Survival of two Orientia tsutsugamushi bacterial strains that infect mouse macrophages with varying degrees of virulence

Microbial Pathogenesis 39 (2005) 177–187 www.elsevier.com/locate/micpath

Survival of two Orientia tsutsugamushi bacterial strains that infect mouse macrophages with varying degrees of virulence Masahiro Fukuharaa, Masahiro Fukazawaa, Akira Tamuraa, Tatsunosuke Nakamuraa, Hiroshi Urakamib,* a

Laboratory of Microbiology, Faculty of Pharmacy, Niigata University of Pharmacy and Applied Life Sciences (NUPALS), 265-1 Higashijima, Niigata 956-8603, Japan b Laboratory of Food Microbiology and Food Safety, Faculty of Applied Life Sciences, Niigata University of Pharmacy and Applied Life Sciences (NUPALS), 265-1 Higashijima, Niigata 956-8603, Japan Received 25 May 2005; received in revised form 2 August 2005; accepted 5 August 2005 Available online 12 September 2005

Abstract Orientia tsutsugamushi, an intracellular parasitic bacterium, comprises numerous strains of differing virulence. When BALB/c mice were infected intraperitoneally with this pathogen, a virulent strain known as Karp was found to multiply in the intraperitoneal macrophages and kill the mouse. In contrast, an avirulent strain, Kuroki, was shown to invade macrophages but be eliminated from the cells, allowing mouse survival. O. tsutsugamushi invades its host cell cytoplasm through phagocytosis and disruption of phagosomal membranes but some bacteria are then killed by phago-lysosomes within 1 h of infection. Microscopic observations could not differentiate the Karp and Kuroki strains during entry and subsequent cell killing by phago-lysosomes. However, the Kuroki cells failed to divide and were markedly deformed following cytoplasmic invasion at several days post-infection. These findings suggest that macrophages have a mechanism to eliminate O. tsutsugamushi in the cytoplasm, if the invading bacteria escape phagosomal clearance, and that it is this mechanism that Kuroki does not survive. Additionally, significant levels of nitric oxide (NO) are produced in macrophages by Kuroki, but not by Karp. An NO synthase inhibitor, however, does not increase the growth of Kuroki, suggesting that NO is induced in a strain-dependent manner but does not effect proliferation. q 2005 Elsevier Ltd. All rights reserved. Keywords: Orientia tsutsugamushi; Macrophage; Virulence

1. Introduction Orientia tsutsugamushi, a causative agent of scrub typhus, is an obligate intracellular parasitic bacterium that multiplies only in the cytoplasm of animal cells. Strains of this microorganism vary in both their antigenicity and virulence in mammalian hosts [1,2]. Many of these strains have now been isolated by inoculation of laboratory mice * Corresponding author. Address: Laboratory of Food Microbiology and Food Safety, Faculty of Applied Life Sciences, Niigata University of Pharmacy and Applied Life Sciences (NUPALS), 265-1 Higashijima, Niigata 956-8603, Japan. Tel.: C81 250 25 5000; fax: C81 250 25 5148. E-mail address: [email protected] (H. Urakami).

0882-4010/$ - see front matter q 2005 Elsevier Ltd. All rights reserved. doi:10.1016/j.micpath.2005.08.004

with clinical blood samples and shown to have different levels of virulence in these animals. Virulent strains, such as Karp, kill the mice even at low doses. On the other hand, a previous report by Kobayashi et al. [3] has described an avirulent strain that does not propagate in mice, which survive without showing symptoms of illness. However, when avirulent strains are injected into either nude mice or euthymic mice treated with an immunosuppressive reagent, they can multiply in mononuclear phagocytes, including intraperitoneal macrophages, and be successfully isolated. Thereafter, strains that are avirulent in mice, such as Kuroki, could be isolated from patients [4–6]. Although it is still unclear whether avirulent strains exist in humans, seroepidemiological studies in endemic areas have shown inapparent infections, which were suggested to be caused by O. tsutsugamushi strains with low virulence in humans

178

M. Fukuhara et al. / Microbial Pathogenesis 39 (2005) 177–187

Special characters a g

‘Alpha’ in Greek ‘Gamma’ in Greek

[7,8]. In the present study, we have compared infection by Karp and Kuroki in laboratory mice as a model for human infection, even though these strains are both virulent in humans. Defense mechanisms against O. tsutsugamushi have previously been studied in cultured cells, laboratory mice and human patients. Furthermore, antibodies have been detected both in the clinic [9] and in residents of endemic areas [10]. The type specific antigen protein of O. tsutsugamushi is abundant in the outer membranes of these bacteria and is immunodominant [11]. Antibodies against these proteins have been detected in sera from patients [12] and found to have inhibitory effects against attachment and penetration of O. tsutsugamushi in mice [13]. A previous report by Shirai et al. analyzed protective effects against infection by two antigenically different strains of O. tsutsugamushi [14]. They revealed in their study that lymphocytes were capable of conferring protection against heterologous strains, but that large quantities of serum from immunized mice failed to protect recipients against a challenge from a different strain. The importance of macrophages in protecting against O. tsutsugamushi has also been demonstrated in mice. When laboratory mice were infected by O. tsutsugamushi, activated macrophages that suppressed bacterial growth were observed to develop before the appearance of circulating antibodies [15]. Another study by Catanzaro et al. analyzed mice infected with O. tsutsugamushi and showed that survival depended on an increase in bactericidal macrophage levels [16]. Cytokines have also been shown to play important roles during O. tsutsugamushi infection. Macrophage activating factor [17], tumor necrosis factoralpha (TNF-a), interleukin-6 [18], chemokines [19] and macrophage colony-stimulating factor [20] are detectable in infected cultures in vitro and also in infected animals and humans. Gamma interferon (IFN-g) and TNF-a were also previously shown to inhibit the growth of O. tsutsugamushi in mouse fibroblasts and macrophages [21,22]. These findings suggest that whilst antibodies have some role in the defense against O. tsutsugamushi infection, cellular immunity and cytokine induction are more important factors in this protection [23]. Nitric oxide (NO) is known to have bactericidal effects against intracellular parasitic bacteria, such as those of the Rickettsiaceae family to which O. tsutsugamushi belongs. Significantly, the replication of Rickettsia conorii in mouse endothelial cells is inhibited by IFN-g and TNF-a, and this growth is recovered by the addition of an NO synthase

m G

’mu’ in Greek Superscript of ‘G’

inhibitor [24]. Moreover, Ehrlichia risticii is eliminated from IFN-g-treated macrophages in an NO-participating manner [25]. These data suggest that NO is an important effector in the defense against these intracellular bacteria. However, little is still known about either the mechanism of induction of NO in infected cells or its effects on O. tsutsugamushi. Observations of the process by which O. tsutsugamushi enters the host cell cytoplasm have revealed aspects of the host defense mechanisms that kill these microorganisms. We previously observed the entry of O. tsutsugamushi into cultivated mouse L929 fibroblasts by electron microscopy [26]. Our results showed that the bacteria were phagocytosed but escaped into the cytoplasm by disruption of phagosomal membranes. Some of the phagocytosed microorganisms, however, failed to escape and as a result were contained within the phagosomes and subsequently appeared to experience a bactericidal mechanism. These bacterial cells contained within phagosomes were plasmolyzed, deformed and eventually killed within the phago-lysosomes within 40 min from the initiation of phagocytosis. Such early killing of O. tsutsugamushi was also reported in guinea pig neutrophils in vitro [27]. It is noteworthy however that Nacy and Meltzer [28] showed not only a rapid killing mechanism by which macrophages kill O. tsutsugamushi within 1 h, but also a slower killing mechanism that acts 24 h after infection. In spite of these previous findings, it is still unclear how the behavior of virulent and avirulent O. tsutsugamushi bacteria differs and also how the host cells in question respond in a different way to strains of dissimilar virulence. In our present study, the entry, multiplication and survival of the two strains with different degrees of virulence that we selected were analyzed in mice in vivo and in macrophages in vitro. We show from our subsequent data that mouse macrophages use at least two separate mechanisms to eliminate O. tsutsugamushi. At the first level of defense, the bacteria are destroyed by phago-lysosomes during the entry process. A second mechanism involves the killing of invading microorganisms in the cytoplasm after they evade lysis by escaping from the phagosomes. In our current comparisons of the survival of the virulent Karp and avirulent Kuroki strains of O. tsutsugamushi in mouse cells, our findings strongly suggest that phago-lysosomes were equally effective in killing both strains, but that the cytoplasmic defense mechanism was only effective against Kuroki.

M. Fukuhara et al. / Microbial Pathogenesis 39 (2005) 177–187

2. Results 2.1. Virulence of the Karp and Kuroki strains of O. tsutsugamushi in mice Serially diluted inocula of Karp and Kuroki strains of O. tsutsugamushi were prepared from infected cultures of mouse L929 cells. These inocula were then used to compare the infectivity of these bacterial strains in vivo and in vitro, that is, 50% lethal dose concentration (LD50) in BALB/c mice and infected cell counting unit (ICU, which is determined by infection of cultivated cells and expressed as 10% infectious dose) in L929 cells (Table 1). The inocula of both strains had comparable ICU titers of around 106/ml and the infected mice were killed by inoculations of Karp diluted to 103 ICU. In contrast, a Kuroki dose of 106 ICU neither killed the animals nor caused any signs of illness, such as erection of the hair and the accumulation of peritoneal fluids. These results indicated that Karp is virulent and that Kuroki is avirulent to BALB/c mice. 2.2. Measurement of both the cell populations and the multiplication of bacterial strains in the peritoneal cavity of mice after infection with the Karp and Kuroki strains of O. tsutsugamushi To analyze the cellular response in mice to O. tsutsugamushi infection, the Karp and Kuroki strains were injected intraperitoneally at a dose of approximate 104 ICU. The population numbers of macrophages, neutrophils and lymphocytes in the peritoneal cavities were then analyzed at various timepoints after injection. All of these intraperitoneal cell types significantly increased in number after injection with Karp. It was striking that the number of macrophages was 30-fold greater in Karp infected mice than in the mock control on day 6 post-infection (pi), and that all of these mice died on day 7 pi (Fig. 1A). It is also significant that Neutrophils, which are rarely found in the cavities of mock-infected mice, increased linearly during the first 5 days pi and peaked at this number by day 6 pi. Lymphocytes also increased in number over the timecourse of this experiment, at about 50% of the macrophage levels. Karp bacteria were detectable in both macrophages Table 1 Infectivity levels of inocula of O. tsutsugamushi Strain Karp Kuroki

ICUa (ml) 6

2.25!10 1.49!106

LD50b (ml) 3

5.62!10 3.16!100

LD50/ICU 2.20!10K3 !2.12!10K6

L929 cells and BALB/c mice were inoculated with 0.25 ml of serially diluted inocula of Karp and Kuroki. Results are expressed as ICU and LD50 in the undiluted inoculum of each strain. a 10% infectious dose in L929 cell culture. b 50% lethal dose in BALB/c mice.

179

and neutrophils (Fig. 1C; solid lines) but few microorganisms were found in lymphocytes. In addition, the numbers of O. tsutsugamushi in macrophages increased exponentially from day 0 (3 h) to day 6 pi. In contrast, the numbers of O. tsutsugamushi in neutrophils was less than 0.01 per cell up to day 1 pi and as low as one per cell until day 3 pi (Fig. 1C). These numbers increased linearly until day 6 pi. The symptoms of illness became severe on day 5 pi in Karp infected mice. In contrast to the data for Karp infection, in mice infected by Kuroki the influx of immune cells into the peritoneal cavity was significantly less (Fig. 1B). This influx increased during the first 4 days pi but decreased gradually to mock infection levels and the mice were all found to survive. In addition, the incremental increases in macrophage numbers in Kuroki infected mice were much greater than in neutrophils, which were always found to be less than 10% of the maximum neutrophil populations in Karp infected animals. Bacterial cells were frequently observed in the macrophages of Kuroki infected mice but few were found in neutrophils and lymphocytes throughout our observation period (Fig. 1C; broken line). Although Kuroki cells were observed in similar numbers to Karp in macrophages at 3 h pi, these numbers remained unchanged for 8 days and then decreased gradually as macrophage numbers declined in the peritoneal cavity. The O. tsutsugamushi numbers in the mouse cells did not appear to change but macrophage numbers increased during the first 4 days pi, suggesting that Kuroki cells did proliferate slowly until this time and decrease thereafter. These observations clearly indicate sharp differences in both bacterial propagation and in the immune responses of mice following infection by the Karp and Kuroki strains. It is noteworthy also that Kuroki invades macrophages as efficiently as Karp but multiplied much more slowly and only during the first 4 days pi. In contrast, the Karp strain was shown to continue to propagate until the death of the host. This suggests that macrophages play an important role in protection against the Kuroki strain. 2.3. In vitro infection by O. tsutsugamushi of exuded mouse macrophages The in vivo immune responses in the mouse to infection by the Karp and Kuroki strains of O. tsutsugamushi were found to be distinctly different, so we designed a new series of experiments that would omit influences from other cell and tissue types on bacterial growth in macrophages. Exuded macrophages were collected following injection of mice with proteose peptone and were subsequently infected in vitro with the Karp and Kuroki strains. As shown in Fig. 2, similar numbers of Karp and Kuroki were found in the macrophages at the initial stage of infection, suggesting that the efficiency of phagocytosis was equivalent for both strains, despite differences in virulence. The results of these

180

M. Fukuhara et al. / Microbial Pathogenesis 39 (2005) 177–187

Fig. 1. Measurement of the mouse cell populations in the peritoneal cavities after infection with O. tsutsugamushi Karp and Kuroki strains and of the multiplication of these microorganisms in the peritoneal cells. BALB/c mice were injected intraperitoneally with Karp and Kuroki. For mock infections, MEM alone was injected into the mice. At various timepoints after injection, the intraperitoneal cells were recovered and counted. These cells were then stained with Giemsa to determine the O. tsutsugamushi levels and to differentiate between the cell types. The data for day 0 were obtained from observations at 3 h postinfection. The numbers of macrophages (circle), neutrophils (triangle) and lymphocytes (square) are shown after infection with Karp (A) and Kuroki (B). (C) Multiplication of Karp (solid lines) and Kuroki (broken line) in the macrophages (circle) and neutrophils (triangle). The populations of Karp and Kuroki in lymphocytes, Kuroki in neutrophils and Karp in neutrophils, until day 2 post-infection, were found to be less than 0.01/cell throughout the experiment. All of the data shown are average readings with standard error for four mice.

experiments showed that Karp had multiplied exponentially within macrophages following invasion, and that the cells began to lyse on day 6 pi. Conversely, Kuroki bacteria did not multiply in the macrophages, but were detectable at constant levels until day 8 pi. This observation indicated that the growth of Kuroki was arrested in macrophages even in vitro without any influences from surrounding tissues. 2.4. Ultrastructural observations of O. tsutsugamushi in intraperitoneal cells We examined infected intraperitoneal cells by electron microscopy to further analyze the characteristics of the growth and elimination of O. tsutsugamushi. Mice were injected intraperitoneally with inocula of approximately 106

ICU, and intraperitoneal cells were harvested at 1 h pi. These intraperitoneal cell preparations consisted mainly of macrophages, and invading bacteria were found not only in the cytoplasm of these cells (Fig. 3A) but also in phagosomes (Fig. 3B–D). Some of the bacterial cells contained within the phagosomes appeared to be intact but many showed considerable levels of plasmolysis. The plasmolyzed microorganisms were considered to be undergoing a bactericidal process by a phago-lysosomal mechanism. Lysosomal fusions were also occasionally observed in phagosomes containing bacteria (Fig. 3B, arrow). We counted the bacterial numbers in both the cytoplasm and in the phagosomes (Table 2). We observed that more than 75% of the invading O. tsutsugamushi had already escaped into the cytoplasm with some remaining in

M. Fukuhara et al. / Microbial Pathogenesis 39 (2005) 177–187

Fig. 2. Growth of Karp and Kuroki in peritoneal exuded macrophages in vitro. Monolayers of exuded macrophages (5!105/chamber) were infected in vitro with Karp (closed symbols) and Kuroki (open symbols) at a multiplicity of 0.07 and 0.16, respectively. All of the data are representative of the average with standard error of four independent experiments.

181

phagosomes. Unlike the situation in phagosomes, few of the bacteria that we detected in the cytoplasms had undergone changes characteristic of plasmolysis. Additionally, no significant differences were found in the ratios between Karp and Kuroki in terms of their localization or the incidence of plasmolysis. This finding indicated that the efficiency of cellular destruction by phago-lysosomes was equivalent between these two differently virulent O. tsutsugamushi strains. It was significant that beyond day 1 pi, neither Karp nor Kuroki were detectable in phagosomes, indicating that bacteria that were contained within these structures had already been digested. The mice that were injected with a 106 ICU inocula of Karp died on day 2 pi, so we used a 105 ICU inoculum to observe the bacterial cells during the later stages of infection. On days 1, 2 and 4 pi, Karp numbers increased in the cytoplasms of both macrophages and neutrophils (Fig. 4A and B). The morphologies of the invading Karp were not significantly different between these species of mouse cells. Several bacteria were usually found in a single mouse cell at day 4 pi. Moreover, Kuroki numbers were equivalent to Karp at the initial stages of infection, with Kuroki found to be present solely in

Fig. 3. Electron micrographs of O. tsutsugamushi in mouse intraperitoneal macrophages at 1 h post-infection. The intraperitoneal cells were recovered from infected mice, which had been injected intraperitoneally with Karp and Kuroki. Bacteria are seen in the cytoplasms of macrophages. (A) Karp cell in the cytoplasm of mouse cells. The phagosomal membrane is no longer evident around the microorganisms. (B) Kuroki cell within a phagosome. The morphology of the microorganism is almost intact and a lysosomal fusion (arrow) is observed. (C and D) Karp and Kuroki cells, respectively, in phagosomes. The microorganisms are plasmolyzed and deformed. Scale bar equals 0.5 mm.

182

M. Fukuhara et al. / Microbial Pathogenesis 39 (2005) 177–187

Table 2 Localization and plasmolysis of O. tsutsugamushi strains in macrophages at 1 hr post-infection Strain

No. of bacteria (%)

Average no. of

In cytoplasm

Karp Kuroki a b

Intact 86.2G5.51b 77.6G11.47

a

In phagosome Plasmolyzed 0.3G0.41 0.2G0.41

Intact 3.3G1.65 4.2G2.79

Plasmolyzed 5.5G2.66 10.9G9.61

Undefined

Bacteria examined

Macrophages examined

2.2G3.66 7.0G3.24

133 134

84.7 219.3

Could not be defined either in cytoplasm or in phagosome by observation of ultrathin sections. Average with standard error for three mice.

macrophages until days 1 and 2 pi (Fig. 4C). It then became difficult to detect Kuroki at subsequent timepoints, even in mice that had received a dose of 106 ICU. The number of Kuroki cells appeared to decrease rapidly over the postinfection timecourse by electron microscopic observations. This suggested that Karp could propagate in the cytoplasm of macrophages, but that Kuroki failed to multiply in these cells despite successful invasion from phagosomes into the cytoplasm. In order to follow the morphological changes that Kuroki cells would undergo at the later stages of infection, we inoculated macrophages that had been cultivated in vitro, as the infection rates could be achieved at a high multiplicity in this system. Kuroki bacteria were always detectable in the cytoplasm of these macrophages immediately after

inoculation, but they acquired a deformed and swollen morphology on days 2 and 4 pi (Fig. 4D and E). Furthermore, all of the bacterial cells that displayed these characteristics were found in the cytoplasm and not in membrane-bound vacuoles. In contrast, Karp cells multiplied in the cytoplasm yet did not show any such deformed or swollen morphology (data not shown). We observed the progression of both strains in the in vitro macrophage cultures until day 8 pi, but Kuroki cells could not be detected at any timepoint later than day 4 pi, even though the number of these bacteria had not noticeably decreased when analyzed by light microscopy (Fig. 2). It is conceivable therefore that Giemsa staining allows for the visualization of considerably deformed microorganisms that electron microscopy does not detect.

Fig. 4. Electron micrographs of O. tsutsugamushi in mouse intraperitoneal cells infected in vivo and in vitro on days 2 and 4 post-infection. (A) Karp cells in an intraperitoneal mouse macrophage. (B) Mouse intraperitoneal neutrophil. Mice were intraperitoneally infected and intraperitoneal cells were harvested and observed on day 4 post-infection. Bacteria with intact morphologies are observed in the cytoplasm. (C) Kuroki cells in an intraperitoneal macrophage. A mouse was infected with Kuroki and observed on day 2 post-infection. The Kuroki bacteria were found solely in the cytoplasm. (D and E) Kuroki in exuded macrophages on days 2 and 4 post-infection, respectively. Macrophages were isolated from the peritoneal cavity of uninfected mice and infected in vitro. The Kuroki bacteria are considerably deformed in the cytoplasms. Scale bar equals 0.5 mm.

M. Fukuhara et al. / Microbial Pathogenesis 39 (2005) 177–187

183

Fig. 5. Growth of O. tsutsugamushi and nitrite production in NG-monomethyl-L-arginine (NMMA)-treated and untreated mouse macrophages. Macrophages on slide chambers were infected by O. tsutsugamushi and the cells were cultivated in medium containing NMMA or in medium alone. Nitrite in the culture supernatants was measured with Greiss reagent. (A and B) Concentration of nitrite in the cultures infected by Karp (filled) and Kuroki (hatched) and uninfected (open). The cells were treated with NMMA (B) and with medium alone (A). (C) Growth of Karp (closed symbols) and Kuroki (open symbols) in NMMAtreated (square) and untreated (circle) macrophages. All data shown are the averages with standard error of four independent experiments.

2.5. Induction of nitrous oxide in O. tsutsugamushi infected mouse macrophages and its effect upon bacterial cell growth Exuded macrophages were infected in vitro with the Karp and Kuroki strains of O. tsutsugamushi to analyze the effects upon the induction of NO. The levels of NO in the cells did not change significantly from control levels following Karp infection, but were markedly increased by Kuroki infection (Fig. 5A). Addition of NG-monomethyl-Larginine (NMMA), an inhibitor of NO synthase, suppressed the induction of NO in the Kuroki-infected cultures to control levels, whereas NO levels in Karp-infected cultures were equivalent to control cultures treated with NMMA (Fig. 5B). However, although NO levels in Kuroki-infected mouse macrophages were reduced to control levels by NMMA treatment, the growth of this bacterium was unaffected (Fig. 5C). 3. Discussion and conclusions Previous studies of early immunological responses to O. tsutsugamushi infection in laboratory mice have shown

that macrophage-mediated cellular immunity plays an important role in protection against this bacterium [14,16, 17,21]. In the present study, we demonstrate clear differences in the growth and elimination of two strains of O. tsutsugamushi in a single strain of mouse macrophages that are dependent on the virulence of the infecting bacteria. When the virulent Karp strain was injected into mice, macrophages, neutrophils and lymphocytes were found to infiltrate the peritoneal cavity. These macrophages and neutrophils were susceptible to invasion and multiplication of Karp, which resulted the death of the animals within several days. In contrast, the avirulent Kuroki strain also attracted the migration of immune cells into the peritoneal cavity, but the number of invading macrophages and neutrophils was less than the response to Karp infection. Kuroki cells were phagocytosed into macrophages as efficiently as Karp, but their multiplication was at a much slower rate and was restricted to in first 4 days pi. These bacteria were finally eliminated in the later stages of infection. Neutrophils and lymphocytes were also shown to infiltrate the peritoneal cavities of mice in response to Kuroki-infection, but subsequent in vitro experiments suggested that macrophages can clear O. tsutsugamushi

184

M. Fukuhara et al. / Microbial Pathogenesis 39 (2005) 177–187

without the help of other immune cells. These observations indicate that macrophages play a crucial role in the elimination of Kuroki, and that other factors, such as cytokines secreted by other cell types, are not essential for this function. It is also noteworthy that this protective mechanism is effective only against the Kuroki strain and not Karp. Infection by O. tsutsugamushi has previously been shown to induce the mouse macrophage cell line J774A.1 to secrete several kinds of chemokines [19], some of which are potent attractants for macrophages, neutrophils and lymphocytes [29]. Based on our current data, neutrophils and lymphocytes are likely to have participated in the protective mechanisms against infection by Kuroki bacteria, although we have not yet fully elucidated this. A previous report has demonstrated that upon O. tsutsugamushi invasion of guinea pig neutrophils, 50% of the bacteria successfully entered the cytoplasm by disrupting the phagosomal membrane, whereas the remainder were digested by phago-lysosomes [27]. Geng and Jerrells also demonstrated in their study that growth of O. tsutsugamushi is inhibited in mouse macrophages by IFN-g, which is one of the principal cytokines that is secreted by lymphocytes [21]. These observations provide highly informative evidence for the mechanisms by which neutrophils and lymphocytes participate in protection against O. tsutsugamushi infection. The majority of the O. tsutsugamushi strains can be classified as either virulent or avirulent in laboratory mice, but the Gilliam strain is known to display varying levels of virulence depending on the mouse host. A mouse gene responsible for susceptibility to infection by Gilliam bacteria has now been identified and mapped to the locus of a cytokine gene, early T-lymphocyte activation 1 [30,31, 32], which affects both the migration and bactericidal activity of macrophages [32,33]. In addition, a study by Nacy and Groves analyzed the characteristics of Gilliam infectivity in both susceptible and resistant strains of mice [17]. They demonstrated that injection of Gilliam intraperitoneally resulted in bacterial growth in the peritoneal macrophages of the susceptible C3H mouse strain, but not in resistant BALB/c mice. When the same experiments were performed in vitro, however, the macrophages from both strains of mice were vulnerable to the invasion and growth of the Gilliam bacteria to the same extent. These findings suggests that the growth of Gilliam in BALB/c macrophages in vivo is suppressed mainly by extracellular factors. In our present study, the growth of Kuroki could be inhibited in vitro by macrophages alone. This indicated that mice have at least two mechanisms that regulate the growth of O. tsutsugamushi in macrophages. The first mechanism, initially described by Nacy and Groves, is dependant upon extracellular factors and is influenced by the mouse host strain [17]. The second protective pathway, described in our current study, is dependent on the specific infecting strain of O. tsutsugamushi and is activated by macrophages alone.

The entry process of O. tsutsugamushi into mammalian cells has been observed previously in mouse spleen cells [34], guinea pig neutrophils [27] and L929 cells [26]. All of these studies revealed that this invasion of bacterial cells consisted of phagocytosis into phagosomes and subsequent escape into the cytoplasm by disruption of phagosomal membranes. Some of the phagocytosed bacteria, however, failed to escape into the cytoplasm and were digested within the phago-lysosomes. Our results also confirmed that the same entry and killing processes occurred in mouse macrophages in vivo. Interestingly, neither light nor electron microscopic observations could detect any significant differences, either in vivo or in vitro, between the total numbers of invading Karp and Kuroki in mouse macrophages immediately after infection. This suggested that the virulence of these strains does not depend on the efficiency of uptake by phagocytosis. We then analyzed these infected mouse macrophages by electron microscopy to compare early killing of the two different O. tsutsugamushi strains. No significant differences were found in the numbers of bacteria present either in the phagosomes, cytoplasm or in the process of plasmolysis. This strongly suggested that the events of the early stages of infection were equivalent for both strains. Differences between the Karp and Kuroki strains, however, were clearly observed in the later stages of infection. Although both strains successfully migrated to the cytoplasm, Kuroki appeared to divide very slowly and only during the first 4 days after which the cells ceased to propagate. In contrast, infecting Karp bacteria multiplied exponentially throughout the post-infection period. It was also evident that Kuroki cells were both swollen and markedly deformed following entry into the macrophage cytoplasm, but that Karp cells appeared to be intact at this stage. This was indicative of a bactericidal function of macrophages against O. tsutsugamushi that is separate from early killing by phago-lysosomes. This additional protective mechanism appeared to act over a range of several days and remove the bacteria from the cytoplasm. In this regard, it is noteworthy that digestion by phago-lysosomes appears to be equally effective against both the Karp and Kuroki strains, but that bacterial killing in the cytoplasm affected mainly the Kuroki cells. Importance of this determinant is still unclear in humans, because both of Karp and Kuroki are virulent to humans, but it is conceivably a key determinant in the virulence levels between these two strains of O. tsutsugamushi evaluated in BALB/c mice. R. conorii is eliminated in the cytoplasm of endothelial cells stimulated by TNF-a and IFN-g. During this killing process, the microorganisms are enveloped with double membraned, auto-phagosomal vacuoles [24]. When antibiotics were added to macrophage-like cultured cells harboring Listeria monocytogenes in their cytoplasms, the metabolically arrested bacteria were also auto-phagocytosed into double membrane-bound vacuoles and eliminated [35]. In the present study, however, we did not find

M. Fukuhara et al. / Microbial Pathogenesis 39 (2005) 177–187

O. tsutsugamushi within double membrane-bound vacuoles and all of the deformed Kuroki cells that we observed were localized in the cytoplasm during the late stages of infection. In a previous study from this laboratory, we treated infected L929 cells with antibiotics that were effective against O. tsutsugamushi, and observed that the bacteria were deformed with swollen morphologies, and that auto-phagocytosis was not evident [36]. The morphologies of the deformed bacteria that we observed previously are very similar to those observed in our present study. These observations suggest that O. tsutsugamushi bacteria in host cell cytoplasms are eliminated by a mechanism other than auto-phagocytosis. Our current data show that Kuroki induce mouse macrophages to produce considerable amounts of NO, but that this effect was not induced by infection with Karp. This clearly indicates that NO is produced in macrophages by infection with O. tsutsugamushi, but that this is bacteria strain-dependent. Treatment with an inhibitor of NO synthase reduced the NO levels in the Kuroki-infected cultures to those of the Karp-infected cells, but the growth of Kuroki bacteria was not promoted by this inhibition. We performed this same experiment using a macrophage-like cell line, RAW 267.4. The results were almost identical except that the growth of Kuroki was slightly improved (data not shown). These data indicate that NO has a negligible or limited effect on the growth of these bacteria, even if it is significantly induced in cells infected by certain strains of O. tsutsugamushi. This is a quite different result from the recovery effects that have been reported for NMMA on the growth of R. conorii in mouse epithelial cells transduced by cytokines [24]. Induction of NO by mouse macrophages infected with Trypanosoma cruzi [37], Francisella tularensis [38] and group B streptococci [39] can be inhibited by anti-TNF-a antibodies in vitro, indicating that this cytokine is secreted by mouse macrophages and stimulates the cells to produce NO in an autocrine manner. Furthermore, O. tsutsugamushi is known to induce mouse macrophages and macrophage-like cell lines to secrete TNF-a [18,40]. We have also now found that Kuroki induces higher levels of TNF-a in macrophages than Karp (Fukuhara et al., unpublished). In the present study, we have compared the Karp and Kuroki strains of O. tsutsugamushi that have different levels of virulence in BALB/c mice. We contend from the results of our analyses that these differences in virulence depend mainly on whether the bacteria can survive killing in the host cytoplasm after they escape from the phagosomes. This difference is also likely to be one of the principal underlying reasons for the variations in virulence between the Karp and Kuroki strains of O. tsutsugamushi in BALB/c mice. However, it is still unclear whether this is a general phenomenon in other mouse strains or with other virulent and avirulent strains of O. tsutsugamushi. We will also need to further elucidate the components of these different protective mechanisms and further studies that have been

185

designed to examine additional bacterial strains and to analyze the effects of different cytokines are now in progress.

4. Materials and methods 4.1. Host cells L929 cells were grown in Eagle’s minimum essential medium (MEM, Nissui pharmaceutical Co., Ltd, Tokyo, Japan) containing 10% fetal bovine serum. For preparation of mouse peritoneal exuded macrophages, 1 ml of 10% proteose peptone solution (Difco Lab., Detroit, MI, USA) was injected intraperitoneally into 7–8-week-old BALB/c mice. Three days after the injection, 10 ml of phosphate buffered saline (PBS) was infused, and the fluid was recovered by suction with syringes. The cells were collected by low speed centrifugation and resuspended in MEM supplemented with 10% serum, 100 U/ml of penicillin and 150 mg/ml of streptomycin. About 0.5 ml of the cell suspensions (1!106 cells/ml) were then added to each chamber of a Lab-Tek 4-chamber slide (Nunc, Inc., Naperville, IL, USA) and incubated for 1 day. Non-adherent cells were removed by vigorous washing with PBS and adherent cells were used as the exuded macrophages in subsequent experiments. 4.2. O. tsutsugamushi. The O. tsutsugamushi Karp and Kuroki strains were cultivated in L929 cells as described previously [41]. Both of these strains multiply efficiently in L929 cells in spite of their differences in virulence in mice. When the growth of O. tsutsugamushi reached approximately 100 microorganisms per cell in smears stained with Giemsa, the cells were harvested and suspended in MEM, either supplemented or not with 10% serum, and homogenized by a stainless-steel Dounce homogenizer (Kontes Glass Co., Vineland, NJ, USA). The homogenates were centrifuged at 200!g for 10 min and the supernatants were used as inocula. MEM with serum was used to inoculate L929 cells, and MEM with serum was used for the exuded macrophages. The microorganisms suspended in MEM without serum were injected into mice. The infectivity of O. tsutsugamushi in the inocula was determined by infection of L929 cells in chamber slides and expressed as 10% infectious dose in 106 cells, designated as ICU [42]. Briefly, L929 cells in the chamber slides were inoculated with serially diluted inocula, and cultivated for 4 days. The cells were then stained by Giemsa to determine the ratios of the infected cells by microscopic observations. The ICU was calculated from dilutions of the inoculum in proximity to the 10% infectious dose and the number of the L929 cells on day 4 pi was determined by a hemocytometer.

186

M. Fukuhara et al. / Microbial Pathogenesis 39 (2005) 177–187

All cultivation of L929 cells and macrophages occurred at 37 8C in a moist atmosphere containing 5% CO2. 4.3. Infection of mice Seven to eight week-old BALB/c female mice were injected intraperitoneally with 0.25 ml of serum-free MEM containing approximate 104–106 ICU of O. tsutsugamushi, prepared as described. For mock infections, the same amount of MEM alone was injected. At various timepoints after injection, the intraperitoneal cells were recovered by the same method for preparation of the exuded macrophages described above. The number of cells in the fluid was determined using a hemocytometer. The ratio of macrophages, neutrophils and lymphocytes was measured by microscopic observation of smears stained by Giemsa. To determine the growth of O. tsutsugamushi in these cells, the cells were stained by Giemsa and classified into five categories based on the number of O. tsutsugamushi in a cell. These categories included 0, 1–10, 11–50, 51–100 and greater than 100 bacteria per cell. More than 300 cells were observed for each preparation, and the total number of microorganisms in each cell was estimated by calculation of the median numbers of the classes (0, 5, 20, 75 and 125), the ratios of the cells in each class and the total number of cells. The virulence of the bacteria in mice was determined by their LD50 in the BALB/c strain by intraperitoneal injection as described previously [43]. 4.4. In vitro infection of exuded macrophages by O. tsutsugamushi Monolayers of exuded macrophages in chamber slides (5!105/chamber) were washed three times with PBS and overlaid with 0.25 ml of the inocula. After incubation for 2 h, the cells were washed and cultivated in 0.5 ml of MEM containing serum. At various times after infection, the cells on the slides were stained with Giemsa for light microscopic observations, and the number of microorganisms in each chamber was estimated by classifying the cells into the five categories described above. The multiplicity of infection was calculated from the ICU values measured in L929 cells and by the number of macrophages. 4.5. Electron microscopy Cells from the intraperitoneal cavities and the macrophages cultivated in vitro were prepared for electron microscopy as reported previously [26,41]. Briefly, samples were double fixed with 2% glutaraldehyde and 1% osmic acid in 0.1 M cacodylate buffer, pH 7.4, stained with 2% uranyl acetate. They were then dehydrated through an ethanol series, and embedded in Quetol-812 (Nissin EM Co., Ltd, Tokyo, Japan). Ultrathin sections were stained with lead staining solution (Katayama Yakuhin Co., Ltd,

Tokyo, Japan) and observed in an electron microscope. For the analysis of the localization and morphology of the microorganisms in the early stage of infection, the macrophages cut at the center of the cells were observed in two to four ultrathin sections prepared from distant parts of the blocks until the number of the microorganisms exceeded 120. 4.6. Analysis of nitric oxide production The infected and uninfected macrophages in the chamber slides were cultivated with or without 0.5 mM NMMA in the culture medium. At various times after infection, the culture fluids were harvested and centrifugal supernatants at 8000!g for 5 min were stored at K80 8C to assay for nitrite, a degradation product of NO. Nitrite in the culture fluids was determined by the Griess reaction [38]. Briefly, 100 ml of the culture fluid was mixed with an equal amount of Griess reagent, consisting of 1% sulfanilamide and 0.02% naphtylethylene diamine dihydrochloride in 2.5% phosphoric acid, incubated at room temperature for 30 min and quantified by measuring in an ELISA reader at 540 nm. Sodium nitrite solutions at known concentration in RPMI 1640 were used as standards.

Acknowledgements This study was supported by the Promotion and Mutual Aid Corporation for Private School of Japan and by the Ministry of Education, Science, Sports and Culture.

References [1] Kawarura Jr A, Tanaka H, Tamura A. In: Kawarura Jr A, Tanaka H, Tamura A, (Eds.). Tsutsugamushi Disease. Tokyo: University Tokyo Press; 1995. [2] Rapmund G. Rickettsial diseases of the Far East: new perspectives. J Infect Dis 1984;149:30–8. [3] Kobayashi Y, Tachibana N, Matsumoto I, Oyama T, Kageyama T. Isolation of very low virulent strain of Rickettsia tsutsugamushi by use of cyclophosphamide-treated mice. In: Kazar J, Ormsbee RA, Tarasevich IN, (Eds.). Rickettsiae and rickettsial diseases. Bratislava: Slovak Academy of Science; 1978. p. 181–6. [4] Ohashi N, Tamura A, Sakurai H, Yamamoto S. Characterization of a new antigenic type, Kuroki, of Rickettsia tsutsugamushi isolated from a patient in Japan. J Clin Microbiol 1990;28: 2111–3. [5] Tamura A, Takahashi K, Tsuruhara T, Urakami H, Miyamura S, Sekikawa H, Kenmotsu M, Shibata M, Abe S, Nezu H. Isolation of Rickettsia tsutsugamushi antigenically different from Kato, Karp, and Gilliam strains from patients. Microbiol Immunol 1984; 28:873–82. [6] Yamamoto S, Kawabata N, Tamura A, Urakami H, Ohashi N, Murata M, Yoshida Y, Kawamura Jr A. Immunological properties of Rickettsia tsutsugamushi, Kawasaki strain, isolated from a patient in Kyushu. Microbiol Immunol 1986;30:611–20.

M. Fukuhara et al. / Microbial Pathogenesis 39 (2005) 177–187 [7] Fan MY, Walker DH, Yu SR, Liu QH. Epidemiology and ecology of rickettsial diseases in the People’s Republic of China. Rev Infect Dig 1987;9:823–40. [8] Kawamura Jr A Mur, ata M, Osono M, Nogami S, Shirasaka A, Tanaka H, Sudo K, Suzuki K, Miyairi T, Kijima H. Studies on inapparent infection of Tsutsugamushi disease in Izu Shichito Islands; seroepidemiology and demonstration of an avirulent rickettsia strain for mice. Jpn J Exp Med 1980;50:91–105. [9] Kularatne SAM, Edirisingha JS, Gawarammana IB, Urakami H, Chenchittikul M, Kaiho I. Emerging rickettsial infections in Sri Lanka: the pattern in the hilly Central Province. Trop Med Int Health 2003;8:803–11. [10] Strickman D, Tanskul P, Eamsila C, Kelly DJ. Prevalence of antibodies to rickettsiae in the human population of suburban Bangkok. Am J Trop Med Hyg 1994;51:149–53. [11] Tamura A, Ohashi N, Urakami H, Takahashi K, Oyanagi M. Analysis of polypeptide composition and antigenic components of Rickettsia tsutsugamushi by polyacrylamide gel electrophoresis and immunoblotting. Infect Immun 1985;48:671–5. [12] Ohashi N, Tamura A, Suto T. Immunoblotting analysis of antirickettsial antibodies produced in patients of Tsutsugamushi disease. Microbiol Immunol 1988;32:1085–92. [13] Seong SY, Kim HR, Huh MS, Park SG, Kang JS, Han TH, Choi MS, Chang WH, Kim IS. Induction of neutralizing antibody in mice by immunization with recombinant 56 kDa protein of Orientia tsutsugamushi. Vaccine 1997;15:1741–7. [14] Shirai A, Catanzaro PJ, Phillips SM, Osterman JV. Host defenses in experimental scrub typhus: role of cellular immunity in heterologous protection. Infect Immun 1976;14:39–46. [15] Nacy CA, Osterman JV. Host Defenses in Experimental Scrub Typhus: Role of Normal and Activated Macrophages. Infect Immun 1979;26:744–50. [16] Catanzaro PJ, Shirai A, Hilderbrandt PK, Osterman JV. Host defenses in experimental scrub typhus: histopathological correlates. Infect Immun 1976;13:861–75. [17] Nacy CA, Groves MG. Macrophages in resistance to rickettsial infections: early host defense mechanisms in experimental scrub typhus. Infect Immun 1981;31:1239–50. [18] Kim MK, Kang JS. Orientia tsutsugamushi suppresses the production of inflammatory cytokines induced by its own heat-stable component in murine macrophages. Microb Pathog 2001;31:145–50. [19] Cho NH, Seong SY, Huh MS, Han TH, Koh YS, Choi MS, Kim IS. Expression of chemokine genes in murine macrophages infected with Orientia tsutsugamushi. Infect Immun 2000;68:594–602. [20] Iwasaki H, Takada N, Nakamura T, Ueda T. Increased levels of macrophage colony-stimulating factor, gamma interferon, and tumor necrosis factor alpha in sera of patients with Orientia tsutsugamushi infection. J Clin Microbiol 1997;35:3320–2. [21] Geng P, Jerrells TR. The role of tumor necrosis factor in host defense against scrub typhus rickettsiae, I. Inhibition of growth of Rickettsia tsutsugamushi, Karp strain, in cultured murine embryonic cells and macrophages by recombinant tumor necrosis factor-alpha. Microbiol Immunol 1994;38:703–11. [22] Hanson B. Susceptibility of Rickettsia tsutsugamushi Gilliam to gamma interferon in cultured mouse cells. Infect Immun 1991;59: 4125–33. [23] Seong SY, Cho MS, Kim IS. Orientia tsutsugamushi infection: overview and immune responses. Microbes Infect 2001;3:11–21. [24] Walker DH, Popov VL, Crocquet-Valdes PA, Welsh CJ, Feng HM. Cytokine-induced, nitric oxide-dependent, intracellular anti-rickettsial activity of mouse endothelial cells. Lab Invest 1997;76:129–38. [25] Park J, Rikihisa Y. L-arginine-dependent killing of intracellular Ehrlichia risticii by macrophages treated with gamma interferon. Infect Immun 1992;60:3504–8.

187

[26] Urakami H, Tsuruhara T, Tamura A. Penetration of Rickettsia tsutsugamushi into cultured mouse fibroblasts (L cells): an electron microscopic observation. Microbiol Immunol 1983;27:251–63. [27] Rikihisa Y, Ito S. Entry of Rickettsia tsutsugamushi into polymorphonuclear leukocytes. Infect Immun 1982;38:343–50. [28] Nacy A, Meltzer MS. Macrophages in resistance to rickettsial infection: macrophage activation in vitro for killing of Rickettsia tsutsugamushi. J Immunol 1979;123:2544–9. [29] Baggiolini M, Dewald B, Moser B. Interleukin-8 and relatedchemotactic cytokines-CXC and CC chemokines. Adv Immunol 1994;55: 97–179. [30] Groves MG, Osterman JV. Host defenses in experimental scrub typhus: genetics of natural resistance to infection. Infect Immun 1978; 19:583–8. [31] Groves MG, Rosenstreich DL, Taylor BA, Osterman JV. Host defenses in experimental scrub typhus: mapping the gene that controls natural resistance in mice. J Immunol 1980;125:1395–9. [32] Patarca R, Freeman GJ, Singh RP, Wei FY, Durfee T, Blattner F, Regnier DC, Kozak CA, Mock BA, Morse HC. Structural and functional studies of the early T lymphocyte activation 1 (Eta-1) gene. Definition of a novel T cell-dependent response associated with genetic resistance to bacterial infection. J Exp Med 1989;170:145–61. [33] Singh RP, Patarca R, Schwartz J, Singh P, Cantor H. Definition of a specific interaction between the early T lymphocyte activation 1 (Eta-1) protein and murine macrophages in vitro and its effect upon macrophages in vivo. J Exp Med 1990;171:1931–42. [34] Ewing Jr EP T, akeuchi A, Shirai A, Osterman JV. Experimental infection of mouse peritoneal mesothelium with scrub typhus rickettsiae: an ultrastructural study. Infect Immun 1978;19:1068–75. [35] Rich KA, Burkett C, Webste P. Cytoplasmic bacteria can be targets for autophagy. Cell Microbiol 2003;5:455–68. [36] Yamaguchi N, Hasegawa T, Tamura A, Urakami H. Morphological analysis of growth inhibition of Rickettsia tsutsugamushi by tetracycline, chloramphenicol and erythromycin. Niigata Prefectural Hosp Med Bul 1984;32:38–40 [in Japanese]. [37] Aliberti JC, Machado FS, Gazzinelli RT, Teixeira MM, Silva JS. Platelet-activating factor induces nitric oxide synthesis in Trypanosoma cruzi-infected macrophages and mediates resistance to parasite infection in mice. Infect Immun 1999;67:2810–4. [38] Green SJ, Nacy CA, Schreiber RD, Granger DL, Crawford RM, Meltzer MS, Fortier AH. Neutralization of gamma interferon and tumor necrosis factor alpha blocks in vivo synthesis of nitrogen oxides from L-arginine and protection against Francisella tularensis infection in Mycobacterium bovis BCG-treated mice. Infect Immun 1993;61:689–98. [39] Goodrum KJ, Dierksheide J, Yoder BJ. Tumor necrosis factor alpha acts as an autocrine second signal with gamma interferon to induce nitric oxide in group B streptococcus-treated macrophages. Infect Immun 1995;63:3715–7. [40] Jerrells TR, Geng P. The role of tumor necrosis factor in host defense against scrub typhus rickettsiae. II. Differential induction of tumor necrosis factor-alpha production by Rickettsia tsutsugamushi and Rickettsia conorii. Microbiol Immunol 1994;38:713–9. [41] Urakami H, Tsuruhara T, Tamura A. Electron microscopic studies on intracellular multiplication of Rickettsia tsutsugamushi in L cells. Microbiol Immunol 1984;28:1191–201. [42] Tamura A, Urakami H. Easy method for infectivity titration of Rickettsia tsutsugamushi by infected cell counting. Nippon Saikingaku Zasshi 1981;36:783–5 [in Japanese]. [43] Ohashi N, Koyama Y, Urakami H, Fukuhara M, Tamura A, Kawamori F, Yamamoto S, Kasuya S, Yoshimura K. Demonstration of antigenic and genotypic variation in Orientia tsutsugamushi which were isolated in Japan, and their classification into type and subtype. Microbiol Immunol 1996;40:627–38.