Immunobiol., vol. 185, pp. 103-120 (1992)

1 Institute of Medical Microbiology and University, Mainz, Germany

2 Department

of Pathology, Johannes Gutenberg

Modulation of Type II Collagen-Induced Arthritis in DBAI1 Mice by Intravenous Application of a Peptide from the Clq-A Chain MARKUS

J. MAEURERt, PETER K. E. TRINDERt, STEfAN STORKEL 2, and

MICHAEL LOOSl

Received February 26, 1992 . Accepted in Revised Form April 9, 1992

Abstract In this report we are able to show that intravenous (i.v.) application (day 0) of a nonapeptide (residues 26-34) from the human C1q A-chain (designated peptide A-C1q) prior to intradermal (i.d.) administration of chicken type II collagen (ClI) in arthritis-susceptible DBA/1 mice (H2q), leads to abrogation of polymorphonuclear neutrophil (PMN) invasion into the joints. This nonapeptide exhibits epitope characteristics and high homology to residues 137-147 of CB11 (a cyanogen bromide fragment of chicken ClI, known to contain both arthritis inducing and suppressing determinants). Arthritis index was lowest in animals pretreated i.v. with CII (as internal control), though animals pretreated i.v. with peptide K (residues 137-147 with an additional glycine residue from CBll) or peptide A-C1q exhibited comparative arthritic indices. Only in the arthritis-positive control group (day 0: PBS i.v.) did i.d. application of CII lead to invasion of PMN into the synovial layer and the joint space. Analysis of antibody (Ab) responses at day 48 after i.v. immunization (day 0) and cn challenge (day 7) revealed IgE-Abs to native CII and also to native C1q. IgG titers to CII were highest in animals pretreated with peptide A-C1q. Abs from this group, exhibiting activity to peptide A-C1q (immunizing antigen), were of mainly IgG1 and IgG3 isotypes. Evaluation of the immune response following i.v. application of peptide A-C1q or ClI, prior to i.d. cn administration, in DBA/1 mice, revealed IgM responses to peptide A-C1q and peptide K, but not to CII. Intravenous application of peptide A-C1q led to generation of IgG3-Abs reacting only with peptide A-C1q and peptide K, but not with native CII. Additionally, i.v. application of peptide A-C1q elicited IgG responses to a pentapeptide, resembling amino acid residues 26-30 (K-G-E-Q-G) of the C1q A-chain. This five residue antigenic determinant is present in peptide K, in chicken and human CII as well as in human C1q. No specific IgE response to any of the antigens tested could be detected. Since a peptide from the C1q A-chain is both capable of eliciting immune responses and modulating crr -induced arthritis in mice, we postulate that the collagen-like

Abbreviations: i. v. = intravenous; ClI = type rr collagen CB11 = cyanogen bromide fragment of type II collagen; peptide K = dodecapeptide, resembling a.a. 137-147 of CBll + an additional glycine residue; peptide A-Clq = nonapeptide, resembling a.a. 26-34 of C1q A-chain; i.d. = intradermal; PBS = phosphate-buffered saline; PMN = polymorphonuclear neutrophils; Ab = antibody; Clqnat = highly purified native C1q; Clqox = C1q oxidized by hydrogen peroxide (0.3 %,30 min, 37°C); IFA = incomplete Freund's adjuvant; PBST = PBS containing 0.05 % Tween 20; TN = solution of 0.01 M Tris and 0.1 M NaCI; MHC = major histocompatibility complex; TH = T helper cells; IL = interleukin; IFN = interferon; TNF = tumor necrosis factor; DTH = delayed-type hypersensitivity; CFA = complete Freund's adjuvant

104 .

M.

J. MAEURER,

P. K. E.

TRINDER, S. STORKEL,

and

M.

Loos

complement component C1q is involved in the development of ClI-induced inflammatory arthritic lesions, and may represent, in vivo, the early antigen responsible for inducing anticollagen antibodies prior to ClI in hyaline cartilage becoming available as antigen.

Introduction Type II collagen (ClI) has been shown to be involved in the pathogenesis of human rheumatoid arthritis, as demonstrated by the presence of complement-fixing anti-CII antibodies (Ab) and CII -reactive T cells in serum and synovial fluid of rheumatoid arthritis patients (1, 2). Intradermal (i.d.) immunization of CII is known to elicit arthritic lesions in certain strains of rodents and primates (3-5). This CII -induced arthritis is characterized by an early proliferative synovitis, influx of polymorphonuclear neutrophils (PMNs) into the joint cavity and, ultimately, articular ankylosis (6, 7). Development of arthritic lesions is accompanied by cellular (8) and humoral (production of complement-fixing mouse IgG2a-Abs) immune responses to CII (9). Recent studies have shown that the immune response to ClI, as well as the development of CII -induced arthritic lesions, can be modulated by intravenous (i.v.) application of native CII or ClI-Abs, ClI-coupled spleen cells, partially purified collagen peptides or monoclonal Abs (mAbs), recognizing putative arthritogenic epitopes on ClI, prior to intradermal application of CII (10-14). Suppression of CII -induced arthritis in mice has also been achieved by oral feeding of native CII (but not altered ClI), as well as by transfer of CD4 + T cells from diseased animals (15-17). Arthritogenic determinants play an important role in the induction of CII -induced arthritis. A cyanogen bromide fragment of CII (CB 11) possesses determinants involved in development of CII-induced arthritis (18), and i.d. application of CB 11 alone is capable of eliciting arthritic lesions. CB 11 has been demonstrated to contain arthritis-suppressing as well as arthritisinducing determinants, one of the former being located between residues 137-147 from the N-terminus of the CBll split-product of the CII alchain (19). Since complement has been shown to play an important role in the induction of arthritic lesions (9, 20), and a predicted epitope on the A-chain of the collagen-like complement component Clq (residues 26-34 from the N-terminus) exhibits high homology to the 137-147 amino acid stretch from the CB 11 fragment of CII (18, 19), we synthesized a nonapeptide, representing residues 26-34 of the Clq A-chain. Using an arthritis-prone mouse strain as an animal model of ClI-induced arthritis, we characterized the humoral responses and histopathological events associated with i.v. application of this peptide. Intravenously administered peptide A-Clq was able to evoke Ab-responses (IgM and IgG3) in arthritis susceptible DBA/I mice (H2q). Antibodies were found to react not only with peptide A-Clq, but also with peptide K (residues 137-147 from the CBll fragment of ClI, + an additional glycine) and with

Modulation of Arthritis by C1q . 105

a five amino acid stretch (K-G-E-Q-G) fully identical to residues 26-30 from the Clq A-chain and to residues 141-145 from the CBll split-product of CII (see Fig. 1). Intravenous application of peptide A-C1q prior to intradermal inoculation of CII, abrogated the invasion of PMNs into the synovial layer and the joint space, an event seen to occur in the positive control group, i.e., i.v. application of PBS followed by i.d. administration of CII.

Materials and Methods Reagents 10 mM phosphate-buffered saline (PBS), pH 7.5, ionic strength 15 milli-Siemens, was used in all assays. Unless stated otherwise, all chemicals were obtained from Sigma (Munich, Germany). ELISA tests were carried out in polystyrene microtiter plates (Nunc-Immuno, Wiesbaden, Germany). Antigens Chicken sternal cartilage ClI, produced from a proteolytic digest, was obtained from Genzyme (Munich, Germany). Human C1q was purified using fast protein liquid chromatography as described by STEMMER and Loos (21). Highly pure native C1q (C1q nat) in PBS was oxidized (C1q ox) using hydrogen peroxide at a final concentration of 0.3 % (30 min, 37°C). Purity of C1q and CII was confirmed by SDS-PAGE, performed in 5-20 % gradient or 6 % polyacrylamide slab gels, respectively, using the method of LAEMMLI (22). Under reducing conditions, SDS-PAGE of C1q ox revealed the typical A-, B-, C-chain pattern expected of native C1q, hydrogen peroxide induced fragmentation could not be detected. A dodecapeptide (residues 137-147, peptide K) from the CBll fragment of ClI, with an additional glycine added at the C-terminal end, a nonapeptide (residues 26-34, peptide A-C1q) and a pentapeptide (residues 26-30, K-G-E-Q-G) from C1q, were kindly synthesized by Prof. Dr. MOLLERESTERL, Dept. of Physiological Chemistry, University of Mainz, Germany, using standard solid phase procedures. Purity was determined by high performance liquid chromatography, and sequences confirmed by amino acid analysis. Protein concentrations These were determined using the method of LOWRY (23), using bovine serum albumin as standard. Comparison of amino acid sequences The amino acid sequence for human type II collagen was obtained from the EMBL Data Library (Heidelberg, Germany, Feb 1992, data submitted by F. RAMIREZ SUNY Health Science Center, Brooklyn, NY, USA, last updated June 1991), and that for chicken CII from MYERS et al. (19). The sequence for human C1q was obtained from SELLAR et al. (24), and amino acid sequence comparison (Fig. 1) was performed using the method of MYERS and MILLER (25) (PC/Gene, Intelligenetics, USA). Mice Inbred male DBAIl mice (9-10 weeks old), were bought from Charles River Wiga (Sulzfeld, Germany), housed in groups of 2 or 3 in cages, and fed standard rodent chow with acidified water (pH 3.0) ad libidum.

106 . M.

J. MAEURER, P. K. E. TRINDER, S. STaRKEL, and M. Loos

Immunization Mice were immunized intravenously (day 0) with 100 [!l PBS or with 100 [!l PBS containing 100 [!g chicken ClI, 500 [!g peptide A-C1q or 500 [!g peptide K. Arthritis was induced by intradermal application (day 7) of 100 [!g chicken ClI in 100 [!l PBS mixed with 50 [!l incomplete Freund's adjuvant (IFA) at several sites on the back. Macroscopic evaluation Mice were examined daily for indications of arthritis. Each limb was graded from 0-4, according to the extent of erythema, severity of swelling and involvement of periarticular tissue. 0 = no change, 1 = slight change in the joints of the digits, 2 = slight edema of the paw or swelling of more than two digits, 3 = moderate swelling of the paw, 4 = severe swelling of the paw. Histologic evaluation Mice were sacrificed at day 48, hind paws were dissected, snap frozen in -70°C isopentane and stored in liquid nitrogen. Paws were embedded in paraffin, serially-sectioned and stained with Mayer's Haematoxylin. At least two sections from each joint were examined for signs of inflammation. Antibody determination Mice were sacrificed on day 7 or day 48 and bled by cardiac puncture. After clotting at room temperature for 1 h, blood was centrifuged at 1500 x g for 10 min and serum removed for storage at -20°C. IgM and IgG serum antibody titers were determined using the following ELISA Microtiter plates were coated with chicken ClI, C1q nat, C1q ox, peptide K, peptide A-C1q or the K-G-E-Q-G peptide overnight at 4°C (10 [!g/ml, 30 [!l!well in PBS). PBS containing 0.05 % Tween 20 (PBST) was used to remove excess antigen, and free sites were blocked with 2 % milkpowder in PBS. Serial dilutions of mouse sera were added to the wells in duplicate; following incubation and washing with PBST, the wells were incubated with peroxidaseconjugated anti-mouse IgM (Nordic Immunological, Tilburg, The Netherlands) or peroxidase-conjugated anti-mouse IgG (Bio Rad, Munich, Germany) diluted in PBST. Washing of the wells was followed by addition of 2,2' -azino-bis(3-ethylbenzthiazoline-6-sulfonic acid) (2 mg/ml in 10 mM sodium citrate, pH 4.5, containing 0.01 % H z0 2 ) to complete the reaction. IgG subclass determination Following antigen coating, blocking with milkpowder in PBS and washing in PBST, wells were incubated with goat anti-mouse IgG1, 2a, 2b or 3 (Nordic Immunological). Wells were then washed with 0.01 M TRIS containing 0.1 M NaCI (TN) and incubated with biotinlabelled rabbit anti-goat IgG in TN. Further washing with TN was followed by incubation with avidin-alkaline phosphatase conjugate in TN. After washing with TN, the wells were further washed with 1 M diethylamine containing 0.5 mM MgClz, pH 9.8. Incubation with p-Nitrophenyl phosphate, dis odium salt (Sigma 104 phosphatase substrate tablets, 1 tablet in 5 ml diethylamine/MgClz) brought the reaction to completion. Assessment of antigen specific IgE levels Following antigen-coating and blocking with milkpowder in PBS, wells were washed with TN. After incubation with monoclonal rat anti-mouse IgE (Dianova, Hamburg, Germany) in TN and washing, also with TN, wells were incubated with biotin-labelled rabbit anti-rat IgG in TN and then washed with TN. Avidin-conjugated alkaline phosphatase and phosphate substrate were applied as for IgG subclasses.

Modulation of Arthritis by Clq . 107

Results

Clinical course of arthritis To evaluate the ability of ClI, peptide K and peptide A-Clq (see Fig. 1) to modulate induction of ClI-induced arthritis, 18 DBA/l (H2q, arthritis susceptible) mice were split into four groups. Mice nos. 1-5 (group I) were immunized i.v. with PBS, mice nos. 6-10 (group II) with chicken CII, nos. 11-15 (group III) with peptide A-Clq and nos. 16-18 (group IV) with peptide K (see Fig. 2). Seven days later, all 18 animals were inoculated i.d. with CII emulsified with IF A. Clinical onset of arthritis, as determined by joint swelling, erythema and limitation of motion, occurred in all group I animals by day 44. The hind paws were affected first, with arthritis spreading from a primary affected foot into all joints. A mean arthritis score of 4.1 ± 1.0 was attributed to animals in group I. By day 46, this had increased to 5.7 ± 1.5 and by day 48, the value was 7.1 ± 1.5 (Fig. 2). First signs of arthritis were not detected in group II animals until between days 46 and 48 (mean arthritis score 0.5 ± 0.5 by day 46, 4.8 ± 1.0 by day 48); whilst in groups III and IV, some signs were observable by day 44 (mean arthritic indices 1.8 ± 1.0 and 1.9 ± 1,0, respectively, compared to 4.1 ± 1.0 group I). The values for these groups had risen to 3.0 ± 1.0 and 3.0 ± 1.5, respectively, by day 46 (group I: 5.7 ± 1.5) and to 4.7 ± 1.0 and 5.0 ± 1.5 by day 48 (group I: 7.1 ± 1.5, group II: 4.8 ± 1.0).

Histologic evaluation Histologic examination of sectioned material from hind paws of sacrificed mice (day 48) from group I (i.v. PBS, i.d. ClI) correlated well with

137 GB11:

147

(I)

I-A-G-F-~-Q-E-Q-Q-P-K

(chicken)

G1q A-

G-L-~-Q-E-Q-Q-E-P-G-A

24 :

chain:

34

Gil

Q-P-G-A-~-Q-E-Q-Q-E-A

(human)

344

(II) (III)

354

Figure 1. Homology between the «immunodominant» epitope present within the chicken type II collagen fragment, CB 11, human type II collagen and an epitope present in the collagen like region of human Clq A-chain. (I): «lmmunodominant» epitope present within type II collagen CBll fragment (residues 137-147) (18, 19). (II): Amino acid residues of Clq A-chain (residues 24-34 from N-terminus) (24). (III): Amino acid residues of human type II collagen (residues 344-354 from N-terminus of aI-chain (data from EMBL Data Library, Heidelberg, Germany, February 1992). Points indicate identical amino acids, underlined Clq A-chain residues indicate the predicted antigenic determinant of Clq.

108 . M.

J. MAEURER, P. K. E. TRINDER, S. STORKEL, and M. Loos 10 .---------------------------- - - - - - - ,

x

8

Q)

-0

c: u

6

+'

.... .r: +' .... 0

4

c: 0

Q)

::::E

2

44

46

48

Days after immunizat io n

Figure 2. Modulation of type II collagen induced arthritis in male DBA/1 mice by intravenous application of CII, peptide A-C1q or peptide K. Male, 6-8 weeks old DBA/1 mice were injected intravenously (day 0) with phosphate-buffered saline (PBS) (Group I [cross-hatching]), 100 flg of CII (Group II [filled bars]), 500 flg of peptide A (Group III [unfilled bars]) or 500 flg of peptide K (Group IV [hatching]), followed (day 7) by i.d. application of 100 flg CII in IFA. Severity of arthritis symptoms of forepaws and hind paws of each mouse was scored on a scale of 0-4, where a = no change, 1 = slight change in the joints of the digits, 2 = slight edema of the paw or swelling of more than 2 digits, 3 = moderate swelling of the paw, 4 = severe swelling of the paw. Sum of the scores of all 4 paws for each mouse represents the arthritis index. Values resemble mean maximum arthritis score of 5 animals and bars show the SEM of each group.

A

Modulation of Arthritis by C1q . 109

B

/

I

\ · ,t

)

-

....... \

-.

---01 .",

U

O® .....

,

-I

.

-~,

@®

..::-- --. _- .

---~-

.

-C

Figure 3. Histologic sections of mice hindpaws. A: Section (group III mice) of transition zone of synovialis (s) and hyaline cartilage (c) at day 48 (day 0: peptide A -C1q i.v., day 7: ClI i.d.), showing no signs of inflammatory activity . This also proved to be characteristic for group 1I (day 0: CII i.v., day 7: cn i.d.) and group IV (day 0: peptide K i.v., day 7: cn i.d.) animals (Haematoxylin and Eosin, HE, original magnification X2S8). B: Positive control (group I mice) ankle section at day 48 (day 0: PBS i.v., day 7: CII i.d.) showing extensive infiltration of polymorphonuclear neutrophils (PMNs) within the synovial layer and extending into the joint space (HE, original magnification x282). C: Section of positive control joint (group I mice) at day 48 (day 0: PBS i.v., day 7: ClI i.d.), PMN (arrow) are distributed within the joint space. This proved to be characteristic for aU group I mice (HE, original magnification x282).

110 . M.

J. MAEURER, P. K. E. TRINDER, S. STORKEL, and M. Loos

clinical observations, as evidenced by massive infiltration of PMN into the synovial layer (Fig. 3b) and into the joint space itself (Fig. 3c). The spectrum of mild to major PMN invasion into the joint is a reflection of serial histological sections representing different levels within the paw joints. Unlike group I animals, analysis of sectioned-material from group III animals (i.v. peptide A-Clq, i.d. CII) failed to reveal any signs of inflammatory activity in synovial tissue or the joint space (Fig. 3a). Material from groups II (i.v. CII, i.d. CII) and IV (i.v. peptide K, i.d. CII) also showed no signs of inflammation (e.g. PMN infiltration) in synovial tissue or the joint space.

Antibody responses to intravenous application of type II collagen or peptide A-Clq Five DBA/l mice each were immunized i.v. (day 0) with PBS, CII or peptide A-Clq, sacrificed on day 7 and bled by cardiac puncture. AU sera were assessed for reactivity with native CII, human Clq in its native (Clq nat) and oxidized (Clq ox) forms, peptide K, peptide A-Clq and the K-G-E-Q-G pentapeptide. The results of these assays are given in Table l. Mice immunized with PBS failed to exhibit significant Ab titers to any of the antigens tested. CII recipients produced IgM Abs that reacted weakly with peptide A-Clq, K-G-E-Q-G, peptide K and Clq ox but, notably, not with CII or Clq nat. Mice receiving peptide A-Clq had IgM Abs that reacted well with peptide A-Clq, K-G-E-Q-G and peptide K, but only weakly with Clq ox and not at all with CII or Clq nat. A weak IgG (lgG3) response was observed, in CII immunized animals, against peptide K, all Table I. IgM, IgG and IgE serum antibody response at day 7 after i. v. application (day 0) of peptide A-Clq (Pep A-Clq) or type II collagen (ClI) in arthritis susceptible DBA/l mice reacting with peptide A-Clq, K-G-E-Q-G peptide, peptide K, CII and Clq in its native and oxidized forms'" Treatment of mice at day 0

Pep A-Clq

K-G-EQ-G

Pep K

20

20 20 (IgG3)

CII i.v. DBA/l mice (H2q)

IgM IgG IgE

20

Pep A-Clq i.v. DBA/l mice (H2q)

IgM IgG IgE

400 400 800 (IgG3) 400

400 100 (IgG3)

CII

Clq nat

Clq ox 20

20 -

". Serum IgM, IgG and IgE titer profiles are expressed as highest reciprocal measurable dilution in ELISA, using peptide A-Clq, the K-G-E-Q-G peptide, peptide K, native CII and Clq in its native and oxidized forms as antigen. Results are given as mean values of 5 animals in each group. IgG subclasses (brackets) were determined using an ELISA-system as described in materials and methods. A group of animals immunized i.v. with PBS, failed to give significant Ab titers with any of the above antigens.

Modulation of Arthritis by Clq . 111

other antigens failed to react with IgG from these animals. Peptide A-Clq recipients, on the other hand, exhibited a strong IgG response (lgG3) to peptide A-Clq, a good response to K-G-E-Q-G, and a somewhat weaker response to peptide K (lgG3). No IgE Abs, to any of the antigens tested, could be detected.

Antibody responses following intradermal administration of type II collagen Animals from groups I-IV (see Materials and Methods) were sacrificed and bled by cardiac puncture on day 48. The sera were individually assessed for reactivity with CII, Clq nat, Clq ox, peptide A-Clq and peptide K. Low levels of IgM serum Abs (Table 2) recognizing peptide K were observed in sera from all five group I animals (i.v. PBS, i.d. CII). IgM from one mouse from this group also exhibited weak binding to peptide A-Clq, Clq ox and CII. All group II mice (i.v. CII, i.d. CII) produced significant levels of IgM reacting with peptide A-Clq, and low IgM levels to peptide K. IgM Abs from 2 out of 5 mice reacted weakly with CII, and IgM Abs

Table 2. Detection of serum IgM antibodies recognizing peptide A-Clq, native Clq, Clq ox, type II collagen and peptide K in male DBAII mice at day 48, pretreated either with PBS, CII, peptide A-Clq or peptide K i.v. (day 0), followed (day 7) by i.d. administraion of CII"" Male DBAII Pep A-Clq mice No.

Clq nat

Clq ox

ClI

20 20 20 20 20

Group I Day 0: PBS i.v. Day 7: ClI i.d.

20

20 20 20 20 20

Group II Day 0: cn i.v. Day 7: ClI i.d.

20 20

20

20 20 20

20 200 20

2 3

4 5

20

6

200 200 200 200 200

7 8 9

10 11 12 13

14 15 16

17 18

20 200

20

20 20 20

20

20 20

20 20 20 20 20 200 200 200

20

20

Pep K

20 20 20 20 20

20 20 20

Group III Day 0: Pep A-Clq i.v. Day 7: ClI i.d. Group IV Day 0: Pep K i.v. Day 7: ClI i.d.

". Serum IgM antibody titer profiles are expressed as highest reciprocal measurable dilution in ELISA, using peptide A-Clq, native Clq, oxidized Clq, type II collagen and peptide K as antigen.

112 . M.

J. MAEURER, P. K. E. TRINDER, S. STORKEL, and M. Laos

from 3 out of 5 mice were found to weakly bind Clq nat, whilst sera from 3 out of 5 mice were found to contain low levels of IgM reacting with Clq ox. All mice from group III (i.v. peptide A-Clq, i.d. CII) produced IgM Abs that weakly recognized peptide A-Clq. IgM from 4 out of 5 sera recognized, weakly, Clq ox, and IgM from 2 out of 5 sera bound weakly to CII. An individual mouse produced IgM to peptide K, whereas none of the sera from this group had IgM Abs capable of binding to Clq nat. Sera from group IV animals (i.v. peptide K, i.d. CII) exhibited a significant IgM response to peptide A-Clq, an intermediate IgM response to peptide K and a weak IgM response to Clq nat, Clq ox and CII. Total IgG responses are shown in Table 3. Sera from group I animals all exhibited significant to high IgG titers to CII but no IgG titers to Clq ox. Two out of 5 mice had IgG Abs weakly recognizing peptide A-Clq, whereas 3 from 5 mice had low IgG titers to Clq nat. Only 2 mice showed any IgG titers to peptide K. Low to high IgG titers were observed against CII in sera from group II mice, though peptide K-binding IgG titers were low in all members of this group. All 5 sera were positive for IgG binding to peptide A-Clq, whilst 4 of 5 mice produced significant IgG titers to Clq Table 3. Detection of serum IgG antibodies recognizing peptide A-Clq, native Clq, Clq ox, type II collagen and peptide K in male DBA/l mice at day 48, pretreated either with PBS, ClI, peptide A-Clq or peptide K i.v. (day 0), followed (day 7) by i.d. administration of CW Male DBA/l Pep A-Clq mice No.

Clq nat

Clq ox

20 20

CII

Group I

200 8000 2000 8000 200

20 20

Day 0: PBS i.v. Day 7: CII i.d.

200 2000 20 20 2000

20 20 20 20 20

Day 0: CII i.v. Day 7: CII i.d.

2 3 4 5

20 20

6 7 8 9 10

200 20 20 20 20

200 200 200

200 200

200

20

11 12 13 14 15

2000 200 200 200 400

2000 200 200 2000 2000

20 20 20 20

32000 32000 2000 16000 4000

16 17 18

200 200 200

200 200 200

20 20 20

2000 2000 200

20

Pep K

Group II

Group III 20

20 200 200

Day 0: Pep A-Clq i.v. Day 7: CII i.d.

Group IV Day 0: Pep K i.v. Day 7: CII i.d.

". Serum IgG antibody titer profiles are expressed as highest reciprocal measurable dilution in ELISA, using peptide A-Clq, native Clq, oxidized Clq, type II collagen and peptide K as antigen.

Modulation of Arthritis by Clq . 113

nat. Only 3 of the 5 sera had IgG that bound to C1q ox. Mice pretreated with peptide A-C1q (group III) produced high to very high levels of ClIbinding IgG. All 5 mice from this group produced significant to high IgG titers to peptide A-C1q and also to C1q nat. Low titers of C1q ox-specific IgG were found in sera from 4 of the 5 mice, but IgG Abs reacting with peptide K were observed only in one individual. Sera from the group IV animals produced IgG Abs reacting with all 5 of the antigens tested. The titers being high against ClI, significant against peptide A-C1q, C1q nat and peptide K, but only weak against C1q ox. Differentiation of IgG subclasses, with respect to peptide A-C1q and ClI, is summarized in Table 4. Within the sensitivity of the assay system available to us (see Materials and Methods), no IgG subclasses could be identified that bound to peptide A-C1q in groups I or II (total IgG reacting w-ith peptide A-C1q was of minimal titer, 1:20, see Table 3). However, group I animals exhibited high IgG 1 levels, slightly lower IgG2a levels, fairly minimal IgG2b and only very nominal IgG3leveis of ClI-specific Ab. Table 4. IgG subclasses in serum recognizing peptide A-Clq and type II collagen in male DBAIl mice at day 48, pretreated either with PBS, CII, peptide A-Clq or peptide K i.v. (day 0), followed (day 7) by i.d. administration of Cll"" Male DBA!1 mice No.

Peptide A-Clq

Collagen II

IgGI

IgGI

IgG2a IgG2b IgG3

+ ++++ +++ ++++ n.d.

+ ++ + +++ n.d.

+ +

+ +

++ n.d.

n.d.

+ ++ ++ n.d. +++

+ ++ ++ n.d.

+ + + n.d.

+

++ + + + ++

++++ ++++ + +++ ++

++ ++

++ ++

+

+ +

+

+ +

+++ +++ ++

+ + +

+ + +

IgG2a IgG2b IgG3

2 3 4 5 6 7 8 9

10 12 13 14 15

++ + + + +

16 17 18

+ + ++

11

+

n.d. +

Group I Day 0: PBS i.v. Day 7: CII i.d. Group II Day 0: ClI i.v. Day 7: CII i.d. Group III Day 0: Pep AClq i.v. Day 7: CII i.d. Group IV Day 0: Pep K i.v. Day 7: CII i.d.

g'Serum IgG subclasses are expressed semiquantitatively, using an ELISA system. Values represent OD after subtraction of background (lgG 1: mean OD 0.2; IgG2a or IgG2b: mean OD 0.1; IgG3: mean OD 0.05). n. d.: not determined. + = OD 0.5 Absorption 405 nm; ++ = OD 1.0; +++ = OD 1.5; ++++ = OD 2.0

114 . M.

J. MAEURER, P.

K. E. TRINDER, S. STORKEL, and M. Loos

Table 5. Detection of serum IgE antibodies recognizing peptide A-Clq, native Clq, Clq ox, type II collagen and peptide K in male DBA/l mice at day 48, pretreated either with PBS, ClI, peptide A-Clq or peptide K i.v. (day 0), followed (day 7) by i.d. administration of Cll"" Male DBA/I mice No. 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18

Pep A-Clq Clq nat

+ +

n.d. +

n.d.

n.d. +

++ ++ + ++ n.d.

n.d. ++ ++ n.d. ++ ++ ++ ++ n.d. ++ + + +

Clq ox

CII

Group I Day 0: PBS i.v. Day 7: CII i.d.

n.d.

+++ ++++ ++++ ++++ n. d. + ++++ ++ n.d. +++

Group II Day 0: ClI i.v. Day 7: ClI i.d.

++++ ++++ +++ n.d. +++

Group III Day 0: Pep A-Clq i.v. Day 7: ClI i.d.

++++ +++ ++++

Group IV Day 0: Pep K i.v. Day 7: CII i.d.

n.d.

n.d.

n.d.

0' Serum IgE antibody titer is expressed semiquantitatively, using an ELISA system. Values represent OD after subtraction of background (mean OD 0.18). + = OD 0.5 Absorption 405 nm; ++ = OD 1.0; +++ = OD 1.5; ++++ = OD 2.0

This distribution of subclasses also held true for Abs to CII from group II animals. IgG Abs from groups III and IV, reacting with peptide A-Clq, were found to be almost exclusively of the IgG 1 and IgG3 subclasses. IgG Abs from groups III and IV, reacting with ClI, were mainly of the IgG 1 subclass, together with some IgG2a and IgG2b. No significant IgG3 activity was detected. Serum IgE Abs profiles - with respect to binding of ClI, Clq nat, Clq ox and peptide A-Clq - are summarized in Table 5. Although IgE activity varied with respect to the different antigens tested, no major differences between the 4 groups of animals could be observed. The highest IgE reactivity was to ClI, followed by Clq nat. Only very low level (4 animals out of 15 tested) or no activity to peptide A-Clq was observed, whilst no IgE capable of binding to Clq ox could be detected. Discussion Several studies have shown that type II c911agen-induced arthritis can be suppressed and/or delayed by i.v. pretreatment with native CII, CII-

Modulation of Arthritis by Clq . 115

peptides, mAbs recognizing putative arthritogenic epitopes of CII or CD4+ T cells from diseased animals (10-14, 16, 17). We therefore immunized DBA/l mice i.v. with a nonapeptide from the Clq A-chain (residues 26-34 from N-terminus) that exhibits epitope characteristics and high homology to residues 137-147 from the N-terminus of a cyanogen bromide splitproduct (CB 11) of chicken CII (19), known to playa role in the suppression of arthritic lesions. We were able to demonstrate that i.v. application of peptide A-Clq in soluble form is alone capable of inducing an immune response (generation of IgM and IgG3 serum antibodies) in arthritis susceptible DBA/1 mice (H2q) (see Table 1). It is currently accepted that synthetic peptides are not able to induce a polyclonal immune response, and furthermore, it is welldocumented that short peptides, eliciting B cell responsiveness, contain CD4+ T cell and B cell epitopes (for review see 26). Thus production of Abs to the nonapeptide of the C1q A-chain may be a result of the interaction of T and B cells specific for epitopes located within the peptide. However, the involvement of CD4+ T cells in Ab-responses to this peptide needs to be further investigated. It has been shown that T cells recognize not native antigen, but immunogenic peptides displayed in association with MHC class II molecules by antigen presenting cells (27). The usual approach for screening for B cell recognition epitopes involves generation of a panel of candidate peptides, their immunization into a suitable animal model and screening of the resultant sera for Abs specific for the immunizing peptide and the native protein. Such epitopes have been demonstrated to be either sequential or conformational. Therefore, potentially three distinct groups of peptide Abs could be generated (for review see 28): (i) anti-peptide Ab reacting with the peptide, but not with the native protein; (ii) anti-peptide Ab reacting both with the peptide and with the native protein (not comparable with Abs raised by immunization with the native protein); (iii) anti-peptide Ab reacting with both the peptide and the native protein that additionally interferes with Abs raised by immunization with native antigen. The IgG Ab-response in DB All mice, following peptide A-Clq application, would appear to fit in group (i) as the Abs fail to react with the native protein, despite reacting with the immunizing peptide (peptide AC1q). Intravenous application of native CII or peptide A-Clq in DBA/l mice, results in generation of peptide A-Clq as well as peptide K specific IgM Abs (peptide K: residues 137-147, + additional glycine, from CB11 splitproduct of chicken CII). This presumably reflects the presence of common amino acid sequence(s) (K-G-E-Q-G) in these peptides (Fig. 1 and Table 1). No reactivity to native CII, or to native C1q, was observed, but reactivity to an altered, oxidized form of Clq (C1q ox) could be demonstrated (Table 1). This suggests that oxidation processes within the C1q molecule may bring about conformational changes and/or generation of (hidden) neoepitopes, which is known to be the case in binding of C1q to the Fc

116 .

M.

J. MAEURER, P.

K. E.

TRINDER, S. STORKEL,

and

M.

Loos

portion of IgG in immune complexes (29, 30), or, following limited proteolysis of the collagen-like C1q molecule (31, 32). The fact that IgM titers from peptide A-C1q immunized animals (group III) are lower on day 48 than on day 7, and total IgG titers from the same animals are higher on day 48 than on day 7, may be a reflection of antigendriven B cell differentiation leading to antibody switching. Whether i.d. application of CII (following i.v. peptide A-C1q administration) affects this peptide specific Ab generation (via presentation of B or T cell epitopes of CII) is still under investigation. All IgG antisera from mice injected with peptide A-C1q are able to bind, in addition to peptide A-C1q, peptide K as well as the common five amino acid stretch (K-G-E-Q-G), shared by these peptides (Fig. 1), suggesting existence of a common B cell recognition site. As immunogenicity of a synthetic peptide and subsequent modulation of immune responses is dependent on a combination of T and B cell recognition, we have yet to fully characterize the humoral responses and T cell activation processes associated with peptide A-C1q recognition. Since i.v. application of peptide A-C1q is able to evoke an immune response that leads to intervention of the typical inflammatory processes associated with CII -induced arthritis (i.e. invasion of PMN into synovial tissue and the joint space; see Fig. 3), we predict that peptide A-C1q interacts with the trimolecular complex of MHC, T cells and immunogenic peptides modulating autoimmune respon7 ses to autologous CII-like structures. Possible mechanisms resulting from i.v. administration of peptide A-C1q may include: (i) MHC-blockade (33); (ii) induction of tolerance; (iii) antigen specific suppression (34) (via induction of suppressor T cells); (iv) clonal dominance (35) (leading to overstimulation of an immune response to an immunogenic competitor peptide). Several groups have shown that immunointervention with peptides can inhibit experimentally induced allergic encephalomyelitis (EAE), a demyelinating disease of rodents in which T cells reacting with self-antigens, such as myelin basic protein, play an important role. Synthetic peptides, some of which are substituted immunogenic analogues of encephalitogenic determinants, have been characterized that inhibit the development of EAE (36-38). It has also been shown that, binding to the appropriate MHC class II molecule is necessary for inhibition of in vivo responses in this model. Various other experimental systems have been used to demonstrate nonresponsiveness to an antigen, transferable by T cells, induced by i.v. application of a high dose of antigen. Pretreatment (i.v.) of animals abrogates the development of autoimmunity induced by administration of the antigen in adjuvant. Specific tolerance to this challenge can be transferred by CD4+ splenic T cells (39). In the CII-induced arthritis model, it is possible to confer protection by transfer of CD4+ T cells (17), thus i.v. application of peptide A-C1q may induce effector mechanisms that involve the interaction of dinstinct T cell subsets. To date, two subsets of mouse CD4 + T cells shave been shown to exist, identified by their unique cytokine

Modulation of Arthritis by C1q . 117

profiles. T H1 cells secrete, amongst other cytokines, IL-2, IFN -y, TNF-a and TNF-~, whilst T H2 cells secrete IL-4, IL-5, IL-6, IL-10 and only low levels of TNF-a. T H2 cells are believed to activate resting B cells, whereas T H1 cells appear important in the induction of delayed-type hypersensitivity (DTH) reactions (40-42). Switching of Ab-isotype production to IgE is dependent on IL-4, secreted by activated T H2 cells. In the study presented here, we report that i.d. application of CII leads to the development of IgE serum Abs that bind both CII and native human C1q (Table 5). This provides confirmation of the recent report by MARCELLETTI et al. (43), that CII -specific IgE-Ab levels increase with the onset of histological signs of CII-induced arthritis in DBA/1 mice, thus promoting CII -specific IgE-mediated mast cell degranulation. MARCELLETTI et al. administered CII in the presence of complete Freund's adjuvant (CF A), which is itself known to evoke DTH responses involving IFN-y. CFA alone appears to stimulate the TH1, rather than the TH2, subset (as evidenced by a lack of IgE induction), i.e. CII application in the presence of CF A should induce stimulation of several distinct T cell subsets. In our studies, CII was emulsified in incomplete Freund's adjuvant, thus providing evidence that i.d. application of CII itself leads to activation of a distinct subset of T cells. To date, it is unclear whether CII-specific serum IgE (plasma half-life 2-5 days) is involved in the induction of CII -induced inflammatory lesions, or whether it may act as a marker for IL-4, known to inhibit IFN-y production. IFN-y stimulates production of complementfixing IgG2a Abs, known to be directly involved in the induction of CIIinduced lesions (9). Since C1q is present in serum, further studies may reveal the significance of the existence of IgE recognizing the collagen-like C1q molecule in animals exposed to CII. Bridging of the bivalent IgE Ab via its specific antigen may lead to distinct effector responses mediated by high-affinity IgE receptors (FcERI) on mast cells and basophils or - in the case of elevated serum concentrations of IgE - by low-affinity (FcERII) receptors on B cells and macrophages (44). The presence of IL-4, leading to IgE as well as to IgG 1 subclass secretion (45), may contribute to the observation that by day 48, animals pretreated i.v. with peptide A-Clq and then exposed i.d. to CII produced peptide A-Clq-binding IgG 1 as well as IgM and IgG3 Abs, whereas by day 7 (before i.d. CII exposure) only IgM and IgG3 Abs could be detected. Taken together, these results lead us to postulate that the collagen-like Clq molecule - containing the amino acid sequence K-G-E-Q-G in the collagenous part of the A-chain - is involved in the concert of autoimmune responses to type II collagen-like structures responsible for modulation of arthritic inflammatory lesions. CII is believed to contain both arthritis inducing and suppressing determinants (18, 19) whereas only one determinant appears to be present in peptide A-C1q, as evidenced by sequence homology to chicken and human CII (Fig. 1). The physiological relevance may become clear in those situations where the Clq molecule is altered by

118 . M. J. MAEURER, P. K. E. TRINDER, S. STaRKEL, and M. Loos

oxidation and/or by proteolytic processes in inflammatory lesions, and where autoantibodies bind to the collagen-like portion of the Clq molecule (46) leading to interference of autoimmune responses against collagen-like structures in individuals with a distinct genetic background. Acknowledgements We wish to thank GABRIELE WINTERNHEIMER for her excellent technical support, and MONICA O'MALLEY for help in preparation of the tables.

References 1. TARKOWSKI, A., L. KLARESKOG, H. CARLSTEN, P. HERBERTS, and W. J. KOOPMAN. 1989. Secretion of antibodies to types I and II collagen by synovial tissue cells in patients with rheumatoid arthritis. Arthritis Rheum. 32: 1092. 2. MORGAN, K., R. B. CLAQUE, I. COLLINS, S. AYAD, S. D. PHINN, and L. P. J. LENNOX HOLT. 1989. A longitudinal study of anticollagen antibodies in patients with rheumatoid arthritis. Arthritis Rheum. 32: 139. 3. TRENTHAM, D. E., A. S. TOWNES, and A. H. KANG. 1977. Autoimmunity to type II collagen: An experimental model of arthritis. J. Exp. Med. 146: 857. 4. COURTENAY, J. S., M. J. DALLMAN, A. D. DAYAN, A. MARTIN, and B. MOSEDALE. 1980. Immunization against heterologous type II collagen induces arthritis in mice. Nature 283: 666. 5. Yoo, T. J., S. Y. KIM, and J. M. STUART. 1988. Induction of arthritis in monkeys by immunization with type II collagen. J. Exp. Med. 168: 777. 6. DESIMONE, D. P., D. B. PARSONS, K. B. JOHNSON, and R. P. JACOBS. 1983. Type II collagen-induced arthritis. A morphologic and biochemical study of articular cartilage. Arthritis Rheum. 26: 1245. 7. TRENTHAM, D. E. 1982. Collagen arthritis as a relevant model for rheumatoid arthritis. Arthritis Rheum. 25: 911. 8. KLARESKOG, L., R. HOLMDAHL, E. LARSSON, and H. WIGZELL. 1983. Role of T lymphocytes in collagen II induced arthritis in rats. Clin. Exp. Immunol. 51: 117. 9. WATSON, W. C. and A. S. TOWNES. 1985. Genetic susceptibility to murine collagen II autoimmune arthritis. Proposed relationships to the IgG-2 autoantibody subclass response, complement C5, major histocompatibility complex (MHC) and non-MHC loci. J. Exp. Med. 162: 1878. 10. WOOLEY, P. H., H. S. LUTHRA, C. J. KRCO,J. M. STUART, and C. S. DAVID. 1984. Type II collagen-induced arthritis in mice. II. Passive transfer and suppression by intravenous injection of anti-type II collagen antibody or free native type II collagen. Arthritis Rheum. 27: 1010. 11. STAINES, N. A., T. HARDINGHAM, M. SMITH, and B. HENDERSON. 1981. Collagen-induced arthritis in the rat: modification of immune and arthritis responses by free collagen and immune anti-collagen antiserum. Immunology 44: 737. 12. SCHOEN, R. T., M. I. GREEN, and D. E. TRENTHAM. 1982. Antigen-specific type II collagen-induced arthritis by collagen coupled spleen cells. J. Immunol. 128: 717. 13. ENGLERT, M. E., M. J. LANDES, A. L. ORONSKY, and S. S. KERWAR. 1984. Suppression of type II collagen-induced arthritis by the intravenous administration of type II collagen or its constituent peptide a-1(II) CB11. Cell. Immunol. 87: 357. 14. KOBAYASHI, S., K. TERATO, Y. HARADA, H. MORIYA, and M. TANIGUCHI. 1991. Suppression of type II collagen-induced arthritis by monoclonal antibodies. Arthritis Rheum. 34: 48. 15. NAGLER-ANDERSON, c., L. A. BOBER, M. E. ROBINSON, G. W. SISKIND, and G. J. THORBECKE. 1986. Suppression of type II collagen-induced arthritis by intragastric administration of soluble type II collagen. Proc. Nat!. Acad. Sci. USA 83: 7443.

Modulation of Arthritis by Clq . 119 16. KRESINA, T. F. and R. W. MOSKOWITZ. 1985. Adoptive transfer of suppression of arthritis in the mouse model of collagen-induced arthritis: evidence for a type II collagen-specific suppressor T cell. J. Clin. Invest. 75: 1990. 17. MEYERS, L. K., J. M. STUART, and A. H. KANG. 1989. A CD4 cell is capable of transferring suppression of collagen-induced arthritis. J. Immunol. 143: 3976. 18. TERATO, K., K. A. HASTY, M. A. CREMER,J. M. STUART, A. S. TOWNES, and A. H. KANG. 1985. Collagen-induced arthritis in mice. Localization of an arthritogenic determinant to a fragment of the type II collagen. J. Exp. Med. 162: 627. 19. MYERS, L. K., 1. M. STUART, 1. M. SEYER, and A. H. KANG. 1989. Identification of an immuno-suppressive epitope of type II collagen that confers protection against collageninduced arthritis. J. Exp. Med. 170: 1999. 20. MORGAN, K., R. B. CLAQUE, M. J. SHAW, S. A. FIRTH, T. M. TwosE, and L. P. J. LENNOX HOLT. 1981. Native type II collagen induced arthritis in the rat. The effect of complement depletion by cobra venom factor. Arthritis Rheum. 24: 1356. 21. STEMMER, F. and M. Loos. 1984. Purification and characterization of human, guinea pig and mouse Clq by fast protein liquid chromatography (FPLC). J. Immunol. Meth. 74: 9. 22. LAEMMLT, U. K. 1970. Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature 227: 680. 23. LOWRY, O. H., N. J. ROSEllROUGH, A. L. FARR, and R. J. RANDALL. 1951. Protein measurement with the folin phenol reagent. J. BioI. Chern. 139: 265. 24. SELLAR, G. L., D. J. BLAKE, and K. B. M. REID. 1991. Characterization and organization of the genes encoding the A, Band C chains of human Clq. The complete derived amino acid sequence of human Clq. Biochem. J. 274: 481. 25. MYERS, E. W. and W. MILLER. 1988. CABIOS (computer application in the biosciences) 14: 11. 26. FAYOLLE, c., E. DERIAUD, and C. LECLERC. 1991. In vivo induction of cytotoxic T cell response by a free synthetic peptide requires CD4+ T cell help. J. Immunol. 147: 4069. 27. UNANUE, E. R. 1984. Antigen-presenting function of the macrophage. Ann. Rev. Immunol. 2: 395. 28. MILICH, D. R. 1990. Synthetic peptides: prospects for vaccine development. Sem. Immunol. 2: 307. 29. GOLAN, M. D., R. BURGER, and M. Loos. 1982. Conformational changes in Clq after binding to immune complexes: detection of neoantigens with monoclonal antibodies. J. Immunol. 26: 163. 30. Loos, M. 1985. The complement system: activation and control. Curro Top. Microbiol. Immunol. 121: 7. 31. HEINZ, H.-P., D. BRACKERTZ, and M. Loos. 1988. Enzymatic alteration of Clq, the collagen-like subcomponent of the first component of complement, leads to crossreactivity with type II collagen. FEBS Lett. 228: 332. 32. HEINZ, H.-P., K. RUBIN, A.-B. LAURFLL, and M. Lo05. 1989. Common epitopes in Clq and collagen type II. Molec. Immunol. 26: 163. 33. ADORINI, L., A. VALLI, and A. GUERY. 1991. Inhibition of T cell activation by blockade of MHC class II molecules. Sem. Immunol. 3: 231. 34. SECARZ, E. and U. KRZYCK. 1991. The distinctive specificity of antigen-specific suppressor T cells. Immunol. Today 12: 111. 35. LAMONT, A. G., M. G. POWELL, S. M. COLON, C. MILES, H. M. GREY, and S. SETTE. 1990. The use of peptide analogues with improved stability and MHC binding capacity to inhibit antigen presentation in virro and in vivo. J. Immunol. 144: 2493. 36. WRAITH, D. c., D. E. SMILEK, D. J. MITCHELL, L. STEINMAN, and H. O. McDEVITT. 1989. Antigen recognition in autoimmune encephalomyelitis and the potential for peptidemediated immunotherapy. Cell. 59: 247. 37. STEINMAN, L. 1989. Prevention of experimental encephalomyelitis with pep tides that block interaction of T cells with major histocompatibility complex proteins. Proc. Natl. Acad. Sci. USA 86: 9470.

120 . M. J. MAEURER, P. K. E. TRINDER, S. STaRKEL, and M. Loos 38. LAMONT, A. G., A. SETTE, R. FU]INAMI, S. M. COLON, C. MILES, and H. M. GREY. 1990. Inhibition of experimental autoimmune encephalomyelitis induction in SLJIJ mice using a peptide with high affinity for IN molecules. J. Immunol. 145: 1687. 39. PARISH, N. M., I. M. ROITT, and A. COOKE. 1988. Phenotypic characteristics of cells involved in induced suppression to murine experimental autoimmune thyroiditis. Eur. J. Immunol. 18: 1463. 40. MOSMANN, T. R. and K. W. MOORE. 1991. The role ofIL-10 in cross regulation of Th1 and Th2 responses. Immunoparasitol. 7: A49. 41. VERCELLI, D. H., H. JABARA, R. P. LAUENER, and R. S. GEHA. 1990. IL-4 inhibits the synthesis of IFN-gamma and induces the synthesis of IgE in human mixed lymphocyte cultures. J. Immunol. 144: 570. 42. SNAPPER, C. M. and W. E. PAUL. 1987. Interferon gamma and B cell stimulatory factor-1 reciprocally regulate Ig isotype production. Science 236: 944. 43. MARCELLETTI, J. F., J. I. OHARA, and D. H. KATZ. 1991. Collagen-induced arthritis in mice. Relationship of collagen-specific and total IgE synthesis to disease. J. Immunol. 147: 4185. 44. METZGER, H., J.-P. KINET, H. BLANK, L. MILLER, and C. RA. 1989. The receptor with high affinity for IgE, in: D. CHADWICK, D. EVERED, and J. WHELAN (eds.), IgE, mast cells and the allergic responses. Ciba Foundation Symposium 147: 93. 45. FINKELMAN, F. D., J. HOLMES, I. M. KATONIA, J. F. URBAN, M. P. BECKMANN, L. S. PARK, K. A. SCHOOLEY, R. L. COFfMAN, T. R. MOSMANN, and W. E. PAUL. 1990. Lymphokine control of in vivo immunoglobulin isotype selection. Ann. Rev. Immunol. 8: 303. 46. ANTES, U., H.-P. HEINZ, and M. Loos. 1988. Evidence for the presence of autoantibodies to the collagen-like portion of C1q in systemic Lupus erythematosus. Arthritis Rheum. 31: 457.

Dr. MICHAEL Loos, Institute of Medical Microbiology, Johannes Gutenberg University, Augustusplatz/Hochhaus, 6500 Mainz, Germany