Autoreactive responses to an environmental factor: 1. phthalate induces antibodies exhibiting anti-DNA specificity

Abstract

Environmental factors have been implicated in the induction of autoimmune disorders. We report here that a common chemical, phthalate, used widely in synthetic polymers and cosmetics induces serum anti-self DNA antibodies in BALB/c, NZB and autoimmune-prone NZB/W F1 mice. The latter group experiences a high mortality, and significantly higher anti-DNA antibody levels along with nephritis and other histopathologic changes in kidney. Comparison of amino acid sequences of an anti-phthalate BALB/c B-cell hybrid, 2C3 with the known database at the National Center for Biotechnology Information reveals a striking homology between the variable regions of 2C3-Ig (γ₁, κ) and an anti-DNA antibody, BV04-01 (γ2b,κ) isolated from the lupus-prone NZB/W F1 mice. The homology is 98% for κ light chain and 70% for γ heavy chain. Like 2C3-Ig, BV04-01 also has specificity for d(pT)4. Furthermore, the light chains of both 2C3-Ig and BV04-01 are products of Vκ1 gene. To understand the nature of anti-phthalate responses in general, hybridomas generated from phthalate–keyhole limpet haemocyanin-primed BALB/c splenocytes were characterized. The study identifies cross-reactive populations that strongly bind phthalate, DNA, or both. Of the 14 hybridomas evaluated, six express the same Vκ1 gene-derived light chain as 2C3, and bind both phthalate and ds and ss-DNA. They specifically recognize the oligonucleotides, d(pT)4, and d(pT)10. Additionally, when antisera raised against idiopeptides corresponding to 2C3-Ig hypervariable regions are allowed to react with 2C3-Ig, their binding is blocked specifically by both d(pT)4 and phthalate. This study clearly demonstrates that phthalate exposure leads to activation of a significant number of autoreactive B-cells, with the consequence of a significant pathogenic progression in susceptible NZB/W F1 mice but not in non-autoimmune-prone BALB/c.

Introduction

The aetiology of autoimmune disorders such as induction of anti-DNA antibodies in systemic lupus erythematosus (SLE) is largely unknown. High titres of anti-double-stranded (ds) DNA antibodies have been considered as the serological hallmark of a systemic autoimmune disease like lupus.^1,2 Anti-DNA antibodies in lupus-prone mice undergo heavy chain class switching to immunoglobulin G (IgG), and display somatic mutations in their variable region gene segments.^3,4 The structure of variable regions of anti-DNA antibodies indicates that they are produced in response to antigen-selective stimulation.⁵

Various environmental, genetic, and hormonal antigenic factors have been implicated in the development of this disease.^6–8 Some studies have implied that molecular mimicry between self and microbial antigens could play a role in triggering production of autoreactive anti-DNA antibody.⁹ It is possible that activation of self-reactive clones of T or B cells may also occur during exposure to chemicals that are abundantly present in the environment or administered as medications.^10,11

Phthalates (i.e. o-benzene dicarboxylates) comprise one such group of chemicals that are widely used as plasticizers in flexible polyvinyl chloride (PVC) polymers used in medical devices, children's toys, and as solvents for cosmetics. Another type of phthalate is widely used in making synthetic fibres. Thus, phthalate compounds in various forms readily occur in day-to-day human contacts. In rodents, these chemicals have often proved harmful,^12–15 although there is as yet no report linking phthalates to autoimmune disorders.

The present study was initiated during our studies with an anti-phthalate monoclonal antibody (mAb), 2C3-Ig.¹⁶ It was noted from a search in the BLAST database of National Center for Biotechnology Information (NCBI) that the amino acid sequence of 2C3-Ig chain variable regions has extensive homology (>98%) with that of BV04-01, an auto anti-DNA mAb from autoimmune-prone NZB/W F1 mice. The homology between the two with respect to heavy chain variable regions is less striking (70%). Indeed, both 2C3 and BV04-01 are cross-reactive with single and double-stranded DNA and have specific affinity for d(pT)4 oligonucleotides.¹⁷ This finding led us to determine if phthalates elicit an autoreactive antibody response, particularly anti-self DNA antibody responses in BALB/c, NZB and autoimmune-prone NZB/W F1 mice. It appears that phthalate–keyhole limpet haemocyanin (KLH) immunization evokes complex immune responses that include anti-DNA antibodies in all these mice, but not so in KLH-immunized control mice. Analysis, at the clonal level using hybridomas from the fusion of phthalate-KLH primed BALB/c splenocytes, reveals that these hybridomas exhibit significant cross-reactivity towards phthalate, and DNA. This finding strongly indicates that phthalate as a conjugate with KLH is capable of inducing antibodies specific not only for the hapten itself but also for DNA, especially, against d(pT)4 oligonucleotides. However, repeated exposure to phthalate affects only the NZB/W F1 mice in terms of mortality. To our knowledge, this is the first report implicating a commercially used chemical phthalate in the induction of autoreactive immune response in BALB/c, NZB and NZB/W F1 mice.

Materials and methods

Animals

BALB/c female and male mice were bred in the animal facility of Indiana State University. Six-week-old NZB/W F1 and NZB female mice were purchased from Harlan Sprague Dawley (Indianapolis, IN). All mice were gender matched and used at the age of 8–12 weeks. All animals were housed in the animal facility of Indiana State University according to principles of laboratory animal care (NIH publication 85–23) followed under a specific protocol approved by the Animal Care and Use Committee (ACUC) of Indiana State University.

Chemicals

Reagents and kits were from the following sources: Dulbecco's modified Eagle's minimal essential medium (DMEM), horse and fetal bovine serum (Gibco/Invitrogen, Grand Island, NY); Calf thymus DNA, oligonucleotide d(pT)4, rabbit anti-mouse immunoglobulin–horseradish peroxidase (Ig-HRP) reagent, o-phenylenediamine (OPD), methylated bovine serum albumin (BSA) and BSA (Sigma Chemical Co., St. Louis, MO); synthetic homopolymers, d(pA)10, d(pT)10, d(pG)10, and d(pC)10 (IDT, Santa Clara, CA); random hexamer primers, pd(N)6 (Amersham Pharmacia Biotech Inc. Piscataway, NJ); KLH (Calbiochem, CA); 4-aminophthlate (Pfaltz and Bauer, Inc., Waterbury, CT); m-maleimidobenzoyl-N-hydroxysuccinimide ester (MBS), and Bradford protein assay kit (Pierce Chemicals, Rockford, IL); RNA isolation system (Biotecx Laboratory, Houston, TX), Maloney murine leukaemia virus reverse transcriptase (MMTV-RT, from Epicenter Technologies, Madison, WI), Taq polymerase (Stratagene, La Jolla, CA). cDNAs from spleen of BALB/c and NZB/W F1 were obtained by RT–polymerase chain reaction (PCR) as previously described.¹⁸ Other reagents used include an antioestriol mAb 1BF7 described before,¹⁹ and normal mouse immunoglobulin M and 2C3-Ig-HRP prepared as described.²⁰ 2C3-Ig V_H and V_L complementarity determining regions and an unrelated control peptide CEMSDMSFSKDWSFYI were obtained from Alpha Diagnostic International, Inc. (San Antonio, TX).

Phthalate–protein carrier conjugate

Phthalate–protein conjugates were prepared by coupling the diazotized 4-aminophthalate to proteins (KLH or BSA) as described by Ghosh et al.²¹ 4-Aminophthalate was diazotized by HNO₂ and the resulting diazonium salt, when added to KLH or BSA, was covalently linked to proteins by azo-coupling.

Cell fusions and hybridomas studied

B-cell hybridomas were established from BALB/c mice immunized twice or more with phthalate-KLH. Primed spleen cells were fused with non-secreting myeloma cells X63-Ag8·653, as described by Bloor et al.²² Several hybridoma clones that tested positive were propagated further and subcloned by the standard limiting dilution technique. Cells derived from a few selected clones were expanded in vitro and then injected intraperitoneally into pristane-primed BALB/c mice for ascites production.

Phthalate-binding hybridomas, L1B3, 1H5, and 2D2, are all IgM, κ obtained from these fusions. The hybridoma clone 2C3 secreting antiphthalate 2C3-Ig(γ1, κ), was generated previously from a similar fusion of phthalate–KLH primed BALB/c splenocytes with X63-Ag8.653 myeloma,^16,21 and has been used extensively as a tumour model,^16,20,23 These and other hybridomas clones were propagated in DMEM containing 10% horse serum (HS). Antibodies secreted in the supernatant or BALB/c ascites fluid were purified by salt fractionation using 50% ammonium sulphate precipitation followed by affinity chromatography on phthalate–sepharose or protein A-agarose columns and elution with 0·1 m citrate buffer (pH 3).

Peptides and peptide–protein conjugates

Idiopeptides, corresponding to 2C3-Ig V_H and V_L complementarity determining regions, and an unrelated control peptide CEMSDMSFSKDWSFYI were conjugated to carrier proteins KLH or BSA using m-maleimidobenzoyl-N-hydroxysuccinimide ester, as previously reported.²⁰ Peptide–KLH conjugate was used to immunize BALB/c mice and peptide-BSA or unconjugated peptide was used in enzyme-linked immunosorbent assay (ELISA).

Immunizations

Experimental mice (five or more) were grouped and primed intraperitoneally (i.p.) with phthalate–KLH conjugates emulsified in complete Freund's adjuvant. Control groups of mice were immunized with either adjuvant or KLH. Boosters in PBS were given at 10 days interval. The mice were bled through retro-orbital veins 5 days after each immunization.

ELISA

Direct and indirect ELISA were performed to assess and correlate different humoral responses as follows:

1. Screening of B-cell hybridomas: 121 hybridomas secreting antibodies were tested for their specificities either to phthalate or to DNA, on polyvinyl 96-well flat bottom plate (Falcon) coated first with methylated BSA, and then with 10 µg/ml of ds-DNA (calf thymus DNA). After blocking with 1% BSA/PBS overnight at 4°, supernatant from hybridoma cell culture was added to each well, and the plates were incubated for 1 hr at 37°. After washing with phosphate-buffered saline-containing 0·05% Triton X, rabbit anti-mouse immunoglobulin–HRP (50 µl) (at 1 : 3000 dilution) was added. Plates were incubated for 1 hr and washed again. Bound rabbit anti-mouse immunoglobulin–HRP was detected by addition of OPD. The reaction was stopped with 50 µl of 10% H₂SO₄, and the intensity of colour was read at 490 nm. For phthalate binding, 10 µg/ml of phthalate (as a conjugate of BSA) was used to coat ELISA plate.

2. Specificity of 2C3-Ig and L1B3-Ig – direct binding: binding of 2C3-Ig to DNA was assessed using ELISA plate coated with ds-DNA (either calf thymus DNA or cDNAs from spleen/salivary gland) or oligonucleotides (random hexamer primer). Subsequent steps included blocking, addition of 2C3-Ig and L1B3-Ig and development of colour with rabbit anti-mouse immunoglobulin–HRP as described above.

3. Specificity of 2C3-Ig: an inhibition assay was performed as described by Shoenfeld et al.²⁴ To evaluate the relative contribution of base composition to dsDNA binding by 2C3-Ig, synthetic homopolymers were employed. Dilutions of d(pT)4, d (pA) 10, d (pT) 10, d (pG) 10, or d (pC) 10 were separately incubated with 10 µg of 2C3-Ig overnight at 4°. The oligonucleotide d(pT)4 was chosen to ascertain if 2C3-Ig resembles in specificity to BV04-01 an auto-antibody from NZB/W F1 previously described.¹⁷ Determination of bound 2C3-Ig to the ds-DNA-coated plates was performed as described above.

4. Inhibition of binding of 2C3-Ig to phthalate by oligonucleotides: 2C3-Ig–HRP (50 µg/ml) as well as different dilutions of phthalate–BSA (1–50 µg) or d(pT)4 (1–100 nm) were added to either d(pT)4 or phthalate–BSA coated plates and incubated for 1 hr at 37°. Finally, determination of bound 2C3-Ig–HRP was performed by colour development with OPD, as described above.

5. Inhibition of binding of 2C3-Ig to anti-idiopeptide antibodies by antigens: anti-idiotype sera from idiopeptide-immunized mice were diluted to 1 : 200 in 0·01 m sodium bicarbonate buffer, pH 8 and added to the microtitre plates. Dilutions of phthalate–BSA or d(pT)4 (10–50 µg) were added to the coated plates along with 2C3-Ig–HRP and incubated for 1 hr. at 37°. After three washes with PBS/0·05% Triton-X, determination of bound 2C3-Ig–HRP was performed as described above.

Analysis of Vκ gene expression

Hybridoma clones were screened for their expression of Vκ1 genes as previously described.²³ Total RNA was isolated from hybridomas (Ultraspec RNA kit. Biotecx Laboratories Houston, TX). First strand cDNA for the κ light chain were synthesized using 4 µg RNA and random hexamer primer. PCR amplification was performed using Taq polymerase. For kappa light chain genes, a Vκ 1, framework region (FR) 1 specific oligonucleotide primer (FR1: CCTGCTAGCCATCCGTTTGATTTCCAGCTTGGTGCC) and the κ constant region primer (Cκ ACACTCATTCCTGTTGAA) were used.

Assessment of signs of severe nephritis

Blood urea nitrogen (BUN) and proteinuria of the control immunized BALB/c and NZB/W F1 mice were tested, according to the manufacturer's protocols, using Azostix (Bayer, Elkart, IN) and Multistix (Bayer, Elkart, IN), respectively.

Histology of kidney

Kidney tissues were fixed in 10% paraformaldehyde, and embedded in paraffin. These were then cut into sections (4-µm-thick) and stained with haematoxylin and eosin (H&E). This procedure was kindly performed by Dr Ronald M. Kohr, Chief of Pathology and Certified Pathologist at the Terre Haute Regional Hospital, Terre Haute, IN. Dr Kohr's analysis of the slides was performed in a double-blind fashion to determine pathology present in the sections.

Statistical analysis

Statistical analyses of all data were done by the paired Student's t-test (Sigma Plot). P < 0·05 was considered significant.

Results

Comparison of amino acid sequences of the variable regions of light and heavy chains of 2C3 and BV04-01 antibodies

We have previously described an anti-phthalate hybridoma 2C3 in the context of idiotype-specific antitumor immune response.^16,20,23 This led us to compare, using the NCBI database, the variable region-sequences of 2C3-Ig with the known sequences of an anti-DNA mAb, BV04-01, from (NZB/W) F1 mice (Fig. 1). This revealed a striking homology between the V_H and V_L sequences of antiphthalate 2C3 mAb from normal BALB/c mice and the corresponding sequences of BV04-01 mAb. The variable region of light chain of 2C3-Ig was found to be 98% homologous to the V_L of BV04-01.¹⁶ Only four amino acids (aspartate, histidine, leucine and alanine) of BV04-01 at positions 1, 34, 96, and 100 were substituted with tyrosine, tyrosine, tryptophan and glycine, respectively, on 2C3-V_L. However, the bases spanning the CDRs of both mAbs are virtually identical except at position 96. The variable regions of heavy chain of 2C3 exhibited 70% homology with the corresponding heavy chain of BV04-01 mAb. In the V_H sequences, there are differences of amino acids at 42 positions, which include numerous substitutions in all three CDRs.

Figure 1.

Comparison of amino acid sequences of the variable regions of light and heavy chains of 2C3 and BV04-01 antibodies. Dots indicate identity of amino acid sequences between BV04-01 with 2C3 and dashes denote regions lacking corresponding amino acid residues. CDR regions are underlined. A general numbering scheme introduced for immunoglobulins by Kabat et al.³⁴ is used.

Binding specificity of 2C3-Ig to ds-DNA and oligonucleotides

The striking sequence homology of 2C3-Ig with the DNA-binding auto-antibody BV04-01 suggests that this anti-phthalate mAb 2C3-Ig might also be cross-reactive with DNA. An inhibition ELISA was performed to determine if 2C3-Ig would have any specificity for DNA and, in particular, for any oligonucleotide (s) that would compete for its DNA or phthalate binding site. When various concentrations (10¹−10⁴ nm nucleotides) of the homopolymers were used, 1 µm d(pT)10 effectively inhibited (65% inhibition) 2C3-Ig binding to dsDNA, whereas the other three homopolymers (d(pA)10, d(pG)10, and d(pC)10) yielded <50% inhibition in 1 µm concentration (Fig. 2). Furthermore, there was also significant binding of 2C3-Ig with d(pT)4 oligonucleotide (Fig. 3a), reportedly the epitope for the auto-antibody, BV04-01.¹⁷

Figure 2.

Inhibition of 2C3-Ig binding to ds-DNA was assessed in the presence of varying concentrations of homopolymers, d(pT)10, d(pA)10, d(pG)10, d(pC)10 and random hexamer primer (ss-DNA). Significant inhibition was raised with d(pT)10.

Figure 3.

2C3-Ig bound not only to phthalate but also to d(pT)4. (a) Direct binding of 2C3-Ig to either phthalate-coated ELISA plate or d(pT)4-coated ELISA plate showed that 2C3-Ig had strong affinity to both phthalate and d(pT)4. (b) Inhibition of 2C3-Ig binding to d(pT)4 by phthalate. ELISA plate was coated with the oligonucleotide, d(pT)4, as described in Materials and Methods. 2C3 Ig-HRP conjugate was added in the absence of inhibitor (•), in the presence of intact 2C3-Ig (□), or phthalate (▴), as inhibitors. (c) The ELISA plate was coated with phthalate instead of d(pT)4, and tested as the same way as in (b).

Inhibition of phthalate binding to 2C3-Ig by oligonucleotides

To assess if the d(pT)4 oligonucleotide would compete with phthalate for the binding site on 2C3-Ig, an ELISA was performed using either phthalate- or d(pT)4-coated plates and the results (Fig. 3b) indicate that 50% inhibition of the binding of 2C3-Ig to d(pT)4 could be achieved with 1 µg phthalate-BSA. Likewise, the binding of 2C3-Ig to phthalate was inhibited by 50% after its adsorption with 10 µg d(pT)4 (Fig. 3c). A similar experiment was done using sera from BALB/c and NZB/W F1 mice that were immunized twice with phthalate–KLH. When these BALB/c sera were assayed after preincubation with various concentrations of phthalate–BSA or calf thymus ds-DNA, in both cases anti-phthalate and anti-DNA titres were abolished. In the case of NZB/W F1 mice also, the decline occurred but not to the same extent (data not shown), possibly because NZB/W F1 mice use a broad spectrum of light chain genes.²⁵

Inhibition of binding of 2C3-Ig to anti-idiopeptide antibodies by antigens

The amino acid sequences of the V_H and V_L of the 2C3-Ig have previously been determined in our laboratory.^20,23 To determine DNA binding specificity of the six CDRs from 2C3-Ig H and L variable regions, synthetic peptides corresponding to these regions were used as KLH conjugates to raise antisera in BALB/c mice. After initial determination of titers, it has been shown that all anti-idiopeptide immune sera can recognize the intact 2C3-Ig.²⁰ The anti-idiopeptide antisera were used at dilutions of 1: 200 to coat ELISA plates, and the assay was performed by addition of 2C3-Ig in the presence or absence of phthalate. Whether there was any binding was assessed using 2C3-Ig-HRP. Figure 4a indicates that phthalate inhibited the binding of 2C3-Ig to all six anti-CDR antisera to the extent of 60–70% (Fig. 4a). A similar range of inhibition was also observed in the presence of random hexamer oligonucleotides (data not shown). However, when assay was performed using d(pT)4 as the inhibitor rather than phthalate, only those peptides corresponding to the variable domains of light chains yielded the most significant inhibition (Fig. 4b). The results summarized in Table 1show that with respect to-V_L peptides, antibodies to V_L1 and V_L2 were inhibited the most by d(pT)4, while lowest inhibition (I₅₀ values) for phthalate was noted with anti-V_H3 and anti-V_L3 antibodies.

Figure 4.

Binding of 2C3-Ig to anti-idiopeptide antibody was inhibited by ligands. ELISA plates were coated with each idiopeptide-specific serum antibody (1 : 200 dilution), to which: (a) 2C3-Ig–HRP conjugate was added along with various concentrations of phthalate as inhibitor. (b) 2C3-Ig–HRP conjugate was added along with various concentrations of d(pT)4 as inhibitor (1 nm d(pT)4 = 1·5 µg).

Table 1.

Distribution of CDR regions in binding of 2C3-Ig to phthalate and d(pT)4

	I50*
	I50*
Anti-idiopeptide antibodies	Phthalate(µg)	d(pT)4 (µg)†
Anti-VH1	9·0	>100‡
Anti-VH2	5·2	>100‡
Anti-VH3	1·1	>100‡
Anti-VL1	4·8	9·5
Anti-VL2	3·2	6·2
Anti-VL3	2·8	52·1

^*Inhibition index (I₅₀) is defined as the amount of ligand required to achieve 50% inhibition of 2C3-Ig binding to anti-idiopeptide antibodies, and extrapolated from Fig. 7(a, b).

^†d(pT)4 1 nm = 1·5 µg.

^‡50% inhibition was not detected within experimental micrograms of inhibitor.

Induction of anti-DNA antibody response with phthalate

The cross-reactivity of anti-phthalate 2C3-Ig with DNA or oligonucleotides led us to determine if it was an isolated case or a general event for any phthalate-induced immune response. To test this, groups of BALB/c, NZB, and NZB/W F1 mice were immunized with phthalate–KLH three times at 10 day-intervals, and their serum titres of anti-phthalate and anti-DNA antibodies were determined using ELISA. It is apparent from the results in Fig. 5(a and b) and that although anti-phthalate antibody levels increased in all groups after each immunization, the anti-DNA antibody response varied considerably among the mouse strains. Notably, anti-DNA antibody level was modest in BALB/c mice after the first two immunizations but it declined after third immunization. However, anti-phthalate response in BALB/c mice increased further after the third immunization (Fig. 5a, b). In the case of NZB parental strain, phthalate immunization led to a modest increase in anti-DNA response that leveled off after the third immunization. In contrast, repeated immunizations of NZB/W F1 mice resulted in significantly higher levels of both anti-phthalate and anti-DNA antibody responses. Moreover, the latter exhibited higher mortality rate than the control group receiving only KLH, as 50% of phthalate-immunized NZB/W F1 mice died by 4·5 months while the control group lived on average longer than 8 months (data not shown). Because the reason for this high mortality was not clear, it was of interest to determine if any clinical signs of nephritis in NZB/W F1 mice would be evident with the development of phthalate-induced anti-DNA antibody response. To assess this, urinary protein and BUN levels were measured 2.5 months after a regimen of the three weekly immunizations with phthalate–KLH. The results presented in Table 2 reveal that urinary protein level was almost fourfold higher in NZB/W F1 mice than in CFA-immunized NZB/W F1 or BALB/c mice. In addition, there was a clear indication of azotomia, as BUN level registered threefold higher in experimental mice (Table 2). For further confirmation of this adverse effect of phthalate, we also examined kidney tissues for histopathological changes in affected mice. As shown in Fig. 6, kidneys from phthalate-immunized NZB/W F1 mice had the most severe infiltration of monocytic cells.

Figure 5.

Immunization with phthalate–KLH conjugates induces both anti-DNA and anti-phthalate antibody response in BALB/c, autoimmune-prone NZB/W F1, and parental NZB mice. The sera were tested at 1 : 200 dilution. The results represent average of two separate experiments (n = 8 mice in two experiments). (a) Anti-phthalate antibody levels in sera of mice immunized with phthalate–KLH were tested on phthalate–BSA-coated ELISA plates. (b) The anti-DNA antibody levels of the same groups. Sera obtained from BALB/c (adjuvant and KLH primed) and NZB/W F1 mice (unprimed and KLH primed) were used as control. *Significant decrease of anti-DNA antibody was observed in the serum of BALB/c mice at third immunization. (P < 0·05). 1. BALB/c immunized with phthalate-KLH conjugate. 2. NZB/W F1 immunized with phthalate-KLH conjugate. 3. NZB immunized with phthalate-KLH conjugate. 4. BALB/c immunized with adjuvant (CFA) only. 5. BALB/c immunized with KLH. 6. NZB/W F1 immunized with KLH.

Table 2.

Clinical tests for renal failure in BALB/c and NZB/W F1 mice that were immunized with phthalate–KLH and emulsified in CFA

	NZB/W F1
	NZB/W F1
Tests	CFA only	Phthalate–KLH with CFA	BALB/c Phthalate–KLH with CFA
*Anti-DNA antibodies (Absorbance @490 nm)	0·2	1·22	0·34
Proteinuria (mg/dl)	76·66	314·31	43·33
BUN (mg/dl)	15·33	52·5	20·1

All results above represent the average of three separate experiments (n = 3) obtained two and a half months following immunization at 7 months of age. The measurements were done as described under Materials and Methods.Serum anti-DNA antibody titre was measured at 1 : 200 dilution.

Figure 6.

Renal histopathology. Kidney tissues of three groups of mice at 7 month of age were examined following three weekly immunizations. (a) BALB/c mice immunized with phthalate–KLH emulsified in CFA; (b) NZB/W F1 mice immunized with CFA only; (c) NZB/W F1 mice immunized with phthalate-KLH emulsified in CFA. There were clear indications of increased leucocytic infiltration only in NZB/W F1 mice that were immunized with phthalate-KLH (c). haematoxylin–eosin stain; magnification, ×100.

Hybridomas secreting mAbs with specificity for phthalate/DNA

To further characterize phthalate-induced anti-DNA response in non-autoimmune prone BALB/c mice, hybridomas were generated in six separate fusions. ELISA was used to screen the resulting clones and subsequent sub clones for anti-phthalate and anti-DNA antibody specificity. Out of a total of 121 phthalate-specific hybridomas generated, 14 stable clones were selected for further analysis (Table 3). Only one mAb, secreted by 3D7 clone was IgG₃, κ; mAbs of the other 13 were IgM, κ.

Table 3.

Summary of six fusions

		Specific clones producing high titre of mAbs recognizing
		Specific clones producing high titre of mAbs recognizing
Fusion number	*Fusion efficiency	DNA	Phthalate	DNA and phthalate
1	72·9	1G6 (IgM, κ)	1H5(IgM, κ)	1A5(IgM, κ)
2	32·29	2G10(IgM, κ)	2D2(IgM, κ)	2G4(IgM, κ)
3	33·33	3B4(IgM, κ)		3D7(IgG3, κ)
4	20·83			4A9(IgM, κ), 4C4(IgM, κ)
5	22·92			5C10(IgM, κ)
6	15·63			6D2(IgM, κ), 6D7(IgM, κ)

^*Fusion efficiency (%) = (Number of antibody-positive wells/total wells) × 100.

These hybridomas were propagated as ascites in BALB/c mice, and the fluids drawn from these mice were assayed for anti-phthalate or anti-DNA activity on phthalate–BSA or only DNA-coated plates. The results shown in Fig. 7 indicate that only three (1G6, 2G10, 3B4) hybridomas showed high specificity for DNA, the two other (6D7, 2D2) primarily for phthalate, and the rest for both phthalate and DNA.

Figure 7.

Ligand (phthalate or DNA) binding with supernatants of hybridoma clones. White bars represent the binding with DNA; black bars represent that with phthalate. Antibodies after salt fractionation using 50% ammonium sulphate were used in ELISA at 50 µg/ml. All samples were assayed for binding to calf thymus DNA and phthalate. Results are average OD₄₉₀ (≥0·4) of two experiments, with 2C3-Ig, as positive control, and isotype-matched 1BF7 as negative control.

Anti-phthalate hybridoma antibodies bind to DNA

To understand if the dual specificity observed above was due to cross-reactivity and not due to presence of contaminating clones, we repeatedly subcloned and analysed one anti-phthalate hybridoma clone L1B3 (µ, κ) and compared with 2C3 (γ1, κ). Both these clones express identical light chains, as determined by serology using anti-2C3-specific Vκ antibody and also by RT–PCR shown as below. As shown in Fig. 8a, anti-phthalate L1B3 antibody reacted with calf thymus DNA albeit to a lesser extent than 2C3. An unrelated antioestriol mAb, 1BF7, used as a control, did not bind to DNA or phthalate. Nor was there any binding of DNA with normal BALB/c mouse sera. It is noteworthy that 2C3-Ig showed specificity not only to calf thymus DNA but also to autologous ds-cDNAs from spleen (Fig. 8b)

Figure 8.

Cross-reactivity of monoclonal antiphthalate antibodies with DNA. (a) 2C3-Ig (γ1, κ) and L1B3-Ig (µ, κ) obtained from antiphthalate mAb producing hybridomas showed specificity to ds-DNA from calf thymus. (b) These two mAbs showed similar results to autologous cDNA from spleen of normal female BALB/C mice. 1BF7 and normal IgM antibodies were used as isotype-matched negative control. Results expressed as mean OD (490 nm) of triplicates ± SD are average of two separate experiments.

Vκ gene usage

Among the other 13 hybrids, five clones expressed light chains encoded by Vκ1 gene (Fig. 9). Two Vκ1 expressing hybridomas, 4A9 and 4C4, reacted equally with both phthalate and DNA. The other three (IG6, 2G10 and 3B4) were essentially specific for DNA.

Figure 9.

Vκ1 gene-specific mRNA expression in hybridoma clones. 13 hybridomas secreting either anti-DNA or anti-phthalate mAbs were screened for Vκ1 gene usage in their light chains. As described, the total RNA prepared from the hybridomas were reverse-transcribed using Vκ1-specific primers by RT–PCR and analysed by agarose gel electrophoresis. Lanes corresponding to each group (A to D) are as follows: A, 1: 100 bp marker, 2: Negative control minus reverse transcriptase, 3: 2C3, 4: L1B3. B, 1: 1A5, 2: 1G6, 3: 2G4, 4: 1H5. C, 1: 3D7, 2: 4A9, 3: 4C4, 4: 2G10. D, 1: 3B4, 2: 2D2, 3: 5C10, 4: 6D2.

Discussion

This study has focused on activation of autoreactive B-cell clones by an environmental hapten in both autoimmune-prone and non-autoimmune-prone mice. The consequence of such activation in susceptible mice appears to be deleterious. Although as a paradigm, the clonal deletion of autoreactive cells in bone marrow is considered an event of central importance, it does not fully account for the absence of protection against immune responses to self-antigens even in apparently non-susceptible individuals. Autoreactive B cells could escape to the periphery and exist as mature and recirculating lymphocytes. They are then amenable to activation as a result of exposure to a vast array of self-mimicking foreign antigens.⁹ Although no unusual set of variable regions in Ig genes is known to be associated with autoantibodies, the presence or induction of anti-DNA antibody is the most evident abnormality of lupus-like disease in susceptible mice.^26,27 These autoimmune disorders could be the results of either defects in B-cell regulation or cross-activation by foreign antigens. Plants and microbes are a likely source of these foreign antigens.^7,8 However, unlike these complex antigens, phthalates are relatively simple chemicals and used widely as solvents in various cosmetics, and as plasticizers to soften rigid PVC in toys and medical products. Although there has been speculation as to whether exposure to phthalate-related compounds can evoke adverse immune reactions,¹⁵ there is, to our knowledge, no report linking phthalate with induction of autoreactive B lymphocytes. This study clearly suggests possible involvement of phthalate in the activation of such B lymphocytes that produce pathogenic anti-DNA antibodies in autoimmune-prone mice. Furthermore, while there is induction of anti-DNA humoral responses to phthalate in all three strains of mice to a varying degree. NZB/W F1 mice respond the best, and BALB/c the least. Most notably, these NZB/W F1 mice develop significant kidney pathology in terms of nephritis within 3–4 months following phthalate administration and also experience a high mortality rate.

Simultaneous elicitation of both anti-DNA and anti-phthalate response suggests that cross-reactive antigens are involved. This is borne out by our analyses of phthalate-immune sera in these mice as well as by specific screening of anti-phthalate hybridomas generated from primed BALB/c mice. Specifically, the findings with two phthalate-specific mAbs, from BALB/c hybridomas, 2C3 and LIB3, provide further support for this conclusion. These mAbs show considerable affinity not only for phthalate but also for DNA. Of the two mAbs, 2C3-Ig clearly exhibits stronger affinity for both phthalate and DNA, and the sequences of its V_H and V_L have already been determined.^20,23

Further support for phthalate as the inducer of anti-DNA autoreactive B-cell clones in BALB/c mice comes from comparisons of the sequences of 2C3-Ig with known sequences in the database of NCBI. There is a striking homology in the Ig gene sequence of 2C3 with that of BV04-01 auto-antibody, particularly in the light chain. In addition, 2C3-Ig, like BV04-01 shows an affinity for d(pT)4 oligonucleotide as it competitively binds both d(pT)4 and phthalate. Further support comes from binding studies with antibodies to the idiopeptides corresponding to CDRs of both 2C3 V_H and V_L. The binding of 2C3-Ig to anti-idiopeptide antibodies is strongly inhibited by phthalate as well as by d(pT)4 nucleotide. This suggests that these anti-idiotope antibodies are binding-site related. The CDR-peptides of the three V_L regions, particularly the CDR1 peptide of Vκ region, represents the contact sites for d(pT)4 nucleotide.

Usually antibody heavy chains have a significantly greater involvement in antigen binding specificity than the corresponding light chain.²⁸ However, based on available evidence from murine models, the binding affinity of anti-DNA antibody may also be determined by the V_L gene usage.²⁹ It was noted out by Corbet et al.³⁰ that the Vκ1 gene family is extensively used in a number of antibodies of different specificities, particularly in anti-DNA antibodies that spontaneously arise in MLR and NZB/W F1 strains. However, investigations by Spatz on B cells expressing transgene-encoded autoantibodies show that anti-DNA expressing B cells from non-autoimmune prone mice exclusively use Vκ1 genes; in contrast, those from NZB/W F1 mice use a broad spectrum of light chain genes.²⁵ In this study, we observed that a number of antiphthalate hybridomas express Vκ1, the same germ-line used in 2C3, implying that identical κ chains represent the common denominator for DNA-specificity of phthalate-specific mAbs. Vκ1 gene usage seems to be common for most anti-phthalate responses in mice.³¹

We conclude on the basis of the evidence presented that the 2C3 and other hybridomas described here represent autoreactive B-cell clones activated by phthalate. B-1 cells often contribute to the production of serum immunoglobulin and natural auto-antibody.³² In mice, ‘B-1 lymphocytes’ represent fetal/neonatal B-cell development; such development and maintenance appear to depend critically on positive selection. Interestingly, in CD22-deficient mice B1-cells are activated producing autoantibodies to ds-DNA without any overt autoimmune disorder.³³ We hypothesize that phthalate-induced anti-DNA antibodies result from activation of an autoreactive B-1 cell population. In a future study, we will isolate DNA from selected CD5⁺ B-cell population of phthalate-immunized BALB/c and other strains, and determine the expression of Vκ1 gene in enriched phthalate-specific B cells.

Finally, it is interesting to note that although phthalate activates autoreactive B-cells in BALB/c mice, they do not develop the symptoms of autoimmune disorder, as do NZB/W F1 or other susceptible mice. Indeed repeated immunizations with phthalate do not augment an anti-DNA response in BALB/c mice. This however, is in quite contrast to the pathologic developments, mentioned earlier, that occur in autoimmune-prone NZB/W F1 mice. This suggests that there are regulatory mechanisms operating in non-autoimmune-prone mice. Indeed, our ongoing study suggests that non-autoimmune prone mice overcome the effect of autoimmune response by regulating the level of an anti-DNA response through mediation of idiotype-specific cytotoxic T lymphocytes.