Mycobacterium ulcerans (Mul) is the causative agent of Buruli ulcer (BU), a debilitating skin-neglected tropical disease (NTD) that predominantly affects communities in rural areas with limited access to healthcare services. Despite its low mortality rate, BU can lead to significant morbidity due to the destruction of skin, soft tissues, and even bones. In the absence of early treatment, this may result in irreversible disfigurement and disability with high socio-economic burden and stigma. BU cases have been reported in 33 countries, with over 80% of global cases in the African Region, where nearly 50% of the affected individuals are children under the age of 15 (1). The current WHO-recommended antibiotic treatment involves an 8-week daily regimen of rifampicin (RIF) and clarithromycin (CLA). Treatment also includes extensive wound care, sometimes surgical intervention, and physiotherapy rehabilitation of movement limitations and disability, with lesion healing potentially lasting for several months. Treatment compliance can be difficult due to socioeconomic factors. A shortened and highly effective regimen would improve the care of BU patients and result in cost reduction. Consequently, the WHO requires evaluating promising medicines to provide new treatment options, including shortening the duration of treatment, as outlined in the WHO NTD Road Map 2021–2030 (1, 2).

In this context, using a drug repurposing approach, Arenaz et al. (3) investigated the in vitro effects of combining various β-lactam antibiotics with RIF and/or CLA. By utilizing checkerboard assays, they demonstrated a strong synergistic effect of amoxicillin/clavulanate (AMX/CLV) with RIF, evidenced by a substantial reduction in the Minimum Inhibitory Concentration (MIC) of RIF and CLA at a fixed time point; these promising findings were subsequently corroborated by in vitro time-kill assays (TKA) (4), which paved the way for the BLMs4BU clinical trial currently underway in several African countries (https://blms4bu.org/) (NCT05169554, PACTR202209521256638) (5). In addition, Saez et al. (4) conducted a comparative analysis of different in vitro readout measurements; this analysis demonstrated the value of a wide variety of biomarkers in BU research, correlating colony-forming units (CFUs) (the in vitro gold standard method for quantifying bacterial load) with optical density at 600 nm, luminescence production, quantification of the IS2404 DNA by quantitative PCR (qPCR), and the 16S rRNA burden by reverse transcriptase qPCR (4). This comparison was motivated by the arduous nature of CFU determination for Mul, which necessitates manipulation in biosafety level 3 laboratories and is hampered by the bacterium's slow growth rate on agar, often requiring from one to three months for visible colonies to appear. Furthermore, the study identified 16S rRNA burden quantification and luminescence measurements as having the strongest correlation between them, and a satisfactory correlation with CFU determination, with luminescence being the most cost-effective method. However, beyond conventional phenotypic metrics such as CFU quantification, there is an emerging consensus on the translational significance of evaluating the molecular metabolic state of bacteria as a proxy for the sterilizing activity of drug regimens. Sterilizing drugs, which can eradicate all viable bacterial organisms, are imperative for achieving effective shortening treatment regimens.

Ribosomal RNA (rRNA) constitutes a major component of the ribosome, comprising 80–85% of the total RNA present in cells. Its synthesis is subject to stringent regulation as it is a critical rate-limiting step for ribosome biogenesis and, consequently, protein synthesis and cell growth. The ribosomal RNA synthesis ratio (RS-ratio) has recently been described as a novel pharmacodynamic marker used to assess the effectiveness of drugs against Mycobacterium tuberculosis (Mtb) (6), which belongs to the same genus as Mul. The RS-ratio is a molecular assay that quantifies the abundance of a rapidly processed spacer region present in the precursor polycistronic rRNA (immature) relative to the stable final rRNA (mature) sequences. The ratio between these two types of rRNA (immature vs. mature) provides a measure of how actively the bacteria are synthesizing new ribosomes, which is a proxy for their physiological activity and replication rate. This provides different information than conventional markers, such as CFU or rRNA burden by 16S rRNA quantification.

The RS-ratio is highly sensitive to the effect of sterilizing drugs, being a promising biomarker for the evaluation of the sterilizing activity of new regimens. In fact, pronounced and sustained reduction in the RS-ratio correlates with improved treatment outcomes and reduced relapse rates in murine models of tuberculosis (TB) (6). However, despite its potential, the application of the RS-ratio as a tool for drug activity evaluation has primarily focused on Mtb, with limited exploration in other mycobacterial species, as evidenced by the lack of reports in the case of Mul.

In this study, we describe for the first time the in vitro development and application of the RS-ratio biomarker to Mul to evaluate the in vitro activity of RIF, CLA, and AMX/CLV and their combinations, currently under clinical investigation for BU treatment shortening. The results underpin the benefits of the RS-ratio as an in vitro biomarker and point to its relevance for future use at the clinical level.

Genetic material. RNA from the clinical Mul isolate ITM 000932 was obtained from previous extractions of TKA experiments described by Saez et al. (4). RNA extracts from the Mtb H37Ra and Mycobacterium smegmatis (Msm) mc²155 strains were used to test amplicon design specificities. Complementary DNA (cDNA) was synthesised using the SuperScript IV VILO Master Mix (Invitrogen) in a 10 µL reaction volume, using 5 µL of RNA as template. Reverse transcription conditions were 25°C for 10 minutes, followed by 50°C for 15 minutes.

Oligonucleotide Design for M. ulcerans rRNA Targets. Primers-probe sets for the quantification of the external transcribed spacer 1 (ETS1), internal transcribed spacer 1 (ITS1), and 23S rRNA of Mul were designed to target homologous regions to those used for RS-ratio calculations in Mtb (6) (Fig. 1A). These sequences were designed, ensuring species specificity, aligning Mul, Mtb and Msm rRNA operon regions using SnapGene v8.0.3 software. To ensure high specificity and prevent cross-amplification, at least one oligonucleotide within each primer and probe set was designed to not align with the rRNA operons from published sequences of Mtb (NCBI accession number NC_018143) and Msm (NCBI accession number NC_008596) (Fig. 1B). Specificity was confirmed using NCBI Nucleotide BLAST. Additionally, we used the primers-probe sequences for detecting the Mul 16S rRNA previously published (7), which already met the specificity criterion (Fig. 1BI). Fluorophore probes were selected to allow differential detection in the QX600 system (Bio-Rad) channels: Channel 1 for ETS1, Channel 2 for 23S rRNA, Channel 3 for 16S rRNA, and Channel 5 for ITS1. BioRad synthesized the ETS1, 23S, and 16S primers-probe sets, and the ITS1 primers-probe set was synthesized by CerTest S.L. (Table 1).

Droplet digital PCR and RS-ratio determination for M. ulcerans. Primers and probe sets for droplet digital PCR (ddPCR) quantification were used in a single multiplexed reaction at Bio-Rad's recommended concentrations (900 nM for primers and 250 nM for probes) with ddPCR SuperMix for Probes (no dUTP) (Bio-Rad). Thermocycling conditions included an initial denaturation step at 95°C for 10 minutes, followed by 40 cycles of denaturation at 96°C for 30 seconds and annealing/extension at 60°C for 90 seconds. A signal stabilization step was performed at 98°C for 10 minutes, followed by a final hold at 4°C for at least 30 minutes before droplet reading and quantification in the QX600 system. The temperature ramp rate was set at 2°C/s for all steps. The number of copies for each target was determined from the partitioned droplets using the QuantaSoft v2.1 software (Bio-Rad). The RS-ratio was calculated as the ratio of the copy number of the precursor rRNA (ETS1 or ITS1) to the copy number of the total rRNA (23S or 16S), multiplied by 10⁴, as previously described (6).

Validation of specificity and ddPCR optimization for M. ulcerans RS-ratio determination. To experimentally confirm their specificity, the newly synthesized oligonucleotide sets were subjected to a series of tests using a multiplexed ddPCR reaction with cDNA extracts from Mul, Mtb, and Msm. The Mul cDNA sample saturated the quantification when undiluted, resulting in 100% positive droplets for all four targets (ETS1, ITS1, 23S rRNA, and 16S rRNA). Crucially, no positive droplets were obtained from the cDNA samples of Mtb and Msm, thus confirming the high specificity of the designed probes and primers for Mul targets (Fig. 2A).

The RS-ratio was determined for the Mul ITM 000932 isolate longitudinally exposed to RIF, CLA, and AMX/CLV at 1/2X, 1X, and 1X MIC values, respectively, both in monotherapy and in two- and three-way combinations (4). To ensure accurate quantification and prevent saturation of the ddPCR reaction, each cDNA extract from the Mul TKA samples (3) was optimized for specific dilution, allowing for quantifiable droplet numbers for all four targets (ETS1, ITS1, 23S rRNA, and 16S rRNA) (Fig. 2B).

Evaluation of dynamic range and selection of the optimal precursor/total rRNA pair. Four distinct RS-ratios were compared employing combinations of a single precursor rRNA (ETS1 or ITS1) and total rRNA (23S or 16S) (Fig. 3). All pairs demonstrated comparable trends across groups. In the conditions where the RS-ratio barely changed over time, such as untreated control, CLA, and AMX/CLV, the different pairs tested (ETS1/23S, ITS1/23S, ETS1/16S, and ITS1/16S) mostly overlapped; however, different dynamic ranges were observed in the conditions in which drug combinations induced a reduction of the RS-ratio. Here, ITS1/16S and ITS1/23S pairs had the lower change, while ETS1/23S and ETS1/16S pairs showed higher change, having the strongest dynamic range the ETS1/23S pair. Thus, the ETS1/23S pair was selected for further comparisons with other biomarkers (Fig. 4).

RS-ratio kinetics reveal differential metabolic responses of M. ulcerans to RIF, CLA, and AMX/CLV combinations. The untreated Mul growth control exhibited an elevated RS-ratio (≈ 2000) (Fig. 4A, black circles), which underwent a slight decrease when the culture reached the stationary growth phase (≈ 900–1500 at day 7) (Fig. 4B, black circles). Exposure to RIF monotherapy resulted in a substantial reduction in the RS-ratio (≈ 200) after 7 days of incubation, with a slight recovery at relatively low levels (≈ 400) until the end of the 28-day assay (Fig. 4A, purple circles). Conversely, CLA monotherapy led to a marginal increasing effect (≈ 3000) on the RS-ratio (Fig. 4A, orange circles), while AMX/CLV did not significantly affect the RS-ratio (Fig. 4A, grey circles). The combination of CLA + AMX/CLV led to an increasing effect in the RS-ratio levels (≈ 4000) at day 3 (Fig. 4A, red circles) that dropped close to levels of the untreated growth control after day 7. On the contrary, RIF-containing combos consistently induced a decrease in the RS-ratio, though the intensity and kinetics of this effect appear to be dependent upon the accompanying drug. The RIF + CLA combo reduced the RS-ratio to levels comparable to those attained with RIF monotherapy (≈ 250) at days 7 and 10 (Fig. 4A, yellow circles). However, after day 14 of incubation, the RS-ratio of Mul treated with RIF + CLA showed a recovery to similar levels to the untreated control (≈ 700–900) that was stronger and faster than the slight increase observed in Mul treated with RIF monotherapy. In contrast, both the RIF + AMX/CLV combination and the triple combination (RIF + CLA + AMX/CLV) (Fig. 4A, green and blue circles, respectively) induced the fastest and most potent decreases of the RS-ratio of the conditions tested. A visible reduction (≈ 200) was already evident at day 3, reaching the maximal inhibition (≈ 50–60) between days 7 and 10. After day 14, the RS-ratio began to recover (≈ 400), but the levels of the untreated sample were not reached even at day 28. Importantly, this recovery was not identified by CFUs or luminescence biomarkers (Fig. 4B and C).

This study has investigated for the first time the implementation of the RS-ratio biomarker to assess the in vitro activity of drugs against Mul. To this end, four sequences of the rRNA operon of Mul were targeted using a primer-probe set previously described (7) and three newly designed primer-probe sets. The implementation of a single multiplexed ddPCR assay confirms the specificity of the different primer-probe sets. Although we used only two closely related mycobacteria (Mtb and Msm) as internal controls for specificity, our results support previous reports for the 16S rRNA Mul set (7). In silico alignment using NCBI BlastN predicted potential cross-amplification only with M. marinum due to high sequence identity in the rRNA operon region. Nevertheless, only lesions in early stages can appear similar, with clearly differential clinical signs and epidemiology. While our work is currently circumscribed to in vitro drug activity evaluations, testing against a panel of potential BU-relevant skin-coinfection species would be beneficial to provide a more comprehensive validation of the primers and probes in clinical samples.

Additionally, we tested various combinations of precursor-to-total rRNA ratios using a multiplexed ddPCR assay, and we observed that the ETS1/23S rRNA pair demonstrated the highest dynamic range (Fig. 3). The observed variations in the dynamic range of the distinct pairs in response to various treatments could be hypothesized to be a consequence of both the direct impact on precursor rRNA transcription and the effect on precursor rRNA maturation pathways. A thorough investigation into the underlying causes of these differences in dynamic range is beyond the scope of this study. However, these findings underscore the importance of empirically exploring various precursor/total rRNA pairs to identify the most suitable one before implementing the RS-ratio biomarker for a new organism.

A substantial reduction in the RS-ratio in Mul treated with RIF monotherapy was observed despite RIF being purposely added at half of its MIC for the corresponding strain (the study performed by Sáez et al. was designed to identify synergistic interactions among RIF, CLA, and AMX/CLV, not to test the full range of activity of the compounds alone). Nonetheless, this effect was not observed with CLA and AMX/CLV, which were also sub-dosed at their respective 1xMIC values. This observed behavior of the RS-ratio in response to different drugs could be explained by the mechanism of action of each drug. RIF, a cornerstone sterilizing drug in the treatment of Mtb and other mycobacterial infections, inhibits transcription by binding to the DNA-dependent RNA polymerase ß-subunit, encoded by the rpoB gene (8–10). In accordance with prior observations in Mtb (6, 11), RIF reduced the RS-ratio of Mul, likely by directly inhibiting the transcription of precursor-rRNAs. Although there are no extant reports on the effect of CLA or AMX/CLV on the RS-ratio of Mtb, our findings are also consistent with documented observed effects of other drugs with a similar mechanism of action in Mtb. The macrolide CLA inhibits protein translation of bacterial mRNAs by binding to the 23S rRNA in the 50S subunit of the ribosome (12). As previously documented, pharmaceutical agents that target translation by binding to the ribosome, such as aminoglycosides (streptomycin) or oxazolidinones (linezolid) (6, 11), cause a modest initial increase in Mtb RS-ratio, similar to our observations with CLA against Mul. This might represent a potential bacterial response to overcome the inhibition of ribosomal activity. AMX, a semisynthetic penicillin derivative from the ß-lactam family of antibiotics, exerts its antimicrobial effect by inhibiting bacterial cell wall synthesis, thereby preventing peptidoglycan cross-linking by transpeptidase proteins (13). However, its activity against mycobacteria is hindered due to the constitutive expression of BlaC (a ß-lactamase in mycobacteria), which is also present in Mul (14). Nevertheless, this natural resistance mechanism can be counteracted by the co-administration of CLV, a ß-lactamase inhibitor (3, 15). Our findings with AMX/CLV are consistent with previous reports showing that other cell wall synthesis inhibitors, such as isoniazid and ethambutol, have a minimal impact on reducing the RS-ratio of Mtb even at concentrations several times above their MIC, despite targeting different pathways in cell wall biosynthesis (6, 11).

Our data on the activity of the different combinations against Mul using the RS-ratio (Fig. 4A) support previous findings reported by Saez et al. (4). RIF + AMX/CLV-containing combinations induced the most pronounced reductions in RS-ratio and were also identified as the most potent regimens by other biomarkers (Fig. 4B and C). This observation is consistent with the notion that the RS-ratio functions as a more expeditious biomarker of sterilizing drug activity. As reported in Mtb (6), stronger reductions in the RS-ratio appear to correlate with enhanced sterilizing effects, a proxy for improved treatment outcomes even with shorter treatments. This finding is in accordance with the ongoing clinical trials aiming to shorten BU treatment from 8 weeks to 4 weeks by the inclusion of AMX/CLV to the WHO RIF + CLA recommended treatment (NCT05169554, PACTR202209521256638) (5, 16).

The experimental model employed for assessing the RS-ratio in Mul in this study has some limitations. In the context of the in vitro TKA samples used to implement the RS-ration in Mul (4), drugs were administered only at the onset of the experiment at sub-optimal concentrations, which implies that drug concentrations do not remain constant and stable throughout the duration of the experiment. In fact, this approach does not consider the dynamic drug exposure and varying half-lives of the compounds in the in vitro assay media, which are key considerations for drug effect in vivo. For instance, RIF has an in vitro half-life of approximately seven days, with about 75% degradation after 14 days of incubation in 7H9 medium at 37°C (17–19). A comparable or even shorter half-life (14.2 h) has been documented for CLA (17), while the half-life of AMX fluctuates extensively (from 27.4 h to 10 days) contingent on media conditions (20–23). Indeed, a slight recovery in the RS-ratio was observed after 10 days of incubation in the different RIF-containing treatments, including those containing AMX/CLV, even as bacterial burden levels continued to decline (Fig. 4B and C) (4). This phenomenon may be attributed to the initial sub-optimal concentration of RIF (1/2X MIC) and the degradation of the compound over time, thereby limiting the efficacy of the drug and enabling bacterial metabolic recovery. In a preceding time-kill assay study on Mtb, even concentrations of RIF 20X MIC were unable to prevent bacterial regrowth at endpoint due to RIF degradation in the assay medium (19).

A detailed comparison between conditions of RIF monotherapy and RIF-CLA combination reveals a faster recovery of the RS-ratio in the presence of CLA after day 10 (Fig. 2A). This is in line with previous observations showing that monotherapy CLA administration results in a modest increase in the RS-ratio. With a similar drug stability in the assay medium, the observed accelerated recovery is likely due to the differential initial dose and a negative impact of CLA on the activity of RIF, allowing for the overcoming of the RS-ratio reducing effect of RIF when its concentrations decline over time. This increase in the RS-ratio anticipates a potential antagonistic effect observed after day 21 with other biomarkers; for example, luminescence from the RIF-CLA combination is reduced more slowly than that from RIF monotherapy (Fig. 4C) (4).

The addition of AMX/CLV to RIF-containing regimens led to the most significant reductions in the RS-ratio (Fig. 4A) in line with observations with other biomarkers (Fig. 4B and C). In a previous study (19), the combination of RIF with beta-lactams and, in particular, first-generation cephalosporins demonstrated a capacity to prevent bacterial regrowth at endpoint, despite the substantial drug degradation of both compounds in the assay medium. At the time, no correlation was performed with the RS-ratio, but together with the present data, it provides the first insights into the sterilizing effect of RIF and beta-lactams (i.e., AMX/CLV) combinations against mycobacterial species. Although RS-ratio recovery patterns were observed in our analysis, the experimental design was only powered to identify synergistic interactions (i.e., drugs were evaluated at sub-optimal concentrations) and not to assess the full capacity of drugs alone or in combination, which would have required higher doses, such as in the above-mentioned study (19). Consequently, RS-ratio recovery patterns observed in this study suggest the existence of a residual bacterial subpopulation that survived the initial treatments. This sub-population could have potentially resumed metabolic activity once drug concentrations dropped below the effective levels, suggesting that culture regrowth might be expected if longer TKA times were provided. As such, the RS-ratio has the potential to function as an early monitoring biomarker, capable not only of ranking potential sterilizing drug combinations but also of detecting the early onset of metabolic bacterial reactivation and regrowth. Further investigations should be done to confirm whether the RS-ratio could also be used for detecting the presence of bacterial resistance or tolerance. Interestingly, even at sub-optimal concentrations, the combination of RIF + AMX/CLV was more effective than both drugs alone, supporting the synergistic interaction which correlates with the shortening potential of the novel regimen.

While the administration of a single dose of suboptimal concentrations of different drugs is useful for the understanding of in vitro synergistic drug interactions against Mul, it does not accurately represent in vivo and clinical exposures needed for translational efforts. Optimal drug concentrations will yield enhanced reductions of the RS-ratio in Mul, in accordance with documented evidence of RIF dose-response effects against Mtb, both in vivo (6) and in vitro (11). Therefore, there is compelling rationale for implementing the RS-ratio analysis in in vivo models, when necessary, and in more complex in vitro models, such as the Hollow Fibre System (HFS), which provides dynamic measures of pharmacokinetic and pharmacodynamic relationships of antimicrobial treatments against a given pathogen. The HFS was endorsed by the European Medicines Agency for the study of Mtb drug treatments (24), and it is currently under further optimization (25).

We have described for the first time the implementation of the RS-ratio biomarker for BU research, measuring bacterial metabolic activity and growth dynamics together with in vitro drug activity against Mul. These activities align with WHO-recommended key areas of development of rapid diagnostic tests for BU that can be used at the primary health care level (1). The RS-ratio offered novel insights into the impact of pharmaceutical agents on persistent bacterial activity, a capability that conventional gold standard methodologies of bacterial load quantification by CFU enumeration do not possess. In addition, research conducted with Mtb has demonstrated that a substantial decrease in the RS-ratio is associated with treatment-shortening properties of novel regimens, which our data support for the RIF-AMX/CLV-based combinations currently under evaluation in two clinical trials (https://blms4bu.org/) (NCT05169554, PACTR202209521256638). This finding suggests that the RS-ratio biomarker has the potential to inform the development of more efficacious and efficient regimens for the treatment of BU, TB, and even other infectious diseases. Furthermore, the results obtained clearly indicate the high potential of the RS-ratio as an in vitro biomarker, justifying further investigation into its prospective use as a novel biomarker in both animal models and clinical settings for the monitoring of BU treatment effectiveness.