and Immunity (2006) 7, 592–599 & 2006 Nature Publishing Group All rights reserved 1466-4879/06 $30.00 www.nature.com/gene

ORIGINAL ARTICLE Identification of intervals from 129 and C57BL/6 mouse strains linked to the development of systemic lupus erythematosus

Y Heidari1,3, AE Bygrave1,3, RJ Rigby1, KL Rose1, MJ Walport1,4, HT Cook2, TJ Vyse1 and M Botto1 1Molecular Genetics and Rheumatology Section, Faculty of Medicine, Imperial College, Hammersmith Campus, London, UK and 2Department of Histopathology, Faculty of Medicine, Imperial College, Hammersmith Campus, London, UK

Systemic lupus erythematosus is an autoimmune disease in which complex interactions between genes and environmental factors determine the disease phenotype. We have shown that genes from the non-autoimmune strains 129 and C57BL/6 (B6), commonly used for generating -targeted animals, can induce a lupus-like disease. Here, we conducted a genome-wide scan analysis of a cohort of (129 Â B6)F2 C1q-deficient mice to identify loci outside the C1qa locus contributing to the autoimmune phenotype described in these mice. The results were then confirmed in a larger dataset obtained by combining the data from the C1q-deficient mice with data from previously reported wild-type mice. Both analyses showed that a 129-derived interval on distal is strongly linked to autoantibody production. The B6 genome contributed to anti-nuclear autoantibody production with an interval on chromosome 3. Two regions were linked to glomerulonephritis: a 129 interval on proximal chromosome 7 and a B6 interval on chromosome 13. These findings demonstrate that interacting loci between 129 and B6 mice can cause the expression of an autoimmune phenotype in gene-targeted animals in the absence of any disrupted gene. They also indicate that some susceptibility genes can be inherited from the genome of non-autoimmune parental strains. Genes and Immunity (2006) 7, 592–599. doi:10.1038/sj.gene.6364335; published online 31 August 2006

Keywords: systemic lupus erythematosus; autoantibodies; rodent; gene-targeting

Introduction clear that several intervals are strain-specific, confirming the genetic complexity of the disease and indicating the Systemic lupus erythematosus (SLE) is a systemic presence of extensive heterogeneity in the genes con- autoimmune disease characterized by the production of tributing to SLE pathogenesis. autoantibodies against a variety of nuclear and cell More recently, gene-targeting technology has allowed surface antigens, resulting in immune-complex mediated researchers to investigate the impact of a single gene on damage to vascular, dermatological, renal, neurological murine physiology. There is, however, accumulating and rheumatological tissues. The aetiology of SLE is evidence that genetic factors other than the actual complex, with a combination of multiple genes and disrupted gene can influence the resulting phenotype environmental factors determining both susceptibility of the knockout mouse. In this regard, it is of note that and disease phenotype. the majority of the gene-targeted strains are initially Spontaneous murine models of SLE, such as the New developed on a hybrid genetic background between 129 Zealand, BXSB and MRL mouse strains, have been widely and C57BL/6 (B6) mice, which has been shown to be used to dissect the complex genetic component of SLE. spontaneously predisposed to development of humoral Comparison of the many linkage studies performed in autoimmunity with low levels of glomerulonephritis.2–5 these lupus-prone strains illustrates the complexity of this The initial knockout strain is then usually backcrossed disease. Numerous disease susceptibility loci have been onto B6 in order to remove as much 129 genome as identified, with some loci, such as distal chromosome 1, possible. However, despite 10 or more generations of mid to distal chromosome 4, proximal chromosome 7 and backcrossing, a considerable 129-derived genome inter- proximal chromosome 17 (the H2 complex) being linked val, flanking the targeted gene, will remain. We have to disease in different murine models.1 However, it is also previously shown that a number of genetic loci, derived from both 129 and B6, are linked to the development of disease in the (129 Â B6) mice, with the most statistically Correspondence: Professor M Botto, Molecular Genetics and significant of these being a 129-derived interval on distal Rheumatology Section, Faculty of Medicine, Imperial College, chromosome 1 linked to increased autoantibody produc- Hammersmith Campus, Du Cane Road, London W12 0NN, UK. tion.6 This region has been consistently linked to E-mail: [email protected] 3 autoimmune traits in a number of lupus-prone strains, These authors equally contributed to the work. 7,8 9 4Current address: The Wellcome Trust, London, UK. including NZB (Nba2), NZM2410 (Sle1) and BXSB 10,11 Received 5 July 2006; accepted 31 July 2006; published online 31 (Bxs3). A congenic strain, comprising the 129-derived August 2006 locus on distal chromosome 1 on a B6 background, a Autoimmunity in 129 and C57BL/6 mice Y Heidari et al 593 combination commonly created by backcrossing onto B6 range 0–10240), anti-chromatin antibodies (Abs) in 58% a knockout strain in which the gene located in that region (median: 2.23, range 0–4.0), anti-double-stranded DNA has been inactivated in 129 embryonic stem cells, also (anti-dsDNA) Abs in 18% (median 0; range 0–2560) and developed an autoimmune phenotype. The humoral anti-single-stranded DNA (anti-ssDNA) Abs in 85% of autoimmunity in this congenic strain was indistinguish- the mice. Histological evidence of glomerulonephritis able to that observed in a mouse carrying a deletion of (above grade I) was found in 32% of the mice. the Apcs gene, located within the lupus-linked genomic Interval mapping demonstrated significant linkage to region on distal chromosome 1 and considered as a ANA (LOD ¼ 4.0, P ¼ 9.9 Â 10À5, Figure 1a), anti-dsDNA candidate gene for murine SLE.6 Therefore, background- Abs (LOD ¼ 5.3, P ¼ 4.9 Â 10À6, Figure 1b) and anti- derived genes can significantly contribute to the pheno- ssDNA Abs (LOD ¼ 5.5, P ¼ 3.1 Â10À6, Figure 1c) at type observed in knockout strains even when the mice a locus approximately 95 cM from the centromere of have been extensively backcrossed onto the B6 strain, chromosome 1 in the (129 Â B6)F2.C1qaÀ/À cohort. Anti- greatly complicating the interpretation of the phenotypic chromatin Abs were also linked to this chromosome 1 analysis of gene-targeted animals. region, but at a more distal locus and with a lower level The influence of background genes on the develop- of significance (LOD ¼ 2.2, P ¼ 6.3 Â 10À4, Figure 1d). All ment or modification of spontaneous autoimmune these loci were derived from 129. disease is well known, especially with respect to the lpr In addition to the linkage observed on distal chromo- and Yaa disease-susceptibility genes.12–14 Not surpris- some 1, ANA titres were linked to a locus on mid-distal ingly important effects of the genetic background on the chromosome 3, albeit with a reduced degree of sig- expression of autoimmunity have also been reported in nificance (LOD ¼ 2.3, P ¼ 5 Â 10À4, Figure 2). This locus gene-targeted mice.5,15,16 For example, C1q deficiency, a was derived from the B6 background. In this context, condition that in humans is strongly associated with the it is of note that QTL analysis of the mid-distal region development of a lupus-like disease,17 in mice appears to of chromosome 4 could not be applied to the have a disease-accelerating effect only in lupus-prone (129 Â B6)F2.C1qaÀ/À mice as this region was of fixed strains, including the (129 Â B6) genetic background.3,16 129 origin. Likewise, inactivation of the FcgRIIB gene results in the Owing to the reduced variability in the glomerulone- development of a lupus-like phenotype in the context of phritis (GN) data compared to the autoantibody titre the B6 genomic background, but not the BALB/c data, the linkage to GN was determined using two genomic background.15 In this example, it is of note that methods – as a regular quantitative trait, and in an two recessive B6-derived loci outside the targeted region analysis of extremes, where only mice that were clearly were linked to the development of the disease pheno- negative or positive were included. Both these analyses type.18 Thus, SLE exists as a complex-trait disorder in showed that GN was linked to a 129-derived region of which specific combinations of susceptibility alleles are proximal chromosome 7 in the (129 Â B6)F2.C1qaÀ/À required for the expression of the full phenotype. cohort. The QTL analysis demonstrated suggestive In order to identify loci outside the C1qa locus that linkage (LOD ¼ 2.2, P ¼ 6.3 Â 10À4) to a 22 cM region modify the autoimmunity observed in (129 Â B6).C1qaÀ/À between D7Mit246 (15 cM) and D7Mit30 (37 cM), and mice we carried out a linkage analysis to both autoanti- the analysis of extremes showed linkage to D7Mit230 body production and nephritis in a new cohort of at 26.3 cM, with a w2 of 8.84. (129 Â B6)F2.C1qaÀ/À mice. In addition, we combined the data from this cohort with data from a previously Confirmation of linkage analysis data in a combined cohort of published cohort of (129 Â B6)F2 wild-type mice,6 in (129 Â B6)F2.C1qaÀ/À and (129 Â B6)F2 mice order to expand the sample size and confirm our Overall the study of the (129 Â B6)F2.C1qaÀ/À cohort observations. We show here that a number of genetic confirmed a number of the 129 and B6 loci previously loci outside the C1q locus on distal chromosome 4 are described as being associated with SLE traits, but failed linked to autoimmunity, both confirming previously to support others.6 In order to increase the power of our described loci and identifying novel regions in this strain study, we repeated the QTL analysis in a larger data set, combination. obtained by combining data from the (129 Â B6)F2.C1qaÀ/À mice presented in this study with data from a pre- viously reported (129 Â B6)F2 intercross,6 resulting in 297 Results female mice. In this analysis, the medial and distal regions of chromosome 4 were not included. The results Mapping of loci predisposing to lupus in the from the IgG anti-ssDNA ELISA and IgG anti-chromatin (129 Â B6).C1qaÀ/À mice ELISA assays were ranked, as described in the Materials The observation that 129.C1qaÀ/À and B6.C1qaÀ/À mice and methods section, in order to compare the data from did not develop any autoimmune traits16 while the the two cohorts. The ranking method was verified by (129 Â B6).C1qaÀ/À mice developed a lupus-like disease3 comparing the patterns of genome-wide linkage to both suggest that the disease-modifying loci may arise as a ranked and unranked ELISA data in a single cohort, and result of interaction between specific combinations of confirming that the linkage patterns were comparable in alleles inherited from both the 129 and B6 parental both position and magnitude of linkage. strains. In order to investigate the genetic contribution of The combined cohort of (129 Â B6)F2.C1qaÀ/À and 129 and B6 genes to the lupus-like disease observed in (129 Â B6)F2 mice confirmed the linkage of ANA the (129 Â B6).C1qaÀ/À mice, we generated a new cohort (LOD ¼ 6.3, P ¼ 5 Â 10À7, Figure 3a), IgG anti-dsDNA of (129 Â B6)F2.C1qaÀ/À animals and monitored these for Abs (LOD ¼ 7.6, P ¼ 2.5 Â 10À8, Figure 3b), IgG anti- a year. In this cohort of C1q-deficient mice, anti-nuclear ssDNA Abs (LOD ¼ 8.3, P ¼ 4.9 Â 10À9, Figure 3c) and antibodies (ANA) were detected in 38% (median 0, IgG anti-chromatin Abs (LOD ¼ 4.7, P ¼ 2 Â 10À5,

Genes and Immunity Autoimmunity in 129 and C57BL/6 mice Y Heidari et al 594

Figure 1 Interval maps showing QTLs on chromosome 1 with ANA (a), anti-dsDNA Abs (b), anti-ssDNA Abs (c) and anti-chromatin Abs (d) in (129 Â B6)F2.C1qaÀ/À mice. Centimorgan positions were deduced by interval mapping, anchoring marker locations according to data from www.informatics.jax.org. Intermittent dashed lines indicate threshold over which linkage is considered suggestive; dashed lines indicate threshold over which linkage is considered significant; dotted lines indicate threshold over which linkage is considered highly significant (see Materials and methods). LOD scores were generated with Map Manager QTb29.

individual cohorts. It is of note that the linkage to IgG anti-chromatin Abs mapped to a more distal region of chromosome 1 than the other autoantibodies in the (129 Â B6)F2.C1qaÀ/À cohort, whereas in the combined cohort analysis it mapped to the same locus as the other autoantibodies (B95 cM from the centromere). Akin to the linkage on chromosome 1, the ANA linkage to mid- distal chromosome 3 was observed in the combined cohort of (129 Â B6)F2.C1qaÀ/À and (129 Â B6)F2 mice. ANA titre was linked to a locus of around 50 cM from the centromere, again with an increased degree of signifi- cance in the combined cohort (LOD ¼ 8.0, P ¼ 9 Â 10À9, Figure 4). To investigate any potential interactions between the B6-derived gene(s) on chromosome 3 and the 129- derived gene(s) on chromosome 1, we grouped the mice Figure 2 Linkage of ANA to a B6-derived region of Chromosome 3 accordingly to their genotype at these two loci and in (129 Â B6)F2.C1qaÀ/À mice. Threshold for suggestive linkage compared ANA titres. We selected a marker at the peak (intermittent dashed line), as determined by 1000 cross- and trait- of the linkage to ANA ((D1Mit206 (95.8 cM) on chromo- À4 specific permutation tests, is indicated (LOD ¼ 2.3, P ¼ 5 Â 10 ). some 1 and D3Mit103 (51.1 cM) on chromosome 3) to define the genotype of the mice. As illustrated in Figure 5, Figure 3d) to a locus on distal chromosome 1 with a peak ANA levels were significantly higher in the mice of around 95 cM. All of the above disease traits were carrying two B6 alleles on chromosome 3 in combination linked to the distal chromosome 1 region with a higher with one or two 129 alleles on chromosome 1 compared degree of significance in the combined than the to all the other genetic combinations. Of note the ANA

Genes and Immunity Autoimmunity in 129 and C57BL/6 mice Y Heidari et al 595

Figure 3 Interval maps showing QTLs on chromosome 1 with ANA (a), anti-dsDNA Abs (b), anti-ssDNA Abs (c) and anti-chromatin Abs (d) in a combined analyses of (129 Â B6)F2.C1qaÀ/À and wild-type (129 Â B6)F2 mice. Centimorgan positions were deduced by interval mapping, anchoring marker locations according to data from www.informatics.jax.org. Intermittent dashed lines indicate threshold over which linkage is considered suggestive; dashed lines indicate threshold over which linkage is considered significant; dotted lines indicate threshold over which linkage is considered highly significant. Thresholds were determined by 1000 cross- and trait-specific permutation tests.

titres were not significantly different between the animals homozygous or heterozygous for the 129- derived segment on chromosome 1, when in combina- tion with B6 homozygosity on chromosome 3 suggesting that the presence of a single 129-derived allele on chromosome 1 was sufficient to drive loss of tolerance to nuclear antigens. This analysis provided further support to the hypothesis that the B6-derived loci on chromosome 3 tend to operate in a recessive manner while the 129-derived loci on chromosome 1 operate in a dominant fashion. As in the (129 Â B6)F2.C1qaÀ/À cohort, the linkage analysis of GN in the combined cohort was carried out using both a quantitative trait analysis and analysis of extremes. Using the quantitative trait analysis method, linkage of GN to a 129-derived locus on proximal Figure 4 Linkage of ANA to a B6-derived region of Chromosome 3 chromosome 7 was confirmed with an increased sig- in a combined analyses of (129 Â B6)F2.C1qaÀ/À and wild-type nificance (LOD ¼ 3.8, P ¼ 1.6 Â 10À5, Figure 6a). This was (129 Â B6)F2 mice. Centimorgan positions were deduced by interval confirmed by the analysis of extremes which showed a mapping, anchoring marker locations according to data from linkage to D7Mit230 (26.3 cM), with a w2 of 11.72, www.informatics.jax.org. Intermittent dashed lines indicate thresh- P ¼ 0.0028 (Table 1). Interestingly, a further linkage to old over which linkage is considered suggestive; dashed lines indicate threshold over which linkage is considered significant; GN, this time derived from a B6 locus, was observed in dotted lines indicate threshold over which linkage is considered the combined cohort. Located on proximal chromosome highly significant (LOD ¼ 8.0, P ¼ 9 Â 10À9). 13, this locus reached the cutoff for suggestive linkage

Genes and Immunity Autoimmunity in 129 and C57BL/6 mice Y Heidari et al 596

Figure 5 ANA titres of various chromosome 1 and chromosome 3 genotype combinations in a pooled cohort of (129 Â B6)F2.C1qaÀ/À and wild-type (129 Â B6)F2 mice. Small symbols represent one mouse; large symbols a variable number of animals as indicated in parentheses. The genotype was established using a single marker at the peak linkages on chromosome 1 (D1Mit206) and chromosome 3 (D3Mit103). The mice carrying two B6-derived alleles on chromo- some 3 and one or two 129-derived alleles on chromosome 1 had significantly higher levels of ANA compared to all the other genetic combinations (Po0.001). The most relevant comparisons are shown. One way ANOVA with Bonferonni’s multiple comparison tests were applied.

(LOD ¼ 2.4, P ¼ 4 Â 10À4, Figure 6b) in the combined cohort, but not in the individual (129 Â B6)F2.C1qaÀ/À or (129 Â B6)F2 cohorts. Like the locus on chromosome 7, this linkage was also observed in an analysis of extremes, Figure 6 Interval maps showing QTLs on chromosome 7 (a) and with GN linked to D13Mit117 at B19 cM, with a w2 of chromosome 13 (b) with GN in a combined analysis of 15.3, P ¼ 4.77 Â 10À4 (Table 1). (129 Â B6)F2.C1qaÀ/À and wild-type (129 Â B6)F2 mice. Centimorgan positions were deduced by interval mapping, anchoring marker locations according to data from www.informatics.jax.org. Inter- mittent dashed lines indicate threshold over which linkage is Discussion considered suggestive; dashed lines indicate threshold over which linkage is considered significant. (LOD ¼ 3.8, P ¼ 1.6 Â 10À5 and Multiple genetic loci are known to contribute to the LOD ¼ 2.4, P ¼ 4 Â 10À4, respectively). development and pathogenesis of SLE in mice and in humans. In this study, we have provided further evidence that epistatic interactions between 129 and B6 Table 1 Linkage of 7 and 13 to GN in the combined mice, even though autoimmunity has not been reported cohort of (129 Â B6)F2.C1qaÀ/À and wild-type (129 Â B6)F2 mice in either of the strains, can lead to a spontaneous lupus- like phenotype. In particular, we confirmed that a 129- Marker Position (cM) Origin w2 P derived region of chromosome 1, when expressed in context of the B6 genome, is strongly linked to autoanti- D7Mit178 0.5 129 4.14 1.26E-01 body production. Consistent with this, the B6 genome D7Mit246 15.0 129 11.04 4.01E-03 contributed to the autoimmune phenotype with an D7Mit158 23.0 129 10.05 6.58E-03 D7Mit230 26.3 129 11.72 2.86E-03 interval on chromosome 3, displaying a highly signifi- D7Mit30 37.0 129 7.57 2.27E-02 cant linkage to anti-nuclear autoantibodies. Interestingly D7Mit253 52.8 129 6.13 4.66E-02 glomerulonephritis was linked to different chromo- somes: a 129 region on proximal chromosome 7 and a D13Mit135 10.0 B6 12.53 1.90E-03 B6 interval on chromosome 13 (Figure 7). D13Mit117 19.0 B6 15.29 4.77E-04 A genome-wide scan of the (129 Â B6)F2.C1qaÀ/À mice D13Mit64 30.0 B6 5.81 5.47E-02 D13Mit248 34.0 B6 3.54 1.70E-01 demonstrated linkage of lupus serological markers D13Mit193 43.0 B6 2.51 2.85E-01 (ANA, anti-dsDNA Abs, anti-ssDNA Abs and anti- D13Mit30 52.0 B6 0.63 7.31E-01 chromatin Abs) to a 129-derived locus on distal chromo- D13Mit262 68.0 B6 1.61 4.48E-01 some 1 – recently named Sle 16 (http://www.informatics. 6 jax.org). This is in agreement with our previous study, in Centimorgan positions were deduced by interval mapping, anchor- which SLE traits were mapped in (129 Â C57BL/6)F2 ing marker locations according to data from www.informatics.jax.- wild-type and (129 Â C57BL/6)F2.ApcsÀ/À mice. The org. w2-values are calculated with a standard (2 Â 2) contingency linkages were markedly increased when we enlarged table with 2 degree of freedom. Suggestive linkages, defined as the sample size by combining data from the described in the material and method section, are underlined.

Genes and Immunity Autoimmunity in 129 and C57BL/6 mice Y Heidari et al 597 1 3

7 13 1,2 1,2 Sle3 Sles3* 3 1,2 Lbw5 GN Bxs5 2 ANA Nba3 GN 1 Sle1 ANA 1,2Yaa1 1,2,4Nba2 anti-ssDNA 1,2Bxs3 anti-dsDNA 1Lbw7 anti-chromatin

Figure 7 A summary of the lupus susceptibility loci mapped in this study on chromosomes 1, 3, 7 and 13. Regions previously linked to SLE in the other lupus models are also indicated. Linkage to SLE traits from this study are shown as open boxes ¼ 129, filled boxes ¼ B6. 1ANA (any), 2GN, 3mortality, 4gp70/gp70 immune complexes, *disease suppressor locus.

(129  B6)F2.C1qaÀ/À mice with data from a previously the 129 gene(s) that predispose to GN in the (129  B6)F2 reported (129  B6)F2 intercross, indicating that this 129 strain could modify or augment the GN observed in a region is the main locus capable of initiating the humoral knockout strain. Recently a knockout of the Bcl-2 autoimmune response to nuclear antigens upon interac- associated  (Bax) gene, located on proximal tion with the B6 genome. The distal chromosome 1 chromosome 7, was reported to have GN as one of its region is associated with a number of potent SLE resulting phenotypes.27 In light of the data presented susceptibility loci including Sle1,9 Nba219 and Bxs3,10 all herein, one cannot exclude the possibility that the of which are strongly associated with the production of observed GN phenotype may have been due to the autoantibodies to nuclear antigens. The significance of surrounding 129 genome, and not the inactive Bax gene. this chromosome 1 locus has also been confirmed by A previously described locus of B6 origin on chromo- several congenic dissection analyses.1,20 Furthermore, some 3,6 was here confirmed to be linked to ANA recent genomic characterization of the Sle1b locus production in both the (129  B6)F2.C1qaÀ/À mice and (located between 171.8 and 173.1 Mbp) has identified a combined cohorts. Unlike the loci on chromosomes 1 and highly polymorphic cluster of Slam/Cd2 family genes, 7, this locus has not been consistently linked to lupus encoding key regulators of lymphocyte function, as the disease traits, the SLE-associated loci Sles3 and Bxs5 strongest candidate genes for mediating the Sle1b being proximal (between 20 and 40 cM) to our observed autoimmune phenotype.21 The autoimmune-associated linkage peak. However, they intriguingly both involve haplotype of the lupus-prone NZM2410 (NZW) strain, non-autoimmune strains; in the case of Sles3, named the Slam/Cd2 haplotype 2, is also present in the NZWxC57BL/6 heterozygosity,1 and in the case of 129/SvJ mice,21 indicating that these two strains may Bxs5, C57BL/10 homozygosity.11 The confirmation of a share the same pathways leading to loss of peripheral B6-derived locus on chromosome 3 that modifies ANA tolerance when in combination with one or more titres (in combination with 129-derived genes) highlights polymorphic genes in the B6 genome. These observations the likelihood of epistatic interactions influencing the give compelling evidence for the presence of a 129 locus phenotype of a knockout strain, in spite of extensive influencing systemic autoimmunity on telomeric chro- backcrossing to B6. mosome 1 as revealed in the context of B6 genome. The possibility of B6 genes influencing the phenotype Whether this 129 lupus locus can contribute to the of a knockout mouse was further underlined by the development of autoimmunity on other genetic back- observation of B6-derived linkage to GN on proximal grounds is still unknown. chromosome 13. This linkage was only observed in In common with the distal chromosome 1 region, the combined analysis of (129  B6)F2.C1qaÀ/À and numerous studies have demonstrated linkage of SLE (129  B6)F2 mice, and had thus a less penetrant effect traits to proximal chromosome 7. In this study, we than the loci on chromosomes 1, 3 and 7. Nonetheless, confirm that there is a 129-derived locus on proximal such minor loci may together have a significant chromosome 7 linked to GN.6 Furthermore, this locus cumulative influence on the phenotype of a knockout colocalizes with a number of loci from various lupus- strain bred onto the B6 background. Hence, it is prone mouse strains, including Sle3 (NZM2401 locus important to identify both the genomic position and linked to GN and ANA);20,22 Lbw5 (NZW locus linked to the resulting phenotypes of these loci. In this context, it is mortality),23 Lmb3 (MRL locus linked to splenomegaly, of note that we had previously reported a B6-derived lymphadenopathy, anti-dsDNA Abs),24 Nba3 (NZB locus suggestive linkage to anti-dsDNA Abs on the mid-distal linked to GN)25 and Nba5.26 As this is an analogous region of chromosome 4.6 However, this observation situation to that on distal chromosome 1, a similar could not be validated in the current study as this region haplotype-based candidate gene identification strategy was of fixed 129-origin in the (129  B6)F2.C1qaÀ/À mice. could be applied, especially as murine single nucleotide Hence, we cannot exclude the potential effect of a minor polymorphism (SNP) databases continue to increase in B6-derived locus on distal chromosome 4. both SNP density and the number of typed strains. The In conclusion, we have identified linkage of lupus presence of a 129-derived lupus locus on proximal disease traits to loci present on chromosomes 1, 3, 13 and chromosome 7 also has impact on the phenotype of 7 in the 129  B6 mice (Figure 7). As this is the strain any targeted gene in this region – it is conceivable that combination of choice for gene-targeted mice, knowledge

Genes and Immunity Autoimmunity in 129 and C57BL/6 mice Y Heidari et al 598 of the influence of parental genetic loci on the disease Schiff reagent and scored for GN. Glomerular histology phenotype is critical if we are to avoid attributing false was graded as follows: grade 0 – normal, grade I – focal positive phenotypes to knockout strains. hypercellularity in 10–25% of the glomeruli, II – hypercellularity involving 450% of the glomerular tuft in 25–50% of glomeruli, grade III – hypercellularity Materials and methods involving 450% of the glomerular tuft in 50–75% of glomeruli, grade IV – glomerular hypercellularity in Mice 475% or crescents in 425% of glomeruli. Histological The C1q-deficient mice, C1qaÀ/À were generated as 3 À/À analysis was performed in a blinded fashion and 50 previously reported and the (129 Â B6)F1.C1qa mice glomeruli per section were analysed. were generated by crossing 129.C1qaÀ/À mice with B6.C1qaÀ/À mice that had been backcrossed onto B6 for À/À Genotypic analysis 10 generations. (129 Â B6)F2.C1qa were obtained by Genotyping of the (129 Â B6)F2.C1qaÀ/À cohort was intercrossing the (129 Â B6)F1.C1qaÀ/À mice. A total of 156 À/À carried out using polymorphic microsatellite markers, a (129 Â B6)F2.C1qa female mice were produced and standard polymerase chain reaction and either 4% monitored for 1 year. The strain- and sex-matched wild- 6 MetaPhor agarose (Cambrex Bioscience Rockland, Rock- type (129 Â B6)F2 cohort was as previously described. land, ME, USA) or 16% polyacrylamide gels stained with Mice were maintained in specific pathogen-free condi- ethidium bromide. The average marker density was one tions, and all procedures were in accordance with marker per 10 centimorgans (cM) across the autosomes, institutional guidelines. the positions and sequences of which were determined from the Mouse Genome Informatics (MGI) database Serological analyses (http://www.informatics.jax.org). The list of markers Serum was collected at 12 months of age and assayed for used is available on request. autoantibodies. IgG ANA and IgG anti-dsDNA Abs were measured by indirect immunofluorescence using Hep-2 Statistical analyses cell and Crithidia luciliae slides (The Binding Site, All linkage analyses and interval mapping were con- Birmingham, UK) respectively.16 Serum samples were ducted using MapManager QTb29 (ftp://mcbio.med.- screened at a 1:80 (ANA) or 1:20 (anti-dsDNA Abs) buffalo.edu/pub/MapMgr/).29 Marker maps were dilution and the positive samples titrated to end point. generated to determine the accuracy of genotyping, with IgG anti-ssDNA Abs and anti-chromatin Abs were all markers mapping to within 3 cM of the marker measured by ELISA as described previously.28 For position in the MGI database. All centimorgans positions determination of anti-chromatin antibodies ELISA plates of markers referred to in this study are from the MGI (Nunc-Immuno MaxiSorp, NUNC, Denmark) were database. coated with 50 ml of PBS Thimerosal (0.1 g/l) containing Anti-ssDNA and anti-chromatin ELISA data were log 0.5 mg/ml nucleohistones from calf thymus (Lorne transformed prior to linkage analysis as this resulted in a laboratories Ltd, Reading, UK) at 41C overnight, then more normalized distribution. As the ELISA assays for blocked with PBS 5% milk powder for 1 h at 371C. Sera these two autoantibodies in the two cohorts of mice had diluted 1/100 in PBS 2% BSA 0.05% Tween20 were been carried on separate occasions and using a different incubated for 1 h at 371C. For measuring anti-ssDNA Abs MRL/Mp.lpr/lpr standard positive sample, for the plates were coated with 50 mlof10mg/ml ssDNA (Sigma combined analysis of the two sets of data the samples Chemical Co., Poole, UK) in sodium carbonate buffer pH in each group were ranked using uncategorized (con- 9.6 at 41C overnight and then blocked with 100 ml PBS tinuous) arbitrary values and the ranked values were 0.5% BSA. Samples were screened at a 1/100 dilution in subsequently used for the quantitative trait locus (QTL) PBS 2% BSA 0.05% Tween20 for 1 h at 371C with 50 ml/ analysis. well. Bound antibodies were detected with alkaline Thresholds for suggestive, significant and highly phosphatase (AP)-conjugated goat anti-mouse IgG significant linkages were determined using cohort- and (g-chain specific) (Sigma-Aldrich, Dorset, UK). The plates trait-specific permutation tests, based on 1000 permuta- were developed using the substrate p-nitrophenyl phos- tions of the data, in Map Manager QTb29.29 A logarithm phate (Sigma Chemical Co., Poole, UK). The OD of the of odds ratio (LOD) of X2.0 (Pp9.9 Â 10À4), of X3.6 reaction mixture at 405 nm wavelength was measured (Pp2.5 Â 10À5) and of X5.7 (Pp2.0 Â 10À6) was indica- using an ELISA plate reader (Titertek Labsystems, tive of suggestive, significant and highly significant Basingstoke, UK). Samples were tested in duplicate with linkage, respectively, in the (129 Â B6)F2.C1qaÀ/À cohort a non-specific binding control and the results were of mice. In the combined cohort of C1q-deficient expressed in arbitrary ELISA units (AEU) relative to and wild-type F2 mice, the threshold for suggestive, serial dilutions of a standard positive sample derived significant and highly significant linkages were from pooled autoimmune MRL/Mp.lpr/lpr serum. Serum LOD X2.1 (Pp7.9 Â 10À4), X3.5 (Pp3.1 Â10À5) and samples were considered positive if above the mean X5.3 (Pp4.9 Â 10À6), respectively. The calculated thresh- 7 2s.d. of the blank. The intra-assay coefficient of olds for suggestive, significant and highly significant variation was between 5 and 8%. linkages were similar across the different traits. Glomerulonephritis score was analysed as both a Histological analysis quantitative trait in Map Manager and in an analysis of All the mice were killed at 1 year of age. Kidney tissue extremes, in which mice with a GN score of II or above was fixed in Bouin’s solution for at least 2 h, transferred were considered as positive for glomerulonephritis, and into 70% ethanol, and then processed into paraffin. mice with a GN score of 0 were considered negative. Sections were cut, mounted, stained with periodic acid- All mice with grade I GN were excluded. Linkage to

Genes and Immunity Autoimmunity in 129 and C57BL/6 mice Y Heidari et al 599 microsatellite markers was determined using a Chi 10 Hogarth MB, Slingsby JH, Allen PJ, Thompson EM, Chandler P, square (w2) test in an Excel (2003) spreadsheet. w2-values Davies KA et al. Multiple lupus susceptibility loci map to of over 12.9 (P ¼ 0.0016) was considered to be indicative chromosome 1 in BXSB mice. J Immunol 1998; 161: 2753–2761. of suggestive linkage.30 11 Haywood ME, Hogarth MB, Slingsby JH, Rose SJ, Allen PJ, Non-parametric data are presented as median, with Thompson EM et al. Identification of intervals on chromo- range of values in parentheses unless otherwise stated. somes 1, 3, and 13 linked to the development of lupus in BXSB Statistics were calculated using GraphPad Prism version mice. Arthritis Rheum 2000; 43: 349–355. 12 Izui S, Kelley VE, Masuda K, Yoshida H, Roths JB, Murphy 3.0 (GraphPad Software, San Diego, CA, USA). One way ED. Induction of various autoantibodies by mutant gene lpr in ANOVA with Bonferonni;s multiple comparison tests several strains of mice. J Immunol 1984; 133: 227–233. were applied for analysis of multiple groups. 13 Izui S, Higaki M, Morrow D, Merino R. The Y chromosome from autoimmune BXSB/MpJ mice induces a lupus-like syndrome in (NZW Â C57BL/6)F1 male mice, but not in Abbreviations C57BL/6 male mice. Eur J Immunol 1988; 18: 911–915. 14 Merino R, Shibata T, De Kossodo S, Izui S. Differential effect of Abs, antibodies; AEU, arbitrary ELISA units; ANA, the autoimmune Yaa and lpr genes on the acceleration of antinuclear antibody; AP, alkaline phosphatase; anti- lupus-like syndrome in MRL/MpJ mice. Eur J Immunol 1989; dsDNA, anti-double stranded DNA; anti-ssDNA, 19: 2131–2137. anti-single stranded DNA; B6, C57BL/6; GN, glomer- 15 Bolland S, Ravetch JV. Spontaneous autoimmune disease in ulonephritis; QTL, quantitative trait locus; SLE, Systemic Fc(gamma)RIIB-deficient mice results from strain-specific lupus erythematosus. epistasis. Immunity 2000; 13: 277–285. 16 Mitchell DA, Pickering MC, Warren J, Fossati-Jimack L, Cortes-Hernandez J, Cook HT et al. C1q deficiency and Acknowledgements autoimmunity: the effects of genetic background on disease expression. J Immunol 2002; 168: 2538–2543. We thank Mrs Margarita Lewis for technical assistance 17 Pickering MC, Botto M, Taylor PR, Lachmann PJ, Walport MJ. with the processing of tissue for histological studies and Systemic lupus erythematosus, complement deficiency, and the staff of the Biological Services Unit at our institution apoptosis. Adv Immunol 2000; 76: 227–324. 18 Bolland S, Yim Y-S, Tus K, Wakeland EK, Ravetch JV. Genetic for the care of the animals involved in this study. This modifiers of systemic lupus erythematosus in FcgRIIBÀ/À work was supported by the Wellcome Trust (Grant mice. J Exp Med 2002; 195: 1167–1174. number 071467). 19 Rozzo SJ, Vyse TJ, Drake CG, Kotzin BL. Effect of genetic background on the contribution of New Zealand black loci to autoimmune lupus nephritis. Proc Natl Acad Sci USA 1996; 93: References 15164–15168. 20 Morel L, Tian XH, Croker BP, Wakeland EK. Epistatic 1 Wakeland EK, Liu K, Graham RR, Behrens TW. Delineating modifiers of autoimmunity in a murine model of lupus the genetic basis of systemic lupus erythematosus. Immunity nephritis. Immunity 1999; 11: 131–139. 2001; 15: 397–408. 21 Wandstrat AE, Nguyen C, Limaye N, Chan AY, Subramanian 2 Obata Y, Tanaka T, Stockert E, Good RA. Autoimmune and S, Tian XH et al. Association of extensive polymorphisms in lymphoproliferative disease in (B6-GIX+ X 129)F1 mice: the SLAM/CD2 gene cluster with murine lupus. Immunity relation to naturally occurring antibodies against murine 2004; 21: 769–780. leukemia virus-related cell surface antigens. Proc Natl Acad Sci 22 Morel L, Rudofsky UH, Longmate JA, Schiffenbauer J, USA 1979; 76: 5289–5293. Wakeland EK. Polygenic control of susceptibility to murine 3 Botto M, Dell’Agnola C, Bygrave AE, Thompson EM, Cook systemic lupus erythematosus. Immunity 1994; 1: 219–229. HT, Petry F et al. Homozygous C1q deficiency causes 23 Kono DH, Burlingame RW, Owens DG, Kuramochi A, Balderas glomerulonephritis associated with multiple apoptotic bodies. RS, Balomenos D et al. Lupus susceptibility loci in New Nat Genet 1998; 19: 56–59. Zealand mice. Proc Natl Acad Sci USA 1994; 91: 10168–10172. 4 Bickerstaff MC, Botto M, Hutchinson WL, Herbert J, Tennent 24 Vidal S, Kono DH, Theofilopoulos AN. Loci predisposing GA, Bybee A et al. Serum amyloid P component controls to autoimmunity in MRL-Fas lpr and C57BL/6-Faslpr mice. chromatin degradation and prevents antinuclear autoimmu- J Clin Invest 1998; 101: 696–702. nity. Nat Med 1999; 5: 694–697. 25 XieS,ChangSH,SedrakP,KaliyaperumalA,DattaSK,MohanC. 5 Santiago-Raber ML, Lawson BR, Dummer W, Barnhouse M, Dominant NZB contributions to lupus in the (SWR Â Koundouris S, Wilson CB et al. Role of cyclin kinase inhibitor NZB)F1 model. Genes Immun 2002; 3 (Suppl 1): S13–S20. p21 in systemic autoimmunity. J Immunol 2001; 167: 4067–4074. 26 Kikuchi S, Fossati-Jimack L, Moll T, Amano H, Amano E, Ida 6 Bygrave AE, Rose KL, Cortes-Hernandez J, Warren J, Rigby RJ, A et al. Differential role of three major New Zealand Black- Cook HT et al. Spontaneous autoimmunity in 129 and C57BL/ derived loci linked with Yaa-induced murine lupus nephritis. 6 mice-implications for autoimmunity described in gene- J Immunol 2005; 174: 1111–1117. targeted mice. PLoS Biol 2004; 2: E243. 27 Takeuchi O, Fisher J, Suh H, Harada H, Malynn BA, 7 Drake CG, Rozzo SJ, Hirschfeld HF, Smarnworawong NP, Korsmeyer SJ. Essential role of Bax, Bak in B cell homeostasis Palmer E, Kotzin BL. Analysis of the New Zealand Black and prevention of autoimmune disease. Proc Natl Acad Sci contribution to lupus-like renal disease. Multiple genes that USA 2005; 102: 11272–11277. operate in a threshold manner. J Immunol 1995; 154: 2441–2447. 28 Burlingame RW, Rubin RL. Subnucleosome structures as 8 Vyse TJ, Rozzo SJ, Drake CG, Izui S, Kotzin BL. Control of substrates in enzyme-linked immunosorbent assays. J Immu- multiple autoantibodies linked with a lupus nephritis suscept- nol Methods 1990; 134: 187–199. ibility locus in New Zealand black mice. J Immunol 1997; 158: 29 Manly KF, Olson JM. Overview of QTL mapping software and 5566–5574. introduction to map manager QT. Mammlian Genome 1999; 10: 9 Morel L, Blenman KR, Croker BP, Wakeland EK. The major 327–334. murine systemic lupus erythematosus susceptibility locus, 30 Lander E, Kruglyak L. Genetic dissection of complex traits: Sle1, is a cluster of functionally related genes. Proc Natl Acad guidelines for interpreting and reporting linkage results. Nat Sci USA 2001; 98: 1787–1792. Genet 1995; 11: 241–247.

Genes and Immunity