The authors state that they have no conflicts of interest.
Extracellular Matrix Made by Bone Marrow Cells Facilitates Expansion of Marrow-Derived Mesenchymal Progenitor Cells and Prevents Their Differentiation Into Osteoblasts†
Version of Record online: 6 AUG 2007
Copyright © 2007 ASBMR
Journal of Bone and Mineral Research
Volume 22, Issue 12, pages 1943–1956, December 2007
How to Cite
Chen, X.-D., Dusevich, V., Feng, J. Q., Manolagas, S. C. and Jilka, R. L. (2007), Extracellular Matrix Made by Bone Marrow Cells Facilitates Expansion of Marrow-Derived Mesenchymal Progenitor Cells and Prevents Their Differentiation Into Osteoblasts. J Bone Miner Res, 22: 1943–1956. doi: 10.1359/jbmr.070725
- Issue online: 4 DEC 2009
- Version of Record online: 6 AUG 2007
- Manuscript Accepted: 31 JUL 2007
- Manuscript Revised: 9 JUL 2007
- Manuscript Received: 27 AUG 2006
- mesenchymal stem cells;
- extracellular matrix;
- cell replication;
- osteoblast differentiation;
- bone morphogenetic protein
We cultured MSCs on an ECM made by bone marrow cells to attempt to reconstitute the MSC niche. This ECM promoted replication of mesenchymal progenitors and retention of their multipotentiality. We conclude that the marrow ECM facilitates expansion of mesenchymal progenitors and hypothesize that it plays an important role in the maintenance of MSC stemness.
Introduction: Mesenchymal colony-forming cells of the bone marrow comprise mesenchymal stem cells (MSCs) and their transit amplifying progeny, which we term mesenchymal colony-forming units (MCFUs). These progenitors undergo self-renewal and can differentiate into many different cell types including osteoblasts. However, they lose their unique properties when cultured on tissue culture plastic. This indicates that a critical feature of the marrow microenvironment that facilitates retention of stem cell properties is missing in such culture systems. In other tissues, the extracellular matrix (ECM) forms part of the specialized niche that controls stem cell behavior. Therefore, we examined whether a marrow cell–derived ECM promotes retention of the stem cell characteristics of MCFUs in vitro.
Materials and Methods: A cell-free ECM was prepared from cultured murine marrow adherent cells. The replication and multipotentiality of murine MCFUs maintained on this marrow cell–derived ECM were examined in vitro and in vivo and compared with the behavior of MCFUs maintained on plastic.
Results: The marrow cell–derived ECM was made up of collagen types I, III, and V, syndecan-1, perlecan, fibronectin, laminin, biglycan, and decorin, similar to the composition of the marrow ECM. This ECM preparation promoted MCFU replication, restrained their “spontaneous” differentiation toward the osteoblast lineage, and preserved their ability to differentiate into osteoblasts or adipocytes. Moreover, transplantation of MCFUs expanded on the marrow cell–derived ECM into immunocompromised mice generated five times more bone and eight times more hematopoietic marrow compared with MCFUs expanded on plastic.
Conclusions: The marrow ECM facilitates expansion of MCFUs in vitro while preserving their stem cell properties. We hypothesize that the ECM made by bone marrow cells plays an important role in the maintenance of MSC function.
Mesenchymal stem cells (MSCs) of the bone marrow give rise to many cell types including osteoblasts, stromal cells that support hematopoiesis and osteoclastogenesis, and adipocytes of the bone marrow.(1) Hence, they are critical for the production of the osteoblasts and osteoclasts needed for bone development and bone remodeling throughout life. As stem cells, MSCs are characterized by their ability to both self-renew and to differentiate into specific cell types in response to appropriate lineage-specific growth factors, for example, to osteoblasts on stimulation with BMP-2. Moreover, MSCs are ideally suited for cell-based tissue engineering, for example, the repair of skeletal tissue in nonunion fractures and reconstructive surgery.(2) However, MSCs are extremely rare in the bone marrow (∼0.01–0.001%),(3,4) and earlier attempts to expand them ex vivo have been largely thwarted because of the loss of their stem cell properties during culture as they undergo senescence or “spontaneously” commit to a particular cell lineage.(5–8) Furthermore, with extensive passaging, the stem cell population is likely diluted by more committed transiently amplifying and differentiated cells. Recently, it has been reported that expansion of human or murine MSCs is accompanied by outgrowth of transformed cells, albeit transformation is less frequent in cultured human MSCs.(9–11) These problems have impaired efforts to expand MSCs in culture for purpose of studying molecular mechanisms that control their behavior and for therapeutic purposes.(8)
The loss of MSC properties in vitro strongly suggests that a critical feature of the marrow environment responsible for the maintenance of MSC stemness is missing in standard culture systems. It is now appreciated that stem cells in other regenerating tissues, for example, hair follicle epidermal stem cells, hematopoietic stem cells, and intestinal stem cells, require a specialized microenvironment or niche that supports their self-renewal capability and maintains their multipotentiality while facilitating differentiation in response to appropriate signals.(12) Such niches often comprise an extracellular matrix (ECM) component, and factors such as BMPs and Wnts made by MSCs themselves or accessory cells in the microenvironment. Direct interactions of stem cells with accessory cells can also be involved.(13)
The bone marrow in which MSCs reside contains hematopoietic cells, stromal cells, adipocytes, vascular elements, and sympathetic nerve cells.(14,15) All these cells are arrayed within a complex ECM composed of fine reticular fibers thought to be made by the stromal cells.(16–18) Analysis of bone marrow, as well as the ECM made by cultured marrow stromal cells, has shown the presence of collagens I, III, IV, V, and VI, fibronectin, laminin, and other adhesive proteins, as well as large molecular weight proteoglycans such as syndecan, perlecan, members of the small leucine-rich proteoglycan family including biglycan and decorin, and hyaluronan, a glycosaminoglycan.(17,19,20) That the marrow ECM plays a critical role in the control of MSC behavior has been strongly indicated by our recent studies showing that mice lacking biglycan exhibit defects in the ability of marrow-derived progenitors to differentiate into osteoblasts.(21,22)
Based on the above, we hypothesized that the ECM is an important component of the MSC niche. As a corollary, lack of an appropriate ECM is responsible for loss of stem cell properties when MSCs are maintained on a standard tissue culture plastic surface. The loss of stemness during growth of MSCs using current culture methods reflects the production of more differentiated progeny with diminished self-renewal capacity, rather than the production of identical daughter stem cells. Study of the impact of the ECM on the self-renewal of MSCs would ideally use markers that distinguish MSCs from their more differentiated progeny, but such markers do not currently exist. Thus, in our studies, we have relied on the ability of MSCs to adhere to culture substratum and form a colony of cells that exhibit a fibroblast-like morphology. These colony-forming cells are called colony forming unit-fibroblast (CFU-F),(1) but they are heterogeneous and comprise MSCs and their transit amplifying progeny.(23) Thus, we have termed this population of cells as mesenchymal colony-forming units (MCFUs). We have previously established that most if not all of CFU-Fs of the murine bone marrow replicate during culture to produce additional CFU-Fs as detected in a subsequent replating assay. Moreover, 50% of these newly formed progenitors differentiated into osteoblasts in response to ascorbic acid (CFU-OB).(23) We report that culture of murine marrow– derived MCFUs on a cell-free ECM made by murine marrow–derived cells promoted replication of MCFUs and dramatically restrained “spontaneous” differentiation. After expansion on this ECM, functional MCFUs were increased as shown by increased formation of bone and hematopoietic marrow tissue after subcutaneous transplantation of in vitro expanded MCFUs into immuno-compromised mice.
MATERIALS AND METHODS
Swiss Webster female mice, 6–8 wk old, were obtained from Harlan (Indianapolis, IN, USA). Animal procedures were approved by the UAMS Institutional Animal Care and Use Committee.
Preparation of cell-free ECM from cultured bone marrow cells
Femoral marrow cells were obtained as previously described(23) and cultured in 6-well plates (Corning, Corning, NY, USA) at 3 × 106 cells/10-cm2 well in 4 ml of a standard culture medium made up of α-MEM (Life Technologies, Grand Island, NY, USA) supplemented with glutamine (2 mM), penicillin (100 U/ml), streptomycin (100 μg/ml; Sigma Chemical Co., St Louis, MO, USA), and 15% preselected FBS (Atlanta Biologicals, Lawrenceville, GA, USA). After 7 days of culture, nonadherent cells were removed by rinsing. The adherent stromal cell layer was dispersed with PBS containing 400 U/ml type II collagenase (Worthington Biochemical, Lakewood, NJ, USA) for 10 min at 37°C. Then 1 × 105 adherent cells were seeded onto a 10-cm2 well of a 6-well plate containing a 24 × 30-mm Thermanox plastic coverslip (Nalge Nunc International, Rochester, NY, USA) and cultured for an additional 15 days. The medium was changed every 3–4 days; ascorbic acid (50 μM; Sigma Chemical Co.) was added during the final 8 days of culture. After extensive washing with PBS, cells were removed from the ECM by incubation with 0.5% Triton X-100 containing 20 mM NH4OH in PBS for 5 min at 37°C, similar to a previously described procedure.(24) The ECM was treated with DNase (100 units/ml; Sigma Chemical Co.) for 1 h at 37°C. The ECM was washed with PBS three times and stored in 2.0 ml of PBS containing penicillin (100 U/ml), streptomycin (100 μg/ml), and fungizone (0.25 μg/ml) at 4°C for up to 4 mo.
Preparation of tissue culture plates coated with fibronectin or type I collagen
One milliliter of 25 μg/ml fibronectin in PBS was added to each well of a 6-well plate and incubated for 1 h at 37°C. After rinsing with PBS, plates were used immediately for cell culture. Type I collagen (Sigma Chemical Co.) was dissolved at 0.1% in 1% acetic acid and diluted 10-fold with PBS. One milliliter of this solution was added to each well of a 6-well plate and incubated for 3 h at 37°C. Plates were rinsed with PBS and dried in the culture hood under UV light.
Scanning electron microscopy
Samples were washed three times with PBS, fixed with 2% glutaraldehyde in 0.1 M sodium cacodylate buffer (pH 7.2) for 1 h, and transferred to 0.1 M cacodylate buffer solution. The specimens were dehydrated in ascending concentrations of ethanol (from 70% to 100%) and embedded in epon resin (Polysciences, Warrington, PA, USA). After dehydration, the coverslips were attached to a stub and sputtered with gold-palladium. The specimens were examined using an FEI/Philips XL30 Field emission environmental scanning electron microscope (Hillsboro, OR, USA).
The preparations were fixed for 30 min with 4% formaldehyde in PBS at room temperature, washed with PBS, and blocked with 5% normal goat serum containing 0.1% BSA in PBS for 1 h. The matrices were incubated with the relevant primary antibodies (1:10 dilution) in 2% goat serum for 2 h. Antibodies against biglycan, collagen type I, III, and V, fibronectin, decorin, perlecan, syndecan-1, and laminin, were purchased from Santa Cruz Biotechnology (Santa Cruz, CA, USA). Nonspecific isotype IgG (1:10 dilution) was used as a negative control. After washing with PBS, samples were incubated with the appropriate horseradish peroxidase–conjugated secondary antibody (1:100 dilution) for 1 h, developed with a 3,3′-diaminobenzidine substrate-chromogen system (Dako, Carpinteria, CA, USA) for 5 min, and counterstained with methyl green.
Determination of CFU-F, CFU-OB, and CFU-AD number
The assay used was a modification of a technique previously described.(23) Freshly isolated murine femoral marrow cells, or marrow cells expanded on the various matrices, were placed into 6-well plates at various seeding densities, incubated for 4 h at 37°C to allow attachment of adherent cells, and washed twice with PBS to remove the nonadherent cells. Then, 3 × 106 irradiated guinea pig feeder cells(23) were added immediately in 4 ml of standard culture medium containing 1 mM l-ascorbate-2-phosphate (Wako Chemicals, Richmond, VA, USA). One half of the medium was replaced every 5 days. After 10–12 days, CFU-F colonies were visualized with crystal violet. For determination of CFU-OB, BMP-2 (30 ng/ml) was added to the cell cultures at day 7. After 25 days of culture, CFU-OB colonies were visualized with von Kossa staining. For determination of CFU-AD, the cells were cultured as above but without l-ascorbate-2-phosphate for 7 days. Rosiglitazone (100 nM) or vehicle (dimethylsulfoxide) was added to the cell cultures. After 10 days, the cultures were stained with Oil red O to visualize adipocytes. Colonies containing >50 cells were counted using a dissecting microscope.
Determination of MCFU replication capacity
The replication of MCFUs (Tables 1 and 2) was determined by comparing the number present in the initial femoral marrow cell isolate to the number present after 6 days of culture on the various matrices, using a previously described replating assay.(23) Freshly isolated bone marrow cells were pooled from six mice, and an aliquot was used to determine CFU-F, CFU-OB, and CFU-AD number as described above. The total number of each type of CFU present in the initial isolate was calculated by multiplying the number of CFUs per cell seeded by the number of cells present in the isolate. Portions of the remaining freshly isolated bone marrow cells were cultured in standard culture medium in 6-well plates at 7 × 106 cells per 10-cm2 well on tissue culture plastic, the marrow cell–derived ECM, or they were incorporated into a type I collagen gel as previously described.(23) After 6 days of culture to allow replication, nonadherent cells were removed; the adherent cells were detached with collagenase. The cells were counted and replated for quantification of CFU-F, CFU-OB, and CFU-AD number using the methods described for the determination of CFUs in the initial marrow isolate. The same number of cells were seeded for determination of CFU number regardless of their substratum used for expansion. The total number of CFUs after expansion (had the entire femoral marrow isolate been cultured on plastic or a particular ECM) was calculated by multiplying the number of CFUs obtained per cell seeded by the number of cells obtained after expansion and dividing the result by the fraction of the initial marrow isolate used for expansion. The fold change in CFU during the expansion was determined by dividing the calculated total number of CFU-F, CFU-OB, and CFU-AD after expansion by the total number of CFU-F, CFU-OB, and CFU-AD present in the initial femoral marrow cell isolate.
Quantification of gene expression during culture of bone marrow cells on plastic or the stromal cell–derived ECM
Freshly isolated murine femoral marrow cells, pooled from six mice, were seeded at 3 × 106 cells/10-cm2 well of a 6-well plate without or with the marrow cell–derived ECM and maintained in standard culture medium for up to 25 days. One half of the medium was replaced every 5 days. To isolate RNA, cells were rinsed three times with ice-cold PBS and extracted using Ultraspec reagent (Biotecx Laboratories, Houston, TX, USA). RNA (2 μg) was reverse-transcribed using a High Capacity cDNA Archive Kit (Applied Biosystems, Foster City, CA, USA). The transcripts of interest and that of the housekeeping gene GAPDH were amplified from cDNA by real-time PCR using TaqMan Universal PCR Master Mix and Assay Demand or Assay by Design primer and probe sets (Applied Biosystems). Amplification and detection were carried out with an ABI Prism 7300 Sequence Detection System (Applied Biosystems) as follows: 5-min denaturation at 95°C for 10 min, 40 cycles of amplification including denaturation at 94°C for 15 s, and annealing/extension at 60°C for 1 min. Gene expression was quantified by subtracting the GAPDH threshold cycle (Ct) value from the Ct value of the gene of interest and expressed as 2−ΔCt, as described by the protocol of the manufacturer.
Measurement of alkaline phosphatase activity and osteocalcin secretion in response to BMP-2
Freshly isolated murine bone marrow cells, pooled from six mice, were seeded on tissue culture plastic or stromal cell–derived ECM at 3 × 106 cells per 10-cm2 well in standard culture medium and cultured for 15 days. For measurement of alkaline phosphatase (ALP) response, FBS was reduced to 2%, and 3–300 ng/ml human recombinant BMP-2 (R&D Systems, Minneapolis, MN, USA) was added. After 48 h, cells were lysed with 20 mM Tris, 0.5 mM MgCl2, 0.1 mM ZnCl2, and 0.1% Triton X. ALP activity was determined using a kit from Sigma Chemical Co. The ALP value was normalized for cell number by the amount of protein in the lysates and was expressed as ALP activity per minute per microgram. For measurement of the osteocalcin response, medium was removed 6 days after addition of BMP-2, and the osteocalcin levels were measured by radioimmunoassay (RIA; Biomedical Technologies, Stoughton, MA, USA).
Measurement of BMP-2
After extensive rinsing, BMP-2 was extracted from the ECM/cell layer using 2 M urea, 2% SDS, 10% glycerol, and 10 mM Tris-HCl, pH 6.8.(25) The amount of BMP-2 in the culture supernatant and the extracts were measured using a murine-specific ELISA Assay Kit (R&D Systems).
Single-cell suspensions were obtained from the expanded cells by collagenase treatment (400 U/ml for 10–15 min at 37°C) followed by two washes in cold PBS containing 5% FCS. For antibody staining, cells (1–2 × 106) were incubated in 100 μl of diluted (10 μg/ml) anti-CD45 antibody (BD Biosciences, San Jose, CA, USA) for 30 min at 4°C. The stained cells were washed twice in staining buffer (PBS containing 5% FCS and 0.01% sodium azide) and incubated in 20 μg/ml of FITC-conjugated goat anti-mouse IgG for 20 min at 4°C, washed twice with staining buffer, and either immediately analyzed by flow cytometry or fixed with 1% paraformaldehyde in PBS and analyzed within 96 h. Cells were stained with isotype IgG as a negative control. The cell suspensions were analyzed using a Becton Dickinson FACStarplus flow cytometer. For each sample, 10,000 events were collected. The percentage of positive-stained cells was derived directly from the fluorescence-activated cell sorting (FACS).
In vivo bone formation
Freshly isolated murine marrow cells, pooled from 15 mice, were seeded at 7 × 106 cells per 10-cm2 well on tissue culture plastic or the marrow cell–derived ECM and cultured for 7 days. After rinsing with PBS, cells were detached with collagenase. The cells (1 × 106) were loaded into hydroxyapatite/tricalcium phosphate (HA/TCP) ceramic powder (Zimmer, Warsaw, IN, USA) and implanted subcutaneously into the dorsal surface of 10-wk-old immunodeficient beige mice (NIH-bg-nu-xid; Harlan Sprague Dawley, Indianapolis, IN, USA), as previously described.(26,27) Cells precultured on tissue culture plastic were implanted on the left side, and cells precultured on the marrow-derived ECM were implanted on the right side of each animal. As a negative control, an HA/TCP vehicle without cells was implanted into a mouse. Transplants were harvested after 4 or 8 wk (three animals for each), fixed in 10% phosphate-buffered formalin at 4°C for 24 h, decalcified with 5% EDTA (pH 8.0) at room temperature for 1–2 wk, and embedded in paraffin. Each ossicle was bisected, and three sections (10 μm thick) were cut starting at the bisection point of each half-ossicle at 100-μm intervals to yield a total of 18 sections for each ossicle. Sections were stained with H&E, and the percentage of the total ossicle area containing new bone or hematopoietic marrow was measured using Osteometrics image analysis software (Ostomeasure version 3.00; Osteometrics, Atlanta, GA, USA).
Data are presented as mean ± SD. Statistically significant effects were detected with Student's t-test or one-way ANOVA, using SigmaStat (Systat Software, Point Richmond, CA, USA). Statistically significant differences among multiple treatment groups were detected after correction by Bonferoni's method. p < 0.05 was considered significant.
Characteristics of the marrow cell–derived ECM
Consistent with earlier studies,(28) adherent cells cultured from murine femoral bone marrow elaborated a fibrillar ECM as revealed by scanning electron microscopy (SEM; Fig. 1A). Some debris remained after removal of cells, as reported previously in the case of epithelial cell-derived ECM preparations.(29) The ECM contained fibers of ∼25–50 nm diameter, similar to that of collagenous fibrils seen in the bone marrow.(16) The ECM was ∼100 μm thick as determined by transmission electron microscopy (data not shown).
When examined before removal of cells, immunostaining revealed the presence of collagen types I, III, and V, syndecan-1, perlecan, fibronectin, laminin, biglycan, and decorin associated with the ECM, and sometimes within adherent cells (Fig. 1B). The composition of the ECM is consistent with that previously reported for the ECM made by cultured marrow stromal cells and are similar to the composition of the bone marrow ECM.(17–19) For the most part, the cell extraction procedure did not seem to affect the composition of the ECM as determined by the semiquantitative immunostaining procedure; however, collagen type V and syndecan-1 immunostaining was significantly reduced after extraction (Fig. 1B).
Culture on marrow cell–derived ECM restrains differentiation and promotes replication of MCFUs
We first determined the effect of the ECM on the adherence of MCFUs capable of forming a colony of fibroblastic cells (i.e., CFU-F). In this experiment, freshly isolated murine marrow cells were allowed to adhere for 4 h to plastic, the cell-derived ECM, plastic coated with fibronectin, or plastic coated with type I collagen. Nonadherent cells were removed, and feeder cells (irradiated guinea pig marrow cells) were added to facilitate colony formation at the low seeding density used for this experiment.(30) The colonies that subsequently formed on the marrow cell–derived ECM appeared larger, and the cells tended to be more densely packed, compared with the colonies that developed on tissue culture plastic or tissue culture plastic coated with fibronectin or type I collagen (Fig. 2A). In addition, there was a 2- to 3-fold increase in the number of CFU-F when seeded on the marrow cell–derived ECM compared with the other matrices tested (Fig. 2B).
We also examined the ability of MSCs to differentiate into osteoblasts in response to addition of BMP-2 or into adipocytes in response to rosiglitazone. MCFUs that gave rise to colonies containing osteoblasts or adipocytes were designated as CFU-osteoblast (CFU-OB) and CFU-adipocyte (CFU-AD), respectively. When cultured in the presence of BMP-2 to stimulate osteoblastogenesis, the number of CFU-OB, as well as the degree of von Kossa staining for mineral within each colony, was increased when the cells were cultured on the ECM (Fig. 2A). We also found that the number of CFU-AD was increased when cells were cultured on the stromal cell–derived ECM, and they contained more Oil red O–stained adipocytes compared with cells cultured on tissue culture plastic or plastic coated with fibronectin or type I collagen. These observations indicate that more MCFUs adhere to the marrow cell–derived ECM and that they gave a stronger response to prodifferentiating factors.
Examination of the morphology of cells within CFU-F colonies by SEM revealed that, after 5 days of culture on tissue culture plastic, cells were round and flat, and there was no evidence of an endogenous ECM. However, cells cultured on the marrow cell–derived ECM were embedded within the matrix, and they exhibited a fibroblastic morphology with extensive cellular processes (Fig. 2C). After 10 days of the culture, some of the cells maintained on plastic had begun to elaborate and become embedded in an ECM; however, they did not exhibit the same morphology as cells cultured on the marrow cell–derived ECM. In both conditions, round cells with a morphology characteristic of hematopoietic cells(31) were present at 10 days of culture.
We next studied whether the marrow cell–derived ECM prevented “spontaneous” differentiation of MCFUs. In this experiment, murine bone marrow cell cultures were established at the seeding density commonly used for expansion of MCFUs (i.e., higher than that used for colony assays). Nonadherent cells were not removed, and exogenous guinea pig feeder cells were not added because, at this seeding density, it is thought that endogenous cells serve this function.(22) Cell number, as reflected by RNA content, progressively increased during the first 15 days of culture (Fig. 3A). However, when examined at day 20, cells were confluent (Fig. 3B). Cells were grouped into nodules when maintained on plastic, whereas cells cultured on the marrow cell–derived ECM were evenly distributed. The expression of the osteoblast markers ALP, osteocalcin, bone sialoprotein, and type I collagen progressively increased during 25 days of culture on plastic (Fig. 3C), consistent with the “spontaneous” differentiation of MCFUs reported previously.(8) In contrast, the marrow cell–derived ECM preparation significantly delayed the appearance of these osteoblast markers. In a separate experiment, there was practically no mineral deposition, as determined by von Kossa staining, when cells were maintained on the marrow cell–derived ECM for 25 days (data not shown).
The restraint of osteoblastogenesis seen in cultures of MCFUs maintained on marrow cell–derived ECM was not caused by increased production of antagonists of the bone morphogenetic proteins (BMPs) or Wnt proteins needed for osteoblast differentiation. Indeed, transcript levels of such antagonists, including Sost, Noggin, Dkk1, Twisted gastrulation, Gremlin, and Chordin, in cultures maintained on plastic was significantly higher than in cultures maintained on this ECM (Fig. 3D). We also noted a transient increase in Gremlin on day 7 in cells cultured on either plastic or the ECM.
We have previously shown that autocrine/paracrine production of BMP-2 and BMP-4 mediate the osteoblastogenesis that occurs when MCFUs are cultured on plastic.(32) Hence, the restraint of osteoblast differentiation observed in cultures maintained on the marrow cell–derived ECM could be caused by decreased synthesis of endogenous BMP-2. However, BMP-2 mRNA levels were similar in both culture conditions (Fig. 4A). In a separate experiment, we determined that cell-free ECM preparations contained no detectable BMP-2 (data not shown). However, the amount of BMP-2 protein was increased by ∼30% in cultures maintained for 15 days on the marrow cell–derived ECM compared with plastic (Fig. 4B). Strikingly, and in agreement with evidence that BMPs bind to components of the ECM,(33) we found that >90% of BMP-2 protein was associated with the cell/matrix layer in cultures maintained on the marrow cell–derived ECM compared with 60% in the case of cultures maintained on the plastic. Moreover, BMP-2 levels in the culture supernatant were 4-fold lower in the ECM cultures compared with cells cultured on plastic. Thus, it is possible that the restraint of osteoblast differentiation when MSCs were cultured on the marrow cell–derived ECM is related to sequestration of BMP-2 by the ECM.
We next examined whether MCFUs retain their osteoblastogenic response to exogenous BMP-2 when grown on the ECM. In this experiment, BMP-2 was added at 15 days after establishment of the cultures. When cultured on the marrow cell–derived ECM, as little as 3–10 ng/ml recombinant human BMP-2 stimulated ALP activity and osteocalcin secretion (Fig. 4C), as well as the level of ALP and osteocalcin mRNA (Fig. 4D). Basal ALP activity was already elevated in cultures maintained on tissue culture plastic compared with the ECM, consistent with the data of Fig. 3C showing an increase in ALP transcripts at the early stage of culture. These findings indicate that MCFUs retained their ability to differentiate into osteoblasts in response to exogenous BMP-2 when cultured on the marrow cell–derived ECM. Addition of exogenous BMP-2 to cells maintained on plastic modestly increased ALP activity and osteocalcin secretion but only at 30–100 ng/ml of added BMP-2 (Fig. 4C). BMP-2 had no effect on ALP mRNA in these cultures, but osteocalcin mRNA was increased at 100 ng/ml BMP-2. Higher levels of exogenous BMP-2 are evidently needed to further enhance osteoblastogenesis beyond that already stimulated by endogenous BMPs when the cells were cultured on plastic.
Marrow cell–derived ECM promotes MCFU replication while retaining multipotentiality
The replication of MCFUs during culture on the various matrices was determined by measuring the increase in CFU number using a replating assay that we have previously described.(23) Freshly isolated bone marrow cells were divided into aliquots for the determination of CFUs present in the initial isolate and after expansion on plastic or the marrow cell–derived ECM, as well as in type I collagen gels, which we have used previously for determination of CFU replication.(23) During the 6-day expansion period, nonadherent cells were not removed, and exogenous guinea pig feeder cells were not added. As shown in Table 1, the number of cells obtained after 6 days of culture on the ECM was increased compared with cells cultured on plastic or in type I collagen gels.(23) The frequency of the MCFUs in the replating assay was ∼50% greater than in cells expanded on plastic or type I collagen gel (Table 1). The total number of CFU-F present in the cultures expanded on the marrow cell–derived ECM was increased 47-fold over the number of CFU-F present in the initial bone marrow isolate (Table 1; Fig. 5). In contrast, CFU-F increased 10- and 27-fold in cultures maintained on plastic and type I collagen gel, respectively. The number of MCFUs capable of differentiating into osteoblasts or adipocytes, after expansion, was also measured by inducing differentiation with ascorbate-2-phosphate and BMP-2, or rosiglitazone, respectively. We found that the increase in the number of CFU-OB and CFU-AD was 2- to 4-fold greater when cultured on the marrow cell–derived ECM compared with cultured on plastic or type I collagen gel. CFU-F replication was greater than that of CFU-OB and CFU-AD, regardless of the matrix used for expansion. Hence, even though the expansion of CFU-F, CFU-OB, and CFU-AD was greater when the cells were cultured on the marrow cell–derived ECM, the culture substratum did not alter the proportion of MCFUs that could differentiate into osteoblasts or adipocytes.
In the above experiment, MCFU number was determined using a standard procedure in which cells were assayed on plastic before or after expansion. However, it is possible that MCFUs expanded on plastic versus the ECM have different adhesion characteristics. Such a difference could influence the estimation of MCFU replication because the number of MCFUs (determined in the standard assay) could be different from the number of MCFUs adhering to the ECM during expansion. To study this issue, we performed an experiment in which the same culture substratum was used for both enumeration and expansion of CFU-Fs. As shown in Table 2, more CFU-Fs in the initial marrow isolate adhered to the ECM compared with plastic, consistent with the data of Fig. 2. Nevertheless, the increase in CFU-Fs during culture on the marrow cell–derived ECM was 2-fold greater than when expanded on plastic. Therefore, although there are differences in CFU-F adherence to plastic versus the ECM, such differences do not unfairly bias determination of CFU-F replication. We also found in this experiment that the majority of the expanded cells comprised CD45+ hematopoietic cells (determined by flow cytometry), regardless of whether cells were cultured on plastic or the marrow cell–derived ECM, and that the number of CD45+ hematopoietic cells present in cultures maintained on the ECM is higher than that maintained on plastic. Thus, the ECM promoted increased replication of both MCFUs and hematopoietic cells.
To show the capacity of MCFUs expanded on the marrow cell–derived ECM to generate skeletal tissue, we used a transplantation assay.(27) After 7 days of culture of bone marrow cells on plastic or on marrow cell–derived ECM, 1 × 106 adherent cells were loaded onto an HA/TCP carrier and implanted subcutaneously into immuno-compromised NIH-bg-nu-xid mice. We found that, whereas little bone was formed at 4 wk after implantation by cells expanded on plastic, there was substantial bone formed by cells expanded on the ECM at this time-point (Fig. 6E). The amount of bone generated at 8 wk after implantation of cells precultured on plastic was ∼3% of the total area of the ossicle. This finding is consistent with previous reports that implantation of 3–5 × 106 murine marrow cells expanded on plastic for at least one passage generated bone ossicles containing ∼5–7% bone tissue.(27,34) There was no bone in implants that were not loaded with cells (data not shown). We also found that there was minimal hematopoietic marrow in ossicles made by cells expanded on plastic, and adipocytes and osteoclasts were rarely observed (Figs. 6A, 6E, and 6F). In contrast, transplantation of 1 × 106 cells expanded on marrow cell–derived ECM generated five times more bone than the cells precultured on tissue culture plastic (Figs. 6B and 6E). The hematopoietic marrow of the ossicles made by MCFUs grown on the ECM was characterized by a large number of adipocytes and was observed at 8, but not 4, wk after implantation (Fig. 6C). The area of hematopoietic marrow was 8-fold higher in ossicles made by cells cultured on the ECM compared with cells cultured on plastic (Fig. 6F). Osteoclasts were also present in ossicles made by cells precultured on the ECM (Fig. 6D).
We showed that culture of murine MCFUs on an ECM made by marrow-derived cells promotes replication of MCFUs, whereas culture on plastic favors production of differentiated progeny. Moreover, the MCFUs expanded on the ECM contained progenitors that could differentiate in vitro into osteoblasts or adipocytes. In contrast, MCFUs grown on tissue culture plastic exhibit diminished replicative capability and increased expression of the osteoblast phenotype. Although the ECM restrained differentiation and promoted replication of MCFUs, it could not prevent loss of multipotentiality of the MCFU population as a whole, as shown by the decrease in the proportion of progenitors that could differentiate into osteoblasts or adipocytes compared with freshly isolated bone marrow cells. Nevertheless, MCFUs expanded on the ECM formed a complete bone-like structure in vivo that contained hematopoietic marrow with adipocytes, as well as stromal cells that support hematopoiesis and osteoclastogenesis. In comparison, cells expanded on plastic made less bone and hematopoietic marrow. The ∼50% increase in CFUs and CD45+ cells after expansion on the marrow ECM, compared with plastic (Table 1), cannot fully explain the 5- to 8-fold increase in bone and bone marrow in ossicles made by cells cultured the stromal ECM. This could be caused by selective expansion of MCFUs on the ECM that have high self-renewal capacity and therefore are able to produce more osteoblasts and stromal cells capable of supporting osteoclast formation and hematopoiesis. However, it is unknown whether such progenitors represent MSCs or early transit amplifying progeny. Alternatively, the ECM may favor production of putative accessory cells such as hematopoietic cells or CD45+ cells that have a positive effect on the maintenance of MSCs in vitro and the formation of skeletal tissue after transplantation into immunodeficient mice.
The maintenance of MCFU properties when cultured on the marrow cell–derived ECM supports the hypothesis that the marrow ECM provides important microenvironmental cues to MCFUs in vivo. Stromal cells are the source of the structural components of the ECM. However, cells of the hematopoietic lineage are also present in the cultures used to make the ECM. It is possible that they influence properties of the ECM by secreting growth factors, cytokines, and matrix metalloproteinases that affect the biosynthetic activity of the stromal cells. Such hematopoietic cell–derived factors may also become incorporated into the ECM.
Involvement of the ECM in the regulation of MCFUs is further supported by evidence previously reported by us that deletion of the ECM components biglycan and decorin has a deleterious effect on responsiveness of marrow-derived osteoblast progenitors to BMPs and TGF-β.(27,33) At this stage, it is unknown how the ECM regulates the behavior of MCFUs. Earlier work has shown that the ECM modulates the activity of growth factors by controlling proteolytic activation of latent factors, as occurs in the case of TGF-β.(35) The ECM also interacts with cell surface receptors to prevent binding of the cognate ligand, as occurs in the case of the epidermal growth factor (EGF) receptor,(36) and sequesters factors such as platelet-derived growth factor (PDGF) and BMPs.(37,38) The ECM may also bind growth-promoting factors from the serum for optimal presentation to MSCs. Finally, the ECM may enhance the function of putative accessory cells that support MCFU replication.
It is well established that MSCs lose their multipotentiality when cultured on tissue culture plastic.(5–7) For example, Banfi et al.(7) calculated that the generation of 1 mm3 of bone required transplantation of 161 CFU-Fs isolated from fresh human bone marrow to immunodeficient mice, whereas transplantation of 5781 CFU-Fs was required after marrow cells had been cultured for 3 wk. Loss of stem cell properties, coincident with so-called “spontaneous” differentiation when MCFUs are cultured on plastic, may actually represent the response of MSCs to growth factors produced endogenously in these cultures. Indeed, we previously showed that autocrine/paracrine production of BMP-2/4 is required for osteoblastogenesis when MCFUs are cultured on plastic.(32) BMPs bind strongly to collagen and small proteoglycans such as biglycan.(33,37) The finding of this study that the ECM sequesters endogenously produced BMP-2 is consistent with such binding and may explain why MCFUs retained an undifferentiated phenotype when cultured on the ECM. Other prodifferentiating proteins may also be sequestered by the ECM. Wnt proteins, a large family of ligands that regulate MSC differentiation through activation of LRP5 and LRP6, are known to bind to glycosaminoglycans of the ECM.(39) Interestingly, inhibition of Wnt signaling by autocrine production of Dkk-1, an inhibitor of LRP5 and LRP6 activation, is required for the replication of human MSCs,(40) presumably by inhibiting their differentiation.
Previous attempts to restrain “spontaneous” MSC differentiation have involved culture on fibronectin matrices under low oxygen tension (3%)(41) to mimic the microenvironment of the bone marrow(42) or culture at low seeding density in low serum in the presence of growth factors.(43,44) However, the ability of such preparations to form skeletal tissue in vivo has not been reported. In addition, fibroblast growth factor (FGF)-2 has been reported to increase the size of human MSC colonies and to restrain their differentiation, but in distinction to our finding of increased adhesion and replication of MCFUs using the marrow cell–derived ECM, FGF-2 reduced colony number.(45) Other investigators have reported that FGF-2 alters the properties of human MSCs and may even enhance osteoblastogenesis while reducing neurogenic capability.(46)
In this study, we focused on the ability of MCFUs to differentiate into osteoblasts and adipocytes. However, marrow MSCs have been reported to differentiate into cardiomyocytes, neurons, and hepatocytes.(1,47–50) It has not been firmly established whether each of these cell types are derived from a common progenitor or whether the bone marrow contains a variety of stem cells with restricted lineage specification. Nevertheless, the marrow cell–derived ECM described herein could also be used to study MSCs that differentiate into these other cell types.
The limitations of this study derive from lack of a specific surface marker to permit detection and isolation of murine MSCs. Thus, the assay we used to detect and quantify MSCs also detects the transit amplifying progeny of MSCs. Moreover, we used an in vitro MCFU replication assay as surrogate for determination of self-renewal capability. This assay is not as rigorous as the classic assay that involves the demonstration that progeny of a single stem cell forms a complete tissue in vivo, followed by isolation of a colony-forming cell from the transplant, and subsequently showing that the progeny of this cell is capable of again forming the complete tissue in vivo.(51) Stringent demonstration of MSCs in vitro not only requires specific surface markers, but also a culture system that allows the maintenance of MSC stemness. The ECM-based culture system that we describe here may set the stage for further advancement in this area.
In summary, our findings provide strong support for the hypothesis that the marrow ECM forms part of the niche that supports MCFUs in the bone marrow and that the ECM regulates the balance between replication and differentiation in response to appropriate signals. Consequently, the marrow cell–derived ECM described herein provides an advantageous environment for the expansion of functional MCFUs for therapeutic applications. This system will also be useful for studying whether aging and/or hormonal changes negatively affect the ability of the marrow ECM to maintain MSC function, and thereby contribute to the development of osteoporosis.
The authors thank A Deloose, L Mommsen, C Skinner, R Wynne, and S Tsai for technical contributions to this study. This work was supported by the Department of Veterans Affairs, a grant from the Department of Internal Medicine at University of Arkansas for Medical Sciences (to X-DC), and the National Institutes of Health [R21 AG025466 (to X-DC), and P01 AG13938 (to SCM)].
- 61999 Propagation and senescence of human marrow stromal cells in culture: A simple colony-forming assay identifies samples with the greatest potential to propagate and differentiate. Br J Haematol 107: 275–281., , , , ,
- 241999 Preparation of extracellular matrices produced by cultured corneal endothelial and PF-HR9 endodermal cells. MorganK (ed.), Current Protocols in Cell Biology, vol UNIT 10.4. John Wiley & Sons, Hoboken, NJ, USA, 10.4.1–10.4.14.
- 291984 Comparison of the ability of basement membranes produced by corneal endothelial and mouse-derived Endodermal PF-HR-9 cells to support the proliferation and differentiation of bovine kidney tubule epithelial cells in vitro. J Cell Biol 99: 947–961., , ,