In vitro activity of 20 antibiotics against Cupriavidus clinical strains Open Access

Sir,

Cupriavidus are Gram-negative non-lactose-fermenting motile bacilli with peritrichous flagella, a number of which were previously and successively classified in the Ralstonia and Wautersia genera. Until the recent expansion of MALDI-TOF MS, Cupriavidus could be mistaken for Burkholderia or Pseudomonas species. They are resistant to heavy metals and have been described from environmental (soil and water) samples, as well as from human samples.^1,Cupriavidus gilardii, Cupriavidus pauculus and Cupriavidus metallidurans are involved in invasive human infections, such as bacteraemia and pneumonia, most of which (though not exclusively) occur in immunocompromised patients.^2–4 Additionally, Cupriavidus species, Cupriavidus respiraculi in particular, are increasingly identified in patients with cystic fibrosis (CF).⁵ However, their clinical relevance in CF is not established. Due to the rare occurrence of Cupriavidus infections, antibiotic susceptibility data are only available from sparse case reports. Therefore, we determined the MICs of 20 antibiotics for a panel of Cupriavidus clinical strains, mainly from respiratory samples of CF patients (82%).

Thirty-seven epidemiologically unrelated clinical isolates of Cupriavidus obtained from the collection of the French Observatoire Burkholderia cepacia and from 11 French hospitals, as well as two type strains from clinical sources, i.e. C. pauculus LMG 3244^T and C. respiraculi LMG 21510^T (Laboratory of Microbiology, Ghent University, Ghent, Belgium),¹ were included. Isolates were identified by amplified ribosomal DNA restriction analysis (ARDRA)⁶ and MALDI-TOF MS (Maldi Biotyper Microflex^®, Bruker Daltonics, Bremen, Germany; IVD 7712). The experimental panel thus comprised 18 C. respiraculi, 6 C. gilardii, 5 C. pauculus, 4 C. metallidurans, 2 Cupriavidus necator, 2 Cupriavidus taiwanensis, 1 Cupriavidus basilensis and 1 unidentified Cupriavidus sp.

The MICs of 20 antibiotics, listed in Table 1, were determined using the broth microdilution method, as recommended by EUCAST (www.eucast.org). Briefly, each strain was inoculated on a blood agar plate (bioMérieux, Marcy-l’Étoile, France) for 16 h at 35°C. Bacterial suspensions in Mueller–Hinton broth (Bio-Rad, Marnes-la-Coquette, France) at concentrations of 5 × 10⁵ cfu/mL were dispensed in 96-well microtitre plates (Dutscher, Brumath, France, 160 μL per well). Antimicrobial agents were added at increasing 2-fold concentrations (40 μL per well). The MICs were determined as the lowest antibiotic concentrations that inhibited visible bacterial growth after an 18 ± 2 h incubation at 35°C in an aerobic atmosphere. Escherichia coli ATCC 25922, Pseudomonas aeruginosa ATCC 27853 and Staphylococcus aureus ATCC 29213 were used as controls. All MICs were determined in triplicate and replicates never differed by more than 2-fold. For each MIC determination, the median of the replicates was recorded as the MIC. The MICs were interpreted according to Pseudomonas EUCAST breakpoints (2019) for colistin, amikacin and tobramycin, according to non-Enterobacteriaceae CLSI breakpoints (2019; https://clsi.org) for minocycline and co-trimoxazole and according to pharmacokinetic/pharmacodynamic (non-species-related) EUCAST breakpoints (2019) for the other antibiotics, except for temocillin, for which the previously proposed breakpoint of 16 mg/L was used.⁷ The EUCAST and CLSI breakpoints that could be used to interpret Cupriavidus MICs are listed in Table S1 (available as Supplementary data at JAC Online).

The susceptibility testing results are summarized in Table 1 and the full MIC distributions of each tested antibiotic are available in Figure S1. Since C. pauculus and C. metallidurans are phylogenetically close species¹ and exhibited the same susceptibility profiles, they were considered as a group. Our collection comprises very few strains of C. basilensis, C. necator, C. taiwanensis and Cupriavidus sp. (maximum n=2 for each species). Moreover, these species have similar antibiotic susceptibility profiles, that's why they were also considered as a group. The susceptibility testing results interpreted with CLSI breakpoints are available in Table S2. The only significant change in our study between the EUCAST and CLSI interpretations concerned piperacillin/tazobactam. Eleven strains had an MIC of 8 or 16 mg/L (nine C. respiraculi). They were classified as susceptible to piperacillin/tazobactam according to CLSI breakpoints, but intermediate according to EUCAST breakpoints.

Nearly all strains were resistant to amoxicillin, amoxicillin/clavulanate, temocillin and aztreonam. Regarding cephalosporins, only 23% of Cupriavidus strains were ceftazidime susceptible, whereas 74% and 82% were susceptible to ceftriaxone or cefotaxime, respectively. The ceftolozane/tazobactam combination also demonstrated good activity, except against C. gilardii. For most strains, ceftazidime/avibactam MICs were similar compared with ceftazidime alone. Cefepime was the most active β-lactam, with 95% of strains being susceptible, whereas only a few strains were susceptible to meropenem (8%). Interspecies differences were observed for piperacillin/tazobactam and imipenem, since they were less active against C. respiraculi and C. gilardii than against the other species. Such discrepancies between meropenem and imipenem activities were previously noticed in case reports.^2,3 Similarly to P. aeruginosa, they could be due to the overexpression of efflux pumps from the resistance-nodulation-division family.⁸ Indeed, a homologue of the MexAB OprM efflux pump that extrudes meropenem in P. aeruginosa has been identified in C. gilardii.⁹

Aminoglycosides were poorly active, in agreement with case reports,^2,3 probably due to efflux pumps and aminoglycoside-modifying enzymes.⁹ Minocycline was the most active antibiotic, with very low MICs and a 100% susceptibility rate. Fluoroquinolones were frequently active, with over 80% of strains being susceptible, except for C. pauculus and C. metallidurans. Interspecies discrepancies were also noticed for co-trimoxazole. It was active against approximatively 80% of C. gilardii and C. respiraculi strains, whereas more than 50% of the strains belonging to other species were resistant.

Over 90% of C. respiraculi and 67% of C. gilardii strains were susceptible to colistin, while strains from the other species were mostly colistin-resistant. Colistin susceptibility was one of the characteristics of the Cupriavidus genus initially described by Vaneechoutte et al.¹ However, Petrou et al.¹⁰ showed that the expression of ArnT was particularly strong in a strain of C. metallidurans. This enzyme catalyses the attachment of the cationic sugar 4-amino-4-deoxy l-arabinose (l-Ara4N) to lipid A phosphate groups. The subsequent reduction of negative membrane charge is responsible for colistin resistance. We detected homologues of arnT (CP000353.2: 1481129–1482725) using BLASTn in C. basilensis, C. necator, C. pauculus and C. taiwanensis sequenced strains (a query cover >80%, an identity >70% and an E value <1 × 10⁻⁴⁰ were chosen as cut-off values for significance), which is in accordance with the high rate of colistin resistance in these species observed in our study. Additionally, C. gilardii appears to be the origin of the gene mcr-5, which is an emerging plasmid-mediated mechanism of colistin resistance in other environmental species such as Salmonella and Pseudomonas.¹¹

In conclusion, our study showed that minocycline and cefepime exhibited the best in vitro activities against Cupriavidus strains. Meropenem, aminoglycosides and polymyxins, often considered antibiotics of last resort against infections caused by Gram-negative bacilli, do not have reliable activity against Cupriavidus. Perhaps resistance to these agents confers a selective advantage to Cupriavidus and therefore it may emerge in clinical scenarios where these agents are used, such as in patients with CF. Imipenem was more active than meropenem and cefotaxime/ceftriaxone was more active than ceftazidime. Ceftolozane/tazobactam had reasonable activity against Cupriavidus, whereas the novel inhibitor avibactam does not seem to add to the activity of ceftazidime. Interspecies variations were observed, especially concerning colistin, co-trimoxazole, fluoroquinolones and piperacillin/tazobactam. Clinical data is now required to establish the optimal treatment of Cupriavidus infections.

Acknowledgements

We would like to thank Dr Aberrane from Créteil Hospital, Dr Belmonte from St Denis de La Réunion Hospital, Dr Dib from Troyes Hospital, Dr Ferroni from Paris University Hospital, Dr De Gialluly from Tours University Hospital and Dr Sansot from Toulon Hospital for providing the French Observatoire Burkholderia cepacia with Cupriavidus strains.

Funding

This study was supported by internal funding.

Transparency declarations

None to declare.

Supplementary data

Tables S1 and S2 and Figure S1 are available as Supplementary data at JAC Online.

References