Introduction

Buruli ulcer is a necrotizing skin disease prevalent in West Africa and Australia [1]. Although culture of M. ulcerans has traditionally been the gold standard for diagnosis, growth of the organism can take as long as 8 weeks, facilities for culture are often not available in endemic areas, and sensitivity is low. Even where available, culture has a sensitivity of only 35 to 50% [2]. Other methods such as staining for acid-fast bacilli and histology are available (sensitivity is 40% and 63 to 90% respectively), but resources for these methods are lacking in most endemic areas. PCR targeting the insertion sequence IS2404 has become a rapid and sensitive tool for diagnosis of Buruli ulcer and is now the gold standard for diagnosis. However, culture, microscopy histology, and IS2404-PCR cannot be used to identify genetic differences between strains.

The inability to discriminate between different strains of M. ulcerans has hampered epidemiological investigations into routes of infection, virulence, and made it impossible to distinguish between relapse and new infection. Molecular epidemiology has benefited from the development of PCR methods based on analysis of restriction fragment length polymorphisms (RFLP), amplified fragment length polymorphisms (AFLP), multi-locus sequence analysis, and PCR of the intragenic regions between insertion sequences IS2404 and IS2606 [3], [4], [5]. While these methods could discriminate between strains isolated in widely separated areas such as Australia, China, Japan, Mexico, and Africa, they failed to discriminate between isolates within a specific geographical locale [3], [4], [5].

More recently analysis of variable numbers of tandem repeats (VNTR) within the M. ulcerans genome has provided insight into strain variability in African isolates of M. ulcerans. PCR targeting two VNTR loci, ST1 and MIRU1, identified three different genotypes among strains isolated from human tissue samples within Ghana [6]. The incorporation of two other loci, locus 6 and locus 19, led to more refined sub-grouping and the finding of a fourth genotype [7]. VNTR typing has also been used successfully to discriminate between strains of M. ulcerans detected in environmental samples collected in Ghana and Benin has also shown strain heterogeneity within aquatic habitats [7], [8], [9].

Genomic data from a number of strains from Ghana has facilitated the development of methods based on the detection of single nucleotide polymorphisms (SNP typing) for studying M. ulcerans transmission pathways and phylogenetic relationships [10], [11]. However, these methods have only been used successfully with pure bacterial cultures [10], [11].

A particular advantage of VNTR profiling has been the ability to not only distinguish between M. ulcerans genotypes, but also to distinguish M. ulcerans from other, recently discovered, mycolactone-producing mycobacteria (MPMs) such as M. liflandii and M. pseudoshottsii which are pathogenic for aquatic vertebrates (Table 1). These MPMs have been isolated from diseased fish and frogs, but their virulence for humans is not known [12], [13], [14]. Finally, VNTR analysis can be used to identify and type organisms in environmental samples [7], [8]. VNTR profiling of DNA from aquatic environmental samples collected from Ghana has also led to the discovery that these MPMs can share the same environments with M. ulcerans [7], [8].

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. VNTR genotypes of M. ulcerans, mycolactone producing M. marinum and M. pseudoshottsii (MPM), and M. liflandii (MPML).

https://doi.org/10.1371/journal.pone.0088007.t001

It has been difficult to understand the very low sensitivity of culture from Buruli ulcer patients because M. ulcerans is readily grown in the laboratory and the lesions typically contain an extremely high bacterial load. One explanation for the low sensitivity of culture is that the organisms are located in discrete foci within lesions and these foci may be missed through sampling error [15]. A second possibility is that some strains of M. ulcerans are not easily cultured.

In order to gain insight into why IS2404-positive patient samples containing massive numbers of M. ulcerans fail to yield a positive culture, we conducted a study to compare VNTR genotypes of M. ulcerans isolated from punch biopsies of Buruli ulcer patients with IS2404-positive, culture negative tissues from Buruli ulcer patients.

Materials and Methods

Two sets of material were used for analysis. The first set of samples contained DNA from punch biopsies of patients with a confirmed diagnosis of Buruli ulcer, but where M. ulcerans was not isolated upon culture. Information as to whether a culture of M. ulcerans was obtained from the punch biopsy was also included (N = 15). A second set of samples (N = 27) contained DNA isolated from pure cultures of M. ulcerans obtained from biopsy material.

From 2006 to 2007, punch biopsy samples were obtained from subjects with Buruli ulcer at the Tepa Government hospital of the Ahafo Ano North District in the Ashanti Region in Ghana. Subjects were provided with study sheets and were recruited only after the study procedures and potential risks associated with participation had been explained and written informed consent obtained. The local ethics committee approved the consent procedure. Ethical approval was obtained from the ethical review committee at the School of Medical Sciences, Kwame Nkrumah University of Science and Technology, Kumasi, Ghana (CHRPE/11/28/06). Samples were processed for diagnostic confirmation by microscopy for acid-fast bacilli, PCR for the IS2404 insertion sequence of M. ulcerans or cultured on Lowenstein-Jensen slopes, as described elsewhere [2].

VNTR analysis targeted four loci: MIRU1, locus 6, ST1, and locus 19. VNTR Primer sequences, PCR thermocycler conditions and controls used are as previously described [6], [7]. The identity of several positive PCR products from each sample, and any ambiguous bands was confirmed by DNA sequencing. PCR products from positive samples were either cloned into the pCR2.1 Topo vector (Invitrogen), or extracted from the agarose gel using QIAquick Spin (Qiagen) according to the manufacturer's instructions in the instance of a doublet band. In this case both bands were extracted. Sequencing was performed using an ABI 3100 automated genetic analyzer (Applied Biosystems). A matrix of VNTR genotypes representing the number of repeats at each designated locus was used for genotype designation as previously described [7] (Table 1).

The ANOVA and independent t-tests were performed using SPSS 19.0 data analysis software. Significance was defined as p<.05.

Results

M. ulcerans isolates from patient tissues exhibited significantly less genetic variability than those identified from culture-negative tissue samples (p = .012).

VNTR typing was successful for 24/27 (89%) M. ulcerans isolates from biopsy tissue from 15 patients. Out of 27 M. ulcerans cultures, 23 produced a genotype matching M. ulcerans genotype C characteristic of the genome strain, M. ulcerans Agy99 (Table 1, Table 2 and 6,7). One sample had a VNTR genotype matching M. ulcerans genotype A (Table 1 and Table 2). Multiple punch biopsies, taken from individual patients (patients 2,8,11, and 14), yielded identical genotypes (Table 2). A genotype for the 3 samples from patient 15 could not be confirmed because no amplification product was obtained at locus 6 (Table 2). However the single copy at MIRU1 excludes the possibility that the patient was infected with M. ulcerans type C. A gel image showing bands from VNTR analysis of representative samples are included as Figure S1.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 2. VNTR genotypes of M. ulcerans isolates from Buruli ulcer punch biopsy tissue.

https://doi.org/10.1371/journal.pone.0088007.t002

DNA samples from IS2404-PCR positive, culture-negative punch biopsies from 15 patients were subjected to VNTR analysis (Table 3). Considerable strain heterogeneity was identified within this sample set. One sample typed as M. ulcerans genotype A (Table 1 and Table 3), three samples typed as M. ulcerans genotype B (Table 1 and Table 3), eight samples typed as M. ulcerans genotype C, and one sample typed as M. ulcerans genotype D. Sample 13 matched a VNTR genotype for mycolactone producing mycobacteria (M. marinum/M. pseudoshottsii, MPM genotype) associated with fish disease. This genotype has not previously been identified in a human patient. It has recently been proposed that the MPM species should be considered ecovars of M. ulcerans. These ecovars have a lower growth temperature requirement than M. ulcerans and cannot be isolated under the conditions used to isolate M. ulcerans [12], [13], [14], [16]. Sample 15 did not match known genotypes for M. ulcerans or MPMs (Table 3). Amplification matched a MPM genotype at three of the four loci, but showed a band higher than control bands at ST1 when viewed on an agarose gel. Sequencing data from this band did not confirm the presence of ST1 DNA.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 3. VNTR genotypes of M. ulcerans culture negative, IS2404 positive punch biopsy tissue.

https://doi.org/10.1371/journal.pone.0088007.t003

Strain heterogeneity was much greater among culture-negative tissue samples than in culturable samples (p = .012), and M. ulcerans genotype C (p = .037) was cultured more often than any other M. ulcerans genotype in the study (p = .037).

Discussion

This is the first investigation to compare genotypes of M. ulcerans from PCR-positive, culture-positive samples with those from PCR-positive, culture negative samples. The low sensitivity of culture for diagnosis [2] has been attributed to the focal distribution of M. ulcerans in infected tissue [15]. However, growing evidence on the genetic heterogeneity of M. ulcerans raises the possibility that some genotypes are more readily cultured than others. Results in this paper provide evidence for both reasons for culture failure.

M. ulcerans genotype C was positively associated with M. ulcerans isolation (p = .037). However, this was also the most frequently identified genotype in culture-negative patient tissues. These data provide some evidence that genotype C may be more readily cultured from patient tissue than other genotypes. However, the possibility that genotype C is present in an environment niche associated with a high frequency of human contact could also explain over-representation of genotype C in patient samples.

The fact that VNTR typing showed significantly greater strain heterogeneity among culture-negative biopsy material than among bacterial cultures obtained from patient tissue (p = .012) is particularly interesting. This could be due to the fact that the specific habitats of strains with genotypes A, B, and D are rarely encountered by humans in the environment, or due to the fact that these genotypes do not grow well under laboratory conditions used to culture M. ulcerans. The distribution of VNTR genotypes from environmental samples from Ghana is quite different than that obtained from the patient samples analyzed here. VNTR analysis of environmental samples from Ghana shows that 68% of environmental samples from invertebrates, macrophytes, biofilm samples, and water filtrand matched M. ulcerans genotype A, 13% matched genotype D, 8% matched M. ulcerans genotype B, and 11% matched M. ulcerans genotype C (7, and work in progress). In contrast, 74% of the DNA samples genotyped in this study from human tissue samples matched genotype C, while only 4.7%, 7.1% and 2.4% matched genotypes A and B and D respectively. These results suggest more frequent human exposure to the environmental niche associated with genotype C. Although genotypes A, B and D are well represented in the environment, human-environmental contact may not favor exposure to these genotypes.

Laboratory diagnosis of M. ulcerans is increasingly based on IS2404 PCR [17]. Although it was initially reported that IS2404 was present only in M. ulcerans, subsequent work led to the identification of IS2404 in a closely related group of organisms in the M. marinum complex: M. liflandii, M. pseudoshottsii and a unique clade of mycolactone-producing M. marinum that are associated with disease in aquatic amphibians and fish [12], [13], [14], [16]. In addition to containing IS2404, these species produce unique forms of the mycolactone toxin. Recent data from whole genome sequencing suggests that all mycolactone-producing mycobacteria should be designated M. ulcerans ecovars [18], [19], [20]. However, the pathogenic potential of M. liflandii, M. pseudoshottsii and mycolactone producing M. marinum for humans is unknown. There is data that mycolactone variants produced by these MPMs have toxicity for human cells [16].

This work provides the first evidence that mycolactone-producing mycobacteria other than M. ulcerans could have pathogenic potential for humans. The finding of a VNTR genotype matching that of mycolactone-producing species associated with fish disease (M. marinum DL strains and M. pseudoshottsii) in IS2404 positive, culture-negative patient tissue points to the potential virulence of these strains for humans. The fact that a single genotype was isolated from each patient rules out the possibility that the lesion containing a MPM genotype was caused by co-infection with M. ulcerans.

A limitation of this work is the small number of patients tested. Our matched sample sets included samples from 15 patients each for a total of 30 patients. Nonetheless, evidence presented here has direct relevance to the poor sensitivity of bacterial culture from M. ulcerans patients and suggests that culture sensitivity might be improved by incubating samples at 25 degrees as well as at 32 degrees.

Tim Stinear, lead author on the M. ulcerans genome paper suggested that the location of mycolactone genes on a plasmid was evidence that there might be numerous mycolactone producing bacterial groups in the environment, only some with potential for causing human disease [21]. Three DNA samples from M. ulcerans isolated from patients with Buruli ulcer had unique VNTR genotypes that did not match known VNTR genotypes. Whether these represent unique mycolactone-producing mycobacteria or simply a failure of amplification could not be determined because of limitations in the amount of sample available.

SNP typing has shown greater strain discrimination among M. ulcerans than VNTR typing [10], [11] but has not yet been successfully used on patient samples or environmental samples. Results presented here support the importance of developing finer molecular tools for identification of M. ulcerans as well as the importance of including a 25°C incubation temperature for improving isolation of pathogenic mycobacterial species from patient samples.

Supporting Information