Introduction
Melioidosis is a serious disease with a wide range of manifestations, depending on the route of infection. The causative agent,
Burkholderia pseudomallei, is a Gram-negative bacillus that is endemic in the tropics [
1,
2]. The bacterium is normally resident in soil and acts as an opportunistic pathogen. There is little evidence of person:person transmission and infection occurs after exposure to contaminated soil/aerosols: increased risk factors for contracting the disease include diabetes and excessive alcohol consumption [
3]. Most notified cases of melioidosis are in South East Asia and northern Australia, but underreporting is likely exacerbated by its similarity to other infectious conditions (notably tuberculosis). Despite recent estimates that it causes more fatalities (89,000 per annum) than Dengue fever [
2], melioidosis has thus been relatively neglected and its prevalence underestimated. Infection can remain dormant for long periods [
4]; there is no effective vaccine and the bacterium is resistant to many antibiotics.
Burkholderia pseudomallei is a facultative intracellular pathogen and can invade a wide range of tissues, giving rise to diverse clinical manifestations including pneumonia, septicaemia, abscesses, acute pyelonephritis, osteomyelitis, and encephalitis [
2]. After cell invasion, the bacteria are able to escape from the endocytic compartment into the cytosol where they replicate and acquire mobility by inducing actin polymerisation [
5,
6]. Unusually amongst bacteria,
B. pseudomallei are able to induce fusion of the infected cells with non-infected cells to form multinucleated giant cells (MNGC) [
7]. Such MNGC or syncytia are observed in the tissue of patients with melioidosis [
8] and the bacterium is also able to induce MNGC formation in vitro in a variety of mammalian cell lines [
9,
10]. MNGC formation can similarly be induced by the closely related but relatively non-pathogenic species,
Burkholderia thailandensis [
10]. In melioidosis, the capacity to form MNGC is thought to be associated with pathogenicity and may facilitate cell:cell spread, evasion of the immune response and might also protect against antibiotics [
5‐
7]. The type VI secretion system 5 (T6SS-5), which is associated with virulence in animal models of infection, is required for cell:cell fusion in both
B. pseudomallei and
B. thailandensis [
11,
12]. It is very likely, however, that host cell factors are also involved in MNGC formation and it has been shown that antibodies to certain host cell surface proteins inhibit cell fusion induced by
B. pseudomallei in human U937 macrophages [
13].
The tetraspanins are a family of evolutionarily conserved membrane proteins with 33 members in humans and a similar number in mice [
14]. They are involved in many basic cell functions and act primarily by associating with and organising the other membrane proteins to form functional microdomains known as TEM (tetraspanin-enriched microdomains) [
15]. Tetraspanins have been implicated in the control of cell:cell fusion, perhaps most notably tetraspanin CD9 in sperm:egg fusion, with female mice showing greatly reduced fertility owing to the inability of the oocytes to fuse [
16]. Loss of the tetraspanin CD81 exacerbates this phenotype [
17]. Tetraspanins have also been shown to be involved in the control of muscle cell fusion [
18], osteoclast formation [
19], mononuclear phagocyte MNGC formation [
20‐
22], and virus-induced syncytium formation [
23‐
25].
Given their involvement in other types of infectious and non-infectious cell:cell fusion, it was, therefore, of interest to determine if tetraspanins might play a role in fusion induced by
Burkholderia species.
B. pseudomallei is classed as Tier 1 bioweapon [
2], but the components of the T6SS-5 machinery involved in promoting MNGC formation are very similar in the rarely pathogenic but closely related species
B. thailandensis [
26]. We, therefore, investigated the role of tetraspanins CD9, CD63, and CD81 in MNGC formation induced by this bacterium in mouse macrophage cell lines. Two
B. thailandensis isolates were used: E264, an environmental isolate [
27], and CDC272, a clinical isolate [
28]. The effects of specific anti-tetraspanin antibodies and recombinant proteins representing the large extracellular region (EC2) on
B. thailandensis-induced MNGC formation were examined. In addition, we investigated MNGC formation in infected mouse macrophages where the tetraspanin CD9 had been ablated or was overexpressed. We also conducted preliminary investigations on other cell surface molecules that have been implicated in mononuclear phagocyte MNGC formation, some of which are known to associate with tetraspanins. Overall, our findings demonstrate that tetraspanins act to regulate MNGC formation induced by
B. thailandensis in mouse macrophages, with CD9 in particular acting as a negative regulator of this process. Given the similarity between MNGC formation induced by
B. thailandensis and.
B. pseudomallei, our findings may have implications for the disease mechanisms involved in melioidosis.
Materials and methods
Mammalian cell lines
The J774.2 mouse macrophage cell line was originally obtained from Prof H. Harris and Dr R. Sutherland (Sir William Dunn School of Pathology Oxford) and cultured in DMEM + 2 mM glutamine + 4.5 g/l glucose (Gibco) with 10% FCS (Biowest). The RAW 264.7 mouse macrophage cell line was from the American Type Culture Collection (ATCC) and cultured in DMEM with GlutaMAX (Gibco) with 10% FCS. J774.2 stably overexpressing mouse CD9-GFP and GFP were generated by transfection with pCMV6-AV-mCD9-GFP or pCMV-AV-GFP (Origene) using electroporation followed by selection in G418-containing medium and further selection by fluorescence activated cell sorting (FACS Aria, Becton–Dickinson). CD9WT and CD9 null macrophage cell lines were derived from CD9 knock-out and the corresponding wild-type C57BL/6 mice [
29] by transformation of peritoneal mouse macrophages using J2-transforming retrovirus [
30]. These cell lines were kindly provided by Dr. Gabriela Dveksler, Dept. Pathology, Uniformed Services University of Health Sciences, Bethesda, MD, US and cultured in DMEM + 2 mM Glutamine + 4.5 g/l glucose + 10% FCS. CD82WT and CD82 null mouse macrophage cell lines were derived from CD82 knock-out and the corresponding wild-type C57BL/6 mice [
31]. These cells were kindly provided by Professor Jatin Vyas, Division of Infectious Disease, Massachusetts General Hospital, Boston, US. All cells were maintained at 37 °C under 5% CO
2.
Burkholderia thailandensis strains
E264, an environmental isolate, sequenced strain [
27,
32] and CDC2721121, a clinical isolate from Louisiana, abbreviated as CDC272 [
28] were kind gifts from the laboratory of Professor Richard Titball, Dept. Biosciences, University of Exeter, UK.
Antibodies
Monoclonal antibodies (mAb) against mouse tetraspanins were CD9 (Cat. No. MCA2749, clone MF1 Bio-Rad), CD63 (Cat. No. 143902, clone NVG-2, Biolegend), CD81 (Cat. No. MCA1846, clone Eat2 Bio-Rad), and matching isotype controls rat IgG2b (Bio-Rad), rat IgG2aκ (Biolegend) and hamster IgG1 (Bio-Rad), respectively. MAb to other cell surface molecules were CD36 (Cat. No. 102602, clone HM36), CD44 (Cat. No. 103002, clone IM7) CD47 (Cat. No. 127502, clone miap301) CD98 (Cat. No. 128202, clone RL388), CD172a (Cat. No. 144002, clone P84) (all from Biolegend), and DC-STAMP (Cat. No. MABF39, clone 1A2 Millipore). Isotype controls were Armenian hamster IgG, rat IgG2bκ, rat IgG2aκ, rat IgG2aκ, rat IgG1κ, (Biolegend), and IgG2aκ (Millipore), respectively. The secondary antibodies used for flow cytometry were anti-rat IgG-FITC (Cat. No. F-9387, Sigma) and anti-hamster IgG-FITC (Cat. No. MCA2357, Bio-Rad). All antibodies were used at saturating binding concentrations.
Recombinant GST-EC2 proteins
The glutathione S-transferase (GST) fusion system was used to produce GST-tagged tetraspanin EC2 proteins as described previously [
22,
33].
Invasion assay
The invasion and intracellular survival of
B. thailandensis was assessed using a modified kanamycin protection assay [
10]. Cells were seeded at 2 × 10
5 cells/ml in 24-well plates and cultured overnight. An overnight culture of bacteria was washed twice with PBS with centrifugation and the pellet suspended to OD ~ 0.4. Cells were infected at a multiplicity of infection (MOI) of 3:1 (determined after optimisation) and were incubated at 37 °C 5% CO
2 for 2 h. Cells were then washed with PBS and incubated with media containing 500 µg/ml kanamycin and 500 µg/ml amikacin for an additional 2 h to eliminate extracellular bacteria. As a negative control, cells were treated with cytochalasin D (Sigma) for 1 h prior to infection to prevent bacterial uptake; also the supernatant from the cells was examined for viable bacteria. Cells were then washed with PBS and lysed in 0.01% Triton X-100 in PBS for 5 min. The lysis mixture was diluted and a selection of dilutions were plated on LB agar plates and then incubated at 37 °C. The number of intracellular bacteria was quantified after 28 h of incubation.
Cells were infected and extracellular bacteria later eliminated as described above. After an appropriate time post-infection, cells were washed with PBS and fixed using acid/ethanol [5% acetic acid (v/v), 5% dH2O, and 90% ethanol (v/v)] for 30 min at RT. Cells were washed with PBS and stained with Giemsa solution (Sigma) (0.1% solution w/v) for 30 min at RT, then washed with dH2O, and allowed to dry. Images were captured with a Nikon light microscope using the 40X objective.
Evaluation of multinucleated giant cell formation
Images were analysed using Image J software. For each experimental condition, images of 10 random fields captured at 400× magnification were analysed and cells with > 3 nuclei were considered to be MNGCs. Data from all 10 fields were combined, and then, the percentage of MNGCs and the average MNGC size were calculated as described previously [
21].
Effect of mAb and GST-EC2 proteins on B. thailandensis-induced MNGC formation
Macrophages were seeded at 2 × 105 in 96-well plates (100 µl/well) and incubated overnight at 37 °C 5%CO2. The cells were treated with mAb at 10 µg/ml (saturating binding concentration as determined by titration) or EC2 proteins at 500 nM for 1 h before the infection. The MNGC assays were then carried out as described above.
Flow cytometric analysis of antigen expression
The expression of tetraspanins and other cell surface molecules was assessed by flow cytometry. Cells were harvested, pelleted, and resuspended with wash buffer (HBSS (Lonza) containing 0.2% BSA (Sigma) and 0.1% sodium azide), and then transferred to flow cytometry tubes (Elkay) at 106/tube. After centrifugation, the cell pellet was incubated with appropriate primary antibody or isotype control at 10 µg/ml for 1 h on ice. After washing twice, the cells were incubated with appropriate FITC-labelled secondary antibody (anti-rat IgG-FITC (AbCam) or anti-hamster IgG-FITC (Abcam)) for 1 h on ice. Cells were then analysed using an FACS LSR II (Becton–Dickinson). Data were analysed with FlowJo, LLC software. In the case of infected cells, cells were fixed with 1% paraformaldehyde for 30 min prior to analysis. Intracellular expression of CD63 was assessed following cell permeabilization using Fix and Perm (Caltag) according to the manufacturer’s instructions.
Statistical analysis
Unless otherwise stated, all data presented represent at least three independent experiments. Graphs were drawn and statistical analyses performed using Graphpad Prism version 8.3.1 (Graphpad Software, San Diego USA). Details of the statistical analyses used are given in the figure legends.
Discussion
The disease melioidosis is caused by the Gram-negative environmental bacterium
Burkholderia pseudomallei, with the highest incidences recorded in South East Asia and northern Australia. Underreporting of the condition is very likely, however, and recent estimates suggest that melioidosis may be globally responsible for 89,000 deaths per year [
2]. The closely related species
B. thailandensis has been reported to cause disease in man in only a handful of cases, but shares many of the features of
B. pseudomallei. B. thailandensis is commonly used as a model for infection and it causes disease in mice [
27]. Of relevance to the present study,
B. thailandensis shares with
B. pseudomallei the ability to induce MNGC formation and the bacterial factors involved in cell:cell fusion, notably the T6SS-5 secretion system, are very similar between the species [
12]. This feature, which is associated with virulence in
B. pseudomallei, is likely to be regulated by host cell factors including cell surface proteins.
Members of the tetraspanin superfamily of membrane proteins have been implicated in naturally occurring cell:cell fusion (e.g., sperm:egg fusion, myoblast, and osteoclast formation) [
41], as well as MNGC formed by mononuclear phagocytes in response to inflammation [
20‐
22] and virus-induced syncytial formation [
24,
25]. We, therefore, investigated the role of tetraspanins reported to regulate cell:cell fusion in these systems in
B. thailandensis-induced MNGC formation in mouse macrophages.
MNGC formation was successfully induced in J774.2 and RAW264.7 macrophages following infection with CDC272 (a clinical isolate) and E264 (an environmental isolate) strains of
B. thailandensis, in line with a previous study [
10]. These macrophage cell lines expressed relatively high levels of the tetraspanins CD9 and CD81, which are reported to regulate non-infectious mononuclear phagocyte fusion. A significant enhancement of MNGC formation following treatment of J774.2 and RAW264.7 cells with anti-CD9 and anti-CD81 mAb prior to infection with
B. thailandensis was observed (Fig.
3). This pattern, which did not relate to an effect on infection per se, was very similar to the effects of antibodies seen on the other forms of MNGC formation in mononuclear phagocytes [
20‐
22]. Conversely, recombinant proteins corresponding to the large extracellular domain (EC2) of CD9, CD63, and CD81 significantly inhibited
B. thailandensis-induced MNGC formation in J774.2 cells, suggesting overall that these tetraspanins act as negative regulators of fusion. To explore this further, we investigated MNGC formation following
B. thailandensis infection of macrophages derived from CD9 null and wild-type mice. A significant enhancement of MNGC formation was observed with CD9 null infected macrophages, although these cells did not appear to show any increase in overall infection. This is again strongly indicative of a negative regulatory role for tetraspanin CD9 in bacterial-induced fusion. Consistent with this, J774.2 macrophages overexpressing CD9 showed significantly reduced MNGC formation on infection. Interestingly, the initial investigations showed that macrophages from CD82 null mice were also more prone to forming MNGC on infection with
B.
thailandensis. To our knowledge, there are currently no mAb specific to mouse CD82, but it would be informative to investigate this further should such reagents become available.
CD9, in common with other tetraspanins, is known to interact with the other cell surface proteins in TEM and to regulate their localisation/activity. To try to explore this further, we examined the expression of other molecules implicated cell:cell fusion in our system. CD36, a member of the class B scavenger receptor family, has been implicated in cytokine-induced mononuclear phagocyte fusion [
40] and myoblast fusion [
42]. A slight reduction in the surface expression of this protein was evident on CD9 null macrophages, but antibodies to CD36 had no effect on
B.
thailandensis-induced MNGC formation and no change in expression was observed on macrophage infection with
B thailandensis. Interestingly, CD36 has been reported to associate closely with CD9 on the surface of mononuclear phagocytes [
43]. CD44 is a widely expressed cell surface protein with roles in cell adhesion and cell:cell interactions. CD44 has also been implicated in cell fusion processes, with ligands and mAb reported to inhibit osteoclast formation [
44,
45], whilst bone marrow cells from CD44 KO mice showed enhanced osteoclast formation in vitro [
46]. We observed that antibodies to CD44 gave some enhancement of
B.
thailandensis-induced MNGC formation; this mirrors the effects of mAb to CD9 and CD81, suggesting that CD44 may also act as a negative regulator of fusion here. Interestingly, CD44 is known to associate with CD9 in TEM [
47]. CD47 is an integrin-associated protein that is reported to be important in macrophage fusion [
48]. Our demonstration that mAb to this protein significantly inhibit
B. thailandensis-induced MNGC formation echoes the finding that mAb to CD47 suppress
B. pseudomallei-induced MNGC formation in human U937 macrophages [
13]. DC-STAMP (dendritic cell-specific transmembrane protein) is a seven transmembrane domain protein that is described as critical for osteoclast and foreign-body giant cell formation, with mAb and DC-STAMP knock-out suppressing these processes [
35,
49]. Our finding that mAb to DC-STAMP inhibits
B. thailandensis-induced MNGC formation in mouse macrophages is consistent with a general role for this protein in controlling mononuclear phagocyte fusion. We also observed some inhibitory effects with antibodies to CD98 (also known as fusion regulatory protein-1 or FRP-1), which is also in line with the anti-CD98 suppression of MNGC formation induced by
B. pseudomallei infection of human U937 macrophages observed by Suparak and co-workers [
13]. Interestingly, CD98 is reported to associate with tetraspanins CD9 and CD81 on the surface of oocytes, where mAb to both CD9 and CD98 inhibited sperm:egg fusion [
50].
Taken together, our results demonstrate that tetraspanins CD9 and CD81 are involved in MNGC formation induced by B. thailandensis in mouse macrophages, with clear evidence that CD9 acts as a negative regulator of this process. Given the similarities between their fusion apparatus, in our view, it is highly likely that CD9 also regulates MNGC formation induced by B. pseudomallei. Our findings may, therefore, have implications for the better understanding of this aspect of pathogenesis in melioidosis and could indicate future treatments aimed at controlling bacterial syncytium formation.
Acknowledgements
This work was supported by PhD scholarships from the Ministry for Higher Education (Libya) and the Ministry of Higher Education and Scientific Research (Iraq). The authors appreciate the skilled technical assistance provided by the Flow Cytometry Unit (Medical School, University of Sheffield) and expert advice provided by Dr Helen Marriott (infection work) in the Dept Infection, Immunity and Cardiovascular Disease. We also thank Drs Marzieh Fanaei, Ibrahim Yaseen, and John Palmer for providing the recombinant EC2 proteins and Jocelyn Pinto for her assistance with work on the CD82 null macrophages (Dept Molecular Biology and Biotechnology, University of Sheffield). We are grateful to Dr. Gabriela Dveksler, Dept. Pathology, Uniformed Services University of Health Sciences, Bethesda, MD, US for providing us with the CD9 null macrophage cell line and to Professor Jatin Vyas, Division of Infectious Disease, Massachusetts General Hospital, Boston, US, for the CD82 null macrophage cell line. We would also like to thank Professor Richard Titbull (Dept. Biosciences, University of Exeter, UK) for providing the B. thailandensis strains used.
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.