• Open Access

Epidermal growth factor (EGF) transfection of human bone marrow stromal cells in bone tissue engineering


  • J. Wallmichrath,

    Corresponding author
    1. Plastic, Hand- and Microsurgery, University Hospital Grosshadern, Munich, Germany
    2. Department of Plastic and Hand Surgery, University Hospital, Freiburg, Germany
    Search for more papers by this author
  • G. B. Stark,

    1. Department of Plastic and Hand Surgery, University Hospital, Freiburg, Germany
    Search for more papers by this author
  • U. Kneser,

    1. Department of Plastic and Hand Surgery, University Hospital, Freiburg, Germany
    2. Department of Plastic and Hand Surgery, University of Erlangen Medical Center, Erlangen, Germany
    Search for more papers by this author
  • C. Andree,

    1. Department of Plastic and Hand Surgery, University Hospital, Freiburg, Germany
    2. Plastic and Aesthetic Surgery, Sana Kliniken Duesseldorf GmbH, Krankenhaus Gerresheim, Duesseldorf, Germany
    Search for more papers by this author
  • M. Voigt,

    1. Department of Plastic and Hand Surgery, University Hospital, Freiburg, Germany
    2. Plastic Aesthetic Surgery Freiburg, Freiburg, Germany
    Search for more papers by this author
  • R. E. Horch,

    1. Department of Plastic and Hand Surgery, University Hospital, Freiburg, Germany
    2. Department of Plastic and Hand Surgery, University of Erlangen Medical Center, Erlangen, Germany
    Search for more papers by this author
  • D. J. Schaefer

    1. Department of Plastic and Hand Surgery, University Hospital, Freiburg, Germany
    2. Department of Plastic, Reconstructive and Aesthetic Surgery, Clinic of Reconstructive Surgery, University Hospital, Basel, Switzerland
    Search for more papers by this author

  • Presented in parts at:
    2nd BioValley Tissue Engineering Symposium, November 25–27, 1999: Schaefer DJ, Wallmichrath J, Looden Z, Humar M, Kiefer T, Merschmann U, Andree C, Walgenbach KJ, Stark GB. Induction of Bone Constructs by Gene Therapy. Cells Tissues Organs, 2000; 166 / B502: 52, Stark GB, Cancedda R (Hrsg)

Correspondence to: Dr. med. Jens C. WALLMICHRATH, Plastische, Hand- und Mikrochirurgie, Klinikum Grosshadern, Marchioninistrasse 15, D-81377 Muenchen, Germany.
Tel.: +49-89-7095-3512
Fax: +49-89-7095-6505
E-mail: Jens.Wallmichrath@med.uni-muenchen.de


A novel therapeutic approach for the treatment of bone defects is gene therapy assisted bone tissue engineering using bone marrow stromal cells (hBMSC). The aim of this study was to investigate the influence of human epidermal growth factor (hEGF) on proliferation and alkaline phosphatase (AP) activity of primary hBMSC in vitro. hBMSC cultures were achieved by explantation culture of bone chips. Following exposure to 0–10 ng recombinant hEGF (rhEGF)/ml cell numbers were determined by automated cell counting and cell bound AP activity was measured spectrophotometrically. hBMSC were transfected with hEGF plasmids and the proliferative effect was studied by cocultivation of transfected and untreated cells using porous cell culture inserts. The persistence of hEGF expression even after cell transfer was studied by the generation of possibly osteogenic constructs introducing transfected hBMSC in fibrin glue and bovine cancellous bone. The maximum increase in proliferation (156 ± 7%) and AP activity (220 ± 34%) was detected after exposition to 10 ng rhEGF/ml. In the separation chamber assay transfected cells produced hEGF concentrations up to 3.6 ng/ml, which induced a mean proliferation increase of 93% which could be significantly inhibited by a neutralizing hEGF antibody. Further, EGF-secretion of transfected hBMSC in 3D-culture was verified. Recombinant and transgenic hEGF stimulate proliferation of primary hBMSC in vitro. Lipotransfection of hBMSC with hEGF plasmids allows the transient and site directed delivery of biologically active transgenic hEGF. The introduction of mitogenic, angiogenic and chemoattractive factors in gene therapy assisted bone tissue engineering is discussed by the example of EGF.


Major bone defects of various aetiology are challenging to treat. Classical reconstruction using autologous bone grafts or implanting synthetic substitutes involve problems such as morbidity of the donor area or material fatigue. These disadvantages could be avoided using tissue engineering [1, 2]. In accordance to the composite construction of natural bone tissue the tissue engineered bone usually consists of viable cells embedded in an avital scaffold and optional supplements. Human bone marrow stromal cells (hBMSC) with their stem cell plasticity are suitable as vital, bone forming components [3–5]. Introduced into bone constructs these cells are able to form vital bone, but significant tissue formation is strongly dependent on vascularisation [6]. Thus, factors enhancing cell proliferation and vascularisation are expected to be useful tools in yielding an osteogenic construct in vivo. Bone marrow stromal cells (BMSCs) are known to respond to the epidermal growth factor (EGF) [7]. The goal of our work was to evaluate the applicability of recombinant and transgenic human EGF (hEGF) for bone tissue engineering initially in the in vitro phase. Therefore, we studied the effects of recombinant hEGF (rhEGF) on proliferation and activity of the osteogenic marker enzyme alkaline phosphatase (AP) of hBMSC in vitro. Another way of administration and a smart tool for delivery of hEGF to the site of interest is gene transfer. Following non-viral gene transfer by lipotransfection of hBMSC with hEGF plasmids we studied the proliferative effect of the transgenic hEGF on primary native hBMSC in a separation chamber assay. Finally, we composed transgenic constructs with viable cells in a scaffold of fibrin and cancellous bone and detected a significant hEGF expression in vitro even after transfer of the lipotransfected cells into the 3D scaffolds.

Materials and methods

Culture medium

The culture medium was MEM Alpha-Medium (GIBCO BRL, Paisley, UK) and BGJB-Medium (GIBCO BRL), both supplemented with 10% foetal calf serum (GIBCO BRL), 100 U/ml penicillin G (Sigma, St. Louis, MO, USA) and 100 μg/ml streptomycinsulfate (Sigma).

Bone marrow stromal cells

Primary cultures of hBMSC were obtained from remaining pelvic bone fragments with informed consent of the patients undergoing bone transplantation. The patients included males (n= 6) and females (n= 3) ranging in age from 29 to 51 years (m= 40 years). The processing mainly followed the protocol of Haynesworth and Caplan [3]. Bone blocks were reduced mechanically to small pieces, vortexed in culture medium and centrifuged for 8 min. at 250 ×g at 4°C. The pellet was resuspended, transferred into culture flasks (Falcon, Franklin Lakes, NJ, USA) and incubated in humidified atmosphere at 37°C and 5% CO2. The first medium change was performed after 5 days, subsequent medium changes every third day. Cell passaging was carried out as usual with trypsin solution (Viralex Trypsin/EDTA-1x-Solution, PAA Laboratories, Linz, A). Cells were replated in six-well plates or six-well plate-cell culture inserts with a pore size of 1 μm (BectonDickinson, Franklin Lakes, NJ, USA).

Recombinant human epidermal growth factor (rhEGF) and hEGF antibody

Lyophilized recombinant hEGF expressed in Escherichia coli (R&D Systems, Wiesbaden, Germany), was reconstituted in sterile 10 mM acetic acid containing 0.1% bovine serum albumin. Lyophilized neutralizing anti-hEGF (R&D Systems) was reconstituted in sterile phosphate buffered saline (PBS), pH 7.4.

pCMV-hEGF plasmid

A distinct description of the plasmid has previously been published [8]. The engaged vector PWRG 1630 for mature hEGF (amino acid 949–1001 of the presursor molecule) is a 4286 bp low copy plasmid. Its transcription is promoted by a CMV promoter and it contains ampicillin resistance and a bacterial promoter. The plasmid stock solution was stored at −20°C and contained 1 μg DNA/μl.


Subconfluent primary hBMSC cultures in the log phase were used for gene transfer. The general procedure was as follows: 24 hrs prior to transfection the cells were replated at 4 × 104 cells per cell culture insert. We used Escort™ Transfection Reagent (Sigma), which is a ready to use liposome formulation containing the cationic liposomes DOTAP and DOPE in a total lipid concentration of 2 μg/μl. We chose a ratio (w/v) of DNA:liposome of 1 μg:2.5 μl. The amount of plasmid-DNA was 6 μg per cell culture insert. The formation of the lipid-DNA-complex was performed in a volume of 230 μl DMEM (GIBCO BRL). The mixture was incubated for 15 min. at room temperature before addition of 2 ml of complete medium and adding to the cells. Furthermore, 1 ml of complete medium was added after 1 hr. After incubation for 16 hrs the cells were rinsed thoroughly with complete medium.

Lipotransfection of corresponding primary hBMSC monolayer cultures for the assessment of transfection efficiency was carried out using β-Gal-plasmids (6.5 kb-Vector pZeoSV-LacZ, Invitrogen, Carlsbad, CA, USA). The procedure was as follows: Lipotransfection (c.f. transfection protocol above), rinsing with PBS, fixation (2% formaldehyde, 0.2% glutaraldehyde in PBS, pH 7.6–7.8) for 5 min. at room temperature, rinsing with PBS, addition of substrate solution (1 mg/ml X-Gal-substrate, Sigma-Aldrich, Deisenhofen, Germany) and incubation at 37°C for 6 hrs. The transfection efficiency was determined visually.


Aliquots of the culture supernatants were stored at −20°C, diluted with culture medium and hEGF levels were determined quantitatively in a solid-phase ELISA (Quantikine hEGF Immunoassay, R&D Systems).

Determination of alkaline phosphatase activity and cytochemical AP staining

The measurement of the cell-surface AP activity followed descriptions of Cassiede and Caplan [9] using a substrate buffer (5 mM Sigma 104 Phophatase Substrate, Sigma; 50 mM glycin, 1 mM MgCl2, pH 10.5). The reaction depends on the cleavage of the chromogen p-nitrophenylphosphate (pNPP) to 4-nitrophenol (pNP). The cytochemical staining followed the descriptions of the manufacturer (Sigma ALP-kit, Sigma-Aldrich).

Cell counting

The cell counting was performed with the Casy TT Culture Counter™ (Schaerfe System, Reutlingen, Germany). The cells were resuspended in a weak electrolyte solution (CASYton™, Schaerfe System). The standard setting was: 150 μm capillary, single counting volume 400 μl, five cycles of measurement, particle size 10.5–30 μm, dilution 40-fold. Periodical parallel countings were performed with a conventional haemacytometer and trypan blue (Trypan Blue Solution, Sigma).

Preparation of 3D-constructs

A processed bovine cancellous bone matrix Tutobone™ (Tutogen, Erlangen, Germany) in cylindrical shape (diameter 5 mm, height 2 mm) served as a scaffold for lipotransfected cells in vitro. Fibrin glue was used for immobilization of osteogenic cells. It was prepared for each assay by mixing a serum-free suspension of 5 × 104 cells 1:5 (v/v) with fibrinogen solution (8 mg/ml human fibrinogen in 0.9% sodium chloride, Sigma). Thrombin solution (1.25 IE/ml thrombin in 40 mM CaCl2, Sigma) was mixed 1:1 (v/v) with ε-aminocaprionic acid (0.1 M, Sigma). This solution was gently mixed 1:6 (v/v) with the cell-fibrin-suspension to yield the final product.

Experimental design

(1) Examination of the effect of rhEGF on primary hBMSC cultures

First passage BMSC cultures of three donors (m, 29; m, 43; f 36) were each replated in 56 35-mm dishes at 4 × 104 cells/dish and incubated for 24 hrs. Medium changes were carried out every 48 hrs with medium containing 0, 0.1, 0.5, 1, 2, 5 or 10 ng rhEGF/ml, a group of eight wells for each concentration and donor. One donor culture (m, 29) yielded only 6 assays (Table 1). After 7 days the assays were split for the determination of cell number and the measurement of AP activity, respectively.

Table 1.  Stimulation of BMSC with rhEGF
 Number of subcultures exposed to 0, 0.1, 0.5, 1, 2, 5 or 10 ng/ml rhEGF over 7 days
Proliferation testAP measurement
  1. Quadruplicate assays (donors f, 36 and m, 43) or triplicate assays (donor m, 29) were used per group of rhEGF-concentration in the culture medium for the determination of proliferation and AP-activity, respectively.

Series 1 (donor f, 36) n= 4 n= 4
Series 2 (donor m, 43) n= 4 n= 4
Series 3 (donor m, 29) n= 3 n= 3

(2) Examination of the effect of transgenic hEGF on hBMSC (separation chamber)

EGF-expressing transfected hBMSC and untreated hBMSC were cocultured using cell culture inserts with a porous PE-membrane allowing diffusion of the transgenic EGF. Cells of three donors were each replated in 35-mm dishes and matching culture inserts (Table 2). Twenty-four hours later the cells in the inserts were either treated only with a medium change (group A), treated only with liposome solution in conformity with the transfection protocol (group B) or were lipotransfected according to the protocol described above (groups C and D). The subcultures in the wells remained untreated. Twenty-four hours later the inserts were transferred into the wells containing untreated cells and a medium change was carried out with complete medium (day 0). No further medium changes were performed. In group D, a neutralizing hEGF antibody was administered to the medium at a dosage of 2 μg/ml*d (series 1) and 12 μg/ml*d (series 2 and 3), respectively. Cell counting was performed at observation of first subconfluence which was in the three series on days 3, 6 and 5, respectively.

Table 2.  Stimulation of BMSC with transgenic EGF
 Group AGroup BGroup CGroup D
(exposure to untreated cells)(exposure to cells treated with liposome)(exposure to EGF-transfected cells)(exposure to EGF-transfected cells plus hEGF antibody)
  1. In the separation chamber experiment, untreated cells in the bottom of the wells were cocultured with cells in culture inserts which were either untreated (Group A), only treated with liposome (Group B), hEGF-transfected (Group C) or hEGF-transfected plus external administration of a neutralizing hEGF-antibody (Group D).

Series 1 (donor f, 39, P1) n= 6 n= 0 n= 6 n= 6
Series 2 (donor m, 36, P2) n= 6 n= 10 n= 10 n= 4
Series 3 (donor f, 51, P1) n= 7 n= 6 n= 7 n= 3

(3) Examination of EGF-expression of transfected hBMSC in 3D-cultures

The expression of transgenic hEGF by lipotransfected hBMSC was measured after seeding of the cells into scaffolds (Table 3). A second passage hBMSC culture (donor: m, 51) in the log phase was lipotransfected and 24 hrs later the cells were aliquoted at 5 × 104 cells in 100 μl medium. The cells were either put in 35-mm dishes for monolayer culture (group I), flushed into Tutobone™ by a micropipette (group II), embedded in a fibrin matrix (group III) or injected as a fibrin-cell-suspension into Tutobone™ blocks (group IV). After a sedimentation period of 30 min., 3 ml of complete medium was added to each assay. Subsequent medium changes were performed every 48 hrs and hEGF levels were measured in the culture supernatants.

Table 3.  Experiment for the analysis of transgenic hEGF-expression by lipotransfected BMSC distributed into four culture types (Group I–IV)
hEGF-transfected hBMSC subcultured in different culture types:
Monolayer culture (group I) n= 3
Tutobone™ (group II) n= 3
Fibrin glue (group III) n= 3
Fibrin glue + Tutobone™ (group IV) n= 3

Statistical analysis

All statistical analyses were performed with the software SAS for statistical analysis (SAS Institute Inc., Cary, NC, USA) with kind support by Professor Dr. J. Schulte-Moenting, Institute for Medical Biometry, University of Freiburg, Germany. An anova was carried out over all groups to assess a linear function of cell number and AP activity from rhEGF levels, and secondly from the logarithm of rhEGF levels. For the data of the separation chamber experiments the Kruskal–Wallis test was applied to all series separately followed by a conclusion analysis. Values of untreated cells were considered as a reference point of 100% and mean and standard deviation were calculated. The level of significance was expressed as ‘*’ for P≤ 0.05 and ‘**’ for P≤ 0.005, respectively.


Cell culture

First hBMSC colonies were observed around day five of culture and cells primarily exhibited their typical fibroblast-like spindle shaped morphology. No adipocytic, round, cuboid or multinuclear cells were observed. The cell colonies incubated with and without EGF exhibited high cytochemical AP activity, particularly cells forming multilayers or nodules: Fig. 1 gives an macroscopic view of primary hBMSC subcultures (Donor m, 36; first passage) stained for AP activity after 16 days of pre-treatment without (Fig. 1A) and with rhEGF (5 ng/ml; Fig. 1B).

Figure 1.

Macroscopic aspect of primary hBMSC cultures grown in basic culture medium (A) or in medium containing 5 ng/ml rhEGF (B) after cytochemical AP-staining. The cultures show a similar appearance with a swirling pattern in monolayer regions and a more intense AP activity (purple) especially in multilayer areas and nodules (dark spots). The cells at the outer rim are partially detached during the fixation process.

Stimulation of hBMSC with rhEGF in vitro

Primary hBMSC subcultures were treated with recombinant EGF-media as shown in Table 1. The resulting cell numbers detected on day seven are demonstrated in Fig. 2. The cells of all three donors responded to increasing administration of rhEGF with increasing cell numbers finding the maximum between 148% and 161% (m= 156%) after exposure to 10 ng rhEGF/ml. Even stimulation with 1 ng rhEGF/ml lead to cell numbers between 140% and 154% (m= 145%) compared with the cells incubated without rhEGF. In the investigated range of rhEGF concentration statistical analysis showed significant dependence of the cell number from the logarithm of rhEGF concentrations (P= 0.012). The analysis of linear dependence was not significant (P= 0.2845).

Figure 2.

Cell numbers increased in the presence of rhEGF (0–10 ng/ml) in the culture medium.

The detected AP activity of the cells in dependence on rhEGF administration is demonstrated in Fig. 3. Cells of all donors responded to increasing rhEGF levels with an increasing AP activity. The maximum substrate cleavage was detected in the group with 10 ng rhEGF/ml with values between 200% and 260% (m= 220%). Even 1 ng rhEGF/ml lead to an increase between 161% and 208% (m= 181%) compared to the controls. The statistical anova over all rhEGF concentrations showed slight linear dependence of AP activity on rhEGF concentrations (P= 0.0485). The dependence of enzyme activity on the logarithm of rhEGF concentrations was highly significant (P= 0.0001). Although an increased AP level does not necessarily mean the cells will grow into bone tissue in vivo, it is an indication of the persistence of the BMSC phenotype under hEGF stimulation.

Figure 3.

Increase in relative AP-activity of hBMSC cultures of three donors following a 7-day exposure to rhEGF concentrations between 0 and 10 ng/ml.

Separation chamber assay

BMSC grown in culture inserts were treated as mentioned above and put in coculture with untreated cells (Table 2: groups A–D). The resulting numbers of the cells in the bottom of the wells are expressed in Fig. 4. The numbers of the cells in group C (transgenic hEGF stimulation) ranged between 156% and 232% (m= 194) which was in all series significantly higher than in the untreated controls (group A, P= 0.005). The addition of the neutralizing hEGF antibody (group D) reduced the cell numbers significantly by 39 ± 9% (P= 0.05). Significant EGF-levels were detected in the supernatants of group C (m= 1.9 ng/ml) at the end of the experiment. The transfection efficiency of the transfer procedure described above was tested using β-Gal-plasmids and was determined visually as <5% of the subconfluent adherent cells.

Figure 4.

Cell numbers of hBMSC subcultures after cocultivation with untreated cells (Group A), liposome-treated cells (Group B, omitted in series 1), hEGF-transfected cells (Group C) and hEGF-transfected cells plus administration of a neutralizing hEGF antibody (Group D), respectively. The EGF-induced increase in cell numbers (Group C) can be attenuated by addition of anti-hEGF (Group D).

Genetically modified hEGF-expressing constructs

Genetically modified constructs were generated involving hEGF-lipotransfected hBMSC and EGF-levels were measured on days 2, 4 and 6 (Table 3). All cultures expressed significant EGF levels (Fig. 5). In all groups maximum levels were detected on day 2. In the medium of cells grown without biomaterial (group I) the level was 5.3 ± 0.4 ng/ml (100%). Cells grown on Tutobone™ (group II) reached 27% of this value, cells embedded in fibrin (group III) 60% and cells immobilized in fibrin, which was additionally injected into Tutobone™ (group IV), reached 41%.

Figure 5.

The expression of hEGF by lipotransfected hBMSC seeded in different scaffolds: 5 × 104 transfected cells were either subcultured in a 35 mm dish (Group I), injected into Tutobone™ (Group II), embedded in fibrin (Group III) or immobilized in fibrin and injected into Tutobone™ (Group IV). The concentration of hEGF was measured in the culture supernatant which was changed every 48 hrs.


hBMSC as cell source for bone tissue engineering

The osteogenic potential of BMSC makes them widely used vital components of artificial bone constructs. In addition their stem cell plasticity makes them interesting tools in the development of different tissues and compound constructs [3, 4, 6, 10]. BMSC have several advantages with regard to regenerative medicine: (i) cell isolation is relatively easy; (ii) the cells adhere to the surface of tissue culture vials and constructs; (iii) they have a high proliferative potential; (iv) their expansion and subculturing is easy in monolayer cultures using standard media; (v) they allow multiple passages with no loss of osteochondral potential [11]; (vi) their default pathway is osteogenic differentiation [12] and (vii) as progenitors of secretory cells they are attractive vehicles in gene transfer applications [13]. Besides bone regeneration BMSC are also implicated in making a substantial contribution to postnatal vasculogenesis, which is important for engineering vascularised bone tissue substitutes [14].

Despite the inherent problems of primary cells with interindividual and interexperimental variability we used them for our experiments with the aim of future clinical applications. Autologous sourcing is most desirable as cells are collected from the affected patient, thereby eliminating the problems of immune rejection. The cells for our studies were gained from waste bone fragments of patients undergoing surgery. The less invasive way of harvesting hBMSC from bone marrow aspirates is already routine [15, 16]. Mesenchymal stem cells can be derived from the peripheral blood but they have the disadvantage of a difficult isolation process and lower proliferation rates compared with BMSC [17]. The isolation of pluripotent stem cells from human embryonal tissue is possible but ethically controversial [18, 19]. A potential new starting point is the extraction of adult stem cells from different non-osseous tissues (e.g. muscle) which are held responsible for ectopic bone formation [11, 20].


EGF is a potent mitogen for many cell types, primarily epithelial and mesenchymal cells. Experiments using a microchemotaxis chamber already showed that EGF induced the migration of hBMSC via its chemotactic effect and also enhanced their proliferation [21, 22]. It was demonstrated by Fan et al. in 2007 that EGF covalently surface tethered to a biomaterial promotes both cell spreading and survival of hMSC in vitro[7]. Also, the migratory response of (murine) BMSC is known to be enhanced by EGF-receptor-transfection [13]. The proliferative effect of rhEGF on BMSC is basically already known [7, 22]. However, the examination of its proliferative effect on our particularly cultured primary hBMSC is required to determine the level of biological effects of rhEGF as a basis. Our experiments with rhEGF in multiple concentration steps revealed that transfected primary cultures should be able to produce transgenic hEGF levels high enough to achieve a significant biological (proliferative) effect. Additionally, we did not find an antiproliferative effect in the concentration range of interest (Fig. 2). Recent studies revealed that human bone morphogenetic protein-6 (BMP-6) exerts its osteoinductive effect on hBMSC (i.e. acceleration of cell differentiation and formation of mineralised nodules; dramatical increase in AP activity), at least in part, through the EGF signalling pathways [23]. The EGF-R-family consists of four transmembrane receptors belonging to the receptor tyrosine kinase superfamily and includes EGF-R (also known as ErbB1(HER-1), ErbB2(Neu(HER-2, ErbB3(HER-3 and ErbB4(HER-4. The ligands include among others EGF and transforming growth factor-α (TGF-α) [24]. It is reported that the EGF-R seems to negatively affect osteoblastic differentiation, which inversely correlates with the proliferation capacity [25]. This reflects the phenotype stabilizing feature of growth factors keeping cells at less differentiated stages with a higher potential for dividing and thus for regeneration.

A problem that is yet unsolved and which limits the size of artificial bone constructs is lack of vascularisation. Adequate blood supply is required for the development of significant amounts of bone tissue in vivo, which has been proven in diffusion chamber experiments [26]. Thus, the two most promising approaches of engineering large bone constructs momentarily are (1) generating an ectopic (extraskeletal) vascularized ‘bone flap’in vivo suitable for transplantation or (2) the generation of an avascular construct in vitro containing undifferentiated dividing osteogenic cells which is implantated and then primarily nourished mainly by diffusion. Vascularization of such constructs can be surgically optimized by integration of surrounding vessels or vascularized tissues [10]. Including hEGF in our constructs, a growth factor with proliferative and angiogenic effects, we aim at an EGF-induced proliferative and angiogenic effect of our constructs when applied in vivo[21, 27, 28]. For all proliferation experiments we used low passages of BMSC cultures which are known to have a higher density of EGF-receptors [27, 29]. In our experiments even the stimulation with 1 ng rhEGF(ml lead to cell numbers of 145% compared with the controls (Fig. 2). The AP activity showed relative higher values compared with cell proliferation (Fig. 3). This was thought to be the result of increased differentiation of cells with increasing cell confluency. EGF has been used by others to direct the differentiation of hBMSC cultures to the epithelial lineage (applying 20–30 ng EGF(ml) [30]. Nevertheless, in our experiments all cultures stained positive for AP activity following exposure to recombinant or transgenic hEGF and also after hEGF transfection which demonstrates the continuance of the osteogenic phenotype. This might be due to the strong osteogenic stimulus of BMSC cultures reaching confluence, the lower applied EGF concentrations, the shorter time period of exposure and (or the missing remaining substances of the growth factor cocktail commonly used for epithelial differentiation [30]. The final consequence of applying hEGF in engineered bone constructs has to be examined in vivo, especially concerning complex processes as bone formation and vascularisation.

Genetic modification of hBMSC by non-viral gene transfer

Viable cells in artificial bone constructs can be influenced by exogenous addition of osteogenic supplements or by using gene therapy. The advantage of gene therapy is the bioexpression of posttranslational eucaryotic modified proteins at the site of interest, e.g. a bony defect. Additionally, BMSC-based gene transfer directly increases the number of osteogenic cells in the defect region. The two principal techniques are viral and non-viral gene transfer. The advantages of the latter are transient bioexpression, easy handling, safety and low immunogenicity [19, 31, 32]. Ex vixo transfection of autologous cells in our opinion has advantages for future clinical applications with regard to non-immunogenicity and efficiency. In contrast to gene therapy of enzyme deficiency disorders such as Duchenne muscular dystrophy our focus is on local and transient expression for the days or weeks of construct integration. The generation of stable transfection is technically complex and may have unforeseen gene-regulatory effects due to integration of the transgene into the host-DNA [33]. Thus, as long as chromosomal integration cannot be controlled our aim is episomal localisation of the transgene. It warrants further examination to establish a more sustained, efficient, safe, reproducible, maybe targeted and injectable in vivo vector systems such as virus-like particles [34]. A new ex vivo alternative is the Nucleofector™ technology which is a non-viral transfection method especially designed for primary cells and hard-to-transfect cell lines. It has successfully been applied to BMSC [35]. Especially osteogenic cells appear to possess ideal preconditions for gene therapy because of their high secretory capability [15]. In our experiments with lipotransfected hBMSC cultures we found the typical peak of transgene expression on days 2–3 with EGF levels in the supernatant in the ng/ml level followed by a time course dependent decrease with finally negligible levels detectable up to day 22. In our experiments the transfected cells produced transgenic hEGF levels comparable to the rhEGF levels which showed biologically effects on hBMSC in our primary studies (Figs 2 and 3). The efficiency of hEGF lipotransfection in our experiments can be estimated by the ratio of the cumulative hEGF concentration in the culture medium (4 ml) at the end of the separation chamber experiment (m= 1.9 ng EGF/ml) to the transfected cell number (4 × 104 cells). The transfection efficiency of the procedure was tested in parallel experiments using β-Gal-plasmids. β-Gal+cells appearing blue represented <5% of the subconfluent hBMSC. However, this immunohistochemical reporter gene assay is known for its underestimation of the transfection efficiency of lipid-mediated gene transfer [36].

Transgenic hEGF has a proliferative effect on primary hBMSC

The scientific proof of the proliferative potency of transgenic hEGF on hBMSC in vitro is an important step in the development of a genetically modified construct. In the separation chamber assay we detected 1.7- to 2.3-fold higher proliferation rates of native hBMSC under transgenic hEGF stimulation which was apparently due to the expression of soluble transgenic hEGF. The detected transgenic EGF-concentrations reached the levels of rhEGF which have been shown to be biologically active on hBMSC (Figs 2 and 3) and the addition of a neutralizing hEGF antibody specific to EGF (Fig. 4, group D) reduced the stimulatory effect significantly (P≤ 0.01). However, neutralization was only partial presumably due to a shortage of antibody concentration or incomplete inhibition of the biological activity of transgenic hEGF. An almost complete neutralization of hEGF activity may be reached by addition of the antibody in a more frequent pulse and/or in a higher concentration or by adding an antibody with receptor-blocking quality.

The shown proliferative effect of transgenic hEGF on hBMSC could be helpful in bone tissue engineering in the in vivo phase as well as the known angiogenic effect of hEGF. A possible potentiation of the EGF effects could be received in co-action with other growth and osteogenic factors such as BMPs, basic fibroblast growth factor or TGF and warrants substantial further efforts [36–38].

Genetically modified 3-D bone constructs expressing hEGF

Biodegradable scaffolds combined with cells or biological molecules have already reached clinical applicability and certain minimum requirements concerning biological, chemical, and physical properties have to be met [2]. Osteogenic constructs containing precursor cells with high proliferative capacity and differentiation after implantation into a bone defect have the advantage of requiring a shorter time period for fabrication. We used hEGF-transfected hBMSC as cellular components for our constructs aiming at a stimulation of proliferation and potentially vascularisation in vivo (Fig. 5). The detected concentration of transgenic hEGF in the supernatant of the constructs depended on the type of biomaterial and perhaps the embedding procedure of the cells. EGF levels were maximally in the group with cells grown in monolayer, which possibly was the consequence of minimal manipulation and unimpaired diffusion (group I). The cumulative EGF-concentration measured in the culture supernatant after 48 hrs was higher than 5 ng(ml and even the concentration during the second period of 48 hrs (i.e. days 5+6) yielded EGF-levels of more than 2 ng/ml. These significant levels represent biologically active concentrations as demonstrated for rhEGF in Figs 2 and 3. The conclusion is that trypsinization and replating of the transfected cells does not seem to have a detrimental effect on the subsequent transgene expression. The concentration of transgenic EGF in the culture medium of groups II, III and IV showed comparatively lower values. The cells grown in Tutobone™ (group II) showed minimal expression levels, possibly due to the detrimental incubation period in a minimal medium volume to achieve sedimentation and cell adhesion. The reduced EGF concentration we measured in groups II–IV can be due to a diminished transgene expression but is rather a consequence of a retention due to a reduced release of EGF because of restricted diffusion and/or adsorption of the growth factor to the scaffold material. This binding or adsorption to fibrin, for instance, can be used as a drug release system. Our data show, that transgenic EGF is expressed by lipotransfected hBMSC even after transfer of the cells into the 3D scaffolds in vitro. The in vitro data cannot be simply transferred into the in vivo situation, but the constructs containing viable hEGF-expressing primary hBMSC could be a useful supplement in bone tissue engineering. New, cell saving methods for cellular embedding have to be developed and solid free-form fabrication techniques for scaffold-based tissue engineering require active investigations [1]. Also, the application of other osteogenic factors such as bone morphogenetic and osteogenic proteins either expressed transgenically or surface tethered to a biomaterial warrants further research [9, 39].


The presented data demonstrate the in vitro analysis and the development of genetically modified, hEGF-expressing viable constructs by combination of gene therapy and tissue engineering. Primary hBMSC monolayer cultures were exposed to rhEGF in vitro which stimulated proliferation and also the activity of the osteogenic marker enzyme AP of the cells. Following lipotransfection of hBMSC with hEGF plasmids the proliferative effect of the transgenic hEGF on hBMSC was studied in a separation chamber assay. To our best knowledge this is the first report of successful lipotransfection of primary hBMSC with hEGF plasmids with the demonstration of the proliferative effect of transgenic hEGF on hBMSC. Further, genetically modified and possibly osteogenic constructs were generated and the persisting transgene expression by lipotransfected cells was shown, which were embedded in fibrin or Tutobone™. Although the proliferative effect of recombinant and transgenic hEGF on hBMSC seems to be comparatively low, the known angiogenic effect of EGF as well as its chemoattractive effect on local hBMSC could be useful in vivo, particularly in co-action with other osteogenic factors. The full consequence for bone tissue engineering can only be tested in vivo, which will be the object of further studies.


The authors thank M. Humar for the valuable technical support and expert assistance and Professor Dr. J. Schulte-Moenting for the processing of the data and the statistical analysis. The project was funded over a 2-year period by the Center of Clinical Research I (Deutsche Forschungsgemeinschaft), University Hospital Freiburg, Germany.