Autoimmunity as a result of escape from RNA surveillance

Abstract

In previous studies, we detected a frame shift mutation in the gene encoding the autoantigen La of a patient with systemic lupus erythematosus. The mutant La mRNA contains a premature termination codon. mRNAs that prematurely terminate translation should be eliminated by RNA quality control mechanisms. As we find Abs specific for the mutant La form in approximately 30% of sera from anti-La-positive patients, we expected that mutant La mRNAs circumvent RNA control and the expression of mutant La protein could become harmful. Indeed, real-time PCR, immunostaining, and immunoblotting data of mice transgenic for the mutant La form show that mutant La mRNAs are not repressed in these animals and are translated to mutant La protein. In addition to the mutant La protein, we detected a minor portion of native human La in the mutant La-transgenic mice. Therefore, ribosomal frame shifting may allow the mutant La mRNA to escape from RNA control. Interestingly, expression of the mutant La mRNA results in a lupus-like disease in the experimental mice. Consequently, escape of mutant La mRNA from RNA control can have two effects: it 1) results in the expression of an immunogenic (neo)epitope, and 2) predisposes to autoimmunity.

Figures

FIGURE 1

Native (nLa) and mutant human La (muLa) protein (a,b). (a) nLa consists of two domains. The N-domain contains a RNP consensus motif. The C-domain contains a nuclear localization signal (NLS). muLa contains a frame shift mutation (X). The resulting mutant protein reading frame is truncated and lacks the NLS. (b) nLa) contains eight (A) residues at position 1051 to 1058. The deletion of one (A)-residue (muLa, arrow) results in a PTC (nt 1092–1094). Thus, the mutant La mRNA encodes a mutant reading frame (muLa) which differs from the native La reading frame (nLa) downstream of the frame shift mutation (double headed arrow). The mutant aa sequence is underlined. The “*” denotes the location of the PTC (c1, c2) According to recent models, mRNAs containing a PTC are recognized and eliminated by NMD as EJCs will remain associated with the mRNA after the first round of translation. As schematically shown in c1, the native human La mRNA consists of 12 exons. Consequently, 11 splice junctions will be formed during splicing. Four of these junctions locate downstream of the PTC (c2). Consequently, these four EJCs should remain associated with the mutant La mRNA after the first round of translation and, thus, the mutant La mRNA should be recognized and destroyed by NMD.

FIGURE 2

Anti-NeoLa Abs in autoimmune patients. (a) The native N-domain (LaN), the mutant N-domain (LaN-Neo), and the native C-domain (LaC) of La protein and full length La protein was recombinantly expressed and purified using Nickel-affinity chromatography. The purified fragments were dialysed against PBS. During dialysis various amounts of the La fragments undergo proteolysis resulting in some cases in truncated proteins. All La fragments were blotted against serum samples of either healthy donors (e.g. lanes 1 to 3) or autoimmune patients (e.g. lanes 4 to 14). In addition, the recombinant proteins were blotted against anti-La mAbs directed to either the N-domain (lane 15, SW5) or the C-domain (lane 15, 4B6) of La protein. (b) 74 anti-La negative and 109 anti-La positive sera were analyzed by ELISA using recombinant human La protein as substrate. (c) The same sera as in (b) were analyzed for anti-NeoLa peptide antibodies.

FIGURE 3

Features of the mutant La transgene. (a) The cloning of the mutant La transgene (La gene mutant) started from a native human La transgene construct (La gene native). We replaced the genomic region from exons 2 to 10 of the human native La transgene with the respective sequence of a mutant La cDNA (mutant exon 1La cDNA). For this purpose we isolated the KpnI/BsTEII fragment from the mutant exon 1 La cDNA construct and inserted it in the respective sites of the native La transgene. (b1, b2) Thereby, we conserved the introns upstream and downstream of the cDNA inserts. A mutant La mRNA transcribed and processed from this construct contains four splice junctions (1 to 4). Two splice junctions locate upstream and two downstream of the PTC. During the first round of translation the scanning ribosome should not pass the exon junctions 3 and 4. Consequently, two EJCs should persist downstream of the PTC and the mutant La mRNA should be recognized and eliminated by NMD.

FIGURE 4

Quantitative analysis of expression of mutant La mRNA in mutant La transgenic mice. Quantitative PCR was performed for RNAs isolated from liver (a) or spleen (b) tissues of either mutant La transgenic (a, M1, M2; b, muLa), or non-transgenic (b, non-tg), or native La transgenic mice (b, nLa). In addition, we estimated the copy number of La mRNAs for commercial multi-tissue–cDNA panels from mouse (b, mMTC) and human (b, hMTC). We estimated the copy numbers for mouse exon 1a and b La (blue bars), mouse exon 1c La (purple bars), human exon 1 La (brown bars), and human exon 1′,1″ La (yellow bars) mRNAs. The copy numbers were normalized to actin. (a) shows the data for liver tissues of six individual animals. The estimations were performed at least as duplicates. The error bars show the deviation of the values. (b) shows the data for spleen tissues. The mean values and the standard deviations of all analzyed animals are given. Similar results were obtained for other tissues.

FIGURE 5

Expression of mutant La protein in transgenic mice. (a) Extracts from mutant La transgenic mice (muLa-tg; M1,M2) and non-transgenic mice (non-tg) were prepared from spleen tissues and analyzed by SDS-PAGE/immunoblotting with the anti-La mAb SW5 which recognizes the N-domain of human La. The extracts were normalized with anti-actin antibodies. (b) Extracts from different tissues of mutant La transgenic mice (muLa-tg) were prepared from thymus, spleen, liver, brain, kidney and intestine, immunoprecipitated with the anti-La mAb SW5 and analyzed by SDS-PAGE/immunoblotting with the anti-La mAb SW5 which recognizes the N-domain of human La and therefore also reacts with mutant human La (muLa). (c) Tissues from mice transgenic for mutant La (muLa-tg) or native La (nLa-tg) or from non-transgenic mice (non-tg) were stained with the human anti-La specific mAb SW5 (see Materials and Methods). Due to the lack of the NLS, mutant La (muLa-tg, SW5) is in the cytoplasm while native La (nLa-tg, SW5) is in the nucleus. SW5 does not stain tissues from non-transgenic mice irrespective of the background (FVB/N or B6). The negative controls (negative) were also not stained.

FIGURE 6

Expression of full length human La in mutant La transgenic mice. Total extracts from liver tissue of mice either transgenic for mutant La (lanes a to d), or native La (lanes e,g,i), or non-transgenic (lanes f,h,j) were immunoprecipitated with an anti-La serum and blotted against the anti-La mAbs 3B9 (lanes a,e,f), SW5 (lanes b,g,h), and 4B6 (lanes c,i,j). The anti-La mAb 3B9 recognizes human and mouse La protein. In contrast, the anti-La mAbs SW5 and 4B6 do not cross-react with mouse La protein. The epitopes recognized by SW5 and 3B9 locate in the N-domain of La while the epitope recognized by 4B6 locates in the C-domain. A mixture of unrelated mAbs did not react with any immunoprecipitated protein (lane d). hLa, human La; moLa, mouse La; muLa, mutant La.

FIGURE 7

Immune complex deposition in kidneys of mutant La-transgenic mice. We stained kidney sections from mice transgenic for mutant La (a to g) and FVB/N control animals (h to j) with anti-Ig Ab (red), anti-complement Ab (green), and DAPI (blue). (a) Three colour overlay of a nephron of a mouse transgenic for mutant La. (b to g) Glomeruli from mice transgenic for mutant La. (h to j) FVB/N controls.

FIGURE 8

Autoantibodies in mice transgenic for mutant La. (a,b) Development of autoantibodies in mice transgenic for mutant La. We analyzed sera from the mutant La-transgenic mouse lines M1 and M2 for Abs to neoLa, La, Ro and dsDNA. The dashed lines represent cutoffs for positive reaction based on mean plus four standard deviations of 15 control FVB/N anti-sera. Nine of the control mice were five to six months of age. Six mice were older than seven months. (c to f) Ab pattern of individual mice. The autoantibody pattern differs between individual mice. Some mice develop Abs to NeoLa, La, Ro and dsDNA (e.g. c), others Abs to NeoLa and La but not to Ro and dsDNA (e.g. d) or to NeoLa, La and Ro but not to dsDNA (e.g. e). In some mice the response started with anti-Ro Abs (e.g. e) in others with anti-Sm Abs (f). The serum samples (1 to 4) in f were taken at the age of 4 to 8 months. (g) One of the analyzed mice was positive for anti-nuclear Abs in immunofluorescence but negative in ELISA for Sm, RNP, Ro, La, dsDNA. (h) FVB/N negative control. (i) secondary Ab control.