The Mtb H37Rv (ATCC 25618) and K-strain (NCCP 15986) were used as references in the drug susceptibility test. The 82 MDR, and 40 XDR isolates were supplied by the Tuberculosis Specimen Bank of Masan National Hospital. The clinical isolates were collected from November 2009 to January 2016. The isolates used in this study did not have drug susceptibility test results for bedaquiline or linezolid, so the definition of XDR before 2021 was used for drug resistance criteria. The MDR was defined as resistant to both isoniazid and rifampin. The XDR was defined as MDR plus resistance to any fluoroquinolone and any second-line injectable drug (capreomycin, kanamycin, and amikacin) [8]. The clinical isolates of Mtb, Middlebrook 7H9 Broth (BBL MGIT Mycobacteria Growth Indicator Tube, Becton-Dickinson, Sparks, MD, USA), and Ogawa II agar (Asanpharm, Seoul, Korea) were sequentially cultured and used. The bacterial colonies in the log phase were taken and applied to determine the minimum inhibitory concentration (MIC), polymerase chain reaction (PCR), and DNA sequencing.

The MIC of the TB strains was measured using the resazurin microtiter assay [9]. Amoxicillin and meropenem (Sigma-Aldrich, St. Louis, MO, USA) were added to 7H9-S broth with final concentrations ranging from 0.031 to 128 mg/L in 96-well plates. Potassium clavulanate (Sigma-Aldrich) was added to 7H9-S broth with a final concentration of 2.5 mg/L in each well. Growth controls containing no antibiotic and sterility controls without inoculation were also included. The inoculum was adjusted to McFarland 1.0 with DensiCHEK plus instrument (bioMereux, Craponne, France) from fresh colonies on Ogawa II agar, and further diluted at a ratio of 1:10 in 7H9-S broth. The MIC was determined as the lowest concentration where no color change occurred by culturing Mtb in 7H9-S broth containing various concentrations of Amoxicillin and meropenem for up to 9 days, then adding resazurin (Fisher Scientific, Waltham, MA, USA) and culturing for up to 48 hours. The Mtb H37Rv (ATCC 25618) and Mtb K-strain (NCCP 15986) were used as references in the drug susceptibility test. Drug susceptibility tests for the isolates that showed MIC >32/2.5 mg/L in to amoxicillin/clavulanate or meropenem/clavulanate, were repeated twice.

Mtb DNA was extracted using a commercial kit, DNeasy UltraClean Microbial Kit (Qiagen, Hilden, Germany) for PCR of the β-lactam resistance and susceptibility-related genes. PCR amplification of the genes was performed using Maxime PCR premix i-StarTaq (iNtRON, Seongnam, Korea). To identify genes related to high amoxicillin and meropenem susceptibility, seven strains with MIC ≤2 mg/L were selected, and DNA was extracted. Five strains with amoxicillin MIC >32 mg/L or meropenem MIC >16 mg/L were selected, and DNA was extracted to identify genes related to high β-lactam antibiotics resistance. Drug resistance data of the clinical Mtb isolates selected for PCR, and DNA sequencing are shown in Table 1.

The blaC, blaI, ldtA, ldtB, dacB2, and crfA genes were amplified by PCR. Genomic DNA from Mtb isolates and reference strains were subjected to PCR amplification. The primer sets for amplification and sequencing of the β-lactam resistance and susceptibility-related genes are shown in Table 2. The PCR amplification products of each gene were performed using Sanger sequencing with the same primers by Bioneer (Daejeon, Korea). To confirm the mutation site of the genes, each Mtb strain was compared and analyzed using Mega 10.2.5 software [10].

The susceptibility rates of Mtb isolates to amoxicillin and meropenem with clavulanate were analyzed using the Chi-square test. A cross-tabulation analysis was used to assess meropenem and amoxicillin susceptibility between MDR and XDR isolates. We used the nonparametric test because the MICs were recorded as discrete ordinal values and were not continuous. Fisher’s exact test and Paired sample t-test were used to compare the susceptible rates of MDR and XDR isolates to meropenem and amoxicillin with clavulanate. A p-value of <0.05 was considered statistically significant.

This study protocol was approved by the Institutional Biosafety Committee (MTHIBC-18-11) and Institutional Review Board (IRB-398837-2018-E-26) of the Masan National Tuberculosis Hospital. The clinical information was retrieved by the medical staff from electronic medical records, to analyze the relationship between the treatment history and progress in patients with β-lactam resistance.

Amoxicillin/clavulanate MIC values of H37Rv were from 32 to 64/2.5 mg/L, and K-strain was from 64 to 128/2.5 mg/L. Meropenem/clavulanate MIC values of H37Rv were from 8 to 16/2.5 mg/L, and K-strain was from 16 to 64/2.5 mg/L. The WHO does not define a critical concentration and breakpoint for amoxicillin/clavulanate and meropenem/clavulanate susceptibility testing for Mtb. Based on a previously described study [11], the tentative susceptible criterion of amoxicillin/clavulanate and meropenem/clavulanate was ≤4/2.5 mg/L and ≤8/2.5 mg/L, respectively. The susceptibilities of drug-resistant isolates to different concentrations of amoxicillin and meropenem with clavulanate are shown in Table 3. Clinical Mtb isolates to 4 mg amoxicillin and 8 mg meropenem with potassium clavulanate at 2.5 mg/L, showed 24.6% (30/122) and 94.3% (115/122), respectively. Among these, MDR and XDR strains were 28.0% (23/82) and 17.5% (7/40) susceptible to amoxicillin-clavulanate, respectively. The susceptibility rates of clinical Mtb isolates were increased on two additional concentrations of amoxicillin and meropenem. MDR and XDR strains were 85.4% (70/82) and 87.5% (35/40) susceptible to meropenem-clavulanate, respectively. The amoxicillin-clavulanate susceptibilities were higher in MDR than XDR, but the difference was not statistically significant (p=0.38). The meropenem-clavulanate susceptibilities were higher in XDR than MDR, but the difference was not statistically significant (p=0.12).

The MICs of the amoxicillin and meropenem with clavulanate for MDR and XDR Mtb isolates are shown in Figure 1. MIC₅₀ and MIC₉₀ values of amoxicillin/clavulanate, which mean 50% and more than 90% inhibition of the target Mtb, were 16/2.5 mg/L and 64/2.5 mg/L, respectively. The MIC₅₀ and MIC₉₀ values of meropenem/clavulanate were 8/2.5 mg/L and 16/2.5 mg/L, respectively. The MIC₅₀ and MIC₉₀ values of amoxicillin/clavulanate were higher than those of the meropenem/clavulanate (p=0.001). XDR-TB is more difficult to treat than MDR-TB. Therefore, we included XDR Mtb, since in vitro susceptibility to amoxicillin/clavulanate and meropenem/clavulanate may provide effective treatment options for patients with XDR-TB. The MIC₅₀ and MIC₉₀ values of amoxicillin/clavulanate, which mean 50% and more than 90% inhibition of the target Mtb, for MDR were 8/2.5 mg/L and 64/2.5 mg/L, respectively. The MIC₅₀ and MIC₉₀ values of amoxicillin/clavulanate for XDR were 16/2.5 mg/L and 32/2.5 mg/L, respectively. The MIC₅₀ values of amoxicillin/clavulanate were higher in XDR than MDR (p<0.001), and The MIC₉₀ values of amoxicillin/clavulanate were higher in MDR than XDR (p<0.001). The MIC₅₀ and MIC₉₀ values of meropenem/clavulanate for MDR were 4/2.5 mg/L and 16/2.5 mg/L, respectively. The MIC₅₀ and MIC₉₀ values of meropenem/clavulanate for XDR were 8/2.5 mg/L and 16/2.5 mg/L, respectively. The MIC₅₀ values of meropenem/clavulanate were higher in XDR than MDR (p<0.001).

The treatment history of the patients and bacterial culture test of isolates are shown in Table 4. The five strains confirmed to be highly resistant to amoxicillin MIC >32 mg/L or meropenem MIC >16 mg/L, were isolated from four patients. Only a patient from whom 23920 strain was isolated, had a history of treatment with amoxicillin-clavulanate for 12 months during inpatient treatment. However, the isolate was cultured from the patient before administering amoxicillin-clavulanate and meropenem. The other patients had no history of treatment with amoxicillin-clavulanate and meropenem. The relationship between drug discontinuation and resistance was low in the β-lactam resistant isolates.

The relationship between mutations in the β-lactam target genes and the drug susceptibility and resistance was investigated. A total of 12 Mtb isolates with blaC, blaI, crfA, ldtA, ldtB, and dacB2 genes were sequenced and analyzed. Gene mutations related to amoxicillin and meropenem resistance could not be identified in five resistant Mtb isolates.

The G514A a sensitivity-related mutation in the blaC gene previously reported concerning β-lactam drug-susceptibility [11], could not identify the same positional mutation in the seven susceptible Mtb isolates in this study. However, the T448G common mutation of the blaC gene was identified in susceptible Mtb isolates, as shown in Figure 2. The T448G mutation of the blaC was observed in four isolates among the seven β-lactam susceptible Mtb isolates. The mutation was detected with a higher frequency in XDR than in MDR among the β-lactam drug-susceptible Mtb isolates.