- Email: [email protected]
A novel trait of naturally occurring anti-DNA antibodies: dissociation from immune complexes in neutral 0.3–0.5 M NaCl
Immunology Letters, 22 (1989) 293- 300
Elsevier IMLET 01296
A novel trait of naturally occurring anti-DNA antibodies: dissociation from immune complexes in neutral 0.3-0.5 M NaC1 Y. K a n a i 1 a n d T. K u b o t a 2 1Department of Molecular Oncology, Institute of Medical Science, University of Tokyo, and 2First Department of Internal Medicine, Tokyo Medical and Dental University, Tokyo, Japan
(Received 28 April 1989; revisionreceived 7 July 1989; accepted 11 July 1989)
1. Summary Monoclonal and polyclonal anti-DNA antibodies from autoimmune mice, and experimentally induced rabbit anti-nucleic acid polyclonal antibodies were tested for stability o f binding to nucleic acids in the presence o f various concentrations o f NaC1 by an enzyme-linked immunosorbent assay (ELISA). Murine monoclonal antibodies 2C10 (IgG2b) and 1A2 (IgG2a), which are known to react specifically with double-stranded (ds) DNA, dissociated completely from their complexes with DNA when washed with a neutral 0.5 M NaCI solution. Another monoclonal antibody (MoAb) (IgM,r), polyreactive with singlestranded (ss) DNA, cardiolipin, and trinitrophenylhapten (TNP), was also dissociated from its complexes with ss DNA, but not from its complexes with TNP, by 0.3-0.5 M NaC1. Similar differences were observed in the binding stability of serum antibodies from autoimmune mice to DNA and TNP. In contrast, anti-nucleic acid polyclonai antibodies induced in rabbits by immunization with poly(I), poly(dT) or poly(ADP-ribose) were not significantly dissociated from their immune complexes with relevant antigens or DNA by 0.5 M NaC1. The finding that nucleic acid antigens were not detached from a solid phase by washing with 0.5 M NaC1 solution indicated that the reduction of binding of anti-DNA
Key words: Naturally occurring anti-DNA antibody; Immune complex; Affinity column Correspondence to: Dr. Y. Kanai, Department of MolecularOncology,Instituteof MedicalScience,Universityof Tokyo,Minatoku, Tokyo 108, Japan.
antibodies in both MoAbs and naturally occurring antibodies was really due to dissociation of the antibodies from immune complexes. This is the first demonstration that DNA epitopes recognized by naturally occurring antibodies in both SLE and its mouse models are sensitive to neutral NaC1 concentrations. This novel trait o f naturally occurring antibodies will be very useful in studies on the nature of immune complexes in sera and kidneys of cases of systemic lupus erythematosus (SLE).
2. Introduction There is extensive evidence that monoclonal antiDNA antibodies are not specific to a given antigen but are polyreactive; i.e., they react with epitopes of RNA phospholipids and of the cytoskeleton [1, 2]. Moreover, studies on a murine lupus model have implicated polyclonal B cell activation (PBA) as the basis for induction o f autoantibodies in SLE [3, 4]. The diversity of autoantibodies in SLE may be explained partly by the polyreactivity of a single immunoglobulin molecule or by PBA, and also partly by some electrostatic interactions between slightly positively charged antibodies and negatively charged antigens such as nucleic acids and acidic phospholipids. To investigate this possibility, we first examined the stability in neutral 0.2-0.5 M NaC1 solution of antigen-antibody complexes formed by oligo- and monospecific anti-DNA MoAbs (IgG2a and IgG2b) and polyreactive anti-DNA IgM,r MoAb. Unexpectedly, we found that the oligo- and monospecific IgG MoAbs were dissociated completely in 0.5 M NaCI solution, but that the polyreactive IgM MoAb was not. Using a series of auto-
0165-2478 / 89 / $ 3.50 © 1989 ElsevierSciencePublishers B.V.(BiomedicalDivision)
293
immune and experimentally induced anti-nucleic acid antibodies, we found that naturally occurring anti-DNA antibodies, in contrast to experimentally induced ones, could be dissociated more or less completely from the immune complexes by a neutral 0.5 M NaC1 solution. Until now, acid or alkaline elution has mainly been used to dissociate conventional immune complexes, including complexes of antiDNA antibodies with DNA [5]. We have also reported a method for elution of anti-DNA antibodies from a DNA affinity column [6]. However, so far there have been no reports on the dissociation of anti-DNA antibodies from immune complexes with neutral NaCI solution at about 0.5 M. The present study demonstrates a novel trait of naturally occurring anti-DNA antibodies and the application o f this trait for the isolation o f naturally occurring antiDNA antibodies from autoimmune sera. The mechanism and importance of this novel trait in systemic autoimmune disease are discussed. 3. Materials and Methods
3.1. Mice M R L / M p - l p r / l p r (MRL/1) and M R L / M p - +/+ (MRL/n) mice, originally purchased from the Jackson Laboratory, Bar Harbor, ME, U.S.A. are maintained in our laboratory. CBA/K1 mice, originally obtained from the Karolinska Institute, Stockholm, Sweden, have been maintained in this Institute since 1969. Mutant mice with general lymphadenopathy were recently detected by Matsuzawa et al. in our CBA/K1 mouse colony. After repeated sisterbrother mating of these mice, we examined autoantibodies such as anti-ss DNA and ds DNA antibodies in their sera. The results showed that all mutant mice o f both sexes produced both anti-ss DNA and antids DNA antibodies with high titers at about 5 months of age, and thereafter showed a switch from IgM to IgG. Details of findings in CBA/K1 mutant mice will be reported separately. 3.2. MoAbs Murine MoAbs were obtained as described previously [7]. The detailed characterization o f MoAbs 2C10 and 1A2, which react specifically with ds DNA, has been reported [8]. MoAbs 3E8 (IgM,K) 294
and 1D8 (IgM,K) were newly obtained from a CBA/K1 mutant mouse (9-month-old female): the former reacted with ss DNA, cardiolipin and TNP, and the latter also reacted with ds DNA and RNA. 3.3. Polyclonal antibodies Anti-poly(ADP-ribose) polyclonal antibodies (PoAbs) were raised in New Zealand white rabbits as described [9]. Anti-poly(I) and anti-poly(dT) PoAbs were obtained similarly: the former was specific for RNA and the latter specific for ss DNA (data not shown). Naturally occurring anti-DNA and antiT N P PoAbs were obtained from a 4-month-old female MRL/1 mouse and from a 9-month-old female CBA/K1 mutant mouse, respectively.
3.4. Determination of antibodies by ELISA Anti-nucleic acid antibodies were measured by ELISA as described [7]. Anti-TNP antibodies were also measured by ELISA. Briefly, 50 ~1 o f a solution containing 50 ng of TNP-bovine serum (BSA) (molar ratio of T N P to BSA was 1:55; a gift from Dr. Katagiri, Department of Immunology of this Institute) in 10 mM carbonate buffer, p H 9.8, was placed in wells of polystyrene microtiter plates (Immulon No. 2, Dynatech) and incubated at room temperature for 2 h. The plates were washed with Trisbuffered saline (TBS; 25 mM Tris/140 mM NaCI, p H 7.4) containing 0.1% Tween 20. The subsequent procedure was described for anti-nucleic acid antibodies [7]. 3.5. Examination of the stability
of antigenantibody complexes in neutral 0.3-0.5 M NaCI solution
Anti-DNA, anti-RNA, anti-poly(ADP-ribose) and anti-TNP antibodies were added to individual antigens in microtiter plates as described above. The wells o f the plates were washed separately four times each with 0.2-0.5 M NaCI in TB and then washed three times with TBS containing 0.05% Tween 20. Residual bound antibody was measured by conventional ELISA [7]. The A405 after washing with TBS without additional NaCI was considered to express 100% bound antibody.
3.6. Dissociation of antibodies from an affinity
Column by 0.5 M NaCI in TB ss DNA- or ds DNA-bound Sepharose was prepared as described [6]. Usually, antibodies bound to the column are eluted with 20 mM carbonate buffer containing 5070 dimethyl-sulfoxide (DMSO), p H 10.5 (carbonate-DMSO), but in this study, the affinity column containing bound anti-DNA antibodies was washed first with 0.5 M NaC1 in TB, and then with carbonate-DMSO. 3.7. Western immunoblotting for heavy and light
chains of mouse immunoglobulins Western immunoblotting was carried out using an SDS-polyacrylamide (12.5070) slab gel with electrophoretic transfer o f proteins to Millipore GVHP nitrocellulose sheets as described [10]. The amount of IgG loaded onto the gel was more than 4 t~g per lane.
B
__A
100 1 ~ _ ~
50 ~
0
00
50
i f i J0 014 0.2 0.3 04 0.5 0.1402 0.3 0.4 0.5 0,4 0!2 013 014 015 NaCI (M3
Fig. 1. Dissociationof immunecomplexesofmurine MoAbsand DNA (A, B and C), but not of those of MoAb 3E8 and TNP (C) by neutral 0.3 -0.5 M NaCI. 100O7obound means about 1 A4o5U antibody activity obtained under conventional conditions. See text for details. A, MoAb 1A2; B, MoAb 2C10; C, MoAb 3E8. o, ss DNA binding; e, ds DNA binding; A, TNP binding.
reacted with ss D N A and TNP. As shown in Fig. 1C, neutral NaC1 solution also dissociated this MoAb from ss DNA, but did not dissociate it from TNP.
4. Results 4.2. Dissociation of immune complexes of natural-
4.1. Dissociation of immune complexes formed between murine MoAbs and DNA by neutral 0.3- 0.5 M NaCI Murine MoAbs 2C10 (IgG2b) and 1A2 (IgG2a) are known to bind with both ds D N A and ss DNA, but the former prefers ds D N A over ss D N A [8]. After primary interaction of these MoAbs with ds D N A or ss DNA, immune complexes formed on the E L I S A plates were washed individually with solutions o f 0 . 2 - 0 . 5 M NaCI in TB, as described in Materials and Methods. As shown in Fig. 1A and B, antibodies bound to D N A were dissociated as a function o f the neutral NaCI concentration used for washing, and complete dissociation was achieved with 0.5 M NaC1. MoAb bound to ss D N A plates was dissociated at a lower NaCI concentration than that bound to ds D N A plates, but there was no significant difference between the NaC1 concentrations required for the dissociations o f 2C10 and 1A2 MoAbs. To determine whether this dissociation from immune complexes by neutral 0 . 3 - 0 . 5 M NaCI was a c o m m o n characteristic of MoAbs, we also examined another MoAb 3E8 [IgM,x] which was derived from a CBA/K1 mutant mouse and
ly occurring anti-DNA PoAbs and DNA by neutral 0.3 - 0.5 M NaCl solution Similar experiments with serum containing high titers of IgG class anti-ss D N A and anti-ds D N A antibodies selected from those of a number of autoimmune MRL/1 mice were carried out next. As this serum had very low titers of anti-TNP antibodies, we also used serum containing IgM antibodies with high T N P reactivity selected from those of a number o f CBA/K1 mutant mice. As shown in Fig. 2A, no significant difference was observed between the dissociations of anti-ss D N A and anti-ds D N A antibodies from DNA, which reached levels of up to 70°7o dissociation. In contrast, IgM PoAbs, like MoAb 3E8, was not dissociated from T N P by neutral solutions of up to 0.5 M NaC1, as shown in Fig. 2B. 4.3. Effect of neutral NaCI solution on immune
complexes formed between experimentallyinduced anti-nucleic acid lgG antibodies and nucleic acids In contrast to naturally occurring PoAbs and 295
A
B 100
100
100
100
50
50
•
e
50
50
o m , 0. 4 0.2 03 0.4 05
0
,,
0
I
0
~
--" 100
0.14 012 01.3 014 0'5 c
i
0.14 0.2 03 04 0.5
100
i
e
012 03 014 015
E ~
NaCI (M) Fig. 2. Dissociation of immune complexes of anti-DNA IgG PoAb from an MRL/1 mouse and DNA (A), but not o f those of anti-TNP IgM PoAb from a CBA/K1 mutant mouse and TNP (B) by neutral 0.3 - 0 . 5 M NaCI. The serum dilution was 1:100; 100% bound means about 1 A405 U antibody activity obtained under conventional conditions. See text for details, o , ss DNA binding; o , ds DNA binding; A, TNP binding.
50
50
0 0.;4 012 013 014 015
0 off4 012 013 014 015
NaCI (M)
MoAbs from MRL/1 and CBA/K1 mutant mice, experimentally induced rabbit IgG PoAbs to poly(I), poly(dT) and poly(ADP-ribose) were not readily dissociated from immune complexes by neutral solutions o f up to 0.5 M NaCI, as shown in Fig. 3, although higher NaC1 concentrations caused 30% dissociation of anti-poly(dT) antibody from calf thymus ss DNA, as shown in Fig. 3C.
Fig. 3. Effects of neutral 0.3-0.5 M NaC1 solution on the immune complexes formed between experimentally induced antinucleic acid rabbit IgG PoAbs and nucleic acids. 100% bound is defined as for Figs. 1 and 2. Alkaline phosphatase-conjugated anti-rabbit IgG antibodies (Sigma) were used in place of alkaline phosphatase-conjugated anti-mouse isotype specific antibodies. A, anti-poly(I) rabbit-antibodies; B, anti-~91y(dT) rabbit antibodies; C, anti-poly(dT) antibodies bound with calf thymus ss DNA; D, anti-poly(ADP-ribose) rabbR.~emtibodies.
1 /It
4.4. Elution of anti-DNA antibodies from a DNA affinity column with neutral 0.5 M NaCI solution and their demonstration by both ELISA and immunoblotting A sample o f 200 #1 of MRL/1 mouse serum containing anti-ss and ds DNA antibodies was applied to a DNA affinity column and the column was washed extensively with TBS alone and then developed with 0.5 M NaC1 in TB. Fractions o f the eluate were found to contain IgG anti-ss and ds DNA antibodies, as shown in Fig. 4A. Although the protein peak eluted with carbonate-DMSO had DNAbinding activity, the relative amount of the fraction was less than 15°70 judging from A280 U recovered. As a positive control, the culture supernatant of the 1A2 or 2C10 hybridoma was treated similarly; the anti-DNA antibodies eluted f r o m t h e column are 296
'B/
7
.< 0.5 Z
£3
ol
o I
i
I
i
I
0.01 0.02 0.05 0.1 0.2 Protein,
8==8===8==8==~ i
r
i
e
I
0.01 0,02 0,05 0.1 0.2 A2ao U/ml
Fig. 4. Dose-dependent binding to DNA of MRL/1 mouse PoAb (A) and MoAb 1A2 (B) eluted from a DNA affinity column with neutral 0.5 M NaC1. o , ss DNA binding; o , ds DNA binding. , IgG antibodies; - - -, IgM antibodies.
shown in Fig. 4B. Confirmation that the fractions from the DNA affinity column eluted with 0.5 M NaC1 really contained IgG and/or IgM was obtained by Western immunoblotting, as described in Materials and Methods. As shown in Fig. 5, anti-IgG (7-chain-specific) and anti-IgM (/~-chain-specific) reactive bands corresponding to sizes of 50 and 72 kDa, respectively, were immunostained. AntiIgG whole-molecule antibodies stained the bands o f the light chains of both IgM and IgG, corresponding to a molecular weight o f 22 kDa. Protein staining demonstrated the existence o f additional bands that reacted with DNA in both MRL/1 mouse serum and the supernatants o f hybridoma cultures. Anti-/~chain-specific antibodies also detected a band of 50 kDa other than the 72-kDa band o f the heavy chain. This was detected with murine IgM MoAb, but not with a myeloma culture supernatant such as Sp2/0-Ag-14 (data not shown). Therefore, the 50-kDa band is probably derived from the heavy chain o f the IgM during denaturation. This 50-kDa band disappeared when the amount of Ig loaded onto the gel was less than 1/zg, while the 72-kDa
Mabcd
B Fig. 5. Western immunoblotting of MoAb 1A2 (A) and an MRL/I mouse PoAb (B) eluted from the DNA affinity column. Lane a, protein staining with Coomassie Blue; lanes b, c and d, immunoblotting with anti-mouse IgG whole molecule (Sigma), anti-mouse it-chain-specificgoat antibodies (Zymed) and antimouse v-chain-specificrabbit antibodies (Pel-Freez),respectively. L, light chain; it, # chain; 3', 3' chain.
band of the heavy chain was clearly visible (data not shown). 5. Discussion
The autoantibody diversity found in SLE may be due to polyreactivity o f a single molecule or socalled PBA. Polyreactivity o f MoAbs has been explained by the shared structures of macromolecules such as DNA and phospholipids, in which similar spacing o f phosphodiester groups occur [1]. We are considering another possibility, namely that it is due to the interactions between slightly positively charged antibodies and negatively charged molecules such as nucleic acids and some phospholipids. We have examined this possibility using an ELISA system [7]. We were surprised to find in this study that MoAb 2C10, which is considered to be a prototype anti-ds DNA antibody [8], was dissociated with 0.3 M NaCI from DNA coated on a solid phase (Fig. 1). In fact, 2C10 was recently proved to be a prototype of anti-DNA antibodies, by gel retardation experiments using a number of synthetic doublestranded oligonucleotides with special base sequences (B. D. Stollar et al., in preparation). This finding raised two questions: firstly, are nucleic acid antigens non-covalently attached to the solid phase stable against washing with 0 . 3 - 0.5 M NaCI? Secondly, if so, are they still stable at these concentrations o f NaC1 after formation of immune complexes? To answer the first question, we washed ds DNA- or ss DNA-coated polystyrene plates with neutral 0.5 M NaCI before carrying out ELISA, but obtained the same results as without pre-washing the plates with 0.5 M NaC1 (data not shown). To answer the second question, we reused plates from which bound antibodies had been dissociated with 0.5 M NaC1, and found that antibody binding was unchanged as compared to fresh plates coated with ds DNA or ss DNA (data not shown). These data indicated that DNA antigen attached to polystyrene plates through poly-0.-lysine) was stable in solutions of up to 0.5 M NaCI. Naturally occurring anti-DNA antibodies from MRL/1 mice were also found to be dissociated by NaCI, although about 30°7o residual binding was observed after washing with 0.5 M NaCI (Fig. 2). This 30°7o residual binding o f polyclonal anti-DNA antibodies of the MRL/1 mice may reflect their poly297
reactivity with a number of related compounds, because MoAb 1D8, which was derived from a CBA/K! mutant mouse, was polyreactive, reacting with ds DNA, ss DNA, cardiolipin, RNA and T N P (data not shown), and it was not dissociated from DNA with neutral 0.5 M NaCl (data not shown). Whether or not 3007o residual binding o f polyclonal anti-DNA antibodies of MRL/1 mice really reflects polyreactivity should await further experiments using many samples obtained by 0.5 M NaCl elution and by alkaline elution with carbonate-DMSO. It should be mentioned here that cross-reaction of 2C10 and 1A2 MoAbs with ds DNA and ss DNA could be due to $1 nuclease-resistant double helical cores in the ss DNA molecules, as described by Stollar and Papalian [11]. However, we found that MoAbs 2C10 and 1A2 dissociated from their immune complexes with ss DNA more readily than from their immune complexes with ds D N A at higher concentrations of NaCl (Fig. 2). This may be due to the structural difference in the double helices of ss DNA and ds DNA: the double helices of the former are more susceptible to higher NaCl concentrations than those of the latter. This difference in susceptibility to NaCl concentration was not clear with naturally occurring anti-DNA antibodies in the sera of an MRL/1 mouse, but was observed with anti-ss DNA antibodies of patients with SLE (data not shown). The conformational structures rather than the double helices common to both ss DNA and ds DNA must be involved in the recognition by naturally occurring anti-ss DNA antibodies, because MoAb 3E8, which binds with ss DNA but not with ds DNA, was also susceptible to higher concentrations of NaC1 (Fig. 1C). In contrast, experimentally induced antibodies to nucleic acids were rather stable to higher concentrations of NaCl: immune complexes of anti-poly(I) and anti-poly(dT) with relevant antigens were quite stable (Fig. 3A and B) but 3007o dissociation of antipoly(dT) antibodies from calf thymus ss DNA was observed in 0.5 M NaCl solution, suggesting that some epitopes recognized by anti-poly(dT) PoAb were altered by exposure to higher concentrations of NaCl, even after formation of immune complexes (Fig. 3C). In normal animals, antibodies are produced against relatively simple antigenic determinants such as haptens and/or primary structures, 298
whereas in SLE these antigenic determinants are more complex helical structures [12]. Therefore, it is possible that the anti-nucleic acid antibodies induced experimentally bind firmly to simple epitopes, whereas those found in SLE are very sensitive to minor changes in nucleic acid structures induced by higher concentration of NaC1. The changes in DNA structures induced by higher concentrations of NaC1 were reversible because after removal of bound antibodies with neutral 0.5 M NaC1 and washing with normal saline, the plates could again bind MoAbs 1A2 and 2C10 (data not shown). After application o f MoAb 1A2 to a DNA affinity column, we demonstrated elution of antibodies with neutral 0.5 M NaC1 solution: the fractions of eluate contained 1A2 MoAb that reacted with DNA dose-dependently (Fig. 4B). We also demonstrated elution of 2C10 MoAb from a DNA affinity column (data not shown). Anti-DNA antibodies in the serum of an MRL/1 mouse could also be eluted from a DNA affinity column with 0.5 M NaC1, but IgM detected by Western immunoblotting did not bind to DNA, as shown by ELISA (Fig. 4A), indicating non-specific binding of IgM to DNA. The band corresponding to the IgG heavy chain, seen by immunoblotting with anti-#-chain antibodies (Fig. 5B, lane C) was not seen, when a s m i l e r amount of IgM was loaded on the gel or when IgG MoAb 1A2 was applied to the gel (Fig. 5A, lane C), indicating that a cleavage product of IgM was formed under the reducing conditions used for electrophoresis. In any case, experiments using a DNA affinity column confirmed that anti-DNA antibodies could be recovered by elution with 0.5 M NaCl and that reduction in the binding of naturally occurring antiDNA antibodies on the solid phase was actually due to dissociation of antibodies from the immune complexes. The naturally occurring anti-DNA antibodies in patients with SLE were also dissociated from complexes by 0.5 M NaC1. Detailed studies on the nature of the dissociated anti-DNA antibodies of SLE patients will be reported separately. This common trait of naturally occurring anti-DNA antibodies in lupus model mice and human SLE will be useful in studies on the nature of the immune complexes in the sera and the kidney of SLE patients.
References [1] Lafer, E. M., Rauch, J., Andrzejewski Jr., C., Mudd, D., Furie, B., Schwartz, R. S. and Stollar, B. D. (1981) J. Exp. Med. 153, 897. [2] Andr6-Schwartz, J., Datta, S. K., Shoenfeld, Y., Isenberg, D. A., Stollar, B. D. and Schwartz, R. S. (1984) Clin. Immunol. Immunopathol. 13, 261. [3] Theofilopoulos, A. N. and Dixon, E J. (1985) Adv. Immunol. 37, 269. [4] Klinman, D. M. and Steinberg, A. D. 0987) J. Exp. Med. 165, 1755. [5] Datta, S. K., Stollar, B. D. and Schwartz, R. S. 0983) Proc. Natl. Acad. Sci. USA 80, 2723.
[6] Kubota, T., Akatsuka, T. and Kanai, Y. (1985) Clin. Exp. Immunol. 62, 321. [7] Kanai, Y., Akatsuka, T., Kubota, T., Goto, S. and Stollar, B. D. (1985) Clin. Exp. Immunol. 59, 139. [8] Kubota, T., Akatsuka, T. and Kanai, Y. (1986) Immunol. Lett. 14, 53. [9] Kanai, Y., Sugimura, T. and Matsushima, T. (1978) Nature 274, 809. [10] Kanai, Y., Yaoita, Y., Imaoka, K. and Honjo, T. (1988) Immunol. Lett. 19, 15. [11] Stollar, B. D. and Papalian, M. (1980) J. Clin. Invest. 66, 210. [12] Stollar, B. D. (1986) CRC Crit. Rev. Biochem. 20, 1.
299
Elsevier IMLET 01296
A novel trait of naturally occurring anti-DNA antibodies: dissociation from immune complexes in neutral 0.3-0.5 M NaC1 Y. K a n a i 1 a n d T. K u b o t a 2 1Department of Molecular Oncology, Institute of Medical Science, University of Tokyo, and 2First Department of Internal Medicine, Tokyo Medical and Dental University, Tokyo, Japan
(Received 28 April 1989; revisionreceived 7 July 1989; accepted 11 July 1989)
1. Summary Monoclonal and polyclonal anti-DNA antibodies from autoimmune mice, and experimentally induced rabbit anti-nucleic acid polyclonal antibodies were tested for stability o f binding to nucleic acids in the presence o f various concentrations o f NaC1 by an enzyme-linked immunosorbent assay (ELISA). Murine monoclonal antibodies 2C10 (IgG2b) and 1A2 (IgG2a), which are known to react specifically with double-stranded (ds) DNA, dissociated completely from their complexes with DNA when washed with a neutral 0.5 M NaCI solution. Another monoclonal antibody (MoAb) (IgM,r), polyreactive with singlestranded (ss) DNA, cardiolipin, and trinitrophenylhapten (TNP), was also dissociated from its complexes with ss DNA, but not from its complexes with TNP, by 0.3-0.5 M NaC1. Similar differences were observed in the binding stability of serum antibodies from autoimmune mice to DNA and TNP. In contrast, anti-nucleic acid polyclonai antibodies induced in rabbits by immunization with poly(I), poly(dT) or poly(ADP-ribose) were not significantly dissociated from their immune complexes with relevant antigens or DNA by 0.5 M NaC1. The finding that nucleic acid antigens were not detached from a solid phase by washing with 0.5 M NaC1 solution indicated that the reduction of binding of anti-DNA
Key words: Naturally occurring anti-DNA antibody; Immune complex; Affinity column Correspondence to: Dr. Y. Kanai, Department of MolecularOncology,Instituteof MedicalScience,Universityof Tokyo,Minatoku, Tokyo 108, Japan.
antibodies in both MoAbs and naturally occurring antibodies was really due to dissociation of the antibodies from immune complexes. This is the first demonstration that DNA epitopes recognized by naturally occurring antibodies in both SLE and its mouse models are sensitive to neutral NaC1 concentrations. This novel trait o f naturally occurring antibodies will be very useful in studies on the nature of immune complexes in sera and kidneys of cases of systemic lupus erythematosus (SLE).
2. Introduction There is extensive evidence that monoclonal antiDNA antibodies are not specific to a given antigen but are polyreactive; i.e., they react with epitopes of RNA phospholipids and of the cytoskeleton [1, 2]. Moreover, studies on a murine lupus model have implicated polyclonal B cell activation (PBA) as the basis for induction o f autoantibodies in SLE [3, 4]. The diversity of autoantibodies in SLE may be explained partly by the polyreactivity of a single immunoglobulin molecule or by PBA, and also partly by some electrostatic interactions between slightly positively charged antibodies and negatively charged antigens such as nucleic acids and acidic phospholipids. To investigate this possibility, we first examined the stability in neutral 0.2-0.5 M NaC1 solution of antigen-antibody complexes formed by oligo- and monospecific anti-DNA MoAbs (IgG2a and IgG2b) and polyreactive anti-DNA IgM,r MoAb. Unexpectedly, we found that the oligo- and monospecific IgG MoAbs were dissociated completely in 0.5 M NaCI solution, but that the polyreactive IgM MoAb was not. Using a series of auto-
0165-2478 / 89 / $ 3.50 © 1989 ElsevierSciencePublishers B.V.(BiomedicalDivision)
293
immune and experimentally induced anti-nucleic acid antibodies, we found that naturally occurring anti-DNA antibodies, in contrast to experimentally induced ones, could be dissociated more or less completely from the immune complexes by a neutral 0.5 M NaC1 solution. Until now, acid or alkaline elution has mainly been used to dissociate conventional immune complexes, including complexes of antiDNA antibodies with DNA [5]. We have also reported a method for elution of anti-DNA antibodies from a DNA affinity column [6]. However, so far there have been no reports on the dissociation of anti-DNA antibodies from immune complexes with neutral NaCI solution at about 0.5 M. The present study demonstrates a novel trait of naturally occurring anti-DNA antibodies and the application o f this trait for the isolation o f naturally occurring antiDNA antibodies from autoimmune sera. The mechanism and importance of this novel trait in systemic autoimmune disease are discussed. 3. Materials and Methods
3.1. Mice M R L / M p - l p r / l p r (MRL/1) and M R L / M p - +/+ (MRL/n) mice, originally purchased from the Jackson Laboratory, Bar Harbor, ME, U.S.A. are maintained in our laboratory. CBA/K1 mice, originally obtained from the Karolinska Institute, Stockholm, Sweden, have been maintained in this Institute since 1969. Mutant mice with general lymphadenopathy were recently detected by Matsuzawa et al. in our CBA/K1 mouse colony. After repeated sisterbrother mating of these mice, we examined autoantibodies such as anti-ss DNA and ds DNA antibodies in their sera. The results showed that all mutant mice o f both sexes produced both anti-ss DNA and antids DNA antibodies with high titers at about 5 months of age, and thereafter showed a switch from IgM to IgG. Details of findings in CBA/K1 mutant mice will be reported separately. 3.2. MoAbs Murine MoAbs were obtained as described previously [7]. The detailed characterization o f MoAbs 2C10 and 1A2, which react specifically with ds DNA, has been reported [8]. MoAbs 3E8 (IgM,K) 294
and 1D8 (IgM,K) were newly obtained from a CBA/K1 mutant mouse (9-month-old female): the former reacted with ss DNA, cardiolipin and TNP, and the latter also reacted with ds DNA and RNA. 3.3. Polyclonal antibodies Anti-poly(ADP-ribose) polyclonal antibodies (PoAbs) were raised in New Zealand white rabbits as described [9]. Anti-poly(I) and anti-poly(dT) PoAbs were obtained similarly: the former was specific for RNA and the latter specific for ss DNA (data not shown). Naturally occurring anti-DNA and antiT N P PoAbs were obtained from a 4-month-old female MRL/1 mouse and from a 9-month-old female CBA/K1 mutant mouse, respectively.
3.4. Determination of antibodies by ELISA Anti-nucleic acid antibodies were measured by ELISA as described [7]. Anti-TNP antibodies were also measured by ELISA. Briefly, 50 ~1 o f a solution containing 50 ng of TNP-bovine serum (BSA) (molar ratio of T N P to BSA was 1:55; a gift from Dr. Katagiri, Department of Immunology of this Institute) in 10 mM carbonate buffer, p H 9.8, was placed in wells of polystyrene microtiter plates (Immulon No. 2, Dynatech) and incubated at room temperature for 2 h. The plates were washed with Trisbuffered saline (TBS; 25 mM Tris/140 mM NaCI, p H 7.4) containing 0.1% Tween 20. The subsequent procedure was described for anti-nucleic acid antibodies [7]. 3.5. Examination of the stability
of antigenantibody complexes in neutral 0.3-0.5 M NaCI solution
Anti-DNA, anti-RNA, anti-poly(ADP-ribose) and anti-TNP antibodies were added to individual antigens in microtiter plates as described above. The wells o f the plates were washed separately four times each with 0.2-0.5 M NaCI in TB and then washed three times with TBS containing 0.05% Tween 20. Residual bound antibody was measured by conventional ELISA [7]. The A405 after washing with TBS without additional NaCI was considered to express 100% bound antibody.
3.6. Dissociation of antibodies from an affinity
Column by 0.5 M NaCI in TB ss DNA- or ds DNA-bound Sepharose was prepared as described [6]. Usually, antibodies bound to the column are eluted with 20 mM carbonate buffer containing 5070 dimethyl-sulfoxide (DMSO), p H 10.5 (carbonate-DMSO), but in this study, the affinity column containing bound anti-DNA antibodies was washed first with 0.5 M NaC1 in TB, and then with carbonate-DMSO. 3.7. Western immunoblotting for heavy and light
chains of mouse immunoglobulins Western immunoblotting was carried out using an SDS-polyacrylamide (12.5070) slab gel with electrophoretic transfer o f proteins to Millipore GVHP nitrocellulose sheets as described [10]. The amount of IgG loaded onto the gel was more than 4 t~g per lane.
B
__A
100 1 ~ _ ~
50 ~
0
00
50
i f i J0 014 0.2 0.3 04 0.5 0.1402 0.3 0.4 0.5 0,4 0!2 013 014 015 NaCI (M3
Fig. 1. Dissociationof immunecomplexesofmurine MoAbsand DNA (A, B and C), but not of those of MoAb 3E8 and TNP (C) by neutral 0.3 -0.5 M NaCI. 100O7obound means about 1 A4o5U antibody activity obtained under conventional conditions. See text for details. A, MoAb 1A2; B, MoAb 2C10; C, MoAb 3E8. o, ss DNA binding; e, ds DNA binding; A, TNP binding.
reacted with ss D N A and TNP. As shown in Fig. 1C, neutral NaC1 solution also dissociated this MoAb from ss DNA, but did not dissociate it from TNP.
4. Results 4.2. Dissociation of immune complexes of natural-
4.1. Dissociation of immune complexes formed between murine MoAbs and DNA by neutral 0.3- 0.5 M NaCI Murine MoAbs 2C10 (IgG2b) and 1A2 (IgG2a) are known to bind with both ds D N A and ss DNA, but the former prefers ds D N A over ss D N A [8]. After primary interaction of these MoAbs with ds D N A or ss DNA, immune complexes formed on the E L I S A plates were washed individually with solutions o f 0 . 2 - 0 . 5 M NaCI in TB, as described in Materials and Methods. As shown in Fig. 1A and B, antibodies bound to D N A were dissociated as a function o f the neutral NaCI concentration used for washing, and complete dissociation was achieved with 0.5 M NaC1. MoAb bound to ss D N A plates was dissociated at a lower NaCI concentration than that bound to ds D N A plates, but there was no significant difference between the NaC1 concentrations required for the dissociations o f 2C10 and 1A2 MoAbs. To determine whether this dissociation from immune complexes by neutral 0 . 3 - 0 . 5 M NaCI was a c o m m o n characteristic of MoAbs, we also examined another MoAb 3E8 [IgM,x] which was derived from a CBA/K1 mutant mouse and
ly occurring anti-DNA PoAbs and DNA by neutral 0.3 - 0.5 M NaCl solution Similar experiments with serum containing high titers of IgG class anti-ss D N A and anti-ds D N A antibodies selected from those of a number of autoimmune MRL/1 mice were carried out next. As this serum had very low titers of anti-TNP antibodies, we also used serum containing IgM antibodies with high T N P reactivity selected from those of a number o f CBA/K1 mutant mice. As shown in Fig. 2A, no significant difference was observed between the dissociations of anti-ss D N A and anti-ds D N A antibodies from DNA, which reached levels of up to 70°7o dissociation. In contrast, IgM PoAbs, like MoAb 3E8, was not dissociated from T N P by neutral solutions of up to 0.5 M NaC1, as shown in Fig. 2B. 4.3. Effect of neutral NaCI solution on immune
complexes formed between experimentallyinduced anti-nucleic acid lgG antibodies and nucleic acids In contrast to naturally occurring PoAbs and 295
A
B 100
100
100
100
50
50
•
e
50
50
o m , 0. 4 0.2 03 0.4 05
0
,,
0
I
0
~
--" 100
0.14 012 01.3 014 0'5 c
i
0.14 0.2 03 04 0.5
100
i
e
012 03 014 015
E ~
NaCI (M) Fig. 2. Dissociation of immune complexes of anti-DNA IgG PoAb from an MRL/1 mouse and DNA (A), but not o f those of anti-TNP IgM PoAb from a CBA/K1 mutant mouse and TNP (B) by neutral 0.3 - 0 . 5 M NaCI. The serum dilution was 1:100; 100% bound means about 1 A405 U antibody activity obtained under conventional conditions. See text for details, o , ss DNA binding; o , ds DNA binding; A, TNP binding.
50
50
0 0.;4 012 013 014 015
0 off4 012 013 014 015
NaCI (M)
MoAbs from MRL/1 and CBA/K1 mutant mice, experimentally induced rabbit IgG PoAbs to poly(I), poly(dT) and poly(ADP-ribose) were not readily dissociated from immune complexes by neutral solutions o f up to 0.5 M NaCI, as shown in Fig. 3, although higher NaC1 concentrations caused 30% dissociation of anti-poly(dT) antibody from calf thymus ss DNA, as shown in Fig. 3C.
Fig. 3. Effects of neutral 0.3-0.5 M NaC1 solution on the immune complexes formed between experimentally induced antinucleic acid rabbit IgG PoAbs and nucleic acids. 100% bound is defined as for Figs. 1 and 2. Alkaline phosphatase-conjugated anti-rabbit IgG antibodies (Sigma) were used in place of alkaline phosphatase-conjugated anti-mouse isotype specific antibodies. A, anti-poly(I) rabbit-antibodies; B, anti-~91y(dT) rabbit antibodies; C, anti-poly(dT) antibodies bound with calf thymus ss DNA; D, anti-poly(ADP-ribose) rabbR.~emtibodies.
1 /It
4.4. Elution of anti-DNA antibodies from a DNA affinity column with neutral 0.5 M NaCI solution and their demonstration by both ELISA and immunoblotting A sample o f 200 #1 of MRL/1 mouse serum containing anti-ss and ds DNA antibodies was applied to a DNA affinity column and the column was washed extensively with TBS alone and then developed with 0.5 M NaC1 in TB. Fractions o f the eluate were found to contain IgG anti-ss and ds DNA antibodies, as shown in Fig. 4A. Although the protein peak eluted with carbonate-DMSO had DNAbinding activity, the relative amount of the fraction was less than 15°70 judging from A280 U recovered. As a positive control, the culture supernatant of the 1A2 or 2C10 hybridoma was treated similarly; the anti-DNA antibodies eluted f r o m t h e column are 296
'B/
7
.< 0.5 Z
£3
ol
o I
i
I
i
I
0.01 0.02 0.05 0.1 0.2 Protein,
8==8===8==8==~ i
r
i
e
I
0.01 0,02 0,05 0.1 0.2 A2ao U/ml
Fig. 4. Dose-dependent binding to DNA of MRL/1 mouse PoAb (A) and MoAb 1A2 (B) eluted from a DNA affinity column with neutral 0.5 M NaC1. o , ss DNA binding; o , ds DNA binding. , IgG antibodies; - - -, IgM antibodies.
shown in Fig. 4B. Confirmation that the fractions from the DNA affinity column eluted with 0.5 M NaC1 really contained IgG and/or IgM was obtained by Western immunoblotting, as described in Materials and Methods. As shown in Fig. 5, anti-IgG (7-chain-specific) and anti-IgM (/~-chain-specific) reactive bands corresponding to sizes of 50 and 72 kDa, respectively, were immunostained. AntiIgG whole-molecule antibodies stained the bands o f the light chains of both IgM and IgG, corresponding to a molecular weight o f 22 kDa. Protein staining demonstrated the existence o f additional bands that reacted with DNA in both MRL/1 mouse serum and the supernatants o f hybridoma cultures. Anti-/~chain-specific antibodies also detected a band of 50 kDa other than the 72-kDa band o f the heavy chain. This was detected with murine IgM MoAb, but not with a myeloma culture supernatant such as Sp2/0-Ag-14 (data not shown). Therefore, the 50-kDa band is probably derived from the heavy chain o f the IgM during denaturation. This 50-kDa band disappeared when the amount of Ig loaded onto the gel was less than 1/zg, while the 72-kDa
Mabcd
B Fig. 5. Western immunoblotting of MoAb 1A2 (A) and an MRL/I mouse PoAb (B) eluted from the DNA affinity column. Lane a, protein staining with Coomassie Blue; lanes b, c and d, immunoblotting with anti-mouse IgG whole molecule (Sigma), anti-mouse it-chain-specificgoat antibodies (Zymed) and antimouse v-chain-specificrabbit antibodies (Pel-Freez),respectively. L, light chain; it, # chain; 3', 3' chain.
band of the heavy chain was clearly visible (data not shown). 5. Discussion
The autoantibody diversity found in SLE may be due to polyreactivity o f a single molecule or socalled PBA. Polyreactivity o f MoAbs has been explained by the shared structures of macromolecules such as DNA and phospholipids, in which similar spacing o f phosphodiester groups occur [1]. We are considering another possibility, namely that it is due to the interactions between slightly positively charged antibodies and negatively charged molecules such as nucleic acids and some phospholipids. We have examined this possibility using an ELISA system [7]. We were surprised to find in this study that MoAb 2C10, which is considered to be a prototype anti-ds DNA antibody [8], was dissociated with 0.3 M NaCI from DNA coated on a solid phase (Fig. 1). In fact, 2C10 was recently proved to be a prototype of anti-DNA antibodies, by gel retardation experiments using a number of synthetic doublestranded oligonucleotides with special base sequences (B. D. Stollar et al., in preparation). This finding raised two questions: firstly, are nucleic acid antigens non-covalently attached to the solid phase stable against washing with 0 . 3 - 0.5 M NaCI? Secondly, if so, are they still stable at these concentrations o f NaC1 after formation of immune complexes? To answer the first question, we washed ds DNA- or ss DNA-coated polystyrene plates with neutral 0.5 M NaCI before carrying out ELISA, but obtained the same results as without pre-washing the plates with 0.5 M NaC1 (data not shown). To answer the second question, we reused plates from which bound antibodies had been dissociated with 0.5 M NaC1, and found that antibody binding was unchanged as compared to fresh plates coated with ds DNA or ss DNA (data not shown). These data indicated that DNA antigen attached to polystyrene plates through poly-0.-lysine) was stable in solutions of up to 0.5 M NaCI. Naturally occurring anti-DNA antibodies from MRL/1 mice were also found to be dissociated by NaCI, although about 30°7o residual binding was observed after washing with 0.5 M NaCI (Fig. 2). This 30°7o residual binding o f polyclonal anti-DNA antibodies of the MRL/1 mice may reflect their poly297
reactivity with a number of related compounds, because MoAb 1D8, which was derived from a CBA/K! mutant mouse, was polyreactive, reacting with ds DNA, ss DNA, cardiolipin, RNA and T N P (data not shown), and it was not dissociated from DNA with neutral 0.5 M NaCl (data not shown). Whether or not 3007o residual binding o f polyclonal anti-DNA antibodies of MRL/1 mice really reflects polyreactivity should await further experiments using many samples obtained by 0.5 M NaCl elution and by alkaline elution with carbonate-DMSO. It should be mentioned here that cross-reaction of 2C10 and 1A2 MoAbs with ds DNA and ss DNA could be due to $1 nuclease-resistant double helical cores in the ss DNA molecules, as described by Stollar and Papalian [11]. However, we found that MoAbs 2C10 and 1A2 dissociated from their immune complexes with ss DNA more readily than from their immune complexes with ds D N A at higher concentrations of NaCl (Fig. 2). This may be due to the structural difference in the double helices of ss DNA and ds DNA: the double helices of the former are more susceptible to higher NaCl concentrations than those of the latter. This difference in susceptibility to NaCl concentration was not clear with naturally occurring anti-DNA antibodies in the sera of an MRL/1 mouse, but was observed with anti-ss DNA antibodies of patients with SLE (data not shown). The conformational structures rather than the double helices common to both ss DNA and ds DNA must be involved in the recognition by naturally occurring anti-ss DNA antibodies, because MoAb 3E8, which binds with ss DNA but not with ds DNA, was also susceptible to higher concentrations of NaC1 (Fig. 1C). In contrast, experimentally induced antibodies to nucleic acids were rather stable to higher concentrations of NaCl: immune complexes of anti-poly(I) and anti-poly(dT) with relevant antigens were quite stable (Fig. 3A and B) but 3007o dissociation of antipoly(dT) antibodies from calf thymus ss DNA was observed in 0.5 M NaCl solution, suggesting that some epitopes recognized by anti-poly(dT) PoAb were altered by exposure to higher concentrations of NaCl, even after formation of immune complexes (Fig. 3C). In normal animals, antibodies are produced against relatively simple antigenic determinants such as haptens and/or primary structures, 298
whereas in SLE these antigenic determinants are more complex helical structures [12]. Therefore, it is possible that the anti-nucleic acid antibodies induced experimentally bind firmly to simple epitopes, whereas those found in SLE are very sensitive to minor changes in nucleic acid structures induced by higher concentration of NaC1. The changes in DNA structures induced by higher concentrations of NaC1 were reversible because after removal of bound antibodies with neutral 0.5 M NaC1 and washing with normal saline, the plates could again bind MoAbs 1A2 and 2C10 (data not shown). After application o f MoAb 1A2 to a DNA affinity column, we demonstrated elution of antibodies with neutral 0.5 M NaC1 solution: the fractions of eluate contained 1A2 MoAb that reacted with DNA dose-dependently (Fig. 4B). We also demonstrated elution of 2C10 MoAb from a DNA affinity column (data not shown). Anti-DNA antibodies in the serum of an MRL/1 mouse could also be eluted from a DNA affinity column with 0.5 M NaC1, but IgM detected by Western immunoblotting did not bind to DNA, as shown by ELISA (Fig. 4A), indicating non-specific binding of IgM to DNA. The band corresponding to the IgG heavy chain, seen by immunoblotting with anti-#-chain antibodies (Fig. 5B, lane C) was not seen, when a s m i l e r amount of IgM was loaded on the gel or when IgG MoAb 1A2 was applied to the gel (Fig. 5A, lane C), indicating that a cleavage product of IgM was formed under the reducing conditions used for electrophoresis. In any case, experiments using a DNA affinity column confirmed that anti-DNA antibodies could be recovered by elution with 0.5 M NaCl and that reduction in the binding of naturally occurring antiDNA antibodies on the solid phase was actually due to dissociation of antibodies from the immune complexes. The naturally occurring anti-DNA antibodies in patients with SLE were also dissociated from complexes by 0.5 M NaC1. Detailed studies on the nature of the dissociated anti-DNA antibodies of SLE patients will be reported separately. This common trait of naturally occurring anti-DNA antibodies in lupus model mice and human SLE will be useful in studies on the nature of the immune complexes in the sera and the kidney of SLE patients.
References [1] Lafer, E. M., Rauch, J., Andrzejewski Jr., C., Mudd, D., Furie, B., Schwartz, R. S. and Stollar, B. D. (1981) J. Exp. Med. 153, 897. [2] Andr6-Schwartz, J., Datta, S. K., Shoenfeld, Y., Isenberg, D. A., Stollar, B. D. and Schwartz, R. S. (1984) Clin. Immunol. Immunopathol. 13, 261. [3] Theofilopoulos, A. N. and Dixon, E J. (1985) Adv. Immunol. 37, 269. [4] Klinman, D. M. and Steinberg, A. D. 0987) J. Exp. Med. 165, 1755. [5] Datta, S. K., Stollar, B. D. and Schwartz, R. S. 0983) Proc. Natl. Acad. Sci. USA 80, 2723.
[6] Kubota, T., Akatsuka, T. and Kanai, Y. (1985) Clin. Exp. Immunol. 62, 321. [7] Kanai, Y., Akatsuka, T., Kubota, T., Goto, S. and Stollar, B. D. (1985) Clin. Exp. Immunol. 59, 139. [8] Kubota, T., Akatsuka, T. and Kanai, Y. (1986) Immunol. Lett. 14, 53. [9] Kanai, Y., Sugimura, T. and Matsushima, T. (1978) Nature 274, 809. [10] Kanai, Y., Yaoita, Y., Imaoka, K. and Honjo, T. (1988) Immunol. Lett. 19, 15. [11] Stollar, B. D. and Papalian, M. (1980) J. Clin. Invest. 66, 210. [12] Stollar, B. D. (1986) CRC Crit. Rev. Biochem. 20, 1.
299