Detection of Mycolactone A/B in Mycobacterium ulcerans-Infected Human Tissue

Abstract

Background: Mycobacterium ulcerans disease (Buruli ulcer) is a neglected tropical disease common amongst children in rural West Africa. Animal experiments have shown that tissue destruction is caused by a toxin called mycolactone.

Methodology/principal findings: A molecule was identified among acetone-soluble lipid extracts from M. ulcerans (Mu)-infected human lesions with chemical and biological properties of mycolactone A/B. On thin layer chromatography this molecule had a retention factor value of 0.23, MS analyses showed it had an m/z of 765.6 [M+Na(+)] and on MS:MS fragmented to produce the core lactone ring with m/z of 429.4 and the polyketide side chain of mycolactone A/B with m/z of 359.2. Acetone-soluble lipids from lesions demonstrated significant cytotoxic, pro-apoptotic and anti-inflammatory activities on cultured fibroblast and macrophage cell lines. Mycolactone A/B was detected in all of 10 tissue samples from patients with ulcerative and pre-ulcerative Mu disease.

Conclusions/significance: Mycolactone can be detected in human tissue infected with Mu. This could have important implications for successful management of Mu infection by antibiotic treatment but further studies are needed to measure its concentration.

Figure 1. Chemical structure of mycolactone A/B showing the lactone core ring and polyketide side chains.

Figure 2. Detection of mycolactone A/B by thin layer chromatography.

A. 20 µl of two-fold serial dilutions of mycolactone A/B at concentrations from 125 to 1 µg/ml were spotted and examined under UV-light and by oxidative charring. The detection limit on TLC was at a concentration of 2–8 µg/ml (8 µg corresponding to 160 ng of mycolactone A/B). B. Each track represents one sample. M is purified mycolactone A/B; tracks 1 and 2 are positive controls with100 µg of purified mycolactone; tracks 3 and 4 are samples extracted from human skin spiked with 100 µg of purified mycolactone; tracks 5 and 6 are negative controls from healthy human skin; tracks 7 to 16 are extracts from infected human skin samples. Mycolactone A/B was the predominant UV-active band with an Rf of 0.23 in positive controls and in ASLs from infected human skin. There were perceptible signals from patients 7, 8, 11, 14 and 15 whereas samples from 9, 10, 12, 13 and 16 were below the detection limit.

Figure 3. Detection of mycolactone by mass spectrometry.

A. MS analysis of ASL from Mu infected human skin showing a molecule with m/z 765.5 which represents the sodium adduct of mycolactone A/B [M+Na⁺]. B. MS-MS analysis of this ion produced the core lactone ring of mycolactone with m/z 429.4 and the polyketide side chain of mycolactone A/B with m/z 359.2.

Figure 4. Cytotoxicity of ASL from human Mu lesions on human embryonic lung fibroblasts.

Cytotoxicity after 48 h culture was assessed using an MTT assay. Negative control 1 is untreated cells, negative control 2 is ASL from uninfected skin. Positive control was purified mycolactone at a concentration of 5 µg/ml. Significant cytotoxicity was observed with all patient samples with ***p<0.001 compared to negative control 1. The apparent difference in percentage cytotoxicity between 5 untreated and 5 antibiotic treated lesions was not statistically significant. HELF cells were treated in quadruplicates and cytotoxicity determined in at least 2 independent experiments. Data are shown as a percentage of untreated cells (negative control 1). Error bars are ±SEM of duplicate assays.

Figure 5. Cytotoxicity of mycolactone on HELF cells.

A and B show the effect of ASL from Mu infected skin on HELF cells after 48 h with detachment of cells from culture plate, numerous apoptotic cells (orange nuclei stained with acridine orange) and few live spindle shaped cells (green nuclei with ethidium bromide). A and B represent ASL from a nodular lesion and an ulcerative lesion respectively. C shows the effect of purified mycolactone at a concentration of 5 µg/ml and D is a negative control demonstrating the effect of ASL from uninfected human skin. Pictures taken at x10 with a DM1 6000B Leica fluorescent microscope.

Figure 6. The effect of acetone soluble lipids from human Mu lesions on TNF-α release by J774 macrophages.

J774 macrophages were stimulated with 0.5 µg/ml of LPS. Negative control 1 is untreated J774 macrophages, negative control 2 is J774 treated with ASL from uninfected skin. Positive control refers to purified mycolactone at a concentration of 500 ng/ml and patient samples refers to all 10 patient lesions. ASL from infected lesions significantly inhibited TNF-α release compared to both negative controls with ***p<0.001. Error bars are ±SEM of duplicate assays. Although TNF-α release by J774 macrophages was significantly inhibited by purified mycolactone and lipid extracts from patient lesions, this occurred without significant cytotoxicity.

Figure 7. Comparison of TNF-α release by J774 macrophages treated with lipid extracts from 5 untreated and 5 antibiotic treated Mu infected human lesions.

TNF-α was inhibited significantly more by lipid extracts from untreated lesions compared to antibiotic treated lesions. ***p<0.05 compared to untreated lesions. Lipid extracts from all lesions had detectable mycolactone signal on mass spectroscopic analysis.