Immunoglobulin V_H-gene usage of autoantibodies in mercuric chloride-induced membranous glomerulopathy in the rat

Abstract

Brown-Norway (BN) and Dorus Zadel Black (DZB) rats develop a T-cell-dependent membranous glomerulopathy (MGP) with high proteinuria and antiglomerular basement membrane (GBM) autoreactive antibodies (Abs), upon exposure to mercuric chloride (HgCl₂). Laminin is an important autoantigenic target of the anti-GBM Abs, absorbing ≈30% of the anti-GBM reactivity. Although many anti-GBM Abs have undergone isotype switching, it is currently unclear whether affinity maturation occurs during the HgCl₂-induced autoimmune response. To address this question we analysed the rearranged immunoglobulin heavy chain variable-region genes (V_HDJ_H regions) of 15 mAbs that were previously obtained from HgCl₂-treated rats. Seven of these mAbs exhibit reactivity towards laminin. Our study showed that the V_H-gene usage of antilaminin mAbs is largely restricted to the PC7183 V_H-gene family (six out of seven). In addition, we demonstrated that at least three out of six laminin reactive and five out of six non-laminin-binding mAbs are encoded by germline V_H genes (a total of eight out of 12 mAbs). Of the eight mAbs that are encoded by germline V_H genes, seven are of a non-immunoglobulin M (IgM) isotype, indicating that isotype switching has occurred in these mAbs in the absence of somatic mutations. The mutations observed in the V_H genes of the four remaining mAbs do not provide strong evidence for antigenic selection. The data support the notion that B cells in this model of MGP are not subjected to affinity maturation and probably result from polyclonal B-cell activation.

Introduction

After repeated administration of low doses of mercuric chloride (HgCl₂), Brown-Norway (BN) and Dorus Zadel Black (DZB) rats develop a membranous glomerulopathy (MGP) with proteinuria and antiglomerular basement membrane (anti-GBM) autoreactive antibodies (Abs).^1,2 The Abs that are produced upon HgCl₂ treatment exhibit reactivity towards exogenous (non-self) antigens, such as sheep red-blood cells and 2,4,6-trinitrophenyl (TNP),³ and/or self-antigens, such as nuclear antigens (i.e. DNA)³ and extracellular matrix components, like laminin, type IV collagen, fibronectin, nidogen and heparan sulphate proteoglycans (HSPG).^1,2,4,5 Although HgCl₂ induces a striking increase in both immunoglobulin G1 (IgG1) and immunoglobulin E (IgE) levels in serum, Abs to renal autoantigens are typically of the immunoglobulin M (IgM), IgG1 or IgG2a isotype, but never of the IgE isotype.^2,3,6–9 Anti-GBM Abs play an important role in the pathogenesis of HgCl₂-induced MGP.^2,5,9,10 Many of the anti-GBM Abs bind to laminin, as illustrated in the study of Aten et al.,² which showed that glomerular-bound IgG in HgCl₂-induced MGP can be absorbed for ≈30% by laminin of the murine Englebreth-Holm-Swarm (EHS) mouse saccoma (i.e. laminin-1). In addition, reactivity to laminin is also frequently observed among anti-GBM monoclonal Abs (mAbs), as shown for the Hg^11,12 and MEC⁹ series of mAbs. The Hg and MEC hybridomas were obtained from BN and DZB rats, respectively, and were selected on the basis of binding to crude collagenase-digested rat GBM⁹ and kidney sections of normal untreated rats.^9,11,12 The preferential binding of IgG anti-GBM Abs to laminin, especially to the P1-fragment of laminin,⁹ suggests that the production of antilaminin Abs in HgCl₂-induced MGP might be antigen-driven.

B cells that are activated in a T-cell-dependent immune response may develop along a pathway leading to the formation of either plasma cells or memory B cells.¹³ In contrast to naive IgM⁺ B cells, memory B cells are selected for high-affinity binding to the antigen. This is achieved by somatic mutations in the immunoglobulin V-region genes and occurs in the microenvironment of the germinal centre.¹³ Under the influence of cytokines secreted by T cells and direct cell–cell contacts (i.e. CD40–CD40 ligand interaction) the germinal centre B cells are urged to switch to downstream isotypes (e.g. IgG).¹⁴ In mice the majority of IgG⁺ B cells have somatically mutated immunoglobulin V-region genes, which designate these B cells as memory cells.¹⁵ In the present study we analysed the immunoglobulin heavy-chain V-region genes (V_HDJ_H rearrangements) encoding anti-GBM Abs in the model of HgCl₂-induced MGP in the rat, i.e. Hg and MEC mAbs. In this manner we could address the question of whether antilaminin autoreactive Abs are encoded by a restricted set of V_H genes, and whether these Abs are encoded by somatically mutated (and antigen-selected) V_H genes or by V_H genes in unmutated (germline) configuration. Similarly to mice and humans, immunoglobulin heavy- and light chains in rats are made after sequential rearrangement of V_H-, D- and J_H-region genes, and V_L- and J_L-region genes, respectively.¹⁶ V_H genes from mouse and rat display a high degree of similarity. We showed previously that V_H genes of the rat can be subdivided into V_H-gene families similar to those of mice.^16–18 At present, sequence analysis has revealed the existence of at least 10 V_H-gene families in rat (PC7183, X24, Vh11, S107, J606, Vh10, Q52, 3609, J558 and VGAM3.8).^16–18 In this report we show that the antilaminin response in HgCl₂-induced MGP in rats seems to be largely restricted to the PC7183 V_H-gene family and that most, if not all, Abs are encoded by germline V_H genes.

Materials and methods

MEC and Hg hybridomas

Generation of the DZB hybridomas MEC1–MEC8, and the BN hybridomas Hg15–Hg18 has been described previously.^9,11,12 The IgE-producing hybridomas Hg12, Hg32 and Hg33, were also generated from HgCl₂-treated rats (BN).¹⁹ The cells were maintained in culture in RPMI-1640 (Life Technologies, Glasgow, Strathclyde, UK) supplemented with (10%) fetal calf serum (FCS; Life Technologies), 60 mg/ml of gentamycin, sodium pyruvate, glutamine, β-mercaptoethanol, hypoxanthine and thymidine, at 37° in an atmosphere of 5% CO₂.

Molecular cloning of rearranged immunoglobulin V_HDJ_H regions of the rat

Total RNA was isolated from ≈10⁷ hybridoma cells, as described by Bos & Meeuwsen,²⁰ or by using TRIzol reagent (Life Technologies), according to the manufacturer's instructions. Primers used for first-strand cDNA synthesis and the polymerase chain reaction (PCR) are listed in Table 1. First-strand cDNA synthesis was performed using 1–5 µg of total RNA with primers specific for the rat µ constant region (Cµ2 or J_H-Cµ1), or a mouse derived γ constant-region primer (Cγ). The cDNA reactions were carried out at 42° for 60 min, in a 20-µl reaction volume containing 50 mm KCl, 20 mm Tris-HCl (pH 8·4), 2·5 mm MgCl₂, 0·1 mg/ml of nuclease-free bovine serum albumin (BSA) (Pharmacia Biotech, Uppsala, Sweden), 20 U of RNAquard (Pharmacia Biotech), 10 U of avian myeloblastosis virus (AMV) reverse transcriptase (HT Biotechnology Ltd, Cambridge, Cambs., UK), and 1 mm each of dGTP, dATP, dTTP and dCTP. The reactions were terminated by incubation at 95° for 5 min and then cooled on ice. First-strand cDNA was then subjected to PCR amplification by using a mouse-derived universal V_H-gene specific primer (VH1BACK) and one of the constant-region primers used for cDNA synthesis (see above). VH1BACK is a V_H-gene consensus primer and is designed to amplify a large number of mouse V_H genes.²¹ The PCR reactions were performed in a 100-µl reaction mixture containing 35 pmol of VH1BACK primer, in combination with 120 pmol of Cµ2 primer or 60 pmol of J_H-Cµ1 primer or 90 pmol of Cγ primer, and 50 mm KCl, 20 mm Tris-HCl (pH 8·4), 2·5 mm MgCl₂, 0·1 mg/ml of nuclease-free BSA (Pharmacia Biotech), 0·5 U of Taq DNA polymerase (HT Biotechnology Ltd) and 0·2 mm each of dGTP, dATP, dTTP and dCTP. The PCR reactions were initiated by denaturation for 2 min at 94°, annealing for 2 min at 42° and elongation for 1 min at 72°, and followed by 39 cycles consisting of 1 min at 94°, 1 min at 55° and 1 min at 72°. PCR products were size-fractionated by SeaKem LE agarose gel electrophoresis (FMC BioProducts, Rockland, ME), purified by using a Biotrap (Schleicher & Schuell GmbH, Dassel, Germany), digested with PstI and BamHI, and finally force-cloned into the polycloning site of pUC19. The PCR products obtained using the Cγ primer were digested with PstI only and ligated into pUC19 (PstI/SmaI). JM83 competent cells were transformed by heat-shock and grown on colour selection plates containing an appropriate antibiotic. Colour-selected colonies were screened for plasmids containing inserts of the appropriate size and these plasmids were subsequently sequenced.

Table 1.

Primers used for cDNA synthesis, the polymerase chain reaction (PCR) and sequencing of immunoglobulin heavy chain variable-region genes of the rat

Primer name	Sequence (5′→3′)*	Position†	Reference
Universal VH primer
VH1BACK	XAGGTSMARCTGCAGSAGTCWGG	23	21
IgH constant-region primers
Cµ2	CTGGATCCTTTGGCCTCGAC	502
JH-Cµ1	GCGGATCCGAGGACTGACTCTCTGARGA	− 6
Cµ1·3	CCCTGGATGACTTCAGTGTTG	138
Cγ	GGGGCCAGTGGATAGAC	22	22
Cγ1·1	GGACAGGGATCCAGAGTTCCA	121
Cε1·1	CCGGATCCATGCTGCTCGTGTTCA	134
FITC-labelled sequencing primers
Cµ(FITC)	XATCAGACAGGGGGCTCT	44
Cγ(FITC)	XGGGGCCAGTGGATAGAC	22	22
Cε(FITC)	XGGGCTTCAAGGGATAGA	23

^*S = C or G; M = A or C; R = A or G; W = A or T; X=FITC or biotin. Restriction sites are underlined.

^†Relative position of the most 3′ nucleotide of the primer. For primer VH1BACK, the nucleotide position is according to ImMunoGeneTics (IMGT) nomenclature.²³ For the constant region and sequencing primers, the position is relative to the first nucleotide of exon 1.

FITC, fluorescein isothiocyanate.

Solid-phase DNA sequencing

First-strand cDNA reactions were performed in a 30-µl reaction volume containing 1–5 µg of total RNA, 5–10 pmol of constant-region primer (isotype-specific primers: Cµ1.3, Cγ1.1 or Cε1.1), 50 mm Tris-HCl (pH 8·3), 75 mm KCl, 3 mm MgCl₂, 10 mm dithiothreitol (DTT), 200 U of SuperScript II (Life Technologies) and 0·3 mm each of dGTP, dATP, dTTP and dCTP, at 45° for 30 min. The Cγ1.1 primer is designed on a homologous region between the exon 1 sequences of mouse Cγ1, Cγ2a, -2b, -2c, Cγ3^24–28 and the Cγ2c region sequence of the rat.²⁹ The Cµ1.3 and Cε1.1 constant-region primers are deduced from rat (MEC7 and MEC8) and mouse³⁰ µ exon 1 constant-region sequences, and from the published rat ε exon 1 constant-region sequence,³¹, respectively. RNA–cDNA hybrids were denatured at 94° for 5 min, extracted once with phenol–chloroform (1:1), once with chloroform and eventually precipitated in ethanol. After centrifugation, the DNA pellet was washed once with 70% ethanol, air-dried and dissolved in 20 µl of water. One-tenth of the cDNA obtained this way was amplified in a 50-µl reaction volume containing 35 pmol of 5′-biotinylated VH1BACK primer, 30–35 pmol of constant-region primer (see above), 20 mm Tris-HCl (pH 8·4), 50 mm KCl, 1·5 mm MgCl₂, 0·05% W-1 (Life Technologies), 2·5 U of Taq DNA polymerase (Life Technologies) and 0·2 mm each of dGTP, dATP, dTTP and dCTP. PCR cycles consisted of 2 min at 94°, 2 min at 55° and 1·5 min at 72°, for one cycle, followed by 34 cycles of 1 min at 94°, 1 min at 55° and 1 min at 72°. PCR products were sequenced at our local sequence facility (Department of Pathology and Laboratory Medicine, Division of Medical Biology, University of Groningen, Groningen, the Netherlands) using solid-phase DNA sequencing (Dynabeads M-280 Streptavidin; Dynal, Oslo, Norway) and 5′-fluorescein isothiocyanate (FITC)-labelled sequencing primers (Table 1). Sequences were determined using the Automated Laser Fluorescent (ALF) DNA Sequencer (Pharmacia Biotech).

V_H-gene analysis

Sequences used for alignments were obtained from the EMBL genebank (EMBL Outstation, European Bioinformatics Institute [EBI], Cambridge, Cambs., UK). The rat germline PC7183 reference sequences (PC series) are described in Dammers et al.³² and are available at the ImMunoGeneTics (IMGT) database (http://imgt.cnusc.fr:8104).²³

Results

General remarks of the mAbs obtained by HgCl₂-induced systemic autoimmunity

In the present study we analysed the rearranged immunoglobulin V_HDJ_H sequences expressed by 15 hybridomas that were previously generated from rats with HgCl₂-induced MGP (the MEC series from a DZB rat and the Hg series from BN rats). The hybridomas MEC1 to MEC8,⁹ and Hg15 to Hg18^11,12 were selected on the basis of reactivity to collagenase-digested rat GBM⁹ and kidney sections of normal untreated rats. Because of their staining pattern on normal kidney sections and their reactivity to a panel of renal proteins,^2,9,11,12 the mAbs MEC1 to MEC8, and Hg15 to Hg18, can be considered as a set of Abs representative for the autoreactive Abs within the glomerular eluate of HgCl₂-diseased rats. For most of these mAbs the reactivity to glomerular and/or tubular epithelium is caused by recognition of the matrix-protein laminin (Table 2). Four out of eight MEC mAbs react with laminin; MEC2, MEC3 and MEC5 bind preferentially to the P1 fragment of laminin-1, while MEC8 shows reactivity to the P1 and E4 fragments of laminin-1. MEC8 and also MEC6 are polyreactive Abs. In addition, most of the kidney-binding Hg mAbs (Hg15 to Hg18) also react with laminin (Hg15, Hg16 and Hg17) (Table 2). For Hg15 and Hg16, the binding to laminin is highly specific, while mAb Hg17 is polyreactive.³³ Interestingly, three of the Hg mAbs (Hg15, Hg16 and Hg17) share cross-reactive idiotypes (CRid⁺) with laminin-specific immunoglobulin deposited along the glomerular capillary wall during HgCl₂-induced MGP.¹² From criss-cross inhibition experiments, two major cross-reactive idiotypes (CRid15 and CRid17) were detected on these three laminin-binding Hg mAbs, which might be indicative for similar V-region gene usage. In addition to the kidney-binding Abs described above, we also analysed the V_H genes expressed by three IgE-producing hybridomas that were obtained in the HgCl₂ model of MGP (Hg12, Hg32 and Hg33).^19,35 mAb Hg12 reacts with DNP, actin and renin. Hg33 binds to DNA, DNP, myosin and tubulin. Hg32 did not bind to any of the antigens tested.

Table 2.

Characteristics of monoclonal antibodies (mAbs) derived from mercuric chloride (HgCl₂)-induced autoimmune membranous glomerulopathy (MGP)

mAb*	Isotype*	CS-kidney*	Reactivity*	VH-gene family†	Germline‡	Mutation frequency§	H-CDR3¶	JH	Accession number
MEC
1	IgG2a	+	?	PC7183	PC-1	0·4	12	2	Z75904
2	IgG1	+	(P1-)laminin-1	PC7183	PC-22	2·5	10	2	Z75905
3	IgM	+	(P1-)laminin-1	PC7183	PC-29	6·3	9	2	X78897
4	IgG2a	+	?	Vh11	YTH655(5)6	0·4	–	–	X78894
5	IgG1	+	(P1-)laminin-1	Vh11	–**	> 10	–	–	X78896
6	IgM	+	actin, etc.	J558	GLVH18	2·7	13	3	X78893
7	IgM	+	?	J558	–	> 10	9	2	X68782
8	IgM	+	(P1-/E4-)laminin-1, actin, etc.	PC7183	PC-22	7·9	10	3	X78895
Hg
12	IgE	–	DNP, actin, renin	PC7183	PC-1	0·4	12	4	Z75897
15	IgG1	+	Laminin	PC7183	PC-21	0	9	2	Z75898
16	IgG1	+	Laminin	PC7183	PC-24	0·8	9	3	Z75899
17	IgG1	+	Laminin, etc.	PC7183	PC-21	0	6	2	Z75900
18	IgM	+	?	Q52	RNIGVJL	0·4	9	2	Z75901
32	IgE	–	?	X24	–	> 10	14	3	Z75902
33	IgE	–	DNP, DNA, myosin, tubulin	PC7183	PC-14	0·8	7	2	Z75903

^*Described previously.⁹,¹²,³³ CS, cryostat sections.

^†Mouse V_H-gene family nomenclature.¹⁶–¹⁸,³²

^‡The rat germline PC7183 V_H reference-genes (PC series) are described by Dammers et al.³² and are available at the IMGT database (http://imgt.cnusc.fr:8104).²³ Genes YTH655(5)6, GLVH18 and RNIGVJL were obtained from the EMBL database (GenBank accession numbers: M87785, X86663 and L07405, respectively).

^§Mutation frequency=(number of mutations÷number of nucleotides compared) × 100%. The Taq DNA polymerase error rate was not subtracted from the values.

^¶The numbers of codons of complementarity-determining region of the heavy chain (H-CDR3) were determined according to Kabat et al.³⁴

^**No potential germline gene could be identified (< 90% identity).

DNP, dinitrophenyl; IgG, immunoglobulin G; IgM, immunoglobulin M; IgE, immunoglobulin E.

Determining immunoglobulin V_HDJ_H-region sequences of the MEC and Hg series of hybridomas

In order to determine the V_HDJ_H-µ-region sequences of IgM Ab-producing MEC hybridomas, we amplified the V_HDJ_H-µ transcripts expressed by MEC7 and MEC8 by reverse transcription–PCR (RT–PCR). This was performed by using a mouse-derived universal V_H-gene primer (VH1BACK)²¹ in combination with primer Cµ2. The Cµ2 primer was designed based on the published immunoglobulin µ C_H2-region sequence of the rat.²⁹ By using this primer combination we were able to establish the rat µ C_H1-region sequence, which has not been described previously (Fig. 1). Based on this µ C_H1-region sequence and the germline J_H-gene sequences of the rat,^29,36 we designed a new 3′-region primer (J_H-Cµ1). The V_HDJ_H regions of MEC3 and MEC6 (both IgM) were subsequently amplified by using primers VH1BACK and J_H-Cµ1. The V_HDJ_H-region sequences of the IgG-producing hybridomas MEC4 and MEC5 were determined after PCR amplification using primers VH1BACK and Cγ. Owing to an internal PstI restriction site in the V_H regions of MEC4 and MEC5, we obtained only part of the V_H-gene sequence of these hybridomas. V_HDJ_H-region sequences of the hybridomas MEC1, MEC2 and the Hg series were determined directly from the PCR fragments using solid-phase DNA sequencing (see the Materials and methods).

Figure 1.

Nucleotide cDNA sequence of the rat immunoglobulin µ C_H1 domain and part of the C_H2 domain with deduced amino acid sequence. Sequence is established from MEC7 and MEC8 (see Table 2 for EMBL accession numbers). The mouse germline µ C_H1-region sequence (EMBL accession number: V00818)³⁰ and the rat germline µ C_H2-region sequence (CH2; EMBL accession number: M13800)²⁹ are aligned for comparison. The µ C_H1-region sequences of rat and mouse share 90·5% identity upon alignment, which is the same extent of homology as previously shown for the µ C_H2 regions of these species (90%).²⁹ Primer Cµ2 is shown in lower case characters, with the BamHI recognition site underlined.

V_H genes encoding MEC and Hg mAbs: the majority of antilaminin mAbs are encoded by PC7183 V_H genes

The members of the MEC and Hg series of hybridomas are derived from distinct B cells, as their V_H-D-J_H junctions represent unique rearrangements of the complementarity-determining region of the heavy chain (H-CDR3) and reveal no clonal relationship (Fig. 2). The V_HDJ_H-region characteristics of the MEC and Hg hybridomas are shown in Table 2. MEC and Hg mAbs are encoded by V_H genes that originate from five different V_H-gene families (PC7183, J558, V_H11, Q52 and X24). Interestingly, seven out of 12 (58%) kidney-binding (anti-GBM) mAbs are encoded by PC7183 V_H genes, and this percentage is even higher among the antilaminin mAbs, where six out of seven mAbs (86%) are encoded by PC7183 V_H genes. Among the antilaminin mAbs, only MEC5 (V_H11) is not encoded by a PC7183 V_H gene. The PC7183 V_H genes of the antilaminin MEC mAbs have less than 95% identity upon comparison to each other, indicating that the V_H-gene utilization of PC7183-encoded antilaminin MEC mAbs is not restricted to very closely related PC7183 members (Fig. 3). In contrast to the antilaminin mAbs, a low frequency of PC7183 V_H-gene usage (25%) is observed among the non-laminin reactive mAbs. Among the non-laminin-binding MEC mAbs, only MEC1 is encoded by a PC7183 V_H gene. The other non-laminin MEC mAbs are encoded by V_H genes that belong to the Vh11 (MEC4) and J558 (MEC6 and MEC7) V_H-gene families.

Figure 2.

Nucleotide sequences (a) and deduced amino acid sequences (b) of the V_H-D-J_H junctions (complementarity-determining region 3 of the heavy chain [H-CDR3]) of the immunoglobulin heavy chain transcripts expressed by the MEC and Hg series of hybridomas. Sequences span the region from the most 3′-terminal end of the V_H gene (TGT = conserved Cys on position 92) to the conserved J_H TGG (Trp on position 103), according to Kabat et al.³⁴ The H-CDR3 is boxed. In some rearrangements, codons 95 (at the start of H-CDR3) and 96 may be encoded by the most 3′ V_H nucleotides. Because the 3′-terminal end of most rat V_H genes have not been resolved at the genomic level, the 3′-terminal end of V_H genes is determined upon sequence alignment. Underscored characters represent potential P nucleotides. N nucleotide additions could not be established in these V_H-D-J_H junctions, because only one D gene (D_Q52) is presently known in the rat²⁹ and D_Q52 is not used among these rearrangements. Nucleotide ambiguities are according to International Union of Biochemistry (IUB) group code; ‘?’ identifies undefined amino acids.

Figure 3.

Nucleotide sequence comparison of the V_H genes expressed by the MEC and Hg series of hybridomas. Numbers in the homology matrix represent the percentage of identity between two individual monoclonal antibodies (mAbs) and is revealed from maximal sequence overlap between nucleotide position 1 and 312, according to IMGT nomenclature.²³ Positions with missing nucleotides are excluded by pairwise deletion. MEC1 is arbitrarily placed at the top. Laminin-binding mAbs and their homology values are shown in bold with a grey-shaded background.

In agreement with the observation of cross-reactive idiotypes among the Hg15, Hg16 and Hg17 mAbs (suggesting that these mAbs are encoded by closely related V-region genes), all three mAbs are encoded by PC7183 V_H genes. More specifically, the V_H regions of Hg15 and Hg17 are nearly identical to each other. Both Hg15 and Hg17 are encoded by the same PC7183 V_H gene and J_H gene (J_H2), although they show no clonal relationship to each other because they differ at their V_H-D-J_H junctions (Fig. 2). Although Hg16 also expresses the same cross-reactive idiotypes as Hg15 and Hg17, and is also encoded by a PC7183 V_H gene, comparison of Hg16 with Hg15/Hg17 reveals a rather low identity (84%) (Fig. 3). Furthermore, Hg16 also expresses a different J_H gene (J_H3). These data clearly indicate that the CRid⁺ Hg mAbs are encoded by closely related, but not necessarily identical, V_H genes. In addition, these CRId⁺ Hg mAbs originate from distinct B cells and are not derived from one clonally expanded progenitor cell.

MAbs encoded by germline V_H genes predominate in the HgCl₂ response

The V_H genes expressed by the MEC and Hg hybridomas were compared to known rearranged and genomic V_H-gene sequences of the rat.^16–18,32 For each hybridoma, the most homologous rat germline V_H gene is shown in Table 2. To MEC5, MEC7 and Hg32 no potential germline V_H gene could be assigned (less than 90% identity with the most homologous V_H gene) and possible mutations in these mAbs could therefore not be determined. The majority of the V_H-gene sequences of the 12 remaining mAbs revealed rather low mutation frequencies (≤ 0·8%, representing approximately two mutations per 250 nucleotides). Mutations in these sequences are probably derived from Taq DNA polymerase errors and/or immunoglobulin H (IgH) haplotype differences. Therefore, we consider V_H-gene sequences with mutation frequencies of ≤0·8% as germline. Using this criterion, our data revealed that at least three out of six laminin-reactive mAbs (MEC5 excluded) and five out of six non-laminin-binding mAbs (MEC7 and Hg32 excluded) are encoded by germline V_H genes (a total of eight out of 12). Interestingly, seven out of eight of these germline V_H gene-encoded mAbs are of a non-IgM isotype, indicating that isotype switching has occurred in these mAbs in the absence of somatic mutations.

Only four (out of 12) remaining mAbs are possibly encoded by V_H genes that are somatically mutated (MEC2, MEC3, MEC6 and MEC8). The number of mutations in these mAbs far exceeds the number of mutations expected from Taq DNA polymerase errors, and therefore these mutations are probably generated somatically. Figure 4 shows the alignment of these four sequences with their most homologous germline V_H gene. The number of nucleotide exchanges ranges from six (MEC2) to 19 (MEC8). For MEC2, MEC3 and MEC6, mutations are observed predominantly within the H-CDRs and H-CDR flanking regions. The mutations in MEC8 seem to be more or less randomly distributed. As a result of antigen-driven selection, the distribution of mutations in V_H genes that encode for high-affinity mAbs generally reveals a statistically higher proportion of amino acid replacement mutations in the H-CDRs (antigen-binding sites) than in the framework regions of the heavy chain (H-FR).³⁷ In order to address whether the mutations in the V_H genes of mAbs MEC2, MEC3, MEC6 and MEC8 are selected by antigen, we computed the replacement (R) over silent (S) mutation ratios (observed R/S) of H-CDRs and H-FRs of the V_H genes expressed by these hybridomas, and compared them to the R/S mutation ratios theoretically expected from a random distribution of mutations. In addition, we calculated the probability (P) that an excess or scarcity of replacement mutations in H-CDRs and H-FRs were the result of coincidence (according to Chang & Casali³⁷). The observed and expected R/S mutation ratios and P-values of the four mutated mAbs are shown in Table 3. Only the mutations in MEC3 and MEC8 resulted in an excess of R mutations in the H-CDRs, and a scarcity of these mutations in the H-FRs (P < 0·05). Both mAbs carry a higher and lower number of R mutations in their H-CDRs and H-FRs, respectively, as expected. This indicates that the distribution of mutations in MEC3 and MEC8 is probably the result of selection. No selection for R mutations is revealed from the H-CDRs and H-FRs of MEC2 and MEC6. These hybridomas may therefore express V_H genes that either originate from another presently unknown germline gene or are somatically mutated without subsequent selection (randomly mutated), or are somatically mutated but do not reveal these R/S mutation characteristics typical for high-affinity mAbs. Thus, out of the 12 V_H genes analysed in the present panel of mAbs, possible evidence for antigen-selected somatic mutations is only found in two V_H genes (MEC3 and MEC8).

Figure 4.

Alignment of the V_H genes expressed by MEC2, MEC3, MEC6 and MEC8 with their most homologous germline counterpart (top sequence). Codon numbering is according to IMGT nomenclature.²³ Letter codes for nucleotide ambiguities are used as proposed by the International Union of Biochemistry (IUB), dashes indicate sequence identity and dots represent gaps occurring as a result of the IMGT classification. The reference germline PC7183 V_H-gene sequences PC-22 and PC-29 are described by Dammers et al.³²

Table 3.

Ratio of replacement (R) over silent (S) mutation (R/S) analysis of mutated V_H genes among the MEC monoclonal antibodies (mAbs)

			R/S mutation ratio‡
mAb	Germline*	Mutations†	Observed			Expected	P
MEC2	PC-22	6	FR:CDR:	4·0∞	(4/1)(1/0)	2·923·86	0·30·4
MEC3	PC-29	15	FR: CDR:	1·3∞	(4/3)(8/0)	2·953·21	0·010·001
MEC6	GLVH18	6	FR: CDR:	∞2·0	(3/0)(2/1)	3·783·42	0·30·2
MEC8	PC-22	16	FR: CDR:	0·6∞	(3/5)(8/0)	2·923·86	0·0010·002

^*The rat germline PC7183 V_H reference-genes (PC series) are described by Dammers et al.³² and are available at the IMGT database (http://imgt.cnusc.fr:8104).²³ Gene GLVH18 was obtained from the EMBL database (GenBank accession no.: X86663).

^†Number of mutations is revealed from nucleotide position 55 (codon 19) up to and including 297 (codon 99), according to the IMGT nomenclature.²³

^‡The R/S mutation ratio is the quotient of replacement (R) to silent (S) mutations. R/S mutation ratios are calculated separately for the complementarity-determining region of the heavy chain (H-CDR) and the framework region of the heavy chain (H-FR). H-CDR and H-FR regions are defined according to IMGT nomenclature.²³ Numbers in parentheses are the actual number of R and S mutations. The theoretical expected (inherent) R/S mutation ratio is the quotient of total possible R to total possible S mutations in the germline gene, as described by Chang & Casali.³⁷ The probability (P) that an excess or scarcity of R mutations in the H-CDR or the H-FR results solely from chance is negated by the significantly low probability values (P < 0·05) calculated according to the binomial distribution.³⁷

Discussion

Although many anti-GBM Abs in HgCl₂-induced MGP have undergone isotype switching, it was at present unclear whether these Abs are produced by B cells that have undergone affinity maturation. In the present study we analysed rearranged immunoglobulin V_HDJ_H-region sequences of a panel of hybridomas that were generated from HgCl₂-induced MGP diseased rats, and investigated whether these mAbs were encoded by somatically mutated and (auto)antigen-selected V_H genes. Our results show that the kidney-binding Abs in HgCl₂-induced MGP are predominantly encoded by V_H genes of the PC7183 V_H-gene family (seven out of 12 mAbs), especially among the laminin-reactive mAbs (six out of seven). Although the majority of the antilaminin mAbs in our panel were encoded by very homologous V_H genes, their H-CDR3 regions did not reveal a common amino acid motif (Fig. 2). It is possible that each laminin-reactive mAb recognizes a different epitope on laminin (also within the P1 fragment). The binding to laminin may also rely on the V_H-gene sequence, as described for the H-FR1 region of mAbs reacting with the I antigen on red blood cells (cold agglutinins).³⁸ The contact residues involved in binding to laminin thus remain to be elucidated. Data obtained in our laboratory indicate that the PC7183 V_H gene family in the rat consists of at least 28 V_H genes³² and probably represents the largest V_H-gene family in this species.¹⁸ As observed in different strains of mice, poly- and autoreactive Abs are very frequently encoded by members of the PC7183 V_H-gene family.³⁹ In this context, it is also of interest to note that the mAbs in our panel that are encoded by PC7183 V_H genes (nine out of 15) carry relatively short H-CDR3 lengths (Table 2; mean number of codons ±SD: 9·3 ± 2·0; n = 9). The mean H-CDR3 length of these PC7183-encoded mAbs is very similar to the mean H-CDR3 length of PC7183 V_HDJ_H-µ transcripts expressed by the subset of splenic marginal zone B (MZ-B) cells in the rat.³² In comparison to other B-cell subsets, PC7183 V_HDJ_H rearrangements of MZ-B cells carry significantly shorter H-CDR3 lengths. The H-CDR3 plays an important role in determining the poly- and autoreactive behaviour of Abs.^40–45 Ichiyoshi & Casali,⁴³ and Crouzier et al.⁴¹ showed, by gene-shuffling experiments, that mAbs lose polyreactivity when their H-CDR3 is replaced by the corresponding region of a monoreactive counterpart (both encoded by highly similar V_H- and V_L-region genes), but not when the polyreactive mAbs were grafted with either the H-FR1-FR3 (V_H) region or with the V_κ light-chain region of that monoreactive counterpart. In addition, relatively short H-CDR3s are documented for poly- and autoreactive mAbs in several cases.^40,46 Furthermore, although this was not reported as such, data obtained by Tornberg & Holmberg⁴⁷ reveal that PC7183 V_HDJ_H-µ transcripts expressed by peritoneal B-1a and B-1b cells in mouse carry relatively short H-CDR3 lengths (mean ±SD: 8·6 ± 3·4 and 8·7 ± 2·9, respectively) in comparison to conventional (B-2) cells from spleen (11·3 ± 2·6). A high frequency of the B-1 cells are known to produce poly- and autoreactive Abs.^48,49 PC7183 V_HDJ_H rearrangements, especially those with relatively short H-CDR3s, may therefore encode for Abs with poly and/or autoreactive behaviour.

As shown for several humoral autoimmune disorders in mice and humans,^50–52 V-region genes encoding autoreactive Abs frequently resemble Abs that emerge in the course of an immune response to foreign antigens (i.e. they are derived from clonally related precursors, are somatically mutated and isotype switched). The somatic mutations in these V-region genes often display a selection of mutations (antigen-driven selection) and the mutations are therefore usually not randomly distributed along the genes.³⁷ Our data indicate, however, that the majority of autoreactive Abs produced in HgCl₂-induced MGP in rats probably derive from B cells that have not undergone (auto)antigen-driven selection, as we showed that the majority of mAbs (eight out of 12) are encoded by germline (unmutated) V_H genes. Despite the fact that most of the mAbs in our panel are isotype switched (non-IgM), their V_H genes are not somatically mutated. This is in line with the observation that isotype switching and somatic mutations are independent, and not necessarily linked, processes.⁵³ Interestingly, germline V_H-gene usage was also found among IgG antitubular basement membrane (TBM) autoreactive Abs, in an experimental model of interstitial nephritis.⁵⁴ In addition, antilaminin and anti-DNA IgG mAbs obtained from lupus-prone MRL/lpr mice are also reported to be encoded by germline V-region genes.^55,56 The usage of germline-encoded V_H genes in HgCl₂-induced MGP strongly supports the notion that the autoreactive Abs in this model are produced by polyclonally activated B cells³ and are not the result of a typical adaptive immune response.

In addition to the predominant usage of germline-encoded V_H genes by the MEC and Hg series of hybridomas, our data revealed that only four out of 12 mAbs may be encoded by somatically mutated V_H genes (MEC2, MEC3, MEC6 and MEC8). From these mAbs, the distribution of mutations in the V_H genes of MEC3 and MEC8 may provide some evidence for (auto)antigen-driven selection. The R mutations observed in the H-FRs and H-CDRs of these two anti(P1)laminin mAbs are unlikely to result solely from chance, and these sequences may thus be the result of (auto)antigen-selection. An antigen-driven selection of antilaminin-producing B cells also corresponds with the observation that HgCl₂ treatment increases the ratio of antilaminin-secreting B cells to total-secreting B cells in spleen and lymph nodes, as determined by enzyme-linked immunospot (ELISPOT) assay (J. Aten, unpublished). Mutations in the V_H genes of these four mAbs do not necessarily conflict with their IgM isotype, as somatically mutated V-region genes of IgM class are described for humans^52,57 and rats.³² However, we would like to emphasize here that currently no definite conclusions can be drawn from the mutations present in mAbs MEC2, MEC3, MEC6 and MEC8. First, it is still possible that the V_H genes of these mAbs are derived from unknown germline genes. Even if this were so, the mAbs may still harbour (antigen-selected) somatic mutations. Second, because the hypermutation mechanism is intrinsically biased towards certain bases within DNA motifs (hotspots),⁵⁸ we cannot completely rely on the statistical method of Chang & Casali³⁷ in order to determine whether the mutations in these mAbs are the result of selection by (auto)antigen.

In conclusion, our results imply that the majority of autoreactive Abs involved in HgCl₂-induced MGP are derived from B cells that probably exist in the preimmune repertoire and are inclined to switch to downstream isotypes without undergoing affinity maturation (germline V_H-gene usage). From our data, however, we cannot fully exclude the possibility that a small fraction of the antibody-producing cells in this autoimune disease model carry somatic mutations in their V-region genes and thus probably result from an (auto)antigen selection process. Our observation that the kidney-binding antibodies encoded by PC7183 V_H genes have a relatively short H-CDR3, suggests that autoreactive Abs in HgCl₂-induced MGP may well be produced by polyreactive B cells. The questions of whether polyreactive B cells in the animal are more sensitive to HgCl₂-induced activation and whether these cells belong to a specific B-cell subset, e.g. splenic MZ-B cells,³² remain to be addressed.

Glossary

BN

Brown-Norway

H-CDR

complementarity-determining region of the heavy chain

DZB

Dorus Zadel Black

H-FR

framework region of the heavy chain

GBM

glomerular basement membrane

mAbs

monoclonal antibodies

MGP

membranous glomerulopathy

R/S

ratio of replacement (R) over silent (S) mutation