CpG DNA induces cyclooxygenase-2 expression and prostaglandin production.

Unmethylated CpG motifs found in bacterial DNA are potent activators of the innate and acquired immune systems, and rapidly induce the production of proinflammatory cytokines. We hypothesized that CpG DNA may also elicit the production of prostaglandins (PG), which are central lipid mediators of the immune and inflammatory response. To test our hypothesis, we stimulated murine spleen cells and RAW 264.7 murine macrophage cells with CpG DNA and assessed the effects on the PG synthesis pathway. Compared to control, DNA-containing CpG motifs induced >5-fold increase in PGE (2) production and rapidly up-regulated cyclooxygenase-2 (COX-2) at both the mRNA and protein level. CpG DNA was an extremely strong inducer of COX-2 as concentrations as low as 3 ng/ml induced COX-2 protein expression. The CpG DNA-induced PGE (2) down-regulated the immune response elicited by CpG. Blockade of PGE (2) production with selective COX-2 inhibitors or neutralizing anti-PGE (2) antibody markedly enhanced IFN-gamma secretion in vitro from CpG DNA-stimulated spleen cells. Moreover, selective COX-2 inhibition increased CpG DNA-induced IFN-gamma secretion in vivo. Inhibition of COX-2 also increased CpG DNA-induced lytic activity of NK cells. Taken together, these data indicate that DNA containing CpG motifs is a potent inducer of COX-2 and PGE (2) production. CpG-induced PG may subsequently down-regulate the immune and inflammatory responses elicited by the CpG DNA.

Specific inhibition of cyclooxygenase 2 restores antitumor reactivity by altering the balance of IL-10 and IL-12 synthesis.

Cyclooxygenase-2 (COX-2), the enzyme at the rate-limiting step of prostanoid production, has been found to be overexpressed in human lung cancer. To evaluate lung tumor COX-2 modulation of antitumor immunity, we studied the antitumor effect of specific genetic or pharmacological inhibition of COX-2 in a murine Lewis lung carcinoma (3LL) model. Inhibition of COX-2 led to marked lymphocytic infiltration of the tumor and reduced tumor growth. Treatment of mice with anti-PGE2 mAb replicated the growth reduction seen in tumor-bearing mice treated with COX-2 inhibitors. COX-2 inhibition was accompanied by a significant decrement in IL-10 and a concomitant restoration of IL-12 production by APCs. Because the COX-2 metabolite PGE2 is a potent inducer of IL-10, it was hypothesized that COX-2 inhibition led to antitumor responses by down-regulating production of this potent immunosuppressive cytokine. In support of this concept, transfer of IL-10 transgenic T lymphocytes that overexpress IL-10 under control of the IL-2 promoter reversed the COX-2 inhibitor-induced antitumor response. We conclude that abrogation of COX-2 expression promotes antitumor reactivity by restoring the balance of IL-10 and IL-12 in vivo.

Attenuation of IL-5-mediated signal transduction, eosinophil survival, and inflammatory mediator release by a soluble human IL-5 receptor.

A soluble form of the human IL-5R alpha-chain (IL-5Ra) that contains the extracellular IL-5 binding domain has been evaluated for its effect on IL-5 binding to and activation of human eosinophils and basophils. The truncated receptor was expressed in Escherichia coli and recovered in biologically active form following renaturation and anion exchange chromatography. The soluble receptor formed a 1/1 complex with IL-5 in solution and bound IL-5 with affinity comparable to that of cell-associated IL-5Ra. Soluble IL-5Ra also competed with IL-5 for binding to the native alpha beta IL-5R on human cells and inhibited IL-5-mediated receptor activation and inflammatory mediator production. In this regard, the soluble receptor prevented IL-5-induced tyrosine phosphorylation of JAK2 kinase and IL-5R beta-chain and inhibited IL-5 priming of leukotriene C4 release by human basophils. However, the E. coli-derived receptor failed to inhibit IL-5 in longer term assays, including eosinophil survival and TF-1 cell proliferation, possibly due to its propensity to aggregate in a time- and temperature-dependent manner. In contrast, we observed that a soluble IL-5Ra derived from baculovirus-infected cells was less prone to aggregate and effectively antagonized IL-5-induced cell proliferation and survival. These findings indicate that the extracellular portion of the human IL-5Ra chain can prevent association of IL-5 with cell surface receptors and can attenuate signal transduction, mediator release, and survival of inflammatory cells. As such, soluble IL-5R may be useful in treating diseases such as human asthma, in which pulmonary injury is associated with the activity of IL-5R-bearing cells.

Inhibition of cyclooxygenase-2 rapidly reverses inflammatory hyperalgesia and prostaglandin E2 production.

PGs derived from cyclooxygenase-2 (COX-2), in particular PGE2, play important roles in the initiation of inflammation and pain. In the present study, we evaluated the role of COX-2-derived PGE2 in an animal model of established hyperalgesia. Inflammation and hyperalgesia were first induced by injection of carrageenan into rat footpads. Then we investigated the effects of subsequent therapeutic treatment with a selective COX inhibitor, with a nonsteroidal anti-inflammatory drug and with anti-PGE2 antibody. Test compounds were administered 1 to 3 hr after carrageenan challenge, and inhibition of pain (hyperalgesia, measured by withdrawal from a thermal stimulus), and changes in paw edema and PG levels were evaluated. The i.v. administration of a nonselective COX inhibitor, ketorolac, caused a rapid reduction in hyperalgesia in the inflamed footpad, returning it to near-normal values within 1 hr. Normal (control) paw response times were not affected. Therapeutic administration of ketorolac prevented most further swelling caused by carrageenan but did not reverse edema already present at the time of dosing. Administered p.o., a selective COX-2 inhibitor (SC-58635) was as efficacious as ketorolac in reducing inflammatory hyperalgesia. Footpad PG levels returned to base line or below within 5 min of dosing with ketorolac, which suggests rapid turnover of PG in the inflamed tissue. Therapeutic treatment with a monoclonal anti-PGE2 antibody also fully reversed the hyperalgesia response. These studies suggest that continuous production of PGE2 by the COX-2 enzyme is a critical element in sustaining the hyperalgesic response at sites of tissue inflammation.

Selective neutralization of prostaglandin E2 blocks inflammation, hyperalgesia, and interleukin 6 production in vivo.

The role of prostaglandin E2 (PGE2) in the development of inflammatory symptoms and cytokine production was evaluated in vivo using a neutralizing anti-PGE2 monoclonal antibody 2B5. In carrageenan-induced paw inflammation, pretreatment of rats with 2B5 substantially prevented the development of tissue edema and hyperalgesia in affected paws. The antibody was shown to bind the majority of PGE2 produced at the inflammatory site. In adjuvant-induced arthritis, the therapeutic administration of 2B5 to arthritic rats substantially reversed edema in affected paws. Anti-PGE2 treatment also reduced paw levels of IL-6 RNA and serum IL-6 protein without modifying tumor necrosis factor RNA levels in the same tissue. In each model, the antiinflammatory efficacy of 2B5 was indistinguishable from that of the nonsteroidal antiinflammatory drug indomethacin, which blocked the production of all PGs. These results indicate that PGE2 plays a major role in tissue edema, hyperalgesia, and IL-6 production at sites of inflammation, and they suggest that selective pharmacologic modulation of PGE2 synthesis or activity may provide a useful means of mitigating the symptoms of inflammatory disease.

Di- and trinucleotide target preferences of somatic mutagenesis in normal and autoreactive B cells.

During Ag-driven development of memory B cells, Ab V genes are modified by somatic mutagenesis. Although V gene somatic mutations have important biologic consequences in both physiologic and autoimmune Ab responses, little is known about the mechanism of mutation, or whether it operates normally in autoreactive B cells. To approach these issues, we analyzed somatic mutations in Ab genes for evidence of sequence-specific target preferences. Our analysis was confined to noncoding segments of V genes so that the intrinsic characteristics of the somatic mutation process could be reliably dissociated from the indirect but substantial influences of cellular selection. We consistently observed that some dinucleotides, GC and TA in particular, mutated at frequencies that were higher than expected based on their frequency of occurrence. Most of the dinucleotide mutation preferences could not be extrapolated directly from mononucleotide mutation preferences. Specific trinucleotides, including AGC, TAC, and their inverse repeats (GCT, GTA), also mutated more frequently than expected. These and other mutation characteristics were virtually indistinguishable in V genes of normal and autoreactive B cells. An analysis of mutations in published flanking sequences confirmed the target preferences, as did an examination of reported "hot spots" within coding V sequences. The shared preferences in coding and noncoding regions of V genes suggests that somatic mutations are generated de novo. Collectively, our findings indicate that the somatic mutation process exhibits sequence-specific preferences, consistent with an untemplated mechanism, and appears to operate similarly in normal and autoreactive B cells.

Characterization of a monoclonal antibody that neutralizes the activity of prostaglandin E2.

We describe the generation of an IgG1 mAb, 2B5, that neutralizes the biologic activity of PGE2 in vitro and in vivo. The Ab was derived from a BALB/c mouse that was immunized with a PGE2-thyroglobulin conjugate. 2B5 is one of the highest affinity and specific anti-PGE2 mAbs reported to date. The Kd for PGE2 was approximately 300 pM and crossreactivity toward eicosanoids other than PGE1 was less than 1%. The addition of 2B5 to [3H]PGE2 blocked the binding of radioligand to cell membranes transfected with the murine EP3 prostaglandin receptor. In functional studies, 2B5 neutralized the capacity of PGE2 to suppress T cell proliferation induced by a mitogenic anti-CD3 Ab in vitro. In contrast, immunosuppression by the phosphodiesterase inhibitor, isobutylmethylxanthine was not affected. In an in vivo model of nociception, 2B5 substantially reduced the dorsoflexion response of mice to phenylbenzoquinone. This response is associated with prostaglandin production and is blocked by inhibitors of prostanoid synthesis. Our findings demonstrate that this nociceptive response is largely due to PGE2. In the absence of antagonists that prevent PGE2 from activating a diverse family of receptor subtypes, neutralizing Abs to PGE2 should represent useful reagents to delineate the biologic properties of this eicosanoid in vivo.

Rapid and selective induction of blood eosinophilia in guinea pigs by recombinant human interleukin 5.

Interleukin 5 (IL-5) has been implicated as a causal mediator in the development of pulmonary eosinophilia and airway hyperreactivity in human asthma. Supportive evidence for a pathogenic role of IL-5 in this disease has been provided by guinea pig models in which antigen-induced lung eosinophilia and bronchial hyperresponsiveness have been specifically attenuated with a neutralizing antibody to IL-5. In the present report, we describe a rapid mechanism-based model of IL-5-induced eosinophilia in the guinea pig. Our results show that intravenous injection of human IL-5 induced a 5-10-fold increase in the percentage and number of eosinophils in blood within 1 hour. In contrast, neutrophils and mononuclear cells were not recruited during this time. The increase in eosinophils was blocked by pretreatment of animals with an anti-IL5 monoclonal antibody (TRFK5) in doses similar to those which inhibit eosinophilia in guinea pig asthma models. Furthermore, dexamethasone, a potent inhibitor of eosinophilia in guinea pig asthma, profoundly suppressed the eosinophilia induced by human IL-5. This rapid model will be useful for elucidating the eosinophil activating properties of IL-5 in vivo and may potentially facilitate the development of IL-5 receptor antagonists for the specific blockade of the eosinophilic inflammation observed in human asthma.

An early post-mutational selection event directs expansion of autoreactive B cells in murine lupus.

We report evidence for a strong selection event directing the outgrowth of autoreactive B cells in spontaneous murine lupus. The event occurred shortly following the induction of the somatic hypermutation process. This conclusion is derived from extensive sequence analyses of VH and VL loci expressed by hybridomas representing two large histone-specific clones (lineages) from an autoimmune (NZB x SWR)F1 mouse. To obtain unambiguous somatic mutational information, we devised a strategy to amplify and sequence the JH and JK clusters that flank expressed V genes. Somatic mutations in V flanking sequences of the two autoreactive clones revealed that in one clone the pattern was relatively simple: the frequency of mutation was low, and only one somatic mutation was shared by all clone members. Members of the second large histone-specific clone contained many somatic mutations in combinations that indicated numerous rounds of selection. Importantly, however, as observed with the first clone, one observed somatic mutation was shared by all clone members. Since, for each clone, all members shared only one visible mutation over extensive sequence tracts, we conclude that the autoreactive clones were derived from single precursors that had just begun to mutate their V genes. The data indicate that a strong selection event had occurred shortly after the initial acquisition of somatic mutation(s) in precursors to each clone, at a stage of development corresponding to that of the germinal center B cell approximately 1 week post immunization.

Histone autoantigens in murine lupus. Definition of a major epitope within an accessible region of chromatin.

In SLE and in the (NZB x NZW)F1 murine model of this disease, IgG autoantibodies are frequently produced to DNA and histones. In the present study, we define a linear epitope on histone H2B that is recognized by (NZB x NZW)F1 mice. IgG antibodies from anti-H2B positive (but not anti-H2B negative) mice bound strongly to a peptide containing the first 15 N-terminal amino acids, a region that is exposed in chromatin. Competitive inhibition studies showed that the binding of autoantibodies to H2B in ELISA as well as the binding to soluble H2B was substantially blocked by this peptide. Studies with smaller peptides mapped the epitope to residues 3-12. Individual mice recognized different residues within this region, and a sequence search did not reveal proteins other than H2B that could elicit this spectrum of antibodies. Interestingly, these autoantibody specificities were not a component of those induced in preautoimmune mice by immunization with H2B/RNA complexes or with H2B peptide 1-30 containing the autoantigenic sequence. These findings argue that recognition of a specific N-terminal region of self histone contributes to the anti-H2B autoantibody response in lupus. Autoreactive B cells with specificity for this sequence seem to develop only after the autoimmune process has been initiated.

Self-reactive T cells in murine lupus: analysis of genetic contributions and development of self-tolerance.

Our understanding of the immune mechanisms that lead to systemic lupus erythematosus has been greatly advanced by the availability of murine models which display both serological and clinical features of the human disease. Studies have demonstrated that CD4+ T cells are required for the full expression of disease in these mice. (NZB X NZW)F1 mice exhibit a lupus-like disease (elevated levels of IgG antinuclear antibodies and a fatal glomerulonephritis) that is not characteristic of either parent. At least three gene loci have been identified in NZW mice that could potentially contribute to a T cell-dependent autoimmune disease, including the T cell receptor alpha- and beta-chain gene complexes and the major histocompatibility complex (MHC). The NZW T cell receptor beta-chain complex appeared to be particularly unusual in that the C beta 1, D beta 2, and J beta 2 gene segments have been deleted. However, an analysis of (NZB X NZW)F1 X NZB back-cross mice revealed no association of disease expression with the presence of this allele. There was also no correlation of disease incidence with the presence of the NZW T cell receptor alpha-chain allele. In contrast, nearly 90% of the backcross mice with the NZW MHC expressed severe autoimmune disease compared with 12% of the mice that did not carry this haplotype. Additional studies strongly suggested that the gene(s) within the NZW MHC is the only dominant NZW genetic contribution to F1 disease. We also determined if self-reactive T cells are able to escape thymic tolerance in autoimmune New Zealand and MRLlpr/lpr mice. In nonautoimmune mice expressing I-E, T cells utilizing V beta 17a and V beta 11 encoded domains have been shown to be clonally eliminated in the thymus. Similarly, V beta 8.1+ and V beta 6+ T cells are tolerized in nonautoimmune mice expressing Mls-1a. These T cell subsets were quantified in the lymph nodes and spleens of (NZB X NZW)F1, (NZB X SWR)F1, and MRL-lpr/lpr mice before and after the development of lupus-like disease. The results indicate that peripheral T cells in these mice, including the massive CD4-, CD8- T cell population in lpr mice, have been modified by normal mechanisms of tolerance such that potential self-reactive V beta specificities have been eliminated in the thymus.

Allogeneic MHC antigen requirements for lupus-like autoantibody production and nephritis in murine graft-vs-host disease.

A graft-vs-host (GVH) reaction of parental T cells in allogeneic F1 mice can lead to an autoimmune disease resembling human SLE. We analyzed the contribution of MHC genes to the development of IgG antinuclear antibody production and immune complex glomerulonephritis in MHC-congenic F1 recipients. DBA/2 T cells elicited IgG antibodies to histone, ssDNA, and dsDNA in all histoincompatible F1 recipients that were studied. The anti-DNA antibody responses were quantitatively similar among the F1 combinations and displayed comparable IgG2a subclass and cationic charge characteristics. In contrast, severe renal disease was manifested only in F1 mice that expressed H-2b encoded class II gene products. Disease susceptibility was associated with a decrease in circulating anti-DNA antibodies and a characteristic localization of immune complexes in the glomeruli. The data suggest that the production of potentially pathogenic IgG anti-nuclear antibodies is not sufficient for the development of renal disease in GVH-induced lupus. Thus, another event separate from autoantibody production is MHC dependent and appears to be critical for the formation and/or deposition of pathologic immune complexes.

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.

Selective production of autoantibodies in graft-vs-host-induced and spontaneous murine lupus. Predominant reactivity with histone regions accessible in chromatin.

The injection of (C57BL/6 X DBA/2)F1 mice with parental DBA/2 lymphoid cells leads to a lupus-like disease in which IgG autoantibodies are targeted to certain nuclear and cell surface antigens. To investigate further the extent of antibody diversity in this graft-vs-host (GVH) model, we studied the specificity of antihistone antibodies induced by the GVH reaction. High levels of IgG antibodies to histones H1 and H2B were detected whereas responses to H2A, H3, and H4 were only marginally elevated above pre-GVH levels. Immunoblotting analysis further revealed that the response to H2B was focused on epitopes that most likely reside in the N-terminal region. In contrast, F1 mice immunized with H2B/RNA complexes in adjuvant produced antibodies to the N terminus as well as to other regions of the H2B molecule. Thus, the antihistone response stimulated by the GVH reaction is only a fraction of the potentially activatable B cell repertoire. We also determined whether antibodies that arise spontaneously in genetically predisposed lupus strains were restricted in their histone reactivity. The response to core histones was highly variable among individual animals of the NZB/NZW and MRL-lpr/lpr strains despite their inbred nature. However, nearly all mice exhibited a preferential reactivity for epitopes in histone regions that are lost after partial trypsin digestion of chromatin. These data demonstrating autoantibody responses that are limited to particular histone regions support a mechanism by which B cells are selectively activated in murine lupus. The predominant production of antibodies to histone regions that are exposed in nucleosomes raises the possibility that chromatin is an antigenic stimulus for histone-specific B cells in this disease.

Anti-histone antibodies in idiopathic and drug-induced lupus recognize distinct intrahistone regions.

We have identified regions within core histones that are antigenic for autoantibodies in systemic lupus erythematosus (SLE) and drug-induced lupus. An immunoblotting technique was used to determine the reactivity of lupus antibodies for intact histones and for trypsin-resistant histone fragments that lack the amino- and carboxyl-terminal amino acids that are normally exposed in native nucleosomes. In SLE, the predominant anti-histone response was restricted to epitopes in the trypsin-sensitive regions. Of 20 SLE sera that had strong antibody activity for multiple intact histones, 17 showed minimal activity with any of the corresponding trypsin-resistant fragments. A markedly different pattern of reactivity was present in sera of patients with procainamide (Pr)-induced lupus in which antibodies to H2A, H2B, and the H2A-H2B complex had strong fragment activity. Interestingly, recognition of trypsin-resistant fragments was also noted in a small number of SLE sera that contained antibodies to the H2A-H2B complex. In contrast to both SLE and Pr-induced lupus, antibodies induced by hydralazine (Hy) reacted primarily with H3 and H4. Furthermore, these antibodies bound equally well to the corresponding trypsin-resistant regions that are thought to be relatively unexposed in native nucleosomes. Thus, the specificities of anti-histone antibodies in SLE, Pr-induced lupus, and Hy-induced lupus are markedly different, but in each disease reactivity appears to be restricted to a limited number of histone determinants. The data raise the possibility that autoantigen in the form of native nucleosomes may be recognized in SLE and possibly in Pr-induced lupus. In contrast, the propensity of Hy to induce autoantibodies to determinants usually not recognized in SLE or Pr-induced lupus may suggest a different immunogenic stimulus in this disease.

Interleukin 2 is a proliferative signal for B cells from autoimmune mice.

T cells from murine lupus strains manifest complex defects in interleukin 2 (IL 2) production and receptor expression. The capacity of B cells from such mice to utilize IL 2 as a growth factor has not been previously reported and is examined herein. Anti-Thy-1.2 plus complement-treated spleen cells from 6-8-week-old autoimmune MRL-lpr/lpr mice and from age and sex-matched immunologically normal CBA/J mice were cultured with lipopolysaccharide (LPS) for 36 h and analyzed for the expression of IL 2 receptors using the monoclonal antibody 7D4. The percentage of B cells expressing IL 2 receptors was comparable in MRL-lpr/lpr and CBA/J mice. In contrast to those from CBA/J, BALB/c and (BALB/c X NZW)F1 mice, LPS-stimulated B cells from MRL-lpr/lpr and from (NZB X NZW)F1 mice were capable of proliferating in response to IL 2. Fractionation of MRL-lpr/lpr B cells using Percoll gradient density separation demonstrated that the IL 2-responsive population consisted predominantly of large cells. In addition, unfractionated B cells from MRL-lpr/lpr mice were found to be substantially more responsive to IL 2 than those from CBA/J and BALB/c mice following activation with anti-immunoglobulin plus LPS. The hyper-responsiveness to IL 2 may be a consequence of the state of activation of autoimmune B cells and is of potential importance in the pathogenesis of systemic lupus erythematosus.

Autoimmunization in murine graft-vs-host disease. I. Selective production of antibodies to histones and DNA.

A lupus-like disease characterized by a severe immune complex glomerulonephritis and IgG autoantibody production was induced in (C57BL/6 X DBA/2)F1 mice by injection of parental DBA/2 lymphoid cells. The ensuing graft-vs-host (GVH) reaction resulted in a 10- and a 100-fold increase in serum IgG antibody levels to denatured DNA and total histones, respectively, compared with that in F1----F1 control mice. The level of anti-DNA antibodies peaked 2 wk after injection of DBA/2 cells and preceded peak anti-histone levels by approximately 2 wk. Anti-histone antibodies were generated predominantly to histones H1, H2A, and H2B, a profile different from that observed in NZB/NZW and MRL-lpr/lpr mice. The marked increase in IgG antinuclear antibodies did not correlate with increases in total IgG serum levels and was not associated with comparable increases in antibodies to transferrin, hemoglobin, fibrinogen, or thyroglobulin. Selective autoantibody production was also observed in vitro, wherein GVH spleen cells produced high levels of IgG antibodies to total histones and denatured DNA but not to these non-nuclear protein antigens. In contrast, spleen cells stimulated in vitro with lipopolysaccharide produced equivalent amounts of antibodies to all antigens tested. Our results are in agreement with those of other investigators and collectively suggest that IgG autoantibodies in GVH disease, and possibly in spontaneous lupus-like disease, are not secondary to a generalized B cell activation, but may be selectively generated in response to self antigens with unique configurational properties.

Rabbits immunized with mixtures of staphylococcal protein A and autologous IgG produce anti-human IgG antibodies.

The results of this study provide evidence that protein A may render IgG immunogenic in the autologous host. Antibodies to human but not rabbit IgG were detected in sera of rabbits immunized with a mixture of autologous serum and protein A. Anti-human IgG antibodies appeared within 2 weeks at which time the antibodies were of the IgM class. Upon further immunization, both IgM and IgG antibodies were produced with the IgG class predominating. The antibodies elicited by a mixture of protein A with autologous IgG resembled those which arise in response to autologous IgG that has been denatured by physicochemical means, in that they react mainly with foreign species IgG and weakly, if at all, with IgG of rabbit origin.