Through their broad differentiation potential, mesenchymal stem cells (MSCs) are candidates for a range of therapeutic applications, but the precise signaling pathways that determine their differentiated fate are not fully understood. Evidence is emerging that developmental signaling cues may be important in regulating stem cell self-renewal and differentiation programs. Here we have identified a consistent expression profile of Wnt signaling molecules in MSCs and provide evidence that an endogenous canonical Wnt pathway functions in these cells.
Wnts bind to Frizzled (Fz) receptors and subsequent canonical signaling inhibits glycogen synthase kinase–3β (GSK-3β), causing β-catenin translocation into the nucleus to induce target gene expression. In human MSCs isolated from bone marrow of different donors, we appear to have identified a common Wnt/Fz expression profile using reverse transcriptase polymerase chain reaction (RT-PCR). Associated Wnt signaling components, including low-density lipoprotein receptor–related protein–5 (LRP-5), kremen-1, dickkopf-1 (Dkk-1), secreted Frizzled-related peptide (sFRP)–2, sFRP3, sFRP4, Disheveled (Dvl), GSK-3β, adenomatous polyposis coli (APC), β-catenin, T-cell factor (TCF)–1, and TCF–4, were also identified. Nuclear β-catenin was observed in 30%–40% of MSCs, indicative of endogenous Wnt signaling. Exposure to both Wnt3a and Li+ ions, which promotes canonical Wnt signaling by inhibiting GSK-3β, reduced phosphorylation of β-catenin in MSCs and increased β-catenin nuclear translocation approximately threefold over that of the controls.
Our findings indicate that autocrine Wnt signaling operates in primitive MSC populations and supports previous evidence that Wnt signaling regulates mesenchymal lineage specification. The identification of a putative common Wnt/Fz molecular signature in MSCs will contribute to our understanding of the molecular mechanisms that regulate self-renewal and lineage-specific differentiation.
Mesenchymal stem cells (MSCs) have great therapeutic potential, with the capacity to influence diverse medical applications, such as tissue engineering and gene therapy. MSCs self-renew and differentiate into a range of mesenchymal tissues, including bone, adipose, cartilage, muscle, marrow stroma, tendon, and ligament, both in vivo and in vitro, under appropriate culture conditions, via a series of increasingly differentiated precursor cells [1–3].
Mesenchymal stem cells were first identified in human bone marrow as cells that were able to self-proliferate and also differentiate into connective tissues, such as bone and cartilage [4–6]. Many groups have since isolated similar cells from bone marrow and expanded colonies in culture, before inducing differentiation into various mesenchymal lineages [7–9]. These cells, capable of extensive self-proliferation and also multilineage differentiation, have subsequently been assigned several terms, such as mesenchymal progenitor cells, marrow stromal cells, and mesenchymal stem cells. In this rapidly growing area of research, uniformity with regard to cell type classification is currently lacking, and no definitive marker of MSCs is currently available. However, several monoclonal antibodies have been raised against MSCs [10–13], providing a panel of markers for characterization of these cells, including SH-2 (CD105, endoglin ), SH-3 and SH-4 (both CD73, ecto-5′-nucleotidase ), SB-10 (CD166, ALCAM ), and, in addition, the cell-surface antigens CD29 (β1-integrin), CD44 (H-CAM), CD90 (Thy-1), and STRO-1 . MSCs are also characterized by their negativity of the hematopoietic markers CD34 and CD45 and have been found to be human leukocyte antigen (HLA) class I positive and HLA class II negative . Here, consistent with previously published research and general consensus in the field, we define MSCs as a discrete population of cells within the bone marrow stromal cell (BMSC) compartment that self-proliferate extensively, express the MSC antigens described above, and exhibit multilineage potential.
It is clearly desirable to determine the signaling mechanisms involved in controlling MSC activity in order to identify novel therapeutic targets. Previously published work has identified intracellular signaling mechanisms involved in osteogenic and adipogenic specification in MSC cultures , but relatively little is known about the regulatory inputs that enable the MSCs to maintain their undifferentiated phenotype or those that regulate its differentiation into specialized cell types. Recent expression profiling studies have identified key developmental signaling molecules, including Wnts, in pluripotent embryonic stem cell populations, suggesting that these fundamental cell–cell communication pathways may be instrumental in controlling “stemness” [20–23]. Here we have identified the expression profile of Wnt signaling molecules in MSCs, providing evidence for their involvement in both the maintenance and differentiation of adult mesenchymal progenitor cells.
Wnt proteins constitute one of the most important families of signaling molecules in development, and these proteins also have vital roles in adult tissues—for example, in the regulation of cell proliferation and motility, generation of cell polarity, and specification of cell fate [24–26]. Recently, Wnt signaling has been implicated in the control of differentiation of hematopoietic stem cells [27–29] and stem cells in skin , as well as in regulating myogenesis [31, 32], chondrogenesis [33, 34], and adipogenesis [35, 36]. A role for Wnt signaling in osteoblast differentiation has also been suggested, predominantly through studies of the Wnt coreceptor, low-density lipoprotein receptor–related protein–5 (LRP-5). Disruptive mutations in this receptor are associated with decreased bone mass, as a consequence of reduced osteoblast proliferation [37, 38], whereas high bone-mass phenotypes and increased osteoblast proliferation are associated with activating mutations in LRP-5 [39–41]. Recently, BMP-2 has also been shown to induce alkaline phosphatase and mineralization via Wnt signaling in mesenchymal cell lines .
Wnts are highly conserved, cysteine-rich secreted ligands, and so far 19 have been identified in humans. Wnt signaling can stimulate at least four different signaling pathways, the best characterized being the canonical pathway, which regulates β-catenin stability, leading to downstream transcription of target genes (Fig. 1; for review, see ). In the absence of a Wnt signal, β-catenin is phosphorylated by glycogen synthase kinase–3β(GSK-3β), in association with axin and adenomatous polyposis coli (APC), which targets β-catenin for ubiquitinylation and subsequent degradation by proteasomes. However, when Wnt ligands bind to Frizzled (Fz) receptors, as well as coreceptors LRP-5 and LRP-6, the cytoplasmic protein Disheveled (Dvl) is activated. Phosphorylation of β-catenin by GSK-3βis inhibited by Dvl, causing β-catenin stabilization and accumulation, before translocation to the nucleus, where it binds with members of the T-cell factor (TCF) and lymphoid enhancer factor (LEF) transcription factor family, to induce expression of target genes. Secreted Fz-related peptides (sFRPs) have recently been identified as negative regulators of Wnt signaling (reviewed in [43, 44]). These are structurally very similar to Fz and are thought to compete with Fz for Wnt ligand binding. Dickkopf-1 (Dkk-1) has also been shown to inhibit Wnt signaling  via its association with LRP-5 and -6. Kremen is the Dkk-1 receptor, which functionally cooperates with Dkk-1 to inhibit Wnt signaling . The first report of a successfully isolated Wnt protein was published only last year , as until then bioactivity was often lost in the purification process. However, lithium ion (Li+) application is often used to mimic canonical Wnt signaling, as it inhibits GSK-3β, therefore stimulating downstream components of the Wnt signaling pathway [48–50]. It is also possible to stimulate Wnt signaling in vitro using conditioned medium collected from a Wnt3a-overexpressing cell line. Here we provide evidence to suggest a consistent Wnt/Fz expression profile across several undifferentiated MSC populations, which appears to support functional, endogenous canonical Wnt signaling in these cells.
Materials and Methods
Isolation and Culture of Primary BMSCs
Tissue culture plastic ware and reagents were purchased from Invitrogen (Paisley, U.K., http://www.invitrogen.com) unless otherwise stated. Primary human BMSCs were isolated from femoral heads following informed consent (with kind authority from York Hospital, York, U.K.), removed during hip-replacement operations. The marrow surrounding the trabecular bone fragments was collected and grown in control medium (alpha minimum essential medium [α-MEM] containing 2 mM glutamax, 100 U/ml penicillin, 100 μg/ml streptomycin, and 15% fetal bovine serum [FBS]) in 75-cm2 flasks at 37°C in 5% CO2 at 95% air atmosphere. After 7 days, the nonadherent cells were removed and medium was replaced every 3 days. On reaching confluency, the cells were subcultivated at a 1:3 ratio and then cultured up to passage 15. Commercially available human mesenchymal stem cells (Cambrex Bio Science, Baltimore, http://www.cambrex.com) were cultured using the same methods. MSCs were used up to passage 6 for functional studies and passage 12 for reverse transcriptase polymerase chain reaction (RT-PCR) expression profiling.
Cell Line Culture Conditions
MC3T3-E1 cells (a primitive osteoblast-like cell line) were grown in α-MEM containing 2 mM glutamax, 100 U/ml penicillin, 100 μg/ml streptomycin, and 10% FBS in 75-cm2 flasks under the same conditions. C2C12 cells (a bipotent myoblastic cell line with both myogenic and osteogenic potential) were also cultured under the same conditions in Dulbecco's Modified Eagle's Medium (DMEM), 100 U/ml penicillin, 100 μg/ml streptomycin, and 10% FBS.
Wnt3a Conditioned Medium Collection
L-Wnt3a cells, a Wnt3a-overexpressing cell line, and the control nontransfected L cells were purchased from American Type Culture Collection (ATCC; Manassas, VA, http://www.atcc.org) and incubated in DMEM with glutamax, 100 U/ml penicillin, 100 μg/ml streptomycin, and 10% FBS under the conditions described above. The L-Wnt3a cell medium was supplemented with 0.4 mg/ml G418 (Sigma-Aldrich Company Ltd., Dorset, U.K., http://www.sigma-aldrich.com). The cells were grown in 75-cm2 flasks and passaged 1:6 when confluent. Conditioned media from both L-Wnt3a cells and the control L cell line were collected according to the manufacturer's instructions. Briefly the cells were passaged 1:10 in 8-ml medium without G148 and left to grow for 10 days. After this time, the first batch of medium was collected from each cell line and replaced with 8 ml fresh medium for a further 3 days. This was then collected as the second batch of conditioned medium and added to the first before being sterile filtered.
Characterization of Human MSCs from Primary Human BMSCs
Flow Cytometric Analysis of Surface Antigen Expression
When confluent, the BMSCs were passaged 1 in 3, and a sample was analyzed for mesenchymal stem cell marker expression by flow cytometry. The cells were washed in phosphate-buffered saline (PBS), and then removed from the flask by incubating in wash buffer (PBS containing 0.2% bovine serum albumin [BSA] (Sigma-Aldrich) and 5 mM EDTA at pH 7.4). 1 × 105 cells were incubated with each mouse monoclonal primary antibody at 4°C for 30 minutes. All antibodies were purchased from Beckton Dickinson (Oxford, U.K., http://www.bd.com), except mouse anti-CD90 and HLA-class I (CHEMICON International, Temecula, CA, http://www.chemicon.com), mouse phycoerythrin (PE)–labeled anti-HLA-DR, and mouse fluorescein isothiocyanate (FITC)–conjugated anti-CD45 (Caltag Laboratories, Burlingame, CA, http://www.caltag.com). All antibodies were titrated to determine optimal concentration prior to use. FITC-labeled anti-CD45 and PE-labeled anti-HLA-DR were used at 1:200 dilution; anti-CD29, anti-CD73, anti-CD90, anti-CD105, and anti-HLA class I were used at 1:100 dilution; PE-Cy5 conjugated anti-CD34 was used at 1:50 dilution; and FITC-labeled anti-CD44 and PE-labeled anti-CD166 were used at 1:10 dilution. The same concentration of mouse immunoglobulin G (IgG) (Sigma-Aldrich) was used as the negative control for each treatment. After washing, cells incubated with unconjugated antibodies were subsequently incubated with secondary FITC-labeled goat antimouse antibody (Sigma-Aldrich; at 1:20 dilution) for 30 minutes in the dark at 4ºC. After a final wash, the cells were resuspended in 300 μl wash buffer and analyzed on a Cyan flow cytometer (Dako Cytomation, Ely, U.K., http://www.dakocytomation.com). Dead cells and doublets were excluded by light scatter properties, and the percentage of live cells expressing the different antigenic markers was determined by setting a gate at 1% on negative control samples.
Standard published protocols were used to promote differentiation of passage 3–6 MSCs along mesenchymal lineages, including osteogenic , adipogenic , and chondrogenic lineages [52, 53]. Briefly, to induce osteogenic differentiation, confluent MSCs were incubated in osteogenic medium (control medium, supplemented with 10 nM dexamethasone, 50 μg/ml L-ascorbic acid phosphate, and 5 mM β-glycerophosphate [all from Sigma-Aldrich]) for 12 days, before being stained for alkaline phosphatase activity and mineralization using the von Kossa stain. Adipogenic differentiation was stimulated by culturing confluent MSCs in adipogenic medium (control medium with 10 nM dexamethasone, 50 μg/ml L-ascorbic acid phosphate, 500 μM isobutylmethyxanthine [IBMX; Sigma-Aldrich] and 60 μM indomethacin [Sigma-Aldrich]) for 12 days, before identifying adipocytes by oil red O staining. MSCs were grown in micromass cultures to induce chondrogenic differentiation. 2 ×105 cells in chondrogenic medium (DMEM containing 110 μg/ml sodium pyruvate, 100 U/ml penicillin, 100 μg/ml streptomycin, 1% human recombinant insulin, human transferrin, and selenous acid [ITS] + Premix [BD Biosciences, Oxford, U.K., http://www.bd.com], 50 μg/ml L-ascorbic acid phosphate, 100 nM dexamethasone, 50 μg/ml proline [Sigma-Aldrich], and 10 ng/ml recombinant human transforming growth factor–β3 [TGF-β3; R&D Systems, Oxon, UK, http://www.rndsystems.com]) were placed in universals and pelleted by centrifugation at 800 rpm for 5 minutes, before incubation for 15 days.
Alkaline Phosphatase Enzyme Histochemistry and von Kossa Staining
To detect alkaline phosphatase activity, the cells were stained with 1 mg/ml Fast red TR (Sigma-Aldrich) and 0.2 mg/ml napthol AS-MX phosphate (Sigma-Aldrich), dissolved in 1 ml N, N dimethylformamide (Sigma-Aldrich) in 0.1 M Tris buffer at pH 9.2, and then fixed in 4% paraformaldehyde. Sites of mineralization were identified using the von Kossa (vK) staining technique by adding 1% silver nitrate (Fisher Scientific UK, Loughborough, U.K., http://www.fisherscientific.com) at room temperature for 60 minutes under strong light, followed by 2.5% sodium thiosulphate (Fisher Scientific) for 5 minutes.
Oil Red O Adipocytic Stain
Cells were fixed in 60% isopropanol for 1 minute, before 60% oil red O stock (0.5% w/v oil red O [Sigma-Aldrich] in isopropanol) diluted in dH2O was added to the cells for 30 minutes, to stain for lipid content. After washing, the cells were counterstained with hematoxylin.
Alcian Blue Chondrogenic Stain
After 15 days in chondrogenic medium, cell pellets were collected, mounted in optimal cutting temperature (OCT) compound on brass chucks and frozen in chilled ethanol. Sections 5-μm thick were cut on a Bright cryostat (Bright Instruments, Huntingdon, U.K., http://www.brightinstruments.com) and collected on electrostatically charged slides (BDH, Dorset, U.K., http://www.merckeurolab.com). After fixing in 70% ethanol for 10 minutes, the sections were washed in PBS before staining in 1% alcian blue 8GX (Sigma-Aldrich) in 3% acetic acid overnight. Following destaining in 70% ethanol, the sections were mounted in glycerol plus PBS for examination.
To act as a positive control for RT-PCRs of Wnt-related genes, total RNA was isolated from whole B6CBA mouse embryos on days 8 and 10 after fertilization using TRIzol (Invitrogen), after the tissue had been pulverized in a microdismembrator (B. Braun Biotech International, Surrey, U.K., http://www.bbraunbiotech.com). TRIzol was also used to extract total RNA from confluent primary undifferentiated MSCs (passages 2–12), taken from three different donors, as well as commercially available MSCs. To compare Wnt gene expression patterns in primary hMSCs to those of commonly used mesenchymal cell lines, RNA was also extracted from confluent MC3T3-E1 cells and C2C12 cells. Genomic DNA from all samples was removed using the DNA-free DNase digestion kit (Ambion Europe Ltd., Cambridgeshire, U.K., http://www.ambion.com) according to the manufacturer's protocol. cDNA was synthesized from 5 μg total RNA using the Superscript II RT-PCR system, and RNA was then digested with RNase H (Invitrogen). The primer sequences used to identify Wnt gene expression are listed in Table 1. For all positively expressed genes, the amplified products have been sequenced and confirmed to represent the correct target gene. RT-PCR was performed with these primers for 35 cycles at 94°C for 10 seconds, 60°C for 20 seconds, and 72°C for 1 minute.
Table Table 1.. Primer sequences used for reverse transcriptase polymerase chain reaction analysis expression of Wnt signaling genes
Immunofluorescent Localization ofβ-catenin in Human MSCs following LiCl and Wnt3a Time-Course Addition
Cells were plated at 2.5 ×104 cells/cm2 in control medium on sterilized cover slips (13 mm diameter) in 24-well dishes. At time 0, fresh medium was added; the medium contained either 20 mM LiCl (Sigma-Aldrich) or 50% Wnt3a conditioned medium and was left for 0–24 hours. Control cells were treated for the same time periods with 20 mM NaCl (BDH) vehicle alone or 50% conditioned medium from the nontransfected control L cell line. The cells were then fixed in 4% paraformaldehyde for 5 minutes. Nonspecific binding was blocked with 10% FBS (in PBS with 0.1% Triton X-100; Sigma-Aldrich) for 30 minutes at room temperature. Rabbit anti-β-catenin antibody (1:1000 dilution; Sigma-Aldrich) was then added to the cells (in PBS with 0.1% Triton X-100) in a humidified box at 4°C overnight. Nonimmune rabbit IgGs were used in place of the primary antibody for the antibody controls. Cells were then incubated in FITC-labeled goat antirabbit IgG (1:100 in PBS) for 45 minutes. Finally, the cells were mounted in Vectashield mounting medium containing 4, 6-diamido-2-phenylindole (DAPI; Vector Laboratories, Peterborough, U.K., http://www.vectorlabs.com) to stain the nuclei. Cells were viewed using a Leica DMLA microscope (http://www.Leica-microsystems.com), from three geometrically predetermined fields, at constant exposure settings. Numbers of DAPI-positive and β-catenin–positive nuclei were determined using an automated program on Leica Quantimet image analysis system (Leica Q500win standard version 2.2 Leica Imaging Systems). For each time point, the percentage of cells with nuclei containing β-catenin was calculated for both Wnt3a and Li+ treatment and L-cell or NaCl controls. The results obtained after Wnt3a or Li+ treatment were then normalized against the relevant controls. Confocal images were taken using LSM 510 meta on an Axiovert 200M microscope (Zeiss, Oberkochen, Germany, http://www.zeiss.com).
Western Blot Analysis of Levels of Phospho-β-catenin in Li+-Treated MSCs
Confluent MSCs grown in 25 cm2 culture flasks were preincubated in 20 mM LiCl, 20 mM NaCl, or vehicle in control media for 4 hours. Cells were further incubated with 50 nM calyculin A (Sigma-Aldrich), a serine/threonine phosphatase inhibitor, for 0, 30, or 60 minutes before being lysed with 0.1% Triton in PBS containing protease inhibitors (protease inhibitor cocktail I [EMD Biosciences, Nottingham, U.K., http://www.cnbi.com]) at 4°C. Protein concentrations were determined using the BCA (bicinchoninic acid) protein assay (Perbio Science UK, Ltd., Cheshire, U.K., http://www.perbio.com), and 10 μg protein of each sample was separated by SDS-polyacrylamide gel electrophoresis (PAGE) before being transferred onto polyvinylidene difluoride (PVDF) membrane (Amersham Biosciences, Amersham, U.K., http://www4.amershambiosciences.com). The membrane was incubated in blocking buffer (4% dried milk in Tris-buffered saline containing 0.05% Tween-20 [TBS-T]) for 1 hour to minimize nonspecific binding. Anti-phospho-β-catenin antibody (New England Biolabs, Beverly, MA, http://www.neb.com/neb/) was used at 1:500 dilution in block solution overnight at 4°C and detected using a horseradish peroxidase (HRP)–labeled goat antirabbit IgG (1:1000; Sigma-Aldrich) for a further 1 hour. Specific antibody binding was visualized using ECL reagent with exposure to hyper film (both Amersham Biosciences). Antibodies were then stripped from the membrane with stripping buffer (62.5 mM Tris, pH 6.8, 0.7% β-mercaptoethanol, 2% SDS), before reprobing with anti-β-catenin antibody (Sigma-Aldrich) at 1:1000 dilution in blocking buffer for 1 hour at room temperature, for loading controls. Antibody detection was performed as described above. Relative antigenic positivity was measured using Alpha Imager 2000 software (Flowgen) coupled to a digital camera. Positivity of phosphorylated β-catenin levels were expressed as a ratio to that of total β-catenin.
Characterization of MSCs from BMSCs
A series of analyses were performed to confirm that cells isolated from the human bone marrow exhibited MSC characteristics. Primary adherent human BMSCs from six donors were cultured in control medium, and a cell sample was analyzed for expression of MSC markers using flow cytometry at each passage (Fig. 2A). The percentage of cells expressing the MSC markers CD29, CD44, CD73, CD90, CD166, and HLA class I was increased as the cells were cultured from passage 0 to passage 6. Cells were also found to be negative for the hematopoietic markers CD34 and CD45 and HLA-DR antigen. When cultured in appropriate differentiation conditions, these cells could also be induced to undergo osteogenic, adipogenic, and chondrogenic differentiation (Fig. 2B). These cells were designated MSCs.
Expression of Components of the Wnt Signaling Pathway by MSCs and Mesenchymal Cell Lines
RT-PCR analysis was used to determine which Wnt-related genes were expressed by primary MSCs, as well as two cell lines commonly used to model MSCs: MC3T3-E1 cells and C2C12 cells. As Wnt signaling is of major importance during development, cDNA from whole mouse embryos was used as a positive control for the Wnt primers when appropriate (Fig. 3). cDNAs from MSCs obtained from femoral heads from three different donors and a commercial source, at different passages, were analyzed, and expression patterns were found to be strikingly similar across all cell populations. Wnt2, 4, 5a, 11, and 16 expression was identified, along with Fz2, 3, 4, 5, and 6, as well as sFRP2, 3, and 4. Other signaling molecules—LRP-5, Dkk-1, kremen-1, Dvl-1, Dvl-2, Dvl-3, GSK-3β, APC, β-catenin, TCF1, and TCF4—were also shown to be expressed by these cells (Fig. 3). Expression of Wnt1, 2b, 3, 3a, 6, 7a, 7b, 8a, 10b, and 14; Fz1, 7, 8, and 10; and LRP-6 was not detected in MSCs using these techniques (data not shown). MC3T3-E1 cells and C2C12 cells were both shown to express Wnt4, 5, 6, 10a, 11, and 14 and Fz1, 4, and 6, with C2C12s also expressing Wnt2, 7b, and 16 and Fz3.
Functional Canonical Wnt Signaling in MSCs
Having determined expression of many components of the Wnt pathway by MSCs, canonical Wnt signaling was then mimicked in these cells using conditioned medium from a Wnt3a-overexpressing cell line or Li+ ions to inhibit GSK-3βand induce β-catenin stabilization. MSCs were incubated in 20 mM LiCl or 50% Wnt3a conditioned medium, and control cells were treated with 20 mM NaCl, vehicle, or 50% L-cell conditioned medium, for 0–24 hours, before immunofluorescent localization of β-catenin with DAPI nuclear staining (Fig. 4). In vehicle-treated, NaCl-treated, and L-cell conditioned medium–treated controls, β-catenin was predominantly immunolocalized at the cell periphery. However, weak β-catenin immunoreactivity was identified in some nuclei (Fig. 4A). In comparison, nuclear accumulation of β-catenin immunostaining was markedly increased in MSCs treated with Li+ and Wnt3a conditioned medium (Fig. 4A). These observations were quantified, and the percentage cells with nuclear β-catenin staining increased approximately threefold after 24 hours incubation with 20 mM LiCl or Wnt3a, compared with controls (Fig. 4B). Levels of phosphorylated β-catenin were increased in control cells following incubation with calyculin A, a serine/threonine phosphatase inhibitor that prevents dephosphorylation of β-catenin, once it has been phosphorylated by GSK-3β(Fig. 4C). However, after 1 hour treatment with calyculin A, levels of phosphorylated β-catenin were decreased in MSCs incubated with LiCl, compared with the NaCl- and vehicle-treated controls.
MSCs offer great therapeutic potential in a number of medical applications, but currently, relatively little is known about the signaling pathways involved in controlling MSC differentiation. Here we provide evidence of a functional role for canonical Wnt signaling in undifferentiated MSCs.
This work has shown expression of a wide range of components of Wnt signaling by MSCs, which suggests this signaling pathway may be of fundamental importance in these cells. The presence of β-catenin in the nucleus of a proportion of MSCs under control conditions, which increased following Wnt 3a treatment and Li+ application to inhibit GSK-3β, provided evidence that canonical Wnt signaling is functional in MSCs. Furthermore, an accumulation of phosphorylated β-catenin was identified in control cells treated with calyculin A, a phosphatase inhibitor, but not in Li+-treated cells, indicating an Li+-induced inhibition of β-catenin phosphorylation by GSK-3βin primary MSCs. Lithium affects development in a wide range of organisms, producing phenotypes consistent with overstimulation of canonical Wnt signaling. Until recently, isolated Wnt proteins have lacked bioactivity, and consequently, lithium has been routinely used to mimic canonical Wnt signaling [48–50]. However, lithium also inhibits the enzyme inositol monophosphatase (IMPase) , causing a reduction in inositol levels and thereby preventing the generation of inositol tris-phosphate (IP3). Despite this, complete inhibition of IMPase, with an inhibitor 1,000 times more potent than lithium, has no effect on the morphogenesis of Xenopus embryos ; moreover, Dictyostelium discoideum mutants lacking the ability to generate IP3 develop normally . Together, these results suggest that any inhibition of IMPase by lithium does not affect developmental processes. It is now clear that more specific techniques to manipulate particular components of the canonical Wnt signaling pathway are necessary to determine more precisely how mesenchymal differentiation is mediated through canonical Wnt signaling in MSCs. Recent evidence suggests that the application of purified bioactive Wnt proteins, or inhibitors such as Dkk-1 [28, 47, 56], combined with molecular manipulation of Wnt-mediated cascades, may provide a sophisticated mechanism for controlling MSC self-renewal and differentiation.
Our findings support previous data, which suggest that Wnt signaling may regulate differentiation pathways in various mesenchymal cell types. Endogenous canonical Wnt signaling in preadipocytes inhibits further differentiation along this lineage and spontaneous differentiation occurs following inhibition of this pathway. Wnt10b has been proposed as the specific Wnt protein responsible for the endogenous signaling activity , as expression levels are high in preadipocytes, but decrease when differentiation is stimulated. Also, overexpression of Wnt10b in preadipocytes retards their differentiation .
However, activated Wnt signaling has also been shown to promote mesenchymal differentiation; for example, Wnt3a and 5b have been implicated in the induction of myogenesis . During chondrogenesis, Wnt signaling acts as both a positive and negative regulator, depending on which specific Wnt protein is responsible for the signaling. Wnt1, 3a, 4, 7a, and 7b promote proliferation and chondrogenic differentiation, whereas Wnt5a and 11 (which we have shown to be expressed by undifferentiated MSCs) retard these processes [33, 57]. The functional diversity of the various Wnt proteins are thought to be due, at least in part, to the specific Fz with which they bind and the particular signaling cascade that is subsequently stimulated within the cell. It now appears likely that individual Wnts are able to bind multiple Fz receptors, and each receptor can bind numerous Wnt proteins, providing a wide range of possible Wnt/Fz pairs (for review, see ). Stimulation of a pathway by a particular Wnt also appears to depend on the experimental system employed; for example, Wnt4 has been proposed to signal via both canonical  and noncanonical  mechanisms. Therefore, the specific intracellular pathways stimulated by each of the Wnt proteins expressed by MSCs need to be established to fully understand the signaling mechanisms involved in regulating differentiation.
However, we appear to have identified a consistent Wnt and Fz expression profile of human primary MSCs isolated from different donors. Expression of Wnt2, 4, 5a, 11, and 16 and Fz2, 3, 4, 5, and 6 was identified in MSCs from all sources, at different passages, and appeared to be unaffected by periods of cryostorage. This common Wnt/Fz expression profile may prove important, not only in the identification of specific Wnt-dependent signaling in MSC proliferation and differentiation programs but also in the characterization of the MSC phenotype. Using mRNA extracted from the mesenchymal cell lines MC3T3-E1 and C2C12, we identified a similar, but not identical, expression pattern, with these cells sharing expression of Wnt4, 5a, and 11 and Fz4 and 6. This appears to support previously published data suggesting differences between mouse and human MSCs .
Previous publications have revealed that some Wnt signaling molecules are differentially expressed by MSCs at different cell densities . Our expression profiling results were obtained using cDNA collected from MSCs on reaching confluency. Although we did not find evidence of LRP-6 expression in MSCs under these conditions, consistent with findings from Gregory et al. , the Wnt signaling pathway could still be functional, as we did find expression of LRP-5. Dkk-1 expression was also found in MSCs at this cell density, although Gregory and co-workers provide evidence to suggest that Dkk-1 expression by MSCs is cell-density dependent .
Once secreted, Wnt proteins remain associated with the cell surface or extracellular matrix of the cell from which they were produced and are involved in autocrine and paracrine signaling mechanisms . Consequently, the target cells for the Wnt proteins expressed by MSCs may be either these cells themselves or other cell types in the bone marrow. In particular, Wnt signaling has been shown to regulate the differentiation of several hematopoietic lineages [29, 61]. Our evidence that MSCs express both Wnt ligands and Fz receptors, with some previously identified Wnt/Fz pairs (such as Wnt5a/Fz5 ), would support our observation of endogenous Wnt signaling in MSCs and suggests that MSC regulation is, at least in part, autocrine. These data also support previous evidence of Wnt5a expression by fetal and adult human BMSCs  and more recent data that Wnt5a is expressed by MSCs at the end of the exponential growth and stationary phases in culture, suggesting a role for this particular Wnt in controlling MSC proliferation .
Our evidence suggests that an endogenous level of Wnt signalling is required for the maintenance and functionality of MSCs. These data indicate that further studies concerning the role of specific Wnt and Fz pairings in the MSC will greatly enhance our current understanding of the signaling mechanisms involved in both stem cell maintenance and mesenchymal differentiation.
This research was supported by BBSRC, Oliver Bird Fund, Smith and Nephew plc and Arthritis Research Campaign.