Type II collagen and glycosaminoglycan expression induction in primary human chondrocyte by TGF-β1

Human cartilage samples were obtained at the time of surgery for excision of a polydactyl finger. The patient was one year old. Cartilage tissues were obtained from the rudimentary excised finger, which was donated for further experiments by parents with written consent. The study was performed in compliance with the ethical guidelines of the Helsinki Declaration, and was approved by the institutional review board at the Inha University Medical School and Hospital.

Excised cartilage tissues were rinsed with sterile phosphate buffered saline (PBS) and treated with collagenase (from Clostridium histolyticum; GIBCO-BRL/Invitrogen, Rockville, MD) in Ham’s F12 medium containing 10 % fetal bovine serum (FBS). After collagenase treatment, the cartilage tissues were strained to collect chondrocyte cells. Chondrocytes were washed with PBS and cultured in Ham’s F12 medium containing 10 % FBS [ 19].

The primary chondrocytes (hChonJ) and the genetically modified clone, hChonJb#7 were expanded with Dulbecco’s modified Eagle’s medium (DMEM) containing 10 % FBS and 1 % L-Glutamine in a 37 °C, 5 % CO ₂ atmosphere.

For micro-mass culture, 3 × 10 ⁵ hChonJ cells (passage 8) or a mixture of 3 × 10 ⁵ hChonJ cells and 1 × 10 ⁵ irradiated hChonJb#7 cells (passage 18) in 20 μl of chondrogenic medium were seeded in each well of type I collagen coated 24 well plates. After cells were allowed to adhere at 37 °C for 1.5 h, 1 ml of chondrogenic medium was carefully added to each well. When required, recombinant human TGF-β1 (25 ng/ mass if not designated, eBioscience, San Diego, CA, USA), purified human type II collagen (25 ng/ mass, EMD Millipore, Temecula, CA, USA), or Proteoglycans (25 ng/ mass ProteaImmun, Berlin, Germany) was added into the seeding cell mixture. Culture medium was changed every 3–4 days with supplemented chondrogenic medium. For chondrogenic medium preparation, 0.1 mM dexamethasone, 0.17 mM ascorbic acid, 0.35 mM proline, 1 mM Sodium pyruvate, and Insulin-transferrin-salenous acid (1×) were supplemented into serum free high glucose DMEM. Micro-masses were harvested on day 7 to use for tests.

For culture and expansion of COS-1 monkey kidney cells (ATCC CRL-1650) and Hep G2 Hepatocellular Carcinoma (ATCC HB-8065), ATCC guidance was followed using 10 % FBS containing DMEM and Eagle’s Minimum Essential Medium (EMEM) respectively.

To investigate of the biological activity of TGF-β1 produced from hChonJb#7, we made four constructs as follows: wild-type TGF-β1, TGF-β1 ^S223/225 containing a serine at position 223 and 225 of the TGF-β1 precursor in place of the wild-type cysteine residue, TGF-β1 ^A84/W255 containing a alanine and tryptophan at position 84 and 255 of the TGF-β1 precursor in place of the wild-type threonine and arginine residue, respectively, and TGF-β1 ^hChonJb#7 containing all of four residue mutations, which is produced from hChonJb#7. To construct vector containing wild-type TGF-β1 and mutants, the coding region of wild-type TGF-β1 and TGF-β1 ^hChonJb#7 was amplified by PCR using cDNA synthesized by the RNA of normal chondrocyte and pKEB1 plasmid which was used to manufacture hChonJb#7 cell line and then inserted into the Xho I and EcoR I sites of pMSCV retroviral vector (Clontech, CA, USA). Afterwards, TGF-β1 ^S223/225 and TGF-β1 ^A84/W255 were respectively constructed from wild-type TGF-β1 and TGF-β1 ^hChonJb#7 constructs using a point-mutagenesis kit according to the manufacturer’s protocol (Agilent Technologies, CA, USA).

To confirm the expression of TGF-β1 constructs, culture supernatant was mixed with reducing or non-reducing (with or without β-mercaptoethanol, respectively) Tris-Glycine sample buffer and separated by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE). Electrotransferred nitrocellulose membrane was probed with anti-TGF-β1 antibody (Cell Signaling Technology, Inc., Beverly, MA, USA).

The four TGF-β1 constructs transfected Cos-1 cell culture medium was used for TGF-β1 source to monitor proliferation inhibition of Hep G2 hepatocellular carcinoma. The concentration of TGF-β1 in the Cos-1 culture medium was measured by Enzyme-linked Immunosorbent Assay (ELISA) after acid activation and a designated amount of TGF-β1, which was expected to be if activated, was applied into Hep G2 cell culture (seeded at a cell density of 5000 cells/well of 96 μ-well plate). After 1 week incubation, the levels of viable cells were measured using Cell Counting Kit-8 (CCK-8) assay (Dojindo Molecular Technologies, Inc., Rockville, MD, USA).

Culture-expanded hChonJ (passage 8) cells were thawed at 37 °C and washed in phosphate buffered saline (PBS) and then resuspended in a chondrogenic medium. Aliquots of 2.0 × 10 ⁵ cells, suspended in 1 mL chondrogenic medium supplemented with 25 ng/mL of recombinant human TGF-β1 (Peprotech, NJ, USA), were centrifuged at 450 × g for 5 min in 15 mL polypropylene conical tubes. Pelleted cells were incubated at 37 °C, 5 % CO ₂ atmosphere with loosened caps to permit gas exchange. The medium was changed every 3–4 days and pellets were harvested on days 21. Safranin-O staining and immunohistochemical staining of type II collagen were performed. This experiment was performed four times, making 3 pellets each time.

At 21 days, pellets were embedded in optimum cutting temperature (OCT) compound and transferred the frozen pellet blocks to −20 °C of a Cryotome™ E cryostat (Thermo Scientific, MA, USA). Blocks of the pellets were cut into 3 μm thick slices using a cryostat and mounted on glass slides. The sections were stained with Safranin-O (S-O) for proteoglycan and immunohistochemistry for type II collagen. For S-O staining, sections were incubated in distilled water for 10–15 min and 0.1 % S-O solution was applied for 5 min, and then washed with distilled water. For immunohistochemistry, we used the Histostatin-Plus Kit (Invitrogen, CA, USA). After sections were incubated in distilled water for 10–15 min, washed with PBS for 10 min, and then, we performed according to the manufacturer’s protocol. The primary antibody against human type II collagen (EMD Millipore) was used and incubated at 4 °C overnight.

The total soluble GAG of cultured micro-masses was measured using a Blyscan sulfated glycosaminoglycan assay kit (biocolor, Carrickfergus, County Antrim, UK). The micro-mass in each well was washed with PBS and then digested with papain extraction reagent for 3 h at 65 °C with 30 min interval mixing. The cell lysate from two wells was combined into one sample and transferred to a micro-centrifuge tube. Total 6 μ-masses were used as triplicated samples (2 μ-masses/ sample) for each experiment. Afterwards, the manufacturer’s protocol was followed.

Total RNA was extracted from 10 to 12 μ-masses using TRIzol reagent (GIBCO-BRL/ Invitrogen by Life Technologies, Grand Island, NY, USA). Reverse transcriptase-polymerase chain reaction (RT-PCR) was performed with a SuperScript II reverse transcriptase (Invitrogen by Life Technologies, Grand Island, NY, USA) and random primer mix (NEB, Ipswich, MA, USA). For qPCR, C1000 thermal cycler system with CFX96 real-time PCR detection systems (Bio-Rad, Hercules, CA, USA) and IQTM SYPR® Green supermix (Bio-Rad, Hercules, CA, USA) were used with the following primers: type II collagen (forward, 5'-CCCTGAGTGGAAGAGTGGAG-3', reverse, 5'-GAGGCGTGAGGTCTTCTGTG-3'), type I collagen (forward, 5'-CGATGGCTGCACGAGTCACAC-3', reverse, 5'-CAGGTTGGGATGGAGGGAGTTTAC-3'), GAPDH (forward, 5'- TCGACAGTCAGCCGCATCTTCTTT-3', reverse, 5'-ACCAAATCCGTTGACTCCGACCTT-3'). The relative mRNA levels were evaluated using the 2 ^ΔΔC(t) method.

The distal end of the femurs were removed and partial osteochondral defects (3 mm × 6 mm, 1 mm deep) were created in the trochlear groove of the femurs of adult rabbits. The cartilages with defects were cultured in serum-free DMEM (Gibco by Life Technologies, Grand Island, NY, USA) supplemented with antibiotic-antimycotic (Gibco by Life Technologies, Grand Island, NY, USA) for 24 h. To determine the length of time for cell attachment to the defect area, the cartilage defect on the femoral condyle was faced upward. CM-DiI (Molecular probes, by Life Technologies, Grand Island, NY, USA) labeled hChonJ and CM-DiO (Molecular probes, by Life Technologies, Grand Island, NY, USA) labeled irradiated-hChonJb#7 were used for the transplantation. The defect was filled with TG-C (mixed hChonJ:hChonJb#7, ratio = 3:1, 2 × 10 ⁶ cells in 10 μL PBS), and was left undisturbed for 7, 15, 20, 30, and 40 min. The femurs were then turned with the defect side down for 10 min in a 50 ml conical tube containing PBS to washout the non-adherent cells in the culture medium. Non-adherent cells in the medium were collected and cell number was counted. The number of cells attached to the cartilage defects was calculated by subtracting the non-adherent cell number from the initial cell number of 2 × 10 ⁶. The cells attached to the cartilage were observed with a fluorescent image analyzer LAS-4000 (FujiFilm, Tokyo, Japan). The use of blue LED with an Y515Di filter and red LED with a R670 filter allows the direct detection of fluorescence from CM-DiO labeled hChonJb#7 and CM-DiI labeled hChonJ, respectively [ 24]. Guidelines from the Institutional Animal Care and Use Committee at the Inha University Medical School and Hospital were followed during all animal procedures.

Cysteine 223 and Cysteine 225 of TGF-β1preprotein are essential for dimerization and mutation of them to Serine was known to enhance spontaneous maturation without activation [ 25]. Based on this observation, we manipulated human TGF-β1 gene to bear the mutation and isolated single clone (hChonJb#7) after transduction into human chondrocyte (hChonJ). While we analyzed the DNA sequence of the transduced gene in hChonJb#7, we detected additional mutations with which the wild-type threonine at position 84 and arginine residue at position 255 of the TGF-β1 precursor had changed into alanine and tryptophan respectively (A84/W255) in addition to the intended mutation of Cysteine 223 and 225 into Serine (S223/255). To determine the effects of the mutations, we prepared pMSCV constructs bearing 4 different TGF-β1variants -wild type (WT), S223/225 mutant, A84/W255 mutant, and hChonJb#7 mutant (both S223/225 and A84/W255 mutated and designated as #7 in figures) - and transfected them into Cos-1 monkey kidney cells to obtain TGF-β1 preproteins released into its culture supernatant.

To confirm their expression, TGF-β1 preproteins in culture supernatants were revealed by SDS-PAGE in both non-reducing and reducing conditions. In non-reducing conditions in which disulfide bonds can be maintained, all four variants showed the mature dimer form of TGF-β1, which is comparable to recombinant human TGF-β1 (rhTGF-β1), and the mature dimer was shifted into monomer form under reducing conditions (Fig.  1a).

To distinguish the activities of TGF-β1 variants quantitatively, we monitored proliferation of hepatocellular carcinoma HepG2 cells when treated with individual variant. This assay was based on the fact that active TGF-β1 inhibits cell proliferation. As shown in Fig.  1b, S223/225 mutant and #7 mutant of TGF-β1 preprotein inhibited proliferation of Hep G2 cells more efficiently (at least 3 times more) than WT or A84/W255 mutant suggesting that S223/225 or #7 mutant was more efficiently activated than WT or A84/W255. TGF-β1 preprotein produced from hChonJb#7 cells also showed similar inhibitory effect on Hep G2 proliferation comparable to that of S223/225 or #7 (Fig.  1b).

In serial dilution studies to determine IC ₅₀, TGF-β1 preprotein in hChonJb#7 cell culture supernatant showed almost identical functional activity (<0.156 ng/ml of IC ₅₀) with rhTGF-β1, which exists in a fully active form (Fig.  1c). Based on these observations, we confirmed that hChonJb#7 cells produce almost fully active TGF-β1.

In previous experiments, we showed that TGF-β1 produced from hChonJb#7 cells is equally functional as rhTGF-β1 in inhibiting carcinoma cell proliferation. Next, we examined whether TGF-β1 from hChonJb#7 cells is similarly effective on human chondrocytes (hChonJ cells).

When hChonJ cells were incubated with rhTGF-β1 in a pellet culture, a shiny white cartilage-like pellet was formed in 2 weeks (Fig.  2a). However, if not supplemented with rhTGF-β1 then pellets do not form well (data not shown). Safranin-O staining showed the presence of proteoglycan in articular cartilage-like pellets. Type II collagen was also detected using immunohistochemistry (Fig.  2b). Proteoglycan and type II collagen are the main components of extra-cellular matrix (ECM) abundant in articular cartilage so they are considered as markers for articular chondrocytes. Rather than staining or using immunohistochemistry, proteoglycan and type II collagen were detected quantitatively with alternative assays such as the GAG assay and qPCR, respectively, following micro-mass culture which is more convenient culture method than pellet culture. As shown in Fig.  2, rhTGF-β1 supplemented hChonJ micro-masses formed approximately 4 times more GAG compared to untreated masses. The TGF-β1 produced by hChonJb#7 exhibited a tendency to induce GAG formation at similar bioactivity levels compared to fully active rhTGF-β1 protein (Fig.  2c). There was a trend in the results of qPCR experiments to detect type II collagen mRNA levels which indicated that TGF-β1 produced from hChonJb#7 cells and rhTGF-β1 both induced type II collagen gene expression and did so at a similar scale exhibiting approximately a 200 fold difference compared to the negative control (Fig.  2d). Although the type II collagen expression level increased, if this was accompanied by an increase in type I collagen expression, it would be difficult to consider the chondrocytes as redifferentiated. Therefore, we examined the ratio of type II collagen mRNA levels over to type I collagen mRNA levels. As shown in Fig.  2e, the ratio also increased compared to the negative control suggesting that type II collagen gene expression was selectively induced.

Because the hChonJb#7 cells themselves produce type II collagen [ 19], we tested whether the extra-cellular matrix of hChonJb#7 including type II collagen and proteoglycan could affect the phenotypic properties of hChonJ cells when they were mixed. To test this possibility, we added human type II collagen or proteoglycan into micro-mass culture as media supplements and compared this to rhTGF-β1. Unlike rhTGF-β1, neither type II collagen nor proteoglycan induced type II collagen gene expression (Fig.  3a). Next, we examined GAG accumulation using the same factors used to test for type II collagen change. hChonJ supplemented with rhTGF-β1 accumulated more GAG than the hChonJ control. However, neither type II collagen nor proteoglycan supplemented hChonJ micro-masses showed any additional GAG accumulation as compared to hChonJ only control (Fig.  3b). From these observations, we could assume the main function of hChonJb#7 in cartilage regeneration is as the source of active TGF-β1 to hChonJ cells.

hChonJb#7 is genetically modified cell line produced via retro-viral infection and recombination. Previous uses of retroviral vectors in the clinic resulted in cases in which the patients who had been treated with retroviral gene therapy developed tumors, and most of these cases were due to the transduction into oncogenes [ 26]. Although the modified TGF-β1 gene was not introduced into any known oncogene in hChonJb#7, we decided to render hChonJb#7 cells replication-incompetent as an extra precaution. For this purpose, the hChonJb#7 cells were irradiated with weak dosage of X-ray irradiation so that the cell would die after completing its life span. However, it was expected to provide TGF-β1while it was alive. To test this, we measured the quantity of TGF-β1 produced by the hChonJb#7 cells after irradiation. As shown in Fig.  4a, a considerable amount of TGF-β1 was detected for approximately 2 weeks while the cells remained viable, confirming that the role of hChonJb#7 cells as TGF-β1 provider to hChonJ cells was not impaired.

To prove that hChonJ cells can be redifferentiated by irradiated hChonJb#7 cells, we mixed hChonJ cells and irradiated hChonJb#7 cells in 3:1 ratio (TG-C) and cultured them to form micro-masses. The 3:1 cell mixture ratio has been used for our pre-clinical animal study and Phase I and II clinical studies proving its efficacy and safety for cartilage regeneration [ 23, 27]. Together with rhTGF-β1, we measured type II collagen mRNA levels and the amount of GAG accumulated in the TG-C micro-masses to determine the hChonJ potency as articular chondrocytes in TG-C. The type II collagen mRNA level was approximately 200 fold higher (Fig.  4b) and the GAG amount was also at least 4 times more than negative control (Fig.  4c). These results were comparable to those of experiments performed with fully active rhTGF-β1 or hChonJb#7 culture supernatant (Fig.  2c and d) and we confirmed that irradiated hChonJb#7 cells also provided sufficient TGF-β1 to redifferentiate the hChonJ cells in TG-C.

To measure the binding ability of TG-C to cartilage, fluorescent labeled hChonJ and hChonJb#7 cells were placed on cartilage with partial knee defects (rabbits). Both CM-DiI labeled hChonJ (red color) and CM-DiO labeled hChonJb#7 (green color) were bound to the defect area compared to the un-disturbed smooth area of the cartilage (Fig.  5a). The number of bound cells increased with time until it reached a plateau after 30 min (Fig.  5b). We also performed the same experiment in human knee joint tissue which was collected after knee joint replacement and observed the similar results (Data not shown). Based on these observations, we conclude that when TG-C is injected onto the damaged cartilage, irradiated hChonJb#7 and hChonJ cells attached to the lesion within 30 min and can play their role as a source of TGF-β1 and cartilage material in order to rebuild cartilage.