Introduction
Cholera is a secretory diarrheal disease which is caused by infection with
Vibrio cholerae, and is associated with high mortality if left untreated. Estimates suggest that there are approximately 2.9 million cases in 69 endemic countries every year, resulting in 95 000 deaths, mainly in Sub-Saharan Africa where approximately 60% of cases and 68% of deaths occur annually [
1]. However, in 2015 World Health Organization (WHO) was notified of only 172 454 cases from 42 countries, resulting in 1304 deaths [
2]. The discrepancy between actual and recorded disease burden is thought to result from low reporting for a variety of reasons, including limitations in environmental, community and public health surveillance systems in endemic countries, and financial and commercial concerns on potential consequences for trade and tourism [
1,
3–
5]. The disease is associated with developing countries and poor infrastructure, especially in relation to sanitation and access to clean drinking water. It has been essentially eradicated in developed countries, with cases arising mainly due to travel to endemic countries [
6,
7]. For example, between 2010 and 2011 an increase in travel-associated cases observed in the United States was attributed mainly to travel to Hispaniola in Haiti, linked to a major cholera outbreak there in the aftermath of a devastating earthquake [
7].
V. cholerae strains are natural inhabitants of the marine environment, found attached to multiple surfaces, with water temperature, saline content and seasonality being vital components in distribution and bacterial counts [
8]. This explains the importance of sanitation and access to clean drinking water in cholera spread and also the importance of natural disasters and climate change in causing “spillover” into the human population. The disease is transmitted by the faecal-oral route, and contaminated food and water act as disease transmission vehicles.
Cholera epidemic in Yemen
Yemen, a country with a population of approximately 27 million people at the southern end of the Arabian Peninsula, is currently in the grip of a devastating cholera epidemic. The outbreak was first announced in October 2016, and a surge in cases has occurred since 27th April 2017. The speed of this surge has been described as “unprecedented” by Dr. Nevio Zagaria, the WHO representative in Yemen. WHO and the United Nations Children’s Fund (UNICEF) have estimated that 172 286 suspected cases arose between 27th April and 19th June 2017, associated with 1170 deaths [
9]; on 26th October there was a cumulative total of 862 858 suspected cases. Aid efforts are currently hampered by closure of all air, sea and land ports. The most vulnerable citizens are suffering most, with 46% of cases occurring in those age under 15 years and 33% of fatalities in those aged over 60 years. The epidemic has been designated as a WHO Grade 3 emergency. Spread has been facilitated by the ongoing war in Yemen, which has left 18.8 million people in need of humanitarian assistance. The infrastructure of the country has been left in ruins, leaving little or no access to safe drinking-water or sanitation for more than 14 million people [
10]. Cholera is a disease of poverty, which flourishes and spreads in such conditions. The war has also destroyed the health system in the country; a WHO survey carried out in November 2016 found that only 45% of the facilities surveyed were fully functional and accessible, while 274 health facilities had been damaged as a result of the war [
11]. There are fears that there is a risk of further spread due to the rainy season, as well as high levels of malnutrition and lack of food security for most Yemeni people [
10]. While WHO has responded to the crisis by beginning to set up 350 cholera treatment centers and 2000 oral rehydration points, as well as tracing infection hotspots, this epidemic underlines the need to increase our understanding of cholera and in particular to ensure that the best possible options are available in terms of vaccination and therapy in outbreak situations. In this review, we give an overview of cholera epidemiology, virulence, antibiotic resistance, vaccination and therapy.
Cholera serology
Serologically,
V. cholerae has been classified into more than 200 O serogroups according to variations of the
V. cholerae“O”antigen [
12–
15]. However, only the O139 and O1 serogroups have been identified as causative agents in epidemics [
15,
16]. The O1 serogroup can be sub-classified into three serotypes, Ogawa, Inaba and Hikojima, based on expression of antigenic factors A, B and C, and into two biotypes, Classical and El Tor [
15]. The O1 serogroup is responsible for more than 99% of global cholera cases [
17]. Stool samples from patients in the current outbreak in Yemen have tested positive for the O1 serogroup, Ogawa serotype [
18,
19]. However, O139 arises from time to time, for example as a major cause of infection among older individuals in Pakistan between 1995 and 2010 [
20], as an important pathogenic strain in Thailand [
21], and sporadically in Bangladesh [
22] and China [
23]. Non-O1/non-O139 serogroups occur rarely and are not considered to be of great clinical significance, although they have been associated with cholera cases in Czechoslovakia in 1960, caused by the O5 serogroup [
24] and Sudan in 1970, caused by the O37 serogroup [
25]. Fig.1 shows an overview of the classification of
V. cholerae by serogroup and biotype. The development of molecular tools to study
V. cholerae, including pulsed field gel electrophoresis (PFGE), ribotyping, polymerase chain reaction (PCR)-based methods, and sequencing-based methods such as variable number of tandem repeats (VNTR) analysis, multi-locus sequence typing (MLST) and multi-locus variable tandem repeat analysis (MLVA), means that we now have a better understanding of the evolution and molecular epidemiology of cholera [
26].
Historically there have been six pandemics of cholera prior to 1961, caused by the Classical O1 biotype which emerged in the Indian subcontinent and spread worldwide between 1817 and 1923 [
27,
28]. Currently, we are in the seventh pandemic, which is believed to have originated in Indonesia in 1961 and is caused by the El Tor biotype; it has spread throughout Asia, Africa and Latin America. It has caused large epidemics such as the 2010 epidemic in Haiti, and major outbreaks in several African countries, including epidemics in Mozambique in 1997−1999 and 2012−2014 and Cameroon in 2010 [
4,
15,
29–
31]. Some commentators suggest that there may also be an ongoing eight pandemic, which began in 1992, based on serogroup O139 and “atypical El Tor” [
15,
26]. These atypical El Tor strains have traits of both classical and El Tor biotypes. Pathogenic
V. cholerae strains carry a prophage, CTXФ, which carries the gene encoding cholera toxin, the major cholera virulence factor. Atypical El Tor strains isolated in Matlab in Bangladesh in 1991 and 1994 carried the
ctxB allele which is characteristic of the classical CTXФ (CTXФ
Cla) and are designated “MB El Tor” [
32]. In fact, all El Tor strains isolated in Bangladesh since 2001 express
ctxB and have been designated “altered El Tor.” Variants isolated in Mozambique in a 2004 outbreak contained two copies of a prophage almost identical to CTXФ
Cla on the small chromosome, but had the El Tor phenotype and was otherwise genotypically El Tor (“Mozambique El Tor”) [
33,
34]. Meanwhile, it was also shown that strain VC44 isolated in Kolkata, India in 1992 also carried the classical
ctxB [
33]. Further atypical El Tor strains termed “hybrid El Tor” carrying classical
ctxB and the
rstR gene, needed for phage gene expression regulation, from both classical and El Tor strains were isolated between 1991 and 2004 in various African and Asian countries. These variants have arisen due to transfer of mobile genetic elements. There is as yet no information on the serogroups or biotypes mainly responsible for the Yemen outbreak.
Epidemiology of cholera
The seventh cholera pandemic is the latest manifestation of a disease that has been recognized since the time of Hippocrates [
35]. As mentioned in Section 2.0, we are currently experiencing the seventh pandemic which began in 1961 when the El Tor biotype, named after the town of El Tor in Egypt where it was first identified in 1905, was the cause of an epidemic in Indonesia. In recent years El Tor and “atypical El Tor” strains have caused several major outbreaks and epidemics throughout Asia, Africa and Latin America. Table 1 shows a timeline of some of the important cholera epidemics and outbreaks within the last decade.
Most cholera cases and deaths occur in Africa, especially in sub-Saharan Africa, including the Great Lakes Region, which has had the highest burden of cholera in the world over the past 20 years. For example, among fishing communities of Uganda, there are recurrent seasonal outbreaks of cholera. Between 2011 and 2015, several outbreaks occurred in these communities with a high case fatality rate, associated with contaminated lake water use, and poor sanitation and hygiene [
36]. Cholera outbreaks also occur regularly in Tanzania, with seven outbreaks reported between 2011 and 2016 due to genetically diverse
V. cholerae O1 isolates [
37]. However, both O1 and O139 have been identified in estuaries in Tanzania, although O1 predominates [
38]. Other outbreaks in the African Great Lakes Region include 2009 and 2011/2012 outbreaks in the Democratic Republic of Congo (DRC) and a 2012 outbreak in Zambia caused by
V. cholerae O1 isolates which clustered closely together genotypically according to Multi-locus VNTR (variable number tandem repeat) Analysis (MLVA) [
39].
An International Coordination Group under the WHO Secretariat manages the global cholera vaccine stockpile, and countries or agencies who need to access the stockpile, for example, in the event of an outbreak, must submit a detailed request to this group. While this system has been designed to ensure that scarce vaccine stocks are targeted to areas where they are most needed, the obstacles that can arise were highlighted in another study from the African Great Lakes Region, this time from Southern Sudan [
40]. The study concerned an outbreak that occurred in the city of Juba in June 2015. Like Yemen, this city was suffering from civil strife, which began in December 2013 and resulted in major movement of people in and out of the city, making Juba a potential submission hub for the whole country. Global shortages of cholera vaccine meant that there were insufficient doses to vaccinate the entire at-risk population, which numbered approximately 1 million in Juba alone, with the usual two-dose regime [
40]. After internal consultations, a request was made to the International Coordination Group two weeks after the epidemic start was declared. Three weeks after the International Coordination Group made a decision to release 270 340 doses, a targeted vaccination campaign was eventually agreed among stakeholders, in which a single dose strategy was used and 140 249 doses were targeted to areas of Juba in which transmission was sustained or increasing, as well as to high-risk groups. Remaining doses were used in a program combining sanitation and hygiene promotion, and for use among internally displaced people outside Juba in places where fighting was ongoing and where there had been recent large population movement [
40]. This study has parallels to the current situation in Yemen, and highlights the issues arising from civil unrest and internal displacement of people. It makes clear the challenges to planning effective vaccination campaigns in the face of outbreaks in areas of civil unrest, the lack of data-based guidance on when, where and how vaccines should be used reactively, the time that can be lost when a country’s surveillance and laboratory capacity is undermined, in applications to the International Coordination Group, and consultations with stakeholders, and the difficulties inherent in prioritising competing needs in the context of a global vaccine shortage [
40]. Another study focused on the issue of cross-border cholera outbreaks in the African Great Lakes Region, for example, on the borders between Malawi and Mozambique and between Uganda and DRC [
41]. These cross-border outbreaks accounted for 603 cases and five deaths in Malawi and Uganda in 2015. Contributing factors included poor sanitation and hygiene, contaminated water and environmental factors such as flooding. Children of school age or aged under five years were the most commonly affected age groups. Typically, there were only unilateral efforts by one affected country to control such outbreaks and they were increased by frequent cross-border movements between countries [
41]. This type of cross-border outbreak could be a risk in the current situation in Yemen, if affected individuals crossed borders into neighboring countries such as Saudi Arabia or Oman. Other studies from the Middle East have shown how travel between neighboring countries can contribute to cholera spread. For example, in a study on cases arising in Iran in 2013, 83% of cases were associated with individuals traveling from Afghanistan or Pakistan, and were dominated by the
V. cholerae O1 Inaba serotype, with evidence of increasing antibiotic resistance compared to previous years [
42].
Outbreaks of cholera in West Africa and in Zimbabwe have highlighted the importance of seasonality and climate [
43,
44]. In a study on cholera outbreaks which occurred in Cameroon between 2000 and 2012, seasonal patterns of the three waves of outbreaks which occurred in this time period differed significantly depending on climate subzone, emphasizing the need to understand local climate and environmental conditions when considering cholera transmission patterns [
43]. Other studies in Cameroon also identified the importance of poor food preservation methods as a risk factor in contracting cholera, as well as the necessity for improvements in hygiene and sanitation and in water infrastructure in tackling current and preventing future epidemics and outbreaks [
31]. This further highlights the difficulties currently faced in Yemen, where the ongoing war is undermining the water infrastructure and leaving citizens without reliable food sources or food preservation resources.
The issue of contamination of water sources is further underlined in studies in Southern Asia, especially in Bangladesh [
45,
46] and India [
47], and in South-East Asia, for example, in the Philippines [
48]. In Bangladesh, cholera is endemic and epidemic, with approximately 450 000 cholera cases every year [
1,
45]. Both O1 and O139 biotypes are found in ponds, river water and irrigation canals [
46], with outbreaks displaying a robust seasonality [
49]. As in other areas, the emergence of increased multi-drug resistant (MDR) strains is causing concern. In a study on
V. cholerae O1 isolates from remote coastal areas of Bangladesh, 93% of isolates were MDR, with evidence of reduced susceptibility to ciprofloxacin and azithromycin among other antibiotics [
50]. Outbreaks and epidemics in India have been associated with both El Tor and atypical El Tor O1 isolates. A tetracycline-resistant O1 El Tor variant was associated with a 2010 epidemic in the tribal area of Odisha in India [
51], and a hybrid O1 El Tor carrying the
ctxB7 allele was associated with two outbreaks that occurred in South-West India in 2012, originating with faecal contamination of the potable water supply leading to the first outbreak, which then spread to other areas and caused the second outbreak [
47]. Another hybrid El Tor variant was responsible for a large outbreak in the Philippines in 2011, which was found to be indigenous and present locally in the aquatic ecosystem [
48]. The Philippine isolates were unique and differed from O1 isolates found in other countries in Asia, or in Africa or Haiti. Again, in India outbreaks tend to be seasonally driven, such as the clusters that occur during the monsoon season every year in Karnataka. During the outbreaks of 2013, an MDR variant of
V. cholerae O1 (Ogawa serotype) was identified in cases in the remote Bagalkot area of Karnataka [
52]. In Mumbai, a waxing and waning of MDR among
V. cholerae O1 El Tor (most commonly Ogawa serotype) has been observed among isolates gathered between 2004 and 2013 [
53]. These studies highlight the need to be vigilant about seasonal changes in Yemen potentially affecting cholera transmission and to be aware of possible MDR arising, making it essential that antibiotic sensitivity testing be facilitated as much as possible. At present it is unclear what the predominant variants are in the Yemen outbreak.
In October 2010, one of the largest ever recorded cholera outbreaks began in Haiti in the aftermath of a devastating 7.0 magnitude earthquake which struck on January 12th, 2010. The cholera epidemic has since caused approximately 700 000 cases and 10 000 deaths to date. The experience of Haiti highlights how this disease flourishes in conditions where infrastructure is damaged, sanitation and hygiene breaks down and the healthcare system is compromised. The Haiti outbreak originated with the introduction in 2010 of the hybrid O1 El Tor carrying the
ctxB7 allele, which also caused outbreaks in India, including the two outbreaks in South-West India in 2012 [
47], as well as outbreaks in Cameroon and Zimbabwe in 2009, and Nepal in 2010 [
54,
55]. Prior to this outbreak, there had been no cases of cholera in Haiti for a century. It was established by genomic epidemiology studies that the epidemic originated due to a single introduction from Nepal, followed by rapid clonal spread [
29]. It has since been widely accepted that the original source of the infection was Nepalese soldiers who arrived in October 2010 as part of the United Nations Stabilization Mission in Haiti (MINUSTAH), although there is still some debate on the matter [
56,
57]. The Artibonite watershed was contaminated with infected sewage, leading to downstream cases within days [
57,
58]. Two years on, toxigenic
V. cholerae of the outbreak strain was still detectable in surface waters [
59], while between April 2013 to March 2014, the number of surface water samples containing culturable
V. cholerae O1 increased more than 5-fold from the previous year, with seasonal water temperatures and precipitation playing a significant role [
60].
The nature of the way in which cholera was introduced to Haiti has resulted in an UN-commissioned independent report with recommendations on pre-deployment interventions to reduce risk of transmission of diseases such as cholera from endemic to non-endemic regions by peacekeepers [
61,
62]. A recent computational modeling analysis suggested that of the three options suggested — screening for
V. cholerae carriage, immunisation with oral cholera vaccines (OCV) or administering prophylactic antibiotic chemotherapies — the latter would be the most efficient and cost-effective, although all would be effective to some extent [
62]. The administration of prophylactic antibiotics would be somewhat controversial in the current era of increased antimicrobial resistance among pathogens.
The altered El Tor variant strain responsible for the Haiti epidemic is part of Wave 3 of the spread of El Tor strains that have evolved since their emergence in 1961. When compared to strains from Wave 1, the Haiti strain is hypervirulent, with greater cholera toxin (CT) and hemolysin production [
63]. It has caused a substantial mortality rate [
4]. Risk factors for transmission in intra-peak periods in Haiti have been identified as close contact with cholera patients, eating food bought from street vendors or drinking/washing dishes with untreated tap water, while protection is associated with prevention messages such as hygiene promotion [
64]. Living in remote areas poorly accessible by road was identified as a mortality risk factor [
65]. This is relevant in the Yemen situation, where the damage to the infrastructure leaves greater numbers of people remote from help.
Cholera virulence genes and CTXФ phage infection process
As mentioned above, pathogenic
V. cholerae strains carry a prophage CTXФ, which expresses the CT-encoding
ctxAB genes. Production of CT is central to
V. cholerae pathogenicity and its ability to cause outbreaks and epidemics, and is responsible for the disease symptoms. CT causes the watery diarrhea associated with cholera by binding to the GM1 ganglioside receptor on erythrocytes, which leads to its internalization and consequent cAMP increases, which then leads to catastrophic water and electrolyte loss in the form of diarrhea [
66,
67]. The non-toxic CT subunit, CTB, from both classical and El Tor biotypes also complexes with blood group determinants when binding to its primary GM1 receptor. This mediates blood-group-dependence of cholera infection; individuals with blood group O suffer the most severe symptoms, probably because blood group H determinant, characteristic of blood group O, binds CT with higher affinity than blood group A-determinant [
68]. CTB can also bind to other mediators beyond as well as GM1 on immune cells such as monocytes, including the Toll-like receptor TLR4, and the immunoglobulin (Ig) superfamily members, TREM2 (triggering receptor expressed on myeloid cells 2) and LMIR5/CD300b (leukocyte mono-immunoglobulin (Ig)-like receptor 5) [
69]. This immunological targeting is implicated in the inflammatory responses induced by
V. cholerae infection.
The CTXФ consists of core and RS2 gene clusters [
15,
70]. The
ctxAB genes are carried on the core cluster, along with other virulence and phage morphogenesis-associated genes including
psh, core-encoded pilin (
cep), pIII
CTX, accessory cholera enterotoxin (
ace) and zonula occludens (
zot). The RS2 cluster contains the genes
rstR,
rstA and
rstB, which are associated with the regulation of CTXФ gene expression, phage replication and phage integration respectively [
15]. In O1 El Tor strains, the CTXФ RS2 and core elements are flanked by satellite phages called RS1, which resemble RS2 but have an extra gene,
rstC, encoding an anti-repressor protein that enhances CTXФ gene transcription [
15].
CTXФ infection and virulence also requires the host-encoded toxin co-regulated pilus (TCP), which is encoded on the VPI-1 pathogenic island and which mediates small intestine colonization. The CTXФ interacts with TCP and TolQRA division proteins for delivery to the host cell cytoplasm, then undergoes rolling circle replication (RCR) as the host replication machinery converts its single-stranded DNA genome into double-stranded [
70,
71]. This allows RCR to proceed, with production of new CTXФ particles [
72]. CTXФ is then irreversibly integrated into
V. cholerae at integration sites called
dif, via the action of host tyrosine recombinases XerC and XerD, part of the chromosomally encoded site-specific recombination machinery [
73]. While integration of non-replicative forms of CTXФ tends to be inefficient, the integration of replicative forms is highly efficient and usually results in multiple tandem integrations. RCR is dependent on the CTXФ HUH endonuclease protein RstA, which is in turn controlled by the CTXФ repressor protein RstR and the host cell SOS machinery [
74]. RstA creates a nick at the CTXФ
ori(+) to give a 5′ phosphotyrosine intermediate and free 3′-OH in replicative CTXФ, and primes the remainder of the replication process to be carried out by host proteins [
74,
75]. Various host proteins have been identified as essential in CTXФ replication including histone-like protein HU, which is needed by RstA to introduce the nick at CTXФ
ori(+), and UvrD, a DNA helicase usually involved in DNA repair [
76]. HU and UvrD are therefore potential candidates for targeting of live attenuated cholera vaccines.
Transcriptional activation of CT and TCP depends on the ToxT protein, a member of the AraC/XylS transcription factor family [
77,
78]. Activation of the
toxT gene depends in turn on synergistic coupling of the membrane-located heterodimers ToxR/ToxS and TcpP/TcpH [
79]. These virulence transcriptional regulator proteins, in particular ToxT itself, are potentially effective targets for cholera therapies. Targeting of the bacterial pathogenesis rather than its survival would have the advantage that resistance would be unlikely to develop.
Cholera therapy
Antibiotic resistance and mobile genetic elements
Therapy for cholera generally relies on rehydration, supplemented by antimicrobial treatment in the case of severe infection and septicaemia, as recommended by WHO guidelines [
80]. A Cochrane systematic review confirmed the improvements that result from antimicrobial treatment, both clinically and microbiologically, with azithromycin and tetracycline emerging as having possibly increased benefits compared to other antibiotics [
81]. There is, however, an increasing problem of emergence of antibiotic-resistant strains of
V. cholerae, for example strains that are resistant to the fluoroquinolones such as norfloxacin, which had previously had consistently high activity against
V. cholerae, and β-lactam-resistant strains which have arisen due to acquisition of extended-spectrum β-lactamases (ESBLs) [
52,
81–
85]. Comprehensive information on antibiotic resistance genes associated with cholera can be sources from the DBDiaSNP database, an open-source facility containing information on mutations and antibiotic resistance genes for a range of diarrhea-causing pathogens including
V. cholerae [
86]. This rise in resistance to previously effective antibiotics has further increased interest in use of other antibiotic classes, such as the macrolide antibiotic azithromycin. A randomized control trial carried out on 120 adult cholera patients in India revealed that a single dose of azithromycin was as efficacious in treatment of watery diarrhea and dehydration as three days of twice-daily doses of norfloxacin [
87]. However, reduced susceptibility to azithromycin has been observed in strains isolated, for example during outbreaks in Bangladesh [
50] and Thailand [
21]. Currently, the tetracycline antibiotic doxycycline is recommended as the first line antimicrobial treatment for adult patients with either
V. cholerae O1 or O139 infection, with erythromycin or azithromycin recommended for children or pregnant women [
84,
88]. However, the rise in antibiotic resistance highlights the importance of reliably establishing antibiotic susceptibility patterns, which is difficult in situations such as the epidemic in Yemen where healthcare services are seriously undermined. There are multiple reasons for the rise in antibiotic resistance in
V. cholerae, of which the most important is horizontal gene transfer via mobile genetic elements such as integrative and conjugative elements (ICEs) of the SXT mobile integron family and plasmids such as the IncA/C plasmid family [
84,
89–
94].
SXT family ICE linear DNA sequences have been particularly important in conferring antibiotic resistance on
V. cholerae strains. Phylogenetic analyses have estimated that between the first and second waves of the seventh pandemic, an SXT ICE of the SXT-R391 family was acquired, coincidental with the dating of the most recent common ancestor (MRCA) of the O1 and O139 biotypes between 1978 and 1984 [
91,
95,
96]. The SXT SXT-R391 family share a common 52 core-gene backbone, which facilitates integration/excision and conjugation, as well as variable regions which include elements such as antibiotic resistance genes [
91]. SXT is named for its conferring of sulfamethoxazole and trimethoprim resistance. Since the first
V. cholerae-associated SXT was identified in O139 MO10 in India (SXT
MO10) [
97], several other SXT elements have been identified in
V. cholerae O1 or O139 strains [
91]. Examples include the hybrid SXT element ICEV
chCHN1307, which was isolated from the ICDC-1307 O1 strain from a patient in China in 1998 [
93]. Along with an IncA/C plasmid, pVC1307 and a chromosomal integron, ICEV
chCHN1307 confers intermediate or complete resistance to 13 antibiotics. Other El-Tor O1-expressed SXT elements include the ICEV
chInd5 and ICEV
chMoz10, both of which have rearrangements in their variable regions when compared to basic SXT [
96,
98]. Genetic analyses indicate that there has been multiple acquisitions as well as homologous recombination between ICE elements, contributing to the diversity in pathogenic
V. cholerae strains [
91]. There is also evidence of mobile genetic elements carrying antibiotic resistance in both environmental
V. cholerae strains and non-O1/non-O139 clinical isolates. For example, in Haiti a mobilizable genomic island (MGI) called MGIV
chHai6 which confers resistance to multiple antibiotics was identified in a non-O1/non-O139 MDR strain HC-36A1, and subsequently all non-O1/non-O139 MDR clinical strains isolated in Haiti [
83]. MGIV
chHai6 can be mobilized by IncA/C plasmids and could therefore be transferred to other
V. cholerae strains in epidemic regions. It has already been identified in strains from India and North and South America [
83]. Meanwhile, in environmental strains class 1 integrons conferring resistance to several antibiotics have been found to persist in environmental non-O1/non-O139 strains from the Amazon region of Brazil, while class 2 integrons conferring resistance to a different range of antibiotics have been found in non-O1/non-O139 environmental strains from India, Bangladesh and Ghana, and one strain of the
V. cholerae O1 Amazonia lineage, suggesting that these elements can be transferred between strains and therefore act as a potential reservoir for further antibiotic resistance in cholera [
99].
Potential cholera therapies
The increase in antimicrobial resistance has also led to increased interest in development of potential new cholera therapies to supplement or replace antibiotic treatment. A summary of some of the pathogen-directed and host-directed agents which have been shown to have potential utility in cholera treatment within the last three years is shown in Table 2. These include use of natural products such as seaweeds and herbal sources, quorum sensing, host-directed therapeutic agents, re-purposing of existing drugs such as ribavirin, use of bacteriophages, and use of toxin inhibitors such as linoleic acid, bithianol, AraC inhibitors, virstatin or dietary minerals [
99–
118].
Use of natural compounds from sources such as seaweed or herbs presents attractive potential source of therapy. Polysaccharides from the seaweed
Gracilaria caudate have been shown to reduce secretory diarrhea, loss of chloride ions and fluid secretion in response to CT in
in vivo studies in mice [
100]. There was evidence that these polysaccharides interacted directly with both CT and its GM1 receptor. In
in vitro studies, the sweet fennel component anethole reduced CT and TCP expression in a
toxT-dependent,
toxR/
toxS-independent manner, via
tcpP/
tcpH repression [
101]. This was mediated by overexpression of the
cyaA and
crp genes of the cAMP-cAMP receptor protein (CRP) system. Anethole also reduced fluid accumulation mediated by
V. cholerae in rabbit ileum, suggesting its possible utility as an anti-cholera agent [
101]. Meanwhile,
in silico screening, absorption, digestion, metabolism, and excretion (ADME) of a range of herbal herbal compounds, with docking on ToxT, indicated that catechin and luteolin were potentially inhibitory of ToxT and had good ADME properties, suggesting that they are potential lead molecules for therapeutic development [
102]. Another compound that has potential utility against ToxT include the unsaturated fatty acid (UFA) linoleic acid, including conjugated linoleic acid (CLA), which is currently commercially available as an over-the-counter weight loss supplement. In
in vitro studies, CLA reduced CT and TCP expression via inhibition of ToxT [
111]. Furthermore,
in vivo studies in an adult rabbit ileal loop model indicated that CT production was reduced in response to CLA after infection with
V. cholerae El Tor strain C6706 and that fluid accumulation was also significantly reduced [
111]. Another ToxT-targeting agent is the synthetic isoquinolone alkaloid virstatin [
113]. It reduces CT expression via ToxT inhibition and reduces intestinal colonisation of infant mice after
V. cholerae infection [
113]. A recent study designed to design, synthesize and characterize a new set of potent ToxT inhibitors used a structure-based approach with the folded conformation of UFAs such as CLA to design model bicyclic compounds [
110]. The resulting “pre-folded” small molecule inhibitors bound more tightly to ToxT than UFAs or virstatin and reduced
in vitro tcp expression; they have potential to be developed as anti-diarrhea therapies both in cholera and in disease caused by other enteric pathogens [
110]. An advantage of all of these ToxT-directed inhibitors is that they target
V. cholerae virulence rather than survival, therefore resistance is less likely to arise. Ribavirin, which is an FDA-approved drug, was recently shown to inhibit the
V. cholerae protein AphB, which is a LysR-type transcriptional regulator (LTTR) involved in
tcpP and
tcpH expression [
106]. Thus, this represents an opportunity to consider an already-approved, safe drug for repurposing in cholera treatment. Dietary minerals including zinc (Zn), manganese (Mn) and selenium (Se) have also been shown to reduce transcription of
V. cholerae virulence genes including
ctxAB,
fliA and
tox R in an
ex vivo mouse intestine model, as well as reducing bacterial motility, Caco-2 cell adhesion and CT expression
in vitro [
114].
Another possible strategy for cholera therapy is use of
V. cholerae lytic phages. They are thought to be important in modulation of
V. cholerae O1 and O139 populations in aquatic environments, with effects on seasonality of cholera outbreaks depending on their levels [
115]. In a study from Mexico, phages with lytic and lysogenic activity were isolated from
V. cholerae O1 strains found in wastewater samples [
108]. Of the 16 phages isolated, only the ØVC8 phage, a member of the VP2-like phage subfamily, had specific lytic activity against
V. cholerae O1 strains. Bioinformatics studies revealed conserved domains within the ØVC8 phage such as immunoglobulin domains that could enable it to bind to mucus substrates, for example in human intestine. Another lytic environmental bacteriophage active against the O1 El Tor Inaba serotype, VPUSM 8, was recently isolated from sewage water in Malaysia, and its complete genome sequence was reported [
109]. It has potential as a therapeutic or biocontrol agent. The potential of such bacteriophages was strikingly demonstrated in a recent
in vivostudy using both infant mouse and rabbit models of cholera showed that administration of a cocktail of three virulent bacteriophages, ICP 1, 2 and 3, up to 24 h before
V. cholerae infection reduced both intestinal tract colonization and diarrhea [
107]. None of the
V. cholerae colonies that survived were resistant to all of the three bacteriophages, while any resistance was mainly due to mutations in genes encoding bacteriophage receptors. This suggests that prophylactic administration of ICP bacteriophages could be an effective method preventing cholera infection, for example in the context of a household of an infected individual [
107]. Bacteriophages have also shown promising results in other animal model studies from India. For example, in a removable intestinal tie–adult rabbit diarrhea (RITARD) model, a mixture of O1 biotype El Tor typing phages reduced diarrhea grade in
V. cholerae MAK 757-infected animals [
116]. Results from another study in rabbits suggested that efficacy of a bacteriophage cocktail in reducing bacterial shedding in a model of
V. cholerae O1 infection depended on oral administration of the phage cocktail after rather than before infection [
117]. Treatment with a phage cocktail subsequent to
V. cholerae MAK 757 infection of mice reduced bacterial colony numbers and inflammatory cytokine levels [
118].
Host-directed therapies are another potential anti-cholera strategy. One example is possible use of cystic fibrosis transmembrane conductance regulator (CFTR) chloride channel inhibitors [
104]. These channels are inappropriately activated in secretory diarrhea associated with cholera. The benzopyrimido-pyrrolo-oxazinedione (R)-BPO-27 inhibits CFTR conductance in epithelial cell cultures and intestine after treatment with CT and also prevented CT-induced fluid accumulation in small intestinal loops in an
in vivo mouse model, with high bioavailability for more than 4 h at non-toxic(R)-BPO-27 concentrations [
104]. Thus re-purposing of CFTR inhibitors is a possible strategy for treatment of secretory diarrhea in cholera. Another approved drug, Bithionol, which was originally approved for helminthic infection treatment, has been found to directly inhibit human caspases-1, -3, -6, -7, and -9, which are involved in apoptosis [
112]. It thereby reduced the
in vitro detrimental effects of CT, as well as a number of other toxins including diphtheria toxin,
Pseudomonas aeroginosa exotoxin A, anthrax lethal toxin, ricin,
Botulinum neurotoxin A and Zika virus [
112]. Entinostat, an aroylated phenylenediamine, was also an effective host-directed therapy used
in vivo in a recent study on
V. cholerae-infected rabbits [
105].
V. cholerae infection reduced the intestinal expression of the antimicrobial peptide CAP-18. Entinostat treatment restored CAP-18 levels and increased the levels of serum zonulin, a gut permeability marker, as well as mediating recovery from cholera and reduction in
V. cholerae stool counts [
105]. Thus etinostat is a potential host-directed therapeutic agent for cholera treatment which could increase CAP-18 and regulate intestinal permeability. Another possible host-directed “therapy” would be manipulation of the gut microbiota in order to exploit the fact that commensal gut microorganisms can apparently antagonise
V. cholerae and reduce its colonisation in an interaction mediated by quorum sensing [
103].
Cholera vaccines
Vaccination is an important element in the battle to prevent cholera transmission, in conjunction with access to safe water, sanitation and hygiene (WASH) [
119,
120]. There are currently three oral cholera vaccines (OCVs) pre-qualified by WHO, namely Dukoral®, Shanchol® and Euvichol®, all of which are killed/inactivated vaccines [
120].
Dukoral®, which is manufactured by SBL Vaccines, contains heat- and formalin-killed
V. cholerae and CT subunit B (CTB) and has shown efficacies ranging between 50% and 88% [
121]. It is licensed for use in individuals over two years of age in many countries. It has also been shown to confer indirect (herd) protection as well as direct protection in a sub-Saharan African setting [
122]
Shanchol®, which is manufactured by Shantha Biotec in India, contains killed whole cells of both
V. cholerae O1 El Tor and Classical biotypes, and of O139 [
120]. Its safety and immunogenicity has been confirmed in both adults and children as young as one year [
123]. It is cheaper to produce than Dukoral® because it does not contain CTB and therefore does not need buffer. It has been used in vaccination programs in many countries. In Haiti, Shanchol® was used to vaccinate 97 774 people as part of a reactive vaccination campaign in 2012 in the context of the ongoing epidemic, as part of a package of measures that also included establishment of cholera treatment centers, promotion of WASH practices, and appropriate use of antibiotics [
124–
126]. This represented the first use of an OCV in an outbreak situation in urban and rural areas since prequalification by WHO. The effectiveness of the measures were confirmed in follow-up studies conducted 37 months after vaccination which demonstrated an efficacy of 97.5% [
127]. It prompted the Ministry of Health in Haiti to incorporate use of OCVs into its national cholera control strategy [
127]. However, despite this demonstration of the effectiveness of a reactive use of Shanchol®, administering such a scheme in Yemen remains complicated by issues such as the local movement of people and the difficulties in achieving WASH measures with a compromised healthcare system and sanitation and hygiene infrastructure, as well as issues such as global shortage of OCVs and the recommendations that the OCVs be administered in a two dose regimen [
38]. Shanchol® was shown to be more effective when delivered as part of routine government services and combined with behavioral interventions aimed at encouraging safe drinking water and hand washing in a large cluster-randomized open-label trial in Bangladesh [
128]. Another double-blind, cluster-randomized, placebo-controlled trial in Kolkata in India confirmed the long-term efficacy during a five-year follow-up study which showed cumulative protective efficacy of 65%, with no evidence of decreased efficacy [
129]. Effectiveness of Shanchol® has also been demonstrated in significantly reducing cholera transmission in settings such as refugee camps [
130].
Euvichol®, which is manufactured by EuBiologics in South Korea, contains killed whole cells of both
V. cholerae O1 El Tor and Classical biotypes, and of O139 and is similar in formulation to Shanchol®. Its safety and immunogenicity was confirmed in a Phase I trial in South Korea [
131], while its efficacy was confirmed in a Phase III trial on 1219 adults and children in the Philippines, in comparison to Shanchol® [
132]. It had similar vibriocidal responses to Shanchol® in both adults and children [
132].
Despite the effectiveness of these prequalified OCVs, there are constraints on their uptake and use. One major issue for developing countries in particular is cost. In considering cost-effectiveness of OCV use in Bangladesh, the Foundation Mérieux and the International Centre for Diarrheal Disease Research (ICDDR) found that use of lower cost OCVs such as Shanchol® would be potentially useful in combination with a range of WASH measures and clinical and rehydration treatments [
45]. Cost-effectiveness was considered greatest for targeting of children aged between 1 and 14 years, while other potentially cost-effective strategies would include targeting everyone aged over 1 year in high risk areas, or targeting urban slums or rural areas with poor water supply quality [
45]. However affordability was identified as a major driver of which strategy would be chosen, while adequate global vaccine availability was also pinpointed as a crucial factor. Data from a vaccination campaign carried out in 2011 using Shanchol® in a high-risk area in Dhaka in Bangladesh was used to determine that total vaccination cost per person was US$3.98, given that 123 661 people received two doses and 18 178 people received one dose [
133]. Vaccine purchase cost accounted for 58% of the cost overall, suggesting that reduction of cost-per-dose would impact significantly on future vaccination campaigns. Cost and global vaccine availability are both factors that have driven studies in which efficacy of using only one vaccine dose, as opposed to the recommended two, has been tested. Clinical trial data from Bangladesh [
134] and from South Sudan [
40] suggest that single-dose regimens can be efficacious. However, efficacy in the Bangladesh study was limited to adults and to children aged over five years, but not younger children [
134]. Another approach would be to consider whether an alternative dosing schedule to the current two-week schedule could be used. In a randomized control trial with 356 participants conducted in Kolkata in India, vibriocidal response rates against
V. cholerae O1 Ogawa in both adults and children were comparable for both two-week and four-week Shanchol® dosing schedules [
135]. Less strict requirements for two-week schedules could improve vaccine uptake in situations where infrastructure is compromised and there is less reliable access to healthcare facilities. It is important that WHO and vaccine manufacturers clarify recommendations on dosing regimens in scenarios such as that currently existing in Yemen, where local logistical issues in exercising adequate surveillance in a country at war are compounding the more general problems. These include the global OCV shortage, the administrative difficulties inherent in submission of requests to the International Coordination Group and accessing sufficient vaccine doses for a reactive campaign in the face of competing demands in other endemic and epidemic regions [
40]. Mathematical modeling has suggested that the optimal strategy for allocation of OCVs from the global stockpile is allocation of most doses to reactive campaigns, unless requests are made late in the epidemic, particularly when stocks are getting low [
136]. This further emphasizes the desirability of having sufficient surveillance mechanisms in place that allow early reporting of epidemics.
Live OCVs have some advantages in some situations over the killed OCVs used in mass vaccination campaigns. Live-attenuated vaccines tend to require only one dose and produce robust and rapid mucosal and protective immunity responses. The live-attenuated vaccine CVD-103-HgR is a single dose vaccine and the only cholera vaccine licensed for use in the United States of America (USA) [
17]. It was approved in 2016 by the Food and Drug Administration (FDA) for use in adults traveling to cholera endemic or epidemic areas. Its efficacy against severe secretory diarrhea has been estimated at 90% ten days post-vaccination and 80% three months post-vaccination [
17]. However, live-attenuated vaccines are expensive to produce, may require refrigeration and may be risky for use in immunocompromised individuals [
120]. Live attenuated vaccine currently under development include the VA 1.3 and VA 1.4 vaccines, both of which have undergone Phase I and II trials in India showing their safety and immunogenicity [
137,
138]. They are based on a non-toxigenic
V. cholerae O1 El Tor strain and provide a potential tool for single-dose vaccination strategies which could be useful in reactive vaccination campaigns and emergencies. However, phase III trials are required to establish their efficacy and reservations remain on the use of live-attenuated vaccines in immunocompromised individuals such as HIV patients. Another live-attenuated cholera vaccine, derived from the 638
V. cholerae O1 El Tor Ogawa strain has also been shown to be safe and immunogenic in a placebo-controlled, double-blind randomized control trial on 120 adults in an endemic cholera area of Mozambique [
139]. Conjugated vaccines are presently at the pre-clinical stage but may eventually provide an option for non-oral vaccination routes, or for supplementation of OCVs in vaccination of children [
140,
141]. Promising results have been obtained
in vivoin mouse models for conjugate vaccines based on the O-specific polysaccharide (OSP) of the lipopolysaccharide (LPS) antigen of O1 Inaba and O1 Ogawa serotype strains [
140,
141].
Summary and perspectives
Cholera is a disease of poverty, as well as a frequent consequence of events such as war and/or natural disasters which undermine hygiene and health services. It remains rampant in many developing countries despite being readily preventable by access to adequate sanitation and hygiene measures and clean drinking water. Its devastating effects have been clear in the epidemic that occurred in Haiti in the aftermath of the 2010 earthquake and in the ongoing epidemic in Yemen, as well as in multiple seasonal outbreaks which occur regularly in sub-Saharan Africa and in Asia. It is readily treated by rehydration therapy if identified in time, although antibiotic treatment in the case of more serious infections has become more complicated in the light of the rise in antibiotic resistance. There are three killed OCVs which have been prequalified by WHO, but there is also a global shortage of vaccines despite the stockpile managed by the International Coordination Group under the WHO Secretariat. Also, in situations of conflict or in the aftermath of natural disasters, problems with disease surveillance complicate early identification of cholera outbreaks and timely applications to the International Coordination Group in the event of an outbreak. While this system has been designed to ensure that scarce vaccine stocks are effectively targeted, the administration process and the vaccine shortage complicates the prompt distribution of vaccines to where they are urgently needed. Vaccination programs are most effective as part of a suite of measures also tackling infrastructure, sanitation and hygiene, which is difficult to establish in conflict situations such as that currently ongoing in Yemen. Clarity is needed on factors such as recommendations on dosing regimens in reactive vaccination campaigns and easier access to the vaccine stockpile in emergency situations. It is also important to be able to clarify which V. cholerae strains are dominant in the Yemen outbreak, and their antibiotic sensitivity and resistance profiles. Meanwhile, research is ongoing on several potential therapeutic options, including therapies aimed at viral virulence, host-directed options, or use of bacteriophages, to try to overcome the rising antibiotic resistance of V. cholerae strains. There are also promising developments in research on live-attenuated vaccines and conjugated vaccines which may be useful in one-dose administration in emergency situations, or for administration by alternative means than the oral route.
Higher Education Press and Springer-Verlag GmbH Germany, part of Springer Nature
PDF
(304KB)
