Individuals with acute melioidosis present symptoms rapidly and succumb to disease (<24 hpi) before antibiotic treatment can be administrated . Previous studies on elucidating the pathogen-host response of melioidosis had focused primarily on a subset of immune response genes [8, 9, 28–30], however, analyses of single gene or limited gene expression patterns is insufficient to dissect the host response to infection globally. We developed an acute melioidosis model in BALB/c mice to get a comprehensive genome wide view of the host transcriptional response during the acute stage of melioidosis. Our analyses clearly demonstrated that the pathogen had intimately engaged the innate immune system at the early onset of infection by rapid induction of numerous inflammatory responses.
The primary response observed was the overwhelming induction of TLR2 to counteract B. pseudomallei, which we propose, subsequently triggered the activation of many inflammation-biased genes important in attracting neutrophils and monocytes to the site of acute inflammation. These cytokines and chemokines also function as central mediators in activating various host defence systems such as apoptosis, JAK/STAT signalling pathway, mitogen activated protein kinase (MAPK) signalling pathway and ultimately trigger the appropriate adaptive immune system. Induction of these genes was previously reported in numerous in-vivo, in-vitro or melioidosis patient studies [2, 13, 28, 30–32]. Hence our study reinforces the consistency of the inflammatory genes expression in response to acute melioidosis. Concomitantly, the host frontline defence system is boosted by increasing the production of granulocytes (Figure 2). Nevertheless, the bacteria are capable of propagating in a tissue environment that is evidently overloaded with high levels of inflammatory-associated proteins (Figure 1). This genome-wide expression study confirms that the production of signals responsible for the activation of pro-inflammatory genes in response to B. pseudomallei infection, are mainly TLR2 dependent. This observation supports a previous finding of improved survival in respiratory infection in TLR2 KO mice with reduced bacterial burden and lung inflammation, as well as less distant organ injury .
The cluster of inflammatory-associated genes consistently highly induced in response to B. pseudomallei acute infection is part of the group designated as "common host immune response". Most of these genes are induced in many different cell types in response to exposure to several different pathogen species such as Escherichia coli, Salmonella typhi, Staphylococcus aureus, Listeria monocytogenes, Mycobacterium tuberculosis, Candida albicans, Bordetella pertussis, Mycobacterium bovis, P. aeruginosa and S. typhimurium [33–39]. Up-regulation of this core set of genes by pathogens might represent a general "alarm signal" for inflammatory infections . Common host genes (such as CCL2, CCL7, IP30 and genes encoding MHC class II related molecules) known to be repressed by pathogens have been identified in PBMCs infected with B. pertussis, E. coli and S. aureus . Surprisingly in our study, these genes were highly induced in response to B. pseudomallei infection and could be a Burkholderia specific response.
The reaction to a given pathogen must be sufficient for bacterial elimination but not so strong as to be harmful to the host . This is particularly true for innate immunity in cases including acute melioidosis where excessive activation of inflammatory genes is commonly associated with septic shock. We did not see up-regulation in the levels of anti-inflammatory signals and TLR negative regulators at 24 hpi, suggesting that the failure to suppress inflammation at this early time point contributes to the excessive inflammation and acute nature of this infection. Nevertheless, at 42 hpi, a significant decrease in expression of these potent inflammatory genes (Figure 5a) was observed and may actually benefit the intracellular pathogen. However, the underlying factors that contribute to the decrease in expression of these inflammatory genes remain unclear as the production of anti-inflammatory cytokines (IL4, IL6 and IL10) was relatively insufficient to counter the high pro-inflammatory responses at 24 hpi.
Acute forms of melioidosis that lead to sepsis, multiple organ failure and death are thought to result from an uncontrolled inflammatory reaction that ultimately leads to excessive inflammation  and eventually tissue injury in the B. pseudomallei -infected host. Activation of proteasomal degradation following tissue injury suggests the production of immunological waste products such as apoptotic cells and immune complexes in the B. pseudomallei -infected host. This could be attributed to a failure in activating the complement system in time, leading to the accumulation of waste and uncontrolled spread of the pathogen (Figure 1). The low levels of the potent anaphyatoxin C5a observed in our study most likely inhibit the downstream terminal complement pathway. As a result, deficient rapid clearance of apoptotic cells resulting in extracellular disintegration of the cell and release of intracellular components triggers inflammatory cytokine production and contributes to "breaking tolerance" by facilitating an immune response to intracellular constituents . This is the first evidence of failure of the downstream complement pathway in acute melioidosis.
The B. pseudomallei -infected host also over express many cell death related genes which suggests that the host initiates various cell death defence responses and disrupts cell regulation to limit a favourable intracellular niche for the pathogens. Elevation of caspase 2, 3, 7 and 8, as well as the BCL-2 family protein BID and TNF-receptor superfamily suggests that the host triggers apoptosis signalling via the death receptor mediated (extrinsic) pathway. In addition, we saw an up regulation of inflammasome related genes (NAIP2, NLRP3, CIITA, NLRP6, interferon activated gene 205 (IFI205)) not previously reported in the B. pseudomallei -infected host. B. pseudomallei virulence factors such as type-three secretion factors (TTSS), flagellin and channel forming toxins like hemolysin could trigger inflammasome-dependent caspase 1 activation [6, 42].
B. pseudomallei is known to interfere with iNOS expression in RAW264.7 macrophages and abrogate nitric oxide (NO) production during the early stages of infection [12, 43]. Arginase 1 and arginase 2 have been reported to compete with NO syntheses for their common substrate, arginine, and prevent NO production in the M. tuberculosis infected bone-marrow derived macrophages as well as Salmonella infected RAW264.7 macrophages [44–46]. Here we report for the first time that B. pseudomallei up-regulates both arginase 1 and arginase 2 isoforms in the host with arginase 2 being more dominant. The expression profiles demonstrate both host nitric oxide synthase 2 (NOS2) and arginase 2 were elevated at a similar magnitude at 24 hpi. This suggests that arginase competes with NOS2 to produce NO from arginine during the infection, leading to the suboptimal antibacterial effect of NOS2 in the B. pseudomallei -infected host.
Certain pathogens evade the host defence by triggering the TLR2-mediated biased anti-inflammatory effects or prevent recognition by TLRs . For example, Yersinia- and Candida-induced TLR2 signalling leads to the release of IL-10, which can lead to immunosuppression. However, the response following recognition of B. pseudomallei via the TLR2 signalling pathway is contrary to the evasion mechanism exploited by Yersina spp. and Candida spp. In addition, some pathogens have developed strategies to either block or avoid their recognition by TLRs and subsequent activation of the innate defence. This study suggests that B. pseudomallei may use specific TLR-mediated signals to escape from the host defence. Future studies will be aimed at determining whether B. pseudomallei utilizes these signals to evade TLR clearance mechanisms.
Tissue injury leads to extracellular matrix breakdown, including the degradation of hyaluronic acid (HA) and resulting oligosaccharides. In this study, the gene encoding hyaluronan synthase 2 (HAS2), the enzyme that produces HA, was induced. In contrast, the genes encoding hyaluronoglucosaminidases (HYAL1, HYAL2), the enzymes that degrade HA, were repressed, indicating that perhaps HA is not degraded during a B. pseudomallei infection. These endogenous signals can also trigger TLR2 and/or TLR4 activation and signals distinct from microbial stimulators, for instance HA but not LPS, signal through TLR4, MD2, and CD44 . Up regulation of TLR2, TLR4 and TLR7 as well as MD2 could indicate B. pseudomallei -infected host responses to endogenous signals released during tissue damage . However, the ability of the engaged TLRs to distinguish between microbial and endogenous signals and subsequently trigger appropriate responses, remains unclear [40, 48, 49]. These observations reflect that the inflammatory response may cause more damage to the host than the microbe. In summary, our work has provided an extensive description of host defence responses to B. pseudomallei during an acute infection.
Changes in host cell metabolism as a consequence of nutrient scavenging by intracellular B. pseudomallei have never been studied. The microarray data presented here provides the first description of changes in the B. pseudomallei -infected host cell metabolism particularly the glycolytic and TCA pathways. The glycolysis pathway and the TCA cycle were both transcriptionally repressed. It remains to be determined if shutting down both these pathways is part of the host response to control the replication of intracellular bacteria or a strategy adopted by the pathogen to survive intracellularly. In addition, we found that expression of 37 cytochrome P450-related genes was suppressed in the liver over the course of infection, most notably at 24 hpi. The expression of the detoxification enzymes amine UDP-glucuronosyltransferases (UGT2B1, UGT2B34) and N-sulfotransferase (NDST1) was also down-regulated. Our data suggests that B. pseudomallei- induced impaired liver detoxifying activity might be a causative factor in liver sepsis. Collectively, the data presented here suggests that hepatocytes, via receptors for many pro-inflammatory cytokines, modify their metabolic pathways (glycolysis, TCA, fatty acid metabolism and various amino acid biosynthesis) in response to B. pseudomallei acute infection.