Vibrio vulnificus: new insights into a deadly opportunistic pathogen
Summary
Vibrio vulnificus is a Gram-negative aquatic bacterium first isolated by the United States (US) Centers for Disease Control and Prevention (CDC) in 1964. This bacterium is part of the normal microbiota of estuarine waters and occurs in high numbers in molluscan shellfish around the world, particularly in warmer months. Infections in humans are derived from consumption of seafood produce and from water exposure. Vibrio vulnificus is a striking and enigmatic human pathogen, yet many aspects related to its biology, genomics, virulence capabilities and epidemiology remain elusive and poorly understood. This pathogen is responsible for over 95% of seafood-related deaths in the United States, and carries the highest fatality rate of any food-borne pathogen. Indeed, infections associated with this pathogen that progress to primary septicaemia have a similar case fatality rate to category BSL 3 and 4 pathogens, such as anthrax, bubonic plague, Ebola and Marburg fever. Interestingly, V. vulnificus infections disproportionately affect males (∼85% of cases) and older patients (> 40 years), especially those with underlying conditions such as liver diseases, diabetes and immune disorders. New insights from molecular studies and comparative genomic approaches have offered tantalising insights into this pathogen. A recent increase and geographical spread in reported infections, in particular wound cases, underlines the growing international importance of V. vulnificus, particularly in the context of coastal warming. We outline and explore here a range of current data gaps regarding this important pathogen, and provide some current thoughts on approaches to elucidate key aspects associated with this bacterium.
Introduction
Vibrio vulnificus is part of the normal microbiota of estuarine waters and occurs in high numbers in molluscan shellfish around the world, particularly in warmer months. This opportunistic pathogen is common in estuarine waters and has been isolated from a range of different environmental sources, including water, sediment and seafood produce (O'Neill et al., 1992; DePaola et al., 1994; Høi et al., 1998; Bisharat et al., 1999; do Nascimento et al., 2001; Baffone et al., 2006; Baker-Austin et al., 2010). Human infections associated with this bacterium originate from two distinct sources: consumption of seafood (primary septicemias) or exposure to seawater or seafood products (wound infections). Septicemia infections typically follow ingestion of raw/undercooked molluscan shellfish, primarily oysters, where it occurs in large numbers (105 g−1 or more). Vibrio vulnificus is a serious human pathogen, responsible for over 95% of seafood-related deaths in the United States (Jones and Oliver, 2009), and carries the highest fatality rate of any food-borne pathogen (Rippey, 1994). A review of 459 U.S. cases reported to the Food and Drug Administration (FDA) between 1992 and 2007 revealed that 51.6% of patients infected with V. vulnificus died (Jones and Oliver, 2009). This striking case fatality rate (CFR) is similar to a range of category Biosafety Level (BSL) 3 and 4 pathogens, such as anthrax, bubonic plague, Ebola and Marburg fever. Wound infections associated with V. vulnificus are usually contracted during recreational activities such as swimming, fishing and handling seafood (in particular shellfish) (Oliver, 2005; Baker-Austin et al., 2012b). Severe wound infection are often characterized by necrotizing skin and soft-tissue infection, including fasciitis and gangrene (Horseman and Surani, 2011). Most cases occur in immuno-compromised males, or patients with underlying conditions resulting in elevated serum iron levels, primarily alcohol-associated liver cirrhosis (Bross et al., 2007). Most strikingly, V. vulnificus infections are characterized by a short incubation period between the onset of symptoms and subsequent clinical outcome (Baker-Austin et al., 2008), typically within 24 h of exposure (Jones and Oliver, 2009). Because this bacterium can cause such severe infections, such as blood-stream and wound infections, prompt antimicrobial therapy is essential (Iwamoto et al., 2010). Thus, the need to accurately and rapidly identify this bacterium in clinical settings is paramount (Baker-Austin and Oliver, 2016), especially where infections in individuals with underlying risk conditions are suspected. Vibrio vulnificus is currently subdivided into three biotypes based on genomic, biochemical and serological features, as well as host range (Tison et al., 1982; Bisharat et al., 1999; Baker-Austin et al., 2012a), which can complicate identification. While an obligate halophile, V. vulnificus is rarely isolated from waters with salinities that exceed 25 parts per thousand, so occurrence and human infections are infrequent in such environments (e.g. the Mediterranean Sea). It also rarely occurs when water temperatures are less than 13°C, so that the vast majority of cases occur in the warmer summer months of May to October (Fig. 1B). Vibrio vulnificus is known to enter a viable but nonculturable state at temperatures below 13°C, thus the inability to culture the organism at low temperatures does not necessarily mean the cells are not present (Oliver, 2015). Within these limits, this pathogen occurs in estuarine environments worldwide, but with an increasing incidence and geographic distribution due to global climate change (Paz et al., 2007; Martinez-Urtaza et al., 2010; Baker-Austin et al., 2012b).
A. Wound infections caused by V. vulnificus are rare, but often life changing. This example shows a patient following a fishing-related infection where significant tissue debridement was necessary. A common characteristic of these wound infections is significant tissue destruction caused by the potent proteolytic/tissue degrading enzymes produced by V. vulnificus. B. Seasonal distribution of Vibrio vulnificus cases reported in the USA, 1988–2010, compared to average Gulf of Mexico Sea Surface Temperature. A clear seasonal distribution of cases, with infections aggregated in the summer months, is clearly evident. [Colour figure can be viewed at wileyonlinelibrary.com]
Numerous characteristics associated with this pathogen are striking, including aspects related to epidemiology (expanding geographical spread of cases, gender and age disparity of cases, low incidence of disease), virulence (rapidity of infections, including the high CFR, lack of defined virulence factors) and genomics (interesting genotypes and host-specificity, sequence heterogeneity). We provide here an overview of some of the most striking aspects related to V. vulnificus and associated diseases, identify key research questions and provide a framework for future studies to address the most important and pressing research areas related to this important and emerging human pathogen.
Key insights and questions from epidemiological studies
Several extremely interesting observations can be gleaned by reviewing epidemiological data associated with this pathogen. A good starting point for this analysis is in the United States, where long-term epidemiological datasets can be used to assess key aspects related to this pathogen. Vibrio vulnificus carries the highest mortality rate of any foodborne pathogen (Rippey, 1994). More recently, we obtained data from COVIS (Cholera and Other Vibrio Information Service), a CDC-dataset that maintains epidemiological data on vibrios in the United States. During 1988–2010, 1874 V. vulnificus cases and almost 600 deaths were reported to COVIS. The most common source of V. vulnificus was wound infections (n = 861, 17.6% fatality rate) followed closely by blood isolations (n = 808 cases, 45.9% fatality rate). We have noticed a sizeable increase in wound infections in the last two decades. An average of 24 wound associated cases were reported each year from 1988 to 1999, compared to 52 per year between 2000 and 2010. Wound infections, although rare, can cause significant and life altering infections (Fig. 1A); these infections also frequently progress to sepsis. A clear and extremely interesting gender disparity is observed with this pathogen. Jones and Oliver (2009) reported that 85.6% of V. vulnificus cases reported to the FDA between 1992 and 2007 were in males. Our data (using a significantly larger dataset) are strikingly similar, with males accounting for 86% of V. vulnificus cases reported to COVIS from 1988 to 2010 (Fig. 2A). This gender disparity is fascinating, with males approximately six times more likely to contract a V. vulnificus infection than females. What factors could account for this skewed gender discrepancy? Estrogen has previously been implicated in the protection of females against the endotoxic activity of V. vulnificus, and may help explain the apparent disproportionately fewer number of females developing this infection compared to males (Merkel et al., 2001). A possible further explanation may be behavioural (e.g. women disproportionately less likely to consume seafood, or engage in recreational activities such as swimming), although we have noticed a similar gender disparity between these two routes of transmission (wounds and sepsis following seafood consumption). As such, it is unlikely that behaviour alone drives this observed risk. A more cogent explanation can be derived from recent epidemiological data, such as liver cirrhosis trends in the United States (Scaglione et al., 2015), which show a similar correlation between gender and age to that of observed V. vulnificus cases, with a disproportionate rate of cirrhosis in males and in older age groups (e.g. > 40 years old). Previous studies have indicated that individuals with compromised immune systems or chronic liver disease such as cirrhosis are up to 80 times more likely than healthy individuals to develop V. vulnificus primary septicaemia (CDC, 1993). The epidemiological data are corroborated by laboratory studies that have investigated the role of iron in disease, as elevated serum iron levels is a notable feature of many infected individuals (Jones and Oliver, 2009). Previous virulence studies using iron overloaded mice models showed a significant increase in the lethality of V. vulnificus compared to controls (Wright et al., 1981; Thiaville et al., 2011). Another startling observation from these data is that both the number of cases and fatality rates associated with V. vulnificus infections are highly aggregated with age, with just one fatality reported in the < 25 years age group (n = 70 cases, 1.4%) compared to a fatality rate of 49.1% in cases aged 40–60-year-old age groups (Fig. 2B). The demographic cohort that represents the greatest risk to V. vulnificus infections (> 45 years of age, male, underlying risk conditions such as liver disease) based on a variety of studies may be simply explained by an increase in underlying risk factors in this group. Similarly, although vibrio epidemiological data is not gathered systematically in other countries, where only fragmentary surveillance data typically exist, most reported infections that have been observed in Europe (Baker-Austin et al., 2010, 2012b), China (Zhao et al., 2015) and South America (Raszl et al., 2016) tend to impact older males with underlying risk conditions.
A. Epidemiological surveillance data of V. vulnificus cases reported in the USA, 1988–2010 (COVIS, CDC) show a clear gender disparity with regards to reported infections, with males making up almost 90% of reported infections. The underlying factors behind this gender disparity are unknown, but a greater proportion of men carrying underlying risk conditions (in particular liver cirrhosis) is thought to play a role. B. Epidemiological surveillance data (COVIS, CDC) reveal that age is also a key contributory factor in driving V. vulnificus infections. The middle age demographic (here shown as 41–60 years of age) contribute almost 50% of infections and carry a 50% mortality rate, highlighting the significant case fatality rate associated with infections. Of note, very few cases are observed in the younger demographic, e.g. individuals aged 0–30 years old. [Colour figure can be viewed at wileyonlinelibrary.com]
A useful approach to improve our understanding of the global burden of infections would be to establish a mechanism whereby epidemiological data are shared in a central repository. This could take the form of an epidemiological database where anonymized case data could be presented to provide important insights into possible exposure routes using a globally representative dataset. Unfortunately, few countries, with the exception of the United States (Newton et al., 2012), maintain dedicated and legally enforced surveillance systems for these pathogens, which greatly hampers our understanding of a global overview of V. vulnificus disease. This is particularly relevant in regions such as Europe where vibriosis is not a notifiable infection (Baker-Austin et al., 2010) but may be emerging with global warming (Baker-Austin et al., 2012b; Le Roux et al., 2015).
Why are infections so rare and are they increasing?
A perplexing issue with regards to the epidemiology of V. vulnificus is the low number of reported infections which, coupled to the high mortality rate, raises important questions regarding why this pathogen is not more prevalent. We know that a range of medical conditions can substantially increase the risk of acquiring a V. vulnificus infection, including chronic liver disease, diabetes, hemochromatosis and compromised immune systems (Klontz et al., 1988; Jones and Oliver, 2009). Recent U.S. FDA estimates suggest that 12–30 million Americans possesses one or more of these risk factors that make these infections more likely (Oliver and Kaper, 2007; Jones and Oliver, 2009). Other factors which are known to be necessary are that the invading cells must be encapsulated (Amako et al., 1984), and that the cells typically must be Biotype 1 and of the ‘C’ (clinical) genotype (see discussion below). It is also likely that a threshold number of cells must be ingested or inoculated into the wound site, although infective doses are not known.
An increase in the incidence of reported infections has been apparent recently (Newton et al., 2012), especially wound infections. The CDC reported Vibrio infections increased by 78% between 1996 and 2006, and in 2005, 121 cases of V. vulnificus disease were confirmed (CDC, 2007). On average, there are around 100 cases (ingestion and wound) reported each year in the United States (C. Baker-Austin, unpubl. data), with wound infections increasingly an issue. The factors underlying this increase in reported infections are complex and likely multifaceted. Several factors may account for this apparent trend. First, an increasingly susceptible population in the United States to vibriosis (i.e. the elderly and individuals with pre-disposing risk factors for more progressive Vibrio infections such as diabetes, immune and blood iron serum conditions) may have played some role in amplifying the numbers of V. vulnificus cases over time. Second, participation in COVIS has varied since its inception in 1988, with a gradual increase in state-by-state involvement until all Vibrio infections became nationally notifiable in the United States in 2007 (Newton et al., 2012). Third, improvements in the knowledge of, as well as clinical diagnosis and identification of, Vibrio pathogens has improved in the last two decades, and are likely to have increased national reporting capacity. Finally, the role of climate change in extending the disease burden associated with these infections should not be discounted. Certainly, heatwave events have been shown to increase risks associated with these bacteria (Paz et al., 2007; Martinez-Urtaza et al., 2010; Baker-Austin et al., 2012b).
The hunt for virulence markers
The reasons why this pathogen is so deadly are poorly understood, and almost certainly multi-faceted. Vibrio vulnificus possesses a wide array of putative virulence factors including acid neutralization, capsular polysaccharide expression, iron acquisition, cytotoxicity systems, motility and expression of proteins involved in attachment and adhesion (Jones and Oliver, 2009). However, unlike related pathogens such as V. cholerae (Waldor and Mekalanos, 1996) and Vibrio parahaemolyticus (Bej et al., 1999), where distinct molecular attributes, such as haemolysin and toxin genes, are normally associated with clinical strains – the hunt for specific molecular markers that can be used to distinguish pathogenic and non-pathogenic V. vulnificus continues. Numerous studies have been published on the putative virulence factors of V. vulnificus, yet it remains to be determined to what extent specific gene products are critical to the ability of this pathogen to cause such rapidly fatal infections. Early efforts to distinguish strains based on virulence potential have included analysis of 16S rRNA sequence homology (Aznar et al., 1994; Nilsson et al., 2003), pilus-type IV assembly proteins (Roig et al., 2010; Baker-Austin et al., 2012a) as well as a randomly amplified polymorphic DNA (RAPD) PCR amplicon associated with clinical isolates, termed the virulence correlated gene (vcg) (Rosche et al., 2005). Unfortunately, no single virulence gene has been identified that reliably correlates with observed virulence. Only the capsule has proven to be absolutely essential, as it is well documented that non-encapsulated cells are readily phagocytosed (Amako et al., 1984). The cell's lipopolysaccharide is an endotoxin, and is the likely cause of human death, a result of significant hypotension and generalized organ failure (Merkel et al., 2001). Also likely critical to successful infections are a MARTX toxin, essential for bacterial dissemination from the intestine (Gavin et al., 2017), with the massive tissue destruction which characterizes both ingestion and wound infections likely resulting from the powerful collagenase, metalloproteases and lipases/phospholipases the bacterium produces (Jones and Oliver, 2009). Both catechol and hydroxymate siderophores are produced (Simpson and Oliver, 1983), but given that the organism requires high serum iron levels in order to produce a successful infection (typically resulting from the chronic liver damage patients suffer), these are likely quite weak. We contend that infections may be driven more by factors associated with host susceptibility than the virulence of the invading bacterium. Complicating our understanding of pathogenesis is the more recent realization that there are three biotypes of this pathogen. Biotype 1 is responsible for both the ingestion cases (‘primary septicemia’) and most wound infections. Biotype 2 is the causative agent of a rapidly fatal septicemia in farmed eels, but only rarely of human infections (Fouz et al., 2007). A further biotype (biotype 3) was discovered in 1996 after an outbreak of V. vulnificus wound infections in an Israeli fish market, and was later found to be a hybrid of biotypes 1 and 2 (Bisharat et al., 1999). To date biotype 3 strains have been geographically restricted to Israel, with the exception of a recent infection reported in Japan (Hori et al., 2017). In addition, cells of all three biotypes are comprised of two genotypes (based on the vcg gene), a clinical (‘C’) genotype, responsible for virtually all septicemia cases, and an environmental (‘E’) genotype, which causes virtually all of the wound infections (Oliver and Baker-Austin, 2016). These genotypes, which reflect substantial differences in DNA sequences (Morrison et al., 2012), correlate well not only with human virulence, but exhibit significant differences in isolation source. Whereas 90% of human clinical isolates are the C-genotype, 85%–90% of cells isolated from environmental sources (shellfish, water, etc.) are the E-genotype. Unfortunately, while whole genome sequencing efforts have been carried out on strains of the two genotypes (Morrison et al., 2012), such studies have to date not definitively elucidated which genes are essential for human virulence or for why E-genotype cells enjoy the significant survival advantage in estuarine environments. Why wound infections, caused by both Biotype 1 and 3 cells, are of the E-genotype, which as a rule do not cause septicemia, is not yet known although whole genome sequencing of more strains will likely point to gene variations that will help elucidate this enigma. The virulence factors required to cause these life-threatening wound infections likely include very powerful tissue-degrading enzymes which lead to the necrotizing fasciitis characterizing these infections (Fig. 1A). Given that wound infections have a gender difference similar to that seen in septicaemia (C. Baker-Austin and J. D. Oliver, unpubl. data), the E-genotype virulence factors are likely similar to those produced by C-genotype cells.
What has whole genome sequencing told us?
The first V. vulnificus genome was published in 2003 (Chen et al., 2003), and these initial insights into the genomic structure of this sequenced strain revealed a number of startling findings. Analysis of the genome identified a range of genes possibly associated with pathogenicity, including an hemolysin (vvhA), the rtx gene cluster for MARTX and three complete secretion systems (Type I, II and VI) as well as iron uptake-related genes. Most interesting from this initial insight was the role of horizontal gene transfer (HGT) in shaping the evolution of the V. vulnificus genome, which was pockmarked with insertion events, integrons and conjugative elements. A significant number of recent efforts has focused on sequencing and comparing strains involved in human infections to those from environmental sources (Gulig et al., 2010; Morrison et al., 2012). These genomic studies have proved fruitful in that they allow a more cohesive understanding of genes associated with human infections and likely necessary for pathogenesis. Genes specific to the clinical strains were identified, including components of sialic acid catabolism, Flp pili, GGDEF proteins, ‘genomic island XII’, mannitol fermentation and a component of a Type IV secretory pathway (VirB4), as well as several other genes with potential significance for human virulence. Genes specific to environmental strains and likely involved in water/oyster survival were also identified. A possible limitation to these studies is the inability to identify appropriate ‘non-pathogenic’ strains. More recent transcriptomic analysis, identifying genes expressed during human infection, potentially provide a more useful approach to determining key pathways of pathogenesis involved in human infections. A recent study by Bisharat identified genes encoding bacterial toxin (RtxA1) and genes involved in flagellar components, Flp-coding region, GGDEF family protein, an iron acquisition system and sialic acid metabolism during infection (Bisharat et al., 2013). Similarly, Williams and colleagues used transcriptomic analysis to identify a ‘virulence profile’ of V. vulnificus grown in human serum (Williams et al., 2014). A number of putative virulence factors were upregulated, including genes involved in intracellular signaling, substrate binding and transport, toxin and exoenzyme production and the heat shock response. Recent sequencing efforts on biotype 3 strains suggest that V. vulnificus biotype 3 evolved from environmental populations and formed a genetically distinct group (Koton et al., 2015). It is clear from the recent gamut of sequencing efforts that the genome of V. vulnificus is highly divergent, impacted by HGT events, yet the key genes involved in pathogenesis remain elusive. Overall, in comparison to other pathogens, very few factors have been tested for a role in pathogenesis across a broad distribution of strain isolates (Phillips and Satchell, 2017). Recently, analysis using whole genome SNP-genotyping data showed formation of the main known evolutionary lineages of V. vulnificus. The genotyping results divided V. vulnificus species into three main phylogenetic lineages and an additional subgroup, clade B, consisting of environmental and clinical isolates from Israel (Raz et al., 2014). This study is particularly noteworthy because using such a large and geographically diverse strain library encompassing both environmental and clinical strains contributes greatly to broadening a basic understanding of the evolution of this human pathogen. We believe that more targeted sequencing efforts, in particular typing and sequencing strains to develop a backbone of shared virulence factors (e.g. genes associated with toxin and exoenzyme production, iron acquisition systems and sialic acid metabolism) in a large number of clinical isolates may be a useful means of identifying key virulence genes which could be used as targets for testing purposes in clinical, regulatory or food microbiology settings.
Concluding remarks
Vibrio vulnificus is an emerging and enigmatic human pathogen. The impact of these infections highlights the need for continued and improved epidemiology and surveillance. It appears that the manifestation of infections has also changed over time, with wound infection cases now predominating, and with cases emerging in areas where Vibrio infections had not previously been identified (Paz et al., 2007; Dechet et al., 2008; Baker-Austin et al., 2012b; Newton et al., 2012). For example, we are now observing infections occur regularly in areas undergoing rapid coastal warming, such as the Baltic Sea (Baker-Austin et al., 2012b). A striking feature of V. vulnificus is the high percentage of fatalities (around 50% of blood infections and around 20% of wound infections) (Jones and Oliver, 2009). The significant fatality rate associated with V. vulnificus infections clearly highlights the need to improve awareness programmes for at-risk groups. Thus, education programs should be amended to highlight the associated risk of these pathogens from recreational exposure as well as from seafood produce. A key issue that remains to be elucidated is the virulence mechanisms associated with this pathogen. Efforts linking genomic and virulence studies and utilising transcriptomic approaches are likely to be fruitful avenues for future work. Linked to this, efforts should be focussed on unambiguously identifying what specific demographic and epidemiological factors drive disease, and use these as useful hypothesis-driven approaches that can be linked to laboratory-based studies.
