Diversity of Achromobacter species recovered from patients with cystic fibrosis, in Argentina

Abstract

Different phenotype-based techniques and molecular tools were used to describe the distribution of different Achromobacter species in patients with cystic fibrosis (CF) in Argentina, and to evaluate their antibiotic resistance profile. Phenotypic identification was performed by conventional biochemical tests, commercial galleries and MALDI-TOF MS. Genetic approaches included the detection of A. xylosoxidans specific marker bla_oxa-114, the amplification and sequencing of the 16S rRNA gene, nrdA and bla_OXA complete sequence, and MLST analysis. Phenotypic approaches, even MALDI-TOF, rendered inconclusive or misleading results. On the contrary, concordant results were achieved with the nrdA sequencing or sequence type (ST) analysis, and the complete bla_OXA sequencing, allowing a reliable discrimination of different Achromobacter species. A. xylosoxidans accounted for 63% of Achromobacter infections and A. ruhlandii accounted for 17%. The remaining species corresponded to A. insuavis, A. dolens, A. marplatensis and A. pulmonis. Antimicrobial susceptibilities were determined by the agar dilution method according to CLSI guidelines. Piperacillin, piperacillin/tazobactam and carbapenems were the most active antibiotics. However, the emergence of carbapenem-resistant isolates was detected. In conclusion, prompt and accurate identification tools were necessary to determine that different Achromobacter species may colonize/infect the airways of patients with CF. Moreover, antimicrobial therapy should be administered based on the susceptibility profile of individual Achromobacter sp. isolates.

Resumen

Se emplearon diversas técnicas fenotípicas y moleculares para describir la distribución de diferentes especies del género Achromobacter en pacientes con fibrosis quística (FQ) en Argentina, y se evaluó el perfil de resistencia a los antibióticos. Se realizó la identificación fenotípica por pruebas bioquímicas convencionales, galerías comerciales y MALDI-TOF MS. El enfoque genético incluyó la detección del marcador especie-específico de A. xylosoxidans bla_oxa-114, la amplificación y la secuenciación de los genes ARNr 16S, nrdA y secuencia completa de bla_OXA, y el análisis por MLST. Los enfoques fenotípicos, incluso la técnica de MALDI-TOF, proporcionaron resultados no concluyentes o erróneos. Por el contrario, se obtuvieron resultados concordantes entre la secuenciación del gen nrdA o el análisis de secuenciotipos (ST) y la secuenciación completa de bla_OXA, lo que permitió una discriminación confiable de las diferentes especies de Achromobacter. A. xylosoxidans representó el 63% de las infecciones por Achromobacter y A. ruhlandii representó el 17%. Las especies restantes correspondieron a A. insuavis, A. dolens, A. marplatensis y A. pulmonis. Se determinó la sensibilidad a antimicrobianos por el método de dilución en agar de acuerdo al CLSI. Los antibióticos más activos fueron piperacilina, piperacilina/tazobactam y carbapenemes. Sin embargo, se detectó la emergencia de aislamientos resistentes a carbapenemes. En conclusión, resultaron necesarias herramientas de identificación rápida y precisas para determinar las diferentes especies del género Achromobacter capaces de colonizar/infectar las vías respiratorias de los pacientes con FQ. Asimismo, la terapia antimicrobiana debería llevarse a cabo en función del perfil de sensibilidad de los aislamientos individuales de Achromobacter spp.

Introduction

Achromobacter spp. are increasingly recognized as emerging pathogens in patients with cystic fibrosis (CF)²⁰. Reported rates of Achromobacter colonization/infection in individuals with CF, vary between 2% and 17.9%, and display a rising tendency worldwide^3,10. Achromobacter xylosoxidans is the most frequent species recovered within this genus, however other species have been associated with human infections^15,17. Moreover, clinical isolates are mostly referred as A. xylosoxidans given that the accurate species identification of Achromobacter isolates is difficult.

Conventional phenotypic methods have been commonly used for bacterial identification of Achromobacter spp. in many clinical laboratories, as their implementation and cost make them more affordable. Since these classical methods yield unclear results, and generally fail to differentiate between species of the genus, molecular methods have been proposed as complementary or alternative procedures.

Amplification and sequencing of the 16S rRNA coding gene, which constitutes a useful method in the identification of numerous microorganisms, is not able to discriminate between species of Achromobacter^5,12. In 2011, Turton et al. proposed the amplification of the bla_OXA-114 gene for rapid and accurate A. xylosoxidans identification^9,21. Moreover, in 2013 we described the presence of bla_OXA-258, bla_OXA-364, and bla_OXA-243 in Achromobacter ruhlandii, Achromobacter dolens, and Achromobacter insuavis, respectively, as species-specific markers useful for bacterial identification^14,19. Simultaneously, Spilker et al. proposed the amplification and sequencing of an inner fragment of the nrdA gene, one of those included in the multilocus sequence typing (MLST) scheme, as a precise method for the identification of different species of Achromobacter¹⁷.

MALDI-TOF MS has emerged as a revolutionary technique for rapid bacterial identification. This method has been shown to be more rapid, accurate and cost-efficient than conventional phenotypic techniques or genotypic approaches. However, its performance is uncertain in infrequent species.

In this study, different phenotype-based techniques and molecular tools were conducted to describe the distribution of different species of Achromobacter in patients with CF in Argentina, and to evaluate their antibiotic resistance profile.

Methods

Isolates

Forty-one non-related Achromobacter spp. clinical isolates, recovered from patients with CF at 6 healthcare centers in Argentina during 1996–2013 were included. Isolates were mainly obtained from respiratory secretions (Table S1, supplementary material).

Phenotypic identification

Phenotypic identification was performed by biochemical tests according to Yabuuchi et al.²³ and Vandamme et al.²² The biochemically-based commercial system, API 20NE (bioMèrieux) was also conducted.

MALDI-TOF MS identification was performed using a Microflex MALDI-OF MS instrument (Bruker Daltonics, GmbH, Germany) and FlexControl 3.0 software (Bruker Daltonics). Identification scores ≥2.0 were accepted for a reliable identification at species level and scores between ≥1.7 and ≤2.0 were accepted for identification at genus level. Scores <1.7 indicated no reliable identification⁸.

Genotypic identification

The 16S rRNA gene was amplified by PCR as previously described¹³. Purified amplicons were sequenced in both strands using an ABI Prism DNA 3700 sequencer and compared with databases using the NCBI's BLAST tool.

The nrdA sequence was achieved according to Spilker et al. and compared with databases (https://pubmlst.org/achromobacter/)¹⁷.

The presence of A. xylosoxidans species-specific marker, bla_OXA-114, was investigated by PCR amplification as previously described by Turton et al.²¹ The complete bla_OXA-114, bla_OXA-258, bla_OXA-364 and bla_OXA-243 sequences intrinsic for A. xylosoxidans, A. ruhlandii, A. dolens, and A. insuavis, respectively, were amplified according to Traglia et al.¹⁹ Amplicon sequences were compared with those of all the different bla_OXA variants deposited in GenBank.

Multilocus sequence typing analysis

A multilocus sequence typing (MLST) scheme was conducted to identify those isolates which could not be unambiguously identified by nrdA sequencing. For this purpose, amplification and sequencing of inner fragments of seven housekeeping genes were performed and the corresponding allele profiles and sequence types (ST) were assigned according to the Achromobacter MLST website (http://pubmlst.org/achromobacter/)¹⁸.

Antimicrobial susceptibility testing

Minimal inhibitory concentrations (MIC) were determined for a representative set of antibiotics: ampicillin, piperacillin, cefoxitin, ceftazidime, cefepime, piperacillin/tazobactam, imipenem, meropenem, ciprofloxacin, levofloxacin, amikacin, kanamycin, gentamicin, trimethoprim-sulfamethoxazole, tetracycline and colistin. Antimicrobial susceptibilities were determined by the agar dilution method according to Clinical and Laboratory Standards Institute (CLSI) recommendations. The susceptibility breakpoints used in this study were those established for other non-Enterobacteriaceae⁶.

Results

Phenotypic identification results are shown in Table S1. According to classical biochemical methods, 27/41 isolates were identified as A. xylosoxidans, the remaining ones being identified as Achromobacter spp. Using the API 20NE commercial gallery, 39/41 corresponded to A. xylosoxidans. Based on MALDI-TOF MS data, all isolates except one were consistent with A. xylosoxidans. Species identification was achieved in 25/40 isolates with score ≥2, while 15/40 displayed scores from 1.7 to 1.9. None of the samples presented score <1.7.

The 16S rRNA sequences obtained for the isolates included in this study displayed about 99% identity with those deposited for different species of Achromobacter, being unable to discriminate among them.

The results of the other genetic approaches performed in this study are shown in Table 1. The amplification of bla_OXA-114 proposed by Turton et al.²¹ yielded positive results in 39/41 isolates, indicating that they corresponded to A xylosoxidans. However discrepancies were observed in 13/39 isolates when comparing these results with those obtained using other molecular methods. Based on these methods, the 13 isolates were finally identified as Achromobacter species other than A. xylosoxidans (Table 1).

Table 1.

Comparison of different genotypic methods for the identification of Achromobacter spp. isolates

Isolate	Year of isolation	nrdA analysis^17 and ST	PCR blaOXA^21	blaOXA^19
A39	2008	A. xylosoxidans	+	blaOXA-114m
A43	2009	A. dolens	+	blaOXA-787 (MK388909)
A45	2011	A. pulmonis	−	–
A46	2011	A. marplatensis	−	–
A49	2012	A. xylosoxidans	+	blaOXA-790 (MK388911)
A50	2012	A. dolens	+	blaOXA-789 (MK388910)
A51	2012	A. xylosoxidans	+	blaOXA-114o
A52	2004	A. xylosoxidans	+	blaOXA-791 (MK388912)
A53	2012	A. dolens	+	blaOXA-364a
A79	2012	A. xylosoxidans	+	blaOXA-114i (MK388913)
A113	2009	A. xylosoxidans	+	blaOXA-790 (MK388911)
A131	2012	A. ruhlandii	+	blaOXA-258a
A134	2013	A. ruhlandii	+	blaOXA-258a
38	2002	ST 148 (A. ruhlandii)^a	+	blaOXA-258a
39	2002	ST 148 (A. ruhlandii)^a	+	blaOXA-258a
67	2004	ST 148 (A. ruhlandii)^a	+	blaOXA-258a
79	2004	ST 64 (A. insuavis)^a	+	blaOXA-243a
80	2004	ST 14 (A. ruhlandii)^a	+	blaOXA-258a
114	2002	ST 165 (A. insuavis)^a	+	blaOXA-243e
319	1996	ST 43 (A. ruhlandii)^a	+	blaOXA-258b
336	2010	ST 164 (A. dolens)^a	+	blaOXA-364a
A1	2004	A. xylosoxidans	+	blaOXA-114g
A2	1995	A. xylosoxidans	+	blaOXA-114i
A3	2006	A. xylosoxidans	+	blaOXA-114f
A4	2003	A. xylosoxidans	+	blaOXA-114g
A5	2002	A. xylosoxidans	+	blaOXA-114l
A6	2007	A. xylosoxidans	+	blaOXA-114f
A7	2007	A. xylosoxidans	+	blaOXA-114c
A8	2008	A. xylosoxidans	+	blaOXA-114f
A9	2008	A. xylosoxidans	+	blaOXA-114l
A10	2008	A. xylosoxidans	+	blaOXA-114f
A11	2008	A. xylosoxidans	+	blaOXA-114f
A12	2008	A. xylosoxidans	+	blaOXA-114e
A13	2008	A. xylosoxidans	+	blaOXA-114l
A14	2001	A. xylosoxidans	+	blaOXA-114c
A15	2001	A. xylosoxidans	+	blaOXA-114c
A16	2007	A. xylosoxidans	+	blaOXA-114f
A17	2006	A. xylosoxidans	+	blaOXA-114a
A18	2006	A. xylosoxidans	+	blaOXA-114a
A19	2007	A. xylosoxidans	+	blaOXA-114i
A20	1996	A. xylosoxidans	+	blaOXA-114o

^a Isolates identified by ST analysis. Data published previously^14,19.

nrdA gene sequencing¹⁷ allowed to identify 33/41 isolates. In the remaining isolates, nrdA sequences corresponded to alleles that were not deposited in pubmlst.org/achromobacter. For these isolates, identification was performed based on the complete MLST Scheme¹⁸. Different alleles of bla_OXA-114 were observed in the 26 isolates identified as A. xylosoxidans by nrdA gene sequencing¹⁷; even, 3 new bla_OXA-114 variants were identified in this study (accession numbers: MK388911, MK388912, MK388913) (Table 1). These new bla_OXA-114 variants were designated as OXA-114i, OXA-790 and OXA-791 by NCBI. Then again, in those isolates identified as A. insuavis, A. dolens and A. ruhlandii by nrdA gene sequencing or ST analysis, the bla_OXA sequences displayed 99–100% identity to bla_OXA-243, bla_OXA-364 and bla_OXA-258, respectively. Two new bla_OXA-364 variants were identified (MK388909, MK388910) (Table 1). These bla_OXA-364 variants were designated as OXA-787 and OXA-789 by NCBI. No bla_OXA amplicons could be obtained in any of the isolates identified by nrdA sequencing¹⁷ as Achromobacter marplatensis and Achromobacter pulmonis (Table 1).

A. xylosoxidans accounted for 63.4% of Achromobacter colonization/infections in this study, while A. ruhlandii accounted for 17.1%. The remaining species corresponded to A. dolens (9.8%), A. insuavis (4.9%), A. marplatensis (2.4%) and A. pulmonis (2.4%). According to previous studies performed in the UK⁷, Spain⁴, France², USA¹⁷, Brazil¹⁶ and Denmark¹¹, A. xylosoxidans was the most frequently recovered species from the airways of patients with cystic fibrosis. A. ruhlandii, the second most prevalent species identified in this study, was only identified in 3/96 patients in the UK and none in Spain and France; however it was the second most prevalent species in the USA and Brazil. The reported prevalence for A. dolens varies between 2 to 17%, accounting for 10% in the present study. Discrepant results were observed in species prevalence in different countries, the local distribution being more similar to that reported in Brazil.

All isolates included in this study were resistant to quinolones and aminoglycosides, while piperacillin, piperacillin/tazobactam and carbapenems were the most active antibiotics, as it had been previously described by Almuzara et al.¹ Fifty percent (50%) of the isolates were resistant to trimethoprim-sulfamethoxazole while 85% were resistant to colistin. Moreover, 3/7 A. ruhlandii isolates and 1/26 A. xylosoxidans were resistant to imipenem. Antimicrobial MIC values for the Achromobacter isolates included in this study are shown in table S2 (supplementary material).

Conclusions

Although A. xylosoxidans was the most common species recovered from the clinical samples of patients with CF, it was not the only species present in those samples, as it was inferred from the phenotypic approaches analyzed in this study. A. ruhlandii was the second species in prevalence, in agreement with the results obtained in Brazil and USA.

MALDI-TOF rendered inconclusive or misleading results understating the presence of species other than A. xylosoxidans, probably due to commercial databases, key components of MALDI-TOF platforms, which are constructed with a low number of isolates mainly corresponding to A. xylosoxidans. Their expansion should be crucial to resolve many of the current inadequate identifications, improving the usefulness of this technique.

On the other hand, concordant results were achieved with nrdA sequencing or ST analysis, and the bla_OXA sequencing approach proposed by Traglia et al.¹⁹, allowing a reliable discrimination among the different Achromobacter species, and demonstrating that a wide diversity of Achromobacter spp. may colonize and/or infect the airways of patients with CF. Although the MLST scheme constitutes an accurate tool, its arduous procedure may not be applicable in moderate complexity laboratories. In this regard, the amplification and sequencing of a single gene, nrdA or bla_OXA, may be more pertinent.

The susceptibility profile of Achromobacter spp. indicated that these microorganisms were resistant to a wide range of antibiotics, including fluoroquinolones, aminoglycosides and the majority of broad-spectrum β-lactams. Carbapenems were the most active antibiotics; however, the emergence of resistant isolates was detected. No correlation could be established among the susceptibility profiles and Achromobacter species. Therefore, the antimicrobial therapy in patients with CF should be conducted based on the susceptibility profile of individual Achromobacter spp. isolates.

Finally, prompt and accurate identification tools should provide an opportunity to understand the clinical impact of the different Achromobacter species on the progression of respiratory infections in patients with CF. In this sense, the development of a robust MALDI-TOF database should be desirable.

Ethical approval

Not required.

Funding sources

This work was partially supported by grants from UBACyT to M. Radice and G. Gutkind; ANPCyT to M. Radice, and PIP to G. Gutkind.

M. Radice and G. Gutkind are members of Carrera del Investigador Científico (CONICET). M. Papalia is recipient of a posdoctoral fellowship from CONICET.

Conflict of interest

The authors declare that they have no conflicts of interest.