Previous Article | Next Article ![]()
Clinical Microbiology Reviews, April 2005, p. 383-416, Vol. 18, No. 2
0893-8512/05/$08.00+0 doi:10.1128/CMR.18.2.383-416.2005
Copyright © 2005, American Society for Microbiology. All Rights Reserved.
Menzies School of Health Research, Charles Darwin University,1 Northern Territory Clinical School, Flinders University,2 Department of Medicine, Royal Darwin Hospital, Darwin, Australia3
SUMMARY INTRODUCTION Background and History EPIDEMIOLOGY Melioidosis in the Australia-Pacific Region Distribution of Melioidosis in Asia Thailand. PDR Laos. Vietnam. Malaysia and Indonesia. Singapore. China, Hong Kong Special Administrative Region, and Taiwan. Other parts of Asia. Areas outside Asia Animals and Melioidosis BACTERIOLOGY AND PATHOGENESIS General Bacteriology Environmental Microbiology and Epidemiology Bacterial Virulence Factors Secretory antigens. Cell-associated antigens. Molecular Epidemiology Genome Sequence of B. pseudomallei and Comparison with That of B. mallei Role of Host Immune Responses Innate immunity. Role of macrophages in immunity. Role of T cells in immunity. Antibody responses. Other host factors. Prospects for Vaccine CLINICAL FEATURES Risk Factors Clinical Syndromes Modes of Acquisition and Incubation Period Melioidosis as a Potential Bioweapon DIAGNOSIS AND MANAGEMENT OF MELIOIDOSIS Diagnosis Culture-based methods. Antigen detection. Antibody detection. Molecular methods. Antibiotic Resistance Antibiotic Treatment Management of Sepsis Syndrome CONCLUSIONS ACKNOWLEDGMENTS REFERENCES
| SUMMARY |
|---|
|
|
|---|
| INTRODUCTION |
|---|
|
|
|---|
This disease, now termed melioidosis, was named from the Greek "melis" (distemper of asses) and "eidos" (resemblance) by Stanton and Fletcher in 1932 (405). During the last century this gram-negative environmental bacterium has been variously known as Bacillus pseudomallei, Bacillus whitmorii (or Bacille de Whitmore), Malleomyces pseudomallei, Pseudomonas pseudomallei, and, since 1992, Burkholderia pseudomallei (496).
In the latter half of the 20th century, melioidosis emerged as an infectious disease of major public health importance in southeast Asia and northern Australia. In Ubon Ratchathani, Thailand, B. pseudomallei accounts for up to approximately 20% of community-acquired bacteremias (74). At the Royal Darwin Hospital, Australia, it has been the most common cause of fatal community-acquired bacteremic pneumonia (112, 147).
Largely due to clinical trials in Thailand, significant improvements have been made in defining the optimal antibiotic therapy for melioidosis. However, the choice of antibiotic regimen has not been shown to have an impact on mortality within the first 48 h of admission (473), and severe melioidosis in Thailand is still associated with a case fatality rate of approximately 50% (472). In Australia, the mortality rate is still significant and approaches 20% among all patients with melioidosis (111).
| EPIDEMIOLOGY |
|---|
|
|
|---|
|
|
Other anomalies that may be related to incomplete ascertainment include a high serological prevalence (7%) of melioidosis in returning American troops stationed in Vietnam (97) but a low rate of disease in the indigenous population (324, 453). In addition, the high prevalence of melioidosis in the Issan region of northeastern Thailand contrasts with the low prevalence in the People's Democratic Republic of Laos (PDR Laos) to the east of the Mekong River and in Cambodia further south (9, 68, 329).
A caveat to this paradox is the uncertainty associated with the seropositivity rates in southeast Asia, which may represent exposure to the less pathogenic Burkholderia thailandensis (472). The worldwide distribution of B. thailandensis is yet to be clearly defined, but it is clear that it comprises the most common soil isolates in northeast Thailand (437); it has not been found in Australia (108).
Although the area where melioidosis is endemic has generally been regarded as restricted to the latitudes 20oS and 20oN in southeast Asia and northern Asia (266), large outbreaks have occurred outside this area in Australia, including the first case in Winton (22oS) and 159 cases of melioidosis in pigs over 3 years in the Burnett River region (25.5oS), which were attributed to a contaminated water supply (237). Autochthonous cases have also occurred outside this area in southwest Western Australia (107, 171), the Brisbane River Valley in Queensland (27oS) (236, 374), Alice Springs, and Mackay. The distribution of endemic cases in 2001 and 2002 and previously reported foci outside of the area of endemicity is shown in Fig. 2.
|
Environmental sampling has revealed widespread isolation of samples from soil, mud, and pooled surface water in northern Australia, including Queensland (427), around Darwin (58, 299), and remote communities in the Northern Territory and Western Australia (107, 215). Two outbreaks have been linked to contamination of the drinking water supply, where disease control measures, such as cleaning of the water supply and pipes, led to a cessation of the outbreaks (115, 212, 213).
Although serological tests have been demonstrated to have poor sensitivity and specificity in clinical situations, seroprevalance is likely to reflect background exposure to B. pseudomallei on a population basis. Serosurveys of populations in northern Australia have demonstrated relatively low rates of seropositivity compared to the rates seen in northeastern Thailand. This was reflected in a study in Queensland, where seropositivity in urban populations (up to 5%) was lower than that in patients residing in rural locations or in patients of Aboriginal or South Pacific origin (up to 10%) (24), which were similar to those found in the primarily indigenous population of Arnhem Land in the Northern Territory (12.8%) (112). These seroprevalences contrast with the much higher rates in immigrants from southeast Asia to Queensland (29%) (24).
At least six cases of melioidosis have been reported from Port Moresby in Papua New Guinea (104, 132, 260, 362), as has one additional case in an ex-serviceman living in Brisbane (240), for whom the place of exposure was not clear. Small serological surveys in the Port Moresby region have not demonstrated antibodies to B. pseudomallei (33, 362). However, a series of clinical cases in Balimo, Western Province, has led to the suggestion that melioidosis may occur elsewhere in the country (112, 469).
|
These findings and the possibility of the existence of other, less virulent strains of Burkholderia spp. may also account for the much higher rates of seropositivity seen in Thailand (230), compared to the areas of endemicity of northern Australia (24).
PDR Laos. Despite the fact that the highest rates in Thailand have been documented in the northeast Issan region, relatively few cases have been reported in adjacent Laos (329). Mahosot Hospital in Vientiane recognized a handful of cases each year, constituting 2% of blood culture isolates at this referral center between 2000 and 2002 (P. Newton, personal communication), despite recovery of B. pseudomallei from soil isolates in the Vientiane region (494a). Outside of Vientiane, microbiological facilities are limited and the epidemiology is undefined.
Vietnam. Melioidosis was first noted in what is now southern Vietnam by Pons and Advier (336), and the first descriptions of the saprophytic niche of B. pseudomallei were made by Vaucel in Hanoi and Chambon in Saigon (now Ho Chi Minh City) (67, 130). With the large numbers of French and later American troops based in Vietnam, with exposure to environmental B. pseudomallei and access to modern clinical and laboratory services, large numbers of cases were described from the 1940s to the 1970s (130). Cases have continued to be observed in returning servicemen for up to 29 years following exposure (93, 173, 285), as have sporadic cases in Vietnamese emigrants and returned travelers to other countries (192, 489).
However, recent attempts at systematic surveillance have not found significant proportions of B. pseudomallei in blood culture isolates or soil around Ho Chi Minh City (324), although it is likely to be found elsewhere in the country. This may be analogous to the situation in Papua New Guinea, with the patchy distribution of B. pseudomallei in the environment associated with clinical cases.
Malaysia and Indonesia. Stanton and Fletcher noted animal cases at the Institute of Medical Research of the Federated Malay States as far back as 1913, and they published reports of these and subsequent human and animal cases in 1932 (404).
Cases from both peninsular and east Malaysia have continued to be described, and most recently, Puthucheary et al. reviewed 50 septicemic cases of melioidosis in 1992 at a single referral center in Kuala Lumpur and noted a total of 85 cases from June 1976 to June 1991 (346). A serosurvey conducted in Malaysia in 1964 to 1966 revealed a 7.3% seropositivity (indirect hemagglutination assay [IHA] titers of >1:40), with the highest rates in recruits from Kedah (peninsular Malaysia) and Sabah (east Malaysia) (412).
Sporadic cases of melioidosis from Indonesia have been reported in the Dutch literature for many years (39, 395), in addition to more recent cases exported to Australia (177) the United Kingdom (120), and the United States (370).
Singapore. In Singapore, melioidosis has been a notifiable disease since 1989; an annual rate of 1.7 cases of melioidosis was documented between 1989 and 1996, with the majority (89%; 337 cases) being culture-confirmed cases (191). More recently, higher numbers have been noted since 1998, when 104 patients were reported (41). During early 2004, 57 cases have been reported, with an unusually higher case-fatality rate (40%), which were attributed to abnormally heavy rains and flooding (O. P. Lim, Abstr. 4th World Melioidosis Congr., abstr. 3, 2004). The case-fatality rate for patients with severe melioidosis otherwise appears in line with those in more developed countries (69).
Serosurveys have consistently demonstrated a low rate of seropositivity in Singapore (0.2% in the general population and 1.6% in construction workers), except in immigrants from Thailand or Malaysia (191, 272). Isolation of B. pseudomallei from surface water appears to be less common than it was in the 1960s (425) and less common than it is in other countries in the region. Unusually, a seasonal pattern or an association with rainfall has not been noted in Singapore, in contrast to the case for most other series (191).
China, Hong Kong Special Administrative Region, and Taiwan. Small numbers of cases of locally acquired melioidosis have been observed in Hong Kong (397, 398, 440, 483), and a seroprevalence of 14% was demonstrated by IHA in a tuberculosis sanatorium (396). An ongoing outbreak in marine mammals in an ocean park has been described (193). In Taiwan, a case series and several sporadic cases, mostly autochthonous, have been described (40, 205, 262-264, 283).
On mainland China, B. pseudomallei has been isolated from 4.2% of soil and water specimens in Hainan Island and adjoining coastal provinces as far north as 25°N (500), as confirmed by human cases and seroprevalences (IHA titers of >1:40) of up to 34% in farmers in the region (499, 501). In the seven isolates from culture-positive cases tested, a high rate of ceftazidime resistance (57%) was observed (501).
Other parts of Asia. Multiple cases have been reported from disparate regions of India but have been largely restricted to a few large medical centers, presumably where identification is possible (87, 223, 225, 231, 292, 348); some of these, including an apparent outbreak of a bubonic plague-like illness attributed to B. pseudomallei (44), have been disputed (43, 121). Cases in travelers returning to Europe from the Indian subcontinent have been reported, suggesting poor ascertainment of cases locally (122, 431). In addition, one serosurvey revealed a seroprevalence of 7% in a rural rice-growing area near Vellore (231).
Although few cases in Sri Lanka have been described (452), a report described a case that was believed to have been acquired in that country (326). Sporadic cases of melioidosis in travelers returning from Bangladesh have also been reported (122, 239, 300), including a series of three patients presenting with septic arthritis following travel to Syhet (201). One case from within the country has been described (414). Although the disease was reported in up to 10% of autopsy deaths in Rangoon, Burma, in the original series, since 1945 the only cases reported were one in a Dutch traveler (268) and a second possible exported case in a Taiwanese traveler (205).
Cases imported into the United Kingdom probably reflect the prevalence of melioidosis within the countries of origin, distorted by the magnitude of immigration from those countries. Human and animal cases have been imported from Bangladesh, Pakistan, India, Indonesia, and the Philippines (120, 122, 205, 370). Serological studies in East Timor suggest exposure to B. pseudomallei, but no culture-confirmed cases have been reported (P. Armstrong, personal communication).
Much debate has focused on the question of indigenous B. pseudomallei in the Americas. In North America, early reported cases were associated with travel (150), had poorly documented travel histories (38, 228), or are disputed (164).
Possibly the most intensely studied organism was the "Oklahoma isolate" from a soil-contaminated wound infection following a farming accident (295), which was identified by the authors as B. pseudomallei but with atypical characteristics that were felt to put this identification in doubt (130). Subsequent molecular studies have been conflicting (495), with phylogenetic analysis placing this isolate in a distinct group apart from both B. thailandensis and B. pseudomallei (170).
A variety of animals have been used in experimental models, including inbred mouse strains (456), chickens (457), and rats and guinea pigs (474). Most recently, the susceptible BALB/c and more resistant C57BL/6 inbred mouse strains have been used extensively in studies of host responses to B. pseudomallei (200).
Epizoontic outbreaks have also been reported among imported animals from areas of endemicity. A cluster of infections is sheep, goats, and pigs living on Aruba (Dutch Antilles) was attributed to B. pseudomallei in 1957 (421). An outbreak in a Paris zoo in the 1970s resulted in the spread to other zoos and equestrian clubs throughout France and to the deaths of at least two humans and a number of animals. This outbreak, referred to as "laffaire du jardin des plantes," was thought to be due to either the importation of horses from Iran or an infected panda donated by Mao Tse-Tung (130, 472).
| BACTERIOLOGY AND PATHOGENESIS |
|---|
|
|
|---|
On culture, the organism demonstrates differing colonial morphology, with mostly smooth colonies initially and dry or wrinkled colonies on further incubation. The clinical significance of the various colony types, including small-colony variants, is being prospectively investigated in Thailand (N. Chantratita, personal communication).
The saprophytic nature of B. pseudomallei was first recognized in 1955 in French Indochina (67). Some early studies implicated the aerosolization of dry dusts as a route of acquisition for American servicemen in Vietnam, based on the high incidence in helicopter crews (204). However, further studies have demonstrated highest yields from moist soils and pooled surface water (306, 413, 492).
The association between surface water and melioidosis is supported by the strong association with monsoonal rains (74, 112, 113, 266) and with occupational and recreational exposure to surface water and mud (74, 112, 266), particularly with flooding of rice paddies and planting at the commencement of the monsoonal season (417). The finding that higher rainfall is significantly associated with sepsis and pneumonia suggests that environmental conditions during the monsoonal season may be associated with inhalation rather than inoculation as the primary mode of acquisition (113).
In particular, moist clay soils seem to be favored by the organism (427), and populations residing on these soil types in Darwin, Australia, have a higher rate of disease (343). Sampling studies in Australia have suggested that bacterial counts increase to a depth of 60 to 90 cm (427, 492), but the finding that dry, shallower soils may be culture negative yet PCR positive has led to the suggestion that the organism may persist in a viable but nonculturable state (58). Although cleared, irrigated areas have been shown to be favored by the organism in Malaysia and Thailand (306, 413, 492), these are not found in the Top End region of Australia; a sampling study of the site of an aborted attempt to grow rice in the Northern Territory at Fogg Dam failed to recover B. pseudomallei (B. Currie, unpublished data).
In a Western Australia outbreak, both B. pseudomallei and the parasite Acanthamoeba sp. were isolated from a potable water source and the environment. Subsequent studies suggested that B. pseudomallei may infect Acanthamoeba trophozoites by coiling phagocytosis, a process described for other pathogens such as Legionella pneumophila (217). The significance of this interaction and those with plant-associated organisms as a reservoir for persistent environmental contamination is being assessed (A. Levy and T. Inglis, personal communication).
Contamination of drinking water supplies, rather than soil, has also been implicated in other outbreaks in Australia (115, 237). Chlorination of the water supply was associated with the termination of one of these outbreaks and appears to be effective against B. pseudomallei in vitro (A. D. Thomas, unpublished data). However, sensitive techniques using flow cytometry do demonstrate the possible presence of viable bacteria in small numbers despite free chlorine concentrations of up to 1,000 ppm (202). The presence of chlorinators does not affect community rates of melioidosis in areas of endemicity (82).
Factors that may influence the distribution of B. pseudomallei in the environment may include physical factors such as rainfall, humidity, UV radiation, and temperature; chemical factors such as soil composition, other vegetation, and the use of fertilizers; and recent soil disturbances such as excavation and plowing (215). The implications of global climate change for the epidemiology of melioidosis are as yet unknown (103).
There has been some interest in the interaction between species in the Burkholderia genus (99, 167). As Burkholderia cepacia can degrade toxic compounds in pesticides and is active against many soilborne pathogens, there has been interest in its use as a crop biological control agent (199). However, numerous insertion sequences within B. cepacia, including for some strains sequences identical to B. pseudomallei insertion sequences (284), and transposable genetic elements in B. pseudomallei (61) have been identified. This justifies concerns that widespread agricultural use of B. cepacia may be a hazard to human health, with the potential for more virulent B. cepacia bacteria to emerge following horizontal transmission of genetic elements (199).
B. pseudomallei produces a glycocalyx polysaccharide capsule that is probably an important virulence determinant (407). This capsule (biofilm, or "slime") allows for the formation of microcolonies in a protective environment in which the organism is phenotypically altered, resulting in significant antibiotic resistance (461). In other bacteria, such as B. cepacia, it is believed that biofilm formation is stimulated by bacterial quorum-sensing mediators such as N-acylhomoserine lactones (468); early work has defined putative signaling pathways in B. pseudomallei that may be virulence factors (E. Valede and Y. Song, Abstr. 4th World Melioidosis Congr., abstr 30 and 31, 2004).
Altered phenotypes such as slow-growing small-colony variants can be observed on primary plates from clinical specimens (V. Wuthiekanun, personal communication) or can be induced by passaging in vivo or in vitro and are also associated with significant antibiotic resistance. These variants may subsequently revert spontaneously to their normal morphology and antibiotic susceptibility (188). The significance of other mutant forms of the organism, such as cell wall-deficient L-forms that can be induced only in vitro by passage through rabbit alveolar cells (241), remains uncertain. This suggests that unusual mechanisms may mediate the survival of B. pseudomallei within the body, such as the "globi" observed within macrophages and giant cells in autopsy specimens (477).
Ultrastructural studies using electron microscopy have shown multiplication within the vacuoles of phagocytes following internalization and subsequent endosome lysis (185). When Acanthamoeba trophozoites are infected with B. pseudomallei, bacterial escape from vacuoles is mediated by coiling phagocytosis, a process described for other pathogens such as L. pneumophila, as discussed above (217). Movement of the bacterium towards one pole of the cell occurs by means of continuous actin polymerization into a "comet-tail" formation similar to that observed with other pathogens such as Listeria (48, 234, 410). Direct cell-to-cell spread is thought to occur by the induction of these cellular protrusions and fusion of cell membranes to form multinucleated giant cells (184, 234, 410), which have also been observed in human tissue (477).
Secretory antigens. The role of secreted antigens is unclear. Many, including proteases (261, 376), phospholipase C (246), and hemolysin, lecithinase, and lipase (26) are probably secreted via the general secretory pathway (type II secretion system) (137). Transposon mutations in the general secretory pathway, resulting in a failure to secrete protease, lipase, or lecithinase, do not appear to result in an attenuation of virulence in an animal model (53). However, the finding that the relationship between the density of bacteremia and mortality is similar in melioidosis and other gram-negative bacteremias suggests that exotoxins do not play a significant role in determining outcome (472).
However, a number of type III secretion systems (TTSS) have been described (349). TTSS in other organisms, such as Salmonella enterica, are activated under specific conditions to allow delivery of effector molecules to host cells in order to facilitate invasion and survival in phagosomes (162, 503). This presumed function in B. pseudomallei is supported by the finding of homology between the SPI-1 pathogenicity island of Salmonella enterica (Inv/Spa/Prg) and TTSS3 of B. pseudomallei (349, 408, 410). Bacterial products secreted by this TTSS (termed Bsa, or Burkholderia secretion apparatus), including BopE, are thought to result in cytoskeletal rearrangements facilitating host cell invasion (408). B. pseudomallei organisms with mutations in the Bsa and BopE system are also confined to the endosome and are unable to gain access to cell actin, suggesting that this system is also important in mediating endosomal membrane lysis (410).
In addition, less virulent B. thailandensis strains do not contain some TTSS (349), and mutations involving the TTSS3 system (as opposed to TTSS1 and TTSS2) attenuate virulence in a hamster model (467). However, B. pseudomallei strains with mutations involving known putative TTSS3 effectors are not significantly attenuated (409, 467). Microarray studies have determined that growth of B. thailandensis in the presence of arabinose results in downregulation of TTSS3 via the putative positive regulator bsaN, suggesting that the loss of the ability to assimilate arabinose is linked to the increased virulence of B. pseudomallei in humans and animals (304).
Cell-associated antigens. A number of cell-associated antigens have been demonstrated to be immunogenic in patients with melioidosis, including capsular polysaccharide (CPS), lipopolysaccharide (LPS) (formerly O-PS II) (195), and flagellin proteins (75, 138, 290, 328).
Antibodies to LPS have been shown to be protective against severe disease in humans (75) and in animals (63). The important role of LPS is supported by studies examining laboratory-inducted mutations in the gene coding for LPS, which demonstrate a susceptibility to alternative complement pathway (139, 486) and an attenuation in virulence in a mouse diabetes model (139, 487).
Capsular polysaccharide appears to have a role in environmental protection (229), immune system evasion (347), and attachment to epithelial cells (3). Capsular polysaccharide appears to provide protection within the phagosomal environment (342, 389), and strains with mutations in this antigen are less virulent than wild-type strains (28). In addition, passive immunization against an exopolysaccharide provided protection against high-dose challenge in a mouse model (227).
Antigenic differences in CPS or other surface proteins may account for the lack of epithelial attachment and pathogenicity of B. thailandensis (233, 356). The ability of B. pseudomallei to attach to and invade epithelial cell lines appears to be growth phase and temperature dependent; the mechanisms underlying this in vitro phenomenon and its clinical relevance are yet to be determined (60). A number of genes for different CPSs, termed CPS I to IV, have recently been described. CPS I (previously thought to represent a component of LPS and known as O-PS I) is found only in B. pseudomallei and is a virulence determinant; CPS II is downregulated in vivo and is thought to be involved in environmental survival (S. Reckseilder-Zenteno, Abstr. 4th World Melioidosis Congr., abstr. 37, 2004).
Outer membrane proteins, such as a protein tyrosine phosphatase, have been defined (245). Although bacteria do not contain the substrate tyrosine phosphate, analogous enzymes in other bacteria such as Yersinia have alternative substrates that are believed to be important in signal transduction (229). However, acid phosphatases do not appear to be a major virulence determinant, as strains with mutations of acpA, resulting in loss of phosphatase activity, retain their virulence (64). Other outer membrane proteins, with molecular masses of 70, 38, 31, 24, and 17 kDa, have also been identified and used in diagnostic tests; the 38-kDa peptidoglycan-associated protein appears to form aggregates and to function as a porin (174, 388).
Other cell-associated antigens include type I pili (encoded by fimA, fimC, and fimD) and a putative type IV pilus complex, and other antigens with strong homology to the pilB, pilC, and pilD products of Pseudomonas aeruginosa have been defined. Flagellin proteins may be important in pathogenesis; flagellin-specific antiserum passively protected diabetic infant rats against B. pseudomallei challenge (52). However, conflicting results have been reported with nonmotile fliC mutants which were less virulent in BALB/c mice (96) but not in Syrian hamsters or diabetic mice (138).
A largely unexplored area has been the role of iron metabolism in determining virulence. It is known that a siderophore, malleobactin, is elaborated by B. pseudomallei and is efficient at acquiring iron at acidic pH (498). This siderophore is regulated by the fur gene, which also regulates superoxide dismutase and peroxidase (275). In other organisms, such as Vibrio parahaemolyticus, iron-restricted conditions result in the induction of siderophore production, which parallels increased virulence (117). The virulence of B. pseudomallei in a mouse model appears to be attenuated by iron-enriched media (441). Although clinical conditions with iron overload, such as hemosiderosis and thalassemia, appear to be associated with increased rates of melioidosis (286, 363), like infections with Vibrio and Salmonella spp, this may suggest that iron plays a more important role in pathogenesis than has been recognized.
It is important to note that these putative virulence factors are mechanisms developed (in evolutionary terms) by the organism to survive in its as-yet-unidentified ecological niche(s). They also happen to allow the bacterium to avoid the host immune responses of humans and animals. However, infection of these hosts is accidental and is not likely to provide an evolutionary advantage for an otherwise environmental organism. This fact is reflected in the poor characterization of bacterial products as being truly virulent in animal studies and in the organism's primary affinity for hosts with impaired immunity. In addition B. pseudomallei generally has low disease-causing potential in healthy hosts despite its ubiquity in the environment. This stands in contrast to other organisms whose ecological niche is in humans and animals, such as Staphylococcus aureus, which can affect otherwise healthy individuals and where virulence factors such as Panton-Valentine leukocidin correlate with severe disseminated disease.
These and other studies have demonstrated considerable diversity in isolates (333). This suggests that introduction of B. pseudomallei into these regions is not a recent event, in contrast to clonal outbreaks in regions where the organism is not endemic (107). Comparisons of typing methods demonstrate that RAPD analysis and PFGE are more discriminatory than ribotyping (180, 216, 449). Multilocus sequence typing appears to have a discriminatory ability similar to that of PFGE (79, 170).
The finding that the antigenically similar B. thailandensis was relatively avirulent led to a suggestion that different strains of B. pseudomallei may have differing virulence. Molecular typing of Australian isolates by PFGE and ribotyping have shown three clusters, termed A, B, and C (107, 180, 441). There was stability in the levels of virulence within two of these clone types, but many isolates in this study did not fall into defined groups (441).
Two other studies have suggested that clinical presentation or outcome may depend on the strain type. Certain ribotypes appeared to be associated with a higher mortality or risk of relapse in one study (333). A second small study (n = 18) using multilocus enzyme electrophoresis and RAPD analysis suggested that soft tissue infections were restricted to one cluster and respiratory and neurological infections were seen in another cluster (317).
More recently, a sequence-based successor to multilocus enzyme electrophoresis, multilocus sequence typing, has been developed for B. pseudomallei. This technique, based on allelic differences present at seven housekeeping genes, is ideal for genetic analysis beyond the outbreak situation due to its reproducibility and the low rate of genetic change in the allelic sites and allows for comparisons between strains typed at different laboratories though an internet-based database (http://www.mlst.net/). This study confirmed the diversity seen in isolates worldwide and placed B. mallei within this wider B. pseudomallei group, both being distinct from B. thailandensis (170). Furthermore, Australian isolates appear to be clustered separately from those from the other countries of southeast Asia where the disease is endemic, suggesting ancient origins that predate the transient land bridges between these regions in the Miocene period (79).
|
) appears to be vital in resistance, as might be expected for an intracellular bacterium, the role of adaptive CD4-mediated immunity remains uncertain, particularly as human immunodeficiency virus (HIV) infection does not seem to be a risk factor for disease; (iv) the contributions of the various bacterial products to infection and, more specifically, the relative contributions of endotoxin and exotoxins to pathogenesis and evasion of the immune system (the finding that the relationship between bacterial counts in blood and mortality is similar to that with other gram-negative organisms suggests that exotoxins do not contribute directly to outcome); and (v) the site from which latent infection may reactivate, as relapse after apparently successful treatment and an extended incubation period after exposure suggest a dormant state similar to that seen in tuberculosis. Innate immunity. B. pseudomallei appears to be resistant to serum bactericidal components (220); although the alternative complement pathway is activated, resulting in phagocytosis, it is resistant to the effects of the terminal complement membrane attack complex (149). Phagocytosis was increased by the addition of specific antibodies and complement (148).
Studies of the role of phagocytes in melioidosis have demonstrated conflicting results. It is established that B. pseudomallei can survive and multiply within professional phagocytes, including macrophage/monocyte and neutrophil cell lines (226, 342). It appears to be able to evade phagosome-lysosome fusion and destroy the phagosome membrane as soon as 15 min after ingestion (185). This is consistent with functional studies demonstrating poor early bactericidal activity in most (149, 195, 289) but not all (354) studies.
The comorbidities recognized as risk factors for melioidosis also may be operating by impairing neutrophil function. Diabetes mellitus has been demonstrated to result in impaired chemotaxis, phagocytosis, oxidative burst, and killing activity (30-32, 166, 287, 288, 305, 353). Similar defects have been observed in association with high alcohol consumption (27, 46, 47, 168, 183, 309, 312, 325, 411, 505), chronic renal failure (190, 194, 207, 338, 366), and thalassemia (476). Reports of patients with chronic granulomatous disease (146, 422) support the role of neutrophils in resistance to melioidosis.
Of further interest is the possibility of reversing these functional defects with granulocyte colony-stimulating factor (G-CSF). Multiple studies of G-CSF in nonneutropenic animal models have demonstrated improvements in outcome (172, 276, 312, 322, 433). In human studies, improvements in laboratory markers are seen following G-CSF use (186, 242, 360), and patients with diabetic foot infection may benefit from the use of G-CSF (134, 175). However, large multicenter clinical trials have failed to demonstrate benefits in pneumonia and severe sepsis, possibly due to late administration of G-CSF in the course of the illness (85, 310, 311, 361). The use of G-CSF has been associated with a fall in mortality in patients with septic shock due to B. pseudomallei; a clinical trial has commenced recently (84).
Role of macrophages in immunity. As mentioned above, B. pseudomallei appears to be able to survive and multiply within professional phagocytes, including those of the macrophage/monocyte lineage (185, 226). Although this is assumed to play a role in the site of latent infection, little is known regarding the precise localization of latent intracellular B. pseudomallei after early host-bacterium interaction and the bactericidal ability of macrophages.
Macrophages exposed to B. pseudomallei do not appear to respond in the same way as they do to other pathogens; in one study, lower levels and slower production of inducible nitric oxide synthase and tumor necrosis factor alpha (TNF-
) were seen in a macrophage cell line compared to those after exposure to Escherichia coli and S. enterica serovar Typhi (446, 447). However, the responsiveness of these cells, as well as their bactericidal activity, was increased by priming with IFN-
, providing an explanation for the important role of this cytokine in mediating resistance (447). Similarly, the low levels of IFN-ß production by macrophages may also mediate poor intracellular control by inducible nitric oxide synthase-dependent mechanisms (444).
In addition, ultrastructural studies of macrophage-B. pseudomallei interactions have compared responses in patients with melioidosis and healthy controls. They suggest that there is less early phagolysosome fusion in macrophages from melioidosis patients, resulting in higher intracellular bacterial concentrations (S. Puthucheary, Abstr. 4th World Melioidosis Congr., abstr. 40, 2004).
TNF-
, produced primarily by macrophages but also by B cells, T cells, and fibroblasts, is an early and potent proinflammatory cytokine. High systemic levels are associated with septic shock, including in patients with melioidosis, where high TNF-
levels were associated with mortality. However, TNF-
is required for the containment of infection; neutralization of TNF-
in a mouse model increased susceptibility to melioidosis (369).
The TNF2 allele, representing a stable mutation in the TNF-
promoter (G
A substitution at base 308), is associated with increased production of TNF-
. It is also associated with increased susceptibility to mucocutaneous leishmaniasis, cerebral malaria, and meningococcal purpura fulminans (249). In 109 of 123 Thai patients with melioidosis, the presence of TNF2 (present in 9% of the study population) was associated with an increased risk of death, septicemia, and multiple and single organ involvement. When the patients were matched to 74 seronegative controls, the presence of TNF2 was associated with a statistically significant relative risk of melioidosis of 2.3 (318).
We have previously speculated that the intracellular niche of this organism and its interaction with hosts with specific comorbidities, such as diabetes, suggest that the innate immune response plays a prime role in the control of this organism (111). Severe impairment of specific cellular immunity, such as with advanced HIV infection, does not appear to be a risk factor for infection. Similarly, the disease may be present even with high antibody titers, indicating that natural humoral immunity does not appear to be protective. The relatively minor role of specific immunity relative to innate immunity in this infection may explain the capacity of this organism for latency as well as the lack of protection against disease despite repeated exposure to the organism.
Role of T cells in immunity. A classic framework to view specific host responses is the type 1/type 2 dichotomy, in which cell-mediated responses are usually inversely proportional to the levels of antibody response. Type 1 responses are generally adaptive against intracellular pathogens; C57BL mice are more resistant to Leishmania major and IL-10 gene knockout (IL-10/) mice are more resistant to Listeria monocytogenes, and their resistance is attributed to their tendency to type 1 responses (403).
In melioidosis, this appears to be supported by the demonstration that type 1 response-prone C57BL6 mice were relatively resistant to B. pseudomallei compared to BALB/c mice (200, 259) and by the demonstration of the key role of IFN-
, a key Th1 cytokine that stimulates phagocytosis, the oxidative burst, and intracellular bacterial killing (368, 369), and its correlation with the severity of illness in humans (59, 256).
Similarly, the sine qua non of Th1 polarization, IL-12, can be demonstrated in patients with melioidosis and in ex vivo blood simulated with heat-killed B. pseudomallei, and neutralization of IL-12 can be demonstrated to result in impaired IFN-
production (256) and increased susceptibility in mice (369). In addition, the ratio of immunoglobulin G1 (IgG1) to IgG2 in humans (454) and mice (200) also suggests that a type 1 response may be adaptive.
However, there is evidence that anti-inflammatory Th2 responses are necessary to balance proinflammatory Th1 responses; severe sepsis may represent an unregulated Th1 response. In septic patients, as in melioidosis, poor outcomes are associated with high levels of both IL-6 (a Th1 cytokine) and IL-10 (a Th2 cytokine) (382). This may suggest that the intensities of both pro- and anti-inflammatory responses merely reflect the magnitude of the inflammatory insult. More extensive cytokine characterization of Th1 and Th2 responses in mouse strains with differential susceptibility has now demonstrated a mixed pattern; levels of mRNA for Th1 cytokines (IFN-
, IL-6, and IL-12)and Th2 cytokines (IL-10) increased in both susceptible BALB/c and resistant C57BL/6 mice (443).
Other studies have implicated effector cells as cytotoxic T lymphocytes and NK cells responding to IFN-
via the CXC chemokines IP-10 and Mig (257). These cells may act to induce apoptosis in infected cells via granzymes (GrA and GrB) (255). NK cells and, unexpectedly, CD8+ T cells have been demonstrated to be a source of IFN-
early in the response to infection with B. pseudomallei (270). This process appears to be a "bystander" phenomenon mediated by IL-12 and IL-18 rather than engagement of the T-cell receptor.
The demonstration of a specific cell-mediated immune response in survivors of acute melioidosis suggests that this may be important in the long-term control and prevention of relapse (34, 235). However, most puzzling is the absence of evidence to suggest that HIV infection is a risk factor for melioidosis, despite its prevalence in Thailand (92).
Antibody responses. As discussed above, many B. pseudomallei components have been demonstrated to be immunogenic in patients with melioidosis, including capsular polysaccharide and lipopolysaccharide (O-PS I and O-PS II) (195), flagellin proteins (75, 138, 290, 328), and other cell wall proteins (386). Of these, antibodies against the LPS component O-PS II and possible an exopolysaccharide and flagellin components have been demonstrated to be protective (52, 75, 227).
Antibodies of all classes against a culture filtrate antigen were demonstrated in patients with prior melioidosis, and the level was highest for IgG, particularly the IgG1 and IgG2 subtypes (86). Antibodies could persist variably for over 3 years (454). The description of a patient with a persistently high IHA titer (>1:5,120) for many years who presented with reactivation in association with staphylococcal endocarditis suggests that persistently high titers may define a group of patients that require close follow-up (355). Evidently, the antibody response resulting from repeated natural exposure to B. pseudomallei and B. thailandensis is insufficient to provoke a protective response for primary infection or relapse.
Other host factors. A possible association between HLA-DRB1*1602 and severe melioidosis, independent of possible confounders such as diabetes mellitus, was described (145), but this has not been borne out by other studies, placing susceptibility genes between the HLA-Cw and HLA-DQ loci of the HLA-B58 haplotype (141).
Studies examining the role of polymorphisms in Toll-like receptor alleles and mannose-binding lectin genes and their correlation with the severity of disease are under way (W. J. Wiersinga, personal communication). It is likely that recognition of bacterial endotoxin or other bacterial components plays a significant role in adaptive and maladaptive responses in melioidosis, as with other septic states.
A promising therapy is the use of the use of the bacterial genomic sequences in the form of unmethylated CpG oligodeoxynucleotide, either as an immunoprotective agent or as a DNA vaccine adjuvant. In a mouse model, CpG was shown to be strongly protective when administered prior to exposure to B. pseudomallei, possibly by increasing bacterial phagocytosis by macrophages (445, 480). Although no significant adverse reactions have been noted in human studies to date, theoretical concerns focus on the possibility of DNA integration into host cells (although CpG itself does not contain a promoter), the induction of adverse cytokine profiles, and the possibility of triggering autoimmune phenomena in susceptible patients (243).
A B. mallei candidate conjugate vaccine, linking the capsular and O-PS lipopolysaccharide antigens to exotoxin A of Pseudomonas aeruginosa, is being investigated in a horse model of glanders (53, 54). Given the cost of such vaccines, the relatively low incidence of the disease, the uncertain duration of protection, and the resource-constrained regions in which such a vaccine might be used, it is unlikely that a vaccine for B. mallei or B. pseudomallei would find use outside military settings.
Although protective footwear is often recommended for patients at high risk of melioidosis, the usefulness of this advice has not been evaluated. The observation that case clusters are associated with sudden, heavy, postcyclonic rainfall in Australia has suggested that public health interventions may best be targeted after these events (80).
| CLINICAL FEATURES |
|---|
|
|
|---|
|
45 years, 2.4 (1.9 to 3.0) for males, 3.0 (2.3 to 4.0) for Aboriginal Australians, 13.1 (9.4 to 18.1) for diabetics, 2.1 (1.6 to 2.6) for those with excess alcohol consumption, 4.3 (3.4 to 5.5) for those with chronic lung disease, and 3.2 (2.2 to 4.8) for those with chronic renal disease. The reason for these specific risk factors is not clear, but many have implicated the effect of these comorbidities on neutrophil function (111, 417), which is known to be important in the pathogenesis of melioidosis (226). The use of steroids is associated with an increased risk of melioidosis; this includes steroid-containing herbal remedies ("yaa chud") in Thailand, the use of which was documented in up to 10% of Thai patients (415). In the Australian series, chronic obstructive pulmonary disease and the consumption of kava and alcohol have also been implicated (111).
Despite cell-mediated immunity being implicated as a mechanism of resistance to melioidosis (34, 235), infection with HIV does not appear to be a major risk factor (92).
A number of case studies have noted an intriguing association with chronic granulomatous disease that may highlight the underrecognized role of neutrophil defects in pathogenesis (146, 422). Similarly, reports of patients with hemosiderosis suggest that impairment of phagocytic cells may be important (286, 363). Case reports of melioidosis and previous or subsequent mycobacterial infection (Mycobacterium tuberculosis, Mycobacterium terrae, or Mycobacterium leprae) may reflect a common host susceptibility to these intracellular pathogens (56, 89, 362, 417).
Important clinical differences have been seen between patients in Australia and Thailand (Table 4): the high incidence of genitourinary infection in Australia, with prostatic abscesses occurring in 18% of males; the absence of suppurative parotitis in Australia, in contrast to a rate of 30 to 40% in Thai children (119); and the distinct but uncommon encephalomyelitis syndrome seen in tropical Australia.
|
Acute suppurative parotiditis accounts for up to 40% of pediatric cases but only small numbers of adult cases in Thailand (119, 282). It seems to arise in patients with no defined risk factors and is generally associated with a good prognosis. It may be bilateral in 10% of patients and may be complicated by rupture or permanent facial nerve palsy. It has been reported only once in Australia (151).
The high proportion of patients with prostatic infection in Australia contrasts with the higher proportions of patients with liver and spleen abscesses seen in Thailand (455). In Australia, the prevalence of prostatic infection (18% of male patients) mandates routine imaging, with drainage commonly required (106, 111). This contrasts with other internal abscesses, which may respond to medical therapy alone (111).
Bone and joint infections are uncommon and may be difficult to differentiate from other causes of infection, except that the systemic features of the illness may be more prominent. Surgical drainage is often required, together with long courses of intravenous antibiotics (337).
Skin and soft tissue infections are a common manifestation of melioidosis and may be the source of systemic infection or result from hematogenous spread. Presentations may be rapidly progressive, similar to necrotizing fasciitis from other organisms (465). Infections involving many other sites have been described, including mycotic aneurysms, mediastinal infection, and thyroid and scrotal abscesses (111). Corneal ulcers were described for a series of three Thai patients following corneal trauma. Extensive ulcers, subconjunctival abscesses, and hypopyon were managed with topical and intravenous ceftazidime with good outcomes (385). Other ocular manifestations include orbital cellulitis with contiguous spread to the sinuses (478). Cardiac involvement is rare; pyopericardium is probably the most common manifestation, but myocardial abscesses (37) and endocarditis (344) have been reported.
Markers of organ dysfunction, including leukopenia (particularly lymphopenia), hepatic dysfunction (raised aspartate aminotransferase, alanine aminotransferase, and bilirubin levels), renal dysfunction (raised urea and creatinine levels), and metabolic derangements (hypoglycemia and acidosis), on admission appear to predict mortality (74, 81, 106).
The response to therapy is often poor, with a mean duration of fever of 9 days documented. Treatment failure, with antibiotic therapy alone, has been defined in studies as fever for longer than 14 days or bacteremia for longer than 7 days (383). Persistently positive cultures from other sites and radiological abnormalities are not uncommon and do not necessarily portend a poorer prognosis (111, 419).
Markers of inflammation such as C-reactive protein (CRP) are often used; a small study demonstrated that CRP levels were elevated in 46 patients with melioidosis, that the CRP level responded within 2 days in patients with uncomplicated treatment courses, and that persistent elevation in four patients was attributed to undiagnosed sites of infection or inadequate treatment. In addition, elevations in CRP levels predicted relapse, even in the absence of fever or leukocytosis (22). However, a review of the use of CRP levels in the larger Darwin prospective series has revealed cases with normal levels or mild elevations on admission with severe or fatal disease, including in relapse (83).
The clinical use of procalcitonin has also been assessed in Thailand; although the level of procalcitonin reflected the clinical severity of illness and the response to therapy, this was not specific to melioidosis. In addition, there was a wide variation in response, with a significant mortality even in patients with low procalcitonin levels (391).
The state of asymptomatic carriage has been suggested (109), but in 1,000 hospitalized children and 4,545 adults (1,011 patients with melioidosis and 3,524 healthy subjects) in Thailand, the positive predictive value of a throat swab for clinical disease was 100% (230, 494). In goats, aymptomatic carriage has been reported (426).
Occasional patients, particularly those with cystic fibrosis, with long-term carriage without signs of overt disease have been described (