The role of dendritic cells (DC) in urinary tract infections (UTI) is unknown. These cells contribute directly to the innate defense against various viral and bacterial infections. Here, we studied their role in UTI using an experimental model induced by transurethral instillation of the uropathogenic Escherichia coli (UPEC) strain 536 into C57BL/6 mice. While few DC were found in the uninfected bladder, many had been recruited after 24 h, mostly to the submucosa and uroepithelium. They expressed markers of activation and maturation and exhibited the CD11b tumor necrosis factor alpha (TNF-␣)-and inducible nitric oxide synthase (iNOS)-producing CD11bINT DC (Tip-DC) were detected, which recently were proposed to be critical in the defense against bacterial infections. However, Tip-DC-deficient CCR2 ؊/؊ mice did not show reduced clearance of UPEC from the infected bladder. Moreover, clearance was also unimpaired in CD11c-DTR mice depleted of all DC by injection of diphtheria toxin. This may be explained by the abundance of granulocytes and of iNOS-and TNF-␣-producing non-DC that were able to replace Tip-DC functionality. These findings demonstrate that some of the abundant DC recruited in UTI contributed innate immune effector functions, which were, however, dispensable in the microenvironment of the bladder.

Treatment of healthy rats and mice with a single intravenous injection of recombinant human tumor necrosis factor-alpha (rHuTNF-a) caused a dose-dependent gastrointestinal inflammation. Within 30 min gastric emptying was blocked and tissue edema occurred in the small and large intestine. In the cecum hemorrhage occurred after 4 h at doses 2 250 ag/kg. The cecum exhibited an acute inflammatory response following rHuTNF-a treatment similar to that seen in tumor necrosis at the same dose. The vascular endothelium became swollen, increased numbers of neutrophils and other leukocytes attached to and penetrated the endothelium, and finally hemorrhage occurred. Treatment of rats with daily injections of rHuTNF-a (250 tig/kg per d) for 3 wk failed to produce cachexia. Within 2448 h rats became resistant to the hemorrhagic effect of rHuTNF-a, however, the cytokine still caused a transitory block of gastric emptying after 10 d of treatment. Treatment at 5-or 10-d intervals produced results similar to the initial injection. These results suggest that maximum hemorrhagic response will occur when rHuTNF-a is administered at intervals of 5-10 d rather than daily.

Highly purified populations of large granular lymphocytes (LGL) have been shown to mediate natural killer (NK) cell activity. The mechanism of target cell killing by NK cells is as yet undefined; however, it has been postulated that such killing may involve soluble cytotoxic factors produced and secreted by NK cells. The data presented show that NK-sensitive, but not NK-resistant, tumor cell lines induce highly purified populations of human LGL to produce factors with cytotoxic and/or cytostatic activities. We have identified one of these factors as tumor necrosis factor-alpha (TNF-alpha), and have shown that production of this factor is enhanced by recombinant human interferon-gamma (rHuIFN-gamma). We have also examined the role of TNF-alpha in the cytotoxic function of NK cells. The data show that although highly purified LGL populations produce low levels of TNF-alpha, the cytotoxic/cytostatic activity of this lymphokine on tumor target cells does not correlate with the cytotoxic activity of highly purified populations of LGL on tumor target cells. Furthermore, NK cell-mediated cytotoxicity is not reliably inhibited by antibodies directed against various epitopes of recombinant human TNF-alpha and/or recombinant TNF-beta (lymphotoxin) or rHuIFN-gamma. These data show that although TNF-alpha is produced by highly purified NK-containing LGL cell populations, this factor does not appear to be responsible for NK cell cytotoxicity against classical NK target cells such as Molt-4 or K562. We suggest that NK function can be attributed to a combination of factors rather than to a single factor alone, and that at least two major phenomena are involved in LGL function: the rapid cytotoxic events which lead to the cell lysis measured in classical in vitro NK assays such as against K562; and the release of factors such as TNF-alpha with cytotoxic/cytostatic activities which would inhibit the growth of invading tumor cells in vivo.