Chondrex: new marker of joint disease

¹ Novadex, Inc., San Diego, CA 92121.

² University of California–San Diego Medical Center, Division of Rheumatology, San Diego, CA 92103.

³ University of Nebraska Medical Center, Department of Internal Medicine, Omaha, NE 68198.

⁴ Metra Biosystems, Inc., 265 N. Whisman Rd., Mountain View, CA 94043.
^a Author for correspondence. Fax 650-903-9500; e-mail jvisor{at}metrabio.com.

	Abstract

Top Abstract Introduction Materials and Methods Chondrex (YKL-40).... Alkaline phosphatase-labeled... assay method study subjects Healthy subjects. Serum... statistical method Results Discussion References

 
Chondrex, a major secretory protein of human chondrocytes and synovial fibroblasts, is increased in serum of patients with joint and cartilage disease. We have developed a sandwich-type ELISA for quantifying chondrex in serum. The interassay CVs were 2.8–3.7% and the average within-run and total CVs were 3.6% and 5.4%, respectively. The limit of detectability by linear dilution was 20 µg/L, recovery upon dilution was 102% ± 5%, and analytical recovery (of added analyte) was 98% ± 11%. The reference interval (central 90% interval) for chondrex in healthy adults was 25–95 µg/L. Chondrex values for patients with active rheumatoid arthritis or osteoarthritis were significantly greater than in healthy adults, inactive rheumatoid arthritis patients, and diabetes patients (P <0.05). In patients treated with disease-modifying antirheumatic drug therapy, decreasing chondrex values reflected the clinical improvement observed in responders, whereas the values were maintained or increased in nonresponders. In conclusion, chondrex may be a useful marker in the clinical investigation of arthritis.

	Introduction

Top Abstract Introduction Materials and Methods Chondrex (YKL-40).... Alkaline phosphatase-labeled... assay method study subjects Healthy subjects. Serum... statistical method Results Discussion References

 
Chondrex (YKL-40, HC gp-39) is a 40-kDa glycoprotein originally identified in the whey secretions of nonlactating cows (1). It is secreted by the human osteosarcoma cell line MG-63 (2), by articular human cartilage chondrocytes (3)(4), and by human synovial fibroblasts but not skin or lung fibroblasts (5). In previous publications, this protein had been called YKL-40 according to the one-letter code for its first three N-terminal amino acids and its apparent molecular mass of 40 kDa as determined by sodium dodecyl sulfate–polyacrylamide gel electrophoresis (2). Hakala et al. (4) referred to this protein as human cartilage glycoprotein-39 (HC gp-39), based on an apparent mass of 39 kDa.

Chondrex mRNA is expressed strongly in chondrocytes and liver; weakly in brain, kidney, and placenta; and at undetectable amounts in heart, lung, skeletal muscle, pancreas, mononuclear cells, and skin fibroblasts (4). In addition, chondrex mRNA is undetectable in normal human cartilage but is prominent in cartilage of patients with rheumatoid arthritis (RA)¹ (4). The amino acid sequence of chondrex shows some homology to the bacterial chitinase protein family, although chitinase activity has not been demonstrated (3)(4). Some postulate that chondrex functions as a glycosidic bond hydrolase involved in the tissue-remodeling process (3)(4)(6).

Based on this information, a RIA was developed and used to measure chondrex concentrations in serum and synovial fluid of patients with various forms of joint disease and in serum of presumably healthy adults (3)(6). The RIA data showed that chondrex concentrations are ~2.5-fold greater in the serum of patients with inflammatory or degenerative joint disease than in healthy adults. Moreover, chondrex concentrations in synovial fluid were 10–15-fold higher than in serum. Serum and synovial fluid chondrex concentrations were highly correlated in patients with joint disease, suggesting that in patients with joint disease, most of the chondrex found in serum may be produced in the joint. These data suggest that chondrex may be a useful new marker for joint disease.

We describe here the development of the Chondrex(TM) enzyme immunoassay for quantifying chondrex in human serum. In the present study we report the analytical performance characteristics of the assay and describe preliminary clinical results. Having established a representative reference interval for serum chondrex in healthy adults, we compare this with the chondrex serum concentrations in individuals with active and inactive RA, osteoarthritis (OA), and diabetes. In addition, serum chondrex concentrations are compared with clinical response variables in RA patients on disease-modifying antirheumatic drug (DMARD) therapy.

	Materials and Methods

Top Abstract Introduction Materials and Methods Chondrex (YKL-40).... Alkaline phosphatase-labeled... assay method study subjects Healthy subjects. Serum... statistical method Results Discussion References

 
reagents
Unless otherwise indicated, all materials were obtained from Sigma Chemical Co.

Chondrex (YKL-40). Chondrex (YKL-40) was obtained from the supernatants of MG-63 cells (from the laboratory of Paul Price, University of California–San Diego) cultured by a modified version of the procedure of Johansen et al. (3). MG-63 production flasks were seeded at 1.8 x 10 to 2.7 x 10 cells/cm in RPMI-1640 (Irvine Scientific) plus newborn calf serum, 100 mL/L (Irvine Scientific), 0.1 mol/L HEPES, and 50 mg/L vitamin C (complete medium). Flasks were incubated at 37 °C with humidity and CO2-enriched (100 mL/L) atmosphere for 6–8 days, replacing spent medium with fresh every 2–3 days. The cultures were then switched to serum-free medium (complete medium minus the newborn calf serum). The supernatants were harvested and the media replaced every 1–3 days for 30 days. Chondrex was purified from the supernatants by concentrating glass-fiber-filtered material 20-fold with a 30-kDa screen channel cassette with tangential flow (Filtron) and then affinity-purifying over a heparin-Sepharose CL-6B column (Pharmacia Biotech) equilibrated with a solution of 10 mmol/L sodium phosphate and 50 mmol/L sodium chloride, pH 7.5. Bound material was eluted with a sodium chloride gradient (from 50 mmol/L to 2 mol/L) in 10 mmol/L sodium phosphate, pH 7.5, and 4-mL fractions were collected and pooled according to: absorbance at 280 nm, chondrex concentration by immunoassay, and purity by sodium dodecyl sulfate–polyacrylamide gel electrophoresis.

Top Abstract Introduction Materials and Methods Chondrex (YKL-40).... Alkaline phosphatase-labeled... assay method study subjects Healthy subjects. Serum... statistical method Results Discussion References

 
Monoclonal antibodies.
Monoclonal antibodies were generated by immunization of A/J mice with purified chondrex. After an initial intraperitoneal boost of 50 µg of chondrex in complete Freund's adjuvant, the mice were allowed to rest for 8 weeks. They were then boosted intraperitoneally at 2-week intervals; first with 25 µg of chondrex in incomplete Freund's adjuvant; then with 10 µg of chondrex in 10 mmol/L phosphate, 150 mmol/L sodium chloride, pH 7.2; and finally with 5 µg of chondrex in 10 mmol/L sodium phosphate, 150 mmol/L sodium chloride, pH 7.2, intravenously. The spleens of two mice were then fused in accordance with standard procedures. Supernatants were screened with a solution-phase ELISA in which chondrex bound to a microplate well was incubated with supernatant to be tested. Sequential incubations with anti-mouse IgG–horseradish peroxidase conjugate and then substrate allowed detection of anti-chondrex antibodies. Four IgG antibodies were subcloned, isotyped, and screened further in sandwich assay format. A pair-search included combinations of all the monoclonal antibodies as well as polyclonal material. The final clone (116) was chosen because of its superior sensitivity in a monoclonal:polyclonal sandwich ELISA format as well as its growth characteristics in cell culture.

Monoclonal antibody 116 was purified from hybridomas grown in HB Pro serum-free media (Irvine Scientific). Terminal culture supernatants were harvested by centrifugation when cell viability fell to <10–20%. Supernatants were purified over fast-flow Protein A–Sepharose (Pharmacia) after 0.2-µm (pore size) filtration and addition of saturated sodium borate and sodium chloride to final concentrations of 100 g/L and 3 mol/L, respectively. Bound antibody was eluted with 0.1 mol/L glycine, pH 3.0, and fractions were neutralized by the addition of 10% (by vol) 1.2 mol/L Tris, pH 8.5. Antibody-containing fractions were pooled and dialyzed into 10 mmol/L phosphate, 150 mmol/L sodium chloride, pH 7.2.

Biotin–Fab conjugate.
Monoclonal antibodies were digested into Fab fragments by incubation with immobilized papain (Pierce Chemical Co.) in accordance with the manufacturer's instructions. The Fab fragments were purified over fast-flow Protein A–Sepharose with use of 10% saturated sodium borate and 3 mol/L sodium chloride. Purified Fab in the unbound material was concentrated and dialyzed into 50 mmol/L sodium carbonate, pH 8.5. Biotinylation was performed with a 15-fold molar excess of D-biotinyl--aminocaproic acid N-hydroxysuccinimide ester (Boehringer Mannheim) in accordance with the manufacturer's instructions.

Polyclonal antibody.
New Zealand White rabbits were immunized every 3 weeks for 90 days with 50 µg of heparin-purified chondrex. The primary immunization was with chondrex in complete Freund's adjuvant, the first boost was in incomplete Freund's adjuvant, and all subsequent boosts were in 10 mmol/L sodium phosphate, 150 mmol/L sodium chloride, pH 7.2. Rabbits were bled 10 days after immunization. Antisera was purified by Protein A–Sepharose (Pierce) chromatography. Bound antibody was eluted with 0.1 mol/L glycine, pH 3.0, and neutralized by addition of 10% (by vol) 1.2 mol/L Tris, pH 8.5. Antibody-containing fractions were pooled and dialyzed into 50 mmol/L sodium phosphate, pH 7.5.

Alkaline phosphatase-labeled antibody. Purified polyclonal antibody was conjugated to alkaline phosphatase by cross-linking with sulfo-succinimidyl 4-(N-maleimidomethyl)-cyclohexane-1-carboxylate (sulfo-SMCC) and N-succinimidyl-3-(2-pyridyldithio)-propionate (SPDP). The antibody was incubated with a 10 mol/L excess of sulfo-SMCC diluted in 50 mmol/L sodium phosphate, pH 7.5, for 60 min at 24 °C with inversion. Purification was through G-25 SF resin (Pharmacia) equilibrated with 100 mmol/L sodium phosphate, pH 6.0. In parallel, alkaline phosphatase (Biozyme Labs) that had been dialyzed into 50 mmol/L sodium phosphate, pH 7.5, was incubated with a 10 mol/L excess of SPDP (Pierce) in acetonitrile for 30 min at 24 °C with inversion and then purified over a G-25 column (1.6 x 29 cm) equilibrated with 100 mmol/L sodium phosphate, pH 6.0. The SPDP-labeled alkaline phosphatase was reduced with dithiothreitol (Pierce) for 30 min and purified over G-25 once more. The derivatized antibody and alkaline phosphatase (3 mol of alkaline phosphatase per mole of antibody) were mixed together and incubated for 60 min at 24 °C with inversion. The reaction was quenched with 0.1 mol/L N-ethylmaleimide in acetonitrile to a final concentration of 2 mmol/L, and the mixture was purified through Superdex 200 (Pharmacia) size-exclusion gel (1.6 x 80 cm) equilibrated with 10 mmol/L sodium phosphate, 150 mmol/L sodium chloride, pH 7.0. Fractions were pooled on the basis of: absorbance at 280 nm, alkaline phosphatase activity, and signal-to-noise ratio.

Top Abstract Introduction Materials and Methods Chondrex (YKL-40).... Alkaline phosphatase-labeled... assay method study subjects Healthy subjects. Serum... statistical method Results Discussion References

 
For use in the Chondrex assay, 1.75 mg/L of the alkaline phosphatase-labeled antibody was added to 20 mg/L of the unlabeled antibody in assay buffer (20 mmol/L Tris, 150 mmol/L sodium chloride, 1 mmol/L magnesium chloride, 1 g/L sodium azide, 30 g/L bovine serum albumin, 1 mL/L Tween 20, 10 mmol/L benzamidine-HCl, pH 7.5). This dilution was optimized for a signal of ~2 absorbance units (2 A) at 405 nm with the highest-concentration calibrator.

Calibrators.
Calibrators were prepared by dilution of heparin-purified chondrex to 0, 20, 50, 100, 200, and 300 µg/L in calibrator base (10 mmol/L sodium phosphate, 100 g/L bovine serum albumin, 1 g/L sodium azide, pH 7.0).

Streptavidin plates.
Streptavidin-coated MaxiSorp(TM) Nunc-Immuno(TM) microtiter plates (from VWR) were prepared by incubating 150 µL (per well) of 10 mg/L streptavidin (Scripps Labs) in 50 mmol/L sodium phosphate, pH 7.2, for 16–24 h at room temperature. The liquid was then aspirated from the wells, and any remaining binding sites were blocked by incubation for 2 h at room temperature with 20 mmol/L Tris, 150 mmol/L sodium chloride, 15 g/L bovine serum albumin, 0.1 mL/L Tween, 1 g/L sodium azide, pH 7.5. After washing 3 times with wash buffer (20 mmol/L Tris, 150 mmol/L sodium chloride, 1 mL/L Tween-20, pH 7.5), plates were coated overnight (16–24 h) at room temperature with 200 µL/well of 150 g/L sucrose in 10 mmol/L sodium phosphate, 150 mmol/L sodium chloride, pH 7.2. The liquid was aspirated and the plates were dried overnight with low humidity at 25 °C. The plates were stored in foil pouches with desiccant at 4 °C until needed.

Other reagents.
Triglyceride solution (Pentex, Triglyceride Superstrate, no. 96–052) was purchased from Bayer. Hemoglobin was prepared from a lysate of washed erythrocytes. The hemoglobin concentration was determined spectrophotometrically at 578 nm ([heme, mg/L] = A_{578 nm}/0.0002278). Hydroxychloroquine was obtained from Geneva Pharmaceuticals.

	assay method

Top Abstract Introduction Materials and Methods Chondrex (YKL-40).... Alkaline phosphatase-labeled... assay method study subjects Healthy subjects. Serum... statistical method Results Discussion References

 
For the Chondrex assay, 20 µL of calibrator, control, or serum sample and 100 µL of the biotin–Fab conjugate (0.75 mg/L in assay buffer) were incubated for 1 h at room temperature in duplicate streptavidin-coated wells. After the wells were washed 4 times with wash buffer, 100 µL of alkaline phosphatase-labeled polyclonal antibody was added to each well and incubated for 1 h at room temperature. After 4 washes with wash buffer, 100 µL of substrate solution (2 g/L p-nitrophenyl phosphate in 1 mol/L diethanolamine, 1 mmol/L magnesium chloride, pH 9.9) was added to each well and incubated for 1 h at room temperature. The reaction was quenched with 100 µL of 1 mol/L sodium hydroxide per well, and the absorbance was read at 405 nm. A calibration curve was generated by using a linear fit of the curve constructed (linear regression by least-squares method), and the chondrex concentration in the samples was calculated from the calibration curve. Clinical samples were assayed in duplicate; those samples with results greater than the calibrator range were diluted 5-fold in calibrator base and retested.

	study subjects

Top Abstract Introduction Materials and Methods Chondrex (YKL-40).... Alkaline phosphatase-labeled... assay method study subjects Healthy subjects. Serum... statistical method Results Discussion References

 
All subjects gave their informed consent before participation.

Healthy subjects. Serum samples were collected from 329 apparently healthy individuals, 226 women and 103 men, mean age 37.1 ± 8.6 years (median 36). The women were ages 24–60 years (mean 37.7 ± 8.9 years, median 36) and the men 21–60 years (mean 35.6 ± 7.9 years, median 35). All subjects were thought to be free from articular, bone, liver, endocrine, or other chronic disorders. None was currently taking any medication known to modify arthritic disease or influence joint metabolism (e.g., slow-acting or DMARD).

Top Abstract Introduction Materials and Methods Chondrex (YKL-40).... Alkaline phosphatase-labeled... assay method study subjects Healthy subjects. Serum... statistical method Results Discussion References

 
Non
-joint-disease control subjects. Serum samples from diabetic patients visiting a physician were used as a control population of individuals with non-joint- and noninflammatory chronic illness. There were 35 diabetes patients: 25 women and 10 men, ages 23–91 years (mean 51 ± 16 years, median 49).

Joint disease subjects
. Serum samples from individuals who had been previously diagnosed with active and inactive RA as well as OA according to American College of Rheumatology criteria were obtained from consecutive individuals visiting a physician during the collection interval. Serum was stored at -70 °C until use. The active RA group consisted of 56 patients, 44 women and 12 men, ages 20–95 years (mean 55 ± 18 years, median 56) at various clinical stages of RA. The inactive RA group consisted of 9 patients, 7 women and 2 men, ages 29–76 years (mean 45 ± 17 years, median 38). The OA group consisted of 27 patients, 22 women and 5 men, ages 26–90 years (mean 65 ± 15 years, median 67).

Joint disease subjects on DMARD therapy
. Serum samples and corresponding clinical data were obtained from 20 patients who had received DMARD therapy (methotrexate alone, sulfasalazine and hydroxychloroquine in combination, or all three drugs) at regular intervals over at least a year. These patients were part of a larger study described in O'Dell et al. (7). Patients were classified by response, according to a modified version of the American College of Rheumatology criteria. Responders had 20% improvement in joint count and 20% improvement in 3 of the 4 remaining variables tested (patient global assessment, physician global assessment, erythrocyte sedimentation rate or C-reactive protein, and duration of morning stiffness). Moderate responders had 20% improvement in joint count and 20% improvement in 1 or 2 of the remaining 4 variables. Nonresponders had <20% improvement in joint count. Two patients could not be classified in any of these categories because of fluctuations in response criteria; they were excluded from further analysis.

	statistical method

Top Abstract Introduction Materials and Methods Chondrex (YKL-40).... Alkaline phosphatase-labeled... assay method study subjects Healthy subjects. Serum... statistical method Results Discussion References

 
The serum chondrex values for the patient groups were expressed as medians, means, and SDs. To determine whether the values were gaussianly distributed in apparently healthy persons, we performed the Shapiro–Wilk test for normality. The distribution of chondrex values for men, women, and combined populations were skewed and nongaussian. The serum chondrex values for the healthy adults were also expressed as medians, means, and SDs. The reference interval of chondrex values for healthy adults was reported as the nonparametric central 90% interval. Comparisons between subjects with joint disease or nonjoint disease and the healthy population were evaluated with the Mann–Whitney U-test because the populations were not gaussianly distributed. For these analyses, P <0.05 (two-tailed test) was considered significant. Statistical calculations were performed by using StatMost software (DataMost Corp.). To assess the clinical performance of the Chondrex assay, receiver operating characteristic (ROC) curves were constructed (8), and ROC curve analyses were performed with GraphROC software (Maxiwatti Oy). The cutoff value (maximal accuracy) was determined by calculating the point that gave the shortest distance from the upper left corner of the ROC curve (9).

	Results

Top Abstract Introduction Materials and Methods Chondrex (YKL-40).... Alkaline phosphatase-labeled... assay method study subjects Healthy subjects. Serum... statistical method Results Discussion References

 
assay characteristics
Calibration curve.
A typical Chondrex assay calibration curve with linear regression fit is shown in Fig. 1 . The nonspecific binding, determined by the absorbance of the zero calibrator, is less than 0.2 A at 405 nm. The minimum detectable concentration, determined by interpolating the mean of 30 replicates of the zero calibrator plus 3 SD, was 8 µg/L. The limit of detection by linear dilution was 20 µg/L.

View larger version (14K): [in this window] [in a new window]   Figure 1. Typical calibration curve of the Chondrex assay. Error bars represent one standard deviation on either side of the mean (y = 0.00593x + 0.142, r2 = 0.999).

Precision.
The interassay precision was determined by triplicate measurements of two serum samples in 27 assays over 11 test days. For chondrex values of 208 ± 7.7 µg/L and 79 ± 2.2 µg/L (mean ± SD), the interassay CVs were 3.7% and 2.8%, respectively. The within-run and total CVs (Table 1 ) were determined by replicate measurements (n = 12) of three serum samples over 3 days. For serum samples with mean chondrex concentrations of 36, 86, and 177 µg/L, the average within-run and total CVs were 3.6% and 5.4%, respectively.

Assay performance.
Three serum samples with various endogenous chondrex values (A, B, and C) and one serum sample to which purified chondrex had been added (D) were diluted with the zero calibrator. There was good agreement between observed and expected values (Table 2 ). The mean ± SD recovery for all samples was 102% ± 5%. Addition of 34 µg/L of chondrex to each of nine serum samples resulted in an analytical recovery of 98% ± 11%.

stability studies
Sample stability.
The freezing and thawing of serum as many as 6 times did not significantly affect the measured chondrex concentration in the 5 samples tested. Recovery after 6 freeze/thaw cycles was 106% ± 6% (mean ± SD) of the initial unfrozen sample values. Storage of serum samples for 9 days at 4 °C did not significantly affect the measured chondrex concentration in the 8 samples tested, the average recovery being 107% ± 8%.

interference studies
Possible interference in the determination of chondrex by common compounds in human serum was investigated by adding 300 mg/L bilirubin, 5 g/L hemoglobin, and 30 g/L triglyceride into low, medium, and high chondrex serum pools and comparing the recoveries with those for controls to which only buffer had been added. No significant interference was seen with the high concentrations of these serum components, the mean recovery of the 3 serum pools being 100% ± 3%, 100% ± 3%, and 96% ± 2% after addition of bilirubin, hemoglobin, and triglyceride, respectively. Total serum protein, tested at 7 concentrations between 30 and 120 g/L, did not affect quantification of chondrex; the mean recovery for the protein concentrations tested was 97% ± 4%.

Possible interference in the determination of chondrex by pharmaceutical reagents commonly prescribed to arthritis patients was investigated by adding various concentrations of 23 drugs to a serum pool. Pharmaceutical interference was determined in an iterative manner, starting at a maximum concentration of 10 g/L unless prohibited by the solubility of the additive. Interference was defined as a >10% change in chondrex recovery in comparison with a solvent-dilution control. The highest concentration of the various drugs that did not cause interference with chondrex recovery is listed in Table 3 . All concentrations listed are at least 3 times (range, 3–5000) the expected peak plasma concentration.

clinical performance
Chondrex values in healthy adults.
The serum from 329 healthy adults (226 women, 103 men) was tested with the Chondrex assay. Fig. 2 illustrates the individual concentration of serum chondrex as a function of age in both men and women. The chondrex values decreased slightly with age in women (P <0.05), but there was no significant trend between age and Chondrex values for the men. As shown in Fig. 3 , the distribution of the values was skewed and nongaussian for both women and men. The nonparametric reference intervals were 25–93 µg/L for women, 24–125 µg/L for men, and 25–95 µg/L for all healthy adults combined (indicated by shading in Fig. 4 ). The median chondrex value for the 103 men was 45.6 µg/L (mean, 55.4 ± 35.0) and for the 226 women was 41.15 µg/L (mean, 48.6 ± 24.2), but the differences were not statistically significant (P = 0.207). The median chondrex concentration for all healthy adults combined was 42.9 µg/L (mean ± SD, 50.7 ± 28.2 µg/L) (Table 4 ).

View larger version (24K): [in this window] [in a new window]   Figure 2. Distribution of Chondrex values in healthy women () and men () by age.

View larger version (19K): [in this window] [in a new window]   Figure 3. Frequency distributions of chondrex values in healthy women (A) and men (B).

View larger version (31K): [in this window] [in a new window]   Figure 4. Chondrex values for individual patients. The shaded area represents the combined reference interval for healthy men and women (central 90% interval). Mean values are indicated by horizontal lines.

Comparisons between groups.
Sera from individuals with active joint disease (RA and OA) and from individuals with inactive RA or nonjoint autoimmune disease (diabetes) were tested with the Chondrex assay. The data are summarized in Table 4 and the individual values are plotted in Fig. 4 . The reference interval for healthy adults is represented by shading (Fig. 4 ). Chondrex values for patients with active RA were significantly higher than in healthy adults, patients with inactive RA, and patients with diabetes (P <0.0001). Chondrex values for patients with OA were also significantly greater than in healthy adults (P <0.0001), inactive RA (P = 0.018), and diabetes patients (P = 0.019). No differences were observed between healthy adults and subjects with inactive RA (P = 0.932) or between active RA and OA subjects (P = 0.487).

ROC curves were constructed to compare chondrex values obtained in active RA and OA subjects with those of healthy adults, and to compare values in active and inactive RA subjects (Fig. 5 ). The area under the ROC curve for active RA patients vs healthy adults was 0.864 (SE 0.030) and that for OA patients vs healthy adults was 0.754 (SE 0.068). The ROC curve for active RA vs inactive RA patients yielded an area under the curve of 0.857 (SE 0.065). A cutoff of 86 µg/L for active RA was chosen by determining the closest point to the upper left corner of the ROC curve. At this cutoff, the Chondrex assay demonstrates a clinical sensitivity of 69.6% (95% confidence interval 57.9–79.6) and a specificity of 91.5% (95% confidence interval 88.5–93.9) for active RA compared with healthy adults.

View larger version (18K): [in this window] [in a new window]   Figure 5. ROC curves for comparisons between chondrex values in active RA (heavy line) or OA (thin line) and healthy subjects, or between active and inactive RA (dotted line).

RA patients treated with DMARD therapy.
Chondrex values were obtained on baseline and follow-up samples from 18 RA patients participating in a clinical trial of the effects of methotrexate alone, sulfasalazine and hydroxychloroquine in combination, or all three drugs. The mean of the chondrex values at baseline, 144.5 ± 85.1 µg/L, was significantly greater than in healthy adults (P <0.0001). As can be seen in Fig. 6 , chondrex values for patients in the responder group decreased to within the reference interval, except for three patients whose chondrex values at baseline exceeded 3 times the mean for healthy adults. All of these exhibited decreased chondrex values at the first time point, and values in 2 patients decreased steadily while on treatment. The mean chondrex value for all responders at the first time point was -21% from baseline, and baseline values were negatively correlated with the degree of change (r = 0.67, P = 0.03). The moderate responder group had a decrease of 13% from baseline in chondrex values at the first time point. Chondrex values in the nonresponder group increased 13% at the first time point, and all of these patients had values exceeding the reference interval at completion of the study.

View larger version (20K): [in this window] [in a new window]   Figure 6. Chondrex values in subjects being treated with methotrexate alone, sulfasalazine and hydroxychloroquine in combination, or all three drugs. Study subjects were classified as responders (A), moderate responders (B), or nonresponders (C) on the basis of modified American College of Rheumatology criteria. Methotrexate was added to sulfasalazine and hydroxychloroquine for one subject (*). The shaded band represents the reference interval for healthy adults.

	Discussion

Top Abstract Introduction Materials and Methods Chondrex (YKL-40).... Alkaline phosphatase-labeled... assay method study subjects Healthy subjects. Serum... statistical method Results Discussion References

 
Chondrex, a major secretory protein of human chondrocytes and synovial cells, has been previously shown to be increased in the synovial fluid and serum of individuals with joint and cartilage disease (3)(6). We have developed a simple, fast, two-site, sandwich-type ELISA that uses a streptavidin-coated microplate well, a biotinylated-Fab monoclonal capture antibody, and an alkaline phosphatase-labeled polyclonal detection antibody. The assay involves three 1-h incubation steps, is carried out at room temperature, and does not require sample pretreatment.

The Chondrex assay has good within-run and between-run precision, good linearity upon dilution of samples, and good analytical recovery. Chondrex concentrations in serum samples are stable at 4 °C for as long as 9 days and as many as 6 freeze/thaw cycles. The measurement of chondrex in serum is not affected by abnormally high concentrations of common serum components such as bilirubin, hemoglobin, triglycerides, and total protein. In addition, pharmaceutical compounds commonly prescribed for the treatment of arthritis did not affect assay performance when added to serum.

We have established reference intervals for a large group of healthy women and men. Because of the lack of availability of radiographs of the study subjects and the prevalence of undiagnosed but radiographically confirmed evidence of OA in the elderly (10), the maximum age in the reference group was restricted to 60 years. Chondrex values in women and men were not statistically different, so the data were combined for analyses. This is consistent with a study by Johansen et al. (6), which found no significant difference in serum chondrex concentration between men and women. In our population of healthy women, chondrex decreased slightly with age, but there was no significant trend between age and chondrex values for the men. Johansen et al. (6) reported no difference in serum concentrations of chondrex between age groups in children and adults younger than 70 years.

Chondrex values in healthy adults were compared with those in individuals with joint disease. Mean chondrex values in patients with active RA were significantly greater than the mean concentrations seen in both healthy adults and patients with inactive RA. These results are consistent with two studies by Johansen et al. (3)(11) and suggest that chondrex may be useful in assessing disease activity in RA patients. ROC analysis indicates very good diagnostic accuracy of the assay for active RA and may facilitate the use of chondrex as an adjunct in the diagnosis of inflammatory disease.

Chondrex values were also significantly higher in OA patients than in healthy adults, and ROC analysis indicates the diagnostic accuracy is good. However, chondrex values in OA patients were not significantly different from patients with active RA. Similar findings have been reported by Johansen et al. (3)(11). This suggests that this marker is not likely to support a differential diagnosis between RA and OA.

It is not entirely unexpected that the concentration of chondrex is increased in both RA and OA patients, given the resulting manifestation of accelerated cartilage degradation and joint damage in both diseases. Hakala et al. (4) reported detecting chondrex mRNA in preparations from cartilage of RA patients but not in preparations from normal articular cartilage. Also, mRNA was detected from surgical specimens of inflamed synovia as well as cartilage specimens from OA patients. The expression of chondrex in diseased joints may be a response to an altered tissue environment. The rapid induction of chondrex production in normal human cartilage after introduction to cell culture conditions (4) supports a function for chondrex in the cartilage-remodeling process. One postulate is that chondrex functions as a glycosidic bond hydrolase involved in tissue remodeling (3)(4)(6)—a postulate supported by amino acid sequence homology of chondrex to the bacterial chitinase protein family (3)(4)(6). Because chitin itself is not found in vertebrates, however, and because chitinase activity of chondrex could not be demonstrated, Johansen et al. (3) speculate that divergent evolution of an ancestral chitinase altered the specificity of the vertebrate enzyme so that it now cleaves a different glycosidic linkage in a yet-unidentified macromolecule found in articular cartilage.

It has also been postulated that chondrex is an autoantigen target of the immune response in RA, in light of the presence of sequences that bind to the DR4 peptide-binding motif and elicit mononuclear cell proliferation (12). Corollary studies in OA patients are not available but, when performed, will perhaps help to clarify immune response differences and the subsequent disease courses of OA and RA.

In patients treated with DMARD therapy, decreasing chondrex values reflected the clinical improvement observed in responders, whereas Chondrex values were maintained or increased in nonresponders. Responders with the highest baseline chondrex values failed to normalize, though all showed a large decrease at the first timepoint examined after the start of treatment. These responses are consistent with those seen in studies of the effects of corticosteroid injections in patients with knee RA (13) and in RA patients treated with methotrexate, sulfasalazine, or penicillamine (14). The largest decreases in chondrex values were observed in the responders with the highest baseline values. The assay may be useful in determining which patients are most likely to benefit from therapy. Whether it can aid in the selection of therapy needs to be addressed in larger studies.

Clinical response in the present study was defined on the basis of improvement in markers of inflammation (joint count, laboratory indicators reflecting systemic inflammation, and so forth). However, chondrex values remained increased in some of the responders in spite of this improvement. As seen here and in previous studies, chondrex values are increased in patients with both degenerative and inflammatory joint disease. These findings suggest that chondrex may reflect aspects of joint destruction in addition to inflammation. Longitudinal studies with measures of articular cartilage and periarticular bone destruction will be required to define the prognostic capabilities of this marker.

	Acknowledgments

 
We thank Robert R. Henry, University of California at San Diego School of Medicine, for providing the serum samples from diabetic patients used in this study.

	Footnotes

 
¹ Nonstandard abbreviations: RA, rheumatoid arthritis; OA osteoarthritis; sulfo-SMCC, sulfo-succinimidyl 4-(N-maleimidomethyl)-cyclohexane-1-carboxylate; SPDP, N-succinimidyl-3-(2-pyridyldithio)-propionate; DMARD, disease-modifying antirheumatic drug.

	References

Top Abstract Introduction Materials and Methods Chondrex (YKL-40).... Alkaline phosphatase-labeled... assay method study subjects Healthy subjects. Serum... statistical method Results Discussion References

 

1. Rejman JJ, Hurley WL. Isolation and characterization of a novel 39 kilodalton whey protein from bovine mammary secretions collected during the nonlactating period. Biochem Biophys Res Commun 1988;150:329-334. [ISI][Medline] [Order article via Infotrieve]
2. Johansen JS, Williamson MK, Rice JS, Price PA. Identification of proteins secreted by human osteoblastic cells in culture. J Bone Miner Res 1992;7:501-512. [ISI][Medline] [Order article via Infotrieve]
3. Johansen JS, Jensen HS, Price PA. A new biochemical marker for joint injury. Analysis of YKL-40 in serum and synovial fluid. Br J Rheumatol 1993;32:949-955. [Abstract/Free Full Text]
4. Hakala BE, White C, Recklies AD. Human cartilage gp-39, a major secretory product of articular chondrocytes and synovial cells, is a mammalian member of a chitinase protein family. J Biol Chem 1993;268:25803-25810. [Abstract/Free Full Text]
5. Nyirkos P, Golds EE. Human synovial cells secrete a 39 kDa protein similar to a bovine mammary protein expressed during the non-lactating period. Biochem J 1990;268:265-268.
6. Johansen JS, Hvolris J, Hansen M, Backer V, Lorenzen I, Price PA. Serum YKL-40 levels on healthy children and adults. Comparison with serum and synovial fluid levels of YKL-40 in patients with osteoarthritis or trauma of the knee joint. Br J Rheumatol 1996;35:553-559. [Abstract/Free Full Text]
7. O'Dell JR, Haire CE, Erikson N, Drymalski W, Palmer W, Eckhoff PJ, et al. Treatment of rheumatoid arthritis with methotrexate alone, sulfasalazine and hydroxychloroquine, or a combination of all three medications. N Engl J Med 1996;334:1287-1291. [Abstract/Free Full Text]
8. Zweig MH, Campbell G. Receiver-operating characteristic (ROC) plots: a fundamental evaluation tool in clinical medicine [Review]. Clin Chem 1993;39:561-577. [Abstract/Free Full Text]
9. Guechot J, Laudat A, Loria A, Serfaty L, Poupon R, Giboudeau J. Diagnostic accuracy of hyaluronan and type III procollagen amino-terminal peptide serum assays as markers of liver fibrosis in chronic viral hepatitis C evaluated by ROC curve analysis. Clin Chem 1996;42:558-563. [Abstract/Free Full Text]
10. Burger H, van Daele PL, Odding E, Valkenberg HA, Hofman A, Grobbee DE, et al. Association of radiographically evident osteoarthritis with higher bone mineral density and increased bone loss with age. The Rotterdam study. Arthritis Rheum 1996;39:81-86. [ISI][Medline] [Order article via Infotrieve]
11. Johansen JS, Hansen M, Stoltenberg M, Hvolris J, Florescu A, Price PA, Hørslev-Petersen K. Does serum YKL-40 reflect disease activity in rheumatoid arthritis and osteoarthritis? [Abstract]. Arthritis Rheum 1995;38(9 Suppl):S217.
12. Verheijden GFM, Rijnders AWM, Bos E, Coenen-de Roo CJJ, van Stavern CJ, Miltenburg AMM, et al. Human cartilage glycoprotein-39 as a candidate autoantigen in rheumatoid arthritis. Rheum 1997;40:1115-1125.
13. Volck B, Johansen JS, Stoltenberg M, Østergaard M, Price PA, Lorenzen I. Serum YKL-40 reflects pharmacological induced reduction of synovial inflammation of the knee joint in patients with rheumatoid arthritis [Abstract]. Arthritis Rheum 1996;39:S155.
14. Johansen JS, Stoltenberg M, Hansen M, Hansen TM, Price PA, Gøtzche PC. Serum YKL-40 reflects disease relapse after withdrawal of SAARD in patients with rheumatoid arthritis [Abstract]. Arthritis Rheum 1996;39:S155.