|Home | About | Journals | Submit | Contact Us | Français|
We examined clinical samples from Nigerian patients with acute watery diarrhea for Vibrio cholerae during the 2010 cholera outbreak. A total of 109 suspected isolates were characterized, but only 57 V. cholerae strains could be confirmed using multiplex real-time PCR as well as rpoB sequencing and typed as V. cholerae O:1 Ogawa biotype El Tor. This finding highlighted the need for accurate diagnosis of cholera in epidemic countries to implement life-saving interventions.
Cholera remains a perennial health problem in Nigeria, presenting in sporadic, endemic, and epidemic proportions and causing significant morbidity and mortality. Cholera outbreaks can be observed almost annually in Nigeria (1), including a large outbreak in 2013 and 2014 (2), and it is well known that the disease is dependent on climate and socioeconomic factors (3). The cholera epidemic in Nigeria in 2010, with 41,787 reported cases and a case fatality rate (CFR) of 4.1%, was one of the largest outbreaks in Africa, with 80% of reported cases attributed to women and children (4, 5). The vast majority of strains associated with epidemic cholera are attributed to the V. cholerae O:1 and O:139 serogroups, producing the cholera toxin (CT) (6). Infection usually leads to profuse, watery diarrhea with a tendency to cause severe dehydration and death if left untreated (7). However, some strains of V. cholerae that lack the ctx genes may also be implicated in human illness, but these illnesses are less severe compared to ctx-encoding strains (8, 9). CT is an enterotoxin encoded by the ctx genes from a CTø prophage and acts by suppressing inflammation and activating adenylate cyclase (10).
A cross-sectional study was conducted by African partners between April and September during the 2010 cholera outbreak in Nigeria. Ethical approval for this study was obtained from the ethical committee of Olabisi Onabanjo University with permission further granted by the various stakeholders in each of the states surveyed. Patients presenting with acute watery diarrhea were enrolled at designated cholera treatment centers in eight affected states (Borno, Bauchi, Abia, Gombe, Kano, Katsina, Ilorin, and Ibadan), covering five of the six geopolitical zones in the country. Clinical samples comprising rectal swabs, feces, and vomit were collected from consenting patients at the treatment camps in the states surveyed. Water samples were taken from affected areas. All samples were transported to the designated laboratories within 6 h, and primary identification of V. cholerae was performed using standard tests. After enrichment of bacterial growth in alkaline peptone water, subculturing of samples was performed on agar plates. After 24 h at 37°C, colonies suspected to be V. cholerae were identified using Gram staining, oxidase test, string test, and Voges Proskauer reaction according to the protocols of the Centers for Disease Control and Prevention (CDC; Atlanta, GA, USA). A total of 103 strains were sent for confirmation and further characterization to the Robert Koch Institute (RKI). Additionally, six V. cholerae O:1 biotype El Tor strains recovered from stool (n = 5) and well water (n = 1) samples during the 2007 cholera epidemic in Kano, Nigeria were obtained from the Central Public Health Laboratory (CPHL), a national repository laboratory for the cholera control program of the Federal Ministry of Health (FMOH). These isolates were also sent to RKI and used for comparative analysis with those of the 2010 cholera epidemic. Isolates were streaked onto thiosulfate-citrate-bile salts-sucrose agar (TCBS) and Columbia sheep blood agar. Sucrose-positive isolates (yellow on TCBS displaying cytochrome oxidase activity) were tested by slide agglutination with polyvalent O:1 and monospecific Ogawa and Inaba antisera.
Extraction of bacterial DNA was performed by using the DNeasy blood and tissue kit according to the manufacturer's recommendations. The purity and concentration of eluted DNA were determined photometrically by using a NanoDrop spectrophotometer.
To confirm a V. cholerae species-specific sequence of the sodB gene (superoxide dismutase), multiplex real-time PCR (5′ nuclease assay) was performed on all suspected isolates. In the same assay, the presence of ctxA, which encodes the CT responsible for the profuse secretory diarrhea caused by toxigenic V. cholerae, was detected. Primers and probes were used as described by Messelhäusser et al. (11) but were slightly modified as shown in Table 1. An internal amplification control (KOMA) was added (12). Control DNA from a V. cholerae O:1 biovar El Tor strain from the strain collection of the RKI was included in each run. Control strains are shown in Table 2.
The sequence analysis of rpoB was used to validate the species determination by real-time PCR and to exclude the presence of other human pathogenic Vibrio species, closely related species of the genus Aeromonas, or those of the family Enterobacteriaceae, which also cause gastroenteritis. The amplification of the specific fragment of rpoB was performed by touchdown PCR according to Tarr et al. (13) (Table 1). Sequence determination was performed by applying dye terminator chemistry (Applied Biosystems), and chromatograms were analyzed with the Lasergene software as described earlier (14), followed by sequence comparison by BLAST search (http://www.ncbi.nlm.nih.gov/).
Conventional PCRs for serotyping and biotyping were performed using primers with the nucleotide sequences listed in Table 1. For the determination of serotype, serogroup-specific gene rfb was amplified to distinguish O:1, O:139, and non-O:1/non-O:139 V. cholerae strains. Amplification of tcpA, encoding the toxin coregulated pilus, as well as hlyA, encoding enterotoxigenic hemolysin, was used for the discrimination of classical and El Tor biotypes of V. cholerae strains depending on fragment size. For amplification controls, V. cholerae standard strain O:1 biovar El Tor was included as well as the O:1 classical biovar and O:139 biovar El Tor from the strain collection of the RKI. We confirmed the PCR results on serotyping by agglutination with polyvalent O:1 and monospecific Ogawa antiserum.
The ability of the confirmed V. cholerae isolates to produce and secrete CT in vitro was determined by using a GM1 ganglioside enzyme-linked immunosorbent assay (ELISA) (15, 16). After growth under oxygen-deficient conditions, culture supernatants were obtained by centrifugation. Confirmation of the β subunit of CT was carried out following CDC guidelines (17) with minor modifications. Supernatants were tested in triplicate as well as control supernatants of Vibrio mimicus (negative control) and a toxin-producing V. cholerae O:1 Ogawa strain from the RKI strain collection. Optical density (OD) values were recorded at 492 nm. Samples were considered positive if they showed an OD greater than the cutoff. The cutoff was calculated from the mean OD of negative-control supernatants plus 2 standard deviations, with a value of 0.18 to 0.35 in every approach tested.
Out of 109 isolates primarily identified as V. cholerae by standard tests, only 57 isolates could be confirmed using real-time PCR targeting a species-specific sequence of the sodB gene and by rpoB sequencing. With these two methods, we could clearly exclude other closely related human pathogenic Vibrio species like Vibrio mimicus that can also cause vomiting and diarrhea (18) and Aeromonas spp. These grow on TCBS and present all or in part as yellow colonies and are cytochrome oxidase positive. These are the typical criteria for the identification of V. cholerae. One other problem causing the very low number of confirmed strains was the cultivation of Proteus spp. from the slants sent, which made it impossible to cultivate V. cholerae. Other slants showed no growth after repeated cultivation. Altogether, we analyzed 92 isolates by rpoB sequencing, all grown on TCBS. Twelve of the isolates tested produced yellow to light greenish growths, and most of these 12 isolates were identified as Enterobacter cloacae. This species belongs to the family Enterobacteriaceae, which normally can be easily identified by the negative cytochrome oxidase, just like Proteus and Providencia spp. that were identified five times and the Shewanella sp., Klebsiella pneumoniae, and Salmonella enterica that were identified once each (data not shown). The remaining isolates were members of Aeromonas spp. The 57 V. cholerae strains identified consisted of the six strains (five clinical isolates and one water isolate) from the Kano outbreak in 2007 and two isolates from water samples.
All confirmed V. cholerae isolates were positive for ctxA and analyzed by real-time PCR. For 41 out of 57 strains, production and secretion of CT into the culture medium was displayed. OD values of >2.0 with a cutoff of 0.2 (data not shown) were given for 15 of the tested isolates, indicating considerably large amounts of more than 0.5 μg/ml CT in the supernatant. In contrast, no CT was detected by capture ELISA in the culture supernatants of 16 Vibrio isolates, although a clear ctxA signal was present in real-time PCR. As the growth of the cultures was comparable to that of other CT-producing isolates, the differences may be caused by influences on ctx expression, the process of protein production, or the secretion of toxin. Regarding the severity of disease in patients, no correlation between illness and the amount of CT production by the isolates could be shown (data not shown). This is probably due to the individual (also diet-related) immune status of affected patients, which is additionally influenced by different factors like medical history and potential coinfections. The 57 V. cholerae isolates of this study belonged to the El Tor biotype and the O:1 Ogawa serotype. Our findings partially corroborate the work of Oyedeji et al. (19) who reported the Ogawa serotype of V. cholerae O:1 as the sole cause of the 2010 cholera outbreaks in Bauchi, Gombe, and Borno. In contrast to the findings by Oyedeji et al. (19) who reported the classical biotype for all of the strains analyzed, we detected the biotype El Tor in all 57 confirmed strains, including 36 strains from Bauchi, Gombe, and Borno, in this study. In both studies, PCR analysis of the tcpA gene was used for the identification of biotypes. In our study, we additionally used PCR analysis of hlyA (20) for biotype differentiation, which confirmed the results obtained by tcpA amplification. This finding is consistent with the assumption that the classical biotype is being replaced by the biotype El Tor in the predominant share of V. cholerae infections (21). Both biotypes represent closely related genetic variants of each other and are not easily distinguished. The emerging predominance of biotype El Tor is due to the longer persistence of the bacteria in the environment. The mechanisms of selective advantage are not yet known (22). It is also known that the biotype El Tor is responsible for milder infections, with a lower mortality rate in contrast to the classical biotype (8, 21). Our data and those of Oyedeji et al. (19) strongly suggest—and are in accordance with each other—that V. cholerae O:1 strains of the Ogawa serotype caused the 2010 cholera outbreak in Nigeria. This is in contrast to cholera outbreaks in northern Nigeria in 1985 in which both Inaba and Ogawa strains were implicated (23) and the 2011 outbreak in Calabar, south Nigeria in which the Hikojima serotype was recovered from fecal samples of infected patients (24). Further typing of the isolates using multilocus variable-number tandem-repeat analysis (MLVA) or single nucleotide polymorphism (SNP) would give further information regarding the molecular and epidemiological relationship (25, 26).
Cholera has remained endemic in Nigeria since its first report in 1971, and in recent years, successive epidemics of the disease have occurred across the six geopolitical zones of the country. Data on pathogen and host factors promoting the pathogenesis of severe cholera are strongly needed to formulate appropriate surveillance, control, and treatment measures against future cholera outbreaks in Nigeria (5). Our findings highlight the need for the accurate diagnosis of V. cholerae during cholera epidemics to guide the implementation of appropriate life-saving interventions. This scenario has been reported during cholera outbreaks in many countries where cholera is endemic, such as Bangladesh, Tanzania, and Sierra Leone (27, 28, 29), and currently during the ongoing outbreak in Tanzania (WHO).
We thank Silke Becker and Petra Lochau (both RKI, Berlin) for excellent technical assistance and Ursula Erikli for copy editing. Moreover, we are very grateful to Julia Assmann (RKI, Berlin) and Nasiba Alimatova (National Public Health Reference Laboratory, Dushanbe, Tadzhikistan) for their support.