Quantitatively analyzing the protective effect of mesenchymal stem cells on cardiomyocytes in single-cell biochips

Abstract

To understand how stem cells benefit native cardiomyocytes is crucial for cell-based therapies to rescue cardiomyocytes (CMCs) damaged during heart infarction and other cardiac diseases. However, the current conclusions on the protective effect of mesenchymal stem cells (MSCs) were obtained by analyzing the overall amount of protein and factor secretion in a conventional co-culture system. These results neglected the heterogeneity of MSC population and failed to determine the importance of cellular contact to the protective effects. To address these issues, we have constructed two biochips by microfabrication methods and laser-guided cell micropatterning technique. Using the biochips, the protective effect of MSCs on CMCs can be quantitatively analyzed at single-cell level with defined cellular contact. The role of cellular contact on protective effect can be clarified according to our statistical results.

Introduction

Myocardial infarction is characterized pathologically by loss of functional cardiomyocytes (CMCs), which have low potential for repair and regeneration within the damaged area. It was reported that spontaneous mobilization of bone marrow derived mesenchymal stem cells (MSCs) after acute myocardial infarction contributed to favorable heart remodeling (Leone et al. 2005). The MSC transplantation benefited to the infarcted myocardium not only by stem-cell differentiation into cardiovascular cells but also by their ability to release large amounts of angiogenic, anti-apoptotic, and mitogenic factors (Hou et al. 2007). These findings on MSC protective effects were confirmed by research involving both in vitro and in vivo studies, which demonstrated that MSCs would secrete cytokines and prosurvival proteins to prevent apoptosis of CMCs in early stages of myocardial infarction and improve cardiac function in later stages (Xu et al. 2007; Abarbanell et al. 2009).

The protective effect of MSCs is partially derived by secreting multiple soluble factors that may act through a reduction in infiltration of inflammatory neutrophils, inactivation of fibrogenic cells and scarring, stimulation of angiogenesis and vascularization, or recruitment and activation of resident cardiac stem cells (Rogers et al. 2011). Other reports suggest that protective effects might depend on the intercellular communications through direct cell–cell contact, in which junctional proteins, including connexins and cadherins, delivered the anti-apoptotic factors and activated the reparative pathways to damaged CMCs (Cselenyak et al. 2010). However, current in vitro results were determined in the conventional culture model in which large numbers of cells were analyzed in a random bulk experiments. These average results opposed to the heterogeneous population of MSCs, whereas only portion of MSCs may contribute to the protective effect. Furthermore, conventional in vitro model involved random cell shape, random spatial distribution, and random cell–cell contact. Undefined mode of cell–cell contact made it difficult to elucidate the importance of cellular contact for MSC protective effect. Therefore, understanding the distribution and statistical significance of protective effect over a whole MSC population is important to classify MSC heterogeneity and to detect variability of protective effect under different contact modes.

Biochips with well-defined cell arrangements have been developed for isolating and analyzing large quantities of individual cells in identical and controlled environments, which allow the assessment of contact-mediated cell–cell interaction in a dynamic and high-throughput fashion (Moroni and Lee 2009; Dworak and Wheeler 2009; Klinger and Bursac 2008). However, the conventional microfabrication-based biochips achieved by trapping individual cells from a massive cell flow for cell seeding cannot provide single cell-based heterotypic micropatterning with high spatial and temporal resolution. We developed a laser-guided cell micropatterning technique to achieve single-cell resolution micropatterning, which has been combined with microfabrication methods to construct biochips for single-cell analysis on neuronal network formation and its electrical activities (Pirlo et al. 2008, 2011b). In this report, two biochips were constructed to promote or prevent cellular contact between individual laser-patterned CMCs and MSCs. Each microwell containing one CMC and one MSC in the biochips can be used to mimic the condition with low nutrients supply and waste exchange. The morphology of CMCs and MSCs in the biochips was compared with the random culture through immunocytochemistry. The dependence of MSC protective effect on the cell–cell contact was systematically and quantitatively assessed.

Materials and methods

Cell preparation

CMCs were isolated and collected from three-day neonatal rats using a two-day protocol (Ma et al. 2011). CMCs were cultured using high glucose Dulbecco's Modified Eagle's Medium (DMEM) supplemented with 20 % (v/v) fetal bovine serum and 1 % (w/v) penicillin streptomycin. Commercial rat MSCs from bone marrow were purchased from ScienCell research laboratories and characterized by the company with antibodies to CD73, CD90, CD105, and oil red staining after adipocyte differentiation. The MSCs were cultured using mesenchymal stem cell medium (MSCM) from the same company and used for cell patterning before the fifth passage. The CMC culture medium was used as the coculture medium for the biochips during and after the laser-patterning procedure. All experiments that involve animal use were performed in compliance with the relevant laws and institutional guidelines. These experiments have been approved by Clemson University's Institutional Animal Care and Use Committee through protocol AUP2010-032.

Biochip construction

The microenvironment with defined geometrical restriction involved a microwell containing one CMC and one MSC. For each biochip, 9 × 9 identical microwells were created by attaching an elastomeric membrane with through holes onto a glass coverslip. Elastomeric membranes were microfabricated using polydimethylsiloxane (PDMS) with the standard procedure of photolithography and soft-lithography. The distance between two neighboring microwells was 200 μm. The microwell promoting the cellular contact between CMC and MSC was rectangular (50 μm × 25 μm), which confined the cell bodies of single CMC and MSC to ensure physical contact. The microwell preventing the cellular contact was designed in a dumbbell shape with two circles (30 μm diam.) connected by one channel (30 μm length × 15 μm width). The curve between the circle and the channel prevented a cell from contacting the cell in the other side. The depth of the microwell was controlled at 40 μm which restricted the cell inside the microwells and obstructed the nutrients transportation. The limited living space for two cells can be used to mimic the condition with low nutrients supply and waste exchange. Before being used, the biochips were cleaned, sterilized, and activated by oxygen plasma. Immediately following the plasma treatment, the biochips were coated with fibronectin to improve cell attachment and biocompatibility on the underlying glass substrate. The biochips were placed inside cell-deposition chambers, one biochip for each chamber, where individual MSCs and CMCs would be laser-patterned into the biochip's microwells.

Laser-guided cell micropatterning

The uniqueness of the laser-guided cell micropatterning technique in heterotypic, single-cell micropatterning with high temporal and spatial resolution made it essential to create these single-cell biochips. The details of the system design and laser-patterning procedures were discussed in our previous paper (Pirlo et al. 2011a). A typical cell-micropatterning procedure began with preparation of the biochip (i.e., a patterning substrate with a matrix of microwells) and the cell suspension. To create a heterotypic cell pair, one CMC was first laser-patterned into the left side of one microwell on the biochip and then one MSC was patterned into the right side of the microwell containing the CMC, shown in Fig. 1a. The orientation of the microwell can be determined bythe triangular features established outside the central region, and thus CMCs could be distinguished from MSCs in the following experiments. Homotypic cell pairs that were created at the same time served as positive (two CMCs in one microwell) and negative (two MSCs in one microwell) controls. Figure 1b, c showed that only one CMC and one MSC inside one microwell right after the laser-patterning process. After one day culture, viable heterotypic cell pairs in the contact-promoting biochips filled the entire microwell and formed broad contact on their cell membranes as shown in Fig. 1d. The noncontact cells in the contact-preventing biochips stayed within their own side of the dumbbell-shape microwell, attached to the wall, and communicated with the other cell in the microwell only through soluble factors, shown in Fig. 1e. In our previous research to pattern cells onto a glass coverslip, the viability of cells with laser-patterning procedure did not show significant difference from the cells without patterning (Rosenbalm et al. 2006).

Fig. 1.

a A schematic image of laser-patterning procedure. A MSC (pink sphere) was being trapped by the laser beam and deposited into the microwell containing a CMC (green sphere); the heterotypic cell pairs in b a contact-promoting biochip and c a contact-preventing biochip right after the laser-patterning procedure; after one day culture, d two cells in a contact-promoting microwell spread and formed broad contact area, and e two cells in a contact-preventing microwell stayed in their compartments and did not contact each other

Immunocytochemistry

The cellular morphology of CMCs and MSCs was examined by immunostaining in the random culture and biochips. Cells were fixed, permeabilized and blocked in 4 % (v/v) paraformaldehyde (10 min), 0.1 % Triton X-100 (15 min) and 2 % (w/v) BSA with 4 % (v/v) goat serum, respectively. CMCs were stained by α-actinin (primary antibody: mouse anti-α-actinin, Invitrogen; secondary antibody: donkey anti-mouse IgG-FITC) and MSCs were stained by MSC marker, CD90, separately (primary antibody: mouse anti-CD90, Santa Cruz Biotechnology Inc.; secondary antibody: goat anti-mouse IgG-TR). After three washes with PBS to remove the secondary antibody, the slides were prepared with ProLong antifade kit mounting medium (Invitrogen) and observed under the fluorescent microscope.

Cell viability evaluation

Immediately after the laser-patterning procedure, the biochips were transferred from cell-deposition chambers to commercial 35 mm Petri dishes containing CMC culture medium. All the conventional cultures and biochips were incubated at 37 °C in a humidified atmosphere of 95 % air/5 % CO₂ to prevent evaporation, and culture medium was changed every two days. We followed the normal cell culture procedures to maintain the same conditions as the conventional cultures, so that the effect of microwell's geometrical limitation can be emphasized.

Typically, freshly isolated CMCs required about 6 h to attach and start spreading on a glass surface coated with fibronectin, whereas MSCs required less than 2 h in the conventional cultures. In the biochips, the cells required more than 12 h for attachment. The cell viability was evaluated on the third day (48 h after laser-patterning procedure), the round cells inside the microwells were considered to be dead cells, which were confirmed by live/dead assays. Thus, one microwell containing two spread cells was considered as a successfully survival event. According to this criterion, survival events within one biochip were counted to calculate the rate of cell viability by the equation (1) for each biochip. For each contact mode, 20 identical biochips, which had 81 cell pairs, were analyzed to collect the statistical results. Data were expressed as means ± SEM. Statistical comparisons of data were performed using Student's t test comparison between two groups. A value of p < 0.05 was considered statistically significant.

$$ViabilityRatio=\frac{NumberofViableCellPairs}{NumberofPatternedCellPairs}$$ (1)

Results

In the random culture, shown in Fig. 2a, b, MSCs and CMCs spread on the glass substrate without any geometrical restrictions. Thus, cell shape was irregular and distribution was undefined. Both contact and noncontact phenomenon existed in one culture. However, in our biochips, cellular contact was highly confined by restricting the cell growth and spread within a limited microenvironment. In contact-promoting biochips, homotypic, Fig. 2c, d, and heterotypic, Fig. 2e, f, cell pairs were forced to contact to form contact mode. In contact-preventing biochips, two cells were separated into two circle compartments to form noncontact mode as shown in Fig. 2g, h. CMCs showed clear sarcomere structure by staining α-actinin, and MSCs showed entire cell shape by staining CD90.

Fig. 2.

a α-actinin staining for CMCs and b CD90 staining for MSCs in the conventional cultures; in contact-promoting microwells, CMCs showed α-actinin positive in c homotypic CMC pair and e heterotypic cell pair, and MSCs showed CD90 positive in d homotypic MSC pair and f heterotypic cell pair; in contact-preventing microwells, a homotypic MSC pair showed CD90 positive staining and a heterotypic cell pair showed α-actinin positive for the CMC and α-actinin negative for the MSC

The viability of freshly isolated CMCs in our conventional cell cultures was typically ~85 %, and the viability of MSCs after each passage was as high as 95 %. It was difficult to count and calculate the viability of CMCs and MSCs in the conventional co-cultures with two types of cells. Laser-patterning only one cell in one contact-promoting microwell, the viability ratio of single CMCs and MSCs was 72 ± 7.1 % (P₁) and 86 ± 4.4 % (P₂), respectively. The significant reduction of cell viability suggested that the limited geometric space and isolated cell connections inside a microwell might obstruct the exchange of nutrients and metabolic waste and thus cause lower cell viability compared to conventional cell culture.

Assuming that viability of each cell was considered as an independent event, the viability ratio of two cells in one contact-promoting microwell can be calculated as $P_1^2=51.84\%$ for homotypic CMC pairs, $P_2^2=73.96\%$ for homotypic MSC pairs, and P₁ × P₂ = 61.92 % for heterotypic cell pairs. However, in our experimental results, the viability ratio was 39 ± 2.6 % (P₃) for homotypic CMC pairs, 67 ± 3.4 % (P₄) for homotypic MSC pairs, and 58 ± 3.1 % for heterotypic cell pairs. This suggests that the independent assumption is a fault, and the viability ratio of the two cells in one microwell is dependent to cell–cell and cell–microenvironment interactions. The dramatic decrease of viability ratio of cell pairs was caused by more nutrients consumption within the microwell for two cell survival. High metabolic characteristics of CMC pairs required even more nutrients to maintain their physiological stability, which induced their lower viability ratio than that of MSC pairs.

Since the calculated viability value of heterotypic cell pairs neglected the extra nutrient consumption inside the microwells, the experimental viability ratio of homotypic cell pairs can be used for further calculations. The viability ratio of single CMC and MSC can be respectively calculated as $P_5=\sqrt{P_3}$ and $P_6=\sqrt{P_4}$ with a premise two cells in one microwell. Therefore, the viability ratio of heterotypic cell pair can be calculated as P₅ × P₆ = 51.12 %, and this calculated value is statistically significantly lower than our experimental value of 58 %, which indicated that MSCs and CMCs in one microwell cannot be simplified as independent event. Therefore, the increasing viability ratio for the CMCs from 39 ± 2.6 % in the homotypic CMC pairs to 58 ± 3.1 % in the heterotypic cell pairs indicated that contacting MSCs can rescue CMCs in a microenvironment with poor nutrients support.

For a contact-preventing microwell, the area of the dumbbell (about 1,950 μm²) was 56 % larger than that of the rectangular contact-promoting microwell (about 1,250 μm²). It was possible that this relatively larger microwell provided comparatively better support for the nutrients and waste transportation. The viability ratio of the noncontact homotypic cell pairs was 42 ± 2.8 % for CMCs and 69 ± 3.9 % for MSCs, which was slightly higher than that in the contact-promoting microwells. However, only 52 ± 3.6 % of noncontact heterotypic cell pairs survived. Calculating the viability ratio of heterotypic cell pairs in contact-preventing microwells, the value of 53.83 % showed no significant difference comparing to experimental value of 52 %, which suggested that two cells can be considered as independent events in the noncontact mode. Therefore, the significant increase of viability ratio from noncontact mode (52 %) to contact mode (58 %) suggested that the ability of MSCs to rescue CMCs was dependent on cellular communication through physical contact. In Table 1, the experimental viability ratio of homotypic and heterotypic cell pairs was summarized for two types of biochips.

Table 1.

Viability ratio of cell pairs in contact-promoting/preventing biochips

Viability ratio	Homotypic CMC pairs (%)	Homotypic MSC pairs (%)	Heterotypic cell pairs (%)
Contact-promoting biochips	39 ± 2.6	67 ± 3.4	58 ± 3.1*
Contact preventing biochips	42 ± 2.8	69 ± 3.9	52 ± 3.6*

^*p < 0.05 based on Student's t test

Discussions and conclusions

Stem cells in the contact co-culture system express several cardiac-specific proteins, including myosin heave chain, troponin I, troponin T, ANP, and connexin 43, which were absent in the stem cells in the noncontact system (Wang et al. 2006; Xu et al. 2004). Cell contact co-culture is therefore necessary for inducing stem cells to acquire the phenotypic characteristics of CMCs; simple chemicals and secreted factors from CMCs are insufficient. On the other hand, the soluble factors secreted by MSCs also may not sufficient to rescue the CMCs within the infarction area only through paracrine fashion. Currently, the evaluation of MSC protective effect in a massive cell sample involved multiple contact modes (e.g., one stem cell contacts many other stem cells or CMCs) including noncontact cases. Therefore, the protective effect through directly cell–cell contact cannot be emphasized in the conventional culture system. Without spatial control of cell arrangement, it is not possible to correlate protective effect to a specific contact mode. In our research, microfabrication methods and laser-guided cell micropatterning technique make it possible to control cellular contact to quantitatively and systematically study the importance of cellular contact on protective effect of MSCs on CMCs in a low nutrient microenvironment.

Our cell-viability study indicated that MSCs in contact-promoting biochips rescued the contacted CMCs in an unhealthy microenvironment. In addition, viability ratio of heterotypic cell pairs in the contact-preventing biochips was 6 % less than that in the contact-promoting biochips. Thus, the protective effect of MSCs appeared to be dependent on cell–cell connections to improve CMC viability. These results suggested that the beneficial effect of stem-cell grafting might be based not only on improved neovascularisation and replacement of lost cells but also on rescuing the damaged host CMCs in a low-nutrients environment under ischemic conditions.