ABSTRACT

INTRODUCTION

Burkholderia pseudomallei is a Gram-negative soil bacterium that causes the potentially fatal disease melioidosis in humans and animals (1, 2). B. pseudomallei is found in the tropics worldwide, with the majority of infections being reported in Southeast Asia and Northern Australia (3). A recent study examining documented human and animal melioidosis cases as well as the environmental presence of B. pseudomallei estimates that the disease is underreported in 45 countries in which it is endemic (4). Underreporting may occur because of the wide variety of manifestations of melioidosis complicating diagnosis of the disease. A common clinical manifestation of an infection with B. pseudomallei is pneumonia (5,–7). However, the capability of B. pseudomallei to infect virtually every tissue produces many other manifestations such as encephalitis, osteomyelitis, and abscesses of the skin and internal organs (5).

B. pseudomallei is a facultative intracellular bacterium that enters and survives in nonphagocytic and phagocytic cells, including neutrophils (8,–14). Following entry into the host cell, the bacteria escape from the endocytic vacuole into the cytosol, where they exhibit intracellular motility based on actin polymerization and flagella (15,–17). A characteristic feature of the intracellular life cycle of B. pseudomallei is the formation of multinucleated giant cells (MNGCs), which result from the fusion of an infected mononuclear cell with one or more neighboring cells (18). Although the physiological role of B. pseudomallei-induced host cell fusion remains to be elucidated, it is conceivable that MNGC formation facilitates localized dissemination of B. pseudomallei, access to nutrients, and escape from extracellular immune defense mechanisms and antibiotics. MNGCs have been observed in vivo in autopsy lung tissue samples of melioidosis patients and in pulmonary lesions of C57BL/6 mice challenged with a low dose of B. pseudomallei (19, 20). Furthermore, a previous analysis of B. pseudomallei-infected cockroaches indicates MNGC formation of hemocytes (21). This finding may suggest that the bacteria are able to induce the fusion of leukocytes during bacteremia in humans.

While the identity of host cellular factors involved in MNGC formation is as yet unknown, recent work has demonstrated that the bacterial type VI secretion system (T6SS) has a central role in this process. The T6SS is a bacteriophage-related secretion apparatus that translocates effector proteins into bacteria or eukaryotic cells by a contraction-based mechanism (22,–25). Inactivation of T6SS-5 (also named T6SS-1) of B. pseudomallei abrogates MNGC formation (17, 26, 27). Similarly, a homologue of this T6SS-5 present in the closely related species Burkholderia thailandensis is essential for host cell fusions (28, 29). B. thailandensis shares the same intracellular life cycle and is used as a low-virulence model organism to study B. pseudomallei-host interactions. Disruption of the T6SS-5 causes a virulence defect of both species in animal models of infection (27, 30, 31). A recent study revealed that the expression of the conserved T6SS-5 gene hcp-5 (and likely therefore the whole T6SS-5 gene cluster) is triggered by the glutathione present in the cytoplasm of the host cell (32). Glutathione is synthesized in all cells of the human body, suggesting that production of T6SS-5 proteins is not cell type specific and may occur in all tissues during infection. Importantly, however, the signal(s) for induction of T6SS-5 effector translocation, presumably via an as yet unidentified host-derived molecule(s) or via contact with host cell membranes, remains elusive. Likewise, the host factor(s) that is targeted by the T6SS-5 to mediate host cell fusions is not known.

The range of host cells that are infected by B. pseudomallei and support intracellular survival during in vivo infection has not been fully explored. In vitro, the bacteria actively invade various nonphagocytic cells. However, to the best of our knowledge, this notion is based on findings obtained with immortalized cells except for two studies using primary alveolar epithelial cells (8, 33,–36). Furthermore, the ability of B. pseudomallei to infect every tissue and the ubiquitous presence of glutathione in human cells raises the question of cell type specificity of T6SS-5 activity and therefore of MNGC formation. Although MNGC formation in B. pseudomallei-infected cells was first described 20 years ago, surprisingly little is known about the types of host cells stimulated to form cell-cell fusions following infection (9). To the best of our knowledge, B. pseudomallei and B. thailandensis T6SS-5-induced MNGC formation has been analyzed only in assays using epithelial, fibroblast, and monocyte/macrophage-like cell lines or HEK293 cells of unknown type (9, 17, 18, 37,–40). In the present study, we investigated the capacity of B. pseudomallei and B. thailandensis to enter and survive in host cells and promote host cell fusions in primary cells that the bacteria are likely to encounter during infection in vivo. To this end, we included human endothelial cells, epithelial cells, fibroblasts, keratinocytes, and mesenchymal stem cells in the study. In addition, we used various epithelial cell lines originating from different organs to test for tissue-specific differences and analyzed infected neutrophils and monocyte-derived macrophages (MDMs) for the formation of MNGCs.

RESULTS

B. pseudomallei is able to invade primary endothelial cells, epithelial cells, fibroblasts, keratinocytes, and mesenchymal stem cells.

We included a selection of normal primary human phagocytic and nonphagocytic cells as well as epithelial cell lines derived from diverse organs in this study (Table 1). The multiplicity of infection (MOI) and postinfection time were adjusted for each cell and cell line to optimize bacterial entry, MNGC formation, and host cell viability. Entry into host cells was determined after infection with B. pseudomallei and antibiotic treatment for 2 h to kill extracellular bacteria. The cells were infected with wild-type B. pseudomallei E8 or B. thailandensis E264 and mutants thereof that harbor a disrupted T6SS-5 (ΔBPSS1504 and ΔBTH_II0857 mutants, respectively). These unmarked mutants were generated and subjected to complementation analysis in previous studies (31, 41). For the sake of simplicity throughout the remainder of this report, we will refer to B. pseudomallei ΔBPSS1504 and B. thailandensis ΔBTH_II0857 as ΔT6SS-5 mutants. In addition, it should be noted that the purpose of this study was to examine the capability of B. pseudomallei to complete the intracellular life cycle in a particular cell type rather than comparing infection efficiencies between different cells. Using antibiotic protection assays, we found that the B. pseudomallei wild-type and ΔT6SS-5 mutant strains are capable of invading vein endothelial cells (HUVEC), bronchial epithelial cells (NHBE), gingival fibroblasts (HGFs), undifferentiated keratinocytes (KCs), and mesenchymal stem cells (hMSC) and were also detected inside MDMs and neutrophils (PMNs) (Fig. 1A). Likewise, B. thailandensis invaded all epithelial cell lines derived from diverse human tissues: adrenal glands (SW 13), bladder (5637), duodenum (HuTu-80), ovary (SK-OV-3), pancreas (PANC-1), and prostate (DU 145) (Fig. 1B).

Open in a separate window

FIG 1

(A) Entry and intracellular survival of B. pseudomallei wild type (wt) and T6SS-5 mutant (ΔT6SS-5) in human primary cells (HUVEC, human vein endothelial cells; NHBE, bronchial epithelial cells; HGF, gingival fibroblasts; KC, undifferentiated keratinocytes; hMSC, mesenchymal stem cells; MDM, monocyte-derived macrophages; PMN, neutrophils). Entry and intracellular survival were determined at 3 h and 17 to 24 h postinfection, respectively, using antibiotic protection assays. The MOI and time of infection were optimized for each cell and are listed in Table 1. (B) Entry and intracellular survival of B. thailandensis wild type (wt) and T6SS-5 mutant (ΔT6SS-5) in human epithelial cell lines derived from different tissues (SW 13, adrenal gland cell line; 5637, bladder cell line; HuTu-80, duodenal cell line; SK-OV-3, ovary cell line; PANC-1, pancreatic cell line; DU 146, prostate cell line). Entry and intracellular survival were measured at 3 h and 14 to 40 h postinfection, respectively, using antibiotic protection assays. The MOI and time of infection were optimized for each cell line and are listed in Table 1. Shown are mean values + standard deviation. a, not detected.

TABLE 1

Primary cells, cell lines, and culture conditions used in this study^a

Cell/cell line	Type/origin (source/reference)	Culture medium (source)	Bacterium/MOI/time p.i. (h)	No. of experiments/ no. of replicates	Range of MOIs tested for MNGC formation
HGF	Fibroblast/gingiva (CLS)	FGM-MG (CLS)	B. pseudomallei/200/24	2/2	50–500
hMSC	Stem cell/bone marrow (ATCC)	Mesenchymal stem cell medium (ATCC)	B. pseudomallei/100/23	1/2	10–400
HUVEC	Endothelial/vascular (Lonza)	EGM-2 (Lonza)	B. pseudomallei/1/21	2/1 or 2	NA
KC	Keratinocyte/epidermal foreskin (healthy volunteers)	KGM Gold, KGM-2 BulletKit (Lonza)	B. pseudomallei/50/21	2/3	50–200
MDM	Peripheral blood mononuclear cells/blood (healthy volunteers)	RPMI 1640 (Biochrom), 10% FBS (Sigma)	B. pseudomallei/10/20	2/2	1–50
NHBE	Epithelial/airways above bifurcation of the lungs (Lonza)	BEGM (Lonza)	B. pseudomallei/1/19	3/3	NA
PMN	Granulocyte/blood (healthy volunteers)	RPMI 1640 (Biochrom), 10% FBS (Sigma)	B. pseudomallei/10/17	2/3	1–50
5637	Carcinoma epithelial/bladder (CLS/51)	RPMI 1640 (Biochrom), 10% FBS (Sigma)	B. thailandensis/10/40	3/3	10–100b
DU-145	Epithelial/brain derived metastatic prostate cancer (CLS/52)	DMEM (Gibco), 10% FBS (Sigma)	B. thailandensis/100/26	2/3	10–100b
HuTu-80	Adenocarcinoma epithelial/duodenum (CLS/53)	DMEM (Gibco), 10% FBS (Sigma), 1% NEA (Biochrom)	B. thailandensis/50/18	2/3	10–100b
PANC-1	Carcinoma epithelial/pancreas (CLS/54)	DMEM (Gibco), 10% FBS (Sigma)	B. thailandensis/100/14	3/3	10–100b
SK-OV-3	Adenocarcinoma epithelial/ovary (CLS/51)	McCoy 5A (Biochrom), 2 mM l-glutamine (Gibco), 1× NEA (Gibco), 5% FBS (Sigma)	B. thailandensis/100/25	2/3	10–100b
SW13	Carcinoma epithelial/adrenal gland (CLS/55)	1:1 DMEM/Ham's F-12 (Gibco), 2 mM l-glutamine (Gibco), 5% FBS (Sigma)	B. thailandensis/50/20	3/3	10–100b

Open in a separate window

^aAbbreviations: ATCC, American Type Culture Collection (www.atcc.org); CLS, Cell lines services (http://www.clsgmbh.de); DMEM, Dulbecco's modified Eagle medium; DSMZ, German Collection of Microorganisms and Cell Cultures (https://www.dsmz.de/); MDM, monocyte-derived macrophages; NEA, nonessential amino acids; NA, not applicable; p.i., postinfection.

^bMNGCs were observed at all MOIs used.

B. pseudomallei stimulates MNGC formation in primary endothelial and epithelial cells.

Next, we analyzed the capacity of B. pseudomallei to induce MNGC formation in the different types of primary cells. For this, the same time point postinfection as for the quantification of intracellular survival was used. To allow for primary cells and cell lines displaying more than three nuclei per cell in the absence of bacteria (which could be the result of mitosis occurring without cytokinesis), MNGC formation was also quantified in the uninfected state for some cells. Using Giemsa stain, MNGC formation was detected in HUVEC and NHBE upon infection with B. pseudomallei (Fig. 2A; see also Fig. 4A). Likewise, B. thailandensis was able to invade, survive in, and induce MNGC formation in HUVEC and NHBE (Fig. 3). Despite the presence of intracellular B. pseudomallei, no host cell fusions were observed in fibroblasts and keratinocytes at MOIs of 200 and 50, respectively (Fig. 2A and ​and4A).4A). The infection of HGFs and KCs at increased MOIs of 500 and 200, respectively, also failed to result in MNGC formation (data not shown). Similarly, infected MDMs and PMNs were monitored for MNGC production until 20 and 17 h postinfection, respectively. However, no MNGCs were detected (Fig. 2A and ​and4A4A).

Open in a separate window

FIG 2

(A) Human primary cells infected with B. pseudomallei wild type (wt) and T6SS-5 mutant (ΔT6SS-5) for 17 to 24 h using antibiotic protection assays and stained with Giemsa. The same MOI and time of infection were used as for the determination of intracellular survival of B. pseudomallei in these cells and are listed in Table 1. MNGC formation was observed in NHBE and HUVEC infected with wild-type B. pseudomallei. (B) Human epithelial cell lines infected with B. thailandensis wild type (wt) and T6SS-5 mutant (ΔT6SS-5) for 14 to 40 h using antibiotic protection assays and stained with Giemsa (see legend to Fig. 1 for abbreviations of cell lines). The same MOI and time of infection were used as for the determination of intracellular survival of B. thailandensis in these cell lines and are listed in Table 1. Wild-type B. thailandensis stimulated MNGC formation in all cell lines tested.

Open in a separate window

FIG 3

MNGC formation in primary endothelial (HUVEC) and bronchial epithelial cells (NHBE) induced by B. thailandensis. (A) Giemsa stain of NHBE and HUVEC infected with B. thailandensis wild type and ΔT6SS-5 mutant at MOIs of 200 and 50, respectively, for approximately 20 h. (B) Quantification of MNGC formation efficiency under the same conditions. (C) Invasion and intracellular survival of B. thailandensis in HUVEC and NHBE at 3 h and 20 h postinfection, respectively. Shown are mean values + standard deviation. a, not detected; b, not determined.

Open in a separate window

FIG 4

(A) Quantification of MNGC formation in primary cells (see legend to Fig. 1 for abbreviations) infected with B. pseudomallei wild type (wt) and T6SS-5 mutant (ΔT6SS-5) or left uninfected. The same MOI and time of infection as for the determination of intracellular survival of B. pseudomallei in these cells were used and are listed in Table 1. (B) MNGC formation in epithelial cell lines infected with B. thailandensis wild type (wt) or ΔT6SS-5 or left uninfected (see legend to Fig. 1 for abbreviations of cell lines). The same MOI and time of infection were used as for the determination of intracellular survival of B. thailandensis in these cells and are listed in Table 1. Shown are mean values + standard deviation. a, not detected; b, not determined.

B. thailandensis stimulated the production of MNGCs in all six epithelial cell lines included in the study irrespective of the organ from which they were derived (Fig. 2B). The degree of MNGC formation differed considerably among the cell lines, ranging between 19.1% for duodenum cells (HuTu-80) and 2.8% for bladder cells (5637) (Fig. 4B). Importantly, disruption of the T6SS-5 in B. pseudomallei or in B. thailandensis abolished MNGC formation in all cell lines (infected with B. thailandensis) and in NHBE and HUVEC (infected with B. pseudomallei). Taken together, the results verify the essentiality of the secretion system for host cell fusions in a wide range of nonphagocytic cell types, including primary epithelial and endothelial cells.

Association of MNGC formation with the number of bacteria residing within host cells.

MNGC formation was observed in NHBE and HUVEC primary cells, which supported efficient intracellular growth of B. pseudomallei and B. thailandensis and was most pronounced in the HuTu-80 cell line, which had the highest bacterial load of all the cell lines studied. These findings led us to examine whether there is an association between the number of intracellular bacteria and cell-cell fusion efficiency. For this, we plotted the average CFU of wild-type B. thailandensis used to determine intracellular survival against the average MNGC formation among HUVEC and NHBE and each of six cell lines (Fig. 5). The correlation between intracellular CFU and MNGC formation was significant (Pearson's correlation coefficient [r] = 0.77; P = 0.024) when all data points were included. However, when the data point of the HuTu-80 cell line was excluded—as it might represent an outlier based on the CFU value, which exceeds that of all the other cell lines more than 1,000-fold—no correlation between the intracellular bacterial burden and MNGC formation was observed (r = 0.34; P = 0.45). In addition, the CFU of wild-type B. thailandensis isolated, for example, from HUVEC were 4.3-fold higher than those of NHBE, yet MNGC formation of HUVEC was 2.8-fold lower. These observations may indicate that host cell-specific factors contribute to cell-cell fusion or that the activity of the T6SS-5 differs between cell types.

Open in a separate window

FIG 5

Intracellular loads of wild-type B. thailandensis in relation to the extent of MNGC formation. The diagram shows average values of final intracellular CFU of wild-type B. thailandensis plotted against average values of MNGC formation (%) for each cell line (black symbols), primary endothelial cells (HUVEC), and epithelial cells (NHBE) (gray symbols).

PMNs engage in heterotypic MNGC formation.

Although viable bacteria were isolated from PMNs after 17 h postinfection, B. pseudomallei did not stimulate MNGC formation in these cells. To test if this was caused by PMN-driven impairment of the intracellular bacteria to induce MNGCs or if this was attributable to an inherent inability of the PMNs themselves to fuse (e.g., absence of a surface protein functioning as a receptor of the T6SS-5), we performed mixed MNGC formation assays. For this, adherent RAW264.7 cells, which facilitate efficient intracellular proliferation of B. thailandensis and form pronounced MNGCs, were mixed with PMNs at a ratio of 1:2 and infected with wild-type B. thailandensis and the ΔT6SS-5 mutant for 22 h. Plasma membrane and DNA of infected cells were stained with wheat germ agglutinin Alexa Fluor 594 and DAPI (4′,6-diamidino-2-phenylindole), respectively, to identify MNGCs. Nuclei from PMNs and macrophages can be morphologically distinguished by their distinct nuclear shapes. Indeed, after wild-type infection, several MNGCs containing multilobed and round nuclei were reproducibly observed (Fig. 6A). No MNGCs were detected in ΔT6SS-5-infected cells (Fig. 6B). These data suggest that PMNs were not inherently unable to fuse and that a mechanism for the absence of MNGC formation observed in PMNs alone may be due to impairment of bacterial functions.

Open in a separate window

FIG 6

Confocal fluorescence microscopy z-stack imaging of mixed PMN and RAW264.7 macrophages infected with wild-type B. thailandensis (A) and B. thailandensis ΔT6SS-5 (B) for 22 h at an MOI of 50. z-stack images were collected at 0.6-μm sections, and the red and green lines indicate the orthogonal planes of the y-z and x-z projections, respectively. Heterotypic MNGC containing macrophage (single-lobed) and neutrophil (multilobed) nuclei were detected upon infection with wild-type B. thailandensis. Shown are representative images of two experiments performed in triplicate. Plasma membrane was stained with wheat germ agglutinin Alexa Fluor 594 conjugate (red), and DNA was stained with DAPI (blue). Arrows mark multilobed nuclei. Bars, 10 μm.

DISCUSSION

Understanding the pathogenesis of melioidosis requires the identification of the cell types susceptible to infection with B. pseudomallei in vivo. In the present study, we performed in vitro infection assays utilizing primary phagocytes and nonphagocytic cells to gain insight into the cell type specificity of the intracellular life cycle of B. pseudomallei. We show that B. pseudomallei has the capacity to invade normal bronchial epithelial cells (NHBE), vein endothelial cells (HUVEC), gingival fibroblasts (HGFs), and keratinocytes (KCs) and to survive in these cells for at least 19 h. The discovery of invasion and intracellular survival of B. pseudomallei in endothelial cells and keratinocytes is a novel finding, which has also not been reported for endothelial or keratinocyte cell lines. Infection of bronchial epithelial NHBE and HGFs is in agreement with previous studies reporting the isolation of viable bacteria from primary alveolar epithelial cells and immortalized fibroblast cells, respectively (9, 34, 36, 42). HGFs and KCs appear to restrict efficient intracellular replication of B. pseudomallei as indicated by decreasing levels of intracellular CFU counts over the course of the infection. However, the cells may constitute a niche for intracellular survival of B. pseudomallei during in vivo infection. Unlike fibroblasts and keratinocytes, NHBE and HUVEC facilitated extensive multiplication of B. pseudomallei (177- and 19-fold increases in CFU, respectively) and B. thailandensis (78- and 14-fold increases in CFU, respectively), which was similar to that observed for some of the cell lines. These results suggest that bronchial epithelial and vein endothelial cells provide a major replication niche for B. pseudomallei in vivo and as such likely play an important role in the pathogenesis of airway and bloodstream infection. Previous work showed that infections of olfactory neurons or phagocytes able to cross the blood-brain barrier (BBB) are routes of access for B. pseudomallei to the brain (43, 44). The ability, therefore, of B. pseudomallei to invade HUVEC could suggest transcellular crossing of the endothelium potentially by cell lysis-mediated disruption of BBB integrity as another mechanism to access the brain. Moreover, invasion of vascular endothelial cells would provide a means of spread of B. pseudomallei from airway tissue into the bloodstream and from the bloodstream into organs as observed for other intracellular pathogens (45, 46).

MNGC formation was detected following B. thailandensis infection in all six epithelial cell lines derived from different organs. Though these are cancer cell lines, the finding might indicate that the T6SS-5 does not exhibit tissue-specific activity. These data are in line with previous studies reporting MNGC formation in, for example, cervical and lung cancer cell lines (47, 48). Using physiologically more-relevant primary NHBE cells, we confirmed the capability of B. pseudomallei to mediate cell-cell fusions in lung epithelial cells. We speculate that epithelial cells of other organs produce MNGCs upon infection with the bacteria as well. Furthermore, we present the novel finding that B. pseudomallei stimulates the fusion of endothelial cells. Importantly, MNGC formation by cells and cell lines of NHBE and HUVEC was strictly dependent on a functional T6SS-5. To date, cell lines but not primary cells have been used to investigate the role of the T6SS-5 in cell-cell fusion (17, 27,–29, 41). Thus, the lack of MNGCs of NHBE and HUVEC infected with B. pseudomallei ΔT6SS-5 provides further evidence for the fundamental importance of the secretion system for mediating the fusion of host cells. These data strongly suggest that MNGCs observed in the tissue of melioidosis patients are induced by the T6SS-5 (20). Previous studies detected MNGCs in B. pseudomallei- or B. thailandensis-infected monocyte/macrophage, epithelial, and fibroblast cell lines, pointing toward the lack of strict cell type specificity of T6SS-5 function (18, 37, 39). Our finding that B. pseudomallei stimulates cell-cell fusions of primary epithelial and endothelial cells verifies that T6SS-5 activity is not restricted to a single cell type. Whether the T6SS-5 employs a host structure(s) found in both cell types for inducing cell-cell fusions or has the ability to target multiple and cell-specific host factors remains to be determined. A secreted protein of the T6SS-5 with effector activity essential for MNGC formation and virulence has been identified; however, the effector has not been experimentally characterized and lacks significant sequence similarity to proteins of known function, precluding speculation as to the host cellular target of the T6SS-5 (27,–29).

Despite the presence of viable intracellular bacteria in PMNs, HGFs, KCs, and MDMs, no MNGCs were observed under the conditions studied. In contrast to HGF cells, fibroblast cell lines were previously shown to be permissive for MNGC formation mediated by B. pseudomallei (8, 9, 18). Thus, our data might point toward distinct differences between primary cells and cell lines or host/tissue-specific differences between the cells influencing the susceptibility for B. pseudomallei-induced host cell fusions. A potential explanation for the basis of this is the lack of expression of host factor(s) targeted by the T6SS-5 to induce fusions or functioning as the signal for activating the T6SS-5. Another possibility for the absence of MNGCs in infected MDMs (but not keratinocytes) is the lower cell density and smaller number of cell-cell contacts than in MNGC-positive cells and cell lines. In addition, this study involved a single B. thailandensis isolate and a single B. pseudomallei isolate, and interstrain variation in, for example, MNGC formation was not determined. However, given that the T6SS-5 is highly conserved in B. pseudomallei, it appears unlikely that the lack of MNGC formation in some of the primary cells is due to strain-specific differences.

Consistent with previous work, we recovered viable bacteria from PMNs after extended incubation times (8, 49, 50). PMNs filled with B. pseudomallei were found adjacent to other PMNs, and no cell-cell fusions were detected. To determine whether the properties of the bacteria inside PMNs or the properties of the PMNs themselves, e.g., the absence of surface structures, account for this, we performed mixed-infection assays with RAW264.7 macrophages, which facilitate efficient intracellular replication of B. thailandensis. Although further studies are required to clarify the mechanism of heterotypic MNGC formation, the detection of MNGCs consisting of both macrophages and PMNs indicates that B. thailandensis located inside macrophages is able to mediate the fusion with adjacent PMNs and suggests that this assay can be used to assess the susceptibility of other host cells to T6SS-5-induced cell-cell fusion.

MATERIALS AND METHODS

ACKNOWLEDGMENTS

REFERENCES