Procedures
Infected Participants
After collection of demographic data using standard World Health Organisation forms (BU01) together with a careful history to establish when early lesions (nodules, plaques and ulcers) were first observed, the type and dimensions of lesions will be documented together with digital photographs and tracings onto acetate sheets. For oedematous lesions only digital photographs will be obtained. Patients will be reviewed at 2 weekly intervals during standard antibiotic treatment (rifampicin 10mg/kg and streptomycin 15mg/kg) with further recordings of clinical data as is done for all routine patients. These measurements will enable calculation of rate of healing and healing time in relation to lesion size and type. Patients will be monitored for paradoxical reactions that occur after the start of treatment in about 8% of patients. These procedures are routinely provided as part of routine care of Buruli ulcer patients.
Samples
The additional samples required from patients for the study are an extra volume of blood (7ml) and 3 swabs taken on occasions (at baseline, week 4, 8, 12 and 16) only when the patient has a lesion at these time points. Swabs/fine needle aspirates are required to determine if organisms detected by polymerase chain reaction (PCR) are still viable, one for M. ulcerans culture, one for measurement of mycolactone concentration, and one for combined 16S ribosomal ribonucleic acid (rRNA) reverse transcriptase / insertion sequence IS2404 Real-Time qPCR assay (1).
For patients that develop paradoxical lesions during therapy an additional blood sample and 3 swabs/fine needle aspirates will be obtained at the time of the reaction for viability markers (M. ulcerans culture, combined 16S rRNA reverse transcriptase / IS2404 Real-Time quantitative polymerase chain reaction (qPCR) assay and mycolactone detection).
For oedematous lesions, one swab/fine needle aspirates will be obtained for mycolactone detection and quantification from lesions that are ulcerated. A 3 mm punch biopsy will be performed under local anaesthetic on non-ulcerated lesions for the same purpose only when the patient is 15 years old or above.
Established practice for routine diagnosis of M. ulcerans infection is to take fine needle aspirates from non-ulcerated lesions for acid-fast bacillus (AFB) detection, culture and PCR for IS2404. This specimen is not large enough for mycolactone quantification, an important part of the investigation into the pathogenesis of oedematous disease which often presents without ulceration. Swabs are used in the diagnosis of ulcers.
Controls For the household contacts of patients in endemic villages and healthy controls in non-endemic villages a 7ml blood sample will be obtained to investigate protein biomarkers by Luminex assay, T cell subsets by flow cytometry and immune responses using Enzyme Linked Immunosorbent Assay(ELISA) in a whole blood assay for comparative studies. No tissue biopsies will be taken.
Definition of paradoxical reaction Paradoxical reaction will be defined as an increase in inflammatory changes with increase in lesion size of greater than 100%, after initial improvement and decrease in size; and/or the appearance of new lesions following or during antimycobacterial treatment.
Investigations Detection of rapid responders: Biomarkers, tissue mycolactone concentration and immune response will be measured at baseline. The rate of healing will be estimated by documenting the initial size of the lesion and the time to complete healing as previously (2).
Detection of Mu: Swabs homogenates will be inoculated onto Lowenstein Jensen slopes for culture and microscopy will be carried out on Ziehl-Neelsen stained smears for AFB. PCR for IS2404 will be carried out as described previously (3).
Biomarkers:
Serum samples will be subjected to multi-analyte profiling using Luminex multianalyte profiling to identify markers of oedematous disease, paradoxical reactions and rapid responders.
We will also use mass spectrometry-based proteomics to generate additional serum/tissue biomarkers taking a three-stage approach. First we will use sample fractionation and mass spectrometry (matrix-assisted laser desorption/ionisation/time of flight/mass spectrometry) to determine the degree of variability of proteomic composition in samples from subjects to be treated with antibiotics. This first screening will allow the identification and removal of subjects showing a partial response and subsequent analysis of those showing the most disparate responses. Stage 2 comprises a deep liquid chromatography mass spectrometry (LC/MS/MS)-based quantitative analysis of pooled samples (3 pools generated randomly from each response cohort) from each of the identified response groups. Molecules that show significant changes will be considered as candidate biomarkers of drug response. In Stage 3 immunoassays (ELISA) will be developed in order to test the validity of each novel biomarker in its ability separately or in combination with existing biomarkers to predict drug response.
Immune response: Whole blood samples will be incubated with Mu antigens overnight at 37 degrees celsius and the supernate will be stored for cytokine assays using ELISA as in our previous studies (4).
Antigen stimulated T cell populations will also be studied by flow cytometry to assess the proportion of Interferon(IFN) gamma, Tumor necrosis factor (TNF) and interleukin-2 (IL2) secreting T cells in patients compared to controls.
Mycolactone detection: Lipids will be extracted from homogenised skin biopsies and swabs//fine needle aspirates and the biological activity of mycolactone will be measured by a cytotoxicity assay as described previously using human embryonic lung fibroblasts as target cells. The presence of mycolactone in tissue samples will be confirmed and quantified by mass spectrometry as previously described (5). Mycolactone production by the strain of Mu isolated from oedematous lesions will be quantified in vitro and the tissue concentration of mycolactone in oedematous lesions will be measured.
Sample size and justification There will be 450 participants in this study. Our previous studies showed a mean difference in TNF alpha at baseline between fast and slow responders to be 15pg/ml and a standard deviation of 33pg/ml. The standardized mean difference computed as d=15/33 was 0.45. The sample size required to detect this difference with a two-sided significance of 5% and with 80% power would be 100 participants in each group (that 200 patients). We expect an attrition rate of 20% bringing the sample size to 250 patients. There will be control group of 200 healthy volunteers
Statistical analysis For the purpose of this study a "rapid responder or fast responder" is defined as a patient with a time to healing of less than 12 weeks or rate of healing of greater than 4mm/week and " slow responder or slow healer" is defined as a patient with time to healing of 12 or more weeks or rate of healing of less than 3 mm/week
Data generated by multianalyte profiling will be analyzed by multivariate analysis, including principal component analysis (PCA) and partial least-squares (PLS)-related methods, using the SIMCA 13 software (Umetrics, Sweden) . Univariate analyses will be performed to further characterize candidate biomarkers. Here, differences between controls and disease patients will be analyzed using the non-parametric Wilcoxon rank sum (Mann-Whitney) test. Paired samples for BUD patients at week 0 and week 12 will be compared using the Wilcoxon signed rank test. Univariate analysis and multiple testing adjustments will be performed using R software (version 2.13.2, R Development Core Team, R Foundation for Statistical Computing, Vienna, Austria) and package QVALUE. GraphPad Prism software will be used for graphical representation.
