series, HAM specimens were embedded in paraffin, sectioned at 3-4 mm on a rotary microtome, deparaffinized in xylene, rehydrated in a descending alcohol series, and stained with hematoxylin and eosin (HE). Five micrometer sections of paraffin-embedded AHAM were stained with primary polyclonal antibodies against human collagen I (SAB4500363; Sigma Aldrich), collagen II (SAB4500366; Sigma Aldrich), collagen III (SAB4500367; Sigma Aldrich), collagen IV (SAB4500375; Sigma Aldrich), laminin (L9393; Sigma Aldrich), TGF-β1 (SAB4502958, Sigma Aldrich) and TGF-β2 (SAB4502960, Sigma Aldrich), and a monoclonal antibody against fibronectin (F0916; Sigma Aldrich) according to the manufacturers' procedures and routine immunohistochemistry methods.
Scanning electron microscopy (SEM) was performed according to the following procedures. Specimens were fixed overnight in 4% (w/v) glutaraldehyde in 0.1 M cacodylate, washed three times for 15 min in 0.1 M cacodylate, and postfixed for 1 h in 1% (w/v) aqueous osmium tetraoxide. Then, specimens were washed three times in distilled water, dehydrated through a graded acetone series (30, 50, 70, 80, 90, 95, 100%, acetone mixture and in pure acetone), dried to the critical point in a drier (Bal Tec, CPD 030) with CO 2 and finally mounted on an aluminum stub and sputtercoated with gold before examination by SEM (S-3500N, Hitachi, Japan).

Quantification of Cell Morphology. At 24 h after seeding
HFFs on glass coverslips or the AHAM, β-actin was detected to visualize the cytoskeleton and morphology by immunofluorescence staining using a primary polyclonal antibody (sc-130656; Santa Cruz Biotechnology, Santa Cruz, CA, USA) according to manufacturers' procedure. The cell axial ratio (cell length/width) and total cell spread area of HFFs were digitally determined as described elsewhere [5,22].

Immunofluorescence Staining and Confocal Microscopy.
Three types of matrices for cell culture were used, namely the AHAM, Matrigel, and glass coverslips. Glass coverslips were coated with either the AHAM or 0.2 mL Matrigel (5 mg/mL) in a 24-well plate (Carolina Biological Supply Co.). The Matrigel layer was 10 μm. Cells were cultured overnight on the stromal side of the AHAM, Matrigel or glass coverslips in DMEM containing 10% (w/v) FBS, 100 U/mL penicillin, and 100 mg/mL streptomycin at 37 • C in a humidified atmosphere with 5% CO 2 . All subsequent steps were performed at room temperature. Briefly, cells were washed twice with PBS, fixed for 5 min with 2% (w/v) formaldehyde in PBS, rinsed three times with 0.5% (w/v) Triton X-100 in PBS for 10 min, and then incubated with polyclonal antibodies against αV integrin (sc-6595; Santa Cruz Biotechnology) and paxillin (sc-5574; Santa Cruz Biotechnology), and a monoclonal antibody against fibronectin (sc-69681; Santa Cruz Biotechnology) for 1 h at 37 • C. After rinsing three times with 0.5% (w/v) Triton X-100 in PBS for 10 min, cells were incubated with appropriate secondary antibodies conjugated to AMCA/TRITC/FITC (50 μg/mL; Jackson ImmunoResearch) in PBS for 1 h and then washed three times with 0.5% (w/v) Triton X-100 in PBS for at least 30 min. Stained samples were mounted in GEL/MOUNT (Biomeda Corp) containing 1 mg/mL 1,-4-phenylenediamine (Fluka) to reduce photobleaching. Immunofluorescence images were obtained with a Laser Scanning Confocal Microscope (TCS-SP5, Leica, Germany).

Time-Lapse
Analysis. HFF migration patterns were observed at 8 h after seeding on glass coverslips or the AHAM. Time-lapse images of cell movements were recorded every 15 min over 6 h with a CCD camera (Nikon, DXM1200C) attached to an inverted microscope (Nikon, TS100) with a 10× objective lens, which was fitted with a 37 • C/5% CO 2 incubator stage consisting of an enclosed chamber with temperature and CO 2 controls. Image stacks were converted using the ImageJ plug-in for smart projector (http://rsbweb.nih.gov/ij/plugins /smart-projector/index.html). Nuclei were tracked to quantify cell motility, and the velocities were calculated in micrometers every 15 min using the ImageJ plug-in (http://rsbweb.nih.gov/ij/plugins/track/track.html).

Statistical Analysis.
All numerical data were presented as the means ± standard deviation. Statistically significant differences between two groups were calculated using the 2tailed unpaired Student's t-test. A value of P < 0.05 was considered as significant.

3.1.
Histology of the AHAM. The HAM was composed of five layers: the amniotic epithelium, basement membrane, compact connective tissue, fibroblast layer, and spongy layer (Figure 1(a)). The epithelial layer and fibroblasts in the stromal side were thoroughly removed for the AHAM that was negative for hematoxylin staining (Figure 1(b)). Under SEM, the stromal layer of the AHAM was composed of a reticular structure with long collagen fibers of various diameters, without interruption or breaks (Figure 1(c)). Many pits and niches (approximately 10 μm in diameter) were observed on the rugged stromal surface. Additional fragments were not found on the bare long collagen fibers, indicating no residual cell debris (Figure 1(c)).
The AHAM was positive for type I, III, IV, and VI collagen, fibronectin, and laminin, as shown by immunohistochemical or immunofluorescence staining (Figures 1(d)-1(i)). Type II collagen, TGF-β1, and TGF-β2 were not observed in the AHAM (data not shown).

Biomechanics of the AHAM.
Results of the biomechanical testing are summarized in Table 1. Although the AHAM was slighter thinner than the fresh control (0.143±0.005 mm versus 0.164 ± 0.006 mm), the AHAM showed no significant differences in transition strain and stress or failure strain, compared with those of fresh samples (Figure 2). The process of decellularization did not compromise the strength of the resulting matrix, as shown by the failure stress/ultimate tensile strength at 0.796 ± 0.045 Mpa.

Morphology of Various Cell Types on the AHAM.
Cell outlines of HFFs on glass coverslips and the AHAM were determined by immunofluorescence staining to visualize the cytoskeleton. HFFs were cultured on glass coverslips or the AHAM for 24 h and then fixed and stained to visualize morphology (Figure 3(a)). Cells cultured on glass coverslips were broader and more flattened, and had more cell protrusions and lamellae than those of cells cultured on the AHAM.  Conversely, HFFs cultured on the AHAM showed a more slender appearance with fewer protrusions than those of cells culture on glass coverslips. The length, width, and total area of HFFs were calculated using MetaMorph software. Consistent with the observed differences in cell morphology, the quantitative differences in the cell axial ratio and total cell spread area of HFFs were digitally calculated using 30-50 cells that were representative of each sample by MetaMorph software (Figures 3(b) and 3(c)). The cell axial ratio was 3.3 ± 0.35 for cells cultured on glass coverslips, and 6.3 ± 0.32 for those on the AHAM. Bovine bone marrow stromal cells and NIH 3T3 cells were also seeded onto the AHAM and showed the same bipolar spindle-shaped morphology. After 96 h of culture

Mode of Cell
Migration on the AHAM. At 8 h after seeding HFFs on coverslips or the AHAM, cell migration was recorded. On the AHAM, HFFs moved with more monoleading processes in an almost straight line, whereas cells migrated on coverslips in a random direction (Figure 6(a)). Cells migrated much faster on coverslips than on the AHAM. Results showed that cells migrated at an average speed of 5 μm/h on the AHAM, whereas the average speed was about 30 μm/h on coverslips ( Figure 6(b)). When cells formed clusters in situ on the AHAM, migration was not observed.

Discussion
The development of optimal biocompatible scaffolds for tissue engineering requires a deep understanding of the interactions between cells and the ECM in vivo. Natural 3D matrices have been adopted as physiological models to analyze cell-matrix interactions, rather than traditional 2D tissue culture. We have developed a new simple method using 2% trypsin and 0.05 mg/mL EDTA to decellularize fresh HAM, in which cells and debris on both stromal and epithelial sides are thoroughly removed, while the favorable cell-matrix adhesion potential and biomechanical properties are preserved. Various attempts have been made to modify the HAM, including chemical and physical cross-linking [23], to enhance its physical properties and remove all cellular components from the HAM in an attempt to produce a biological substrate for seeding various cell types [24][25][26][27][28]. Recently, cells have been removed from the stromal side of the HAM by SDS. However, SDS tends to disrupt the native tissue structure and causes a loss of collagen integrity [29]. Alternatively, trypsin/EDTA does not affect the amount of collagen in tissues [30]. Moreover, in our study, no denaturing of the collagen was observed on the stromal side of the AHAM by SEM, and type I, III, IV and VI, collagens were detected by immunohistochemistry in the AHAM treated by our method. It has been reported that prolonged exposure to trypsin/EDTA greatly decreases the amount of glycosaminoglycans (GAGs). Because tensile strength is closely related to the amount of GAGs, the loss of GAGs contributes to a decrease of tensile strength [31,32]. In the present study, the AHAM showed no differences in transition strain and stress, or failure strain, compared with those at predigestion, suggesting that the loss of GAGs may be negligible.
Trypsin substantially reduces the laminin and fibronectin content of the ECM [30]. On the surface of the AHAM before seeding HFFs, both fibronectin and laminin were not observed by immunofluorescence (data not shown). Surprisingly, at 40 min after seeding HFFs, the cells attached and fibronectin were detected by immunofluorescence. It is likely that HFFs self-assembled fibronectin on the collagen surface of the AHAM. The appearance of self-assembled fibronectin was not observed in the control group. In fact, fibronectin has not been detected in any of the current fabricated matrices [33].
HFFs in the AHAM tended to show a more spindleshaped morphology, fewer branched terminal processes, or fewer protrusions and lamellipodia, as well as a smaller total cell spread area than those of cells cultured on a glass substrate. The cell axial ratio of HFFs cultured on the AHAM was quite similar, whereas HFFs cultured on glass coverslips had a flattened morphology. The shape of cells cultured on the AHAM resembles that of fibroblasts and mesenchymal cells in vivo [5,34,35].
Cells plated on the AHAM showed colocalization of αV integrin, paxillin, and fibronectin. Even in regions of the AHAM without fibronectin, colocalization of both αV integrin and paxillin was observed. Paxillin combines with various integrins to recruit regulatory and structural proteins that control the dynamic changes of cell adhesion, cytoskeletal reorganization, and gene expression that are necessary for cell migration and survival [36,37]. Cultured NIH 3T3 cells have been removed from dishes to produce a self-assembled matrix that, together with Matrigel, shows the features of 3D adhesion in vitro [2,5]. However, such as self-assembled matrix and Matrigel are hydrogels. Moreover, the self-assembled matrix is relatively thin at only 8 μm, which limits the biomechanical properties to study cells of interest in culture under tensile strength. In our study, the AHAM maintained its biomechanical properties, and various cell types, including bone marrow stromal cells, HFFs, and NIH 3T3 cells, were verified to proliferate and invade the AHAM.
For many cell types, such as fibroblasts and macrophages, it is important to migrate through the ECM to carry out functions such as tissue repair and remodeling. We compared the cell migration of HFFs cultured on the AHAM and glass coverslips. HFFs migrated on the AHAM with monoleading processes in a straight line at an average speed of 5 μm/h, which is different from the migration speed on selfassembling matrices [5]. On coverslips, cell migration was multidirectional, and the speed was similar to that observed in previous studies [38]. This observation demonstrated the validity of our cell migration protocol. It should be noted that the speed and direction of adherent cell migration in vivo has not yet been defined. Thus, there is no standard to evaluate cell migration with 3D adhesion. It is clear that the biomechanical environment is crucial for cell migration, thus our production of an AHAM will be an alternative matrix on which to study cell migration with 3D adhesion in vitro.
We removed endogenous fibroblasts from the AHAM that lost the properties of anti-inflammation and inhibition of regeneration. This finding could lead to an alternative use of the AHAM as a matrix for bridging regeneration or loading cells for transplantation into the body. Complexes of αV integrin, paxillin and fibronectin, a bipolar cell shape, and multilayer cell clusters, which are characteristics of cells in vivo [39][40][41], were verified for HFFs cultured on the stromal side of the AHAM. In summary, we have developed a rapid cost-effective method to produce AHAMs with a favorable adhesion potential, while preserving the biomechanical properties. Thus, our AHAM could be a an ideal candidate as a cell-matrix adhesion substrate to study cell adhesion and invasion as well as other functions in vitro under a tensile force that mimics the in vivo environment.