IN THIS ISSUE Free

Food poisoning caused by Salmonella typhimurium is characterized by an acute inflammatory reaction. However, the bacterial component responsible for acute gastroenteritis has not been identified. Zeng et al. (p. 3668 ) used cDNA microarray analyses to profile gene expression of an in vitro polarized epithelial cell monolayer in response to a 4-h exposure to live S. typhimurium. A subset of 29 induced genes discriminated between the expression profiles seen after exposure of the monolayer to S. typhimurium and after exposure to non-proinflammatory pathogens. The S. typhimurium response was dependent on direct contact between bacteria and cells. Regulatory and structural mutations affecting flagellin protein synthesis rendered S. typhimurium incapable of up-regulating the proinflammatory gene cluster. Application of purified flagellin to the basolateral side, but not to the apical side, of the monolayer was as effective as live bacteria in eliciting the proinflammatory profile, indicating that flagellin must cross the intestinal epithelium to be proinflammatory. The authors found that proinflammatory strains of S. typhimurium, but not non-proinflammatory strains, activated the NF-κB and mitogen-activated protein kinase signal transduction pathways. Flagellin is the first bacterial determinant of inflammation to be identified using expression profiling.

Collagen can induce the conversion of immature dendritic cells (iDC) to mature DC (mDC) in the periphery. However, the collagen receptor involved in DC maturation has not been identified. Matsuyama et al. (p. 3520 ) found high levels of discoidin domain receptor 1 (DDR1) expressed during growth factor-induced maturation of human iDC to mDC in vitro. Treatment of iDC with an agonistic anti-DDR1 mAb resulted in up-regulation of CD83 and CD86 compared with an Ab control; further DC maturation occurred with the addition of TNF-α or LPS. The authors primed CD8⁺ T cells with DCs pulsed in vitro with lysates from HLA-compatible human melanoma cells. CTLs primed with mDCs induced with TNF-α and anti-DDR1 mAb killed 3-fold more tumor cells than CTLs primed with mDCs induced with TNF-α and a control Ab. Anti-DDR1 mAb, but not the control Ab, induced autophosphorylation of both DDR1 and p38α mitogen-activated protein kinase (p38αMAPK) in iDC and mDC. p38αMAPK coprecipitated with TNFR-associated factor 6 (TRAF6) and TGFβ-activated kinase 1 binding protein 1β (TAB 1β) in mDC activated by anti-DDR1 Ab. Pretreatment of iDC with an inhibitor of p38α MAPK autophosphorylation inhibited TNF-α-induced expression of mDC markers and abrogated T cell activation. The results indicate that activation of DDR1 may present a way of enhancing DC maturation and anti-tumor activity through the TRAF6/TAB 1β/p38α MAPK signaling pathway.

T cells from 37% of patients with systemic lupus erythematosus (SLE) lack protein kinase A II activity. At the same time, those cells have increased amounts of the β isoform of the RII subunit (RIIβ) of protein kinase A II in their nuclei. It is not clear if the higher levels of RIIβ contribute to SLE pathology. Elliott et al. (p. 3636 ) transiently transfected a RIIβ-deficient mouse cell line and a normal human T cell line with an RIIβ expression plasmid. RIIβ accumulated in the nuclei of both cell types and coprecipitated with the cAMP response element (CRE) binding protein (CREB). Transfection of the defective mouse cells with a series of RIIβ deletion mutants pinpointed the region of interaction to amino acids 50–135 at the N terminus of the RIIβ protein. This region was required to inhibit transcription of both a CREB-fusion gene and a CRE-reporter gene in human T cells activated by treatment with forskolin, anti-CD28, and anti-CD3; the expression level of each construct was 67% of that seen with empty vector controls. Expression of endogenous c-fos, a gene regulated by CREB, was inhibited 37% in the activated T cells. The experiments show that over-expression of RIIβ is responsible for inhibition of CREB-dependent transcription of some genes. This inhibition could contribute to the T effector cell dysfunction seen in SLE.

Current models of T cell polarization state that IL-12 produced by dendritic cells (DC) favors a Th1 outcome and that LPS suppresses the Th2 response. However, there are conflicting data about the roles of IL-12 and LPS. Kuipers et al. (p. 3645 ) studied in vitro stimulation of bone marrow-derived DC with LPS and found that expression of both IL-12 subunits peaked at 12 h but died out by 24 h. OVA-pulsed-DC (OVA-DC) and OVA-DC pretreated with LPS (OVALPS-DC) were cocultured with fluorescent-labeled T cells from OVA-TCR transgenic mice. The major difference between the two cocultures was a decrease in the secretion of Th2 cytokines from the T cells responding to OVALPS-DC. The decrease in Th2 cytokines and an increase in IFN-γ were seen in T cells recovered from OVA-TCR transgenic mice injected i.t. with OVALPS-DC, but not in animals injected with OVA-DC. OVALPS-DC produced more IL-12p40 subunit after migration to the mediastinal lymph nodes compared with OVA-DC. OVALPS-DC injection reduced the numbers of eosinophils in lungs of mice challenged with OVA, whereas mice injected with OVA-DC developed eosinophilic airway inflammation. The decrease in eosinophilia was accompanied by a decrease in Th2 cytokines. Wild-type mice injected with IL-12p40^−/− OVALPS-DC also had reduced eosinophilia and lower Th2 cytokine levels compared with mice injected with IL-12p40^−/− OVA-DC. The data show that the effect of LPS stimulation of DC in eosinophilic airway inflammation is independent of IL-12.

Inflammatory responses in the lung have different outcomes. But little is known about the genes responsible for Th1-mediated cytotoxicity and for Th2-mediated fibrosis. Sandler et al. (p. 3655 ) developed gene expression profiles in the Schistosoma mansoni egg model of mouse lung inflammation. Lymph node cells from egg-sensitized, Th2-polarized (IL-10/IL-12 knockout (KO)) mice produced higher levels of IL-4 and IL-13 and no IFN-γ 14 days after egg challenge compared with cells from sensitized wild-type (WT) mice. Cells from egg-sensitized, Th1-polarized (IL-10/IL-4 KO) mice produced no IL-4 or IL-13 and high levels of IFN-γ. Lung granulomas were large in Th2-polarized mice, small in Th1-polarized mice, and intermediate in size in WT mice. Five distinct patterns of gene expression were seen among the three strains of mice. Three patterns were seen in all three strains: genes involved with the inflammatory response were expressed by day 4; by days 8 and 14, genes involved in organogenesis were expressed; and, a set of diverse genes was down-regulated early after egg challenge. A fourth pattern, specific to the Th1-polarized phenotype, involved genes for the acute phase reaction and apoptosis, whereas the fifth pattern, specific to the Th2-polarized response, involved genes for collagens and enzymes involved in tissue repair. WT mice had expression profiles intermediate to those of Th1- and Th2-polarized mice. The authors demonstrate that the use of KO strains of mice and gene expression profiling are useful in defining the divergent pathways of inflammation and pathology.

Destruction of the acetylcholine receptor (AChR) at the neuromuscular junction in experimental autoimmune myasthenia gravis (EAMG), the animal model for MG, is mediated by complement and anti-AChR Abs. Yet it is not known whether anti-AChR Abs activate complement by the classical or alternative pathway. Tüzün et al. (p. 3847 ) immunized C57BL/6J mice that were homozygously deleted for either C3 or C4, were heterozygous for either C3 or C4, or were wild-type with AChr in CFA. The C3 and C4 knockout mice were highly resistant to EAMG induction, whereas the heterozygotes had intermediate levels of susceptibility. Additionally, the C4^−/− mice had diminished levels of complement-binding IgG2_b Abs compared with the other groups. Immunostaining demonstrated IgG, C3, and membrane attack complex (MAC) deposits in muscle sections from C3^+/+ and C4^+/+ animals, but an absence of C3 and MAC in those from C3^−/− and C4^−/− mice. In addition, C3^−/− mice produced less IL-6, IL-10, and IFN-γ than mice from the other groups. C3, and to some extent C4, knockout mice expressed increased amounts of FasL and CD69. The experiments provide direct genetic evidence of the involvement of the classical complement pathway in the development of EAMG. They further suggest that inhibition of C4 may be a useful approach to treatment of MG.

Early thymic T cell progenitors can generate T cells, NK cells, and dendritic cells (DC). The major question is whether this tripotentiality is a property of an individual cell or whether it reflects a population potential. Shen et al. (p. 3401 ) cultured 12-days-postcoitum (12-dpc) fetal thymus (FT) cells of the phentoype CD44⁺CD25⁻FcR⁻ (FcR⁻) with a cytokine cocktail at one cell per well in 96-well plates. 30% of the cells developed into DC after seven days. Flow cytometry of the progeny of single daughter FcR⁻ cells showed that about half of the progenitors of DC also generated T and NK cells. Fluorescence analyses of tagged cells grown on stromal cell monolayers showed that only FcR⁻ FT cells capable of generating both T and NK or only T cells were able to generate DC cells. Cells that generated NK cells only had lost the DC potential. The authors showed that T/NK/DC tripotentiality existed in 12 of 40 older (14-dpc) FT cells at the CD44⁺CD25⁻ stage and in 6 of 40 of the 14-dpc FT cells at the CD44⁺CD25⁺ stage. All tripotential clones at 14-dpc generated mature T cells. These single cell analyses show that individual FT progenitors are able to generate DC as well as T and NK cells.

Tumor-derived heat shock protein gp96 (HSP96) protects mice against tumor challenge and causes the regression of established tumors. These activities are thought to result from the ability of HSP96 to chaperone tumor peptides and deliver them to APC for stimulation of specific CD8⁺ T cells. Rivoltini et al. (p. 3467 ) studied HSP96 prepared from human melanoma and colon carcinoma cells. T cells that recognized a specific melanoma epitope were incubated with monocytes from healthy donors and pulsed with HSP96 purified from melanoma cells. The T cells released high levels of IFN-γ that were blocked by a mAb specific for an HLA class I molecule, but not by one for an HLA class II molecule. Comparable results were seen with T cells and HSP96 from colon carcinoma patients and cells, respectively. Three of five melanoma patients (of a total of 28) and two of five colon carcinoma patients (of a total of 35) vaccinated with autologous tumor-derived HSP96 had increased tumor-specific CD8⁺ T cell reactivity and expressed higher levels of IFN-γ in response to tumor-specific peptides. The percentages of CD8⁺ T cells stained with HLA/tumor Ag tetramers increased significantly during the vaccination protocol. Two of the three patients with melanoma experienced complete tumor regression; the two colon carcinoma patients remain tumor free after 30 mo. The results suggest that tumor-derived HSP96 can be effective in treating human cancer.

Summaries written by Dorothy L. Buchhagen, Ph.D.