Bone formation by three-dimensional stromal osteoblast

Bone formation by three-dimensional stromal osteoblast

(Parte 1 de 3)

Bone formation by three-dimensional stromal osteoblast culture in biodegradable polymer scaffolds

Susan L. Ishaug,1,† Genevieve M. Crane,1 Michael J. Miller,2 Alan W. Yasko,3 Michael J. Yaszemski,4,‡ and Antonios G. Mikos1,* 1Cox Laboratory for Biomedical Engineering, Institute of Biosciences and Bioengineering, Rice University, P.O. Box 1892, Houston, Texas 77251; 2Department of Reconstructive and Plastic Surgery, University of Texas MD Anderson Cancer Center, Houston, Texas 77030; 3Department of Orthopaedic Surgery, University of Texas MD Anderson Cancer Center, Houston, Texas 77030; 4Department of Orthopaedic Surgery, Wilford Hall and Medical Center, Lackland AFB, Texas 78236

Bone formation was investigated in vitro by culturing stromal osteoblasts in three-dimensional (3-D), biodegradable poly(DL-lactic-co-glycolic acid) foams. Three polymer foam pore sizes, ranging from 150–300, 300–500, and 500–710 m, and two different cell seeding densities, 6.83 × 105 cells/cm2 and 2.1 × 105 cells/cm2, were examined over a 56-day culture period. The polymer foams supported the proliferation of seeded osteoblasts as well as their differentiated function, as demonstrated by high alkaline phosphatase activity and deposition of a mineralized matrix by the cells. Cell number, alkaline phosphatase activity, and mineral deposition increased significantly over time for all the polymer foams. Osteoblast foam constructs created by seeding 6.83 × 105 cells/cm2 on foams with 300–500 m pores resulted in a cell density of 4.63 × 105 cells/cm2 after 1 day in culture; they had alkaline phosphatase activities of 4.28 × 10−7 and 2.91 × 10−6 mmol/cell/min on Days 7 and 28, respectively; and they had a cell density that increased to 18.7 × 105 cells/cm2 by Day 56. For the same constructs, the mineralized matrix reached a maximum penetration depth of 240 m from the top surface of the foam and a value of 0.083 m for mineralized tissue volume per unit of cross sectional area. Seeding density was an important parameter for the constructs, but pore size over the range tested did not affect cell proliferation or function. This study suggests the feasibility of using poly(a-hydroxy ester) foams as scaffolding materials for the transplantation of autogenous osteoblasts to regenerate bone tissue. © 1997 John Wiley & Sons, Inc.

Skeletal reconstruction is required in cases involving large defects created by tumor resection, trauma, and skeletal abnormalities.1 Presently, grafts and flaps of autogenous tissue are two of the most successful means of reconstruction because they allow for the transplantation of bone containing bioactive molecules, live cells, and, frequently, a vascular supply that will allow the transplant to survive and remodel even in hostile radiated environments.2 Other current therapies involve the use of allograft bone, nondegradable bone cement, metals, and ceramics.3 All of these options have their associated problems and limi- tations: only a minimal amount of tissue can be harvested for autografts, and it can be very difficult to form into the desired shapes; allografts have the potential of transferring pathogens; and synthetic implants may result in stress shielding to the surrounding bone or fatigue failure of the implant. These shortcomings have inspired a search for improved methods of repairing skeletal defects.

The ideal bone substitute would approximate the autograft, requiring minimally that it be biocompatible and osteoconductive, contain osteoinductive factors to enhance new bone ingrowth, and contain osteogenic cells to begin secreting new extracellular matrix.1 Bone regeneration by autogenous osteoblast transplantation meets these requirements and thus holds promise as an improved method of skeletal reconstruction. The scaffolding material used in this approach must allow for the attachment of osteoblasts because they are anchorage-dependent cells that require a supportive matrix in order to survive.4 The

*To whom correspondence should be addressed. †Present address: Los Alamos National Laboratory, MS

M888, CST-1, Los Alamos, New Mexico 87545.

‡Present address: Department of Orthopedic Surgery, Mayo Clinic, Rochester, Minnesota 55905.

Journal of Biomedical Materials Research, Vol. 36, 17–28 (1997) © 1997 John Wiley & Sons, Inc. C 0021-9304/97/010017-12 material also must provide an appropriate environment for the proliferation and function of osteoblasts and allow for the ingrowth of vascular tissue to ensure the survival of the transplanted cells. Finally, it should be biodegradable with a controllable rate of degradation into molecules that easily can be metabolized or excreted, and it should be processable into irregular 3-D shapes.

Poly(a-hydroxy esters), including poly(DL-lacticco-glycolic acid) (PLGA) copolymers, satisfy many, if not all, of these material requirements.5 The ability of two-dimensional (2-D) PLGA films to support osteoblast attachment, growth, and function already has been established in our laboratory.6 Recently, we also demonstrated that osteoblasts on polymer films migrate from isolated osteoblast cultures and bone chips as a monolayer of cells and continue to proliferate to form multilayers.7 Two-dimensional cultures such as these were necessary to establish the potential for using poly(a-hydroxy esters) as a supportive material for osteoblast growth and function, but they are not the ideal form for transplanting into defect sites. Skeletal defects vary in size and shape and require a 3-D graft material to fill or replace the missing tissue. Poly(a-hydroxy esters) can be fabricated into 3-D foams that can serve as a supportive scaffold for the culture and transplantation of osteoblasts,8 and they have the potential of filling in skeletal defects of various sizes and shapes. These poly(a-hydroxy esters) foams already have been shown to allow for the penetration of vascular tissue,9 which is essential to supporting the metabolic needs of the transplanted cells. Osteoblasts should populate the constructs by proliferation of the transplanted cells and the migration of cells into the construct from the surrounding tissue while the polymer scaffold gradually degrades. Eventually the construct will be filled with calcified extracellular matrix secreted by the osteoblasts and will be devoid of the synthetic biodegradable polymer.

The transplanted osteoblasts may be obtained by a variety of methods for in vitro and in vivo studies, including migration from bone chips and enzymatic digestion of harvested bone.10 However, the most desirable method would be to obtain osteoblasts percutaneously from the patient’s bone marrow. This would avoid the need for open surgery, with its possible donor-site complications of pain, infection, and damage to nerves and blood vessels. In addition, because the cells would be of autogenous origin, there would be no risk of immune rejection and little change of pathogen transfer. Several studies have demonstrated the feasibility of obtaining bone-forming cells from human, rabbit, and rat bone marrow.1–13 These techniques involve in vitro expansion of the mesenchymal stem cells present in the marrow by the addition to culture media of the appropriate factors to enhance osteoblast differentiation and function.

An investigation of the effects of polymer foam morphology and culture conditions on cell proliferation and function was needed to elucidate the important parameters in the design of an in vitro osteoblast foam culture system before osteoblast transplantation could be attempted in vivo. Such an investigation is the focus of the present study. Rat stromal osteoblasts have been seeded onto highly porous PLGA foams of different pore sizes and cell seeding densities and cultured over a 56-day period. This study addresses: 1) whether polymer foam pore size in the range of 150– 710 m affects osteoblast proliferation and function in vitro, 2) whether osteoblast seeding density on polymer foams affects cell attachment, proliferation, and function in vitro, and 3) whether a 3-D osseous tissue can be formed by culturing osteoblasts in polymer foams in vitro.

Polymer foam fabrication

Polymer foams of three different pore sizes were fabricated by a solvent-casting particulate-leaching technique with NaCl as the leachable component.14 NaCl was sieved into particles ranging in diameter from 150–300, 300–500, or 500–710 m and combined with 75:25 PLGA (Birmingham Polymers, Birmingham, AL) dissolved in chloroform to make 90% porous foams of various pore sizes. The foams were approximately 1.9 m thick and cut into 7 m diameter disks. They were stored under vacuum until use. Prior to cell seeding, the foams were prewetted with ethanol for 30 min to sterilize and enhance their water uptake.15 The ethanol was removed by soaking with agitation for1hi n three changes of phosphate-buffered saline (PBS) and then for3hi nt wo changes of media.

Stromal cell isolation, seeding, and culture

Stromal osteoblastic cells were obtained from the marrow of young adult male (6-weeks old, 150–170 g) Sprague Dawley rats.13 Following euthanasia by ethyl ether inhalation, femora were aseptically excised, cleaned of soft tissue, and washed in Dulbecco’s Modified Eagle medium (DMEM) (Life Technologies, Grand Island, NY) containing 250 mg/mL gentamicin sulfate (GS) (Sigma Chemical, St. Louis, MO). This concentration of antibiotics is 10 times the normal amount used in cell culture and was used as a precautionary measure to avoid contamination during the femora harvest. The metaphyseal ends then were cut off and the marrow flushed from the midshaft with 5 mL of primary media [DMEM containing 10% fetal bovine serum (FBS, Hyclone, Logan, Utah) and 25 mg/ mL GS] using a syringe equipped with a 2-gauge


needle and collected in a sterile petri dish. Cell clumps were broken up by repeatedly pipetting the cell suspension. The cells then were centrifuged at 400 g for 10 min. The resulting cell pellets were resuspended in 12 mL of primary media and plated in T-75 flasks (cells from two femurs per flask). After 3 days, hematopoietic cells and other unattached cells were removed from the flasks by repeated washes with PBS. With confluent monolayers were reached (yielding approximately2×1 06 cells/femur) the cells were enzymatically lifted from the flask using a 625 mg/mL solution of trypsin.6 The cells were concentrated by centrifugation at 400 g for 10 min and resuspended in a known amount of media. Cells were counted with a Coulter Counter and diluted to concentrations of either 56,587,0 cells/mL (high density) or 17,525,0 cells/ mL (low density) in complete media [consisting of DMEM supplemented with 10% FBS, 8 mg/mL GS, 10 mM Na b-glycerol phosphate (Sigma), and 50 mg/mL L-ascorbic acid (Sigma)] containing 10 nM dexamethasone (Sigma) to promote the osteoblastic phenotype of marrow stromal cells.16,17 Aliquots of 15 mLo fe ither the high- or the low-density cell suspensions were seeded onto the top of prewetted foams placed in the wells of 24-well plates, resulting in seeding densities of 2.1 × 105 cell/cm2 and 6.83 × 105 cells/cm2 (or 849,0 and 263,0 cells/foam), respectively, when normalized to the top surface area of the foams. The foams were left undisturbed in an incubator for3ht o allow the cells to attach to the foams, after which time an additional 1 mL of complete media containing 10 nM dexamethasone was added to each well. Medium was changed every 2–3 days. Cells also were seeded into 35 m tissue culture polystyrene (TCPS) dishes as a conventional 2-D control culture to monitor the functionality of the cells. The cells were seeded at a density of 2.1 × 104 cells/cm2 in the TCPS dishes and used as a comparison for the alkaline phosphatase activity and for a qualitative assessment of mineralization.

DNA assay

Cell numbers were determined by a fluorometric quantification of DNA in the foams by an assay adapted from West et al.18 After 1, 7, 14, 28, and 56 days in culture, the foams and 2-D control cultures were harvested by rinsing with PBS and freezing at −80°C until assayed. Cell standards were prepared and frozen at −80°C, as described. First, a stock solution of rat marrow osteoblasts was prepared by trypsinizing confluent monolayers of cells and resuspending them in DMEM to a final concentration of 106 cells/mL. Aliquots of 25–625 mL of the stock cell suspension were centrifuged at 400 g for 10 min, washed once in PBS. The resulting cell pellets (cell standards) were stored frozen at −80°C. DNA standards were prepared at the time of the assay from a stock of DNA solution. The aqueous stock solution of highly polymerized calf thymus DNA (type I, Sigma) was prepared by adjusting it spectroscopically to a concentration of 50 mg/mL (50 mg of DNA in 1.0 mL distilled

H2O produces 1.0 absorbance unit across a 1.0-cm light path at 260 nm). Aliquots of the DNA stock so- lution ranging from 0–160 mL were used as DNA standards. After thawing, the foams were homogenized in 1.4 mL of cold 10 mM ethylenediaminetetraacetic acid (EDTA) solution (pH 12.3), and the cell layers of the 2-D control cultures were scraped from the well bottoms using disposable cell scrapers (Fisher Scientific, Springfield, NJ), into 1.4 mL of the EDTA solution. An equal amount of cold EDTA solution was added to thawed cell standards and DNA standards. All samples and standard were placed in a 37°C bath for 20 min and subsequently cooled on ice. The pH of the samples and standards was adjusted to 7.0 by adding

200 mLo f1 M KH2PO4 prior to the addition of 1.5 mL of the fluorescent dye solution [200 ng/mL Hoechst

33258 dye (Polysciences, Warrington, PA), 100 mM NaCl, and 10 mM Tris adjusted to pH 7.0]. Supernatant sample fluorescence emissions at 455 nm were read with the excitation set at 350 nm on a fluorescence spectrophotometer (series 2, SLM Amino Bowman, Urbana, IL). Cell number and DNA content in the foams were determined by comparing sample fluorescent results to the DNA and cell standard curves.

Alkaline phosphatase assay

Production of alkaline phosphatase (ALPase) was measured spectroscopically. Osteoblast-seeded foams and 2-D osteoblast cultures of 7, 14, and 28 days were washed with PBS and then frozen. Upon thawing, the foams were homogenized with 1 mL Tris buffer (pH 8.0, Sigma), and the osteoblasts of the 2-D control cultures were scraped from the well bottoms, using disposable cell scrapers, into 1 mL Tris buffer. Both sets of samples were sonicated (Ultrasonik 300, J. M. Ney, Bloomfield, CT) for 4 min at 110 watts (50/60 Hz) in ice. Aliquots of 20 mL were incubated with 1 mL of a p-nitrophenyl phosphate solution (16 mmol/L, Diagnostic Kit 245, Sigma) at 30°C for up to 5 min. The production of p-nitrophenol in the presence of ALPase was measured by monitoring light absorbance by the solution at 405 nm at 1 min increments. The slope of the absorbance versus time plot was used to calculate the ALPase activity.

Confocal microscopy preparation

Osteoblast foam constructs were prepared for confocal microscopy by staining the viable cells with the

19BONE FORMATION IN POLYMER SCAFFOLDS fluorescent dye 28,7 8-bis-(2-carboxyethyl)-5-(and-6)- carboxyfluorescein, acetoxymethyl ester (BCECFAM). The 3-D cultures were incubated with 5 mg/mL BCECF-AM in complete media for 1 h. Following a rinse in PBS and replacement with fresh media, the cell cultures were examined using a confocal microscope (Zeiss LSM Axiovert, Carl Zeiss, Germany). Depth projection micrographs were constructed from 16 horizontal image sections through the cultures.

Histological preparation

Osteoblast foam constructs were prepared for histology after 1, 14, 28, and 56 days in culture. The samples were fixed and stored in 10% neutral buffered formalin until ready for embedding. Standard dehydration in sequentially increasing ethanol solutions to 100% ethanol was performed, followed by immersion in Hemo-De (Fisher Scientific, Pittsburgh, PA), in paraffin-saturated Hemo-De, and finally, in molten paraffin. Tissue blocks were sectioned at 5 m and stained either by hematoxylin and eosin for visualization of cells and demonstration of tissue formation or by von Kossa’s silver nitrate staining method for demonstration of matrix mineralization.


The volume of mineralized tissue per initial plating surface area and the penetration depth of mineralized tissue were calculated by histomorphometry. Von Kossa-stained tissue cross-sections were projected on an RGB color video camera (JVC, Model TK 107-OU) with a Jena 250-CF stereo light microscope (Micro- Tech Instruments, Dallas, TX). Output from the camera was routed to a digital image acquisition system (Quick Capture, Data Translation, Marlboro, MA) and analyzed with the public domain NIH Image software, version 1.54. Images of the central crosssectional region, 640 pixels or 3.9 m wide, were captured for each sample to be used for analysis. The mineralized tissue was distinguishable from the nonmineralized tissue and polymer material by its graylevel intensity and thus was selected. The resulting measured areas in pixels were converted to areas in mm2 following calibration of the system using a micrometer slide to determine pixel/m ratio. The mineralized tissue area per length was calculated by dividing the mineralized tissue area by the width of the selected cross-sectional region. The standard ste- reological relationship—that the area/length of a given phase in a two-dimensional field approximates the volume/surface area of that phase in threedimensions—was used to provide an estimate of the volume/surface area.

The penetration depth of mineralized tissue in the foams was calculated by manually measuring on the same captured images used above the maximum depth from the top surface of the polymer foam for which mineralized tissue was detected. The average of seven measurements taken at evenly spaced points along the width of the selected region was used to calculate the average mineralized penetration depth for each sample.

Polymer characterization

The number and weight average molecular weights of the polymer foams were determined over the course of the experiment by gel permeation chromatography (Waters, Milford, MA) equipped with a differential refractometer (Waters, Series 410). The precipitates of homogenized foams from the DNA assay were collected and dried overnight in a laminar flow hood. They were further dried and stored under high vacuum (50 m Hg) to remove any remaining water. The homogenized and dried foams were dissolved in HPLC grade chloroform and filtered with glass wool to remove any insoluble components. The solubilized samples then were eluted in a series configuration through a Phenogel 5-guard column (series 106337G, 50 × 7.8 m, Phenomenex, Torrance, CA) and a Phenogel 10 linear column (series 106338, 300 × 7.8 m, Phenomenex) at a flow rate of 1 mL/min. Polystyrene standards were used to construct a primary calibration curve.

Statistical analysis

All measurements were collected in triplicate and expressed as means ± standard deviations. Single factor analysis of variance (ANOVA) was employed to assess the statistical significance of results for all the osteoblast foam constructs whose single factor was pore size or seeding density. Scheffe’s method was used for multiple comparison tests at a significance level of 95% and 9%. In addition, a two-tailed unpaired t test was used to evaluate the significance of the cell-seeding density effect on percent of cell attachment after 1 day of culture.

Polymer foams

Polymer foams of controlled porosity, pore size, and thickness were fabricated with 75:25 PLGA by the sol-


vent-casting particulate-leaching technique using salt particles as the leachable porogen. The resulting foam pore sizes were dictated by the size of the salt particles used in the fabrication process. Foams processed with salt particles sieved into the size ranges 150–300 m [Fig. 1(a)], 300–500 m [Fig. 1(b)], and 500–710 m

[Fig. 1(c)] exhibited pore sizes comparable to the size of the salt particles used in the fabrication process. An interconnected pore morphology was apparent in all foams, which was possible because of the high porosity of the scaffolds.

Cell proliferation

Osteoblasts seeded onto 75:25 PLGA foams attached to the pore surfaces and continued to proliferate over the 56-day in vitro culture period on all the samples producing 3-D osteoblast foam constructs. Scanning electron micrographs reveal the pore morphology of a foam created using salt particles 300–500 m in diameter before [Fig. 1(b)] and after [Fig. 1(d)] osteoblast seeding. Osteoblasts can be seen covering the pore surfaces after 1 day in culture [Fig. 1(d)].

Only a fraction of the seeded cells remained attached to the polymer foams. The initial high-seeding density of 2.1 × 105 cells/cm2 resulted in only 1.8 × 105 cells/cm2 remaining attached to the 300–500 m foams after one day in culture, giving a percent attachment of 53 ± 1% after this 24-h period. For the lower seeding density, 4.63 × 105 cell/cm2 of the 6.83 × 105 cells/cm2 seeded onto the 300–500 m foams remained attached after 1 day in culture, giving a higher (p < 0.01) percent attachment of 68 ± 5%.

Confocal depth projection micrographs demonstrated the initial rapid growth achieved when osteoblasts were seeded at a low density on polymer foams having pore sizes 300–500 m in diameter [Fig. 2(a– c)]. Proliferation results, as determined by quantification of DNA in the polymer foams, also indicated that osteoblasts grew more rapidly in the foams seeded with a lower cell density (p < 0.01 at day 1), eventually reaching comparable cell numbers to the foams seeded with a high cell density by Day 7 [Fig. 3(a)]. Comparable cell numbers also were found in the foams after 14 days in culture; however, by 28 and 56 days in culture, the number of osteoblasts found in the foams seeded with a lower cell density was lower (p < 0.05) than that for the high-density seeded foams after the same culture time. Osteoblast proliferation leveled off in all the osteoblast foam constructs studied following 28 days in culture, with no significant change in cell numbers between Day 28 and Day 56 [Fig. 3(a,b)]. Osteoblasts proliferated equally well on polymer foams of all pore sizes studied [Fig. 3(b)]. The rate of osteoblast proliferation on TCPS should not be com- pared to the osteoblast foam construct rates of proliferation because they were seeded at a much lower cell density. ALPase activity for these cultures will be compared to activity results of the osteoblast foam constructs in the next section.

Alkaline phosphatase activity

All osteoblast foam cultures expressed high alkaline phosphatase activity that increased substantially over time in culture [Fig. 4(a,b)]. Osteoblasts seeded at a low-cell density on polymer foams of 300–500 m appeared to express higher levels of ALPase activity compared to foams seeded with a higher cell density at all days in culture; nevertheless, these results were not significantly different [Fig. 4(a)]. Measurement of ALPase activity of osteoblasts cultured on standard TCPS were included for comparison to the 3-D substrates. The cells in the constructs expressed levels of comparable ALPase activity to the standard 2-D osteoblast cultures after 7 and 14 days in culture [Fig. 4(a)], with the exception of Day 28, where slightly higher activity (p < 0.05) was observed for osteoblasts cultured on TCPS than in osteoblast foam constructs initially seeded with a high-cell density. Polymer foam pore size did not affect the expression of ALPase activity of the osteoblasts at any culture period [Fig. 4(b)].


In addition to growth, the osteoblasts began to lay down osteoid, demonstrated by the pink regions of H and E-stained sections of a construct after 56 days in culture [Fig. 5(a,c)]. Osteoblasts appeared to be embedded in the newly formed tissue matrix, which is characteristic of the natural osteoblast differentiation and their progression into osteocytic cells. Von Kossa’s staining of parallel histological sections revealed that portions of the tissue matrix had been mineralized, with the mineral deposits predominantly covering the surface of the construct [Fig. 5(b,d)].

(Parte 1 de 3)