Pathogenic anti-DNA idiotype-reactive IgG in intravenous ...

Clin Exp Immunol 1994; 97:19-25

Pathogenic anti-DNA idiotype-reactive IgG in intravenousimmunoglobulin preparations

F. SILVESTRIS, P. CAFFORIO & F. DAMMACCO Department of Biomedical Sciences and Human Oncology,University ofBari Medical School, Section of Internal Medicine and Human Oncology, Bari, Italy

(Acceptedfor publication 22 March 1994)

SUMMARY

This study was addressed to explore the reactivity of natural anti-idiotypes from commerciallots of immunoglobulins to several idiotypes (Ids), usually expressed by anti-DNA molecules inlupus nephritis. Eleven intravenous immunoglobulin (IVIG) preparations and nine (threepolyvalent and six hyper-immune) intramuscular IgG were investigated for specific content ofanti-DNA, anti-F(ab')2 and antibodies reacting with several anti-DNA IgG Ids. Two samples(nos 6 and 11) showed high reactivity with allogeneic F(ab')2 and with F(ab')2 of myelomaproteins bearing the anti-DNA Id 3I+ and the 8.12+. Since both 31 and 8.12 Id markers are

known to characterize pathogenic anti-DNA IgG in systemic lupus erythematosus (SLE), anti-Id antibodies to these markers were obtained by absorbing the IVIG samples nos 6 and 11 toSepharose columns coupled with pooled F(ab')2 fragments of 3I+-F4+-8.12+-myeloma proteins.Inhibition experiments showed that anti-8.12 Id-eluted IgG induced a selective suppression ofthe DNA-reactive antibodies derived from patients with active lupus nephritis to their sub-strate, suggesting the involvement of 8.12+ molecules in the SLE glomerular damage. Since8.12+ anti-DNA are nephritogenic antibodies, the occurrence of anti-8.12+ Id in commercialIVIG may be of potential therapeutic relevance in modulating the pathogenic SLE Id network.Previous variable results of IVIG treatment in SLE, such as resolution of proteinuria or

worsening nephritis, could be related to variable enrichment of different lots of IVIG insuppressive anti-pathogenic Id antibodies.

Keywords autoimmunity idiotypic networksystemic lupus erythematosus

INTRODUCTION

The beneficial effect of high dose intravenous immuno-globulins (IVIG) has been reported in the therapy of auto-immune peripheral cytopenias [1-3], as well as in otherdisorders asso-ciated with pathogenic autoantibodies, includ-ing myasthenia gravis, chronic inflammatory demyelinatingpolyneuropathy, the anti-cardiolipin syndrome and anti-factor VIII auto-immune disease [4-6]. Systemic lupus erythe-matosus (SLE) has also been treated by IVIG, althoughcontroversial effects have been reported, ranging from resolu-tion of proteinuria [7] to actual exacerbation of glomerulone-phritis [8]. The immunomodulation on autoantibodymolecules has been postulated as a specific effect of thistreatment [9,10].

Modulation of pathogenic idiotypes (Ids) in SLE by IVIG

Correspondence: Franco Dammacco MD, DIMO, Sez. MedicinaInterna e Oncologia Clinica, Universiti di Ban, Piazza Giulio Cesare,11, 70124 Ban, Italy.

immunomodulation anti-F(ab')2 antibodies

implies the expression in healthy donors of anti-Id antibodieswith potential regulatory function affecting the autoimmunenetwork via V region-dependent mechanisms [11]. Differentanti-DNA markers have been shown to occur in SLE [12].Several of them, including as many as 18 common Ids [13], mayalso be detectable in a lower proportion of further autoimmuneconditions [14] and on normal IgG from healthy subjects [15].However, because of the anionic sites of the human glomerularbasement membrane [16], nephritogenic properties of anti-DNA molecules are usually paralleled by their complementaryisoelectric cationic charge [17,18]. Thus, production of anti-Idsinhibiting the potentially harmful autoimmune repertoire mayarise as a normal result of the natural regulatory networkdirected at autoantibodies secreted in normals by CD5 Bcells [19,20].

Other investigators have supported the efficacy of IVIG inmodulating autoimmune states because of their ability toimpose normal function in a defective autoimmune networkby regulatory anti-Id molecules [21,22]. IVIG could modulatethe autoimmune phenomena by restoring or stimulating the

19

F. Silvestris, P. Cafforio & F. Dammacco

physiological control mechanisms of autoimmunity, ratherthan by merely suppressing pathogenic Ids. In addition,activation by IVIG of the regulatory network would takeadvantage of the Id variability of the IVIG pool derived froma large number of normal donors. Finally, antibodies to theCD5 molecule have been detected in IVIG [23], thus postulat-ing their therapeutic potential through both modulation of Tcell function and regulation of B cell subsets in autoimmunity,via the CD5 receptor.

Recent work by our group indicates that significant increaseof serum anti-F(ab')2 antibodies in inactive SLE patients andtheir relatives may reflect a general activation of the anti-Idcontrol mechanism(s) on autoantibody production [24,25].Moreover, SLE-derived anti-F(ab')2 are capable of suppres-

sing production of autologous anti-DNA antibody by patients'B cells in vitro [26]. The content of anti-F(ab')2 antibodiesoccurring in some IVIG preparations would probably reflecttheir potential capability to modulate the expression of patho-genic anti-DNA Ids in SLE patients receiving such therapy.

Data presented in this study provide evidence that the levelsof anti-F(ab')2 antibodies are quite variable in different batchesof IVIG preparations, and that specific reactivity to pathogenicanti-DNA Ids can be found in a minority of instances. Inaddition, the occurrence of anti-idiotypes reactive to cationicanti-DNA Id within the IVIG commercial pools will beemphasized.

MATERIALS AND METHODS

Commercial immunoglobulin preparationsIVIG lots included 11 preparations obtained from differentdrug manufacturers. In addition, three polyvalent and sixhyperimmune intramuscular IgG (IMIG) were also studied.Details of these preparations are listed in Table 1.

Antigens and ELISA testsA panel of sequential ELISA assays was employed in order toevaluate the reactivities of each immunoglobulin preparation tovarious substrates. The preliminary ELISA included analysis ofboth anti-DNA and anti-F(ab')2 levels by using previouslydescribed methods [24,25]. Briefly, microtitre plates were sepa-

rately coated with 1 ,g/ml of calf thymus dsDNA and 5 pg/ml ofpurified F(ab')2 fragment obtained from pepsin-digested humanIgG (Cohn fraction II). Both IVIG and IMIG preparations were

diluted to identical concentrations (300 jig/ml) in PBS andincubated as described [26]. Detection of DNA- and F(ab')2-reacting IgG was completed using goat peroxidase-conjugatedFc-y-specific antiserum, diluted 1: 50000 (Jackson Immunore-search Labs., West Grove, PA). Finally, the o-phenylendiamine(OPD) solution was used for colour development, and therelative absorbance of each sample was read at OD 492 in a

micro-ELISA reader (Titertek, Flow Labs, Irvine, UK). Eachassay included two IgGk myeloma proteins as negative con-

Table 1. Properties and commercial synonyms of 11 intravenous immunoglobulins (IVIG), three conventional and six hyperimmune intramuscularIgG (IMIG) preparations

Sample Registered name and manufacturer Molecular condition and treatment*

Polyvalent IVIG1 Gammavenin P Pepsin-digested IgG(Behring Inst., Scoppito, L'Aquila, Italy)

2 Endobulin (Immuno, Pisa, Italy) Polyethylenglycol-stabilized IgG3 Venogamma Polivalente Polyethylenglycol-stabilized IgG

(Alfa Wassermann, Milan, Italy)4 Isiven (ISI, Naples, Italy) Whole IgG, pH 4 by acidic treatment5 Intraglobin F o3-propiolactone treatment

(Biotest, Trezzano sul Naviglio, Milan, Italy)6 Biaven (Farma Biagini, Pisa, Italy) Whole IgG, pH 4 by acidic treatment7 Venimmun (Behring Inst.) Whole IgG, sulphitolysis8 Venoglobulina Plasmin-digested IgG(Mrieux Italia, Rome, Italy)

9 IG Vena N (Sclavo, Siena, Italy) Dithiothreitol-iodoacetamide treatment10 Sandoglobulina (Sandoz, Milan, Italy) Whole IgG, pH 4 by acidic treatment11 Globuman (Berna, Como, Italy) Polyethylenglycol-stabilized IgG

Polyvalent IMIG12 Uman-Gamma (Farma Biagini) Stabilized whole IgG13 IG-Gamma (Sclavo) Stabilized whole IgG14 Gammabulin (Immuno) Stabilized whole IgG

Hyperimmune IMIGAnti-hepatitis B 15 Uman-Big (Farma Biagini) Stabilized whole IgGAnti-varicella zoster 16 Uman-Vzig (Farma Biagini) Stabilized whole IgGAnti-German measles 17 Rosol-Gamma (Farma Biagini) Stabilized whole IgGAnti-mumps 18 IG-Parotite (Sclavo) Stabilized whole IgGAnti-cytomegalovirus 19 Uman-Cig (Farna Biagini) Stabilized whole IgGAnti-tetanus toxoid 20 Tetuman (Berna) Stabilized whole IgG

* Communicated by the manufacturers.

20

IgG to pathogenic anti-DNA idiotypes in IVIG

trols. Each determination was performed in triplicate and theabsorbance values were expressed as mean + s.d. Samplesshowing mean OD values higher than the anti-F(ab')2 con-fidence interval (M + 2 s.d.) from all IVIG and IMIG lots wereconsidered as high-titered preparations.

Further analysis focused on the reactivity of selected IVIGpreparations with purified F(ab')2 fragments expressing recur-rent anti-DNA idiotypes. To this purpose, we used purified IgGmyeloma proteins which had previously been typed usingmonoclonal anti-Id reagents [27] as showing three Id markersassociated with IgG anti-DNA antibody molecules, namely theF4 [28], 31 [29] and the 8.12 [30] Ids. The myeloma protein FU-647-40 was 3I+, the J-10-42 was F4+, whereas the DC-305-39was found to express both 8.12 and F4 markers.

Purified F(ab')2 fragments from FU-647-40, J-10-42 andDC-305-39 myeloma proteins were obtained as reported [27]and used as antigens in subsequent ELISA tests.

Immunoabsorption experimentsAnti-DNA Id-reactive molecules from the IVIG preparationswere isolated by immunoabsorption. Thus, 70mg of each IVIGlot were first absorbed with DNA-Sepharose to remove anypossible anti-DNA activity [24], and then passed over a columnof DNA-absorbed IgG (derived from Cohn fraction II) byusing a recycling pump in order to deplete both anti-F(ab')2and anti-Fc reactivities. Moreover, these sequentially absorbedIgG preparations were further passed over the anti-DNAaffinity column containing an idiotypic pool obtained bycoupling 20 mg of each Id+ myeloma protein to Sepharose.

Finally, eluates from this last column of anti-DNA-Id+ mye-loma proteins were obtained by acid elution using a pH 2 2,0 I M glycine-HCl buffer and promptly adjusted to neutral pH.The protein content of the eluates was assessed by standardSDS-PAGE electrophoresis.

Inhibition experimentsThe eluates from the final anti-DNA-Id+ myeloma proteincolumn were tested for their ability to inhibit the anti-DNAbinding properties of sera from active SLE patients. Briefly,solutions containing 70 Ag/ml of affinity-isolated anti-DNA anti-bodies from seven SLE patients were separately incubated over-night at 40C with serial amounts (5, 10, 20, 40, and 80 pg/ml) ofeach eluate in the final volume of 2-35 ml of PBS. Residual anti-DNA affinity of these preparations was subsequently defined byELISA as previously described [24]. Parallel control tests wereinstalled to evaluate the inhibiting properties of both eluates onpurified anti-tetanus toxoid IMIG from preparation no. 20.

Statistical analysisANOVA in each ELISA was assessed by the Wilcox-Wilcoxon(multiple comparison) test.

RESULTS

Detection of anti-DNA and anti-F(ab')2 activities in IVIG andIMIGFigure 1 shows data obtained in a preliminary ELISA test. Ageneral pattern of low reactivities to both DNA and F(ab')2

ECMcv

a

0

0 4

0*3[

*4

- - t--- - - - - - - - - - - - ---

Sample no. 2 3 4 5 6 7 8 9 10 11IVIG

.

12 13 14 15 16 17 18 1920IMIG

Anti-F(abi)2 M + 2 s.d.

A-71 A-74Control IgGK

myeloma proteinsCommercial immunoglobulin preparations

Fig. 1. Levels of anti-dsDNA and anti-F(ab')2 IgG in 20 commercial immunoglobulin preparations, detected by ELISA. Two IgGkmyeloma proteins, i.e. A-71 and A-74, were also tested as negative controls. Values are expressed as mean of the absorbance ± s.d. Thebroken line shows the upper limit of the confidence interval of all intravenous immunoglobulin (IVIG) and intramuscular IgG (IMIG)anti-F(ab')2 reactivities. Substrates: 0, dsDNA; *, allogeneic F(ab')2.

21

40

6 6.t

F. Silvestris, P. Caforio & F. Dammacco

Table 2. Immunochemical properties of three myeloma proteins withanti-DNA markers used in ELISA

Anti-DNAMyelomas Isotype ic/A pI marker*

1. FU-647-40 IgG X 8-6 3I+2. J-10-42 IgG K 8-6 F4+3. DC-305-39 IgG A 8-5 F4+-8.12+

* Detected by ELISA using specific monoclonal anti-Ids [271.

0-6k

05k

E

cJ

0)

0-2k

0-1kantigens was revealed in the various IVIG samples. However,much lower values of the anti-DNA binding compared with theanti-F(ab')2 reactivity were found in most instances. In theIMIG preparations (samples 12-20) the pattern of reactivitiesto both substrates appeared lower than in the IVIG (samples1-11). A striking result of this assay concerned the high anti-F(ab')2 titres revealed in selected IVIG samples. IVIG prepara-

tion no. 3 showed levels of anti-F(ab')2 near the upper limit,whereas the values of IVIG nos 6 and 11 were higher than theconfidence interval detected in all IVIG and IMIG samples. Inaddition, each anti-F(ab')2 reactivity was significantly elevatedcompared with the relative low affinity to DNA (P < 0 001 ineach instance).

Detection of anti-DNA Id(s)-reactive IgG in IVIGThe IVIG samples nos 3, 6 and 11 with high anti-F(ab')2 titreswere next studied for their affinity to a panel of anti-DNA-associated Ids, including 31, F4 and 8.12 markers. The antigensin this ELISA included a panel of IgG myeloma proteinsdisplaying the mentioned anti-DNA-related Ids (Table 2).

Figure 2 shows that a variable degree of affinity to singleantigens could be detected in the three IVIG preparations.Indeed, sample no. 3 was found to react almost equally with theF(ab')2 fragment from both 3I+ and F4+-myeloma proteinsand, furthermore, with the allogeneic F(ab')2. In addition, thereactivity with the F4+-8.12+-F(ab')2 was similar, althoughslightly lower than the others (P > 0 1). By contrast, IVIGsample no. 6 appeared to react differently with several sub-strates. The highest activity of this preparation was directed tothe allogeneic F(ab')2. However, the titres of anti-3I+-F(ab')2reactivity were higher (P < 0 01) than both anti-F4+ and anti-8.12+-F4+-F(ab')2 levels (Fig. 2). IgG molecules from sampleno. 11 displayed low reactivity to the 3I+ and F4+-F(ab')2preparations, whereas their binding to the F4+-8.12+-F(ab')2was significantly higher (P < 0 01). In addition, antibody levelsto allogeneic F(ab')2 were similarly high in this batch of

No. 3 (Venogammo P) No. 6 (Blaven) No II (Globuman) A-71

Commercial IVIG samples

A-74

IgGK myeloma proteins

Fig. 2. ELISA analysis of the F(ab')2 reactivity of intravenous immu-noglobulin (IVIG) nos 3, 6 and 11 to different F(ab')2 fragmentsexposing 31, F4 and 8.12 anti-DNA idiotypic markers. A specificreactivity to the 3I+-F(ab')2 was found in IVIG no. 6, whereas IVIGno. 11 appeared more specific to the F4+-8. 122+-F(ab')2. This specificitywas presumably directed to the 8.12 marker, since the affinity to theF4+-F(ab')2 was significantly lower (P < 0 01). Control myelomas A-71 and A-74 were unreactive to the substrates used. Results are opticaldensity (M ± s.d.). Lj, 3I+-F(ab')2; F], F4+-F(ab')2; *, F4+-8.12-F(ab')2; El, allogeneic F(ab')2.

commercial immunoglobulin. Conversely, we failed to detectanti-F(ab')2 reactivity in the two unrelated IgGk monoclonalproteins, namely A-71 and A-74 myelomas, used as controls inthe ELISA.

Purification of anti-Id(s) from absorbed IVIGThe commercial preparations nos 6 and 11 were furtherabsorbed on dsDNA and allogeneic Cohn fraction II IgGcolumns in order to deplete reactivities to DNA, F(ab')2 andFc. Table 3 shows the residual antibody activity of thesamples to the various substrates, indicating removal of anti-F(ab')2 reactivities in these absorbed IgG preparations. Sub-sequent immunoabsorption of each sample was then per-formed on a 3I+-F4+-8. 12+-F(ab')2-coupled Sepharosecolumn. Acid elution of these columns provided two eluates:El-6 and El-l1, respectively. Acid elution of the original70mg from IVIG samples nos 6 and 11 produced a differentcontent of anti-Id: 2-8 mg of IgG were present in El-6, and575mg in El-li.

Table 3. Absorbance values (mean + s.d.) detected by ELISA methods before and after absorption ofintravenous immunoglobulin (IVIG) nos 6 and 11 with DNA and allogeneic IgG columns

IVIG no. 6 IVIG no. 11

Reactivity to Before After Before After

dsDNA 0-187±214 0-129± 121 0131 ±27-3 0082± 10-4F(ab')2 0-494 ± 76 3 0 174 ± 11 0 0 409 ± 81 3 0-107 ± 9 2

22

0 4t-

0.3:

IgG to pathogenic anti-DNA idiotypes in IVIG

DISCUSSIONEl-lI (8.12+ -Id-reoctive 1g9)

By using a number of ELISA tests we found that levels of anti-DNA antibodies are low in most IVIG samples, whereassignificant amounts of F(ab')2-reacting IgG were detected inseveral IVIG preparations. High levels of anti-F(ab')2 anti-bodies were also paralleled by independent reactivities to 31, F4and 8.12 anti-DNA markers. Furthermore, IgG eluted from aspecific anti-DNA-Ids affinity column were capable of inhibit-ing the reactivity of purified anti-DNA antibodies from SLEpatients with lupus nephritis. Since the original IVIG sampleswere depleted of any detectable anti-DNA, anti-Fc and anti-normal F(ab')2 reactivities, the anti-DNA inhibitory effect isprobably mediated by true Id/anti-Id binding.

Previous studies indicate that several IVIG preparationsinclude anti-Id molecules to recurrent anti-DNA Ids [9,31]because of their relatively frequent occurrence within normaldonors [11] and SLE healthy kindreds [15,32]. In addition, theseanti-Id molecules from IVIG have demonstrated a specificinhibitory effect in vitro to the anti-DNA secreting cells fromSLE patients [33]. In this context, antibodies to cellular CD5molecules recently identified in commercial lots of IVIG [23]would cooperate to modulate the hyperactivity ofCD5' T cellsand B cells committed to the natural autoreactivity [19]. In ourexperiments, and in line with previous observations [31], wefound a low expression of natural anti-DNA molecules in allIVIG samples, whereas undetectable levels were demonstratedin the hyperimmune IgG samples.

Analysis of the Id specificities of the three normal F(ab')2-reactive IVIG samples revealed a high affinity of IVIG samplesnos 6 and 11 to (FU-647-40)-3I1 and to (DC-305-39)-F4+-8.12+ myeloma proteins, respectively. However, since thereactivity of IVIG no. 11 to F4+-8.12+-(DC-305-39) myelomasample was significantly higher than that to F4+-(J-10-42)monoclonal protein (Fig. 2), it appeared that the prevalentspecificity of affinity-purified IgG from IVIG no. 11 wasprobably directed to the 8.12 anti-DNA marker.

The 31+, F4+ and 8.12+ myeloma proteins used as antigensin the present study were previously shown [27] to possess ahigh isoelectric point (pI> 8-0) because of their cationicmigration in isoelectric focusing analysis. Recent investiga-tions [18,34] support previous work [35], suggesting thatcationized anti-DNA molecules, including the IdX [36], IdGN1 and Id GN2 [37] and 8.12 Ids, are nephritogenic becauseof their complementary electric charge to the anionic glomer-ular basement membrane. Conversely, the 31 Id is not specifi-cally related to pathogenic anti-DNA molecules since it hasbeen found to occur in 85% of SLE healthy family members[38]. The distribution of these Ids within the normal populationmay account for the presence of their corresponding anti-Ids insome IVIG preparations. The occurrence of anti-cationic IgGreactivities in both normal subjects and SLE patients has beenrecently reported by ourselves [27]. However, IgG absorbed on3I+ and 8.12+ myeloma proteins, namely El-6 and El-ll, weretrue anti-Id molecules, in that the binding of anti-DNAantibodies from four patients with SLE nephritis to theirsubstrate was highly inhibited (Fig. 3).

The inhibition of anti-DNA by El-Il, namely 8.12 Idreactive IgG, suggests the expression of 8.12 by such autoanti-bodies. Conversely, the lack of inhibition by El-6 may reflect alow expression of 31 Id. Although both 31 and 8.12 Ids are

70wg/ml 5 10 20 40 80

Inhibitor concentration (pg/mr)

Fig. 3. Inhibition experiments of purified anti-DNA antibodies fromseven systemic lupus erythematosus (SLE) patients with (0) andwithout (0) lupus nephritis, by increasing amounts of E1-6 and El-11. A significant dose-dependent inhibition to the anti-DNA bindingfrom SLE patients nos 2, 3, 5, and 7 was induced by El- I 1, namely the8.12-reactive IgG (P < 0-01 in all four instances when comparing theoriginal and the inhibited anti-DNA activities using the inhibitor at40Ipg/ml or more). The low DNA reactivity of control IgG was notaffected by the inhibitors. Values are mean of absorbance ± s.d.

DNA-binding inhibition testBoth anti-Id-enriched IgG preparations were next tested fortheir their specific inhibiting properties in functional ELISAwhich employed affinity-purified anti-DNA antibodies. Thelatter were obtained from serum samples of seven patientswith active SLE. Five of these were female patients, aged 22-29 years, with rapidly progressive lupus nephritis, whereaskidney function was apparently unaffected in the remainingtwo patients. Increasing amounts of anti-Id IgG from El-6and El-Il were then incubated overnight with 70 pg of theisolated anti-DNA preparations. Profiles of these experimentsare illustrated in Fig. 3, which indicates a general dose-dependent inhibition of the anti-DNA preparations to theirsubstrate. The highest inhibition of the DNA binding wasfound in the anti-DNA molecules from four nephriticpatients (nos 2, 3, 5 and 7) incubated with IgG of the El- l.In these instances, although a slight decrease of absorbancewas detected in the presence of concentrated IgG from El-6(80 gg/ml), the difference between the OD values of theoriginal and the El-i 1-inhibited anti-DNA activities was sig-nificant (P < 0 05) by using the inhibitor at 40 jig/ml or more.In contrast, anti-DNA antibodies from the two patients withspared kidney function were very little inhibited to theirantigen using both inhibitors at various concentrations. Asexpected, control IgG from Cohn fraction II incubated withEl-6 and El-Il showed no changes in their minimal, non-specific affinity to DNA (Fig. 3). In addition, parallel ELISAexperiments using hyperimmune anti-tetanus toxoid (TT)IMIG (commercial preparation no. 20) proved the inabilityof El-6 and El-Il to inhibit the binding of such anti-TT IgGto its substrate (data not shown).

DNA reactivity

ISLE potients EI-6 (31+-id-rwctive IgG)

EcJ

0)

a0

23

~~~~no3:~~~~n.Iano.ana72no.7t

Ia

24 F. Silvestris, P. Cafforio & F. Dammacco

located on A light chains of anti-DNA IgG, a number ofsequences of 8.12 are apparently located near to the firstcomplementarity determining region (CDR1) of the molecule[30]. Thus, it seems likely that the high inhibition of anti-DNAdetected by El-l is related to the complementarity of its 8.12-reactive IgG with the paratope to DNA.

Based on these results, effective anti-Ids to nephritogenicanti-DNA molecules such as the 8.12 marker may occasionallyoccur in certain IVIG preparations. Therefore, individualanalysis of the anti-DNA Id repertoire probing the anti-Idcomplementarity of selected IVIG batches should be performedto predict their putative therapeutic efficacy. Idiotypic charac-terization of anti-DNA antibodies [13] from additional SLEpatients is now in progress to define the potential clinical use ofanti-8.12+ Id-enriched IVIG preparations. Also, the datapresented here appear of interest when compared with thoseconcerning the 8.12 Id in SLE patients [20]. Indeed, theexpression of anti-Ids to 8.12 marker in the healthy populationand in the pool of donors of the IVIG preparation no. 11,strongly suggests that 8.12 is a germline gene-encoded determi-nant, and that anti-8. 12 molecules may also arise in normals bya mechanism which differs from the specific antigen-drivenselection of autoantibodies, as in SLE [30].

Results from the present study indicate that suppression ofpathogenic anti-DNA antibodies may occur only in the presenceof specific Id-related antibodies. However, the specificity ofpurified anti-Ids to pathogenic anti-DNA molecules should bebetter explored by using synthetic peptides related to the primarysequence of each Id determinant, as recently proposed for the V-88 anti-DNA Id [39]. Previous unsuccessful therapeuticapproaches with IVIG in SLE [40] may probably be accountedfor by a relative depletion of specific anti-pathogenic idiotypeantibodies in the IVIG preparations used. Alternatively, occa-sional exacerbation of proteinuria following IVIG treatment ofSLE nephritis might be explained by deposition of immunecomplexed, polyreactive, non-Id-specific IgG within the glomer-ular structures. Indeed, unless the IVIG preparations actuallycontain elevated levels of well defined anti-Id activity, noimpressive therapeutic benefit might be expected in SLE patients.

ACKNOWLEDGMENTS

The authors wish to thank Dr Betty Diamond, Albert Einstein Schoolof Medicine, New York, NY, for typing the myeloma IgG for anti-Ids3I, F4, and 8.12, and Alessandra Romito, Maria A. Grizzuti and PorziaFanelli for their excellent technical assistance. This work was supportedin part by the Italian Ministry of Universities and Scientific andTechnological Research (MURST), 60% funds, 1993.

REFERENCES

1 McIntyre EA, Linch DC, Macey MG et al. Successful response tointravenous immunoglobulin in autoimmune hemolytic anemia. BrJ Haematol 1985; 60:287-92.

2 Pollak S, Cunningham-Rundles C, Smithwick EM et al. High doseintravenous -y-globulin for autoimmune neutropenia. N Engl J Med1982; 307:243-4.

3 Dammacco F, Iodice G, Campobasso N. Treatment of adultpatients with idiopathic thrombocytopenic purpura with intra-venous immunoglobulin: effects on circulating T cell subsets andPWM-induced antibody synthesis in vitro. Br J Haematol 1986;62:125-31.

4 Gadjos PH, Outin H, Elkharrat D et al. High-dose intravenous-y-globulin for myasthenia gravis. Lancet 1984; i:406-7.

5 Carreras LO, Perez GN, Vega HR et al. Lupus anti-coagulant andrecurrent fetal loss: successful treatment with -y-globulin. Lancet1988; ii:393-5.

6 Sultan Y, Kazatchkine MD, Maisonneuve P et al. Anti-idiotypicsuppression of autoantibodies to factor VIII by high-dose intra-venous immunoglobulin. Lancet 1984; ii:765-8.

7 Ashaki K, Nagasawa K, Mayumi T. Successful treatment ofrefractory systemic lupus erythematosus with intravenous immuno-globulins. J Rheumatol 1990; 17:375-82.

8 Jordan SC. Intravenous -y-globulin therapy in systemic lupus ery-thematosus and immune complex disease. Clin Immunol Immuno-pathol 1989; 53:S164-9.

9 Rossi F, Kazatchkine MD. Anti-idiotypes against autoantibodies inpooled normal human polyspecific Ig. J Immunol 1989; 143:4104-12.

10 Dietrich G, Kazatchkine MD. Normal immunoglobulin G (IgG)for therapeutic use (intravenous Ig) contain anti-idiotypic specifi-cities against an immunodominant, disease-associated, cross-reactiveidiotype of human anti-thyroglobulin auto-antibodies. J Clin Invest1990; 85:620-9.

11 Madaio MP, Schattner A, Schattner M et al. Lupus serum andnormal human serum contain anti-DNA antibodies with the sameidiotypic marker. J Immunol 1986; 137:2535-40.

12 Buskila D, Shoenfeld Y. Anti-DNA idiotypes: their pathogenic rolein autoimmunity. In: Cruse JM, Lewis RE Jr, eds. Clinical andmolecular aspects of autoimmune diseases. Basel: S. Karger AG1992:85-113.

13 Isenberg D, Williams W, Axford J et al. Comparison of DNAantibody idiotypes in human sera: an international collaborativestudy of 19 idiotypes from 11 different laboratories. J Autoimmun1990; 3:393-414.

14 Shoenfeld Y, Ben-Yehuda 0, Messinger Y et al. Autoimmunediseases other than lupus share common anti-DNA idiotypes.Immunol Lett 1988; 17:285-90.

15 Karounos DG, Grudier JP, Pisetsky DS. Spontaneous expression ofantibodies to DNA of various species origin in sera of normals andpatients with SLE. J Immunol 1988; 140:451-5.

16 Vernier RL, Klen DJ, Sisson SP et al. Heparan sulfate-rich anionicsites in the human glomerular basement membrane: decreasedconcentration in congenital nephrotic syndrome. N Engl J Med1983; 309:1001-9.

17 Brinkman K, Termaat R, Berden JH et al. Anti-DNA antibodiesand lupus nephritis: the complexity of crossreactivity. ImmunolToday 1990; 11:232-5.

18 Suzuki N, Harada T, Mizushima Y et al. Possible pathogenic role ofcationic anti-DNA autoantibodies in the development of nephritisin patients with systemic lupus erythematosus. J Immunol 1993;151:1128-36.

19 Casali P, Notkins AL. CD5+ lymphocytes, polyreactive antibodiesand the human B-cell repertoire. Immunol Today 1989; 10:364-8.

20 Livneh A, Or G, Many A et al. Anti-DNA antibodies secreted byperipheral B cells of lupus patients have both normal and lupus-specific features. Clin Immunol Immunopathol 1993; 68:68-73.

21 Roux KH, Tankersley DL. A view of the human idiotypic reper-toire: electron microscopic and immunologic analyses of sponta-neous idiotype-anti-idiotype dimers in pooled human IgG. JImmunol 1990; 144:1387-92.

22 Dietrich G, Kaveri SV, Kazatchkine MD. Modulation of auto-

immunity by intravenous immune globulins through interactionwith the function of the immune/idiotypic network. Clin ImmunolImmunopathol 1992; 62:S73-81.

23 Vassilev T, Gelin C, Kaveri SV et al. Antibodies to the CD5molecule in normal human immunoglobulins for therapeutic use

(intravenous immunoglobulins, IVIg). Clin Exp Immunol 1993;92:369-72.

IgG to pathogenic anti-DNA idiotypes in IVIG 25

24 Silvestris F, Bankhurst AD, Searles RP et al. Studies of anti-F(ab')2antibodies and possible immunologic control mechanisms in sys-temic lupus erythematosus. Arthritis Rheum 1984; 27:1387-96.

25 Silvestris F, Searles RP, Williams RC Jr et al. Distribution of anti-F(ab')2 antibodies and 16/6 idiotype in SLE probands and kindreds.J Clin Immunol 1989; 9:462-8.

26 Silvestris F, Williams RC Jr, Frassanito MA et al. In vitro inhibitionof anti-DNA producing cells from patients with systemic lupuserythematosus by autologous anti-F(ab')2 antibodies. Clin Immu-nol Immunopathol 1987; 42:50-62.

27 Silvestris F, Yancey WB, Malone C et al. Parallelism of serum anti-F(ab')2 and anti-cationic IgG reactivities in patients with systemiclupus erythematosus. Clin Immunol Immunopathol 1991; 59:256-70.

28 Aranow C, Manheimer-Lory A, Diamond B. Molecular analyses ofF4W anti-DNA antibodies in SLE. Arthritis Rheum 1992; 35:S53-59.

29 Davidson A, Smith A, Katz J et al. A cross-reactive idiotype onanti-DNA antibodies defines a H chain determinant present almostexclusively on IgG antibodies. J Immunol 1989; 143:174-81.

30 Paul E, Iliev AA, Livneh A et al. The anti-DNA-associated idiotype8.12 is encoded by the V lambda II gene family and maps to thevicinity of L chain CDR1. J Immunol 1992; 149:3588-95.

31 Evans MJ, Suenaga R, Abdou NI. Detection and purification ofanti-idiotypic antibody against anti-DNA in intravenous immuneglobulin. J Clin Immunol 1991; 11:291-5.

32 Abdou NI, Suenaga R, Hatfield M et al. Regulation of anti-DNAantibodies by anti-idiotypic antibodies in sera of families of lupus

patients. In: Reichlin M, Capra JD, eds. Idiotypes. New York:Academic Press, 1986:95-107.

33 Evans M, Abdou NI. In vitro modulation of anti-DNA secretingperipheral blood mononuclear cells of lupus patients by anti-idiotypic antibodies of pooled human intravenous immuneglobulin. Lupus 1993; 2:371-5.

34 Zouali M, Kalsi J, Isenberg D. Autoimmune diseases at themolecular level. Immunol Today 1993; 14:473-6.

35 Gauthier VJ, Mannik M, Stricker GE. Effect of cationized anti-bodies in preformed immune complexes on deposition and persis-tence in renal glomeruli. J Exp Med 1982; 156:766-77.

36 Hahn BH, Ebling FM. Idiotype restriction in murine lupus: highfrequency of public idiotypes in serum IgG in nephritic (NZB/NZW)Fj mice. J Immunol 1987; 138:2110-7.

37 Kalunian K, Panosian-Sahakian N, Ebling FM et al. Idiotypiccharacteristics of immunoglobulins associated with systemic lupuserythematosus: studies of antibodies deposited in glomeruli ofhumans. Arthritis Rheum 1989; 32:373-82.

38 Halpern R, Davidson A, Lazo L et al. Familial SLE-presence of across-reactive idiotype in healthy family members. J Clin Invest1985; 70:731-6.

39 Staines NA, Ward FJ, Denbury AN et al. Primary sequence andlocation of the idiotope of V-88, a DNA-binding monoclonalautoantibody, determined by idiotope scanning with syntheticpeptides on pins. Immunology 1993; 78:371-8.

40 Barron KS, Sher MR, Silverman ED. Intravenous immunoglobulintherapy: magic or black magic. J Rheumatol 1992; 19:94-97.