Induction and immunochemical properties of a novel anti-antibody

Mobcular Immunology,Vol. 27. No. 1I, pp. I 119-l 126, 1990 Printed in Great Britain.

INDUCTION

0161-5890/90$3.00+ 0.00 Pergamon Press plc

AND IMMUNOCHEMICAL PROPERTIES A NOVEL ANTI-ANTIBODY

OF

JANASIMECKOVA-ROSENBERG and L. SCOTTRODKEY Department of Pathology and Laboratory Medicine, University of Texas Medical School, P.O. Box 20708, Houston, TX 77225-0708, U.S.A. (First received

18 January 1990; accepted in revised form I7 April 1990)

Abstract-This paper describes results which characterize an induced antibody in normal outbred rabbits which we have, for convenience, called parareactant (PR). PR resulting from autoimmunization of rabbits with either keyhole limpet hemocyanin-anti-tetanus toxoid F(ab’), or with tetanus toxoid-anti-tetanus toxoid F(ab’), complexes was studied. PR activity was directed solely to autologous, homologous or heterologous F(ab’): fragments regardless of their specificity. PR failed to react with intact antibodies or with antigen-antibody complexes consisting of homologous antibody bound to specific antigen. Radioimmunoassay and ELISA inhibition assays showed that reactivity between PR and autologous anti-tetanus toxoid F(ab’), or homologous anti-bovine serum albumin F(ab’), fragments was specifically inhibited with antigen. Anti-allotypic antibodies specific for a2 and b6 markers strongly inhibited binding of “‘I-anti-micrococcal carbohydrate F(ab’), (a2, b6) with PR (a3, b4, b5). PR specificity thus appears to be directed against non-idiotypic determinants present in Fv regions. Affinity immunoblotting was used to analyze clonality of PR in the sera collected from individual rabbits during the course of an active immune response. PR-positive sera displayed clonally restricted spectrotype patterns. PR molecules were predominantly IgG with isoelectric points of 5.9-6.8. These results strongly suggest that these PR molecules are coded by a small number of V region genes.

INTRODUCHON Antibodies (Ab) specific for epitopes on other Ab have been recognized in both healthy and diseased individuals for many years. Numerous anti-Ab specificities for both Fc and Fab regions have been described. An anti-Fab antibody termed homoreactant was first described as a gamma-globulin present in normal rabbit sera which reacts with enzymatic fragments of homologous IgG (Mandy et al., 1966; Mandy, 1967). This was demonstrated by the agglutination of Fab or F(ab’),-sensitized human 0 Rh( +) erythrocytes upon the addition of pooled or individual normal rabbit sera. Because homoreactant was unreactive with intact IgG molecules (Mandy, 1966), it was concluded that its specificity was directed against buried determinants of IgG, which are exposed after enzymatic digestion. In this respect homoreactant was similar to human antibodies reacting with F(ab’), fragments, initially called pepsin

Abbreviations: Ab, antibody; Ag, antigen; Ag-Ab, antigenantibody complex; BSA, bovine serum albumin; BSB, borate-saline buffer; CFA, complete Freund’s adjuvant; ELISA, enzyme-linked immunosorbent assay; Fv, variable fragment of immunoglobulin; IEF, isoelectric focusing; IFA, incomplete Freund’s adjuvant; Id, idiotype; KLH, keyhole limpet hemocyanini MCO, micrococcal cell wall carbohydrate; NC, nitrocellulose; PBS, phosphate-saline b&er; PBS-T, phosphate-salink buffer with Tween; PR, parareactant; PR-GA, glutaraldehydepolymerized parareactant; RIA, radioimmunoassay; TT, tetanus toxoid; TCA, trichloroacetic acid; V region, variable region.

agglutinators (Osterland et al., 1963). Although pepsin agglutinators could be demonstrated in many normal human sera (Birdsall and Rossen, 1982; Morgan ef al., 1979; Persselin et al., 1984), they have been found most frequently in high titers in patients with autoimmune diseases and in chronic illnesses such as malignancies and infections (Birdsall and Rossen, 1982; Davey and Korngold, 1982; Nasu et al., 1982; Morgan et al., 1982; Silvestris et al., 1984; Nasu et al., 1980). The mechanism for pepsin agglutinator production as well as the biological function and significance of these antibodies is not clear. In addition to pepsin agglutinator, human antibodies with specificities for the different portions of the Fab and F(ab’), molecules were reported (Nasu et al., 1982, 1980; Birdsall and Rossen, 1983; Nasu et al., 1981) and thought to be directed to idiotypes (Nasu et al., 1982; Birdsall and Rossen, 1983; Nasu et al., 1981) or previously unknown allotypes (Nasu et al., 198 1). Thus, anti-F(ab’), antibodies represent a heterogenous group of immunoglobulins with diverse antigenic specificities (Silvestris et al., 1986; Ling and Drysdale, 1981) and they thereby may also possess multiple functions, including the potential for interactions in the idiotype network (Jerne, 1974; Birdsall and Rossen, 1984). In this report, rabbit antibodies specific for epitopes on F(ab’), fragments of IgG were induced by autoimmunization using F(ab’), fragments of autologous antibodies chemically coupled to carrier protein or bound into an antigen-antibody (Ag-Ab) complex. We established that the resulting antibodies

1119

JANASIMECKOVA-ROSENBERG and L. SCOTTRODKEY

1120

reacted with epitope(s) in or near Ag binding sites. However, PR did not possess anti-idiotypic properties unlike human anti-F(ab’), antibodies previously described (Nasu et al., 1982; Birdsall and Rossen, 1983; Nasu et al., 1981). Furthermore, evidence is provided that PR specificity is directed against epitopes in Fv regions. The clonal restriction of PR suggests that PR molecules are coded by a small number of V region genes. MATERIALSAND

METHODS

Antigens and immunization

Tetanus toxoid (TT) was obtained from Connaught Laboratories, Swiftwater, PA. Micrococcus lysodeikticus cells were purchased from Worthington Biochemical Corporation, Freehold, N.J. Micrococcal cell wall carbohydrate (MCO) was prepared by lysozyme digestion as described previously (Brown and Rodkey, 1979). Anti-TT F(ab’), fragments (1 mg/ml) were mixed with KLH (1 mg/ml PBS) at a ratio of 4: 1. Freshly diluted glutaraldehyde was then added to a final concn of 0.05%. The reaction was allowed to proceed until the solution was opalescent. The reaction was stopped by addition of lysine to a final concn of 0.05 M. This was followed by extensive dialysis against 0.1 M phosphate-saline buffer (PBS), pH 7.2. TT-anti-TT F(ab’), complex for injection was prepared from anti-TT F(ab’), fragments (2 mg/ml) mixed with TT (1 mg/ml) at the ratio 5: 1 and stirred for 4 hr at 37°C. The mixture was then stirred at 4°C for 2 days. The first group (8 rabbits) was given three injections of TT (1 mg; 0.5 mg; 1 mg) at three week intervals and serum was collected weekly for a period of 168 days. The second group (3 rabbits) was immunized subcutaneously with 150 pg of TT (25 pg/site, 6 sites). Anti-TT antibodies from individual rabbits were purified from a pool of serum collected during the first 26 days post-inoculation by affinity chromatography and their F(ab’)? fragments were isolated after pepsin digestion. Part of the anti-TT F(ab’), preparation was coupled to KLH for reinjection. Two hundred microgrammes of KLH-anti-TT F(ab’), was emulsified in either complete (CFA) or incomplete (IFA) Freund’s adjuvant (Difco Laboratories, Detroit, MI) and injected back into the same individual on days 122 and 137. The third group (2 rabbits) was administered two rounds of injections. First, the rabbits received single injections of TT (25 pg/site, 6 sites). During the second round of immunization 200 pg of TT-anti-TT F(ab’), complex in CFA or IFA was injected back into the same rabbit on days 122 and 137. Purification of antibodies and pepsin digestion

Tetanus toxoid was dialyzed against borate-saline buffer (BSB), pH 8.0 and coupled to Sepharose 4B

(Pharmacia Fine Chemicals, Piscataway, NJ) using the cyanogen bromide method (Cuatrecasas, 1970) at a ratio of IO-13 mg TT per 1 ml packed beads. Rabbit or human anti-m serum was passed over a Sepharose 4B-TT column and after extensive washing with BSB, the anti-TT antibodies were eluted with 1 M propionic acid and dialyzed exhaustively against BSB, pH 8.0. Immune sera collected from rabbits immunized with a vaccine of Micrococcus lysodeikticus cells (Osterland et al., 1966; Strosberg et al., 1974) were used for purification of anti-micrococcal carbohydrate antibody (anti-MCO) by affinity chromatography on a Sepharose 4B MC0 column (Brown and Rodkey, 1979). Specifically purified rabbit and human antibodies were dialyzed against 0.1 M acetate buffer, pH 4.3 and digested with pepsin (Sigma Chemical Co., St Louis, MO) at a 50: 1 protein-to-enzyme ratio for 18 hr, at 37°C. The resulting digest was separated on Sephadex G-l 50 followed by chromatography on protein A-Sepharose 4B (Pharmacia Fine Chemicals, Inc.) to remove any residual undigested IgG. Reagents for immunoassays

(a) Biotinylation. F(ab’), fragments or anti-b4 IgG (1 mg/ml) were dialyzed overnight at 4°C against 0.1 M NaHCO,, pH 8.0. The N-hydroxysuccinimido biotin was dissolved in dimethylsulfoxide (1 mg/ml) and was added to dialyzed samples (at a weight ratio of activated biotin: Ig = 1: 8). The mixture was incubated at room temp for 4 hr and dialyzed overnight at 4°C against PBS, pH 7.2. (b) Iodination. F(ab’), fragments were iodinated using carrier ICI (redistilled) and “‘1 (New England Nuclear, Boston, MA) according to McFarlane (1958). Incorporation of i2’I was between 1.4 and 1.6 atoms I/molecule F(ab’), . Iodinated samples were dialyzed free of non-TCA precipitable radioactivity, and stock iodinated F(ab’), preparations were assayed for TCA precipitable counts before each radioimmunoassay experiment. (c) PR polymerization. A globulin fraction (10 mg/ml) prepared from PR-positive sera by sodium sulfate precipitation was mixed with an equal volume of 0.1% glutaraldehyde (Fisher Scientific Company, St Louis, MO) and stirred for 30 min at room temp. After washing the precipitate with PBS, lysine was added to a final 0.1 M concn and the mixture was stirred for 30 min. The polymerized PR (PR-GA) was washed three times with BSB and homogenized to a final 50% suspension. Radioimmunoassay (RZA) (1) Screening of PR positive sera

One hundred microlitres of ‘2sI-F(ab’), (- 10,000 cpm/25 ng) was added to plastic tubes which had been previously coated with 2 ml of 2% bovine serum albumin (BSA) to prevent nonspecific binding. Five

Anti-antibodies microlitres of test serum, or the same volume of both PR-negative preinoculation serum and BSB (negative controls), was then added. After incubation at 37°C for 2 hr, 25 ~1 of a 50% suspension of protein A-Sepharose (Pharmacia Fine Chemicals) was added. The mixtures were rotated at room temp for one hr, then centrifuged at 3000 rpm for 15 mins. The supernatants were discarded, the protein A-Sepharose beads were washed three times with BSB and the radioactivity was counted in a gamma counter (LKBWallac, Model 1282). All determinations were done in duplicate. In all calculations the background counts were subtracted. (2) Inhibition ‘*‘I-anti-TT

of the F(ab’),

reactivity

(a) By autologous, homologous IgG and their F(ab’), fragments.

between

PR

and

and heterologous

Inhibition assays were done using a quantity of PR (usually 5-10~1) which was pre-titrated to bind 50% of labeled target F(ab’)* This was incubated with different amounts of inhibitors in a constant volume of 10 ~1 or with BSB only for 2 hr at 37”C, after which 100 ~1 radiolabeled anti-TT F(ab’), fragments (_ 10,000 cpm) was added and incubation was continued at 37°C for 2 hr. The rest of the experiment was performed as described above. When IgG was used as inhibitor, the quantities of protein A-Sepharose added to tubes were adjusted to accommodate the increase of IgG in reactions. Preinoculation serum or BSB were used in place of PR as negative controls. The percentage inhibition was calculated as

l-

cpm bound by PR in the presence of inhibitor x 100 cpm bound by PR without inhibitor

(b) By antigen (TT). The inhibition effect of TT on reactivity between PR and “‘1-anti-TT F(ab’), fragments was assessed by preincubation of different concns of TT (25-5OO~g/lOO~l) with 100 ~1 of ‘*‘I-anti-TT F(ab’), (_ 10,000 cpm) at 37°C for 2 hr. Then 10 ~1 of PR positive serum was added and the experiment was performed as described in RIA above. (3) Inhibition of PR binding by anti-allotype

antibodies

One hundred microlitres (w 10,000 cpm) aliquots of ‘*‘I-anti MC0 F(ab’), fragments (a3, b6 allotype) were mixed with increasing quantities of anti-allotype antiserum specific for either a2 or b6 allotypic determinants or with BSB only. After incubation at 37°C for 2 hr, a volume of 50% suspension of PR-GA (a3, b4, b5 allotype) which bound 50% of labeled F(ab’), fragments in the absence of inhibitor was added to each tube and the mixture was agitated for 2 hr at room temp. The samples were centrifuged, washed three times with BSB and the radioactivity bound to insoluble PR-GA was measured as described above. All determinations were done in triplicate . To check that neither anti-a2 nor anti-b6 antibodies bind

1121

nonspecifically to PR-GA (a3, b4, b5 allotype) the highest amount of anti-allotypic antibodies used for inhibition was mixed with a 50% suspension of PR-GA and agitated for 2 hr at room temp. After exhaustive washing with BSB, 100~1 of “‘I-antiMC0 F(ab’)* fragments (N 10,000 cpm) was added and the mixture was agitated for another 2 hr at room temp, then washed with BSB and the radioactivity was counted. ELISA

The ELISA method was used to evaluate the titre of PR in the serum samples collected from rabbits at different times during the immune response. Microtiter wells (Dynatech) were coated with 50 PI/well containing 5 pg/ml of human F(ab’), fragments previously absorbed with TT by affinity chromatography on a Sepharose 4B-TT column. After overnight incubation, the plates were washed three times with PBS-0.05% Tween (PBS-T) and incubated in PBS-T at room temp for 1 hr to block nonspecific binding. Fifty microlitres of test sera serially diluted in PBST or the same volume of either PR-negative preinoculation serum or BSB-T (negative controls) was added to the wells. After incubation for 1 hr at 37”C, plates were washed three times with PBS-T and 50 ~1 of peroxidase-conjugated goat anti-rabbit Fc fragment antiserum (Cooper Biomedical, Malvem, PA.) diluted 1: 10,000 in PBS-T was added. Plates were incubated for 1 hr at room temp, washed three times with PBS-T followed by one wash with PBS. Fifty microlitres of o-phenylenediamine substrate was then added and the plates were incubated in the dark for 10min at room temp. The reaction was stopped by addition of 50 ~1 of 2N H,SO, per well. The color was read at 492 nm in a Titertek” Multiskan ELISA Microreader. Coating of ELISA plates with human F(ab’), fragments was routinely verified using peroxidase-conjugated goat antihuman F(ab’), fragment antiserum diluted 1: 1000 in PBS-T. Sandwich

ELISA

Preliminary titration experiments were first performed to determine the optimal Ag:Ab concn ratio. Microtiter wells were coated with MC0 (0.25 pg/ well). The plates were washed and 50~1 aliquots containing either lOpg/ml anti-MC0 F(ab’), or 15 pg/ml affinity purified anti-MC0 IgG from a rabbit of allotype a2,b6 was added. Following incubation at 37°C for 2 hr the plates were washed and 50 ~1 of PR-positive serum from a rabbit of allotype a3,b4,b5 serially diluted in PBS-T was added. The plates were incubated at 37°C for 2 hr, washed three times with PBS-T and 50~1 of biotinylated anti-b4 antibodies was added. After incubation at 37°C for 1 hr and washing with PBS-T, 50 ~1 of peroxidase-conjugated avidin (Cooper Biomedical, Malvern, PA) diluted 1:2000 in PBS-T was added and the mixture was incubated

1122

JANA

SIMECKOVA-ROSENBERG and L. SCOTT RODKEY

for 1 hr at 37°C. The final steps were completed as described above. Several controls were included in this experiment as follows: (a) S~ificity of biotinylated anti-b4 antibodies and peroxidase-conjugated avidin was verified by incubation of these reagents with wells containing either MC0 or MCO-anti-MC0 F(ab’), (a2,b6 allotype) complexes. (b) The absence of anti-MC0 Ab from PR-positive sera was verified. Fifty microliters of PR-positive serum (a3,b4,bS allotype) was incubated with MCOcoated wells and after washing was developed with biotinylated anti-b4 antibodies, peroxidase-conjugated avidin and substrate. (c) Successful coating of ELISA plates with MC0 was verified using rabbit anti-MC0 F(ab’), fragments, biotinylated goat anti-rabbit F(ab’), antibodies, peroxidase-conjugated avidin and substrate. Isoelectric focusing (ZEF) Polyacrylamide gels (25 x 12 x 0.2 cm) consisted of 5.3% T, 3% C, 2% ampholyte and 0.1% Tween 20. Ampholytes used for IEF were synthesized according to Binion et nl. (1983). Electrode wicks were saturated with 1 M NaOH (cathode) and 1 M H,PO, (anode). Ten microlitres of PR-positive sera were applied to the gel surface using silicone rubber applicators approximately 3 cm from the anode. IEF was performed using constant voltage of 100 V for 15 min followed by raising the voltage to 200 V for 10 additional min. The power supply was then switched to 20 W constant power for 20 min and to 30 W for an additional 90 min. Immediately after IEF, the pH gradient in the gel was measured.

PR clonotype distribution was assayed as previously described (Knisley and Rodkey, 1986). Briefly, nitrocellulose (NC) sheets (0.45 pm, Schleicher and Schuell, Keene, NH) were coated with protein-A (10 ~g/ml) in OS M sodium bicarbonate by overnight

agitation. The protein-A coated NC was rinsed with PBS and rocked in PBS-l% Tween for 15 min prior to immunoblotting. Following the IEF run, the NC sheet was blotted on filter paper to remove excess buffer and placed onto the gel surface over the lane containing the focused PR. After incubation in a humid chamber at 37°C for 15 min, the NC sheet was rinsed with PBS-0.05% Tween for 20 min. The immunoblot was then developed using biotinylated rabbit anti-MC0 F(ab’), fragments, avidin-horseradish peroxidase conjugate and substrate (0.6 mg/ml diaminobenzidine, 0.03% HzOZ in PBS). RESULTS

Screening for PR activity Individual sera collected from three groups of rabbits were tested for PR by independent RIA and ELISA assays. Rabbit and human F(alQz fragments were used to measure PR activity. The controls, Group I, which had been given 3 injections of TT at 21 day intervals contained virtually undetectable PR in pre- and post-inoculation sera. The PR titers of most of these sera were 1:2S and never exceeded 1:SO (data not shown). Table 1 shows the immunization schedule and the PR titers for Group II and Group III. Only trace amounts of PR were found in the preinoculation sera and in sera from Groups II and III after a single injection of TT. However, a dramatic rise in PR titers after autoimmuni~tion with either KLH-anti-TT F(ab’), (Group II) or with TT-anti-TT F(ab’)z complex (Group III) was obtained (Table 1). PR Specificity The immune set-a with the highest PR concns were chosen for RIA inhibition assays to examine the fine specificity of PR. Autologous, homologous and heterologous intact IgG or pepsin fragments were used as inhibitors of the reaction between PR and autologous radiolabeled anti-?-T F(ab’), fragments. The results of these experiments (Tables 2 and 3)

Table I. PR titers in sera from individual rabbits determined bv ELISA Titer of parareactant Group II Days 0

Treatment inject TT

is 20 26 96 122 137 147 158

172

186 202

646 hi:50 I:.50 1:lOO I:100 I:100

647 1:25 I:25 I:50 I:25 1:50

648 1:25


Group 111 644

645

< i:25 I:25 I:25 I:25 I:100

1:25 I:25 I:25 I:25 I:25

‘Injections ?njections I:1600 I:1600 I:1600 l:1600 1:1600

I:1600 I:1600 I :3200

1:i2,800 I:1600

1:200 I:1600 I :3200 I :3200 1:6400

I:1600

1:1600

I:800 1:aoo 1:1Mfo I :3200

I:400 1:zOO I:100 I:200

“Microtiter wells were coated with human F(ab’), fragments absorbed with TT by affinity chromatography on a Sepharose 4B-TT column. “PR titers in preinoculation sera. ‘Group II was injected with 200 pg of KLH-anti-TT F(ab’), dGroup II1 was injected with 200 pg of TT-anti/F@&‘), complexes.

Anti-antibodies Table 2. Inhibition Inhibitors

of PR and autologous F(ab’), reaction

‘2sI-anti-TT

Amount

Inhibition

b59”

WI

SOUXC”

1123

Anti-TT

IgG

3.700 0.370 0.037

6.1 0 0

Anti-TT

F(ab’),

2.500 0.250 0.025

88.5 65.8 18.7

Anti-MC0

IgG

3.700 0.370 0.037

0.9 0 0

Anti-MC0

F(ab’),

2.500 0.250 0.025

86.0 73.6 23.0

0.0

1.5

1.0

0.5

PR Reciprocal Dilution (Hundreds)

“PR, anti-TT IgG, radiolabeled and unlabeled anti-TT F(ab’), were from rabbit 645 (a3,b4,b5). ‘pg of inhibitors in a constant volume of IO ~1. Anti-MC0 IgG and anti-MC0 F(ab’), were from rabbit 102 (al&l). Note: Equimolar amounts of affinity purified IgG antibodies and their F(ab’), fragments were used for the RIA-inhibition assay.

showed that PR activity is directed only to F(ab’), fragments regardless of their specificity and species origin. The human F(ab’), molecules inhibited with slightly lower efficiency (Table 3). Furthermore, the results clearly confirmed that PR cannot be autoanti-idiotype Ab because of lack of inhibition with intact autologous anti-TT IgG. Additional experiments were performed to determine whether these epitopes can be exposed as a result of non-digested IgG antibodies binding to their specific antigens. Immune complexes composed of MCO-anti-MC0 IgG and MCO-anti-MC0 F(ab’), were prepared. The binding of homologous PR on the pre-formed immune complexes was measured using sandwich ELISA. The data showed (Fig. 1), that PR failed to react with either intact antibodies or their F(ab’), fragments when either were bound to antigen. PR reacted only with noncomplexed, soluble anti-MC0 F(ab’), fragments. These results suggested that PR reacts with epitopes within or closely associated with the antigenbinding site of IgG antibodies. To insure that the inhibition result was not biased due to either the antigen used or the assay method, two assays and two Ag-Ab systems were studied. The inhibiting effect of antigen (TT) on the reactivity of PR with autologous 1251-anti-TT F(ab’), fragments was assessed by an RIA inhibition assay (Fig. 2). The binding of PR to autologous radiolabeled anti-TT F(ab’), fragments

Fig. 1. ELISA binding assay of PR (a3,b4,b5) to (a) homologous anti-MC0 F(ab’), (0 solid line); (b) MC&anti-MC0 F(ab’), complex (1 dashed line); (c) MCGanti-MC0 IgG complex (+ dotted line).

-rfW) Fig. 2. RIA inhibition assay for PR and autologous ‘%anti-TI F(ab’), inhibited with IT.

was significantly inhibited with TT. However, TTmediated inhibition was never complete and reached a plateau (maximal inhibition about 58.5%) at a concn of 24Opg of TT (Fig. 2). Similar inhibition (data not shown) was found for a heterologous system in which the cross-reactivity between rabbit PR and human “‘1-anti-TT F(ab’), fragments was inhibited with TT. The capacity of bovine serum albumin (BSA) to inhibit the reactivity between rabbit PR and homologous anti-BSA F(ab’), fragments was assayed by ELISA. As shown in Fig. 3, BSA nearly completely blocked (89.1%) the binding loo,

I

90 SO g

‘O

so

Table 3. Inhibition Amount

reaction of PR and ‘*‘I-anti-n F(ab’\, fraements

of inhibitors (pg)” 25 2.5 0.25

anti-TT

F(ab’), by human

Inhibition (%) F(ab’), nonspecific

77.3 24. I 17.5

ND = not done. “The same human serum was used as a source F(ab’), and non-specific F(ab’), fragments. ‘pg of inhibitors in a constant volume of IO ~1.

F(ab’),”

86.9 45.5 ND of both anti-TT

0’ 50

g E

40

=

30 20 10 I 0

2

4

BSA 0w)

Fig. 3. ELISA inhibition assay for PR and homologous anti-BSA F(ab’), inhibited with BSA.

JANA SIMECKOVA-ROSENBERGand L. SCOTT RODKEY

1124

of PR to anti-BSA F(ab’), fragments. These results suggest that PR reacts with epitopes located near the antibody paratope. Further evidence related to the location of the epitope target of PR specificity was obtained by taking advantage of our experimental model, i.e. that rabbits possess well-characterized a-locus allotypic markers located within the V-region of Ig heavy chains. The RIA reaction between PR from a rabbit of allotype a3,b4,b5 and an “‘I-anti-MC0 F(ab’)z fragment from an a2,b6 rabbit was inhibited by adding anti-a2 antibodies. The results shown in Fig. 4 indicate that the anti-a2 antibodies effectively inhibited interaction between “‘I-anti-MC0 F(ab’), and PR. As little as 1 ~1 undiluted anti-a2 antiserum caused strong inhibition (64.5%). Strong inhibition of PR reaction with F(ab’), fragments was also observed (Fig. 5) in the same assay, when anti-b6 antiserum specific for b6 allotypic markers located on the light chain (K type) was used as an inhibitor. Maximal inhibition (75.3%) was reached using 40 ~1 of undiluted anti-b6 serum. CIonality of PR The clonal heterogeneity of PR was examined by IEF affinity immunoblotting. These results (Fig. 6) showed that PR displayed considerably restricted clonal spectrotype patterns. Early in the responses, PR consisted predominantly of molecules with isoelectric points between 5.9-6.8 (about 6 bands). New sets of PR-producing clones appeared to initiate production at two discrete times. New acidic clonotypes were noted beginning on day 265 and a set of basic clonotypes appeared on day 328. PR activity was not detected in preinoculation serum by IEF immunoblotting, confirming the results obtained with the ELISA and RIA methods.

lo OX

I/ 0

I

10

20

Anti-b6 Serum

30

40

(uL)

Fig. 5. RIA inhibition assay for PR (a3,b4,b5) and oSI-antiMC0 F(ab’), (a2,b6) with anti-b6 serum.

F(ab’), fragments, but were unreactive with intact IgG (Table 2). Antibodies of similar specificities were not elicited using TT alone for immunization. These auto-antibodies seemed initially to be homoreactant, i.e. a naturally occurring serum factor which reacts with determinants on the Fab or F(ab’)* fragments ordinarily buried in the undigested IgG (Mandy, 1967). However, our studies differ from findings of Woolsey and Mandy (1970), who reported that the levels of homoreactant activity in normal adult rabbits were not affected by immunization with autologous intact IgG or with purified Fab and F(ab’), fragments. These investigators suggested that the unsuccessful attempts to enhance homoreactant titers in adults may be due to the fact that adults are already hyperimmune with respect to homoreactant synthesis and therefore the exposure to the additional amounts of antigen would not enhance the homoreactant titers. This failure could have been due to inappropriate immunization protocols. Therefore, PR is obviously distinguishable from homoreactant on the

DISCUSSION

After autoimmunization of rabbits with either KLH-anti-TT F(ab’)* or with TT-anti-TT F(ab’), complex anti-globulin antibodies were produced (Table 1) which had specificity for epitope(s) on

8

7 I n 90

6

60

- 70 e\p 60 .s 60 .% 2 40 +

30

I

20 10 0

2

0

4

6

6

10

Anti-a2 Serum (uL)

Fig. 4. RIA inhibition MC0

assay for PR (a3,b4,b5) and “‘l-antiF(ab’)z (a2,b6) with anti-a2 serum.

Fig. 6. Spectrotype pattern of PR-positive sera collected from rabbit 645 at different times during the immune response. Sera were separated by IEF and blotted on protein A-coated NC. The blot was probed with biotinylated rabbit anti-MC0 F(ab’)z fragments followed by peroxidase-conjugated avidin.

Anti-antibodies

of PR being readily and routinely inducable in adults. Inhibition studies offered insight into the specificity of PR and drew a clear distinction between PR and homoreactant. The reactivity between PR and autologous 12jI-anti-TT F(ab’), fragments was inhibited by autologous, homologous or heterologous F(ab’), fragments regardless of their Ag specificity, but not with undigested IgG antibodies (Table 2). PR activity could also be absorbed with autologous anti-TT F(ab’), or unrelated F(ab’), fragments and remained undiminished following absorption with intact autologous anti-TT IgG or pooled rabbit IgG. Based on these results it appears that PR is specific for epitope(s) uncovered by pepsin digestion of IgG antibodies rather than for idiotypes. Although subsets of human anti-F(ab’), antibodies have been reported to react with idiotypes in pooled human F(ab’), (Nasu et al., 1982; Birdsall and Rossen, 1983; Nasu et al., 1981), we found no measurable anti-idiotypic antibodies in PR. Unlike Milgrom factor (anti-antibody which reacts only with altered IgG bound to its antigen) (Milgrom et aI., 1956) PR failed to combine with homologous MCO-anti-MC0 IgG complexes (Fig. I). Evidently, these PR specificities are restricted to “hidden” determinants only revealed by enzymatic treatment of IgG Ab and not by binding of IgG Ab to its antigen. Since PR did not react with homologous anti-MC0 F(ab’), fragments when they were bound to antigen (Fig. l), it seems that PR activity is directed to epitope(s) within or near antigen-combining sites. Alternatively, the binding of anti-MC0 F(ab’), fragment to antigen characteristic

could induce a conformational change and alter antigenicity of F(ab’), fragments. The results ob-

tained by RIA and ELISA inhibition assays strongly supported the first suggestion. The reactivity between PR and autologous anti-TT F(ab’), or homologous anti-BSA F(ab’), fragments was specifically inhibited with antigen (Figs 2 and 3). Thus PR, like human pepsin agglutinator (Silvestris et al., 1986) recognizes epitopes within variable region (probably framework epitopes), located very near the antigen binding site. Additionally, we have shown that reactivity between PR (a3,b4,b5) and homologous ‘251-anti-MC0 F(ab’), fragments (a2,b6) was strongly blocked with anti-a2 allotypic serum (Fig. 4). The results suggest that PR is directed against common epitopes present within Fv regions of F(ab’), molecules. The results show further (Fig. 5) that the reactivity between PR and homologous F(ab’)z fragments is strongly inhibited with anti-b allotype serum. These data are in accordance with the finding of the association of Fab homoreactant activity with the group b kappa chain allotype markers (Hagen et al., 1973). This could indicate the presence of additional specificities in PR sera or steric interference. What is the physiological role of PR? At this time we can only speculate on the function PR may have in uiuo, preliminary results suggest that PR could

1125

represent auto-antibodies with immunoregulatory properties. In some respects, the relationship between PR and F(ab’), molecules is similar to Id-anti-Id interactions. Some investigators reported that the levels of Id and anti-Id fluctuate reciprocally (Brown and Rodkey, 1979; Binion and Rodkey, 1982; Kelsoe and Cemy, 1979; Geha, 1982; Zouali and Eyquem, 1983) and that auto-anti-Id can modulate Id expression (Brown and Rodkey, 1979; Binion and Rodkey, 1982; Fernandez and Moller, 1979; Schrater et al., 1979; Goidl et al., 1979). In our preliminary experiments serum samples taken from one of the rabbits at different times during the course of active anti-TT response were analyzed by IEF for anti-TT clonotype distribution. We found (data not shown) that at the time when PR titer increased to the maximal level, a substantial shift in the distribution of individual anti-m Ab clonotypes occurred. Some anti-TT clonotypes which were predominant at the beginning of the immune response diminished later while new clonotypes were expressed. The observed phenomena, i.e. correlation between the appearance of PR and disappearance of specific clonotypes of the original Ab response are reminiscent of the behavior and relationship between Id and anti-Id. Furthermore, the clonally restricted pattern of PR (Fig. 6) suggests that an extremely limited endogenous or exogenous antigenic repertoire may be driving the production of PR. It also could indicate, that PR might bind to “conserved” epitope(s) on F(ab’), molecules. These “conserved” epitopes might be defined by a recurring gene-encoded sequence, e.g. framework or a specific D-region shared by a large number of heterogenous populations of immunoglobulins (Gerard0 et al., 1988). D region sequences contributing to cross-reactive idiotypes have been reported previously (Clevinger et al., 1980; Gridley et al., 1985; Rudikoff et al., 1983). Thus, like idiotypes and anti-idiotypes, PR-F(ab’), interactions might have some regulatory role in the immune network. The fact that PR failed to react with intact immunoglobulin in vitro may argue against a regulatory function. The suggestion that the role of PR is simply to facilitate removal of normal catabolically produced F(ab’)2 during immune responses must also be considered. Acknowledgements-We thank Dr W. J. Mandy for a critical review of the manuscript. This work was supported, in part by NIH grant AI20590 and by NASA contract NAS 9-17403 from the University of Texas Health Science Center at Houston through The Bioprocessing Research Center at Houston. REFERENCES

Binion S. B. and Rodkey L. S. (1982) Naturally induced auto-anti-idiotypic antibodies: induction by identical idiotypes in some members of an outbred rabbit family. .I. exp. Med. 156, 860-872. Binion S. B., Rodkey L. S., Egen N. B. and Bier M. (1983) Properties of narrow-range ampholytes isolated from widerange ampholyte preparations. Anal. Biochem. 12s, 71-76.

1126

JANASIMECKOVA-ROSENBERG and L.

Birdsall H. H. and Rossen R. D. (1982) Production of antibodies specific for Fc, Fab’, and streptokinasestreptodornase in vitro by peripheral blood cells from patient with rheumatoid arthritis and normal donors: identification of immune complexes in culture supernatants containing hidden antibodies reactive with Fab’ fragments of imm&oglobulin G. J. c/in. Invest. 69,75-84. Birdsall H. H. and Rossen R. D. (1983) Characterization of anti-Fab’ antibodies in human sera: identification of soluble immune complexes that contain hidden anti-KLH and blocking anti-immunoglobulins following immunization with keyhole limpet hemocyanin. Clin. exp. Immun. 53, 497-504. Birdsall H. H. and Rossen R. D. (1984) Regulation of antibody release by naturally occurring anti-immunoglobulins in cultures of pokeweed mitogen-stimulated human peripheral blood lymphocytes. J. Immun. 133, 1257-1264. Brown J. C. and Rodkey L. S. (1979) Autoregulation of an antibody response via network induced auto-antiidiotype. J. exp. Med. 150, 67-85. Clevinger B., Schilling J., Hood L. and Davie J. M. (1980) Structural correlates of cross-reactive and individual idiotypic determinants on murine antibodies to a-(1-> 3) dextran. 1. exp. Med. 151, 1059-1070. Cuatrecasas P. (1970) Protein purification by affinity chromatography. Derivatizations of agarose and polyacrylamide beads. J. bioi. Chem. 245, 3059-3065. Davey M. P. and Korngold L. (1982) Association of antiF(ab’), antibodies (pepsin agglutinators) with immune complexes as determined by enzyme-linked immunosorbent assays. Inr. Archs Allergy appl. Immun. 61, 278-283. Fernandez C. and Moller G. (1979) Antigen-induced strainspecific autoantiidiotypic antibodies modulate the immune response to dextran B512. Proc. nam Acad. Sci. U.S.A. 76, 59445947.

Geha R. S. (1982) Presence of auto-anti-idiotypic antibody during normal human immune response to tetanus toxoid antigen. J. Immun. 129, 139-144. Gerard0 S. H., Persselin J. E., Keld B. and Stevens R. H. (1988) Recognition bv anti-Fab antibodies in rheumatoid arthritis of s‘iructure(s) widely distributed on human Fab molecules. &and. J. Immun. 28, 613625. Goidl E. A., Schrater A. F., Siskind G. W. and Thorbecke G. J. (1979) Production of auto-anti-idiotypic antibody during the normal immune response to TNP-Ficoll. II. Hapten-reversible inhibition of anti-TNP plaque-forming cells by immune serum as an assay for auto-anti-idiotypic antibody. J. exp. Med. 150, 154165. Gridley T., Margolies M. N. and Gefter M. L. (1985) The association of various D elements with a singleimmunoglobulin VH gene segment: influence on the expression of a major cross-reactive idiotype. J. Immun. 134, 1236-l 244.

Hagen K., Kulkarni S. S. and Mandy W. J. (1973) A new serum factor in normal rabbits. IX. Association with group b immunoglobulin allotypic specificities. J. Immun. 111, 546553. Jerne N. K. (1974) Towards a network theory of immune system. Annls Immun. (Inst. Pasteur) 125c, 373-389. Kelsoe G. and Cerny J. (1979) Reciprocal expansions of idiotypic and anti-idiotypic clones following antigenic stimulation. Nature 279, 333-334. Knisley K. A. and Rodkey L. S. (1986) Affinity immunoblotting. High resolution isoelectric focusing analysis of antibody clonotype distribution. J. Immun. Meth. 95, 79-87.

Ling N. R. and Drysdale P. (1981) Antibodies in human sera to F(ab’), fragments of monoclonal and polyclonal IgG. Int. Archs Allergy appl. Immun. 66, 459463.

Mandy W. J. (1966) A new serum factor in normal rabbits. II. Reaction with buried antigenic determinants revealed after papain digestion. J. Immun. 97, 8766884.

SCOTT RODKEY

Mandy W. J. (1967) A new serum factor in normal rabbits, III. Specificity for antigenic determinants uncovered by papain or pepsin digestion. J. Immun. 99, 815-824. Mandy W. J., Fudenberg H. H. and Lewis F. B. (1966) A new serum factor in normal rabbits. I. Identification and characterization. J. Immun. 95, 501-509. McFarlane A. S. (1958) Efficient trace-labeling of proteins with iodine. Nature 182, 53. Milgrom F., Dubiski S. and Wozniczko G. (1956) Human sera with “anti-antibody”. VOX Sang. 1, 172-183. Morgan A. C., Rossen R. D., McCormick K. J., Stehlin J. S.. Jr. and Giovanella B. C. (1982) “Hidden” cvtotoxic antibodies that react with alldgeneic cultured fetal and tumor cells contained in soluble immune complexes from normal human serum. Cancer Res. 42, 881-887. Morgan A. C., Jr, Rossen R. D. and Twomey J. J. (1979) Naturally occurring circulating immune complexes: Normal human serum contains idiotype-anti-idiotype complexes dissociable by certain IgG antiglobulins. J. Immun. 122, 167221680.

Nasu H., Chia D. S., Iwaki Y., Barnett E. V. and Terasaki P. I. (1981) Cytotoxicity of anti-Fab antibodies against B lymphocytes in rheumatoid arthritis. Arthrifis Rheum. 24, 1278-1282.

Nasu H., Chia D. S., Knutson D. W. and Barnett E. V. (1980) Naturally occuring human antibodies to the F(ab’), portion of IgG. Ciin. exp. Immun. 42, 378-386. Nasu H., Chia D. S., Taniguchi 0. and Barnett E. V. (1982) Characterization of anti-F(ab’), antibodies in SLE patients. Evidence for cross-reacting auto-anti-idiotypic antibodies. Clin. Immun. Immunopaih. 25, 80-90. ._ Nisonoff A.. Markus G. and Wissler F. C. (1961) Senaration of univalent fragments of rabbit antibody by reduction of a single, labile disulphide bond. Narure-189, 2933295. Osterland C. K.. Harboe M. and Kunkel H. G. (1963) Anti_ y-globulin factors in human sera revealed by enzymatic splitting of anti-Rh antibodies. VOX Sang. 8[ 1331152. Osterland C. K.. Miller E. J.. Karakawa W. W. and Krause R. M. (1966) Characterization of streptococcal groupspecific antibody isolated from hyperimmune rabbits. J. exp. Med. 123, 599614. Perssehn J. E., Louie J. S. and Stevens R. H. (1984) Clonally restricted anti-IgG antibodies in rheumatoid arthritis. Arthritis Rheum. 27, 137881386. Rudikoff S., Pawlita M., Pumphrey J., Mushinski E. and Potter M. (1983) Galactan-binding antibodies. Diversity and structure of idiotypes. J. exp. Med. 158, 1385-1400. Schrater A. F., Goidl E. A., Thorbecke G. J. and Siskind G. W. (1979) Production of auto-anti-idiotypic antibody during the normal immune response to TNP-Ficoll. I. Occurrence in AKR/J and BALB/c mice of haptenaugmentable, anti-TNP plaque-forming cells and their accelerated appearance in recipients of immune spleen cells. J. exp. Med. 150, 1388153. Silvestris F., Bankhurst A. D., Searles R. P. and Williams R. C., Jr. (1984) Studies of anti-F(ab’), antibodies and possible immunologic control mechanisms in systemic lupus erythematosus. Arthrilis Rheum. 27, 138771396. Silvestris F., Williams R. C., Jr. and Searles R. P. (1986) Human anti-F(ab’), antibodies and pepsin agglutinators react with Fv determinants. Stand. J. Immun. 23, I

499-508.

Strosberg A. D., Hamers-Casterman C., Van Der Loo W. and Hamers R. (1974) A rabbit with the allotypic phenotype ala2a3b4b5b6. J. Immun. 113, 1313-1318. Woolsey M. E. and Mandy W. J. (1970) A new serum factor in normal rabbits. V. Induction in neonatal rabbits. J. Immun. 104, 475482.

Zouali M. and Eyquem A. (1983) Idiotypic-antiidiotypic interactions in systemic lupus erythematosus: demonstration of oscillary levels of anti-DNA autoantibodies and reciprocal anti-idiotypic activity in a single patient. Annls Immun. (Inst. Pasteur) 134C, 377-391.