Mesenchymal stem cells (MSCs) have become a critical addition to all facets of tissue engineering. While most
Tissue engineering integrates the application of engineering and biological principles to study, design, develop, and repair biological structures. It is an iterative, objective-driven process which has spurred the production of artificial skin [
Bone marrow-derived mesenchymal stem cells (MSCs) are less controversial to obtain and easier to control than embryonic stem cells [
While these advances are significant to the field, current research is lacking in relation to the three-dimensional culture of MSCs, their metabolic state, and how both of these considerations may affect their terminal differentiation. It is important to consider how cells will respond in three-dimensional culture compared with two-dimensional culture because it has been shown that extracellular matrix interactions cause many different cellular reactions related to differentiation, proliferation, growth, motility, and gene expression [
Consequently, the overall motivation of this series of experiments was to investigate the metabolic state of MSCs in culture in response to variations in glucose and fetal bovine serum (FBS) concentrations in both two-dimensional and three-dimensional culture conditions.
Murine MSCs (ATCC Number CRL-12424) were used for these experiments (ATCC, Manassas, VA, USA). MSCs were cultured in maintenance medium composed of glutamine-free Dulbecco’s modification of Eagle’s medium (DMEM) (90-113-PB, Mediatech, Herndon, VA) supplemented with 10% FBS (Invitrogen, Carlsbad, Calif, USA), 1% penicillin-streptomycin (10 mg/mL, Sigma-Aldrich, St. Louis, Mo, USA), and 0.1% amphotericin-B (250
Circular collagen gels with an area of 2.0 cm2 and a height of 0.5 mm were used as three-dimensional scaffolds for culturing the MSCs. Collagen was harvested from porcine tendon and placed in 0.5 M acetic acid (Mallinckrodt Baker, Phillipsburg, NJ). The acid soluble collagen was neutralized with 2.0 M sodium hydroxide (NaOH) (Mallinkrodt Baker, Phillipsburg, NJ). High-density gels at 5.0 mg/mL and low-density gels at 2.5 mg/mL were prepared in 24-well plates. Each gel was formed using 100
The 24-well plates were used for cell number determination for both two- and three-dimensional cultures. Briefly, Hoechst dye (Invitrogen, Carlsbad, Calif, USA) was added to DMEM for a final concentration of 5.0
The same 24-well plates used to assess cell population were used to observe cell viability qualitatively using fluorescence imaging. Briefly, both calcein-AM and ethidium bromide (Fluka, St. Louis, MO) were added to DMEM for a final concentration of 2.5
Medium samples were tested for glucose concentration using a fluorometric assay kit (Invitrogen, Carlsbad, CA). The kit indicates the presence of glucose by activating fluorescent resorufin, which has absorption maxima at 571 nm and emission maxima at 585 nm. Samples were diluted with deionized water so that the glucose concentration would fall within the optimum detection range of the kit, between 3.0
Medium samples were tested for lactate concentration using a spectrophotometric assay kit (Instruchemie, The Netherlands). The kit determines the concentration of lactate by detecting an increase in the absorbance of reduced nicotinamide adenine dinucleotide (NAD) at 340 nm, indicating the amount of lactate originally present in the sample. The NAD absorbance at 340 nm was measured with a spectrophotometer (Thermo Electron, Finland). Extracellular samples were diluted with an equal amount of deionized water so that the lactate concentration would fall within the optimum detection range of the kit, between 0.22 mM and 13.32 mM. Background measurements for extracellular samples were taken using ordinary DMEM. These background measurements were subtracted from their respective absorption measurements. Background-subtracted measurements were converted to lactate concentrations using a calibration curve.
Extracellular and intracellular samples were tested for pyruvate concentration using a spectrophotometric assay kit (Instruchemie, The Netherlands). The kit determines the concentration of pyruvate by detecting a decrease in the absorbance of oxidized nicotinamide adenine dinucleotide
Pyruvate measurements were taken using both an intracellular and extracellular approach. The extracellular approach was performed in both the two- and three-dimensional conditions with six samples accorded to each condition. The extracellular approach consisted of measuring the pyruvate concentration within the same medium samples used for glucose and lactate measurements. The intracellular approach was only performed for the two-dimensional case. Briefly, each well of the 6-well plate was rinsed twice with PBS after the medium sample was extracted for extracellular measurements. Next, the cells were removed from the wells using cell scrapers, 0.5 mL of cold perchloric acid was added to each well, and the cell/acid solution was placed in microcentrifuge tubes. Next, the samples were subjected to vortex for thirty seconds, placed in a −20°C freezer for five minutes, and then centrifuged for ten minutes at 1700 ×g (Beckman Coulter, Fullerton, CA). The resulting supernatant was used to assay for pyruvate concentration. The samples did not have to be diluted in either the extracellular or intracellular case. Background measurements for intracellular samples were taken using deionized water samples. Background measurements for extracellular samples were taken using ordinary DMEM. These background measurements were subtracted from their respective absorption measurements. Background-subtracted measurements were converted to lactate concentrations using a calibration curve.
Viscous drag coefficients were obtained for trypan blue (Sigma-Aldrich, St. Louis, MO) diffusing through 2.5 mg/mL and 5.0 mg/mL collagen gels. Clear 1.0 cm diameter test tubes were used to hold 2.0 mL of each gel concentration (
StatView version 5.0.1 was used to apply a one factor analysis of variance (ANOVA) and Tukey-Kramer post-hoc test to the data to evaluate statistical significance between the groups at a significance level of
The cell population at the start of the two-dimensional experiment was even between the different experimental groups (Figure
Cell population (mean ± standard deviation) over the course of six days while in two-dimensional culture (
Cell population (mean ± standard deviation) over the course of six days while in three-dimensional culture (
The two-dimensional, day zero cell viability indicates high cell number and confluency before the cells were exposed to any experimental treatments (images not shown). However, by day two the cells in the 0.5 mM and 1.0 mM environments exhibited low viability while those in the 5.0 mM and 25.0 mM environments exhibited high viability (Figure
Fluorescence microscopy images indicating the viability of BMSCs in two-dimensional culture after the second day (increasing glucose concentrations from left to right; increasing serum concentrations from top to bottom). Calcein-AM indicates living cells by emitting green light, while ethidium bromide indicates dead cells by emitting red light. The 0.5 mM and 1.0 mM environments resulted in low viability and sparse cell distribution. The 5.0 mM and 25.0 mM environments resulted in high viability and the maintenance of high confluency. Varying the serum concentration did not produce a noticeable difference in the viability. This result was consistent throughout the course of the week (20x objective; scale bar represents 250
Fluorescence microscopy images indicating the viability of BMSCs in three-dimensional culture after the second and sixth days (increasing glucose concentrations from left to right; increasing collagen densities from top to bottom). Calcein-AM indicates living cells by emitting green light, while ethidium bromide indicates dead cells by emitting red light. After day 2, the 0.5 mM environment resulted in low viability, while the 25.0 mM environment resulted in high viability. By day 6, the 0.5 mM environment and the 25.0 mM−2.5 mg/mL group resulted in low viability, while the 25.0 mM−5.0 mg/mL group resulted in high viability and confluency. (40x objective; scale bar represents 100
In both the two- and three-dimensional cases, glucose consumption was the highest in the 25.0 mM groups (data not shown). For the two-dimensional case, there were three cases of significant differences between serum groups and a number of significant differences between glucose groups. The three-dimensional case did not yield a significant difference between collagen groups or glucose groups until day six, when the consumption was higher for the cells in the 5.0 mg/mL collagen matrix than the cells in the 2.5 mg/mL collagen matrix.
Groups with higher glucose consumption would be expected to yield higher lactate production as the endpoint of glycolysis. This behavior was shown in the lactate production data, with lactate production increasing with increasing glucose consumption in both two- and three-dimensional cultures (data not shown). At all FBS levels on days 4 and 6 the 0.5 mM and 1.0 mM glucose groups exhibited significantly lower lactate production than the 15 mM group at the
As pyruvate is not likely to accumulate, being converted to either lactate or acetyl coenzyme A, the amount of pyruvate produced would not be expected to have a consistent pattern amongst the experimental glucose groups. This random behavior was shown in both the intracellular and extracellular, two-dimensional culture data, especially in terms of statistical significance (Figure
Extracellular pyruvate production after the final two days of two-dimensional culture (
Lactate-to-pyruvate ratios for two-dimensional culture showed an increasing trend with an increase in the starting glucose concentration, indicating that higher glucose cultures were more anaerobic (Table
Summary of two-dimensional L/P ratios and L/G ratios. L/P ratios indicate that groups with a higher starting glucose concentration exhibited a more anaerobic characteristic. This trend was similar in the L/G ratios for the 2% serum group, but not for either the 5% or 10% serum groups. (a: significantly different from corresponding value in the 5.0 mg/mL group; b: significantly different from the 25.0 mM value within the same collagen group; f: significantly different from 5.0 mM value within same serum group;
Two-dimensional lactate-phosphate (L/P) ratios.
Glucose conc. 2% serum | 5% serum | 10% serum | ||||||||
Mean | St. Dev. | Mean | St. Dev. | Mean | St. Dev. | |||||
Day 4 | 0.5 mM | 2.41 | 2.36 | 3 | 2.85 | 2.92 | 6 | 5.92 | 3.48 | 3 |
1.0 mM | 0.59b | 0.93 | 9 | 4.77b | 4.18 | 10 | 5.61 | 3.02 | 7 | |
5.0 mM | 8.79 | 8.52 | 12 | 23.46 | 17.01 | 10 | 52.48 | 34.08 | 8 | |
25.0 mM | 14.98 | 13.12 | 12 | 36.05 | 34.25 | 12 | 50.01 | 32.62 | 11 | |
Day 6 | 0.5 mM | 0.17 | 0.34 | 8 | 9.57 | 17.89 | 4 | 4.15 | 4.83 | 4 |
1.0 mM | 10.06 | 25.28 | 11 | 26.58 | 11.34 | 6 | 31.76 | 37.43 | 4 | |
5.0 mM | 51.37 | 17.22 | 6 | 134.35 | 111.21 | 8 | 234.68 | 549.5 | 10 | |
25.0 mM | 71.49 | 59.71 | 10 | 113.18 | 76.5 | 12 | 114.8 | 152.36 | 9 |
Two-dimensional lactate-glucose (L/G) ratios.
Glucose conc. 2% serum | 5% serum | 10% serum | ||||||||
Mean | St. Dev. | Mean | St. Dev. | Mean | St. Dev. | |||||
Day 4 | 0.5 mM | 0.72a | 0.45 | 8 | 1.47 | 0.09 | 6 | 0.57 | 0.62 | 12 |
1.0 mM | 0.52 | 0.55 | 12 | 1.81b | 1.44 | 12 | 2.71 | 3.65 | 10 | |
5.0 mM | 1.19 | 0.89 | 12 | 2.2 | 1.42 | 12 | 2.09 | 1.44 | 12 | |
25.0 mM | 1.53 | 1.56 | 12 | 1.9 | 1.49 | 12 | 2.85 | 2.67 | 12 | |
Day 6 | 0.5 mM | 0.65f | 0.49 | 9 | 1.94 | 2.52 | 8 | 0.59 | 0.63 | 12 |
1.0 mM | 0.47f | 0.49 | 12 | 3.65 | 2.6 | 12 | 13.53 | 22.57 | 9 | |
5.0 mM | 2.39 | 1.75 | 12 | 2.99 | 2.29 | 12 | 2.69 | 2.07 | 12 | |
25.0 mM | 1.56 | 1.2 | 12 | 1.95 | 1.45 | 12 | 2.32 | 1.91 | 12 |
Summary of three-dimensional metabolic ratios. There was no clear trend in the data. (a: significantly different from corresponding value in the 5.0 mg/mL group; b: significantly different from the 25.0 mM value within the same collagen group;
Experimental Group L/P Ratio | L/G Ratio | ||||||
Mean | St. Dev. | Mean | St. Dev. | ||||
Day 4 | 0.5 mM/low density | 2.30b | 5.66 | 12 | 1.93 | 2.29 | 6 |
25.0 mM/low density | 11.92a | 5.41 | 7 | 0.92a | 0.97 | 12 | |
0.5 mM/high density | 2.15b | 2.42 | 12 | 15.47b | 11.20 | 8 | |
25.0 mM/high density | 18.73 | 4.95 | 6 | 1.11 | 0.88 | 12 | |
Day 6 | 0.5 mM/low density | 74.75a,b | 77.85 | 3 | 17.30a | 47.92 | 12 |
25.0 mM/Low Density | 8.49 | 9.85 | 12 | 2.34 | 2.44 | 8 | |
0.5 mM/high density | 2.92 | 1.15 | 6 | 11.04b | 8.60 | 8 | |
25.0 mM/high density | 6.32 | 7.23 | 12 | 1.02 | 1.22 | 12 |
Drift velocities and accelerations indicate that the trypan blue molecules traveled quicker through the 2.5 mg/mL gel than the 5.0 mg/mL gel (Table
Summary of results pertaining to trypan blue diffusion through collagen gels. The diffusivity for both cell densities is roughly the same. *Relative drag coefficient and relative diffusivity values were calculated by normalizing against that of water (there was no significant difference between the diffusivity values of the 2.5 mg/mL and 5.0 mg/mL groups).
Water | 2.5 mg/mL | 5.0 mg/mL | ||||
Mean | Range | Mean | Range | Mean | Range | |
Drift velocity (m/s) | 0.04 | N/A | ||||
Accel. (m/s2) | 0.0 | N/A | ||||
Relative drag coef.* | 1.0 | N/A | ||||
Relative diffusivity* | 1.0 | N/A |
The overall purpose of this research was to investigate possible differences in the metabolic state of MSCs when cultured in ordinary two-dimensional culture versus collagen-based three-dimensional culture. The two-dimensional experiments consisted of subjecting the cells to various medium types that differed in both glucose and serum concentrations. The three-dimensional experiments subjected cells to two different glucose concentrations and two collagen densities. The three-dimensional cultures were only subjected to variation in glucose concentration as glucose appeared to have the most critical effect on cell growth, viability, and metabolic state. The 5% serum concentration was chosen because of its moderation between the 2% and 10% serum groups. The results highlight relative differences in the metabolic state of MSCs when cultured within these differing conditions. Overall, the two most important factors in affecting the cell population were starting glucose concentrations and the addition of a three-dimensional matrix.
This result could have important implications for stem cell therapies, as differences in metabolic state during culture may affect the terminal differentiation of the MSCs. Human adipose-derived stem cells have already shown evidence of this as their differentiation into osteogenic cells is altered by changing their metabolic conditions [
Anaerobic glycolysis and the aerobic citric acid cycle are the two primary means for cells to obtain energy. They can be distinguished by the fate of the glucose consumed by the cells. In glycolysis, the glucose is converted to pyruvate, which is then converted to lactate as a dead-end product. However, in the citric acid cycle, the pyruvate produced at the conclusion of glycolysis enters into the citric acid cycle in the form of acetyl coenzyme-A. The relative metabolic state—whether aerobic or anaerobic—of a cell culture can be assessed after measuring the glucose consumed and lactate produced; the higher the amount of lactate, the more likely that the relative metabolic state anaerobic.
These data elucidate the ability of growth medium glucose and serum concentrations to alter the metabolic state of the MSCs. It has already been shown that adipose-derived stem cells from rabbits [
The setup of these experiments had several advantages. The primary advantage was the use of the same cell type for both two-and three-dimensional culture so as to avoid potential significant differences in metabolism between different cell types. This reality was highlighted by Sander and Nauman, as it was found that culturing rat astrocytes produced lower lactate-to-glucose ratios than both porcine lamina cribrosa cells and porcine scleral fibroblasts after one week of culture in environments with different glucose and oxygen levels [
These experiments also possessed a number of limitations. The purpose of the trypan blue diffusion experiments was to investigate the possibility that the metabolites were trapped within the collagen matrix, thus avoiding detection by the assays. However, trypan blue—molecular weight of ~960 g/mol—was able to seep approximately 1.0 mm through both the 2.5 mg/mL and 5.0 mg/mL collagen gels in about fifteen minutes (data not shown). Therefore, it is reasonable to assume that glucose, lactate, and pyruvate—all with molecular weights of less than 200 g/mol—were able to traverse the collagen matrix and be subject to detection. There are two other circumstances that indicate that the thickness of the gels did not prevent adequate transport of lactate out of the gels. The first is the recognition that calcein-AM—molecular weight ~995 g/mol—did not have any trouble diffusing into the gel within fifteen minutes in order to take fluorescence images of the cells. The second comes from the result of the nutrient supply study by Horner and Urban [
The results of this series of experiments serve as a baseline characterization of MSCs in culture before being subjected to either an extracellular matrix or a differentiation scheme, and they will be especially beneficial to the tissue engineering field. While it is well documented that extracellular matrices have significant effects on several cell types [
While, there is always great risk associated with transitioning from in vitro to in vivo studies because the environments are different in many ways, from practical and ethical standpoints, it would be advantageous to have a means of investigating a set of experimental variables on an artificial tissue analogue prior to implementing the same variables in an animal system; this intermediate step could accelerate research findings and prevent the unnecessary deaths of the animals should the results from this in vitro tissue reveal dangerous effects that the two-dimensional case alone did not reveal. Consequently, this intermediate step could potentially decrease costs associated with in vitro-to-in vivo experimentation.
The results of this study indicate that the amount of glucose provided and the presence of an extracellular matrix do indeed affect the metabolic state and viability of cells in culture. Providing larger concentrations of glucose may be a way to push the cells toward the anaerobic state, all else being equal. Additionally, the results demonstrated that the three-dimensional culture of MSCs may have to take place within a relatively stiff matrix or a composite of collagen hydrogel and collagen fibers [
The authors would like to acknowledge the funding from the National Science Foundation Graduate Research Fellowship Program. They would also like to thank Dr. Sherry Voytik-Harbin in the Weldon School of Biomedical Engineering at Purdue University for her donation of the MSCs used in these experiments.