Microbes and Infection 12 (2010) 864e874 www.elsevier.com/locate/micinf
Original article
Malaria exacerbates experimental mycobacterial infection in vitro and in vivo Michael Hawkes a,1, Xiaoming Li b,1, Maryanne Crockett a, Angelina Diassiti a, W. Conrad Liles a,c,d,e,f, Jun Liu b,**, Kevin C. Kain a,c,d,e,f,* a Institute of Medical Sciences, University of Toronto, Toronto, Canada Department of Molecular Genetics, University of Toronto, Toronto, Canada c Department of Medicine, University of Toronto, Toronto, Canada d Sandra A. Rotman Laboratories, McLaughlin-Rotman Centre for Global Health, Toronto, Canada e McLaughlin Centre for Molecular Medicine, Toronto, Canada f Tropical Disease Unit, Toronto General Hospital, Toronto, Canada b
Received 19 February 2010; accepted 31 May 2010 Available online 11 June 2010
Abstract Tuberculosis (Mtb) and malaria are among the most important infectious causes of morbidity and mortality worldwide, causing an estimated 1.5 million and 1 million deaths every year, respectively. Here we demonstrate a biological interaction between malaria and mycobacteria in vitro and in vivo. Murine macrophages co-incubated with Plasmodium falciparum parasitized erythrocytes demonstrated impaired control of intracellular Mtb replication, and reduced production of reactive nitrogen species in response to mycobacteria. Infection of C57BL/6 mice with Plasmodium species exacerbated the course of acute mycobacterial infection (57% increase in peak splenic CFU, p ¼ 0.043 for difference over time course of infection), induced disruption of the structural integrity of established granulomas, and caused reactivation of latent mycobacterial infection (2.6-fold increase in peak splenic CFU, p ¼ 0.016 for difference over time course of reactivation). Malaria pigment deposition within the granulomas of co-infected mice suggested that the influx of dysfunctional hemozoin-laden monocytes into the locus of mycobacterial control may contribute to impaired containment of mycobacteria. Collectively, these results point to malaria-induced dysregulation of innate and adaptive anti-mycobacterial defences, and suggest that the interaction of these globally important pathogens may potentiate Mtb infection and transmission. Ó 2010 Institut Pasteur. Published by Elsevier Masson SAS. All rights reserved. Keywords: Malaria; Tuberculosis; Mycobacteria; Co-infection
1. Introduction Tuberculosis and malaria are leading causes of infectious disease associated with morbidity and mortality globally. Mycobacterium tuberculosis (Mtb) infects one-third of the * Corresponding author. Eaton North 13-214, 200 Elizabeth Street, Toronto, ON M5G 2C4, Canada. Tel.: þ1 416 340 3535; fax: þ1 416 595 5826. ** Corresponding author at: 4382 Medical Sciences Building, 1 King’s College Circle, Toronto, ON M5S 1A8, Canada. Tel.: þ1 416 946 5067; fax: þ1 416 978 6885. E-mail addresses:
[email protected] (J. Liu),
[email protected] (K.C. Kain). 1 Contributed equally to this work.
world’s population and accounts for 1.5 million deaths annually, more than any other bacterial pathogen [1]. Plasmodium falciparum is the leading parasitic cause of mortality worldwide, causing approximately 300 million new infections and one million deaths per year [2]. Both diseases are endemic in tropical and impoverished areas of the world, and co-infection is likely to occur in individuals in these zones of intense transmission [3]. In addition to the socio-economic health determinants that account in part for the significant geographic overlap of both pathogens, biological interactions within the host may play a role in malariaetuberculosis co-infection. Establishing a role for P. falciparum in exacerbating the course of mycobacterial infection, analogous to its potentiating effect
1286-4579/$ - see front matter Ó 2010 Institut Pasteur. Published by Elsevier Masson SAS. All rights reserved. doi:10.1016/j.micinf.2010.05.013
M. Hawkes et al. / Microbes and Infection 12 (2010) 864e874
on the spread of HIV in Africa [4], could have important public health implications. Control of mycobacterial infection requires a highly coordinated host response that may be vulnerable to dysregulation by co-infecting pathogens. In human tuberculosis, following inhalation of aerosolized tubercle bacilli, alveolar macrophages internalize Mtb, initiating a cascade of cellular migratory events that lead to lymphohematogenous dissemination. In the majority of immunocompetent hosts, cell-mediated immunity results in the containment of mycobacteria; however, acute infection may occasionally progress to miliary or meningeal disease. Risk factors for disseminated disease are incompletely understood, but include factors associated with varying degrees of immune compromise such as young age, HIV-infection, and corticosteroid use [5]. Moreover, the protective effect of BCG vaccination against infant and childhood miliary and meningeal disease suggest a role for adaptive immunity in limiting distal spread of mycobacteria after primary exposure [6,7]. The histological hallmark of mycobacterial infection is the granuloma, a highly structured yet dynamic unit, comprised of a wide array of immune effector cells including macrophages, CD4þ and CD8þ T-lymphocytes, and B-lymphocytes acting in concert to restrict mycobacterial replication, while allowing viable mycobacteria to persist for decades [8]. Reactivation of bacterial replication from foci of latent infection is associated with disruption of the granuloma structure [9], and occurs in approximately 10% of infected hosts over their lifetime. Determinants of reactivation have not been completely defined, although recrudescence of chronic mycobacterial infection can be induced through depletion of CD4þ lymphocytes or blockade of the TH1 cytokine tumor necrosis factor (TNF) in experimental animal models [9e12], or by administration of corticosteroids or TNF neutralizing agents in humans [13]. Thus, challenges to host defences may affect the course of both acute disseminated mycobacterial infection and chronic or reactivation disease. Malaria has known immunomodulatory effects [14e17], which may impair host responses to mycobacterial infection. The macrophage, which plays a central role in the internalization and intracellular control of mycobacteria, is also the primary phagocyte for clearance of parasitized erythrocytes (PEs). Circulating monocytes as well as tissue macrophages and dendritic cells that have phagocytosed PEs accumulate toxic parasite hemozoin, with consequent functional impairment [18]. Furthermore, malaria alters the balance of circulating cytokines that are instrumental to the control of mycobacterial infection [19]. Given that malaria modulates elements of host immune response that play an important role in the control of mycobacterial infection, we hypothesized that malaria co-infection would exacerbate the course of acute and reactivation mycobacterial infection, and tested this hypothesis using murine infection models in vitro and in vivo. Several experimental systems have been used to study the pathogenesis and immunology of tuberculosis and malaria, each of which has its advantages and limitations. One wellestablished murine model of disseminated tuberculosis utilizes the closely related mycobacterial species Mycobacterium bovis BCG. C57BL/6 mice develop systemic infection
865
following BCG inoculation via the intraperitoneal route, developing the characteristic granulomatous lesions of human tuberculosis in multiple organs including the liver and spleen [8]. Mycobacterial replication is ultimately controlled by adaptive host immune responses, but viable bacilli persist at low levels in infected organs, mirroring the course of latent tuberculosis in humans [20,21]. Thus, both acute and latent tuberculous infection can be modeled with a single mycobacterial species in the same murine host. In keeping with clinical observations [13,22], cytokines including TNF and IFN-g are critical regulators of susceptibility and pathology in murine models of Mtb [23,24], and intracellular killing mechanisms including reactive nitrogen intermediates are conserved in both mice and humans [25e27]. In addition, infection of C57BL/6 mice with PccAS produces a transient parasitemia that recapitulates human blood stage malaria infection [19,21] and the immune dysregulation observed in humans is paralleled in murine malaria [28e30]. Thus, murine BCG and PccAS effectively simulate important immunological aspects of human tuberculosis and malaria, respectively. Here we show a deleterious effect of malaria co-infection on the course of experimental acute and reactivation mycobacterial infection in an established murine model of disseminated disease [12,20], and present in vitro and histopathological evidence that granuloma disruption and macrophage dysfunction secondary to ingested parasite hemozoin may contribute to this observation. 2. Materials and methods 2.1. Mice, mycobacteria and Plasmodium parasites C57BL/6 mice were bred and kept in the animal facility at the University of Toronto. Animal protocols were approved by the Animal Care Committee of the University of Toronto. M. tuberculosis strain H37Rv (TMC no. 102), M. bovis BCGPasteur strain, and Mycobacterium marinum type strain 1218R (ATCC 927) were grown at 37 C or 30 C in Middlebrook 7H9 broth (BD Biosciences; Franklin Lakes, NJ USA) supplemented with 0.2% glycerol and 10% OADC (Oleic Acid, Albumin, Dextrose, Catalase; BD Bioscience) or on Middlebrook 7H11 agar (BD Biosciences) supplemented with 0.5% glycerol and 10% OADC. Infections in experimental animals were initiated by intraperitoneal injection of 1.5 107 M. bovis BCG. Plasmodium chabaudi chabaudi AS (PccAS) parasites from stock frozen at 80 C were passaged through wild type C57BL/6 mice prior to intraperitoneal injection in experimental animals. P. falciparum (ITG and 3D7 strains) used for in vitro studies was cultured as previously described in Ref. [31]. Mature stage schizonts were isolated by density gradient centrifugation using 80% Percoll. A ratio of 20 PEs to each macrophage was used in all in vitro experiments. 2.2. In vivo infection and histopathology Mice (C57BL/6) were inoculated intra-peritoneally with 1.5 107 CFU of M. bovis BCG (day 0). For the acute
866
M. Hawkes et al. / Microbes and Infection 12 (2010) 864e874
co-infection model, mice were inoculated intra-peritoneally with 1.0 106 PccAS PEs 7 days later, such that the peak microbial burdens of mycobacteria and malaria would coincide temporally. For the reactivation experiments, latent mycobacterial infection was established by 84 days after IP inoculation of BCG, at which time mice were challenged with 1.0 106 PccAS PEs IP. Blood was collected by cardiac puncture and spleens, livers and lungs were harvested from mice after euthanasia by CO2 inhalation at various times over a course of infection. Half of each organ was homogenized and plated on 7H11 agar (BD Biosciences) at appropriate dilutions, and incubated for 21 days at 37 C in order to determine BCG CFUs. The remaining half was preserved in 10% formalin, embedded in paraffin, processed in 5 mm sections, and stained with H&E for histopathological analysis. Immunohistochemistry for the murine macrophage marker F4/ 80 was performed on paraffin embedded liver sections using a rat anti-mouse F4/80 IgG primary antibody (AbD Serotec; Raleigh, NC), followed by vector-blue conjugated goat anti-rat IgG secondary antibody (AbD Serotec), as previously described in Ref. [9]. Image analysis was performed by enumeration of malaria pigment (Adobe Photoshop) expressed as a proportion of the total pixel area [32]. 2.3. Infection of macrophages Thioglycollate-elicited macrophages from C57BL/6 mice were co-incubated with mycobacteria at a MOI of 10:1 (M. tuberculosis), 10:1 (M. bovis BCG) or 1:1 (M. marinum) for 3 h. Non-adherent cells were washed away. Following overnight incubation, cells were co-cultivated with parasitized erythrocytes (PEs) or uninfected erythrocytes (as controls). Co-infected macrophages were incubated in medium containing gentamicin (RPMI 1640, with 10% fetal bovine serum, and 2.5 mg/L gentamicin) at 37 C for 1, 3, 5 or 7 days. Cell lysates were plated on 7H11 medium (BD Biosciences) and mycobacterial colony forming units (CFUs) were quantified after incubation at 37 C for 21 days (M. tuberculosis and M. bovis BCG) or at 32 C for 7 days (M. marinum). 2.4. Cytokines and nitric oxide assay Blood was collected from euthanized mice by cardiac puncture, allowed to clot, and cleared by centrifugation. Serum was stored at 80 C and later assayed for cytokines using a cytometric bead array assay (Mouse Inflammation Kit, BD Biosciences) according to manufacturer’s instructions. Murine thioglycollate-elicited peritoneal macrophages were isolated from C57BL/6 mice, and plated at a density of 200,000 cells per well in 200 mL of media (RPMI 1640 supplemented with 10% fetal bovine serum). Macrophages were exposed to PEs or control conditions for 12e16 h, primed with interferon-gamma (IFN-g), then co-incubated with BCG or M. marinum (MOI of 50:1). Culture supernatants were collected 24 h later and assayed for nitrite concentration as an index of nitric oxide production using the Greiss reaction
(Greiss Reaction Kit, Promega) as well as TNF by ELISA (eBioscience; San Diego, CA). 3. Results 3.1. Co-infection with malaria exacerbates acute mycobacterial infection in vivo Given the known immunomodulatory effects of malaria infection, we hypothesized that malaria would exacerbate mycobacterial infection in a mammalian host. Using a well characterized experimental model of disseminated mycobacterial infection [3,21,33], we challenged C57BL/6 mice with BCG and quantified the burden of mycobacteria in the liver, spleen and lung over a time course of infection (63 days). Coinfection with PccAS was initiated intra-peritoneally 7 days after BCG infection such that the peak pathogen burden of mycobacteria and malaria parasites were co-incident. Control groups consisted of mice infected with BCG alone or PccAS alone. All mice survived the infection with one or both pathogens. As previously described in Ref. [3], BCG counts in the liver and spleen rose to peak levels two weeks after infection, declining thereafter to low but non-sterile levels (Fig. 1AeB). In mice co-infected with PccAS, higher mycobacterial counts were observed over the time course of infection in the spleen (Fig. 1A, p ¼ 0.043) and the liver (Fig. 1B, p ¼ 0.010). In addition, clinically relevant correlates of disease severity including weight loss (Fig. 1C, p < 0.0001) and splenomegaly (Fig. 1D, p ¼ 0.0034) were more pronounced in co-infected mice than mice infected with BCG alone. Liver mass was equivalent in both groups (Fig. 1E). TH1 cytokine responses are critical for granuloma formation and control of mycobacterial infection, with TNF and IFN-g playing a central role, as evidenced by the severity of disease observed in TNF- and IFN-g-deficient mice [23,24,34], in humans with inherited disorders of the IL-12-IFN-g axis [22] and individuals receiving therapeutic TNF antagonists [13]. We measured a panel of pro-and anti-inflammatory cytokines in the sera of co-infected mice and observed that TNF and IFN-g levels rose and declined in parallel to the mycobacterial burden (Fig. 1F and G). Co-infected mice had circulating TNF levels higher than mice infected with PccAS alone (Fig. 1F, p ¼ 0.036), but similar to BCG infected mice (Fig. 1F, p ¼ 0.31). The levels of IFN-g (Fig. 1G) as well as IL-6, IL-10, IL-12p70, or MCP-1 (data not shown) were not significantly different in co-infected mice and controls. 3.2. Parasite hemozoin co-localizes with macrophages within mycobacterial granulomas Intra-erythrocytic Plasmodium catabolizes host hemoglobin and sequesters the released toxic heme as crystalline hemozoin, which appears as a brown pigment under light microscopy [18]. PccAS infection led to hemozoin accumulation in reticulo-endothelial organs, resulting in darkly pigmented livers and spleens on gross examination, compared to the (normal) tan appearance of formalin-fixed livers from mice
M. Hawkes et al. / Microbes and Infection 12 (2010) 864e874
A
B
spleen BCG BCG+PccAS *
8.0
liver
5.0 total liver CFU (×10 )
6.0 4.0 2.0 0.0
4.0 3.0 2.0 1.0 0.0
1
7
10
14 21 35 time (days)
49
1
63
D
total body mass
14 21 35 time (days)
49
63
10
* liver mass (mg)
* 20
liver 2000
600 spleen mass (mg)
500 400 300 200
1500 1000 500
100
TNF
100
S cc A +P C G
+P PccAS BCG BCG+PccAS
G
50 25 0
IFN-γ
100
* IFN- γ (pg/mL)
75
B
B
C G
C G B
F
C G
S cc A
S cc A +P
B
0
C G
0
C G
0
B
total body mass (g)
10
E
spleen
700
30
TNF (pg/mL)
7
BCG PccAS
BCG PccAS
C
BCG BCG+PccAS *
B
total spleen CFU (×10 )
10.0
867
PccAS BCG BCG+PccAS
75 50 25 0
1
7
10
14 21 35 time (days)
49
63
BCG PccAS
1
7
10
14 21 35 time (days)
49
63
BCG PccAS
Fig. 1. Co-infection with murine malaria exacerbates acute mycobacterial infection in vivo. A and B. Bacillary loads in spleen (A; p ¼ 0.043, 2-way ANOVA) and liver (B; p ¼ 0.010, 2-way ANOVA) homogenates were significantly higher in C57BL/6 co-infected with BCG (1.5 107 CFU intra-peritoneally) and P. chabaudi chabaudi AS (PccAS; 1 106 parasites IP, 7 days later) than control mice infected with BCG alone. Data are shown as boxplots (median, inter-quartile range and range) with 4 mice per group at each time point. One representative experiment of two is shown. C. Weight loss was exacerbated in co-infected mice at peak of infection (day 14; 4 mice per group, p < 0.0001). D and E. Splenomegaly was more pronounced in co-infected mice at day 14 (D, *p ¼ 0.0034), but differences in liver weights were not statistically significant (E, p ¼ 0.068). F. Serum levels of TNF rose and declined in parallel with the mycobacterial burden, reaching a local maximum at days 14e21. TNF levels were higher in co-infected mice (grey boxplots) than in mice infected with PccAS (black boxplots, p ¼ 0.036, 2-way ANOVA), but similar to mice infected with BCG alone (white boxplots, p ¼ 0.31). Data are shown as boxplots (median, inter-quartile range, range) with 3e4 mice at each time point in each group. G. IFN-g levels during acute BCG infection and BCG þ PccAS co-infection also appear to be maximal at day 21, but no significant differences between groups were observed. *p < 0.05; TNF: tumor necrosis factor; IFN-g: interferon-gamma; BCG: M. bovis BCG; PccAS: Plasmodium chabaudi chabaudi AS.
infected with BCG alone (Fig. 2AeC). Histopathological examination of the livers of mice infected with PccAS demonstrated scattered deposition of pigment throughout the liver parenchyma (Fig. 2D). On the other hand, granulomas, the histological hallmark of mycobacterial infection, were apparent in liver sections of mice infected with BCG (Fig. 2E). Interestingly, in co-infected mice, enhanced pigment
deposition within granulomas was evident (Fig. 2F). The micro-architecture and size of granulomas was not affected by PccAS co-infection (Fig. 2GeI). Immunohistochemical staining for murine macrophages (F4/80 antigen) demonstrated co-localization of macrophage aggregates and malaria pigment in liver sections of co-infected mice (Fig. 2JeL). Quantitative morphometric analysis demonstrated similar
868
M. Hawkes et al. / Microbes and Infection 12 (2010) 864e874
Fig. 2. Malaria pigment co-localizes with macrophages within mycobacterial granulomas. AeC. Livers harvested at day 35 post-infection were darkly pigmented in PccAS infection (A), and BCG þ PccAS co-infection (C), but not in BCG infected mice (B). DeI. Histopathological examination of liver sections (day 35, H&E stain) revealed scattered pigment deposition (black arrowhead), representing parasite hemozoin, associated with PccAS infection (D and G), and granulomas associated with BCG infection (E and H). Of note, in co-infected animals, malaria pigment (black arrowhead) was observed within the granuloma, the locus of mycobacterial containment (F and I), although granuloma micro-architecture was not altered. JeL. Labeling the murine macrophage F4/80 antigen (blue) identified tissue resident macrophages (Kupffer cells) in association with pigment in PccAS infection (J), and macrophage aggregates (granulomas) in BCG infection (K) and BCG þ PccAS co-infection (L) which co-localized with malaria pigment (L). M. Granuloma size was equivalent in BCG and BCG þ PccAS coinfected mice ( p ¼ 0.66). N. Quantification of pigment using image analysis demonstrated that pigment density, expressed as a fraction of cross-sectional area, was higher within the granuloma than surrounding liver parenchyma, and well above the negligible levels in granulomas of mice infected with BCG alone ( p < 0.0001). Each point represents one granuloma or surrounding parenchyma, bars represent median; a total of 100 granulomas from 4 mice in each group were analysed. *p < 0.05; BCG: M. bovis BCG; PccAS: Plasmodium chabaudi chabaudi AS.
M. Hawkes et al. / Microbes and Infection 12 (2010) 864e874
granuloma size (Fig. 2M) but increased levels of pigment, concentrated within the granulomas of co-infected animals, above background levels in the surrounding liver parenchyma (Fig. 2N, p < 0.0001), and well above the negligible level in control images of granulomas of BCG infected mice (Fig. 2N, p < 0.0001). This result suggests that the influx of dysfunctional hemozoin-laden monocytes into granulomas may contribute to impaired containment of mycobacteria.
869
Latent infection with M. bovis BCG was established in C57BL/6 mice by 49 days after intraperitoneal inoculation, as evidenced by stably low levels of viable mycobacteria in the homogenized organs of infected mice (Fig. 1A, B). Latently infected animals were challenged with PccAS at day 84, resulting in a significant rise in the mycobacterial counts in the spleen relative to control mice over 4 weeks following infection with PccAS (Fig. 3A, p ¼ 0.036). All mice survived and ultimately controlled the reactivation mycobacterial infection, and the secondary (reactivation) peak mycobacterial burden was more than 10-fold lower than the primary (acute) peak mycobacterial burden. Body mass was reduced (Fig. 3B), spleen mass was increased (Fig. 3C), and liver mass was reduced (Fig. 3D) in mice challenged with PccAS.
3.3. Malaria reactivates latent mycobacterial infection in a murine model Up to one-third of the world’s population is latently infected with M. tuberculosis, which can resume replication from a dormant state to cause active disease in approximately 10% of infected persons over their lifetime. Reactivation of Mtb can be experimentally triggered in mice through depletion of CD4þ lymphocytes or TNF blockade [9e11]; however, natural precipitating factors of reactivation are not well defined. We examined whether malaria, a prevalent infection in many TB endemic areas, can trigger reactivation of latent mycobacterial infection in vivo.
3.4. Malaria-induced reactivation of latent mycobacterial infection is characterized by hepatic and splenic inflammation and disorganization of granulomas Reactivation tuberculosis is associated with disruption of the ordered architecture of the granuloma [9]. We examined
A total spleen CFU (×104)
5.0
latent BCG PccAS challenge
4.0
*
3.0 2.0 1.0 0.0 84
91
98
105
time (days) BCG
10 5 0
500 250
ch al le ng e
G
S
Pc cA
la te nt B C
ch al le ng e
G
S
Pc cA
500
0
0
la te nt B C
1000
ch al le ng e
15
*
G
20
750
S
25
1500
* liver mass (mg)
*
spleen mass (mg)
1000
30
Pc cA
35
total body mass (g)
D
C
la te nt B C
B
PccAS
Fig. 3. Malaria co-infection induces reactivation of latent mycobacterial infection in vivo. A. In a murine model of latent mycobacterial reactivation, mice were infected with M. bovis BCG, and followed for 84 days, at which time mycobacteria are controlled, but not eliminated, by adaptive immune mechanisms. Mice were then infected with PccAS by IP injection (1 106 parasites), resulting in a transient rise in viable mycobacteria isolated from the spleens of co-infected mice ( p ¼ 0.016). One representative experiment of two is shown. Data are displayed as boxplots (median, inter-quartile range, range). BeD. Weight loss (B, p ¼ 0.0039) and splenomegaly (C, p < 0.0001) were more pronounced, and liver mass was reduced (D, p ¼ 0.0006) in mice challenged with PccAS at peak of reactivation (day 98). *p < 0.05; BCG: M. bovis BCG; PccAS: Plasmodium chabaudi chabaudi AS.
870
M. Hawkes et al. / Microbes and Infection 12 (2010) 864e874
histopathological liver sections during malaria-induced reactivation of BCG infection and observed diffuse mononuclear cell infiltration, as previously described with other forms of inflammatory challenge in BCG infected mice [35,36] (Fig. 4AeD). Poorly organized, loose collections of mononuclear cells were observed, often in association with malaria pigment, compared to the compact pigment-free granulomas of mice with latent BCG (Fig. 4C,D). In spleens of mice with latent BCG infection, typical histological features were observed including lymphoid follicles and intervening red pulp (Fig. 4E) in contrast to mice challenged with PccAS, where intense mononuclear cell infiltrate obscured the splenic micro-architecture (Fig. 4F). Morphometric analysis revealed more numerous (Fig. 4G, p ¼ 0.0079) and smaller (Fig. 4H, I, p < 0.0001) mononuclear cell aggregates in co-infected mice, reflecting a generalized inflammatory infiltrate [35] rather than the quiescent chronic granulomatous inflammation observed in mice latently infected with BCG alone. 3.5. Co-cultivation with P. falciparum parasitized erythrocytes impairs the control of M. tuberculosis replication by macrophages Having observed that malaria exacerbates mycobacterial infection in vivo, we next explored the mechanism underlying this phenomenon using macrophage co-cultivation experiments in vitro. The macrophage is a principal effector cell for the control of both malaria and mycobacteria. We hypothesized that internalization of malaria parasitized erythrocytes by macrophages would impair their ability to restrict intracellular mycobacterial replication. After 10 days co-cultivation in vitro, Mtb proliferated to levels approximately two-fold higher in macrophages co-cultivated with either of two laboratory clones of P. falciparum than control macrophages co-incubated with uninfected erythrocytes (Fig. 5, 3D7 and ITG, p ¼ 0.0097 and 0.0073, respectively). To examine whether these findings were generalizable to other mycobacterial species, we repeated co-cultivation experiments with M. marinum and observed similar results (Fig. S1AeF). 3.6. P. falciparum alters macrophage-mediated anti-mycobacterial defences We next investigated possible mechanisms for macrophage dysfunction that might account for the impaired mycobacterial control. We hypothesized that intracellular killing mechanisms may be impaired, and examined the production of nitric oxide, an important effector molecule for macrophage antimicrobial activity [26]. In vitro, murine macrophages co-cultivated with BCG (MOI 50:1) produced higher levels of nitric oxide than macrophages treated with media alone (Fig. 6A). Nitric oxide production was diminished in macrophages co-incubated with PEs but not uninfected erythrocytes (Fig. 6A). Similar results were observed using M. marinum (MOI 50:1) as the mycobacterial stimulus (data not shown). Under similar conditions, TNF production in response to BCG was not affected by co-incubation with PEs (Fig. 6B).
4. Discussion Here we show that malaria co-infection exacerbates the course of mycobacterial infection in vitro and in vivo. We provide histopathological evidence that malaria pigment co-localizes with macrophages within the granuloma, and that macrophages ingesting malaria parasites or hemozoin have defects in nitric oxide production and in their ability to restrict intracellular mycobacterial growth. Taken together, these findings provide a cohesive description of innate and adaptive immune dysregulation in the setting of malaria infection with consequences for the pathogenesis of experimental mycobacterial infection. Malaria co-infection is known to exacerbate viral [37,38], parasitic [28], and intracellular bacterial [29,39] infections in experimental animals and humans. Malaria impairs humoral and cell-mediated immune responses to heterologous vaccines [14,40] and other antigens, including the mycobacterial purified protein derivative [16,17]. Plasmodium yoelii 17XL co-infection has been previously shown to increase bacillary load in experimental murine infection with Mtb [33]. Our results extend these findings in a distinct model system using the malaria species PccAS and the mycobacterial strain M. bovis BCG, with intraperitoneal initiation of systemic infection. The histopathological observation of pigmented granulomas in co-infected mice led us to postulate that trafficking of hemozoin-laden monocytes or macrophages to the granuloma may result in loss of mycobacterial control. The granuloma is a highly dynamic unit, as demonstrated by [3H]thymidine labelling of macrophage bone marrow precursors to track their flux in and out of rabbit skin granulomas [41] and intravital microscopy of nascent granulomas in transparent zebrafish embryos [42] as well as BCG-induced murine liver granulomas [12]. Previous investigators have shown that super-infecting mycobacteria home to existing granulomas in zebrafish [43], amphibian [43] and mammalian [44] models, transported as intracellular cargo within macrophages. Although macrophages harbouring mycobacteria seem preferentially directed to the granuloma, antigenically heterologous latex beads and Salmonella enterica subsp. arizonae were also carried to pre-formed granulomas [43]. By extension, we propose that circulating monocytes and/or tissue resident macrophages bind and internalize PEs, acquiring and accumulating hemozoin, and subsequently transit to established foci of chronic inflammation [42e44]. Influx into the granuloma of hemozoin-containing monocytes may thus disturb the hostepathogen equilibrium that maintains mycobacterial quiescence. In our model of mycobacterial reactivation following malaria challenge, we observed increased mycobacterial burden in infected organs, in association with hepatic and splenic mononuclear infiltrates, which contrasted markedly with the structured granulomas seen in latent BCG. Previous studies have documented hypersensitivity hepatitis and splenic injury in mice infected with BCG and subsequently challenged with various inflammatory stimuli, in a process that is dependent on IFN-g and associated with intense TNF production [34,35,45].
M. Hawkes et al. / Microbes and Infection 12 (2010) 864e874
871
Fig. 4. Malaria-induced reactivation of latent mycobacterial infection is characterized by granuloma disruption. AeD. The compact, organized granulomas in livers of mice with latent BCG (A and C) are contrasted with mice challenged with PccAS (B and D). Disordered aggregates of mononuclear cells were observed in association with malaria pigment (black arrowhead), suggesting a breakdown of granuloma architecture. E and F. The spleens of mice with latent BCG show typical lymphoid follicles (arrowhead) and red pulp (E) compared with the intense mononuclear infiltrate in spleens of mice challenged with PccAS (F). GeI. In contrast to the compact granulomas in livers of mice latently infected with BCG, mononuclear cell aggregates in mice challenged with PccAS were scattered, ill-defined and variable in size. Morphometric analysis demonstrated more numerous (G, p ¼ 0.0079), and smaller (H, p < 0.0001 and I, p < 0.0001) mononuclear cell aggregates in contrast to the structured granulomas of latent BCG infection. All mononuclear cell aggregates (collection of >10 mononuclear cells) from eighty random low-power (10) fields of liver sections from 4 mice per group were counted (>250 aggregates). *p < 0.05; BCG: M. bovis BCG; PccAS: Plasmodium chabaudi chabaudi AS. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)
872
M. Hawkes et al. / Microbes and Infection 12 (2010) 864e874
*
Mtb CFU ( × 104/mL)
15
control replication of Mtb in vitro. Similarly, previous investigators have shown that macrophages internalizing parasitized erythrocytes have reduced microbicidal activity against bacterial and fungal pathogens [46]. The parasite by-product hemozoin alters macrophage function in vitro, including inhibitory effects on respiratory burst, NADPH oxidase activity, protein kinase C activity, expression of ICAM-1, integrin CD11c, MHC class-II expression, and differentiation and maturation into functional dendritic cells [18]. Furthermore, hemozoin and its synthetic analog b-hematin inhibit nitric oxide synthesis by murine macrophages in response to inflammatory stimuli [30,47e49]. We extend these findings to mycobacteriaemalaria co-infection in vitro, showing that PEs inhibit nitric oxide synthesis by murine macrophages in response to stimulation with BCG. This may contribute to the impaired macrophage anti-mycobacterial defences observed in vitro and possibly in vivo, particularly within the granuloma, where hemozoin-laden macrophages were noted. Despite speculation based on previous work [33], we did not find differences in TNF production by co-infected macrophages in vitro, or serum TNF or IFN-g in co-infected mice. Thus, differences in mycobacterial control in the context of malaria coinfection do not appear to be due to major alterations in the TH1 cytokine profile. Despite the well-recognized role of TNF in host defence against mycobacteria [9,10,13,50], other independentlyregulated processes including nitric oxide synthesis may account for deficiencies in anti-mycobacterial defences in the setting of malaria co-infection. Our experimental models have several limitations. In attempting to recapitulate human tuberculous disease, murine models are imperfect. Relevant examples include the non-caseating BCG granuloma in the mouse, whereas human granulomas form necrotic caseum, which may play a role in bacterial persistence [51]. The aerogenic route is the common mode of transmission for tuberculosis in humans whereas we used an intraperitoneal route to simulate systemic disease. Parasitemias in excess of those seen in typical human disease may limit the ability to generalize findings in PccAS infection to human malaria.
*
10
5
0 media RBC
3D7
ITG
P.falciparum PEs Fig. 5. Co-incubation with P. falciparum inhibits innate control of M. tuberculosis by macrophages in vitro. A. Co-incubation with P. falciparum parasitized erythrocytes (PEs; laboratory strains 3D7 and ITG, grey bars) led to increased growth of M. tuberculosis (MOI ¼ 10:1) within murine macrophages. Significant differences ( p ¼ 0.0097 and p ¼ 0.0073 for 3D7 and ITG, respectively) were observed in the mycobacterial counts 10 days after infection, compared with control macrophages infected with Mtb and cultivated with uninfected erythrocytes (white bar). Data are shown as mean SEM (4 replicate wells at each time point). *p < 0.05; PE: P. falciparum parasitized erythrocytes; RBC: red blood cell.
Given the critical role of tightly coordinated local cytokine and cellular networks in maintaining granuloma integrity and mycobacterial dormancy [12], we postulate that the generalized inflammation associated with inflammatory challenge may disrupt ordered host defences, allowing mycobacterial replication to resume, albeit temporarily, as reflected by disorganized granulomas and transiently elevated burden of viable mycobacteria in our study. This finding contrasts with the effect of malaria on acute mycobacterial infection, where we hypothesize that influx of dysfunctional hemozoin-laden macrophages leads to impaired mycobacterial control without alterations in granuloma micro-architecture. We demonstrated that macrophages phagocytes ingesting parasitized erythrocytes are dysfunctional in their ability to
A
B 7000
15
10
*
5
TNF (pg/mL)
nitrite ( μ M)
6000 5000 4000 3000 2000 1000 0
PE
C G G +R B C B C G +P E
B
C
B
R
B
C
a ed i m
+P
E
C B
C G B
+R
C G B
C G B
PE
B C R
m
ed
ia
0
Fig. 6. P. falciparum alters anti-mycobacterial defences in murine macrophages in vitro. A. Defective nitric oxide production by murine macrophages in malaria-BCG co-infection. Reactive nitrogen species were assayed in macrophage culture supernatants (Greiss reaction) in response to uninfected erythrocytes (RBC, light grey), P. falciparum PEs (black) (MOI ¼ 20:1), BCG (MOI ¼ 50:1), individually and together (dark grey). Nitric oxide synthesis in response to BCG was diminished in macrophages co-incubated with PEs ( p < 0.0001 for BCG þ PE vs BCG þ RBC). B. Levels of the pro-inflammatory cytokine TNF assayed in culture supernatants (ELISA) were unchanged during co-infection. *p < 0.05; PEs: P. falciparum parasitized erythrocytes; RBC: red blood cell; BCG: M. bovis BCG.
M. Hawkes et al. / Microbes and Infection 12 (2010) 864e874
Nonetheless, our findings of a biologically plausible malariaemycobacterial interaction suggest that malaria may play an underappreciated role in the course of acute and reactivation infection with M. tuberculosis. Given that 300 million clinical episodes of malaria occur annually, in areas of the world where tuberculosis prevalence frequently exceeds one in three individuals, these findings may have important implications for public health globally. Acknowledgements This study was funded in part by a CIHR Team Grant in Malaria (KCK), CIHR MT-13721 (KCK), Genome Canada through the Ontario Genomics Institute (KCK), CIHR MOP15107 (JL), MOP-82772 (JL), and CIHR Canada Research Chairs (WCL, KCK). The funding agencies had no role in study design, data collection and analysis, decision to publish or preparation of the manuscript.
Appendix. Supplementary material Murine macrophage-like RAW 264.7 cells were plated at a concentration of 5 104 cells/well in 12-well polystyrene plates and maintained at 37 C in 5% CO2 in RPMI 1640 (BD Biosciences) supplemented with 10% heat inactivated fetal bovine serum (BD Biosciences). Cells were infected with M. marinum at a multiplicity of infection (MOI) of 1:1, and 3 h later were thoroughly washed with RPMI and pulsed for 2 h with amikacin (200 mg/mL; Sigma; St. Louis, MO) in order to eliminate extracellular mycobacteria. Following overnight incubation, cells were co-cultivated with parasitized erythrocytes (PEs) or uninfected erythrocytes (as controls) and cell lysates were plated on 7H11 culture medium at 30 C after varying durations of incubation (1, 2, 3 and 4 days) for mycobacterial colony counts after 7-days incubation. The supplementary data associated with this article can be found in the online version at doi:10.1016/j.micinf.2010.05. 013. References [1] World Health Organization, WHO report 2008. Global tuberculosis control - surveillance, planning, financing. www.who.int/tb. [2] World Health Organization, Global Malaria Programme. http://www. who.int/malaria [3] K.R. Page, A.E. Jedlicka, B. Fakheri, G.S. Noland, A.K. Kesavan, A.L. Scott, N. Kumar, Y.C. Manabe, Mycobacterium-induced potentiation of type 1 immune responses and protection against malaria are host specific. Infect. Immun. 73 (2005) 8369e8380. [4] L.J. Abu-Raddad, P. Patnaik, J.G. Kublin, Dual infection with HIV and malaria fuels the spread of both diseases in sub-Saharan Africa. Science 314 (2006) 1603e1606. [5] Z. Yang, Y. Kong, F. Wilson, B. Foxman, A.H. Fowler, C.F. Marrs, M.D. Cave, J.H. Bates, Identification of risk factors for extrapulmonary tuberculosis. Clin. Infect. Dis. 38 (2004) 199e205. [6] L.C. Rodrigues, V.K. Diwan, J.G. Wheeler, Protective effect of BCG against tuberculous meningitis and miliary tuberculosis: a meta-analysis. Int. J. Epidemiol. 22 (1993) 1154e1158.
873
[7] G.A. Colditz, C.S. Berkey, F. Mosteller, T.F. Brewer, M.E. Wilson, E. Burdick, H.V. Fineberg, The efficacy of bacillus CalmetteeGuerin vaccination of newborns and infants in the prevention of tuberculosis: meta-analyses of the published literature. Pediatrics 96 (1995) 29e35. [8] D.O. Co, L.H. Hogan, S.I. Kim, M. Sandor, Mycobacterial granulomas: keys to a long-lasting host-pathogen relationship. Clin. Immunol. 113 (2004) 130e136. [9] S.D. Chakravarty, G. Zhu, M.C. Tsai, V.P. Mohan, S. Marino, D.E. Kirschner, L. Huang, J. Flynn, J. Chan, Tumor necrosis factor blockade in chronic murine tuberculosis enhances granulomatous inflammation and disorganizes granulomas in the lungs. Infect. Immun. 76 (2008) 916e926. [10] V.P. Mohan, C.A. Scanga, K. Yu, H.M. Scott, K.E. Tanaka, E. Tsang, M. M. Tsai, J.L. Flynn, J. Chan, Effects of tumor necrosis factor alpha on host immune response in chronic persistent tuberculosis: possible role for limiting pathology. Infect. Immun. 69 (2001) 1847e1855. [11] C.A. Scanga, V.P. Mohan, K. Yu, H. Joseph, K. Tanaka, J. Chan, J.L. Flynn, Depletion of CD4(þ) T cells causes reactivation of murine persistent tuberculosis despite continued expression of interferon gamma and nitric oxide synthase 2. J. Exp. Med. 192 (2000) 347e358. [12] J.G. Egen, A.G. Rothfuchs, C.G. Feng, N. Winter, A. Sher, R.N. Germain, Macrophage and T cell dynamics during the development and disintegration of mycobacterial granulomas. Immunity 28 (2008) 271e284. [13] J. Keane, S. Gershon, R.P. Wise, E. Mirabile-Levens, J. Kasznica, W.D. Schwieterman, J.N. Siegel, M.M. Braun, Tuberculosis associated with infliximab, a tumor necrosis factor alpha-neutralizing agent. N. Engl. J. Med. 345 (2001) 1098e1104. [14] W.A. Williamson, B.M. Greenwood, Impairment of the immune response to vaccination after acute malaria. Lancet 1 (1978) 1328e1329. [15] T.G. Theander, L. Hviid, Y.A. Abu-Zeid, N.H. Abdulhadi, B.O. Saeed, P.H. Jakobsen, C.M. Reimert, S. Jepsen, R.A. Bayoumi, J.B. Jensen, Reduced cellular immune reactivity in healthy individuals during the malaria transmission season. Immunol. Lett. 25 (1990) 237e242. [16] L. Hviid, T.G. Theander, Y.A. Abu-Zeid, N.H. Abdulhadi, P.H. Jakobsen, B.O. Saeed, S. Jepsen, R.A. Bayoumi, J.B. Jensen, Loss of cellular immune reactivity during acute Plasmodium falciparum malaria. FEMS. Microbiol. Immunol. 3 (1991) 219e227. [17] T.G. Theander, M. Svenson, I.C. Bygbjerg, A. Kharazmi, S. Jepsen, B.J. Andersen, P.B. Larsen, Inhibition of human lymphocyte proliferative response by serum from Plasmodium falciparum infected patients. Acta Pathol. Microbiol. Immunol. Scand. C. (1987) 257e263. [18] E. Schwarzer, O.A. Skorokhod, V. Barrera, P. Arese, Hemozoin and the human monocyte e a brief review of their interactions. Parassitologia 50 (2008) 143e145. [19] J. Langhorne, S.J. Quin, L.A. Sanni, Mouse models of blood-stage malaria infections: immune responses and cytokines involved in protection and pathology. Chem. Immunol. 80 (2002) 204e228. [20] R.V. Blanden, M.J. Lefford, G.B. Mackaness, The host response to CalmetteeGuerin bacillus infection in mice. J. Exp. Med. 129 (1969) 1079e1107. [21] A.L. Leisewitz, K. Rockett, D. Kwiatkowski, BCG-malaria co-Infection has paradoxical effects on C57BL/6 and A/J mouse strains. Parasite Immunol. 30 (2008) 1e12. [22] J.L. Casanova, L. Abel, Genetic dissection of immunity to mycobacteria: the human model. Annu. Rev. Immunol. 20 (2002) 581e620. [23] J.L. Flynn, J. Chan, K.J. Triebold, D.K. Dalton, T.A. Stewart, B.R. Bloom, An essential role for interferon gamma in resistance to Mycobacterium tuberculosis infection. J. Exp. Med. 178 (1993) 2249e2254. [24] J.L. Flynn, M.M. Goldstein, J. Chan, K.J. Triebold, K. Pfeffer, C.J. Lowenstein, R. Schreiber, T.W. Mak, B.R. Bloom, Tumor necrosis factor-alpha is required in the protective immune response against Mycobacterium tuberculosis in mice. Immunity 2 (1995) 561e572. [25] J.L. Flynn, C.A. Scanga, K.E. Tanaka, J. Chan, Effects of aminoguanidine on latent murine tuberculosis. J. Immunol. 160 (1998) 1796e1803.
874
M. Hawkes et al. / Microbes and Infection 12 (2010) 864e874
[26] E.D. Chan, J. Chan, N.W. Schluger, What is the role of nitric oxide in murine and human host defense against tuberculosis? Current knowledge. Am. J. Respir. Cell Mol. Biol. 25 (2001) 606e612. [27] J. Chan, K. Tanaka, D. Carroll, J. Flynn, B.R. Bloom, Effects of nitric oxide synthase inhibitors on murine infection with Mycobacterium tuberculosis. Infect. Immun. 63 (1995) 736e740. [28] R.S. Phillips, G.R. Selby, D. Wakelin, The effect of Plasmodium berghei and Trypanosoma brucei infections on the immune expulsion of the nematode Trichuris muris from mice. Int. J. Parasitol. 4 (1974) 409e415. [29] J.R. Murphy, M.J. Lefford, Host defenses in murine malaria: evaluation of the mechanisms of immunity to Plasmodium yoelii infection. Infect. Immun. 23 (1979) 384e391. [30] J. Prada, J. Malinowski, S. Muller, U. Bienzle, P.G. Kremsner, Effects of Plasmodium vinckei hemozoin on the production of oxygen radicals and nitrogen oxides in murine macrophages. Am. J. Trop. Med. Hyg. 54 (1996) 620e624. [31] W. Trager, J.B. Jensen, Human malaria parasites in continuous culture. Science 193 (1976) 673e675. [32] L. Mulrane, E. Rexhepaj, S. Penney, J.J. Callanan, W.M. Gallagher, Automated image analysis in histopathology: a valuable tool in medical diagnostics. Expert Rev. Mol. Diagn. 8 (2008) 707e725. [33] C.P. Scott, N. Kumar, W.R. Bishai, Y.C. Manabe, Short report: modulation of Mycobacterium tuberculosis infection by Plasmodium in the murine model. Am. J. Trop. Med. Hyg. 70 (2004) 144e148. [34] R. Kamijo, J. Le, D. Shapiro, E.A. Havell, S. Huang, M. Aguet, M. Bosland, J. Vilcek, Mice that lack the interferon-gamma receptor have profoundly altered responses to infection with Bacillus CalmetteeGuerin and subsequent challenge with lipopolysaccharide. J. Exp. Med. 178 (1993) 1435e1440. [35] J. Ferluga, Tuberculin hypersensitivity hepatitis in mice infected with Mycobacterium bovis (BCG). Am. J. Pathol. 105 (1981) 82e90. [36] J.W. Shands Jr., V.C. Senterfitt, Endotoxin-induced hepatic damage in BCG-infected mice. Am. J. Pathol. 67 (1972) 23e40. [37] A.M. Moormann, K. Chelimo, O.P. Sumba, M.L. Lutzke, R. PloutzSnyder, D. Newton, J. Kazura, R. Rochford, Exposure to holoendemic malaria results in elevated Epstein-Barr virus loads in children. J. Infect. Dis. 191 (2005) 1233e1238. [38] M.R. Thursz, D. Kwiatkowski, M.E. Torok, C.E. Allsopp, B.M. Greenwood, H.C. Whittle, H.C. Thomas, A.V. Hill, Association of hepatitis B surface antigen carriage with severe malaria in Gambian children. Nat. Med. 1 (1995) 374e375. [39] D.C. Mabey, A. Brown, B.M. Greenwood, Plasmodium falciparum malaria and Salmonella infections in Gambian children. J. Infect. Dis. 155 (1987) 1319e1321.
[40] B.M. Greenwood, A.K. Bradley, I.S. Blakebrough, H.C. Whittle, T.F. Marshall, H.M. Gilles, The immune response to a meningococcal polysaccharide vaccine in an African village,. Trans. R Soc. Trop. Med. Hyg. 74 (1980) 340e346. [41] A.M. Dannenberg Jr., Macrophage turnover, division and activation within developing, peak and “healed” tuberculous lesions produced in rabbits by BCG. Tuberculosis (Edinb) 83 (2003) 251e260. [42] J.M. Davis, L. Ramakrishnan, The role of the granuloma in expansion and dissemination of early tuberculous infection. Cell 136 (2009) 37e49. [43] C.L. Cosma, O. Humbert, L. Ramakrishnan, Superinfecting mycobacteria home to established tuberculous granulomas. Nat. Immunol. 5 (2004) 828e835. [44] C.L. Cosma, O. Humbert, D.R. Sherman, L. Ramakrishnan, Trafficking of superinfecting Mycobacterium organisms into established granulomas occurs in mammals and is independent of the Erp and ESX-1 mycobacterial virulence loci. J. Infect. Dis. 198 (2008) 1851e1855. [45] R. Guler, M.L. Olleros, D. Vesin, R. Parapanov, C. Vesin, S. Kantengwa, L. Rubbia-Brandt, N. Mensi, A. Angelillo-Scherrer, E. Martinez-Soria, F. Tacchini-Cottier, I. Garcia, Inhibition of inducible nitric oxide synthase protects against liver injury induced by mycobacterial infection and endotoxins. J. Hepatol. 41 (2004) 773e781. [46] P.L. Fiori, P. Rappelli, S.N. Mirkarimi, H. Ginsburg, P. Cappuccinelli, F. Turrini, Reduced microbicidal and anti-tumour activities of human monocytes after ingestion of Plasmodium falciparum-infected red blood cells. Parasite Immunol. 15 (1993) 647e655. [47] N. Morakote, D.E. Justus, Immunosuppression in malaria: effect of hemozoin produced by Plasmodium berghei and Plasmodium falciparum. Int. Arch. Allergy Appl. Immunol. 86 (1988) 28e34. [48] D. Taramelli, N. Basilico, E. Pagani, R. Grande, D. Monti, M. Ghione, P. Olliaro, The heme moiety of malaria pigment (beta-hematin) mediates the inhibition of nitric oxide and tumor necrosis factor-alpha production by lipopolysaccharide-stimulated macrophages. Exp. Parasitol. 81 (1995) 501e511. [49] D. Taramelli, S. Recalcati, N. Basilico, P. Olliaro, G. Cairo, Macrophage preconditioning with synthetic malaria pigment reduces cytokine production via heme iron-dependent oxidative stress. Lab. Invest. 80 (2000) 1781e1788. [50] R.S. Wallis, M.S. Broder, J.Y. Wong, M.E. Hanson, D.O. Beenhouwer, Granulomatous infectious diseases associated with tumor necrosis factor antagonists. Clin. Infect. Dis. 38 (2004) 1261e1265. [51] B.M. Saunders, W.J. Britton, Life and death in the granuloma: immunopathology of tuberculosis. Immunol. Cell Biol. 85 (2007) 103e111.
