Abstract
(SWR × New Zealand Black (NZB))F1 (or SNF1) mice succumb to lupus nephritis. Although several NZB lupus susceptibility loci have been identified in other crosses, the potential genetic contributions of SWR to lupus remain unknown. To ascertain this, a panel of 86 NZB × F1 backcross mice was immunophenotyped and genome scanned. Linkage analysis revealed four dominant SWR susceptibility loci (H2, Swrl-1, Swrl-2, and Swrl-3) and a recessive NZB locus, Nba1. Early mortality was most strongly linked to the H2 locus on chromosome (Chr) 17 (log likelihood of the odds (LOD) = 4.59 − 5.38). Susceptibility to glomerulonephritis was linked to H2 (Chr 17, LOD = 2.37 − 2.70), Swrl-2 (Chr 14, 36 cM, LOD = 2.48 − 2.71), and Nba1 (Chr 4, 75 cM, LOD = 2.15 − 2.23). IgG antinuclear autoantibody development was linked to H2 (Chr 17, LOD = 4.92 − 5.48), Swrl-1 (Chr 1, 86 cM, colocalizing with Sle1 and Nba2, LOD = 2.89 − 2.91), and Swrl-3 (Chr 18, 14 cM, LOD = 2.07 − 2.13). For each phenotype, epistatic interaction of two to three susceptibility loci was required to attain the high penetrance levels seen in the SNF1 strain. Although the SWR contributions H2, Swrl-1, and Swrl-2 map to loci previously mapped in other strains, often linked to very similar phenotypes, Swrl-3 appears to be a novel locus. In conclusion, lupus in the SNF1 strain is truly polygenic, with at least four dominant contributions from the SWR strain. The immunological functions and molecular identities of these loci await elucidation.
Murine models of lupus have contributed significantly to our understanding of this systemic autoimmune disease (reviewed in Refs. 1, 2, 3, 4). In particular, the BWF1 (F1 intercross between New Zealand Black (NZB) and New Zealand White (NZW)) (3) and NZM2410 (an inbred strain comprised genomically of 75% NZW and 25% NZB) lupus strains have been intensively studied with respect to the genomic locations of disease susceptibility loci (5, 6, 7, 8, 9, 10, 11, 12). In addition, genetic studies in the BXSB and MRL/lpr strains have also shed light on additional loci predisposing to lupus (13, 14, 15, 16, 17, 18). Among all these loci, four certainly appear to tower above the rest, as they light up in most of the studies. These include loci on chromosomes (Chr)3 1 (Sle1/Nba2), 4 (Sle2/Lbw2/Sbw2 and Nba1), 7 (Sle3/Sle5), and 17 (H2). More recently, congenic studies have also shed light on the respective roles of these loci in disease pathogenesis (19, 20, 21, 22, 23, 24, 25).
Of the loci that have been identified in the NZM2410 and BWF1 models, several are NZW derived (such as Sle1, Sle3, and part of Sle2), whereas others such as Nba1 and Nba2 are NZB encoded. Thus, it is clear that the epistatic interaction of NZB-derived and NZW-derived loci is required for the genesis of autoimmunity in these models. However, the NZW strain is not the only strain that can engender lupus in epistasis with NZB; SWR is another strain that does so (26, 27). Thus, (SWR × NZB)F1 (or SNF1) mice develop lupus nephritis that is clinically very similar (in onset, severity, female bias, and pathology) to the BWF1 model. Indeed, studies in this model have clearly demonstrated the relative contributions of Ag-specific and Ag-nonspecific modalities of T cell:B cell crosstalk toward disease pathogenesis (reviewed in Refs. 28 and 29). However, the genetic contributions of the SWR genome to disease in this model are poorly understood. Extrapolating from the findings in the BWF1 and NZM2410 models, it is reasonable to expect several SWR-derived loci to be important for disease development in this model. To elucidate the dominant SWR contributions to lupus, we have chosen to examine a panel of 86 (NZB × SNF1) backcross (BC) mice to ascertain whether the development of autoantibodies, glomerulonephritis (GN), or early mortality could be linked to any specific (dominant) SWR and (recessive) NZB loci, using a panel of 122 microsatellite markers, spanning all 19 autosomes.
Materials and Methods
Mice and phenotyping
The derivation of the 86 (NZB × SNF1) BC mice has previously been detailed (27, 30). Thus, all loci in the progeny from this cross bear at least one NZB allele (derived from the NZB parent), while the other allele is either SWR or NZB in origin, as it arises from the F1 parent. Mice were sero-tested at 9–12 mo of age, or earlier if moribund. Seropositive mice were monitored for evidence of proteinuria, at weekly intervals, using albustix strips. Mice that exhibited persistent proteinuria for 2 consecutive wk, or appeared moribund, were sacrificed. Upon sacrifice, their kidneys were examined for evidence of GN. The glomeruli were screened for evidence of hypertrophy, proliferative changes, hyaline deposits, and/or basement membrane thickening, in a blinded fashion. The severity of GN was graded on 0–4 scale, in which the grades 1, 2, 3, and 4 were accorded when 1–10%, 11–25%, 26–50%, and >50% of the glomeruli were affected, respectively, as detailed elsewhere (31).
The following phenotypes were used for linkage analysis: serum IgG antinuclear autoantibodies (ANAs), IgM Abs, histological GN, and age of death. The ANA specificities that were examined are detailed below. For all phenotypes studied, the mice were arbitrarily classified as being negative, low positives, or strong positives, following the criteria listed in Table I⇓. As detailed below, these phenotypic categories were classified in three different ways, for the purpose of linkage analysis. No significant sex bias was seen in any of the phenotypes studied. Therefore, the data from both the male and female progeny were combined for analysis.
Criteria used to define whether a progeny (n = 86 backcross mice) was classified as being negative, low positive, or high positive, for any given phenotype
ELISA for total Ig
Total serum IgM and IgG levels were assayed using a sandwich ELISA. Briefly, goat anti-mouse IgM or IgG (Boehringer Mannheim, Indianapolis, IN) was first coated onto Immulon I plates (Dynatech, Chantilly, VA) and blocked. Serum was diluted serially and added to the plates for 2 h at room temperature. Bound Ig was revealed with alkaline phosphatase-conjugated goat anti-mouse IgM or IgG Abs (Boehringer Mannheim), using p-nitrophenyl phosphate as a substrate. Serial dilutions of isotype-specific Ig standards were also added to each plate for quantitation and interplate standardization.
ELISA for autoantibodies
IgM anti-ssDNA, IgM anti-histone, and IgM anti-DNP/keyhole limpet hemocyanin Abs were assayed as described (20). The anti-dsDNA, anti-histone, and anti-histone/DNA ELISAs were conducted as detailed elsewhere (32). Briefly, for the anti-dsDNA ELISA, Immulon II plates (Dynatech) precoated with methylated BSA were coated overnight with 50 μg/ml dsDNA (Sigma-Aldrich, St. Louis, MO), dissolved in PBS, and filtered through cellulose acetate before use. For the anti-histone/DNA ELISA, the dsDNA-coated plates were then postcoated with 10 μg/ml total histones (a mixture of all histones; purchased from Boehringer Mannheim) overnight at 4°C. After blocking with PBS/3% BSA/0.1% gelatin/3 mM EDTA, 1/100 dilutions of the test serum or 1/2 dilutions of culture supernatants were incubated in duplicate for 2 h at room temperature. Bound IgG was detected with alkaline phosphatase-conjugated anti-mouse IgG (Jackson ImmunoResearch Laboratories, West Grove, PA), using p-nitrophenyl phosphate as a substrate. Raw OD was converted to U/ml, using a positive control serum derived from an NZM2410 mouse, arbitrarily setting the reactivity of a 1/100 dilution of this serum to 100 U/ml. Sera with reactivities stronger than the test standard were diluted further and reassayed. The ANA serotiters in the BC mice were compared with sera from a panel of twelve 9- to 12-mo-old SWR control mice. Serotiters that exceeded the SWR means by 3 or 8 SDs were classified as being low positives, or strong positives, respectively, as indicated in Table I⇑.
Genotyping
Oligonucleotide primers flanking microsatellite repeats were synthesized commercially (Integrated DNA Technologies, Coralville, IA). A panel of 122 primers that readily distinguished SWR from NZB alleles was utilized. Together, they spanned 1500 cM of the autosomal genome (covering all 19 Chr), with an average intermarker distance of 10 cM. The sequences of the selected primers are publicly available (http://www-genome.wi.mit.edu/). The chromosomal positions of the primers with respect to the acromere are reported in accordance with the Mouse Chromosome Committee Reports obtained through the Encyclopedia of the Mouse Genome, Mouse Genome Database, The Jackson Laboratory (Bar Harbor, ME) (http://www.imformatics.jax.orgl). Genotyping was performed using tail DNA. PCR amplification was performed in an Eppendorf MasterCycler (Eppendorf Scientific, Westbury, NY), generally using 30 cycles of 30 s at 94°C, 1 min at 55°C, and 30 s at 72°C. For certain primers, the annealing temperature was ramped up to 60°C or ramped down to 50°C, for optimal results. Amplified products were electrophoresed onto 5% agarose gels and visualized by ethidium bromide staining and UV transillumination. The mice were then scored as being “N” (or NN, homozygous for the NZB allele), or “H” (heterozygous for SWR/NZB alleles).
Linkage analysis and statistics
Linkage analysis was performed using two different software programs, MapManager.QTX (http://mapmgr.roswellpark.org/mapmgr.html) and Bymarker (http://www.infosci.coh.org/jal/bymarker02/bymarker02.htm), as described (33). These two programs use very different algorithms. The Bymarker software performs marker-specific linkage analysis and calculates the log likelihood of the odds (LOD) scores and p values at each locus, using the Pearson’s χ2 test. In contrast, MapManager.QTX is a QTL mapping program that was used to perform interval mapping for each entire Chr, using the Kosambi mapping function. As evident from the results, there is good concordance between these two algorithms. The phenotype data were classified and analyzed in three different modes. In mode A, all offspring were classified as being either negative or positive (i.e., including both low positives and strong positives), following the criteria outlined in Table I⇑. In classification mode B, only the high positives were considered to be positive, whereas all other progeny were classified as negative. Finally, classification mode C entailed extreme phenotype analysis (negatives and strong positives only), essentially excluding the low positives from analysis. Thus, whereas classification modes A and B utilized all 86 progeny, mode C used only ∼80–90% of the progeny, having excluded the borderline positives. As is evident in Results, all three classification modes were successful in identifying most of the loci.
Since genome-wide search essentially entails multiple hypotheses testing, a threshold for suggestive linkage was set at LOD > 1.9, p < 0.0034 (χ2 > 8.6, 1 df), based on the recommendation of Lander and Kruglyak (34). The threshold for significant or probable linkage was set at a LOD > 3.3, p < 1 × 10−4 (χ2 > 15.1, 1 df). In this communication, all loci showing evidence of linkage to any of the tested lupus traits at p < 0.01 are reported. SWR-derived loci that were identified at least at the suggestive level (LOD > 1.9) by both the Bymarker linkage analysis algorithm and the MapManager interval mapping algorithm were assigned names.
Correlation between serum ANA, GN, and mortality was performed using the χ2 test after classifying individual mice as being high positive, low positive, or normal for each given phenotype. Likewise, epistatic interactions between loci were tested for using χ2 test, with Yate’s correction for continuity where necessary.
Results
Correlation between serology and disease
As illustrated in Table II⇓, ∼20–40% of BC progeny exhibit various component lupus phenotypes, including autoantibodies, GN, or early mortality. This is consistent with the notion that two or more loci, in various epistatic combinations, are likely to be responsible for each of the studied phenotypes. It is also clear that IgG anti-dsDNA, anti-ssDNA, and anti-histone/DNA ANAs in these mice correlate with GN scores and early mortality. Indeed, both the latter phenotypes themselves show strong correlation with each other (p = 0.001). These observations are consistent with the notion that the early mortality in these mice is largely due to IgG ANA-induced GN. However, it is noteworthy that the IgM Abs assayed do not correlate with the GN scores or early mortality.
Distribution of autoimmune phenotypes among 86 (NZB × SNF1) BC progeny
Mapping of loci for early mortality and GN
As illustrated in Table III⇓, the most potent locus affecting early mortality is the H2 locus on proximal Chr 17 (D17mit36, 20 cM, LOD = 4.59 − 5.38), with a dominant contribution from the SWR allele, as identified by both algorithms. This allele is clearly linked to several other phenotypes, as discussed below. The second largest contribution to mortality arises from the recessive NZB allele on mid-Chr 4 (D4mit147, 55 cM, MapManager LOD = 1.48, Bymarker LOD = 2.21), which qualifies as a suggestive locus, at least by the Bymarker algorithm. It is to be noted that this locus maps to the same position as the Lbw2/Sbw2 locus (also of NZB origin) described by Kono et al. (7) and the Sle2 locus (NZB/NZW origin) described by Morel et al. (11). Finally, a weak NZB recessive contribution is noted from a locus on mid-Chr 10.
Mapping of loci predisposing to early mortality in NZB × SNF1 mice
Interestingly, of the four loci that show some evidence of linkage to GN, three originate from the SWR strain, as summarized in Table IV⇓. Once again, the H2 locus plays a dominant role in GN susceptibility (MapManager LOD = 2.37, Bymarker LOD = 2.70). There is also a dominant SWR contribution to GN from proximal Chr 14 (D14mit37, 36 cM, MapManager LOD = 2.48, Bymarker LOD = 2.71). This locus reaches the suggestive level of significance using both analysis algorithms and is accorded the name Swrl-2 (SWR lupus locus 2). A weaker contribution to GN is noted from the SWR allele on mid Chr 6 (MapManager LOD = 1.5, Bymarker LOD = 1.43). This locus has been noted in previous NZB/B6 crosses (12) and NZB/NZW crosses (Lbw4, Ref. 7). Finally, the NZB allele on distal Chr 4 also lights up as a suggestive locus for GN (75 cM, MapManager LOD = 2.15, Bymarker LOD = 2.23). Based on its position, strain of origin, and linked phenotype, this locus is very likely to be the Nba1 locus, described by Kotzin and coworkers (6, 8). It is to be noted that the three strongest GN-susceptibility loci, H2, Swrl-2, and Nba1, are all identified as being suggestive, both the MapManager and the Bymarker analysis algorithms, almost irrespective of the phenotype classification mode selected (Table IV⇓).
Mapping of loci predisposing to GN in NZB × SNF1 mice
Mapping of loci for autoantibodies
There are several genetic contributions to IgG ANA formation, as listed in Table V⇓. Importantly, the three strongest contributions all arise from the SWR strain. The strongest SWR contribution arises from the H2 locus (MapManager LOD = 5.48, Bymarker LOD = 4.92). A second important SWR contribution arises from distal Chr 1, mapping similarly as Nba2/Sle1 (6, 9, 11), with a LOD score of 2.89 − 2.91. Interestingly, this locus shows the strongest linkage to IgG anti-histone/DNA ANAs, rather than anti-dsDNA ANAs, with little evidence of linkage to IgG anti-ssDNA ANAs, GN, or mortality, as diagrammed by the interval mapping chromosomal scan in Fig. 1⇓. This locus is assigned the name Swrl-1 (SWR lupus locus 1). Another SWR locus that impacts IgG anti-histone/DNA, anti-dsDNA, and anti-ssDNA ANAs in a dominant fashion is located on proximal Chr 18 (LOD > 2). This locus is accorded the name Swrl-3 (SWR lupus locus 3). This locus appears to be a novel locus implicated in murine lupus. In addition, as summarized in Table V⇓, there are also weaker contributions from the NZB genome to IgG ANA formation, notably on proximal Chr 5 and telomeric Chr 10. Once again, it is clear that the three strongest ANA -susceptibility loci, H2, Swrl-1, and Swrl-3, are all identified as being suggestive, both the MapManager and the Bymarker analysis algorithms, irrespective of the phenotype classification mode selected (Table V⇓). Finally, several NZB and SWR loci contribute to the formation of IgM Abs with specificities for ssDNA, histones, or DNP/keyhole limpet hemocyanin, as summarized in Table VI⇓. These include suggestive loci on proximal Chr 1 (similarly positioned as BXSB-derived Bxs1) and proximal Chr 6.
Swrl-1 on Chr 1 is linked to IgG ANA production. Interval mapping scans for GN (▴), early mortality (dashed line), IgG anti-dsDNA ANAs (thin solid line), and IgG anti-histone/DNA ANAs (thick solid line) were performed using MapManager. QTX.LOD scores depicted on the y-axis were calculated at 2-cM intervals. The microsatellite marker loci used and their respective chromosomal positions (in cM from acromere) are also depicted. Total progeny examined = 86.
Mapping of loci predisposing to IgG ANAs in NZB × SNF1 mice
Mapping of loci predisposing to elevated serum IgM antibodies in NZB × SNF1 mice
Taking all the above data into consideration, the five strongest contributions to the various disease phenotypes identified by the locus-specific Bymarker linkage analysis algorithm, as well as the MapManager interval mapping algorithm, include the following: H2 (Chr 17, SWR dominant), Swrl-1 (Chr 1, SWR dominant), Swrl-2 (Chr 14, SWR dominant), Swrl-3 (Chr 18, SWR dominant), and the recessive NZB locus, Nba1, on Chr 4, as summarized in Table VII⇓. Also indicated in Table VII⇓ are the penetrance rates of the different phenotypes in offspring with the indicated susceptibility genotypes, as well as a listing of potential candidate genes within these loci. It should be pointed out that none of these loci are absolutely required for disease or ANA formation, because at least one phenotype-positive animal could be identified bearing the nonsusceptible genotype at each of these five susceptibility loci.
The major SWR/NZB loci conferring susceptibility to lupus, and the potential candidate genes located within these loci
Epistatic interactions between susceptibility loci
We next examined for evidence of epistatic interactions between the different loci, with respect to the different phenotypes studied. Although the H2 locus on Chr 17 and the Sle2/Lbw2/Sbw2 locus on Chr 4 both confer susceptibility to early mortality, we could not demonstrate any statistically significant epistatic interactions between these two loci. However, we could demonstrate epistatic interactions between the loci that were linked to the other phenotypes. Among the loci that contribute to GN, those on Chr 4 and Chr 14 demonstrate strong epistatic interaction, as depicted in Table VIII⇓. Thus, mice that bear both susceptibility alleles (i.e., NZB homozygous at Nba1 on Chr 4 and NZB/SWR heterozygous at Swrl-2 on Chr 14) have significantly higher incidence of GN (penetrance = 66%), compared with progeny with any of the other three potential genotypes at these loci. However, neither of these two loci shows any epistatic interactions with the H2 locus. In other words, progeny that are NZB/SWR heterozygous at H2 have a higher incidence of GN (penetrance = 51%, Table VIII⇓), irrespective of their genotypes at the Nba1 and Swrl-2 loci.
Epistatic interactions between loci predisposing to GN or ANAs, among NZB × SNF1 BC mice (n = 86)
We next examined the three loci that confer susceptibility to IgG anti-dsDNA and IgG anti-histone/DNA ANAs: H2 on Chr 17, Swrl-1 on Chr 1, and Swrl-3 on Chr 18. Although we could not demonstrate any statistically significant epistatic interactions between Swrl-1 and either of the other two loci, the H2 and Swrl-3 loci show strong epistatic interactions with each other. Thus, mice that are heterozygous at both these loci exhibit IgG anti-dsDNA and anti-histone/DNA ANAs with a penetrance of ∼67–71%, as detailed in Table VIII⇑. This is significantly higher than the penetrance of ANAs seen in progeny bearing any of the other three two-locus genotypes. Finally, Fig. 2⇓ charts the incremental contribution to disease and phenotype penetrance as increasing numbers of susceptibility loci are inherited. Thus, mice that inherit three strong susceptibility loci exhibit penetrance rates of GN and IgG ANAs that parallel the observed penetrance of these phenotypes in the original SNF1 lupus strain.
Penetrance of component lupus phenotypes is dictated by the genetic load. Penetrance rates were calculated using all positive mice (i.e., pooling both the low and high positives). For the penetrance of GN (dashed line), the genotypes at D4mit190 (Nba1), D14mit37 (Swrl-2), and D17mit36 (H2) were considered. A total of 14 mice had none of these three GN-susceptible genotypes, 28 had one, 34 had two, and 10 progeny had the GN-susceptible genotypes at all three loci. For the penetrance of IgG anti-dsDNA (dotted line) and IgG anti-histone/DNA ANAs (solid line), the genotypes at D17mit36 (H2), D1mit15 (Swrl-1), and D18mit17 (Swrl-3) were considered. A total of 20 mice had none of these three ANA-susceptible genotypes, 26 had one, 22 had two, and 18 progeny had the ANA-susceptible genotypes at all three loci. Plotted on the x-axis are the numbers of susceptibility loci inherited by a progeny. Thus, “2” refers to all progeny bearing any two of the three susceptibility loci for the given phenotype. Total number of progeny = 86.
Discussion
Given the lessons learned from the NZM2410 and NZB/NZW genetic studies, we hypothesized that several dominant SWR loci must be contributing to disease in the SNF1 lupus model. The present study supports this hypothesis and points to the existence of at least four dominant SWR contributions: H2 on Chr 17, Swrl-1 on Chr 1, Swrl-2 on Chr 14, and Swrl-3 on Chr 18, as summarized in Table VIII⇑. Of all these loci, the locus that clearly towers above the rest is the H2 locus on Chr 17. Early mortality, GN, and IgG ANA formation are all strongly facilitated by the SWR/NZB genotype at this locus with LOD scores in excess of 4 (Tables III-V). This is also the one locus that has been reproducibly identified in almost all murine lupus genetic studies to date (1, 2, 3, 4, 7, 8, 11, 35, 36, 37, 38). Indeed, heterozygosity at the H2 locus has also been implicated in disease by earlier restriction-fragment length polymorphism studies in the SNF1 model (30).
More recently, the same locus has been mapped as the strongest suppressor locus in NZW × B6.Sle1 crosses (39). The later study demonstrates that the H2z/z genotype (or Sles1NZW/NZW) strongly shuts down Sle1-triggered ANA production. Indeed, the findings of the present study are consistent with all of the above reports. Thus, mice that genotype as H2SWR/NZB exhibit significantly higher penetrance of autoantibodies and disease, compared with mice that type as H2NZB/NZB. Moreover, among mice that are heterozygous at Swrl-1 (i.e., mice bearing the Chr 1 susceptibility locus that maps similarly to Sle1), having the H2NZB/NZB genotype on Chr 17 totally shuts down ANA production. All these observations support the recognition of the H2 locus as one of the most potent determinants of murine lupus susceptibility.
Although it is unequivocally accepted that the H2 locus is a major genetic determinant, the culprit gene(s) within this locus still remains undetermined. In most of these studies, heterozygosity at the H2 locus (e.g., d/q, d/z, or b/z genotypes) is most strongly associated with disease. The H2 molecules themselves are attractive candidates, and several interesting ideas and potential mechanisms have been put forth to explain how and why heterozygosity at this locus might confer disease susceptibility (40, 41, 42). More recently, elegant I-A or I-E transgene reconstitution experiments conducted by Kotzin and coworkers (43, 44) have demonstrated that these molecules do not really confer disease susceptibility, thus dampening the earlier enthusiasm for these molecules. The TNFα gene within this locus still remains as an attractive candidate (45, 46, 47). Importantly, the H2 locus is richly packed with several genes of immunological interest, a sampling of which is listed in Table VIII⇑. Based on these studies, it is reasonable to expect the culprit gene not to be the H2 molecule itself, but a linked gene, still shrouded in mystery.
Swrl-1 maps right into a chromosomal interval that is also certainly not new to lupus researchers. It maps similarly to Sle1 mapped in NZM2410/NZW strains (11), and Nba2 mapped in the NZB strain (6, 9). Likewise, a BXSB locus at this position has also been implicated in disease (13). Most excitingly, a susceptibility locus for human SLE has been recognized to be syntenic to these murine Chr 1 loci (48, 49). It is certainly conceivable that the culprit molecule(s) within these different murine (and human) loci may be polymorphic variants of the same gene, but this awaits molecular confirmation. In the meantime, congenic studies of Sle1, on the normal B6 background, have revealed that Sle1 is a potent locus that breaches tolerance, leading to high levels of anti-histone/DNA (but not anti-dsDNA) ANAs, with little attendant pathology (21). In this context, it is very interesting to note that the phenotype that shows the strongest linkage to Swrl-1 is IgG anti-histone/DNA ANAs, not anti-dsDNA, anti-ssDNA ANAs, early mortality, or GN, as illustrated in Fig. 1⇑. These observations suggest that Swrl-1 might function analogously to Sle1. This hypothesis is currently being tested by the derivation of congenic strains bearing Swrl-1.
In contrast to the H2 and Swrl-1 loci, the Swrl-2 and Swrl-3 loci map to genomic positions that have been infrequently implicated in murine lupus studies. Interestingly, Rozzo et al. (9) have linked GN susceptibility to a similarly positioned NZB locus on Chr 14, notably in H2z-positive progeny. In view of the colocalization and the similarity in the associated phenotype (i.e., GN), Swrl-2 is likely to be a genuine GN susceptibility locus, in which the SWR allele might be more potent that the NZB counterpart. This hypothesis needs to be verified by congenic dissection analysis. In contrast, Swrl-3 appears to be a novel locus, mapping to a chromosomal position not previously implicated in murine lupus susceptibility. Although it surfaces as a suggestive locus using both the analysis algorithms, its authenticity certainly warrants confirmation in additional, independent data sets. Since it is linked only to IgG ANAs (but not to any of the other tested phenotypes), it might play an important role in breaking tolerance to nuclear Ags. Although there are several potential candidates within this region (see Table VIII⇑), the functional role of this locus would become clearer once congenic strains bearing this locus become available.
In addition to these four SWR loci, the NZB genome appears to be contributing perhaps two recessive loci, both localized to Chr 4. The NZB locus positioned similarly to Sle2/Lbw2/Sbw2 is linked to early mortality, but not to any of the other tested phenotypes studied (Table III⇑). However, this locus has been linked to several other phenotypes, including ANAs, splenomegaly, and GN, in crosses performed with other strain combinations (7, 9, 11, 13, 17). These observations suggest that the expression of the locus on mid-Chr 4 (55 cM) is heavily influenced by the rest of the genomic context. However, it should be cautioned that this locus did not reach suggestive levels of significance using the MapManager interval mapping algorithm. In contrast, the more distal, NZB-derived Nba1 locus on Chr 4 (identified as being a suggestive locus by both mapping algorithms) is linked to GN, but not ANAs, in the present study, as well as in previous mapping studies by other investigators (6, 8). Indeed, it is tempting to speculate that this locus on distal Chr 4 may be impacting local events in the end organs, but not systemic immunity. These predictions need to be verified by congenic studies.
As depicted in Table VIII⇑ and Fig. 2⇑, it takes the epistatic interaction of two to three susceptible loci to attain the high penetrance rates seen in SNF1 mice. Thus, mice with the resistant genotype at all three susceptibility loci have <10% penetrance of GN or IgG ANAs (Fig. 2⇑). With the inheritance of each additional susceptibility locus, the penetrance (or chance of expressing this phenotype) escalates by ∼20–30%. Thus, although lupus pathogenesis in SWR/NZB mice is truly polygenic, the epistatic interaction of two to three potent loci is all that may be required for orchestrating disease. These findings parallel the observations in the other murine models of lupus (1, 2, 3, 4) and are likely to represent a generalized paradigm for both murine and human lupus. Indeed, congenic reconstitution studies have provided experimental support for this “2–3 gene” paradigm for lupus susceptibility (4, 24, 25).
Mapping of these five loci (H2, Swrl-1, Swrl-2, Swrl-3, and Nba1) is a first step toward understanding the genetic basis of lupus in the SNF1 model. Although the above studies highlight the most potent dominant SWR loci contributing to SNF1 disease, the (NZB × SNF1) BC may not have been optimal in uncovering all potential loci. For instance, recessive SWR loci and loci that require certain other recessive loci for full expression of disease will not be mapped in this BC panel. Linkage analysis with further panels of (SWR × SNF1) and/or (SWR × NZB)F2 mice will be necessary to uncover these additional loci. Finally, to fathom the functions of these loci, it is instructive to introgress these individual loci onto a normal genetic background for detailed immunophenotyping analysis, as has been accomplished for the NZM2410-derived susceptibility loci. Undoubtedly, the ultimate challenge is to decipher the culprit genes within these loci. Charting out the myriad pathways and genes that can potentially pave the way to lupus, using the different, naturally occurring, spontaneous mouse models of lupus, stands to boost our understanding of this complex, systemic autoimmune disease.
Acknowledgments
We thank Drs. Laurence Morel and Edward Wakeland for helpful discussions and technical assistance.
Footnotes
-
↵1 Work in our laboratory is funded by grants from the National Institutes of Health (AR44894, AI41985, and AI47460) and the Arthritis Foundation. C.M. is a recipient of the Robert Wood Johnson Jr. Arthritis Investigator Award.
-
↵2 Address correspondence and reprint requests to Dr. Chandra Mohan, Simmons Arthritis Research Center, Department of Internal Medicine/Rheumatology, University of Texas Southwestern Medical School, Mail Code 8884, Y8.204, 5323 Harry Hines Boulevard, Dallas, TX 75390-8884. E-mail address: chandra.mohan{at}utsouthwestern.edu
-
↵3 Abbreviations used in this paper: Chr, chromosome; ANA, antinuclear autoantibody; BC, backcross; GN, glomerulonephritis; LOD, log likelihood of the odds.
- Received June 5, 2001.
- Accepted October 3, 2001.
- Copyright © 2001 by The American Association of Immunologists
References
In this issue
