Quantitative MODS-Wayne assay for rapid detection of pyrazinamide resistance in Mycobacterium tuberculosis from sputum samples

Abstract

Tuberculosis (TB) remains a significant global health challenge, exacerbated by the emergence of drug-resistant strains, such as those resistant to pyrazinamide (PZA). The current scarcity of affordable and precise quantitative diagnostic tests for PZA resistance underscores the urgent need for more accessible diagnostic tools. We evaluated PZA susceptibility in 264 TB-positive samples by quantifying pyrazinoic acid (POA) production, using both the MODS-Wayne qualitative assay and our newly developed quantitative approach (MODS-WQ). The MODS-WQ was assessed in 7H9 medium (MODS-WQ_7H9) or citrate buffer (MODS-WQ_CB), with POA levels measured via spectrophotometry against a calibration curve. PZA susceptibility determinations were based on a composite reference standard. Associations between POA levels and pyrazinamidase mutations were explored. The composite standard detected PZA resistance in 23.5% of the samples, which accounts for 62.8% of the multidrug-resistant (MDR) samples. The MODS-WQ established specific POA cutoffs of 123.25 µM for MODS-WQ_7H9 and 664.7 µM for MODS-WQ_CB, with sensitivities of 81.3% and 92.3% and specificities of 77.2% and 95.9%, respectively. Notably, samples with mutations in the pyrazinamidase metal-binding site exhibited significantly lower POA levels compared with mutations in the enzyme periphery. Furthermore, a significant correlation was found between POA production and PZA resistance, Bactec Growth Index, and minimum inhibitory concentration (MIC) values. This study presents a novel, direct, and accessible susceptibility test for PZA resistance that quantifies POA, enhancing the detection capabilities for this condition. The citrate-buffered MODS-WQ assay demonstrated high sensitivity and specificity for quantifying POA, confirming that POA production is a reliable indicator of PZA resistance.

Importance: PZA susceptibility testing continues to be a challenge, particularly in countries with high TB incidence. In response to this pressing need, we have developed a quantitative MODS-Wayne (MODS-WQ) assay. This approach offers a direct and cost-effective solution representing a significant advancement in TB diagnostics, particularly benefiting resource-limited laboratories, primarily in developing regions. The MODS-WQ assay stands out for its ability to quantify pyrazinoic acid (POA) production, as a reliable indicator of PZA resistance. Unlike traditional qualitative assays, MODS-WQ eliminates the inherent subjectivity in interpretation, providing more accurate and actionable results. Moreover, the MODS-WQ approach accounts for critical factors influencing PZA resistance, including enzymatic efficiency and efflux pump activity. By integrating these factors into the detection process, our methodology offers a comprehensive understanding of PZA resistance levels, enabling tailored treatment strategies for patients.

Keywords: MODS-Wayne; Tuberculosis; pyrazinamide; pyrazinoic acid; resistance.

Fig 1

Overview of study design and experimental approach. Only samples testing positive for acid-fast smear were selected for further analysis. These samples underwent evaluation through four distinct experimental assays: MODS-MDR for tuberculosis multidrug-resistant characterization, MODS-Wayne for qualitative assessment of pyrazinamide (PZA) resistance, MODS-WQ as the quantitative variant of the assay, and the composite reference standard derived from the conventional Wayne test, pncA gene sequencing, and the BACTEC MGIT960 PZA susceptibility testing. The diagnostic accuracy of both the MODS-Wayne and MODS-WQ assays was determined in comparison to the composite reference standard.

Fig 2

Evaluation of MODS-WQ variants' performance. (A) Standard curves for pyrazinoic acid (POA) in two distinct buffer matrices, based on three replicates and their standard deviations. The formula for calculating POA concentration (POAc) is also presented. (B) Violin plot illustrating the distribution of sample classifications according to the composite reference standard, alongside measured POA concentrations in μM for both MODS-WQ_7H9 and MODS-WQ_CB variants. Median POA concentrations for each group are indicated. (C) ROC curves and corresponding P-values for MODS-WQ_CB and MODS-WQ_7H9 variants highlight statistical differences between them. (D) Violin plot demonstrating sample classifications per MODS-WQ_7H9 and MODS-WQ_CB, including the established cutoff point for each. Statistical significance is denoted by asterisks, with * indicating P < 0.05, **P < 0.01, ***P < 0.001, and ****P < 0.0001.

Fig 3

Impact of pyrazinamidase mutations on pyrazinoic acid (POA) production. (A) 3-D structure of pyrazinamidase, highlighting its active sites in distinct colors for clarity: green for the PZA binding site, composed of amino acids Phe13 and Trp68 (A.1); red for the catalytic triad, consisting of Asp8, Lys96, and Cys138 (A.2); and purple for the metal-binding site (Fe+), involving Asp49, His51, and His57 (A.3). The enzyme core area adjacent to active sites and the periphery distant from these sites are also represented. (B) Violin graph presenting the distribution of POA concentrations across various mutation sites compared with wild-type (WT) samples, with the MODS-WQ_BC assay cutoff value (665 µM) delineated by a dashed line. Categories are labeled as EC: enzymatic core, N = 22; MBS: metal-binding site, N = 11; ACS: active catalytic site, N = 2; periphery, N = 4; and WT: wild type, N = 116. Statistical significance is indicated by asterisks: *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001, with “ns” denoting not significant, based on the Mann-Whitney test."

Fig 4

Complexity of PZA susceptibility phenotype. Variability in susceptibility phenotype is a result of different factors, such as the lack of specific PZA target, the large number of potential mutations throughout the coding and promoter sequence of pncA, the diversity of phenotypic methods, and errors that they entail, among others. In this figure, we can observe the four types of isolates according to the susceptibility phenotype.