Serotype Distribution, Antimicrobial Susceptibility, and Multilocus Sequencing Type (MLST) of Streptococcus pneumoniae From Adults of Three Hospitals in Shanghai, China

Background: Streptococcus pneumoniae, a main causative agent associated with invasive and non-invasive infection in elderly population, is a major global health problem. After pneumococcal conjugate vaccines (PCV) and pneumococcal polysaccharide vaccines (PPV) were introduced, the distribution of S. pneumoniae serotypes has changed. There was currently limited data on epidemiology and status of antimicrobial resistance of S. pneumoniae in Shanghai. Objective: To determine the serotype distribution, antimicrobial susceptibility and molecular epidemiology of S. pneumoniae isolated from adults in Shanghai. Method: A total of 75 S. pneumoniae isolates consecutively collected from 2015 through 2017 were serotyped by conventional multiplex-PCR. The antimicrobial susceptibility was determined by broth microdilution method. The multilocus sequence type (MLST) was performed to estimate the molecular epidemiology. Results: The predominant serotypes among the isolates were 19F (20.00%), 3 (16.00%), 23F (9.33%), 14 (8.00%), and19A (5.33%). The prevalence of pneumococcal strains with serotypes targeted by vaccines PCV7, PCV10, PCV13, and PPV23 was 44, 45.33, 66.67, and 80%, respectively. Penicillin non-susceptible S. pneumoniae (PNSSP) accounted for 16% of the isolates examined and resistance to erythromycin, azithromycin, tetracycline, clindamycin, cefaclor and trimethoprim-sulfamethoxazole were found in 92.00, 90.67, 86.67, 81.33, 54.67, and 54.67% of isolates, with most isolates (78.67%) presenting multidrug-resistance. The top three sequence types (STs) were ST271 (17.33%), ST180 (9.33%), and ST81 (8.00%). The international resistance clone complexes Spain23F-1 (n = 4), Netherland3-31 (n = 8), and Taiwan19F-14 (n = 14) were identified. Conclusions: The S. pneumoniae isolates showed high genetic diversity in Shanghai and the prevalence of antimicrobial resistance was also high among S. pneumoniae isolates, most of which were multidrug-resistant. The spread of international resistance clones might contribute to the increase of resistant isolates. The PPV23 could protect against most pneumococcal capsular serotypes causing infection of adults in Shanghai.


INTRODUCTION
Streptococcus pneumoniae is an opportunistic pathogen that causes both invasive pneumococcal diseases (IPD) including meningitis, bacteremia and non-invasive infection diseases such as otitis media, sinusitis, pneumonia (Engholm et al., 2017). The rate of IPD and the case fatality rate increased with increasing age. Although only 27.3% of the cases were in those aged ≥ 65 years, they accounted for 48% of the deaths (Marrie et al., 2018). The burden of non-invasive pneumococcal disease in adults is mainly determined by community-acquired pneumonia. Pneumococcal pneumonia also shows high incidence rates and carries a high mortality risk, especially in the elderly (Drijkoningen and Rohde, 2014). Chronic diseases and immunocompromise (IC), common in elderly population, are considered as two important risk factors in pneumococcal disease (PD) (Gil-Prieto et al., 2016). It has been presented as a significant cause which leads to substantial economic burden in at-risk and high-risk adults (Zhang et al., 2018).
Vaccines targeting pneumococcus are beneficial to the susceptible population. A study in the US reported that use of the pneumococcal conjugate vaccines (PCV) among children since 2000 has dramatically reduced IPD burden among adults (Moore et al., 2015;Pilishvili and Bennett, 2015). Multiple pneumococcal conjugate vaccines (PCV) and pneumococcal polysaccharide vaccines (PPV) have been developed and introduced to numerous counties (Ochoa-Gondar et al., 2015;Akata et al., 2017;Jauneikaite et al., 2017;Lyu et al., 2017). In China, the 7valent PCV were first licensed in 2008 and PCV13 and PPV23 are currently on the market (Liang et al., 2016). So far, they have not been covered by the official mandatory vaccination system as a secondary category vaccine. The vaccination is given only on voluntary basis at own expense, which maybe the cause of low vaccination rate in targeting population.
Antimicrobial is used as the most common anti-infection therapy and beta-lactam antibiotic is the first choice for the treatment of pneumococcal pneumonia. The increasing MIC value of penicillin worldwide has been observed and high-level cephalosporin-resistance caught most concern (Hakenbeck et al., 2012). On the other hand, high resistance rate (over 90%) of macrolide and clindamycin, tetracycline has been reported frequently in both pediatric and adult patients in mainland of China (Zhao et al., 2017;Wang et al., 2019). The resistance status is an important reference for antimicrobial selection for antiinfection therapy. The antibiotic-resistance varies in different regions so that local drug resistance monitoring is essential to provide evidence for clinical practice.
Surveillance of serotypes distribution and prevalence of drug-resistance strains in general population is critical for development of appropriate prevention and treatment protocol for S. pneumoniae infection. Shanghai is one of the most denselypopulated municipals with more than two million people aging over 65 years old. Although the serotype distribution of S. pneumoniae and the status of antimicrobial resistance in children has been reported in Shanghai, there has been only limited data for adults in this area (Pan et al., 2015). This multicenter study tried to survey the serotype distribution, antimicrobial resistance and molecular epidemiology of S. pneumoniae among adults in Shanghai.

Patients and Clinical Isolates
The study was conducted at Ruijin Hospital, Renji Hospital and Zhongshan Hospital, which are university-affiliated institutions with over 2000 licensed beds. We performed a retrospective lab-based study of adult patients (over 18 years old) with PD between January 2015 and December of 2017. Cases of PD were identified according to the database in the Department of Clinical Microbiology of each hospital. If the strain was separated from aseptic specimens such as blood and cerebrospinal fluid, it was regarded as IPD. If the strain was separated from nonaseptic specimens such as sputum, it was regarded as non-IPD. Seventy-five episodes of PD were enrolled in this study, only the first isolate of the same type of specimen from the same patient was reviewed and recorded. On basis of colony, microscopic morphology and the production of catalase, optochin test (5 µg in a 6mm disk) was carried out to detect the susceptibility of isolates. If the diameter was over 14mm, the isolate was identified as S. pneumoniae. If the diameter ranged from 9 to 13mm, bile-solute test was performed to identity the isolate. Meanwhile, isolates enrolled were also identified using matrix-assisted laser desorption ionization-time of flight mass spectrometer (bioMérieux, Marcy-l'Étoile, France).
This study was approved by the Ethics Committee of Ruijin Hospital affiliated with the School of Medicine at Shanghai Jiao Tong University. The Review Board exempted requirement for informed consent because this retrospective study used only the bacterial samples and had no any negative impact on the patients.

DNA Preparation
Rapid DNA extraction was performed by boiling 200µl of bacteria suspension of 1.0 MCF for 10 min. Then the suspension was centrifuged at 12,000rpm/min for 10 min. The supernatant was transferred into a second sterile tube and stocked at −20 • C until required.

Serotyping
Serotyping was performed by sequential multiplex-PCR and primers targeting cpsA were used as a positive control in each reaction (Pai et al., 2006). Serotype 6A and 6B were distinguished as previously described (Jin et al., 2009). Serotype 2 and 9N/L were identified by PCR according to previous reported (Dias et al., 2007;da Gloria Carvalho et al., 2010). Other non-vaccine serotypes were detected as published by the Centers for Disease Control and Prevention (CDC) 1 . If serotype that could not identified by conventional multiplex-PCR above, the isolate was classified as untypeable. All primer pairs were listed in Table 1. The prevalence of pneumococcal strains with serotypes targeted by current vaccines were calculated by the percentage of the isolates that the vaccines could protect against.

Multilocus Sequence Typing (MLST)
Seven housekeeping genes (aroE, gdh, gki, recP, spi, xpt, and ddl) were amplified, sequenced, and analyzed. Alleles and sequence types (STs) were determined according to the PubMLST database (https://pubmlst.org/spneumoniae/). Sequence and STs that could not be found in the database were submitted to the curator of the database. The clustering of related STs was analyzed by eBURST Version 3.0.

Statistical Analysis
Data in this study were analyzed by SAS 8.2 (SAS Institute Inc., Cary, NC, USA). For categorical variables, such as rates of antimicrobial resistance and serotype, the chiq-square test or Fisher test was used to compare the disparity between different groups. P < 0.05 was considered statistically significant.

Clinical Data
Seventy-five isolates were collected from January 2015 to December of 2017 (Supplementary Table 1), more (65.33%) were from male than female (34.67%). The age of those patients 1 http://www.cdc.gov/streplab/pcr.html  ranged from 28 to 90 years old, of which the mean was 58y and the median was 60 y. The proportion of patients aged over 65 was 37.33%. Among all isolates, 54 were from sputum specimen, 10 from blood and 4 from bronchial lavages and others were from nasal swab, bile, pus, exudates, pleuroperitoneal fluids and drainage-fluid (Figure 1), of which 21.33% were separated from aseptic specimens regarded as invasion, 78.67% from non-aseptic specimens regarded as non-invasion. The proportion of IPD in patients over 65 years old (28.57%) was higher than that in the younger group (17.02%). The common initial diseases in this study was malignant tumor (28.00%) and diseases of respiratory system (16.00%) ( Table 2).

Antimicrobial Susceptibility
The rates of susceptible, intermediate and resistance of S. pneumoniae islolates were listed in Table 3. According to the parenteral breakpoint of CLSI guideline, no isolate was resistant to penicillin (MIC ≥ 8 µg/mL), and penicillin-intermediate S.
pneumoniae (PISP, MIC = 4 µg/mL) accounted for 16.00% of all isolates. For oral breakpoint, the resistant (MIC ≥ 2 µg/mL) and intermediate (MIC = 0.12-1 µg/mL) rates were up to 42.67 and 9.33%, respectively. Most isolates (more than 80%) were resistant to erythromycin, clindamycin, azithromycin, and tetracycline. The resistant rates of cefaclor, cefuroxime, and trimethoprim/sulfamethoxazole were about 50%. There were 14.67 and 17.33% of all isolates resistant to cefepime and ceftriaxone. In penicillin non-susceptible isolates (PNSSP, MIC ≥ 4 µg/mL), the resistant rates of erythromycin, clindamycin, azithromycin, tetracycline and trimethoprim/sulfamethoxazole were 100%, which in PSSP were 90.47, 77.78, 88.89, 84.13, and 46.03%, respectively. Most isolates (over 95.00%) were susceptible to moxifloxacin, levofloxacin, chloramphenicol and all isolates were susceptible to linezolid and vancomycin. As described in Table 4, two thirds of serotype 19F isolates were non-susceptible to penicillin and more resistant to antibiotics than other serotypes. Moreover, serotype 23F, 14 and 19A showed high resistant rates of cefuroxime. Serotype 3 isolates were more susceptible to these antibiotics than others. The rates of resistance to cefuroxime, trimethoprim/sulfamethoxazole and meropenem in PCV13 targeting isolates were significantly higher than non-PCV13 isolates (the values of χ 2 were 16.67, 4.90, and 7.41 respectively, P < 0.05). A similar phenomenon was also observed in ceftriaxone and cefepime in PCV13 group and non-PCV13 group (both χ 2 = 13.61 P < 0.05). There was no significant difference in erythromycin, clindamycin, azithromycin, tetracycline and levofloxacin (P > 0.05).
There were two clonal complexes (CC) which the primary founders were ST271 with three single-locus variants (SLVs) and ST180 with two SLVs (Figure 3). We defined a group as at least six of seven MLST alleles in common. The ST4113 and ST14606 could be considered to belong to a single clonal complex, so could ST1263 and ST280. And there was no SLV and DLV of ST81.

DISCUSSION
S. pneumoniae is the principle pathogen that usually caused respiratory infection and invasive infection in children under five years old and senior citizen over 65 years old (Bogaert et al., 2004). Generally, it colonizes the nasopharynx as part of commensal flora in the health person. Once immune system is weaken or ectopic infection occurs, it can caused pneumococcal disease (Bogaert et al., 2004). IPD and pneumococcal pneumonia are common and carry a significant morbidity and mortality. In UK, the incidence of PD (20.58 per 100,000) in those aged > 65 years was significantly higher than that in all aged adults (Chalmers et al., 2016). In this study, chronic diseases and malignant tumor were observed in most patients enrolled, which indicated that chronic diseases and IC were two risk factors of pneumococcal infection. As a previous study in the United States, chronic liver disease and chronic obstructive pulmonary disease (COPD) were the highest risk for IPD. For person with non-IC chronic conditions, IPD risk increased with each additional condition (Baxter et al., 2016). Immunosuppressed patients due to immunosuppressive therapies and immunodeficiency were also at risk of pneumococcal infection. The relative risk (RR) of IPD could be multiplied by 20 to 50 in those with chemotherapy or HIV (Kyaw et al., 2005). Therefore, the prevention of pneumococcal infection in the population is crucial. Vaccines are the most effective tool to reduce the mortality and economic burden caused by pneumococcus (Naucler et al., 2017).
Since 2000, the 7-, 10-, and 13-valent pneumococcal conjugate vaccines (PCV) and 23-valent pneumococcal polysaccharide vaccines were introduced in many countries and reduced PD burden dramatically (Pilishvili and Bennett, 2015). In China, the 7-valent PCV were first licensed in 2008 and PCV13 and PPV23 are currently on the market (Liang et al., 2016). Although PCV13 has been listed in China, it has not been approved for the elderly. The pneumococcal vaccine approved in China for the elderly is PPV23. However, the overall vaccination rate is still unclear in Shanghai. Only a survey reported that in Qingpu district of Shanghai, the vaccination rate of 65-79 years old was 1.6%, the vaccination rate of 80 years old and above was 4.4% in 2014 (A Chinese Expert-workshop Recommendation for Elderly Influenza and Pneumococcal Vaccination, 2018). According to expert recommendation in China, PPV23 is recommended for the elderly aged 60 years and above, and 1 dose for basic inoculation. It is not recommended for the people with normal immune function to be inoculated again. However, if there is a risk factor for severe pneumococcal infection and the first vaccination has been more than 5 years, another vaccination is recommended. People aged 65 and over who have not been vaccinated within 5 years (including those under 65 at the time of the previous vaccination) may be vaccinated again. Due to insufficient safety data for 3 or more doses of PPV23, it is generally not recommended to vaccinate after the second dose (A Chinese Expert-workshop Recommendation for Elderly Influenza and Pneumococcal Vaccination, 2018). It was different from the Centers for Disease Control and Prevention (CDC) of the United States 2 . Although Shanghai has launched a project since 2013 to vaccinate the elderly aged 60 years and above with PPV23 free of charge, the overall vaccination rate in elderly population was still unclear in recent years. As the vaccination project progresses, the vaccination rate in elderly 2 https://www.cdc.gov/vaccines/vpd/pneumo/hcp/PCV13-adults.html population may have increased. In addition, the distribution of S. pneumoniae capsular serotypes in a certain region are associated with local status of vaccines (Akata et al., 2017). So seventy-five S. pneumoniae isolates from adults were analyzed to reveal the current prevalence state in Shanghai.
Usually, the introduction of vaccine was followed by a decrease in the rate of vaccine serotype accompanied by an increase in the diversity of non-vaccine serotype (Morales et al., 2018). For example in Japan, PCV13 was licensed in 2013 and the non-PCV13 serotype isolates increased significantly from 2012 to 2014 (Nakano et al., 2016). In a previous study of China from 2011 to 2016, serotype 19F (25.7%), 19A (14.0%), 15 (6.8%), 6B (3.6%), 6A (3.0%), and 17 (2.8%) were prevalent in adults. The prevalence of PCV7-and PCV13-serotypes were 37.5 and 58.3%, respectively (Zhao et al., 2017). The current study showed that the prevalence of serotype 19 group decreased while serotype 3 and 23F increased and the proportion of PCV13-serotypes in Shanghai was higher (66.67%). In patients ≥ 65 years old, the prevalence of PCV7 and PCV10 targeting serotypes were significant higher but those of PCV13 and PPV23 were not obviously different in adult-patients under and over 65y. Although PPV23 covered more serotypes than PCV13, the proportion in current study had no significant difference (P = 0.0648 > 0.05) in current study due to the prevalence of serotype 3 and 19A as a report from 2006 to 2016 (Lyu et al., 2017), The current vaccines could still cover most serotypes of the strains causing PD in adults, therefore, the vaccination rate in adult population, especially in the elderly, need to be promoted in the future. In addition, the epidemic data of S. pneumoniae from children in Shanghai has been reported previously (Pan et al., 2015). In this surveillance study, we found that the prevalence of 7-, 10-, and 13-valent PCV targeting serotypes were lower than that in children (Pan et al., 2015;Zhao et al., 2019), and it seems that PCV13 was more appropriate for pneumococcal infection protection in children.
Penicillin is normally the first choice for PD treatment. However, as the MIC value of penicillin increases, the breakpoint of CLSI have changed based on the clinical data in 2008. The former breakpoint categorizing susceptible, intermediate and resistant were ≤0.06, 0.12-1, and 2 µg/mL, respectively, which are still valid for oral treatment with penicillin and same with the EUCAST criteria. The new breakpoint, ≤2, 4, and 8 µg/mL, has  established for patients without meningitis but with intravenous treatment (Hakenbeck et al., 2012). In this retrospective study, we found that PNSSP isolates accounted for 16% based on parenteral breakpoint, this rate would have been 52.00% based on oral breakpoint, which was different from that previously reported from Beijing, 8.0% for parenteral breakpoint and 91.6% for oral breakpoint (Lyu et al., 2016). In addition, alterations in the penicillin-binding proteins (PBPs) have been recognized as a major resistance mechanism in S. pneumoniae (Hakenbeck et al., 2012). Altered PBP1a, PBP2x and PBP2b are the most important PBPs for β-lactam resistance amongst clinical isolates. It has been found that PBP2b mutation is mainly related to penicillin resistance. And PBP2x mutation can lead to moderate resistance to cephalosporins, while PBP1a mutation can cause high levels of resistance to both penicillin and cephalosporins (Cafini et al., 2006;Izdebski et al., 2008;Davies et al., 2012). In this study, ceftriaxone-resistant isolates (17.33%) were not resistant to penicillin and serotyped as 19F. PBPs mutation of these isolates are required in further research. The resistant rates of erythromycin, clindamycin, azithromycin and tetracycline were higher than 80%. As reported previously, the resistant rate of erythromycin in China was the highest in Asia (Kim et al., 2012). No matter in children or in adults, resistant rates of S. pneumoniae to erythromycin, tetracycline and trimethoprim/sulfamethoxazole were high in China (Yao and Yang, 2008;Zhao et al., 2017;Wang et al., 2019), more serious than that in Europe (Yahiaoui et al., 2016). Levofloxacin and moxifloxacin showed high susceptible rates and no isolates were resistant to vancomycin and linezolid, which was similar with reports in other regions in China .
S. pneumoniae with multidrug-resistance (MDR, resistant to three or more classes of antimicrobial) was widely existed in clinical infection and greatly threaten the treatment of PD. The incidence of MDR in invasive PD was higher than that in non-invasive PD . In our study, 78.67% of isolates showed MDR. According to a study from Asian Network for Surveillance of Resistant Pathogens (ANSORP), the overall S.pneumoniae MDR rate was 59.3%, with the highest MDR rate 83.3% in China (Kim et al., 2012). Vaccine-targeting serotypes demonstrated higher resistance compared with uncovered serotypes. In current study, antimicrobial resistant rate in the most prevalent serotypes (19F, 19A, 14, and 23F) was higher than other serotypes but serotype 3 was much more susceptible. Extensive beta-lactam antibiotic resistance was observed more in 19F. In addition, From 2007 to 2013 in Canada, MDR rate of was only 2. 9% and 19F and 19A were also the most common serotype for resistance to more than five antimicrobial classes (Golden et al., 2015).
In previous report of China, CC271 was more common in children, whereas singletons were more prevalent in adults, and serotype 19F and 19A possessed the homogeneous genetic background, together with high resistance to antibiotics (Zhao et al., 2013). The molecular analysis of S. pneumoniae in this study showed that ST271, ST180, ST81, ST320, and ST876 were the top five prevalent STs. CC271 was also prevalent in adults, accounted for 25.33% (19/75).The serotype 19F-ST271 was dominating and covered most PNSSP isolates as reported previously (Lyu et al., 2016). The spread of international resistant clones played a predominant role in increasing resistance in Shanghai. The international clone complexes Taiwan 19F -14, Spain 23F -1, Spain 6B -2, and Taiwan 23F -15 had been reported previously in Shanghai . In our study, Taiwan 19F -14 was still predominant in the same region. However, the prevalence of Netherland 3 -31 (ST180) clone was first reported. The international resistance clone Netherland 3 -31 accounted for 10.67% of S. pneumoniae from adult, following the Taiwan 19F -14.
In conclusion, serotype 19F, 3, 23F, 14, 19A targeted by PCV13 and PPV23 were most prevalent in adult patients of Shanghai associated with ST271, ST180, ST81, ST320, and ST876. Taiwan 19F -14 was predominant of three international resistant clone complexes followed by Netherland 3 -31 (ST180) clone, which was first reported in Shanghai. The incidence of MDR was high and MDR existed in most isolates covered by current vaccines. And the epidemiology of S. pneumoniae requires further long-termed investigation on the basis of this study.

DATA AVAILABILITY STATEMENT
All datasets generated for this study are included in the article/Supplementary Material.

ETHICS STATEMENT
This study was approved by the Ethics Committee at Ruijin Hospital affiliated with the School of Medicine at Shanghai Jiao Tong University. The Review Board exempted requirement for informed consent because this retrospective study used only the bacterial samples and had no any negative impact on the patients.

AUTHOR CONTRIBUTIONS
X-XL and S-ZX conceived and designed the experiments, performed the experiments, and wrote the paper. X-XL, S-ZX, F-FG, and S-YZ analyzed the data. QX, Z-KS, Y-XN, J-MQ, and L-ZH contributed reagents, materials, analysis, and tools. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.

ACKNOWLEDGMENTS
We would like to thank all the technicians of clinical microbiology of Ruijin Hospital for their assistance in routine AST work.