Quantitative trait loci
Potentially, autoimmune diseases develop from a combination of multiple genes with allelic polymorphisms. An MRL/Mp-Faslpr/lpr (MRL/lpr) strain of mice develops autoimmune diseases, including lupus nephritis, but another lpr strain, C3H/HeJ-Faslpr/lpr (C3H/lpr) does not. This indicates that MRL polymorphic genes are involved in the development of the diseases. By quantitative trait loci (QTL) analysis using 527 of the (MRL/lpr × C3H/lpr)F2 mice, we identified a novel locus for susceptibility to lupus nephritis at map position D5Mit115 on chromosome 5, the same alias of the osteopontin (Opn) gene (LOD score =4.0), susceptible in the MRL allele. In functional analyses of the MRL and C3H Opn alleles using synthetic osteopontin (OPN) made with a new method “cell-free system” with wheat germ ribosomes, the MRL-OPN induced higher expression and production of immunoglobulins as well as cytokines including TNF-α, IL-1β and IFN-γ in splenocytes and/or macrophages than that of the C3H allele. These findings suggest that allelic polymorphism of OPN causes the functional differences in antibody production and macrophage activation between MRL and C3H strains, possibly involved in the development of lupus nephritis.
Systemic autoimmune diseases represented by SLE, are thought to be multi-factorial disorders associated with genetic and environmental factors. Genetic analyses using animal models of SLE provided evidence that multiple gene loci are responsible for the development of autoimmune traits. In human, a number of gene polymorphisms have been shown to be associated with autoimmune diseases including SLE. This indicates that the genetic factors in such diseases consist of sets of polymorphic genes, which may act in the development of disease by the differential cumulative effects of polymorphic gene products 1.
Studies using progenies of autoimmune-prone mice have attempted to identify gene loci for susceptibility to glomerulonephritis (GN), which is a major disease in SLE. A study with (MRL/Mp-Faslpr/lpr (MRL/lpr) × Cast/Ei)F1 × MRL/lpr backcross mice (N2) indicated GN susceptibility loci in Chr 7 and 12 2. Several studies with (MRL/lpr × C57BL/6-Fas lpr/lpr (B6/lpr))F2 mice, BXSB × (B10 × BXSB)F1 backcross mice (N2), (NZB × NZW)F2 mice, NZW × (NZW × C57BL/6)F1 (N2), and (BXSB × NZW)F2 mice indicated susceptibility loci in Chr 10, designated Lmb43, Chr 1 and 3 4, Chr 1 and 4 5, and Chr 6 6, respectively. It is noteworthy that the loci identified are varied and seem to depend on the parental strain of mice. Thus, the question arises as to how such a variety of loci can take part in the mechanism of the onset of GN in a strain-dependent manner. To address this question, it would be important to know the causative polymorphic gene in each of the loci of interest. However, a very limited number of genes in this concern have been proposed previously.
To identify gene polymorphisms associated with GN, we conducted genetic studies using large numbers of (MRL/lpr × C3H/lpr)F2 mice. In this report, we present the findings from the study with (MRL/lpr × C3H/lpr)F2 mice which indicate three GN susceptibility gene loci, one of which is a novel locus, involving the osteopontin gene (Opn = Spp1) not reported previously. We demonstrate functional difference of allelic osteopontin (OPN), and propose that this difference contributes to the disease-specific polymorphism located in this novel locus.
Incidence of GN in parental, F1 and F2 mice
The incidence of GN in MRL/lpr, C3H/lpr, (MRL/lpr × C3H/lpr) F1 (MC F1) and 527 (MRL/lpr × C3H/lpr)F2 intercross (MCF2) mice was determined according to the histopathological grading (see Materials and methods) (Fig. 1). The results are summarized in Table 1. GN-positive mice corresponding to grade 2 or 3 were found in 68.2% of MRL/lpr mice, and in only 5.1% of C3H/lpr mice. The incidence of GN in the F1, and F2 mice were 24.0% and 28.8%, respectively. There were no significant sex differences. The relative decrease in the incidence in the F1 or F2 mice suggested that recessive or semi-dominant GN susceptibility genes might exist in the MRL strain.
GN susceptibility loci
First, selective genome wide screening was performed using 80 MCF2 mice which consisted by randomly selected 40 individuals with grade 0 and the other 40 with grade 3 or grade 2 from a total of 527 MCF2 mice. As a result, Chr 4, 5, 13 and 17 were selected as candidate chromosomes which had markers with confirmative linkage (p<0.05) in a chi-square analysis (see Materials and methods) 7.
In quantitative trait loci (QTL) analyses on Chr 4, 5, 13 and 17, using 266 MCF2 mice, candidate loci were manifested on Chr 4 and 5, at D4Mit241 and D5Mit115, respectively, both of which showed a suggestive linkage with GN, susceptible in an MRL allele. Then, precise QTL analysis was performed on each chromosome using a total of 527 (MRL/lpr × C3H/lpr)F2 mice. The plot of LOD scores on Chr 4 (Fig. 2A) showed a significant linkage with a peak between markers D4Mit89 (19.8 cM) and D4Mit241 (24.1 cM) in a recessive susceptible mode (LOD score =8.2). A composite interval mapping (Fig. 2C) revealed another significant linkage between D4Mit187 (49.0 cM) and D4Mit147 (57.6 cM) in an additive (semi-dominant) susceptible mode (LOD score =5.5). On Chr 5 (Fig. 2B), a susceptibility gene locus was mapped around D5Mit115 (56.0 cM) in an additive (semi-dominant) susceptible mode (LOD score =4.0). This locus was also confirmed by composite interval mapping (Fig. 2D). Hereafter, the susceptibility loci on Chr 4 are designated Autoimmune GN in MRL mice (Agnm) 1 at 19.8–24.5 cM, and Agnm2 at 49.6–57.6 cM, and that on Chr 5 as Agnm3 at 54.0–65.0 cM.
Allelic differences of OPN in autoimmune strains of mice
Considering our previous results with the same MCF2 progenies of mice, Agnm1 and Agnm2 could share two susceptibility loci to systemic vasculitis, Arvm1 and Arvm2, respectively 8. On the other hand, Agnm3 is a novel locus for susceptible to GN in an MRL allele. In this region, we chose the Opn (Spp1) gene as a candidate for the reasons described in the discussion. Previously, we had reported the allelic difference of Opn between an MRL and a C3H strain of mice to a high extent 9. According to the genome-database search, D5Mit115 locates on the promoter region of OPN, 345 bp upstream of the transcription start site of Opn. Therefore, it seems that the allotype of OPN based on PCR-restriction fragment length polymorphism (RFLP) should be consistent with that of D5Mit115. RFLP analyses on Opn clearly showed that the allelic form of B6/lpr, NZB or NZW strains were the same type as that of MRL/lpr while that of the BXSB strain was the same type as that of C3H/lpr (see Materials and methods, data not shown).
Expression level of OPN mRNA in spleen
To examine the quantitative difference of the two allelic Opn, we evaluated expression level of OPN mRNA by real-time RT-PCR assay by using randomly selected MCF2 mice. As shown in Fig. 3, the representative results of quantitative RT-PCR analysis manifested that expression of OPN mRNA in the splenocytes derived from mice with MRL allele homozygote of Opn gene was significant higher than those from mice with MRL/C3H heterozygote (p<0.05) and those from mice with C3H homozygote (p<0.01).
Generation of two types of allelic OPN
To examine the functional difference of the two allelic Opn products, we generated synthetic OPN of the MRL and C3H types using a cell-free protein synthesizing system. This system allowed us to obtain large amounts of recombinant proteins without any contaminating mitogen such as LPS, which could influence greatly on immunological assays especially in our case. SDS-PAGE, autoradiogram of SDS-PAGE performed with synthetic OPN using 14C-labeled leucine and Western blot showed that both types of OPN proteins could be prepared with the same quality and quantity (see Materials and methods, data not shown).
Functional differences of allelic OPN in activation of macrophages and B cells
Previous studies demonstrated that OPN affects in vitro macrophage and polyclonal B cell activation 10. Therefore, we examined the functional differences of synthetic OPN peptides derived from an MRL/lpr and a C3H/lpr strain using M-CSF-dependent bone marrow-derived macrophages (BMMΦ) and splenocytes prepared from an MRL/Mp-Fas+/+ (MRL/+) and a C3H/HeJ-Fas +/+ (C3H/+) strain. Activation profiles were obtained by determining the mRNA expression of TNF-α, IL-1β and IFN-γ mRNA using a real time RT-PCR assay, and by measuring the secretion of TNF-α and Ig using ELISA (see Materials and methods). Macrophage activation was observed in response to both types of synthetic OPN in a dose-dependent manner irrespective of the responder origin (data not shown), but MRL-OPN induced significantly higher expression of TNF-α and IL-1β in BMMΦ (Fig. 4A–D), and IFN-γ in splenocytes than C3H-OPN (Fig. 4E). TNF-α secretion was parallel to mRNA expression (Fig. 4F). Next, we examined the differences in induction of Ig secretion from splenocytes by the two allelic OPN. MRL-OPN induced a significantly higher amount of IgG3 which was previously found to be nephritogenic in MRL/lpr mice 11, as well as IgG2a and IgM secretion (Fig. 5). In particular, IgM production by MRL/+ splenocytes was increased more than tenfold at 72 h of culture (Fig. 5E).
Several genetic studies with autoimmune-prone mice have shown that a number of loci are associated with autoimmune diseases, but there has been little evidence for the involvement of gene polymorphism in the mechanism of autoimmune disease.
In the initial step of this study, three loci were found to be associated significantly with GN, and one of them, Agnm3, turned out to be a novel locus that did not share a chromosomal region with any lupus susceptibility loci reported previously 2–6, 8, although some phenotypic loci referring to immune traits were reported such as Lprm4 (lymphoproliferation modifier 4, 54 cM) that was designated by spleen weight in MCN2 mice 12 and Tsz1 (Thymus size 1, 56 cM) that was designated by thymus size 13. It should be noted that another study using progenies from B6/lpr and MRL/lpr mice did not show any linkage of GN to Agnm3, even though we employed the same conditions for mouse breeding and histopathologic GN grading as in this study (Zhang et al., manuscript in preparation). Disease-specific loci defined by a genetic linkage study seem to depend largely on the combination of parental strains, but the involvement of unknown environmental effects cannot be ruled out. Thus, it seems reasonable to postulate that disease-specific loci do not generally represent a diseased strain-specific polymorphism, but just a polymorphism that expresses a functional difference affecting on the autoimmune pathway. In this regard, a functional polymorphism in an immunological context and a parental combination-restricted polymorphism had been assumed for the gene corresponding to Agnm3.
In this region, there are a number of candidate genes which might be involved in immune responses or inflammation processes, including chemokine (C-X-C) ligand 5, 9 and 10, bone morphogenetic protein 3, fibroblast growth factor 5, SPARC-like 1, down-regulator of transcription 1, heat shock 27kDa protein 8, platelet selectin ligand, and OPN etc.
Among them, as far as we know, only OPN had been manifested polymorphic between MRL/Mp and C3H/HeJ. Thus, we selected OPN as a candidate gene for its polymorphisms and its relationship to immune responses. Prior to this study, the Opn product, termed the early T lymphocyte activation 1 (Eta-1) or secreted phosphoprotein 1 (Spp1), had been known for its cytokine activities affecting migration of macrophages to an inflammatory site 14 and polyclonal B cell activation and differentiation 10, 15. With respect to autoimmune disease, Patarca et al. 16 reported dysregulated expression of Eta-1 in a T lymphocyte subset (CD4– CD8–) during the development of autoimmune disease in MRL/lpr mice. We reported an extensive coding polymorphism of Opn between MRL/lpr and C3H/lpr mice 9. In the present study, the highest linkage revealed by LOD score was found around D5Mit115, and the association analysis with the chi-square test also demonstrated the most probable association at this marker position (p<0.00072). A genome database search showed that the 3′ end of D5Mit115 is located 345 bp upstream of the transcription start site of the Opn gene. Previous and present data are consistent with our working hypothesis that OPN polymorphism is involved in the development of GN in MRL/lpr mice, and its suppression in C3H/lpr mice.
Opn genotyping in several strains of mice revealed that the Opn genotype of NZB, NZW, B10 and C57BL/6 correspond to the MRL-type (allele a), whereas that of BXSB correspond to the C3H type (allele b). Previous linkage analyses using autoimmune (NZB, MRL and BXSB) and non-autoimmune strains (NZW, C57BL and B10) did not test the combination of OPN allotypes as in the present study (MRL vs. C3H) 2, 3, 5 except of a few studies 4, 6, 17. In the study of BXSB × B10 crosses 4, 17, to detect the loci with additive mode of inheritance in disease susceptibility might be difficult since authors used BXSB × (B10 × BXSB)F1 or B10 × (BXSB × B10)F1 backcrosses, not F2 crosses. Furthermore, in these studies, BXSB strain was used as a lupus-prone strain, while BXSB type OPN might be disease resistant. For these reasons, OPN locus might not be detected. Kono et al. 6 used (BXSB × NZW)F2 with a different MHC haplotype combination (H-2b and H-2z, respectively), which might affect the difference of set of susceptibility loci possibly influenced by MHC 18. These may be the reason why previous studies did not conclude that a locus for autoimmune diseases including glomerulonephritis and vasculitis shares a chromosomal region with Agnm3.
Interestingly, the same locus as Agnm3 have been identified in linkage studies on the susceptibility for experimental Lyme arthritis 19 and experimental encephalomyelitis 20. This chromosomal region contains polymorphic OPN gene. Teuscher's group had worked with the allele b OPNvs. allele a OPN as a candidate for arthritis. However, they finally excluded this possibility based on the experiment using OPN–/– mice which showed more severe Lyme arthritis than OPN+/+ mice 21. Nevertheless, the conclusion of Teuscher's group is not to deny the possibility that the allelic polymorphism of OPN is a candidate of susceptibility for other autoimmune diseases since different immune and inflammatory processes are involved in autoimmune phenotypes associated with host background genes. However, there are still controversies around OPN deficiency and onset of autoimmune diseases including experimental encephalomyelitis (EAE) [22–24]. A certain number of reports have recently confirmed the view that OPN deficiency has no influence on arthritis and EAE 21 and suggested possible implication of polymorphic gene(s) closely linked to OPN. On the other hand, some papers suggested the involvement of OPN in the pathogenesis of multiple sclerosis; that is sustaining autoimmune responses by assisting in maintenance of Th1 immunity during EAE 22, 25, manifesting the relationship between polymorphism of OPN and disease course of multiple sclerosis in human 26, and showing increased plasma OPN levels in the chronic progressive phase of multiple sclerosis 27. Furthermore, Steinman et al. 24 demonstrated that immunization with a DNA encoding OPN alters the chronic course of EAE. Thus these issues certainly do not exclude a role of OPN in autoimmune diseases. Further studies using MRL/lpr-OpnC3H/C3H or MRL/lpr-Opn–/– mice should be performed.
In autoimmune disease-prone MRL/lpr mice, the level of OPN in serum was found to be increased in association with the onset of autoimmune diseases and expansion of CD3+/CD4–/CD8– T-cells which express OPN mRNA 16. Several studies have demonstrated the link between OPN overexpression and autoimmune diseases including lupus nephritis 28, 29, interstitial nephritis 30, rheumatoid arthritis 31 and pulmonary fibrosis 32. These findings suggest that there is a quantitative contribution of OPN activity to autoimmune diseases. We also presented here the expression data of OPN mRNA in MCF2 mice, indicating that MRL allele induced higher expression than that of C3H allele. This was consistent with the issues. Nevertheless, there has been no report about polymorphism in the promoter region of Opn as far as we know. Thus, we analyzed the sequence of the Opn promoter region in MRL/lpr and C3H/lpr. Except for a 12 bp insertion in C3H/lpr which designates the SSLP of D5Mit115, we found at least a SNP (deletion in C3H/lpr) at 880 bp upstream of the translation start site which resulted in deletion of the alpha IFN.2 element (unpublished data). Of course, the matters of expression differences and potential promoter polymorphisms of OPN are much more complicated. Giacopelli, et al. 33 suggested that polymorphisms in the osteopontin promoter affect its transcriptional activity in human, and another group reported that the expression of OPN might be controlled by not only its promoter polymorphism but also by trans-activators 34. Moreover, a report suggested that allelic polymorphism of Opn qualitatively affects on β-glucan-induced granuloma formation, and allele b OPN defected in induction of cell migration 35. Here, we like further to suggest that there might be a qualitative contribution of OPN to autoimmune diseases. Thus, the polymorphism of Opn affecting the protein's function may be of interest in terms of its relation to human autoimmune diseases.
Allelic differences of OPN between MRL and C3H strains are found in 10 amino acid residues (9 substitutions and 1 insertion). The insertion of a lysine is at the 14th residue of the mature peptide, and the substitutions are Asn126 to Asp, Asp155 to Tyr, Asp172 to Asn, Arg208 to Ser, Gln216 to His and Tyr261 to His. It should be noted that these substitutions may affect some of the functional sites of OPN. Indeed, the Tyr261 and the Asp172 correspond to a heparin binding site and a phosphorylation site overlapped by the homologous Itgα4β1 binding site of fibronectin, respectively 15. Asn126 neighbors the RGDS motif that serves as a binding site for Itgαvβ136, αvβ337, and αvβ538. The critical substitutions for the functional differences of allelic OPNs need to be identified.
Opn was reported to be a candidate gene linked to the Ric locus, which determines the susceptibility to Rickettsia Tsutsugamushi (RT) infection in mice 39. It was concluded that OPN is an effector molecule responsible for the protection against RT infection. The present results indicate that RT resistance is associated genetically with an autoimmune disease, whereas RT susceptibility is not. It is interesting to speculate that acquisition of immunity against infection resulted in the acquisition of susceptibility to autoimmune diseases.
We conclude that the polymorphism of OPN may play important roles in the development of GN in MRL/lpr mice via activation of macrophages and accelerating polyclonal B-cell activation. Further studies are needed to elucidate the relevant polymorphism site in OPN in addition to in vivo analyses using recombinant congenic mice.
Materials and methods
MRL/lpr, C3H/lpr, MRL/+ and C3H/+ mice were originally purchased from The Jackson Laboratory (Bar Harbor, ME). These mice were maintained in the Institute of Experimental Animals, Tohoku University School of Medicine, Sendai, Japan. Eighty-eight MRL/lpr, 78 C3H/lpr, 50 MCF1 and 527 MCF2 intercross mice were prepared in the same institute and partly in the laboratory animal center of Kawashima Co. Ltd. (Gifu, Japan). All mice were treated according to the guideline for laboratory animal experiments.
Histopathological grading of glomerular lesions
Mice were dissected under ether anesthesia at 4.5 to 5 months of age. Kidney specimens were fixed with 10% buffered formalin and embedded in paraffin. Serial 2-μm sections were stained with hematoxylin-eosin (HE) or periodic acid Schiff (PAS) for histological examination by light microscopy.
Glomerular lesions were graded according to the following criteria: grade 0, no recognizable lesion in glomeruli; grade 1, mild cell proliferation and/or cell infiltration; grade 2, the same as grade 1 with mesangial proliferation, lobulation and hyaline droplet, associated with macrophage infiltration; grade 3, the same as grade 2 with crescent and granuloma formation and/or hyalinosis. An individual positive for GN was defined as one with more than 50% grade 2 and /or 3 glomeruli of more than 20 renal glomeruli examined. These criteria indicate the progressive stage of GN in MRL/lpr mice, resembling lupus nephritis (Fig. 1).
Genome wide screening
At first, selective genome wide screening using 80 MCF2 mice which consisted by randomly selected 40 individuals with grade 0 and the other 40 with grade 3 or grade 2 was performed by a simple sequence length polymorphism (SSLP) assay using a tail DNA sample of each mouse as genomic DNA. Microsatellite markers purchased from Research Genetics Inc were used to determine the genotypes of each locus by PCR. The map positions of the microsatellite loci were based on the Mouse Genome Database (MGD) 13. At first, more than 400 microsatellite loci were checked for their polymorphisms between alleles of MRL/lpr and C3H/lpr strains. Finally, 106 microsatellite loci were selected to provide the full coverage of the mouse autosomes. They were an average of 12 cM apart, with a maximum distance between any two markers of 37 cM.
Association studies between the incidence of GN (positive or negative) and the genotypes of microsatellite loci (MRL/MRL homozygote, C3H/C3H homozygote or MRL/C3H heterozygote) were carried out with a chi-square analysis using a standard [2×3] contingency matrix. As recommended by Lander and Kruglyak 7, candidate chromosomes with loci showing a confirmative linkage (p<0.05) were pursued with linkage studies.
The linkage map for the MCF2 mice was created with Mapmaker3 40. Quantitative trait loci (QTL) were analyzed with WinQTL Cartographer ver 2.0 41 based on the histopathological grades of glomerular lesions as the indicators of phenotype. LOD scores were calculated using the LRmapqtl program. Then, the ZMapqtl module of this software implemented composite interval mapping, under model 6, with options set at 2-cM intervals, 10 cM window size considering background loci that include unlinked loci and linked loci and five background parameters 6. Permutation tests with 10,000 permutations were also performed to assess the probability of the data for each locus comparisonwise (α<0.01) 42.
RFLP analysis of OPN
To distinguish the Opn genotype, RFLP analyses on PCR amplicons were carried out by the following procedure. PCRs on mouse genomic DNA were performed with the sense primer OPNL 5′-TTGATTGATAGAAGCTGACT-3′ which is specific for bases 516–535 in intron 5 of Opn and the antisense primer OPNR 5′-GTGTTTCCAGACTTGGTTCA-3′ which is specific for bases 152–171 in exon 6. This results in a 344 bp PCR amplicon. There is an Nde I cleavage site exists in the PCR amplicon from the C3H allele, but not in that from the MRL allele 9. After the digestion by Nde I, the products were electrophoresed on 3% agarose gel, and the genotype was determined by the bands of the PCR amplicon. A single 334 bp band and two bands of 187 and 157 bp indicated the MRL allele and C3H allele, respectively. To determine the allelic distribution of the Opn, genomic DNA of B6/lpr, BXSB, NZB and NZW strains were analyzed above.
Evaluation of OPN mRNA in splenocytes
To evaluate the expression of OPN mRNA in splenocytes of MCF2 mice, quantitative RT-PCR of OPN was carried out employing the TaqMan® system (Applied Biosystems, Foster City, CA). Quantitative PCR was performed with an ABI PRISM® 7700 sequence detection system (Applied Biosystems) by SYBR green assay, employing primers mSPP1-TAF 5′-ACTTTCACTCCAATCGTCCCTA-3′ specific for bases 175–195 in exon 5 and mSPP1-TAR 5′-TGTGGCATCAGGATACTGTTCAT-3′ specific for bases 278 in exon 5 to bases 1–15 in exon 6. The quantity was normalized to 18s-ribosomal RNA using the formula of the 2-ΔΔCt method 43. Evaluation was performed in triplicate. Means and standard deviations of relative expression of OPN mRNA were calculated, then statistically analyzed among groups of OPN genotype (MRL homozygote, heterozygote and C3H homozygote).
Synthesis of polymorphic OPN
OPN derived from the MRL and C3H alleles were prepared with a cell-free protein synthesis system using wheat germ ribosomal RNA, with which there is efficient protein synthesis without risk of LPS contamination 44. In brief, Opn cDNA derived from an MRL/lpr or C3H/lpr mouse was inserted into pGEX-6P-1® expression vector (Amersham Bioscience, Piscataway, NJ) containing a glutathione S-transferase (GST) tag region. Capped mRNA encoding each allele of OPN were prepared by in vitro transcription of linearized plasmid pGEX-6P-1-MRL-OPN and pGEX-6P-1-C3H-OPN, under SP6 RNA polymerase promoter control. The cell-free translation system in a dialysis tube contained purified wheat germ extract, Hepes/KOH (pH 7.8), ATP, GTP, creatine phosphate, creatine kinase, dithiothreitol, spermidine, 20 amino acids, magnesium acetate, potassium acetate, deacylated tRNA prepared from wheat embryos, NP-40, E-64, NaN3, and OPN mRNA. Incubation was done at 26°C for 24 h in feeding buffer containing elements above except for wheat germ extract and OPN mRNA.
Synthetic GST-fusion OPN was purified with a GST purification module® (Amersham Bioscience). It was detached from column-conjugated GST using PreScission® protease (Amersham Bioscience). The purity of the OPN fraction was confirmed by SDS-PAGE, autoradiography of SDS-PAGE with 14C-labeled synthetic OPN using a BAS-2000 Phosphoimager (Fuji, Tokyo, Japan) 44 and by Western blotting analysis using rabbit anti-murine OPN antibody (Cosmo Bio, Tokyo, Japan). OPN concentration was measured with an ELISA kit (Mouse osteopontin detection kit®, Immuno-Biological Laboratories, Fujioka, Japan).
Bioassay of synthetic OPN in vitro
To examine the functional differences between synthetic polymorphic OPN proteins, a BMMΦ stimulation assay and an induction of immunoglobulin production assay on whole splenic cells were performed as follows. Responder cells used in each assay were from MRL/+ mice and C3H/+ mice to avoid the effect of high intrinsic production of OPN by lpr immunocytes, possibly due to dysregulated expression in CD4–/CD8– lymphocytes 15.
Macrophage stimulation assay
Femoral bone marrow cells from 8–12-week-old mice were harvested and suspended at 3×105 cells/ml in DMEM containing 10% FCS, 2 mM glutamine, 0.37% (w/v) NaHCO3, 100 U/ml penicillin-G, 100 μg/ml streptomycin and 1 ng/ml M-CSF (Upstate Biotechnology, Charlottesville, VA). The cells were cultured for 6 days at 37°C under 5% CO2. Attached cells were harvested and suspended in DMEM. The cells at 2×106 cells/ml were dispensed into 24-well culture plates (1 ml/well). Synthetic MRL- or C3H-OPN was added to each well as indicated except for the negative control wells. Experiments were carried out in triplicate per condition.
After 24, 30, and 36 h cells were harvested for quantitative RT-PCR of TNF-α and IL-1β as indicators of macrophage activation, employing the TaqMan® system as described above. Primers and probes used for TNF-α mRNA and IL-1β mRNA were designed according to Overbergh et al. 45. The quantification was carried out in triplicate on each sample.
ELISA of TNF-α employing OptEIA® murine TNF-α Set (BD PharMingen) was performed with the culture supernatant from each well in triplicate.
Assay of induction of Ig production
Splenocytes from 8–12 week-old mice were cultured at 4×106 cells/ml in 24-well plates in RPMI1640 containing 10% heat -inactivated FCS and 10 μM indomethacin (Sigma-Aldrich). Synthetic MRL- or C3H-OPN was added at a concentration of 100 ng/ml. Negative controls lacked synthetic peptide. Experiments were carried out in triplicate per condition.
Culture supernatants were used for ELISA of IgG3, IgG2a and IgM after 24, 36 and 72 h of culture. Rabbit IgG anti-mouse Ig involving IgG, IgA and IgM (both heavy and light chain) (Zymed Laboratories Inc., South San Francisco, CA) were used for capturing Ig in the culture supernatants. Alkaline phosphatase-labeled rabbit anti-mouse IgG2a, IgG3 or IgM antibody (Zymed) was used as the secondary antibody for visualizing with Sigma 104® (Sigma-Aldrich). For quantitation, a standard curve was made with dilutions of mouse Ig reference serum (ICN). Evaluation was carried out in triplicate.
In addition, the cells harvested after culture for 24 and 36 h were used for quantitative real time RT-PCR of TNF-α, IL-1β and IFN-γ as in the macrophage stimulation assay. The primers for IFN-γ have been described 45.
Differences in the incidence of GN among progenies and between sexes were analyzed by the chi-square test, using standard [2×2] contingency matrices. Statistical analysis of the expression level of OPN mRNA was analyzed with Mann-Whitney's U-test, and those of cytokines and Ig were analyzed with the Student t-test. p<0.05 was considered significant.
We are grateful to Ms. Miho Terada, Dr. Tomoko Hyodo, Dr. Hiroto Mitsui, Mr. Masashi Maune and Dr. Akihiro Yamada for experimental assistance, and to Dr. Sachiko Hirose for the use of their genomic DNA samples of autoimmune strains of mice. We also thank Dr. Herb Schulman for reviewing the manuscript. This study was supported by grants from the Research Funds of the Ministry of Health and Welfare of Japan, and by Grant-in Aid for Scientific Research of the Ministry of Education, Science, Sports and Culture of Japan (#11670217, #13557018 and #15590346).