Microbial Pathogenesis 1992 ; 12 : 3 33-341
Local induction of tumor necrosis factor as a molecular mechanism of mucosal damage by gonococci Zell A . McGee,' * Christopher M . Clemens,' Robert L . Jensen,' John J . Klein,' Lee R . Barley' and Gary L . Gorby2 'Center for Infectious Diseases, Diagnostic Microbiology, and Immunology, University of Utah School of Medicine, Salt Lake City, Utah 84132, U .S.A . and 2 Omaha V.A . Medical Center and Department of Medical Microbiology, Creighton University School of Medicine, Omaha, Nebraska, U.S .A .
(Received October 29, 1991 ; accepted in revised form December 14, 1991)
McGee, Z. A . (Center for Infectious Diseases, Diagnostic Microbiology, and Immunology, University of Utah School of Medicine, Salt Lake City, Utah 84132, U .S .A .), C . M . Clemens, R . L . Jensen, J . J . Klein, L . R . Barley and G . L. Gorby. Local induction of tumor necrosis factor as a molecular mechanism of mucosal damage by gonococci . Microbial Pathogenesis 1992 ; 12 :333-341 . Tumor necrosis factor (TNF) is an endogenously produced cytokine that plays a critical role in mediating septic shock and multi-organ failure, but previous studies of the role TNF in disease have not examined its role in mucosal disease processes . In an experimental model of acute gonococcal salpingitis, gonococcal infection of human fallopian tube mucosa resulted in increased mucosal production of TNF . Recombinant human TNF-a damaged fallopian tube mucosa in a dose-response manner and produced epithelial damage with the same ultrastructural features as those observed in gonococcal infection . Blocking production of TNF during gonococcal infection diminished the extent of damage to fallopian tube mucosa . In addition to mediating systemic disease, such as septic shock, TNF is also produced locally, and can play a critical role in mediating mucosal disease processes, such as acute gonococcal salpingitis . Key words : Neisseria gonorrhoeae ; salpingitis ; tumor necrosis factor ; cytokines ; pathogenic mechanisms .
Introduction Gonococcal infection of the fallopian tubes (acute gonococcal salpingitis) is responsible for a majority of the economic cost, threat to life (from rupture of ectopic pregnancy) and infertility associated with gonococcal infections ;' , 2 however, the molecular mechanisms by which gonococci damage the fallopian tube mucosa have not hitherto been determined . Because the species specificity of gonococcal infections precludes the use of experimental animals to study gonococcal molecular mechanisms of pathogenicity, 3 we have studied these mechanisms using an experimental model of gonococcal salpingitis which employs human fallopian tubes in organ culture . 4 '5 The pathologic and pathogenetic features of gonococcal infection in the model s closely parallel those of naturally occurring gonococcal infections in men s and women,' as well as those of experimental gonococcal infections in men .' The fallopian tube mucosal epithelium is comprised primarily of ciliated and non" Author to whom requests for reprints should be addressed . 0882-4010/92/050333+09 $03 .00/0
© 1992 Academic Press Limited
334
Z . A . McGee et al.
Fig . 1 . Effect of gonococcal infection or rHuTNF-a on the ultrastructure of human fallopian tube mucosa in organ culture . (A) untreated control (normal) human fallopian tube mucosa (22 h after the organ cultures were established) . Note ciliated and nonciliated cells ; (B) human fallopian tube mucosa 22 h after infection with the same clone of N . gonorrhoeae, strain 2686, used in the experiments shown in Figs 2 and 4 . Note : (1) two sloughing ciliated cells (center) with an apparently full complement of cilia ; and (2) gonococci attached almost exclusively to the microvilli of nonciliated cells, but not attached to sloughing ciliated cells ; (C) human fallopian tube mucosa exposed to 50 µg/ml rHuTNF-a for 22 h . Note ciliated cell sloughing (center) in a manner analogous to those in gonococcus-infected mucosa (B) . Scale bar, 5µm .
ciliated cells' (Fig . 1A) . We have previously shown that gonococcal infection of human fallopian tube mucosa in organ culture results in damage to the epithelium that is characterized by dysfunction and sloughing of the ciliated cells, which begins approximately 18 h post-infection s (Fig . 113) . The gonococcal outer membrane component, lipopolysaccharide (LPS/endotoxin), which is elaborated by gonococci
335
Local induction of tumor necrosis factor as a molecular mechanism
during the course of infection, also causes species-specific sloughing of ciliated cells in this model ."' Peptidoglycan monomers, another type of molecule that gonococci elaborate into their milieu, also damage the fallopian tube mucosa ." The observation that gonococcal LPS-induced structural damage to the fallopian tube mucosa s was identical in appearance to peptidoglycan-induced damage" suggested that the damage might have a common molecular mechanism . Both LPS and peptidoglycan induce production of TNF,' 1,13 which can reproduce multisystem failure and septic shock in experimental animals ." Therefore, we developed and tested the hypothesis that rather than damaging the fallopian tube mucosa directly, gonococci induce the fallopian tube mucosa to produce TNF, which in turn damages the fallopian tube mucosal epithelium .
Results Gonococcal induction of mucosal TNF production Assays of uninfected fallopian tube mucosal organ cultures for TNF indicated a low, but detectable, level of TNF production, which varied little during 30 h of observation (Fig . 2) . In contrast, paired, gonococcus-infected human fallopian tube mucosa showed a significant increase in TNF production following gonococcal infection . The difference between the infected and uninfected (control) groups across all time points was significant (P < 0 .0001) by the Wilcoxon signed-rank test (two-tailed), paired t test (two-tailed) and regression analysis . The observed increase in TNF production occurred prior to, and during, the onset of mucosal damage, which was characterized by dysfunction and sloughing of ciliated cells (Fig . 1 B) . In several experiments the L929-cell-cytotoxicity of material that tested positive in the TNF bioassay of the gonococcus-infected fallopian tube organ culture homogenates was neutralized by treatment with rabbit anti-TNF-a antibody (data not shown) . The L-929 cell assay used to quantitate TNF measures only biologically active TNF; it was more sensitive than ELISA in detecting and quantitating TNF in the low range of concentrations (30 U/ml, respectively) . The specificity of the L-929 cell TNF assay was indicated by the observation that the cytotoxity of gonococcal-infected organ culture homogenates was neutralized by treatment with anti-TNF-a antibody as described above . Although the site of TNF production was not assessed directly in these studies, the concentrations of TNF in the medium of gonococcus-infected organ cultures were compared to the TNF concentrations in the supernatant fluid of homogenized, gonococcus-infected organ cultures . The TNF concentrations were regularly higher if the tissue in a gonococcus-infected organ culture was homogenized in its own organ culture medium than if the medium alone was assayed . TNF damage of the mucosa When fallopian tube mucosa was exposed to human recombinant TNF-a (rHuTNF-a) in a range of concentrations (Fig . 3), tissues treated with rHuTNF-a had significantly lower percent of the periphery fallopian tube pieces with ciliary activity (PPCA ; fewer ciliated cells remaining) compared to untreated control tissues [P < 0 .0001 ; analysis of variance (ANOVA)] . Such differences were observed with as little as 5 U/ml rHuTNF-a (P= 0 .0025 ; t test) and with all other concentrations of rHuTNF-a tested (P < 0 .0001 ; t test, even with Bonferroni correction for multiple comparisons) . The extent of damage increased as the dose of TNF increased . Thus, rHuTNF-a damaged the mucosa in a dose-response manner and at the highest concentration of rHuTNF-a used, the organ culture fluid contained a maximum of 1 .8 ng of Eschericia co/i LPS/ml of organ culture fluid . However, in dose-response experiments with up to 100 ng of E. coli LPS/ml, no concentration of E. coli LPS damaged the mucosa (data not shown) . The ultrastructural damage in rHuTNF-a-treated tissues, which featured sloughing of ciliated cells (Fig . 1C) was indistinguishable from that in gonococcus-infected tissues (Fig . 113) ; except that gonococci could be observed in the latter tissues . Blocking TNF production Whereas in gonococcus-infected, untreated tissues the mean TNF concentration was 11 .9 U/ml 22 h after infection, at the same time period in gonococcus-infected tissues
80 70 60 50 40 30 20 10 0 0
5
50
500 500050,000 5,000,000 500,000 TNF concentration (U/mi)
Fig . 3 . Effect of increasing concentrations of recombinant human TNF-a (rHuTNF-a, Genentech) on damage to human fallopian tube epithelium . The height of the bars represents the mean results of three or more experiments+SEM . Lower PPCA values indicate greater sloughing of ciliated cells .
Local induction of tumor necrosis factor as a molecular mechanism
337
treated with dexamethasone, a potent inhibitor of both TNF mRNA transcription and 1,15 the TNF concentration was only 0 .46 U/ml . The PPCA in infected translation," mucosal epithelium was significantly less than that in control (uninfected) epithelial tissues (P 0 .05 with Bonferroni correction for multiple comparisons) . There was no statistically significant difference between the PPCA of uninfected, untreated (control) mucosal tissue and of gonococcus-infected, but dexamethasonetreated mucosal tissue. Thus, the decrease in TNF production which occurred with dexamethasone treatment was associated with a statistically significant decrease in mucosal damage relative to infected tissues without dexamethasone . Discussion The data presented here support the hypothesis that gonococci, rather than damaging the human fallopian tube mucosa directly, elaborate one or more products, such as LPS 9 or peptidoglycan, 11 that may elicit the production of TNF by mucosal cells . The concentrations of TNF produced by the mucosa in response to gonococcal infection were equal to or greater than the lowest concentration of rHuTNF-a which produced significant damage to the mucosa . Because cytokine action is often the result of more than one cytokine and usually the result of a variety of agonists and antagonists, 16-18 the locally-produced TNF is likely to act either alone, or in concert with other cytokines such as interleukin-1 (IL-1), 16 .1719 IL-6, 17 and perhaps IL-4, 18 to cause tissue damage . However, the observations that TNF alone damaged the mucosa in relatively low
100
Tissue control
Infected tissue
Infected Tissue +I /.LM dex . +I µM dex .
Experimental group Fig . 4 . Effect of blocking TNF production with 1 pM dexamethasone on damage to the fallopian tube epithelium during gonococcal infection . The height of the bars represents the mean results of two or more experiments+SEM . The mean TNF concentration in homogenates of gonococcus-infected organ cultures not treated with dexamethasone (second column) was 11 .9 U/ml, whereas the mean TNF concentration of identically infected organ cultures treated with dexamethasone (third column) was 0 .46 U/ml . The presence of dexamethasone and decreased TNF production (column three) was associated with significantly diminished damage in gonoccus-infected organ cultures (third column vs second column, P < 0 .001) .
338
Z . A . McGee et al.
concentrations (Fig . 3), which were comparable to the concentrations produced by the mucosa (Fig . 2), and that pharmacologically decreasing TNF production decreased damage to the mucosa (Fig . 4), suggests that TNF alone was necessary and sufficient to cause the mucosal damage observed ; however, TNF induction of IL-6 and other cytokines" cannot be excluded . Whereas immunologic assays of human TNF elicited by an infectious process may give false high values because they measure biologically inactive as well as biologically active TNF, 20 the assay used to quantitate TNF in the studies presented here was the standard L-929 cell bioassay for TNF, which measures only biologically active TNF . The observation that the cytotoxicity of the gonococcus-infected organ culture homogenates was neutralized by anti-TNF-a antibody indicates that the assay was specific and TNF-a was the major or only actively cytotoxic moiety present . The site of mucosal production of TNF was not directly determined in these studies ; however, the finding that homogenates of gonococcus-infected organ cultures regularly had higher concentrations of TNF than the same organ culture medium without tissue homogenization suggests that the TNF may be produced by cells yet to be identified in the subepithelial tissues . If TNF is produced in subepithelial tissues, efforts to block its action with antibody added to the organ culture medium may not be successful . Having shown that gonococcal infection induced the production of TNF by human fallopian tube mucosa, it was important to determine if human TNF could damage the mucosa . Whereas there was substantial damage to the mucosa by rHuTNF-a, there was a very low concentration of LPS in the E. co/i-produced rHuTNF-a preparation (up to 1 .8 ng/ml in the highest concentration of rHuTNF-a used) . However, in doseresponse experiments, up to 100 ng/ml of E. co/i LPS failed to damage the fallopian tube mucosa; thus, the damaging effect of the rHuTNF-a preparation could be attributed to the rHuTNF-a per se . Previous studies in our laboratory demonstrated that gonococcal LPS in concentrations as low as 0 .15 tg/ml reproducibly damaged human fallopian tube mucosa .' Furthermore, the damage, like gonococcal infections per se, was highly species- specific ; 10 the genital mucosa of humans, but not that of lower animals, was susceptible ." Thus, the observation that E. coil LPS did not damage the fallopian tube mucosa suggests that human genital mucosa may respond differently to the LPS from different bacterial species . Such differences in the biologic response of human cells to LPS from different bacterial species has been well documented by Greisman and Hornick,21 but the basis for the differences in response is unclear . Because gonococci elaborate LPS and peptidoglycan into their milieu 9• 1 1 and these compounds damage fallopian tube mucosa,' , " the finding that dexamethasone, which blocked production of TNF, also blocked damage to the mucosa, suggests that gonococcal products such as LPS and peptidoglycan monomers are not intrinsically damaging, but depend on eliciting TNF production for the damage they cause . The action of the ciliated cells in moving the ovum from the ovary down to the ampullary portion of the fallopian tube where it can be fertilized, 22 is critical for conception, and movement of the fertilized ovum onto the uterus for implantation is important in avoiding ectopic (tubal) pregnancy . Therefore, the pathophysiologic process described here, by causing sloughing of ciliated cells, may be responsible for the infertility and ectopic pregnancy that frequently follow gonococcal salpingitis .' •2 Although there is compelling evidence that systemic production and action of TNF is a molecular mechanism of bacterial sepsis, shock and multiorgan failure" and local production of TNF in the central nervous system may be a mechanism of nerve damage in meningitis, 23,24 we are not aware of any other evidence that local mucosal production of TNF is a molecular mechanism of bacterial pathogenicity . However, the current
Local induction of tumor necrosis factor as a molecular mechanism
339
studies demonstrate the importance of the local production of TNF in an experimental model previously shown to represent the events in naturally occurring and experimental gonococcal infections of human genital mucosa ." The molecules, such as those of pili,' that bacteria use to cause infection (establishment of the organism in the host) are different from those molecules, such as LPS, 9.14 that bacteria use to cause disease (dysfunction and damage of tissues) . The findings reported here suggest that rather than focusing strategies to develop bacterial vaccines only on blocking infection, or enhancing elimination of bacteria by phagocytosis, attention should also be directed to blocking the molecular mechanisms, such as those'described here, whereby bacteria cause disease .
Materials and methods Microorganisms . A colony type 1 (piliated), Tr/Op 5 clone of N. gonorrhoeae, strain 2686, was used in all studies involving gonococcal infection of organ cultures . The organisms were grown and used to prepare organ culture inocula as previously described,' so that after inoculation the concentration in the organ culture medium was 2 x 10 5 cfu/ml . Human fallopian tube organ cultures. Human fallopian tubes were obtained from unpregnant, pre-menopausal women with no history of salpingitis or other gonococcal infection who were undergoing hysterectomy and bilateral salpingooophorectomy for surgical indications (usually fibromas of the uterus) . After the tubes were freed of adventitial tissues, they were opened longitudinally and mucosal pieces with similar or identical epithelial surface area were cut using a skin biopsy punch with an internal diameter of 2 .5 mm . The mucosal pieces included the epithelium, which is comprised of ciliated and nonciliated cells ; the mucosal pieces also contained sub epithelial tissues, including connective tissue, arteries, veins and lymphatics as well as scattered mononuclear round cells . Such mucosal pieces contain antibody-producing cells, as evidenced by their production of IgA in response to gonococcal infection .21,2' Three pieces were put into each tissue culture grade polystyrene dish as previously described,' , ' except that the dishes were 35x 10 mm and contained 1 ml of Hepes-MEM (minimal Eagle's medium),' which was supplemented to 5% v/v with fetal calf serum (GIBCO) except in studies with rHuTNF-a, when the Hepes-MEM was supplemented to 0 .2% wt/v with bovine serum albumin (Sigma) to help block adsorption of TNF to the dish . In each individual experiment the tissues that comprised the organ cultures in the control and all experimental groups were from the same subject . Quantitation of damage to fallopian tube epithelium. Damage to the epithelium was assessed by determining the percent of the periphery of fallopian tube pieces with ciliary activity (PPCA) remaining at a particular time period relative to the pretreatment value for the same segments of the epithelium as described previously .` PPCA is an indirect measure of the number of functional ciliated cells remaining in the epithelium .' Thus, the tissues were assessed prior to infection or treatment with rHuTNF-a and 22 h later . All assessments of PPCA were performed without knowledge of the treatment group of the tissues examined . Enumeration of gonococci . Gonococci in the inoculum and in organ cultures were enumerated
as previously described .' Quantitation of TNF . All three mucosal pieces from each organ culture were homogenized in their own organ culture medium using a 'tissuemizer' (Tekmar, Cincinnati, OH), and the homogenates were filtered separately through 0 .45 ym pore size membrane filters prior to TNF quantitation . TNF was quantitated using the L-929 cell assay described by Zacharchuk et al- 21 which was modified by incubating the L-929 cells for 18 h prior to addition of actinomycin D, TNF standards and samples to be tested . Early in the studies of TNF induction by gonococcal infection, the TNF concentrations in organ culture homogenates were quantitated for purposes of comparison using both the L-929 cell bioassay and by ELISA, using a commercially prepared ELISA kit (Endogen) .
340
Z . A . McGee
et al.
Electron microscopy studies . Tissues to be examined by scanning electron microscopy were prepared and examined as previously described .' Effect of gonococcal infection on production of TNF by fallopian tube mucosal organ cultures . Duplicate human fallopian tube mucosal organ cultures were prepared, and one set was infected with N. gonorrhoeae, whereas the other set served as uninfected control tissues . Immediately following infection (0 time), and at intervals ranging over a 30 h period, the mucosal organ cultures were homogenized and assayed for TNF . In several experiments, the filtrates for TNF bioassay were tested with and without the addition of rabbit anti-TNF-a antibody (Endogen) .
Effect of rHuTNF-a on fallopian tube mucosa . Organ culture media were supplemented to various final concentrations of rHuTNF-a (Genentech) ranging from 5 to 5x 10 6 U/ml . Damage was assessed by determining the PPCA remaining 22 h after addition of rHuTNF-a to the organ culture fluid . In a sub-set of these studies, E. coli LPS (Associates of Cape Cod, Inc .) was suspended in Hepes-MEM using sonification as previously described' and added to fallopian tube organ cultures to final concentrations ranging from 1 to 100,000 pg/ml . The PPCA was determined just prior to addition of the LPS and after incubation of the organ cultures for 22 h following the addition of LPS. Effect of blocking TNF production . Four groups of organ cultures were designated . One group remained uninfected ; the second group was infected with N. gonorrhoeae in the standard manner;' the third group was infected as the second group, but the organ culture medium was supplemented with a blocker of TNF production, dexamethasone, 14,15 (Sigma) to a final concentration of 1 pm just prior to infection ; the fourth group was only treated with 1 µM dexamethasone . Epithelial ciliary activity was assessed 22 h later by determining the PPCA remaining in each of the four groups . After assessment of damage in two of the experiments in this series, the organ cultures were homogenized in their own medium and assays for TNF were performed in the standard manner on filtrates of the homogenates .
These studies were supported by grant AI-26285 from NIAID, NIH, U .S . Department of Health and Human Services, and in part by Public Health Services research grant no . M01-RR00064 from the National Center for Research Resources . We are indebted to Drs Thomas Abbott, John Andrews, Wallace Bryner, George Cannon, Dan Chichester, Philip Clark, Kent Farnsworth, Robert Flinner, Jeanne Gemmell, Jeff Gliedman, Tom Jones, Jeff Labrum, Ron Larkin, Steve Prescott, Alan Rappleye and Kent Rasmussen, without whose help these studies would not have been possible . The first author is especially grateful to Dr Porter Anderson (University of Rochester School of Medicine and Dentistry) whose personal encouragement was important in the genesis and completion of these studies .
References 1 . McGee ZA . Gonococcal pelvic inflammatory disease. In : Holmes KK, Mardh PA, Sparling PF, Wiesner PJ, eds . Sexually transmitted diseases . New York : McGraw-Hill Book Company, 1984; 220-29 . 2 . Cates W Jr, Rolts RT Jr, Aral SO . Sexually transmitted diseases, pelvic inflammatory disease, and infertility : an epidemiologic update . Epidemiol Rev 1990 ; 12 : 199-220 . 3 . Hill JH . Experimental infection with Neisseria gonorrhoeae. II . Animal inoculations . Am J Syph Gonorrhea and Vener Dis 1944 ; 28 : 334-78, 471-510 . 4 . McGee ZA, Johnson AP, Taylor-Robinson D . Human fallopian tubes in organ culture : preparation, maintenance, and quantitation of damage by pathogenic microorganisms . Infect Immun 1976; 13 : 608-18 . 5.
6. 7. 8. 9.
McGee ZA, Johnson AP, Taylor-Robinson D . Pathogenic mechanisms of Neisseria gonorrhoeae: observation on damage to human fallopian tubes in organ culture by gonococci of colony type 1 or type 4 . J Infect Dis 1981 ; 143 : 413-22 . Ward ME, Watt PJ . Adherence of Neisseria gonorrhoeae to urethral mucosal cells : an electronmicroscopic study of human gonorrhea . J Infect Dis 1977 ; 126 : 248-55 . Pollock CE, Harrison LW . Gonococcal Infections. London : Oxford University Press, 1912 ; xi-xiii . Kellogg DS Jr, Peacock WL Jr, Deacon WE, Brown L, Pirkle CI . Neisseria gonorrhoeae. I . Virulence genetically linked to clonal variation . J Bacteriol 1963; 85 : 1274-9 . Gregg CR, Melly MA, Hellergvist CG, Coniglio JG, McGee ZA. Toxic activity of purified lipo-
Local induction of tumor necrosis factor as a molecular mechanism
10 . 11 . 12 . 13 . 14 . 15 . 16 .
17 . 18 .
19 . 20 . 21 . 22 . 23 .
24 . 25 .
26 .
27 .
341
polysaccharide of Neisseria gonorrhoeae for human fallopian tube mucosa . J Infect Dis 1981 ; 143 : 432-9 . Gregg CR, Johnson AP, Taylor-Robinson D, Melly MA, McGee ZA . Host species-specific damage to oviduct mucosa by Neisseria gonorrhoeae Iipopolysaccharide . Infect Immun 1981 ; 34 : 1056-8. Melly MA, McGee ZA, Rosenthal RS . Ability of monomeric peptidoglycan fragments from Neisseria gonorrhoeae to damage human fallopian tube mucosa . J Infect Dis 1984; 149 : 378-86 . Mannel DN, Falk W . Optimal induction of tumor necrosis factor production in human monocytes requires complete S-form lipopolysaccharide . Infect Immun 1989 ; 57 : 1953-8 . Parant M . TNF as a mediator of LPS- and MDP-induced enhancement of mouse non-specific immunity . J Leukocyte Biol 1987 ; 42 : 576 . Beutler B, Cerami A . Cachectin : More than a tumor necrosis factor . New Eng J Med 1987 ; 316 : 37985 . Beutler B, Krochin N, Milsark IW, Luedke C, Cerami A . Control of cachectin (tumor necrosis factor) synthesis : mechanisms of endotoxin resistance . Science 1986 ; 232 : 977-9 . Okusawa S, Gelfand JA, Ikejima T, Connolly RJ, Dinarello CA. Interleukin 1 induces a shock-like state in rabbits: synergism with tumor necrosis factor and the effect of cyclooxygenase inhibition . J Clin Invest 1988; 81 : 1162-72 . Elias JA, Lentz V . IL-1 and tumor necrosis factor synergistically stimulate fibroblast IL-6 production and stabilize IL-6 messenger RNA . J Immunol 1990 ; 145(1) : 161-6 . Hart PH, Gerard FV, Burgess DR, Whitty GA, Piccoli DS, Hamilton JA . Potential antiinflammatory effects of interleukin 4 : suppression of human monocyte tumor necrosis factor x, interleukin 1, and prostaglandin E 2 . Proc Natl Acad Sci USA 1989 ; 86 : 3803-7 . Last-Barney K, Homon CA, Faanes RB, Merluzzi VJ . Synergistic and overlapping activities of tumor necrosis factor-a and IL-1 . J Immunol 1988 ; 141 (2) : 527-30 . Moller B, Mogensen SC, Wendelboe P, Bendtzen K, Petersen CM . Bioactive and inactive forms of tumor necrosis factor-a in spinal fluid from patients with meningitis . J Infect Dis 1991 ; 163 : 886-9 . Greisman SE, Hornick RB . Comparative pyrogenic reactivity of rabbit and man to bacterial endotoxin . Proc Soc Exp Biol Med 1969; 131 : 1154-8 . Blandau RJ . In : Hafez ESE, Blandau RJ, eds . The mammalian oviduct : comparative biology and methodology . Chicago: University of Chicago Press, 1969; 133-9. Waage A, Halstensen A, Shalaby R, Brandtzaeg P, Kierulf P, Espevik T . Local production of tumor necrosis factor alpha, Interleukin 1, and Interleukin 6 in meningococcal meningitis : relation to the inflammatory response . J Exp Med 1989; 170 : 1859-67 . Lebel MH, Freij BJ, Syrogiannopoulos GA et al. Dexamethasone therapy for bacterial meningitis : results of two double-blind, placebo-controlled trials . N Engl J Med 1988, 319 : 964-71 . Gorby GL, Clemens CM, Barley LR, McGee ZA . Effect of human chorionic gonadotropin (hCG) on Neisseria gonorrhoeae invasion of and IgA secretion by human fallopian tube mucosa . Microb Pathogen 1991 ; 10 : 373-84 . Cooper MD, McGee ZA, Mulks MH, Koomey JM, Hindman TL . Studies on attachment and invasion of human fallopian tube mucosa by a gonococcal IgA1 protease deficient mutant and its wild type parent . J Infect Dis 1984 ; 150 : 737-44 . Zacharchuk CM, Drysdale B-E, Mayer MM, Shin HS . Macrophage-mediated cytotoxicity : role of a soluble macrophage cytotoxic factor similar to lymphotoxin and tumor necrosis factor . Proc Natl Acad Sci USA 1983 ; 80 : 6341-5 .
Local induction of tumor necrosis factor as a molecular mechanism of mucosal damage by gonococci Zell A . McGee,' * Christopher M . Clemens,' Robert L . Jensen,' John J . Klein,' Lee R . Barley' and Gary L . Gorby2 'Center for Infectious Diseases, Diagnostic Microbiology, and Immunology, University of Utah School of Medicine, Salt Lake City, Utah 84132, U .S.A . and 2 Omaha V.A . Medical Center and Department of Medical Microbiology, Creighton University School of Medicine, Omaha, Nebraska, U.S .A .
(Received October 29, 1991 ; accepted in revised form December 14, 1991)
McGee, Z. A . (Center for Infectious Diseases, Diagnostic Microbiology, and Immunology, University of Utah School of Medicine, Salt Lake City, Utah 84132, U .S .A .), C . M . Clemens, R . L . Jensen, J . J . Klein, L . R . Barley and G . L. Gorby. Local induction of tumor necrosis factor as a molecular mechanism of mucosal damage by gonococci . Microbial Pathogenesis 1992 ; 12 :333-341 . Tumor necrosis factor (TNF) is an endogenously produced cytokine that plays a critical role in mediating septic shock and multi-organ failure, but previous studies of the role TNF in disease have not examined its role in mucosal disease processes . In an experimental model of acute gonococcal salpingitis, gonococcal infection of human fallopian tube mucosa resulted in increased mucosal production of TNF . Recombinant human TNF-a damaged fallopian tube mucosa in a dose-response manner and produced epithelial damage with the same ultrastructural features as those observed in gonococcal infection . Blocking production of TNF during gonococcal infection diminished the extent of damage to fallopian tube mucosa . In addition to mediating systemic disease, such as septic shock, TNF is also produced locally, and can play a critical role in mediating mucosal disease processes, such as acute gonococcal salpingitis . Key words : Neisseria gonorrhoeae ; salpingitis ; tumor necrosis factor ; cytokines ; pathogenic mechanisms .
Introduction Gonococcal infection of the fallopian tubes (acute gonococcal salpingitis) is responsible for a majority of the economic cost, threat to life (from rupture of ectopic pregnancy) and infertility associated with gonococcal infections ;' , 2 however, the molecular mechanisms by which gonococci damage the fallopian tube mucosa have not hitherto been determined . Because the species specificity of gonococcal infections precludes the use of experimental animals to study gonococcal molecular mechanisms of pathogenicity, 3 we have studied these mechanisms using an experimental model of gonococcal salpingitis which employs human fallopian tubes in organ culture . 4 '5 The pathologic and pathogenetic features of gonococcal infection in the model s closely parallel those of naturally occurring gonococcal infections in men s and women,' as well as those of experimental gonococcal infections in men .' The fallopian tube mucosal epithelium is comprised primarily of ciliated and non" Author to whom requests for reprints should be addressed . 0882-4010/92/050333+09 $03 .00/0
© 1992 Academic Press Limited
334
Z . A . McGee et al.
Fig . 1 . Effect of gonococcal infection or rHuTNF-a on the ultrastructure of human fallopian tube mucosa in organ culture . (A) untreated control (normal) human fallopian tube mucosa (22 h after the organ cultures were established) . Note ciliated and nonciliated cells ; (B) human fallopian tube mucosa 22 h after infection with the same clone of N . gonorrhoeae, strain 2686, used in the experiments shown in Figs 2 and 4 . Note : (1) two sloughing ciliated cells (center) with an apparently full complement of cilia ; and (2) gonococci attached almost exclusively to the microvilli of nonciliated cells, but not attached to sloughing ciliated cells ; (C) human fallopian tube mucosa exposed to 50 µg/ml rHuTNF-a for 22 h . Note ciliated cell sloughing (center) in a manner analogous to those in gonococcus-infected mucosa (B) . Scale bar, 5µm .
ciliated cells' (Fig . 1A) . We have previously shown that gonococcal infection of human fallopian tube mucosa in organ culture results in damage to the epithelium that is characterized by dysfunction and sloughing of the ciliated cells, which begins approximately 18 h post-infection s (Fig . 113) . The gonococcal outer membrane component, lipopolysaccharide (LPS/endotoxin), which is elaborated by gonococci
335
Local induction of tumor necrosis factor as a molecular mechanism
during the course of infection, also causes species-specific sloughing of ciliated cells in this model ."' Peptidoglycan monomers, another type of molecule that gonococci elaborate into their milieu, also damage the fallopian tube mucosa ." The observation that gonococcal LPS-induced structural damage to the fallopian tube mucosa s was identical in appearance to peptidoglycan-induced damage" suggested that the damage might have a common molecular mechanism . Both LPS and peptidoglycan induce production of TNF,' 1,13 which can reproduce multisystem failure and septic shock in experimental animals ." Therefore, we developed and tested the hypothesis that rather than damaging the fallopian tube mucosa directly, gonococci induce the fallopian tube mucosa to produce TNF, which in turn damages the fallopian tube mucosal epithelium .
Results Gonococcal induction of mucosal TNF production Assays of uninfected fallopian tube mucosal organ cultures for TNF indicated a low, but detectable, level of TNF production, which varied little during 30 h of observation (Fig . 2) . In contrast, paired, gonococcus-infected human fallopian tube mucosa showed a significant increase in TNF production following gonococcal infection . The difference between the infected and uninfected (control) groups across all time points was significant (P < 0 .0001) by the Wilcoxon signed-rank test (two-tailed), paired t test (two-tailed) and regression analysis . The observed increase in TNF production occurred prior to, and during, the onset of mucosal damage, which was characterized by dysfunction and sloughing of ciliated cells (Fig . 1 B) . In several experiments the L929-cell-cytotoxicity of material that tested positive in the TNF bioassay of the gonococcus-infected fallopian tube organ culture homogenates was neutralized by treatment with rabbit anti-TNF-a antibody (data not shown) . The L-929 cell assay used to quantitate TNF measures only biologically active TNF; it was more sensitive than ELISA in detecting and quantitating TNF in the low range of concentrations (30 U/ml, respectively) . The specificity of the L-929 cell TNF assay was indicated by the observation that the cytotoxity of gonococcal-infected organ culture homogenates was neutralized by treatment with anti-TNF-a antibody as described above . Although the site of TNF production was not assessed directly in these studies, the concentrations of TNF in the medium of gonococcus-infected organ cultures were compared to the TNF concentrations in the supernatant fluid of homogenized, gonococcus-infected organ cultures . The TNF concentrations were regularly higher if the tissue in a gonococcus-infected organ culture was homogenized in its own organ culture medium than if the medium alone was assayed . TNF damage of the mucosa When fallopian tube mucosa was exposed to human recombinant TNF-a (rHuTNF-a) in a range of concentrations (Fig . 3), tissues treated with rHuTNF-a had significantly lower percent of the periphery fallopian tube pieces with ciliary activity (PPCA ; fewer ciliated cells remaining) compared to untreated control tissues [P < 0 .0001 ; analysis of variance (ANOVA)] . Such differences were observed with as little as 5 U/ml rHuTNF-a (P= 0 .0025 ; t test) and with all other concentrations of rHuTNF-a tested (P < 0 .0001 ; t test, even with Bonferroni correction for multiple comparisons) . The extent of damage increased as the dose of TNF increased . Thus, rHuTNF-a damaged the mucosa in a dose-response manner and at the highest concentration of rHuTNF-a used, the organ culture fluid contained a maximum of 1 .8 ng of Eschericia co/i LPS/ml of organ culture fluid . However, in dose-response experiments with up to 100 ng of E. coli LPS/ml, no concentration of E. coli LPS damaged the mucosa (data not shown) . The ultrastructural damage in rHuTNF-a-treated tissues, which featured sloughing of ciliated cells (Fig . 1C) was indistinguishable from that in gonococcus-infected tissues (Fig . 113) ; except that gonococci could be observed in the latter tissues . Blocking TNF production Whereas in gonococcus-infected, untreated tissues the mean TNF concentration was 11 .9 U/ml 22 h after infection, at the same time period in gonococcus-infected tissues
80 70 60 50 40 30 20 10 0 0
5
50
500 500050,000 5,000,000 500,000 TNF concentration (U/mi)
Fig . 3 . Effect of increasing concentrations of recombinant human TNF-a (rHuTNF-a, Genentech) on damage to human fallopian tube epithelium . The height of the bars represents the mean results of three or more experiments+SEM . Lower PPCA values indicate greater sloughing of ciliated cells .
Local induction of tumor necrosis factor as a molecular mechanism
337
treated with dexamethasone, a potent inhibitor of both TNF mRNA transcription and 1,15 the TNF concentration was only 0 .46 U/ml . The PPCA in infected translation," mucosal epithelium was significantly less than that in control (uninfected) epithelial tissues (P 0 .05 with Bonferroni correction for multiple comparisons) . There was no statistically significant difference between the PPCA of uninfected, untreated (control) mucosal tissue and of gonococcus-infected, but dexamethasonetreated mucosal tissue. Thus, the decrease in TNF production which occurred with dexamethasone treatment was associated with a statistically significant decrease in mucosal damage relative to infected tissues without dexamethasone . Discussion The data presented here support the hypothesis that gonococci, rather than damaging the human fallopian tube mucosa directly, elaborate one or more products, such as LPS 9 or peptidoglycan, 11 that may elicit the production of TNF by mucosal cells . The concentrations of TNF produced by the mucosa in response to gonococcal infection were equal to or greater than the lowest concentration of rHuTNF-a which produced significant damage to the mucosa . Because cytokine action is often the result of more than one cytokine and usually the result of a variety of agonists and antagonists, 16-18 the locally-produced TNF is likely to act either alone, or in concert with other cytokines such as interleukin-1 (IL-1), 16 .1719 IL-6, 17 and perhaps IL-4, 18 to cause tissue damage . However, the observations that TNF alone damaged the mucosa in relatively low
100
Tissue control
Infected tissue
Infected Tissue +I /.LM dex . +I µM dex .
Experimental group Fig . 4 . Effect of blocking TNF production with 1 pM dexamethasone on damage to the fallopian tube epithelium during gonococcal infection . The height of the bars represents the mean results of two or more experiments+SEM . The mean TNF concentration in homogenates of gonococcus-infected organ cultures not treated with dexamethasone (second column) was 11 .9 U/ml, whereas the mean TNF concentration of identically infected organ cultures treated with dexamethasone (third column) was 0 .46 U/ml . The presence of dexamethasone and decreased TNF production (column three) was associated with significantly diminished damage in gonoccus-infected organ cultures (third column vs second column, P < 0 .001) .
338
Z . A . McGee et al.
concentrations (Fig . 3), which were comparable to the concentrations produced by the mucosa (Fig . 2), and that pharmacologically decreasing TNF production decreased damage to the mucosa (Fig . 4), suggests that TNF alone was necessary and sufficient to cause the mucosal damage observed ; however, TNF induction of IL-6 and other cytokines" cannot be excluded . Whereas immunologic assays of human TNF elicited by an infectious process may give false high values because they measure biologically inactive as well as biologically active TNF, 20 the assay used to quantitate TNF in the studies presented here was the standard L-929 cell bioassay for TNF, which measures only biologically active TNF . The observation that the cytotoxicity of the gonococcus-infected organ culture homogenates was neutralized by anti-TNF-a antibody indicates that the assay was specific and TNF-a was the major or only actively cytotoxic moiety present . The site of mucosal production of TNF was not directly determined in these studies ; however, the finding that homogenates of gonococcus-infected organ cultures regularly had higher concentrations of TNF than the same organ culture medium without tissue homogenization suggests that the TNF may be produced by cells yet to be identified in the subepithelial tissues . If TNF is produced in subepithelial tissues, efforts to block its action with antibody added to the organ culture medium may not be successful . Having shown that gonococcal infection induced the production of TNF by human fallopian tube mucosa, it was important to determine if human TNF could damage the mucosa . Whereas there was substantial damage to the mucosa by rHuTNF-a, there was a very low concentration of LPS in the E. co/i-produced rHuTNF-a preparation (up to 1 .8 ng/ml in the highest concentration of rHuTNF-a used) . However, in doseresponse experiments, up to 100 ng/ml of E. co/i LPS failed to damage the fallopian tube mucosa; thus, the damaging effect of the rHuTNF-a preparation could be attributed to the rHuTNF-a per se . Previous studies in our laboratory demonstrated that gonococcal LPS in concentrations as low as 0 .15 tg/ml reproducibly damaged human fallopian tube mucosa .' Furthermore, the damage, like gonococcal infections per se, was highly species- specific ; 10 the genital mucosa of humans, but not that of lower animals, was susceptible ." Thus, the observation that E. coil LPS did not damage the fallopian tube mucosa suggests that human genital mucosa may respond differently to the LPS from different bacterial species . Such differences in the biologic response of human cells to LPS from different bacterial species has been well documented by Greisman and Hornick,21 but the basis for the differences in response is unclear . Because gonococci elaborate LPS and peptidoglycan into their milieu 9• 1 1 and these compounds damage fallopian tube mucosa,' , " the finding that dexamethasone, which blocked production of TNF, also blocked damage to the mucosa, suggests that gonococcal products such as LPS and peptidoglycan monomers are not intrinsically damaging, but depend on eliciting TNF production for the damage they cause . The action of the ciliated cells in moving the ovum from the ovary down to the ampullary portion of the fallopian tube where it can be fertilized, 22 is critical for conception, and movement of the fertilized ovum onto the uterus for implantation is important in avoiding ectopic (tubal) pregnancy . Therefore, the pathophysiologic process described here, by causing sloughing of ciliated cells, may be responsible for the infertility and ectopic pregnancy that frequently follow gonococcal salpingitis .' •2 Although there is compelling evidence that systemic production and action of TNF is a molecular mechanism of bacterial sepsis, shock and multiorgan failure" and local production of TNF in the central nervous system may be a mechanism of nerve damage in meningitis, 23,24 we are not aware of any other evidence that local mucosal production of TNF is a molecular mechanism of bacterial pathogenicity . However, the current
Local induction of tumor necrosis factor as a molecular mechanism
339
studies demonstrate the importance of the local production of TNF in an experimental model previously shown to represent the events in naturally occurring and experimental gonococcal infections of human genital mucosa ." The molecules, such as those of pili,' that bacteria use to cause infection (establishment of the organism in the host) are different from those molecules, such as LPS, 9.14 that bacteria use to cause disease (dysfunction and damage of tissues) . The findings reported here suggest that rather than focusing strategies to develop bacterial vaccines only on blocking infection, or enhancing elimination of bacteria by phagocytosis, attention should also be directed to blocking the molecular mechanisms, such as those'described here, whereby bacteria cause disease .
Materials and methods Microorganisms . A colony type 1 (piliated), Tr/Op 5 clone of N. gonorrhoeae, strain 2686, was used in all studies involving gonococcal infection of organ cultures . The organisms were grown and used to prepare organ culture inocula as previously described,' so that after inoculation the concentration in the organ culture medium was 2 x 10 5 cfu/ml . Human fallopian tube organ cultures. Human fallopian tubes were obtained from unpregnant, pre-menopausal women with no history of salpingitis or other gonococcal infection who were undergoing hysterectomy and bilateral salpingooophorectomy for surgical indications (usually fibromas of the uterus) . After the tubes were freed of adventitial tissues, they were opened longitudinally and mucosal pieces with similar or identical epithelial surface area were cut using a skin biopsy punch with an internal diameter of 2 .5 mm . The mucosal pieces included the epithelium, which is comprised of ciliated and nonciliated cells ; the mucosal pieces also contained sub epithelial tissues, including connective tissue, arteries, veins and lymphatics as well as scattered mononuclear round cells . Such mucosal pieces contain antibody-producing cells, as evidenced by their production of IgA in response to gonococcal infection .21,2' Three pieces were put into each tissue culture grade polystyrene dish as previously described,' , ' except that the dishes were 35x 10 mm and contained 1 ml of Hepes-MEM (minimal Eagle's medium),' which was supplemented to 5% v/v with fetal calf serum (GIBCO) except in studies with rHuTNF-a, when the Hepes-MEM was supplemented to 0 .2% wt/v with bovine serum albumin (Sigma) to help block adsorption of TNF to the dish . In each individual experiment the tissues that comprised the organ cultures in the control and all experimental groups were from the same subject . Quantitation of damage to fallopian tube epithelium. Damage to the epithelium was assessed by determining the percent of the periphery of fallopian tube pieces with ciliary activity (PPCA) remaining at a particular time period relative to the pretreatment value for the same segments of the epithelium as described previously .` PPCA is an indirect measure of the number of functional ciliated cells remaining in the epithelium .' Thus, the tissues were assessed prior to infection or treatment with rHuTNF-a and 22 h later . All assessments of PPCA were performed without knowledge of the treatment group of the tissues examined . Enumeration of gonococci . Gonococci in the inoculum and in organ cultures were enumerated
as previously described .' Quantitation of TNF . All three mucosal pieces from each organ culture were homogenized in their own organ culture medium using a 'tissuemizer' (Tekmar, Cincinnati, OH), and the homogenates were filtered separately through 0 .45 ym pore size membrane filters prior to TNF quantitation . TNF was quantitated using the L-929 cell assay described by Zacharchuk et al- 21 which was modified by incubating the L-929 cells for 18 h prior to addition of actinomycin D, TNF standards and samples to be tested . Early in the studies of TNF induction by gonococcal infection, the TNF concentrations in organ culture homogenates were quantitated for purposes of comparison using both the L-929 cell bioassay and by ELISA, using a commercially prepared ELISA kit (Endogen) .
340
Z . A . McGee
et al.
Electron microscopy studies . Tissues to be examined by scanning electron microscopy were prepared and examined as previously described .' Effect of gonococcal infection on production of TNF by fallopian tube mucosal organ cultures . Duplicate human fallopian tube mucosal organ cultures were prepared, and one set was infected with N. gonorrhoeae, whereas the other set served as uninfected control tissues . Immediately following infection (0 time), and at intervals ranging over a 30 h period, the mucosal organ cultures were homogenized and assayed for TNF . In several experiments, the filtrates for TNF bioassay were tested with and without the addition of rabbit anti-TNF-a antibody (Endogen) .
Effect of rHuTNF-a on fallopian tube mucosa . Organ culture media were supplemented to various final concentrations of rHuTNF-a (Genentech) ranging from 5 to 5x 10 6 U/ml . Damage was assessed by determining the PPCA remaining 22 h after addition of rHuTNF-a to the organ culture fluid . In a sub-set of these studies, E. coli LPS (Associates of Cape Cod, Inc .) was suspended in Hepes-MEM using sonification as previously described' and added to fallopian tube organ cultures to final concentrations ranging from 1 to 100,000 pg/ml . The PPCA was determined just prior to addition of the LPS and after incubation of the organ cultures for 22 h following the addition of LPS. Effect of blocking TNF production . Four groups of organ cultures were designated . One group remained uninfected ; the second group was infected with N. gonorrhoeae in the standard manner;' the third group was infected as the second group, but the organ culture medium was supplemented with a blocker of TNF production, dexamethasone, 14,15 (Sigma) to a final concentration of 1 pm just prior to infection ; the fourth group was only treated with 1 µM dexamethasone . Epithelial ciliary activity was assessed 22 h later by determining the PPCA remaining in each of the four groups . After assessment of damage in two of the experiments in this series, the organ cultures were homogenized in their own medium and assays for TNF were performed in the standard manner on filtrates of the homogenates .
These studies were supported by grant AI-26285 from NIAID, NIH, U .S . Department of Health and Human Services, and in part by Public Health Services research grant no . M01-RR00064 from the National Center for Research Resources . We are indebted to Drs Thomas Abbott, John Andrews, Wallace Bryner, George Cannon, Dan Chichester, Philip Clark, Kent Farnsworth, Robert Flinner, Jeanne Gemmell, Jeff Gliedman, Tom Jones, Jeff Labrum, Ron Larkin, Steve Prescott, Alan Rappleye and Kent Rasmussen, without whose help these studies would not have been possible . The first author is especially grateful to Dr Porter Anderson (University of Rochester School of Medicine and Dentistry) whose personal encouragement was important in the genesis and completion of these studies .
References 1 . McGee ZA . Gonococcal pelvic inflammatory disease. In : Holmes KK, Mardh PA, Sparling PF, Wiesner PJ, eds . Sexually transmitted diseases . New York : McGraw-Hill Book Company, 1984; 220-29 . 2 . Cates W Jr, Rolts RT Jr, Aral SO . Sexually transmitted diseases, pelvic inflammatory disease, and infertility : an epidemiologic update . Epidemiol Rev 1990 ; 12 : 199-220 . 3 . Hill JH . Experimental infection with Neisseria gonorrhoeae. II . Animal inoculations . Am J Syph Gonorrhea and Vener Dis 1944 ; 28 : 334-78, 471-510 . 4 . McGee ZA, Johnson AP, Taylor-Robinson D . Human fallopian tubes in organ culture : preparation, maintenance, and quantitation of damage by pathogenic microorganisms . Infect Immun 1976; 13 : 608-18 . 5.
6. 7. 8. 9.
McGee ZA, Johnson AP, Taylor-Robinson D . Pathogenic mechanisms of Neisseria gonorrhoeae: observation on damage to human fallopian tubes in organ culture by gonococci of colony type 1 or type 4 . J Infect Dis 1981 ; 143 : 413-22 . Ward ME, Watt PJ . Adherence of Neisseria gonorrhoeae to urethral mucosal cells : an electronmicroscopic study of human gonorrhea . J Infect Dis 1977 ; 126 : 248-55 . Pollock CE, Harrison LW . Gonococcal Infections. London : Oxford University Press, 1912 ; xi-xiii . Kellogg DS Jr, Peacock WL Jr, Deacon WE, Brown L, Pirkle CI . Neisseria gonorrhoeae. I . Virulence genetically linked to clonal variation . J Bacteriol 1963; 85 : 1274-9 . Gregg CR, Melly MA, Hellergvist CG, Coniglio JG, McGee ZA. Toxic activity of purified lipo-
Local induction of tumor necrosis factor as a molecular mechanism
10 . 11 . 12 . 13 . 14 . 15 . 16 .
17 . 18 .
19 . 20 . 21 . 22 . 23 .
24 . 25 .
26 .
27 .
341
polysaccharide of Neisseria gonorrhoeae for human fallopian tube mucosa . J Infect Dis 1981 ; 143 : 432-9 . Gregg CR, Johnson AP, Taylor-Robinson D, Melly MA, McGee ZA . Host species-specific damage to oviduct mucosa by Neisseria gonorrhoeae Iipopolysaccharide . Infect Immun 1981 ; 34 : 1056-8. Melly MA, McGee ZA, Rosenthal RS . Ability of monomeric peptidoglycan fragments from Neisseria gonorrhoeae to damage human fallopian tube mucosa . J Infect Dis 1984; 149 : 378-86 . Mannel DN, Falk W . Optimal induction of tumor necrosis factor production in human monocytes requires complete S-form lipopolysaccharide . Infect Immun 1989 ; 57 : 1953-8 . Parant M . TNF as a mediator of LPS- and MDP-induced enhancement of mouse non-specific immunity . J Leukocyte Biol 1987 ; 42 : 576 . Beutler B, Cerami A . Cachectin : More than a tumor necrosis factor . New Eng J Med 1987 ; 316 : 37985 . Beutler B, Krochin N, Milsark IW, Luedke C, Cerami A . Control of cachectin (tumor necrosis factor) synthesis : mechanisms of endotoxin resistance . Science 1986 ; 232 : 977-9 . Okusawa S, Gelfand JA, Ikejima T, Connolly RJ, Dinarello CA. Interleukin 1 induces a shock-like state in rabbits: synergism with tumor necrosis factor and the effect of cyclooxygenase inhibition . J Clin Invest 1988; 81 : 1162-72 . Elias JA, Lentz V . IL-1 and tumor necrosis factor synergistically stimulate fibroblast IL-6 production and stabilize IL-6 messenger RNA . J Immunol 1990 ; 145(1) : 161-6 . Hart PH, Gerard FV, Burgess DR, Whitty GA, Piccoli DS, Hamilton JA . Potential antiinflammatory effects of interleukin 4 : suppression of human monocyte tumor necrosis factor x, interleukin 1, and prostaglandin E 2 . Proc Natl Acad Sci USA 1989 ; 86 : 3803-7 . Last-Barney K, Homon CA, Faanes RB, Merluzzi VJ . Synergistic and overlapping activities of tumor necrosis factor-a and IL-1 . J Immunol 1988 ; 141 (2) : 527-30 . Moller B, Mogensen SC, Wendelboe P, Bendtzen K, Petersen CM . Bioactive and inactive forms of tumor necrosis factor-a in spinal fluid from patients with meningitis . J Infect Dis 1991 ; 163 : 886-9 . Greisman SE, Hornick RB . Comparative pyrogenic reactivity of rabbit and man to bacterial endotoxin . Proc Soc Exp Biol Med 1969; 131 : 1154-8 . Blandau RJ . In : Hafez ESE, Blandau RJ, eds . The mammalian oviduct : comparative biology and methodology . Chicago: University of Chicago Press, 1969; 133-9. Waage A, Halstensen A, Shalaby R, Brandtzaeg P, Kierulf P, Espevik T . Local production of tumor necrosis factor alpha, Interleukin 1, and Interleukin 6 in meningococcal meningitis : relation to the inflammatory response . J Exp Med 1989; 170 : 1859-67 . Lebel MH, Freij BJ, Syrogiannopoulos GA et al. Dexamethasone therapy for bacterial meningitis : results of two double-blind, placebo-controlled trials . N Engl J Med 1988, 319 : 964-71 . Gorby GL, Clemens CM, Barley LR, McGee ZA . Effect of human chorionic gonadotropin (hCG) on Neisseria gonorrhoeae invasion of and IgA secretion by human fallopian tube mucosa . Microb Pathogen 1991 ; 10 : 373-84 . Cooper MD, McGee ZA, Mulks MH, Koomey JM, Hindman TL . Studies on attachment and invasion of human fallopian tube mucosa by a gonococcal IgA1 protease deficient mutant and its wild type parent . J Infect Dis 1984 ; 150 : 737-44 . Zacharchuk CM, Drysdale B-E, Mayer MM, Shin HS . Macrophage-mediated cytotoxicity : role of a soluble macrophage cytotoxic factor similar to lymphotoxin and tumor necrosis factor . Proc Natl Acad Sci USA 1983 ; 80 : 6341-5 .
