In the present study, we investigated whether mycolactone was detectable in easily accessible samples of BU patients. We used two chemical approaches, both requiring the extraction of total lipids by organic solvents. The efficacy of this extraction step, as measured by addition of pure mycolactone to control samples, was mediocre and reduced dramatically the sensitivity of the following TLC-Fluo and HPLC determinations. An explanation for such a limited yield of extraction may be that mycolactone associates with biomolecules preventing solvent access, a possibility that we are currently testing by studying the impact of various thermal and enzymatic treatments. If the efficacy and selectivity of mycolactone extraction can be improved, the recently described and field-friendly TLC-Fluo detection method may still be an option for mycolactone-based point-of-care diagnostic tests. If not, it will be necessary to design alternative approaches that do not require this purification step.
Using HPLC/MS/MS, we could demonstrate that mycolactone gains access to the peripheral blood of human patients. In previous studies in the mouse model, we detected structurally intact mycolactone in mononuclear cell fractions of pooled blood samples harvested from mice subcutaneously injected with mycolactone, or experimentally infected with M. ulcerans 
. In the present work, mycolactone could not be identified in blood mononuclear cells. This cell subpopulation was isolated from 10 ml whole blood, and we estimated the maximum yield of mycolactone extraction from mononuclear cell pellets to 20%. If mycolactone effectively reaches blood mononuclear cells in human patients, its cellular concentration may be too low to be detected in the accessible volume of blood. Alternatively, mycolactone may be unstable in the conditions used in the present study to isolate and store mononuclear cell pellets.
In contrast, we were able to demonstrate the presence of structurally intact mycolactone in the serum of 3/5 newly diagnosed BU patients (). This novel information provides the essential proof of concept for the design of BU diagnostic tests based on mycolactone detection in peripheral blood. Since mycolactone is extracted from serum samples with low efficacy, the number of positive samples and the calculated concentrations of circulating mycolactone are probably underestimated in this study. Whether mycolactone kinetics in serum could be employed to monitor the response to antibiotic treatment will certainly be interesting to investigate further. Our preliminary results suggest that, in spite of a sustained presence in ulcer exudates (), mycolactone concentration () showed a tendency to decrease in the serum during antibiotic therapy. If confirmed by longitudinal studies, the decay of circulating mycolactone during antibiotic therapy would provide an explanation to the recovery of cellular immune responses during treatment 
and after surgical excision of BU lesions 
Here we considered patients at ulcerative stages of the disease. BU is usually diagnosed on the basis of clinical symptoms, as the identification of M. ulcerans
by means of cultures or PCR requires dedicated facilities and specialized equipment (reviewed in 
). Common differential diagnoses of BU include other tropical ulcers (venous, phagedenic, neurogenic), leishmaniasis, yaws and squamous cell carcinoma. The presence of biologically active mycolactone was recently demonstrated in skin biopsies of BU patients 
. Here we show that mycolactone can be detected in ulcer exudates obtained non-invasively from wound swabs, in a structurally intact form and at concentrations in the 50–200 ng/ml range, which strongly suggests that mycolactone detection in exudates may be of interest for differential diagnosis of the ulcerative forms. Whether mycolactone is present at pre-ulcerative stages in fine-needle aspirates and in the peripheral blood is currently under investigation.
We observed significant amounts of mycolactone in exudates at the end of antibiotic therapy. In some instances, antibiotic pressure may not efficiently block, and could even enhance the production of a toxin. For instance, in patients infected with E
H7, the use of antimicrobials has been discouraged because it stimulates toxin production and augments the risk of detrimental, or even fatal complications 
. Whether antibiotics promote the expression of mycolactone by M. ulcerans
is unknown. However, their efficacy at killing M. ulcerans
is not questionable with 4 weeks of treatment leading to culture negativity 
. Although a stimulatory effect of antibiotics on mycolactone production cannot be excluded, our observation suggests that mycolactone persists in cutaneous tissues after the demise of M. ulcerans
. Since mycolactone displays inherent ulcerative properties 
, this phenomenon may explain why some BU take considerable time to heal despite remaining culture negative.
The administration of antibiotics is sometimes associated with adverse reactions, excessive inflammation and pain sensation, that paradoxically make the symptoms of infection worse 
. It is possible that in BU, like in the Jarish-Heixheimer reaction in syphilis or in relapsing fevers, efficient killing of M. ulcerans
by antibiotics leads to the sudden and massive release of bacterial antigens locally that act as immuno-stimulants. Mycolactone displays potent immunosuppressive properties in vitro
that are thought to contribute to the cellular response defects of BU patients 
. The rapid decline of mycolactone from the systemic circulation during treatment may thus provoke exuberant inflammatory responses at the level of the lesions, and cause the paradoxical reactions observed in BU disease