Collagen Type II–Derived Telopeptides Cause Tissue Damage in Human Cartilage Cultures Established From Total Arthroplasty

Abstract

Objective

To determine whether telopeptides of collagen type II could induce osteoarthritic tissue damage via receptor for the native protein by using human articular cartilage.

Methods

Cartilage slices were harvested from patients receiving total arthroplasty. Cartilage tissue cultures or primary chondrocyte cultures were treated with 30 µM N- or C-telopeptide (NT or CT) for 7 days or for 24 h. Loss of proteoglycan (PG) from cartilage was evaluated with 1,9-dimethylmethylene blue (DMMB) assay. Conditioned media or cell lysates were measured for levels of matrix metalloproteinases (MMPs), MMP-3 and MMP-13, or integrin beta-1 (ITGB1) with Western blotting or real-time polymerase chain reaction (PCR).

Results

Either NT or CT could induce significant loss of PG from cartilage than did phosphate-buffered saline (PBS), the delivery vehicle (18.45 ± 6.58 or 15.50 ± 4.91 µg PG/mg wet cartilage treated by NT or CT vs. 25.61 ± 4.14 µg PG/mg wet cartilage treated by PBS; P = 0.037 for NT, P = 0.004 for CT). Upregulation of MMP-3 and MMP-13 was induced by either NT or CT at 24 h (chondrocyte cultures) or Days 4 and 7 post-treatment (cartilage cultures). CT-induced stronger expression of ITGB1 in chondrocytes than did NT.

Conclusion

Telopeptides of collagen type II could damage human articular cartilage and upregulate MMP-3 and MMP-13. The catabolic effect of CT might be mediated by ITGB1.

Introduction

Type II collagen accounts for over 90% of collagen fibers in mature articular cartilage which can be found inside weight-bearing joints, such as hip and knee joints. Unlike type I collagen that is widely distributed in the body and consists of 2 alpha1 and 1 alpha2 peptide chains, type II collagen is only found in cartilage of weight-bearing joints and consists of 3 identical alpha1 peptide chains. The network weaved by type II collagen fibers is a structural basis for the elasticity of articular cartilage which can divert stress from subchondral bones inside a weight-bearing joint and function as a protective cushion.^1,2

However, when aging or intraarticular fracture occurring, irreversibly degeneration of articular cartilage is induced and its protective function is compromised. Inside a weight-bearing joint with damaged cartilage, bone rubs against each other resulting in pain, stiffness, and even loss of joint motion. Those are characteristic manifestations of osteoarthritis (OA), the leading cause for disability in adults. ³ A crucial pathological change of osteoarthritic cartilage is extracellular matrix (ECM) breakdown, such as fragmentation of type II collagen fibers, driven by matrix metalloproteinases (MMPs).

Numerous studies have shown that fragments of type II collagen cleaved by MMPs, especially C-terminal telopeptide (CT) also denoted as CTX-II, were elevated in body fluids of OA animal models or patients diagnosed with OA. Matyas et al. ⁴ reported that significantly increased levels of CT were observed in synovial fluids, serum, and urine of canines at 12-week post-surgical transection of the anterior cruciate ligament. Dam and colleagues discovered that urinary level of CT in patients with radiographic knee OA was 56% higher than that in control subjects. Furthermore, they found that baseline level of CT could be applied to prediction of longitudinal cartilage loss determined by magnetic resonance imaging (MRI) measurements over a 21-month follow-up. ⁵

Those fragments of type II collagen can not only act as biomarkers for OA progression but also exert biological effect on chondrocytes to promote cartilage degradation. In 2000, Jennings et al. first reported that those fragments could inhibit synthesis of collagen in bovine or human chondrocyte monolayer cultures, which interfered with the repairing of damaged cartilage. Moreover, they also discovered that those fragments at 1 mg/ml could induce significant tissue damage in human knee or ankle cartilage explant cultures. Coincidently, upregulation of MMP-2 expression was observed in culture media of those explant cultures. ⁶

The MMP-2 upregulation by fragments of type II collagen was further confirmed in bovine or human healthy articular cartilage explant or chondrocyte cultures. In addition, those authors reported that synthetic N-terminal telopeptide (NT) of collagen type II could induce upregulation of MMP-3 and MMP-13 at messenger RNA (mRNA) level up to 5- and 18-fold, respectively. By analyzing culture media of human healthy ankle cartilage explant cultures with enzyme-linked immunosorbent assay (ELISA), they found that NT could induce significant release of MMP-3 from cartilage at time- and dose-dependent fashion. ⁷ This catabolic effect of NT could be mediated by annexin V which is a type of membrane receptor for collagen type II and also acts as a calcium channel in human articular chondrocytes. ⁸

The potential role of NT or CT from collagen type II in osteoarthritic cartilage damage was extensively studied by G. Homandberg’s laboratory which first reported that fragments degraded from fibronectin, another ECM protein, could aggravate cartilage breakdown via their pro-inflammatory effect on chondrocytes. ⁹ In bovine articular chondrocyte or cartilage explant cultures, they demonstrated that NT or CT could upregulate expression of several cartilage-damaging metalloproteinases, including MMP-1, MMP-3, MMP-13, and ADAMTS-5. This effect was comparable to that of fibronectin fragments (Fn-fs) but requiring much higher concentration. Also, those telopeptides could induce significant proteoglycan (PG) loss from cultured cartilage explants, indicating that they were catabolic agents as Fn-fs involved in OA pathogenesis. However, in terms of stimulating release of pro-inflammatory cytokines from chondrocytes, Fn-fs showed much stronger effect than NT or CT. ¹⁰

Although current literature has records of NT or CT cleaved from collagen type II on cartilage degrading bovine or human healthy cartilage, studies on how those telopeptides act on human osteoarthritic articular cartilage have never been reported. Hence, we hypothesized that NT or CT could act through receptors for collagen type II, integrin beta-1 (ITGB1), to upregulate cartilage-damaging MMPs in human osteoarthritic chondrocytes and to promote inflammatory degeneration of the tissue harvested from total joint replacement surgery. Our aims were (1) to examine whether NT or CT could induce expression of MMP-1 and MMP-13 in cartilage tissue cultures and chondrocyte monolayer cultures; (2) to determine whether NT or CT could deplete PG from cartilage causing tissue degradation; (3) to examine the effect of NT or CT on ITGB1 expression in chondrocyte monolayer cultures.

Methods

Acquisition of Human Articular Cartilage

Full-thickness cartilage slices were harvested from either femoral head or tibial plateau of patients (N = 34, 26 women and 8 men, age ranging from 55 to 95 years) who were diagnosed with femoral neck fracture (N = 21), or late stage knee OA (N = 11), or femoral head necrosis (N = 2) and received total hip or knee replacement surgery at a local hospital with institutional review board approval (reference number: LS2019001). Cartilage samples collected from 5 individual patients were used for tissue culture experiments and samples from 29 patients were used for establishment of primary chondrocyte cultures. Characteristics of those cartilage samples are summarized in Suppl. Table 1.

Preparation of Telopeptides of Type II Collagen

Peptides including telopeptides of type II collagen and corresponding control peptides were synthetized by GL Biochem (Shanghai, China) using sequences reported in previous studies.^6,8,10 The features of those peptides are summarized in Table 1. Immediately prior to being used in the experiments, lyophilized peptides were dissolved in sterile 1× phosphate-buffered saline (PBS) to make a 10 mg/ml stock solution. Based on the amino acid count of each peptide and the average molecular weight of amino acids (110 Da), a series of dilutions of stock solution were made to achieve 30 µM final concentration of each peptide in culture media. Equal volume of sterile 1× PBS was added into designated cultures as vehicle control.

Table 1.

Characteristics of Synthesized Peptides Used in the Study.

Full Name	Acronym	Amino Acid Count	Amino Acid Sequence	Location on α1-Chain of Procollagen Type II
N-terminal telopeptide	NT	31	QMAGGFDEKA GGAGLGVMQG PMGPMGPRGP P	Residues 182-212
C-terminal telopeptide	CT	24	IDMSAFAGLG PREKGPDPLQ YMRA	Residues 1218-1241
Scrambled N-terminal telopeptide	SN	31	GPGAGQPGKG RGPAPLQFGM AMMDMADPGE V	N/A (control peptide of NT)
Scrambled C-terminal telopeptide	SC	24	MARFPAMLGP ARDPISYQKE GDGL	N/A (control peptide of CT)
Helical peptide	HP	24	GPEGAQGPRG EPGTPGSPGP AGAS	Residues 384-407

Establishment of Cartilage Tissue Culture and Collection of Conditioned Media and Cartilage Slices

As illustrated in Figure 1 , immediately after being shaved from osteochondral specimens derived from total hip or knee replacement surgery, full-thickness cartilage slices were aseptically weighed and evenly distributed into a 12-well culture plate at ~100 mg wet weight cartilage slices per well containing 1.0 ml serum free culture media (DMEM/F12/1% pen-strep). First, cartilage slices were pre-equilibrated for 2 days and media were changed daily. Conditioned medium samples collected during pre-equilibration were denoted as Days −2 and −1. At Day 0, cartilage slices were either treated with 30 µM telopeptides or left untreated. Cartilage cultures treated with 1× PBS or 10 ng/ml rhIL-1b served as a negative or positive control, respectively. Media were replenished every other day and conditioned medium samples were collected at Days 2, 4, and 7. At Day 7, experiments were ended and cartilage slices were collected. Conditioned media and cartilage slices were analyzed for expression of MMP-3, MMP-13, and for PG depletion, respectively.

Figure 1.

Scheme of experimental design. Human articular cartilage tissue or primary human articular chondrocyte cultures were employed to determine whether telopeptides of Col-2 could cause cartilage damage through upregulating MMP expression and inducing PG depletion. In addition, the effect of those telopeptides on expression of membrane receptor ITGB1 was examined.

Establishment of Primary Chondrocyte Culture and Collection of Conditioned Media and Cell Lysates

As illustrated in Figure 1 , chondrocytes were isolated from cartilage slices digested sequentially with pronase E (Sigma-Aldrich Corp, St. Louis, MO) and collagenase IA (Sigma-Aldrich Corp) as reported.^11,12 Chondrocytes were plated in monolayer at a high density of 2.5 × 10⁵ cells/cm² and cultured in DMEM/F12 supplemented with 10% fetal bovine serum (FBS) and 1% pen-strep for 4 to 5 days till reaching 90% or above confluence. At Day 5 or 6 post-seeding, cultures were switched to serum free media. After 24 h of serum deprivation, cultures were treated with telopeptides, PBS, or rhIL-1b. After another 24 h, conditioned media were collected for examination of MMP expression at protein level and cells were lysed for examination of expression of MMPs and ITGB1 at protein and mRNA level.

Determination of PG Depletion From Cartilage Slices With 1,9-Dimethylmethylene Blue Assay

Cartilage slices collected at Day 7 post-treatment were first digested with 0.5 mg/ml papain (Sigma-Aldrich Corp) digestion buffer at 65°C for 4 h. Sulfated glycosaminoglycan polysaccharide from shark fin cartilage (Sigma-Aldrich Corp) served as PG standard for the establishment of a standard curve. PG content in each cartilage digest sample or PG standards first reacted with 1,9-dimethylmethylene blue (DMMB) reagent, resulting blue-colored product. The absorbance of each sample at 530 nm wavelength was measured with a BioTek microplate reader (Agilent Technologies, Inc., Santa Clara, CA). PG content in each cartilage digest sample was calculated from absorbance and standard curve equation and normalized to the wet weight of cartilage slices.

Determination of Expressions of MMPs and ITGB1 at Protein Levels With Western Blotting

To determine expression of MMPs, conditioned medium samples were first dialyzed with SnakeSkin dialysis tubing (MWCO = 10 kDa) (ThermoFisher Scientific, Waltham, MA) against deionized water at 4°C for 48 h until phenol red color disappeared. Next, dialyzed conditioned medium samples were concentrated with a centrifugal vacuum concentrator (Labconco Corporation, Kansas City, MO) until liquid portion in each sample was invisible to naked eye. Immediately prior to examination with Western blotting, each concentrated sample was reconstituted with 50 µL of deionized water.

To determine ITGB1 expression, cultured chondrocytes were lysed with RIPA lysis and extraction buffer (ThermoFisher Scientific) supplemented with 1:100 diluted protease inhibitor cocktail (ThermoFisher Scientific) and then cell lysates were centrifugated at 8,000 rpm for 10 min at 4°C to obtain supernatants containing intracellular proteins. Protein concentration of each cell lysate sample was measured with a Pierce BCA Protein Assay kit (ThermoFisher Scientific).

Either reconstituted conditioned medium sample or cell lysate sample was prepared with 5× loading buffer, reduced with 0.5 M DTT, and boiled for 5 min. Equal volume of each medium sample or equal total proteins of cell lysate sample was loaded onto a 4% SDS-PAGE concentrating gel and proteins were separated in a 10% separating gel. After proteins were blotted onto a nitrocellulose membrane, 5% non-fat dry milk in 1× TBST was used to block the membrane. Proteins on the membrane were then probed with anti-MMP-1, or MMP-13, or ITGB1 antibody (Cell Signaling Technology, Danvers, MA, USA) 1:1,000 dissolved in 5% BSA/TBST overnight at 4°C. After primary antibody incubation, the membrane was incubated in anti-rabbit IgG conjugated with HRP (Cell Signaling Technology) 1:3,000 diluted in 5% BSA/TBST for 1 h at room temperature.

To visualize proteins bands, SuperSignal West Femto Maximum Sensitivity Substrate (ThermoFisher Scientific) was applied to the membrane and chemiluminescence signals were captured with the ChemiDoc MP imaging system (BioRad, Hercules, CA, USA). Densitometric analysis of protein bands was performed with ImageJ.

Determination of Expression of ITGB1 or MMPs at mRNA Level With Real-Time PCR

After 24-h stimulation, chondrocytes were lysed with TRIzol Reagent (ThermoFisher Scientific). Intracellular total RNAs were isolated with chloroform and purified with isopropanol and ethanol. Concentration and purity of each purified RNA sample were measured with a NanoDrop One/One^C Microvolume UV-Vis spectrophotometer (ThermoFisher Scientific). Up to 1,000 ng of total RNAs were reversely transcribed into cDNA with a PrimeScript RT Master Mix (Perfect Real Time) kit (Takara Bio Inc., Shiga, Japan). Expression levels of MMP-1, MMP-13, and ITGB1 were measured with a FastStart Essential DNA Green Master kit (Roche LifeScience, Basel, Switzerland) and nucleotide sequence encoding 18S rRNA was used as a reference gene.

Ct values of target genes were read and recorded with LightCycler 480 Ⅱ (Roche LifeScience, Basel, Switzerland). All primers were synthesized by Exsyn-Biotechnology Co. Ltd. (Shanghai, China). The sequences of primers used for real-time PCR experiments are listed in Table 2 . The Ct values of MMP-1, MMP-13, and ITGB1 genes were compared with that of 18S rRNA, respectively. The relative mRNA expression level of each target gene was determined by 2^–∆∆Ct method.

Table 2.

Primer Sequences for Target Genes.

Genes	Sequences of Primers
MMP-3	Forward	AGGCTGTATGAAGGAGAGGCTGAT
	Reverse	AGTGTTGGCTGAGTGAAAGAGAC
MMP-13	Forward	AGCATCTGGAGTAACCGTATTG
	Reverse	CCCGCACTTCTGGAAGTATT
ITGB1	Forward	GGCAGTGCATGTGACTGTT
	Reverse	CTGAACACATTCTTTATGCTC
18S rRNA	Forward	CGGCTACCACATCCAAGGAA
	Reverse	GCTGGAATTACCGCGGCT

Statistical Analyses

The difference of PG content expressed as mean (µg/mg wet cartilage) ± standard deviation (SD) in each experimental group was compared with that in either PBS or control peptide treated group with unpaired t-test (t-Test: Two-Sample Assuming Equal Variances). The relative mRNA expression of MMPs or ITGB1 to 18S rRNA among experimental groups was calculated with 2^–∆∆Ct method and the fold increase value was also compared by using unpaired t-test. The same test was employed to compare the difference of ITGB1 expression determined by gray values of protein bands among experimental groups. The difference between 2 compared groups is considered as statistically significant when P value is less than 0.05.

Results

Both NT and CT Significantly Induced PG Depletion From Hip or Knee Cartilage. This Effect Was Comparable to That of IL-1b

After 7 days of continuous stimulation by 30 µM NT, there was only 18.45 ± 6.58 µg PG/mg wet cartilage remaining. This PG content was much lower than that in PBS (vehicle)-treated (25.61 ± 4.14 µg PG/mg wet cartilage; P = 0.037) or SN (control peptide)–treated tissue (25.31 ± 2.80 µg PG/mg wet cartilage; P = 0.04) (Fig. 2A and C). Similarly, cartilage treated with 30 µM CT for 7 days only contained 15.50 ± 4.91 µg PG/mg wet cartilage which was significantly lower than that in cartilage treated either with vehicle (25.61 ± 4.14 µg PG/mg wet cartilage; P = 0.004) or control peptide (SC) (20.78 ± 3.25 µg PG/mg wet cartilage; P = 0.04) (Fig. 2B and C).

Figure 2.

Comparison of PG content in cartilage slices between telopeptide treatments and control treatments at Day 7 post-treatment. (A) PG content in NT-treated or control cartilage slices (N = 5). (B) PG content in CT-treated or control cartilage slices (N = 5). (C) PG content in NT or CT-treated cartilage slices was compared with that in HP- (N = 5) or IL-1b-treated tissue (N = 4). Error bar represents standard deviation. Each dot represents PG content in cartilage slices of an individual patient. Data were generated from the following 5 patients: 77/M/knee OA, 68/F/hip osteonecrosis, 71/M/femoral head fracture, 71/F/knee OA, 82/F/femoral neck fracture.

Among 5 patients examined, NT or CT induced the most PG loss from tibial plateau cartilage of a 71-year-old female patient diagnosed with knee OA late stage. After 7-day treatment, there was 7.98 or 8.56 µg PG/mg wet cartilage remaining in NT- or CT-treated cartilage. The least effect of NT or CT on PG loss was observed in femoral head cartilage isolated from an 82-year-old female patient diagnosed with femoral neck fracture. There was 23.17 or 21.80 µg PG/mg wet cartilage remaining in the tissue after 7-day treatment with NT or CT (Fig. 2A-C).

In addition, we observed that the average PG content in tibial plateau cartilage (N = 2) was 12.63 or 11.63 µg PG/mg wet cartilage after 7-day treatment with NT or CT. This level was lower than that in femoral head cartilage (N = 3) which was 22.33 ± 2.92 µg PG/mg wet cartilage or 18.08 ± 3.71 µg PG/mg wet cartilage, respectively (Suppl. Fig. 1). Moreover, in femoral head cartilage cultures established from a 68-year-old female patient diagnosed with hip osteonecrosis, NT did not induce obvious PG loss from the tissue while CT did (24.75 vs. 14.37 µg PG/mg wet cartilage) compared with untreated control (24.13 ± 2.62 µg PG/mg wet cartilage) (Fig. 2A and B).

In terms of causing PG depletion from cartilage, either NT or CT showed similar effect to rhIL-1b. Cartilage treated by rhIL-1b for 7 days contained 14.46 ± 4.14 µg PG/mg wet cartilage, a level showing insignificant difference from that in NT- or CT-treated tissue (P = 0.164 or P = 0.372). Peptide derived from the helical region of type II collagen (HP) also induced significant PG loss from cartilage compared with vehicle control (PBS) (17.30 ± 3.38 µg PG/mg wet cartilage by HP vs. 25.61 ± 4.14 µg PG/mg wet cartilage by PBS; P = 0.004). However, this effect was slightly weaker than that of CT (15.50 ± 4.91 µg PG/mg wet cartilage) but was comparable to that of NT (18.45 ± 6.58 µg PG/mg wet cartilage) ( Fig. 2C ).

Strong Expression of Cartilage-Damaging MMPs in Femoral or Tibial Cartilage was Induced by NT or CT at Day 4 Post-Treatment and the Effect was Still Obvious at Day 7 Post-Treatment

One day prior to the beginning of the experiments, that is, pre-equilibration Day −1, neither MMP-3 nor MMP-13 protein band was clearly detected in conditioned media collected from femoral head cartilage (male, 71 years old, diagnosed as femoral neck fracture) cultures in all experimental groups. However, at Day 4 post-treatment with NT or CT, strong MMP-3 or MMP-13 signal was observed in conditioned media. By contrast, neither of this signal was detectable in control groups. Although at Day 7 post-treatment MMP-3 or MMP-13 signal was seemingly weaker in conditioned media of cultures treated with NT or CT than that at Day 4 post-treatment, the strength of either signal was still much stronger than that in the control groups (Fig. 3A and B).

Figure 3.

Time-dependent MMP upregulation by telopeptides in cartilage cultures. Protein expression of MMP-3 and MMP-13 in conditioned media was examined at Days −1, 4, and 7. (A) Expression of MMPs in NT-treated or control cultures of femoral head cartilage collected from a 71-year male patient diagnosed with femoral neck fracture. (B) Expression of MMPs in CT-treated or control cultures of femoral head cartilage collected from a 71-year male patient diagnosed with femoral neck fracture. (C) Expression of MMPs in NT-treated or control cultures of tibial plateau cartilage collected from a 77-year male patient diagnosed with late-stage OA. (D) Expression of MMPs in CT-treated or control cultures of tibial plateau cartilage collected from a 77-year male patient diagnosed with late-stage OA.

Similar results were observed in cultures of cartilage from tibial plateau of a 77-year male patient diagnosed with late stage knee OA (Fig. 3C and D). We also examined the effect of telopeptides on MMP upregulation in femoral head cartilage of a 68-year-old female patient diagnosed with femoral head necrosis and found results described above were reproducible (Suppl. Fig. 2A and B).

To verify that those upregulated MMPs in conditioned media was produced by chondrocytes in cultured cartilage slices, we deliberately killed all chondrocytes in some cartilage slices by repeating freeze-thawing cycle 3 times (freezing cartilage slices at −80°C for 3 min and then thawing them at 37°C for 5 min). At Day 4 post-treatment by telopeptides, MMP-3 signal was only detected in live cartilage slices treated with NT or CT, whereas the signal was undetectable in freeze-thawed cartilage slices treated with NT or CT (Suppl. Fig. 3).

In chondrocyte monolayer cultures, we did not observe significantly increased MMP mRNA expression at 24-h post-treatment by either NT or CT (data not shown). However, in conditioned medium samples collected at the same time point, we observed much higher expression level of MMP-3 and MMP-13 in CT-treated culture than in SC-treated one, whereas the expression of those 2 MMPs was barely detectable in untreated or PBS-treated cultures. In contrast, NT stimulated moderate MMP-3 upregulation in only 1 out of 3 replicates and MMP-13 upregulation in 2 out of 3 replicates. HP as a peptide control did not induce any detectable secretion of MMP-3 or MMP-13 from cultured chondrocytes (Suppl. Fig. 4A and B).

The Effect of Telopeptides on MMP Upregulation was Still Weaker Than That of IL-1b. HP Could Also Upregulate MMP Expression and This Effect was Comparable to Telopeptides

We compared the effect of telopeptides on MMP upregulation in cartilage cultures to that of IL-1b which is a well-studied pro-inflammation factor in OA pathogenesis. At Day 4 or 7 post-treatment, the strongest MMP-3 or MMP-13 signal was detected in cartilage treated with 10 ng/ml rhIL-1b, whereas MMP signals in NT- or CT-treated cartilage cultures were moderately weaker but with similar strength (Fig. 4).

Figure 4.

Comparison of MMP expression between telopeptides and IL-1b or HP at Day 4 or 7 post-treatment. Cartilage slices harvested from femoral head of a 71-year male patient diagnosed with femoral neck fracture were continuously treated with telopeptides or IL-1b or HP for 7 days. Conditioned media collected at Day 4 or 7 were examined for expression of MMP-3 and MMP-13 with Western blotting.

Interestingly, we also observed that peptide derived from the helical region of type II collagen (HP) upregulated expression of MMP-3 or MMP-13 in cartilage cultures at Day 4 or 7 post-treatment. Seemingly, when compared with telopeptides, HP showed slightly stronger effect on upregulating MMP-13 expression while slightly weaker effect on upregulating MMP-3 expression (Fig. 4).

Expression of Type II Collagen Membrane Receptor, ITGB1, in Chondrocytes was Moderately Elevated by CT but not by NT Stimulation. The Effect of CT on Upregulation of ITGB1 was Comparable to That of IL-1b

When examined at 24-h post-treatment, the mRNA expression level of ITGB1 in chondrocytes treated with NT was similar to that in non-treated control cells or treated with vehicle reagent or scrambled control peptide (N = 5; Fig. 5A ). In CT-treated chondrocytes, ITGB1 expression was insignificant from that in non-treated controls (N = 3; P = 0.129; Fig. 5B ). As a positive control, rhIL-1b did not stimulate significant ITGB1 mRNA expression either (N = 2; Fig. 5C ). As a peptide derived from the helical region of collagen type II, HP exhibited little effect on ITGB1 expression at mRNA level, which was similar to NT (N = 3; Fig. 5C ). Nonetheless, expression of ITGB1 at protein level was slightly upregulated by either telopeptide with CT showing a stronger effect than did NT ( Fig. 5D ).

Figure 5.

Comparison of ITGB1 expression in chondrocytes at 24-h post-treatment. Total intracellular RNAs or proteins were extracted from cultured chondrocytes either non-treated (NC) or treated with vehicle reagent (PBS) or with telopeptides (NT or CT) or with scrambled telopeptides (SN or SC) or with helical peptide (HP) or with rhIL-1b. After cDNA conversion or total intracellular protein extraction, real-time PCR or Western blotting was performed to determine expression level of ITGB1 in each treatment group. The mean value and each data point of ITGB1 mRNA expression level relative to that of 18S rRNA in NT (N = 5) or CT (N = 3) groups were plotted (A and B). The average expression level of ITGB1 mRNA in NT or CT groups was also compared with that in HP or rhIL-1b treated group (N = 3 for HP; N = 2 for IL-1b) (C). Representative blots showing protein expression levels of ITGB1 in NT or CT groups and plots with each data point showing comparison results of average band intensity in each treatment group are presented (N = 3 for NT; N = 2 for CT) (D). Error bars indicate standard deviation.

Discussion

We are the first to report that telopeptides derived from collagen type II, NT or CT, could cause significant PG loss from femoral or tibial cartilage of patients who received total joint replacement surgery. This telopeptide-induced cartilage damage was accompanied by upregulated expression of MMP-3 and MMP-13, two MMPs heavily involved in osteoarthritic cartilage degradation.^13,14 These results were consistent with what Guo and colleagues ¹⁰ reported in a study that employed bovine articular cartilage harvested from metacarpophalangeal joints. Furthermore, we discovered that CT was more effective than NT in terms of inducing PG depletion and MMP upregulation from cartilage of patients who were either diagnosed with OA or were at high risk of having OA.

In our human cartilage tissue cultures, we discovered that either NT or CT at 30 µM could induce remarkable PG depletion from the tissue after a 7-day continuous treatment. This dramatic effect of telopeptides was previously observed in bovine cartilage tissue cultures by Guo and colleagues. Their 6-day dose-response experiments indicated that NT or CT at 30 µM exhibited the strongest PG depletion effect. ¹⁰ That was why we tested this dose in our human cartilage cultures. By examining the effect of NT or CT on PG depletion from cartilage samples of 5 patients receiving total hip or knee replacement surgery, we discovered that those 2 telopeptides not only caused significant PG loss but also exhibited similar effect to that of IL-1b. In addition, CT showed stronger effect than did NT.

An earlier study conducted by Jennings et al. revealed that bovine collagen type II fragment mixture generated by bacterial collagenase could induce significant PG loss in human ankle or knee cartilage explant after 3-week culturing. The authors stated that their bovine collagen fragment mixture was enriched in N- or C-telopeptides and the MW was smaller than 10 kDa. ⁶ The effective dose of this fragment mixture was 1 mg/ml, equivalent to 200 µM if calculated with 5 kDa as average MW. However, this study did not use defined telopeptide of collagen type II as what we did in our study. We clearly showed that N-telopeptide corresponding to residues 182 to 212 or C-telopeptide corresponding to residues 1218 to 1241 of alpha1 chain of human collagen type II had cartilage-damaging bioactivity. Furthermore, we detected this catabolic activity of telopeptides on cartilage at 1 week of treatment, which was 2 weeks earlier than what they reported. Last, we examined the effect on cartilage of Asian patients diagnosed with femoral neck fracture, hip osteonecrosis, or knee OA which was not studied by them. They harvested macroscopically normal knee or ankle cartilage from 3 Caucasian corpses with a wider range of age (39-71 years).

To explore the mechanism by which NT- or CT-induced cartilage damage, we examined the levels of 2 MMPs that can cause PG depletion and are heavily involved in OA pathogenesis, MMP-3 and MMP-13,^14,15 in conditioned media harvested from cartilage explant cultures. Compared with medium samples collected at 1 day prior to telopeptide treatments that showed little MMP signal, samples at Day 4 or 7 post-treatment with either NT or CT showed not only detectable but also much stronger signals of MMP-3 and MMP-13. The upregulation of those 2 MMPs by telopeptides was also observed in the conditioned media of our human chondrocyte monolayer cultures at 24-h post-treatment, a much earlier time point than that in cartilage cultures. This faster response of chondrocytes in monolayer cultures than in cartilage cultures might be due to lesser ECM surrounding the cell, which enabled faster penetration of telopeptides through ECM to reach to chondrocytes. Since those 2 MMPs have proved to be the driving force for cartilage PG depletion, this observation well explained why NT or CT induced significant loss of PG from cartilage ECM at Day 7 post-treatment observed in our study.

In earlier studies, NT- or CT-induced upregulation of MMPs in conditioned media was only observed in cartilage explant cultures established from bovine metacarpophalangeal joints or human healthy ankle joints. They did not examine this effect of those 2 telopeptides on OA-prone or OA cartilage derived from human hip or knee joints that was used in our study.^7,10 Moreover, Guo et al. ¹⁰ reported that NT or CT at 30 µM could stimulate significant MMP-3 and MMP-13 release from bovine cartilage at 1-day post-treatment. We discovered that the same dose of those 2 telopeptides could greatly upregulate MMP expression in human chondrocytes at Day 4 post-treatment and this effect could still be detected at Day 7. Fichter and colleagues only reported the dose-dependent effect of NT on MMP-3 induction in human healthy ankle cartilage cultures. The highest dose they examined was around 300 µM. But this dose induced less MMP-3 release than did 30 µM, which implied that 30 µM might be the most effective dose in terms of inducing MMP-3 upregulation in human chondrocytes. ⁷

Furthermore, we discovered that a peptide derived from the triple helical region of collagen type II (HP) could also induce upregulation of MMP-3 and MMP-13 but with lesser effect when compared with telopeptides. This result may explain why HP induced lesser PG loss from cartilage explants than did NT or CT in our 7-day experiments. However, in terms of upregulating those 2 MMPs, either telopeptides or HP still showed much weaker effect than did IL-1b. Upregulated MMP-3 and MMP-13 by IL-1b may cleave cartilage ECM collagen type II to generate bioactive telopeptides and HP which can induce more MMPs.^8,16 By contrast, neither of the telopeptides could induce IL-1b expression in bovine chondrocyte cultures. ¹⁰ Therefore, the accumulated action of IL-1b and collagen fragments was certainly stronger than that of any collagen fragment working alone.

Next, we examined the effect of those collagen fragments on expression levels of major membrane receptors for collagen type II to test our hypothesis that the catabolic action of telopeptides or HP was mediated by receptors for their native protein. Integrins are membrane receptors for cartilage ECM proteins and each integrin is a heterodimer composed of an alpha and a beta subunit. Studies have shown that ITGB1 family are major membrane receptors for collagen type II in human articular cartilage and its expression level is elevated in osteoarthritic cartilage.^17,18 Although statistical significance was not achieved in our study, the 2^–∆∆Ct value of ITGB1 mRNA at 24 h in CT-treated chondrocytes was 2.06 ± 1.16, whereas the value was 0.91 ± 0.38 in NT-treated chondrocytes. In addition, slight elevation of ITGB1 protein expression was observed in CT-treated human articular chondrocytes at 24 h while not in NT-treated cells. However, to determine whether or not the catabolic action of CT on human chondrocytes is mediated by ITGB1, a bigger sample size, more time points and specific inhibitors are required in future studies.

Our observation was consistent with what Lucic and colleagues had reported in a study examining how those synthetic peptides derived from collagen type II interacted with chondrocytes. By applying those peptides to chondrocytes isolated from human talus cartilage, they discovered that annexin V, a calcium ion channel on chondrocyte membrane, was most likely the binding receptor for NT and suggested that ITGB1 family might be receptors for CT or HP. ⁸ In our study, after 24-h treatment, we observed slight upregulation of ITGB1 expression induced by CT, not by NT, in human femoral or tibial chondrocytes. Nonetheless, to confirm the involvement of ITGB1 in CT-mediated cartilage degradation, samples collected at earlier time points, such as 4-, 8-, 12-h post-treatments, are required for the examination of ITGB1 expression level and the effect of ITGB1 inhibitors on CT-induced cartilage PG depletion and MMP upregulation need to be determined in our future studies.

A limitation of our study is that cartilage samples were not collected from the same anatomical location or patients with the same diagnosis. Studies have shown that human knee cartilage reacted differently from ankle cartilage to pro-inflammatory stimulation. Kang et al. reported that knee cartilage was more susceptible than ankle cartilage to fibronectin fragment-induced tissue damage. This was further confirmed by Eger and colleagues who reported that knee cartilage was more responsive to IL-1beta-induced catabolic effect while less active in restoration than ankle cartilage following removal of IL-1beta.^19,20 Also, differential expression levels of various MMP mRNAs in cartilage was reported between patients diagnosed with OA and the ones with fracture. ²¹ Although our data showed that cartilage responded to NT or CT stimulation by depleting PG and upregulating MMPs regardless of anatomical location or donor diagnosis, statistical comparison between hip and knee cartilage or between OA and femoral head fracture or hip osteonecrosis could not be made due to a small sample size in each group. This issue will be investigated and solved in our future study.

In summary, we established human articular cartilage explant cultures and primary human chondrocyte monolayer cultures to test our hypothesis that telopeptides derived from human collagen type II could induce osteoarthritic cartilage damage by depleting PG content from the tissue and upregulating MMP expression. This catabolic action could be mediated by membrane receptors for the native protein of those telopeptides. Our data supported this speculation by revealing that significant PG loss was induced by NT or CT at 7 days post-treatment and significant upregulation of MMP-3 and MMP-13 was induced at 4 days post-treatment which was 3 days earlier than the time point when significant PG loss was detected. Although significant upregulation of ITGB1 expression in chondrocytes treated by CT or NT for 24 h was not detected in this study, future studies employing a bigger sample size, multiple treatment time points, and specific inhibitors to ITGB1 or annexin V will be conducted to determine the role of those membrane receptors in CT- or NT-induced human articular cartilage damage.

Here, we would like to clarify that the No. 195 amino acid in the sequence of NT used in this study was Gly (G) which should be Gln (Q) according to protein database (GenBank: KAI4065584.1 or UniProtKB/Swiss-Prot: P02458.3). We cited this amino acid sequence from a published paper authored by Guo et al. ¹⁰ This sequence was also used for synthesizing NT in a study conducted by Chowdhury et al. They observed similar catabolic effect of NT on porcine cartilage to what we observed in human cartilage. ²² This indicated that the Q to G substitution at No. 195 position of NT primary structure did not affect the secondary structure of this peptide. This may be explained by the non-polar nature shared by those 2 amino acids. The corrigendum regarding to this NT amino acid sequence error will be published in Inflammation Research journal in which the original paper containing this amino acid sequence error was published.