A comparison of different purification methods of aggrecan fragments from human articular cartilage and synovial fluid

Matrix Biology 29 (2010) 74–83

Contents lists available at ScienceDirect

Matrix Biology j o u r n a l h o m e p a g e : w w w. e l s e v i e r. c o m / l o c a t e / m a t b i o

A comparison of different purification methods of aggrecan fragments from human articular cartilage and synovial fluid André Struglics ⁎, Staffan Larsson Department of Orthopaedics, Clinical Sciences Lund, Lund University, Lund, Sweden

a r t i c l e

i n f o

Article history: Received 31 March 2009 Received in revised form 18 July 2009 Accepted 28 August 2009 Keywords: Aggrecan Cartilage sGAG Purification Synovial fluid

a b s t r a c t In the study of aggrecan fragmentation several methods to extract and purify aggrecan from cartilage and synovial fluid (SF) are used. This work compares and evaluates the effectiveness for purification of aggrecan of the most commonly used methods by the ratio of sulfated glycosaminoglycan (sGAG) to protein and by fragment analysis by Western blot. A novel method for purification of aggrecan fragments from SF by boiling (Boiled SF) is also presented. Of the sGAG extracted from cartilage by guanidinium, 66% was recovered by associative–dissociative cesium chloride density gradient centrifugation (A1D1–D3) with a 9 times higher ratio of sGAG to protein in the A1D1 fraction. Although less enriched in aggrecan, the Western blot aggrecan pattern of the guanidinium extracted sample resembled that of the combined patterns of the A1D1, A1D2 and A1D3 fractions. The recoveries of sGAG from SF purified by anion chromatography and Alcian blue precipitation were around 50%, while the recoveries were over 80% in the associative or dissociative density gradient fractions (A1 and D1) and Boiled SF. The purification compared to neat SF ranged from 9 times in boiled SF to 1800–1900 times in Alcian blue and D1 samples. To obtain reliable results when analyzing synovial fluid aggrecan fragments by Western blot, purification was necessary. The immuno-pattern of anion chromatography purified SF resembled the patterns of A1 and D1, while the pattern of Boiled SF resembled the D1 sample. This work suggests that aggrecan fragments extracted from cartilage by guanidinium need no further purification to be analyzed by Western blot, whereas aggrecan fragments in SF are best analyzed in the A1 and D1 fractions or in the Boiled SF sample. © 2009 Elsevier B.V. All rights reserved.

1. Introduction Aggrecan is one of the major macromolecules in articular cartilage and plays a central role in cartilage compressibility and elasticity. In arthritis and joint injury there is destruction of aggrecan by proteolysis, mainly by matrix metalloproteases (MMPs) and aggrecanases (Sandy, 2003), which generates a complex aggrecan fragmentation pattern (Sandy and Verscharen, 2001; Struglics et al., 2006a; Sumer et al., 2007a).

Abbreviations: AEBSF, 4-(2-aminoethyl)-benzenesulfonyl fluoride; ADAMTS, a disintegrin and metalloproteinase with thrombospondin motifs; CS, chondroitin sulfate region; DEAE, diaminoethylcellulose anion; DMEM, Dulbecco's modified Eagle's medium; DMMB, dimethyl-methylene blue; EACA, 6-aminohexonic acid; Ft, column flow through; sGAG, sulfated glycosaminoglycan; Gu-PG, guanidinium extracted proteoglycan; IGD, interglobular domain; KS, keratan sulfate region; MMP, matrix metalloproteinase; Mw, molecular weight; Na, not applicable; Nd, not determined; OA, osteoarthritis; PMSF, phenylmethylsulfonyl fluoride; rS, Spearman rank's order correlations; Seph, Sepharose; SF, synovial fluid; TCA, trichloroacetic acid; WB, Western blot; wt, weight. ⁎ Corresponding author. Department of Orthopaedics, Lund University, BMC C12, SE-221 84 Lund, Sweden. Tel.: +46 46 222 0762; fax: +46 46 211 3417. E-mail address: [email protected] (A. Struglics). 0945-053X/$ – see front matter © 2009 Elsevier B.V. All rights reserved. doi:10.1016/j.matbio.2009.08.006

Quantitative analyses of human and animal aggrecan proteolytic fragments have been conducted by ELISA (Moller et al., 1994; Fosang et al., 1996; Lohmander et al., 1999; Valiyaveettil et al., 2005; Pratta et al., 2006a 14; Carter et al., 2007; Sumer et al., 2007b), while for qualitative analysis of cartilage and synovial fluid aggrecan fragments immunohistochemistry (Lark et al., 1997; Lee et al., 1998; Bayliss et al., 2001) and Western blot methods (Sandy and Verscharen, 2001; Malfait et al., 2002; Struglics et al., 2006a) have been applied. Purification is necessary for analysis of aggrecan fragments in cartilage, and sometimes also for analysis in synovial fluid. The most common methods for these purifications are: guanidinium extraction alone or together with associative and/or dissociative CsCl density gradient centrifugation (Hascall and Kimura, 1982; Heinegard and Sommarin, 1987; Hascall et al., 1994; Vilim and Fosang, 1994), and anion chromatography (Hascall and Kimura, 1982; Heinegard and Sommarin, 1987; Sandy et al., 1991; Hascall et al., 1994; Vilim and Fosang, 1994; Sandy and Verscharen, 2001). A specific precipitation by Alcian blue has also been suggested as a method for purifying proteoglycans (Karlsson and Björnsson, 2001). In the continued work towards a better understanding of cartilage destruction in arthritis and joint injuries, methods that allow a detailed mapping of aggrecan fragment patterns in animal and human samples are needed. We here evaluate different aggrecan purification

A. Struglics, S. Larsson / Matrix Biology 29 (2010) 74–83

75

Table 1 Purification of aggrecan fragments from synovial fluid. Purification methods

Purified samples

sGAG recovery (%)

Protein recovery (%)

Purification level (µg sGAG/ mg protein)

Purification factor

– Alcian bluea Boiling SFb Blue Sepha Q-Sephb Blue Seph and Q-Sephb Blue Seph and Mono-Qa Density centrifugationb

SF pools Precipitate Boiled SF Ft 2 M eluate 2 M eluate 3 M eluate D1 A1

100 56 83 67 15 48 53 86 92

100 0.03 9 6.2 1 0.1 1 0.05 0.3

3.1, 5.1 6500 25 28 28 925 96 5770 732

1 1940 9 10 9 288 49 1798 300

Aggrecan fragments were purified from human synovial fluid OA-pools SF-1 (neat SF, 3.1 µg sGAG/mg protein) or SF-2 (lavage, 5.1 µg sGAG/mg protein) by different methods. The sGAG and protein recoveries and the purification factor of the purified samples are normalized against the SF OA-pools. Except for the D1 preparation showing average values (n = 3), the data are from single experiments. 2 M/3 M eluate, 2 M or 3 M NaCl batch eluate of fraction; A1; associative bottom fraction; D1, dissociative bottom fraction; Ft, column flow through; Seph, Sepharose. a Purification from SF-2 OA-pool. b Purification from SF-1 OA-pool.

methods from human cartilage and synovial fluid, and present a new method for quick aggrecan purifications from synovial fluid. The results presented can be used as a guide for the choice of purification method in qualitative and quantitative analysis of aggrecan fragments from cartilage and synovial fluid. 2. Results 2.1. Purification of aggrecan fragments from human synovial fluid Of the different purification methods tested the highest sGAG yields (>80%) were obtained in the D1 and A1 samples, and in the Boiled SF sample, while purification by chromatography or by Alcian blue precipitation gave moderate (15–67%) sGAG yields (Table 1). Even though the sGAG yield increased when running synovial fluid over a Blue-Sepharose column prior to anion chromatography compared to without this cleaning step (mainly removing albumin), the sGAG yields were still moderate. Using urea as a chelating agent during the Q-Sepharose chromatography steps did not considerably change the sGAG yields or the purification levels (data not shown). The sGAG concentration was used as a measure of the amount of high molecular weight (Mw) aggrecan fragments, and a sGAG/total protein ratio was used as a measure of the purification level of these fragments in different samples. Alcian blue precipitation and the D1 fraction had the highest purification levels (6500 and 5770 µg sGAG/ mg protein, respectively), and these samples were 1800–1900 times purified compared to neat synovial fluid (Table 1). Aggrecan purification using either Blue Sepharose or Q-Sepharose methods separately resulted in low purification levels (28 µg sGAG/mg protein), while combining these methods resulted in a 33-fold higher purification level. A similar high purification level was found in the A1 fraction. Purification with either Blue Sepharose in combination with Mono-Q chromatography, or boiling the SF resulted in only modest

purification levels of 96 and 25 µg sGAG/mg protein, respectively (Table 1). 2.2. Purification of aggrecan fragments from human cartilage Guanidinium extraction alone or in combination with density gradient centrifugation was used for purification of aggrecan fragments from human cartilage. The 4 M guanidinium extracted proteoglycan (Gu-PG) sample contained 29 mg sGAG/g wet weight cartilage in average (Table 2), ball mix extraction resulted in 24 mg sGAG/g wet weight cartilage and the polytron extraction gave 34 mg sGAG/g wet weight cartilage. According to previous findings guanidinium extractions contain approximately 70% of the total sGAG (Vilim and Fosang, 1994) and 80– 85% of the uronic acid (Bayliss and Ali, 1978) from the cartilage tissue. The sGAG yield for the A1D1 fraction was surprisingly moderate, with an average of 55% for the ball mix and polytron extraction methods (Table 2). Even though some of the sGAG was in the A1D2 fraction, the total sGAG recovery was only 66% in average (Table 2), where the ball mix extraction gave 80% sGAG recovery and the polytron extraction 52%. The highest aggrecan purification level from the cartilage was observed for the A1D1 fraction (5518 µg sGAG/mg protein) followed by the A1D2 (1775 µg sGAG/mg protein) and the Gu-PG sample (641 µg sGAG/mg protein) (Table 2). The purification of aggrecan fragments from knee cartilage was increased 9-fold and 3-fold, respectively, in the A1D1 and A1D2 fractions compared to the Gu-PG sample, while it was decreased in the A1D3 fraction (Table 2). 2.3. Western blot immuno-pattern of aggrecan fragments in purifications from synovial fluid and cartilage The aggrecan fragments in the purified samples were analyzed further by Western blot.

Table 2 Purification of aggrecan fragments from cartilage. Purification methods

Purified samples

Dry wt/wet wt cartilage (mg/g)

sGAG/wet wt cartilage (mg/g)

sGAG recovery Protein recovery Purification level Purification (%) (%) (µg sGAG/mg protein) factor

Extraction Extraction and density centrifugation

Gu-PG A1D1 A1D2 A1D3

Na 22.7 10.2 1.8

29.0 (34, 24)a 15.5 2.9 0.1

100 55.4 10.3 0.4

100 5.8 4.0 1.3

641 5518 1775 183

1 8.7 2.8 0.3

Aggrecan fragments were purified from the human OA knee cartilage pool by guanidinium extraction alone, collecting the guanidinium extracted proteoglycan sample (Gu-PG), or together with CsCl density centrifugation collecting the A1D1, A1D2 and A1D3 fractions. The purification factor (calculated from the purification levels) and the sGAG and protein recoveries of the CsCl density samples are normalized against the Gu-PG sample. Average data from two separate experiments (polytron and ball mill extractions) are presented. Na, not applicable. wt, weight. a Average (values for polytron and ball mill extractions, respectively).

76

A. Struglics, S. Larsson / Matrix Biology 29 (2010) 74–83

The detected aggrecan fragments were either verified both N- and Cterminally by aggrecan neoepitope or structural recognizing antibodies, or by estimation using a calculation model (Struglics et al., 2006b).

2.3.1. Synovial fluid aggrecan fragments The synovial fluid aggrecan Western blot immuno-pattern obtained with different purifications was compared, and related to the pattern from cartilage explant media. No clear aggrecan immunobands was obtained for the neat SF, Blue Sepharose or the Alcian blue purified samples (data not shown). The Q-Sepharose and the Blue Sepharose/ Mono-Q purified aggrecan samples had in general the same immunopattern as was observed in the Blue Sepharose/Q-Sepharose sample (data not shown). The Boiled SF and the Blue Sepharose/Q-Sepharose sample contained high Mw (>250 kDa) sGAG containing aggrecan fragments: G1-KEEE, IGD-KEEE, ARGS-CS1 (288 kDa) and ARGS-SELE (Fig. 1A–D), although no clear high Mw G1 fragments was seen in the Boiled SF sample. The 340 kDa IGD-KEEE fragment was N-terminally characterized by a positive IGD immunoreaction and by no reaction with G1, ARGS and FFGV antibodies (data not shown). The Boiled SF and the Blue Sepharose/Q-Sepharose sample also contained medium size (100–200 kDa) fragments: ARGS-CS1 (130–160 kDa), CS1-KEEE, GRGT-G3, GLGS-G3 and AGEG-G3 (Fig. 1B, D, E). These large and medium size fragments were also found in the SF A1 and D1 fractions, and in the explant media. The Boiled SF contained no low (<100 kDa) Mw aggrecan fragments, while the Blue Sepharose/Q-Sepharose sample contained: G1-IPEN, G1-TEGE and ARGS-KS fragments (Fig. 1A, B), confirmed with IPEN and TEGE antibodies (data not shown). The low Mw ARGS-KS fragments corresponded to 4% of the

total ARGS signal in the Blue Sepharose/Q-Sepharose sample, and these fragments were also found in similar low amounts in the A1 fraction (1%) and in the explant media (2%) (Fig. 1B). These data show that the SF aggrecan immuno-pattern of the Blue Sepharose/Q-Sepharose purified sample is similar to the combined patterns of the SF A1 and D1 fractions, and to the pattern obtained from the cytokine stimulated cartilage explant media (Fig. 1, Table 3). These data also show that except for some of the G1 fragments the aggrecan immuno-pattern of the Boiled SF is similar to the pattern of the SF D1 sample.

2.3.2. Cartilage aggrecan fragments The cartilage aggrecan Western blot immuno-pattern from different purifications was compared, and related to the pattern obtained from protease in vitro digested cartilage A1D1 samples. The Gu-PG sample contained full length G1–G3 monomer (Fig. 2A, E), and other high Mw (>250 kDa) sGAG containing aggrecan fragments: G1-SELE, G1-KEEE, G1-CS2, G1-CS1 and IGD-KEEE (Fig. 2A, C, D). These high Mw sGAG containing fragments were also found in the A1D1 and A1D2 fractions. ARGS fragments were not detected in any of the cartilage in vivo samples (Gu-PG and A1D1-D3) although these fragments were found in the ADAMTS-4 in vitro digested A1D1 sample (Fig. 2B, C). The GuPG sample also contained some medium size (100–200 kDa) fragments: GLGS-G3 (Fig. 2E) and LGQR-G3 (detected with LGQR antibody, data not shown). These and other medium size G3 fragments were found in the A1D2 fraction as well as in ADAMTS-4 in vitro digested A1D1 samples (Fig. 2E, F). Further, the Gu-PG sample contained low Mw aggrecan fragments (i.e. <100 kDa with no or few sGAG substitutions): G1-IPEN, G1-TEGE and G1-KS. These low Mw fragments were also found in the

Fig. 1. Western blot of purified synovial fluid samples. Human synovial fluid aggrecan fragments were purified by CsCl density centrifugation (collecting D1 and A1 fractions), boiling SF collecting the supernatant and by Blue Sepharose/Q-Sepharose chromatography. Medium from 48 h cytokine stimulated cartilage explant cultures was used for comparison. Deglycosylated samples (1–7 µg sGAG/lane) were separated by SDS-PAGE and probed by different antibodies. Several Western blot experiments were conducted and different exposure times (5 s–10 min) on film or in a luminescence image analyzer were used to develop clear images (strips). The images show representative Western blots from full size blotted gels. The two different D1–A1 strips probed with ARGS antibodies have different exposure times. Bands marked by ⁎ are considered to be false-positive immunobands since they could not be blocked by the corresponding immunogen peptide. G1-TEGE (A) dimer was confirmed by corresponding antibodies (data not shown).

A. Struglics, S. Larsson / Matrix Biology 29 (2010) 74–83

77

Table 3 Aggrecan fragments purified from human knee cartilage and synovial fluid by different methods analyzed by Western blot.

a Plain text = aggrecan domains and regions (G1, IGD, KS, CS1, CS2 and G3; in bold), and amino acid sequences verified by N- and C-terminal antibodies in at least one of the samples. Italic text = domains or regions estimated by the calculation model. Gu-PG, guanidinium extracted proteoglycans. Seph, Sepharose. Grey boxes = fragments detected in significant amounts.

A1D3 and A1D2 fractions, and in in vitro digested A1D1 samples (Fig. 2A), and was confirmed with IPEN and TEGE antibodies (data not shown). These data show that the cartilage aggrecan immuno-pattern of the Gu-PG sample is similar to the combined patterns of the A1D1–D3 fractions, and to the pattern of in vitro digested A1D1 samples (Fig. 2, Table 3). 2.4. Evaluation of the boiling method Diluting the SF six times, keeping the proteins at physiological NaCl concentration and at pH 6 during the boiling step, gave the best sGAG and ARGS recoveries (data not shown). After the boiling step, acetone precipitation (concentrating the aggrecan fragments) gave the best sGAG and ARGS recoveries as compared to if using TCAdeoxycholate precipitation or speed vac concentration (data not shown). No clear G1 fragments were seen by Western blot after concentrating the Boiled SF sample by acetone precipitation (Fig. 1A), or with any of the other concentration methods (data not shown). Qualitative results of the boiling method using the SF pool sample are presented above. To further analyze this purification method, and to evaluate the potential of the boiling method as a preparation of SF aggrecan for screening studies, a quantitative evaluation was done using SF from human subjects. The average sGAG recovery for Boiled SF preparations from the osteoarthritis (OA) pool was 83% (SD = 15%, n = 5 preparations), and had a coefficient of variation of 18%. Boiled SF samples of human subjects (n =23) were analyzed and compared with their corresponding SF D1 samples; and the sGAG recovery in these Boiled SF samples was in

average 116% (SD = 43%), which is higher than found in the corresponding SF D1 samples (= 92%, SD = 29%). Purification of aggrecan fragments in Boiled SF samples was between 8 and 376 µg sGAG/mg protein, and the purification factors was in average 16×, (SD =8). The large variation in sGAG/protein ratios in the Boiled SF samples could to some part be explained by the biological variation in their corresponding SF, which is indicated by a high correlation (rS = 0.86, p < 0.0001, n = 23) of the sGAG/protein ratio between SF and Boiled SF samples (Fig. 3A). There was no correlation between subjects SF sGAG concentrations and the corresponding sGAG yields in the prepared boiled SF samples (rS = 0.32, p = 0.13, n = 23), suggesting that the purification of sGAG containing aggrecan fragments are not dependent on the SF sGAG concentrations. Measuring the total ARGS concentration, by Western blot, the inter preparation CV for the Boiled SF of the OA-pool was 23% (n = 4), which is similar to what has been found for the corresponding SF D1 sample (Struglics et al., 2009). The total ARGS concentration, measured by Western blot, showed strong correlation (rS =0.83, p<0.0001) between the Boiled SF samples and their corresponding SF D1 samples (n=23), although the Boiled SF samples gave in average 40% (SD=66%) of the ARGS mol values obtained with the SF D1 samples (Fig. 3B). Also, using Boiled SF samples (n=18 subjects), ARGS Western blot quantification showed strong correlation (rS =0.81, p<0.0001) with measurements with an ARGS ELISA method (Pratta et al., 2006a) (Fig. 3C). Based on ARGS measurements, these data show that Boiled SF samples are similar to SF D1 samples concerning purification of high Mw aggrecan fragments, and suggest that the boiling method is suitable for ELISA and Western blot screenings.

78

A. Struglics, S. Larsson / Matrix Biology 29 (2010) 74–83

Fig. 2. Western blot of purified cartilage samples. Human cartilage aggrecan fragments were purified by guanidinium extraction (Gu-PG) and by CsCl density centrifugation (collecting A1D1, A1D2 and A1D3 fractions). MMP-3 in vitro digested (16 h) A1D1 and ADAMTS-4 in vitro digested (0.25–24 h) A1D1 samples were used for comparisons. Deglycosylated samples (24 ng–8 µg sGAG per lane) were separated by SDS-PAGE and probed by different antibodies. Several Western blot experiments were conducted and different exposure times (1 s–15 min) on film or in a luminescence image analyzer were used to develop clear images (strips). The images show representative Western blots from full size blotted gels. Bands marked by ⁎ are considered to be false-positive immunobands since they could not be blocked by the corresponding immunogen peptide. G1-TEGE (A) and AGEG-G3 (F) dimers were confirmed by corresponding antibodies (data not shown).

3. Discussion The choice of method when purifying aggrecan from cartilage and SF depends on the origin of the sample, how pure you need the sample to be, and the size and extent of glycosylation of the aggrecan fragments you intend to study. For instance, to get high purity of aggrecan from cartilage (e.g. for in vitro studies) the cumbersome and time consuming cesium chloride density gradient centrifugation yields the best results. If your intention instead is to quantify fragments in a patient material (e.g. by SF ELISA or Gu-PG Western blot), a less pure sample could be more suitable, due to less losses than in an extended preparation. This paper is, to the best of our knowledge, the first large scale comparison of different methods of aggrecan purification from cartilage and SF where their effectiveness is evaluated by sGAG and protein concentrations, and by aggrecan fragment patterns in Western blots. Approximately 0.3% of the total protein in SF consists of aggrecan (our unpublished data), and the SF has approximately 100–200 times lower sGAG/protein ratio compared to the cartilage Gu-PG sample (Tables 1 and 2). Therefore, to obtain reliable Western blot analysis it is necessary to purify the aggrecan fragments from SF. Alcian blue precipitation of SF aggrecan fragments had (together with the D1 fraction) the highest purification level of aggrecan. The disadvantage of this method was that it gave low sGAG recovery (56%, Table 1) and diffuse immuno-bands on Western blots. Low sGAG yield could be due to difficulties in solubilization of the Alcian blue precipitate in buffer (Karlsson and Björnsson, 2001), which is avoided when solving the precipitate in a guanidinium/propanol solution as in

the method of sGAG analysis (Björnsson, 1993). These data indicate that even though this method gives very high aggrecan purification levels the sample is still not very useful as a preparation of aggrecan due to solubilization problems and low sGAG yields. Purifying SF aggrecan by anion chromatography gave variable purification levels of aggrecan, mainly depending on whether albumin was removed (by Blue Sepharose chromatography) prior to the anion chromatography run or not (Table 1). Purifying SF aggrecan at acidic pH by anion chromatography minimizes the negative charge density of amino acid side chains while the sGAG remains negatively charged at this pH and therefore binds to the column (Heinegard and Sommarin, 1987). Even though this method was applied, the sGAG yield in the anion chromatography purified samples was only around 50% (Table 1). sGAG recoveries of 80–90% have been reported for purification of aggrecan from SF by diaminoethylcellulose anion (DEAE) chromatography (Sandy and Verscharen, 2001). These differences in sGAG yields could be explained by Sandy and Verscharen who used the dimethyl-methylene blue (DMMB) dye method for sGAG determination, which is less specific for sGAG than Alcian blue (Björnsson, 1993). Purification of aggrecan fragments by chromatography, as the Blue Sepharose/Q-Sepharose method presented here, are still very useful since the purified sample contains high, medium and low Mw aggrecan fragments similar to those found in the A1 fraction (Fig. 1, Table 3). Associative or dissociative CsCl density gradient centrifugation is the gold standard for purifications of synovial fluid aggrecan fragments. The D1 fraction had an 1800 times higher sGAG to protein

A. Struglics, S. Larsson / Matrix Biology 29 (2010) 74–83

Fig. 3. Characterizing the Boiled SF preparation by analysis of human synovial fluid samples. From synovial fluid, donated from subjects with different arthritis diseases or from patients with knee injuries (Struglics et al., 2009), SF D1 and Boiled SF samples were prepared. (A) Correlation of sGAG/protein ratio (in µg/mg) between Boiled SF and corresponding SF samples (n=23). (B) Correlation of total ARGS concentrations (in pmol/ml SF), measured by quantitative Western blot (WB), between Boiled SF and corresponding SF D1 samples (n=23). (C) Correlation of total ARGS concentration (in pmol/ml SF), measured in Boiled SF samples (n=18), between ARGS ELISA (Pratta et al., 2006a) and ARGS quantitative Western blot methods. Mean values are presented. The solid line shows first order regression. Spearman rank's order correlations (rS) are given for each relationship with p<0.0001. Note the logarithmic scales.

ratio than neat SF, but the low Mw fragments (G1-IPEN and G1-TEGE) with no or few sGAG chains were not present in this dissociative high density fraction (Fig. 1). To analyze this type of fragments, we found

79

an associative A1 fraction more suitable. This fraction had high sGAG to protein ratio and contained the largest variety of SF aggrecan fragments as detected by Western blot (Fig. 1, Table 3). Western blot are nowadays also used for quantitative measurements (Feissner et al., 2003; Mathrubutham and Vattem, 2005; Heidebrecht et al., 2009), and recently we developed Western blot quantification for aggrecan fragments (Struglics et al., 2009). In patient screenings where purification of aggrecan fragments is necessary, D1 preparations give high sGAG recoveries and reliable quantitative Western blot immuno data (Struglics et al., 2009). Although technically suitable, CsCl fractionation is a demanding and cumbersome method not favorable for large scale screenings, and therefore we developed the Boiled SF purification. This method is based on that sGAG containing aggrecan fragments have high concentration of negatively charged sulfate groups and will stay in solution while other proteins in the SF (e.g. albumin) will denaturize and precipitate due to the boiling and centrifugation. As expected, low Mw aggrecan fragments with no or few sulfate groups (e.g. G1-IPEN and G1-TEGE) precipitated and were not purified by this method. The general aggrecan immuno pattern of the Boiled SF sample was mainly similar to the high Mw pattern obtained in SF samples purified by chromatography or CsCl density fractionation (Fig. 1, Table 3). The Boiled SF method was tested with different salt concentrations, pH and sample concentration methods (precipitation and speed vac) and the procedure described herein gave the highest recoveries of sGAG and ARGS fragments. The sGAG recoveries in the Boiled SF of subjects were similar to those found in their corresponding SF D1 samples. The Boiled SF samples showed in average 2.5 times lower ARGS concentrations compared to corresponding SF D1 samples and the SD between these two methods was high. This could indicate of large variations between the two purification methods concerning analysis of high Mw aggrecan fragments, although the high correlation between the Boiled SF and SF D1 samples measuring ARGS concentration suggests similarities between these purification methods (Fig. 3B). A high correlation between ARGS ELISA and ARGS Western blot methods using Boiled SF samples also suggests that the Boiled SF purification can be used in preparations of SF samples for different analysis methods (Fig. 3C). Although, the differences seen (Fig. 3B, C) indicates that in screening analysis it is equally important to use the same purification method (e.g. Boiled SF) for preparing samples as it is to use same analysis methods (e.g. ELISA). Approximately 50% of the dry weight of joint cartilage consist of proteoglycans (mainly aggrecan) (Hascall and Kimura, 1982), and therefore only an extraction (solubilization) of the proteoglycans was needed to obtain a sample with enough aggrecan fragment purity for Western blot analysis. For further purification of aggrecan from cartilage, the most common method is associative and/or dissociative CsCl density gradient centrifugation of the extract. The cartilage A1D1 fraction had the highest aggrecan purification level (sGAG/protein ratio), and therefore this sample is well suited as an aggrecan standard in for instance ELISA (Pratta et al., 2006a) or Western blot quantification (Struglics et al., 2009). The A1D1 fraction had a moderate sGAG yield (55%), which indicates a loss of sGAG containing proteoglycans to other fractions, such as the A2 and A3 fractions. Indeed, cartilage A1, A2 and A3 fractions have sGAG yields of approximately 70%, 10% and 2%, respectively (our unpublished data), and similar modest sGAG yields for articular cartilage for the high density associative fraction has been described (Vilim and Fosang, 1994). In contrast, more than 90% of the proteoglycans extracted from bovine nasal cartilage was recovered in an A1D1 fraction as measured by uronic acid analysis (Hascall and Sajdera, 1969). The discrepancies in sGAG yields between these studies might be due to differences in species, cartilage (nasal versus articular) and methodology, but it could also be due to that measurements by Alcian blue precipitation is specific for sulfated glycosaminoglycans (mainly aggrecan in cartilage) while uronic acid analysis recognizes both

80

A. Struglics, S. Larsson / Matrix Biology 29 (2010) 74–83

sulfated and non-sulfated glycosaminoglycans which include hyaluronic acid. The cartilage Gu-PG sample had a high purification level of aggrecan, and the preparation of this sample is a relative easy process and less time consuming compared to associative and dissociative CsCl density gradient centrifugation. Both high (sGAG containing), medium (GLGS-G3) and low (G1-IPEN and G1-TEGE) Mw aggrecan fragments were present in the cartilage Gu-PG sample as shown by Western blot (Fig. 2, Table 3). Taken together, these data suggest that the purity of the Gu-PG sample is sufficient for qualitative and quantitative analysis of aggrecan fragments, and since the A1D1–D3 fractions have low sGAG yields the Gu-PG is most likely a better sample in mirroring the cartilage in vivo content than the more purified samples. Even though the antibodies were thoroughly checked with positive controls and peptide blocking in Western blots, still some samples reacted false-positively against these antibodies. This indicates the importance of checking all the samples for false-positive Western blot immuno-signals. In addition to the different interpretations and differences compared to other studies, this study has some general limitations: (1) We used two different SF OA pools for the purifications; neat SF (SF-1) and joint lavage (SF-2). (2) The SF OA pools and cartilage OA pool were not from the same patients. (3) The cartilage explant media was analyzed from two patients. (4) The quantitative Western blot analysis of Boiled SF from patients was only done for ARGS fragments. With these data we summarize our experiences of the different applications for the aggrecan purification methods tested (Table 4). We suggest that for Western blot and ELISA analysis of cartilage aggrecan, the Gu-PG sample is the most beneficial sample to use due to the content of high, medium and low Mw fragments and due to the simplicity of the sample preparation. Boiling SF collecting high Mw negatively charged aggrecan fragments in the supernatant after centrifugation is a simple and quick procedure to increase the concentration of high Mw aggrecan fragments. We suggest that this method and the CsCl density gradient preparations are the most favorable purifications from synovial fluid for Western blot and ELISA analysis. This comparative study will facilitate the choice of purification method in future analysis of aggrecan fragments in the search for better understanding of joint degradation.

4. Experimental procedures 4.1. Reagents Alcian blue 8GS (C.I. no 742240) was from Chroma-Gesellschaft (Köngen, Germany). 4-(2-aminoethyl)-benzenesulfonyl fluoride (AEBSF), 6-aminohexonic acid (EACA), Benzamidine–HCl, BSA, Na4EDTA, N-Ethylmaleimide (NEM) and Phenylmethylsulfonyl fluoride (PMSF) were from Sigma. Chondroitinase ABC (EC 4.2.2.4), hyaluronidase Streptomyces hyalurolyticus (EC 4.2.2.1), keratanase (EC 3.2.1.103) and keratanase II (Bacillus sp. Ks36) were from Seikagaku. Molecular weight markers 10–250 kDa (Precision Plus Protein Standards) were from BioRad. Activated recombinant human matrix metalloproteinase 3 (MMP-3) was provided by Merck. Human recombinant ADAMTS-4 (a disintegrin and metalloproteinase with thrombospondin motifs, aggrecanase-1) was a kind gift from M Pratta and Dr. S Kumar (GlaxoSmithKline). ECL Plus detection was from Amersham Biosciences, and ECL SuperSignal West Femto was from Pierce. Mini gels (Tris-acetate 3–8%), LDS sample buffer, SDS running buffers (Tris-acetate buffer), transfer buffer were all NuPAGE from Invitrogen. PVDF membranes were from Invitrogen or BioRad. Non-fat dry milk (Semper) was from the local supermarket. Neo-specific antiKEEE, anti-LGQR, anti-TEGE and anti-IPEN sera were prepared at Merck, and have been characterized previously (Lark et al., 1995; Struglics et al., 2006a). Neo-specific anti-ARGS monoclonal antibody (mab OA-1) and neo-specific anti-AGEG and anti-SELE sera were kind gifts from M Pratta and Dr. S Kumar (GlaxoSmithKline); the ARGS and SELE antibodies have been characterized previously (Pratta et al., 2006a; Struglics et al., 2006a). Monoclonal anti-interglobular domain (IGD) antibody (6-B-4), recognizing the IGD sequence EPEEPFTFAPEI, was from Abcam (Cambridge, UK). Monoclonal anti-FFGV (BC-14), recognizing the MMP generated neoepitope FFGVGGE (immunogen), was from Novus Biologicals (Littleton, CO, USA). Polyclonal anti-G3 and anti-G1 domain antibodies were from Affinity BioReagents (Golden, CO., USA), and neo-specific anti-GLGS and anti-GRGT sera were made by Innovagen (Lund, Sweden). Specificity of antibodies was confirmed by complete blocking of their immunoreactions by the immunogen peptides (data not shown): FFGVGGEEDITVC (BC-14), CATEGQVRVNSIYQKVSL (G1), CDGHPMQFENWRPNQPDN (G3), GLGSVELSC and GRGTIGISC. Peroxidase conjugated secondary

Table 4 Pros and cons of aggrecan purification methods.

Purifications from synovial fluid None Alcian blue Boiling Blue Sepharose Blue Sepharose + Q-Sepharose A1 D1 Purifications from cartilage Guanidinium extraction A1D1 A1D2 A1D3 Explant culture media

Preparation

Sample requirement for preparation

Purity of sGAG fragments

Mw of sample fragments

sGAG recovery

Quantitative screening

Na Laborious Easy Laborious Laborious

Na Large Small Large Large

Low Very high Low Low High

Low-high High Highb Nd Low-high

Na Low High Low Low

−/++a − ++ − +

Laborious Laborious

Large Large

High Very high

Low-high High

High High

++ ++

Easy Laborious Laborious Laborious Na

Small Large Large Large Na

High Very high High Low Nd

Low-high High High and medium Low Low-high

Na Low Low Na Na

+++ + + + +++

Based on our results and on the literature herein, we present our view on the applications of different aggrecan purification methods from human cartilage and synovial fluid. Advantage features for the different methods are marked in bold. We grade the quantitative screening (ELISA and/or Western blot) as satisfactory (+++, ++, +) or unsatisfactory (−) depending on the presented features. Sample requirement for preparation: small, <100 µl SF or <1 g cartilage. Purity of sGAG fragments (in µg sGAG/mg protein): very high, >5000; high, >600. Molecular weight (Mw) of fragments in sample: high, >250 kDa; low, <100 kDa. sGAG recovery: high, >80%. Na, not applicable. Nd, not determined. a Depends on assay (i.e. ELISA). b Except for some G1 fragments.

A. Struglics, S. Larsson / Matrix Biology 29 (2010) 74–83

antibodies were from Cell Signaling Technology (goat anti-mouse IgG) and KPL (goat anti-rabbit IgG). 4.2. Human cartilage and synovial fluid Knee cartilage was obtained from ten patients undergoing joint replacement surgery due to OA. All remaining cartilage from the tibia plateau and femur condyles was removed from the knee joints, diced and pooled for aggrecan extraction. Undiluted (neat) SF was collected from subjects described previously (Struglics et al., 2009) and was used in individual analysis. Two different SF pool samples, from patients with end-stage knee OA or post-injury cartilage changes, were used for purification of aggrecan fragments: (1) neat SF collected from 47 patients (SF-1 pool, age range 16–89 years) and (2) diluted synovial fluid lavage from more than 100 patients (SF-2 pool, similar age range as SF-1 pool). Concentrations of sGAG were 129 µg/ml for the SF-1 pool and 16 µg/ml for the SF-2 pool, and the GAG/protein ratio for these SF pools were 3.1 and 5.1 µg/mg, respectively. Samples were stored at −80 °C. The synovial fluids and the cartilage samples were not from the same subjects, and the individual SF samples were not included in the two SF-pools. All procedures were approved by the ethics review committee of the Medical Faculty of Lund University.

81

fragments were collected, by slicing the tubes (Beckman Coulter Tube slicer kit no. 303811), from the lower 2/5 part of the tubes (A1 fractions, density of 1.7 g/ml). To each A1 fraction guanidinium–HCl was added (adjusting concentration to 3.6–4 M), and the samples (starting density of approximately 1.6 g/ml) were centrifuged for 72 h as previously. Samples were collected, by slicing the tubes, from the upper 1/5 part (A1D3 fractions, density of 1.4 g/ml), the middle 2/5 part (A1D2 fractions, density of 1.5 g/ml) and the lower 2/5 part of the tubes (A1D1 fractions, density of 1.6 g/ml), and these fractions were dialyzed (10 kDa cut off Slide-A-Lyzer; Pierce) at 4 °C extensively against Millipore-water prior to freeze drying. Protein and sGAG analysis of each (polytron and ball mix) A1D1–D3 fractions were conducted. The A1D1–D3 fractions prepared by the polytron method were then deglycosylated and used for Western blot. 4.5. Hyaluronidase digestion of synovial fluid for chromatography Synovial fluid (SF-1 or SF-2 pool samples) was first filtered (0.8 µm and 0.2 µm) to remove turbid matter. Buffer, protease inhibitors and hyaluronidase (final concentrations; 20 mM Na-acetate, 136 mM NaCl, 5 mM Na4-EDTA, 0.3 mM AEBSF, pH 6, 0.037 hyaluronidase turbidity reducing unit per µg sGAG) were added and samples were incubated for 3 h at 60 °C.

4.3. Human explant media Human knee OA cartilage (172 mg) from two patients were cultured (2.4 ml) in Dulbecco's modified Eagle's medium (DMEM) for 48 h in the presence of IL-1 (20 ng/ml) and oncostatin M (10 ng/ml) (Pratta et al., 2006b). The explant medium was analyzed for sGAG and deglycosylated for SDS-PAGE. 4.4. Purification of cartilage aggrecan fragments by guanidinium extraction and CsCl density gradient centrifugation Guanidinium extraction was done on a human knee cartilage pool (ten OA patients) by two separate methods: (1) In the first experiment, 10 g wet weight cartilage was sliced into 1 mm pieces into 90 ml Extraction buffer (50 mM Na-acetate, 4 M guanidinium–HCl, 10 mM EDTA, 100 mM EACA, 10 mM NEM, 5 mM benzamidine–HCl, 1 mM PMSF, pH 6) and was homogenized at 4 °C by a polytron. The homogenate (100 ml) was then stirred overnight at 4 °C, and thereafter the guanidinium extract was centrifuged (29 000 ×g for 30 min at 4 °C) collecting the supernatant enriched in proteoglycans (sample called GuPG). To the remaining cartilage pellet an additional guanidinium extraction (2 h stirring at 4 °C in 20 ml Extraction buffer) was conducted, following centrifugation the Gu-PG samples were pooled. (2) In a separate experiment, 4 g wet weight cartilage (1–5 mm pieces) was frozen in liquid nitrogen and pulverized in a ball mill grinder (Retsch mixer mill MM301, Retsch Gmbh & Co. Germany) for 30 s (25 shakes/s). The sample was collected in 48 ml Extraction buffer and mixed for 24 h at 4 °C, and thereafter the guanidinium extracted proteoglycans (Gu-PG) were collected by centrifugation. A minor part of each Gu-PG sample (by polytron and ball mix, respectively) was acetone precipitated and proteins were dissolved in 50 mM Tris-acetate, 50 mM Na-acetate, pH 7.6 for protein and sGAG analysis. The Gu-PG sample prepared by the polytron method was then deglycosylated and used for Western blot. The major part of each Gu-PG sample (polytron respectively ball mix) was dialyzed (12–14 kDa cut off; no 132706, Spectrapor) overnight at 4 °C against 9 volumes Buffer A (50 mM Na-acetate, 10 mM EDTA, 100 mM EACA, 10 mM NEM, 5 mM Benzamidine–HCl, 0.4 mM PMSF, pH 6), and after this equilibrium dialysis (i.e. 0.4 M guanidinium–HCl in Buffer A), CsCl was dissolved (1 g/g solution) in the samples resulting in a starting density of 1.55–1.60 g/ml. The samples were then put in Quick Seal tubes (Beckman Coulter no. 342413), centrifuged (162 000 ×g for 72 h at 16 °C), and the aggrecan

4.6. Removal of albumin from synovial fluid by Blue-Sepharose chromatography Hyaluronidase digested SF-1 or SF-2 pool samples were dialyzed (10 kDa cut off; Slide-A-Lyzer, Pierce) against Binding buffer (20 mM Na-phosphate, pH 7), and applied at room temperature (22 °C) with a pump (2 ml/min, Model 301 Gilson) or by a syringe (2–5 ml/min) on a 5-ml HiTrap Blue HP column (Blue-Sepharose, GE Healthcare) equilibrated with Binding buffer. The column was washed with Binding buffer with approximately 10 column volumes. Albumin was bound to the column and the unbound flow through was collected, applied a second time on the column, and the sample was collected again in the flow through. The sample was concentrated approximately 10 times in a Centriprep spin column (10 kDa cut off; YM10, Amicon). After sGAG and protein determination the sample was either further purified by anion-chromatography (below) or deglycosylated for SDS-PAGE. 4.7. Purification of synovial fluid aggrecan fragments on Mono-Q column Blue-Sepharose cleaned SF-2 pool sample (10 ml, 160 µg sGAG, 9.8 mg protein) was dialyzed against Buffer A (20 mM PIPES-NaOH, 70 mM NaCl, pH 5), filtered (0.22 µm) and applied (0.5 ml/min, Gilson pump Model 301) at room temperature on a 1-ml Mono-Q anion column 5/5 (GE Healthcare) equilibrated with Buffer A. The column was washed with Buffer A until UV detector (280 nm, SPD-2A Shimadzu) reached baseline. The bound polypeptides were batch eluted at 0.5 ml/min by 5-ml 20 mM PIPES-NaOH, 3 M NaCl, pH 5. The eluted fraction was dialyzed (10 kDa cut off; Slide-A-Lyzer, Pierce) against 100 mM NH4-acetate (pH 6.4) and dried in a speed vac. A small part of the sample was dissolved in 20 mM Nphosphate, pH 7 for sGAG and protein analysis, while the remaining part was deglycosylated for SDS-PAGE. 4.8. Purification of synovial fluid aggrecan fragments on Q-Sepharose column A Blue-Sepharose cleaned SF-1 pool sample (5.6 ml, 105 µg sGAG, 4 mg protein) was dialyzed against Buffer A1 (20 mM PIPES-NaOH, 150 mM NaCl, pH 5) or Buffer A2 (20 mM PIPES-NaOH, 6 M Urea, 150 mM NaCl, pH 5), and applied to 1-ml HiTrap Q-Sepharose anion columns (GE Healthcare) equilibrated with Buffer A1 or A2. The columns were washed with 10 ml Buffer A1 or A2 and the bound polypeptides were then batch

82

A. Struglics, S. Larsson / Matrix Biology 29 (2010) 74–83

eluted with 7-ml Buffer B1 (20 mM PIPES-NaOH, 2 M NaCl, pH 5) or Buffer B2 (20 mM PIPES-NaOH, 6 M Urea, 2 M NaCl, pH 5). The columns were run at room temperature using a syringe at approximately 1 ml/min flow rate. The eluted fractions were dialyzed (10 kDa cut off Slide-A-Lyzer; Pierce) against 50 mM NH4-acetate (pH 6.4). A small part of each sample was dried (speed vac) and dissolved in 20 mM N-phosphate, pH 7 for sGAG and protein analysis, while the remaining part was deglycosylated for SDS-PAGE. 4.9. Purification of synovial fluid aggrecan fragments by CsCl density gradient centrifugation SF D1 and A1 samples were prepared from the SF-1 pool as follows: SF (500–800 µg sGAG) was cleared by centrifugation, and buffer (52 mM Naacetate, pH 6.8), protease inhibitors (10 mM EDTA, 0.41 mM AEBSF, 1.0 µM Pepstatin A, 5.2 µM E64) and CsCl (3.1 M) was added. For the dissociative preparation guanidinium chloride (3.5 M) was also added. Samples were loaded in 12.5 ml Quick Seal tubes (Beckman Coulter no. 342413) centrifuged at 162000 ×g (72 h at 16 °C) and the D1 fraction (lower 2/5 part of the tube with density of 1.5–1.65 g/ml) and the A1 fraction (lower 2/5 part of the tube with density of 1.45–1.6 g/ml) were collected by slicing the tubes (Beckman Coulter Tube slicer kit no. 303811). Synovial fluid D1 samples from individual subjects were prepared by a mini CsCl preparation as described previously (Struglics et al., 2009). Briefly, SF was cleared by centrifugation and 30–90 µg sGAG was loaded in 2-ml Quick Seal tubes (Beckman Coulter no. 344625) in the presence of guanidinium chloride, buffer and proteinase inhibitors. Samples were centrifuged at 162000 ×g (72 h at 16 °C), and the D1 fractions were collected (by slicing the tubes) from the lower half of the tubes (density of 1.46–1.54 g/ml). The A1 and D1 samples were dialyzed thoroughly against Millipore purified HPLC-grade water containing proteinase inhibitors (0.8 mM EDTA, 0.4 mM PMSF, 1 µM Pepstatin A, 20 µM iodoacetamide, 2 mM o-phenanthroline). 4.10. Purification of synovial fluid aggrecan fragments by Alcian blue precipitation The purification was done according to the large scale precipitation method of Karlsson and Björnsson (2001). Briefly (final concentrations are shown); SF-2 pool sample (0.96 mg sGAG) was cleared by centrifugation, guanidinium–HCl (0.7 M) and Triton X-100 (0.44% weight/volume) was added, and the sample was incubated for 15 min at room temperature. The sample was Alcian blue precipitated (0.05% weight/volume) overnight at 4 °C in the presence of 0.4 M guanidinium– HCl, 0.25% H2SO4 and 0.25% Triton X-100. The precipitate was collected by centrifugation, dissolved in 2.67 M guanidinium–HCl, 0.1% H2SO4 and 0.25% Triton X-100, and mixed for 15 min at room temperature. Undissolved material was removed by centrifugation and the soluble aggrecan containing sample was Alcian blue precipitated (0.085% dye, 0.96 M guanidinium–HCl, 0.1% H2SO4 and 0.25% Triton X-100) a second time (2 h at 4 °C). The precipitate was collected by centrifugation, dissolved in 40% DMSO, 50 mM MgCl2 and collected by centrifugation. The precipitate was then dissolved by mixing in 4 M guanidinium–HCl, 33% propanol, 0.25% Triton X-100, followed by an increase in the propanol concentration to 66%, and incubation at 4 °C for 1 h. The sample was then collected as a pellet after centrifugation. A second propanol precipitation (75% propanol, 0.1 M Tris-acetate, pH 7.4) was done and the aggrecan sample was collected as a pellet by centrifugation. 4.11. Purification of synovial fluid aggrecan by boiling Two to forty microgram sGAG of SF-1 pool or individual neat SF samples was digested for 3 h at 60 °C by hyaluronidase (0.01 turbidity reducing unit/µl SF) in the presence of 50 mM Na-acetate, 10 mM

EDTA, 0.25 mM AEBSF, pH 6. The samples were then deglycosylated; first for 2 h at 37 °C by chondroitinase ABC (1 mU/µg sGAG , 50 mM Tris-acetate, 50 mM Na-acetate, 10 mM EDTA, 0.15 mM AEBSF, pH 7.6), then for 1 h at 37 °C by keratanase (1 mU/µg sGAG, 50 mM Trisacetate, 50 mM Na-acetate, 10 mM EDTA, 1.1 mM AEBSF, 10 mM NEM, pH 7.6), and finally for 1 h at 37 °C by keratanase II (0.1 mU/µg sGAG, 49 mM Tris-acetate, 49 mM Na-acetate, 9.8 mM EDTA, 0.9 mM AEBSF, 8.5 mM NEM, pH 6). The hyaluronidase-treated and deglycosylated samples were diluted (corresponding to a total of 6 times of the starting volume of synovial fluid) with a NaCl solution (giving final concentration: 41 mM Tris-acetate, 41 mM Na-acetate, 150 mM NaCl, 8.2 mM EDTA, 0.8 mM AEBSF, 7.1 mM NEM, pH 6), and then boiled for 5 min (100 °C). The supernatants (called Boiled SF) containing aggrecan fragments were collected after a 10 min centrifugation (20 000 ×g, 4 °C), and the samples were acetone precipitated and collected by centrifugation. 4.12. In vitro digestion of cartilage aggrecan ADAMTS-4 in vitro digestion of cartilage A1D1 sample was conducted at 37 °C for 0.25–24 h (Struglics et al., 2009), and MMP-3 digestion was conducted as described previously (Struglics et al., 2006a). 4.13. sGAG and protein analysis The method for quantitation of sGAG (by Alcian blue precipitation) and protein (Pierce Micro BCA™ Assay) was conducted as described previously (Björnsson, 1993; Struglics et al., 2006a). 4.14. Western blot analysis Deglycosylated samples (Struglics et al., 2006a) were denatured in sample buffer under reducing conditions (50 mM DTT) by boiling for 5 min. Samples were separated on 3–8% Tris-acetate SDS mini-gels according to manufacturer's instructions. Proteins were electrophoretically transferred onto PVDF membranes (Struglics et al., 2006a) and probed with antibodies. The immunobands were visualized using either ECL Plus or ECL SuperSignal West Femto together with either film or luminescence image analyzer (Fujifilm LAS-1000). The dilutions of the primary antibodies were as follows: ARGS (1:2000, 5 µg/ml), AGEG (1:500), FFGV (1:1000), G1 (1:400), G3 (1:500), GRGT (1:1000), GLGS (1:750), IGD (1:400), IPEN (1:10 000), KEEE (1:1000), LGQR (1:500), SELE (1:2000) and TEGE (1:3000). The dilutions of the secondary peroxidase conjugated antibodies were for goat anti-mouse IgG 1:25 000 and 1:50 000 (20 and 10 ng/ml) and for goat anti-rabbit IgG 1:25 000–1:400 000 (40–2.5 ng/ml). Except for the AGEG, IGD and LGQR N-terminal fragments all the immunobands mentioned in the text were verified by Western blot immunopeptide blocking experiments, where the peptides (2–10 µM) were pre-incubated with the antibodies for 10 min at 22 °C before addition to the PVDF membrane (Struglics et al., 2006a). False-positive immunobands (i.e. bands not blocked by the corresponding immunopeptide) are marked ⁎ in figures. 4.15. Quantification of ARGS fragments by Western blot Quantification of synovial fluid ARGS fragments by Western blot has been described previously (Struglics et al., 2009). Briefly, ARGS standards was made by a complete digest (24 h at 37 °C) of 0.5 mg dry weight A1D1 (from human knee cartilage pool) by 0.5 µg recombinant human ADAMTS-4 in buffer (50 mM Tris–HCl, 100 mM NaCl, 10 mM CaCl2, pH 7.5). The enzymatic reaction was stopped by addition of EDTA and deglycosylated. This sample was used as ARGS standard in the Western blot quantification, assuming 0.667 nmol ARGS fragments per mg aggrecan dry weight and aggrecan molecular

A. Struglics, S. Larsson / Matrix Biology 29 (2010) 74–83

weight of 1.5 × 106 g/mol (i.e. total protein sequence plus glycosaminoglycan). Boiled SF and SF D1 samples from subjects, and ARGS standards (three different concentrations per gel) were separated by electrophoresis and transferred to PVDF membranes, and probed with ARGS antibodies. Quantification of the ARGS signal from subjects and standard was conducted in luminescence image analyzer (Fujifilm LAS-1000) in the linear range of the imaging system. The positioning and analysis of regions of interest of individual bands were done using Fujifilm software Image Gauge version 4.0. Average ARGS concentrations for subjects were calculated from 2 to 5 separate Western blot experiments. 4.16. Estimation of aggrecan fragment identities by a calculation model Most aggrecan fragments were immuno-identified at both N- and C-terminal ends (Table 3). Fragments not immuno-verified at both ends were characterized by a calculation model based on the SDSPAGE Mw and the Mw of the carbohydrate substitutions (Struglics et al., 2006b). Acknowledgements We thank Michael Pratta and Dr. Sanjay Kumar for the kind gift of ADAMTS-4, antibodies (ARGS, AGEG and SELE) and for the explant culturing samples. We also thank Maria Hansson for the excellent technical assistance, and Dr. Stefan Lohmander for critical reading of the manuscript. Funding for this work was provided by grants from: The Kock Foundation, the Magnus Bergvalls Foundation and the Alfred Österlunds Foundation. References Bayliss, M.T., Ali, S.Y., 1978. Isolation of proteoglycans from human articular cartilage. Biochem. J. 169, 123–132. Bayliss, M.T., Hutton, S., Hayward, J., Maciewicz, R.A., 2001. Distribution of aggrecanase (ADAMts 4/5) cleavage products in normal and osteoarthritic human articular cartilage: the influence of age, topography and zone of tissue. Osteoarthritis Cartilage 9, 553–560. Björnsson, S., 1993. Simultaneous preparation and quantitation of proteoglycans by precipitation with Alcian blue. Anal. Biochem. 210, 282–291. Carter, Q.L., Dotzlaf, J., Swearingen, C., Brittain, I., Chambers, M., Duffin, K., Mitchell, P., Thirunavukkarasu, K., 2007. Development and characterization of a novel ELISA based assay for the quantitation of sub-nanomolar levels of neoepitope exposed NITEGE-containing aggrecan fragments. J. Immunol. Methods 328, 162–168. Feissner, R., Xiang, Y., Kranz, R.G., 2003. Chemiluminescent-based methods to detect subpicomole levels of c-type cytochromes. Anal. Biochem. 315, 90–94. Fosang, A.J., Last, K., Maciewicz, R.A., 1996. Aggrecan is degraded by matrix metalloproteinases in human arthritis. Evidence that matrix metalloproteinase and aggrecanase activities can be independent. J. Clin. Invest. 98, 2292–2299. Hascall, V.C., Sajdera, S.W., 1969. Proteinpolysaccharide complex from bovine nasal cartilage. The function of glycoprotein in the formation of aggregates. J. Biol. Chem. 244, 2384–2396. Hascall, V.C., Kimura, J.H., 1982a. Proteoglycans: isolation and characterization. Methods Enzymol. 82 Pt A, 769–800. Hascall, V.C., Calabro, A., Midura, R.J., Yanagishita, M., 1994. Isolation and characterization of proteoglycans. Methods Enzymol. 230, 390–417. Heidebrecht, F., Heidebrecht, A., Schulz, I., Behrens, S.E., Bader, A., 2009. Improved semiquantitative Western blot technique with increased quantification range. J. Immunol. Methods 345, 40–48. Heinegard, D., Sommarin, Y., 1987. Isolation and characterization of proteoglycans. Methods Enzymol. 144, 319–372.

83

Karlsson, M., Björnsson, S., 2001. Quantitation of proteoglycans in biological fluids using Alcian blue. Methods Mol. Biol. 171, 159–173. Lark, M.W., Gordy, J.T., Weidner, J.R., Ayala, J., Kimura, J.H., Williams, H.R., Mumford, R.A., Flannery, C.R., Carlson, S.S., Iwata, M., Sandy, J.D., 1995. Cell-mediated catabolism of aggrecan. Evidence that cleavage at the “aggrecanase” site (Glu373-Ala374) is a primary event in proteolysis of the interglobular domain. J. Biol. Chem. 270, 2550–2556. Lark, M.W., Bayne, E.K., Flanagan, J., Harper, C.F., Hoerrner, L.A., Hutchinson, N.I., Singer, I.I., Donatelli, S.A., Weidner, J.R., Williams, H.R., Mumford, R.A., Lohmander, L.S., 1997. Aggrecan degradation in human cartilage. Evidence for both matrix metalloproteinase and aggrecanase activity in normal, osteoarthritic, and rheumatoid joints. J. Clin. Invest. 100, 93–106. Lee, E.R., Lamplugh, L., Leblond, C.P., Mordier, S., Magny, M.C., Mort, J.S., 1998. Immunolocalization of the cleavage of the aggrecan core protein at the Asn341– Phe342 bond, as an indicator of the location of the metalloproteinases active in the lysis of the rat growth plate. Anat. Rec. 252, 117–132. Lohmander, L.S., Ionescu, M., Jugessur, H., Poole, A.R., 1999. Changes in joint cartilage aggrecan after knee injury and in osteoarthritis. Arthritis Rheum. 42, 534–544. Malfait, A.M., Liu, R.Q., Ijiri, K., Komiya, S., Tortorella, M.D., 2002. Inhibition of ADAM-TS4 and ADAM-TS5 prevents aggrecan degradation in osteoarthritic cartilage. J. Biol. Chem. 277, 22201–22208. Mathrubutham, M., Vattem, K., 2005. http://www.piercenet.com/files/AN0012.pdf. Moller, H.J., Larsen, F.S., Ingemann-Hansen, T., Poulsen, J.H., 1994. ELISA for the core protein of the cartilage large aggregating proteoglycan, aggrecan: comparison with the concentrations of immunogenic keratan sulphate in synovial fluid, serum and urine. Clin. Chim. Acta. 225, 43–55. Pratta, M.A., Su, J.L., Leesnitzer, M.A., Struglics, A., Larsson, S., Lohmander, L.S., Kumar, S., 2006a. Development and characterization of a highly specific and sensitive sandwich ELISA for detection of aggrecanase-generated aggrecan fragments. Osteoarthr. Cartil. 14, 702–713. Pratta, M., Hanning, C.R., Newman-Tarr, T., Wang, F., Steplewski, K., Connor, J.R., Slathia, S., West, M.R., Lohmander, L.S., Kumar, S., 2006b. Synergistic stimulation of aggrecan degradation by IL-1α and oncostatin m in human articular cartilage explants is associated with the specific induction of ADAMTS-4 expression. Trans. Orthop. Res. Soc. (abstract) 31, 1469. Sandy, J.D., Boynton, R.E., Flannery, C.R., 1991. Analysis of the catabolism of aggrecan in cartilage explants by quantitation of peptides from the three globular domains. J. Biol. Chem. 266, 8198–8205. Sandy, J.D., Verscharen, C., 2001. Analysis of aggrecan in human knee cartilage and synovial fluid indicates that aggrecanase (ADAMTS) activity is responsible for the catabolic turnover and loss of whole aggrecan whereas other protease activity is required for C-terminal processing in vivo. Biochem. J. 358, 615–626. Sandy, J.D., 2003. Proteolytic degradation of normal and osteoarthritic cartilage matrix. In: Brandt, K.D., Doherty, M., Lohmander, L.S. (Eds.), Osteoarthritis. Oxford University Press, Oxford, pp. 82–92. Struglics, A., Larsson, S., Pratta, M.A., Kumar, S., Lark, M.W., Lohmander, L.S., 2006a. Human osteoarthritis synovial fluid and joint cartilage contain both aggrecanaseand matrix metalloproteinase-generated aggrecan fragments. Osteoarthr. Cartil. 14, 101–113. Struglics, A., Larsson, S., Lohmander, L.S., 2006b. Estimation of the identity of proteolytic aggrecan fragments using PAGE migration and Western immunoblot. Osteoarthr. Cartil. 14, 898–905. Struglics, A., Larsson, S., Hansson, M., Lohmander, L.S., 2009. Western blot quantification of aggrecan fragments in human synovial fluid indicates differences in fragment patterns between joint diseases. Osteoarthr. Cartil. 17, 497–506. Sumer, E.U., Qvist, P., Tanko, L.B., 2007a. Matrix metalloproteinase and aggrecanase generated aggrecan fragments: implications for the diagnostics and therapeutics of destructive joint diseases. Drug Dev. Res. 68, 1–13. Sumer, E.U., Sondergaard, B.C., Rousseau, J.C., Delmas, P.D., Fosang, A.J., Karsdal, M.A., Christiansen, C., Qvist, P., 2007b. MMP and non-MMP-mediated release of aggrecan and its fragments from articular cartilage: a comparative study of three different aggrecan and glycosaminoglycan assays. Osteoarthr. Cartil. 15, 212–221. Valiyaveettil, M., Mort, J.S., McDevitt, C.A., 2005. The concentration, gene expression, and spatial distribution of aggrecan in canine articular cartilage, meniscus, and anterior and posterior cruciate ligaments: a new molecular distinction between hyaline cartilage and fibrocartilage in the knee joint. Connect. Tissue Res. 46, 83–91. Vilim, V., Fosang, A.J., 1994. Proteoglycans isolated from dissociative extracts of differently aged human articular cartilage: characterization of naturally occurring hyaluronan-binding fragments of aggrecan. Biochem. J. 304, 887–894.