Molecular and phenotypic characterization of Streptococcus pneumoniae isolates in a Japanese tertiary care hospital

This study aimed to investigate the bacterial characteristics of pneumococcal isolates obtained from a tertiary care hospital in Japan. We analyzed the antimicrobial susceptibility, possession of macrolide resistance genes, pneumococcal serogroup/serotype, and sequence type (ST) of pneumococcal isolates from patients aged 15 years or older between 2011 and 2020 at Nagasaki University Hospital. Of the 73 isolates analyzed, 86.3% showed resistance to macrolides, and 28.8%, 46.6%, and 11.0% harbored mefA, ermB, and both, respectively. Of the isolates possessing ermB, 97.6% showed high levels of macrolide resistance [minimal inhibitory concentration (MIC) range, > 16 µg/mL]. Solithromycin (MIC range, 0.03–0.25 µg/mL), regardless of the presence of macrolide resistance genes, and lascufloxacin (MIC range, 0.06–0.5 µg/mL) showed potent in vitro activity against pneumococci. Serotype 19A was the most prevalent (six isolates), followed by serotypes 10A, 15A, and 15B/C (five isolates each). Four serotypes (11A, 19A, 22F, and 23B) and five STs (36, 99, 433, 558, and 3111) were significantly correlated with the presence of macrolide resistance genes. All four isolates with serotype 11A/ST99 and three isolates with serotype 19A/ST3111 harbored both mefA and ermB. No macrolide resistance genes were detected in either of the two isolates with serotype 22F/ST433, while all ten isolates with serogroup 15 (serotypes 15A and 15B/C, five isolates each) possessed ermB alone. Our study revealed the bacterial characteristics of the pneumococcal isolates obtained from our hospital. In vitro activity of solithromycin and lascufloxacin against these isolates was confirmed.


Introduction
Streptococcus pneumoniae is a leading cause of community-acquired pneumonia (CAP) and nursing and healthcare-associated pneumonia (NHCAP) and causes invasive pneumococcal diseases (IPDs) such as bacteremia and meningitis (The JRS, 2017).Implementation of pneumococcal conjugate vaccine (PCV) has reduced CAP and IPD caused by vaccine preventable serotype (Bonten et al., 2015).However, it was observed that the proportions of vaccine serotypes decreased, but those of non-vaccine serotypes increased after PCV introduction (Hanquet et al., 2022).Additionally, the resistance to commonly used antimicrobials, including penicillin, macrolides, and fluoroquinolones is of great concern (Li et al., 2023).High rates of macrolide resistance have been reported in Japan and other Asian countries (Kawaguchiya et al., 2020;Li et al., 2023).Macrolide resistance is mediated by several mechanisms such as methylation of the ribosomal macrolide target site and macrolide efflux (Wierzbowski et al., 2007).The ermB gene, one erythromycin ribosomal methylase (erm) family gene, is a common determinant of macrolide resistance in pneumococci (Schroeder and Stephens, 2016).Furthermore, the efflux mechanisms encoded by the macrolide efflux (mef) genes cause resistance to only 14and 15-membered macrolides (Littauer et al., 2005).
Previous studies have reported relationships between the presence of macrolide resistance genes, capsular typing, and sequence type (ST).In a study conducted in Korea, most strains with serotypes 19F (77.2%) and 19A (87.5%) possessed both mefA and ermB (Bae and Lee, 2009).In a report from Taiwan, of the azithromycin-resistant isolates carrying both mefA and ermB, 64% and 24% belonged to serotypes 19F and 19A, respectively, and 33% were ST320, 32% were ST236, and 12% were ST271 (Safari et al., 2014).In a study conducted for Japanese children with IPD between 2010 and 2013, the proportions of macrolide-resistant isolates harboring mefA or ermB (or both) decreased in vaccine serotypes after PCV7 introduction, whereas the rates of macrolide-resistant isolates mediated by ermB in the non-vaccine serotypes 15A, 15B, 15C, and 24 increased (Chiba et al., 2014).A study conducted for isolates from adult patients between 2008 and 2016 in Aichi, located in the middle of Japan, reported that 100% of isolates with serotypes 14, 15A, and 23A harbored ermB alone and 89.7% of those with serotype 19F possessed mefA alone, while 73.0% of those with serotype 23F harbored both mefA and ermB (Furuya et al., 2017).In addition, a high prevalence (> 90%) for ermB was observed in serotypes 15A, 23A, 15B, 15C, 33F, and 12F and the prevalence for both mefA/E and ermB was highest in serotype 11A/11D (53.8%), according to a study conducted between 2018 and 2019 in Hokkaido, located in northern Japan (Kawaguchiya et al., 2020).However, these relationships have not been fully investigated in our region, located in western Japan.
Solithromycin, a novel fluoroketolide, and lascufloxacin, a novel fluoroquinolone are two antimicrobial agents with potent activity against S. pneumoniae.Following development, both these antibiotics are expected to become treatment options against pneumococcal pneumonia.A study reported that solithromycin was very active [minimal inhibitory concentration (MIC) range, ≤ 0.001-1 µg/mL; MIC 90 , 0.25 µg/mL] against 996 pneumococcal isolates collected from North America, Europe, Asia-Pacific, and other regions (Hawser et al., 2017).
Fluoroquinolones inhibit DNA synthesis by binding to DNA gyrases (GyrA and GyrB) and topoisomerase IV (ParC and ParE) in S. pneumoniae.One of the main causes of fluoroquinolone resistance is the gradual accumulation of mutations in ParC and GyrA (Cornick and Bentley, 2012).Lascufloxacin has been reported to show in vitro activity against pneumococci with GyrA, ParC, and both mutations (Kishii et al., 2017).It also exhibits lower selectivity of resistant strains after drug exposure to strains with either GyrA or ParC mutation (first-step mutants to fluoroquinolone resistance), compared to levofloxacin and garenoxacin (Murata et al., 2018).In addition, a study reported that lascufloxacin exhibited excellent antibacterial activity against penicillinsusceptible and resistant S. pneumoniae isolates (MIC range, 0.03-0.06µg/mL; MIC 90 , 0.06 µg/mL) collected in Japanese hospitals (Kishii et al., 2017).
This study aimed to investigate the bacterial characteristics of pneumococcal isolates obtained from our hospital, including their susceptibility to a novel fluoroketolide and fluoroquinolone.

Study design
This retrospective study was conducted using pneumococcal isolates obtained from the Nagasaki University Hospital, a tertiary care hospital, between 2011 and 2020.We extracted pneumococcal isolates from respiratory specimens and blood of patients aged 15 years or older from our clinical database.Only the first pneumococcal isolate from individual patients was selected for analysis during the study period.Pneumonia was defined as the presence of new infiltrate on chest radiograph or computed tomography image along with the presentation of clinical symptoms.Isolates from respiratory specimens and blood of patients with pneumonia were stored at −80°C in Microbank or at room temperature after freeze drying.In addition, we listed isolates from respiratory specimens, which were collected from patients without pneumonia and evaluated to be purulent, according to the Miller and Jones classification (Miller, 1963).Of these, we used the isolates stored at −80°C in Microbank for further analysis.
Clinical features, including baseline characteristics, antimicrobial therapy, and outcomes, were collected from the medical records.Pneumonia was classified into three categories, such as CAP, NHCAP, and hospital-acquired pneumonia (HAP), according to the definition proposed by the Japanese Respiratory Society (The JRS, 2017).The severity of pneumonia was assessed using A-DROP, a five-factor scoring system for communityacquired pneumonia (age, dehydration, respiratory failure, orientation disturbances, and low blood pressure), which is a modified version of the CURB-65 and was proposed by the Japanese Respiratory Society (Kosai et al., 2014;The JRS, 2017).The study protocol, including the waiver of informed consent, was approved by the Institutional Review Board of Nagasaki University Hospital (approval number:21091310).

Antimicrobial susceptibility testing
Antimicrobial susceptibility testing (AST) was performed using a MIC plate customized by Eiken Chemical Co., Ltd.The prepared MIC plate was incubated in 5% CO 2 at 35 ± 2°C for 20-24 h.The results were interpreted according to CLSI document M100-ED33.

Identification of S. pneumoniae and detection of macrolide resistance genes
DNA was extracted from the bacteria using the boiling method as previously described (Yamakawa et al., 2019;Kikuchi et al., 2023), with minor modifications.Multiplex PCR for mefA, ermB, and lytA for S. pneumoniae identification was performed using a QIAGEN Multiplex PCR Kit (QIAGEN) under the following conditions: 15 min at 95°C, 30 cycles of 15 s at 94°C, 15 s at 53°C, and 15 s at 72°C, and 10 min at 72°C.The primers used (NIID, 2021) were shown in Supplementary Table 1.

Multilocus sequence typing
Multilocus sequence typing (MLST) was performed targeting seven housekeeping genes (aroE, gdh, gki, recP, spi, xpt, and ddl).PCR for each housekeeping gene was performed using TaKaRa Ex Taq DNA Polymerase (Takara Bio Inc.) and the primers described in PubMLST website (https://pubmlst.org/organisms/streptococcus-pneumoniae/primers) or CDC website (https://www.cdc.gov/streplab/pneumococcus/resources.html) under the following conditions: 2 min at 94°C, 30 cycles of 10 s at 98°C, 15 s at 55°C, and 30 s at 68°C , and 5 min at 72°C.The PCR products were purified using ExoSAP-IT PCR Product Cleanup Reagent (Applied Biosystems).Sequencing reactions were performed using a BigDye Terminator v3.1 Cycle Sequencing Kit (Applied Biosystems).The products were purified using a BigDye Xterminator Purification Kit (Applied Biosystems) and analyzed using a SeqStudio Genetic Analyzer (Applied Biosystems).Allele sequences and STs were determined using the P u b M LS T d at a ba s e ( h t t p s : / / p u b m l s t .o r g / o r g a n i s m s / streptococcus-pneumoniae).

Detection of quinolone resistancedetermining region mutations
Quinolone resistance-determining region (QRDR) mutations were detected by Sanger sequencing.The primers used (Pan et al., 1996) were shown in Supplementary Table 1.PCR was performed using TaKaRa Ex Taq DNA Polymerase (Takara Bio Inc.) under the following conditions: 5 min at 94°C, 30 cycles of 40 s at 94°C, 40 s at 55°C, and 1 min at 72°C, and 5 min at 72°C (Roychoudhury et al., 2016).The subsequent procedures were as described in MLST.The QRDR nucleotide and amino acid sequences were compared to those of S. pneumoniae R6 using the National Center for Biotechnology Information Blast program (https://blast.ncbi.nlm.nih.gov/Blast.cgi).

Determination of pneumococcal serogroup/serotype
The pneumococcal serogroup/serotype was determined by the latex agglutination method using the ImmuLex Pneumotest Kit (SSI Diagnostica), according to the manufacturer's instructions with minor modifications.Briefly, a drop (approximately 10 µL) of ImmuLex reagent and the prepared bacterial suspension were applied separately next to each other onto the reaction card.They were mixed using a mixing stick and spread over a circular area on the card.Agglutination was determined after the cards were rocked slowly.Furthermore, additional serotyping was performed for serotypes classifiable by PCR.Briefly, the PCR assays were carried out using QIAGEN Multiplex PCR Kit (QIAGEN) and the primers described in the literature (Pai et al., 2006;Carvalho Mda et al., 2009;Pimenta et al., 2009;da Gloria Carvalho et al., 2010) or CDC website (https:// www.cdc.gov/streplab/pneumococcus/resources.html)under the following conditions: 15 min at 95°C, 35 cycles of 30 s at 94°C, 90 s at 54°C, and 60 s at 72°C, and 10 min at 72°C (da Gloria Carvalho et al., 2010).

Statistical analysis
Numerical or categorical variables were analyzed using the Kruskal-Wallis test or Fisher's exact test among groups.GraphPad Prism 10 (GraphPad Software) was used for the analyses, and the results were considered statistically significant at p < 0.05.

Bacterial characteristics of pneumococcal isolates
A total of 86 patients with pneumonia, aged 15 years and older, were confirmed during the study period and 41 isolates (32 in Microbank and 9 freeze-dried isolates) were available as isolates from patients with pneumonia.These isolates were obtained from 34 sputum, six tracheal aspirates, and one blood.In addition, 32 isolates from purulent respiratory specimens (24 sputum, seven tracheal aspirates, and one aspirate obtained using bronchoscopy), which were collected from patients without pneumonia, were available.In total, 73 isolates were included in this study.
Table 1 shows the MICs for the antimicrobial agents, while Figure 1 shows the presence of macrolide resistance genes in the 73 pneumococcal isolates.The MIC 90 s for all three macrolides (14membered, erythromycin and clarithromycin; 15-membered, azithromycin) were > 16 µg/mL.Of the 73 isolates, 63 (86.3%) harbored either mefA or ermB and were resistant to all three macrolides.Of the 73 isolates, 21 (28.8%),34 (46.6%), and eight (11.0%)tested positive for mefA, ermB, and both, respectively.The relationship between macrolide MIC and the presence of macrolide resistance genes is shown in Figure 2.Among the 42 isolates possessing ermB, 41 (97.6%) (33 isolates with ermB alone and eight with both mefA and ermB) showed higher MICs (> 16 µg/ mL) for all three macrolides.In contrast, for isolates harboring mefA alone, the MICs were distributed 2-> 16 µg/mL for erythromycin and clarithromycin and 4-> 16 µg/mL for azithromycin.High MIC elevation (> 16 µg/mL) for three macrolides were significantly correlated with the presence of macrolide resistance genes (p < 0.0001) (Table 2).The MICs for solithromycin ranged 0.03-0.25 µg/ mL regardless of the presence of macrolide resistance genes and tended to be lower than those for macrolides (Table 1).Meanwhile, although the MICs for solithromycin ranged 0.03-0.06µg/mL for isolates without mefA and ermB, this range was wider and higher (0.06-0.25 µg/mL) for isolates harboring both the genes (Figure 3).The highest MIC (0.25 µg/mL) for solithromycin were significantly correlated with the presence of macrolide resistance genes (Table 2).
Table 3 shows the serogroup/serotype and ST of the pneumococcal isolates classified according to the presence of macrolide resistance genes.Table 4 shows the relationships between serogroup/serotype or ST and the presence of macrolide resistance genes.Among the isolates identified as serotypes, serotype 19A was the most prevalent (six isolates), followed by serotypes 10A, 15A, and 15B/C (five isolates each), and serotypes 11A, 19A, 22F, and 23B were significantly correlated with the presence of macrolide resistance genes (p < 0.05, respectively) (Table 4).Of the isolates showing significant correlations between serotypes and the presence of macrolide resistance genes, all four isolates with serotype 11A harbored both macrolide resistance genes, whereas all two isolates with serotype 22F harbored no macrolide resistance gene.Of the six isolates with serotype 19A, three (50.0%) each harbored mefA alone or both mefA and ermB.Among the four isolates with serotype 23B, three (75.0%)and one (25.0%)possessed no macrolide resistance genes and ermB alone, respectively.
Among the isolates with identified STs, ST558 was the most frequent (seven isolates), followed by ST63 (six isolates) and ST5236 (five isolates).ST36, 99, 433, 558, and 3111 were significantly correlated with the presence of macrolide resistance genes (p < 0.05, respectively) (Table 4).Neither the ST36 nor ST433 isolates harbored macrolide resistance genes, whereas all ST99 and ST3111 Presence of macrolide resistance genes in pneumococcal isolates.

Clinical features of patients with pneumococcal pneumonia
Table 5 shows the clinical features of patients with pneumococcal pneumonia classified according to the presence of macrolide resistance genes in the isolates.Sex was significantly correlated with the presence of macrolide resistance genes.However, other variables, including classification of pneumonia, comorbidities, pneumonia severity, and mortality, showed no significant correlation with the presence of macrolide resistance genes in the isolates.
Among the three isolates from two patients with severe pneumonia (A-DROP score 3) and one patient with extremely severe pneumonia (A-DROP score 5), two harbored ermB alone and one no resistance genes.Their serogroups/serotypes were 3, 6C, and 22F.Among the five isolates from patients with bacteremia, four harbored ermB alone and one mefA alone.Their serogroups/ serotypes were 3 (two isolates), 6C, 19A, and 13/28.

Discussion
Our results demonstrated high rates of macrolide resistance, which was conferred by the presence of macrolide resistance genes.However, a novel ketolide, solithromycin, possessed potent in vitro  Relationship between the macrolide MICs and the presence of macrolide resistance genes.MIC, minimum inhibitory concentration; EM, erythromycin; CAM, clarithromycin; AZM, azithromycin.
activity, even in the presence of macrolide resistance genes.In addition, lascufloxacin, a novel fluoroquinolone, has demonstrated strong activity against pneumococcal isolates.Bacterial typing revealed the distribution of capsular and sequence types and clarified the relationships between the possession of macrolide resistance genes, serogroups/serotypes, and sequence types.We observed a high macrolide resistance rate (86.3%) in this study, which is similar to previous studies from Japan, in which the resistance rate to erythromycin was 82.6% (Kawaguchiya et al., 2020) and only 2.6-8.1% and 2.1-4.8%showed susceptibility for clarithromycin and azithromycin, respectively (Okade et al., 2014).In this study, of the isolates that tested positive for ermB, including those harboring ermB alone and both ermB and mefA, 97.6% (41/42) showed higher MICs for macrolides (> 16 µg/mL) compared to isolates that tested positive for mefA alone (2-> 16 µg/mL).This confirmed previous reports that high-and intermediate-level resistance to erythromycin is conferred by ermB-mediated target modification and mef-mediated efflux, respectively (Lynch and Martinez, 2002;Xu et al., 2010).Additionally, the MICs for solithromycin were lower (0.03-0.25 mg/ mL) than those for other macrolides, including against isolates harboring mefA and ermB.This result is consistent with previous reports showing that solithromycin is active against macrolide-resistant pneumococci, including those with erm-and mef-mediated resistance (McGhee et al., 2010).The activity of solithromycin against Relationship between the solithromycin MICs and the presence of macrolide resistance genes.MIC, minimum inhibitory concentration.pneumococci harboring erm may be due to its ability to bind to ribosomes methylated by erm (Llano-Sotelo et al., 2010).However, although the MICs for solithromycin was low (0.03-0.25 µg/mL), 62.5% (5/8) of isolates with both ermB and mefA showed the highest MIC (0.25 µg/mL) (Figure 3, Table 2).Accumulation of macrolide resistance mechanisms might contribute to MIC elevation for solithromycin.In addition, a recent study reported novel mutations related to decreased susceptibility for solithromycin, including mutations in the intergenic regions of the macrolide resistance locus mefE/mel and in the vicinity of the ribosome binding site of ermB (Gingras et al., 2023).Because S. pneumoniae can acquire solithromycin resistance, surveillance of antimicrobial susceptibility for the drug should be continued.The present study showed the frequencies of serogroups/ serotypes and STs, and significant correlations between four serotypes (11A, 19A, 22F, and 23B) or five STs (ST36, 99, 433, 558, and 3111) and the presence of macrolide resistance genes.All ten isolates (100%) with serogroup 15 (serotypes 15A and 15B/C, five isolates each) possessed ermB alone, which was similar to a previous report from Japan describing a high prevalence (> 90%) for ermB in isolates with serotypes 15A, 15B, and 15C (Kawaguchiya et al., 2020).In this study, all four isolates (100%) with serotype 11A/ST99 harbored both mefA and ermB and no macrolide resistance genes were detected in any of the two isolates with serotype 22F/ST433.All three isolates (100%) with serotype 19A/ ST3111 possessed both macrolide resistance genes, but all three isolates with serotype 19A/ST2331 harbored mefA alone.
In contrast to the high rate of macrolide resistance (86.3%), the rates of resistance to levofloxacin and moxifloxacin were only 2.7% 11A ( 4) 99 ( 4) Values in parentheses indicate numbers of isolates.ND, not determined.
TABLE 4 Relationship between serogroup/serotype or sequence type and the presence of macrolide resistance genes.et al., 2001;Brueggemann et al., 2002), that were consistent with our results.Meanwhile, the MICs for lascufloxacin against levofloxacinand moxifloxacin-resistant isolates were low (0.5 µg/mL) but highest among the isolates tested in this study.Therefore, we need continuous monitoring of antimicrobial susceptibility for this drug.

Variable
In our study, no significant differences were observed between clinical factors, excluding sex, and the presence of macrolide resistance genes in the isolates.All five bacteremia episodes were observed in patients with pneumonia caused by macrolide-resistant isolates (four with ermB alone and one with mefA alone) and their serogroups/serotypes were 3, 6C, 19A, or 13/28.A previous study, comparing hospitalized patients with pneumonia caused by macrolide-resistant isolates with those by macrolide-sensitive isolates, reported that the rate of bacteremia was significantly lower in patients with macrolide-resistant isolates.In this previous study, the rate of patients with pneumonia caused by macrolide-susceptible isolates was high (78.4%)compared to our study and serotypes of isolates that caused bacteremia were not shown (Cilloniz et al., 2015).Further studies are required to confirm the characteristics of macrolide-resistant S. pneumoniae, including pathogenicity, and the clinical features of patients with pneumonia caused by the bacteria.
This study has some limitations.The study was conducted in a single hospital in western Japan, which might not be representative of other regions or countries.The sample size was relatively small, potentially limiting the generalizability of the results.We did not investigate factors such as prior antibiotic use which might influence resistance patterns.
In conclusion, this study revealed the antimicrobial susceptibility, capsular type, and ST of the pneumococcal isolates obtained from our hospital.Macrolide resistance and possession of resistance genes were frequently observed, and specific capsular types and STs were correlated with the presence of macrolide resistance genes.In vitro activity of solithromycin and lascufloxacin against pneumococcal isolates was confirmed.Data expressed as median (interquartile range) or numbers (column percentages).CAP, community-acquired pneumonia; NHCAP, nursing-and healthcare-associated pneumonia; HAP, hospital-acquired pneumonia.
a One patient with ventilator-associated pneumonia is included.

FIGURE 3
FIGURE 3 numbers (row percentages).a Of the serogroups/serotypes determined as a single serogroup/serotype or determined sequence types, those which include two or more isolates are listed.See Supplementary The author(s) declare financial support was received for the research, authorship, and/or publication of this article.This study was partially supported by a grant from the Japan Agency for Medical Research and Development (AMED) for the Research and Development of Diagnostic Methods and Therapies for Antimicrobial-Resistant Bacteria (JP22fk0108133).

TABLE 2
Relationship between MIC elevation and the presence of macrolide resistance genes in 73 pneumococcal isolates.
Data expressed as numbers (column percentages).MIC, minimum inhibitory concentration.

TABLE 3
Serogroup/serotype and sequence type of 73 pneumococcal isolates.

TABLE 3 Continued
Table 2 for those which include only one isolate.

TABLE 5
Clinical features of 41 patients with pneumococcal pneumonia classified according to the presence of macrolide resistance genes in the isolates.