ABSTRACT
Background:
In this study, we aimed to describe the mutations associated with first-line drug resistance in Mycobacterium tuberculosis complex (MTBC) isolates from São Paulo, Brazil, between 2019 and 2021.
Methods:
Mutations in the coding regions of rpoB and katG genes and in the promoter region of the inhA gene in MTBC clinical isolates were detected using the GenoType MTBDRplus assay (LPA). All mutations inferred by LPA were sequenced.
Results:
Of the 13,489 MTBC isolates with valid LPA results, 657 (4.9%) harbored mutations. The overall prevalence rates of rifampicin-resistant (RIF-R) tuberculosis (TB), isoniazid-resistant (INH-R) TB, and multidrug-resistant (MDR) TB were 1.5, 2.0, and 1.2%, respectively. A significant proportion of RIF-R isolates presented inferred rpoB mutations (89.1%), most of which were the borderline H445N mutation. The inhA promoter C-15T mutation was predominant among the INH-R isolates (52.8%). Most MDR isolates presented rpoB S450L + katG S315T1 mutations. Gene sequencing identified mutations not included in the catalogue of mutations published by the World Health Organization. Phenotypic drug susceptibility testing on isolates with inferred rpoB mutations revealed that the 0.5 µg/mL critical concentration of RIF failed to detect most borderline mutations when using the BACTEC MGIT 960 system.
Conclusions:
These findings emphasize the need for continuous surveillance and the integration of molecular and phenotypic methods to ensure an accurate detection and management of drug-resistant TB in high-burden settings.
Keywords: Drug susceptibility test, Gene sequencing, GenoType MTBDRplus , Multidrug-resistant tuberculosis, Mycobacterium tuberculosis
INTRODUCTION
Despite being a preventable and curable disease, in 2023, tuberculosis (TB) regained its position as the leading cause of death worldwide from a single infectious agent, after being temporarily surpassed by COVID-19 1 . In 2023, 175,923 individuals were diagnosed and treated for multidrug-resistant (MDR) or rifampicin-resistant (RIF-R) TB, representing only 44% of the estimated 400,000 cases of MDR/RIF-R-TB globally 1 .
Brazil is ranked among the 30 high-burden countries, which together accounted for 87% of the global cases of TB in 2023 1 , wherein Brazil reported over 80,000 new TB cases.In 2022, the country recorded approximately 5,800 TB-related deaths 2 . The state of São Paulo contributes significantly to this burden, accounting for approximately 25% of the national burden, with 19,571 new TB cases in 2023 2 .
Line probe assays (LPAs) are molecular tests performed directly on pulmonary samples or clinical isolates. The GenoType MTBDRplus assay (Bruker-Hain Lifescience GmbH, Nehren, Germany) helps detect Mycobacterium tuberculosis complex (MTBC) and associated resistance to isoniazid (INH) and RIF, i.e., MDR-TB.
There is limited information on the molecular mutations underlying drug resistance in TB in Brazil, with most published studies describing local distribution patterns 3 - 9 . In this study, we aimed to describe the mutations associated with RIF-R and INH-R in MTBC isolates and assess the prevalence of first-line drug-resistant TB in São Paulo.
METHODS
Ethics statement
This study was approved by the Technical Scientific Council (approval number: 45M/2020) and Research Ethics Committee (approval number: 6.110.210) of our institution.
Clinical isolates
The network of TB laboratories in São Paulo performs the Xpert MTB/RIF Ultra assay (Cepheid, Sunnyvale, CA, USA), culture, and smear microscopy when indicated. Mycobacterial cultures are then sent to our laboratory for species identification and drug susceptibility testing (DST). Most laboratories provide Xpert results on the requisition form accompanying the submitted cultures. In our laboratory, isolates are initially assessed based on their growth characteristics to presumptively determine whether they belong to MTBC or are non-tuberculous mycobacteria 10 , 11 . All presumptive MTBC isolates from patients meeting the DST criteria established in Brazil 12 , routinely tested using LPA between January 2019 and December 2021, were included in this study. Additional isolates from the same patient were included only when their resistance results differed.
GenoType MTBDRplus v. 2.0
DNA extraction and PCR were performed according to the instructions specified in the GenoType MTBDRplus assay kit manual 13 . The PCR products were hybridized with probes immobilized on strips using the GT-Blot 48 automated system (Bruker-Hain Lifescience GmbH, Germany). Results were interpreted using the GenoScan software (Bruker-Hain Lifescience) and released by a professional.
An inconclusive result was defined when the TUB (probe that identifies the MTBC) band was positive, but the wild-type (WT) bands were faint, and no mutation (MUT) bands developed owing to low bacterial load, which hindered drug resistance detection. When the results remained inconclusive after repeated testing, a new culture was requested.
A mutation was inferred in the absence of hybridization with both the WT and corresponding MUT probe(s). As LPA targets the 81-bp RIF-R-determining region (RRDR) of rpoB, isolates with inferred rpoB mutations were classified as RIF-R 14 . Similarly, isolates with mutations in katG and the inhA promoter were classified as INH-R, even when the mutation was not listed in the catalogue of mutations published by the World Health Organization (WHO) 15 .
An unexpected result occurred when two rpoB WT probes were expected to be positive, but only one was (e.g., WT3 positive, WT4 absent), which is not anticipated by the manufacturer. These cases were considered RIF-R when the clinical specimen was confirmed as RIF-R in the Xpert Ultra assay (Cepheid, USA). Otherwise, a cautionary note suggesting a silent mutation was included in the report. If the Xpert test result was unavailable, a phenotypic DST (pDST) was performed. Doubtful results were obtained when both the WT and corresponding MUT band(s) were faintly positive.
Gene sequencing
Sanger sequencing was performed to characterize the inferred mutations in the LPA gene targets and confirm doubtful results. For rpoB, the pair of primers amplified a 350-bp fragment covering the RRDR (positions 1,184-1,533 from the start codon). katG was sequenced using five pairs of primers from position −135 upstream of the katG start codon to +431 nucleotides downstream from the gene ending. The primers for the inhA promoter amplified a 248-bp fragment spanning positions −168 to +80 from the start codon of fabG1 7 .
The reaction mixes (25 µL) contained 1× buffer, 1.75 mM MgCl2 for rpoB (1.5 mM for katG/inhA), 0.75 U Taq polymerase (GoTaq G2 Flexi; Promega Corporation, Madison, WI, USA), 200 µM of each dNTP (dNTP mix; Promega, USA), 10 pmol primers for rpoB (5 pmol for katG/inhA), approximately 50 ng of DNA template, and nuclease-free water (Qiagen GmbH, Hilden, Germany). Amplification included initial denaturation at 94°C for 5 min, followed by 40 cycles of denaturation at 94°C for 1 min, annealing at 60°C for rpoB (62°C for katG; 64°C for inhA) for 1 min, and extension at 72°C for 1 min, with a final extension at 72°C for 10 min. Primers and unincorporated nucleotides were enzymatically removed (ExoSAP-it; Affymetrix, Inc., Santa Clara, CA, USA), and the amplicons were sequenced using the BigDye Terminator v3.1 (Applied Biosystems, Foster City, CA, USA). The sequence data generated on an ABI 3130×L Genetic Analyzer (Applied Biosystems, USA) were analyzed using BioEdit v7.2.5, MUBII-TB-DB, and BLAST tools 7 . The DNA sequences of all sequenced isolates were deposited in GenBank (https://www.ncbi.nlm.nih.gov) under the identification PRJNA888434.
pDST using BACTEC MGIT 960
pDST was performed to complement the LPA and Xpert results, when available, for interpreting RIF resistance. MTBC cultures were routinely tested for drug susceptibility using the BD BACTEC MGIT 960 IR Kit (Becton, Dickinson and Company, Sparks, MD, USA) 16 . INH was tested at the critical concentration of 0.1 µg/mL. From January 2019 to May 2021, RIF was tested at the critical concentration of 1.0 µg/mL; however, the critical concentration was later adjusted to 0.5 µg/mL for the June-December 2021 period, following WHO recommendations 17 .
In this study, we present only the pDST results for the isolates tested at the WHO-recommended RIF critical concentration of 0.5 µg/mL. Based on the WHO recommendations to report any rpoB mutation detected by LPA as RIF-R, not all isolates with inferred mutations in this study were subjected to pDST 18 .
Declaration of the use of generative AI and AI-assisted technologies in the writing process
During the preparation of this manuscript, the authors used ChatGPT (OpenAI) to assist in improving the clarity and readability of the text. All outputs from the tool were carefully reviewed, revised, and approved by the authors, who assume full responsibility for the content of the final published article.
RESULTS
Between January 2019 and December 2021, our laboratory received 23,696 cultured isolates: 8,710 (36.7%) in 2019, 7,526 (31.8%) in 2020, and 7,460 (31.5%) in 2021. Among these, 15,724 (66.4%) isolates were analyzed using LPA. Of these, 5,855 (37.2%) were received in 2019, 4,857 (30.9%) in 2020, and 5,012 (31.9%) in 2021. Among the tested isolates, 13,489 (85.8%) were identified as MTBC with valid resistance results, 411 (2.6%) were non-MTBC, 20 (0.1%) were inconclusive, and 1,804 (11.5%) were additional isolates from patients already included in the study. Of the 13,489 MTBC isolates, 657 (4.9%) harbored mutations, with 254 (38.7%) exhibiting 258 inferred mutations in any of the three genes. Fourteen of the 254 isolates were not sequenced because their DNA samples were unavailable, resulting in 240 sequenced isolates. The LPA and sequencing results for all inferred mutations are shown in Table 1.
TABLE 1: Mutations identified by DNA sequencing of the rpoB and katG coding regions and the inhA promoter region in 240 of 254 Mycobacterium tuberculosis complex isolates presenting 258 inferred mutations detected using the Genotype MTBDRplus assay.
| Genotype MTBDRplus Genes and mutation probes | Total N/No. | % | Mutations identified by gene sequencing* | Total N/No. | % |
|---|---|---|---|---|---|
| rpoB gene | 230/258 | 89.1 | 216/244 | 88.5 | |
| WT1 neg | 3/230 | 1.3 | T427T (1281 c>t) - silent | 3/216 | 1.4 |
| WT2 neg | 22/230 | 9.6 | L430P (1289 t>c)c | 21/216 | 9.7 |
| L430P (1289 t>c)c + F433dup (1297_1298ins ttc)a | 1/216 | 0.5 | |||
| WT2,3 neg | 4/230 | 1.7 | L430Lfs*10 (1290_1299del gagccaattc)e | 1/216 | 0.5 |
| Q432K (1294 c>a)a | 1/216 | 0.5 | |||
| Q432P (1295 a>c)a | 1/216 | 0.5 | |||
| Q432R (1295 a>g)e | 1/216 | 0.5 | |||
| WT2,3,4 neg | 1/230 | 0.4 | Q432Q (1296 a>g, silent) + F433_D435del (1297_1305del ttcatggac)b | 1/216 | 0.5 |
| WT3,4 neg, WT7+MUT2A pos (H445Y HR) | 2/230 | 0.9 | Q432H (1296 a>c) + F433_D435del (1297_1305del ttcatggac)e | 2/216 | 0.9 |
| WT3 neg | 30/230 | 13.0 | F433F (1299 c>t) - silent | 23/216 | 10.6 |
| F433dup (1297_1298ins ttc)a | 1/216 | 0.5 | |||
| F433dup (1297_1298ins ttc)a + P454S (1360 c>t)d | 2/216 | 0.9 | |||
| N/D | (4) | - | |||
| WT3,4 neg | 4/230 | 1.7 | M434Qfs (1300_1305del atggac)e | 1/216 | 0.5 |
| D435Y (1303 g>t)c | 3/216 | 1.4 | |||
| WT3,4,7 neg | 2/230 | 0.9 | M434I (1302 g>a)b + H445N (1333 c>a)c | 2/216 | 0.9 |
| WT5 neg‡ | 9/230 | 3.9 | P439L (1316 c>t)b | 7/216 | 3.2 |
| N/D | (2) | - | |||
| WT6 neg | 1/230 | 0.4 | S441S (1323 g>a) - silent | 1/216 | 0.5 |
| WT6 neg, WT3,4+MUT1 weak (D435Y HR) | 1/230 | 0.4 | L443S (1328 t>c)b | 1/216 | 0.5 |
| WT7 neg§ | 130/230 | 56.5 | T444T (1332 c>g, silent) + H445S (1333 ca>tc)c + K446Q (1336ª>c)b | 1/216 | 0.5 |
| H445C (1333 ca>tg)a | 2/216 | 0.9 | |||
| H445L (1334 a>t)c | 3/216 | 1.4 | |||
| H445N (1333 c>a)c | 115/216 | 53.2 | |||
| H445R (1334 a>g)a | 1/216 | 0.5 | |||
| H445S (1333 ca>ag)c | 1/216 | 0.5 | |||
| N/D | (7) | - | |||
| WT8 weak, MUT2B pos (H445D)a | 1/230 | 0.4 | H445D (1333 c>g)a | 1/216 | 0.5 |
| WT8 neg† | 20/230 | 8.7 | S450L (1349 c>t)a | 5/216 | 2.3 |
| S450W (1349 c>g)a | 6/216 | 2.8 | |||
| Q409R (1226 a>g)d + S450W (1349 c>g)a | 1/216 | 0.5 | |||
| L452P (1355 t>c)a | 7/216 | 3.2 | |||
| N/D | (1) | - | |||
| katG gene | 17/258 | 6.6 | 17/244 | 7.0 | |
| WT neg | 14/17 | 82.4 | S315G (943 a>g)d | 3/17 | 17.6 |
| S315I (944 g>t)b + A379V (1136 c>t)d | 1/17 | 5.9 | |||
| S315N (944 g>a)a | 4/17 | 23.5 | |||
| S315N (944 g>a)a + A379T (1135 g>a)d | 5/17 | 29.4 | |||
| S315T (944 gc>cg)a | 1/17 | 5.9 | |||
| Loccus, WT, MUT1, MUT2 neg | 3/17 | 17.6 | katG del (partial matching with reference)e | 3/17 | 17.6 |
| inhA promoter region | 11/258 | 4.3 | 11/244 | 4.5 | |
| WT1 neg | 4/11 | 36.7 | g-17tb | 2/11 | 18.2 |
| g-19td | 1/11 | 9.1 | |||
| c-20ge | 1/11 | 9.1 | |||
| WT2 neg | 7/11 | 63.3 | t-8gb | 6/11 | 54.5 |
| WT | 1/11 | 9.1 |
WT: wild type; neg: negative, i.e: lack of binding to a probe; pos: positive signal; HR: heteroresistance; N/D: not done; underlined are mutations for which MTBDRplus has specific probes; in bold are mutations outside the rifampicin-resistance determining region; ‡ an isolate P439L had additional weak signals at wt7-mut2a bands; § three H445N HR results, an isolate H445N had additional wt3,4 weak and another one, wt4,8 weak bands; † one S450L sample had additional wt3,4-mut1 and wt7-mut2a weak bands; one S450W was HR. * Mutation classification based on the catalogue of mutations published by the World Health Organization (2023): aassociated with resistance, bassociated with resistance - interim, cassociated with resistance (borderline), duncertain significance, enot on catalogue.
Most inferred mutations were detected in rpoB (89.1%, 230/258). In the LPA, the most frequently negative rpoB probes were WT7 (56.5%; 130/230), WT3 (13.0%; 30/230), and WT2 (9.6%; 22/230) (Table 1). Of the 244 sequenced targets, 216 (88.5%) were rpoB. The most common rpoB mutation was H445N (missing the WT7 band), accounting for 53.2% (115/216) of the inferred rpoB mutations by LPA and 47.1% (115/244) of all these mutations. Four mutations were not listed in the WHO catalogue, including indels (L430Lfs*10, Q432Hfs, and M434Qfs) and point mutations (Q432R). In addition to F443F (WT3-negative), the S441S silent mutation was an unexpected LPA result (WT6-negative). Conversely, the silent mutation T427T was expected (WT1-negative).
Six mutation types were identified in 6.6% isolates presenting inferred katG mutations (17/258). Most mutations (14/17, 82.4%) occurred at codon 315. Mutations in three (17.6%) isolates were likely due to large genomic deletions, as their katG sequences only partially matched the reference genome: in the first isolate, a fragment covering nucleotides 408-2,075 was identified; the second isolate yielded a sequenced fragment covering nucleotides 1,404-1,707; whereas the third isolate yielded no fragment of the katG gene (Table 1).
In LPA, a WT2-negative probe (7/11; 63.3%) was the most common among the inferred mutations in the inhA promoter, accounting for 4.3% (11/258) of these mutations. Among the four changes identified, C-20G (WT1-negative) was not listed in the WHO catalogue (Table 1).
The mutation patterns detected in the rpoB andkatG genes and the inhA promoter in the 657 isolates are presented in Table 2. According to the LPA, 234 (35.6%) isolates were classified as RIF-R, with the majority (195/234; 83.3%) exhibiting inferred mutations. Of the 14 isolates that were not sequenced (13 RIF-R and 1 MDR), 7,4, and 2 isolates exhibited WT7-, WT3-, and WT5-negative results, respectively, and one isolate exhibited WT8-negative results. Among the 222 RIF-R isolates with identified mutations, the most common were H445N (51.4%), F433F, and S450L (10.4% each) (Table 2). Twenty-seven (12.2%) isolates harbored single silent mutations.
TABLE 2: Pattern of mutations in the rpoB and katG coding regions and the inhA promoter regions in Mycobacterium tuberculosis complex isolates, São Paulo, Brazil, between 2019 and 2021.
| Resistance detected using LPA | Mutated gene | N | N HR | Mutation profile of the isolates | Discordances in resistance after | ||
|---|---|---|---|---|---|---|---|
| rpoB* | katG* | inhA promoter* | sequencing | ||||
| INH-R (n = 265) | katG (n = 116) | 1 | - | WT | S315Ib + A379Vd | WT | |
| 3 | - | WT | S315Gd | WT | |||
| 1 | - | WT | S315Na | WT | |||
| 106 | 6 | WT | S315T1 a | WT | |||
| 2 | - | WT | S315T2 a | WT | |||
| 1 | - | WT | S315T3a | WT | |||
| 2 | - | WT | katG dele | WT | |||
| inhA promoter (n = 148) | 1 | - | WT | WT | t-8c a | ||
| 5 | - | WT | WT | t-8gb | |||
| 139 | 13 | WT | WT | c-15t a | False-WT by sequencing (n = 1, HR) | ||
| 1 | 1 | WT | WT | a-16g a | |||
| 1 | - | WT | WT | g-19td | |||
| 1 | - | WT | WT | c-20g | |||
| katG + inhA promoter (n = 1) | 1 | - | WT | S315T1 a | c-15t a | ||
| RIF-R (n = 234) | 1 | - | Q409R d + S450Wa | WT | WT | - | |
| 3 | - | T427T | WT | WT | False-RIF-R by LPA (n=3) | ||
| 1 | - | L430Lfs*10e | WT | WT | - | ||
| 19 | - | L430Pc | WT | WT | - | ||
| 1 | - | L430Pc + F433dupa | WT | WT | - | ||
| 1 | - | Q432H fse | WT | WT | - | ||
| 1 | - | Q432Ka | WT | WT | - | ||
| 1 | - | Q432Q + F433_D435delb | WT | WT | - | ||
| 1 | - | Q432Re | WT | WT | - | ||
| 23 | - | F433F | WT | WT | False-RIF-R by LPA (n=23) | ||
| 1 | - | F433dupa | WT | WT | - | ||
| 1 | - | M434Q fse | WT | WT | - | ||
| 2 | 1 | D435V | WT | WT | - | ||
| 2 | - | D435Yc | WT | WT | - | ||
| 7 | - | P439Lb | WT | WT | - | ||
| 1 | - | S441S | WT | WT | False-RIF-R by LPA (n=1) | ||
| 8 | 1 | H445D a | WT | WT | - | ||
| 1 | 1 | H445D a + H445Y | WT | WT | - | ||
| 114 | 3 | H445Nc | WT | WT | - | ||
| 5 | - | H445Y | WT | WT | - | ||
| 23 | 4 | S450L a | WT | WT | - | ||
| 1 | - | S450Wa | WT | WT | - | ||
| 3 | - | L452Pa | WT | WT | - | ||
| 4 | - | WT3 neg | WT | WT | N/S | ||
| 2 | - | WT5 neg | WT | WT | N/S | ||
| 7 | WT7 neg | WT | WT | N/S | |||
| MDR (n = 158) | rpoB + katG (n = 115) | 1 | - | Q432H fse | S315T1 a | WT | - |
| 1 | - | Q432Pa | S315T1 a | WT | - | ||
| 2 | - | F433dupa + P454S d | S315T1 a | WT | - | ||
| 5 | - | D435V | S315T1 a | WT | - | ||
| 1 | - | D435Y | S315T1 a | WT | - | ||
| 1 | - | K443Sb | S315T1 a | WT | - | ||
| 1 | - | T444T + H445Sc + K446Qb | S315T1 a | WT | - | ||
| 2 | - | H445Ca | S315T1 a | WT | - | ||
| 4 | - | H445D a | S315T1 a | WT | - | ||
| 2 | - | H445D a + H445Y | S315T1 a | WT | - | ||
| 2 | - | H445Lc | S315T1 a | WT | - | ||
| 1 | - | H445Ra | S315T1 a | WT | - | ||
| 1 | - | H445Sc | S315T1 a | WT | - | ||
| 5 | 2 | H445Y | S315T1 a | WT | - | ||
| 3 | - | S450L a | S315Na | WT | - | ||
| 74 | 2 | S450L a | S315T1 a | WT | - | ||
| 1 | - | S450L a | katG dele | WT | - | ||
| 4 | - | S450Wa | S315T1 a | WT | - | ||
| 3 | - | L452Pa | S315T1 a | WT | - | ||
| 1 | - | WT8 neg | S315T1 a | WT | N/S | ||
| rpoB + inhA promoter (n = 30) | 2 | - | L430Pc | WT | c-15ta | - | |
| 2 | 1 | H445Y | WT | c-15ta | - | ||
| 1 | - | H445Da | WT | WT2 neg | False-WT by sequencing (n = 1, HR) | ||
| 1 | 1 | H445Y + S450L a | WT | c-15t a | - | ||
| 22 | 3 | S450L a | WT | c-15t a | - | ||
| 1 | 1 | S450Wa | WT | t-8gb | - | ||
| 1 | - | L452Pa | WT | c-15t a | - | ||
| rpoB + katG + inhA promoter (n = 13) | 1 | - | H445D a | S315T1 a | g-17tb | - | |
| 2 | - | H445Y | S315T1 a | c-15t a | - | ||
| 1 | - | H445Lc | S315T1 a | g-17tb | - | ||
| 1 | - | H445Nc | S315T1 a | c-15t a | - | ||
| 2 | - | M434Ib + H445Nc | S315T1 a | c-15t a | - | ||
| 5 | - | S450L a | S315Na + A379Td | c-15t a | - | ||
| 1 | - | S450L a | S315T1 a | c-15t a | - | ||
| No resistance | - | 12832 | - | WT | WT | WT | - |
LPA: Genotype MTBDRplus line probe assay; INH: isoniazid; RIF: rifampicin; HR: heteroresistance; N/S: not sequenced; HR: heteroresistance; WT: wild type; neg: negative (lack of binding to a probe); underlined are mutations for which MTBDRplus has specific probes; in bold are mutations outside the rifampicin-resistance determining region. * Mutation classification based on the catalogue of mutations published by the World Health Organization (2023): aassociated with resistance, bassociated with resistance - interim, cassociated with resistance (borderline), duncertain significance, enot on catalogue.
Regarding the mutations in 265 (40.3%) INH-R isolates, 148 (55.8%) and 116 (43.8%) isolates exhibited mutations in the inhA promoter and katG, respectively; however, one (0.4%) isolate exhibited mutations in both genes. The most common were the inhA promoter C-15T (52.8%) and katG S315T (g>c) (40.4%) substitutions (Table 2).
Among the 158 (24.1%) MDR isolates detected by LPA, most mutations (115/158; 72.8%) occurred in rpoB and katG, with the most common rpoB_S450L/katG_S315T being identified in 74 (46.8%) MDR isolates. Inferred katG mutations were detected in four isolates with the rpoB S450L mutation, whereas inferred mutations in rpoB were identified in 22 isolates with the katG S315T mutation. Among the 30 (19.0%) MDR isolates with mutations in both rpoB and inhA promoter detected by LPA, the most common was rpoB_S450L/inhA_C-15T, identified in 22 (13.9%) isolates, including three heteroresistant rpoB cases (S450L+WT). Four inferred mutations were detected in rpoB and one in the inhA promoter. Additionally, 13 MDR isolates (8.2%) exhibited mutations in all three target genes. The most frequent profile in this group was rpoB_S450L/katG_S315N+A379T/inhA_C-15T, observed in five isolates (Table 2).
Certain mutations exhibited markedly different distributions between RIF-R and MDR isolates. Notably, the H445N mutation was detected in 48.5% (114/235) of RIF-R isolates but only 1.9% (3/158) of MDR isolates (p<.0001), whereas the S450L mutation occurred in 67.7% (107/158) of MDR isolates and 9.8% (23/235) of RIF-R isolates (p<.0001). Silent mutations occurring as single mutations were observed exclusively in RIF-R isolates (Table 2).
As shown in Table 2, the proportion of heteroresistance detected by LPA was low in our study, being found in 10/234 (4.3%) RIF-R, 20/265 (7.5%) INH-R, and 10/158 (6.3%) MDR isolates. The sequencing of two doubtful results revealed a mix of WT and S315T mutation in an isolate with katG WT-MUT weak positive bands in LPA, but missed the C-15T mutation in a heteroresistant culture with WT1-MUT1 weak positive bands in LPA.
Assuming that all mutations found in this study were associated with RIF-R or INH-R, excluding silent mutations, the drug resistance prevalence in the 3-year study period was 1.5% (207/13,489) for RIF-R-TB, 2.0% (265/13,489) for INH-R-TB, and 1.2% (157/13,489) for MDR-TB. RIF-R-TB prevalence remained steady at 1.5% (76/5,174) in 2019, 1.5% (61/4,060) in 2020, and 1.6% (70/4,255) in 2021. INH-R-TB showed slight fluctuations and was 2.0% (105/5,174) in 2019, 1.7% (68/4,060) in 2020, and 2.2% (92/4,225) in 2021. MDR-TB prevalence remained stable at 1.2% (62/5,174) in 2019, 1.1% (44/4,060) in 2020, and 1.2% (52/4,255) in 2021.
Overall, 123 among the 230 isolates with inferred rpoB mutations (53.5%) underwent RIF testing using MGIT, of which 111 showed valid results: 76 (68.5%) and 35 (31.5%) at 1.0 and 0.5 µg/mL, respectively. Only the isolates tested at RIF 0.5 µg/mL are shown in Table 3, which describes pDST results for 39 isolates with inferred mutations in LPA (35 in rpoB, 1 in katG, and 3 in inhA). Among the 35 isolates presenting inferred rpoB mutations, 21 were borderline mutations (19 H445N and 2 L430P), and all tested RIF susceptible on MGIT. Of the two isolates with rpoB mutations associated with interim resistance the isolate with the Q432R mutation tested MDR on MGIT (notably, this isolate was INH susceptible on LPA), and the isolate with the P439L mutation was RIF susceptible on MGIT. The two isolates with mutations associated with RIF-R according to the WHO catalogue (rpoB Q432K - this isolate did not show mutations in either the katG coding region or the inhA promoter; and H445L - this isolate also had the inhA promoter G-17T mutation) were phenotypically MDR.
TABLE 3: Results for the phenotypic drug susceptibility testing (DST) using the BACTEC MGIT 960 system of Mycobacterium tuberculosis clinical isolates with mutations inferred in the GenoType MTBDRplus line probe assay (LPA).
| LPA results (rpoB/katG/inhA profile) | Gene sequencing results | WHO mutations catalogue (2023) | N | Phenotypic DST results |
|---|---|---|---|---|
| RIF R, INH S (WT2 neg/WT/WT) | rpoB L430P | Assoc. w R (borderline) | 2 | RIF S, INH S |
| RIF R, INH S (WT2,3 neg/WT/WT) | rpoB Q432K | Assoc. w R | 1 | MDR |
| rpoB Q432R | Assoc. w R - interim | 1 | MDR | |
| RIF R, INH S (WT3 neg/WT/WT) | rpoB F433F (silent) | Not assoc. w R | 8 | RIF S, INH S |
| RIF R, INH S (WT5 neg/WT/WT) | rpoB P439L | Assoc. w R - interim | 1 | RIF S, INH S |
| RIF R, INH S (WT7 neg/WT/WT) | rpoB H445N | Assoc. w R (borderline) | 18 | RIF S, INH S |
| MDR (WT6 neg, WT3,4+MUT1 faint /S315T1/WT) | rpoB L443S | Assoc. w R - interim | 1 | RIF S, INH R |
| MDR (WT7 neg/S315T1/WT) | rpoB H445R | Assoc. w R | 1 | MDR |
| MDR (WT7 neg/S315T1/WT1 neg) | rpoB H445L, inhA G-17T | Assoc. w R (rpoB), uncertain significance (inhA) | 1 | MDR |
| MDR (WT8 neg/S315T1/WT) | rpoB S450W | Assoc. w R | 1 | MDR |
| INH R (WT/Locus, WT neg/WT) | katG does not align to reference | Not on catalogue* | 1 | INH R |
| INH R (WT/WT/WT1 neg) | inhA G-19T | Uncertain significance | 1 | RIF S, INH S |
| inhA C-20G | Not on catalogue | 1 | INH R | |
| INH R (WT/WT/WT2 neg) | inhA T-8G | Assoc. w R - interim | 1 | RIF S, INH S |
| Total | - | - | 39 | - |
WHO: World Health Organization; INH: isoniazid; RIF: rifampicin; neg: negative (absence of the respective band); WT: wild type; S: susceptible; R: resistant; Assoc. w R: mutation associated with resistance; Not assoc. w R: mutation not associated with resistance; underlined are inferred mutations; *probably a loss-of-function mutation due to gene deletion
One isolate with an inferred katG mutation (locus and WT bands absent) was INH-R, as confirmed by pDST, and harbored a katG deletion, likely a loss-of-function mutation associated with resistance. Three isolates exhibited inferred mutations in the inhA promoter (C-20G, G-19T, and T-8G), and only the isolate harboring the C-20G mutation (not in the catalogue) tested INH-R using MGIT (Table 3). The analysis of rpoB mutations in 392 isolates (230 with inferred and 162 with canonical mutations) across the state of São Paulo revealed that inferred mutations were more frequent in Baixada Santista (60.7%) and the São Paulo Metropolitan Region (66.2%), whereas canonical mutations predominated in the countryside (Figure 1).
FIGURE 1: Distribution of canonical (orange) and inferred (blue) rpoB mutations in Mycobacterium tuberculosis isolates across the state of São Paulo, Brazil. In Baixada Santista (n=56) and the Metropolitan region (n=234), inferred mutations were more frequent, representing 60.7% (n=34) and 66.2% (n=155) of cases, respectively, while canonical mutations accounted for 39.3% (n=22) and 33.8% (n=79). In contrast, in the Countryside (n=102), canonical mutations predominated (59.8%, n=61) compared to inferred mutations (40.2%, n=41). These data reveal geographic heterogeneity in mutation profiles, with inferred mutations more common in urbanized, high-incidence areas and canonical mutations more prevalent in less densely populated regions.
DISCUSSION
This study provides a comprehensive molecular characterization of RIF-R, INH-R, and MDR MTBC clinical isolates from São Paulo over a 3-year period, overlapping the COVID-19 pandemic. In total, 13,635 isolates were analyzed using LPA, revealing valuable insights into the prevalence of drug resistance. In 2019, the number of mycobacterial isolates analyzed was higher than in 2020 and 2021 owing to the impact of the COVID-19 pandemic on TB diagnosis. Pontes et al. (2024) reported a significant decline in TB notifications in São Paulo, highlighting the disruptive effects of the pandemic on healthcare services 19 .
The overall prevalences of RIF-R-TB, INH-R-TB, and MDR-TB were 1.5, 2.0, and 1.2%, respectively. Although RIF-R and MDR prevalences remained stable, that of INH-R-TB fluctuated slightly, with a decrease in 2020, followed by an increase in 2021. Brazil is considered a high-burden country for TB and TB/HIV co-infection; however, drug resistance rates remain relatively low, a trend also reflected in the resistance rates observed in São Paulo.
Mutational analysis revealed that a significant proportion (89.1%) of RIF-R isolates (excluding 27 isolates with silent mutations) harbored inferred mutations, most of them identified as the borderline H445N. The L430P borderline mutation was also frequent. The high occurrence of borderline rpoB mutations observed in our setting has been noted since the implementation of Xpert MTB/RIF in 2014, when these mutations were linked to discrepancies between molecular and pDST results 4 , 7 . The frequency of the S450L mutation, identified in 9.6% of RIF-R cases in this study, aligns with previous findings 7 , 20 .
Discordant results for RIF, characterized by genotypically resistant but phenotypically susceptible M. tuberculosis strains, have frequently been reported and present significant challenges for treatment and clinical management. These discordances are often caused by borderline mutations in rpoB. Studies have demonstrated that such mutations can be clinically relevant and may significantly affect treatment outcomes in patients with RIF-R or MDR-TB 21 , 22 . Therefore, detecting borderline mutations is as critical as identifying the mutations associated with high-level resistance.
A previous study by our group reported that the prevalence of MTBC isolates with genotypic resistance to RIF but discordant phenotypic susceptibility was 55.1% in São Paulo between 2014 and 2017. Among them, 75.5% of the sequenced and restriction fragment length polymorphism-genotyped isolates were grouped into molecular clusters, mostly associated with rpoB mutations. The most prevalent mutation was the borderline H445N, which was frequently found in isolates circulating in Baixada Santista and the São Paulo Metropolitan Region 4 . Although molecular typing of the isolates was not performed in the present study, the inferred mutations were observed more frequently in these two regions.
A study conducted in Namibia using whole-genome sequencing (WGS) identified the L430P borderline mutation in clustered isolates, suggesting a high transmission potential for these strains 23 . Lempens et al. 21 , also using WGS, concluded that isolates harboring borderline mutations are transmitted at rates comparable to those of isolates with high-level-resistance-associated rpoB mutations.
A high frequency of the silent F433F mutation was also observed, interpreted as RIF susceptible in Xpert Ultra, but as RIF-R in LPA, which detected this mutation as inferred. Isolates with synonymous rpoB mutations not hybridizing to the WT1, WT3, and WT6 probes highlighted the need to identify mutations with these LPA patterns to avoid false RIF-R classification.
Among INH-R isolates, inhA promoter mutations were more prevalent (55.8%) than katG mutations (43.8%). The most common mutation was inhA C-15T, detected in 52.8% of INH-R cases. Previous studies have also reported a higher prevalence of inhA mutations in INH-R than in MDR isolates 7 , 9 , 24 . The frequency of inferred mutations in INH-R isolates (5.6%) was significantly lower than in RIF-R isolates. This finding is consistent with those of other studies 7 , 20 . Furthermore, in our study, the most frequent mutations found in MDR isolates were rpoB S450L/katG S315T. Both mutations are associated with MDR-TB 3 , 7 , 24 .
The higher prevalence of the C-15T inhA mutation relative to the katG S315T mutation is probably explained by the greater number of INH-R than of MDR cases in our study population. A previous investigation revealed a higher prevalence of molecular clustering among isolates carrying inhA promoter mutations than among those with katG mutations in São Paulo, indicating the sustained transmission of these strains in the community 9 . Hazbón et al. 24 reported that katG S315 mutations increase in frequency as M. tuberculosis isolates progress from INH-R to MDR. This pattern may reflect two potential mechanisms: katG mutations arise as secondary events following initial INH resistance, or they confer a fitness advantage facilitating survival, transmission, and evolution into MDR strains. Supporting the latter, inhA promoter mutations were less frequent in the highly resistant clustered isolates, possibly because of the fitness cost associated with inhA overexpression.
We identified three isolates exhibiting no katG locus control and WT bands in LPA. Subsequent sequencing revealed that these isolates harbored distinct deletions in the katG gene. Although rare, such deletions have been described in clinical settings. A complete deletion of katG was previously reported in a clinical isolate from an HIV-infected patient in Italy, conferring high-level resistance to isoniazid (MIC >25.6 µg/mL) without significant impairment in bacterial fitness 25 . Another study has detected small deletions in the 5′ region of katG, including a 12-bp deletion in isolates from multiple geographic regions 26 . A study conducted in Brazil reported frequent deletions in katG in MDR isolates, supporting the clinical relevance of such structural mutations 27 . Together, these findings suggest that deletions in katG, although uncommon, represent a critical mechanism of INH resistance that may evade detection in conventional molecular assays and must be considered in settings with discordant or atypical results.
DST was performed on 39 isolates with inferred mutations, providing a valuable correlation between genotypic and phenotypic resistance. Among these, 21 isolates harbored borderline rpoB mutations 15 , all confirmed as RIF susceptible in pDST, suggesting that the critical concentration of 0.5 µg/mL may not be appropriate for detecting low-level resistance caused by these mutations in MGIT, also reported by Rupasinghe et al. (2024) 28 . However, an isolate with the rpoB Q432R mutation and another with the rpoB P439L mutation were both classified as associated with interim resistance and tested MDR and RIF-susceptible in pDST, respectively.
Notably, we detected katG deletion in three isolates exhibiting no locus and WT bands in LPA, and the isolate tested on pDST was INH-R, confirming the loss-of-function nature of the deletion and its association with resistance 15 . Among the three isolates with inferred mutations in the inhA promoter (G-19T, C-20G, and T-8G) tested using pDST, only that with the C-20G mutation, not included in the catalogue, was INH-R. These findings highlight the importance of confirming drug resistance through phenotypic testing.
This study has some limitations. A larger number of isolates need to be analyzed using pDST to ensure a comprehensive elucidation of the variability in genotypic and phenotypic resistance patterns. The estimated prevalence was based solely on laboratory data, and the sociodemographic data of people with drug-resistant TB were not assessed. The results from the Xpert MTB/RIF Ultra assay were evaluated only in cases of unexpected LPA results and when available. Finally, neither the minimum inhibitory concentrations of RIF nor WGS were applied, which could have provided deeper insights into the resistance mechanisms, particularly for isolates with atypical mutations, e.g., large katG deletions, and isolates with mutations not included in the WHO catalogue.
Overall, this study highlights the complexity of drug resistance in TB and the critical role of molecular diagnostics, likely requiring a composite of phenotypic methods and other approaches, including sequencing, to guide effective diagnosis and treatment strategies. The detection of novel mutations requires continuous surveillance and adaptation of diagnostic tools to ensure the timely and accurate detection of resistance.
ACKNOWLEDGEMENTS
The authors thank the staff of their laboratory and institution for their technical assistance and support.
Funding Statement
Fundação de Amparo à Pesquisa do Estado de São Paulo (FAPESP) [grant number 2017/16082-7]
Footnotes
Financial support: Fundação de Amparo à Pesquisa do Estado de São Paulo (FAPESP) [grant number 2017/16082-7].
REFERENCES
- 1.World Health Organization (WHO) Global tuberculosis report. 2024
- 2.Ministério da Saúde. Brasil. Boletim Epidemiológico - Tuberculose. 2024. 72 p [Google Scholar]
- 3.Brandao AP, Pinhata JMW, Oliveira RS, Galesi VMN, Caiaffa-Filho HH, Ferrazoli L. Speeding up the diagnosis of multidrug-resistant tuberculosis in a high-burden region with the use of a commercial line probe assay. J Bras Pneumol. 2019;45(2):e20180128. doi: 10.1590/1806-3713/e20180128. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Brandao AP, Pinhata JMW, Simonsen V, Oliveira RS, Ghisi KT, Rabello MCS, et al. Transmission of Mycobacterium tuberculosis presenting unusually high discordance between genotypic and phenotypic resistance to rifampicin in an endemic tuberculosis setting. Tuberculosis. 2020;(25):102004–102004. doi: 10.1016/j.tube.2020.102004. [DOI] [PubMed] [Google Scholar]
- 5.Matsui T, Pinhata JMW, Rabello MCDS, Brandão AP, Ferrazoli L, Leão SC, et al. Frequency of first and second-line drug resistance-associated mutations among resistant Mycobacterium tuberculosis clinical isolates from São Paulo, Brazil. Mem Inst Oswaldo Cruz. 2020;(115):e200055. doi: 10.1590/0074-02760200055. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Malaquias TDSS, Ribeiro EP, Dutra TCP, Ricardo M, Salvato R, Bhering M, et al. The performance of Xpert MTB/RIF and MTBDRplus within a Programmatic setting at TB Laboratory in Rio de Janeiro, Brazil. Rev Soc Bras Med Trop. 2024;57:e004162024. doi: 10.1590/0037-8682-0167-2024. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Pinhata JMW, Brandao AP, Mendes F de F, Rabello MC da S, Ferrazoli L, de Oliveira RS. Correlating genetic mutations with isoniazid phenotypic levels of resistance in Mycobacterium tuberculosis isolates from patients with drug-resistant tuberculosis in a high burden setting. Eur J Clin Microbiol Infect Dis. 2021;40(12):2551–2561. doi: 10.1007/s10096-021-04316-0. [DOI] [PubMed] [Google Scholar]
- 8.Pinhata JMW, Brandao AP, Gallo JF, Oliveira RS de, Ferrazoli L. GenoType MTBDRsl for detection of second-line drugs and ethambutol resistance in multidrug-resistant Mycobacterium tuberculosis isolates at a high-throughput laboratory. Diagn Microbiol Infect Dis. 2023;105(2):115856–115856. doi: 10.1016/j.diagmicrobio.2022.115856. [DOI] [PubMed] [Google Scholar]
- 9.Pinhata JMW, Ferrazoli L, Mendes F de F, et al. A descriptive study on isoniazid resistance-associated mutations, clustering and treatment outcomes of drug-resistant tuberculosis in a high burden country. Eur J Clin Microbiol Infect Dis. 2024;43(1):73–85. doi: 10.1007/s10096-023-04693-8. [DOI] [PubMed] [Google Scholar]
- 10.Pinhata JMW, Felippe IM, Gallo JF, Chimara E, Ferrazoli L, de Oliveira RS. Growth characteristics of liquid cultures increase the reliability of presumptive identification of Mycobacterium tuberculosis complex. J Med Microbiol. 2018;67(6):828–833. doi: 10.1099/jmm.0.000734. [DOI] [PubMed] [Google Scholar]
- 11.Simeão FC dos S, Chimara E, Oliveira RS, Yamauchi JU, Latrilha FO, Telles MA da S. Cord factor detection and macroscopic evaluation of mycobacterial colonies: an efficient combined screening test for the presumptive identification of Mycobacterium tuberculosis complex on solid media. J Bras Pneumol. 2009;35(12):1212–1216. doi: 10.1590/s1806-37132009001200008. [DOI] [PubMed] [Google Scholar]
- 12.Ministério da Saúde . Manual de recomendações para o Controle da Tuberculose no Brasil. 2019. 364 p [Google Scholar]
- 13.Hain Lifescience GenoType MTBDRplus VER 2.0 - Molecular Genetic Assay for Identification of the M. tuberculosis Complex and its Resistance to Rifampicin and Isoniazid from Clinical Specimens and Cultivated Samples. Hain lifescience. 2015;(02):1–10. 1. Mtbdr G. [Google Scholar]
- 14.World Health Organization (WHO) Technical manual for drug susceptibility testing of medicines used in the treatment of tuberculosis. 2018. 52 p [Google Scholar]
- 15.World Health Organization (WHO) Catalogue of mutations in Mycobacterium tuberculosis complex and their association with drug resistance. 2nd edition. 2023. 140 p [Google Scholar]
- 16.Adami AG, Gallo JF, Pinhata JMW, Martins MC, Giampaglia CMS, de Oliveira RS. Modified protocol for drug susceptibility testing of MGIT cultures of Mycobacterium tuberculosis by the MGIT 960. Diagn Microbiol Infect Dis. 2017;87(2):108–111. doi: 10.1016/j.diagmicrobio.2016.10.023. [DOI] [PubMed] [Google Scholar]
- 17.World Health Organization . Technical Report on critical concentrations for drug susceptibility testing of isoniazid and the rifamycins (rifampicin, rifabutin and rifapentine) 2021. 107 p [Google Scholar]
- 18.Van Deun A, Decroo T, Aung KJM, Hossain MA, Gumusboga M, De Rijk WB, et al. Mycobacterium tuberculosis borderline rpoB mutations: Emerging from the unknown. Eur Respir J. 2021;58(3):2100783–2100783. doi: 10.1183/13993003.00783-2021. [DOI] [PubMed] [Google Scholar]
- 19.Pontes TAA, Fernandez-Llimos F, Wiens A. Impact of COVID-19 on tuberculosis notifications. Rev Inst Med Trop Sao Paulo. 2024;7(66):e37. doi: 10.1590/S1678-9946202466037. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Salvato RS, Reis AJ, Schiefelbein SH, Gómez MAA, Salvato SS, da Silva LV, Costa ERD, et al. Genomic-based surveillance reveals high ongoing transmission of multi-drug-resistant Mycobacterium tuberculosis in Southern Brazil. Int J Antimicrob Agents. 2021;58(4):106401–106401. doi: 10.1016/j.ijantimicag.2021.106401. [DOI] [PubMed] [Google Scholar]
- 21.Lempens P, Van Deun A, Aung KJM, Hossain MA, Behruznia M, Decroo T, et al. Borderline rpoB mutations transmit at the same rate as common rpoB mutations in a tuberculosis cohort in Bangladesh. Microb Genom. 2023;9(9):001109–001109. doi: 10.1099/mgen.0.001109. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Mvelase NR, Pillay M, Sibanda W, Ngozo JN, Brust JCM, Mlisana KP. rpoB Mutations Causing Discordant Rifampicin Susceptibility in Mycobacterium tuberculosis: Retrospective Analysis of Prevalence, Phenotypic, Genotypic, and Treatment Outcomes. Open Forum Infect Dis. 2019;6(4):ofz065–ofz065. doi: 10.1093/ofid/ofz065. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Claassens M, Dreyer V, Nepolo E, Mokomele Q, van Rooyen G, Ruswa N, et al. Whole-Genome Sequencing for Resistance Prediction and Transmission Analysis of Mycobacterium tuberculosis Complex Strains from Namibia. Microbiol Spectr. 2022;10(5):e0158622. doi: 10.1128/spectrum.01586-22. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Hazbón MH, Brimacombe M, Bobadilla del Valle M, Cavatore M, Guerrero MI, Varma-Basil M, et al. Population genetics study of isoniazid resistance mutations and evolution of multidrug-resistant Mycobacterium tuberculosis. Antimicrob Agents Chemother. 2006;50(8):2640–2649. doi: 10.1128/AAC.00112-06. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.De Maio F, Cingolani A, Bianco DM, Salustri A, Palucci I, Sanguinetti M, et al. First description of the katG gene deletion in a Mycobacterium tuberculosis clinical isolate and its impact on the mycobacterial fitness. Int J Med Microbiol. 2021;311(4):151506–151506. doi: 10.1016/j.ijmm.2021.151506. [DOI] [PubMed] [Google Scholar]
- 26.Pretorius GS, van Helden PD, Sirgel F, Eisenach KD, Victor TC. Mutations in katG gene sequences in isoniazid-resistant clinical isolates of Mycobacterium tuberculosis are rare. Antimicrob Agents Chemother. 1995;39(10):2276–2281. doi: 10.1128/aac.39.10.2276. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Siqueira HR de, Freitas FAD de, Oliveira DN de, Barreto AMW, Dalcolmo MP, Albano RM. Isoniazid-resistant Mycobacterium tuberculosis strains arising from mutations in two different regions of the katG gene. J Bras Pneumol. 2009;35(8):773–779. doi: 10.1590/s1806-37132009000800009. [DOI] [PubMed] [Google Scholar]
- 28.Rupasinghe P, Ashraf A, Barreda N, Parveen S, Zubair M, Calderon R, et al. Reduced critical concentration might not have improved MGIT- based DST’s sensitivity to rifampicin. Antimicrob Agents Chemother. 2024;68(5):e0170123. doi: 10.1128/aac.01701-23. [DOI] [PMC free article] [PubMed] [Google Scholar]