The in vitro effect of new combinations of carbapenem–β-lactamase inhibitors for Mycobacterium abscessus

ABSTRACT

As new treatment alternatives for Mycobacterium abscessus complex (MABC) are urgently needed, we determined the minimum inhibitory concentrations (MICs) for novel carbapenem combinations, including imipenem-relebactam and tebipenem-avibactam against 98 MABC isolates by broth microdilution. The MIC50 was reduced from 16 to 8 mg/L by adding relebactam to imipenem, while the addition of avibactam to tebipenem showed a more pronounced reduction from 256 to 16 mg/L, representing a promising non-toxic, oral treatment option for further exploration.

INTRODUCTION

Non-tuberculous mycobacteria (NTM) are increasing globally, of which Mycobacterium abscessus complex (MABC) is particularly difficult to treat. Despite long treatment durations with multiple drugs, treatment success in Europe is only 30%–50% (1) and worse for subspecies abscessus (33%) compared to subspecies massiliense (56%) (2).

Due to intrinsic and acquired resistance of MABC to first-line drugs such as macrolides (erm and rrl genes) and aminoglycosides (rrs gene), few treatment alternatives exist (3), and repurposed drugs are currently highlighted for possible treatment options (4). MABC is highly resistant to most β-lactam antibiotics, and the current guidelines include only imipenem and cefoxitin (5). The major challenge to the use of β-lactam antibiotics against MABC is the β-lactamase BlaMab, requiring the addition of a β-lactamase inhibitor (BLI) (6). The novel intravenous combination of imipenem-relebactam is approved for resistant Gram-negative infections (7), and the oral carbapenem tebipenem was recently shown effective against complicated urinary tract infections (8). Relebactam is structurally related to avibactam, which has been shown to inhibit BlaMab (7).

We have previously shown an association between rough colony morphology and poor treatment outcome of MABC infection (9), and we hypothesized that this could be due to a possible association with drug resistance. Therefore, we have determined the in vitro effect of imipenem-relebactam and tebipenem-avibactam against MABC isolates, exploring the potential effect of MABC subspecies, erm and colony morphology on minimum inhibitory concentration (MIC) levels.

Frozen MABC clinical isolates from Sweden collected during 2009–2021 (n = 92) and six isolates from an external quality panel (Instand, Germany) 2020–2021 were thawed and cultured on Löwenstein-Jensen medium at 30°C. Data were available regarding cording (by fluorescent microscopy), MABC subspecies, i.e., abscessus, massiliense, and bolletii, as well as erm, rrl, and rrs genotype (Genotype NTM-DR, HAIN LifeSciences, Germany) (10). MABC colony morphology, i.e., smooth, rough, or mixed, was determined by light microscopy as previously described (9).

MICs of carbapenems alone and in combination with BLIs were determined by broth microdilution (BMD) in cation-adjusted BBL Mueller-Hinton II broth (MH). Imipenem and avibactam (MedChemExpress, USA), tebipenem (Spero Therapeutics, USA), and relebactam (Merck, Germany) compounds were solved in H2O, phosphate buffer, or dimethyl sulfoxide (DMSO) according to manufacturer’s instructions, and aliquoted stock solutions were frozen in −80°C. On the day of testing, stock solutions were diluted in MH broth, and 100 µL was dispersed into 96 well plates, where carbapenems were distributed by stepwise 1:2 dilutions from 256 to 0.5 mg/L, and BLIs were distributed at a fix final test concentration of 4 mg/L, informed by previous studies (11 – 14). MABC colonies from growing cultures were suspended in MH broth to ca. 5 × 10⁵ CFU/mL, and 100 µL was distributed in all wells. MIC was defined as the lowest drug concentration without visible bacterial growth after 3–5 days of incubation at 30°C. Reference strains Mycobacterium peregrinum ATCC 700686 and M. abscessus ATCC 19977 were included as quality controls (QC). Mann-Whitney or Kruskal-Wallis one-way ANOVA was used to compare MIC values between groups. P-value <0.05 was considered statistically significant.

Characteristics of MABC isolates are summarized in Table 1. MIC including 50% of the isolates (MIC50) was reduced one dilution step from 16 to 8 mg/L by adding relebactam to imipenem, while the addition of avibactam to tebipenem showed a pronounced reduction of the MIC50 from 256 to 16 mg/L (Fig. 1a and b) (P < 0.001). For QC M. peregrinum and M. abscessus, median MICs for imipenem were 2 and 16 mg/L, while for tebipenem, median MICs were 1 and >256 mg/L, respectively, with ranges within ±1 dilution step.

TABLE 1.

Characteristics of MABC isolates ^a

MAB subspecies	Total (n)	Cording n (%)	Morphology n (%)			erm(41) n (%)	rrl n (%)	rrs n (%)
		Detected	Smooth	Rough	Mixed	28T	Mutation	Mutation
M. abscessus	53	19 (36%)	22 (42%)	22 (42%)	9 (17%)	40 (75%)	0	0
M. massiliense	26	6 (23%)	17 (65%)	6 (23%)	3 (12%)	Inactive	1 b (4%)	0
M. bolletii	19	2 (11%)	14 (74%)	5 (26%)	0	19 (100%)	0	0

^a Number of isolates and percentage of the total number of respective subspecies (within brackets) with regards to the formation of cording, colony morphology, and resistant genotype of erm(41), rrl, or rrs. Mixed morphology was defined as the occurrence of smooth and rough colonies in the same isolate.

^b rrl A2058G/T.

Fig 1.

MIC distribution of (a) imipenem (IMI) ± relebactam (REL) and (b) tebipenem (TEB) ± avibactam (AVI). QC MPER, quality control M. peregrinum ATCC 700686; QC MAB, quality control M. abscessus ATCC 19977. The majority of the 98 isolates were clinical, whereas 6 isolates were part of an external control panel (Instand, Germany).

For imipenem alone, similar MIC levels were found in a large study including over a thousand of MABC isolates as determined by BMD in MH (15). In a smaller study including 28 MABC isolates, the addition of relebactam to imipenem resulted in the same reduction in modal MIC from 16 to 8 mg/L as in our study (BMD 7H9) (16), a difference that may, however, be explained by the innate variability of any MIC determination (17).

Our results regarding tebipenem are corroborated by previously published studies. Kaushik et al. identified avibactam as the most effective BLI in lowering the MABC MICs of tebipenem, in a study including eight different carbapenems combined with avibactam, sulbactam, or tazobactam, respectively (BMD 7H9). The modal tebipenem MIC of 256 mg/L was reduced to 8 mg/L (range 4–16 mg/L) when combined with avibactam (n = 28) (11). Avibactam has been shown to inhibit BlaMab more rapidly than relebactam in a hydrolysis kinetics study (18). Only a modest reduction has been seen by the addition of clavulanic acid 5 mg/L for M. abscessus ATCC 19,977, as compared with the effect on BlaC expressed by Mycobacterium tuberculosis (19, 20).

As for the high MIC values of tebipenem tested singly, similar MIC50s of 256 to >256 mg/L have been previously reported (11, 16, 21). In another study including 19 clinical MABC isolates, the MIC of the majority of isolates was above the upper test concentration of 256 mg/L, with only two isolates with MICs below 256 mg/L (22).

Regarding colony morphology, there was a tendency of lower tebipenem MICs for rough compared to smooth morphology (MIC50 128 vs 256 mg/L); however, this was not statistically significant (P = 0.064). Otherwise, there were no significant differences in MIC50s for any of the carbapenem alone or in combination with BLI with regard to (a) MABC subspecies, (b) erm genotype, (c) formation of cording, or (d) morphology (see https://doi.org/10.6084/m9.figshare.23589786.v1), which is in line with previous research (23 – 28). From our data, a rough colony morphology does not appear to be a major mechanism of resistance to carbapenems (9).

In vivo studies and clinical breakpoints for imipenem-relebactam or tebipenem-avibactam are lacking for MABC. The MIC breakpoints of imipenem are defined by Clinial & Laboratory Standards Institute (CLSI) as ≤4 mg/L for susceptible, 8–16 mg/L for intermediate, and ≥32 mg/L for resistant (29). Based on our results, the combination of imipenem-relebactam means that the majority of MABC isolates would fall into the intermediate category (MIC50 8 mg/L) and as such, still two dilution steps above the non-species related PK/PD breakpoint (MIC ≤2 mg/L) as suggested by the European Committee on Antimicrobial Susceptibility Testing (EUCAST), indicating that combination therapies are still needed to enhance activity.

As for tebipenem, the mean C _max values for tebipenem was 15.1 mg/L following 600 mg of a prodrug of tebipenem thrice daily for 2 wk in 30 healthy volunteers (8). In a mural model, the bacterial growth increased as soon as the free plasma concentrations of tebipenem were less than the MIC of the isolate (30). An oral prodrug of avibactam (AVI-006) is included in phase I clinical trial (ClinicalTrials.gov NCT03931876). Promising in vivo activity has been shown in a mouse model by this combination, claiming satisfying PK/PD data and likely bactericidal effect of MABC pulmonary infection (13).

In conclusion, the role of the all-oral combination tebipenem-avibactam for MABC may deserve further exploration.