Molecular epidemiology and genomic dynamics of Pseudomonas aeruginosa isolates causing relapse infections

ABSTRACT Pseudomonas aeruginosa (P. aeruginosa) is one of the leading causes of chronic infections, including reinfection, relapse, and persistent infection, especially in cystic fibrosis patients. Relapse P. aeruginosa infections are more harmful because of repeated hospitalization and undertreatment of antimicrobials. However, relapse P. aeruginosa infection in China remains largely unknown. Herein, we performed a 3-year retrospective study from 2019 to 2022 in a tertiary hospital, which included 442 P. aeruginosa isolates from 196 patients. Relapse infection was identified by screening clinical records and whole-genome sequencing (WGS). We found that 31.6% (62/196) of patients had relapsed infections. The relapse incidence of carbapenem-resistant P. aeruginosa infection (51.4%) is significantly higher than that of carbapenem-susceptible P. aeruginosa infection (20.2%, P < 0.0001). These isolates were assigned to 50 distinct sequence types and sporadically distributed in phylogeny, indicating that relapsed infections were not caused by certain lineages. Fast adaptation and evolution of P. aeruginosa isolates were reflected by dynamic changes of antimicrobial resistance, gene loss and acquisition, and single-nucleotide polymorphisms during relapse episodes. Remarkably, a convergent non-synonymous mutation that occurs in a pyochelin-associated virulence gene fptA (T1056C, M252T) could be a considerable target for the diagnosis and treatment of relapse P. aeruginosa infection. These findings suggest that integrated utilization of WGS and medical records provides opportunities for improved diagnostics of relapsed infections. Continued surveillance of the genomic dynamics of relapse P. aeruginosa infection will generate further knowledge for optimizing treatment and prevention in the future. IMPORTANCE Pseudomonas aeruginosa is a predominant pathogen that causes various chronic infections. Relapse infections promote the adaptation and evolution of antimicrobial resistance and virulence of P. aeruginosa, which obscure evolutionary trends and complicate infection management. We observed a high incidence of relapse P. aeruginosa infection in this study. Whole-genome sequencing (WGS) revealed that relapse infections were not caused by certain lineages of P. aeruginosa isolates. Genomic dynamics of relapse P. aeruginosa among early and later stages reflected a plasticity scattered through the entire genome and fast adaptation and genomic evolution in different ways. Remarkably, a convergent evolution was found in a significant virulence gene fptA, which could be a considerable target for diagnosis and treatment. Taken together, our findings highlight the importance of longitudinal surveillance of relapse P. aeruginosa infection in China since cystic fibrosis is rare in Chinese. Integrated utilization of WGS and medical records provides opportunities for improved diagnostics of relapse infections.


Subject inclusion and exclusion criteria
This study was approved by the Guangdong Provincial Hospital of Chinese Medicine, Guangzhou, China.The clinical isolates were routinely collected and whole-genome sequenced in the hospital.To identify the subjects with P. aeruginosa infections, we screened the clinical records of subjects from 1 September 2019 to 31 December 2022.The subjects were included if clinical samples were evidenced by a positive P. aeruginosa culture, and the patients had typical clinical infectious symptoms and abnormal laboratory markers of infections (Table S1).The patients were excluded if the clinical samples were positive for the culture of P. aeruginosa but had no infectious symptoms, which were regarded as colonization.The source of infection was determined as pneumonia, urinary tract infection, surgical site infection, intra-abdominal infection, catheter-related infection, or bacteremia as defined by the Centers for Disease Control and Prevention (17).

Identification of relapse P. aeruginosa infections
Recurrence infection was defined as the return of P. aeruginosa infections after documen tation of negative cultures of clinical samples or clinical improvement after completing a course of antimicrobial therapy, according to previous studies (8,18,19).Therefore, all patients with recurrent episodes (≥2 episodes that occurred ≥30 days apart with clinical sample culture results negative for P. aeruginosa in the interim) were identified.Their medical records were reviewed to verify clearance (resolution of all clinical signs of infection) between recurrences and to determine the clinical characteristics of the episodes.Recurrence was further subclassified as reinfection or relapse infection: if genomes of P. aeruginosa isolates from the patient were closely related to and distributed into the same lineage of the phylogeny, the recurrence was considered as a relapse infection.Conversely, if the genomic contents are greatly different and the isolates were sporadically distributed on phylogeny, the recurrent episode was considered to be reinfection.

Isolation of P. aeruginosa strains
Clinical samples and cultured isolates were collected as part of clinical management and hospital surveillance in the Guangdong Provincial Hospital of Chinese Medicine.Clinical samples (urine, blood, sputum, and wound samples) from patients with infections were plated on Columbia Blood Agar (CBA) with 5% sheep blood (Luqiao, Beijing, China), except blood samples that were injected into blood culture bottles (BACT/ALERT FN/FA Plus) processing in the BacT/ALERT 3D System (bioMerieux Inc.).After the culture at 37°C in 24 hours or positive alarm of blood culture, if P. aeruginosa is identified as the most abundant species in clinical samples, three P. aeruginosa colonies were selected and stored.DNA was extracted using the boiling method.P. aeruginosa isolates were cultured by adding 3 mL of nutrient broth and incubating for 18-24 hours at 37°C.Subsequently, total DNA was obtained from the supernatant by boiling 1 mL of the broth at 100°C for 10-15 minutes.Species identification was first confirmed by MALDI-TOF MS (BrukerDaltonik GmbH, Bremen, Germany).Where the species could not be reliably assigned by MALDI-TOF, 16S rDNA sequencing was applied.After these processes, we randomly choose one representative P. aeruginosa isolate for the subsequent analysis including MIC testing and whole-genome sequencing.

Antimicrobial susceptibility testing
The antimicrobial susceptibilities of the P. aeruginosa isolates were determined by using the Vitek 2 Compact system (bioMérieux, France) according to the manufacturer's instructions.The minimum inhibitory concentrations (MICs) of cefepime, ceftazidime, imipenem, meropenem, ticarcillin-clavulanate, levofloxacin, ciprofloxacin, tobramycin, and amikacin were determined, and the results were interpreted according to Clini cal and Laboratory Standards Institute (Clinical and Laboratory Standards Institute, document M100-S29) guidelines.The MICs of colistin and polymyxin E (colistin) were determined using the broth dilution method (EUCAST breakpoints, version 9.0; CLSI M100-S29).The antibiotic susceptibilities of the P. aeruginosa isolates are provided in Table S2.

Whole-genome sequencing and data processing
A total of 163 P. aeruginosa isolates from 65 patients with relapse infections (identi fied using MALDI-TOF MS ± 16s rDNA sequencing) were subjected to whole-genome sequencing.The Qiagen Blood and Tissue Kit (Qiagen, Hilden, Germany) was used to extract genomic DNA for the P. aeruginosa isolate according to the manufacturer's instructions.DNA libraries were constructed with 350-bp paired-end fragments and sequenced using an Illumina HiSeq 2000 platform (Illumina Inc., USA).The sequencing yielded >100-fold coverage sequences (~1G) of raw paired-end (150 bp) per isolate.Adaptor, unreliable reads, and low-quality bases of raw reads were trimmed using fastp with default parameters (20).SPAdes v3.13.1 with the "--careful" parameter and automatically determined k-mer values (21,33,55,77) were used to perform draft genome de novo assembly (21).Contigs less than 500 bp in size were excluded for further analysis.WGS-based species identification was performed using JSpeciesWS v3.2.7 (22).
Isolates not confirmed as P. aeruginosa using WGS-based species identification, and contaminated/mixed sequences, were manually excluded from subsequent analyses.

Statistical analysis
Statistical analysis was conducted using R software v3.5.3.Statistical significance between groups was calculated using a two-sided Wilcoxon rank sum test and a Fisher's exact test.P values less than 0.05 were considered as statistically significant.

Identification of patients with relapse P. aeruginosa infections
A total of 442 non-duplicate P. aeruginosa isolates were identified in 196 patients at the Guangdong Provincial Hospital of Chinese Medicine from September 2019 to December 2022 (Table S1).Through screening of clinical records and inclusion criteria, we found that 33.2% (65/196) of patients had at least twice P. aeruginosa infections during the study period, which include relapsed infection and reinfection (Fig. 1).
We hypothesize that the relapse P. aeruginosa isolates should be closely related, meaning the differences in genomes should be minor and the isolates are distributed into the same lineage on phylogeny.To precisely identify the relapse P. aeruginosa infections, we performed WGS and phylogenetic analysis for 163 isolates from 65 patients; we identified that 156 isolates from 62 patients were distributed into the same lineage, whereas seven isolates from three patients were distributed sporadically (Fig. 1C), which indicated that 62 patients (95.4%, n = 65) had relapse infections and three patients (4.6%) had reinfections (Fig. 1A and B).
Virulence-associated genes, antimicrobial resistance genes, and plasmid of relapse P. aeruginosa isolates Among 156 isolates from patients with relapse P. aeruginosa infection, a total of 60 kinds of virulence-associated genes and 2 kinds of plasmid replicons were identified against VFDB and PlasmidFinder databases (Tables S4 and S5).The average number of virulence-associated genes for each isolate is 48.4 (standard deviation [SD]: ±3.7).A total of 45 isolates harbored IncFIA plasmid, and 12 isolates harbored IncQ plasmid.No VFs and plasmid replicons were lost or acquired among different time points of infections in all patients.A total of 25 ARGs were detected among relapse P. aeruginosa isolates (Table S3).The average number of ARGs for each isolate is 5.7 (SD: ±1.1).We found that 41 patients had the same ARGs among different time points of infections.Notably, the P. aeruginosa isolates of nine patients lost ARGs at the later stages of infections, such as aph(3')-IIa (n = 5) and aac (3)-IV (n = 5), while P. aeruginosa isolates of 10 patients acquired new ARGs at a later stage, such as catA1 (n = 7) and aac (3)-IV (n = 5).Acquisition and loss of the entire genes are important for evolution and adaptation during infection, which is commonly connected with horizontal transfer of mobile genes and transposons that can rapidly introduce large genomic and phenotypic changes, conferring antimicrobial resistance, virulence, or fitness advantage (12,13,35).

Genomic characterizations of relapse P. aeruginosa isolates
These relapse P. aeruginosa isolates were assigned into 50 distinct STs, of which 8 new STs were identified among 16 isolates (Fig. 4; Table S6).Notably, our results showed that four patients were infected with ST357 isolates and three patients with ST1182 isolates, while 40 STs were detected in only one patient and 8 STs in two patients, which indicated that the relapse infections were not caused by certain lineages of P. aeruginosa.
Pan-genome analysis of 156 relapse P. aeruginosa isolates from 62 patients identified 4,785 core genes representing a 4.5-Mb alignment in ≥99% of genomes.hierBAPS analysis of the relapse P. aeruginosa population structure based on 59,631 cgSNPs identified 10 sequence clusters (SCs), of which SC7 was polyphyletic in the ML phylogeny and bifurcated as two distinct monophyletic clades (Fig. 3).SC7 was the most prevalent cluster containing 46.8% (73/156) of isolates (Fig. 4).We observed that the relapse P. aeruginosa isolates were distributed into different branches, and the distribution of ARGs, VFs, and plasmid replicons was scattered, indicating no lineage specificity in relapse P. aeruginosa isolates (Fig. 3).

Pairwise genetic distance and evolution of relapse P. aeruginosa isolates
Previous studies demonstrated that micro-evolution and adaptation could happen in P. aeruginosa isolate under the selective pressure of antimicrobials and host immune systems, which could promote virulence and antimicrobial resistance of P. aeruginosa isolates (5,(12)(13)(14)(15) .Therefore, we performed a genomic comparative analysis to reveal the evolution of relapse P. aeruginosa isolates.The results of pairwise SNP distance of P. aeruginosa isolates for each patient showed that SNPs of the isolates from 15 patients occurred during early and subsequent stages of relapse infections (ranging from one to seven SNPs), whereas 47 patients occurred with no SNP among different time points (Fig. 5A).These SNPs were located on 25 sites of P. aeruginosa chromosome, and 24 of them are detected in only one patient.
Principal component analysis (PCA) showed that P. aeruginosa isolates among the first episode and later episodes were divided into two parts, which may reveal that there is a convergent evolution during the episodes of relapse infections (Fig. 5B).Remarkably, we found an SNP, which causes a non-synonymous mutation in the fptA gene (T1056C, M252T), was detected in relapse P. aeruginosa of six patients (Fig. 5C).FptA, a TonBdependent transporter, promotes the utilization of iron that permits the high-affinity binding and transport of Fe(III)-pyochelin complex across the outer membrane, which is associated with the virulence of P. aeruginosa (36,37).The mutation of the fptA gene has also been detected in patients with long-term P. aeruginosa infections, which indicated that convergent microevolution occurs in relapse P. aeruginosa isolates (38).

DISCUSSION
Relapse P. aeruginosa infection is responsible for repeated admissions and adverse outcomes of patients.In European and American countries, relapse P. aeruginosa infection commonly happened in CF patients.However, the prevalence and characteris tics of relapse P. aeruginosa infections in China remain unknown since CF is a rare disease in China (16).In this study, we performed a 3-year retrospective study to investigate the epidemiological and genomic characteristics of P. aeruginosa isolates causing relapse infections in a tertiary hospital in China.
Whole-genome sequencing is a precise method for recognition among relapse infection, persistent infection, and reinfection of a certain pathogen (6)(7)(8).Our results showed a high accuracy (95.4%, 62/65) in identifying relapse infections by screening clinical records using the standards in this study.However, the distinguishment between relapse and persistent infections relied on both the genotyping by WGS and clinical symptoms.In many previous studies, relapse and persistent infections were collectively analyzed since the clinical symptoms during the treatment periods were ignored, which underestimated the influences of entirely different dynamics of pathogen virulence and the host's immune status (10,11,14,(38)(39)(40).
Our results revealed that the incidence of relapse P. aeruginosa infection in this study (31.6%, 62/196) is high but relatively lower than the results from European and American countries, which could be attributable to the reason that CF patients were not found in our cohort, which is an important risk factor for colonization and infection of P. aeruginosa (13,41).Relapse P. aeruginosa infections are commonly detected in the lungs (91.9%, 57/62), indicating that the specific host lung environment favors P. aeruginosa colonization and persistence.Notably, 30.8% (20/65) of the patients were diagnosed with chronic lung diseases, but no direct indicator was found.Besides, we found that most of the patients are all elderly and with prolonged hospitalization which are significant risk factors for P. aeruginosa colonization (42).Therefore, we propose that decolonization could be an efficient therapy for the treatment and prophylaxis of relapse P. aeruginosa infection.
We revealed that the incidence of relapse in CRPA isolates (51.4%, 37/72) is signif icantly higher than that in CSPA isolates (20.2%, 25/124, P < 0.0001), which could be attributed to the reason that CRPA is commonly resistant to most of the antimi crobials and hard to eradicate completely (34).Decreased susceptibility of several antimicrobials was observed in 27% (7/63) of patients in the late stage of relapse, including carbapenems, β-lactams, and aminoglycosides.Besides, we found that two aminoglycoside resistance genes (aph(3')-IIa and aac(3)-IV) underwent dynamic events of loss or acquisition in the late stage of relapse infection.As previously confirmed, the acquisition of ARGs could confer resistance, whereas the loss of ARGs could confer a fitness advantage, which is important for evolution and adaptation during infec tion (13,35).However, only one-half of patients (4/10) have received aminoglycoside treatment, which may imply that phenotypical and genotypical dynamics of antimi crobial resistance reflect fast adaptation and genomic evolution not just in antimicro bial resistance but also in changing fitness, resulting in incomplete eradication of P. aeruginosa isolates or increased bacterial fitness by lost genes with a burden.
The identification of pathoadaptive genes, in which mutations accumulate with higher frequencies than expected statistically during relapse infection, gave hope to finding genetic markers that might predict the potential to establish chronic infection in the clinic.In our study, the results showed that relapse infections are not caused by specific lineages of P. aeruginosa isolates since these isolates are assigned to distinct STs and sporadically distributed in the phylogenetic tree.Therefore, we proposed that the relapse infection could impute to non-lineage-associated factors, such as genetic evolutions of horizontal gene transfer, mutation, or relapse-associated functional virulence-associated genes.To test our hypothesis, we additionally analyzed the genetic evolution at different times of relapse P. aeruginosa isolates.We found several mutations of cgSNPs in the later time of relapse infections, reflecting a dramatic genetic evolution during relapse infections.Notably, we found a convergent evolution of a non-synon ymous mutation in the fptA gene (T1056C, M252T).FptA is a TonB-dependent trans porter that permits the high-affinity binding and transport of Fe(III)-pyochelin complex, driving the import of the bacteriocin across the outer membrane and promoting the virulence of P. aeruginosa, especially in the pulmonary and urinary tract infections (43,44).Consistently, another study has found that the fptA gene was hypermutated in P. aeruginosa isolates collected from prolonged ICU inpatients; even the epidemiological and genomic information of these isolates were entirely different from the isolates in this study (38).Since the host lung microbiome is able to guard against colonization and infection of pathogens by competing for the nutrient metal iron (40), the P. aeruginosa is warranted to increase its metal iron acquisition function to overcome the iron limitation which could be a selective pressure facilitating the microevolution of fptA gene, thus in favor of the fitness and adaptation of P. aeruginosa isolates in relapse and prolonged infections.Previous studies have indicated that FptA confers increased susceptibility to siderophore-drug conjugates and is considered a novel target for treating multidrugresistant P. aeruginosa (37).According to the above evidence, we speculate that FptA could be a considerable target for diagnostics, decolonization, and treatment of relapse and prolonged infections by P. aeruginosa isolates.
Our study has several limitations.First, we included the patients with P. aeruginosa infections in a single clinical center, and our epidemiological findings may not be generalizable.We realize that more wide-ranging studies involving the surveillance and genomic investigations are needed to clarify the prevalence and genomic characteriza tions of relapse P. aeruginosa infection since CF is a rare disease in China.Secondly, we are not able to trace the clinical records of patients with relapse infections who were hospitalized in other medical centers, which could underestimate the incidence of relapse P. aeruginosa infections.Thirdly, we were not able to consider the influence of selective pressure from the host immune environment on the convergent evolution of relapse P. aeruginosa isolates.Above two limitations were restricted by the retrospective study design.A prospective cohort study and the collection of comprehensive clinical samples would supply strong evidence to address these gaps.Finally, we were not able to assess the causality between our genomic findings and relapse infection based on an epidemiological study.Further metatranscriptomics and experimental evidence would be helpful to address this noteworthy question and is part of our future work.
Despite these limitations, our data provide a comprehensive understanding of the high prevalence and genomic dynamics of relapse P. aeruginosa isolates across different stages.Integrated utilization of WGS and medical records provides opportu nities for improved diagnostics of relapse infections.Relapse P. aeruginosa infections were not caused by a certain lineage of isolates but could be attributable to the convergent evolution of the siderophore-encoding gene.Continued surveillance of the genomic dynamics of relapse P. aeruginosa infection will generate further knowledge for optimizing the treatment and prevention of relapse infection in the future, and our approach could potentially be used as a template for monitoring and diagnosing relapse infection of such prolonged infection by pathogens more widely.

FIG 1
FIG 1 Identification of relapse infection by screening of clinical data and whole-genome sequencing.(A) Timelines of relapse P. aeruginosa infection for each patient.Data are shown for isolates of 62 patients (1-62) from 1 September 2019, to 31 December 2022.The blue dots represent the positive cultures with P. aeruginosa isolates.The red forks represent negative culture with P. aeruginosa isolates.The docs connected with solid lines represent multiple times (>2) of positive cultures with P. aeruginosa isolates.(B) Timelines of reinfection of P. aeruginosa for each patient.Sequence type (ST) of the isolate was annotated beside the dot.(C) Phylogenetic tree of 163 P. aeruginosa isolates from 65 patients with multiple times of P. aeruginosa infections.The tree was constructed using core genome SNPs with maximum likelihood method.The isolates labeled with red frame indicate the relapse infection, while no frame indicates reinfection.

FIG 2
FIG 2 Density of relapse times for CRPA and CSPA isolates The x-axis represents the number of relapse times, while the y-axis represents the relapse density of the patients.Vertical dotted line represents the mean value of relapse times.CRPA, carbapenem-resistant P. aeruginosa; CSPA, carbapenem-susceptible P. aeruginosa.

FIG 4
FIG 4 MLST, ARGs, virulence-associated genes, and plasmid replicons for each sequence cluster based on phylogeny.(A) The distribution of MLST.The pink bar represents the number of patients of which ST is only found in one patient.(B) The proportion heatmap of ARGs, VFs, and plasmid replicons for each sequence cluster.The proportions of ARGs, virulence-associated genes, and plasmid replicons were represented by blue, orange, and green heatmaps, respectively.

FIG 5
FIG 5 Dynamic of SNPs in P. aeruginosa genome during the episodes of relapse infection (A) Number of SNPs in P. aeruginosa core genome among pair-wise comparisons during relapse infection of 62 patients.(B) PCA analysis of cgSNPs among first time and later times of relapse P. aeruginosa isolates.(C) Timelines for each patient with SNPs during the episodes of relapse infections.Data are shown for isolates of 15 patients from 1 September 2019, to 31 December 2022.The blue dots represent the positive cultures with P. aeruginosa isolates.The blue dots represent the genotype of P. aeruginosa isolate for each patient.The red dots labeled with red text represent SNP found in the fptA gene (T1056C).The green, purple, and orange dots represent the SNPs found in other SNP sites.

TABLE 1
Clinical characteristics of 196 patients infected by P. aeruginosa a a Data are presented as percentage of patients with the corresponding number in parentheses.UTI, urinary tract infection.b Statistical analysis with Fisher's exact test.BSI, bloodstream infection.

TABLE 2
Antimicrobial susceptibility of 442 Pseudomonas aeruginosa isolates a a The numbers in the table refer to the percentages of isolates showing susceptibility, intermediate susceptibil ity, and resistance to each antimicrobial.The numbers in parentheses indicate the number of isolates.-, not applicable.