Anti-subnucleosome reactivities in systemic lupus erythematosus (SLE) patients and their first-degree relatives

Abstract

Antibodies specific for dsDNA appear to have different genetic origins and pathogenic consequences, compared with histone/dsDNA-specific antibodies, in a recently described murine model. The purpose of this study was to examine if this is also true in human lupus. Sera from 40 SLE families (comprising 40 probands and 153 first-degree relatives), and 45 normal adult controls were assayed for the levels of anti-dsDNA, anti-H1/dsDNA, anti-H2A/H2B/dsDNA, and anti-H3/H4/dsDNA autoantibodies by ELISA. Both the probands and the first-degree relatives exhibited significantly increased levels of antinuclear antibodies (ANA) targeting the different subnucleosomal epitopes. Importantly, probands with anti-dsDNA antibodies had a significantly higher incidence of renal disease compared with those with just anti-H2A/H2B/dsDNA antibodies, in resonance with murine studies. The frequency of anti-dsDNA and anti-H2A/H2B/DNA ANA among the first-degree relatives was 11·8% and 18·3%, respectively. Surprisingly, whereas probands with anti-dsDNA ANA had families with several seropositive members, first-degree relatives of patients with anti-H2A/H2B/DNA ANA (but not anti-dsDNA ANA) were uniformly ANA-free. These findings suggest that anti-dsDNA ANA in lupus may not only have worse disease associations, they may also have very different genetic origins, compared with anti-H2A/H2B/DNA (or anti-nucleosome) ANA.

Introduction

A wide spectrum of immunological aberrations is known to underlie lupus pathogenesis [1,2]. The genesis of serum antinuclear antibodies (ANA) is one of the diagnostic hallmarks of this disease, both in man and in mice. However, ‘ANA’ constitute a very heterogeneous group of autoantibodies, and not all ANA are associated with pathogenicity [3–5]. In contrast to dsDNA-reactive ANA, anti-nucleosome (or chromatin-specific) ANA react preferentially to conformational determinants on histone/DNA complexes. Though the presence of these two distinct specificities has been documented in human lupus, the relative pathogenic (i.e. nephritogenic) potential of anti-nucleosome (anti-chromatin ANA specific for H2A/H2B/DNA subnucleosomes, in particular) versus anti-dsDNA ANA remains to be solved [6–10]. Interestingly, patients with drug-induced lupus exhibit high titres of H2A/H2B/DNA-reactive ANA in the absence of dsDNA-reactive ANA [11,12]. Importantly, these patients do not generally develop glomerulonephritis (GN), suggesting that H2A/H2B/DNA subnucleosome-specific ANA may not be a direct pathogenic mediator of GN. However, under certain experimental conditions, ANA with the same specificities have been demonstrated by other investigators to target renal glomeruli via antigenic bridges composed of nucleosomes and anionic basement membrane antigens [13–15]. Work in murine models has contributed to our understanding of the molecular features and antigenic specificities of pathogenic ANA. Interestingly, in both murine and human lupus, the subnucleosome specificity of ANA shows an age-dependent evolution, targeting H2A/H2B/DNA subnucleosomes initially, and then ‘spreading’ to involve other chromatin epitopes as the disease progresses [6,9,16,17]. These serological findings, and experimental evidence regarding the T cell specificities in lupus, suggest that this disease arises as a consequence of autoimmunization with chromatin [18–21].

Recent studies of murine lupus reveal that ANA with different subnucleosomal specificities have different genetic origins (summarized in Fig. 1). The H2 locus, as well as three non-H2 encoded loci, Sle1, Sle2, and Sle3, confer lupus susceptibility in the NZM2410 strain [22]. Interestingly, C57Bl/6 (B6) mice (which are lupus-free) rendered congenic for these individual susceptibility intervals reveal distinct immunophenotypes [23–28]. Of relevance to this study are B6 mice congenic for Sle1 and/or Sle3. B6.NZMc1 mice, congenic for Sle1 on murine chromosome 1, exhibit high titres of ANA directed against the most exposed determinants on chromatin, i.e. H2A/H2B/DNA subnucleosomes, with little reactivity to other chromatin epitopes including histone-free dsDNA [25]. This strain is healthy, and does not develop severe GN. Sle3, by itself, leads to a spectrum of T cell aberrations resulting in activated T cells with elevated CD4:CD8 ratios [27]. In contrast, the epistatic interaction of Sle1 with Sle3 leads to a spectrum of autoimmune phenotypes not seen in the monocongenics [27,28]. These bicongenic mice, named ‘B6.NZMc1|c7’ (Fig. 1) exhibit splenomegaly, with significantly expanded populations of activated B and CD4 T cells, and a robust ANA response targeting all subnucleosomal epitopes (including dsDNA), glomeruli and basement membrane antigens. As one might expect, these mice exhibit highly penetrant severe GN. Interestingly, these phenotypes are all more prominent in female mice. Thus, in this system, loci such as Sle3 appear to facilitate the ‘pathogenic maturation’ of ANA, once tolerance to chromatin is breached by loci such as Sle1.

Fig. 1.

Two sets of genes underlie the genesis of dsDNA-reactive, nephrophilic autoantibodies in murine lupus. Sle1, Sle2, and Sle3 are three non-H2 loci that confer lupus susceptibility in the NZM2410 murine lupus model [22]. When Sle1 is bred onto the normal (B6) genetic background, it triggers the formation of a very restricted set of antinuclear antibodies (ANA) that do not appear to be pathogenic [25]. Sle3 by itself impacts T cell activation and expansion [27]. In contrast, the epistatic interaction of Sle3 with Sle1 facilitates the ‘pathogenic maturation’ of ANA, leading to dsDNA-reactive, nephrophilic autoantibodies [28].

These studies allude to the presence of at least two classes of lupus susceptibility loci with differing impacts on the subnucleosome specificities of ANA. Whereas loci such as Sle1 may function primarily to breach tolerance to chromatin, the ‘pathogenic maturation’ of the ANA response appears to require additional input from other loci (such as Sle3). If indeed similar classes of susceptibility loci also existed in humans, one could make the following two predictions. First, one would expect lupus patients with anti-dsDNA ANA to have a higher incidence of renal disease, compared with those with just anti-H2A/H2B/DNA ANA. Second, one would also expect to see an increased prevalence of H2A/H2B/DNA subnucleosome-specific ANA in SLE families (over and above the prevalence rates of dsDNA-specific ANA), compared with the baseline levels in the normal, healthy population. Thus, for instance, relatives harbouring Sle1-like loci in the absence of Sle3-like loci would be predicted to exhibit serum anti-H2A/H2B/DNA ANA, in the absence of anti-DNA ANA. The purpose of this study was to verify these two predictions. To ascertain this, a cohort of 40 SLE patients and their first-degree relatives (n = 153) were serotested for the presence of ANA reactive with dsDNA, H1/dsDNA, H2A/H2B/dsDNA, or H3/H4/dsDNA subnucleosomes. In this study, these ANA are referred to as anti-DNA, anti-H1/DNA, anti-H2A/H2B/DNA and anti-H3/H4/DNA, respectively.

Subjects and methods

Study subjects

The 40 lupus patients studied here met the ARA criteria for a diagnosis of definite SLE [29], and comprised 18 African-American, eight Hispanic, and 14 Caucasian patients. The clinical data pertaining to the SLE patients were obtained from careful review of each patient's out-patient and in-patient records. Diagnosis of lupus nephritis was based on presence of persistent proteinuria, erythrocytes and/or casts in the urine, and in all but two cases, by kidney biopsy. Most of the patients studied were followed by one of us (R.C.W.) for periods of 2–8 years. If no proteinuria and/or renal insufficiency were observed over these extended periods, the patients were classified as not having any evidence of nephritis.

The first-degree relatives of the above patients were also included in this study, with ages ranging from 12 to 78 years. Their relationships to the affected probands were as follows: 24 were the proband's mothers, 19 were fathers, 11 sons, 19 daughters, 51 sisters, and 29 brothers. The average family size was 4·8, including the proband. Each relative was personally interviewed at the time the blood sample was collected and asked about symptoms related to lupus, rheumatoid arthritis (RA) or other connective tissue diseases. Four of these first-degree relatives (in four different families) were also known to have been previously diagnosed as having lupus, and this information was verified by communication with their attending physicians wherever possible. The study controls were normal healthy adults (without clinical features of SLE) with an age and ethnic distribution similar to that of the patient population.

Elisa

The anti-dsDNA and anti-subnucleosome ELISAs were carried out as described before [30]. Briefly, Immulon II plates (Dynatech, Chantilly, VA) precoated with methylated bovine serum albumin (mBSA) were coated overnight with 50 μg/ml dsDNA (purchased from Sigma Chemical Co. (St Louis, MO), dissolved in PBS, and filtered through cellulose acetate before use). The latter filtration step served to remove denatured DNA. The plates were then post-coated with 10 μg/ml of either histone H1, a 1:1 mixture of histones H2A:H2B, or a 1:1 mixture of histones H3:H4, overnight at 4°C. All histones were purchased from Roche/Boehringer Mannheim. Thus, ‘H2A/H2B/DNA’ refers to dsDNA post-coated with a mixture of histones H2A and H2B, with a similar nomenclature being adopted for the other subnucleosomes. These different nucleosomal antigens (dsDNA, H2A/H2B/DNA, H3/H4/DNA, and H1/DNA) were coated onto serial wells of the same ELISA plates, to facilitate the comparison of the relative reactivities of any given serum to these different epitopes. The concentrations of antigens used in these ELISAs have been shown to be sufficient to saturate all available binding sites [30]. After blocking with PBS/3% BSA/0·1% gelatin/3 mm EDTA, 1:100 dilutions of the test sera (or further dilutions) were incubated in duplicate for 2 h at room temperature. Bound IgG was detected with alkaline phosphatase-conjugated anti-human IgG/IgM/IgA (Caltag, Burlingame, CA), using pNPP as a substrate. Raw optical densities (Ods) were converted to U/ml, using a positive control lupus serum (Biochemed, Lilburn, GA), arbitrarily setting the reactivity of a 1:100 dilution of this serum to 100 U/ml. This control serum showed equally strong (OD) reactivities to all the tested subnucleosomal antigens, again allowing the relative reactivities of the test sera to the different antigens to be compared with each other. Sera exhibiting reactivities > 2 s.d. in excess of the mean levels in the control population were deemed positive.

Statistical analysis

Statistical comparisons between groups were performed using Student's t-test, or the χ² test, using SigmaStat software. The seroreactivities to the four different subnucleosomal complexes were used to classify individual sera as having ‘Profile A’ (seronegative for all chromatin epitopes), ‘Profile B’ (raised ANA preferentially targeting H2A/H2B/DNA subnucleosomes, rather than dsDNA), or ‘Profile C’ (raised ANA levels to all four subnucleosomal epitopes, including dsDNA). Likewise the different families were classified as follows, based on their serum anti-subnucleosome patterns: (i) family pattern A, relatives of probands are seronegative for all four ANA; (ii) family pattern B, seropositive relatives of probands have ANA preferentially targeting H2A/H2B/DNA subnucleosomes, rather than dsDNA; (iii) family pattern C, seropositive relatives of probands have significantly increased ANA to all four subnucleosomal epitopes, including dsDNA.

Results

The anti-subnucleosomal reactivities in the SLE families are depicted in Table 1. Compared with the normal controls, the SLE patients had significantly higher levels of ANA targeting all epitopes of chromatin studied. In particular, SLE patients exhibited significantly higher levels of anti-dsDNA (P < 0·001) and anti-H2A/H2B/DNA (P < 0·001) ANA. Interestingly, the first-degree relatives of the SLE patients exhibited significantly higher levels of anti-dsDNA and anti-H2A/H2B/DNA ANA, compared with the controls. This is illustrated in greater detail in Fig. 2. Though Table 1 and Fig. 2 illustrate the prevalence of the various anti-subnucleosome ANA in SLE patients and their first-degree relatives, these do not depict the relative levels of seroreactivities to the different subnucleosomal components in individual patients or relatives. In order to analyse this, it is important to classify individual sera according to their relative subnucleosomal seroreactivities. To achieve this, we have taken advantage of the observation that the sera fell into three general groups, based on their relative reactivities to the different subnucleosomal epitopes (as illustrated in Fig. 3): Profile A: no significant reactivity to dsDNA, or any of the other histone/DNA subnucleosomes; Profile B: significant reactivity to H2A/H2B/DNA subnucleosomes, but not to dsDNA; Profile C: significant reactivity to dsDNA, as well as to other subnucleosomal epitopes. It should be noted that these three seroprofiles are akin to the seroprofiles seen in normal B6 mice, B6.NZMc1 (Sle1) monocongenics, and B6.NZMc1|c7 bicongenics (Sle1 and Sle3), respectively, as discussed above.

Table 1.

Anti-subnucleosome antinuclear antibodies (ANA; shown in ELISA units/ml) in SLE patients, their first-degree relatives, and normal human controls

	Anti-DNA	Anti-H1/DNA	Anti-H2A/H2B/DNA	Anti-H3/H4/DNA
Normal human sera (n = 45)
Mean ± s.e.m.	13 ± 1·7	29 ± 3·0	24 ± 2·3	35 ± 2·9
Positives, %	2·2	2·2	2·2	2·2
SLE probands (n = 40)
Mean ± s.e.m.	50 ± 5·3***	48 ± 5·5**	71 ± 5·1***	55 ± 5·4**
Positives, %	52·5	35	72·5	35
First degree relatives (n = 153)
Mean ± s.e.m.	21 ± 6·3*	24 ± 5·8	31 ± 5·5*	32 ± 6·3
Positives, %	11·8	9·8	18·3	7·8

Positives, %, the number of individuals in each category exhibiting ANA levels that exceed 2 s.d. beyond the mean levels seen in the normal control sera. Significant differences from normals are indicated.

^*P < 0·05

^**P < 0·01

^***P < 0·001.

Fig. 2.

Anti-subnucleosome seroreactivities in SLE patients and their first-degree relatives. SLE patients (n = 40) exhibit significantly higher levels of anti-dsDNA (labelled as anti-DNA, P < 0·001) and anti-H2A/H2B/DNA antinuclear antibodies (ANA) (P < 0·001), compared with normal controls (n = 45). The first-degree relatives also exhibit significantly higher levels of both these ANA compared with controls (P < 0·05, Table 1). Each dot represents the level of ANA (U/ml) assayed in individual sera. The dotted line represents the cut-off for defining positives, set at 2 s.d. above the mean ANA level assayed in the normal controls.

Fig. 3.

Anti-subnucleosome seroprofiles used to classify individual sera. Individual antinuclear antibody (ANA)-positive members in this study were classified according to whether they had no subnucleosome-specific ANA (Profile A), had ANA that reacted preferentially to H2A/H2B/DNA subnucleosomes (Profile B), or had ANA that were reactive to dsDNA, as well as the other subnucleosomal epitopes (Profile C). □, Mean levels of the respective ANA seen in the normal controls (n = 45). ***2 s.d. above the respective means. ▪, ANA levels measured in three individual SLE probands, one exhibiting Profile A, one Profile B, and the other, Profile C. This chart illustrates graphically what the three seroprofiles look like, relative to normal sera. The actual numbers of sera exhibiting these three different seroprofiles are detailed in Table 2.

As detailed in Table 2, nine of the 40 SLE sera were negative for all anti-subnucleosomal ANA studied. The remaining 31 sera fell into two broad seroreactivity groups: 10 sera reacted strongly with H2A/H2B/DNA subnucleosomes, but not significantly with dsDNA (Profile B, as illustrated in Fig. 3), and 21 sera reacted strongly with all subnucleosomal epitopes (Profile C, as illustrated in Fig. 3). In most cases, Profile B sera also showed little or no reactivity to H1/DNA and H3/H4/DNA subnucleosomes. Interestingly, a significantly greater fraction of the patients with ANA Profile C (15 out of 21) exhibited renal disease, compared with the probands with Profile B (two out of 10, χ² test, P < 0·01). Table 2 also shows that about 11·8% (18 out of 153) of the first-degree relatives exhibited Profile C (raised anti-DNA ANA, and also ANA targeting the other epitopes), whereas about 6·5% (10 out of 153) had raised anti-H2A/H2B/DNA ANA rather than anti-dsDNA ANA (Profile B).

Table 2.

Anti-subnucleosome seroprofiles in 40 lupus patients (probands) and their 153 first-degree relatives

	Profile A	Profile B	Profile C
Lupus probands	9	10	21
Number with nephritis (%)	0 (0)	2 (20)	15 (71·4)
First-degree relatives	125	10	18

We next examined the different families to see if there were any interesting patterns of these anti-subnucleosome seroprofiles within individual families. As described in SUBJECTS and METHODS, the 40 lupus families were classified as having family (seroreactivity) patterns A–C, depending on the ANA seroprofiles exhibited by the members within each family. Representative families exhibiting these different patterns are illustrated in Fig. 4. Twenty-six families exhibited family pattern A, whereas the remaining 35% of the lupus families (14 out of 40 families) had seropositive members (other than the probands). Of these, eight families exhibited family pattern C (increased ANA to all epitopes), this being the most common pattern in seropositive families. Five families exhibited family pattern B (elevated anti-H2A/H2B/DNA ANA but not anti-DNA ANA). A representative family with the latter pattern is Family RR depicted in Fig. 4. In this family, the father, three sisters and the brother of the proband had significantly high ANA levels targeting H2A/H2B/DNA (and H3/H4/DNA) subnucleosomes, whereas only the proband developed high titres of ANA targeting dsDNA, and clinical features of lupus. Surprisingly, only one family exhibited mixed serological patterns among its relatives (not shown).

Fig. 4.

Anti-subnucleosome seroreactivity patterns in SLE families. Twenty-six out of 40 SLE families exhibited family pattern A (all relatives were seronegative for all the four subnucleosomal epitopes assayed). Five families exhibited family pattern B (seropositive relatives had elevated anti-H2A/H2B/DNA, but not anti-dsDNA antinuclear antibodies (ANA)). Eight families exhibited family pattern C (seropositive family members all exhibited dsDNA reactivities). Finally, one family exhibited mixed profiles (not shown). Shown in Figure are three representative families, each illustrating a different family pattern of subnucleosomal seroreactivity. In the top panels, the ELISA seroreactivities to the four subnucleosomal components are diagrammed for each family member. In each chart, the proband is indicated by the continuous line with black dots, whereas the relatives are indicated using dashed lines. In the bottom panels, the respective family trees of these three families are shown. Lupus-afflicted individuals are shaded black, while seropositive relatives are hatched in these family trees.

Finally, we attempted to correlate the ANA seroprofiles of the SLE patients and the ANA patterns of their respective families. If the fine specificities of ANA were under tight genetic control, and if one could extrapolate from the murine model, one would predict (i) families to which Profile B SLE patients belong, to exhibit ANA patterns A or B, but not C, and (ii) families to which Profile C SLE patients belong, to exhibit all three ANA patterns, including C. Figure 5 demonstrates this relationship for the 40 SLE families studied. It is clear that many of the probands with Profile C (i.e. with anti-dsDNA ANA) belong to families with patterns A (10/21), B (3/21), or C (8/21). Surprisingly, in contrast, none of the Profile B SLE patients had any first-degree relatives with ANA, i.e. they all belonged to families with pattern A. Indeed, the distribution of family ANA seropatterns between Profile B probands and Profile C probands was statistically different (χ² = 8·1, P = 0·017, two degrees of freedom).

Fig. 5.

Correlation between proband antinuclear antibody (ANA) seroprofiles and family ANA patterns. The SLE probands were segregated according to their ANA profiles, and their respective families' ANA patterns. The height of each bar represents the total numbers of families. Shown are 39 (out of the 40) families; the remaining family exhibited mixed ANA patterns. The difference in the ANA family patterns between Profile B probands and Profile C probands was statistically significant (χ² = 8·1, P = 0·017, two degrees of freedom).

Discussion

That human SLE has a strong genetic basis has long been acknowledged [31–33], but virtually nothing is known about the culprit genes. Though earlier studies had primarily focused onHLA–disease associations, more recent genome-wide scans have clearly demonstrated the multigenic origins of this disease [33–35]. These studies have been fortified by a series of murine mapping studies highlighting several different disease susceptibility loci, as reviewed [36–39]. In addition to knowing which regions of the genome harbour disease genes, we are also beginning to understand how these different genes might be contributing to pathology. In particular, the non-MHC loci that confer disease susceptibility in the NZM2410 lupus strain, Sle1, Sle2 and Sle3, have been shown to impact the immune system in very different ways. As discussed earlier, in contrast to the selective loss of tolerance to chromatin triggered by Sle1, the combined action of Sle1 and Sle3 leads to splenomegaly with robust activation of T and B cells, high levels of serum autoantibodies targeting all epitopes of chromatin and glomerular antigens, and severe GN [25,27,28].

These observations prompted us to ask if loci equivalent in function to Sle1 and Sle3 might also exist in humans, and whether their presence might be detected serologically in SLE patients and their relatives. Indeed, it has long been recognized that relatives of SLE probands have a higher incidence of ANA [40–44]. Most of these studies have been conducted using immunofluorescence with Hep2 cells as substrates. However, the immunofluorescence assays do not distinguish between anti-dsDNA and anti-H2A/H2B/DNA reactivities. Of special interest is the observation that susceptibility to human lupus also maps to an interval on human chromosome 1 that is syntenic to murine Sle1 [33,35,45–49]. Furthermore, the syntenic interval on human chromosome 1 is also linked to the development of anti-histone/DNA ANA [45]. Based on these collective observations, we reasoned that if loci equivalent in function to Sle1 and Sle3 also existed in humans, the anti-subnucleosome antibody profiles of SLE patients and their relatives could potentially be revealing. Specifically, we asked the following questions: (i) do patients with Profile C ANA have a higher incidence of renal disease compared with patients with Profile B ANA?; (ii) is the incidence of Profile B ANA significantly higher in lupus families, compared with the incidence of Profile C ANA?; and (iii) is there a suggestion of a genetic relationship between the ANA seroprofiles of the patients, and the ANA patterns of their respective families?

Examination of the seroprofiles of the SLE probands in this study uncovered an interesting pattern. About 71% of the SLE patients with anti-DNA ANA (Profile C) had clinical evidence of renal disease. In contrast, most of the probands with Profile B ANA had joint manifestations as the predominant clinical feature, in the absence of renal disease. This is indeed very similar to the observations in the B6.Sle congenic murine models, where anti-DNA ANA (but not anti-H2A/H2B/DNA ANA) correlate best with glomerular pathology. This is also consistent with the observation that patients with drug-induced lupus develop anti-H2A/H2B/DNA ANA, but not renal disease [11,12]. These findings also have important diagnostic and prognostic implications. Since defining ANA positivity by immunofluorescent nuclear staining (e.g. nuclear staining of Hep2 cells) does not distinguish between anti-DNA and anti-H2A/H2B/DNA ANA, it is important to utilize immunoassays that discriminate between these different ANA specificities.

As shown in Table 1, 18·3% of the first-degree relatives of SLE patients harboured anti-H2A/H2B/DNA ANA, whereas 11·8% of them had anti-DNA ANA. Though these figures are consistent with prevalence rates of ANA in SLE families previously reported [40–44], these are likely to be underestimates of the actual prevalence rates for several reasons. First, since some of the relatives were as young as 12 years old, it is possible that wemight have studied them before eventual ‘seroconversion’. Second, the murine studies also point to the existence of strong ‘suppressor’ genes/alleles that can actually shut down ANA production, even in the presence of Sle1 and/or Sle3 [50]. Thus, relatives harbouring Sle1 may be serologically negative if they also happen to possess any such ‘suppressor’ alleles with protective function. This additional level of complexity is likely to render genetic dissection of human SLE even more arduous.

If indeed loci that were equivalent in function to Sle1 and Sle3 were operative in human lupus, one would expect a greater proportion of the relatives to have Profile B ANA, rather than Profile C ANA. However, it is evident from Fig. 5 that a greater fraction of the relatives exhibit Profile C ANA (11·8%), compared with those that are just seropositive for anti-H2A/H2B/DNA ANA (6·5%). These frequencies may suggest that the factors/genes controlling the genesis of the different anti-subnucleosome ANA in humans are likely to be far more heterogeneous and complex than the simple two-step, two-loci (Sle1/Sle3) model suggested by the murine studies. It is also possible that the ‘maturation’ of the initial anti-H2A/H2B/DNA seroresponse to acquire dsDNA reactivity occurs much more rapidly in humans than in the mouse models, due to unknown genetic factors.

A truly unexpected finding of these studies was the observation that the families of SLE patients with only anti-H2A/H2B/DNA ANA (Profile B) were relatively free of ANA (Fig. 5). In contrast, the incidence of ANA (i.e. family patterns B and C) in the first-degree relatives of Profile C probands was significantly higher. This is a strong indication that lupus with anti-H2A/H2B/DNA (or anti-nucleosome) ANA (accompanied by less severe renal pathology), and lupus with anti-dsDNA ANA (accompanied by a higher incidence of nephritis) may be genetically distinct entities. It also seems likely that the genesis of anti-dsDNA ANA (and renal disease) may be far more polygenic in origin, compared with the genesis of anti-nucleosome ANA. Finally, one extreme possibility is that lupus characterized by anti-nucleosome ANA (without any anti-dsDNA ANA) may not even be genetically based, possibly being quite similar to drug-induced lupus in origin. These possibilities need to be carefully evaluated with further epidemiological and genetic studies.

In conclusion, these family-based studies reveal that human lupus shows some of the immunogenic features illustrated by the Sle1/Sle3 lupus congenic mice. Given the observation that probands with anti-dsDNA ANA have a higher incidence of renal disease than those with just anti-H2A/H2B/DNA ANA (in resonance with the findings from the murine studies), it is important to distinguish between these ANA fine specificities in clinical diagnostics. In contrast to the simplicity of the B6.Sle congenic models (where the interaction of just two loci appears to determine whether an individual develops anti-chromatin ANA, or nephritogenic, anti-dsDNA ANA), the factors that dictate these fine specificities in patients appear to be somewhat different. Indeed, the development of anti-H2A/H2B/DNA (or anti-nucleosome) ANA, and anti-dsDNA ANA may very well depend on genetically distinct origins. It is thus important for genetic studies of human lupus to factor in this observation. Further studies focusing on the possibly diverse cellular and molecular origins of anti-nucleosome and anti-dsDNA ANA are certainly warranted.