Scaffold degradation elevates the collagen content and dynamic compressive modulus in engineered articular cartilage

Osteoarthritis and Cartilage (2009) 17, 220e227 ª 2008 Published by Elsevier Ltd on behalf of Osteoarthritis Research Society International. doi:10.1016/j.joca.2008.06.013

International Cartilage Repair Society

Scaffold degradation elevates the collagen content and dynamic compressive modulus in engineered articular cartilage K. W. Ng Ph.D.y, L. E. Kugler B.S.y, S. B. Doty Ph.D.x, G. A. Ateshian Ph.D.z and C. T. Hung Ph.D.y* y Cellular Engineering Laboratory, Department of Biomedical Engineering, Columbia University, 1210 Amsterdam Avenue, 351 Engineering Terrace, MC 8904, New York, NY 10027, USA z Musculoskeletal Biomechanics Laboratory, Department of Mechanical Engineering, Columbia University, New York, NY 10027, USA x Analytical Microscopy Laboratory, Hospital for Special Surgery, New York, NY, 10021, USA Summary Objective: It was hypothesized that controlled, scaffold removal in engineered cartilage constructs would improve their collagen content and mechanical properties over time in culture. Design: Preliminary experiments characterized the effects of agarase on cell-free agarose disks and cartilage explants. Immature bovine chondrocytes were encapsulated in agarose, cultured to day 42, and incubated with 100 units/mL agarase for 48 h. After treatment, constructs were cultured to day 91. The compressive Young’s modulus and dynamic modulus of the constructs were determined every 2 weeks and immediately after agarase treatment. Post-mechanical testing, constructs were processed for biochemistry and histology. Results: Agarase treatment on explants had no detrimental effect on the cartilage matrix. Treatment applied to engineered constructs on day 42 did not affect DNA or collagen content. Agarase treatment decreased tissue GAG content (via GAG loss to the media) and Young’s modulus, both of which recovered to control values over time in culture. By day 91 agarase-treated constructs possessed w25% more DNA, w60% more collagen, and w40% higher dynamic modulus compared to untreated controls. Conclusions: Scaffold degradation increased construct collagen content and dynamic mechanical properties, affirming the experimental hypothesis. The mechanism may lie in increased nutrient transport, increased space for collagen fibril formation, and cellular response to the loss of GAG with agarase treatment. The results highlight the role of the scaffold in retaining synthesized matrix during early and late tissue formation. This work also shows promise in developing an engineered tissue that may be completely free of scaffold material for clinical implantation. ª 2008 Published by Elsevier Ltd on behalf of Osteoarthritis Research Society International. Key words: Cartilage, Tissue engineering, Agarose, Biomaterials, Scaffold, Degradation, Mechanical properties, Agarase.

regardless of bioreactor conditions and scaffold material (e.g.,6,7). The formation of a mechanically competent tissue in vitro may be clinically relevant to prevent iatrogenic damage peri-operatively, along with host cell infiltration and fibrocartilage formation in vivo. In tissue engineering, scaffolds present a threedimensional environment to support cell phenotype, increase diffusional surface area, and entrain synthesized matrix products that form tissue8. Both degradable and non-degradable scaffolds have been adopted for cartilage tissue engineering (e.g.,9e11). Premature scaffold degradation can permit escape of cell products to the surrounding media before adequate tissue formation can occur. In contrast, delayed scaffold degradation may result in insufficient nutrient transport and/or physical space that impede formation of a fully competent tissue. Agarose has been used extensively as a model scaffold for cartilage tissue engineering research10,12e15. The concentration of the gel dictates the scaffold structure, mechanical properties, and transport properties16e18 and this can be utilized to direct tissue development in cartilage engineering19,20. As agarose is not degradable by chondrocytes10, changes to the construct properties can be directly attributed to chondrocyte anabolic

Introduction Engineered articular cartilage cultured with a temporal application of transforming growth factor b3 (TGF-b3) possesses a physiologic compressive Young’s modulus and glycosaminoglycan content1. The application of dynamic loading to these constructs further increases their mechanical properties2, which may be due to loading-induced convective transport3. However, the collagen content and dynamic modulus of these engineered tissues, even with applied loading, remain sub-physiologic. Given the relationship between collagen and mechanical properties in cartilage4,5, improving the collagen network in engineered cartilage may be necessary to develop a functional, clinical replacement for damaged cartilage. This low collagen content is a pervasive problem in cartilage engineering, *Address correspondence and reprint requests to: Dr Clark T. Hung, Cellular Engineering Laboratory, Department of Biomedical Engineering, Columbia University, 1210 Amsterdam Avenue, 351 Engineering Terrace, MC 8904, New York, NY 10027, USA. Tel: 1-212-854-6542; Fax: 1-212-854-8725; E-mail: cth6@columbia. edu Received 22 January 2008; revision accepted 18 June 2008.

220

221

Osteoarthritis and Cartilage Vol. 17, No. 2 and catabolic activities. However, agarose can be enzymatically degraded by the commercially available enzyme agarase21e23 and the controlled degradation of the scaffold could be used to beneficially modulate engineered cartilage tissue development in culture. It is hypothesized that following an initial period of growth where the scaffold provides suitable retention of matrix products, the degradation of the scaffold will improve the collagen content and mechanical properties of engineered cartilage over time in culture. In Study 1, agarose disks and articular cartilage explants were exposed to agarase under culture conditions (37 C, 5% CO2) to assess the enzyme’s effects on agarose, as well as any potential harmful effects on cartilage extracellular matrix (ECM). In Study 2, agarase was added to engineered cartilage after matrix had developed and a nutrient diffusion limitation would be present based on previous experience with this model system24. A novel, agarose assay was devised to quantify the amount of scaffold degradation achieved with this enzymatic technique. Finally, based on the results of the engineered cartilage study, Study 3 was devised to evaluate the removal of glycosaminoglycans on the efficacy of the agarase enzyme. The results from these studies aim to shed light on the possible benefits of scaffold degradation in the engineering of articular cartilage. Method STUDY 1: AGARASE CHARACTERIZATION Molten agarose (2% weight/volume, Type VII, Sigma Aldrich, St. Louis, MO) was cast between two glass plates and cooled at room temperature for 20 min. Disks (B4.0  2.3 mm) were cored and either lyophilized and weighed or incubated under sterile conditions with 0, 10, 50, and 100 units/mL agarase (Sigma) in phosphate buffered saline (PBS, Sigma) (n ¼ 4 per group). After this digestion period, agarose disks were washed, lyophilized, and weighed to determine the amount of agarose digested. Frozen, sterile, devitalized cartilage explants from bovine shoulder joints were incubated with 50 units/mL agarase in PBS for 48 h at 37 C and 5% CO2 (n ¼ 4 per group). After incubation, the explants were washed, digested in proteinase K, and analyzed for biochemical content as described below to assess matrix degradation due to agarase.

STUDY 2: EFFECTS OF AGARASE ON ENGINEERED CARTILAGE CONSTRUCTS

Creation and culture of engineered constructs Articular bovine chondrocytes were isolated from calf carpometacarpal joints (n ¼ 6) via an 11 h digestion of full thickness cartilage slices with 390 units/mL collagenase (Type V, Sigma) in 7.5 mL/g tissue of high-glucose Dulbecco’s Modified Essential Medium (hgDMEM) supplemented with 5% fetal bovine serum (FBS), amino acids, buffers, and antibiotics6. Cells were resuspended and mixed with molten 4% type VII agarose (Sigma) in PBS at 40 C to yield a 2% agarose suspension with 30  106 chondrocytes/mL. This suspension was cast between two glass plates and cooled for 20 min. Disks were cored (B4.0  2.3 mm) and cultured at 37 C and 5% CO2 in 35 mL of chondrogenic media (hgDMEM, 1% ITSþ, 0.1 mM dexamethasone, 110 mg/mL sodium pyruvate, 50 mg/mL L-proline, 50 mg/mL ascorbate-2-phosphate, sodium bicarbonate, and antibiotics1). Media was changed bi-daily. For the first 14 days, 10 ng/mL of TGF-b3 (R&D Systems, Minneapolis, MN) was added with each media change1. Day 0 mechanical testing was performed prior to TGF-b3 treatment. On day 42, half of the constructs were removed and cultured in chondrogenic media with 100 units/mL agarase for 48 h. Constructs were then washed with media 3 and then cultured in chondrogenic media for the rest of the study.

1 Hz. Compressive Young’s modulus (EY) was determined from the equilibrium stress/strain response of the stress relaxation test. Dynamic modulus (G*) at 1 Hz was calculated from the stress/strain response during the dynamic test. Following mechanical testing, samples were halved, with each half either fixed for histology or frozen at 80 C for biochemistry.

Biochemical analysis The samples were thawed, weighed wet, lyophilized, reweighed dry, and digested for 16 h at 56 C with 1 mg/mL proteinase K (EMD Biosciences, San Diego CA) in 50 mM Tris buffered saline with ethylenediaminetetraacetic acid (EDTA), iodoacetamide, and pepstatin A (Sigma Aldrich)26. Using these digests, glycosaminoglycan (GAG) content was determined via the dimethylmethylene blue (DMMB) dye-binding assay27, DNA content via the PicoGreen assay (Invitrogen, Carlsbad, CA), and collagen content via the orthohydroxyproline (OHP) colorimetric assay28. Collagen content was calculated by assuming a 1:10 OHP-to-collagen mass ratio29. Assays were adapted for use in 96-well, micro-titer plates. GAG and collagen content was normalized to the construct wet weight (% ww). The type II collagen content of samples was determined from the proteinase K digests via a custom enzyme-linked immunosorbent assay (ELISA) described previously24,26. The type II collagen content was normalized to the total collagen content of the disks as determined by the OHP assay.

Agarose quantification A novel agarose quantification assay was devised to measure the galactose produced via enzymatic degradation of agarose with agarase (Sigma). Proteinase K digests were vortexed and centrifuged at 100g for 30 s to sediment undigested agarose while avoiding the pelleting of soluble matrix proteins. The supernatant was removed and frozen for storage and the pellet was washed and centrifuged 3 with PBS to remove any remaining matrix proteins. These pellets were then liquefied in PBS at 65 C and cooled to 43 C, at which 100 units/mL of agarase was added to each sample and incubated overnight. Then the galactose content of these samples was determined via the Amplex Red kit (Invitrogen). Values obtained from control and agarase-treated constructs were normalized to day 0 disks.

Histology Samples were fixed in acideethanoleformalin30 for 48 h at 4 C, dehydrated, cleared, embedded in Tissue Prep embedding media (Fisher Scientific, Pittsburgh, PA), and sectioned at 6 mm. Sections were then stained in Safranin O, and Picrosirius Red to study cell proliferation/distribution, proteoglycans, and collagen.

Statistical analysis Statistics were performed using the Statistica (Statsoft, Inc., Tulsa, OK) software package. Groups were examined using multivariate analysis of variance with EY, GAG, collagen, DNA, agarose, and type II collagen as the dependent variables and culture time and agarase treatment as the independent variables. Fisher’s least-significant different (LSD) post hoc tests were carried out with a ¼ 0.05, n ¼ 4e5 per group.

STUDY 3: EFFECTS OF PROTEOGLYCAN REMOVAL ON AGAROSE DIGESTION Day 91 engineered constructs from Study 2 were incubated for 5 h at 37 C in either chondrogenic media (untreated control) or in sterile 2.0 M CaCl2 solution (treated)31e34 to extract proteoglycans (PGs). This CaCl2 technique was previously found to extract >60% of total proteoglycans in engineered cartilage tissue (not shown). After incubation, disks were washed 3 with media for 5 min under gentle agitation. Treated and untreated constructs were then incubated with control media or 100 units/mL agarase as described above. This resulted in four groups (n ¼ 4 per group): control, agarase only, CaCl2-only, both treatments (CaCl2eagarase). These constructs were analyzed for biochemical and agarose content was evaluated as described above. Constructs were prepared for imaging using scanning electron microscopy. Samples were halved and fixed for 24 h at 4 C in an electron microscopy (EM) fixative (0.5% gluteraldehyde, 2% paraformaldehyde, 0.05 M cacodylate buffer; pH 7.2). Then samples were washed in distilled water and viewed in a FEI Quanta environmental scanning electron microscope (ESEM) with beam strength of 10e15 kV at 500e1000 magnification.

Mechanical testing Constructs (n ¼ 4e5 per group) were removed from culture on day 0, 14, 28, 42, 44 (immediately post-agarase treatment), 63, 77, and 91 for mechanical testing. Testing was performed in unconfined compression as previously described25. Constructs were placed in a custom testing device, equilibrated under a creep tare load, then tested via a stress relaxation test (ramp velocity: 1 mm/s) to 10% strain (based on post-creep thickness). After equilibrium was reached, a sinusoidal displacement of 40 mm amplitude was applied at

Results STUDY 1: AGARASE CHARACTERIZATION

Increased agarase concentration progressively increased the degradation of the agarose hydrogel, with 100 units/mL

222

K. W. Ng et al.: Scaffold removal for engineered cartilage Articular Cartilage

Cell-free Agarose Disks n=4 per group; * p<0.05 vs. control

A 1.2

n=4 per group; p=0.99

B

9.0%

45%

8.0%

40%

7.0%

35%

6.0%

30%

0.6

0.4

Collagen %ww

*

0.8

GAG %ww

Normalized Dry Weight

1.0

5.0% 4.0% 3.0%

25% 20% 15%

2.0%

10%

1.0%

5%

0.2

0%

0.0%

0.0 Control

10u/mL

50u/mL 100u/mL

Control Agarase

Control Agarase

Fig. 1. Results of Study 1 e Agarase characterization: Agarase enzyme was found to significantly reduce the dry weights of cell-free agarose disks, with increasing degradation found with increasing agarase concentration (A). The agarase enzyme was found to have no significant effects on the matrix composition of frozen cartilage explants (B).

STUDY 2: EFFECTS OF AGARASE ON TISSUE ENGINEERED CARTILAGE

Agarase treatment removed approximately w45% of the agarose in the constructs, with 55  3% of the agarose remaining (Fig. 2). Agarose content in both controls and in agarase-treated tissues was found to be unaffected with extended culture time, consistent with previous findings10. Control construct EY and G* increased over time to w600 kPa and w2.75 MPa, respectively, but plateaued after day 63 (Fig. 3). GAG content and collagen content normalized to wet weight reached values of w6.5% ww and w2.5% ww, with no significant increases in GAG after day 63 and no significant increases in collagen after day 42 (Fig. 4). Dry weight increased from 2.4  1.0 to 9.6  2.9 mg after 91 days in culture. DNA content was initially 4154  57 ng/construct and increased w2 by day 28 (8643  178 ng/construct, P < 0.05), with no significant changes observed after day 28. Agarase treatment did not result in construct disintegration, as constructs appeared to maintain their shape and could be easily manipulated. A significant decrease in EY and G* was observed immediately post-treatment on day 44, with additional decreases observed on day 63 (Fig. 3). A similar trend was observed in GAG content [Fig. 4(A)], with GAG release to media measured at a rate of w320 mg/construct/day during the 2 days incubation. Immediately after this incubation period, this rate diminished to w130 mg/construct/day and returned to control levels of w45 mg/construct/day after day 63. After agarase treatment on day 44, constructs were significantly smaller than control tissue (Fig. 5, P < 0.05) and possessed significantly lower dry weights vs. controls (4.9  0.3 vs 7.3  1.2 mg,

P < 0.05). After 91 days in culture, agarase-treated tissue dry weights recovered to 7.7  1.5 mg, with no significant differences in size or weight to control constructs at this time point. No DNA or collagen release to media was detected after agarase treatment. However, DNA content significantly increased by an additional w25% with agarase treatment from day 44 to 77 and plateaued by day 91 (day 91 value: 10,584  231 ng/construct, P < 0.05 vs control). Both GAG [Fig. 4(A)] and EY [Fig. 3(A)] of treated constructs recovered to control values (P ¼ 0.9) by day 91 and G*

1.4

*p<0.05 vs. d0 & control

Control Agarase

1.2

Normalized Agarose Content

found to completely digest the agarose during the 48 h incubation period [Fig. 1(A)]. Cartilage disks were found to be unaffected by agarase treatment, with a slight but insignificant decrease noted in GAG content [Fig. 1(B)]. Biochemical values obtained for cartilage [GAG w6.5% ww, collagen w22% ww, Fig. 1(B)] agreed well with reported values in literature35.

1.0

0.8

0.6

*

* 0.4

0.2

0 0

45

91

Days Fig. 2. Agarose quantification of engineered cartilage in Study 2: Agarose quantification found that agarase treatment reduced agarose content to 55  3% of control values. Agarose content was unaffected with time in culture, confirming previous findings by Mauck et al10.

223

Osteoarthritis and Cartilage Vol. 17, No. 2

** ***

900 800 700 600 500 400 300 200 100 0

B

* ‡

Control

***

Agarase

5

** *** G* 1Hz (MPa)

Young's Modulus (kPa)

A 1000

‡

† †

** ***

4

*

** ***

Control

‡

Agarase

***

3 2

†‡

†

1 0

0

14

28

42

44

63

77

0

91

14

28

Days

42

44

63

77

91

Days

Fig. 3. Young’s modulus and dynamic modulus (G* 1 Hz) of control and agarase-treated constructs in Study 2. Control constructs (white) increased in compressive Young’s modulus (A) and compressive dynamic modulus measured at 1 Hz (G* 1 Hz, B) over time in culture. After addition of agarase on day 42, tissue mechanical properties initially decreased post-agarase treatment when measured on day 44, but recovered to control levels after additional culture time (A, black bars). The G* 1 Hz of agarase-treated tissue recovered to control levels on day 77 and surpassed control tissues by day 91 (B, black bars). *P < 0.05 vs d0, **P < 0.05 vs d14, ***P < 0.05 vs d28, 6P < 0.05 vs d42, yP < 0.05 vs respective control, zP < 0.05 vs previous time point within same group.

GAG (%ww)

A 10%

** ***

9% 8% 7% 6% 5% 4% 3% 2% 1% 0%

* ***

Control

** *** ‡

Agarase

†

Histology of agarase-treated samples on day 44 (Fig. 6) found qualitative GAG loss throughout the constructs, consistent with the quantitative results. The greatest GAG loss, indicated by the absence of the Safranin O dye, was at the periphery of constructs. Picrosirius Red showed a qualitative increase in the collagen content of agarase-treated samples.

STUDY 3: EFFECTS OF PG REMOVAL ON AGAROSE DIGESTION

CaCl2 extraction significantly removed w80% of the total proteoglycans as measured by the DMMB assay (control: treated: 1126  168 mg GAG/construct vs CaCl2 195  14 mg GAG/construct, P < 0.05). Agarase treatment alone was also found to comparably reduce GAG content (agarase only: 147  38 mg GAG/construct, P < 0.05 vs control). Both treatments (CaCl2eagarase) resulted in a >95% reduction in GAG, with these constructs possessing only 27  4 mg GAG/construct (P < 0.05 vs all groups). Agarose and collagen content normalized to control tissue

B Collagen (%ww)

[Fig. 3(B)] was significantly greater than that of control (3.94  0.20 vs 2.77  0.12 MPa, P < 0.05). Collagen content normalized to wet weight [Fig. 4(B)] did not significantly increase after agarase treatment (day 44), but did significantly increase compared to controls on days 77 (3.46  0.14 vs 2.44  0.14% ww, P < 0.05) and 91 (3.90  0.72 vs 2.45  0.18% ww, P < 0.05). Collagen content normalized to dry weight significantly increased after agarase treatment compared to controls (day 44: 20.78  3.76 vs 13.17  3.27% dw, P < 0.05), which is partially due to GAG and agarose loss decreasing the overall mass of agarase-treated constructs. However, the significant difference in collagen/dry weight between agarase and control groups continued through the entire study (day 91: 26.70  3.69 vs 16.60  2.82% dw, P < 0.05). Collagen content normalized to DNA (ng/ng) showed slight, but significant increases between agarase and control groups only on day 77 (74.01  3.63 vs 62.39  4.77, P < 0.05) and day 91 (75.13  4.06 vs 64.84  1.92, P < 0.05). Type II collagen, as determined via ELISA, represented w68% of the total collagen and did not change with time in culture or with agarase treatment.

*

6% 5%

** ***

†

Control

4%

Agarase

†‡

3% 2% 1% 0%

0

14

28

42

44

Days

63

77

91

0

14

28

42

44

63

77

91

Days

Fig. 4. GAG content and collagen content of control and agarase-treated constructs in Study 2. GAG content (A) and collagen content (B) of control tissues (white bars) increased early in culture, but reached a plateau by day 42. Agarase-treated constructs (black bars) exhibited lower GAG content immediately after treatment (day 44) that continued until day 63 before recovering to control values. Collagen content was not affected by agarase treatment and continued to increase with time in culture, surpassing control values on day 77 and day 91. *P < 0.05 vs d0, **P < 0.05 vs d14, ***P < 0.05 vs d28, 6P < 0.05 vs d42, yP < 0.05 vs respective control, zP < 0.05 vs previous time point within same group.

224

K. W. Ng et al.: Scaffold removal for engineered cartilage

Discussion

Day 44 Gross Morphology

Diameter: 4.85+/-0.05 mm

4.55+/-0.14 mm

Control

Thickness: 3.02+/-0.11 mm

Agarase

2.70+/-0.04 mm

Fig. 5. Day 44 gross morphology of control and agarase-treated constructs from Study 2. Immediately after agarase treatment, treated constructs possessed smaller diameters and thicknesses than untreated controls. This trend continued through time in culture.

was found to be unaffected by CaCl2 extraction (93  14% and 95  11% of control values, respectively). Agarase treatment significantly reduced the measured agarose content, but did not fully remove the agarose scaffold, consistent with results for Study 2 (58  7% agarose remaining, P < 0.05 vs control). Combined CaCl2 and agarase treatment further lowered agarose content below control and agarase alone values (27  12%, P < 0.05 vs all groups). All samples contracted under ESEM viewing due to the eventual drying of the samples. The control tissue exhibited relatively smooth surfaces with small bumps that may be the embedded chondrocytes [Fig. 7(A)]. Proteoglycan extraction led to the formation of large holes in the construct, with fibers appearing to span these holes [Fig. 7(B)]. Agarase treatment led to much larger contraction of construct with drying while viewed under ESEM and possessed a bumpy texture [Fig. 7(C)]. Combined treatments revealed a contracted specimen with a more fibrous appearance compared to agarase alone and/or control samples [Fig. 7(D)].

The primary purpose for agarose scaffold removal was to increase the engineered cartilage collagen content and mechanical properties. Constructs were cultured for 42 days and possessed a physiologic Young’s modulus (EY) and GAG content prior to agarase treatment at this time point. After 91 days in culture (49 additional days post-agarase treatment), agarase treatment led to a w60% increase in collagen content and a w40% increase in the dynamic modulus, affirming the proposed hypothesis. The rationale for this approach was chosen so that the engineered tissue could first utilize the agarose scaffold for retaining matrix products and then rely on the de novo matrix for structural support upon digestion of the agarose scaffold. Agarase scaffold degradation did not lead to construct disintegration, cell loss, or collagen loss, validating this approach. Though agarose was utilized for its known benefits in maintaining chondrocyte phenotype, the encouraging results from the current studies are relevant to other scaffolds where degradation can also be controlled (e.g., alginate11 and polyethylene glycol [PEG] hydrogels36). From the results, the compressive properties of the agarose scaffold, which do not degrade or change over time10, can contribute only minutely to the overall mechanical properties of the engineered cartilage tissue after physiologic matrix formation [i.e., w10 kPa vs w600 kPa, Fig. 3(A)]. Therefore, the changes in the construct mechanical properties with agarase treatment represent an interplay between the functional role of GAG and collagen and the tissue’s hydraulic permeability. Agarase treatment led to GAG loss that led to a lower compressive EY, as expected based on the functional mechanical role of GAG molecules37. The values for both retained GAG and EY, however, were still significantly greater than day 0 values; this retention may be attributed to the entrapment of the larger proteoglycans in the remaining agarose and newly synthesized collagen matrix, as well as potential binding of proteoglycan molecules to the collagen matrix. Based on theoretical models of cartilage matrix38e40, this decrease in GAG content should result in a concomitant increase in hydraulic permeability due to changes in steric and electrochemical interactions. This change in the permeability is supported by the

Fig. 6. Safranin O and Picrosirius Red histology for Study 2 constructs. Agarase treatment was found to lead to GAG loss throughout the construct, as indicated by decreased staining by Safranin O (left) compared to control tissues, especially at the construct periphery. Collagen appeared to increase in density, as indicated by the increases in Picrosirius Red staining (right). These results appear to correlate well with the quantitative data in Fig. 4.

Osteoarthritis and Cartilage Vol. 17, No. 2

225

Fig. 7. ESEM images for Study 3 constructs. Under environmental scanning electron microscopy (ESEM), all constructs exhibited tissue contraction due to water evaporation over time during scanning. Untreated control tissues exhibited relatively smooth surfaces with small bumps (arrows) that may be chondrocytes (A). Proteoglycan extraction via CaCl2 led to the formation of large holes with fibers spanning the holes (B). Constructs after agarase treatment exhibited greater contraction with drying during ESEM viewing and possessed a very bumpy texture (C). Combined treatments revealed a contracted specimen with a more fibrous appearance compared to agarase alone and/or control samples (D). Magnification ¼ 1000, scale bar ¼ 100 mm.

histological findings and would improve nutrient/waste transport, partially explaining the increases in matrix synthesis following agarase treatment41,42. An increase in hydraulic permeability would explain the decrease in dynamic modulus post-agarase treatment as the dynamic compressive properties of hydrated cartilage matrix depend on tensile stiffness (imparted by collagen) and interstitial fluid pressurization, which is inversely related to permeability43,44. With GAG content and permeability restored to control levels with time in culture, the higher dynamic modulus of agarase-treated constructs over controls is likely due to the increased collagen content, based on its known functional role4,5. To validate these proposed mechanisms, further experiments are needed to directly measure the permeability of engineered constructs pre/post-agarase treatment along with the effects of GAG loss without agarose loss (e.g., via chondroitinase-ABC). The removal of the scaffold did not result in any apparent loss of the de novo collagen matrix or construct disintegration, indicating the presence of an interconnected collagen meshwork by 42 days. With the removal of the agarose scaffold, the tissue engineered construct would be mostly supported by its collagen matrix. The slight but significant

contraction of the construct, noted by the decrease in size and the apparent increase of collagen density observed with histology, is likely due to the reduction in osmotic swelling pressure from the loss of GAG molecules45e47 and not to a collapse of the matrix structure. Indeed, agarase led to improved matrix accumulation and tissue properties that reflect favorable changes in chondrocyte microenvironment post-treatment. Osmotic swelling pressure generated by proteoglycans appears to be one of the major mechanisms behind the size increase of chondrocyte-laden agarose constructs over time in culture and the stresses generated can eventually lead to construct cracking48. After agarase treatment, the GAG molecules that remain in the construct are likely those that have bonded with the collagen meshwork of the tissue; thus it is possible that the post-agarase synthesized GAG molecules that are retained are likely to be better aggregated and integrated with the collagen network than those in the control tissue. The mechanism of proteoglycanecollagen interaction may be related to other cartilage proteins such as COMP49 and this is worth future verification. The increased collagen content in engineered cartilage after agarase treatment can be explained by the

226 combination of increased chondrocyte number and increased collagen accumulation per cell. The mechanisms behind these changes are likely a combination of physical and biological factors: The removal of the agarose scaffold and unbound GAG molecules may have lead to increased permeability (see above) and thereby improve nutrient/ waste transport allowing for increased matrix synthesis. In addition, the loss of the agarose scaffold in the construct would increase the space in which newly synthesized matrix could occupy. The mechanism behind the cell proliferation observed from day 0 to 28 is likely due to the collagenase digestion protocol, which has been found to elicit transient proliferation in explant culture50. After agarase treatment, the disruption of the extracellular matrix and chondrocyte environment by the loss of scaffold and GAG molecules may have triggered the second stage of chondrocyte proliferation, consistent with findings in explants50,51. As matrix content was restored, this transient cell proliferation was observed to cease. The use of type II collagen ELISA confirmed that the chondrocytes appeared to possess their proper phenotype as evidenced by a physiological type II collagen fraction. The increase in collagen synthesis may be triggered by the GAG loss occurring with agarase treatment. In explant cultures, Asanbaeva et al. have found that chondroitinase-ABC treatment of cartilage explants improved the collagen content, GAG synthesis (via sulfate incorporation), and tensile mechanical properties over time in culture52. They proposed that these changes were possibly due to chondrocyte response to changes in collagen pre-stress and/or organization with enzyme treatment, but the exact mechanisms remain unclear. Recent work on chondrocyte-seeded agarose constructs53 and scaffold-free constructs54 has found similar changes with GAG depletion. A novel quantification technique was devised to determine the agarose content in the constructs post-agarase treatment. The difficulty arose from the lack of existing methodologies to quantify agarose along with the need to separate the agarose from the matrix proteins after proteinase K digestion. It was observed that the enzymatic digestion to isolate matrix molecules did not appear to have any effect on the agarose scaffold consistent with the known effects of proteinase K to digest proteins (product literature, EMD Biosciences). Therefore, the undigested matter was pelleted, digested with agarase, and then the resulting galactose released was determined as a means to measure the agarose content. This agarose characterization of engineered cartilage constructs treated with agarase found an incomplete digestion that was unlike the complete digestion found with cell-free disks. As the penetration of agarase in to the construct relies on diffusion, there is likely a greater effect of agarase on the construct periphery. Incomplete penetration of the enzyme may not have greatly affected the tissue properties as the peripheral region exhibits the greatest matrix development24 and would be the region most limiting to nutrient diffusion. In addition, as GAG molecules have been shown to impair tissue integration55, the removal of agarose and GAG molecules at the construct edges may improve construct integration with surrounding host tissue. The proteoglycan extraction via calcium chloride further decreased the agarose content with agarase digestion. However, though the agarose digestion was improved with proteoglycan extraction, the exact mechanism behind the decreased efficacy of agarase remains unclear. One possibility is that the galactose bonds present in GAG molecules56e58 may interfere with the action of agarase. The

K. W. Ng et al.: Scaffold removal for engineered cartilage efficacy may also be tied in with permeability, restricting diffusion of the enzyme to the inside of the engineered tissues. This mechanism is possible as the calcium chloride extraction appeared to increase the pores in the tissue (Fig. 7) and improve the agarose digestion. This indicates that complete removal of the agarose may require additional enzymatic treatments to remove GAG molecules. Future experiments need to determine the exact reason behind the observed diminishment of the agarase enzyme efficiency on the agarose in the engineered cartilage constructs. Another issue to be addressed with future research is the batch-to-batch variations that may arise in the production of the agarase enzyme. The results of this study highlight the importance of controlled scaffold degradation in the formation of engineered cartilage. Controlled degradation of the agarose scaffold employed in the current studies was achieved via the enzyme agarase. Agarase treatment was found to remove scaffold material and GAG molecules from the engineered construct with no detrimental effects on chondrogenic activity. In addition, agarase-treated constructs possessed increased collagen content and dynamic mechanical properties relative to control over time in culture. This demonstrates an important benefit of controlled scaffold degradation in cartilage tissue engineering, and may be explained by increased nutrient transport, increased space for collagen fibril development, and stimulatory effects of GAG loss. Future work will investigate the feasibility of combining established bioreactor culture with scaffolds (e.g., applied dynamic loading10, prescribed mechanical inhomogeneity20) along with degradation to improve tissue properties and reduce scaffold content prior to clinical implantation. Such an engineered tissue would combine the advantages of current scaffold and scaffold-free approaches and be free of possible biocompatibility issues that may arise with host blood/tissue interactions.

Conflict of interest All authors were supported from the following sources: NIH RO1 AR46568 (CTH). There are no conflicts of interest or sources of bias in the research contained in this manuscript.

Acknowledgements This study was supported with funding from the National Institute of Arthritis and Musculoskeletal and Skin Diseases (AR 46568). References 1. Byers BA, Mauck RL, Chiang IE, Tuan RS. Transient exposure to transforming growth factor beta 3 under serum-free conditions enhances the biomechanical and biochemical maturation of tissue-engineered cartilage. Tissue Eng Part A. e-publication 2008; Jul 8. 2. Lima EG, Bian L, Ng KW, Mauck RL, Byers BA, Tuan RS, et al. The beneficial effect of delayed compressive loading on tissue-engineered cartilage constructs cultured with TGF-beta3. Osteoarthr Cartil 2007; 15:1025e33. 3. Mauck RL, Hung CT, Ateshian GA. Modeling of neutral solute transport in a dynamically loaded porous permeable gel: implications for articular cartilage biosynthesis and tissue engineering. J Biomech Eng 2003;125:602e14. 4. Williamson AK, Masuda K, Thonar EJ, Sah RL. Growth of immature articular cartilage in vitro: correlated variation in tensile biomechanical and collagen network properties. Tissue Eng 2003;9:625e34. 5. Wong M, Ponticiello M, Kovanen V, Jurvelin JS. Volumetric changes of articular cartilage during stress relaxation in unconfined compression. J Biomech 2000;33:1049e54.

227

Osteoarthritis and Cartilage Vol. 17, No. 2 6. Ng KW, Defrancis JG, Kugler LE, Kelly TA, Ho MM, O’Conor CJ, et al. Amino acids supply in culture media is not a limiting factor in the matrix synthesis of engineered cartilage tissue. Amino Acids 2008;35:433e8. 7. Seidel JO, Pei M, Gray ML, Langer R, Freed LE, Vunjak-Novakovic G. Long-term culture of tissue engineered cartilage in a perfused chamber with mechanical stimulation. Biorheology 2004;41:445e58. 8. Langer R, Vacanti JP. Tissue engineering. Science 1993;260:920e6. 9. Freed LE, Vunjak-Novakovic G, Biron RJ, Eagles DB, Lesnoy DC, Barlow SK, et al. Biodegradable polymer scaffolds for tissue engineering. Biotechnology (NY) 1994;12:689e93. 10. Mauck RL, Soltz MA, Wang CC, Wong DD, Chao PH, Valhmu WB, et al. Functional tissue engineering of articular cartilage through dynamic loading of chondrocyte-seeded agarose gels. J Biomech Eng 2000; 122:252e60. 11. Augst AD, Kong HJ, Mooney DJ. Alginate hydrogels as biomaterials. Macromol Biosci 2006;6:623e33. 12. Benya PD, Shaffer JD. Dedifferentiated chondrocytes reexpress the differentiated collagen phenotype when cultured in agarose gels. Cell 1982;30:215e24. 13. Buschmann MD, Gluzband YA, Grodzinsky AJ, Kimura JH, Hunziker EB. Chondrocytes in agarose culture synthesize a mechanically functional extracellular matrix. J Orthop Res 1992;10:745e58. 14. Lee DA, Knight MM. Mechanical loading of chondrocytes embedded in 3D constructs: in vitro methods for assessment of morphological and metabolic response to compressive strain. Methods Mol Med 2004; 100:307e24. 15. Mouw JK, Case ND, Guldberg RE, Plaas AH, Levenston ME. Variations in matrix composition and GAG fine structure among scaffolds for cartilage tissue engineering. Osteoarthr Cartil 2005;13:828e36. 16. Gu WY, Yao H, Huang CY, Cheung HS. New insight into deformationdependent hydraulic permeability of gels and cartilage, and dynamic behavior of agarose gels in confined compression. J Biomech 2003; 36:593e8. 17. Johnson EM, Berk DA, Jain RK, Deen WM. Hindered diffusion in agarose gels: test of effective medium model. Biophys J 1996;70: 1017e23. 18. Maaloum M, Pernodet N, Tinland B. Agarose gel structure using atomic force microscopy: gel concentration and ionic strength effects. Electrophoresis 1998;19:1606e10. 19. Ng KW, Mauck RL, Statman LY, Lin EY, Ateshian GA, Hung CT. Dynamic deformational loading results in selective application of mechanical stimulation in a layered, tissue-engineered cartilage construct. Biorheology 2006;43:497e507. 20. Ng KW, Wang CC, Mauck RL, Kelly TA, Chahine NO, Costa KD, et al. A layered agarose approach to fabricate depth-dependent inhomogeneity in chondrocyte-seeded constructs. J Orthop Res 2005;23:134e41. 21. Carlsson J, Malmqvist M. Effects of bacterial agarase on agarose gel in cell culture. In Vitro 1977;13:417e22. 22. Malmqvist M. Purification and characterization of two different agarosedegrading enzymes. Biochim Biophys Acta 1978;537:31e43. 23. Morrice LM, McLean MW, Long WF, Williamson FB. Beta-agarases I and II from Pseudomonas atlantica. Substrate specificities. Eur J Biochem 1983;137:149e54. 24. Kelly TA, Ng KW, Wang CC, Ateshian GA, Hung CT. Spatial and temporal development of chondrocyte-seeded agarose constructs in free-swelling and dynamically loaded cultures. J Biomech 2006;39: 1489e97. 25. Mauck RL, Wang CC, Oswald ES, Ateshian GA, Hung CT. The role of cell seeding density and nutrient supply for articular cartilage tissue engineering with deformational loading. Osteoarthr Cartil 2003;11:879e90. 26. Riesle J, Hollander AP, Langer R, Freed LE, Vunjak-Novakovic G. Collagen in tissue-engineered cartilage: types, structure, and crosslinks. J Cell Biochem 1998;71:313e27. 27. Farndale RW, Sayers CA, Barrett AJ. A direct spectrophotometric microassay for sulfated glycosaminoglycans in cartilage cultures. Connect Tissue Res 1982;9:247e8. 28. Stegemann H, Stalder K. Determination of hydroxyproline. Clin Chim Acta 1967;18:267e73. 29. Hollander AP, Heathfield TF, Webber C, Iwata Y, Bourne R, Rorabeck C, et al. Increased damage to type II collagen in osteoarthritic articular cartilage detected by a new immunoassay. J Clin Invest 1994;93:1722e32. 30. Lin W, Shuster S, Maibach HI, Stern R. Patterns of hyaluronan staining are modified by fixation techniques. J Histochem Cytochem 1997;45: 1157e63. 31. Anderson HC, Sajdera SW. The fine structure of bovine nasal cartilage. Extraction as a technique to study proteoglycans and collagen in cartilage matrix. J Cell Biol 1971;49:650e63. 32. Glaser C, Putz R. Functional anatomy of articular cartilage under compressive loading quantitative aspects of global, local and zonal

33. 34. 35.

36. 37. 38. 39. 40. 41. 42.

43. 44. 45.

46. 47. 48. 49. 50. 51. 52. 53. 54. 55. 56. 57. 58.

reactions of the collagenous network with respect to the surface integrity. Osteoarthr Cartil 2002;10:83e99. Rosenberg LC, Pal S, Beale RJ. Proteoglycans from bovine proximal humeral articular cartilage. J Biol Chem 1973;248:3681e90. Sajdera SW, Hascall VC. Proteinpolysaccharide complex from bovine nasal cartilage. A comparison of low and high shear extraction procedures. J Biol Chem 1969;244:77e87. Mankin HJ, Mow VC, Buckwalter JA, Iannotti JP, Ratcliffe A. Articular cartilage structure, composition, and function. In: Buckwalter JA, Einhorn TA, Simon SR, Eds. Orthopaedic Basic Science. Biology and Biomechanics of the Musculoskeletal System. Rosemont: American Academy of Orthopaedic Surgeons 2000:443e70. Rice MA, Anseth KS. Controlling cartilaginous matrix evolution in hydrogels with degradation triggered by exogenous addition of an enzyme. Tissue Eng 2007;13:683e91. Maroudas A, Bayliss MT, Venn MF. Further studies on the composition of human femoral head cartilage. Ann Rheum Dis 1980;39:514e23. DiSilvestro MR, Suh JK. Biphasic poroviscoelastic characteristics of proteoglycan-depleted articular cartilage: simulation of degeneration. Ann Biomed Eng 2002;30:792e800. Gu WY, Yao H. Effects of hydration and fixed charge density on fluid transport in charged hydrated soft tissues. Ann Biomed Eng 2003; 31:1162e70. Stahurski TM, Armstrong CG, Mow VC. Variation of the intrinsic aggregate modulus and permeability of articular cartilage with trypsin digestion. Biomech Symp Trans ASME AMD 1981;43:137e40. Heywood HK, Sembi PK, Lee DA, Bader DL. Cellular utilization determines viability and matrix distribution profiles in chondrocyteseeded alginate constructs. Tissue Eng 2004;10:1467e79. Heywood HK, Bader DL, Lee DA. Glucose concentration and medium volume influence cell viability and glycosaminoglycan synthesis in chondrocyte-seeded alginate constructs. Tissue Eng 2006;12: 3487e96. Soltz MA, Ateshian GA. Interstitial fluid pressurization during confined compression cyclical loading of articular cartilage. Ann Biomed Eng 2000;28:150e9. Lai WM, Mow VC, Roth V. Effects of nonlinear strain-dependent permeability and rate of compression on the stress behavior of articular cartilage. J Biomech Eng 1981;103:61e6. Basser PJ, Schneiderman R, Bank RA, Wachtel E, Maroudas A. Mechanical properties of the collagen network in human articular cartilage as measured by osmotic stress technique. Arch Biochem Biophys 1998;351:207e19. Chahine NO, Chen FH, Hung CT, Ateshian GA. Direct measurement of osmotic pressure of glycosaminoglycan solutions by membrane osmometry at room temperature. Biophys J 2005;89:1543e50. Lai WM, Hou JS, Mow VC. A triphasic theory for the swelling and deformation behaviors of articular cartilage. J Biomech Eng 1991;113: 245e58. Bian L, Ng KW, Lima EG, Ateshian GA, Hung CT. Influence of nutrient channels on the development of engineered cartilage properties. Proc ASME BIO2007 2007;176234. Chen FH, Herndon ME, Patel N, Hecht JT, Tuan RS, Lawler J. Interaction of cartilage oligomeric matrix protein/thrombospondin 5 with aggrecan. J Biol Chem 2007. Lee DA, Bentley G, Archer CW. The control of cell division in articular chondrocytes. Osteoarthr Cartil 1993;1:137e46. Lee DA, Bentley G, Archer CW. Proteoglycan depletion alone is not sufficient to stimulate proteoglycan synthesis in cultured bovine cartilage explants. Osteoarthr Cartil 1994;2:175e85. Asanbaeva A, Masuda K, Thonar EJ, Klisch SM, Sah RL. Mechanisms of cartilage growth: modulation of balance between proteoglycan and collagen in vitro using chondroitinase ABC. Arthr Rheum 2007;56:188e98. Bian L, Williams DY, Mao DQ, Xu D, Ateshian GA, Hung CT. Influence of temporary chondroitinase ABC-induced GAG suppression on maturation of tissue engineered cartilage. Trans Orthop Res 2008;33:601. Natoli RM, Revell CM, Athanasiou KA. Chondroitinase-ABC treatment results in increased tensile properties in a scaffold-less, serum-free model of articular cartilage tissue engineering. Trans Orthop Res 2008;33:329. Obradovic B, Martin I, Padera RF, Treppo S, Freed LE, VunjakNovakovic G. Integration of engineered cartilage. J Orthop Res 2001;19:1089e97. Helting T, Roden L. Biosynthesis of chondroitin sulfate. I. Galactosyl transfer in the formation of the carbohydrateeprotein linkage region. J Biol Chem 1969;244:2790e8. Schmidt M, Dmochowski A, Wierzbowska B. Galactose and xylose as structural components of vertebrate chondroitin sulfateeprotein complexes. Biochim Biophys Acta 1966;117:258e61. Nieduszynski IA, Huckerby TN, Dickenson JM, Brown GM, Tai GH, Morris HG, et al. There are two major types of skeletal keratan sulphates. Biochem J 1990;271:243e5.