Printer Friendly

Factors Affecting Collagen Gel Contraction During Osteogenic Differentiation of Mesenchymal Stem Cells.


Collagen is the most abundant protein of the human body that forms the basis of the extracellular matrix (ECM) in most tissues and is a key component of healing wounds [1], [2].

Type I collagen, comprising 70-90% of the bone ECM, in combination with the deposited calcium phosphate crystals impart the characteristic hardness to bone while allowing it to retain some degree of flexibility thereby defining the mechanical properties of the bone [3]. The regulatory role of the ECM in cellular function has been receiving increasing attention in the recent years [1], [2]. Signal transduction following interaction of cell surface receptors with ECM, transmission of mechanical forces, regulation of diffusion of growth factors, and influence on cell adhesion, migration and invasion are some of the functions that are orchestrated by the macromolecular components of the ECM. Osteoblasts can bind to collagen type I using two types of receptors, integrins (aiai and a2a1) and discoidin domain receptors (DDR2) [4] [5]. ReceptorECM interaction can transduce mechanical stimuli and initiate outside-in-signaling cascades that are important for osteogenic differentiation (mechanotransduction). These findings highlight and support the central role of collagen type I in the tissue functions of bone.

Comparing various hydrogels as in vitro models for osteogenesis, it was found that collagen type I gel is superior to fibrin glue, alginate or pluronic F127 gels [6]. Seeding efficiency of hMSC is higher and hMSC proliferate faster on collagen scaffolds than on any other, including those that are calcium phosphate based or are formed using demineralized bone matrix. [7] [8] hMSC cultured on collagen scaffolds expressed higher levels of differentiation markers than hMSC differentiating on calcium phosphate ceramics. [8]

Cross-linking collagen type I and/or altering its native conformation to create a biomaterial with more favorable mechanical properties is an approach proposed and adopted by increasing number of tissue engineering groups. However, this is implemented at the expense of decline in biocompatibility since the cross-linking renders collagen susceptible to pathologic calcification and degeneration [9]. Degradation of collagen by matrix metalloproteinases (collagenases) is important not only for remodeling of ECM and cell migration but also for osteogenic differentiation of hMSC [5] [10]. The cleavage of collagen produces peptides that themselves serve as signaling molecules and release ECM bound latent growth factors like TGF-b. Cells do not migrate into and remodel modified non-native collagen matrix as effectively as they interact with native fibrillar collagen [11]. The inability of cells to degrade collagen that contains inert cross-links or has been locked into non-native conformation, interferes with the process of osteogenic differentiation of hMSC [11]. hMSC on native collagen tend to differentiate along osteogenic lineages while those on denatured collagen substrata differentiate along adipogenic lineage [12]. The above observations strongly emphasize the need for the use of native collagen type I in developing tissue engineered cartilage/bone grafts.

In addition to the native status of collagen, contraction of the collagen gel by the resident cells also needs to be considered when engineering tissue grafts or 3-D models based on collagen type I. The contraction process reduces the overall size of the construct by expulsion of water from the gel and increases in the matrix density [13]. In vitro. the reduced porosity of the contracted matrix (gel) may restrict the delivery of nutrients to the components cells and significantly reduce the number of viable cells. In vivo, ECM contraction assists wound closure but when it is excessive it is responsible for intractable scar formation [14]. Because of the poor cell supporting environment dense scars are impermeable to cell invasion and cannot vascularise and remodel to more viable tissue. Previous research in our laboratory led to the development of a non-contracting connective tissue equivalents generated using native collagen type I gel [14]. When fibroblasts are cultured in this 3-D construct under appropriate culture conditions, they remain quiescent and do not contract the collagen gel. In the present study we examined and report on various factors that affect native collagen type I gel contraction that may occur during the process of osteogenic differentiation of hMSC.

Materials and Methods

Collagen Type I Gel

Cold (4[degrees]C) porcine collagen type I solution (3mg/ml, Cellmatrix[R], Wako Chemicals, Richmond, VA,) (8 parts v/v) was mixed thoroughly with cold (4[degrees]C) solution of MEM-a (10X, non-buffered, serum free) (1 part, v/v) and then neutralized (pH 7.4) with cold (4[degrees]C) reconstitution buffer (sodium hydroxide 0.5N, sodium bicarbonate 22 g/L, HEPES free acid 47.7 g/L) (1 part v/v). After each addition, careful and thorough mixing was required to prevent air bubble formation in the viscous solution. All mixing steps were conducted under sterile conditions at 4[degrees]C (on ice). When incubated at 37[degrees]C and at 5% C[O.sub.2] the solution gelled within 30 min. All experiments were performed using this collagen type I gel forming process.


Human bone marrow derived mesenchymal stem cells (hMSC) (Lonza, Switzerland) from a male donor were cultured in MEMa (GIBCO, Invitrogen, Carlsbad, CA) containing fetal bovine serum (FBS, 10%, Atlanta Biologicals, Lawrenceville, GA) and FGF (10 nM, R&D Systems, Minneapolis, MN), with medium changes every second day. When 80-90% confluent hMSC were sub-cultured by harvesting with trypsin (0.05%) with 0.53 mM EDTA in HBSS (GIBCO, Invitrogen, Carlsbad, CA), and neutralization of the enzymatic activity with trypsin inhibitor (Soy protein, GIBCO, Invitrogen, Carlsbad, CA). Cells were counted using hemocytometer and Trypan blue (Sigma-Aldrich, St. Loius, MO) dye exclusion viability stain and then plated at 3000 cells/[cm.sup.2]. Cells were used in either passage 4 (early passage is passage less than 5) or passage 10 (late passage is more than 10).

Osteogenic differentiation to produce human osteoblasts (hOST) was conducted by culturing hMSC in osteogenic differentiation medium (ODM). Medium was prepared using MEMa (GIBCO, Invitrogen, Carlsbad, CA) containing FBS (10%, Atlanta Biologicals), ascorbate-2-phosphate (200 mM, Sigma-Aldrich, St. Loius, MO), dexamethasone (10nM, Sigma-Aldrich), and b-glycerol phosphate (10 mM, Sigma-Aldrich). Medium was changed every second day. When nearly confluent cells were harvested and counted as described above for hMSC.

Telomerase transduced human mesenchymal stem cells (TMSC) (a kind gift from Dr. Dario Campana, St. Jude Children's Research Hospital, Memphis, TN) (15) were cultured as described for hMSC. Osteogenic ally differentiated TMSC (TOST) were obtained by culturing TMSC in ODM for 10 weeks as described above for hMSC.

Collagen Gel Contraction

Collagen solution was prepared as described above. Appropriate number of cells, (150,000 to 2,400,000) were pelleted (15 ml centrifuge tube) and the pellet suspended in cold (4[degrees]C) neutralized collagen type I solution (3 mL). Cells suspension was dispersed by gently passing through a syringe (5 mL) and needle (21 G, 1.5 inches). The resulting cold (4[degrees]C) single cell suspension (250 mL) was dispensed into each well of a 48-well plate. After gelling (incubation at 37o C and 5% CO2 for 30 min) the appropriate medium (MEMa or ODM) was carefully added to each well (1 ml/well). At specific time points (day 1, 7, 14 and 28) the medium was carefully aspirated from the wells, the gels were transferred from the wells to plastic weighing boats and weighed (analytical balance, OHAUS GA110 Pine Brook, NJ). The fluid released from the gel during this handling was also weighed.

MTT Assay

TMSC and TOST were plated in 12-well plates at densities of 2500, 5,000, 10,000, 20.000, 40,000, 80,000, 100,000 cells per well, using biological triplicates for each seeding density value. After allowing the cells to attach, medium was removed, cells washed with phosphate buffered saline (0.256g/L Na[H.sub.2]P[O.sub.4] [H.sub.2]O, 1.19g/L [Na.sub.2]HP[O.sub.4], 8.76g/L NaCl, pH 7.4, in distilled water) (PBS) and incubated with MTT [500 mL per well of a 0.2 mg per ml solution of 3-(4,5-Dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide, Sigma-Aldrich in PBS] for two hours at 37[degrees]C. MTT solution was aspirated and the formazan formed in the cells recovered using formazan solubilization solution (250 ml, Sigma-Aldrich), with rocking for 90 minutes. Aliquots of dissolved formazan (100 mL) were transferred to wells of a 96-well plate, as two statistical replicates for each cell density. The 96-well plate was scanned at 570 nm using a plate reader (Molecular Devices Spectramax 340 PC, MDS Analyticals, Toronto, Canada) and the results recorded.

Collagen gels were populated with cell densities from 3 X [10.sup.4] cells per ml (60K, 125K, 250K, 500K) to 1 X [10.sup.6] cells per ml. Following overnight incubation the whole gels were transferred from their original wells to wells of a 24-well plate, washed with PBS, (1 mL per well), and analyzed using the MTT assay described above. The same assay technique was used to assess the cell number at various time points.

Tissue Processing for Immunochemistry

Collagen gels were fixed in 10% buffered formalin (formaldehyde 10mL, [Na.sub.2]HP[O.sub.4] 1.6g and Na[H.sub.2]P[O.sub.4].[H.sub.2]O 0.4g in 100 mls of distilled water, all from Fisher Scientific, Pittsburg, PA, USA),S at 4[degrees]C for 24 h, dehydrated through a series of ethanols and xylenes and embedded in paraffin. Embedded tissues were sectioned and the sections (~30i) were deparaffinized by incubations in xylenes and ethanols. After rehydration for 30 min in PBS and distilled water washes (3 x 10 min), the tissue sections were subjected to immunohistochemistry (by indirect immunofluorescence). Mounted specimens were examined on Olympus AX70 fluorescent microscope (Olympus, Center Valley, PA) using SPOT[R] TWAIN software (Microsoft, Issaquah, WA).


Slides were stained according to manufacturer's recommendations for IHC kit (Prohisto, Columbia, SC, USA). Rehydrated tissue sections were washed in Amplifying IHC wash buffer (lx, IHC Kit, Prohisto, Columbia, SC), 3 mL per slide for 5 minutes, rocking at room temperature (RT), in the amplifying chamber (IHC, Prohisto). Slides were then placed in epitope unmasking solution (1x, IHC Kit, Prohisto), using sufficient volume to cover all slides in the slide rack, and the slide rack was placed in boiling water for 20 minutes. After cooling to room temperature, slides were transferred to amplifying chamber (IHC Prohisto) in which all the subsequent steps were performed. Slides were washed with amplifying wash buffer (lx, 3mls/slide, IHC Kit, Prohisto), rocking for 5 minutes at RT and then blocked with 1% BSA (3ml per slide, EMD Chemicals, Gibbstown, NJ, USA) in PBS, rocking for 30 minutes at RT. Primary antibodies were diluted in amplifying antibody dilution buffer (lx, 3 ml per slid IHC Kit, Prohisto) as follows: mouse anti-osterix antibody (Novus biologicals, Littleton, CO, USA), to 1:500 and rabbit anti-osteopontin (Abcam Inc., Cambridge, MA, USA) to 1:800. The slides were incubated with rocking for 24 hours at 4[degrees]C. The slides were then washed 3 times in amplifying wash buffer (1x, 3 ml per slide, IHC Kit, Prohisto), rocking for 5 minutes at RT. Alexa Fluor 594 labelled goat anti mouse and Alexa Fluor 488 labelled goat anti rabbit (both from, Invitrogen, Eugene, OR, USA) secondary antibodies were diluted in PBS to 1:4000 and slides were incubated with 3 ml antibody solution per slide for 24 hours at 4oC on a rocker. Slides were then washed 3 times in amplifying wash buffer (1x, 3 ml per slide, IHC Kit, Prohisto), rocking for 5 minutes, at RT, followed by distilled 3 water washes, 3 ml per slide rocking for 5 minutes at RT. Coverslips were mounted on the slides using Prolong Gold antifade reagent with DAPI (Invitrogen, Eugene, OR, USA). After allowing the slides and coverslips to dry overnight the slides were examined by fluorescent microscopy as described earlier.

Statistical Analysis

GraphPad Prism 4 (GraphPad Software, LaJolla, CA) was used to graphically represent and statistically analyze the results. Two-way ANOVA was used for analysis of gel contraction and was followed by Bonferroni post comparison tests to compare individual groups. In all the experiments the "n" indicates biological replicates. Area under the curve (AUC) was calculated using trapezoid formula in order to compare the effect of cell density (for early and late passage cells) on collagen contraction. It served as a representation of the total contraction over 28 days. AUC = 1/2 [Weight on day 1 + 2* Weight on day 14 + Weight on day 28] * 4

One-way ANOVA was applied for AUC analysis and this was followed by Bonferroni post comparison tests, which were utilized to compare individual groups.


Determination of Cell Number in 3-D

Absorbance of light by formazan plotted against cell number is a linear relation with [r.sup.2] values above 0.99 for both telomerase transfected mesenchymal stem cells (TMSC) and their differentiated progeny, telomerized osteoblasts (TOST) (n = 3 for both) when cells are cultured as a monolayer (Figure 1A).

When light absorbance by formazan is plotted against number of hMSC in collagen gel a linear relationship (with spearmans corelation statistic of 1.000) is observed for cell numbers ranging from 0.75 X [10.sup.4] cells to 25 X [10.sup.4] cells (corresponding to cell densities of 3 X [10.sup.4] cells/ml to 100 X [10.sup.4] cells/ml collagen gel) (Figure 1b).

Effect of Cell Number on Collagen Gel Contraction

A rapid contraction phase during the first 24 hours was observed for all seeding densities of hMSC (passage 4), with the gels contracting by approximately 20% (Figure 2a). After this initial period gels with 10 X [10.sup.4] hMSC per mL did not contract further while gels with 20 X [10.sup.4] and 40 X [10.sup.4] hMSC per mL continued to contract (40 X [10.sup.4] > 20 X [10.sup.4] > 10 X [10.sup.4]) over the first fourteen days in culture. Thereafter the contraction stopped even in high cell density gels. This relationship between initial seeding density and contraction remained constant though the cell proliferation in the gels continued (Figure 2b). When average contraction over a period of one month is considered (area under the curve) cell density is a significant factor in the collagen gel contraction process by early as well as late passage cells (Figure 2c).

In collagen gels containing a low cell density there was a change in morphology of the cells from bipolar to rounded/dendritic (osteoblast like) (Supplementary Figure 1A and 1B) along with expression of osterix (marker of differentiation) at day 28 (Supplementary Figure 1c and 1d).

Effect of Serum on Collagen Gel Contraction

With early passage hMSCs (passage < 5) are used, gel contraction in the presence of high (10%) and low (4%) serum culture medium was not significantly different for 10 X [10.sup.4] hMSC per mL (F igure 3A). Similar "contractile insensitivity" to differences in serum content were observed with collagen gels containing 40 X [10.sup.4] hMSC per mL. But significant difference in collagen gel contraction was observed when comparing 10 X [10.sup.4] cells per ml to 40 X [10.sup.4] cells per ml regardless of serum concentrations.

Effect of Serum on Cell Proliferation

hMSC populating collagen gels and cultured under high serum conditions show proliferation of cells over a period of one month irrespective of initial seeding density (Figure 2B). On the other hand when hMSC populating collagen gels are cultured under low serum conditions, gels seeded initially at 40 X [10.sup.4] and 20 X [10.sup.4] cells per mL show proliferation for 14 days but a decrease in the cell number from 14 to 28 days (Figure 3B). Collagen gels initially populated with 10 X [10.sup.4] hMSC per mL showed cell proliferation over 28 days when cultured in low serum medium

Effect of Cell Passage on Collagen Gel Contraction

Late passage (> 10) and early passage (< 5) hMSC contracted the collagen gels similarly (20%) during the first 24 in culture in the presence of 10% serum (F igure 4A). However over a period of 28 days in culture late passage hMSC caused significantly higher collagen gel contraction then early passage cells. Approximately 6% higher contraction was observed irrespective of the initial cell seeding density (10 X [10.sup.4] - 40 X [10.sup.4] cells per mL) with late passage hMSC. Early passage hMSC proliferated over a culture period of 28 days (Figure 2B) but late passage hMSC did not proliferate over the same period

Late passage hMSC in low (4%) serum conditions (Figure 4B) differed from early passage hMSC (Figure 3A). When serum concentration was reduced to 4% there was no significant contraction of the collagen gels over a one-month culture period in case of late passage MSC whereas early passage hMSC had contracted the collagen gel under low serum concentrations.

Effect of Constitutive Expression of Telomerase on Collagen Gel Contraction

During the first 24 hours of culture in 10% serum containing ODM telomerised MSCs, (TMSC, 10 X [10.sup.4] cells/ml) contracted collagen gels by 10% (Figure 5 A). Cultured under the same conditions collagen gels populated with hMSC a significantly greater contraction (20%) was observed. At the end of 14 days TMSC were significantly less (13%) contractile than hMSC cultured under identical conditions (FIG 5A).

Telomerised human osteoblasts (TOST) were derived by osteogenic differentiation of TMSC over a period of 12 weeks. When populating collagen gels (10 X [10.sup.4] cells/ml) at the end of 24 hours culture period in 10% serum containing ODM, TOSTs caused a significantly higher collagen gel contraction (20%) than TMSC (10%) (Figure 5B). Over a period of 14 days in culture TOST populated collagen gel contraction was significantly higher (by 8%) than that caused by TMSC. The proliferation of hMSC, TMSC and TOST in the collagen gel were similar and conform to the same equation (supplementary figure 2)


Collagen is the most important structural protein in the body and is therefore a natural choice for tissue engineering applications [1]. Fibrillar collagens, types I, II and III are basic components of connective tissues, in which collagen is synthesized by resident cells. Collagen fibrils are formed from 300 nm long and 1.5 nm wide parallel tropocollagen arrays. The use of collagen gels as matrices for the construction of 3-D tissue models was first reported by Elsdale in 1972 [16], who described a method of preparation of type I collagen gels as an environment in which fibroblast can be cultured. In 1979 Bell et al observed that contraction of collagen gel by fibroblasts caused extrusion of water that could be detected as a decrease weight [17]. It was concluded that the presence of fibroblasts in the collagen gels was responsible for contraction since a-cellular collagen gels, cultured under similar conditions did not contract. In our study a 10% loss in weight was observed when a-cellular collagen gels extruded water during physical handling (e.g. transfer from the casting wells to the weighing boats). If the extruded water is taken into consideration there was no change in weight. Thus in all our experiments involving collagen gels the loss in weight was compensated to account for mechanical perturbation during handling and were compared with the weight of collagen gel that was initially dispensed.

Grinnell compared collagen gel contraction, in two gel formats, "free floating" gels and "anchored" gels [18] that were similar with respect to tractional remodeling, integrin expression and extracellular forces. In anchored gels tension develops anisotropically and contracted gel is under tension as long as it remains anchored. The free-floating gels were proposed to be more in vivo like because they were not constrained and responded freely to mechanical perturbation. However, in vivo all tissues are under homeostatic tension and in vitro cells can proliferate in anchored gels but do not proliferate, are not as responsive to growth factors and show reduced collagen synthesis in free-floating gels. The cells in anchored gels align with lines of tension and the anisotropic contraction resembles that observed in granulation tissue in vivo while the cellular functions within free-floating gels resembles cell behavior within scar tissue. Therefore use of anchored collagen gels to study contraction during osteogenic differentiation is appropriate.

Use of MTT to Track Cell Number

Resident cell viability is clearly an issue in 3-D matrices including collagen gels. Following cell viability using live-dead assays Sumanasinghe et al [19] acknowledged the limitations of this protocol and the difficulties in accurate determination of the cell number in a 3-D collagen gel. The use of collagenase digestion of the collagen gels was proposed to lead to cell death and was the contributing factor to the low cell numbers. In the present study we showed that MTT conversion to formazan is a simple and effective assay to determine the cell number in collagen gels and other 3-D matrices (collagen based scaffolds, unpublished data). For example the response of the assay, graphical plot of absorbance against cell number, when applied to gels populated with freely proliferating TMSC is not significantly different from the response of TOST, and the same equation can be used to describe both responses (Figure 1A). This indicated that osteogenic differentiation (differentiation of TMSC into TOST) did not affect the performance of the MTT assay. This indicated that the MTT assay was appropriate protocol for studies of cell proliferation during osteogenic differentiation of hMSC in 3-D collagen matrix. A Spearmans correlation statistic of 1.000 indicates a strong corealation between formazan formation (light absorbance by dissolved formazan) and cell number (Figure 1B) in collagen gel. This standard curve can be used to calculate the number of cells from the observed absorbance. The correlation between number of cells in a collagen gel and the absorbance was statistically significant and linear over the range 30k - 1 million hMSC per mL, an appropriate range for our studies.

Effect of Cell Density

Bell et al observed that collagen gel contraction by fibroblasts was initially dependent on the number of cells populating the gel matrix. After extended culture time and after a certain cell density threshold is exceeded, the contractions were independent of starting cell density [17]. Below this poorly defined contractile sensitivity threshold, collagen gel contraction was linearly dependent on cell density. Nishiyama also observed similar cell density dependence of collagen gel contraction by fibroblasts [20] .

During the first 16 hours of culture, Awad et. al. observed that the rate of contraction for collagen gels populated by 1 million rabbit MSC cells/mL was slower than for gels with 4 and 8 million cells [21]. By measuring the initial and final diameter of the constructs they determined that after 72 hours in culture the collagen gel contracted by 60-77% but thereafter the contractions rate was independent of cell number. Using rabbit MSC and extending the study period, Nirmalanandhan [22] showed that the threshold of dependence of collagen gel contractions on cell number was ~500k cells per mL, 100k rMSC per ml collagen contracted the collagen gel by 70% determined by measuring the initial area of the construct and comparing with the area at 7 days, while 1 milion cells resulted in 95% contraction. In our present study using human primary cells we observed collagen gel contraction of 20% after the first 24 hours in culture independent of initial seeding density (Figure 2A). Collagen gels prepared in accordance with our patented technology [14] and seeded with 400,000 cells per ml contracted by 30% which is far less than that previously reported by other groups [13] [19] [21] [22] .

Rat calvarial osteoblasts proliferation was reduced when in 3-D (collagen gels) when compared with that in monolayers cultures [23]. Unfortunately the low amount of collagen used makes this and a number of other studies an example of cell behavior in thin collagen films and not 3-D matrices. Primary human osteoblasts in collagen gels did not proliferate and a significant decrease in cell viability was reported after day 16 [24]. The diameter of gel containing the osteoblasts decreased by 50% in 24 hours (a decrease in volume of 87.5%) and the diameter of the construct stabilized at one-third the original for the remaining 10 days of the experiment. Although it was proposed that the cells became quiescent (arrest in proliferation) it is more likely that there was a loss in cell viability and cell death subsequent to inadequate nutrient supply due to severe increase in matrix density. In our study, minimized contraction allowed continued cell proliferation over 28 days at all seeding densities in the presence of 10% FBS supplementation (Figure 2B) and proved to be a suitable environment for osteogenic differentiation.

In our experiment with hMSC and TMSC the gel contraction at the end of 24 hours was similar for all gels regardless of the initial seeding density (Figure 2A) and the contraction over one month (Figure 2A and 2C) was only marginally increased (9% difference, p < 0.05) for high initial seeding density when early and late passage cells were used. Holy et al demonstrated that bone tissue formation in polymeric scaffolds does not depend on initial cell seeding density [25]. Morphological changes (H&E staining) and expression of osterix showed that in our experiments the cells populating collagen gels at low seeding density (efficient use of therapeutically important cells) were able to differentiate (Supplementary figure 1). The use of a low cell seeding density to prevent collagen gel contraction during generation of tissue engineered grafts may be a viable option to overcome the problems associated with gel contraction.

Effect of Serum

The supply and free access of nutrients is critical in all phases of the wound healing in all tissues and is equally important in bone regeneration. In vitro increasing the serum concentration (10% to 20%) increased collagen gel contraction by rat MSC [13] a similar response to that exhibited by fibroblast populating collagen type I gels [20] [26]. The diameter of native gels with rat MSC contracted by 20% in 24 hours and by 76% in 14 days at 10% serum oncentration [13]. Gels modified by addition of polymerized dehydrated collagen fibers contracted by 45% in 14 days. Comparing contractions of collagen gels is challenging because of the variety of measurement methods used. Thus although in Lewus study a plateau in contraction was noted at 7 days, the reduction of 30% in diameter translates to a 66% decrease in volume. Furthermore addition of polymerized dehydrated collagen fibers would lead to previously discussed (Introduction) drawbacks associated with cross-linked, nonnative collagen. Since in our gels acid solubilized native collagen is used we have shown a complete resorption without adverse reactions after implantation in vivo [27]. We observed that decreasing serum concentration to 4% did not have any effect on the overall collagen contraction (Figure 3 A) but inhibited proliferation after 14 days in culture (Figure 3B). This supports the observation that overall collagen contraction is dependent on the initial seeding density for hMSC as both the gels in which cells proliferated (high serum, Figure 2B) and in which the cells could did not (low serum, Figure 3B) contracted to the same extent. Thus decreasing serum in culture medium may be a feasible strategy to decrease collagen gel contraction for fibroblasts [27] but is not applicable for hMSC.

Cytoskeletal changes, particularly the expression of alpha smooth muscle actin (aSMA), in a substantial population of the contractile fibroblast phenotype--the myofibroblasts, has been the hall-mark of collagen gel contraction. Since contraction of collagen gels by MSC was linearly related to aSMA content of the cells, the aSMA content of cells has been proposed as an important determinant of gel contraction [28]. Human mesenchymal stem cells and osteoblasts both express aSMA and its expression in MSC was not affected when cells were cultured under different serum concentrations for five days [28]. This may explain why we did not observe any effect of the serum concentration on collagen gel contraction by early passage cells (Figure 3A).

Effect of Passage

Late passage fibroblasts were more contractile compared to early passage cells [17]. We observed similar results with late passage hMSC contracting the gel significantly more than early passage hMSC over a culture period of 28 days (Figure 4A). The late passage cell population probably contains a larger number of osteoblasts that are more contractile by virtue of greater SMA content [28].

The response of late passage cells to a decrease in serum was different from that of early passage cells. Lowering the serum concentration to 4% prevented contraction of collagen gels after day one, at all initial seeding densities of hMSC (Figure 4B). Thus use of late passage cells under low serum conditions may provide an alternative method of preventing collagen gel contraction when a high number of cells is required in a potential engineered graft.

Effect of Telomerase

It was reported that transformed fibroblasts contract the collagen gels less than normal cells [26]. We found that telomerase transformed MSC (TMSC) contracted the gel significantly less after 24 hours as well as after 14 days when compared to normal hMSC (Figure 5A). We hypothesize that fewer transformed cells adopt a contractile phenotype and develop a contractile phenotype later than normal cells.

It has been reported that culture of hMSC in ODM caused greater contraction when compared to MSCGM [19], indicating greater contraction by osteoblasts. We observed that TOST contracted the collagen gel significantly more than TMSC during the first 24 hours (10% difference) and over a period of 14 days (8% difference) (Figure 5B). The cell numbers for hMSC, TMSC and TOST doubled after two weeks of culture in collagen hydrogel under osteogenic differentiation conditions (supplementary figure 2). Thus the difference in contraction observed is likely due to increase in the population of the contractile phenotype of the cells. A greater contraction by TOST than TMSC is likely to be due to difference in expression of a-SMA, a contractile phenotype marker, which is upregulated during osteogenic differentiation [28].


In summary we have demonstrated that high initial cell seeding density, late passage cells, non-transformed cells, and differentiated hMSCs cells all increase collagen gel contraction. In contrast a low initial cell seeding density, early passage hMSC (non-differentiated) low serum culture medium, telomerase transfected cells, all minimize collagen contraction. Also allowing non-differentiated early passage hMSC to differentiate in situ could be an additional strategy to prevent collagen gel contraction of collagen type I hydrogels used as 3-D tissue culture models or tissue engineering grafts. A non-contracted matrix will provide a permeable matrix composed of loosely spaced native fibrils. These features would facilitate faster cell invasion (early vascularization), extensive remodelling and greater survival of implanted grafts leading to better host-graft integration and superior graft performance.


The authors would like to thank the Department of Molecular Cardiovascular Physiology for supporting this work by providing a doctoral stipend to Anupam Sule.


[1.] Kielty CM, Hopkinson I, Grant ME. Collagen--the family: structure, assembly and organization in the extracellular matrix In: Connective Tissue and its Heritable Disorders. Wiley-Liss; p. 103--147 (1993).

[2.] Gelse K, Poschl E, Aigner T. Collagens--structure, function, and biosynthesis. Adv Drug Deliv Rev; 55(12):1531-46. (2003)

[3.] ROBINSON RA, WATSON ML. Collagen-crystal relationships in bone as seen in the electron microscope. Anat Rec; 114(3):383-409. (1952)

[4.] Mizuno M, Fujisawa R, Kuboki Y. Type I collagen-induced osteoblastic differentiation of bone-marrow cells mediated by collagen-alpha2beta1 integrin interaction. J Cell Physiol; 184(2):207-13.(2000)

[5.] Buxton PG, Bitar M, Gellynck K, Parkar M, Brown RA, Young AM, Knowles JC, Nazhat SN. Dense collagen matrix accelerates osteogenic differentiation and rescues the apoptotic response to MMP inhibition. Bone; 43(2):377-85. (2008)

[6.] Weinand C, Pomerantseva I, Neville CM, Gupta R, Weinberg E, Madisch I, Shapiro F, Abukawa H, Troulis MJ, Vacanti JP. Hydrogel-beta-TCP scaffolds and stem cells for tissue engineering bone. Bone; 38(4):555-63. (2006)

[7.] Kasten P, Luginbuhl R, van Griensven M, Barkhausen T, Krettek C, Bohner M, Bosch U. Comparison of human bone marrow stromal cells seeded on calcium-deficient hydroxyapatite, beta-tricalcium phosphate and demineralized bone matrix. Biomaterials; 24(15):2593-603. (2003)

[8.] Niemeyer P, Krause U, Fellenberg J, Kasten P, Seckinger A, Ho

AD, Simank HG. Evaluation of mineralized collagen and alpha-tricalcium phosphate as scaffolds for tissue engineering of bone using human mesenchymal stem cells. Cells Tissues Organs; 177(2):68-78. (2004)

[9.] Levy RJ, Schoen FJ, Sherman FS, Nichols J, Hawley MA, Lund SA. Calcification of subcutaneously implanted type I collagen sponges. effects of formaldehyde and glutaraldehyde pretreatments. Am J Pathol; 122(1):71-82. (1986)

[10.] Engler AJ, Sen S, Sweeney HL, Discher DE. Matrix elasticity directs stem cell lineage specification. Cell; 126(4):677-89. (2006)

[11.] Lutolf MP, Lauer-Fields JL, Schmoekel HG, Metters AT, Weber FE, Fields GB, Hubbell JA. Synthetic matrix metalloproteinase-sensitive hydrogels for the conduction of tissue regeneration: Engineering cell-invasion characteristics. Proc Natl Acad Sci U S A; 100(9):5413-8. (2003)

[12.] Mauney J, Volloch V. Progression of human bone marrow stromal cells into both osteogenic and adipogenic lineages is differentially regulated by structural conformation of collagen I matrix via distinct signaling pathways. Matrix Biol; 28(5):239-50. (2009)

[13.] Lewus KE, Nauman EA. In vitro characterization of a bone marrow stem cell-seeded collagen gel composite for soft tissue grafts: Effects of fiber number and serum concentration. Tissue Eng; 11(7-8):1015-22. (2005)

[14.] Dimitrijevich SD, Gracy RW, inventors Non-contracting tissue equivalent. 6471958. 10/29/2002.

[15.] Mihara K, Imai C, Coustan-Smith E, Dome JS, Dominici M, Vanin E, Campana D. Development and functional characterization of human bone marrow mesenchymal cells immortalized by enforced expression of telomerase. Br J Haematol; 120(5):846-9. (2003)

[16.] Elsdale T, Bard J. Collagen substrata for studies on cell behavior. J Cell Biol; 54(3):626-37. (1972)

[17.] Bell E, Ivarsson B, Merrill C. Production of a tissue-like structure by contraction of collagen lattices by human fibroblasts of different proliferative potential in vitro. Proc Natl Acad Sci USA; 76(3):1274-8. (1979)

[18.] Grinnell F. Fibroblasts, myofibroblasts, and wound contraction. J Cell Biol; 124(4):401-4. (1994)

[19.] Sumanasinghe RD, Osborne JA, Loboa EG. Mesenchymal stem cell-seeded collagen matrices for bone repair: Effects of cyclic tensile strain, cell density, and media conditions on matrix contraction in vitro. J Biomed Mater Res A; 88(3):778-86. (2009)

[20.] Nishiyama T, Tominaga N, Nakajima K, Hayashi T. Quantitative evaluation of the factors affecting the process of fibroblast-mediated collagen gel contraction by separating the process into three phases. Coll Relat Res; 8(3):259-73. (1988)

[21.] Awad HA, Butler DL, Harris MT, Ibrahim RE, Wu Y, Young RG, Kadiyala S, Boivin GP. In vitro characterization of mesenchymal stem cell-seeded collagen scaffolds for tendon repair: Effects of initial seeding density on contraction kinetics. J Biomed Mater Res; 51(2):233-40. (2000)

[22.] Nirmalanandhan VS, Levy MS, Huth AJ, Butler DL. Effects of cell seeding density and collagen concentration on contraction kinetics of mesenchymal stem cell-seeded collagen constructs. Tissue Eng; 12(7):1865-72. (2006)

[23.] Talley-Ronsholdt DJ, Lajiness E, Nagodawithana K. Transforming growth factor-beta inhibition of mineralization by neonatal rat osteoblasts in monolayer and collagen gel culture. In Vitro Cell Dev Biol Anim; 31(4):274-82. (1995)

[24.] Rattner A, Sabido O, Le J, Vico L, Massoubre C, Frey J, Chamson A. Mineralization and alkaline phosphatase activity in collagen lattices populated by human osteoblasts. Calcif Tissue Int; 66(1):35-42. (2000)

[25.] Holy CE, Shoichet MS, Davies JE. Engineering three-dimensional bone tissue in vitro using biodegradable scaffolds: Investigating initial cell-seeding density and culture period. J Biomed Mater Res; 51(3):376-82. (2000)

[26.] Steinberg BM, Smith K, Colozzo M, Pollack R. Establishment and transformation diminish the ability of fibroblasts to contract a native collagen gel. J Cell Biol; 87(1):304-8. (1980)

[27.] Dimitrijevich SD, inventor; The University of North Texas Health Science Center at Fort Worth (Fort Worth, TX), assignee. Pericardial anti-adhesion patch. United States Patent 6599526.

[28.] Kinner B, Zaleskas JM, Spector M. Regulation of smooth muscle actin expression and contraction in adult human mesenchymal stem cells. Exp Cell Res; 278(1):72-83. (2002)

Anupam Ashutosh Sule (1, 3), Arvind Nana (2), S. Dan Dimitrijevich (1)

(1) Department of Molecular Cardiovascular Physiology, Graduate School of Biomedical Sciences, (2) Department of Orthopeadic Surgery, Bone and Joint Center, University of North Texas Health Science Center, Fort Worth, TX, USA

(3) Department of Internal Medicine, St Joseph Mercy Oakland, Pontiac, MI, USA

Received 30 May 2017; Accepted 11 November 2017; Published online 31 December 2017

* Coresponding author: Dr. Anupam Ashutosh Sule;


Caption: Figure 1: Number of cells in a collagen hydrogel as determined by MTT Assay. (a) Shows linear relationship between formazan absorbance and the number of cells ranging from 0.25 X [10.sup.4] to 10 X [10.sup.4] cells. The formazan formation is similar for TMSC and TOST. "Runs test" of significance indicates that both lines are not significantly different from one another confirming that MTT assay can be used to determine cell numbers during osteogenic differentiation (Mean and SEM, n = 3). (b) Formazan formation is linearly related to the number of cells in collagen hydrogel over a range of 3 X [10.sup.4] cells per ml to 100 X [10.sup.4] cells per ml. A spearmans correlation statistic of 1 indicates a strong correlation between formazan formation and number of cells in collagen hydrogel. (Mean and SEM, n = 3)

Caption: Figure 2: Effect of cell seeding density on collagen hydrogel contraction. (a) Collagen hydrogels contract to about 80% of their original weight within 24 hours, regardless of the cell seeding density in the presence of 10% serum. Over the next 14 days collagen gel with higher seeding density contract significantly more than those with lower seeding density. The gels do not contract significantly after 14 days. (Mean and SEM, n = 3, p < 0.05). (b) Shows that hMSC continue to proliferate over a period of 28 days under osteogenic differentiation conditions at all seeding densities in the presence of 10% serum. (Mean and SEM, n = 3, p < 0.05). (c) Early and late passage cells contract the collagen gel significantly more at higher cell seeding densities under osteogenic differentiation conditions with 10% serum. A lower area under the contraction curves (graphed in figures 2A and 2B) indicates greater contraction has occurred. When this area is graphed as a bar diagram in figure 2C it is evident that 10 X [10.sup.4] cells per ml have a significantly greater area under the curve (9ndicating lesser contraction) than 40 X [10.sup.4] cells per ml regardless of whether early or late passage cells are used. (Mean and SEM, n = 3, p < 0.05)

Caption: Figure 3: Effect of serum on collagen hydrogel contraction. (a) When cultured in the presence of either 10% or 4% serum, collagen hydrogel contraction follows a similar course at 10 X [10.sup.4] cells per ml and at 40 X [10.sup.4] cells per ml. Thus lowering the serum supplementation does not prevent collagen hydrogel contraction. (Mean and SEM, n = 3, p < 0.05). (b) At low serum concentration in the medium, cells in 20 X [10.sup.4] cells per ml and 40 X [10.sup.4] cells per ml gels, proliferated for 14 days after which the cell number decreased. At 10 X [10.sup.4] cells per ml proliferation in the gels continues to occur over 28 days. Thus low serum culture conditions do not support cell survival or proliferation in early passage cells over a long culture period. (Mean and SEM, n = 3, p < 0.05)

Caption: Figure 4: Effect of cell passage on collagen gel contraction. (a) Early and late passage cells contract the collagen gel in a similar fashion over the first 24 hours. At the end of 28 days late passage cells contracted collagen gel significantly more than early passage cells at 10 X [10.sup.4] cells/ml and 40 X [10.sup.4] cells/ml seeding densities. Late passage cells did not proliferate over the duration of the experiment. (Mean and SEM, n = 3, p < 0.05). (b) Decreasing serum concentration in the culture medium (4%) for late passage cells prevented collagen gel contraction at all seeding densities. This differs from early passage cells where serum did not have any effect on collagen gel contraction. (Mean and SEM, n = 3, p > 0.05)

Caption: Figure 5: Effect of telomerase expression on collagen gel contraction. (a) Telomerase transformed mesenchymal stem cells (TMSC) contracted collagen gels significantly less as compared to primary mesenchymal stem cells (hMSC) at 24 hours and at the end of 14 days in culture. (40 X [10.sup.4] cells/ml, Mean and SEM, n = 3, p < 0.05). (b) Effect of differentiation on collagen gel contraction. Telomerase transformed human osteoblasts (TOST) contracted the gel significantly more at 24 hours and 14 days as compared to TMSC. (40 X [10.sup.4] cells/ml, Mean and SEM, n = 3, p < 0.05)
COPYRIGHT 2017 Society for Biomaterials and Artificial Organs
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 2017 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Title Annotation:Original Article
Author:Sule, Anupam Ashutosh; Nana, Arvind; Dimitrijevich, S. Dan
Publication:Trends in Biomaterials and Artificial Organs
Date:Jul 1, 2017
Previous Article:Tissue Engineering of Skin: A Review.
Next Article:Decellularised and Chemically Cross linked Human Umbilical Vein as a Small Caliber Conduit--Scaffold for Vascular Tissue Engineering.

Terms of use | Privacy policy | Copyright © 2019 Farlex, Inc. | Feedback | For webmasters