Microb Ecol (2017) 74:350–361
DOI 10.1007/s00248-017-0946-6
ENVIRONMENTAL MICROBIOLOGY
Seasonal Pattern ofMycobacterium ulcerans, the Causative Agent
of Buruli Ulcer, in the Environment in Ghana
Samuel Yaw Aboagye1,2 & Kobina Assan Ampah1,3,4 & Amanda Ross3,4 & Prince Asare1 &
Isaac Darko Otchere1 & Janet Fyfe5 & Dorothy Yeboah-Manu1
Received: 19 December 2016 /Accepted: 31 January 2017 /Published online: 25 February 2017
# The Author(s) 2017. This article is published with open access at Springerlink.com
Abstract This study aimed to contribute to the understanding of MU along the Densu River basin was 89/434 (20.5 %),
of Mycobacterium ulcerans (MU) ecology by analysing both whilst that of the Offin River basin was 50/1166 (4.3 %).
clinical and environmental samples collected from ten com- The DNA was detected mainly in snails (5/6, 83 %), moss
munities along two major river basins (Offin and Densu) as- (8/40, 20 %), soil (55/586, 9 %) and vegetation (55/675,
sociated with Buruli ulcer (BU) at different seasons. We col- 8 %). The proportion of MU positive samples recorded was
lected clinical samples from presumptive BU cases and envi- higher during the months with higher rainfall levels (126/
ronmental samples from ten communities. Following DNA 1175, 11 %) than during the dry season months (13/425,
extraction, clinical samples were confirmed by IS2404 PCR 3 %). This study indicates for the first time that there is a
and environmental samples were confirmed by targeting MU- seasonal pattern in the presence of MU in the environment,
specific genes, IS2404, IS2606 and the ketoreductase (KR) which may be related to recent rainfall or water in the soil.
using real-time PCR. Environmental samples were first
analysed for IS2404; after which, IS2404-positive samples Keywords Mycobacterium ulcerans . Buruli ulcer . Ghana
were multiplexed for the IS2606 and KR gene. Our findings
indicate an overall decline in BU incidence along both river
basins, although incidence at Densu outweighs that of Offin. Background
Overall, 1600 environmental samples were screened along
Densu (434, 27 %) and Offin (1166, 73 %) and MU was Buruli ulcer (BU), caused byMycobacterium ulcerans (MU), is
detected in 139 (9 %) of the combined samples. The positivity the third most important mycobacterial disease of public health
importance globally after tuberculosis and leprosy [1]. The dis-
ease has been reported in 32 countries worldwide mostly in the
Electronic supplementary material The online version of this article tropical regions with the greatest burden experienced in West
(doi:10.1007/s00248-017-0946-6) contains supplementary material,
which is available to authorized users. African countries along the Gulf of Guinea [2]. Buruli ulcer,
which affects the skin and its underlying soft tissues, begins
* Dorothy Yeboah-Manu usually as a painless papule or nodule under the skin at the site
DYeboah-Manu@noguchi.ug.edu.gh of trauma, but in some individuals, more severe diffuse forms
occur such as a plaque and/or oedema. Failure to treat these early
1 Bacteriology Department, Noguchi Memorial Institute for Medical forms results in gradual erosion of the skin leaving a well-
Research, University of Ghana, P.O. Box LG 581, Legon, demarcated ulcer with wide undermined edges resulting from
Accra, Ghana
the cytopathic action of the plasmid-encoded macrolide toxin,
2 Institute of Environmental and Sanitation Studies, University of mycolactone [3–5].
Ghana, Accra, Ghana
The epidemiology of BU in endemic countries is not fully
3 Swiss Tropical and Public Health Institute, Basel, Switzerland understood. It has a focal distribution of cases where endemic
4 University of Basel, Basel, Switzerland and non-endemic communities are separated by few
5 Victorian Infectious Diseases Reference Laboratory, kilometres [6]. Nevertheless, various studies have linked high
Melbourne, VIC, Australia BU incidence to slow-flowing or stagnant waters and
Seasonality Influences the Distribution of Mycobacterium ulcerans 351
disturbed environment [7, 8]. Rapid changes in landscape [9] because extensive disease and sero-epidemiological studies
such as deforestation, flooding, construction of dams and ar- have shown high exposure of community member to the
tificial lakes for irrigation, mining activities and extending M. ulcerans 18 KDa heat shock protein 65 [6, 33, 34], and
swamps for growing rice and fish breeding have been associ- unlike other communities in Ghana which depend on passive
ated with the emergence of the disease in some communities case report, these sites are active in reporting BU cases to the
[8, 10–13]. national BU control programme (NBUCP) [35].
One of the factors limiting the prevention and control of Following the Ampah et al. [36] method, ten communities
BU is the lack of understanding of the ecology and mode of were selected by simple randomization from a total of 199
transmission of the causative agent. Many features of MU using a randomization tool embedded within the ArcGIS
ecology, including distribution within the environment, niche 10.0 software. Along the Densu River basin, the three ran-
adaptation and host range(s), are still not fully known [9, domly selected communities were Ntabea in the Akim East
14–16]. Theories that have been proposed to explain the which lies upstream of the river, Ashongkrom in the Akwapim
mechanism ofMU transmission include (1) inhalation of aero- South at midstream both in the Eastern region and
solized MU from contaminated water [5], (2) acquisition of Domesampaman in the Ga-West Municipality of the Greater
MU through an insect or vector bite [5] and (3) contamination Accra region. In Offin, the seven randomly selected commu-
of an existing wound or site of trauma by the environment nities for the study were Ntobroso in the Atwima district
such as soil, vegetation andwater among others [17–20]; how- which lies upstream of the river, Akomfore in the Amansie
ever, none of these theories have been confirmed. West district and Achiase, Wromanso and Keniago all in the
A major factor that has limited the understanding of MU Atwima district which lie midstream of the river and down-
ecology is the inability to culture viable organisms from the stream of the river are Mfantsiman and Pokukrom in the
environment [10, 20, 21]. However, the completion of theMU Upper Denkyira district.
genome sequence provided specific targets for DNA-based Sampling was conducted within the two major climatic
detection methods, such as the insertion sequence IS2404, seasons in Ghana, the rainy and dry season. The months of
IS2606 and the plasmid encoding mycolactone, May to July are classified as the major rainy season and
ketoreductase-B domain (KR) [22]. Such methods have been August to October as the minor rainy season with the remain-
used to elucidate tree-dwelling possums and mosquitoes as ing months constituting the dry season. The communities
possible reservoirs [23] and potential vectors [24], respective- along the Densu River basin were predominantly hamlets with
ly, in South-Eastern Australia. In a previous study, certain smaller coverage area, whilst those at the Offin River basin
water bugs were cited as possible vectors in hosting MU in were large communities with extensive coverage area.
its salivary glands [25]; however, until now, no potential res-
ervoirs have been identified in Africa, which harbours most of Buruli Ulcer Active Case Search
BU disease burden.
Seasonal changes are cyclic and represent a major source of We conducted active case search to monitor the emergence of
external variation influencing human and other natural systems BU cases using community outreach programme and monthly
[26–30] and affect diseases such as malaria [31] and diarrhoea household visits by community volunteers. Community out-
[32]. It is not clearly known whether the incidence of BU is reach education was conducted once every 3 months in all the
seasonal. Understanding the local seasonal drivers of MU in selected communities. During the outreach programme, we
the environment which could influence incidence of BU disease educated community members on the transmission, early case
may be of importance in improving the control strategies in detection and treatment of BU by showing BU documentaries
Ghana. In this study, we surveyed the presence of MU in differ- and interacted with community members through questions
ent environmental sources using DNA-based assays and looked and answers. The following morning, the inhabitants were
at seasonality and also possible risk factors within the environ- screened and those with clinically suspected BU lesions were
ment and retrospectively characterized the occurrence of BU sampled for laboratory confirmation. We also employed a
cases for each community based on active surveillance data. monthly household visit-based surveillance as an additional
tool to the surveillance by the community outreach pro-
gramme. We trained and equipped one community-based sur-
Methods veillance volunteer (CBSV) from each of the communities
with android phones (HTC wildfire S) pre-loaded with a BU
Study Site surveillance questionnaire. The questionnaire used for the
study was designed as previously described [36]. Starting
The study was conducted in ten communities associated with from August 2013 to December 2014, we mandated the
BU along two major river basins (Densu and Offin) of Ghana CBSVs to visit all households monthly and record any pre-
(Fig. 1). These river bodies were selected for the study sumptive case using the mobile application and a notebook.
352 Aboagye S. Y. et al.
Fig. 1 Map of Ghana showing study communities along the Densu and Offin River basins
Presumptive cases were then sampled by a local health staff, We estimated the prevalence of BU within each of the com-
and the samples were sent to the Noguchi Memorial Institute munities using passive data obtained from the local health
for Medical Research (NMIMR) for laboratory confirmation. facilities and data from the national active case surveillance.
Seasonality Influences the Distribution of Mycobacterium ulcerans 353
Environmental Sample Collection Biofilms were prepared from vegetation parts using a modi-
fied version of the method described by Gryseels et al. [37].
Two different sampling methods (convenient and random) Samples were emptied into sterile plastic resealable bags, and
were employed along the two river basins. Convenience sam- 50 ml of PBS was added to each bag. The contents of the bags
pling was conducted to collect environmental samples from were vigorously agitated to dislodge the biofilms into solu-
the three communities (Ntabea, Ashongkrom and tion. The suspensions were poured into sterile 50-ml Falcon
Domesampaman) along the Densu River basin in 2011, tubes and centrifuged at 4000 rpm for 30 min, to sediment all
2013 and 2014. We purposely used convenient sampling tech- suspended bacteria. The supernatant was decanted and the
nique to capture specific zones within the environment where resulting pellet was suspended in 10 ml of PBS for the anal-
there is likelihood of human interactions. During the conve- ysis. Water samples were vortexed tomix homogeneously and
nient sampling, we walked through the communities and col- centrifuged at 4000 rpm for 30 min to sediment all suspended
lected environmental samples from sites of frequent human bacteria. The supernatant was decanted and the resulting pellet
activities such as water sources including hand dug wells, was suspended in PBS.
ponds, streams and boreholes which are regularly utilized,
communal bathing areas, school compounds, agricultural Screening of Samples by Real-Time PCR
farms, market grounds and community centres. Samples were
collected from 239 different locations, and at each sample Genomic DNAwas extracted directly from 1600 environmen-
location, we collected soil sample and any other sample within tal samples using the FastDNA SPIN kit for soil with the
1 m reach. Any other samples that were more than 1 m away FastPrep-24TM instrument (MP Biomedicals) according to
from the location point were excluded from the collection. At the manufacturer’s instructions. Negative controls were in-
each sampling location, a distance of about 10 m was allowed cluded at each point of DNA extraction. Detection of MU
between sampling points or about 5 m where space was lim- DNA from the environment has been based solely on
ited and GPS coordinates were taken at each collection point. IS2404 PCR due to the large copy numbers present in the
Solid samples such as animal faeces (sheep, lizard and chick- MU genome [38]. However, there are other organisms that
en), terrestrial insects (using insect net), snails, soil and water also harbour this IS2404 sequence making it non-specific to
were collected aseptically into 50-ml Falcon tubes and vege- MU [39]. In this study, three independent gene targets,
tation parts were collected and pressed into 50-ml Falcon IS2404, IS2606 and KR, within the MU genome were
tubes and then released into sealable plastic bags from which screened. The extracted DNAwas first screened for the inser-
biofilms were prepared. All samples were clearly labelled im- tion sequence IS2404 by real-time PCR using Rotor Gene Q
mediately, kept in a cool pack at 4 °C after collection and (Qiagen). Primers and TaqMan MGB probes from Applied
transported to the laboratory and kept frozen until analysis. Biosystems that were selected from regions of the sequences
Along the Offin River basin, sampling was conducted in for IS2404, IS2606 and KR present on the plasmid
2013 and 2014. Random sampling from a grid of locations pMUM001 were used [22]. Probes IS2404TP and KRTPwere
was used for the collection of environmental samples from labelled with the fluorescent dye 6-carboxyfluorescein (FAM)
seven communities (Achiase, Akomfore, Keniago, at the 5′ end and a nonfluorescent quencher at the 3′ end.
Mfantsiman, Ntobroso Pokukrom and Wromanso). All the Probe IS2606TP was labelled with the fluorescent dye VIC
communities were mapped and divided into grids, and 487 at the 5′ end and a nonfluorescent quencher at the 3′ end [22].
sampling points were randomly selected using a randomiza- The IS2404 real-time PCR mixtures contained 1 μl of tem-
tion tool embedded within the ArcGIS 10.0. Samples were plate DNA, 0.9 μM concentrations of each primer, a 0.25 μM
collected from each of the 487 points generated and treated concentration of the probe, SensiFast (500 nM) mix (Bioline)
in the same way as those from the Densu River basin and and TaqMan exogenous internal positive control (IPC) re-
transported to the laboratory. We collected rainfall data from agents (Applied Biosystems) in a total volume of 20 μl.
the Ghana Meteorological Agency, Accra after monthly rain- IS2606 and KR assays were performed as a multiplex assay
fall level data from the meteorological substations within the (without IPC) for all IS2404-positive DNA with CT value
study sites have been reported. below 35. At each PCR run, two each of negative and positive
controls were added. Amplification and detection were per-
Sample Processing formed using the Gene Q sequence detection system (Qiagen)
according to the following programme: 1 cycle of 50 °C for
Snail and faecal samples were diced with sterile disposable 2 min, 1 cycle of 95 °C for 15 min and 40 cycles of 95 °C for
surgical blades and homogenized using sterile porcelain and 15 s and 60 °C for 1 min. DNA extracts were tested in at least
pestle and suspended in phosphate-buffered saline (PBS). Soil duplicate, and negative controls were included in each assay.
samples were shaken vigorously in sterile distilled water and All DNA samples that were positive for IS2404, IS2606 and
centrifuged at 600 rpm for 5 min to sediment soil particles. KR were classified as MU confirmed. We also determined the
354 Aboagye S. Y. et al.
IS2404/IS2606 copy number ratio which differentiates Of the laboratory confirmed cases, 19 (59.4 %) were males
M. ulcerans from the other mycolactone=producingmycobac- and 13 (40.6 %) were females, aged between 3 and 70 years,
terium following Fyfe et al. [22]. with a mean age of 26. Lesions presented by cases were in the
early stages; 25 (78.2 %) were detected with pre-ulcerative
Laboratory Confirmation of BU Cases lesion and 7 (21.9 %) presented with category II lesions.
We confirmed presumptive BU lesions by collecting two swab
specimens from the undermined edges of ulcerative lesions Detection and Identification of M. ulcerans by Real-Time
and one fine needle aspirate (FNA) collected into 500 μl PCR
PBS as previously described [40] for pre-ulcerative lesions.
Samples were transported to NMIMR at 4 °C and confirmed A total of 1600 environmental samples from ten communities
by a positive IS2404 PCR laboratory test as previously de- associated with Buruli ulcer along both the Densu (434, 27%)
scribed [41]. and the Offin (1166, 73 %) river basins were collected, cate-
gorized and screened for MU DNA using three independent
molecular markers. We sampled from 239 locations in three
Statistical Analysis communities along the Densu and 487 locations in seven
communities in the Offin River basins. The median number
The data collected were entered into a Microsoft Excel 2010 of samples per location was 1 for the communities along the
spreadsheet and analysed using R statistical software [42]. We Densu River basin and 2 for those along Offin River basin.
used logistic regression and accounted for the cluster sampling Overall, 139 (9 %) samples were positive for MU DNA
by including random effects for location and community. We (Table 2).
estimated the proportion positive with 95 % confidence inter- We found MU DNA to be broadly distributed in all the
vals, overall and for different categories of the explanatory communities along both river basins.
variables. In analyses stratified by site or adjusting for calen- Two different sampling techniques were used in the study
dar month, the numbers were too small to allow the random for both the Densu and Offin River basin. M. ulcerans posi-
effects model to converge and so we did not adjust for tivity was significantly higher with the convenience sampling
clustering. method conducted along the Densu River basin (89, 21 %)
than the random sampling method at Offin (50, 4 %) river
body (p < 0.001) (Table 2).
Results Along the Densu River basin, sample positivity for
M. ulcerans DNAwas high at Ntabea (21/37, 57 %), followed
Demographic Characteristics and Identified BU Cases byAshongkrom (61/293, 22%) and the least at Domesampaman
(7/104, 7 %) as shown in S1. Among the seven communities
An overall population of 10,851 inhabitants from communi- along the Offin River basin, Wromanso (16/142, 11 %) had the
ties along Densu (1217, 11.2 %) comprising 45.6 % (n = 555) highest MU positivity and the least MU positivity was recorded
females and 54.4 % (n = 662) males and Offin (9634, 88.8 %) for Akomfore (2/139, 1%), Keniago (2/177, 1%) and Pokukrom
comprising 49.6 % (n = 4785) females and 50.4 % (n = 4849) (2/88, 2 %), respectively (S1).
males were studied. From samples collected along the Densu River basin, we
Within the study communities, we detected 84 presumptive found MU DNA positives among all nine sample types sam-
BU cases; 56 (66.7 %) from Densu and 28 (33.3 %) from pled with a positivity ranging from 83 % among snails to 9 %
Offin River basins, respectively. Thirty-two (38.1 %) were in animal faecal samples (Table 2). However, in the Offin
laboratory confirmed by IS2404 PCR, of which 24 (75 %) River basin, detections were observed among only moss, soil,
were from Densu and 8 (25 %) from the Offin River basin. vegetation and water (Table 2).We detectedMU in at least one
At Densu, 15 (62.5 %) of the confirmed cases were from sample each of vegetation and soil at every sampling period in
passive reports. The confirmed cases were detected in both river basins. Among the samples confirmed to contain
Ashongkrom (16, 66.7 %) and Domesampaman (8, 33.3 %) MU DNA, we found the highest proportion from agricultural
with a prevalence of 14.3 and 3.6 %, respectively, whilst no farms (45 %), followed by water sources within the commu-
case was detected in Ntabea. At Offin, seven (87.5 %) of the nities (36 %) and the least near household (19 %).
confirmed cases were actively reported. The eight cases were Of the 139 samples confirmed to contain MU DNA, the
detected in three communities: Achiase (4, 62.5 %), Ntobroso IS2404/IS2606 copy number ratio, which differentiates
(3, 37.5 %) and Akomfore (1, 12.5 %). The recorded preva- M. ulcerans from the other mycolactone producing mycobac-
lence was 3.8% for Achiase, 1.6% for Ntobroso and 3.0% for terium, was found in most of the samples to be around 2.4, the
Akomfore (Table 1). expected ratio for MU (S2). This IS2404/IS2606 copy number
Seasonality Influences the Distribution of Mycobacterium ulcerans 355
Table 1 Demographic characteristics and identified BU cases in studied communities
Variables Community
Densu Offin
A B C D E F G H I J
Population (n) 378 512 327 1900 1016 3350 303 1949 900 216
No. of households 52 64 48 295 157 413 85 251 121 41
Sex: females, n (%) 184 (48.6) 228 (44.5) 143 (43.7) 939 (49.4) 499 (49.1) 1719 (51.3) 141 (46.5) 972 (49.9) 409 (45.4) 106 (49.1)
Presumptive BU cases 29 27 0 6 3 3 3 7 4 2
Lab confirmed 16 8 0 4 1 0 0 3 0 0
BU prevalence (%) 14.3 3.6 0 3.1 3 1.2 8.9 1.6 1.3 2.3
Prevalence rate is given as the total prevalence comprising both active and healed lesions
A Ashongkrom, B Domesampaman, C Ntabea, D Achiase, E Akomfore, F Keniago, G Mfanstiman, H Ntobroso, I Pokukrom, J Wromanso
ratios were found in 73/89 (88 %) and 37/50 (74 %) positive with higher rainfall levels (126/1175, 11 %) than during the
samples along the Densu and Offin River basins, respectively. dry season months (13/425, 3 %; p < 0.001).
Along the Densu River basin in 2011, we recorded high
MU positivity for the minor and major rainy seasons,
Rainfall Pattern and MU Positivity September 23/34 (65 %) and October (13/17, 76 %), respec-
tively, while MU positivity declined during the dry season,
To better understand the seasonal drivers for the distri- December (2/15, 13 %), as indicated (Table 3 and Fig. 2)
bution of MU in the environment, we compared month- (R = 0.94). In both 2013 and 2014, we observed a drastic
ly MU positivity with monthly rainfall levels using data decline in MU positivity in both the major and minor rainy
obtained from the meteorological substations within the seasons as well as the dry season (Table 3) (p = 0.0002, 95 %
study communities after approval from the Ghana CI = 1.7–6.4).
Meteorological Agency, Accra. Along the Offin River basin, our findings suggested that
We detected MU DNA in at least one sample in all the rainfall over a low threshold was associated with increased
sampling periods throughout the study. The proportion of MU positivity (Table 4 and Fig. 3), but the linear trend ob-
MU positive samples recorded was higher during the months served in the Densu was not apparent. In both 2013 and 2014,
Table 2 M. ulcerans DNA
positivity among samples Samples Densu Offin Total MU confirmed
analysed from communities along
Densu and Offin River basins No. of MU No. of MU No. of MU
samples confirmed samples confirmed samples confirmed
(%) (%) (%)
Detritus 2 1 (50.0) 21 0 (0.0) 23 1 (4.3)
Faeces 44 4 (9.1) 36 0 (0.0) 80 4 (5.0)
Fungi 8 1 (12.5) 3 0 (0.0) 11 1 (9.1)
Insect 8 1 (12.5) 28 0 (0.0) 36 1 (2.7)
Moss 35 7 (19.4) 4 1 (25.0) 40 8 (20.0)
Water 67 7 (10.4) 76 2 (2.6) 143 9 (6.2)
Soil 143 30 (20.9) 443 24 (5.2) 586 54 (9.0)
Vegetation 121 33 (27.5) 555 23 (3.9) 675 56 (8.1)
biofilm
Snails 6 5 (83.3) – – 6 5 (83.3)
Total 434 89 (20.5) 1166 50 (4.2) 1600 139 (8.7)
356 Aboagye S. Y. et al.
Table 3 Rainfall levels and
monthly MU DNA positivity Period of sampling Rainfall levels (mm) Densu
along the Densu River basin
No. of samples MU positive MU positivity rate (%)
September 2011 268.1 34 22 65
October 2011 130.1 17 13 77
December 2011 22.2 15 2 13
July 2013 176.6 24 13 54
August 2013 20.6 18 2 11
October 2013 146.2 20 6 30
December 2013 26.7 22 2 9
July 2014 103.3 28 6 21
August 2014 108.9 36 8 22
September 2014 102.4 42 6 14
October 2014 45.3 61 2 3
November 2014 95.5 41 5 12
December 2014 26.2 76 2 3
Total 434 89 20.5
MU positivity for the minor and major rainy seasons was also A number of infectious diseases including vector, air and
low as observed along the Densu River basin. waterborne diseases are seasonal [26–29]. At present, there
are no clear indications of association between BU and sea-
sonality due to the long incubation period of the disease which
is estimated between 2 and 4 months [43]. Nevertheless, few
Discussion reports from some endemic countries indicate that there may
be differences in occurrence of BU between wet and dry sea-
In this study, we looked for (1) the presence of MU DNA in sons in the tropics. In Cameroun and Papua NewGuinea, high
the environment to contribute to understanding the pathogen rates of BU occur during the dry season [44–46], while in both
ecology, (2) the associations between MU in the environment Ghana and Côte d’Ivoire, the peak incidence of the BU dis-
and other variables including BU endemicity and rainfall pat- ease has also been reported to be at the end of the rainy season
terns and (3) the occurrence of BU cases within some selected [47, 48]. Our finding therefore agrees with these reports as we
communities. We found for the first time that there is a sea- detected more human BU cases during the monthly active
sonal pattern in the presence of MU DNA in the environment, case search surveillance conducted in Ashongkrom which
possibly related to rainfall and also more human BU cases are peaked after the rainy months (Fig. 4). We are of the view that
most likely to be detected after the raining seasons. MU cells that might have been dormant/or buried in the
Fig. 2 a, b Rainfall levels and
M. ulcerans distribution along the
Densu River basin
Seasonality Influences the Distribution of Mycobacterium ulcerans 357
Table 4 Rainfall levels and
monthly MU DNA positivity Period of sampling Rainfall levels (mm) Offin
along Offin River basin
No. of samples MU positive MU positivity rate (%)
August 2013 114.5 64 8 13
October 2013 149.4 292 21 7
December 2013 24.1 208 1 0.5
July 2014 270.9 292 10 2
September 2014 176.4 78 8 10
October 2014 39.5 98 1 1
December 2014 38.3 63 1 2
Total 1166 50 4.1
environment particularly in the soil are exposed during the ratios for the insertion sequences, IS2404/IS2606, for most of
rainfall season due to the erosion and other environmental the positive samples were about 2.4. According to Fyfe et al.,
disturbances that occur during the rainfalls. The exposure rates the average ΔCT (IS2606–IS2404) for M. ulcerans is 2.37;
that occur during these months probably account for the high while for the other mycolactone-producing mycobacteria, it is
numbers of BU cases at the end of the rainy season (in the BU 7.60 [22]. We found 79 % of the analysed samples in the
treatment facility at the AmasamanHospital, Accra, Ghana) or expected range for M. ulcerans, an indication that
during the dry months due to increased agricultural activities, M. ulcerans was being detected (S2). However, the environ-
and this also supports the suggested incubation time [43–45, mental abundance of MU seems to outweigh the observed
47]. This samemechanismmay account for the observation of disease occurrence. This could be due to the presence of
high MU positivity rate from soil and vegetation biofilm sam- non-human pathogenic strains of MU from the environment
ples that were detected at each sampling period (S2). Even probably resulting from mixed populations of bacteria each
though rainfall seems to influence the distribution of MU in carrying PCR targets but not necessarily MU itself. The broad
the environment, our findings suggest that rainfall within a distribution of MU in the environmental setting also clearly
certain threshold was associated with MU positivity. This contrasts the focal distribution of human BU cases. Ntabea for
may mean that water in the ground is important probably for instance has reported no BU case even though community
suspension but not so much water to wash away the bacteria. members are exposed to the 18-kDa shsp-specific antibodies
M. ulcerans has been described as an environmental path- of MU [6]. The monthly active case surveillance confirmed
ogen; this was confirmed in this study by finding MUDNA in this as we detected no human BU case during the study period;
all the communities along both river basins. The copy number however, environmental MU presence was high. This may
Fig. 3 a, b Rainfall levels and
M. ulcerans distribution in along
the Offin River basin
358 Aboagye S. Y. et al.
Fig. 4 Monthly BU case
surveillance for Ashongkrom and
Achiase
imply that community member has high inborn or genetic DNA can harbour metabolically active bacterial communities
protection that has a reduced attack rate with BU. in its gut [52] during feeding on plants and soil [53, 54].
Furthermore, our finding suggests other pathogens and host Considering the high MU positivity rate among the snail sam-
variables may be important for the occurrence of the disease. ples, a larger collection of snails from both river basins to
Moreover, various studies by our group have shown that not further explore the presence of MU in the environment will
all exposed individuals develop overt disease [6, 49]. The be essential. Our finding therefore is consistent with earlier
pathogen variables that will be interesting to explore will be studies by Stinear et al. [49].
the genomic difference that could lead to differences in viru- This study focused on the presence ofMU in the environment
lence between isolates obtained from the study sites. rather than whether it was the source of human infection.
Currently, methods for the in vitro isolation of MU from the However, case–control studies conducted in BU-burdened com-
environment have recently been published [50]. This will pave munities have identified wearing short and lower-body clothing
the way for comparative studies between environment and while farming [55] as risk factors for BU and covering limbs
clinical samples. during farming [56] as protective for BU. The proportions of
Studies have postulated that MU may be preferentially vegetation biofilm and soil samples (S3) with confirmed MU
adapting to specific ecological niches such as plant biofilms DNA from this studymay be relevant particularly for agricultural
due to its inability to produce light inducible carotenoids that farmers in the tropics who often engage in activities without
serve as a shield against incident sunlight [49]. In our study, protective clothing due to high temperatures.
we found high proportions of MU present in snail, moss, Also in most BU-burdened communities, children may
vegetation and soil along the Densu River basin whilst MU swim in stagnant waters that collect during rainfall that might
was more restricted to vegetation and soil at the Offin River contain MU which may expose susceptible hosts to MU. The
basin sites. Aquatic snails have been reported to transiently findings from our study confirm the presence of MU in the
harbour MU without offering favourable conditions for its environment as the sero-epidemiological studies conducted
growth and survival [51]. In contrast, the edible land African along the river basins indicated that sera of individuals above
giant snail (Achatina fulica) from which we detected MU age 5 contained significant amounts of 18 kDa shsp-specific
Seasonality Influences the Distribution of Mycobacterium ulcerans 359
antibodies of MU. This indicates exposure to MU from the Infectious Disease Reference Laboratory (VIDRL), Melbourne,
environment as children below 5 years showed no immune Australia.
response [6, 33, 34].
Competing Interests The authors declare that they have no competing
This study is limited by the use of two different sampling interests.
techniques for the two study sites which may have influenced
the high positivity observed along the Densu site. However, Open Access This article is distributed under the terms of the
Creative Commons Attribution 4.0 International License (http://
we observed a similarity in higher positivity in both vegetation creativecommons.org/licenses/by/4.0/), which permits unrestricted
and soil samples irrespective of sampling technique, which use, distribution, and reproduction in any medium, provided you give
again underpins the MU adaptation preferences for these par- appropriate credit to the original author(s) and the source, provide a link
ticular sample types. The study is limited also by sampling to the Creative Commons license, and indicate if changes were made.
during only certain months of the year. It would be desirable
for sampling to cover a whole 12 months.
In conclusion, the study provides information on the pres- References
ence of MU in the environment and for the first time indicates
the influence of rainfall on its presence in the environment. In 1. Asiedu K, Scherpbier R, Raviglione M (2000) Buruli ulcer,
Mycobacterium ulcerans infection. WHO, Switzerland
addition, we found clinical BU cases peaking after the raining
2. Hotez PJ, Molyneux DH, Fenwick A, Kumaresan J, Sachs SE,
seasons. We therefore recommend that further work on spe- Sachs JD, Savioli L (2007) Control of neglected tropical diseases.
cific environmental sources may lead to human infection and N. Engl. J. Med. 357:1018–1027
potential protective measures are needed. 3. George KM, Chatterjee D, Gunawardana G, Welty D, Hayman J,
BU Buruli ulcer,MUMycobacterium ulcerans, PCR poly- Lee R, Small PL (1999) Mycolactone: a polyketide toxin from
Mycobacterium ulcerans required for virulence. Science 283:854–
merase chain reaction, IPC internal positive control, PBS 857
phosphate-buffered saline, NBUCP National BU Control 4. Portaels F (2000) Diagnosis of Mycobacterium ulcerans disease.
Programme, CBSV community-based surveillance volunteer WHO, Geneva
5. Portaels F, Meyers WM, Ablordey A, Castro AG, Chemlal K, de
Rijk P, Elsen P, Fissette K, Fraga AG, Lee R, et al. (2008) First
Acknowledgements The authors express their profound gratitude to cultivation and characterization of Mycobacterium ulcerans from
Maria Globan and Franca Azzato of the Victorian Infectious Disease the environment. PLoS Negl. Trop. Dis. 2:e178
Reference Laboratory for the real-time PCR training. We further ac- 6. Yeboah-Manu D, Roltgen K, Opare W, Asan-Ampah K, Quenin-
knowledge Adwoa Asante-Poku, Zuleihatu and Grace of the Fosu K, Asante-Poku A, Ampadu E, Fyfe J, Koram K, Ahorlu C,
Department of Bacteriology, Noguchi Memorial Institute for Medical et al. (2012) Sero-epidemiology as a tool to screen populations for
Research for their assistance sample processing. exposure to Mycobacterium ulcerans. PLoS Negl. Trop. Dis. 6:
e1460
Authors’ Contributions SYA, DYM, JF, AR, PA, IDO, and KAA 7. Ross BC, Johnson PD, Oppedisano F, Marino L, Sievers A, Stinear
analyzed the data. DYM, SYA, JF and KAA conceived and designed T, Hayman JA, Veitch MG, Robins-Browne RM (1997) Detection
the experiment. SYA, JF, and PA performed the experiments. SYA, of Mycobacterium ulcerans in environmental samples during an
DYM, JF, AR, PA, IDO and KAA analyzed the data. SYA, DYM, JF, outbreak of ulcerative disease. Appl. Environ. Microbiol. 63(10):
and AR wrote the manuscript. 4135–4138
8. Weir E (2002) Buruli ulcer: the third most common mycobacterial
Compliance with Ethical Standards infection. CMAJ 166:1691
9. van Ravensway J, Benbow MEE, Tsonis AA, Pierce SJ, Campbell
Ethics Statement Ethical clearance for the studywas obtained from the LP, Fyfe JAM, Hayman JA, Johnson PDR,Wallace JR, Qi J (2012)
institutional review board of the Noguchi Memorial Institute for Medical Climate and landscape factors associated with Buruli ulcer inci-
Research (NMIMR) (Federal-wide Assurance number FWA00001824). dence in Victoria. Australia. PLoS ONE. 7:e51074
Written informed consent was obtained from all individuals that partici- 10. Merritt RW, BenbowME, Small PLC (2005) Unraveling an emerg-
pated in the study. Parents or guardians provided written consent on ing disease associated with disturbed aquatic environments: the
behalf of all participants below 18 years of age. case of Buruli ulcer. Front. Ecol. Environ. 3:323–331
11. Merritt RW, Walker ED, Small PL, Wallace JR, Johnson PD,
Benbow ME, Boakye DA (2010) Ecology and transmission of
Consent for Publication The manuscript does not contain any individ-
Buruli ulcer disease: a systematic review. PLoS Negl. Trop. Dis.
ual person’s data in any form. The image contained in this manuscript was
4:e911
generated by the authors.
12. Garchitorena A, Roche B, Kamgang R, Ossomba J, Babonneau J,
Landier J, Fontanet A, Flahault A, Eyangoh S, Guégan JF, et al.
Availability of Data and Material The dataset analysed during this (2014) Mycobacterium ulcerans ecological dynamics and its asso-
study are included in this manuscript as supplementary information files ciation with freshwater ecosystems and aquatic communities: re-
except one dataset which will be made available from the corresponding sults from a 12-month environmental survey in Cameroon. PLoS
author on request. Negl. Trop. Dis. 8:e2879
13. Bratschi MW, Ruf MT, Andreoli A, Minyem JC, Kerber S,
Funding This work was supported by the Stop Buruli Initiative funded Wantong FG, Pritchard J, Chakwera V, Beuret C, Wittwer M,
by the UBS-Optimus Foundation. Travel grant was received from the et al. (2014) Mycobacterium ulcerans persistence at a village water
Holgar P lhman Foundation for real-time PCR training at the Victorian source of Buruli ulcer patients. PLoS Negl. Trop. Dis. 8:e2756
360 Aboagye S. Y. et al.
14. Doig KD, Holt KE, Fyfe JA, Lavender CJ, Eddyani M, Portaels F, 31. Cairns ME,Walker PG, Okell LC, Griffin JT, Garske T, Asante KP,
Yeboah-Manu D, Pluschke G, Seemann T, Stinear TP (2012) On Owusu-Agyei S, Diallo D, Dicko A, Cisse B (2015) Seasonality in
the origin of Mycobacterium ulcerans, the causative agent of Buruli malaria transmission: implications for case-management with long-
ulcer. BMC Genomics 13:258 acting artemisinin combination therapy in sub-Saharan Africa.
15. Ebong SM, Eyangoh S, Marion E, Landier J, Marsollier L, Guégan Malar. J. 14:321
JF, Legall P (2012) Survey of water bugs in Bankim, a new Buruli 32. Rebaudet S, Sudre B, Faucher B, Piarroux R (2013) Environmental
ulcer endemic area in Cameroon. J Trop Med. 123843:8 determinants of cholera outbreaks in inland Africa: a systematic
16. Williamson HR, Benbow ME, Campbell LP, Johnson CR, Sopoh review of main transmission foci and propagation routes. J Infect
G, Barogui Y, Merritt RW, Small PL (2012) Detection of Dis. 208(Suppl 1):S46–S54
Mycobacterium ulcerans in the environment predicts prevalence 33. Röltgen K, Bratschi MW, Ross A, Aboagye SY, Ampah KA, Bolz
of Buruli ulcer in Benin. PLoS Negl. Trop. Dis. 6(1):e1506 M, Andreoli A, Pritchard J, Minyem JC, Noumen D (2014) Late
17. Duker AA, Portaels F, Hale M (2006) Pathways of Mycobacterium onset of the serological response against the 18 kDa small heat
ulcerans infection: a review. Environ. Int. 32:567–573 shock protein of Mycobacterium ulcerans in children. PLoS Negl.
18. Johnson PD, Azuolas J, Lavender CJ, Wishart E, Stinear TP, Trop. Dis. 8:e2904
Hayman JA, Brown L, Jenkin GA, Fyfe JA (2007) 34. Ampah KA, Nickel B, Asare P, Ross A, De-Graft D, Kerber S,
Mycobacterium ulcerans in mosquitoes captured during outbreak Spallek R, Singh M, Pluschke G, Yeboah-Manu D (2016) A sero-
of Buruli ulcer, southeastern Australia. Emerg. Infect. Dis. 13(11): epidemiological approach to explore transmission of
1653–1660 Mycobacterium ulcerans. PLoS Negl. Trop. Dis. 10:e0004387
19. Quek TY, Athan E, Henry MJ, Pasco JA, Redden-Hoare J, Hughes 35. Amofah G, Bonsu F, Tetteh C, Okrah J, Asamoa K, Asiedu K,
A, Johnson PD (2007 Nov) Risk factors for Mycobacterium Addy J (2002) Buruli ulcer in Ghana: results of a national case
ulcerans infection. Southeastern Australia. Emerg Infect Dis. search. Emerg. Infect. Dis. 8:167–170
13(11):1661–1666. doi:10.3201/eid1311.061206 36. Ampah KA, Asare P, Binnah DD, Maccaulley S, Opare W, Röltgen
20. Carolan K, Garchitorena A, García-Peña GE, Morris A, Landier J, K, Pluschke G, Yeboah-Manu D (2016) Burden and historical trend
Fontanet A, Le Gall P, Texier G, Marsollier L, Gozlan RE (2014) of Buruli ulcer prevalence in selected communities along the Offin
Topography and land cover of watersheds predicts the distribution River of Ghana. PLoS Negl. Trop. Dis. 10:e0004603
of the environmental pathogen Mycobacterium ulcerans in aquatic 37. Gryseels S, Amissah D, Durnez L, Vandelannoote K, Leirs H, De
insects. PLoS Negl. Trop. Dis. 8:e3298 Jonckheere J, SilvaMT, Portaels F, Ablordey A, Eddyani M (2012)
21. Wu J, Tschakert P, Klutse E, Ferring D, Ricciardi V, Hausermann H, Amoebae as potential environmental hosts for Mycobacterium
Oppong J, Smithwick EA (2015) Buruli ulcer disease and its asso- ulcerans and other mycobacteria, but doubtful actors in Buruli ulcer
ciation with land cover in southwestern Ghana. PLoS Negl. Trop. epidemiology. PLoS Negl. Trop. Dis. 6(8):e1764
Dis. 9:e0003840 38. Qi W, Kaser M, Roltgen K, Yeboah-Manu D, Pluschke G (2009)
22. Fyfe JA, Lavender CJ, Johnson PD, Globan M, Sievers A, Azuolas Genomic diversity and evolution of Mycobacterium ulcerans re-
J, Stinear TP (2007) Development and application of two multiplex vealed by next-generation sequencing. PLoS Pathog. 5:e1000580
real-time PCR assays for the detection of Mycobacterium ulcerans 39. Mve-Obiang A, Lee RE, Umstot ES, Trott KA, Grammer TC,
in clinical and environmental samples. Appl. Environ. Microbiol. Parker JM, Ranger BS, Grainger R, Mahrous EA, Small PL
73:4733–4740 (2005) A newly discovered mycobacterial pathogen isolated from
23. Fyfe JA, Lavender CJ, Handasyde KA, Legione AR, O’Brien CR, laboratory colonies of Xenopus species with lethal infections pro-
Stinear TP, Pidot SJ, Seemann T, Benbow ME, Wallace JR, et al. duces a novel form of mycolactone, the Mycobacterium ulcerans
(2010) Amajor role for mammals in the ecology ofMycobacterium macrolide toxin. Infect. Immun. 73:3307–3312
ulcerans. PLoS Negl. Trop. Dis. 4:e791 40. Eddyani M, Fraga AG, Schmitt F, Uwizeye C, Fissette K, Johnson
24. Lavender CJ, Fyfe JA, Azuolas J, Brown K, Evans RN, Ray LR, C, Aguiar J, Sopoh G, Barogui Y, Meyers WM, et al. (2009) Fine-
Johnson PD (2011) Risk of Buruli ulcer and detection of needle aspiration, an efficient sampling technique for bacteriologi-
Mycobacterium ulcerans in mosquitoes in southeastern Australia. cal diagnosis of nonulcerative Buruli ulcer. J. Clin. Microbiol. 47:
PLoS Negl. Trop. Dis. 5:e1305 1700–1704. doi:10.1128/JCM.00197-09
25. Marion E, Eyangoh S, Yeramian E, Doannio J, Landier J, Aubry J, 41. Lavender CJ, Fyfe JAM (2013) Direct detection of
Fontanet A, Rogier C, Cassisa V, Cottin J, et al. (2010) Seasonal Mycobacterium ulcerans in clinical specimens and environ-
and regional dynamics of M. ulcerans transmission in environmen- mental samples. Methods Mol Biol Clifton NJ. 943:201–
tal context: deciphering the role of water bugs as hosts and vectors. 216. doi:10.1007/978-1-60327-353-4_13
PLoS Negl. Trop. Dis. 4:e731 42. R Development Core Team (2008) R: a language and environment
26. Altizer S, Dobson A, Hosseini P, Hudson P, Pascual M, Rohani P for statistical computing. R Foundation for Statistical Computing,
(2006) Seasonality and the dynamics of infectious diseases. Ecol. Vienna
Lett. 9:467–484 43. Trubiano JA, Lavender CJ, Fyfe JA, Bittmann S, Johnson PD
27. Moura FE, Perdigao AC, Siqueira MM (2009) Seasonality of influ- (2013) The incubation period of Buruli ulcer (Mycobacterium
enza in the tropics: a distinct pattern in northeastern Brazil. ulcerans infection). PLoS Negl. Trop. Dis. 7:e2463
Am.J.Trop. Med. Hyg. 81:180–183 44. Radford A (1974)Mycobacterium ulcerans infection in Papua New
28. Fuhrmann C (2010) The effects of weather and climate on the Guinea. Papua N Guinea Med J. 17:145–149
seasonality of influenza: what we know and what we need to know. 45. Ravisse P (1977) Skin ulcer caused by Mycobacterium ulcerans in
Geography Compass. 4:718–730 Cameroon. I. Clinical, epidemiological and histological study. Bull
29. Luis AD, Douglass RJ, Mills JN, Bjornstad ON (2010) The effect Soc Pathol Exot Filiales. 70:109–124
of seasonality, density and climate on the population dynamics of 46. Landier J, de Magny CG, Garchitorena A, Guégan JF, Gaudart J,
Montana deer mice, important reservoir hosts for Sin Nombre han- Marsollier L, Le Gall P, Giles-Vernick T, Eyangoh S, Fontanet A,
tavirus. J. Anim. Ecol. 79:462–470 et al. (2015) Seasonal patterns of Buruli ulcer incidence, Central
30. Morris A, Gozlan RE, Hassani H, Andreou D, Couppié P, Guégan Africa, 2002-2012. Emerg. Infect. Dis. 21:1414
JF (2014) Complex temporal climate signals drive the emergence of 47. Amofah GK, Sagoe-Moses C, Adjei-Acquah C, Frimpong EH
human water-borne disease. Emerging Microbes and Infections. 3: (1993) Epidemiology of Buruli ulcer in Amansie West district,
1–9 Ghana. Trans. R. Soc. Trop. Med. Hyg. 87:644–645
Seasonality Influences the Distribution of Mycobacterium ulcerans 361
48. Marston BJ, Diallo MO, Horsburgh Jr CR, Diomande I, Saki MZ, Bacterial diversity in different regions of gastrointestinal tract
Kanga JM, Patrice G, Lipman HB, Ostroff SM, Good RC (1995) of giant African snail (Achatina fulica). Microbiologyopen. 1:
Emergence of Buruli ulcer disease in the Daloa region of Cote 415–426
d’Ivoire. Am J Trop Med Hy. 52:219–224 53. Rauth AM, Barker GM (2002) Achatina fulica Bowdich and other
49. Stinear TP, Seemann T, Pidot S, Frigui W, Reysset G, Garnier T, Achatinadae as a pest in tropicultural agriculture. In: L. R, L RH
Meurice G, Simon D, Bouchier C, Ma L, et al. (2007) Reductive (eds) Molluscs as crop pest. KABI International Publishing, New
evolution and niche adaptation inferred from the genome of Zealand, p. 474
Mycobacterium ulcerans, the causative agent of Buruli ulcer. 54. Chevalier L, Le Coz-Bouhnik M, Charrier M (2003) Influence of
Genome Res. 17:192–200 inorganic compounds on food selection by the brown garden snail
50. Aboagye SY, Danso E, Ampah KA, Nakobu Z, Asare P, Otchere Cornu aspersum (Müller) (Gastropoda: Pulmonata). Malacologia
ID, Röltgen K, Yirenya-Tawiah D, Yeboah-Manu D (2016) 45:125–132
Isolation of nontuberculous mycobacteria from the environment 55. Pouillot R, Matias G, Wondje CM, Portaels F, Valin N, Ngos F,
of Ghanian communities where Buruli ulcer is endemic. Appl. Njikap A, Marsollier L, Fontanet A, Eyangoh S (2007) Risk factors
Environ. Microbiol. 82(14):4320–4329 for Buruli ulcer: a case control study in Cameroon. PLoS Negl.
51. Marsollier L, Sévérin T, Aubry J, Merritt RW, Saint André JP, Trop. Dis. 1:e101
Legras P, Manceau AL, Chauty A, Carbonnelle B, Cole ST 56. Kenu E, Nyarko KM, Seefeld L, Ganu V, Käser M, Lartey M,
(2004) Aquatic snails, passive hosts of Mycobacterium ulcerans. Calys-Tagoe BN, Koram K, Adanu R, Razum O, et al. (2014)
Appl. Environ. Microbiol. 70:6296–6298 Risk factors for Buruli ulcer in Ghana—a case control study in
52. Pawar KD, Banskar S, Rane SD, Charan SS, Kulkarni GJ, the Suhum-Kraboa-Coaltar and Akuapem South Districts of the
Sawant SS, Ghate HV, Patole MS, Shouche YS (2012) eastern region. PLoS Negl. Trop. Dis. 8:e3279