Journal of Autoimmunity (1991) 4,341-356

Comparison of Natural Antibodies to Autoantibodies Arising During Lupus in (NZB x NZW)Fl Mice

Basma Hentati, The&se Ternynck, Stratis Avrameas and BCatrice Payelle-Brogard Uniti d’lmmunocytochimie,

URA 359 du CNRS,

Institut Pasteur, Paris, France

(Received 253~1~ 1990 and accepted 15 October 1990)

Autoantibodies arising in (NZB x NZW)F, (B/W) mice during the lupuslike syndrome were studied and compared to natural antibodies present in normal mice. The antibody activities were tested in sera, circulating immune complexes (CIC) and kidney eluates, using an enzyme immunoassay against a panel of self and non-self antigens: actin, myosln, tubulln, DNA, myoglobln, spectrin and trinitrophenylated bovine serum albumin (TNP/BSA). In the B/W mouse sera, IgM antibodies reacting with all the panel of antigens (PAg) and comparable to those of normal mice, increased moderately from 5 to 9 months and markedly during the last stage preceding death (10 months), when particularly high levels of anti-DNA, anti-tubulin and anti-myoglobin antibodies were noted. Polyreactlve IgM antibodies present in CIC were moderately increased while those present in complexes deposited in kidneys were strongly enhanced after the 8th month. IgG antibodies showed an early increase (2 months) in B/W sera for anti-TNP activity, which remained more or less constant until death, while a later (5-6 months) and greater increase of activity, mainly directed against DNA but also against the other antigens of the panel, was observed. In CIC, IgG, mainly anti-DNA but also and-TNP, were enhanced at the end of the disease while at the same time IgG reacting with all the PAg were found in kidney deposits. Isolation of antibodies from sera on a DNAimmunoadsorbent demonstrated that eluted IgM reacted with all the PAg but mainly with DNA, while IgG reactivity was more restricted to DNA and to a lesser degree to TNP. The D23 idlotype, characteristic of natural polyspecific antibodies, was expressed on IgM and IgG autoantibodies from B/W mice and was enhanced, particularly in kidneys, at the end of the disease.

Correspondence to: Basma Hentati, UnitC d’Immunocytochimie, Institut Pasteur, 25 rue du Dr Roux, 75724 Paris Cedex 15, France.

Dipartement

d’Immunologie,

341 0896-841 l/91/020341

+ 16 $03.00/O

0 1991 Academic Press Limited

342

Basma Hentati et al. These results demonstrate that natural antibodies are a part of the population of increased autoantibodies in this disease and could participate with IgG anti-DNA antibodies in lupus.

Introduction

Systemic lupus erythematosus (SLE) is an autoimmune disease characterized by the production of autoantibodies directed against various antigens, in particular antibodies to single- and double-stranded DNA [ 1,2]. These antibodies form immune complexes which are deposited in various tissues and especially in the kidneys. Experiments performed with monoclonal antibodies from autoimmune mice suffering from a lupus-like syndrome or humans with SLE showed that a large proportion of the anti-DNA antibodies was polyreactive and recognized determinants such as phospholipids [3], polysaccharides [4], cytoskeletal proteins [5] and cell membrane proteins [6]. Polyreactivity seems to be the main characteristic of the natural antibodies present in the sera of normal humans and other animal species [7] that most likely constitute the antibody repertoire of the normal immune system. Several studies suggested that autoantibodies present in SLE are part of the repertoire of the normal immune system [8] and that somatic mutation generates the pathologic autoantibodies [9]. In order to study the relationships between the natural antibodies and the autoantibodies implicated in the pathology of SLE, using a panel of self and non-self antigens, we analysed, at different stages of the disease, the autoantibodies present in B/W mouse serum and kidneys. We found that, compared to normal mice, the polyreactive IgM antibodies were only moderately increased during the disease; however, they were abruptly enhanced during its terminal phase, at approximately 10 months, a few days before the death of the animals. IgG began to increase earlier than IgM but presented the same peak of enhanced activity at the end of the disease. The IgG antibodies reacted with all the antigens tested, but some seemed to be more specific to DNA. Expression of the D23 idiotype, described by Lymberi et al. [lo] as being characteristic of natural antibodies, was enhanced on both these antibody populations during disease, particularly in the kidneys. These results show that polyreactive IgM and IgG antibodies, similar to natural antibodies present in normal mice, increase progressively during lupus. Furthermore they indicate that at least some of the autoantibodies present at the final phase of the disease, which react strongly with DNA and share the D23 idiotypic determinants, possess characteristics similar to those of natural antibodies and may play a pathogenic role in the lupus-like syndrome. Materials

and methods

Mice

Female NZB mice were obtained from Institut Pasteur’s animal colony and male NZW mice from the Centre de Selection et d’Elevage des Animaux de Laboratoire

Autoantibodies during murine lupus (CNRS, facilities study.

343

Orleans, France). (NZB x NZW)F, mice (B/W) were bred in the animal at Institut Pasteur. Female B/W mice, 1 to 10 months were used for this

Antigens and antibodies Actin and myosin from BALB/c mouse muscle were prepared, according to the methods of Spudish and Watt [ 1 l] and Whalen et al. [ 121, respectively. Mouse brain tubulin was prepared using the method of Shelansky et al. [ 131 and mouse red blood cell spectrin was purified following the protocol of Ungewickell and Gratzer [ 141. Native dsDNA and sperm whale myoglobin were purchased from Sigma Chemical Co., St Louis, MO, USA. Trinitrophenylated bovine serum albumin (TNP/BSA) was prepared according to the procedure of Little and Eisen [ 151. Rabbit antibodies to mouse IgM or IgG, specific to the l.t or y heavy chain, were prepared using, respectively, IgG or IgM immunoadsorbents [16]. The purified antibodies were coupled to E. coli 8-galactosidase (Sigma) with glutaraldehyde in a one-step procedure [ 171. Enzyme immunoassays (EIA) Analysis of antibody activity in the sera and other preparations EIA were performed as reported elsewhere [18]. Briefly, 96-well polystyrene plates (CML, Nemours, France) were coated with actin, myosin, tubulin, myoglobin, TNP/BSA (5 pg/ml), spectrin (2 pg/ml) or DNA (10 pg/ml) in 0.1 M carbonatebicarbonate buffer (pH 9.5) for 2 h at 37°C and kept at 4°C until used. Before the assay, the antigen-coated plates were incubated for 30 min at 37°C with OS:/, gelatin in PBS and then washed three times with PBS containing 0.19, Tween 20 (PBSTween). Serum samples (final dilution l/100, or l/500 when tested against TNP/ BSA), purified immune complexes (IC) (final dilution l/20), eluates from DNA or TNP immunoadsorbents (final dilution l/30) or from kidneys (final dilution l/10) in PBS-Tween containing 0.5% gelatin were incubated on the antigen-coated plates for 2 h at 37°C. The plates were then washed three times with PBS containing O.l’?() Tween 20 (PBS-Tween) and incubated with 8-galactosidase-labelled anti+ or anti-y conjugates at 1 pg/ml. After appropriate washes, 100 l.tl of o-nitrophenyl+Dgalactopyranoside (ONPG, Sigma) at 0.8 mg/ml in 0.1 M phosphate buffer, pH 7, containing 0.00569& (v/v) 8-mercaptoethanol were added to the wells. The optical densities (OD) were read at 414 nm using an automatic reader (D ynatech, Marnes-laCoquette, France). Detection of circulating immune complexes (CIC) CIC in serum were detected by an EIA using immobilized anti-c,, adapting a procedure described for radioimmunoassay [21]. Briefly, 10 ~1 of sheep anti-mouse C, F(ab’), antibody fragments, prepared by pepsin digestion of sheep anti-mouse C, globulins (ICN, UK), were added at 40 pg/ml in borate buffer saline (10 mM borate, 150 mM NaCl, pH 7.4), to the wells of polystyrene plates (CML, Nemours, France) and incubated overnight at 4°C. The plates were washed three times with PBS and unreacted sites were blocked by incubation with l”/b BSA in PBS at 37°C for 1 h.

344

Basma Heatati et al.

Wells were then washed and filled with 100 ul of serum, diluted 1j20 in PBS containing 1 o/0 BSA and 10 mu EDTA, and incubated for 2 h at 37°C. After washing, the plates were incubated for 2 hat 37°C with P-galactosidase-labelled sheep anti-mouse Ig (1 l.tg/ml). The P-galactosidase activity was revealed by ONPG as described above. Analysis of the presence of idiotype (Id) 023 Polystyrene plates were coated with optimal concentrations of anti-D23 antibodies. Serum samples, IC and eluates were then incubated, following the procedure described above, and the Id-anti-Id binding was revealed using j3_galactosidaselabelled anti-p or anti-y conjugates. Immune complexes CIC were isolated from B/W or control BALB/c sera by previously published methods [19]. Briefly, 400 l.d of 5% polyethylene glycol (PEG) (MW 6,000) (BDH, UK) in 150 mM NaCl, 20 mM potassium phosphate buffer, pH 7.4, were added to an equal volume of serum diluted l/l with phosphate-buffered saline (PBS), pH 7.4, and the mixture was incubated overnight at 4°C. The precipitate was isolated by centrifugation (1,500 x g, 20 min, 4”C), washed twice with 2.5O, PEG and dissolved by incubation for 30 min at 37°C in 200 ~1 of 150 mM phosphate buffer, pH 7.5, 10 mM ethylene diamine tetraacetic acid (EDTA), containing 0.05”” Tween 20 and 0.1 yi, merthiolate. In order to dissociate the complexes [20], 50 ~1 of cold 0.2 M HCl-glycine, pH 2.8, were added to 100 ~1 of the preparation, brought to 4°C and the reaction mixture was incubated at 4°C for 15 min. The solution was then neutralized with 25 l.d of 1 M K,HPO,, diluted l/20 by adding 1,825 ~1 of the diluent buffer (1” 0 BSA in PBS) and tested by ETA, as described above, within 30 min. Zdiotype 023 D23, a murine natural polyreactive monoclonal IgM antibody, was derived from the splenocytes of 12-week-old unprimed BALB/c mice [ 181. Anti-Id antibodies were produced in a rabbit against D23 and the antiserum obtained was rendered specific to the idiotopes of D23 according to Lymberi et al. [lo]. The anti-Id D23 antibodies were characterized as reacting with mouse polyspecific IgM and with IgG2b. The presence of D23 Id in normal BALB/c mouse serum correlated with the presence of the natural antibody activity [lo]. Isolation of antibodies Afinity purification of antibodies on DNA-immunoadsorbent Ten milligrams of dsDNA was coupled to 5 ml of glutaraldehyde-activated ACA 3.4, Ultrogel [16]. Anti-DNA antibodies were purified on this immobilized DNA according to the method of Kubota et al. [22]. Briefly, the immunoadsorbent (1 ml of beads in an Eppendorf tube) was first equilibrated at room temperature with Trisbuffered saline (25 mM Tris, 140 mM NaCl, pH 7.4; TBS). One millilitre of l/5 diluted serum or l/2 diluted kidney eluate in TBS containing 5 mu EDTA (EDTATBS) was incubated with the immunoadsorbent for 1 h at room temperature with gentle agitation. The immunoadsorbent was then washed extensively with EDTA-

Autoantibodies

during murine lupus

345

TBS and bound antibodies were eluted with a 20 mM sodium carbonate buffer, pH 10.5, containing 5% v/v dimethyl sulfoxide. The eluted antibodies were promptly mixed with an equal volume of 1 M Tris-HCl (pH 7), dialysed against TBS, and the IgM and IgG contents were determined by EIA. Afinity purij%ation of antibodies on TNP-immunoadsorbent TNP was immobilized on a Sepharose column equilibrated with 0.2 M boric acidNaCl buffer, pH 8, as previously described [23]. One millilitre of l/5 diluted serum or l/2 diluted kidney eluate in PBS was incubated with 1 ml of the immunoabsorbent for 1 h at room temperature with gentle agitation. The column was then washed with the borate buffer until the absorbance at 280 nm was less than 0.05; the antibodies adsorbed were eluted with 0.1 M DNP-glycine (Sigma) in borate buffer, pH 8.6. All yellow fractions were pooled, concentrated and dialysed overnight at 4°C against the above buffer, until the yellow colour disappeared. Elution of antibodies from kidneys Antibodies were eluted from kidneys of B/W or normal BALB/c mice according to the method of Saxena et al. [24]. Briefly, kidney cortices (1 g) from two mice were homogenized in Tris buffer (0.05 M Tris-HCl, 0.15 M NaCl, pH 7.5) containing 1 mM phenylmethanesulfonyl fluoride, 1 mM tosylleucylchloromethylketone as proteinase inhibitors. The homogenates were washed and centrifuged six times until Ig were no longer detectable in the supernatants tested by EIA. The insoluble material was incubated for 10 min successively with cold 0.1 M glycine, 0.5 M NaCl buffer pH 2.8 and pH 2.2. After centrifugation, the supernatant pH was adjusted to 7.4 with 1 M Tris, dialysed overnight against Tris buffer, pH 7.5, and completed to 8 ml. The total IgM and IgG contents of the eluates were measured by EIA. Expression of results Sera from ten female B/W mice were tested individually against the panel antigens (PAg) and compared to a pool of sera from BALB/c mice. They were tested monthly from 1 to 10 months of age and expressed as individual percentages (i.e., the ratio of the absorbance from the test sample divided by the absorbance of the normal pool serum x 100) or means of percentages. ‘When results are expressed as means of percentages, they were derived from a typical experiment which had been run three separate times. Antibody activity in IC or eluates from DNA- or TNP-immunoadsorbents, or kidney eluates, against the PAg is expressed as OD because the very low values of the normal control IgG activity did not allow the calculation of percentages. Id D23 results are also expressed as OD. Results Analysis of kinetics of antibody activity in serum from (NZB x NZW)

F, mice

F, mice were tested monthly from 1 to 10 months, for antibody activity against the panel antigens (PAg) actin, myosin, tubulin, DNA, spectrin, myoglobin and TNP/ BSA. Sera from ten mice at each age were tested individually and the mean values obtained were compared to those of the pool of normal BALB/c mouse sera. Results indicate three stages in the evolution of antibody activity (Figure 1).

346

Basma Hentati et 01.

2

4

6

8

‘0~“Ipi-j

2

IO

4

6

6

IO

Spectr1n

c

d

s

1000

d

6 8

L&J 3

100

_

4 6

8

IO

2

4

6

6

IO

2

4

6

6

10

2

4

6

6

IO

Age (months)

Figure 1. Kinetics of antibody activity in B/W mouse sex-a. Sera from 10 mice were tested individually against the PAg at 1: 100 dilution (1500 for TNP). Mean OD values obtained with 10 sera are expressed as percentages of the IgM ( n ) and IgG (0) values obtained with a pool of normal BALB/c mouse sera (loo”, corresponds to the control BALB/c sera). The absence of value corresponded either to an identical or to a decreased value compared to that of the control.

From 1 to 4 months. Compared to the control, no basic changes appeared in IgM and

IgG antibody activities, except for IgG activity to DNA which increased at 4 months, and TNP/BSA which increased abruptly at 2 months and then remained more or less constant until the death of the animal. From 5 to 8 months. The IgM activity was moderately. increased against myosin, tubulin, DNA and TNP/BSA, while a marked increase in the IgG activity was observed for anti-DNA as well as for all the PAg, except anti-tubulin and antispectrin which were only moderately increased. At 9 and 10 months. The last stage preceding the death of the animals was characterized by a strong increase in both IgM and IgG activities; the IgM reacting with DNA, tubulin myoglobin and spectrin and the IgG recognizing all the PAg were considerably increased. Compared to that of the control, the IgG anti-DNA titers, were the most markedly increased (26 times at 9 months and 43 times at 10 months). When mouse sera were considered individually (Figure 2) a relatively homogeneous IgM response was observed, whereas a heterogeneous IgG response to all the PAg was noted after the 5th month. Analysis of serum antibodies isolated on DNA-

or TNP-immunoadsorbent

Pools of sera from 10 B/W or normal BALB/c mice were passed through a DNA- or TNP-immunoadsorbent column and the antibodies bound were eluted using an

Autoantibodies during murine lupus I month

5 months

347

9 months

Figure 2. Kinetics of antibody activity in B/W mouse sera tested individually. Sera from 10 mice at I,5 or 9 months of age were tested individually against the PAg. Each value is expressed as percentages of the IgM ( l) and IgG (0) value obtained with a pool of normal BALB/c mouse sera. Horizontal lines indicate the mean value.

alkaline medium for the DNA-column or DNP-glycine for the TN&column. When antibodies eluted from DNA-immunoadsorbent were tested against the PAg (Figure 3a), the IgM had marked anti-TNP activity starting at the 2nd month and low reactivity with all the other PAg, including DNA, similar to the normal control. IgG present in eluations obtained from normal mouse sera did not interact with any of the PAg, whereas IgG obtained from sera of 2-month-old B/W mice recognized TNP; those obtained from mice older than 6 months were mainly directed against DNA and to a lesser extent against TNP. When antibodies eluted from TNPimmunoadsorbent were tested (Figure 3b), the IgM reacted strongly with TNJ? and weakly with other PAg, both similar to the normal control, except for anti-DNA activity, which increased as of the 6th month. On the other hand, IgG showed marked anti-TNP and anti-DNA reactivities at 6 and 8 months while none could be detected against the other PAg. In the same experimental conditions, it was not possible to retain any antibody activity from kidney eluates on these two immunoadsorbents. It thus appeared, in B/W mouse sera, that among antibodies isolated for their anti-DNA or anti-TNP affinity, those of the IgM isotype recognized the other antigens of the panel, while those of IgG isotype showed more restricted specificity to DNA and TNP. Antibody activity in immune complexes isolatedfrom sera

CIC were isolated from the pooled sera of ten B/W mice, monthly from 1 to 10 months. The amount of IC, compared to that isolated from normal sera,

348

Basma Hentati et al. (0)

z

0

P

Q

No

2

4

No

6

2

4

6

6

IO

(b)

0.6

i-

0.5

0.4 a3 0.2 0.1 0 6

Jd

NO

8

he

4

6

6

h lonths)

Figure 3. Activity of antibodies isolated from B/W and normal control BALB/c mouse (No) sera on DNA-immunoadsorbent (a) or TNP-immunoadsorbent (b). Results are expressed as OD at 414nm. w Actin; 12 myosin; q tubulin; q DNA; El myoglobin; M spectrin; q TNP.

progressively increased from the 2nd to the 5th month and reached a level approximately four times higher than the normal serum value after the 6th month (data not shown). As shown in Figure 4, the antibody activities in IC isolated from normal (No) and B/W sera were compared and, starting at 5 or 6 months, a marked increase in IgM activity for all the PAg was noted with the exception of tubulin and myoglobin. In contrast, IgG activity for all the PAg was very low. Starting at 8 months, however, a strong activity directed not only against DNA but also against TNP, although to a lesser degree, was noted. Analysis of antibodies isolatedfrom kidneys

Immune complexes were eluted from kidneys at 2-month intervals. As shown in Table 1, the amounts of IgM and IgG antibodies eluted from B/W kidneys up to and

Autoantibodies during murhe lupus 349 0.4

Actin

Myosin

Tubulin

0.4 : 0.3 d 5 0.2 n o: 0.1 0

No 2

4

6 8 IO

No 2 4

6

8 IO

No 2

4

6

8

IO

No 2

4

6

8

If

Age (months1

Figure 4. Antibody activity in CIC isolated from B/W and normal control BALB/c (No) mouse sera. Results are expressed as OD at 414 nm for IgM ( n ) and IgG (Cl).

Table 1. Elution of immunoglobulin deposits from the kidneys of B/W mice at different*ages andfrom control BALBjc mice

Origin of kidneys BALB/c B/W 2months 4 months 6 months 8 months 10 months

Ig concentration IgM 2.36’ 2.20 2.13 2.10 4.15 5.32

@g/ml) I& 1.96’ 1.65 1.13 1.27 13.40 12.26

The IgM and IgG contents of kidney eluates were determined by EIA using a sheep anti-mouse u-chain or y-chain antibody labelled with 8-galactosidase. The Ig were suspended in 8 ml of buffer corresponding to the elution of 1 g of kidney cortex. ‘Mean value of IgM or IgG contents of kidney eluates of BALB/c mice at different ages (2 to 10 months old): no change in accordance with age was observed in BALB/c mice.

including 6 months of age were similar to those eluted from normal BALB/c kidneys. An approximately two-fold increase in IgM was observed at 8 and 10 months, while at the same times, IgM had increased 6.4 and 5 times, respectiveiy. No increase was observed when Ig were eluted from kidneys of BALB/c mice 2 to 10 months old. The

350

Basma Hentati et al. 1.4 /- Actin

2

Ill :,d IC1

2

4

6

6

8

IO

1

Myoglobin

8

4

8

IO

Age

2

4

6

8

IO

2

4

6

8

IO

(months)

Figure 5. Kinetics of antibody activity in B/W kidney eluates. Results are expressed as OD at 4 14 m-n for IgM(H)andIgG(Cl).

IgM antibody activity against the PAg increased from the 4th month (Figure 5), with a peak at 8 to 10 months, while the increase in IgG antibody activity against all the PAg was detectable at 6 months and reached maximal values at 8 to 10 months. During this final stage, both IgM and IgG eluted from the kidneys reacted with all the PAg. Although the amount of IgG eluted from kidneys at this stage was apptoximately three times higher than that of IgM (Table I), the IgM activity in OD was more elevated (Figure 5). This could be due to the pentameric structure of the IgM molecule, which binds more antibody conjugates directed to heavy chains than IgG molecules. Antibodies from Bl W mice expressed the 023 idiotypic determinants characteristic of natural antibodies D23 idiotype expression was sought on IgM and IgG serum fractions, eluates from DNA-immunoadsorbent, purified IC and kidney eluates (Figure 6). The D23 idiotype titers associated with the IgM fraction of normal BALB/c mouse serum were similar to those of B/W mouse sera, regardless of the stage of the disease. The D23 associated with the IgG fraction started to increase at 2 months and persisted throughout the experimental period. In the preparations obtained after DNA immunoadsorption, a similar but relatively larger increase in the D23 associated IgG fraction of B/W mice was noted: 3.6 times the control at 8 and 10 months. A less marked increase in the D23 associated with the IgM fraction was found.

Autoantibodies

0)

during murine lupus

351

(b)

1.0I-

0.6I-

0.6

IllII 0.4

0.2 ,

L

*

NO

5I

2

4

6

6

0

10

2

4

6

E

10

r(d)

d

6

No

I .c

0.E

0.6

a4

0.5

C

!ullJ

dl¶k 2

4

6

8

IO

Age (months1

Figure 6. D23 idiotype expression on IgM ( W) and IgG (0) in B/W mouse sera (a), antibodies isolated from DNA-immunoadsorbent (b). CIC (c), and kidney eluates (d). No =normal control BALB/c mice.

In IC isolated from sera, no increase in Id D23 associated with the IgM fraction was noted. In contrast, an increase in the Id associated with the IgG fraction was observed, starting at 4 and peaking at 8 to 10 months. In kidney eluates, Id D23 expression was not detectable in normal BALB/c and young B/W (2 month old) mice, but started increasing sharply at 4 months for IgM and at 6 months for IgG, with peaks at 8 to 10 months for both isotypes.

Discussion

In lupus, the presence of antibodies reacting with various antigens including DNA, cell surface molecules, cytoskeletal proteins and T-lymphocytes is well documented [2,25,26]. Anti-DNA antibodies have long been considered characteristic of lupus [ 11. Although both IgM and IgG anti-DNA antibodies are detected in serum [27, 281, it was found that IgG correlate well with disease activity and the severity of nephritis, whereas IgM seem to be more closely associated with milder clinical states [27, 29, 301. The present study was undertaken in order to examine, during the evolution of the lupus-like syndrome in B/W mice, the relationships between IgM

352 Basma Hentati et al.

and IgG anti-DNA antibodies as well as between these antibodies and natural autoantibodies. Using a panel of self and non-self antigens, we analysed IgM and IgG antibodies present in the sera and deposited in the kidneys of B/W mice at different ages. We also examined the association of these antibodies with the D23 idiotype, previously shown to be present on all murine natural autoantibodies which are polyreactive [lo]. IgM antibodies recognizing all the PAg were found in the sera of B/W mice at all ages. Until 5 months, the level of these antibodies was similar to that found in sera of normal BALB/c mice. They were moderately increased (two to three times) from 5 to 9 months but increased abruptly and markedly at 10 months. At this stage, particularly high levels of anti-DNA, anti-tubulin and anti-myoglobin antibodies were noted. These latter results as well as the kinetics obtained with IgM antibodies do not support the hypothesis of simply polyclonal stimulation of IgM synthesizing cells during the evolution of the lupus-like syndrome in B/W mice. Rather, and more clearly at the final stage of the disease, they seem to favour a more or less specific antigen(s) stimulation, possibly involving injury of given tissue(s). In any case, since similar results were obtained in three different series of experiments, each using 10 different mice, it appears that these IgM kinetics are not chance observations but rather that they reflect significant biological events. These IgM antibodies seem to share similar characteristics with the natural IgM polyspecific autoantibodies found in normal BALB/c mice. They bear the D23 idiotypes carried by all murine polyreactive autoantibodies [lo]. Furthermore, when IgM antibodies were isolated from B/W mouse sera, using a DNA-immunoadsorbent, they were found to be polyspecific and to react mainly with TNP and DNA but also with several other PAg. Passage of B/W mouse sera through a TNPimmunoadsorbent enabled the isolation of polyspecific IgM antibodies recognizing principally TNP and DNA, but also other antigens of the panel. IgM antibodies reacting with all the PAg were present in the CIC isolated from the sera of B/W mice of different ages. Beginning at 5 months, a moderate increase in these antibodies in CIC was noted, becoming more evident at 8 to 10 months. It is interesting to note that, as in the serum, all antibody specificities were not equally increased: anti-actin, anti-DNA, anti-TNP and anti-spectrin specificities were more enhanced than the others. Similarly, IgM antibodies reacting with all the PAg, probably forming glomerular deposits, were eluted from the kidneys of B/W mice; these antibodies were considerably increased after the 8th month. IgM antibodies both in CIC and in kidneys bore the D23 idiotype, and this idiotype increased to a higher level in the kidney than in serum during the progression of the disease. Whether IgM antibodies play a pathogenic role in the lupus-like syndrome observed in B/W mice has not so far been resolved. On the contrary, it was shown that some IgM antibodies were not pathogenic [29], and it was even postulated that IgM antibodies may play a protective role against autoimmune diseases [31]. Furthermore, most studies on antibodies from kidney eluates concerned IgG, and only occasionally IgM. When IgM was compared with IgG, it was reported to be present in small amounts [32,33]. In contrast, in our study, IgM reacting with all the PAg, thus polyreactive and most probably reflecting an increase in serum IgM, was found in kidneys in substantial amounts and especially at 8 to 10 months of age (IgM to IgG ratio 1:3). Thus, on the whole, the results obtained with IgM in this investigation

Autoantibodies during murhe lupus

353

seem to indicate that at least some of these antibodies possess characteristics similar to those of the natural polyreactive IgM antibodies, and by forming CIC and deposits in the kidney play a pathogenic role in the murine lupus-like syndrome. The implication of these IgM in the disease can be further strengthened if one considers that, in normal mice, IgM antibodies exert an idiotype-like regulation of autoreactive IgG, as demonstrated by Adib et al. [34], that is probably modified in B/W mice because of the marked and unequal increase in IgM and IgG antibodies. The kinetics of IgG antibody appearance in the serum of B/W mice demonstrate an early increase, at 2 months of anti-TNP antibodies that remained more or less constant throughout the experimental period, followed at 5 months by a progressive increase not only in anti-DNA antibodies, but also antibodies recognizing the other antigens of the panel. However, this increased antibody activity did not equally affect all the PAg; anti-tubulin and anti-spectrin antibodies were only moderately increased until 8 months, but they considerably increased at 9 and 10 months. Recently, Klinman et al. [35] demonstrated the presence of IgM in young MRL-lpr/ 1pr mice, and in old animals of IgG anti-Sm antibodies reacting with self proteins similar to those used in the present study. They noted an increase in IgG antibodies at 12 weeks of age but we were unable to detect such an increase before the 5th month. This difference might be as a result of the earlier appearance of the lupus disease in MRL-lpr/lpr mice than in B/W mice. Nonetheless, we were able to detect an increase in IgG anti-TNP activity even at 2 months. If IgM anti-TNP can be considered a marker of polyclonal activation [36], it can be deduced that during lupus there is early polyclonal activation. Consequently, the IgG production, like the IgM kinetics we observed, can be interpreted on the basis of the concept, already advanced [35], that in lupus a first moderate polyclonal stimulation is followed by more intense specific responses, probably antigen driven. In such a view, DNA seems to be the privileged antigen, as suggested by Marion et al. [37], but others, such as actin and myosin, might also contribute significantly. The IgG anti-DNA antibodies isolated on DNA-immunoadsorbent from B/W mice sera, unlike the IgM antibodies, possessed restricted specificity and reacted essentially with DNA and TNP. Similarly, the IgG isolated on TNP-immunoadsorbent bound only to TNP and DNA. Finally, a majority of IgM and IgG anti-DNA antibodies expressed the D23 idiotype. Therefore, it would appear that during lupus a switch from polyspecific IgM anti-DNA to IgG with a more restricted specificity occurs. However, as previously reported [26, 38, 391, the anti-DNA activity seems almost always to include anti-TNP activity. The IgG antibodies found in the CIC of B/W mouse sera mainly recognized only DNA and TNP. This contrasts with the IgG eluted from kidneys that reacted with all the PAg, and increased with age. These results are in accordance with those of Pankewycz et al. [32] and indicate that polyreactivity is found to a much greater extent among nephritogenic Ig than serum Ig. This polyreactive feature may influence the capacity of autoantibodies to form immune deposits because of their cross-reactions with antigenic determinants within the glomerular basement membrane, as advanced by Madaio et al. [40]. The expression of similar idiotypes, on B/W and normal mouse anti-DNA antibodies [41], as well as on human anti-DNA antibodies from SLE patients and normal individuals [42], has been reported. Nevertheless, the pathogenic role of the human

354

Basma

Hentati

et al.

16/6 Id in SLE has been clearly demonstrated [reviewed in 431 and its expression is maintained and enhanced during the disease, in contrast with other anti-DNA antibody Id, for which a shift in expression was observed during the evolution of lupus [44]. Moreover, the pathogenicity of the 16/6 Id was confirmed by induction of an SLE-like disease in mice after immunization with the 16/6 Id [45]. In B/W mice, Ebling et al. [46] reported pathogenic idiotypes specific for the disease which are not expressed in normal mice and which differed in serum and in kidneys. In contrast, in the present study, the D23 idiotypic marker of natural polyspecific antibodies was present in both normal and B/W mice, on IgM and IgG populations. An increase in Id D23 expression was demonstrated, comparatively more intense than the increase in the amount of Ig during the disease. The Id D23 was expressed in both serum and kidneys from B/W mice, in contrast with the Id markers IdGNl and IdGN2 described by Ebling et al. [46] only on the nephritogenic subsets of antibodies. This difference could be owing to the fact that the anti-Id D23 used in this study were polyclonal anti-idiotype antibodies which could recognize several idiotypes on the natural antibody D23. These results demonstrate that natural antibodies are a part of the population of autoantibodies increased in this disease, consistent with the results of Mahana et al. [47] who demonstrated a decrease in anti-DNA antibodies in MRL-lpr/lpr after treatment with anti-Id D23 antibodies.

Acknowledgements This work was supported by grant 6899 from the Association pour la Recherche sur le Cancer. We would like to thank Mrs J. Grttgoire for her technical assistance. The secretarial assistance of Mrs C. Core1 and the re-reading of the manuscript by Mrs J. B. Jacobson are greatly appreciated.

References 1. Koffler, D., R. I. Carr, V. Agnello, R. Thoburn,

polynucleotides

and H. G. Kunkel. 1971. Antibodies to in human sera. Antigenic specificity and relation to disease. J. Exp. Med.

134: 294-312

2. Tan, E. M. 1982. Autoantibodies to nuclear antigens: their immunobiology and medicine. Adv. Immunol. 33: 167-240 3. Lafer, E. M., J. Rauch, C. Andrzejewski, D. Mudd, B. Furie, B. C. Furie, R. S. Schwartz, and B. D. Stellar. 1981. Polyspecific monoclonal lupus autoantibodies reactive with both polynucleotides and phospholipids. 3. Exp. Med. 153: 897-909 4. Naparstek, Y., D. Duggan, A. Schattner, M. Madaio, F. Goni, B. Frangione, B. D. Stollar, E. A. Kabat, and R. S. Schwartz. 1985. Immunochemical similarities between monoclonal antibacterial WaldenstrGm’s macroglobulins and monoclonal anti-DNA lupus autoantibodies. 3. Exp. Med. 161: 1525-1538 J., S. K. Datta, Y. Shoenfeld, D. A. Isenberg, B. D. Stollar, and 5. And&Schwartz, R. S. Schwartz. 1984. Binding of cytoskeletal proteins by monoclonal anti-DNA lupus autoantibodies. Clin. Immunol. Immunopathol. 31: 261-271 6. Jacob, L., F. Tron, J. F. Bach, and D. Louvard. 1984. A monoclonal anti-DNA antibody also binds to cell-surface protein(s). Proc. Natl. Acad. Sci. (USA) 81: 3843-3845 7. Avrameas, S. 1986. Natural autoreactive B cells and autoantibodies: the ‘know thyself’ of the immune system. Ann. Zmmunol. Inst. Pasteur 137D: 150-156

Autoantibodies during murine lupus

a. 9.

10.

11. 12.

13. 14. 15. 16. 17. 18.

19. 20.

21.

22.

23.

24. 25.

26. 27.

28.

29.

355

Schwartz, R. S. and B. D. Stollar. 1985. Origins of anti-DNA antib0dies.J. Cl&. Znvesr. 75: 321-327 Schlomchik, M. J., A. Marshak-Rothstein, C. B. Wolfowicz, T. L. Rothstein, and M. G. Weigert. 1987. The role of clonal selection and somatic mutation in autoimmunity. Nature 328: 805-8 11 Lymberi, P., G. Dighiero, T. Ternynck, and S. Avrameas. 1985. A high incidence of cross-reactive idiotypes among murine natural autoantibodies. Eur. J. Zmmunol. 15: 702-707 Spudish, J. A. and S. Watt. 1971. The regulation of rabbit skeletal muscle contraction. J. Biof. Chem. 246: 4866-4892 Whalen, R. G., G. S. Butler-Browne, and F. Gros. 1978. Identification of a novel form of myosin light chain present in embryonic muscle tissue and cultured muscle cells. J. Mol. Biol. 126: 415-420 Shelansky, M. L., F. Gaskin, and C. Cantor. 1973. Microtubule assembly in the absence of added nucleotides. Proc. Natl. Acad. Sci. (USA) 70: 765-768 Ungewickell, E. and W. Gratzer. 1978. Self-association of human spectrin. A thermodynamic and kinetic study. Eur. J. Biochem. 88: 379-385 Little, J. R. and H. N. Eisen. 1966. Preparation and characterization of antibodies specific for the 2,4,6_trinitrophenyl group. Biochemistry 5: 3385-3395 Ternynck, T. and S. Avrameas. 1976. Polymerization and immobilization of proteins using ethylchloroformate and glutaraldehyde. Stand. 3’. Zmmunol. 3: 29-35 Avrameas, S., T. Ternynck, and J. L. Guesdon. 1978. Coupling of enzymes to antibodies and antigens. Stand. 3. Zmmunol. 8 (suppl. 7): 7-23 Dighiero, G., P. Lymberi, J. C. Mazik, S. Rouyre, G. S. Butler-Browne, R. G. Whalen, and S. Avrameas. 1983. Murine hybridomas secreting natural monoclonal antibodies reacting with self antigens. 3. Zmmunol. 131: 2267-2272 Ohlson, S. and K. Zetterstrand. 1985. Detection of circulating immune complexes by PEG precipitation combined with ELISA. 3. Zmmunol. Methods 77: 87-93 Louzir, H., T. Ternynck, Y. Gorgi, K. Ayed, and S. Avrameas. 1988. Enzyme immunoassay analysis of antibody specificities present in the circulating immune complexes of selected pathological sera. 3. Zmmunol. Methods 114: 145-153 Pereira, A. B., A. N. Theofilopoulos, and F. J. Dixon. 1980. Detection and partial characterization of circulating immune complexes with solid-phase anti-C3.3. Zmmunol. 125: 763-770 Kubota, T., T. Akatsuka, and Y. Kanai. 1985. DNA affinity column chromatography: application in the isolation of distinct antibody populations from SLE sera. Clin. Exp. Zmmunol. 62: 321-328 Jaffe, B. M., H. N. Eisen, E. S. Simms, and M. Potter. 1969. Myeloma proteins with antihapten antibody activity s-2,4_dinitrophenyl lysine binding by the protein produced by mouse plasmacytoma MOPC 460.3. Zmmunol. 103: 872-874 Saxena, R., P. Bygren, R. Butkowski, and J. Weislander. 1989. Specificity of kidneybound antibodies in Goodpasture’s syndrome. Clin. Exp. Zmmunol. 78: 31-36 Shoenfeld, Y., J. Rauch, H. Massicotte, S. K. Datta, J. Andr&Schwartz, B. D. Stollar, and R. S. Schwartz. 1983. Polyspecificity of monoclonal lupus autoantibodies produced by human hybridomas. N. Engl. 3. Med. 308: 414-420 Matsiota, I’., P. Druet, P. Dosquet, B. Guilbert, and S. Avrameas. 1987. Natural autoantibodies in systemic lupus erythematosus. Clin. Exp. Zmmunol. 69: 79-88 Pennebaker, J. B., J. N. Gillian, and M. Ziff. 1977. Immunoglobulin classes of DNA binding activity in serum and skin in systemic lupus erythematosus. 3. Clin. Invest. 69: 1331-1338 Steward, M. W. and F. C. Hay. 1976. Changes in immunoglobulin class and subclass of anti-DNA antibodies with increasing age in NZB/W Fl hybrid mice. Clin. Exp. Zmmunol. 26: 363-370 Ebling, F. and B. H. Hahn. 1980. Restricted subpopulations of DNA antibodies in kidneys of mice with systemic lupus comparison of antibodies in serum and renal eluates. Arthritis Rheum. 23: 392-398

356

Basma

Hentati

et al.

30. Talal, N., R. J. Pillarisetty, R. J. DeHoratius, and R. P. Messner. 1976. Immunologic regulation of spontaneous antibodies to DNA and RNA. I. Significance of IgM and IgG antibodies in SLE patients and asymptomatic relatives. Clin. Exp. Zmmunol. 25: 377-382 31. Cohen, I. R. and A. Cooke. Natural autoantibodies might prevent autoimmune disease. Zmmunol. Today 7: 363-364 32. Pankewycz, 0. G., P. Migliorini, and M. P. Madaio. 1987. Polyreactive autoantibodies are nephritogenic in murine lupus nephritis. J. Immunol. 139: 3287-3294 33. Slack, J. H., L. Hang, J. Barkley, R. J. Fulton, L. D’Hoostelaere, A. Robinson, and F. J. Dixon. 1984. isotypes of spontaneous and mitogen-induced autoantibodies in SLE-prone mice. y. Zmmunol. 132: 1271-1275 34. Adib, M., J. Ragimbeau, S. Avrameas, and T. Ternynck. 1990. IgG autoantibody activity in normal mouse serum is controlled by IgM. J. Zmmunol. 145: 3807-3813 35. Klinman, D. M., R. A. Eisenberg, and A. D. Steinberg. 1990. Development of the autoimmune B cell repertoire in MRL-lpr/ lpr mice. 3. Zmmunol. 144: 506-5 11 36. Anderson, J. O., 0. Sjoberg, and G. Moller. 1972. Mitogens as probes for immunocyte activation and cellular cooperation. Transplant Rev. 11: 13 1-l 77 37. Marion, T. N., A. L. M. Bothwell, D. E. Briles, and C. A. Janeway, Jr. 1989. IgG antiDNA autoantibodies within an individual autoimmune mouse are the products of clonal selection. 3. Zmmunol. 142: 4269-4274 38. Serban, D., C. Rordorf-Adam, Y. Z. Sun, and J. Gordon. 1985. Monoclonal and serum anti-DNA antibodies cross-reactive with 2,4,6_trinitrophenyl derivatives. 3. Zmmunol. 135: 3122-3127 39. Mahana, W., I’. Matsiota, and S. Avrameas. 1987. Both natural and induced murine antiTNP antibodies possess anti-DNA activity. Ann. Zmmunol. (Inst. Pasteur) 138: 805-814 40. Madaio, M. P., J. Carlson, J. Cataldo, A. Ucci, PI Migliorini, and 0. Pankewycz. 1987. Murine monoclonal anti-DNA antibodies bind directly to glomerular antigens and form immune deposits. 3. Immunol. 138: 2883-2889 41. Jacob, L., F. Tron, M. A. Lety, and J. F. Bach. 1986. Idiotypes of monoclonal anti-DNA antibodies produced in autoimmune B/W mice expressed in normal mice. Clin. Exp. Zmmunol. 63: 402-407 42. Madaio, M. P., A. Schattner, M. Schattner, and R. S. Schwartz. 1986. Lupus serum and normal human serum contain anti-DNA antibodies with the same idiotypic marker. 3. Zmmunol. 137: 2535-2540 43. Schoenfeld, Y., H. A. Teplizki, S. Mendlovic, M. Blank, E. Mozes, and D. A. Isenberg. 1989. The role of the human anti-DNA idiotype 16/6 in autoimmunity. Clin. Zmmunol. Immunopathol. 51: 313-325 44. Suenaga, R., P. A. Munoz, S. W. Bright, and N. I. Abdou. 1988. Spontaneous shift of anti-DNA antibody idiotypes in systemic lupus erythematosus. 3. Zmmunol. 140: 3508-35 14 45. Mendlovic, S., S. Brocke, Y. Shoenfeld, M. Ben-Bassat, A. Meshorer, R. Bakimer, and E. Mozes. 1988. Induction of a systemic lupus erythematosus-like disease in mice by a common human anti-DNA idiotype. Proc. Natl. Acad. Sci. (USA) 85: 2260-2264 46. Ebling, F. M. and B. H. Hahn. 1989. Pathogenic subsets of antibodies to DNA. Intern. Rev. Zmmunol. 5: 79-95 47. Mahana, W., B. Guilbert, and S. Avrameas. 1987. Suppression of anti-DNA antibody production in MRL mice by treatment with anti-idiotypic antibodies. Clin. Exp. Immunol. 70: 538-545