A phenotype survey of 36 mutant mouse strains with gene-targeted defects in glycosyltransferases or glycan-binding proteins.

The consortium for functional glycomics (CFG) was a large research initiative providing networking and resources for investigators studying the role of glycans and glycan-binding proteins in health and disease. Starting in 2001, six scientific cores were established to generate data, materials and new technologies. By the end of funding in 2011, the mouse phenotype core (MPC) submitted data to a website from the phenotype screen of 36 mutant mouse strains deficient in a gene for either a glycan-binding protein (GBP) or glycosyltransferase (GT). Each mutant strain was allotted three months for analysis and screened by standard phenotype assays used in the fields of immunology, histology, hematology, coagulation, serum chemistry, metabolism and behavior. Twenty of the deficient mouse strains had been studied in other laboratories, and additional tests were performed on these strains to confirm previous observations and discover new data. The CFG constructed 16 new homozygous mutant mouse strains and completed the initial phenotype screen of the majority of these new mutant strains. In total, >300 phenotype changes were observed, but considering the over 100 assays performed on each strain, most of the phenotypes were unchanged. Phenotype differences include abnormal testis morphology in GlcNAcT9- and Siglec-H-deficient mice and lethality in Pomgnt1-deficient mice. The numerous altered phenotypes discovered, along with the consideration of the significant findings of normality, will provide a platform for future characterization to understand the important roles of glycans and GBPs in the mechanisms of health and disease.

Motheaten and viable motheaten mice have mutations in the haematopoietic cell phosphatase gene.

Mice with the recessive motheaten (me) or the allelic viable motheaten (mev) mutations express a severe autoimmune and immunodeficiency syndrome. We have shown that the basic defect in these mice involves lesions in the gene which encodes haematopoietic cell phosphatase (HCP). These mice thus provide excellent models for investigating the roles of phosphatases in haematopoiesis and the nature of the genetic and cellular events linking impaired haematopoiesis to severe immunodeficiency and expression of systemic autoimmunity.

[Mice are not Men and yet how humanized mice inform us about human infectious diseases]. / Les souris ne sont pas des hommes et pourtant ce que les souris humanisées nous apprennent sur les maladies infectieuses.

The study of human pathologies is often limited by the absence of animal models which are robust, cost-effective and reproduce the hallmarks of human infections. While mice have been frequently employed to study human diseases, many of important pathogens display unique human tropism. These last two decades the graft of human progenitor cells or tissues into -immunodeficient mice has allowed the elaboration of so called humanized mice. Humanized mouse technology has made rapid progress, and it is now possible to achieve high levels of human chimerism in various organs and tissues, particularly the immune system and the liver. The review briefly summarizes the different models of humanized mice available for in vivo experiments. With a focus on lymphotropic, monocytotropic and hepatotropic viruses, we here discuss the current status and future prospects of these models for studying the pathogenesis of infectious diseases. Furthermore, they provide a powerful tool for the development of innovative therapies.

The trophoblast is a component of the innate immune system during pregnancy.

Systemic infection with Listeria monocytogenes, a Gram-positive intracellular bacterium, has been used extensively to analyze the innate immune response. Macrophages are central to this response, acting as both the host for and principal defense against this bacterium. During pregnancy L. monocytogenes has a predilection for replication at the maternal-placental interface and consequently is an important cause of fetal morbidity and mortality. However, macrophages are mostly excluded from the murine placenta with neutrophils acting as the main immune effector cell against this bacterium. Colony stimulating factor (CSF)-1, a macrophage growth factor, is synthesized in high concentrations by the uterine epithelium during pregnancy, where it is targeted to trophoblast bearing CSF-1-receptors. To define the involvement of CSF-1 in placental immunity, we infected pregnant mice either homozygous or heterozygous for an inactivating recessive mutation in the gene for CSF-1 (osteopetrotic; Csfmop) with L. monocytogenes. CSF-1 was required to recruit neutrophils to the site of listerial infection in the decidua basalis, and infection by Listeria remained unrestrained in its absence. CSF-1 acted by inducing the trophoblast to synthesize the neutrophil chemoattractants (KC) and macrophage inflammatory protein (MIP)-2. Thus, during pregnancy, trophoblast responsive to CSF-1 acts to organize the maternal immune response to bacterial infection at the utero-placental interface. This previously unknown function indicates that the trophoblast acts as a pregnancy-specific component of the innate immune system.

McH-lpr/lpr-RA1 mice: A novel spontaneous mouse model of autoimmune sialadenitis.

Many studies of the autoimmune disease Sjögren's syndrome have been performed using spontaneous mouse models. In the present study, we describe the characteristics of McH/lpr-RA1 mice and propose their use as a novel murine model of autoimmune sialadenitis. The McH/lpr-RA1 mouse is a recombinant congenic strain derived from generation F54 or more of MRL-Faslpr x (MRL- Faslpr x C3H- Faslpr) F1. We show for the first time that this mouse spontaneously develops autoimmune sialadenitis and vasculitis in submandibular gland tissues. Sialadenitis was accompanied by extensive inflammatory cell infiltration and tissue destruction. Immunohistochemical studies revealed that the salivary gland lesions strongly expressed four sialadenitis-related molecules: SSA and SSB (autoantigens of Sjögren's syndrome), gp91phox (an accelerator of reactive oxygen species production) and single strand DNA (a marker of apoptotic cells). In contrast, expression of aquaporin-5 (AQP5), which stimulates salivary secretion was weak or negligible. Statistical correlation analyses indicated that the apoptosis of salivary gland cells provoked by oxidative stress contributed to the severe sialadenitis and reduced expression of AQP5. Our study has demonstrated that McH/lpr-RA1 mice spontaneously develop the pathognomonic features of autoimmune sialadenitis and thus could be used as a new animal model of Sjögren's syndrome.

The lpr gene causes an intrinsic T cell abnormality that is required for hyperproliferation.

The lpr gene induces marked lymphoproliferation characterized by the massive accumulation of T cells of an unusual phenotype and concomitant autoimmune disease. To clarify the mechanism of the lpr effect, bone marrow cells from B6-lpr/lpr (Ly-1.2) and B6-+/+ (Ly-1.1) mice were transferred into lethally irradiated B6-lpr/lpr mice. As has been previously reported, recipients of the B6-lpr/lpr bone marrow showed the typical lpr phenotype with marked lymphadenopathy, splenomegaly and increased levels of autoantibodies; while the recipients of B6-+/+ bone marrow had normal sized lymph nodes and spleen and no autoantibodies. A third group of mice received an equal mixture of bone marrow cells from the B6-lpr/lpr and B6-+/+ donors. These mice showed both lymphadenopathy and autoantibody production comparable to that of recipients of the B6-lpr/lpr marrow alone. Immunofluorocytometric analysis of the lymphoid populations in these mixed bone marrow recipients established that the T cells from the lpr/lpr and +/+ donors were equivalently represented in the peripheral blood and thymus. In striking contrast, the T cells that accumulated in abnormally large numbers in the lymph nodes were almost entirely from the lpr donor. Their surface phenotype was Thy-1+(dull), Ly-1.2+(dull), Lyt-2-, L3T4-, 9F3+, and 3A1+, which is consistent with that found in intact lpr mice. These results indicate that the lpr gene causes an intrinsic defect directly within the T cells that accumulate in large numbers in lpr mice. In addition, the presence of the +/+ T cells cannot prevent the expression of the lpr abnormalities.

Specificities and V genes encoding monoclonal autoantibodies from viable motheaten mice.

Several hundred hybridomas were obtained from 1-2-mo-old viable motheaten (mev) mice. Among the Ig-secreting hybridomas tested, greater than 50% (17/33) exhibited reactivity for autoantigens, supporting the idea that the Ly-1 B cells that predominate in mev mice contain frequent precursors of autoantibody-forming cells. Certain of the specificities of these autoantibodies correlated with the documented pathophysiology of mev mice (antithymocyte, -erythrocyte, -skin, -kidney, and -IgG); others were specific for autoantigens not previously observed in motheaten mice but though to be involved in other autoimmune diseases (e.g., intrinsic factor, transferrin, myelin basic protein, and thyroglobulin). About 2 of 3 (11/17) of the self-reactive antibodies exhibited multispecific binding activity for various autoantigens. Analysis by Northern blotting of the V gene families used in mev autoantibodies showed a random usage of VH families and a biased usage of four Vk gene families. Of 16 autoantibodies tested, 12 used a Vk gene from the Vk1, 4, 10, or 19 families. These patterns of Vk gene usage differ from nonautoimmune control animals. Overall, an immunoregulatory defect operating at a more generalized level than the VH or Vk loci, and due to a single gene mutation, appears to be responsible for the multiple immune abnormalities of mev mice.

Rejection of bone marrow allografts by mice with severe combined immune deficiency (SCID). Evidence that natural killer cells can mediate the specificity of marrow graft rejection.

C.B-17 scid (H-2d) mice are homozygous for the gene that causes severe combined immune deficiency (SCID). These mice have no T or B cell function, yet display normal natural killer (NK) activity. Irradiated SCID mice were challenged with marrow grafts to determine if antibodies are necessary for marrow allograft rejection. SCID mice rejected H-2/Hh-1 allogeneic marrow grafts. Moreover, this rejection capability could be adoptively transferred using SCID marrow as a source of NK progenitors infused into irradiated B6 (H-2b) hosts. We conclude that NK cells can mediate marrow allograft reactivity in the absence of immunoglobulin. It follows that NK cells probably have specific receptors for Hh antigens.

Antiglomerular basement membrane nephritis in beige mice. Deficiency of leukocytic neutral proteinases prevents the induction of albuminuria in the heterologous phase.

Antiglomerular basement membrane (GBM) nephritis with massive albuminuria can be induced in mice by injection of heterologous antibodies against mouse GBM. The albuminuria and the glomerular lesions in this model are not mediated by complement, but are dependent on the presence of polymorphonuclear granulocytes (PMN) in the glomeruli. Neutral serine proteinases and reactive oxygen metabolites produced by activated PMN have been implicated as agents contributing to tissue damage. We examined the role of leukocytic neutral proteinases by comparing the glomerular damage and albuminuria after injection of rabbit anti-mouse GBM antibodies in normal control mice (C57BL/6J, +/+) and in beige mice (C57BL/6J,bg/bg) in which PMN are deficient of the neutral proteinases elastase and cathepsin G. The dose-dependent albuminuria that occurred in control mice after injection of 1.4-22 mg of anti-GBM antibodies was not observed in beige mice, despite a comparable influx of PMNs in the glomeruli. By electron microscopy both strains showed a similar attachment of PMN to the denuded GBM together with swelling and necrosis of endothelial cells. Elastase activity of extracts from PMN of beige mice was only 10-15% of the activity of control mice. In vitro, GBM degradation by PMN extracts of beige mice was 70% lower than that seen in control experiments. PMNs of beige and control mice showed no differences in superoxide production. In addition, administration of scavengers of reactive oxygen metabolites, such as catalase and desferrioxamine, did not prevent the albuminuria in this model. These findings support the important contribution of leukocytic neutral proteinases to the induction of albuminuria in the acute phase of anti-GBM nephritis in the mouse.

Antitopoisomerase I monoclonal autoantibodies from scleroderma patients and tight skin mouse interact with similar epitopes.

We have generated for the first time monoclonal antibodies (mAbs) specific for topoisomerase I (topo I) from scleroderma patients, and tight skin mice which develop a scleroderma-like syndrome. The epitope specificity of these antibodies has been determined using a series of fusion proteins containing contiguous portions of topo I polypeptide. Western blot analysis demonstrated that both human and mouse mAbs bound strongly to fusion protein C encompassing the NH2-terminal portion of the enzyme, and weakly to fusion proteins F and G containing regions close to the COOH-terminal end of the molecule. This crossreactivity is related to a tripeptide sequence homology in F, G, and C fusion proteins. It is interesting that a pentapeptide sequence homologous to that in fusion protein C was identified in the UL70 protein of cytomegalovirus, suggesting that activation of autoreactive B cell clones by molecular mimicry is possible. Both human and mouse mAbs exhibiting the same antigen specificity, also share an interspecies cross-reactive idiotope. These data suggest that B cell clones producing antitopo autoantibodies present in human and mouse repertoire are conserved during phylogeny, and are activated during the development of scleroderma disease.

Fcgamma receptor IIB-deficient mice develop Goodpasture's syndrome upon immunization with type IV collagen: a novel murine model for autoimmune glomerular basement membrane disease.

The combination of hemorrhagic pneumonitis and rapidly progressive glomerulonephritis is a characteristic feature of Goodpasture's syndrome (GPS), an autoimmune disease resulting from the interaction of pathogenic anti-collagen type IV (C-IV) antibodies with alveolar and glomerular basement membranes. Lack of a suitable animal model for this fatal disease has hampered both a basic understanding of its etiology and the development of therapeutic strategies. We now report a novel model for GPS using mice deficient in a central regulatory receptor for immunoglobulin (Ig)G antibody expression and function, the type IIB Fc receptor for IgG (FcgammaRIIB). Mutant mice immunized with bovine C-IV reproducibly develop massive pulmonary hemorrhage with neutrophil and macrophage infiltration and crescentic glomerulonephritis. The distinctive linear, ribbon-like deposition of IgG immune complex seen in GPS was observed along the glomerular and tubulointerstitial membranes of diseased animals. These results highlight the role of FcgammaRIIB in maintaining tolerance and suggest that it may play a role in the pathogenesis of human GPS.

The severe combined immunodeficiency (scid) mouse. A laboratory model for the analysis of Lyme arthritis and carditis.

We report that the spirochete B. burgdorferi induces progressive polyarthritis and carditis in mice with severe combined immunodeficiency syndrome (scid) but not in normal C.B-17 mice. The onset and severity of the disease were dependent on (a) the viability; (b) the infectivity; and (c) the dose of inoculated B. burgdorferi organisms. Infective spirochetes were isolated from both blood and joints of inoculated scid mice. These findings suggest that B. burgdorferi-induced chronic arthritis and carditis in mice develops independently of lymphocyte function and makes the scid mouse an attractive laboratory model to study the role of the immune system in experimental Lyme Borreliosis.

Double-negative T cells from MRL-lpr/lpr mice mediate cytolytic activity when triggered through adhesion molecules and constitutively express perforin gene.

The lpr gene induces in mice, accumulation of large numbers of CD4-CD8- (double negative [DN]) T lymphocytes which bear adhesion molecules not characteristic of normal resting T cells. These cells fail to acquire interleukin 2 (IL-2) receptors, produce IL-2, and proliferate when activated with mitogens or monoclonal antibodies (mAbs) against the T cell receptor (TCR). Because of these poor functions in vitro, the nature and significance of DN T cells in the autoimmune disease process is not clear. In the current study, we describe a surprising finding that mAbs against CD3-TCR-alpha/beta complex can strongly trigger the lytic activity of the DN T cells to induce redirected lysis of Fc receptor-positive targets. Similar redirected lysis was also inducible using mAbs against CD44 and gp90MEL-14, molecules involved in the binding of lymphocytes to endothelial cells. The spontaneous cytotoxic potential of the DN T cells was further corroborated by demonstrating that the lpr DN T cells constitutively transcribed perforin gene but failed to express granzyme A. The current study suggests that DN T cells are capable of mediating lysis of autologous cells bearing the specific ligands for adhesion molecules involved in the signaling of cytotoxicity. These findings provide a novel insight into the functional significance of DN T cells in lpr mice and their potential role in the pathogenesis of autoimmune disease.

Evidence of functional lymphocytes in some (leaky) scid mice.

Although the majority of severe combined immune deficiency (scid) mice lack functional lymphocytes, some (2-23%) appear to develop a limited number of B and T cells between 3 and 9 mo old. Most of these leaky scid mice were shown to contain very few clones (less than or equal to 3) of Ig-producing plasmacytes. Clonal progeny were distributed unevenly in the lymphatic tissues and appeared as discrete plasmacytic foci. In many cases, individual clones persisted for several months and produced abnormally high concentrations of Ig that included multiple isotypes. Functional T cells were inferred from the ability of leaky mice to reject allogeneic skin grafts, a T cell-dependent reaction. Interestingly, approximately 40% of leaky mice developed thymic lymphomas. In other respects, leaky mice resembled regular scid mice; e.g., their splenic cells failed to express common lymphocyte antigens (Ly-5[B220], Ly-1) and to proliferate in response to lymphocyte mitogens. Histologically, their lymphoid tissues retained the same general pattern of severe lymphocytic deficiency as scid mice.

Contrasuppression in autoimmunity. Abnormal contrasuppression facilitates expression of nephritogenic effector T cells and interstitial nephritis in kdkd mice.

We have used the murine model of spontaneous autoimmune interstitial nephritis in kdkd mice to examine the importance of abnormal immunoregulation in the expression of disease. T cells from naive congenic CBA/Ca mice suppress both histologic renal injury in the kdkd strain as well as the DTH reactivity to CBA/Ca renal tubular antigens mediated by lymphocytes from nephritic kdkd mice. These antigen-specific suppressor T cells are Lyt-2+, L3T4+, I-Jk+, genetically dominant and I-Jk restricted. Unfractionated spleen cells from young, prenephritic kdkd mice also demonstrate such suppressor function. Shortly preceding disease onset, however, net suppression is functionally bypassed by emergent contrasuppressor T cells. These regulatory cells are also Lyt-2+ and I-Jk+, and adhere both to the Vicia Villosa lectin and CBA/Ca TBM. By admixing these contrasuppressor cells with spleen cells from non-disease-prone CBA/Ca mice we were able to demonstrate the presence of DTH-reactive and nephritogenic effector cells in the latter population. Such nephritogenic effector cells could also be simply demonstrated after depletion of the suppressor cells with anti-I-Jk mAbs and complement. These findings support a role for contrasuppressor cells in the abrogation of tolerance to parenchymal self-antigens.

Possible role of V beta T cell receptor genes in susceptibility to collagen-induced arthritis in mice.

Arthritis was induced by immunization of type II collagen in adjuvant in mice from H-2q-bearing crosses between SWR (H-2q/q) and B10 (H-2b/b mice), two strains known to be resistant to collagen-induced arthritis (CIA). The resistance of B10 is known to be due to its MHC haplotype, but it was postulated that the resistance of SWR mice which expresses the susceptible MHC haplotype could be due to the deletion of close to 50% of the V beta genes of the T cell receptor (TCR) in them. 17% of the F1 hybrids, 33% of the SWR backcrosses, 68% of the B10 backcrosses, and 52% of the F2 hybrids developed arthritis on follow-up to 5 mo after primary immunization with collagen. There was no significant difference in anti-type II collagen antibody titers between the arthritic and nonarthritic mice in each of these crosses. The segregation of the TCR genes with arthritis was determined in the F2 population by typing with F23.1 mAb that reacts with T cells using V beta 8 subfamily genes in their TCRs. SWR mice are F23.1- as V beta 8 genes are deleted in them. All six of arthritic mice homozygous for H-2q, and thus with an H-2 haplotype similar to SWR mice, expressed the F23.1 marker. These studies indicate that for complete susceptibility to collagen-induced arthritis, not only is a susceptible MHC haplotype (H-2q) important, but possibly also the presence of a subset of T cells using certain specific V beta genes in their TCRs. Other background genes may, however, modulate the severity of arthritis.

Development of B lymphocytes in mice heterozygous for the X-linked immunodeficiency (xid) mutation. xid inhibits development of all splenic and lymph node B cells at a stage subsequent to their initial formation in bone marrow.

CBA/N mice were crossed with CBA/Ca-Pgk-1a to produce female F1 hybrids that were heterozygous for both xid and the phosphoglycerate kinase 1 (PGK-1) allozymes. PGK acted as a quantifiable marker for the frequency of cells in which the xid-bearing X chromosome was active in lymphocytic and other cell populations. In adults, such cells (termed xid cells) were virtually absent in FACS-sorted splenic and lymph node B cells, and in all three splenic subpopulations distinguished on the basis of their relative expression of membrane mu and delta chains. Thus, the xid mutation appeared to compromise the development of all B cells. Erythrocytes, thymocytes, T cells, and granulocytes were unaffected. Selection against xid cells was less pronounced in the spleens of 2-6-wk-old mice. In the bone marrow, there was evidence for selection against xid in the production of B cells (except at 2 wk of age), but not at the pre-B cell level. These data suggest that, in competition with normal non-xid cells, newly-formed xid B cells were less likely to be incorporated into the peripheral B cell pool.

Abnormalities induced by the mutant gene, lpr. Patterns of disease and expression of murine leukemia viruses in SJL/J mice homozygous and heterozygous for lpr.

SJL/J mice heterozygous or homozygous for the lpr mutation were compared with SJL/J-+/+ mice for longevity, histopathology, antigenic characteristics of lymphocytes and expression of murine leukemia viruses (MuLV). In comparison to +/+ mice, lpr homozygotes had a markedly shortened life span, died with infiltrative pulmonary disease, but little or no renal disease, and expressed high levels of infectious ecotropic MuLV in lymphoid tissues. SJL-lpr/+ mice had a life span intermediate between SJL-+/+ and -lpr/lpr mice, died with lymphomas that histologically resembled the neoplasms of +/+ mice, and sometimes expressed high levels of ecotropic MuLV. The lymphomas of lpr/+ could be transplanted to +/+ recipients in 78% of cases, and continuous in vitro lines were established from some of them. Similar effects on virus expression or lymphoma development were not observed in other strains homozygous or heterozygous for the lpr mutation. These results indicate that the diseases expressed by mice homozygous for the lpr mutation are highly strain-dependent, and that this gene can have an effect in the heterozygous state in SJL mice.

Fate of bone marrow-derived cultured mast cells after intracutaneous, intraperitoneal, and intravenous transfer into genetically mast cell-deficient W/Wv mice. Evidence that cultured mast cells can give rise to both connective tissue type and mucosal mast cells.

Both connective tissue mast cells and mast cells grown in vitro are derived from multipotential hematopoietic stem cells, but these two mast cell populations exhibit many differences in morphology, biochemistry, and function. We investigated whether the phenotype of cultured mast cells or their progeny was altered when the cells were transferred into different locations in vivo. Cultured mast cells were immature by ultrastructure, and stained with alcian blue but with neither safranin or berberine sulfate, a fluorescent dye that binds to the heparin of connective tissue mast cell granules. By contrast, mast cells recovered from the peritoneal cavity of congenitally mast cell-deficient (WB X C57BL/6)F1-W/Wv (WBB6F1-W/Wv) mice 10 wk after intraperitoneal injection of cultured WBB6F1-+/+ or C57BL/6-bgJ/bgJ mast cells stained with both safranin and berberine sulfate. Staining with berberine sulfate was prevented by treatment of the cells with heparinase but not chondroitinase ABC, suggesting that the adoptively transferred mast cell population had acquired the ability to synthesize and store heparin. Furthermore, the recovered mast cells were indistinguishable by ultrastructure from the normal mature peritoneal mast cells of WBB6F1-+/+ mice, and contained substantially more histamine than mast cells studied directly from culture. Intravenous injection of cultured mast cells resulted in the development of safranin-and berberine sulfate-positive mast cells in the peritoneal cavity, spleen, skin, and glandular stomach muscularis propria. Mast cells also developed on the glandular stomach mucosa, but these cells stained with alcian blue rather than safranin, and did not stain with berberine sulfate. This result suggests that cultured mast cells can give rise to mast cells of either the connective tissue type or mucosal phenotype, depending on anatomical location. Furthermore, transplantation of cultured mast cells into WBB6F1-W/Wv mice had no measurable effect on the anemia of the recipient mice, suggesting a possible strategy for repairing the mast cell deficiency of WBB6F1-W/Wv mice without affecting other bone marrow-derived populations such as erythrocytes. Intravenous injection of representative connective tissue type mast cells (30-50% pure peritoneal mast cells derived from WBB6F1-+/+ mice) gave results similar to those obtained with cultured mast cells: mast cells developing in the peritoneal cavity, skin, spleen, and glandular stomach muscularis propria of WBB6F1-W/Wv recipients stained with safranin and berberine sulfate, whereas mast cells developing in the mucosa of the glandular stomach stained only with alcian blue.(ABSTRACT TRUNCATED AT 400 WORDS)

The contribution of L3T4+ T cells to lymphoproliferation and autoantibody production in MRL-lpr/lpr mice.

The current study examines the role of the L3T4 T cell subset in the development of lupus-like autoimmunity and lymphoproliferation in lpr-bearing mice. Chronic treatment of MRL-lpr/lpr mice with anti-L3T4 antibody beginning at 4 wk old was found to markedly decrease the production of IgG anti-DNA and antihistone antibodies, while having no effect on IgM autoantibodies. A dramatic reduction in splenomegaly and lymphadenopathy was also observed coincident with a decrease in the percentage and total number of Thy-1+, B220+ cells. Together, the data suggest an important role for L3T4+ T cells in the pathogenesis of disease in lpr mice and provide further evidence that a requirement for the L3T4 subset may be a common feature of murine autoimmunity.