SEARCH

SEARCH BY CITATION

Keywords:

  • Mesenchymal stem cells;
  • Mitogen-activated protein kinase;
  • Phospholipase C;
  • Signal transduction;
  • Regenerative medicine;
  • Adipogenesis;
  • Osteogenesis;
  • Ex vivo expansion

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Disclosures
  9. Acknowledgements
  10. References

Bone marrow mesenchymal stem cells (BMMSCs) are pluripotent cells capable of differentiating into several cell types and are thus an attractive cell source for connective tissue engineering. A challenge in such a use is expansion and directed seeding in vitro, requiring proliferation and survival, and directed migration, respectively, prior to functional differentiation. The epidermal growth factor (EGF) receptor (EGFR) is the prototypal growth factor receptor and elicits these responses from a wide variety of stromal, epithelial, and endothelial cells. Ligands for this receptor are appealing for use in tissue engineering because they are relatively resistant to biological extremes and amenable to high-volume production. Therefore, we determined whether an EGFR ligand, EGF, could be used for ex vivo expansion of BMMSCs. EGF stimulated motility in rat and immortalized human BMMSCs. EGF-induced proliferation was observed in immortalized human BMMSCs but was not apparent in rat BMMSCs under our experimental conditions. EGF did not, however, rescue either type of BMMSC from apoptosis due to lack of serum. During our examination of key signaling intermediaries, EGF caused robust phosphorylation of extracellular signal-regulated protein kinase (ERK) and protein kinase B/akt (AKT) but only minimal phosphorylation of EGFR and phospholipase C-γ in rat BMMSCs, whereas in the human BMMSCs these intermediaries were all strongly activated. EGF also induced robust ERK activation in primary porcine mesenchymal stem cells. EGF pretreatment or cotreatment did not interfere with secondarily induced differentiation of either type of BMMSC into adipogenic or osteogenic lineages. Platelet-derived growth factor (PDGF) effects were similar to but not additive with those elicited by EGF, with some quantitative differences; however, PDGF did interfere with the differentiation of these BMMSCs. These findings suggest that EGFR ligands could be used for ex vivo expansion and direction of BMMSCs.


Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Disclosures
  9. Acknowledgements
  10. References

Regenerative medicine holds the promise of using cell transplantation to ameliorate a variety of devastating clinical disorders. One approach is to use the patient's own stem cells so as to avoid immunological barriers and concomitant morbidity. Bone marrow mesenchymal stem cells (BMMSCs), originally isolated as single-cell suspensions of bone marrow colonies of fibroblast-like cells that adhere to plastic [1], are pluripotent cells capable of differentiating into several cell lineages, including osteoblasts, chondrocytes, adipocytes, skeletal and cardiac myocytes, endothelial cells, and neurons in vitro and in vivo after transplantation [29]. Although the in vivo pathophysiologic functions of BMMSCs are still being deciphered, the pluripotency of BMMSCs in vitro prompts many to postulate important roles in pathophysiologic processes, including wound healing and tissue regeneration or rejection, especially after transplantation [10, 11]. These cells have also been shown to be incorporated and differentiated into various tissues upon implantation [3, 7, 12]. These characteristics make BMMSCs good candidates for cell transplantation therapies for various tissue regenerations and vehicles of gene therapies [13, 14].

One limitation to using stem cells, particularly for autologous cellular transplantation, is that the small number that can be readily obtained requires extensive expansion for therapeutic utility. Two general routes to this expansion are in vivo and ex vivo. The former can be used for autologous implantation with little to no manipulation or for simple intraoperative separations, such as proposed for bone wound healing [15]. This has been driven in part by the limited success in expanding human BMMSCs [16]. The ex vivo approach has growing appeal for a variety of other applications in which cell manipulation ex vivo is acceptable or desirable, such as using BMMSCs for cardiac repair, for delivery of targeted anticancer agents, or for cases of allogeneic transplantation [14, 17]. Proliferation and migration are important cell processes required to expand and direct such cells both in vivo and ex vivo and to differentiate and regenerate tissues and/or organs in vivo. Current ex vivo expansion strategies generally rely on the use of serum or conditioned media, which not only carry inherent disease risks [18] but hinder standardization that is critical to establishing a broad clinical adoption.

There is thus a strong motivation to identify factors that might be used in serum-free formulations to expand BMMSCs ex vivo without influencing differentiation capacity or to release locally in vivo to influence growth of BMMSCs within the therapeutic site. In vivo approaches to influencing bone regeneration have primarily focused, with mixed success, on the family of bone morphogenetic proteins [19]. The role of these factors, particularly in more generalized BMMSC expansion, is not clear. The epidermal growth factor (EGF) receptor (EGFR), the prototypal growth factor receptor, exerts various actions, including cell migration and proliferation, on a wide variety of cell types [20, 21]. The expression of EGFR in BMMSCs has been reported [22]. What makes this signaling system attractive from a tissue engineering and production standpoint is that EGF is inexpensive, amenable to high-volume production under good manufacturing processes, remarkably stable under a wide range of conditions, and easy to manipulate [21, 23, 24]. Very recently, it was reported that human primary BMMSCs can be stimulated to proliferate by the EGFR ligand heparin-binding EGF-like growth factor (HB-EGF), which interacts with both erbB4 and EGFR [25], although this EGFR ligand interferes with cellular differentiation when present as an additive or in a paracrine mode from feeder cells. Little is known, however, about how this signaling system directs BMMSC survival and migration or about the activation and regulation of key intracellular signaling pathways that regulate these cellular functions in response to EGFR activation.

If EGFR ligands, particularly the robust EGF ligand, broadly promote migration, proliferation, and survival without either inducing differentiation or preventing further differentiation by other signals, these readily available and well characterized ligands would hold some promise for expanding ex vivo autologous BMMSCs for therapeutic use. In this study, we investigated whether the prototypical EGFR ligand, EGF, could be used for ex vivo expansion of BMMSCs. EGF treatment preserves pluripotency of BMMSCs of both human and rat origin and did not inhibit differentiation. EGFR signaling drove proliferation and migration of human BMMSCs while not adversely affecting cell survival. Furthermore, we demonstrated that EGF activates EGFR signaling pathways in primary porcine MSCs, a preferred model for human craniofacial bone regeneration. These data suggest the possibility of applying EGFR ligands to ex vivo expansion of BMMSCs.

Materials and Methods

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Disclosures
  9. Acknowledgements
  10. References

Materials

Human recombinant EGF was obtained from BD (Franklin Lakes, NJ, http://bd.com). Platelet-derived growth factor (PDGF)-BB and basic fibroblast growth factor (bFGF) were from Peprotech (Rocky Hill, NJ, http://www.peprotech.com). Human recombinant HB-EGF and anti-alkaline phosphatase antibody were from R&D Systems Inc. (Minneapolis, http://www.rndsystems.com). The EGFR kinase inhibitor PD153035 was from Calbiochem (San Diego, http://www.emdbiosciences.com). Phospho-AKT (protein kinase B/akt) antibody, phospho-PLC (phospholipase C)-γ-1 (Tyr783) antibody, and phospho-P44/42 extracellular signal-regulated protein kinase/mitogen-activated protein kinase (ERK/MAPK) (Thr202/Tyr204) antibody were purchased from Cell Signaling Technology (Beverly, MA, http://www.cellsignal.com). Anti-EGFR and anti-phospho-EGFR (Tyr 1173) were from Upstate USA Inc. (Charlottesville, VA, http://www.upstate.com). Anti-pan ERK was obtained from BD Transduction Laboratories (Lexington, KY, http://www.bdbiosciences.com). Anti-Adipocyte Fatty Acid Binding Protein (anti-ALBP, aP2) was from ProSci Incorporated (Poway, CA, http://www.prosci-inc.com), and the anti-osteocalcin antibody was from Biomedical Technologies, Inc. (Stoughton, MA, http://www.btiinc.com). Anti-neurofilament antibody was from Chemicon International, Inc. (Temecula, CA, http://www.chemicon.com). Anti-β-actin antibody was from Sigma-Aldrich (St. Louis, http://www.sigmaaldrich.com) and anti-α-tubulin antibody was from Cal-biochem (San Diego, http://www.emdbiosciences.com). Horse-radish peroxide (HRP)–conjugated secondary antibodies for enhanced chemiluminescence (ECL) were from Promega (Madison, WI, http://www.promega.com), Biosource International (Camarillo, CA, http://www.biosource.com), or Jackson ImmunoResearch Laboratories, Inc. (West Grove, PA, http://www.jacksonimmuno.com). Advanced Dulbecco's modified Eagle's medium (DMEM) or minimum essential medium (MEM) culture medium and fetal bovine serum (FBS) were from Gibco BRL (Carlsbad, CA, http://www.gibcobrl.com), and charcoal-treated FBS was from Gemini Bio-Products (Woodland, CA, http://www.gembio.com). All of the remaining cell culture media and supplements were from Cellgro (Kansas City, MO, http://www.cellgro.com) unless otherwise stated.

Cell Culture

Cultured rat BMMSCs derived from Lewis rats were provided courtesy of Dr. Darwin J. Prockop (Tulane University, New Orleans) and were cultured in MEM-α supplemented with 20% FBS and 2 mM L-glutamine [26]. Three different collections of rat BMMSCs were queried with identical results. Cultured immortalized human BMMSCs (AOH) were the kind gift of Dr. Junya Toguchida (Kyoto University, Kyoto, Japan) [27]. Cultured porcine BMMSCs were provided courtesy of Dr. Joseph P. Vacanti [28] (Massachusetts General Hospital, Boston). All of these cells were cultured in DMEM supplemented with 10% FBS.

Because of a lack of specific markers or criteria for identification of BMMSCs [12], we first confirmed their identities as BMMSCs by establishing the optimal condition of adipogenic, osteogenic, neurogenic, and chondrogenic differentiation of rat, porcine, and immortalized human BMMSCs. Osteogenic differentiation for rat BMMSC medium included 10 nM dexamethasone, 10 mM β-glycerophosphate, and 200 μM ascorbic acid in the full MEM-α culture medium above [26], and the osteogenic differentiation medium for porcine BMMSCs contained 10 nM dexamethasone, 10 mM β-glycerophosphate, and 284 μM ascorbic acid in the full DMEM culture medium above [28]. For osteogenic differentiation of immortalized human BMMSCs, the full DMEM culture medium above was supplemented with 100 nM dexamethasone, 10 mM β-glycerophosphate, and 50 μM ascorbic acid [27]. One micromolar dexamethasone, 100 μg/ml 3-isobutyl-1-methylxanthine, 5 μg/ml insulin, and 60 μM indomethacin were added to the full MEM-α medium above for rat and human BMMSC adipogenic differentiation and to the full DMEM medium above for porcine BMMSC adipogenic differentiation [29]. All of these supplemental compounds or reagents were from Sigma-Aldrich. Chondrogenic differentiation medium from Cambrex (East Rutherford, NJ, http://www.cambrex.com) was used for immortalized human BMMSCs [27]. For neurogenic differentiation, rat BMMSCs were cultured in MEM-α culture medium supplemented with 5% charcoal-treated FBS and 10 nM bFGF [30].

Cells were cultured for 2–4 weeks in adipogenic differentiation, for 4–6 weeks in osteogenic medium, and for 1 week in neurogenic medium; cell pellets were cultured for 4 weeks in chondrogenic differentiation medium. Adipogenic differentiation was confirmed by Oil Red O staining, and osteogenic differentiation was confirmed by Von Kossa staining. In brief, after a rinsing with phosphate-buffered saline (PBS), cells were fixed with 10% formalin in PBS for 30 minutes. Cells were stained with filtered 0.4% Oil Red O solution for confirmation of adipogenic differentiation. For Von Kossa staining, fixed cells were incubated with 1% silver nitrate for 60 minutes under UV light, washed with 2.5% sodium thiosulfate, and counter-stained with Neutral red. Chondrogenic differentiation was confirmed by Alcian Blue staining. Neurogenic differentiation was confirmed by cell morphology, and the expression of neurofilament was detected by immunoblotting. Each of these cell types underwent osteogenic, adipogenic, neurogenic, and chondrogenic differentiation as expected (Fig. 1; supplemental online Figs. 1 and 2), and therefore these undifferentiated mesenchymal cells fulfilled the criteria and were confirmed as BMMSCs [12]. These studies were approved by the Institutional Review Board (under exemption 4e) and Institutional Animal Care and Use Committee of the University of Pittsburgh.

Proliferation Assay

Cell proliferation was determined by actual enumeration of cells to ensure an increase in cell number. Cells were seeded on collagen-coated 12-well plates at a density of 8 × 104 cells per cm2 (2.74 × 105 cells per well) and attached on the plate overnight. Cell culture medium was then switched to Advanced MEM or DMEM containing 2% FBS and the test agent with or without PD153035 (or other inhibitors) as indicated. The medium was changed within 48 hours with the same medium, and cell numbers of each well were measured 96 hours after starting stimulation by Coulter Cell Counter Z2 (Beckman Coulter, Inc., Fullerton, CA, http://www.beckmancoulter.com).

Apoptosis Assay

Apoptosis was assessed with APOPercentage APOPTOSIS Assay (Biocolor Ltd., Newtownabbey, Northern Ireland, http://www.biocolor.co.uk). In brief, cells in the regular media were plated in a 96-well plate at 5 × 105 cells per cm2 and allowed to attach on the plate overnight. The cells were challenged with culture media containing APOPercentage dye, low serum, and various agents. The transfer of phosphatidylserine to the outer surface of the membrane permits the transfer of the dye into the cell, and apoptotic cells incorporating the dye were detected by measuring absorbance at 550 nm with TECAN SPEC-TRAFLUOR microplate reader (TECAN Austria GmbH, Grödig, Austria, http://www.tecan.com).

Motility Assay

Cell motility was assessed by migration into a denuded area in a two-dimensional in vitro wound-healing assay [31]. In brief, after 24 hours of quiescence with MEM-α or DMEM supplemented with 0.5% dialyzed FBS, a confluent cell monolayer on a six-well plate was denuded by rubber policemen and then stimulated with the test agent in the presence of mitomycin with or without EGFR tyrosine kinase inhibitor. Digital images were captured at 0 and 24 hours, and the relative distance traveled by the cells was determined under ×40 magnification with Photo-shop (Adobe Systems Inc., San Jose, CA, http://www.adobe.com).

Immunoblotting

Cells were grown to confluence in six-well plates. After 24 hours of quiescence in MEM-α or DMEM supplemented with 0.5% dialyzed FBS, cells were treated with test agents in the presence and absence of EGFR tyrosine kinase inhibitor. Cells were lysed with sodium dodecyl sulfate (SDS)–sample buffer containing 0.1 M Tris-HCl, 4% SDS, 0.2% Bromophenol Blue, and 5% β-mercaptoethanol. Cell lysates were separated by SDS-PAGE and transferred to a polyvinylidene difluoride (PVDF) membrane (Millipore, Billerica, MA, http://www.millipore.com). Blots were probed by primary antibodies before visualizing with HRP-conjugated secondary antibodies followed by development with an ECL kit (Amersham Biosciences, Piscataway, NJ, http://www.amersham.com) or SuperSignal West Femto (Pierce, Rockford, IL, http://www.piercenet.com).

EGFR Expression Levels

The expression level of EGFR was determined by a standard Scatchard binding assay as described previously [32]. Briefly, cells were grown to subconfluence in 12-well plates coated with collagen I and washed twice with binding buffer (DMEM with 1% bovine serum albumin [Fraction V; Sigma-Aldrich] and HEPES), and 0.1 nM [125I]EGF (Amersham Biosciences) was added to unlabeled EGF (0–100 nM) in binding buffer. Plates were incubated for 2 hours at 4°C, and then the unbound-labeled EGF was collected. Cells were lysed with lysis buffer (Tris-buffered saline with 1% SDS). Both unbound and bound radioactivity were counted by a gamma-counter (Beckman Coulter). The number of binding sites was calculated by Scatchard analysis using linear regression.

Statistical Analysis

Cell enumeration, motility, and apoptosis were analyzed using paired t tests. Significance was set at p < .05 or more stringent as noted in the text and figure legends.

Results

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Disclosures
  9. Acknowledgements
  10. References

EGFR Signaling Drives Motility of Rat and Immortalized Human BMMSCs

EGFR is a prototypal receptor tyrosine kinase present on most adherent cells, including stromal cells [21], and likely is expressed on BMMSCs as demonstrated by reported cellular responses [22, 25]. Scatchard analyses provided evidence for accessible and specific binding of EGF at levels of less than 5,000 EGFRs per cell in rat BMMSCs (reliable limits of calculation by Scatchard), approximately 7300 ± 390 EGFRs per cell in immortalized human BMMSCs, and approximately 11,000 ± 2000 EGFRs per cell in porcine BMMSCs (n = 3). The binding had KD measurements in the range expected for specific EGF/EGFR interactions (2.9 nM for immortalized human BMMSCs and 8.5 nM for porcine BMMSCs; the rat BMMSCs were too low in number to derive reliable KD measurements) [33].

With the presence of specific EGF ligand binding confirmed, we determined whether this ligand-receptor interaction could affect cell behavior. EGF promoted rat BMMSC motility in a dose-dependent manner as determined in a two-dimensional in vitro wound-healing assay (Fig. 2A; supplemental online Fig. 3A). EGF enhanced motility significantly but only at half the level induced by PDGF; the EGF enhancement was noted at 0.1 nM and reached the maximal level at 1 nM (supplemental online Fig. 3A). The effects of EGF and PDGF on rat BMMSC motility were not additive when each was applied at maximal stimulatory concentrations (supplemental online Fig. 4A), presumably because the cells had already reached maximum motility with PDGF stimulation. EGF also significantly promoted motility of immortalized human BMMSCs (Fig. 2B) as strongly as PDGF. However, the maximal induction of either EGF or PDGF on motility was less in the immortalized human BMMSCs than in the rat BMMSCs. This might be due to a much higher level of basal (nonstimulated) motility in the immortalized human BMMSCs (643.7 ± 31.4 μm for human BMMSCs vs. 89.1 ± 15.2 μm for rat BMMSCs, n = 3; p < .001). EGF-induced motility of both BMMSC populations was prevented by the EGFR tyrosine kinase inhibitor PD153035, whereas PDGF-induced cell motility was not, confirming the involvement of EGFR in EGF-induced cell motility. Interestingly, PD153035 did not reduce the high basal motility of the immortalized human BMMSCs, strongly suggesting that this was not secondary to an EGFR-mediated autocrine signaling loop that is present in many differentiated cell lineages.

EGFR Signaling Stimulates Human but Not Rat BMMSC Proliferation

Because EGFR signaling also promotes proliferation and limits apoptosis in stromal cells [21], we examined the effect of EGF on these processes in BMMSCs. After 96 hours, the total number of cells in culture increased approximately fourfold in rat and 15-fold in human from the starting number in Advanced MEM or DMEM containing 2% FBS. The addition of EGF or PDGF-BB increased this proliferation by an additional 30% in human BMMSCs, similar to the increase in proliferation reported for HB-EGF [25], but did not further increase the number of rat BMMSCs (Fig. 3A, 3B; supplemental online Fig. 5). The same experiments with regular MEM-α or DMEM, rather than Advanced media, showed lower baseline proliferations (approximately 1.5-fold in both cell types) with higher variances due to apparent higher cell deaths rates, and yielded the same minimal (but statistically significant for human BMMSCs) increases in cell number by growth factors (data not shown). The EGFR inhibitor blocked the EGF-induced, but not PDGF-induced, increase in cell number in human BMMSCs. The basal level was unchanged by this inhibitor, also ruling out an EGFR autocrine loop as the reason for high basal proliferation.

EGF-induced proliferation of human BMMSCs was noted at < 1/10 KD (0.1 nM) and reached maximal stimulation near KD (1 nM); however, EGF-induced proliferation was blunted at > 10 KD (100 nM) of EGF (supplemental online Fig. 3B), presumably due to the well described effects of ligand-induced EGFR downregulation [34]. Mitogenesis is more sensitive to this due to both the time course of the cellular response (migration occurring from within 2 hours whereas proliferation requires at least 8 hours of EGFR signaling) and differential spatiotemporal aspects of EGFR signaling for proliferation and migration. Again, there was no synergy or additivity of EGF and PDGF when each was applied at maximal stimulatory concentrations, presumably because the cells were already maximally stimulated by either factor alone (supplemental online Fig. 4B).

EGF Does Not Protect BMMSCs from Apoptosis Induced by Low Serum

This limited EGF-induced proliferation of BMMSCs may be due to either near maximal induction of proliferation by even only 2% FBS in Advanced media or concomitantly increased apoptosis. In such conditions, apoptosis was minimal in both rat and human BMMSCs (Fig. 4). At lower levels of FBS that induce quiescence but not apoptosis in stromal cells [35], apoptosis was evident, and addition of EGF or PDGF did not save cells from apoptosis (Fig. 4). At FBS concentrations of 2% and higher, there was little apoptosis either in the presence or absence of growth factor. Advanced media further decreased apoptosis in these conditions in both cell types. Therefore, we suspect that the limited EGF and PDGF induction of proliferation in rat BMMSCs is due to maximal induction of proliferation by 2% FBS in Advanced media, and the moderate increase (30% at 10 nM EGF) in proliferation of immortalized human BMMSCs also reflects the strong influence of serum.

EGFR Ligands Activate ERK, AKT, and PLC-γ in BMMSCs

Key molecular switches, the activation status of which can be surmised from the phosphorylation state, have been defined for EGFR-mediated pathways. Phospho-PLC-γ supports motility, phospho-ERK contributes to both motility and proliferation [20], and phospho-AKT promotes cell survival. In immortalized human BMMSCs, EGF and HB-EGF led to EGFR phosphorylation, which was suppressed by the EGFR kinase inhibitor PD153035; but EGFR phosphorylation was not apparent in rat BMMSCs, most likely due to the low levels of EGFR accessible on these cells, and was minimal in porcine BMMSCs (Fig. 5). ERK demonstrated similar patterns in each of these lineages, in that EGF and HB-EGF caused robust ERK activations that were suppressed by PD153035, whereas PDGF provided robust activations that were not suppressed by PD153035. Interestingly, PLC-γ was activated strongly by PDGF in each of these lineages. However, EGF-induced PLC-γ phosphorylation was noted only in the immortalized human BMMSCs; this differential activation correlates with the ability of EGF to induce motility as strongly as PDGF in human BMMSCs but half as strongly as PDGF in rat BMMSCs (Fig. 2). Although EGF or PDGF did not protect BMMSCs from apoptosis induced by low serum (Fig. 4), these reagents activated protein kinase B (PKB)/AKT, a molecular switch for cell survival [4].

EGF-induced phosphorylation of another signaling pathway related to proliferation and motility, signal transducer and activator of transcription (STAT)-3 [36], was not detected in either rat or immortalized human BMMSCs (data not shown). This might correlate with the minimal or absent affects of EGF on cell proliferation.

EGFR Signaling Does Not Interfere with Subsequent Differentiation of BMMSCs

Although expansion of the BMMSC population is necessary, it must be accomplished in a manner that does not induce differentiation or interfere with the subsequent ability to differentiate into the desired cell type. We determined whether the cells can still be differentiated by routine methods after 5 days of EGF or PDGF exposure. EGF or PDGF pretreatment did not interfere with rat or immortalized human BMMSC differentiation by adipogenic or osteogenic medium; however, PDGF pretreatment might delay the subsequent differentiations (Fig. 6).

We also evaluated whether differentiation could occur in the presence of growth factors. EGF or PDGF alone did not drive BMMSCs into adipogenic or osteogenic lineages. On the other hand, EGF or PDGF did not interfere with neurogenic differentiation of rat BMMSCs (supplemental online Fig. 2B). The situation for other differentiations illuminated differences between the growth factors. The addition of EGF did not inhibit adipogenic differentiation of rat or immortalized human BMMSCs. Interestingly, PDGF enhanced adipogenic differentiation of rat BMMSCs but inhibited that of immortalized human BMMSCs (Fig. 7). The addition of EGF to osteogenic medium did not inhibit osteogenic differentiation of BMMSCs; however, the addition of PDGF minimally promoted osteogenic differentiation in immortalized human BMMSCs but did not promote osteogenic differentiation in rat BMMSCs (Fig. 7). The addition of PDGF to the regular medium also enhanced the expression of nonspecific alkaline phosphatase, an early marker of osteogenic differentiation in immortalized human BMMSCs (Fig. 7) [37], but it did not yield typical morphology of osteogenic differentiated cells.

These data suggested that EGFR alone or the addition of EGF to differentiation media does not interfere with pluripotency of BMMSCs or drive BMMSCs into specific lineages. This would validate the use of EGF for ex vivo expansion of BMMSCs.

Discussion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Disclosures
  9. Acknowledgements
  10. References

The data presented here suggest that EGF promotes proliferation and migration without inhibiting the pluripotency of BMMSCs. Although cell transplantation therapies with BMMSCs hold great promise for various tissue regenerations and vehicles of gene therapies, a major stumbling block is the low numbers of such BMMSCs necessitating ex vivo expansion. Our findings suggest that EGF could be used for ex vivo expansion of BMMSCs prior to cell transplantation and may potentially be useful in vivo if appropriate delivery strategies are deployed.

Engineered cellular transplantation requires an increase in both number and migration prior to terminal differentiation. Compared with serum alone, EGF increased the number of human BMMSCs to a limited extent. Although the increase in cell number at 96 hours (or three population doublings) was only approximately one third, such an increase over 20 population doublings can be quite large (more than sixfold), as seen with similar increases in proliferation driven by HB-EGF [25]. Interestingly, this was not the case for rat BMMSCs. However, the proliferation rate in the absence of ligand was robust in Advanced media with FBS. A possible reason for this limited proliferation is the low level of EGFR on human BMMSCs and even lower level on the rat cells. The proliferation-related molecular switch, ERK, was activated in both cell types, although STAT3, which is also related to proliferation, was not induced in either cell type. It was of interest to see whether EGF could replace serum in supporting cell proliferation. In low serum, apoptosis was high and could not be rescued with EGF or PDGF. Although EGF and PDGF activated PKB/AKT, a key molecular switch for survival (Fig. 5), anti-apoptotic effects of EGF and PDGF were not apparent in either BMMSC population (Fig. 4). The presence of autocrine/paracrine factor was suggested as a BMMSC survival factor against apoptosis [38], but conditioned media did not rescue cells from apoptosis in our study (data not shown). However, we observed that BMMSCs in high-density culture are more resistant to low-serum conditions than in low-density culture (data not shown), a finding that agrees with one report stating that BMMSCs in low-density culture decreased the expression of p21 and p27 cyclin-dependent kinase inhibitors, which induces apoptosis of BMMSCs [39]. The upstream signal is beyond the scope of the present study, but these could be explained by the presence of paracrine, juxtacrine, or matrix protein survival factor(s).

Both EGF and PDGF stimulated motility of both rat and immortalized human BMMSCs (Fig. 2). In immortalized human BMMSCs, EGF stimulated cell motility as strongly as PDGF, and both EGF and PDGF robustly activated ERK as well as PLC-γ (Fig. 5). In rat BMMSCs, EGF enhanced cell motility only at half the level induced by PDGF, and EGF robustly activated ERK but not PLC-γ, whereas PDGF robustly activated both ERK and PLC-γ (Fig. 5). PLC-γ is a key molecular switch for cell migration [20], and the relatively weak motogenic potency of EGF on rat BMMSCs might come from the apparent lack of PLC-γ activation.

The effects of EGF and PDGF on adipogenic differentiation have been controversial. The addition of EGF or PDGF to adipogenic medium promoted adipogenesis of 3T3-L1 preadipocytes in one report [40], whereas the addition of PDGF inhibited adipogenic differentiation of 3T3-L1 preadipocytes in another report [41]. In our data, the addition of EGF to adipogenic medium did not promote or inhibit adipogenic differentiation of rat or immortalized human BMMSCs, but the concomitant addition of PDGF promoted adipogenic differentiation in rat BMMSCs and inhibited it in immortalized human BMMSCs (Fig. 7A). The reason for this differential effect of PDGF on adipogenic differentiation is unclear. It is of interest to note that HB-EGF has recently been reported to reversibly block BMMSC differentiation [25]. The difference in differentiation blockage between HB-EGF in the recent report and EGF herein may be due to differential receptor downregulation (Fig. 5). These discrepancies require further exploration in studies that lie beyond these initial communications.

EGF was reported to inhibit collagen synthesis and alkaline phosphatase activity [42, 43], and EGFR signaling was proposed as a negative regulator of osteogenic differentiation in BMMSCs and preosteoblastic MC3T3 E1 cells [22, 44]. Moreover, a biphasic effect of EGF on the formation of mineralized nodules by rat calvarial cells was reported: the continuous exposure of EGF caused a dose-dependent inhibition of mineralized nodule formation, whereas a short exposure of EGF (4–48 hours) increased it [45]. In our experimental conditions, EGF did not interfere with osteogenic differentiation of either BMMSC population (Fig. 7). PDGF promoted osteogenesis in rabbit and rat calvarial defect models [46, 47], and PDGFR expression was associated with bone forming in BMMSCs [22]. Indeed, the addition of PDGF to the osteogenic medium minimally enhanced osteogenic differentiation of immortalized human BMMSCs (Fig. 7). However, PDGF was also reported to have negative effects on osteogenic differentiation of BMMSCs in one report [48].

The EGF pretreatment of rat and immortalized human BMMSCs did not affect the adipogenic or osteogenic differentiation of BMMSCs after EGF pretreatment (Fig. 6), confirming that ex vivo expansion of BMMSCs by EGF should not interfere with subsequent differentiations, and thus it holds promise for EGF use in ex vivo and in vivo expansion of BMMSCs. Although minimal, the PDGF pretreatment appeared to delay the subsequent osteogenic and adipogenic differentiation of BMMSCs (Fig. 6). PDGF might be a stronger mitogen and motogen for BMMSCs than EGF, but the potential interference with subsequent differentiation of BMMSCs by PDGF exposure should be considered cautiously if PDGF is considered for ex vivo expansion of BMMSCs.

It was recently reported that, as a constituent in osteogenic differentiation media, EGF promoted this phenotypic transformation of immortalized human BMMSCs whereas PDGF did not [49]. At early stages in the osteogenic induction procedure, inclusion of EGF accelerated the phenotypic transformation. Using a proteomic approach, Kratchmarova et al. found that only limited activation of phosphatidylinositol 3-kinase (PI3-K) was responsible for this differential effect. Our data appear superficially to be somewhat at odds with this communication although the differences most likely relate to the time point at which the system was queried; we examined osteogenesis at the end stage (28 days), whereas Kratchmarova et al. investigated during the first week. Interestingly, even at short time periods, we do note AKT phosphorylation (a downstream event initiated via PI3-K) by EGF, even if less rigorously than by PDGF (Fig. 5), whereas Kratcharova et al. do not report this molecule as being phosphorylated. However, the fundamental difference is that our goal was to define factors that can be used to expand stem cell populations ex vivo prior to directed differentiation by other defined media and inducers. Thus, the data presented here are not contradicted and are of value in the generation of these stem cell pools.

Conclusion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Disclosures
  9. Acknowledgements
  10. References

EGFR ligand signaling promotes proliferation and migration of rat and immortalized human BMMSCs without inhibiting their pluripotency after EGFR ligand exposure. EGFR ligands could be used for ex vivo expansion of BMMSCs prior to cell transplantation therapies for various tissue regenerations and vehicles of gene therapies, or via controlled release in vivo to influence BMMSC proliferation in situ.

thumbnail image

Figure Figure 1.. Differentiation of rat (A) and immortalized human (B) BMMSCs into osteogenic and adipogenic cells. Differentiation of BMMSCs into osteogenic cells was confirmed by Von Kossa staining after 4 weeks of culturing in osteogenic media, and differentiation into adipogenic cells was confirmed by Oil Red O staining after 2 weeks of culturing in adipogenic medium. Normal media served as control cultures for Von Kossa staining and Oil Red O staining. The photographs are 330 (A) or 165 (B) μm square. Shown are representative experiments from three experiments. Abbreviation: BMMSC, bone marrow mesenchymal stem cell.

Download figure to PowerPoint

thumbnail image

Figure Figure 2.. Cell motility induced by EGF and PDGF in rat (A) and immortalized human (B) BMMSCs. BMMSCs were exposed to diluent, EGF (10 or 100 nM), or PDGF (0.4 nM) in the absence (open bars) or presence (closed bars) of PD153035 to block EGFR signaling. An arbitrary unit corresponds to 89.1 ± 15.2 μm for rat BMMSCs and 643.7 ± 31.4 μm for human BMMSCs. The differences in motility were compared between growth factor and diluent exposed (*p < .05) and between treated cells based on PD153035 inhibition of response (#p < .05). The experiments were performed three times, each in duplicate; shown are mean ± SEM. Abbreviations: BMMSC, bone marrow mesenchymal stem cell; EGF, epidermal growth factor; PDGF, platelet-derived growth factor.

Download figure to PowerPoint

thumbnail image

Figure Figure 3.. Cell proliferation induced by EGF and PDGF in rat (A) and immortalized human (B) BMMSCs. The cell numbers of BMMSCs increased approximately fourfold in rat BMMSCs and approximately 16-fold in immortalized human BMMSCs in culture medium supplemented with 2% fetal bovine serum (FBS) in a 96-hour time period. The addition of EGF or PDGF gave an extra increase in cell counts in immortalized human BMMSCs (B) but not in rat BMMSCs (A). Only the immortalized human BMMSCs were stimulated in the absence (open bars) or presence (closed bars) of PD153035 to block EGFR signaling (B), as there was no noted growth factor–induced increase in the rat BMMSCs. Shown are mean ± SEM of three experiments, each performed in triplicate. The differences in proliferation were compared between growth factor and diluent exposed (*p < .01) and between treated cells based on PD153035 (#p < .01). There was no statistical difference between EGF and PDGF effects in either graph. Abbreviations: BMMSC, bone marrow mesenchymal stem cell; EGF, epidermal growth factor; PDGF, platelet-derived growth factor.

Download figure to PowerPoint

thumbnail image

Figure Figure 4.. Apoptosis in the presence of EGF and PDGF in rat (A) and immortalized human (B) bone marrow mesenchymal stem cells (BMMSCs). BMMSCs were challenged with culture media containing low serum, various agents, and APOPercentage dye, which was unidirectionally incorporated into apoptotic cells with phosphatidylserine flipping outer leaflet of lipid bilayer. EGF (10 or 100 nM) or PDGF (0.4 nM) was present throughout the challenge. Shown are mean ± SEM of three experiments, each performed in triplicate. *p < .05 compared with no growth factor in the presence of 0.5% DFBS. Abbreviations: Adv., advanced; DFBS, dialyzed fetal bovine serum; EGF, epidermal growth factor; FBS, fetal bovine serum; PDGF, platelet-derived growth factor.

Download figure to PowerPoint

thumbnail image

Figure Figure 5.. EGFR signal transduction in rat (A), immortalized human (B), and porcine (C) bone marrow mesenchymal stem cells (BMMSCs). BMMSCs were treated for 10 minutes with EGF, HB-EGF, or PDGF in the absence or present of the EGFR inhibitor PD153035. Phosphorylation status of key signaling intermediaries was determined by immunoblotting of EGFR (pY1173), AKT (pS473), PLC-γ (pY783), and ERK (pT202pY204). Shown are representative experiments from three experiments. In immortalized human BMMSCs, EGF or HB-EGF led to EGFR phosphorylation, which was suppressed by the EGFR kinase inhibitor PD153035, whereas EGF caused minimal EGFR phosphorylation in a PD153035-sensitive manner in porcine BMMSCs, and EGFR phosphorylation by EGF or HB-EGF was not apparent in rat BMMSCs. AKT and ERK demonstrated similar patterns in each of these lineages, with PDGF providing robust activations that were not suppressed by PD153035. Interestingly, PLC-γ was activated strongly by PDGF in each of these lineages, but EGF-induced PLC-γ phosphorylation was noted only in immortalized human BMMSCs. Abbreviations: EGFR, epidermal growth factor receptor; EGF, epidermal growth factor; HB-EGF, heparin-binding EGF-like growth factor; PDGF, platelet-derived growth factor; AKT, protein kinase B/akt; PLC-γ , phospholipase C-γ ; ERK, extracellular signal-regulated protein kinase; pAKT, phospho-AKT; pEGFR, phospho-EGFR; pERK, phospho-ERK; pPLC-γ , phospho-PLC-γ.

Download figure to PowerPoint

thumbnail image

Figure Figure 6.. Differentiation of rat (A) and immortalized human (B) BMMSCs into adipogenic and osteogenic lineages after EGF or PDGF treatment. Rat and immortalized human BMMSCs were treated with EGF (10 nM) (middle panel), PDGF (0.4 nM) (right panel), or diluent (left panel) for 5 days prior to exposure to adipogenic or osteogenic medium. Differentiation of BMMSCs into adipogenic cells was confirmed by Oil Red O staining after 10 days, and differentiation of BMMSCs into osteogenic cells was confirmed by Von Kossa staining after 4 weeks. Each photograph is 220 μm square. Shown are representative experiments from three experiments. Abbreviations: BMMSC, bone marrow mesenchymal stem cell; EGF, epidermal growth factor; PDGF, platelet-derived growth factor; EGFR, epidermal growth factor receptor.

Download figure to PowerPoint

thumbnail image

Figure Figure 7.. The effects of EGF on adipogenic (A) and osteogenic (B) differentiation of rat and immortalized human bone marrow mesenchymal stem cells (BMMSCs). Rat and immortalized human BMMSCs were cultured in adipogenic media (A) or osteogenic media (B) supplemented with EGF (10 nM) or PDGF. Differentiation of BMMSCs into adipogenic cells (A) was evaluated by immunoblotting with anti-adipocyte fatty acid-binding protein (ALBP) after 2 weeks, and differentiation into osteogenic cells (B) was evaluated by immunoblotting with anti-alkaline phosphatase (for rat and human BMMSCs) or anti-osteocalcin (for human BMMSCs) after 4 weeks. Shown are representative experiments from three experiments. Note that EGF alone did not induce adipogenic or osteogenic differentiation. The inset in (A) shows the same immunoblotting of rat BMMSCs in adipogenic medium with anti-ALBP to delineate each band clearly. Abbreviations: EGF, epidermal growth factor; PDGF, platelet-derived growth factor; Alk Phos, alkaline phosphatase.

Download figure to PowerPoint

Acknowledgements

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Disclosures
  9. Acknowledgements
  10. References

These studies were supported by grants from the National Institute of General Medical Sciences (A.W. and L.G.G.) and National Institute of Aging (H.C.B.). K.T. is a College of American Pathologist Foundation Scholar. We thank the members of the Wells laboratory for comments and suggestions, Dr. George Muschler (Cleveland Clinic, Cleveland) for graciously supplying primary human BMMSCs, Dr. Junya Toguchida (Kyoto University, Kyoto, Japan) for the immortalized human BMMSCs, Dr. Darwin J. Prockop (Tulane University, New Orleans) for rat BMMSCs, Dr. Joseph P. Vacanti (Massachusetts General Hospital, Boston) for porcine BMMSCs, Dr. Kazuro Lee Fujimoto (University of Pittsburgh, Pittsburgh) for experimental support, and Diane George and Dr. Mona Melhem (Pittsburgh VA Medical Center, Pittsburgh) for histochemical techniques and advice.

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Disclosures
  9. Acknowledgements
  10. References