Development and Validation of a PCR Assay To Detect the Prairie Epidemic Strain of Pseudomonas aeruginosa from Patients with Cystic Fibrosis

Abstract

The monitoring of epidemic Pseudomonas aeruginosa is important for cystic fibrosis (CF) infection control. The prairie epidemic strain (PES) is common in western Canadian CF clinics. Using whole-genome sequencing, we identified a novel genomic island and developed a PCR assay for PES. Against a collection of 186 P. aeruginosa isolates, the assay had 98% sensitivity and 100% specificity.

TEXT

Multiple epidemic strains of Pseudomonas aeruginosa have been described as infecting patients with cystic fibrosis (CF) (1–4). Many of these strains are associated with a poor prognosis, manifesting as increased rates of lung function decline, exacerbation frequency, and treatment burden; decreased quality of life; and hastened progression to end-stage lung diseases (5, 6). We recently described a novel epidemic P. aeruginosa strain, the prairie epidemic strain (PES) (multilocus sequence type 192) (7). PES was found in one-quarter of patients attending the Calgary Adult CF Clinic (CACFC) and from patients who were transferred from other Prairie-based clinics, suggesting broad endemicity in western Canada. Furthermore, its prevalence has been unchanged through 3 decades of patients who transitioned to the CACFC. PES is more likely to be antibiotic resistant than are unique strains of P. aeruginosa, and infection with PES has been associated with increased rates of lung function decline, pulmonary exacerbation frequency, and progression to death and/or transplantation (7, 8).

The importance of identifying epidemic P. aeruginosa strains in CF has long been recognized. Typing strategies that are commonly currently employed include pulsed-field gel electrophoresis (PFGE), multilocus sequence typing (MLST), and random amplified polymorphic DNA PCR (RAPD-PCR) (7, 9, 10). These modalities have, to varied degrees, the ability to distinguish bacterial strains from one another, identify known epidemic strains, and determine if novel strains are potentially shared. However, they are costly and laborious, precluding routine use in clinical laboratories. Accordingly, we sought to develop a simple PCR assay to distinguish PES from non-PES P. aeruginosa organisms for clinical and research purposes in areas where this strain is prevalent (western Canada). Similar assays have been developed for other epidemic P. aeruginosa strains, including Liverpool epidemic strain (LES), Midlands-1, Manchester, and Australia epidemic strain-1 (AUS-1), and they are used prospectively in infection control and research (1–4).

We performed whole-genome sequencing on 8 nonsequential isolates of PFGE and MLST-confirmed PES from our prospectively maintained biobank. The isolates were sequenced on the Illumina HiSeq 2500 (150 bp, paired-end) and assembled using SPAdes 2.5.0 (11). These strains were sequenced as a part of a large sequencing study, and a more detailed description will be given in a forthcoming manuscript (M. G. Surette, unpublished data). We identified a novel 40-kb genomic island (GI) using IslandViewer (12) and GView (13) that was not present in other published P. aeruginosa genomes. Using the Primer3 software (14), we developed a pair of primers (forward, ACTGGAACCGAAGCGTCATT; reverse, GGTATCGGCATGTCCATCGT) to amplify a segment within the GI.

To perform the assay, boil preparations of each isolate were performed. For the assay, the following conditions were used: from each boil preparation, 1 μl of template was diluted to 50 μl total volume with 25 mU/μl Taq DNA polymerase (Invitrogen), 1× Taq buffer, 1× Enhancer buffer (catalog no. 52391; Invitrogen), 200 μM deoxynucleoside triphosphates (dNTPs), 600 nM (each) primer, and 1.5 mM MgCl₂. The amplification conditions consisted of an initial 3 min at 95°C, followed by 30 cycles of denaturation at 95°C for 30 s, annealing for 45 s at 61°C, and extension for 1 min at 72°C). A final extension consisting of 1 min at 61°C, followed by 5 min at 72°C, was then performed.

We then assessed the sensitivity and specificity of these primers against a collection involving P. aeruginosa isolated from local CF patients, other CF transmissible strains (oCFTS), common laboratory strains (LAB), and local control populations from patients with community-acquired bacteremia (CAB) (isolation within 48 h of hospitalization) and local environmental isolates (ENV). P. aeruginosa was confirmed using standard techniques and strains genotyped by PFGE, supplemented with MLST. Common non-P. aeruginosa CF-derived pathogens were similarly included as further controls, including Staphylococcus aureus, Stenotrophomonas maltophilia, and Achromobacter xylosoxidans. Whole-sputum samples from patients with documented PES chronic infection and patients without documented PES infection were also assessed (15).

The GI was characterized by a number of predicted phage genes, including integration host factor (IHF), an integrase, and a putative phage two-component histidine kinase (M. G. Surette, unpublished data). Using the PES novel primer pair, a PCR product of 1,032 bp in length was amplified and sequenced, corresponding to the novel GI of PES (Fig. 1).

FIG 1.

Comparison of the PES-PCR product from P. aeruginosa derived from a subset of isolates included in the study; isolates from different patients infected with PES (lanes 2 to 5), nontransmissible P. aeruginosa (lanes 6 to 10), oCFTS, including LES and strain A/B (lanes 11 to 14), and non-CF isolates, including LAB (lane 15), CAB (lanes 16 and 17), and ENV (lane 18).

The primers were assessed against an established and characterized biobank consisting of PFGE/MLST-confirmed PES isolates (n = 57) from 41 different CF patients, and unique isolates (n = 39) from 31 different CF patients with unique nonclonal strains (UNI); oCFTS (n = 23), including (LES) (n = 6)/strain A (n = 1), strain B (n = 3), the Midlands strain (Md-1) (n = 2), the Manchester epidemic strain (n = 1), Australian epidemic strains AUST-01 (n = 2), AUST-02 (n = 2), AUST-03 (n = 1), AUST-04 (n = 1), AUST-6 (n = 1), clone C (n = 2), and isolate 0207 (n = 1) (6, 7), and LAB (n = 5; P. aeruginosa PAO1, PAK, PA103, PA14, and ATCC 27853), CAB (n = 37), and ENV (n = 25). A PCR product of 1,032 bp in size was identified in 56/57 (98%) PFGE/MLST-confirmed PES and in 0/39 (0%) UNI, 0/23 (0%) oCFTS, 0/5 (0%) LAB, 0/37 (0%) CAB, and 0/25 (0%) ENV samples. The primer pair had a sensitivity of 98% and a specificity of 100% for correctly identifying PES relative to the gold standard. Ten isolates each of S. aureus, S. maltophilia, and A. xylosoxidans derived from different CF patients were screened, and none produced an amplification product. Of the whole-sputum samples from five patients with chronic PES infection, all produced a 1,032-bp product, whereas whole-sputum samples from 10 patients without chronic PES infection failed to produce an amplification product.

Here, we have identified a small target within a novel GI in PES that can be used with a high degree of sensitivity and specificity to identify PES. We will use this on an annual basis for all P. aeruginosa chronically infected CF patients (not previously infected with PES) to monitor for superinfection events, and on all newly infected patients. This assay can be adopted for use in other North American CF clinics to identify the point prevalence of PES, something particularly relevant in the Prairie Provinces of Canada, given evidence of endemicity and the adverse prognosis associated with PES infection (8). Importantly, PCR-based assays are inherently limited, in that they are able to assess only for the epidemic strain being sought. However, within our patient population, we previously identified that PES is the dominant epidemic strain, and that superinfection events with other strains, including PES, are very low, with none having been documented in the last 5 years.

As epidemic P. aeruginosa infections carry a grave prognosis in CF, the prevention of cross-infection is key. Multiple studies have demonstrated that epidemic P. aeruginosa can persist in the air for extended periods of time after coughing (16–18). However, observed incident superinfection events have very rarely been documented (6, 7). Measures to prevent the transmission of epidemic P. aeruginosa are thus not evidence based (19). Accordingly, the proposed strategies to prevent incident infections are increasingly complex and varied (20–23).

With this assay, we have the ability to retrospectively study superinfection events, something not possible through prospective studies, owing to their very low incidence. Unique to the CACFC, we maintain a comprehensive biobank of each and every bacterial isolate from all attendees of the CACFC (n = ∼38,000 since 1978). As part of our work, we previously characterized at least the first and last isolate from our entire cohort. We can now retrospectively identify the first moment in time when patients experienced PES superinfection and assess the circumstances that culminated in superinfection events.

Finally, we have confirmed that PES is a CF-specific strain of P. aeruginosa. Despite screening a large population of control strains from local environments and patients with CAB and non-CF bronchiectasis (7), no PES cases have been identified outside CF.