ARCHIVES

Vol.

OF BIOCHEMISTRY

296, No. 2, August

AND

BIOPHYSICS

1, pp. 410-418,

1992

Antigenic Properties of Keratan Sulfate: Influence of Antigen Structure, Monoclonal Antibodies, and Antibody Valency Markus

J. Seibel,*

Department

Received

William

of Orthopaedic

November

Surgery,

Macaulay,

Richard

Orthopaedic

13, 1991, and in revised

form

March

Jelsma,

Columbia

Saed-Nejad, University,

and Anthony

Ratcliffel

New York, New York 10032

23, 1992

The influence of (a) antigen structure, (b) type of monoclonal antibody, and (c) antibody bivalency on the immunochemical detection and quantification of keratan sulfate (KS) from aggrecan has been studied. Apparent KS epitope levels were determined by immunoglobulin G (IgG)-enzyme-linked immunosorbent assay (ELBA) in preparations of human aggrecan and in a defined series of lower molecular weight proteoglycan preparations generated by proteolytic and alkali treatment of aggrecan. Gel filtration chromatography showed KS epitope to be preferentially detected in the higher molecular weight fragments of the preparations. In single KS chains the epitope was detected in the chains of higher Mr. The ability of the proteoglycan to inhibit in the IgG-ELISA decreased with a reduction in proteoglycan fragment size, ranging between 6- and 260-fold, depending on the antibody used. This was considered to be a cooperative binding, effect. With most antibodies, the sensitivity of the IgG-ELISA (represented by the steepness of the inhibition slope) was also reduced with smaller inhibitor sizes. The lowest limit of detectability (the amount of KS required to generate 20% inhibition) varied by up to 60fold depending on the antibody used. The use of monovalent Fab fragments instead of the whole IgG anti-KS antibody in the ELISA showed that the bivalency of the antibody also affected the quantitation of the assay. In the Fab-ELISA the assay was found to have an increased detectability (by 9.5-fold with aggrecan as the inhibitor), and the proteoglycan fragments and aggrecan all generated parallel inhibition curves. Although the FabELISA was somewhat influenced by the structural presentation of the KS, this was not apparent for small frag* Present address: Department of Medicine I, Division of Endocrinology, University of Heidelberg, Luisenstrasse 5, D-6900 Heidelberg, FRG. 1 To whom correspondence should be addressed at Orthopaedic Research Laboratory, Columbia Presbyterian Medical Center BB1412,630 West 168th St., New York, NY 10032. 410

Fatemeh

Research Laboratory,

ments and single chains. Thus the effects of cooperative binding and antibody valency could be overcome and quantitative data could be obtained for all samples, using papain-digested samples and the Fab-ELISA. Application of this assay to analysis of body fluids showed the KScontaining fragments in synovial fluid, serum, and urine were of different sizes and could be quantified. o lssz Academic

Prew, Inc.

The glycosaminoglycan keratan sulfate (KS)2 is characterized by a repeating dissacharide unit of (1-3)-p-Dgalactose-(l-4)-@-D-N-acetylglucosamine, with a sulfation level of approximately one sulfate group per disaccharide. KS is present in the extracellular matrix of various connective tissues with the majority being located in cartilage and intervertebral disc (Type II or skeletal KS) (3-6). There is also a significant concentration of KS in the cornea1 stroma of the eye (Type I or cornea1 KS) (1, 2). In cartilage KS is one of the two types of glycosaminoglycans of the large aggregating proteoglycan, aggrecan. The protein core of aggrecan contains several distinct domains (7, 8). One of these is a KS-rich region which appears to contain at least 50% of the KS of the molecule, although KS is distributed elsewhere on the protein core (9,25). The molecular structure and composition of skeletal and cornea1 KS have been shown to be variable (ll13). The length of the chains is not uniform, and the regions of monosulfated and disulfated disaccharide may vary in size (12). * Abbreviations used: KS, keratan sulfate; S-GAG, sulfated glycosaminoglycan; TPCK, L-1-p-tosylamino-2-phenylethyl chloromethyl ketone; 5D4, antibody 1/20/5-D-4; 2D3, antibody 1/20/2-D-3; lB4, antibody 4/8/1-B-4; 12.1, antibody 1110-S-12.1; 4A4, antibody ET-4-A-4; ELISA, enzyme-linked immunosorbent assay; PBS, phosphate-buffered saline; AlDl, aggrecan; KSRP 1, chondroitinase ABC and trypsin-digested aggrecan; KSRP 2, papain-digested KSRP 1; KSSC, KS single chains; IgG, immunoglobulin G. 0003-9861/92 $5.00 Copyright 0 1992 by Academic Press, Inc. All rights of reproduction in any form reserved.

ANTIGENIC

PROPERTIES

In recent years, a number of monoclonal antibodies against glycosaminoglycan determinants of proteoglycan, particularly KS, have been described (14-22). The antibodies to KS bind to pentasulfated hexasaccharides and larger related oligosaccharides of KS (23). Within a KS chain the repeating disaccharide sequence offers the possibility of multiple epitopes per chain (11,24). The structure of the proteoglycan adds to this complexity. Many KS chains are attached to the protein core and therefore it is likely that numerous epitopes will be present on one proteoglycan monomer. Many of the epitopes (but not necessarily all (25)) may be clustered in close proximity within the KS-rich region. The epitope recognition by anti-KS antibodies has previously been shown to be dependent on the degree of sulfation and on chain length (26, 27). The close proximity of the individual KS chains on the protein core and the bivalency of the antibody can further affect antibody binding (19, 26). These factors can all influence the quantitation of KS by immunoassay. The usefulness of monoclonal antibodies as sensitive and specific probes for KS has been demonstrated in studies of proteoglycan structure, localization, and metabolism (13,25,28,29). Recent studies indicate that the determination of KS in body fluids by immunoassay can provide information on cartilage catabolism (20, 30-37). The objective of this study was to assessthe contributions of (a) the structural presentation of the KS, (b) the type of monoclonal antibody used, and (c) the bivalency of the antibody, on the detection and quantitation of skeletal proteoglycan KS. EXPERIMENTAL

PROCEDURES

Materials. Trypsin (from bovine pancreas, TPCK-treated, EC 3.4.21.4), papain (Type III from papaya latex, EC 3.4.22.2), trypsin inhibitor (Type I-S), sheep anti-mouse IgG alkaline phosphatase conjugate (whole antibody), and guanidine-HCl were obtained from Sigma Chemical Co. (St. Louis, MO). Chondroitinase ABC (Proteus uulguris, EC 4.2.2.4), chondroitin 6-sulfate (from bovine trachea), and KS (Type I, from bovine cornea) were from ICN ImmunoBiologicals (Lisle, II). Sepharose CL-6B and CL-IB, CNBr-activated Sepharose 4B, DEAESephacel, and Protein G-Sepharose 4 Fast Flow were from Pharmacia LKB Biotechnology (Piscataway, NJ). 1,9-Dimethylmethylene blue was from Serva (Heidelberg). ELISA immunoplates were from Nunc InterMed Laboratories (Kampstrup, Denmark). All other reagents were analytical grade. Purification of human articulur cartilage oggrecun. Human articular cartilage aggrecan was purified from macroscopically normal patellae that were obtained at autopsy from six individuals aged 35 to 62 years (generously provided by Dr. J. F. Woessner, University of Miami). The tissue was pooled, and the proteoglycans were extracted with 4 M guanidine-HC1/0.05 M sodium acetate, pH 5.8, in the presence of protease inhibitors (39). Aggrecan was prepared by equilibrium gradient centrifugation (38) as described previously (39,40) using a Beckman 55 ultracentrifuge and a T170 rotor, first under associative conditions to yield an Al preparation and then under dissociative conditions to give an AlDl preparation. This was then exhaustively dialyzed and freeze-dried. In addition, a human articular cartilage proteoglycan monomer (AlDl) preparation from normal femoral condyle (generously provided by Dr. P. Roughley and A. R. Poole, Montreal, Canada) was used to prepare the single chains; characterization studies indicated that the antigenic

OF

KERATAN

SULFATE

411

properties of this preparation were indistinguishable from the pooled preparation. Preparation of KS-contuiningproteoglycun core protein. Lyophilized AlDl was dissolved in 4 M guanidine-HCl at a concentration of 5 mg/ ml and extensively dialyzed against 10 mM Tris-HCl buffer, pH 8.0, at 4°C (41). Chondroitinase ABC was added (0.02 U/mg of proteoglycan), and the AlDl was digested for 6 h at 37°C. The digest was then dialyzed against water and freeze-dried. The KS-containing protein Preparation of KS-contuiningpeptides. core prepared above was dissolved to 4 mg/ml in 0.1 M sodium phosphate buffer, pH 7.6, and digested with TCPK-trypsin (1 kg/mg protein) for 6 h at 37°C (9, 41). After dialysis against 0.1 M NaCl (all dialyses involving the preparation of the KS-containing fragments were performed with dialysis tubing with a molecular weight cutoff of 3500), the digested material was applied to a DEAE-Sephacel ion-exchange column (10 X 300 mm) equilibrated with the same buffer, and the column was eluted with 0.2 and 1.5 M NaCl, respectively (four bed volumes each). Analysis of the fractions for total sulfated glycosaminoglycans (S-GAG) and KS epitope (see below) showed that 97% of the KS epitope-containing fragments were eluted with 1.5 M NaCl. These fractions were pooled, dialyzed against water, and freeze-dried. This fraction was termed KSRP 1. KSRP 1 (5 mg/ml) in 0.05 M EDTA, 0.1 M sodium acetate, and 10 mM cysteineHCl at pH 6.0 was digested with papain (0.5 mg/mg S-GAG) at 60°C for 16 h, followed by heat inactivation of the enzyme (95”C, 30 min). This preparation was termed KSRP 2. Preparation of single KS chains. Lyophilized AlDl from bovine nasal and human articular cartilage was dissolved in 0.2 M Tris-HCl, pH 7.4, at a concentration of 10 mg/ml, and chondroitinase ABC was added at a ratio of 0.1 U/mg of proteoglycan. After digestion for 24 h at 37”C, the material was dialyzed extensively against water, freeze-dried, dissolved to 5 mg/ml in 50 mM NaOH and 1.0 M NaBH,, and incubated for 48 h at 45°C (42). The pH was neutralized with glacial acetic acid, and the material was chromatographed on a Sepharose CL-4B column (120 X 1.0 cm) eluted with 2 M guanidine-HCl and 50 mM Tris-HCl, pH 7.3. The eluate was analyzed to determine protein, S-GAG, and KS epitope concentrations. The fractions containing the S-GAG were pooled, dialyzed against water, and freeze-dried. This preparation was termed KSSC. The relative size distributions of the Gel filtration chromatography. AlDl and the proteoglycan fragments were determined by gel filtration chromatography on a column (1.0 X 120 cm) of Sepharose CL-6B eluted with 2 M guanidine-HCl. The eluate was analyzed to determine S-GAG and KS epitope concentrations. The relative sizes of the proteoglycan fragments containing KS epitope in the synovial fluid, serum, and urine were determined by gel filtration chromatography on a column (1.0 X 120 cm) of Sepharose CL-6B eluted with 0.5 M sodium acetate, pH 5.6. Monoclonal antibodies used. The following mouse monoclonal antiKS antibodies were studied: 1/20/5-D-4 (5D4), 1/20/2-D-3 (2D3), 4/8/ l-B-4 (lB4) (all provided by Dr. B. Caterson, University of North Carolina, Chapel Hill), 1110-S-12.1 (12.1) (provided by Ciba-Geigy Ltd., Basel, Switzerland), ET-4-A-4 (4A4) (provided by Dr. E. J. Thonar, Rush-Presbyterian-St. Luke’s Medical Center, Chicago), and MZ15 (provided by the Kennedy Institute, London, UK). Antibodies lB4, 2D3, and 4A4 were raised against bovine nasal cartilage AlDl (26, 43, 44); antibody 12.1 was raised to bovine nasal cartilage Al (20). Antibody 5D4 was raised against human articular cartilage AlDl (14), and antibody MZl5 was raised against pig laryngeal chondrocytes (22). All antibodies used in this study were of the IgG class and have been shown to react specifically with epitopes present on cornea1 and skeletal KS. Antibodies 5D4, 1B4, and MZ15 recognize pentasulfated hexasaccharides and larger oligosaccharides of KS. 2D3 was shown to have similar properties (26). Antibody 12.1 has been shown to recognize a decasulfated hexadecasaccharide and larger KS oligosaccharides (20). Previous studies with antibody 4A4 indicate that it has properties similar to 5D4 (35). Preparation of Fab fragments. Antibody 5D4 was first purified. The ascitic fluid containing the monoclonal antibody was applied to a column

412

SEIBEL

of Protein G-Sepharose (Pharmacia) equilibrated with 20 mM phosphate buffer, pH 7.0, and incubated at room temperature for 1 h. The column was then washed with the equilibration buffer, and bound antibody was eluted from the column using 0.1 M glycine, pH 2.75. This eluate was immediately neutralized, dialyzed against deionized water, and lyophilized. Fab fragments of the purified IgG were prepared using the Avid Chrom Kit (Bioprobe, Tustin, CA). Briefly, IgG was digested with papain, and the digestion mixture applied to a column of Protein G-Sepharose 4B and eluted with PBS. The Fc fragments bound with the protein G. The Fab-containing eluate was neutralized, dialyzed against PBS, 0.05% Tween, and 0.05 M EDTA, and stored at -2O’C. Enzyme-linked immurwsorbent assay (ELISA) procedure. A solidphase noncompetitive inhibition ELISA essentially as described by Thonar et al. (36) was used. When whole IgG was used as the anti-KS antibody the assay was termed IgG-ELISA. When Fab fragments of the primary antibody were used the assay was termed Fab-ELISA. In brief, 60 ~1 of sample or standard, diluted in PBS and 0.05% Tween, pH 7.3, was incubated overnight at 4°C with an equal volume of monoclonal anti-KS antibody diluted in PBS, 0.05% Tween, 0.05 M EDTA, and 1% bovine serum albumin, pH 7.3. One hundred microliters of this mixture was then transferred into the wells of flat-bottom microtiter plates, previously coated with chondroitinase ABC-digested bovine nasal cartilage aggrecan (200 ~1 of a 5 pg/ml solution per well) and incubated for 1 h at 4’C. The wells were washed with PBS and 0.05% Tween and incubated with sheep anti-mouse IgG-alkaline phosphatase conjugate (diluted 1: 1000 in PBS, pH 7.3) for 1 h at 37’C. Finally, bound conjugate was detected with p-nitrophenyl phosphatase (1 mg/ml) as substrate. AlDl &g/ml) was used as standard. For the IgG-ELISA the AlDl was digested with chondroitinase ABC prior to assay; for the Fab-ELISA the AlDl was digested with papain before assay. All determinations were done in duplicate, and the results presented are means -t SD of three separate experiments. In the IgG-ELISA the antibodies were used at the following dilutions: 12.1, l/40,000; 4A4, l/40,000; 2D3, l/100,000; 5D4, l/100,000; lB4, l/40,000; and MZ15, l/40,000. These dilutions allowed absorbance change of approximately 1.00 in 1 h in the absence of inhibitor. When the Fab fragments were used in the ELISA, the same protocol was used with Fab-specific anti-mouse IgG-alkaline phosphatase conjugate as with the second antibody (diluted to l/500). In these studies the following criteria and definitions were used. The inhibitory capacity of a preparation was assessed as the amount of sample required to produce 50% inhibition. The lower limit of detectability was defined as the amount of inhibitor that was required to produce 20% inhibition. The sensitivity of the assay was described by the slope of the inhibition curve and can be defined as the change in inhibition per picomole of inhibitor; thus, a steep slope indicates high sensitivity, and a shallower slope indicates a lower sensitivity. Chemical determinations. Total S-GAG content was determined spectrophotometrically using the 1,9-dimethylmethylene blue dye binding assay (45), adapted for use in microtiter plates (33). A negative absorbance change is obtained at 600 nm and a positive absorbance change is given at 530 nm (46); the absorbance was therefore determined at 530 and 600 nm, and the value for the change in absorbance was given by (change in absorbance at 530 nm + change in absorbance at 600 nm). KS from bovine cornea was used as a standard for KS and chondroitin 6-sulfate from shark cartilage was used as the standard for aggrecan. N-Acetylglucosamine and N-acetylgalactosamine were determined by the method of described by Perini and Peters (48). Briefly, samples were extensively dialyzed against water, hydrolyzed in 6 M HCl at 100°C for 3 h, evaporated under nitrogen, and dried. Samples were then reconstituted in 0.2 N sodium citrate, pH 3.25 (40 pg in 100 rl), and analyzed using a microcolumn amino acid analyzer with Auorometric detection. Amino acids in the hydrolysis do not interfere with the analyses (48). Body fluid samples. Normal human sera and urine were obtained from healthy volunteers (age 22-36). Synovial fluid was obtained from the knee joint of a patient with osteoarthritis (grade III). All samples

ET AL. were first digested Fab-ELISA.

with

papain

(as described

above)

prior

to analysis

by

RESULTS

Effect of Proteoglycan Fragment Size on the Antigenic Properties of KS by IgG-ELISA Gel filtration chromatography of human aggrecan and the KS-containing fragments, and analysis of the eluate fractions for S-GAG showed the preparations to be of a steadily decreasing average size (Fig. 1): AlDl > chondroitinase ABC digested AlDl > KSRP 1 > KSRP 2 > KSSC. The preparations KSRP 1 (aggrecan that has been digested with chondroitinase ABC followed by trypsin) and KSRP 2 (KSRP 1 after papain digestion) were heterogeneous in size, and KSSC (KS single chains prepared by sodium borohydride cleavage of chondroitinase ABCdigested aggrecan) eluted as a well-defined unimodal peak. Analysis of the eluates by IgG-ELISA (Fig. 1) using the antibodies 12.1,5D4, and MZ15 consistently showed that KS epitope was preferentially in the high M, fractions of each preparation. The difference between the elution profiles shown by the S-GAG analysis and the detectable KS epitope determination was most apparent with KSRP 1, where the first S-GAG peak contained more than 80% of the detectable epitope. However, analysis of the KSSC also showed that the detectable KS epitope eluted predominantly with chains of higher M,. The inhibitory capacities of the proteoglycan fragment preparations in the IgG-ELISA were found to be dependent on the size of the fragments (Fig. 2). Treatment of aggrecan with chondroitinase ABC had little effect on its inhibitory capacity. In contrast, proteolytic cleavage of the chondroitinase ABC-digested aggrecan with trypsin (to produce KSRP 1) and papain (to produce KSRP 2) or treatment with alkaline borohydride (to generate KSSC) resulted in a successive loss of inhibitory capacity (Fig. 2, Table I). The extent of this was dependent on the antibody used in the IgG-ELISA. Compared with aggrecan, the inhibitory capacity of KSRP 1 was decreased by as much as threefold, KSRP 2 was 4- to Ill-fold less inhibitory, and KSSC was 6- to 261-fold less inhibitory. The IgG-ELISA with antibody 12.1 was the least susceptible to the different sizes of the inhibitors, and the assay with antibody lB4 was affected the most by the difference in the size of the inhibition fragments (Table I). Sensitivity and Detectability of the IgG-ELISA The lower limit of detectability (the amount of aggrecan that was required to achieve 20% inhibition) was found to vary as a function of the antibody used. Greatest detectability was achieved with antibodies 12.1 and 4A4, and antibodies lB4 and MZ15 provided the lowest level of detectability (Table II). Maximum inhibition of approximately 100% was achieved with aggrecan or chon-

ANTIGENIC 200

PROPERTIES

OF

KERATAN

200

130

413

SULFATE

ld.-s C n .-"

130

,,.,..**0'

SE s .-

gg

60

,,..O ,..=d

MZ15 184

l.

SD4

l,..2; ,.*..*.-*..*’ ..” f

104-

ld.

60

O... . . . . . . . . . . .& :e

0

;;g

0

lo'

110

I

Id-AGGRECAN

70

=

30

E a 2

0

s

$

KSRPl

A2rAz;AU-

KSRPP

KSSC

FIG. 2. Inhibitory capacity of proteoglycan fragments in the IgGELISA. The amount of proteoglycan fragment required to produce a 50% inhibition in the IgG-ELISA was determined. Values represent the mean + standard deviation of three independent experiments. AggrecanCHase, chondroitinase ABC-digested aggrecan; KSSC, chondroitinase ABC-digested aggrecan treated with alkaline borohydride to prepare KS single chains. The antibodies used were 12.1 (A), 4A4 (O), 2D3 (II), 5D4 (m), lB4 (A), and MZ15 (0).

droitinase ABC-digested aggrecan as inhibitor. However, this was not necessarily achieved when the proteoglycan fragments were used as the inhibitor. The extent of this was dependent on the antibody used and the fragment used as the inhibitor (Fig. 3). The sensitivity of an assay is described by the slope of the inhibition curve (49). The sensitivity of the IgGELISA was found to be dependent both on the inhibitor and on the antibody. When aggrecan was used as the inhibitor, antibodies 12.1, 4A4, 2D3, and 5D4 provided in60

60

0

TABLE

0 0

0.2

0.4

0.6

0.8

1

I

The Effect of Proteoglycan Fragment Size on the Relative

Kav

Inhibitory

Capacity

of KS-Containing

Proteoglycan Fragments Human FIG. 1. Gel filtration of human (a-d) proteoglycan fragments. The proteoglycan preparations chondroitinase ABC-digested AlDl (a), KSRP 1 (b), KSRP 2 (c), and KSSC (d) were chromatographed on a column of Sepharose CLGB, and the eluate fractions were analyzed to determine S-GAG (0) and by ELISA with antibodies 12.1 (A), 5D4 (O), and MZ15 (0). The S-GAG profiles showed the AlDl to elute at the void volume (results not shown), and the chondroitinase ABC-treated AlDl (a) appeared to be reduced in size, eluting at and near the V, (Kav = 0.0 to 0.2/0.3). The trypsin-prepared fragment KSRP 1 (b) eluted over a broad M, range (K,” = 0.1 to 0.8) and displayed a bimodal elution pattern with two distinct S-GAG peaks (K,, = 0.25 and 0.5) of similar proportions. The papain-derived KSRP 2 fragment (c) from both species was eluted as a single, but relatively broad, peak (K., = 0.5 to 0.8, peak 0.6). The borohydride-treated material (d) eluted later than KSRP 2, near the V, of the column, as a well-defined unimodal peak with a Kav between 0.65 and 0.9 (peak 0.75-0.8).

12.1 CORE KSRP KSRP KSSC

1 2

4A4

0.9 1.3

0.9 1.9

4.3 6.1

5.8 15.8

articular

cartilage

proteoglycans

2D3

5D4

lB4

MZ15

0.9

0.9

0.9

0.9

1.5 5.0 9.3

1.6 5.5 31.7

3.0 111.4 260.7

3.3 33.4 79.9

Note. All results are compared relative to aggrecan, which was given the value of 1.0. The values given are (concentration of fragment (pmol NAcGlcNH,/ml))/(concentration of AlDl (pmol NAcGlcNHJml)) required to achieve 50% inhibition in the IgG-ELISA, with each of the monoclonal antibodies. The data are derived from Fig. 3. With all antibodies, the inhibitory capacity decreased with decrease in size of the proteoglycan fragments. The assay using the antibody 12.1 showed the least effect of fragment size, and the assay with antibody lB4 showed the greatest susceptibility to fragment size.

414

SEIBEL TABLE The

Lower

Antibody

II

Limit of Detectability of KS Aggrecan, by IgG-ELISA

Epitope

in

N-Acetylglucosamine (pmol/ml) for 20% inhibition

12.1 4A4 2D3 5D4 lB4 MZ15

0.8 0.9 2.2 2.5 9.4 50.2

Note. The lower limit of detectability was determined by the concentration of Nacetylglucosamine (pmol/ml) in aggrecan required to achieve 20% inhibition. Antibody 12.1 was found to provide the greatest detectability in the IgG-ELISA, and antibody MZ15 provided the poorest level of detectability.

hibition curves with a slightly steeper slope, and therefore higher sensitivity, than the inhibition curves obtained with antibodies lB4 and MZ15 (Figs. 3a-3c). Aggrecan and the chondroitinase ABC-digested aggrecan generated inhibition curves that were parallel to each other. However, the inhibition curves for the proteoglycan fragments generated by proteolytic digestion were often less steep than the respective slopes obtained with aggrecan, and the magnitude of this effect was found to be dependent on the antibody used. This reduction in the sensitivity of the assay with a decrease in fragment size was most pronounced using antibodies 5D4 and 2D3, whereas antibodies 12.1 and 4A4 showed changes of intermediate degree, and this effect was almost absent when antibodies lB4 and MZ15 were employed (Figs. 3a-3c).

Sensitivity

and Detectability

ET

AL.

arations when used in the IgG-ELISA (31.7-fold) (Table I). Importantly, there was no detectable difference in the inhibitory capacities of the KSRP 2 and KSSC preparations, unlike that found in the IgG-ELISA (Fig. 3, Table I).

The Analysis of Body Fluids with the Fab-ELISA Analysis of aggrecan, synovial fluid, serum, and urine by gel filtration chromatography and analysis of the eluate (after papain digestion) by Fab-ELISA showed the samples to contain epitope-bearing proteoglycan fragments of different sizes (Fig. 6). Aggrecan eluted at the V,,, and the urine eluted primarily as a single peak of small molecular size (K,, = 0.51). The synovial fluid and the serum

(a) 1

60. 40.

of the Fab-ELISA

The use of human aggrecan as inhibitor in the FabELISA (replacing the 5D4 whole antibody with the monovalent Fab fragments of antibody 5D4) generated an inhibition curve with a similar slope to the inhibition curve generated in the IgG-ELISA. The lower limit of antigen detectability of the Fab-ELISA, with aggrecan as inhibitor, was found to be improved by 9.5-fold compared to the IgG-ELISA (Fig. 4).

loo MZ1 60 60 40 20

Effect of Proteoglycan Fragment Size on the Antigenic Properties of KS in the Fab-ELISA Human aggrecan, KSRP 2, and KSSC were used as inhibitors in the Fab-ELISA (Fig. 5). The inhibition curves generated by aggrecan and the aggrecan fragments were all parallel, except at the highest concentrations of inhibitor used. Aggrecan was found to have the highest inhibitory capacity, and the KSRP 2 and KSSC preparations both had 5%fold less inhibitory capacity. This was less than the difference observed between these prep-

0! lo-2

FIG. 3. Comparison of the inhibition curves obtained using the human proteoglycan preparations and the proteoglycan fragments as inhibitors in the IgG-ELISA. Known concentrations of proteoglycan monomer and proteoglycan fragments were used as inhibitors in the immunoassays with antibodies 12.1 (a), 5D4 (b), and MZ15 (c) to allow comparison of the inhibition curves. The preparations used were aggrecan (O), chondroitinase ABC-digested aggrecan (O), KSRP 1 (O), KSRP 2 (m), and KSSC (A). The concentration of inhibitor is expressed as pmol N-acetylglucosamine/ml.

ANTIGENIC

PROPERTIES

OF

KERATAN

415

SULFATE 9

= 'z

8

2

6

7

r

5 3

4

B

3

.Z0

2

E s

1 0 0

0.2

0.4

0.6

0.8

1

Kav N-acetylglucosamine(pmolelml) FIG. 4. Comparison of the sensitivity and detectability of the IgGELISA and the Fab-ELISA. AlDl was used as inhibitor in the IgGELISA (0) and Fab-ELISA (0) using antibody 5D4. The Fab-ELISA was found to have its detectability level increased by 9.5fold, and both assays had parallel inhibition curves using AlDl as inhibitor.

eluted over a broad range, including fragments that eluted near the V, of the column, and fragments that eluted at the same K,, as the KS-containing fragments detected in the urine. Samples of papain-digested human aggrecan, synovial fluid, serum, and urine were analyzed with the FabELISA. Inhibition curves with parallel slopes were obtained for all the samples analyzed (Fig. 7), and by using the papain-digested aggrecan as the standard, the level of apparent epitope in the samples could be calculated (Table III). These results indicate that the levels of KS epitope in the serum and urine are similar, whereas the

FIG. 6. Gel filtration chromatography of KS epitope in body fluids. AlDl (O), synovial fluid (O), serum (Cl), and urine (B) were applied to a Sepharose CLGB column, and the eluate fractions were digested with papain and analyzed by Fab-ELISA. The chromatography indicates that the different body fluids each contain KS epitope-containing fragments of different sizes.

level of KS epitope in the synovial fluid is likely to be much higher (86-fold higher in the example presented in Table III). DISCUSSION

The heterogeneity of the structure and organization of KS on the protein core of aggrecan makes KS an unusual antigen. Previous studies have indicated that certain structural characteristics of aggrecan influence the quantitation of KS (19, 26). This present study has described the contributions that the structural presentation of the

1 100

80 80. 60

Oo N- acetylglucosamine

(pmole/ml)

FIG. 5. Comparison of the inhibition curves obtained using the human proteoglycan preparations and proteoglycan fragments as inhibitors in the Fab-ELISA. Known concentrations of AlDl (0) and the proteoglycan fragments KSRP 2 (0) and KSSC (A) were used as inhibitors in the immunoassay with the Fab fragments of antibody 5D4.

1

2 log

3 4 (l/dilution)

5

FIG. 7. Analysis of body fluids by Fab-ELISA. Human 4 M guanidineHCl extract of articular cartilage (Cl), synovial fluid (O), serum (m), and urine (0) were digested with papain and used as inhibitors in the FabELISA. All samples generated inhibition curves that were parallel to each other and to the papain-digested human aggrecan which was used as standard.

416

SEIBEL TABLE

The Concentrations Determined Sample Synovial fluid ( n = 1) Serum (n = 10) Urine (n = 10)

III

of KS Epitope in Body Fluids, Using the Fab-ELISA KS epitope (ng AlDl/ml) 3642 61.7 + 21.4 47.6 + 27.8

Note. The samples were first digested with papain, and then the level of KS epitope in each of the samples was determined using the ELISA with the Fab fragment of antibody 5D4. The sera and urine samples were paired and were obtained from healthy individuals. The synovial fluid was obtained from the knee joint of a patient with osteoarthritis. Since only synovial fluids from degenerative joints were available only one was analyzed as an example of the levels that can be present in synovial fluid.

KS, the choice of antibody, and the bivalency of the antibody make to the apparent antigenic properties of KS. These studies have resulted in the development of a protocol which allows the quantitative determination of KS epitope while minimizing the effects of the structural presentation of the KS. The contribution of the structural presentation of KS in its quantitation in the IgG-ELISA was studied using characterized fragments of aggrecan. Trypsin digestion of aggrecan has previously been used to produce the KSrich region of proteoglycan, as this enzyme cleaves the core protein into large peptide fragments (9, 50). The presence of both large and small KS-epitope-containing fragments in the preparation KSPR 1 may suggest that the larger components are likely to be derived from the KS-rich region, and the smaller fragments may include KS chains from the chondroitin sulfate attachment region of the protein core. Papain digestion is considered to achieve a more thorough digestion of the core protein, and gel filtration chromatography of KSRP 2 indicated that these fragments consisted of small peptides bearing only a few KS chains, as well as a minor proportion containing only KS single chains. Alkaline borohydride treatment has been shown to release O-linked oligosaccharides and polysaccharide from the peptide linkage region (42, 51). This material therefore was considered to contain principally single KS chains. In these studies the apparent epitope concentration was always highest in the preparations containing the largest fragments. This phenomenon, previously termed a cooperative binding effect (26), is created by the presence of adjacent KS chains bound to the protein core, providing a high local concentration of epitopes and thus enhancing the potential for the antibody to bind to the epitope. This general observation was true for any antibody used and in both the IgG-ELISA and, to a lesser extent, the FabELISA. This phenomenon could be observed even within one preparation; gel filtration chromatography and anal-

ET

AL.

ysis of the fractions by ELISA showed each preparation (except the KSSC) to be heterogeneous in size, and the epitopes were found to be present preferentially in the eluate fractions containing the higher molecular size fragments and thus those fragments with the highest number of KS chains. The degree of the effect was dependent on the antibody used and could be as little as 6-fold (antibody 12.1) and as much as 261-fold (antibody lB4). The variation of structure within a single chain offers a further level of complexity to the analysis. Keratan sulfate chains may be variable in terms of their length, degree of sulfation, and sulfational organization along the chain (11, 12). The epitope was previously shown to be dependent on a minimum of five adjacent saccharides each containing a sulfate group (23) and this structural organization is likely to occur only toward the nonreducing end of the longer KS chains. The presence of a series of fully sulfated oligosaccharides offers the possibility of more than one epitope being present on the longer chains (26). Analysis of the KS single chains by immunoassay of the eluate fractions showed that the immunodetermination of KS epitope was biased to the larger KS chains, with a high proportion of the smaller KS chains showing little or no immunoreactivity. This phenomenon is referred to as polyvalency (19, 26). Previous studies (19, 26) using competitive radioimmunoassays with monovalent Fab fragments of anti-KS antibodies show that the bivalency of the anti-KS antibody contributes to differences in the apparent epitope level detected in a particular sample. It has been shown that bivalency contributes to the phenomenon of cooperative binding (19). In the present studies a decrease in the difference between the apparent concentration of epitope was observed between the samples containing the large and the small fragments, confirming the previous study. The bivalency of the antibody has also previously been shown to contribute to the phenomenon of polyvalency (26). Importantly, in the Fab-ELISA, the inhibition curves generated by the samples were all parallel, in contrast to the inhibition slopes obtained with the IgGELISA, and no difference was apparent between the two small fragment preparations. Therefore the difference in apparent epitope levels in the two small fragment preparations is primarily due to the bivalency of the antibody contributing to the cooperative binding. The results show that all of the monoclonal antibodies

tested have different These

can each influence

antigen-binding

characteristics.

the effects of cooperative

binding

and polyvalency. This may be due to (a) the difference in epitope structure recognized by the different antibodies or (b) the flexibility of the IgG molecule at the hinge region, where the two Fab regions can move relative to each other. The effect of the latter can be removed by using the Fab fragments, and suggeststhat results obtained with

ANTIGENIC

PROPERTIES

different monoclonal antibodies may show more uniformity if the Fab protocol was used. The cooperative binding and polyvalency effects, contributed by multiple KS chains per protein core and the bivalency of the antibody in the assay, could therefore be overcome by the combination of two approaches. First, the use of the monovalent Fab fragments, in place of the IgG, can delete the differential cooperative binding effects for small fragments. Second, all samples could be reduced in size, by papain digestion, to one below which the cooperative binding effects are not detectable. The improved detectability of the Fab-ELISA compared to the IgGELISA resulted in an assay which allowed the determination of KS epitope in papain-digested samples at lower concentrations than was previously possible. In this study we applied this approach (Fab-ELISA and papain digestion of the samples) to the analysis of different body fluids which were shown to contain KS-containing fragments of various sizes. The inhibition curves obtained were all parallel and provided quantitative data on the levels of KS epitope in these samples. Thus the approach of using digested samples and Fab-ELISA appears to overcome the major problems of the quantitation of KS epitope, at least to the level of the single KS chain. It should be noted that the presence of multiple epitopes on a single chain and possible variations in epitope structure for different monoclonal antibodies, remain as obstacles to the definitive determination of KS epitope in a given sample.

We thank Dr. B. Caterson, Dr. E. Thonar, The Kennedy Institute of Rheumatology, and Ciba Geigy (Switzerland) for the generous gifts of the monoclonal antibodies. Also, we thank Dr. Perini for performing the amino sugar analyses, Dr. A. R. Poole and Dr. P. Roughley for the gift of human AlDl, and Dr. F. Woessner for providing human articular cartilage. Finally we thank Dr. Paul Beauvais for valuable assistance and Ms. Veronica Hlibczuk for excellent technical assistance. This work was supported by National Institutes of Health Grant AR40032 and Deutscher Akadekischer Austauschdienst 312 024 501 7.

REFERENCES K., Linker, A., Davidson, Chem. 205,611-616.

2. Muir, H., and Hardingham, Ser. One 5, 153-220. 3. Hascall, V. C., and Riolo, 4538. 4. Hascall, 4930.

V. C., and Sajdera,

E. A., and Weissman,

T. (1975)

MTP

R. L. (1972) S. W. (1970)

B. (1953)

Znt. Reu. Sci. Biochem.

J. Biol.

Chem.

247,

4529-

J. Biol.

Chem.

245,4920-

5. Heinegard, D., and Axelsonn, I. J. (1977) CellBioZ. 252,1971-1979. 6. Stevens, R. L., Ewins, R. J. F., Revell, P. A., and Muir, H. (1979) Biochem.J. 179,561-572. 7. Doege, K., Sasaki, M., Horigan, E., Hassell, J. R., and Yamada, Y. (1987) J. Biol. Chem. 262, 17,757-17,767. 8. Doege, K., Sasaki, M., Kimura, T., and Yamada, Y. (1991) J. Biol. Chem. 266,894-902. 9. Heinegard,

166,427-442.

D., and Hascall,

KERATAN

417

SULFATE

10. Sheehan, J. K., Ratcliffe, (1987) &o&em. J. 247,

A., Oates, 267-276.

K.,

and Hardingham,

T. E.

11. Oeben, M., Keller, R., Stuhlsatz, H. W., and Greiling H. (1987) Biochem. J. 248,85-93. 12. Stuhlsatz, H. W., Keller, R., Becker, G., Oeben, M., Lennarts, L., Fisher, D. C., and Greiling, H. (1989) in Keratan Sulfate Chemistry, Biology, and Chemical Pathology (Greiling, H., and Scott, J. E., Eds.), pp. l-11, Biochemical Society, London. 13. Thornton, D. J., Morris, H. G., Cockin, G. H., Huckerby, T. N., Neiduszynski, I. A., Carlstedt, I., Hardingham, T. E., and Ratcliffe, A. (1989) Biochem. J. 260, 277-282. 14. Caterson, B., Christner, J. E., and Baker, J. R. (1983) J. Bid. Chem. 258,8848-8854. 15. Caterson, B., Mahmoodian, F., Sorrell, M. J., Hardingham, T. E., Bayliss, M. T., Carney, S. L., Ratcliffe, A., and Muir, H. (1990) J. Cell Ski. 97, 411-417. 16. Conrad, G. W., Ager-Johnson, Chem. 257,464-471. 17. Couchman, J. R., Caterson, (1984) Nature 307, 650-652. 18. Funderburgh, (1983) Curr.

P., and Woo, B., Christner,

J. L., Stenzel-Johnson, Eye Res. 2, 769-776.

19. Poole, A. R., Webber, C., Reiner, Biochem. J. 260, 849-856.

M.-L.

(1982)

J. Biol.

J. E., and Baker,

P. R., and Chandler, A., and Roughley,

J. R. J. W.

P. J. (1989)

20. Seibel, M. J., Towbin, H., Braun, D. G., Kiefer, B., Mueller, W., and Paulsson, M. (1989) in Keratan Sulfate Chemistry, Biology, and Chemical Pathology (Greiling, H., and Scott, J. E., Eds.), pp. 191-198, Biochemical Society, London. 21. SunderRaj, N., Wilson, J., Gregory, Curr. Eye Res. 4, 49-51.

J. D., and Damle,

22. Zanetti,

F. (1985)

M., Ratcliffe,

A., and Watt,

S. P. (1984)

J. Cell Biol.

101,53-

59. 23. Mehmet, H., Scudder, P., Tang, P. W., Hounsell, E. F., Caterson, G., and Feizi, T. (1986) Eur. J. Biochem. 157, 385-391.

ACKNOWLEDGMENTS

1. Meyer, J. Biol.

OF

V. C. (1974)

Arch.

Biochem.

Biophys.

24. Scudder, P., Tang, P. W., Hounsell, E. F., Lawson, F. H., and Feizi, T. (1986) Eur. J. Biochem. 157,

A. M., Mehmet,

365-373.

25. Sheehan, J., Ratcliffe, A., Oates, K., and Hardingham, T. E. (1987) Biochem. J. 247, 267-276. 26. Caterson, B., Brooks, K., Sattsangi, S., Ratcliffe, A., Hardingham, T., and Muir, H. (1989) in Keratan Sulfate Chemistry, Biology, and Chemical Pathology (Greiling, H., and Scott, J. E., Eds.), pp. 199203, Biochemical Society, London. 27. Tang, P. W., Scudder, P., Mehmet, H., Hounsell, E. F., and Feizi, T. (1986) Eur. J. Biochem. 160, 537-545. 28. Funderburgh, J. L., Caterson, B., and Conrad, G. W. (1987) J. Biol. Chem. 262, 11,634-11,640. 29. Ratcliffe, A., Tyler, J. A., and Hardingham, T. E. (1986) Biochem. J. 238,571-580. 30. Leipold, H. R., Goldberg, R. L., and Lust, G. (1989) Arthritis Rheum. 32, 312-321. 31. Mehraban, F., and Moskowitz, R. W. (1989) Arthritis Rheum. 32, 129331299. 32. Pavelka, K., and Seibel, M. J. (1989) Z. Rheumatol. 48, 294-300. 33. Ratcliffe, A., Doherty, M., Maini, R. N., and Hardingham, (1988) Ann. Rheum. Dis. 47, 826-832.

T. E.

34. Ratcliffe, A., Billingham, M. E. J., Saed-Nejad, F., Muir, H., and Hardingham, T. E. (1992) J. Orthop. Res. 10, 350-358. 35. Sweet, M. B., Coelho, A., Schnitzler, C. M., Schnitzer, T. J., Lenz, M. E., Jakin, I., Kuettner, K. E., and Thonar, E. J.-M. A. (1988) Arthritis Rheum. 31,648-652. 36. Thonar, E. J.-M. A., Lenz, M. E., Klintworth, G. K., Caterson, B., Pachman, L. M., Glickman, P., Katz, R., Huff, J., and Kuettner, K. E. (1985) Arthritis Rheum. 28, 1367-1377.

418

SEIBEL

37. Israel, H. A., Saed-Nejad, ilbfac Surg. 49, 708-711. 38. Hascall,

V. C., and Sajdera,

39. Hardingham, 40. Ratcliffe, 378.

T. E. (1979) A., and Hardingham,

F., and Ratcliffe,

A. (1991)

J. Oral Max-

D. M. (1968)

AL.

45. Farndaie,

R. W., Sayers,

C. A., and Barrett,

A. J. (1982)

Connect.

D. J., and Barrett,

A. J. (1986)

Biochem.

Tiss. Res. 9,247-248. S. W. (1969) &&em.

J. Biol.

J. 177, T. E. (1983)

Chem.

46. Farndale,

244,2384.

237-247. Biochem.

J. 213,371-

J. Biol. Chem. 243,616-626.

43. Caterson, B., Baker, J. R., Christner, J. E., Lee, Y., and Lentz, (1985) J. Biol. Chem. 260,11,348-11,356. 44. Thonar, E. J.-M. A., Meyer, R., Dennis, R., Lens, M. E., Maldonado, B., Hassall, J., Hewitt, A. T., Stark, W. J., Stock, L., and Kuettner, K. E. (1986) Am. J. Ophthalmol. 102, 561-569.

R. W., Buttle,

Biophys. Acta 883,173-177.

41. Heinegard, D., and Sommarin, Y. (1987) in Methods in Enzymology (Cummingham, L. W., Ed.), Vol. 144, pp. 305-372, Academic Press, Orlando. 42. Carlson,

ET

M.

47. Lowry, (1951) 48. Perini, 49. Fraser,

0. H., Rosebrough,

N. J., Farr,

A. L., and Randall,

R. J.

J. Biol. Chem. 193,265-275. F., and Peters, B. P. (1982) And. Biochem. 123, 357-363. C. G., and Browning, M. C. K. (1986) Clin. Chem. 32,569.

50. Hardingham, T. E. (1981) B&hem. Sot. Trans. 9,489-497. 51. Lohmander, L. S., DeLuca, S., Nilsson, B., Hascall, V. C., Caputo, C. B., Kimura, J. H., and Heinegard, D. (1980) J. Biol. Chem. 255, 6084-6091. 52. Williams, J. M., Downey, C., Thonar, E. J.-M. A. (1988) Arthritis Rheum. 31,557-560. 53. Wit&, J., Roughley, P. J., Webber, C., Roberts, N., Keystone, E., and Poole, A. R. (1987) Arthritis Rheum. 30,519-529.

Antigenic properties of keratan sulfate: influence of antigen structure, monoclonal antibodies, and antibody valency.

The influence of (a) antigen structure, (b) type of monoclonal antibody, and (c) antibody bivalency on the immunochemical detection and quantification...
1MB Sizes 0 Downloads 0 Views