Microbial Pathogenesis 1990 ; 8 : 83-90
Expression of host resistance to Salmonella typhi and Salmonella typhimurium : bacterial survival within macrophages of murine and human origin Ion-Rusan Vladoianu, Hernan R . Chang* and Jean-Claude Pechere Department of Microbiology, University of Geneva Medical School, Centre Medical Universitaire, 1211 Geneve 4, Switzerland (Received September 7,1989 ; accepted in revised form November 8, 1989)
Vladoianu, I .-R . (Dept of Microbiology, University of Geneva Medical School, Centre Medical Universitaire, 1211 Geneve 4, Switzerland), H . R . Chang and J .-C . Pechere . Expression of host resistance to Salmonella typhi and Salmonella typhimurium: bacterial survival within macrophages of murine and human origin . Microbial Pathogenesis 1990 ; 8 : 83-90 . Cell-association of various strains of Salmonella typhi and Salmonella typhimurium with different populations of macrophages was studied . Macrophages were infected, exposed to gentamicin, washed, and counts of viable bacteria protected from gentamicin killing were made . J774A.1 cells, a continuous macrophage-like cell line, were the most permissive, all strains tested achieving similar high recoveries . Virulent S . typhimurium 779C-SmS, but not avirulent S. typhimurium 779C-SmD, survived well in mouse peritoneal macrophages and human monocyte-derived macrophages . Virulent S. typhi Ty2 were killed by mouse peritoneal macrophages, but were able to survive within human monocyte-derived macrophages . Viable counts of clinical isolates of S . typhi within the human monocyte-derived phagocytes were lower as compared with those of S . typhi Ty2 . Phagocytosis of opsonized and non-opsonized virulent S. typhi Ty2 and S. typhimurium 779C-SmS by mouse peritoneal macrophages failed to trigger their respiratory burst as assessed by the intracellular reduction of nitroblue tetrazolium dye (NBT) . These experiments support the view that the intracellular survival of Salmonella is in part host dependent and specific in nature . They also suggest that virulence influences the survival and intracellular multiplication of Salmonella within macrophages, and that their ultimate fate within macrophages may not be related to oxygen-dependent mechanisms . Key words : Salmonella ; macrophage ; host resistance .
Introduction Salmonella are enteropathogenic bacteria, with marked host specificity . Thus, S . typhi and S . paratyphi A, strictly pathogenic for humans (or higher primates), are not naturally virulent for mice . To kill mice, a high inoculum, i .e . 50 to
10OX10 6 bacteria
intraperitoneally (i .p .), is necessary ; death occurs within 1 to 4 days following the development of a toxic syndrome .' On the other hand, S . typhimurium is less pathogenic for an immunocompetent adult man, but highly virulent for mice . In mice only a few bacteria administered by the i .p . route will produce severe septicemia and death .' This variable behavior as a function of the host implies that, at a certain level, the defense mechanisms are host specific . As it is not possible, for ethical reasons, to work directly
" Author to whom correspondence should be addressed . 0882-4010/90/020083+08 $03 .00/0
c~
84
I .-R . Vladoianu et al.
with S. typhi and humans, most of the laboratory experimental work up to now has been performed using S . typhimurium and murine models . This means that almost no work on the physiological basis and genetic determination of host specificity of S . typhi has been done .' Moreover, the basis of host specificity among the serovars of the genus Salmonella has not been elucidated . Cell-mediated immunity seems, however, to play an important role in host resistance to Salmonella infection .' Here, we have examined the possible role of survival of the bacteria within macrophages in the host specificity of Salmonella .
Results Survival of Salmonella within macrophages Murine macrophages. The number of virulent S . typhimurium 779C-SmS recovered was similar at T5 (5 h after infection) and T 24 , the yield being about 3 log 10 per well [Fig . 1, panel (a)], whereas the number of avirulent S . typhimurium 779C-SmD recovered from the wells was about 2 log o lower than the parental strain . In contrast, S . typhi Ty2, which actively invaded murine macrophages as shown by the number of cfu obtained at T 5 , was completely killed by these macrophages at T 24 (3 loglo cfu/well) of Salmonella was high . The viable counts were similar at T24 for the three strains tested [Fig . 1, panel (b)] . Human monocyte-derived macrophages. Two experiments were performed with different inocula [Fig . 1, panels (c) and (d)] . S. typhimurium 779C-SmS showed the best survival in both experiments, followed by S . typhi Ty2 with intermediate (> 1 log 10 to < 3 log 10 cfu/well) to high viable counts and S. typhimurium 779C-SmD [panel (c) only] which showed low (< 1 log 10 cfu/well) to intermediate survival . The clinical isolates of S. typhi (6290 and 6553) also survived within these macrophages, ranging from low survival, when a low bacteria-cell ratio (6 :1 to 12 :1) was used, to intermediate survival when a high bacteria-cell ratio (60 :1 to 120 :1) was used . Antibiotic susceptibility of Salmonella strains In vitro susceptibility of the Salmonella strains to gentamicin measured before and after gentamicin exposure was the same, showing that the surviving bacteria had not acquired resistance to gentamicin killing (data not shown) . Moreover, control experiments carried out under similar conditions in the absence of macrophages yielded no bacterial growth at T 5 and T24 (data not shown) . Phagocytosis and intracellular reduction of nitroblue tetrazolium (NBT) dye upon phagocytosis of Salmonella by murine peritoneal macrophages Because oxygen-dependent mechanisms seem to play a role in the microbicidal activity of macrophages we tested the intracellular reduction of NBT (an assay for the production of the superoxide anion) upon phagocytosis of S. typhi Ty2 and S. typhimurium 779C-SmS . We used opsonized heat-killed C. albicans as a positive control and as a negative control, the protozoan parasite Toxoplasma gondii. Phagocytosis was inoculum dependent. When bacteria-cell ratios of 13 :1 and 19 :1 were used (Table 1), 39 and 47% of the macrophages contained intracellular S . typhi Ty2 and S . typhimurium 779C-SmS, respectively; when ratios of 190 :1 and 110 :1
Expression of host resistance to Salmonella
85
5
4
3
67
s 2
(c)
24
5
24
Time after infection ( h )
Fig . 1 . Survival of Salmonella within different populations of macrophages as assessed by the number of bacteria recovered from lysed macrophages (log 10 cfu per well) . Bacterial inocula are given for each experiment as cfu/well . (a) Murine peritoneal macrophages (2x10 5 cells/well) ; inocula : S . typhi Ty2, 4 .5 x 10 6 (o) ; S . typhimurium 779C-SmS, 3 .1 x 106 (o) ; and S . typhimurium 779C-SmD, 4 .4x 106 (a) . (b) J774A .1 macrophage-like cell line (2x10 5 cells/well) ; inocula : S . typhi Ty2, 5 .3x10 6 (o) ; S. typhimurium 779C-SmS, 3x106 (o) ; S . typhimurium 779C-SmD, 3 .1 x10 6 (U) . (c) Human-monocyte derived macrophages (5x10 ° cells/well) ; inocula : S . typhi Ty2, 5 .8x105 (o) ; S. typhi 6290, 2 .5x105 (o) ; S. typhi 6553, 3 .1 x10 5 (o) ; S . typhimurium 779C-SmS, 4 .2x105 (o) ; S . typhimurium 779C-SmD, 4 .4x10 5 (0) . (d) Human-monocyte derived macrophages (5x10 ^ cells/well) : inocula : S . typhi Ty2, 5 .6x106 (o) ; S. typhi 6290, 2 .5xl06 (A) ; S . typhi 6553, 3 .1 x10 6 (o) ; S . typhimurium 779C-SmS, 4 .2xl06 (o) . Results for each point represents the mean ±SEM of four wells .
86
Table 1
I .-R . Vladoianu et al .
Intracellular location of reduced NBT by murine peritoneal macrophages Intracellular location of reduced NBT
Challenge of macrophage? TT gondii' HK C. albicans° S. typhi Ty2` S. typhi Ty2° S. typhimurium 779C-SmS` S. typhimurium 779C-SmS ° S. typhi Ty2d S. typhimurium 779C-SmS °
Opsonization
Infected macrophages (%)
Infected macrophages with formazan-stained organisms (%)
No Yes No No No No Yes Yes
92±4 95+3 39±3 96±2 47±2 98±1 96+1 95±2
3+1" 91+4 2+1 1 3±1' 2+1 3±1e 3+1' 3±1e
Results are the mean±SEM from three replicates . adherent mouse peritoneal macrophages . 63x106 organisms of T. gond/i or opsonized heat-killed (HK) C . albicans in M199-10% FCS . `1 .3 x 10 7 cfu of S , typhi Ty2 or 1 .9X107 cfu S . typhimurium 779C-SmS in M1 99-10% FCS . d1 .9 x 10 8 cfu of opsonized or non-opsonized S . typhi Ty2 ; or 1 .1 x 10' cfu of opsonized or non-opsonized S. typhimurium 779C-SmS in M199-10% FCS . 'P < 0 .05, compared with heat-killed C. albicans . '1 x 10 6
were used, almost 100% of the macrophages contained intracellular S . typhi Ty2 and S. typhimurium 779C-SmS . Ninety-five percent of the macrophages contained intracellular C . albicans and 92% contained intracellular T. gondii after 60 min of incubation . The number of formazan-stained organisms on phagocytosis of both strains of Salmonella was around 3% against 91% with opsonized heat-killed C. albicans and 3% with T. gondii. In experiments in which S. typhi and S . typhimurium were opsonized prior to their contact with macrophage monolayers, we observed that for both strains around 95% of the macrophages presented intracellular bacteria after 60 min of contact but the number of formazan-stained organisms was not enhanced as compared with the experiments using non-opsonized organisms . Discussion and conclusions Survival of Salmonella within macrophages was dependent on host specificity and the virulence of the strain . We are not aware of other studies in which the intracellular survival of S . typhi and S. typhimurium have been compared within such a range of macrophages . The exact mechanism by which virulent S . typhi fail to survive within mouse peritoneal macrophages (as noted in previous work s ) still remains unclear and may explain why S . typhi is not virulent for mice . The fact that S. typhi was recovered from the infected monolayers at T 5 indicates that phagocytosis does take place . Therefore, virulence of S . typhi might depend on its intracellular survival and multiplication, at least in the case of macrophages . Recently, some studies have concluded that survival within the macrophage is essential for virulence of S . typhimurium' and that its ability to survive or grow intracellularly varies according to the source of macrophages, with higher rates of survival in splenic than in peritoneal macrophages .' The low survival of avirulent S . typhimurium 779C-SmD within murine macrophages observed here correlates well with the previously reported lack of virulence of this strain for mice . This lack of virulence seems to be due to an intrinsic, genetic quality of this mutant, rather than to its streptomycin dependence . 8 .9 • 1 ° It would be of interest to know whether there is a correlation between lack of
Expression of host resistance to Salmonella
87
virulence and poor intracellular survival when using avirulent strains of S . typhi and human macrophages . If there is, our in vitro assay of bacterial survival within human monocyte-derived macrophages may help to select avirulent mutants of S . typhi, which could be potential candidates for live oral antityphoid vaccines . Also, the possibility of using this assay to measure expression of cell-mediated immunity against S . typhi in subjects during vaccine trials is attractive and should be explored . Intracellular Salmonella lie within the phagolysosomes of phagocytes ." A major factor associated with virulence and resistance to intracellular killing of Salmonella seems to be the composition of its lipopolysaccharide (LPS) ; notably, smooth LPS confers resistance to lysosomal attack . 12 .13 Recently, a S . typhimurium locus (phoP), necessary for virulence in the mouse and survival within macrophages, has been described ." This locus is necessary for the resistance to defensins which are constituents of granule extracts previously known as phagocyte's lysosomal cationic proteins ." The construction of S . typhimurium phoP mutants has shown that these strains were avirulent but able to induce a protective immune response ." There is evidence to indicate that the phagocytosis of virulent S . typhi by polymorphonuclear neutrophils is not accompanied by enhanced oxygen consumption 17 (which would reflect a respiratory burst") . It has also been shown that avirulent S . typhi cannot block the respiratory burst but that both avirulent and virulent strains of S . typhi are equally killed ." Therefore, the lack of respiratory burst in PMN upon phagocytosis of S . typhi was probably not associated with the intracellular survival of this bacterium . Our data showing that mouse peritoneal macrophages (which killed S . typhi) did not respond with an enhanced respiratory burst (as assessed by the NBT studies) during phagocytosis of S . typhi confirms and extends these findings . If so, however, the ability of S . typhi to survive well within J774A .1 cells, of murine origin, would not be related to the recognized deficiency of this cell line in the production of toxic oxygen metabolites upon phagocytosis ." Besides, J774A .1 cells are also characterized by an enhanced capacity for phagocytosis 21,22 which accounts for the high viable counts encountered with all the Salmonella in our microassays . Taken together these results support the view that the intracellular survival of Salmonella is in part host dependent . They also suggest that bacterial virulence influences the survival and intracellular multiplication of Salmonella within macrophages, and that their ultimate fate within macrophages may not be related to oxygendependent mechanisms .
Materials and methods Salmonella strains and preparation of inocula . The following smooth strains of Salmonella were used : S . typhi Ty2, a virulent strain, widely employed for the preparation of anti-typhoid vaccines, and used by us in previous studies ;23 S . typhi 6290 and S. typhi 6553 strains recently isolated from patients with typhoid fever (Laboratory of Bacteriology, University Hospital of Geneva, Geneva, Switzerland) ; the streptomycin-sensitive 779C-SmS strain of S . typhimurium, a virulent strain (LD 50 i .p . for mice : 2x102 organisms) also used by us in previous studies ; 8' 10 avirulent S . typhimurium 779C-SmD, a one-step streptomycin-dependent mutant obtained from the former 779C-SmS strain . $ An 18 h culture at 37°C on tryptone soya agar (Oxoid Ltd, Basingstoke, U .K . ; simple or if necessary containing 500 µg/ml of streptomycin) was used . A suspension in medium 199 (M199 ; Seromed, Munich, F .R .G .) supplemented with 10% heat-inactivated (60 min, 56°C) fetal calf serum (FCS ; Seromed) or normal serum, corresponding to a turbidity of 10 international opacity units/m1 24 was made . Using this suspension a series of 10-fold dilutions was then prepared in the same medium . Some of these dilutions served as inocula while others were used to count the viable bacterial cells as colony forming units (cfu) by a pour-plate technique .
88
I .-R . Vladoianu et al.
Bacteria were not opsonized prior to contact with macrophage monolayers . In experiments to assess intracellular nitroblue tetrazolium (NBT) dye reduction, some suspensions were opsonized with mice serum (30 min, 37°C) prior to use . Other microoganisms . The highly virulent RH strain of Toxoplasma gondii was used . Purified tachyzoites were obtained as previously described 25 and adjusted to a concentration of 3x10 6 organisms/ml in M199-10% FCS . Candida albicans (ATCC 18804) was prepared as previously described ." Heat-killed (60 min, 100°C) C. albicans was opsonized with mice serum (30 min, 37°C) suspended in M199-10% FCS at a concentration of 3x10 6 microorganisms/ml . Murine macrophages. Resident peritoneal macrophages were obtained from pathogen-free Swiss-Webster mice as previously described ." Briefly, purified mononuclear cells were seeded on 96-well microtiter plates (Costar, Cambridge, Massachusetts, U .S .A.) at a concentration of 2x10 5 cells per well in M199-10% FCS . The monolayers were incubated overnight, washed with prewarmed M199 to remove non-adherent cells, and then used for experiments . For experiments in which NBT reduction was assessed, macrophages were placed in four-chamber glass slides (Lab-Tek, Miles, Naperville, U .S .A .) allowed to adhere overnight, washed with prewarmed M199 and used for experiments . J774A . 1 macrophage-like cell line . This continuous cell line derived from a recticulum cell sarcoma from BALB/c mice, 21 kindly provided by Dr Paul M . Tulkens (Catholic University of Louvain, Brussels, Belgium) was grown in M199-10% FCS and used at concentration of 2 x 10 5 per well (96-well microtiter plates) . Human monocyte-derived macrophages. Human mononuclear cells were obtained from the blood of human volunteers, serologically negative for B hepatitis, syphilis and HIV, and without history of typhoid fever or anti-typhoid vaccination . The heparinized blood was overlaid on Ficoll-Paque (Pharmacia, Uppsala, Sweden) and centrifuged at 450xg for 30 min . The purified mononuclear cells were seeded into the wells of 96-well microtiter plates and after overnight incubation the non-adherent cells were discarded by washing with prewarmed M199-10% normal serum . The medium was replaced every 3 days and the monolayers were used after 1214 days of incubation when the macrophages showed complete differentiation . The number of macrophages per well was approximately 5x10'. Incubations. Unless indicated otherwise, the incubations were performed at 37°C in a humid atmosphere of 5% C02-95% air . The pH of the media was adjusted to 7 .2 . In vitro assay for bacterial survival within macrophages . This was similar for all types of macrophage and was a modification of previously reported assays . 1,27 Briefly, 2 x 105 adherent murine macrophages and J774A .1 macrophage-like cells were washed once with 100 ill of prewarmed M199 and infected with 1100 MI of a suspension, in M199-10% FCS containing 3 to 6x10' cfu/well of the Salmonella strains . The bacteria-cell ratio was between 15 :1 and 30 :1 . The cells were incubated for 120 min at 37°C to allow phagocytosis, and the free bacteria were removed by three washes with M199 . M199-10% FCS (200 ul) containing 200 pg/ml of gentamicin (Seromed) was added, and the monolayers were incubated for 3 h . At this time (T 5 ) some of the wells were washed three times with M199 and the monolayers were lysed with 100 pl of 0 .1% Triton X-100 (4°C) for 15 min . The lysate was aspirated and kept in Eppendorf (Netheler and Hinz GmbH, Hamburg, F .R .G .) microtubes on ice . Four replicate samples were plated individually and the mean and SEM were calculated for each time point . The remaining monolayers were washed only once with 100 pl of M199 at T 5 and received 100 pi of M199-10% FCS containing 10 pg/ml of gentamicin for 19 h . These wells were processed in a similar way as the samples taken at T 5 . Gentamicin was used to kill extracellular bacteria as it has been demonstrated previously that gentamicin does not accumulate within phagocytes . 27,28 With human monocyte-derived macrophage monolayers the same procedure was used but the number of macrophages in the wells was approximately 5 x 10' cells (assessed by microscopic counting and by measuring the protein content of the wells against a previously constructed standard curve) and the bacterial inocula consisted of 100 pl of M199-10% normal serum containing 3 to 6 x 10 5 cfu/well or 3 to 6 x 10 6 cfu/well of the Salmonella strains . The bacteriacell ratio was thus 6 :1 to 12 :1 or 60 :1 to 120:1, respectively . Opsonization and centrifugation
Expression of host resistance to Salmonella
89
were not used to enhance cell-association of the bacteria in our assays . Rounded cells, which would indicate detachment of the monolayers were not observed at any time during experiments . Antibiotic susceptibility of Salmonella strains. The 90% minimal inhibitory concentrations (MIC 90 ) of gentamicin against the Salmonella strains measured using an in vitro assay, in the absence of macrophages, were found to be (µg/ml) : 0 .25 (S. typhi 6290) ; 0 .30 (S. typhi Ty2 and 6553) ; and 0 .35 [S . typhimurium 779C-SmS and 779C-SmD (in the presence of streptomycin)] . Intracellular location of NBT reduction . These studies were performed as previously described . 25 Murine macrophage monolayers were challenged with T. gondii (3 x 106 organisms) or opsonized heat-killed C. albicans; others received unopsonized or opsonized S . typhi Ty2 or S . typhimurium 779C-SmS . The microbial inocula were added in M199-10% FCS and all the monolayers received 300 yl of NBT (1 mg/ml in M199) for 1 h . Monolayers were fixed and 100-200 infected cells were assessed microscopically for the presence of formazan-stained microorganisms within vacuoles . Controls for phagocytosis were performed in parallel to the NBT experiments, these monolayers were assessed microscopically for the percentage of infected cells and number of intracellular microorganisms . Statistics. Data are reported as the mean +SEM . Significance was assessed by Student's ttest or ANOVA . A P value
Expression of host resistance to Salmonella typhi and Salmonella typhimurium : bacterial survival within macrophages of murine and human origin Ion-Rusan Vladoianu, Hernan R . Chang* and Jean-Claude Pechere Department of Microbiology, University of Geneva Medical School, Centre Medical Universitaire, 1211 Geneve 4, Switzerland (Received September 7,1989 ; accepted in revised form November 8, 1989)
Vladoianu, I .-R . (Dept of Microbiology, University of Geneva Medical School, Centre Medical Universitaire, 1211 Geneve 4, Switzerland), H . R . Chang and J .-C . Pechere . Expression of host resistance to Salmonella typhi and Salmonella typhimurium: bacterial survival within macrophages of murine and human origin . Microbial Pathogenesis 1990 ; 8 : 83-90 . Cell-association of various strains of Salmonella typhi and Salmonella typhimurium with different populations of macrophages was studied . Macrophages were infected, exposed to gentamicin, washed, and counts of viable bacteria protected from gentamicin killing were made . J774A.1 cells, a continuous macrophage-like cell line, were the most permissive, all strains tested achieving similar high recoveries . Virulent S . typhimurium 779C-SmS, but not avirulent S. typhimurium 779C-SmD, survived well in mouse peritoneal macrophages and human monocyte-derived macrophages . Virulent S. typhi Ty2 were killed by mouse peritoneal macrophages, but were able to survive within human monocyte-derived macrophages . Viable counts of clinical isolates of S . typhi within the human monocyte-derived phagocytes were lower as compared with those of S . typhi Ty2 . Phagocytosis of opsonized and non-opsonized virulent S. typhi Ty2 and S. typhimurium 779C-SmS by mouse peritoneal macrophages failed to trigger their respiratory burst as assessed by the intracellular reduction of nitroblue tetrazolium dye (NBT) . These experiments support the view that the intracellular survival of Salmonella is in part host dependent and specific in nature . They also suggest that virulence influences the survival and intracellular multiplication of Salmonella within macrophages, and that their ultimate fate within macrophages may not be related to oxygen-dependent mechanisms . Key words : Salmonella ; macrophage ; host resistance .
Introduction Salmonella are enteropathogenic bacteria, with marked host specificity . Thus, S . typhi and S . paratyphi A, strictly pathogenic for humans (or higher primates), are not naturally virulent for mice . To kill mice, a high inoculum, i .e . 50 to
10OX10 6 bacteria
intraperitoneally (i .p .), is necessary ; death occurs within 1 to 4 days following the development of a toxic syndrome .' On the other hand, S . typhimurium is less pathogenic for an immunocompetent adult man, but highly virulent for mice . In mice only a few bacteria administered by the i .p . route will produce severe septicemia and death .' This variable behavior as a function of the host implies that, at a certain level, the defense mechanisms are host specific . As it is not possible, for ethical reasons, to work directly
" Author to whom correspondence should be addressed . 0882-4010/90/020083+08 $03 .00/0
c~
84
I .-R . Vladoianu et al.
with S. typhi and humans, most of the laboratory experimental work up to now has been performed using S . typhimurium and murine models . This means that almost no work on the physiological basis and genetic determination of host specificity of S . typhi has been done .' Moreover, the basis of host specificity among the serovars of the genus Salmonella has not been elucidated . Cell-mediated immunity seems, however, to play an important role in host resistance to Salmonella infection .' Here, we have examined the possible role of survival of the bacteria within macrophages in the host specificity of Salmonella .
Results Survival of Salmonella within macrophages Murine macrophages. The number of virulent S . typhimurium 779C-SmS recovered was similar at T5 (5 h after infection) and T 24 , the yield being about 3 log 10 per well [Fig . 1, panel (a)], whereas the number of avirulent S . typhimurium 779C-SmD recovered from the wells was about 2 log o lower than the parental strain . In contrast, S . typhi Ty2, which actively invaded murine macrophages as shown by the number of cfu obtained at T 5 , was completely killed by these macrophages at T 24 (3 loglo cfu/well) of Salmonella was high . The viable counts were similar at T24 for the three strains tested [Fig . 1, panel (b)] . Human monocyte-derived macrophages. Two experiments were performed with different inocula [Fig . 1, panels (c) and (d)] . S. typhimurium 779C-SmS showed the best survival in both experiments, followed by S . typhi Ty2 with intermediate (> 1 log 10 to < 3 log 10 cfu/well) to high viable counts and S. typhimurium 779C-SmD [panel (c) only] which showed low (< 1 log 10 cfu/well) to intermediate survival . The clinical isolates of S. typhi (6290 and 6553) also survived within these macrophages, ranging from low survival, when a low bacteria-cell ratio (6 :1 to 12 :1) was used, to intermediate survival when a high bacteria-cell ratio (60 :1 to 120 :1) was used . Antibiotic susceptibility of Salmonella strains In vitro susceptibility of the Salmonella strains to gentamicin measured before and after gentamicin exposure was the same, showing that the surviving bacteria had not acquired resistance to gentamicin killing (data not shown) . Moreover, control experiments carried out under similar conditions in the absence of macrophages yielded no bacterial growth at T 5 and T24 (data not shown) . Phagocytosis and intracellular reduction of nitroblue tetrazolium (NBT) dye upon phagocytosis of Salmonella by murine peritoneal macrophages Because oxygen-dependent mechanisms seem to play a role in the microbicidal activity of macrophages we tested the intracellular reduction of NBT (an assay for the production of the superoxide anion) upon phagocytosis of S. typhi Ty2 and S. typhimurium 779C-SmS . We used opsonized heat-killed C. albicans as a positive control and as a negative control, the protozoan parasite Toxoplasma gondii. Phagocytosis was inoculum dependent. When bacteria-cell ratios of 13 :1 and 19 :1 were used (Table 1), 39 and 47% of the macrophages contained intracellular S . typhi Ty2 and S . typhimurium 779C-SmS, respectively; when ratios of 190 :1 and 110 :1
Expression of host resistance to Salmonella
85
5
4
3
67
s 2
(c)
24
5
24
Time after infection ( h )
Fig . 1 . Survival of Salmonella within different populations of macrophages as assessed by the number of bacteria recovered from lysed macrophages (log 10 cfu per well) . Bacterial inocula are given for each experiment as cfu/well . (a) Murine peritoneal macrophages (2x10 5 cells/well) ; inocula : S . typhi Ty2, 4 .5 x 10 6 (o) ; S . typhimurium 779C-SmS, 3 .1 x 106 (o) ; and S . typhimurium 779C-SmD, 4 .4x 106 (a) . (b) J774A .1 macrophage-like cell line (2x10 5 cells/well) ; inocula : S . typhi Ty2, 5 .3x10 6 (o) ; S. typhimurium 779C-SmS, 3x106 (o) ; S . typhimurium 779C-SmD, 3 .1 x10 6 (U) . (c) Human-monocyte derived macrophages (5x10 ° cells/well) ; inocula : S . typhi Ty2, 5 .8x105 (o) ; S. typhi 6290, 2 .5x105 (o) ; S. typhi 6553, 3 .1 x10 5 (o) ; S . typhimurium 779C-SmS, 4 .2x105 (o) ; S . typhimurium 779C-SmD, 4 .4x10 5 (0) . (d) Human-monocyte derived macrophages (5x10 ^ cells/well) : inocula : S . typhi Ty2, 5 .6x106 (o) ; S. typhi 6290, 2 .5xl06 (A) ; S . typhi 6553, 3 .1 x10 6 (o) ; S . typhimurium 779C-SmS, 4 .2xl06 (o) . Results for each point represents the mean ±SEM of four wells .
86
Table 1
I .-R . Vladoianu et al .
Intracellular location of reduced NBT by murine peritoneal macrophages Intracellular location of reduced NBT
Challenge of macrophage? TT gondii' HK C. albicans° S. typhi Ty2` S. typhi Ty2° S. typhimurium 779C-SmS` S. typhimurium 779C-SmS ° S. typhi Ty2d S. typhimurium 779C-SmS °
Opsonization
Infected macrophages (%)
Infected macrophages with formazan-stained organisms (%)
No Yes No No No No Yes Yes
92±4 95+3 39±3 96±2 47±2 98±1 96+1 95±2
3+1" 91+4 2+1 1 3±1' 2+1 3±1e 3+1' 3±1e
Results are the mean±SEM from three replicates . adherent mouse peritoneal macrophages . 63x106 organisms of T. gond/i or opsonized heat-killed (HK) C . albicans in M199-10% FCS . `1 .3 x 10 7 cfu of S , typhi Ty2 or 1 .9X107 cfu S . typhimurium 779C-SmS in M1 99-10% FCS . d1 .9 x 10 8 cfu of opsonized or non-opsonized S . typhi Ty2 ; or 1 .1 x 10' cfu of opsonized or non-opsonized S. typhimurium 779C-SmS in M199-10% FCS . 'P < 0 .05, compared with heat-killed C. albicans . '1 x 10 6
were used, almost 100% of the macrophages contained intracellular S . typhi Ty2 and S. typhimurium 779C-SmS . Ninety-five percent of the macrophages contained intracellular C . albicans and 92% contained intracellular T. gondii after 60 min of incubation . The number of formazan-stained organisms on phagocytosis of both strains of Salmonella was around 3% against 91% with opsonized heat-killed C. albicans and 3% with T. gondii. In experiments in which S. typhi and S . typhimurium were opsonized prior to their contact with macrophage monolayers, we observed that for both strains around 95% of the macrophages presented intracellular bacteria after 60 min of contact but the number of formazan-stained organisms was not enhanced as compared with the experiments using non-opsonized organisms . Discussion and conclusions Survival of Salmonella within macrophages was dependent on host specificity and the virulence of the strain . We are not aware of other studies in which the intracellular survival of S . typhi and S. typhimurium have been compared within such a range of macrophages . The exact mechanism by which virulent S . typhi fail to survive within mouse peritoneal macrophages (as noted in previous work s ) still remains unclear and may explain why S . typhi is not virulent for mice . The fact that S. typhi was recovered from the infected monolayers at T 5 indicates that phagocytosis does take place . Therefore, virulence of S . typhi might depend on its intracellular survival and multiplication, at least in the case of macrophages . Recently, some studies have concluded that survival within the macrophage is essential for virulence of S . typhimurium' and that its ability to survive or grow intracellularly varies according to the source of macrophages, with higher rates of survival in splenic than in peritoneal macrophages .' The low survival of avirulent S . typhimurium 779C-SmD within murine macrophages observed here correlates well with the previously reported lack of virulence of this strain for mice . This lack of virulence seems to be due to an intrinsic, genetic quality of this mutant, rather than to its streptomycin dependence . 8 .9 • 1 ° It would be of interest to know whether there is a correlation between lack of
Expression of host resistance to Salmonella
87
virulence and poor intracellular survival when using avirulent strains of S . typhi and human macrophages . If there is, our in vitro assay of bacterial survival within human monocyte-derived macrophages may help to select avirulent mutants of S . typhi, which could be potential candidates for live oral antityphoid vaccines . Also, the possibility of using this assay to measure expression of cell-mediated immunity against S . typhi in subjects during vaccine trials is attractive and should be explored . Intracellular Salmonella lie within the phagolysosomes of phagocytes ." A major factor associated with virulence and resistance to intracellular killing of Salmonella seems to be the composition of its lipopolysaccharide (LPS) ; notably, smooth LPS confers resistance to lysosomal attack . 12 .13 Recently, a S . typhimurium locus (phoP), necessary for virulence in the mouse and survival within macrophages, has been described ." This locus is necessary for the resistance to defensins which are constituents of granule extracts previously known as phagocyte's lysosomal cationic proteins ." The construction of S . typhimurium phoP mutants has shown that these strains were avirulent but able to induce a protective immune response ." There is evidence to indicate that the phagocytosis of virulent S . typhi by polymorphonuclear neutrophils is not accompanied by enhanced oxygen consumption 17 (which would reflect a respiratory burst") . It has also been shown that avirulent S . typhi cannot block the respiratory burst but that both avirulent and virulent strains of S . typhi are equally killed ." Therefore, the lack of respiratory burst in PMN upon phagocytosis of S . typhi was probably not associated with the intracellular survival of this bacterium . Our data showing that mouse peritoneal macrophages (which killed S . typhi) did not respond with an enhanced respiratory burst (as assessed by the NBT studies) during phagocytosis of S . typhi confirms and extends these findings . If so, however, the ability of S . typhi to survive well within J774A .1 cells, of murine origin, would not be related to the recognized deficiency of this cell line in the production of toxic oxygen metabolites upon phagocytosis ." Besides, J774A .1 cells are also characterized by an enhanced capacity for phagocytosis 21,22 which accounts for the high viable counts encountered with all the Salmonella in our microassays . Taken together these results support the view that the intracellular survival of Salmonella is in part host dependent . They also suggest that bacterial virulence influences the survival and intracellular multiplication of Salmonella within macrophages, and that their ultimate fate within macrophages may not be related to oxygendependent mechanisms .
Materials and methods Salmonella strains and preparation of inocula . The following smooth strains of Salmonella were used : S . typhi Ty2, a virulent strain, widely employed for the preparation of anti-typhoid vaccines, and used by us in previous studies ;23 S . typhi 6290 and S. typhi 6553 strains recently isolated from patients with typhoid fever (Laboratory of Bacteriology, University Hospital of Geneva, Geneva, Switzerland) ; the streptomycin-sensitive 779C-SmS strain of S . typhimurium, a virulent strain (LD 50 i .p . for mice : 2x102 organisms) also used by us in previous studies ; 8' 10 avirulent S . typhimurium 779C-SmD, a one-step streptomycin-dependent mutant obtained from the former 779C-SmS strain . $ An 18 h culture at 37°C on tryptone soya agar (Oxoid Ltd, Basingstoke, U .K . ; simple or if necessary containing 500 µg/ml of streptomycin) was used . A suspension in medium 199 (M199 ; Seromed, Munich, F .R .G .) supplemented with 10% heat-inactivated (60 min, 56°C) fetal calf serum (FCS ; Seromed) or normal serum, corresponding to a turbidity of 10 international opacity units/m1 24 was made . Using this suspension a series of 10-fold dilutions was then prepared in the same medium . Some of these dilutions served as inocula while others were used to count the viable bacterial cells as colony forming units (cfu) by a pour-plate technique .
88
I .-R . Vladoianu et al.
Bacteria were not opsonized prior to contact with macrophage monolayers . In experiments to assess intracellular nitroblue tetrazolium (NBT) dye reduction, some suspensions were opsonized with mice serum (30 min, 37°C) prior to use . Other microoganisms . The highly virulent RH strain of Toxoplasma gondii was used . Purified tachyzoites were obtained as previously described 25 and adjusted to a concentration of 3x10 6 organisms/ml in M199-10% FCS . Candida albicans (ATCC 18804) was prepared as previously described ." Heat-killed (60 min, 100°C) C. albicans was opsonized with mice serum (30 min, 37°C) suspended in M199-10% FCS at a concentration of 3x10 6 microorganisms/ml . Murine macrophages. Resident peritoneal macrophages were obtained from pathogen-free Swiss-Webster mice as previously described ." Briefly, purified mononuclear cells were seeded on 96-well microtiter plates (Costar, Cambridge, Massachusetts, U .S .A.) at a concentration of 2x10 5 cells per well in M199-10% FCS . The monolayers were incubated overnight, washed with prewarmed M199 to remove non-adherent cells, and then used for experiments . For experiments in which NBT reduction was assessed, macrophages were placed in four-chamber glass slides (Lab-Tek, Miles, Naperville, U .S .A .) allowed to adhere overnight, washed with prewarmed M199 and used for experiments . J774A . 1 macrophage-like cell line . This continuous cell line derived from a recticulum cell sarcoma from BALB/c mice, 21 kindly provided by Dr Paul M . Tulkens (Catholic University of Louvain, Brussels, Belgium) was grown in M199-10% FCS and used at concentration of 2 x 10 5 per well (96-well microtiter plates) . Human monocyte-derived macrophages. Human mononuclear cells were obtained from the blood of human volunteers, serologically negative for B hepatitis, syphilis and HIV, and without history of typhoid fever or anti-typhoid vaccination . The heparinized blood was overlaid on Ficoll-Paque (Pharmacia, Uppsala, Sweden) and centrifuged at 450xg for 30 min . The purified mononuclear cells were seeded into the wells of 96-well microtiter plates and after overnight incubation the non-adherent cells were discarded by washing with prewarmed M199-10% normal serum . The medium was replaced every 3 days and the monolayers were used after 1214 days of incubation when the macrophages showed complete differentiation . The number of macrophages per well was approximately 5x10'. Incubations. Unless indicated otherwise, the incubations were performed at 37°C in a humid atmosphere of 5% C02-95% air . The pH of the media was adjusted to 7 .2 . In vitro assay for bacterial survival within macrophages . This was similar for all types of macrophage and was a modification of previously reported assays . 1,27 Briefly, 2 x 105 adherent murine macrophages and J774A .1 macrophage-like cells were washed once with 100 ill of prewarmed M199 and infected with 1100 MI of a suspension, in M199-10% FCS containing 3 to 6x10' cfu/well of the Salmonella strains . The bacteria-cell ratio was between 15 :1 and 30 :1 . The cells were incubated for 120 min at 37°C to allow phagocytosis, and the free bacteria were removed by three washes with M199 . M199-10% FCS (200 ul) containing 200 pg/ml of gentamicin (Seromed) was added, and the monolayers were incubated for 3 h . At this time (T 5 ) some of the wells were washed three times with M199 and the monolayers were lysed with 100 pl of 0 .1% Triton X-100 (4°C) for 15 min . The lysate was aspirated and kept in Eppendorf (Netheler and Hinz GmbH, Hamburg, F .R .G .) microtubes on ice . Four replicate samples were plated individually and the mean and SEM were calculated for each time point . The remaining monolayers were washed only once with 100 pl of M199 at T 5 and received 100 pi of M199-10% FCS containing 10 pg/ml of gentamicin for 19 h . These wells were processed in a similar way as the samples taken at T 5 . Gentamicin was used to kill extracellular bacteria as it has been demonstrated previously that gentamicin does not accumulate within phagocytes . 27,28 With human monocyte-derived macrophage monolayers the same procedure was used but the number of macrophages in the wells was approximately 5 x 10' cells (assessed by microscopic counting and by measuring the protein content of the wells against a previously constructed standard curve) and the bacterial inocula consisted of 100 pl of M199-10% normal serum containing 3 to 6 x 10 5 cfu/well or 3 to 6 x 10 6 cfu/well of the Salmonella strains . The bacteriacell ratio was thus 6 :1 to 12 :1 or 60 :1 to 120:1, respectively . Opsonization and centrifugation
Expression of host resistance to Salmonella
89
were not used to enhance cell-association of the bacteria in our assays . Rounded cells, which would indicate detachment of the monolayers were not observed at any time during experiments . Antibiotic susceptibility of Salmonella strains. The 90% minimal inhibitory concentrations (MIC 90 ) of gentamicin against the Salmonella strains measured using an in vitro assay, in the absence of macrophages, were found to be (µg/ml) : 0 .25 (S. typhi 6290) ; 0 .30 (S. typhi Ty2 and 6553) ; and 0 .35 [S . typhimurium 779C-SmS and 779C-SmD (in the presence of streptomycin)] . Intracellular location of NBT reduction . These studies were performed as previously described . 25 Murine macrophage monolayers were challenged with T. gondii (3 x 106 organisms) or opsonized heat-killed C. albicans; others received unopsonized or opsonized S . typhi Ty2 or S . typhimurium 779C-SmS . The microbial inocula were added in M199-10% FCS and all the monolayers received 300 yl of NBT (1 mg/ml in M199) for 1 h . Monolayers were fixed and 100-200 infected cells were assessed microscopically for the presence of formazan-stained microorganisms within vacuoles . Controls for phagocytosis were performed in parallel to the NBT experiments, these monolayers were assessed microscopically for the percentage of infected cells and number of intracellular microorganisms . Statistics. Data are reported as the mean +SEM . Significance was assessed by Student's ttest or ANOVA . A P value
