Tuberculosis (TB) is an infectious disease caused by the Mycobacterium tuberculosis Complex (MtbC) that remains one of the 13 leading causes of death worldwide. After the COVID-19 pandemic, TB returned as the principal cause of death for a single infection agent, with 10.8 million new cases and 1.25 million deaths worldwide (Global Tuberculosis Report, 2024).

TB can be caused by drug-susceptible (DS) or drug-resistant (DR) strains. DR-TB treatment has the worst prognosis, which in general results in longer treatment durations, increased adverse effects for patients, and higher costs. The rise of rifampicin-resistant TB (RR-TB) is a global threat that has led the World Health Organization (WHO) to classify it as a priority pathogen (World Health Organization, 2024b). DR-TB can lead to Multidrug Resistance (MDR-TB), defined by WHO as Resistance to Rifampicin (RR) and isoniazid (H), or Extensive Drug Resistance (XDR-TB), defined as resistant to rifampicin (may also be resistant to isoniazid) plus a fluoroquinolone (levofloxacin or moxifloxacin), and either bedaquiline or linezolid. Worldwide incidence of MDR/RR-TB has decreased from an estimated 580,000 cases (UI 460,000 -580,000) in 2015 to 400,000 (UI 360,000 – 440,000) in 2023, mainly because of better diagnosis and prompt detection. However, the incidence of Pre-XDR and XDR TB has increased 3.8-fold in this same time period (Global Tuberculosis Report, 2016; Global Tuberculosis Report, 2024).

Rapid and accurate detection of resistant profiles is crucial for successful treatment, reducing transmission, as well as the risk of an increased proportion of resistant strains (Roberts et al., 2024). Overall, bacteria employ diverse mechanisms to resist antibiotic effects, including the enzymatic degradation of antibiotic molecules, acquisition of mutations within or around drug targets, reduced membrane permeability, and overexpression of efflux pumps. These resistance mechanisms leave distinctive genetic signatures that can be localized in specific genomic regions or distributed throughout the entire genome. Next-generation sequencing technologies, including amplicon sequencing and whole-genome sequencing (WGS), offer the potential to identify resistance markers and generate comprehensive antimicrobial resistance profiles (World Health Organization, 2024a; Pérez, 2023).

The identification and validation of DR markers are the result of extensive genomic studies coupled with microbiological tests that have been performed mainly in regions with high disease prevalence. These studies focus mainly on drug targets and drug metabolism, with less emphasis on efflux pumps (The CRyPTIC Consortium, 2022; World Health Organization, 2023). The importance of efflux pumps in antibiotic resistance is becoming more apparent in recent years (Li et al., 2024; Gaurav et al., 2023). They help maintain cellular homeostasis by expelling toxic molecules, which keeps antibiotic concentrations below therapeutic levels. Therefore, efflux pumps should be included in any genetic analysis of antibiotic resistance.

While these global studies encompass a few strains from Latin America, very few come from Mexico. Although the WHO considers Mexico a low-burden country for TB, it ranks among the top three countries with the most DR-TB in the Americas, together with Brazil and Peru (Tuberculosis en las Américas, 2021; Global Tuberculosis Report, 2024). DR-TB cases in Mexico have increased 281% in recent years (from 283 in 2015 to 796 reported in 2023; National Center for Preventive Programs and Disease Control, 2024), stressing the importance of antibiotic surveillance in Mexico.

The majority of genomic epidemiological studies from Mexican TB patients are centered on mutations for first-line drugs, including rifampicin, isoniazid, ethambutol, pyrazinamide, aminoglycosides, and fluoroquinolone, but do not explore mutations for drugs recently introduced in treatment regimens by WHO, such as bedaquiline, linezolid, pretomanid and delamanid, which are now a central part of TB treatment (Zenteno-Cuevas et al., 2009; Lopez-Alvarez et al., 2010; Flores-Treviño et al., 2015; Zenteno-Cuevas et al., 2014; Juarez-Eusebio et al., 2017; Madrazo-Moya et al., 2019; Zenteno-Cuevas et al., 2020; Mónica et al., 2021; Barbosa-Amezcua et al., 2022; Mejía-Ponce et al., 2023). In this study, we aim to characterize both known and novel mutations in target genes and efflux pump genes associated with resistance in Mexican patients with pulmonary-resistant tuberculosis using WGS. For strains with available phenotypic data, we analyze genotype–phenotype correlations. A solid knowledge of the genetic characteristics of resistance will help develop precise and maybe individual treatments to better help patients with DR-TB.

The increasing incidence of DR-TB has reduced the number of effective antibiotics, thus complicating the efforts to combat this persistent global health problem. Drug resistance phenotypes are acquired through the accumulation of mutations (single-nucleotide polymorphisms (SNPs), insertions, or deletions) in genes that code for drug targets or enzymes that activate drugs. This study presents WGS characterization of 49 resistant clinical isolates with at least first-line pDST collected from central and northern Mexico from 2014 to 2023. Among these isolates, 44.8% were classified as MDR-TB and 22.4% as Pre-XDR.

The rate of mutation and the effect of these variants on drug resistance are sometimes dependent on the genetic background of the Mtb strain (Ford et al., 2013; Nimmo et al., 2022). Therefore, it is crucial to characterize the resistance patterns in various regions, with their corresponding prevalent strains, to evaluate and design effective treatment strategies. Latin America has few studies examining resistance variants, although more interest has been given in recent years to our population (Zenteno-Cuevas et al., 2014; Madrazo-Moya et al., 2019; Jiménez-Ruano et al., 2021; Santos-Lazaro et al., 2021; Barbosa-Amezcua et al., 2022; D’Souza et al., 2023; Mejía-Ponce et al., 2023; Morey-León et al., 2023; Conceição et al., 2024; Puyén et al., 2024). It is important to emphasize that DR-TB reporting in Mexico has increased significantly over the last 4 years, after obvious underreporting due to health services being redirected toward COVID-19 pandemic response. DR-TB cases increased from 283 to 796 from 2015 to 2023, in contrast with global DR-TB trends that have remained relatively stable or decreased over the same period (National Center for Preventive Programs and Disease Control, 2023; Global Tuberculosis Report, 2024). Thus, a robust epidemiological surveillance system, based on local data, is crucial for accurate diagnosis and enhanced treatment, which could enable the identification of areas for improvement of health policies to prevent transmission.

The available evidence consistently shows lineage 4 as the dominant clade (>90% of the isolates) in Mexico. National surveys based on WGS and genotyping confirm the expansion of Haarlem (4.1.2.1), X-type (notably sub-lineage 4.1.1.3, “X3”), and 4.8 across several Mexican states. Published work (e.g., Jiménez-Ruano et al., 2021; Mejía-Ponce et al., 2023) reports that X3 is primarily restricted to Mexico and is disproportionately enriched for multidrug-resistant strains. Other L4 sub-lineages such as 4.1.2.1 and 4.8 have also been linked to elevated drug-resistance frequencies in regional cohorts (Molina-Torres et al., 2022; Alvarez-Maya et al., 2025). These findings indicate that, while L4 predominance likely reflects historical introduction and successful local expansion, certain Mexican sub-lineages, especially X3, carry a higher burden of resistance and warrant continued genomic surveillance. Surprisingly, 98% of our samples contain mutations classified as “not associated with resistance” in gyrA E21Q and 81% also carry S95T and G668D mutations, perhaps reflecting the frequent use of fluoroquinolones in the country.

The most frequent drug resistance-conferring mutations identified in our dataset have already been reported elsewhere (World Health Organization, 2023; The CRyPTIC Consortium, 2022), and include katG_S315T, rpoB_S540L, and embB_D354A. The similarity between previous reports in Mexico and WHO global data reflects common adaptative mechanisms against anti-TB drugs while maintaining fitness, suggesting that existing genotyping tests that include these mutations remain useful for drug-resistant prediction and disease management.

However, we did identify novel potential resistance-conferring mutations: 89 variants in 30 genes and promoters; they include polymorphisms in genes previously associated with resistance, including katG for isoniazid, embB for ethambutol, gyrA and gyrB for fluoroquinolones, and rrs for aminoglycosides, as well as others in genes so far not associated with resistance.

Among these, we found 18 variants on ahpC. The ahpC gene encodes alkyl hydroperoxide reductase C, which is part of the bacterial antioxidant defense system. The AhpC protein helps protect against oxidative damage produced by isoniazid, thus variants in ahpC can serve as compensatory changes that help bacteria survive the oxidative stress generated by isoniazid activity. Although none of the specific variants have been associated with resistance, changes in the expression or activity of AphC may increase resistance to isoniazid, which is why WHO considers them as secondary mutations and labeled as “candidate resistance genes” (World Health Organization, 2023).

Another interesting example is the variant found in the dprE2 gene. dprE2 is part of an operon that encodes dprE1-E2, which are essential for the synthesis of the arabinogalactan and lipoarabinomannan components of the bacterial cell wall. Recently (Abrahams et al., 2023), DprE2 was identified as the target of activated Pretomanid and Delamanid. Although not yet validated, it is identified as a potential antimycobacterial target of these drugs and thus susceptible to resistance. All these variants should be validated with phenotypic tests, and may provide new insight into resistance variants in the region.

Additionally, we found that 40.9% (9 out of 22) of rpoB_S450L mutations had compensatory mutations in rpoC, which has been reported to improve the fitness of an rpoB rifampicin-resistant strains and associated with resistance outbreaks (Comas et al., 2012; Conkle-Gutierrez et al., 2023). Consistent with these findings, four of these strains were associated with poor clinical outcomes, and two samples had an increased resistance profile. However our numbers are too small to do proper statistical analysis.

We found a high level of concordance between genotypic and phenotypic analysis in strains containing the most common resistant variants, and a lower concordance when all the samples are included. The discrepancy between genotypic and phenotypic analysis is not particular to our study (Mejía-Ponce et al., 2023; Hasan et al., 2024). Although phenotypic tests are still considered the “golden standard,” it is also subject to interpretation and execution. Critical concentration values have changed through the years; specifically, tests for some of our samples were conducted before 2021, with altered critical concentration values for rifampicin, isoniazid, and fluoroquinolones, or no microbiological test for certain drugs. Furthermore, more analysis is still needed to determine all the specific genes involved in resistance to newer drugs such as bedaquiline, linezolid, delamanid, etc., and whether they vary by region or previous treatments.

Additionally, we identified five strains with heteroresistance: three strains to fluoroquinolones, one to rifampicin, and one to a combination of rifampicin, isoniazid, and pyrazinamide. Although we do not have the therapeutic outcome for all these patients, the majority of them had poor clinical outcomes. Heteroresistance could reveal the inherent complexity of resistance evolution and possible epistatic effects and must be further analyzed.

On the other hand, efflux pumps are emerging as critical players in antibiotic resistance not only in tuberculosis but in all infectious diseases. Efflux pumps represent a critical intersection between essential cellular maintenance systems and barriers to antimicrobial therapy. The presence of multiple efflux pumps in diverse bacteria, including Gram positive and Gram negative genera and fungi, is associated with increased antibiotic resistance (Adams et al., 2021; Han et al., 2024; Leconte et al., 2024; Li et al., 2024). The vast majority of the variants on efflux pumps found in our sample population are on Mycobacterial membrane protein Large (MmpL) genes; 22 on mmpL8, 12 on mmpL3, followed by ABC transporter Rv1458c with 18 variants. MmpL proteins are essential for cell wall biosynthesis and lipid transport, which also transport antibiotic compounds out of the cell. MmpL8 is required for biosynthesis and transport of SL-1 sulfated glycolipid that is involved in host-pathogen interactions during early infections. MmpL3 is key for trehalose monomycolate transport, an indispensable component of the mycobacterial cell wall. MmpL3 is the target of an ethambutol analog, SQ109, which is now in phase 2 clinical trials (Chaitra et al., 2023).

Both by modifying the impermeability of the cell wall and exporting antibiotic compounds, efflux pumps can work synergistically with other resistance mechanisms to increase antimicrobial resistance. Furthermore, several studies have demonstrated that antibiotic treatment upregulates these integral membrane proteins, thereby maintaining sublethal intracellular antibiotic concentrations and selecting for antibiotic-resistant mutants (Hao et al., 2011; Laws et al., 2022; Remm et al., 2022). Thus, the administration of inhibitors may play an important role in enhancing the efficacy of antibiotics in the treatment of bacterial and fungal infections (El Meouche and Dunlop, 2018; Grimsey et al., 2020; de Melo Guedes et al., 2024; Han et al., 2024; Leconte et al., 2024; Li et al., 2024; Ye et al., 2024). In particular, the use of verapamil, chlorpromazine, reserpine, and other efflux pump inhibitors, as adjunct therapy in the treatment of TB and other infectious diseases is a promising strategy to combat antibiotic resistance (Rodrigues et al., 2020; Tran et al., 2020; Remm et al., 2022; Rao and Bhosale, 2024).

In this study, we found that 73% of the analyzed strains had efflux pump variants, the majority of which belonged to the ABC and RND superfamilies, previously associated with antibiotic resistance (Anthony Malinga and Stoltz, 2016; Laws et al., 2022; The CRyPTIC Consortium, 2022; Hasan et al., 2024). It is important to follow these patients through their treatment, as variations in efflux pumps may be the origin of resistance before fixation of mutations in drug targets (Laws et al., 2022). Future research on antibiotic resistance must include efflux pumps to gain a comprehensive understanding of resistance.

Our study has several limitations, including the limited number of samples and the fact that DNA extraction of all our samples was performed on the first subculture after Mtb isolation from the patient, which may have altered the resistant populations. Additionally, not all strains have complete first and second-line pDST. Future studies must include pDST for the new variants found; furthermore, we think that following patients with efflux pump mutations throughout their treatment can shed light into the connection between these proteins and the development of resistance.

In this study, we identified several of the most prevalent mutations conferring antibiotic resistance in patients with TB, as well as new variants that require validation in drug susceptibility tests to assess their relevance to resistance. This study substantially increases the national variant registry and, together with phenotypical drug sensitivity testing, provides valuable insight into the epidemiological landscape of the country, specifically for Pre-XDR and XDR strains. Additionally, we identified variants in efflux pumps, which may be part of the resistance mechanism that needs further investigation.

Furthermore, our findings confirm the need for timely and comprehensive drug susceptibility testing combined with WGS analysis before initiating treatment, which in Mexico, has unfortunately relied primarily on empirical approaches until now. As sequencing becomes more accessible, WGS should become standard practice in both diagnostic and follow-up protocols. This technology enables precise identification of all drug resistance mutations, which is particularly critical for Pre-XDR and XDR tuberculosis strains where treatment options are already limited. The comprehensive resistance profiling will allow clinicians to design personalized treatment regimens based on each patient’s specific resistance profile, moving away from empirical therapy. This precision is especially important for extensively drug-resistant cases, where selecting inappropriate drugs can significantly worsen patient outcomes. Additionally, WGS helps preserve the effectiveness of remaining active drugs—such as bedaquiline, delamanid, pretomanid and linezolid by ensuring they are used appropriately and in optimal combinations to prevent further resistance development.