Characterization of Escherichia coli type 1 pilus mutants with altered binding specificities.

PCR mutagenesis and a unique enrichment scheme were used to obtain two mutants, each with a single lesion in fimH, the chromosomal gene that encodes the adhesin protein (FimH) of Escherichia coli type 1 pili. These mutants were noteworthy in part because both were altered in the normal range of cell types bound by FimH. One mutation altered an amino acid at a site previously shown to be involved in temperature-dependent binding, and the other altered an amino acid lining the predicted FimH binding pocket.

Genetic characterization of Escherichia coli type 1 pilus adhesin mutants and identification of a novel binding phenotype.

Five Escherichia coli type 1 pilus mutants that had point mutations in fimH, the gene encoding the type 1 pilus adhesin FimH, were characterized. FimH is a minor component of type 1 pili that is required for the pili to bind and agglutinate guinea pig erythrocytes in a mannose-inhibitable manner. Point mutations were located by DNA sequencing and deletion mapping. All mutations mapped within the signal sequence or in the first 28% of the predicted mature protein. All mutations were missense mutations except for one, a frameshift lesion that was predicted to cause the loss of approximately 60% of the mature FimH protein. Bacterial agglutination tests with polyclonal antiserum raised to a LacZ-FimH fusion protein failed to confirm that parental amounts of FimH cross-reacting material were expressed in four of the five mutants. The remaining mutant, a temperature-sensitive (ts) fimH mutant that agglutinated guinea pig erythrocytes after growth at 31 degrees C but not at 42 degrees C, reacted with antiserum at both temperatures in a manner similar to the parent. Consequently, this mutant was chosen for further study. Temperature shift experiments revealed that new FimH biosynthesis was required for the phenotypic change. Guinea pig erythrocyte and mouse macrophage binding experiments using the ts mutant grown at the restrictive and permissive temperatures revealed that whereas erythrocyte binding was reduced to a level comparable to that of a fimH insertion mutant at the restrictive temperature, mouse peritoneal macrophages were bound with parental efficiency at both the permissive and restrictive temperatures. Also, macrophage binding by the ts mutant was insensitive to mannose inhibition after growth at 42 degrees C but sensitive after growth at 31 degrees C. The ts mutant thus binds macrophages with one receptor specificity at 31 degrees C and another at 42 degrees C.

Woodchuck lymphotoxin-alpha, -beta and tumor necrosis factor genes: structure, characterization and biological activity.

We cloned and characterized the woodchuck tumor necrosis factor (TNF) and lymphotoxin-alpha, -beta (LT-alpha, -beta) cDNAs, genes and proteins to facilitate study of the functions of these cytokines during the course of woodchuck hepatitis virus (WHV) infection. Woodchuck cDNA and genomic DNA libraries were screened with woodchuck-specific DNA probes to isolate the cDNA and gene clones for TNF, LT-alpha and LT-beta. The cDNAs for woodchuck TNF, LT-alpha and LT-beta code for proteins of 233, 205 and 310 amino acids respectively. The polypeptide encoded by each gene among woodchucks, humans and mice can differ: the human TNF, LT-alpha and LT-beta genes encode polypeptides of 233, 205 and 244 amino acids respectively, whereas the mouse TNF, LT-alpha and LT-beta genes encode polypeptides of 235, 202 and 306 amino acids respectively. In the woodchuck, there are four exons for TNF, four exons for LT-alpha and three exons for LT-beta. The RNA splicing patterns for TNF, LT-alpha and LT-beta genes are identical among woodchucks, humans and mice, except that the human LT-beta gene contains four exons. The woodchuck TNF gene promoter contains consensus sequences for binding of AP-1, AP-2, C/EBPbeta, CRE, Egr-1, Ets, NF-AT, NF-kappaB and SP-1 transcription factors. LT-alpha has AP-2, Ets, NF-kappaB, SP-1 and STAT binding sites, and LT-beta has Egr-1/SP-1, Ets and NF-kappaB binding sites. The bacterially expressed woodchuck TNF and LT-alpha proteins exhibited cytotoxic activities on both mouse L929B and woodchuck A2 cells in the presence of actinomycin D. The specific activities of TNF and LT-alpha were 2.62x10(8) units/mg and 2.22x10(3) units/mg respectively for L929B cells, and 1.05x10(9) units/mg and 3.56x10(4) units/mg respectively for A2 cells. However, only woodchuck TNF showed cytotoxic activity on human HepG2 cells, with a specific activity of 6.55x10(7) units/mg in the presence of actinomycin D. The data obtained from this study will be useful to future investigations of the TNF and LT antitumor and anti-viral activities, and their therapeutic potential in the woodchuck model for human hepatitis B virus (HBV).

Host and bacterial factors involved in the innate ability of mouse macrophages to eliminate internalized unopsonized Escherichia coli.

In an effort to better understand genetic and cellular factors that influence innate immunity, we examined host and bacterial factors involved in the nonopsonic phagocytosis and killing of Escherichia coli K-12 by mouse macrophages. Unelicited (resident) peritoneal macrophages from five different mouse strains, BALB/c, C57BL/6, CD-1, C3H/HeJ, and C3H/HeN, were employed. Additional macrophage populations were obtained from CD-1 mice (bone marrow-derived macrophages). Also, for BALB/c and C57BL/6 mice, peritoneal macrophages elicited with either thioglycolate or proteose peptone, bone marrow-derived macrophages, and macrophage-like cell lines derived from the two strains were employed. Two E. coli K-12 strains that differed specifically in their abilities to produce type 1 pili containing the adhesive protein FimH were examined. The parameters used to assess macrophage bacteriocidal activity were (i) the killing of internalized (gentamicin-protected) E. coli during the approximately 4-h assay and (ii) the initial rate at which internalized E. coli were eliminated. Data on these parameters allowed the following conclusions: (i) unelicited or proteose peptone-elicited peritoneal macrophages were significantly better at eliminating internalized bacteria than thioglycolate-elicited peritoneal macrophages, bone marrow-derived macrophages, or macrophage cell lines; (ii) the host genetic background had no significant effect upon the ability of unelicited peritoneal macrophages to kill E. coli (even though the mouse strains differ widely in their in vivo susceptibilities to bacterial infection); and (iii) the FimH phenotype had no significant effect upon E. coli survival once the bacterium was inside a macrophage. Additionally, there was no correlation between the bacteriocidal effectiveness of a macrophage population and the number of bacteria bound per macrophage. However, macrophage populations that were the least bacteriocidal tended to bind higher ratios of FimH(+) to FimH(-) E. coli. The effect of gamma interferon, fetal calf serum, and the recombination proficiency of E. coli were examined as factors predicted to influence intracellular bacterial killing. These had no effect upon the rate of E. coli elimination by unelicited peritoneal macrophages.

The mucosal phase of Listeria infection.

Listeria monocytogenes is an enteroinvasive bacterial pathogen of man and animals. Listeriae have been shown capable of infecting the host by translocating from the intestinal lumen through Peyer's Patches (PP), however, results of experiments now indicate that these facultative intracellular parasites may also translocate through PP-independent routes. With regards to this, on occasion we observed that listeriae were absent from the PP of mice inoculated intragastrically with L. monocytogenes, but were present in the mesenteric lymph nodes of these same mice. These observations suggested that PP were not necessary for listerial translocation from the intestinal lumen. Two experimental approaches were used to determine whether luminal listeriae could indeed infect the host through PP-independent routes. First, since it is known that: 1) following the intragastric inoculation of L. monocytogenes, listeriae rapidly transit the length of the gastrointestinal tract and reside in the colonic lumen for up to a week, 2) the colon lacks PP, and 3) the descending colon and rectum are drained exclusively by the caudal lymph node (CLN), it was determined whether colonic listeriae could access the CLN. Inoculation of listeriae into the rectum of mice resulted in the infection of the CLN which indicated that PP were not required for listerial translocation. Second, since germfree SCID mice lack PP, it was determined whether listeriae could translocate from the intestinal lumen and infect these immunoincompetent mice. Shortly after the intragastric inoculation of L. monocytogenes into germfree SCID mice, listeriae were found in the mesenteries, livers and spleens. These results also indicate that PP are not required for listerial translocation from the intestinal lumen. One possible route of translocation from the intestinal lumen might occur by listeriae entering enterocytes. Results were obtained showing that listeriae were capable of entering cultured mouse small intestine enterocytes. Internalized listeriae were observed to multiply and spread intracellularly between enterocytes.

Roles for tumor necrosis factor and gamma interferon in resistance to enteric listeriosis.

Listeria monocytogenes normally infects the host by translocating from the intestinal lumen. Experiments were carried out to determine if, when, and where tumor necrosis factor (TNF) and gamma interferon (IFN-gamma) function in antibacterial resistance during enteric listeriosis. Groups of normal mice and severe combined immunodeficient (SCID) mice were injected with neutralizing monoclonal antibodies (MAb) specific for each cytokine and then inoculated intragastrically with L. monocytogenes. The course of infection was monitored by enumerating listeriae in gut-associated lymphoid tissues, livers, and spleens. By the third day of infection, bacterial numbers in infected tissues and organs were greatly exacerbated in all mice treated with anti-TNF MAb, whereas bacterial numbers in the organs of mice treated with anti-IFN-gamma MAb did not differ from those present in the respective organs of control mice. However, by the fifth day of infection, bacterial numbers in the organs of anti-IFN-gamma MAb-treated normal mice and SCID mice were much greater than in the corresponding organs of control mice. Experiments with Listeria-immune mice revealed that TNF and IFN-gamma are involved in the expression of anti-Listeria memory immunity; however, it was also found that the anti-IFN-gamma MAb was relatively ineffective in inhibiting the expression of anti-Listeria immunity, whereas a polyclonal anti-IFN-gamma was quite effective.

Role of tumor necrosis factor and gamma interferon in acquired resistance to Cryptococcus neoformans in the central nervous system of mice.

Although naive C.B-17 and BALB/cBy mice die of meningoencephalitis within 5 weeks of intravenous infection with an opportunistic strain of Cryptococcus neoformans, immunized mice express an acquired, CD4+ T-cell-dependent immunity and survive an intravenous infection. Infusion of lymphocytes from immune mice into severe combined immunodeficiency (SCID) mice renders these mice more resistant to cryptococcal brain infection than uninfused controls. We have investigated the role of gamma interferon (IFN-gamma) and tumor necrosis factor (TNF) in acquired resistance to C. neoformans. Neutralization of either IFN-gamma or TNF impaired resistance of immune BALB/cBy or C.B-17 mice to cryptococci. At 10 days postinfection, there were approximately 10 times as many yeast cells in the brains of mice treated with either anticytokine antibody as in the brains of mice treated with control antibody. Simultaneous neutralization of IFN-gamma and TNF further exacerbated infection. Neutralization of IFN-gamma or TNF also impaired resistance in immune lymphocyte-infused SCID mice, resulting in significantly higher yeast burdens in brains of cytokine-neutralized mice than in brains of controls. Concurrent neutralization of IFN-gamma and TNF rendered SCID recipients of immune cells equivalent to uninfused SCID mice with respect both to brain yeast burdens at 10 days and to survival. Anti-TNF treatment alone also curtailed survival. Histological examination of the brains of cytokine-neutralized mice revealed deficiencies in ability to focus inflammatory cells at brain lesions. These data demonstrate that both IFN-gamma and TNF are important mediators of acquired resistance to cryptococcal meningoencephalitis.

Gamma interferon-dependent temporary resistance to acute Toxoplasma gondii infection independent of CD4+ or CD8+ lymphocytes.

It is well established that resistance to acute primary Toxoplasma gondii infection is mediated by a gamma interferon (IFN-gamma)-dependent mechanism. The present in vivo experiments were undertaken to investigate the cellular basis for this resistance. We show here that immunocompetent T. gondii-infected C57BL/6 (B6) mice treated with anti-IFN-gamma or with anti-Thy-1 or anti-asialo-GM1 antibodies die sooner than infected mice treated with antibodies that deplete both CD4+ and CD8+ T lymphocytes. Thy-1+ CD4- CD8- cells accumulated in the peritoneal cavities of B6 mice during the early stages of an intraperitoneal infection but did not accumulate in sham-infected control mice, and substantial numbers of Thy-1+ CD4- CD8- cells were recovered from the peritoneal cavities of infected B6 mice treated with antibodies that depleted CD4+ and CD8+ lymphocytes. Depletion of Thy-1+ cells reduced IFN-gamma to undetectable levels, whereas depletion of CD4+ and CD8+ cells did not reduce IFN-gamma levels. Thus T. gondii infection in immunocompetent B6 mice elicits Thy-1+ CD4- CD8- cells which either produce protective IFN-gamma themselves or control its production by other cells. It is likely that the function of these Thy-1+ CD4- CD8- cells is to control T. gondii tachyzoites during the early stages of primary infection before specific CD4(+)- and/or CD8(+)-dependent immunity develops.

Mice that lack the interferon-gamma receptor have profoundly altered responses to infection with Bacillus Calmette-Guérin and subsequent challenge with lipopolysaccharide.

Mice with a targeted disruption of the interferon gamma receptor gene (IFN-gamma R0/0 mice) and control wild-type mice were inoculated with the Bacillus Calmette-Guérin (BCG) strain of Mycobacterium bovis. BCG infection was not lethal for wild-type mice whereas all IFN-gamma R0/0 mice died approximately 7-9 wk after inoculation. Histological examination at 2 and 6 wk after BCG inoculation showed that livers of IFN-gamma R0/0 mice had higher numbers of acid-fast bacteria than wild-type mice, especially at 6 wk. In parallel, the livers of IFN-gamma R0/0 mice showed a reduction in the formation of characteristic granulomas at 2 wk after inoculation. Injection of lipopolysaccharide (LPS) 2 wk after BCG inoculation was significantly less lethal for IFN-gamma R0/0 mice than for wild-type mice. Reduced lethality of LPS correlated with a drastically reduced production of tumor necrosis factor alpha (TNF-alpha) in the IFN-gamma R0/0 mice. Interleukin 1 alpha (IL-1 alpha) and IL-6 levels in the serum were also significantly reduced in the IFN-gamma R0/0 mice after BCG infection and LPS challenge. The greatly reduced capacity of BCG-infected IFN-gamma R0/0 mice to produce TNF-alpha may be an important factor in their inability to resist BCG infection. These results show that the presence of a functional IFN-gamma receptor is essential for the recovery of mice from BCG infection, and that IFN-gamma is a key element in the complex process whereby BCG infection leads to the sensitization to endotoxin.

Gamma interferon functions in resistance to Cryptosporidium parvum infection in severe combined immunodeficient mice.

Severe combined immunodeficient (SCID) adult mice are relatively resistant to Cryptosporidium parvum infection, even though they are deficient in both T- and B-cell function. The requirement for gamma interferon (IFN-gamma) in this resistance was examined by treatment of these mice with monoclonal antibody to IFN-gamma. SCID mice injected intraperitoneally with monoclonal anti-IFN-gamma 4 h before and three times weekly after challenge with C. parvum had heavy intestinal infections 3 weeks postchallenge. SCID mice similarly injected with irrelevant antibody were not infected. Furthermore, SCID mice receiving a single injection of anti-IFN-gamma either 2 h before or 18 h after challenge were also susceptible to infection. Although IFN-gamma was not detected in SCID mouse intestinal samples, it was found in the supernatant of SCID mouse splenocyte cultures after stimulation with C. parvum antigens. On the other hand, SCID mice receiving multiple injections of antibodies against tumor necrosis factor remained resistant to infection. These data indicate that the resistance of SCID mice to C. parvum infection is IFN-gamma dependent, whereas tumor necrosis factor appears not to play a significant role.

Listeria monocytogenes-induced interferon-gamma primes the host for production of tumor necrosis factor and interferon-alpha/beta.

Mice acquired an enhanced capacity for the production of tumor necrosis factor (TNF) and the interferons (IFN)-alpha, and -beta shortly after intravenous injection of viable Listeria monocytogenes. By the end of the first day of a sublethal infection, mice were primed to produce 100-1000 times more endotoxin-induced serum TNF than is produced by normal mice. Acquisition of the augmented capacity for TNF production was due to L. monocytogenes-induced IFN-gamma. IFN-gamma also primed infected mice for IFN-alpha/beta production. However, in addition to IFN-gamma, other L. monocytogenes-induced mechanisms endowed the host with an enhanced potential for the production of IFN-alpha/beta. Antibody-mediated depletion of various cell types in vivo revealed that CD8+ cells and NK cells are required for the production of L. monocytogenes-induced IFN-gamma during the first day of listeriosis.

Endotoxin-induced tumor necrosis factor alpha synthesis in murine embryo fibroblasts.

Murine embryo fibroblasts (MEF) were found to secrete tumor necrosis factor (TNF) in response to stimulation with endotoxin. Endotoxin-induced TNF production by MEF was inhibited by cycloheximide. However, reversal of the effect of this inhibitor on protein synthesis results in TNF being secreted in amounts equivalent to those produced by endotoxin-induced MEF not treated with cycloheximide. Actinomycin D treatment of MEF blocked the production of endotoxin-induced TNF. Maximal production of TNF required MEF gene transcription during the first 6 h of incubation with endotoxin. To determine whether endotoxin-induced TNF alpha (TNF-alpha) and/or TNF beta were produced by MEF, cDNA was synthesized from the total RNA isolated from endotoxin-induced MEF and amplified by the polymerase chain reaction in the presence of oligonucleotide primers specific for each cytokine. On the basis of the polymerase chain reaction analysis, it was determined that TNF-alpha mRNA levels were increased in endotoxin-induced MEF. Thus, production of TNF-alpha by fibroblasts in response to the endotoxin component of bacterial cell walls is likely to contribute to the expression of TNF-mediated effects occurring in fibroblast-rich tissues infected with gram-negative bacteria.

Interleukin-6 production in a murine model of Pneumocystis carinii pneumonia: relation to resistance and inflammatory response.

The production of interleukin-6 (IL-6) and its possible relationship to host resistance and inflammatory response to Pneumocystis carinii infection were examined in mice with severe combined immunodeficiency (SCID mice). IL-6 activity was detected in the serum and lungs of P. carinii-infected mice but not in mice free of P. carinii. Moreover, the IL-6 levels in P. carinii-infected mice increased markedly after spleen cell reconstitution but then decreased to an undetectable level after the clearance of P. carinii. However, neutralization of IL-6 activity in spleen cell-reconstituted SCID mice by treatment with anti-IL-6 immunoglobulin G (IgG) resulted in no significant effect on the clearance of P. carinii (P > 0.05). Both the serum and lungs of treated mice contained an excess amount of anti-IL-6 IgG and lacked detectable IL-6. These results suggest that failure to inhibit the P. carinii clearance by anti-IL-6 treatment was not due to insufficient administration of antibody or incomplete neutralization of IL-6 activity. However, compared with mice receiving rat control IgG, mice treated with anti-IL-6 IgG had significantly higher numbers of neutrophils and lymphocytes (particularly CD8+ cells) in the lung lavage fluids (P < 0.05 for both) at day 19 after reconstitution. In addition, the levels of both total IgG (P < 0.001) and P. carinii-specific antibodies (P < 0.05) in the serum of mice treated with anti-IL-6 were significantly higher than those in control mice. These results indicate that although P. carinii infection causes both local and systemic production of IL-6 in SCID mice, IL-6 does not appear to play a crucial role in the clearance of P. carinii. However, it appears that during resolution of P. carinii pneumonia, IL-6 plays a role in the regulation of pulmonary inflammation and antibody responses.

Interleukin 1: an important mediator of host resistance against Pneumocystis carinii.

The importance of endogenous interleukin 1 (IL-1) in resistance to Pneumocystis carinii infection was examined in a SCID mouse model. Naturally acquired pulmonary infection of P. carinii in SCID mice was completely cleared by reconstitution of the infected mice with immunocompetent spleen cells. IL-1 activity in the lung homogenate supernatant of these mice increased significantly after reconstitution and returned to baseline level after the clearance of P. carinii. Treatment of reconstituted SCID mice with 35F5, a monoclonal antibody against murine type I IL-1R almost completely inhibited the clearance of P. carinii. In contrast, treatment with control rat immunoglobulin G had no detectable effect. Further study revealed that for the complete clearance of P. carinii, IL-1 must be present at the early stage of immune responses induced by reconstitution, since clearance could be blocked by a single injection of 35F5 into SCID mice at 2 d, but not at either 8 or 13 d postreconstitution. Furthermore, pulmonary recruitment of neutrophils, macrophages, and lymphocytes was significantly inhibited in mice that received 35F5 treatment. These findings strongly suggest that, in reconstituted SCID mice, endogenous IL-1 is important in host resistance to P. carinii infection and that IL-1 may function early in the host response possibly by the recruitment of inflammatory cells into the lungs.

Importance of endogenous tumor necrosis factor alpha and gamma interferon in host resistance against Pneumocystis carinii infection.

C.B-17 scid/scid (SCID) mice that have acquired natural pulmonary infection with Pneumocystis carinii clear these organisms by 19 days after reconstitution with spleen cells from immunocompetent mice and therefore serve as a model for studying the pathogenesis of and immunity to P. carinii pneumonia. The present study examined the importance of endogenous tumor necrosis factor alpha (TNF-alpha) and gamma interferon (IFN-gamma) in the clearance of P. carinii by treatment of reconstituted SCID mice with anti-TNF-alpha and anti-IFN-gamma immunoglobulin G (IgG). Treatment of reconstituted mice with monospecific rabbit anti-TNF-alpha IgG almost completely inhibited the clearance of P. carinii from the lungs. In contrast, treatment with either anti-IFN-gamma antibody (polyclonal or monoclonal) or control IgG had no detectable effect on the clearance of P. carinii. The importance of endogenous TNF-alpha in the clearance of P. carinii was further supported by the finding of TNF-alpha but not IFN-gamma in lung homogenate supernatants from reconstituted SCID mice. Further study revealed that for the complete clearance of P. carinii, TNF-alpha must be present at the early stage of reconstitution, since clearance could be blocked by a single injection of anti-TNF-alpha IgG into SCID mice at day 0 but not at day 6 and/or day 12 after reconstitution. These results strongly suggest that, in reconstituted SCID mice, endogenous TNF-alpha is important in host resistance against P. carinii infection, whereas IFN-gamma appears not to play a significant role.

Type I IL-1 receptor blockade exacerbates murine listeriosis.

It was found that IL-1 is produced in livers and spleens of mice shortly after the i.v. injection of a sublethal or lethal Listeria monocytogenes inoculum. In sublethally infected mice, IL-1 was present in infected livers and spleens by the end of the first day of infection. Thereafter, the amounts of IL-1 in these organs increased and decreased in concordance with bacterial numbers. IL-1 was not present in the peripheral circulation of mice during sublethal listeriosis, but was present in the blood late in lethal infection. Evidence showing that IL-1 plays a role in antibacterial resistance early in listeriosis was obtained through the use of 35F5 mAb that binds to the murine type I IL-1R and functions to block IL-1 alpha and IL-1 beta actions. Blockade of the type I IL-1R by the 35F5 mAb results in greatly enhanced bacterial growth in the livers and spleens of mice that had received a sublethal Listeria inoculum. Consistent with the exacerbation of listeriosis caused by 35F5 mAb, but in contrast to the effect of 35F5 mAb in other murine models, 35F5 mAb-treated mice exhibit markedly elevated levels of IL-6 in their circulation and infected organs.

A role for tumor necrosis factor in poly(I:C)-induced hemorrhagic necrosis and T-cell-dependent regression of a murine sarcoma.

It was found that intravenous injection of the synthetic double-stranded ribonucleic acid, polyinosinic-polycytidylic acid [poly(I:C)], which is a well-studied interferon (IFN)-inducing agent, can result in extensive hemorrhagic necrosis of the center of an established murine sarcoma and in subsequent complete regression of the surviving rim of the tumor. The poly(I:C)-induced intratumor hemorrhagic reaction was associated with production of appreciable quantities of tumor necrosis factor (TNF) by the host. Neutralization of TNF by treatment with anti-rTNF immunoglobulin G (IgG) caused substantial inhibition of hemorrhagic necrosis and prevented tumor regression from proceeding. Tumor regression was prevented in all mice by depleting them of CD8+ T cells 1 day before poly(I:C) was given. Taken as a whole, the results indicate that the antitumor action of poly(I:C), like that of endotoxin, is based on its capacity to induce the host to make enough TNF to cause a hemorrhagic reaction extensive enough to reduce the tumor burden to a size capable of being dealt with by an underlying host antitumor immune response.

Tumor necrosis factor-independent IL-6 production during murine listeriosis.

We report that TNF, IL-6, and IFN-alpha/beta are produced by mice during either sublethal or lethal Listeria monocytogenes infections. The quantities of these cytokines in infected spleens increase and decrease in concordance with bacterial numbers in these organs. While all of these cytokines were present in Listeria-infected spleens, only IL-6 and IFN-alpha/beta were found in the peripheral circulation. Inasmuch as TNF has been reported to be responsible for the production of IL-6 in vivo following the inoculation of a lethal dose of the Gram-negative bacterium, Escherichia coli (Fong et al., 1989. J. Exp. Med. 170: 1627), experiments were undertaken to determine whether IL-6 production elicited by the Gram-positive bacterium, L. monocytogenes, was also TNF-dependent. It was found that the passive immunization of mice with neutralizing antibodies specific for TNF shortly before i.v. injection of a lethal or sublethal Listeria inoculum resulted in the complete neutralization of endogenously produced TNF, and in the progressive multiplication of bacteria in infected organs. It was also found that the anti-TNF IgG treatment resulted in a progressive increase in the amounts of Listeria-induced IL-6 present in spleen and blood, until the death of the host. These findings indicate that Listeria-induced IL-6 production in mice occurs primarily through a TNF-independent pathway, and correlates directly with the severity of the infection.