Abstract
Background
Pseudomonas aeruginosa is a multidrug-resistant pathogen causing recalcitrant pulmonary infections in people with cystic fibrosis (pwCF). Cystic fibrosis transmembrane conductance regulator (CFTR) modulators have been developed that partially correct the defective chloride channel driving disease. Despite the many clinical benefits, studies in adults have demonstrated that while P. aeruginosa sputum load decreases, chronic infection persists. Here, we investigate how P. aeruginosa in pwCF may change in the altered lung environment after CFTR modulation.
Methods
P. aeruginosa strains (n = 105) were isolated from the sputum of 11 chronically colonized pwCF at baseline and up to 21 months posttreatment with elexacaftor-tezacaftor-ivacaftor or tezacaftor-ivacaftor. Phenotypic characterization and comparative genomics were performed.
Results
Clonal lineages of P. aeruginosa persisted after therapy, with no evidence of displacement by alternative strains. We identified commonly mutated genes among patient isolates that may be positively selected for in the CFTR-modulated lung. However, classic chronic P. aeruginosa phenotypes such as mucoid morphology were sustained, and isolates remained just as resistant to clinically relevant antibiotics.
Conclusions
Despite the clinical benefits of CFTR modulators, clonal lineages of P. aeruginosa persist that may prove just as difficult to manage in the future, especially in pwCF with advanced lung disease.
Keywords: CFTR modulators, cystic fibrosis, elexacaftor-tezacaftor-ivacaftor, ETI, Pseudomonas aeruginosa
We found that clonal strains of Pseudomonas persist after elexacaftor-tezacaftor-ivacaftor treatment with the same “chronic” phenotypes that are just as clinically challenging in people with cystic fibrosis. We further identified commonly mutated bacterial genes that may drive future adaptation to the CFTR (cystic fibrosis transmembrane conductance regulator) modulated lung.
Cystic fibrosis (CF) is caused by mutations in the CFTR gene (CF transmembrane conductance regulator) that result in defective cellular chloride transport [1]. CF is a multisystem disorder, although the predominant cause of morbidity and mortality is due to severe progressive pulmonary disease as a consequence of recurrent and chronic respiratory infections [1].
A principal opportunistic pathogen causing recalcitrant pulmonary infections in people with CF (pwCF) is Pseudomonas aeruginosa. This gram-negative and highly multidrug-resistant bacterium can adapt to the hostile CF lung environment to establish persistent infections that are extremely difficult to eradicate [2]. Development of chronic infection is characterized by the downregulation of bacterial virulence factors, such as oligosaccharide antigen (O-antigen), and the formation of mucoid colonies that overproduce the exopolysaccharide alginate that provides stability for the formation of biofilms [2, 3]. Established clonal communities often persist and continue to diversify for decades [2, 4]; however, highly transmissible P. aeruginosa strains can displace existing chronic communities or co-colonize pwCF, causing mixed-strain infections that further complicate treatment regimens [4–7]. P. aeruginosa infections are correlated with increased pulmonary exacerbations, accelerated lung function decline, impaired quality of life, and early death in pwCF [8].
The advent of effective CFTR modulators in CF has radically changed the landscape of CF health care. The latest triple-combination therapy, elexacaftor-tezacaftor-ivacaftor (ETI; Trikafta), has been found to have the greatest clinical benefits, with up to 90% of pwCF having a genotype responsive to this therapy [9, 10]. Demonstrated benefits of CFTR modulator therapy in pwCF include improvement in lung function and nutritional health indices, a reduced incidence of pulmonary exacerbations, and improved quality of life [10, 11].
Despite the clear benefits to health in pwCF, the impact of CFTR modulators on chronic P. aeruginosa airway infection remains undercharacterized. Observational studies show a reduction in P. aeruginosa–positive sputum cultures by up to 55% in the first year of CFTR modulator treatment [12–14], but this apparent eradication was mainly in individuals who were intermittently colonized. In chronic P. aeruginosa infection, modulators have been associated with a reduction in P. aeruginosa bacterial load but not complete eradication [15–17].
In this study, we aimed to characterize the changes in genotype and phenotype of persisting P. aeruginosa strains collected for up to 21 months after CFTR modulation in pwCF receiving mainly ETI or tezacaftor-ivacaftor (TI). Understanding how CFTR modulators affect CF respiratory infections is critical to direct future research and clinical care, especially for P. aeruginosa, which is notoriously difficult to treat.
METHODS
Participant Recruitment
Thirty-one adults with CF and chronic P. aeruginosa infection commencing TI (n = 9) or compassionate access ETI (n = 22) were recruited between September 2019 and October 2020 after human research and ethics committee (HREC/2019/QPCH/46169) approval and participant written consent (Figure 1). Chronic infection was defined as a participant having P. aeruginosa present in >50% of the last six sputum cultures collected over a minimum of 12 months [18]. Fifteen patients were not able to provide follow-up samples due to a combination of becoming unproductive for sputum following commencement of therapy and/or failure to attend face-to-face follow-up due to the coronavirus pandemic response. This led our study to include pwCF with more severe lung disease and a smaller sample size than desired. One participant, although productive, became culture negative for P. aeruginosa and therefore could not be studied. Fifteen participants provided spontaneously expectorated sputum samples before and at various time points up to 21 months after CFTR modulator commencement. Of these 15 patients, a final cohort of 11 pwCF receiving ETI (n = 8), TI (n = 1), and TI followed by ETI (n = 2) were investigated in detail (Figure 1). Participants one to 11 are labeled with the code CF participant (CFP) throughout the study.
Figure 1.
Workflow of participant recruitment, cystic fibrosis transmembrane conductance regulator (CFTR) modulator therapy, and Pseudomonas aeruginosa (PA) collection. Thirty-one people with cystic fibrosis (pwCF) were initially recruited. TI/ETI indicates that isolates were collected prior to participants starting tezacaftor-ivacaftor (TI) before switching to elexacaftor-tezacaftor-ivacaftor (ETI).
Clinical Data Collection
Clinical details of the 11 participants were collected for the 12 months prior to and 12 months after modulator commencement. Data collected included best forced expiratory volume in 1 second percent predicted (FEV1pp; calculated with Global Lung Function Initiative standards), body mass index (BMI), and number of exacerbations/intravenous antibiotic use. Sweat chloride was collected with the Macroduct sweat collection system per the manufacturer's instructions at the time of therapy commencement and between 1 and 12 months posttreatment.
Bacterial Isolation and Culture
A small plug of sputum was inoculated onto horse blood Columbia agar (PP2001; Thermo Fisher Scientific) and MacConkey agar (base 3, CM0115; OXOID). Plates were incubated for up to 72 hours at 37 °C ± 5% CO2. Morphologically distinct colonies were selected, grown in lysogeny broth at 37 °C, shaken at 220 revolutions per minute overnight, and stored in 20% glycerol at −80 °C. Bacterial species were confirmed by MALDI-TOF (matrix-assisted laser desorption/ionization–time of flight) VITEK mass spectrometry.
Whole Genome Sequencing
All 105 P. aeruginosa isolates were cultured in lysogeny broth by overnight shaking, and genomic DNA was extracted with the Qiagen DNA Mini Kit per the manufacturer's instructions. Genomic DNA was sent to Microbes NG (https://microbesng.com) for Illumina sequencing with 2 × 250–bp paired-end reads with a minimum 30× depth coverage. See supplementary data 1 for methods on genome assembly and quality assessment.
Comparative Genomics
Genome sequences were queried in PubMLST.org to determine multilocus sequence type (MLST) profiles that differentiated strains by sequence variation (alleles) of seven housekeeping genes [19]. Sequence types that are close together in numerical value do not indicate closely related strains. Assembled contigs were queried with PAst V1.0, classifying P. aeruginosa isolates into one of 11 serogroups based on BLAST analysis of the O-specific antigen gene cluster (covering the 20 International Antigenic Typing Scheme serotypes) [20]. We utilized Snippy 4.6.0 (https://github.com/tseemann/snippy) to call variants, including single-nucleotide polymorphisms and indel mutations, from the whole genome shotgun sequences of 90 isolates from persisting clonal lineages (14 clonal lineages from 11 patients) against the P. aeruginosa PAO1 reference genome (assembly accession GCF_000006765.1 [21]). See the supplementary data 1 for detailed methods of the supplementary figures and variant calling performed. Figures were created in R Studio with the ggplot and pheatmap packages and BioRender.com.
O-antigen Expression
O-antigen expression was determined through sodium dodecyl sulfate–polyacrylamide gel electrophoresis of patient isolate lipopolysaccharide and by Western blotting with polyvalent P. aeruginosa O-antigen–specific antiserum (200372; Denka Company Limited). See supplementary data 1 for detailed methods.
Antibiotic Susceptibility Testing
To determine antibiotic susceptibility profiles of P. aeruginosa isolates, disc diffusion was performed per the Clinical Laboratory Standards Institute guidelines and standards. See supplementary data 1 for detailed methods.
Statistical Analyses
All statistical analyses were performed in Prism (version 9; GraphPad). A Wilcoxon signed rank paired test was performed for comparing clinical parameter outcomes, an unpaired Mann-Whitney test for hypermutator isolate analysis, and a Fisher's exact test for pathoadaptive and antibiotic gene enrichment and for comparison of number of isolates with chronic phenotypes (O-antigen expression, mucoid morphology, antibiotic susceptibility). P < .05 was considered significant.
Supplementary Material
Notes
Acknowledgments. We thank the individuals with CF and their families for participating in this study. We are very grateful for all staff at the Adult Cystic Fibrosis Centre at the Prince Charles Hospital, Brisbane, Australia, for aiding in the collection of research samples for this study. We acknowledge Dr Kay Ramsay for her assistance with methodology and the Paterson group at the University of Queensland Centre for Clinical Research for MALDI-TOF services. This research was carried out at the Translational Research Institute, Woolloongabba, Australia, which is supported by a grant from the Australian Government.
Author contributions. Conceptualization: E. L. L., T. J. W., D. J. S., D. W. R. Methodology: E. L. L., T. J. W., D. J. S., D. W. R., J. J. T., M. M., J. B. G. Participant recruitment, sample, and data collection: D. J. S., D. W. R., E. L. L., M. E. W., P. E. W. Investigation: E. L. L., J. J. T. Writing: E. L. L. Reviewing and editing: E. L. L., T. J. W., D. J. S., D. W. R., J. B. G., M. M., J. J. T. All authors have read and agreed to the published version of the manuscript.
Financial support . This work was supported by the Cystic Fibrosis Foundation (00849I221 to T. J. W., J. B. G., and D. J. S.); The Common Good Prince Charles Hospital Foundation (NI2020-61 to E. L. L. and D. J. S.); and Australian Cystic Fibrosis Research Trust (postgraduate studentship grant 2021 to E. L. L.).
Contributor Information
Emma L Ledger, Frazer Institute, Faculty of Medicine, The University of Queensland, Brisbane, Australia.
Daniel J Smith, Northside Clinical Unit, The University of Queensland, Brisbane, Australia; Adult Cystic Fibrosis Centre, The Prince Charles Hospital, Brisbane, Australia.
Jing Jie Teh, Frazer Institute, Faculty of Medicine, The University of Queensland, Brisbane, Australia.
Michelle E Wood, Adult Cystic Fibrosis Centre, The Prince Charles Hospital, Brisbane, Australia.
Page E Whibley, Adult Cystic Fibrosis Centre, The Prince Charles Hospital, Brisbane, Australia.
Mark Morrison, Frazer Institute, Faculty of Medicine, The University of Queensland, Brisbane, Australia; Australian Infectious Diseases Research Centre, Brisbane, Australia.
Joanna B Goldberg, Department of Pediatrics, Division of Pulmonary, Asthma, Cystic Fibrosis, and Sleep, Emory University School of Medicine, Atlanta, Georgia, USA.
David W Reid, Northside Clinical Unit, The University of Queensland, Brisbane, Australia; Australian Infectious Diseases Research Centre, Brisbane, Australia; QIMR Berghofer Medical Research Institute, Brisbane, Australia.
Timothy J Wells, Frazer Institute, Faculty of Medicine, The University of Queensland, Brisbane, Australia; Australian Infectious Diseases Research Centre, Brisbane, Australia.
Data Availability
All P. aeruginosa genomes are available on NCBI BioProject PRJNA1023362 (see Supplementary Table 1 for corresponding BioSample accession numbers). All isolate phenotypic data are available in Supplementary Table 2, clinical data in Supplementary Table 3, and all filtered mutations in Supplementary Data 2 and 3.
RESULTS
Cohort Characteristics
Of the 11 participants who provided multiple P. aeruginosa–positive samples, each had at least one copy of the F508del CFTR mutation. The cohort had a median age of 30.1 years and an almost equal proportion of males to females (Table 1). Eight patients received ETI, one TI, and two initially commenced TI but then switched to ETI six and 12 months after commencement (Figure 1). Almost half (45%) were utilizing lumacaftor-ivacaftor at the beginning of the study. At baseline, the median sweat chloride of 95 mmol/L was above diagnostic levels for CF (>60 mmol/L). The median best FEV1pp in 12 months prior to commencement was 40%, reflecting a cohort of participants with advanced lung disease. A total of 105 P. aeruginosa isolates were collected. Only 1 patient was co-colonized with another CF pathogen, Staphylococcus aureus, prior to modulator commencement.
Table 1.
Cohort Characteristics at Baseline
| No. (%) or Median (IQR) | ||||
|---|---|---|---|---|
| Characteristic | Whole Cohort | ETI | TI | TI/ETIa |
| Patients | 11 | 8 | 1 | 2 |
| F508del homozygous genotype | 6 (55) | 4 | 1 | 1 |
| F508del heterozygous genotype | 5 (45) | 4 | … | 1 |
| Age, y | 30.1 (26.7–32.1) | 30.3 (28.1–31.8) | 26 | 36 (28.8–43.2) |
| Female:male | 5:6 | 4:4 | 0:1 | 1:1 |
| Best FEV1pp 12 mo prior | 39.8 (35.9–42.9) | 41.7 (37.1–43.3) | 39 | 32.3 (28.6–36.0) |
| Best BMI 12 mo prior, kg/m2 | 20.8 (19.7–23.6) | 20.2 (19–21.5) | 37 | 29.5 (25.9–33.0) |
| Sweat chloride, mmol/Lb | 95 (85–100) | 97 (87–100.5) | 85 | 90 |
| No. of exacerbations 12 mo prior | 3 (2–4) | 3 (2–4) | 1 | 4 (3–5) |
| Previous CFTR modulator use | 5 (45) | 4 | 1 | … |
| No. of patients with pulmonary colonization | ||||
| Mucoid P. aeruginosac | 5 | 5 | … | … |
| Nonmucoid P. aeruginosa | 10 | 7 | 1 | 2 |
| Staphylococcus aureus | 1 | … | … | |
Abbreviations: BMI, body mass index; ETI, elexacaftor-tezacaftor-ivacaftor; FEV1pp, forced expiratory volume in 1 second percent predicted; P. aeruginosa, Pseudomonas aeruginosa; TI, tezacaftor-ivacaftor.
aBaseline characteristics are prior to naive TI commencement.
bBaseline sweat chloride was not available for all patients (n = 9).
cTwo additional patients had cultured mucoid P. aeruginosa in previous recent sputum samples prior to study isolation (not included in table).
Clinical Improvements After CFTR Modulation
Clinical outcomes in the 12 months prior to and 12 months after CFTR modulation were compared. In patients receiving ETI, there was a median 5.1% increase in lung function (FEV1pp; Table 2, Supplementary Figure 1A). The median BMI increased from 20.5 to 22.7 kg/m2 (Supplementary Figure 1B). Median sweat chloride fell by 51.5 mmol/L, with all except three patients dropping below the diagnostic level for CF (>60 mmol/L; Supplementary Figure 1C), and the median number of pulmonary exacerbations decreased from three to zero (Supplementary Figure 1D). Similar improvements in clinical parameters were seen in people who transitioned to ETI from an alternate modulator as compared with those commencing ETI for the first time. Of three participants receiving TI, two had a 2.5% and 2.9% increase in FEV1pp but a decrease in BMI and variable change in pulmonary exacerbations. One TI participant (CFP9; age, 50 years) deteriorated throughout the study duration. Overall, we observed that all except one participant had significant improvements across multiple clinical parameters after CFTR modulation.
Table 2.
Median Changes in Clinical Parameters From the 12 Months Prior to the 12 Months After CFTR Modulator Therapy in People With Cystic Fibrosis
| Best FEV1pp in 12 mo | Best BMI in 12 mo, kg/m2 | Sweat Chloride, mmol/L (n = 9)a | No. of Exacerbations in 12 mo | |||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Prior | After | Median Change | Prior | After | Median Change | Baseline | After | Median Change | Prior | After | Median Change | |
| All ETI (n = 10)b | ||||||||||||
| Median | 40.7 | 46.4 | 5.1 | 20.5 | 22.7 | 1.1 | 96.0 | 44.5 | –51.5 | 3.5 | .0 | –2.5 |
| IQR | 34.8–43.0 | 42.0–47.7 | 2.3–6.2 | 19.4–23.0 | 20.6–23.8 | 0.4–2.1 | 86.5–99.5 | 37.8–56.8 | –56.3 to –18.8 | 2.2–4.8 | 0–0 | –4 to –1.3 |
| P value | .0098 | .0391 | .0078 | .0039 | ||||||||
| Naive-ETI (n = 4) | ||||||||||||
| Median | 41.1 | 46.4 | 2.5 | 21.9 | 23.1 | 1.8 | 95.0 | 40.0 | –53.0 | 2.5 | 0.0 | –2.0 |
| IQR | 37.1–44.6 | 44.9–46.7 | 1.7–5.5 | 19.1 –23.6 | 21.5–24.3 | 1.2–2.5 | 87–96 | 33–57 | –55 to –37 | 2–3.3 | 0–0.8 | –2.5 to –1.5 |
| P value | .125 | .125 | .25 | .25 | ||||||||
| LI/TI-ETI (n = 6)b | ||||||||||||
| Median | 40.7 | 45.9 | 6.0 | 20.5 | 22.0 | 0.6 | 99.0 | 46.0 | –50.0 | 4.5 | 0.0 | –3.5 |
| IQR | 32.4–42.3 | 38.2–48.2 | 4.7–6.2 | 19.4–21.2 | 20.6–23.3 | 0.1–1.7 | 89–101 | 43–51 | –56 to –12 | 3.3–5 | 0–0 | –4.75 to –1.5 |
| P value | .0625 | .3125 | .0625 | .0312 | ||||||||
| All TI (n = 3) | ||||||||||||
| Median | 39.3 | 41.7 | 2.5 | 36.5 | 36.2 | –0.4 | 90.0 | 89.0 | –1.0 | 2.0 | 3.0 | –1.0 |
| IQR | 32.1–39.5 | 31.2–42.2 | –0.9 to 2.7 | 29.5–36.8 | 28.8–36.4 | –0.7 to −0.4 | 90–90 | 89–89 | –1 to –1 | 1.5–4 | 1.5–4 | 2 to –1 |
| P value | NT | NT | NT | NT | ||||||||
A Wilcoxon matched-pairs signed rank test was performed comparing clinical parameters before and after CFTR modulator commencement. Bold values indicate P < .05.
Abbreviations: BMI, body mass index; CFP, cystic fibrosis participant; CFTR, cystic fibrosis transmembrane conductance regulator; ETI, elexacaftor-tezacaftor-ivacaftor; FEV1pp, forced expiratory volume in 1 second percent predicted; LI, lumacaftor-ivacaftor; NT, no test performed due to small sample size; TI, tezacaftor-ivacaftor.
aEither no sweat chloride data at baseline or no matched postmeasurements were available for four patients (CFP 4, 6, 10, and 11).
bTwo patients started the study taking TI and later transitioned to ETI; therefore, their clinical parameters were represented twice and included in both the ETI and TI groups accordingly.
Clonal P. aeruginosa Persists After CFTR Commencement
To determine whether participants were colonized with the same clonal P. aeruginosa lineages after CFTR modulation, we sequenced all 105 P. aeruginosa strains and compared their sequence similarity and MLST profile that differentiated strains based on variation within the housekeeping genes to determine clonality before and after treatment. At baseline, six participants were colonized with a single MLST while five participants were colonized with two P. aeruginosa sequence types (Figure 2). P. aeruginosa MLST 801 was shared among seven participants and MLST 775 between two participants, which both corresponded to the profile of two endemic strains (AUST-O6, AUST-O2) circulating in Australian CF centers [5, 6]. Sequence types 649 (AUST-O1), 822 (AUST-11), and 262 (AUST-O7) have also been isolated frequently from the same CF centre as this study [6].
Figure 2.
Clonal Pseudomonas aeruginosa sequence types persist up to 21 months after CFTR modulator commencement in people with cystic fibrosis. Bacteria were isolated at baseline and various time points after CFTR modulator initiation. Each bacterial strain isolated is represented by a circle, and the color indicates the multilocus sequence type (MLST). Circles with three intersecting lines indicate that the isolate was of mucoid morphology. Shaded background indicates the CFTR modulator therapy characteristics. Cystic fibrosis participants (CFP 9 and 10) who commenced tezacaftor-ivacaftor (TI) and then switched to elexacaftor-tezacaftor-ivacaftor (ETI) are indicated by the line intersecting the x-axis at approximately six and 12 months. Figure created in BioRender. CFTR, cystic fibrosis transmembrane conductance regulator; LI, lumacaftor-ivacaftor.
For all participants, at least one of the P. aeruginosa lineages with which they were initially colonized persisted throughout the entire study duration, as reflected in MLST and phylogeny (Figure 2, Supplementary Figure 2). Only two patients (CFP 6 and CFP 8) had a sequence type isolated at baseline (MLST 801, AUST-O6) that was not isolated again after CFTR modulator therapy. New sequence types were isolated from two participants (CFP 4 and CFP 10) after 12 months of therapy. The two sequence types isolated posttreatment from CFP 4 (MLST 4054 and 4055) had the same central genotype that differed in variation at only two of seven loci compared with MLST 775 isolated at baseline, and therefore clustered in the same clonal complex (Supplementary Figures 2 and 3). In contrast, the new sequence type from CFP 10 (MLST 4056) did not cluster with other baseline strains and had a unique O-antigen serotype representing a novel strain.
Genetic Adaption of P. aeruginosa After CFTR Modulation
Next, we wanted to determine any indication of adaptation to the new modulated lung environment by identifying mutations unique to clonal strains isolated after CFTR modulation. Fourteen clonal lineages from 11 participants persisted after CFTR modulation and accumulated 19 to 3400 mutations (nonsynonymous, modifier microindel, or frameshift) within 3565 genes in the PAO1 reference genome (Table 3). In 11 of 14 lineages, mutations were more frequent in previously described pathoadaptive genes identified in longitudinal P. aeruginosa isolates in pwCF [2, 4, 22–26]. In contrast, we observed only four lineages having mutations occurring more often in genes involved in antibiotic resistance [27–29].
Table 3.
Mutations Accumulated in Lineages After CFTR Modulation in People With Cystic Fibrosis
| No. Isolates Analyzed | Mutations Unique to Isolates Post–CFTR Modulationa | No. of Mutated Genes | No. of Mutated Genes | ||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Patient_MLST | Time Since Modulation, mo | Baseline | After | Total | SNP | Indel | Unique genes | Pathoadaptive (n = 159) | Other | P Value | Antibiotic Resistant (n = 185) | Other | P Value |
| CFP1_801 | 18 | 3 | 1 | 90 | 69 | 21 | 72 | 4 | 68 | .1398 | 3 | 69 | .5062 |
| CFP1_775b | 18 | 1 | 5 | 418 | 276 | 142 | 354 | 50 | 304 | <.0001 | 26 | 328 | <.0001 |
| CFP2_822 | 21 | 2 | 1 | 202 | 142 | 60 | 190 | 9 | 181 | .1104 | 6 | 184 | >.9999 |
| CFP3_801 | 16 | 2 | 2 | 92 | 59 | 33 | 84 | 10 | 74 | <.0001 | 2 | 82 | >.9999 |
| CFP4_775 | 17 | 3 | 14 | 3400 | 2652 | 748 | 2176 | 100 | 2076 | <.0001 | 85 | 2091 | .0309 |
| CFP5_262 | 14 | 2 | 2 | 82 | 49 | 33 | 72 | 7 | 65 | .0037 | 1 | 71 | .7310 |
| CFP6_260 | 16 | 2 | 1 | 19 | 11 | 8 | 18 | 2 | 16 | .0880 | 0 | 18 | >.9999 |
| CFP7_262 | 8 | 3 | 1 | 183 | 148 | 35 | 178 | 15 | 163 | .0001 | 6 | 172 | .8294 |
| CFP7_801 | 8 | 1 | 2 | 1042 | 816 | 226 | 629 | 42 | 587 | <.0001 | 30 | 599 | .0305 |
| CFP8_829 | 20 | 1 | 4 | 664 | 437 | 227 | 568 | 55 | 513 | <.0001 | 26 | 542 | .0611 |
| CFP9_845 | 13 | 2 | 5 | 1711 | 1342 | 369 | 1282 | 72 | 1210 | <.0001 | 54 | 1228 | .0311 |
| CFP9_801 | 13 | 1 | 15 | 522 | 357 | 165 | 444 | 30 | 414 | <.0001 | 20 | 424 | .1234 |
| CFP10_1637 | 15 | 2 | 6 | 70 | 27 | 43 | 54 | 6 | 48 | .0036 | 3 | 51 | .2541 |
| CFP11_801 | 12 | 3 | 3 | 122 | 86 | 36 | 112 | 11 | 101 | .0003 | 4 | 108 | .7844 |
P values represent a Fisher's exact test of the number of mutated pathoadaptive genes or antibiotic-resistant genes (Supplementary Tables 5 and 6) in each lineage versus other genes mutated within the genome. Bold values indicate P < .05.
Abbreviations: CFP, cystic fibrosis participant; CFTR, cystic fibrosis transmembrane conductance regulator; MLST, multilocus sequence type; SNP, single-nucleotide polymorphism.
aNonsynonymous and modifier indel or frameshift mutations.
bThis lineage includes MLSTs 775, 4054, and 4055.
To determine if the new lung environment was selecting for mutations in specific genes, we identified mutations in genes shared across a large proportion of lineages. We found that 37 genes were mutated in >42% (6/14) of lineages (Figure 3A, Supplementary Table 4). Of these genes, 13 (35%) were previously described as pathoadaptive and three were associated with antibiotic resistance. Lineages that clustered closely in number of genes commonly mutated accumulated a large amount of overall mutations (>400) after CFTR modulation and had significantly more mutations in mutS, mutL, and urvD genes responsible for DNA mismatch repair, which have been shown to cause hypermutation [30]. The two genes mutated in the most lineages (9/14, 64%) were PA1874, which encodes for a hypothetical protein previously described as pathoadaptive, and cdrA, an extracellular matrix protein [31]. The functional class of “secreted factors (toxins, enzymes, alginate)” had the highest frequency of mutated genes relative to the number of genes in the corresponding class, including algG, lepA, pchE, cdrA, and tle5b [31] (Figure 3B). Other classes with high frequency of mutated genes were “cell wall/lipopolysaccharide/capsule” and “adaption, protection” which contained additional commonly mutated genes, such as mucA, flgK, pslA, pelA, pcoA, ppkA, pvdA, and pvdL. As all these genes were mutated in multiple lineages, they may represent selection of mutations driven by the new modulated lung environment.
Figure 3.
Genes mutated after CFTR modulation in >42% of lineages from people with cystic fibrosis. A, Presence-and-absence heat map of 37 commonly mutated genes (≥6/14, 42%) with at least one nonsynonymous modifier indel or frameshift mutation accumulated in lineages after CFTR modulation in people with cystic fibrosis. Heat map is annotated with gene length and whether the gene was previously described as being pathoadaptive or involved in antibiotic resistance. Lineages are shaded by CFTR modulator therapy characteristics. B, PseudoCAP functional class of 37 commonly mutated genes (number and percentage of genes relative to number of genes in class). Figure 3A created in R Studio and BioRender. CFTR, cystic fibrosis transmembrane conductance regulator; ETI, elexacaftor-tezacaftor-ivacaftor; LI, lumacaftor-ivacaftor; LPS, lipopolysaccharide; TI, tezacaftor-ivacaftor.
Chronic P. aeruginosa Phenotypes Are Sustained
Last, we wanted to determine whether classically chronic P. aeruginosa phenotypes were sustained after CFTR modulation. To do this, baseline isolate characteristics were compared with isolates from the latest time point for each participant. There were no significant changes observed in the proportion of isolates with mucoid morphology (Figure 2), O-antigen expression, and antibiotic resistance (Figure 4). Slight changes were due to small differences in the number of strains isolated per participant at each time point. Between 57% and 90% of isolates were resistant to antibiotics routinely used to treat this cohort of participants, including tobramycin, ceftazidime, aztreonam, meropenem, cefepime, and piperacillin-tazobactam. Five participants were colonized with at least 1 isolate that was resistant to all antibiotics tested except colistin. Thus, pwCF remain colonized with clonal lineages of P. aeruginosa with the same chronic phenotypes that are notoriously difficult to manage clinically.
Figure 4.
Chronic Pseudomonas aeruginosa phenotypes are sustained after CFTR modulation in people with cystic fibrosis. P. aeruginosa (A) mucoid colony morphology, (B) O-antigen expression, and (C) antibiotic resistance at baseline and the latest time point available after CFTR modulator therapy. Resistant, resistant to antibiotics; intermediate, sensitive, although increased antibiotic dosing is advised; sensitive, completely sensitive to antibiotics. No significant differences were found. CFTR, cystic fibrosis transmembrane conductance regulator.
DISCUSSION
P. aeruginosa is a problematic pathogen associated with increased morbidity and mortality in pwCF [32]. While previous studies have demonstrated that CFTR modulators can drive decreases in P. aeruginosa bacterial load in sputum, complete eradication is not achieved [15, 16]. In our cohort of 11 pwCF with advanced lung disease and chronic P. aeruginosa infection, we demonstrated that despite improvements in health, the same clonal P. aeruginosa isolates persisted up to 21 months after ETI or TI commencement. Mutations unique to strains isolated after CFTR modulation were more often in previously described pathoadaptive genes in 11 of 14 lineages, while only four lineages accumulated mutations more frequently in antibiotic-resistant genes. Additionally, 37 genes were identified to be mutated across a sizable proportion of lineages; however, chronic P. aeruginosa phenotypes were sustained, such as mucoid morphology, lack of O-antigen expression, and antibiotic resistance.
Multiple clinical improvements were observed in our study, although increases in lung function were modest as compared with those seen in clinical trials. Average increases of up to 13.9% in FEV1pp were reported in people receiving ETI [10, 11, 33] and up to 6.8% in people receiving TI [34, 35], where such increases were often as significant regardless of previous modulator treatment [11, 33]. The modest increases in lung function seen in our study are likely attributed to participants having more severe lung disease at baseline when compared with patients in pivotal clinicals. Additionally, the decision to select the “best” FEV1pp in the 12 months prior to commencement for comparison with postcommencement values may not recognize deterioration within the year before CFTR modulator commencement; therefore, less drastic increases in lung function may have been observed. Nevertheless, the large reduction in exacerbation frequency and sweat chloride and increase in BMI posttreatment were comparable to literature [10, 33] and suggest significant clinical improvement in this cohort after CFTR modulator commencement.
Investigations evaluating changes in CF lung microbiology after ETI treatment are limited to a number of relatively small studies that often lack consistency due to cohort demographics. These studies have reported increases in lung microbiota alpha diversity and decreases in P. aeruginosa relative abundance and load to varying levels that imply a shift in bacterial ecology [17, 33, 36, 37]. However, all studies consistently show that P. aeruginosa is not eradicated in all or a large proportion of patients [33, 36], as corroborated in observational studies reporting that 54% to 77% of pwCF remain P. aeruginosa culture positive up to 1 year of ETI modulation [14, 38, 39].
In this study, we have further demonstrated that not only do pwCF remain colonized with P. aeruginosa after CFTR modulation but the same clonal lineages persist, as opposed to the eradication of preexisting strains and establishment of new infections [7]. From our knowledge, only 1 other investigation has reported the persistence of P. aeruginosa lineages after 2 years of treatment with the first available modulator ivacaftor for treatment of G551D mutations [16]. Despite those authors performing a more comprehensive analysis of MLST frequencies at a population level, our findings of sequence type persistence are consistent with their study.
A likely explanation for P. aeruginosa persistence is irreversible structural damage that continues to cause defective pathogen clearance, especially in our cohort of pwCF with advanced lung disease. A study of 13 pwCF with severe lung disease reported reductions in pulmonary damage and bronchial destruction, yet signs of structural damage in the airways, such as scarring lesions and bronchiectasis, remained 1 year after treatment with ETI [38]. Individuals with other structural lung conditions, including non-CF bronchiectasis and chronic obstructive pulmonary disease, are commonly colonized by P. aeruginosa [40, 41]. Additionally, there have been small incidences where P. aeruginosa is acquired after CFTR modulator commencement, suggesting that lung damage remains a risk factor for infection [15, 17].
P. aeruginosa is well known for its hypermutable nature and ability to adapt to the harsh CF lung environment [30]. While structural damage is often sustained [38], multiple studies have reported important changes in the CF lung after modulation, such as changes in sputum load, viscosity, metabolomic composition, and pH [17, 36]. We found that P. aeruginosa accumulated mutations more frequently in previously described pathoadaptive genes but less often in antibiotic-resistant genes after CFTR modulation. This may indicate continual mutation of common chronic genotypes in these heterogenetic P. aeruginosa populations, as well as retention and lack of accumulation of additional antibiotic-resistant mechanisms due to reduced antibiotic exposure. The identification of genes mutated across a large proportion of lineages may indicate selection of adaptive mutants. Key genes were identified that play important roles in bacterial aggregation, alginate production, and biofilm formation: cdrA, ppkA, mucA, pslA, pelA, and algG [31, 42]. Genes encoding secreted effectors, including tle5b and lepA, were identified, as well as pvdA, pvdL, and pchE, which contribute to pyoverdine synthesis and iron acquisition [31]. Mutation of these genes may allow for continued survival and competition in the modulated CF lung. Despite the potential impact of the mutations reported, chronic P. aeruginosa phenotypes were sustained after CFTR modulation, including mucoid morphology for promoting biofilm formation, lack of O-antigen expression for immune evasion, and antibiotic resistance. Therefore, these P. aeruginosa lineages are not phenotypically changing or reverting back to more virulent acute phenotypes in this changed lung environment within this study time frame.
Limitations to this study include the small cohort size and the inconsistency of time points and duration, which were attributed to the coronavirus pandemic, where participants did not come to outpatient clinics consistently and research sampling could not always be undertaken. Only spontaneous sputum samples were collected from pwCF receiving compassionate access ETI, which resulted in this cohort including individuals with more advanced lung disease. Future studies with induced sputum sampling will allow for pathogen surveillance on a larger proportion of pwCF with varying disease severities, allowing conclusions to be drawn on P. aeruginosa persistence after CFTR modulation from a more representative cohort. Larger multicenter as well as international studies will further determine whether our findings are exclusive to common endemic strains found in Australia or relevant to all P. aeruginosa infections in pwCF after CFTR modulation.
One scientist performed bacterial isolations from CF sputum to limit variability in P. aeruginosa isolation, yet P. aeruginosa populations may not have been completely captured by taking only morphologically distinct colonies per time point. A recent study showed that 75 P. aeruginosa colonies from the same CF sputum sample can have between 31 and 4592 unique single-nucleotide polymorphisms within a population, highlighting the genomic diversity of P. aeruginosa and limitations in single-isolate sampling [43]. Therefore, collection of more P. aeruginosa colonies per time point are required to make more robust conclusions on P. aeruginosa adaptation and to investigate the impact of these mutations further. Despite this, we were still able to show that dominating clonal lineages with the same chronic phenotypes were persisting after CFTR modulation.
Despite the small sample size, the findings in this study have given important insights into the nature of chronic P. aeruginosa infection in pwCF with advanced lung disease after CFTR modulation. Persistence of the same chronic clonal lineages may indicate that initial improvements in health and systemic inflammation may slow lung disease progression; however, existing structural damage will allow for infection persistence and continued lung injury. Therefore, future research and clinical care should remain focused on P. aeruginosa as well as other lung infections in this post–modulator era. CFTR modulators will be instrumental for younger pwCF to prevent pathogen colonization and subsequent progressive lung disease.
Supplementary Data
Supplementary materials are available at The Journal of Infectious Diseases online (http://jid.oxfordjournals.org/). Supplementary materials consist of data provided by the author that are published to benefit the reader. The posted materials are not copyedited. The contents of all supplementary data are the sole responsibility of the authors. Questions or messages regarding errors should be addressed to the author.
References
- 1. Elborn JS. Cystic fibrosis. Lancet 2016; 388:2519–31. [DOI] [PubMed] [Google Scholar]
- 2. Smith EE, Buckley DG, Wu Z, et al. Genetic adaptation by Pseudomonas aeruginosa to the airways of cystic fibrosis patients. Proc Natl Acad Sci U S A 2006; 103:8487–92. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3. Mayer-Hamblett N, Ramsey BW, Kulasekara HD, et al. Pseudomonas aeruginosa phenotypes associated with eradication failure in children with cystic fibrosis. Clin Infect Dis 2014; 59:624–31. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4. Marvig RL, Johansen HK, Molin S, Jelsbak L. Genome analysis of a transmissible lineage of Pseudomonas aeruginosa reveals pathoadaptive mutations and distinct evolutionary paths of hypermutators. PLoS Genet 2013; 9:e1003741. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5. Kidd TJ, Ramsay KA, Hu H, et al. Shared Pseudomonas aeruginosa genotypes are common in Australian cystic fibrosis centres. Eur Respir J 2013; 41:1091–100. [DOI] [PubMed] [Google Scholar]
- 6. Tai AS, Sherrard LJ, Kidd TJ, et al. Antibiotic perturbation of mixed-strain Pseudomonas aeruginosa infection in patients with cystic fibrosis. BMC Pulm Med 2017; 17:138. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7. Williams D, Fothergill JL, Evans B, et al. Transmission and lineage displacement drive rapid population genomic flux in cystic fibrosis airway infections of a Pseudomonas aeruginosa epidemic strain. Microb Genom 2018; 4:e000167. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8. Kosorok MR, Zeng L, West SE, et al. Acceleration of lung disease in children with cystic fibrosis after Pseudomonas aeruginosa acquisition. Pediatr Pulm 2003; 32:277–87. [DOI] [PubMed] [Google Scholar]
- 9. Ahern S, Salimi F, Caruso M, et al. Australian Cystic Fibrosis Data Registry: annual report 2021. Melbourne: Monash University, 2022. [Google Scholar]
- 10. Middleton PG, Mall MA, Dřevínek P, et al. Elexacaftor-tezacaftor-ivacaftor for cystic fibrosis with a single Phe508del allele. N Engl J Med 2019; 381:1809–19. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11. Heijerman HGM, McKone EF, Downey DG, et al. Efficacy and safety of the elexacaftor plus tezacaftor plus ivacaftor combination regimen in people with cystic fibrosis homozygous for the F508del mutation: a double-blind, randomised, phase 3 trial. Lancet 2019; 394:1940–8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12. Sheikh S, Britt RD Jr, Ryan-Wenger NA, et al. Impact of elexacaftor-tezacaftor-ivacaftor on bacterial colonization and inflammatory responses in cystic fibrosis. Pediatr Pulmonol 2023; 58:825–33. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13. Heltshe SL, Mayer-Hamblett N, Burns JL, et al. Pseudomonas aeruginosa in cystic fibrosis patients with G551D-CFTR treated with ivacaftor. Clin Infect Dis 2014; 60:703–12. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14. Gushue C, Eisner M, Bai S, et al. Impact of elexacaftor-tezacaftor-ivacaftor on lung disease in cystic fibrosis. Pediatr Pulmonol 2023; 58:2308–16. [DOI] [PubMed] [Google Scholar]
- 15. Nichols DP, Morgan SJ, Skalland M, et al. Pharmacologic improvement of CFTR function rapidly decreases sputum pathogen density, but lung infections generally persist. J Clin Invest 2023; 133:e167957. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16. Hisert KB, Heltshe SL, Pope C, et al. Restoring cystic fibrosis transmembrane conductance regulator function reduces airway bacteria and inflammation in people with cystic fibrosis and chronic lung infections. Am J Respir Crit Care Med 2017; 195:1617–28. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17. Sosinski LM, CM H, Neugebauer KA, et al. A restructuring of microbiome niche space is associated with elexacaftor-tezacaftor-ivacaftor therapy in the cystic fibrosis lung. J Cyst Fibros 2022; 21:996–1005. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18. Pressler T, Bohmova C, Conway S, et al. Chronic Pseudomonas aeruginosa infection definition: EuroCareCF Working Group report. J Cyst Fibros 2011; 10:S75–8. [DOI] [PubMed] [Google Scholar]
- 19. Jolley KA, Bray JE, Maiden MCJ. Open-access bacterial population genomics: BIGSdb software, the PubMLST.org website and their applications. Wellcome Open Res 2018; 3:124. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20. Thrane SW, Taylor VL, Lund O, Lam JS, Jelsbak L. Application of whole-genome sequencing data for O-specific antigen analysis and in silico serotyping of Pseudomonas aeruginosa isolates. J Clin Microbiol 2016; 54:1782–8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21. Stover CK, Pham XQ, Erwin AL, et al. Complete genome sequence of Pseudomonas aeruginosa PAO1, an opportunistic pathogen. Nature 2000; 406:959–64. [DOI] [PubMed] [Google Scholar]
- 22. Marvig RL, Sommer LM, Molin S, Johansen HK. Convergent evolution and adaptation of Pseudomonas aeruginosa within patients with cystic fibrosis. Nat Genet 2015; 47:57–64. [DOI] [PubMed] [Google Scholar]
- 23. Feliziani S, Marvig RL, Luján AM, et al. Coexistence and within-host evolution of diversified lineages of hypermutable Pseudomonas aeruginosa in long-term cystic fibrosis infections. PLoS Genet 2014; 10:e1004651. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24. Sommer LM, Alanin MC, Marvig RL, et al. Bacterial evolution in PCD and CF patients follows the same mutational steps. Sci Rep 2016; 6:28732. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25. Fischer S, Klockgether J, Gonzalez Sorribes M, Dorda M, Wiehlmann L, Tümmler B. Sequence diversity of the Pseudomonas aeruginosa population in loci that undergo microevolution in cystic fibrosis airways. Access Microbiol 2021; 3:000286. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26. Eklöf J, Misiakou MA, Sivapalan P, et al. Persistence and genetic adaptation of Pseudomonas aeruginosa in patients with chronic obstructive pulmonary disease. Clin Microbiol Infect 2022; 28:990–5. [DOI] [PubMed] [Google Scholar]
- 27. López-Causapé C, Sommer LM, Cabot G, et al. Evolution of the Pseudomonas aeruginosa mutational resistome in an international cystic fibrosis clone. Sci Rep 2017; 7:5555. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28. Colque CA, Albarracín Orio AG, Feliziani S, et al. Hypermutator Pseudomonas aeruginosa exploits multiple genetic pathways to develop multidrug resistance during long-term infections in the airways of cystic fibrosis patients. Antimicrob Agents Chemother 2020; 64:e02142–19. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29. Cabot G, Zamorano L, Moyà B, et al. Evolution of Pseudomonas aeruginosa antimicrobial resistance and fitness under low and high mutation rates. Antimicrob Agents Chemother 2016; 60:1767–78. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30. Hall KM, Pursell ZF, Morici LA. The role of the Pseudomonas aeruginosa hypermutator phenotype on the shift from acute to chronic virulence during respiratory infection. Front Cell Infect Microbiol 2022; 12:943346. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31. Winsor GL, Griffiths EJ, Lo R, Dhillon BK, Shay JA, Brinkman FS. Enhanced annotations and features for comparing thousands of Pseudomonas genomes in the Pseudomonas genome database. Nucleic Acids Res 2016; 44:D646–53. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32. Kerem E, Corey M, Gold R, Levison H. Pulmonary function and clinical course in patients with cystic fibrosis after pulmonary colonization with Pseudomonas aeruginosa. J Pediatr 1990; 116:714–9. [DOI] [PubMed] [Google Scholar]
- 33. Nichols DP, Paynter AC, Heltshe SL, et al. Clinical effectiveness of elexacaftor/tezacaftor/ivacaftor in people with cystic fibrosis: a clinical trial. Am J Respir Crit Care Med 2022; 205:529–39. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34. Rowe SM, Daines C, Ringshausen FC, et al. Tezacaftor-Ivacaftor in residual-function heterozygotes with cystic fibrosis. N Engl J Med 2017; 377:2024–35. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35. Taylor-Cousar JL, Munck A, McKone EF, et al. Tezacaftor-Ivacaftor in patients with cystic fibrosis homozygous for Phe508del. N Engl J Med 2017; 377:2013–23. [DOI] [PubMed] [Google Scholar]
- 36. Schaupp L, Addante A, Völler M, et al. Longitudinal effects of elexacaftor/tezacaftor/ivacaftor on sputum viscoelastic properties, airway infection and inflammation in patients with cystic fibrosis. Eur Respir J 2023; 62:2202153. [DOI] [PubMed] [Google Scholar]
- 37. Pallenberg ST, Pust M-M, Rosenboom I, et al. Impact of elexacaftor/tezacaftor/ivacaftor therapy on the cystic fibrosis airway microbial metagenome. Microbiol Spectr 2022; 10:e0145422. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 38. Migliorisi G, Collura M, Ficili F, et al. Elexacaftor-tezacaftor-ivacaftor as a final frontier in the treatment of cystic fibrosis: definition of the clinical and microbiological implications in a case-control study. Pharmaceuticals (Basel) 2022; 15:606. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 39. Schnell A, Hober H, Kaiser N, et al. Elexacaftor-tezacaftor-ivacaftor treatment improves systemic infection parameters and Pseudomonas aeruginosa colonization rate in patients with cystic fibrosis a monocentric observational study. Heliyon 2023; 9:e15756. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 40. Eklöf J, Sørensen R, Ingebrigtsen TS, et al. Pseudomonas aeruginosa and risk of death and exacerbations in patients with chronic obstructive pulmonary disease: an observational cohort study of 22 053 patients. Clin Microbiol Infect 2020; 26:227–34. [DOI] [PubMed] [Google Scholar]
- 41. Finch S, McDonnell MJ, Abo-Leyah H, Aliberti S, Chalmers JD. A comprehensive analysis of the impact of Pseudomonas aeruginosa colonization on prognosis in adult bronchiectasis. Ann Am Thorac Soc 2015; 12:1602–11. [DOI] [PubMed] [Google Scholar]
- 42. Jennings LK, Dreifus JE, Reichhardt C, et al. Pseudomonas aeruginosa aggregates in cystic fibrosis sputum produce exopolysaccharides that likely impede current therapies. Cell Rep 2021; 34:108782. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 43. Vanderwoude J, Azimi S, Read T, Diggle SP. The role of hypermutation and collateral sensitivity in antimicrobial resistance diversity of Pseudomonas aeruginosa populations in cystic fibrosis lung infection. mBio 2024; 2:e03109-23. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Data Availability Statement
All P. aeruginosa genomes are available on NCBI BioProject PRJNA1023362 (see Supplementary Table 1 for corresponding BioSample accession numbers). All isolate phenotypic data are available in Supplementary Table 2, clinical data in Supplementary Table 3, and all filtered mutations in Supplementary Data 2 and 3.