Association between Pneumonia Development and Virulence Gene Expression in Carbapenem-Resistant Acinetobacter baumannii Isolated from Clinical Specimens

We investigated the virulence gene expression of carbapenem-resistant Acinetobacter baumanii (CRAB) isolated from the respiratory samples of patients with CRAB pneumonia and those with CRAB colonization to identify the virulence genes contributing to CRAB pneumonia's development and mortality. Patients with CRAB identified from respiratory specimens were screened at a tertiary university hospital between January 2018 and January 2019. Patients were classified into CRAB pneumonia or CRAB colonization groups according to predefined clinical criteria. A. baumannii isolated from respiratory specimens was examined for the expression levels of ompA, uspA, hfq, hisF, feoA, and bfnL by quantitative reverse-transcription polymerase chain reaction. Among 156 patients with CRAB from respiratory specimens, 17 and 24 met the criteria for inclusion in the pneumonia and colonization groups, respectively. The expression level of ompA was significantly higher in the pneumonia group than in the colonization group (1.45 vs. 0.63, P=0.03). The expression levels of ompA (1.97 vs. 0.86, P=0.02), hisF (1.06 vs. 0.10, P < 0.01), uspA (1.62 vs. 1.01, P < 0.01), and bfnL (3.14 vs. 2.14, P=0.03) were significantly higher in patients with 30-day mortality than in the surviving patients. Elevated expression of hisF (adjusted odds ratio = 5.93, P=0.03) and uspA (adjusted odds ratio = 7.36, P=0.02) were associated with 30-day mortality after adjusting for age and the Charlson score. uspA and hisF may serve as putative targets for novel therapeutic strategies.


Introduction
Acinetobacter baumannii is the leading cause of ventilatorassociated pneumonia in patients in intensive care units (ICU); however, it can also cause community-acquired pneumonia [1,2].Among A. baumannii, carbapenem resistance is increasing in A. baumannii, reaching 50-70% depending on the region [3][4][5][6].Carbapenem-resistant A. baumannii (CRAB) has doubled the mortality rate of carbapenem-sensitive A. baumannii [7] because of the lack of optimal treatment options.Tus, CRAB is a critical global priority, according to the World Health Organization [8].
Several host factors such as ICU stay, recent surgery, comorbidities, immunosuppressive therapy, and tracheostomy have been suggested as risk factors for A. baumannii invasive infections [9,10].In addition to host factors, understanding the microbiological factors associated with pneumonia is important because these factors may contribute to developing novel treatment strategies and early diagnostic tools for A. baumannii pneumonia [11].
Several virulence mechanisms contribute to A. baumannii lung infections, including bioflm production, attachment to biotic surfaces, host cell invasion, apoptosis, oxidative stress resistance, and iron regulation [8].Several virulence genes have been reported to contribute to this process.ompA encoding the outer membrane protein (OmpA) enhances cell death via mitochondrial and nuclear targeting [12][13][14][15].Universal stress protein A (UspA) is responsible for resistance to oxidative stress and low pH [16].Te RNA chaperone Hfq contributes to resistance to stresses such as temperature, pH, osmotic pressure, oxidative stress, adhesion, and invasion of host cells and modulates fmbriae production [17].HisF decreases innate immunity and the infammatory responses [18].FeoA plays roles in cell ftness, adhesion, bioflm formation, and growth [19].bfnL, which belongs to the baumanoferrin cluster, is involved in bioflm formation, attachment, and ftness under ironlimiting condition [14].Te deletion of these genes leads to a loss of virulence, as mice infected with the knockout strain showed improved survival [14,16,18,20].Te elevated expression of ompA, feoA, bfnL, and hisF has been observed in murine pneumonia infections [14].Because this knowledge is based on in vitro experiments and animal studies, it is necessary to investigate the virulence gene expression of A. baumannii in human infections.
Tis study aimed to investigate the virulence gene expression of CRAB isolated from respiratory samples of patients with CRAB pneumonia and those with CRAB colonization and to demonstrate the virulence genes contributing to the development of pneumonia and subsequent mortality.

Study Design.
Te study subjects were inpatients at a tertiary university hospital between January 2018 and January 2019, for whom CRAB was identifed from respiratory specimens.Te inpatients included patients admitted to the ICU.Respiratory specimens included sputum, endotracheal aspirates, and bronchoalveolar lavage fuid.Respiratory cultures were performed according to the order of the attending physician as part of usual clinical practice.Te expectorated sputum specimens submitted for culture were evaluated for their quality using the modifed Murray-Washington grouping system [21].Only group 4-6 specimens were considered acceptable and proceeded for culture.Transtracheal aspirate through an endotracheal tube or bronchial aspirate through bronchoscopy were considered as lower respiratory tract specimens, regardless of the squamous epithelial cell count.Te respiratory specimens were inoculated on the blood agar plate (BAP) and the MacConkey agar and incubated in 5% CO 2 at 36 °C for 1-2 days.A. baumannii was identifed using VITEK MS (bioMérieux, France).Te carbapenem resistance of A. baumannii was tested using VITEK2 (bioMérieux, France) with an AST-N225 card (bioMérieux, France).A minimum inhibitory concentration (MIC) of ≥8 µg/mL for imipenem and meropenem was considered to indicate carbapenem resistance, according to the Clinical and Laboratory Standards Institute guidelines [22].Te isolated CRAB was stored at −80 °C until further analysis.
Te screened patients were retrospectively determined to have CRAB pneumonia if they met the following criteria: (a) radiologic evidence of pneumonia at the time of CRAB isolation, (b) ≥5 days of efective antibiotic treatment initiated within three days of identifcation of A. baumannii, and (c) no other microorganisms identifed from blood or sputum within one month of isolation of A. baumannii.In criterion (b), fve days represent the minimal duration needed to treat pneumonia, and three days represent a reasonable interval to consider that the antibiotic prescription was intended to treat the isolates.Te criterion (c) was used to exclude polymicrobial infections.Additionally, patients who satisfed criteria (a) and (c) but died before receiving fve or more days of efective antibiotic treatment, with pneumonia as the primary cause of death, were considered to have CRAB pneumonia.
Screened patients were determined to have CRAB colonization if they met (d) no radiologic evidence of pneumonia within seven days of the isolation of A. baumannii and (e) a lack of coexistence of A. baumannii bacteremia.
CRAB isolated from patients with pneumonia and colonization was used for further molecular analysis.If patients met the above criteria for more than one CRAB isolate, only the frst isolate was included in the analysis.

Results
A total of 156 patients with CRAB identifed from respiratory specimens were screened during the study period.Among these, 68 patients with polymicrobial infections and 47 patients who did not meet the criteria for pneumonia or colonization were excluded from the study.Finally, 17 and 24 patients met the criteria for pneumonia and colonization and were included in the analysis.ompA expression levels were signifcantly higher in the pneumonia group than in the colonization group (1.45 vs. 0.63, P � 0.03).Te expression levels of uspA, hfq, hisF, feoA, and bfnL were not signifcantly diferent between groups (Table 1, Figure S1).After adjusting for age and the Charlson score, none of the virulence gene expression levels was signifcantly associated with pneumonia (Table 2).
Nine of the 17 patients in the pneumonia group and one of the twenty-four patients in the colonizer group died within 30 days of CRAB isolation.Te cause of death in the pneumonia group was CRAB pneumonia in eight patients and sepsis other than CRAB in one patient.One patient in the colonization group died of pneumonia caused by microorganisms other than CRAB.

Discussion
We observed higher ompA expression in the CRAB pneumonia patients than in those with colonization.In addition, higher ompA, hisF, uspA, and bfnL expression in CRAB was observed in patients with 30-day in-hospital mortality than in the surviving patients.Te association between virulence gene expression and pneumonia was not statistically signifcant after adjusting for age and the Charlson score.hisF and uspA were signifcantly associated with 30-day mortality after adjusting for age and the Charlson score.
Given that antibiotic-resistant A. baumannii poses a considerable threat to health in nosocomial settings, diverse eforts have been made to reduce the burden of CRAB infections in healthcare settings, such as introducing an antimicrobial restriction system to guide the appropriate use of carbapenem [27].Many studies have attempted to determine the optimal dose, duration, and combination of traditional antibiotics for treating CRAB.Despite these efforts, there is no precise "standard-of-care antibiotic regimen for CRAB infection yet [28][29][30].
Te management of CRAB infections is complicated for several reasons.First, CRAB is mainly recovered from respiratory specimens.Still, it is not always clear whether an isolate is a colonizing organism or whether CRAB represents a true pathogen, leading to uncertainty regarding the need for antibiotic therapy.Second, once A. baumannii exhibits carbapenem resistance, it generally acquires resistance to most other antibiotics expected to be active against wild-type A. baumannii.With the emergence of multidrug-and pandrug-resistant A. baumannii, there is an urgent need to identify new methods for treating drug-resistant A. baumannii infections [15,31].
Identifying the virulence gene expression of CRAB in patients with pneumonia and colonization is valuable for two reasons.If there is a signifcant diference in gene expression between the two, it can be used as an indicator to distinguish between pneumonia and colonization.Second, the overexpression of virulence genes during pneumonia suggests they could be targets for novel therapeutic developments [15,20].For example, AOA-2 is a synthetic polypeptide designed to interact with OmpA.AOA-2 inhibits the adhesion of A. baumannii to biotic and abiotic surfaces and enhances the sensitivity of A. baumannii to antibiotics [32][33][34][35].OmpA is also a potential target for developing vaccines and monoclonal antibodies [15].Except for the relatively well-known ompA, investigations on the virulence genes of A. baumannii have been limited to in vitro and animal studies.Tis is the frst study to report the expression levels of hisF, uspA, bfnL, feoA, and hfq in clinical isolates, which may provide fundamental data for selecting novel therapeutic targets.
OmpA is a surface-exposed porin protein with a β-barrel structure embedded in the outer membrane [36].Te amino acids of A. baumannii OmpA from a variety of clinical isolates Canadian Journal of Infectious Diseases and Medical Microbiology     Canadian Journal of Infectious Diseases and Medical Microbiology are highly conserved (>89%), but are not homologous to the human proteome [37].Tis is a favorable feature of OmpA as a potential therapeutic target.OmpA is involved in the adherence to the epithelia [13], translocation into the epithelial cell nucleus [38], and induction of the epithelial cell death [39].
OmpA is also involved in bioflm formation [13] and binds to factor H, allowing A. baumannii to develop serum resistance [40].If sufcient data are accumulated to set a threshold for predicting the virulence of A. baumannii, qRT-PCR monitoring of A. baumannii ompA may be a promising tool for detecting the development of pneumonia.uspA has been detected in A. baumannii as an open reading frame (ORF) on an AbaR resistance island [41,42].
In vitro studies have shown that UspA is involved in resistance to H 2 O 2 .Because generation of reactive oxygen species is essential to the host's innate immune response, resistance to oxidative stress helps A. baumannii survive the host defense mechanism, leading to invasive infection.In an animal study, infection with uspA-deleted A. baumannii resulted in a reduced bacterial load in the lungs and improved the survival of mice, suggesting its role in the lethality of A. baumannii.Tis is consistent with our results in which elevated uspA expression was associated with 30-day mortality.Te UspA protein sequence is highly conserved among diferent A. baumannii strains, indicating that it may be a plausible therapeutic target [16].
Te HisF is involved in histidine and de novo purine biosynthesis.HisF is involved in inhibiting the recruitment of innate immune cells and decreasing the production of the proinfammatory cytokine IL-6.Tus, HisF decreases innate immunity and infammatory responses, enabling A. baumannii to survive host defense mechanisms and cause invasive infections.Overexpression of hisF has been established in a mouse model of pneumonia [14,18].In a murine pneumonia model, mice infected with the hisF-deleted mutant strain showed a lower bacterial burden in the lungs and an improved survival rate, suggesting that HisF could be a therapeutic target to improve the survival of A. baumannii infection.Tis was consistent with our results, in which elevated hisF expression was associated with 30-day mortality.
Belonging to the baumanoferrin cluster, bfnL encodes an N-acetyltransferase associated with ferric siderophore synthesis [14,43].bfnL is involved in bacterial attachment and bioflm formation.Bioflm-forming A. baumannii is particularly difcult to treat because it is more resistant to antimicrobial agents and the host immune system than planktonic bacteria.Tis enables A. baumannii to persist on biotic surfaces and diverse medical devices [44].Deletion of bfnL resulted in reduced adherence of A. baumannii to human alveolar epithelial cells and decreased bioflm formation.Mice infected with the bfnL mutant strain showed improved survival in a pneumonia model [14].We also identifed higher expression of bfnL in patients with 30-day mortality, although the association was not signifcant in multivariable analysis.
We observed a higher ompA expression in patients with pneumonia, which is consistent with the results of previous clinical and animal studies.Te virulence gene expression efect on pneumonia was insignifcant after adjusting for age and the Charlson score.Nevertheless, it is difcult to conclude that virulence gene expression is not associated with pneumonia because the sample size in this study was too small to achieve statistical signifcance in multivariate analysis.Furthermore, when categorizing patients according to mortality, signifcantly higher expression of ompA, hisF, uspA, and bfnL was noted in patients with 30-day mortality than in those who survived.hisF and uspA remained signifcantly associated with 30-day mortality after adjusting for age and the Charlson score.Te elevated expression of hisF, uspA, and bfnL in patients with 30-day mortality but not in patients with pneumonia raises two hypotheses.First, the possible misclassifcation of colonization as pneumonia might have attenuated the diferences between the groups.Based on the criteria for pneumonia used in this study, identifying pneumonia on chest radiography and prescribing antibiotics against CRAB were based on clinical decisions that can vary depending on the physician.However, mortality is an objective outcome, and the misclassifcation of deceased and surviving patients is impossible.A large-scale prospective study in which an assigned investigator determines pneumonia according to predefned criteria combining patients' symptoms and laboratory and imaging fndings may improve the grouping and reliability of the multivariable analysis.Second, the expression of these genes may be higher in advanced pneumonia than in mild pneumonia.Te analysis of gene expression levels depending on the severity of pneumonia may provide supporting evidence for this hypothesis.
Tis study has several limitations.First, this was a singlecenter study with a small sample size.Because we excluded polymicrobial infections and patients who did not meet CRAB pneumonia or colonization criteria, the fnal number of patients analyzed was small.Second, the multivariate analysis included only age and the Charlson score as covariates.Te sample size of this study was too small to include many covariates.We did not include ICU stay and mechanical ventilation in the multivariable analysis because it was uncertain whether these were the result of the pneumonia or the predisposing factors for the pneumonia.In addition, multicollinearity was expected between ICU stay and mechanical ventilation.A prospective study design may help clarify this temporal relationship.Tird, we adopted the prescription of antibiotics against CRAB as one of the criteria for pneumonia, assuming that the attending physician decided to treat pneumonia based on symptoms, physical examination fndings, and laboratory and imaging studies.However, clinical decisions can vary depending on the physician, which may afect consistency in the defnition of pneumonia.A prospective study design with a single assigned investigator judging pneumonia or colonization may improve grouping consistency.Fourth, gene expression analysis was performed under controlled conditions (in TSB broth at 37 °C), which did not represent the environment of the living tissue.Whether the overexpression of virulence genes is strain-specifc or a response to various environmental stressors needs to be determined.Fifth, unlike in animal studies, A. baumannii isolates may be obtained from diferent time points of pneumonia in each patient, which may partially afect gene expression levels.
Canadian Journal of Infectious Diseases and Medical Microbiology

Conclusions
Te elevated expression of hisF and uspA in CRAB is associated with 30-day mortality, suggesting that these genes may be putative targets for novel therapeutic strategies.Higher ompA expression in the CRAB of patients with pneumonia than in those with colonization and higher ompA and bfnL expression in CRAB in patients with 30-day in-hospital mortality than in surviving patients necessitate a large-scale prospective study to establish the usefulness of these genes as targets for novel diagnostic and therapeutic strategies.

2
Canadian Journal of Infectious Diseases and Medical Microbiology a housekeeping gene to normalize gene expression, and A. baumannii American Type Culture Collection (ATCC) 19606 was used as the calibrator.

Table 1 :
Expression levels of Acinetobacter baumannii virulence genes and clinical characteristics of patients with pneumonia and colonization.

Table 2 :
Association of Acinetobacter baumannii virulence gene expression levels and pneumonia adjusted for age and Charlson score.

Table 3 :
Expression levels of Acinetobacter baumannii virulence genes and clinical characteristics of deceased and surviving patients.

Table 4 :
Association of Acinetobacter baumannii virulence gene expression and 30 day in-hospital mortality adjusted for age and Charlson score.