Chronic Airway Colonization by Achromobacter xylosoxidans in Cystic Fibrosis Patients Is Not Sustained by Their Domestic Environment

Achromobacter-specific PCR in silico performances.Based on the Proteobacteria 16S rRNA gene aligned sequences (n = 4,592), only three regions allowing primer design (20 to 30 bp) showed sequences conserved among members of the Achromobacter genus and distinct from those of other genera. The two most distant sequences were chosen to design forward primer 638-F (5′-CGCAGGCGGTTCGGAAAGAAA-3′) and reverse primer 810-R (5′-GCCTCCTGGGATAACACTGA-3′), allowing the amplification of a 174-bp product. Both primers exactly anneal to the sequences for type strains of 16 Achromobacter species available in RDP-II but also to sequences for type strains of 8 Bordetella species. If 2 nucleotide mismatches per primer are allowed, 9 other genera were detected: Advenella, Alcaligenes, Kerstersia, Pelistega, Pigmentiphaga, Pusillimonas, Taylorella, Candidimonas, and Eoetvoesia. Using the primer targeted on the third consensual region (amplification product of 120 bp) gave no gain in specificity (data not shown).

Achromobacter detection by specific PCR and cultivation on selective medium.The performance of the Achromobacter selective agar medium described by Amoureux et al. has not been fully validated, i.e., evaluated for all species in the genus Achromobacter or a variety of strains within a species (15). Therefore, to overcome a potential culture bias, the strategy performed in this study was to screen cultures obtained on a nonselective medium by applying the genus-specific PCR before cultivating the samples onto the Achromobacter selective agar medium. Among the 273 cultivable communities obtained on Trypticase soya (TS) agar, 92 showed a positive Achromobacter-specific PCR result, indicating that the prevalence of members of the genus Achromobacter was 33.7% in the collection of 273 samples from CF patient houses.

From the 92 corresponding cultures on the Achromobacter selective medium, a total of 316 colonies corresponding to different morphotypes were identified by matrix-assisted laser desorption ionization–time of flight mass spectrometry (MALDI-TOF MS) and 21 isolates from 15 samples were affiliated with the genus Achromobacter (13 with an identification score between 1.70 and 1.99 and 8 strains with a score of ≥2 using the Microflex Biotyper database [Bruker, USA]). In 11 of these 15 samples, Achromobacter was cocultured on the selective medium with members of other bacterial genera, including Bordetella (Table 1). Achromobacter species strains were isolated by culture from only 16.5% of the PCR-positive samples.

The low rate of isolation is possibly due to the partial selectivity of the medium, allowing the growth of some other genera than Achromobacter that may hide the growth of Achromobacter in the case that the inoculum is of low density. Moreover, the growing capacity of the selective medium was not evaluated for all of the 19 species currently described for the Achromobacter genus (15). False-positive PCR by cross-reaction with 10 predicted genera is unlikely to explain the observed level of discrepancy between PCR and culture. Particularly, Bordetella was isolated from only 10% of PCR-positive samples and coisolated with Achromobacter strains in only 3 samples.

As generally observed for targeted approaches of bacterial diversity, the full recovery of all the members of the genus was not warranted and the diversity of Achromobacter spp. may be higher than described. Metabarcoding that avoids cultural bias can be proposed for accurate detection of Achromobacter in domestic environments as previously performed for indoor community description (16). However, coupling detection by specific PCR and cultivation on selective medium as described here provide a useful tool for studies aiming for prevalence assessment in the environment while also recovering living material for further genotyping and phenotyping.

Cartography of Achromobacter in patients' homes.We explored a limited number of CF patients' homes (n = 3) (patients P2, P5, and P12) but analyzed a large number of sites per home to accurately delineate Achromobacter reservoirs. Positive Achromobacter-specific PCR was used for mapping Achromobacter habitats. The majority of PCR-positive samples came from indoors (82.5%), mostly from wet sites (78%). The majority of positive samples came from the kitchen for P5 and P12 (28.5% and 50%, respectively) and from the garden for P2 (35%) (Fig. 1). Twenty percent of all the PCR-positive samples came from the bathroom. Wastewater drainage systems were particularly contaminated, and 100% of 11 water samples from siphons were positive by Achromobacter-specific PCR (Fig. 1). Achromobacter was also highly prevalent in soils and plants samples, at 75% of positive samples.

• Open in new tab
• Download powerpoint

FIG 1

Samples positive by Achromobacter-specific PCR: total, per location, and in the wastewaters for the 3 homes studied (those of P2, P5, and P12). The number at the top of each bar represents the sample number.

Considering Achromobacter strains cultivated on the selective medium, the origin of the isolates was roughly similar to that highlighted by PCR mapping, and most strains came from the indoor environment, mainly from the bathroom in moist sites, including the bath mat, toothbrush glass, water evacuation of household appliances, siphons (washbasin, dish sink, or shower), and water from the toilet bowl (Table 1). Achromobacter strains were also detected in the washing machine drain in two houses. Outdoors, Achromobacter grew from three soil samples and one rhizosphere (Table 1).

In our study, water-containing evacuation systems were contaminated by Achromobacter spp. but also by several other known CF opportunistic pathogens (17–20), including Stenotrophomonas maltophilia, Bordetella, Ochrobactrum, and Agrobacterium/Rhizobium. As shown in previous studies, wastewater drains are the primary habitat of P. aeruginosa (10–13); we confirm that siphons and other similar draining systems host a large number of opportunistic pathogens. These devices are more and more frequently described as reservoirs for pathogens and multidrug-resistant bacteria in both hospitals and domestic environments (21–23). Moreover, aerosolization and cross-transmission of pathogens from siphons have been demonstrated in recent carefully designed experimental studies (24, 25). Our results reinforced recommendations to CF patients for home cleaning that should be targeted to the bathroom, kitchen, and other wet sites, with particular emphasis on wastewater drains, including hidden drains in domestic appliances. We proposed regular biocidal decontamination of siphons of water points of use (daily for the more exposing devices or weekly) and annual change of wastewater drains to reduce the global exposure to opportunistic pathogens, including multidrug-resistant bacteria.

High Achromobacter diversity in CF patients' homes.Achromobacter genotyping performed by multilocus sequence typing (MLST) allowed the accurate identification of species and the description of the sequence type (ST) diversity in CF domestic environments but also the comparison of results to international data. The 21 Achromobacter sp. isolates from patient homes belonged to 7 species and 15 STs (Table 1). Only 3 STs exactly matched available STs in the PubMLST database. Therefore, we found in 3 CF patient's homes 12 new genotypes within the genus Achromobacter.

Achromobacter mucicolens was the major species identified in this study (8 isolates from 5 samples) and the only one identified in every home. A. mucicolens was isolated from various types of samples, e.g., flowerpot, siphons, and toothbrush glass. Five STs were identified among the 8 A. mucicolens strains, and strains of 2 STs differing by 2 alleles were coisolated in a soil sample (P5-74) (Table 1).

Five A. xylosoxidans isolates were identified in 4 samples in only one home (P5): the first-flush water from the sink faucet in bathroom, the cleaning bucket, the floor cloth, and residual water in the reservoir of the tumble dryer (Table 1). A. xylosoxidans ST175 was identified in the 4 samples and associated with A. xylosoxidans ST169 in the floor cloth water sample; the 7 alleles of the MLST scheme differed between these two STs. ST175 was identified in the CF respiratory tract of another patient of the Montpellier CF center in 2014 and has been described worldwide from CF and non-CF samples (https://pubmlst.org). ST169 was described only for blood of a non-CF patient with a long-term line for dialysis in United Kingdom.

Other isolates belonged to Achromobacter aegrifaciens (2 samples in 2 houses, 3 isolates of 2 STs, including ST143, described for environmental samples in Belgium), Achromobacter spanius (2 isolates of different STs in 2 samples from 2 houses), Achromobacter marplatensis, Achromobacter animicus, and Achromobacter sp. genogroup 9 (one isolate each) (Table 1). Of note, the species A. mucicolens and A. animicus were until then isolated from human samples only. Most species identified from the domestic environment in our study were considered minor CF-associated species in several studies (8, 26–28). For example, A. mucicolens, a species described in 2013 on the basis of five human clinical isolates, including one from CF sputum (29), and currently only scarcely reported for CF, was the major species identified in domestic environments in this study. In contrast, A. xylosoxidans, the major species identified in CF patients, and chronically colonizing 80% of the CF patients in our center (7), was recovered from only one house.

We observed a high within-species genotype diversity among and within homes, and within samples, as well as one unknown species (Achromobacter genogroup 9) and numerous as-yet-unreported genotypes, highlighting a still unrecognized diversity of members of the genus Achromobacter.

Antimicrobial susceptibility profile of domestic Achromobacter isolates.Isolates were fully susceptible to penicillins, ceftazidime, carbapenems, colistin, levofloxacin, ciprofloxacin, and co-trimoxazole but resistant to cefoxitine, cefpodoxime, aztreonam, and fosfomycin (Table 2). A. xylosoxidans and A. aegrifaciens showed higher resistance to aminoglycosides and tetracycline than other species (Table 2). Isolates investigated in this study thus exhibited patterns similar to those observed for innately resistant A. xylosoxidans, i.e., resistance to many antibiotics, including cephalosporins (except ceftazidime) and aztreonam, with no additional acquired resistances. Only rare studies described the antimicrobial susceptibility profiles of nonhuman Achromobacter isolates; they reported resistance to aminoglycosides in Achromobacter sp. strains (unidentified species) recovered from non-CF domestic environments in Japan (14) and ciprofloxacin- and/or imipenem-resistant A. xylosoxidans strains from non-CF domestic environments in France (100% and 10% of 9 isolates, respectively) (15).

No link found between house and airways isolates.Patients P2, P5, and P12 were chronically colonized by A. xylosoxidans ST328, ST27, and ST327, respectively. No A. xylosoxidans isolates were found in the domestic environments of patients P2 and P12. The 5 A. xylosoxidans strains isolated in the home of patient P5 belonged to ST175 or ST169. These STs, respectively, differed by 4 and 7 alleles out of the 7 alleles of the MLST scheme from ST27, the ST of the strains recovered from the patient airways. Compared to ST27, each distinct allele differed by 10 to 31 single nucleotide polymorphisms (SNPs) for ST169 and by 2 to 5 SNPs for ST175.

The results suggested either the absence of an environmental reservoir for the strain colonizing the patients or the failure to detect an existing reservoir. Despite the intensive sampling performed, the presence of an undetected reservoir could not be totally ruled out even if a massive environmental reservoir is unlikely. Moreover, the absence of Achromobacter on patient personal items such as toothbrush glasses (i) suggested a low rate of transmission from patient to environment, (ii) reinforced the hypothesis of the absence of an environmental reservoir, and (iii) disproved the hypothesis that a domestic reservoir could be involved in the persistence of airway colonization by frequent reseeding. The correlated hypothesis that patients recontaminate their environment could also be rejected. These results suggested that after initial colonization from an environmental source and the subsequent persistence probably associated with specialization to the CF airways, the survival of CF airway-adapted A. xylosoxidans strains in the environment could be impaired. This hypothesis is supported by results of previous studies on other CF pathogens. Environmental investigations in the homes of CF patients who were chronically colonized by Burkholderia cepacia complex bacteria revealed the presence of diverse Burkholderia species but did not find the bacterial genotypes colonizing the patients (30). Analysis of the toothbrushes of CF patients colonized by P. aeruginosa, of siblings of these patients, and of volunteers showed no statistical relationship between pulmonary colonization and toothbrush contamination (31). Furthermore, studies showed that P. aeruginosa strains adapted to CF airways lose their ability to colonize a new host in a murine model and that hypermutable CF strains lose their survival capacity in tap water (32, 33).

Finally, we showed that environmental strains had wild-type antimicrobial resistance phenotypes, while the 3 patients were colonized by isolates displaying higher resistance toward several antimicrobial agents, including ceftazidime, ertapenem, aminoglycosides, and fluoroquinolones (Table 2), suggesting that no diffusion of transmissible antimicrobial resistance determinants occurred between environmental and clinical Achromobacter isolates of different genotype or species within domestic environment.