There is considerable interest in the biology and therapeutic potential of adult stem cells from bone marrow stroma, variously referred to as mesenchymal stem cells or marrow stromal cells (MSCs). Human MSCs can expand rapidly in culture, but the rate of expansion and the yields of multipotential progenitors are inversely related to the plating density and incubation time of each passage. We have defined conditions for optimizing the yields of cultures enriched for early progenitors. Also, we developed a simple method for assessing the quality of the cultures by phase-contrast microscopy and image analysis or by forward light scatter in a flow cytometer. The cells expanded most rapidly on day 4 after plating, with a minimum average doubling time of about 10 hours for cells initially plated at 10 or 50 cells/cm2. After plating the cells at 1 to 1,000 cells/cm2, the cultures underwent a time-dependent transition from early progenitors, defined as thin, spindle-shaped cells (RS-1A), to wider, spindle-shaped cells (RS-1B), and to still wider, spindle-shaped cells (RS-1C). Assays for adipogenesis demonstrated that the adipogenic potential of cultures was directly related to their ability to generate single-cell-derived colonies and their enrichment for RS-1A cells. In contrast, cultures enriched for RS-1B cells showed the greatest potential to differentiate into cartilage in a serum-free system. The results indicate that, when preparing cultures of human MSCs, it is necessary to compromise between conditions that provide the highest overall yields and those that provide the highest content of early progenitor cells.
One strategy for cell and gene therapy is to use adult stem cells from bone marrow stroma [1-4], referred to as mesenchymal stem cells or marrow stromal cells (MSCs). Human cells are relatively easy to obtain from a small aspirate of bone marrow. Also, they are relatively easy to expand in culture under conditions in which they retain some of their potential to differentiate into multiple cell lineages that include osteoblasts , adipocytes , chondrocytes , myoblasts , and early progenitors of neural cells . However, as cultures of the cells are expanded under standard conditions, they lose their proliferative capacity and their potential to differentiate into lineages such as adipocytes and chondrocytes [10, 11]. We previously observed that the human cells proliferate most rapidly and maximally retain the multipotentiality if passed by plating the cells at low densities [12, 13].
From these observations and the observations of previous investigators [2-4, 14, 15], it is apparent that a large number of variables and parameters must be considered in expanding MSCs for experimental and clinical purposes. First, there is a large sampling error in terms of the yield and quality of MSCs in bone marrow aspirates, even in aspirates obtained from the same donor at the same time [10, 12, 15]. Therefore, each preparation of MSCs must be standardized. Second, the morphology and other properties of the cells change as single-cell-derived colonies are expanded. In particular, small and rapidly self-renewing cells (RS cells) that have the highest multipotentiality are gradually replaced by slowly replicating large, and apparently mature, cells (mMSCs) that have lost most of their multipotentiality but can still differentiate into osteoblasts as a default pathway [10-12].
In the experiments reported here, we examined several of these variables and parameters for culturing human MSCs and defined improved conditions for obtaining standardized preparations of the cells.
Materials and Methods
Isolation and Cultures of Human MSCs
To isolate human MSCs, bone marrow aspirates were taken from the iliac crest of normal adult donors after informed consent and under a protocol approved by an Institutional Review Board. Nucleated cells were isolated with a density gradient (Ficoll-Paque; Pharmacia; Peapack, NJ; http://www.pnu.com) and resuspended in complete culture medium: alpha minimal essential medium (αMEM; GIBCO/BRL; Carlsbad, CA; http://www.invitrogen.com), 20% fetal bovine serum (FBS) lot selected for rapid growth of MSCs (Atlanta Biologicals, Inc.; Norcross, GA; http://atlantabio.com/default.htm), 100 units/ml penicillin, 100 μg/ml streptomycin, and 2 mM L-glutamine (GIBCO/BRL). All of the nucleated cells (25 to 71 million) were plated in 20 ml medium in a 180-cm2 culture dish and incubated at 37°C with 5% CO2. After 24 hours, nonadherent cells were discarded, and adherent cells were thoroughly washed twice with phosphate-buffered saline (PBS). The cells were incubated for 4-11 days, harvested with 0.25% trypsin and 1 mM EDTA for 5 minutes at 37°C, and replated at 3-50 cells/cm2 in an intercommunicating system of culture flasks (6,300 cm2; Cell Factory, Nunc; Naperville, IL). After 7 to 12 days, the cells were harvested with trypsin/EDTA, suspended at 1 × 106 cells/ml in 5% dimethylsulfoxide and 30% FBS, and frozen in 1-ml aliquots in liquid nitrogen (passage 1 cells). To expand a culture, a frozen vial of MSCs was thawed, plated in a 60 cm2 culture dish, and incubated for 4 days (passage 2 cells).
Culture Density and Proliferation
MSCs were cultured at 10 cells/cm2, 50 cells/cm2, 100 cells/cm2, and 1,000 cells/cm2 in 60-cm2 dishes (Corning; Corning, NY; http://www.corning.com). Cell morphology was then observed, and pictures were taken over the next 12 days under phase-contrast microscopy. Each day, cells from three plates from each culture density were harvested and counted with a hemocytometer.
Measurements of Cell Areas and Widths
Photographs were taken of typical areas of culture plates. The areas and the maximal widths perpendicular to the long axes of individual cells were measured quantitatively using a computerized imaging system (Image-Pro Plus 4.1; MediaCybernetics; Silver Spring, MD; http://www.mediacy.com). Mitotic cells were ignored.
MSCs were detached with EDTA/trypsin, suspended in PBS, and assayed in a flow cytometer (FACS Vantage SE; Becton Dickinson; Franklin Lakes, NJ; http://www.bd.com). The average forward scatter was estimated using WinMidi software (Scripps Research Institute; San Diego, CA).
MSCs were cultured for 12 days, and then, 100 cells were plated on 60-cm2 dishes. The cultures were incubated for 14 days, the media were removed, and the dishes were stained with 0.5% Crystal Violet (Sigma; St. Louis, MO; http://www.sigmaaldrich.com) in methanol for 5 minutes. The cells were washed twice with distilled water, and the number of colonies was counted. Colonies less than 2 mm in diameter and faintly stained colonies were ignored.
Adipogenesis After High-Density Plating Assay
MSCs were plated at 50 cells/cm2 or 1,000 cells/cm2 and cultured in complete culture media for 4, 7, or 12 days in 60 cm2 dishes. Then, the cells were replated at 5,000 cells/cm2 in six-well culture dishes (Falcon; BD Biosciences; http://www.bdbiosciences.com) and incubated in adipogenic media that consisted of complete medium supplemented with 0.5 μM dexamethasone (Sigma), 0.5 mM isobutylmethylxanthine (Sigma), and 50 μM indomethacin (Sigma) . After 21 days, the adipogenic cultures were fixed in 10% formalin for over 1 hour and stained with fresh Oil Red-O solution for 2 hours. The Oil Red-O solution was prepared by mixing three parts stock solution (0.5% in isopropanol; Sigma) with two parts water and filtering through a 0.2-μm filter. Plates were washed three times with PBS and dried. In order to obtain quantitative data, 1 ml of isopropyl alcohol was added to the stained culture dish. After 5 minutes, the absorbance of the extract was assayed by a spectrophotometer at 510 nm after dilution to a linear range .
Adipogenesis in a Colony-Forming Assay
MSCs were plated at 50 cells/cm2 or 1,000 cells/cm2 and cultured in complete media for 12 days. About 100 MSCs were then transferred into 60-cm2 dishes and cultured in complete media for 12 days. The cells were transferred to adipogenic media for an additional 21 days. The adipogenic cultures were fixed in 10% formalin, stained with fresh Oil Red-O solution, and the number of Oil Red-O-positive colonies was counted. Colonies less than 2 mm in diameter or faint colonies were ignored. The same adipogenic cultures were subsequently stained with crystal Violet, and the number of total cell colonies was counted.
MSCs were plated at 50 cells/cm2 and cultured in complete media for 4, 7, or 12 days. For chondrocyte differentiation, a micromass culture system was used . Approximately 200,000 MSCs were placed in a 15-ml polypropylene tube (Falcon; Bedford, MA) and pelleted into micromasses by centrifugation at 450 g for 10 minutes. The pellet was cultured for 21 days in chondrogenic media that contained 500 ng/ml BMP-6 (R&D Systems; Minneapolis, MN; http://www.rndsystems.com) in addition to high-glucose (25 mM) Dulbecco's modified Eagle's medium supplemented with 10 ng/ml transforming growth factor beta 3, 10−7 M dexamethasone, 50 μg/ml ascorbate-2-phosphate, 40 μg/ml proline, 100 μg/ml pyruvate, and 50 mg/ml ITS+™Premix (Becton Dickinson) [11, 17]. For microscopy, the pellets were embedded in paraffin, cut into 5-μm sections, and stained with 1% toluidine blue (Richard Allan Scientific; Kalamazoo, MI; http://www.rallansci.com) and 1% sodium borate (Sigma) for 5 minutes.
Effect of Plating Density on Expansion of MSCs in Culture
To select a preparation of MSCs for detailed examination, bone marrow aspirates were obtained from five volunteers (Table 1). There was no apparent correlation between yield of nucleated cells and volume of marrow obtained with an aspirate. Also, there was no apparent correlation between the rates at which the nucleated cells expanded and their sensitivity to initial plating density (Fig. 1A and 1B). We selected cells from one donor (89L) for further study, primarily because they expanded at an average rate and appeared to be slightly less sensitive to plating density than cells from the other samples.
Table Table 1.. Summary of data from 5 samples of bone marrow
Donor age (years)
Aspirated volume of bone marrow (ml)
Yield of nucleated cells (millions)
Culture periods for passage 0 cells (days)
Total number of passage 0 cells (millions)
Plating density for passage 1 (cells/cm2)
Culture periods (days)
Yield of cells per a cell-factory (millions)
Fold increase of passage 1 cells
Passage 3 (89L) cells plated at densities of 10 cells/cm2 expanded approximately 500-fold in 12 days, whereas the same cells plated at 1,000 cells/cm2 expanded approximately 30-fold (Fig. 2A). However, cells plated at 1,000 cells/cm2 yielded 1.6 million cells per 60-cm2 dish, whereas cells plated at 10 cells/cm2 yielded only 0.3 million cells (Fig. 2B). At day 7, there was no significant difference in fold expansion between cells plated at 10 and 100 cells/cm2. The peak doubling rate per day was seen on day 4 (Fig. 3). The doubling rate for cells plated at either 10 or 50 cells/cm2 was about 2.5 in 24 hours, indicating that the average doubling time on day 4 was about 10 hours. The doubling rate per day was less in cells plated at 100 or 1,000 per cm2, but the peak rate was still observed on day 4. The potential for the cells to generate single-cell-derived colonies (colony-forming units [CFUs]) was higher for cells expanded by initial plating at the lower densities (Fig. 4).
Change in Cell Morphology with Plating Density and Incubation Time
Visual evaluation of cultures of human MSCs by phase-contrast microscopy has long been used to demonstrate changes in the morphology of the cells with time in culture and with passage of the cells . In reevaluating cultures as a function of both initial plating density and incubation time, we observed a transition among three morphologically distinct cell types (Fig. 5A and 5B): thin spindle-shaped cells (RS-1A); wider, spindle-shaped cells (RS-1B); and still wider, spindle-shaped cells (RS-1C). At much later times in culture and with later-passage cells, very large and flat cells (mMSCs) appeared [12, 13]. The three distinct kinds of spindle-shaped cells appeared in the cultures in a time-dependent manner, and the transitions occurred more rapidly with higher initial plating densities. After brief training periods, three independent observers consistently scored the same cultures as summarized in Figure 5B.
In order to quantitate the transitions among the three cell types, photographs were taken after 1 to 12 days of culture for cultures initially plated at 50 cells/cm2, and an image analyzing program was used to measure cell areas and the maximal widths of cells in the cultures. No significant differences were found between the average values for area and the maximal widths of cells in the cultures, apparently because all of the cultures contained a large proportion of small, spindle-shaped cells (not shown). However, significant differences were found when the 10 largest cells (Fig. 6A) in typical fields from cultures were compared. The cell area of the 10 largest cells was greater at day 6 than at day 4 (Fig. 6B), and the maximal widths of the cells were progressively greater between day 4 and day 6 and between day 6 and day 10.
As a second assay for morphological changes, the cells were assayed by light scattering in a cell sorter. After incubation for 5 days, analysis by forward light scattering indicated that cells from cultures plated at 50 cells/cm2 were smaller than those from cultures plated at 1,000 cells/cm2 (Fig. 7A and 7B), an observation consistent with the visual evaluation of the cultures (Fig. 5B). There was no difference in cell size observed at day 9 (Fig. 7A and 7B), an observation again consistent with the visual evaluation of the cultures (Fig. 5B). The differences in side scatter among the same samples were less apparent (not shown).
Adipogenic Potential as a Function of Conditions for Expansion of MSCs
To define the adipogenic potential of the expanded MSCs, cells were plated at 50 or 1,000 cells/cm2 in complete culture medium and expanded for 4, 7, or 12 days before replating at 5,000 cells/cm2 in adipogenic medium and culturing for 21 days (Fig. 8A). As assayed by the amount of Oil Red-O absorption, MSCs plated at 50 cells/cm2 and expanded for 4 days generated more adipocytes than cells expanded for 7 days and 12 days (Fig. 8B). The number of adipocytes per culture was lower when the cells were expanded in complete medium for 12 days instead of 4 or 7 days. Furthermore, cells initially plated at a density of 1,000 cells/cm2 produced less adipocytes than those plated at 50 cells/cm2 regardless of how long they were expanded. Therefore, the results indicate that the greatest number of adipocytes was generated from cultures that were plated at lower densities for shorter time periods and that were enriched for RS-1A cells.
Correlation Between Colonies of Adipocytes and CFUs
Standard assays for adipogenic differentiation of MSCs are complicated by the fact that the cells are replated at near confluency before exposure to adipogenic medium . Here, we developed an assay for adipogenesis in single-cell-derived colonies of MSCs. Samples were initially plated at either 50 or 1,000 cells/cm2 for 12 days, replated at 1.5 cells/cm2 for 12 days to generate single-cell-derived colonies, and then transferred to adipogenic medium and cultured for 21 days (Fig. 9A). All samples generated colonies of adipocytes. The adipocytic colonies from both samples were of about the same size, but the cells initially plated at 50 cells/cm2 generated a larger number of adipocytic colonies (Fig. 9B). As expected (Fig. 3), staining of the same plates with Crystal Violet showed that the cells initially plated at 50 cells/cm2 generated a larger total number of colonies (Fig. 9B and 9C). Of special interest was that the fraction of colonies that became adipocytes was the same with both samples (Fig. 9C). Therefore, the results demonstrate that, with both samples, about 60% of the cells were capable of generating single-cell-derived colonies with adipogenic potential.
Correlation Between Conditions for Expansion and Chondrogenic Potential of MSCs
To assay the chondrogenic potential of the cells, MSCs were plated at 50 cells/cm2, expanded for 4, 7, or 12 days, and pelleted into micromasses of about 200,000 cells each before exposure to chondrogenic medium for 21 days (Fig. 10A). The cultures that were expanded for 7 days and that contained RS-1B cells formed larger cartilage pellets than both the cultures that were expanded for 4 days that were enriched for RS-1A cells and those that were expanded for 12 days that contained RS-1C cells (Fig. 10B). Also, the cultures that were expanded for 12 days and that contained RS-1C cells formed larger cartilage pellets than cells expanded for 4 days that contained RS-1A cells. The presence of cartilage extracellular matrix was measured by the weight of each pellet. Pellets derived from 7-day precultures were the heaviest among the three groups, and the weight of pellets derived from 4-day precultures was the lightest (Fig. 10C). The results suggest that the cell cultures with the greatest chondrogenic potential were cultures containing RS-1B and, therefore, those cultures that contained later progenitors than the cultures with the greatest potential to generate adipocytes.
Human MSCs, or closely related cells, are currently being tested in a number of animal models for human diseases [18, 19], and several clinical trials with the cells have been initiated [20-22]. For most of these experiments and trials, the MSCs were prepared with a standard protocol in which nucleated cells were isolated from a bone marrow aspirate with a density gradient, and then both enriched and expanded in the presence of fetal calf serum by their tight adherence to plastic tissue culture dishes. Cultures of human MSCs, unlike murine cells , become free of hematopoietic precursors after one or two passages and can be extensively expanded before they senesce. However, cultures of human MSCs are morphologically heterogeneous, even when cloned from single-cell-derived colonies [12, 24]. Moreover, the cultures undergo subtle changes as they are expanded, with a marked decrease in the rate of proliferation and multipotentiality [10, 24].
The difficulties in expanding human MSCs in culture are compounded by the fact that there is no consensus as to the characteristic surface epitopes that can be used to identify the cells. A series of antibodies to surface epitopes have been employed by several investigators [25-28], but none have come into general use. We ourselves have now screened over 200 antibodies but have not found any that efficiently distinguish RS cells from mMSCs (A. Perry et al., in preparation). Therefore, it is difficult to compare the results that different groups of investigators obtain either in animal models for disease or in clinical trials. The results here define several parameters that must be considered in preparing frozen stocks of human MSCs either for laboratory experiments or clinical trials: A) variations in the quality and number of MSCs obtained from different bone marrow aspirates, even when obtained from the same donor at the same time ; B) the yield of cells required as frozen stocks for subsequent experimentation and trials; C) the quality of the cultures in terms of their content of early progenitor cells that replicate most rapidly and have the greatest potential for multilineage differentiation, and, probably D) the number of cell doublings the cells have undergone before they are harvested and frozen.
The results presented here provide general guidelines as to how plating densities and incubation times can be varied to reach a compromise between the total yields of MSCs and the quality of the cells in terms of their content of early progenitors. Also, the simple procedure of scoring the cultures by phase-contrast microscopy provides a rapid method of assessing the cultures. In our experience, the visual scoring of the cultures has proven extremely useful in predicting the values for quality of cultures obtained with more time-consuming assays, such as CFUs or differentiation in vitro into osteoblasts, adipocytes, and chondrocytes. This study also provides the surprising result that, although cultures enriched for the earliest progenitors (RS-1A) have the greatest potential for differentiation into adipocytes, cultures with somewhat later progenitors (RS-1B) have the greatest potential to differentiate into chondrocytes. One possible explanation for this observation is that the later progenitors more readily undergo the condensation step that occurs in the initial phase of chondrogenesis .
Supported by National Institutes of Health grants AR47796 and AR44210, the Oberkotter Foundation, HCA, The Healthcare Company, and the Louisiana Gene Therapy Research Consortium.