Phenotypic and genetic differences among group B Streptococcus recovered from neonates and pregnant women in Shenzhen, China: 8-year study

Background Group B Streptococcus (GBS) is a leading cause of early-onset disease (EOD) and late-onset disease (LOD) in infants. We sought to investigate the antibiotic susceptibility profiles, resistance genes, virulence-related genes, serotype distribution and genotypic characteristics of GBS recovered from infected or colonized neonates and pregnant women in a tertiary teaching hospital in Shenzhen, China, from 2008 to 2015. Results High resistance rates of erythromycin (66.7–100%) were detected among early-onset GBS (EOGBS), late-onset GBS (LOGBS), neonatal colonizing GBS (NCGBS) and maternal colonizing GBS (MCGBS). 89.5–100% of four groups of GBS isolates showed resistance to tetracycline. More than 90 % of erythromycin resistant isolates of EOGBS (8/8, 100%), LOGBS (16/17, 94.1%) and NCGBS (10/11, 90.9%) harbored ermB, while only 9.1–17.6% harbored mefA/E. By contrast, 55.8% (24/43) and 62.8% (27/43) of erythromycin resistant MCGBS isolates carried ermB and mefA/E genes, respectively. The tetO gene was more common in tetracycline resistant EOGBS (10/11, 90.9%), LOGBS (17/17, 100%) and NCGBS (10/11, 90.9%), compared to tetracycline resistant MCGBS (12/51, 23.5%). Additionally, the tetM gene accounted for 90.9% (10/11), 76.5% (13/17), 45.5% (5/11) and 80.4% (41/51) of four groups of isolates, respectively. Serotype III was the most predominant in EOGBS (8/12, 66.7%) and LOGBS (15/17, 88.2%), while serotype Ib accounted for 50.0% (6/12) of NCGBS, and serotype Ia and III accounted for 45.6% (26/57) and 33.3% (19/57) of MCGBS, respectively. Sequence type 17 (ST17) was the most common in EOGBS (6/12, 50%) and LOGBS (12/17, 70.6%), while ST12 was predominant in NCGBS (5/12, 41.7%), and five STs (ST19, ST23, ST12, ST103 and ST485) accounted for 66.7% (38/57) of the MCGBS. All serotype III-ST17 isolates recovered from neonates were associated with invasive infections. Conclusions This study shows the meaningful differences in molecular mechanisms of resistance to erythromycin and tetracycline, and the prevalence of serotypes and STs among GBS recovered from neonates and pregnant women. ST17 is predominant in neonatal invasive GBS, but rare in NCGBS and MCGBS. Electronic supplementary material The online version of this article (10.1186/s12866-019-1551-2) contains supplementary material, which is available to authorized users.


Background
Group B Streptococcus (GBS) is the primary pathogen of neonatal infections, and the common colonizer of human genitourinary and gastrointestinal tract [1]. Recently, resistance to erythromycin among invasive GBS from neonates and adults has been increasing gradually, which is mainly mediated by various macrolide resistance genes, e.g., ermB, ermA, ermTR and mefA/E [2][3][4]. In 2010, Centers for Disease Control and Prevention (CDC) issued the updated guideline for prevention of GBS infection in perinatal period, which recommended universal GBS screening for GBS colonization at 35-37 weeks' gestation or among women with threatened preterm delivery and unknown colonization status, as well as intrapartum antimicrobial prophylaxis (IAP) in labor among those GBS colonized. A significant decrease in the incidence of early-onset GBS (EOGBS) has been reported since the implementation of IAP policies [5].
Serotyping is a traditional phenotypic method. Currently, there are 10 serotypes (Ia, Ib and II to IX) according to capsular polysaccharide (CPS) of GBS. It is believed that 40% of early-onset disease (EOD) and 60% of late-onset diseases (LOD) in the world are related with CPS III [1]. Multilocus sequence typing (MLST) has been widely applied in epidemiologic surveillance of GBS, as a relatively reliable tool for the comparison of genetic profiles of isolates recovered from various geographic areas.
Here, we have characterized GBS isolates responsible for invasive infections or colonization in newborns and pregnant women in Shenzhen, China, from 2008 to 2015. Additionally, we also investigated the status of implementation of prenatal GBS screening and IAP policies in our hospital during 8-year period.

Associations between ST and serotype
Among neonatal invasive GBS, all ST17 isolates (n = 18) belonged to CPS III. By contrast, only one isolate was identified as ST17 among NCGBS, whose serotype was Eighteen CPS III-ST17 neonatal invasive isolates were recovered from septicemia (n = 16), bacteremia (n = 1) and omphalitis (n = 1). Sixteen out of eighteen isolates harbored ermB gene with resistance to erythromycin and clindamycin, and all isolates carried tetO gene with resistance to tetracycline (Table 3).

Clinical and biological features of GBS neonatal infections
A total of twenty-nine GBS neonatal infections (12 EOD and 17 LOD cases) were confirmed, and bloodstream infection (including sepsis and bacteremia) was the most common clinical manifestation, with 83.3% (10/12) of EOD and 94.1% (16/17) of LOD cases, respectively. Omphalitis was identified in two EOD and one LOD cases. One LOD case was simultaneously associated with sepsis and meningitis caused by GBS. The clinical signs of pneumonia and meningitis were reported in 58.3% (7/12) and 16.7% (2/12) of EOD cases, respectively, compared with 76.5% (13/17) In the MST, circles represent the STs, and the areas of each of circles indicate the prevalence of the ST in the input data. Lines are used to show the radial links from the founder to each of its single-locus variants (SLVs). The primary founder of a group is defined as the ST that differs from the largest number of other STs at only a single locus, and it is coloured blue. The user-selected primary founder is coloured red. A ST that appears to have diversified to produce multiple SLVs is called a subgroup founder, which is coloured yellow ermB (16), mefA/E (2), lnuB (5) tetO (18), tetM (13), tetL (1) hylB (18), lmb (18), scpB (17) SE (16), BA (1), OM (1) hylB (4), lmb (4), scpB (4) SE (4) hylB (1)

Prenatal GBS screening and IAP
The complete delivery records were obtained from mothers of early-onset cases (n = 9), of late-onset cases (n = 9), of colonized neonates (n = 11), colonized pregnant women (n = 47), and infected pregnant woman (n = 1), along with incomplete data for mothers of early-onset cases (n = 3) and late-onset cases (n = 8

Discussion
Resistance to erythromycin in neonatal invasive GBS has been reported worldwide. Martins ER et al. [2] reported that erythromycin resistance in GBS recovered from neonatal invasive infections in Portugal increased from  [8], in contrast, other lnu genes have been rarely identified in GBS, although lnuC has been reported in this species in a study [9]. In contrast with lnuA in Staphylococcus aureus and lnuB in GBS, both of which confer resistance to clindamycin, lnuD in Streptococcus uberis confers resistance to lincomycin but not to clindamycin. The reason for the difference in the resistance phenotypes remains unexplained [10]. Our study showed the similar result that lnuD was detected only in two clindamycin intermediate and four clindamycin susceptible MCGBS isolates. To the best of our knowledge, this is the first time that the lnuD gene has been reported among GBS isolates recovered from pregnant women in China. In addition, neither erm nor lnu gene was detected in one clindamycin resistant isolate recovered from colonized pregnant women, though several attempts were conducted. The present study showed that more than 90 % of tetracycline resistant isolates recovered from infants harbored tetO, while only 23.5% of those from mothers harbored it. Additionally, tetM was more prevalent in tetracycline resistant EOGBS, LOGBS and MCGBS, compared to NCGBS (76.5-90.9% versus 45.5%). It suggests that there are significant differences for the prevalence of tetO and tetM among four groups of tetracycline resistant GBS isolates.
The bac gene encodes the β-C protein, which is associated with epithelial cell invasion and resistance to phagocyte clearance, whereas, the bca gene encodes the α-C protein, which is associated with epithelial cell adherence [11]. In our study, higher occurrence of bac and bca genes were observed for those isolates recovered from infants, compared those from colonized mothers (24.4% versus 7.0%) and (12.2% versus 1.8%), respectively. Of noted, the differences for the occurrence of bac and bca genes were strongly associated with the serotypes of isolates. Our study demonstrated that the bac gene was only detected in the CPS Ib isolates recovered from infants (10/10, 100%) and from colonized mothers (4/6, 66.7%), as well as bca gene identified mainly in the CPS Ib isolates recovered from infants (5/10, 50%) and occasionally in the CPS Ia isolates from colonized mother (1/26, 3.8%). Our findings are agreement with the results of Lindahl' study [12]. Interestingly, 50% (5/10) of CPS Ib isolates collected from infants harbored simultaneously both bac and bac genes, compared none of those from colonized mothers. The reason is not yet known.
As reported by Madrid L et al. [13], five serotypes (Ia, Ib, II, III and V) accounted for 97% of neonatal invasive GBS isolates in all regions with serotype data, according to a meta-analysis of serotype prevalence. CPS III was the most common serotype worldwide and nearly half (47%) of EOGBS cases and 73.0% LOGBS cases were caused by it. In our study, CPS III was responsible for 66.7% (8/12) of EOGBS cases and 88.2% (15/17) of LOGBS cases. Interestingly, CPS Ib accounted for 50.0% (6/12) of NCGBS isolates in our study, followed by III and Ia. Similarly, five serotypes (Ia, Ib, II, III and V) have been estimated to encompass approximately 98% of the serotypes identified during maternal colonization, according to a recent metaanalysis of maternal colonization [14]. In the research of Lu et al. [6], CPS III was identified in 41.8% of maternal colonizing isolates, followed by Ia (21.4%), V (14.9%), Ib (11.9%) and II (7.0%). In the study of Teatero S et al. [7], the most frequently identified serotypes were III (25%), Ia (23%), and V (19%) among the maternal colonizing isolates. In the present study, Ia (26/57, 45.6%) and III (19/57, 33.3%) were the most common serotypes among MCGBS, and other serotypes (Ib, II and V) accounted for 21.1% (12/57) of the remaining isolates.
Associations between ST and serotype have been reported in the literature, with some showing a strong correlation [14]. In the current study, 95% (19/20) of ST17 and 91.3% (21/23) of ST19 GBS isolates represented CPS III, and 92.3% (12/13) of ST12 isolates represented Ib. Several studies showed the correlation between CPS III-ST17 (CC17) GBS and neonatal invasive infection [2,3,14]. As reported by Shabayek [15], a major clone responsible for a large proportion of invasive neonatal infections are the CC17 strains mostly belonging serotype III. The CC17 strains are reported to be hypervirulent accounting for more than 80% of the GBS late-onset neonatal infections. Our study showed the similar results, CPS III-CC17 isolates accounted for 65.5% (19/29) invasive neonatal infections, however, none of twelve and one of fifty-seven (1.8%) isolates recovered from neonatal and maternal colonization, respectively, belonged to CPS III-CC17. It should be noted that the low overall incidence of invasive GBS disease in infants (data available in the subsequent discussions) is probably associated with the low colonization rates of CPS III-CC17 strains in neonates and pregnant women in our hospital. A significant reduction in the incidence of EOGBS has been reported since the introduction of universal GBS screening and IAP policies [5,16,17]. In our hospital, neither national nor local IAP policy is available presently, only those pregnant women with threatened preterm delivery are required for GBS culture of vaginal or cervix swabs, and receive empirical or targeted antibiotic treatments other than IAP. In the present study, 71.4% (20/28) and 93.1% (27/29) of mothers of infected neonates had vaginal delivery and full-term labor, respectively. 94.4% (17/18) of mothers of infected neonates received neither antenatal GBS screening nor IAP at labor. These results indicate the importance of the implementation of the universal antenatal GBS screening and the IAP policies among pregnant women with full term labor and vaginal delivery. Of note, LOD accounted for 58.6% (17/29) of neonatal invasive infections in our study, which is still hard to prevent by present IAP policies. It is an urgent need to develop an efficacious maternal GBS vaccine against LOD.
In our study, the low incidence of EOGBS and LOGBS among pre-term and very pre-term/low birth weight neonates was found, and the reason remains unknown. In our hospital, more medical cares are given to the pregnant women with threatened preterm delivery, including the screening for GBS and empirical or targeted antibiotic treatments prior to labor, in spite of the absence of local IAP policy. We consider it important to reduce the incidence of EOGBS among pre-term and very preterm/low birth weight neonates in our hospital. Also, the empirical treatments with broad-spectrum antibiotics for the complicates among pre-term and very pre-term/low birth weight neonates at the early stage of pre-term neonatal growth is common in our NICU, which could be useful to decrease the incidence of LOGBS among this population. Guan et al. investigated the epidemiology of invasive GBS collected from three large urban tertiary hospitals in South China from Jan 2011 to Dec 2014, and reported the overall and the EOGBS incidence of 0.55, 0.39 per 1000 live births, respectively [18]. In present study, with the data provided by the statistical department of our hospital, total 18 cases (9 EOGBS and 9 LOGBS) were identified from 54,192 live births born in our hospital from 2008 to 2015, giving the overall incidence of 0.33 per 1000 live births, along with the EOGBS and LOGBS incidence of 0.17 per 1000 live births, respectively. Both the overall and EOGBS incidence in our study was markedly lower than the global average estimation (0.49, 0.41 per 1000 live births, respectively) and those reported in the study from Guan et al. [13,18]. The reason remains unclear.
The present study has a limitation of small scale of GBS isolates, with the lack of pre-term neonatal and maternal invasive GBS isolates. To the best of our knowledge, data on the epidemiology of neonatal GBS infection and bacteria resistance in China are currently very limited. Shenzhen is a rapidly growing city of immigrant from all over the country, with a large of population about thirteen million people. Shenzhen People's Hospital is a Third-level center hospital and one of the largest urban tertiary hospitals in Shenzhen, and receives many transfers from other urban, district and rural hospitals. Our findings were representative of a large population of China. However, it should be noted that a national prospective observational study is urgently needed to investigate the incidence of invasive GBS disease in infants, antimicrobial resistance, the common resistance genes, and the distribution of serotypes and genotypes among invasive/ colonizing GBS recovered from neonates and pregnant women.

Conclusions
This study shows the meaningful differences in molecular mechanisms of resistance to erythromycin and tetracycline, and the prevalence of serotypes and STs among GBS recovered from neonates and pregnant women. ST17 is predominant in neonatal invasive GBS, but rare in NCGBS and MCGBS.

Bacterial isolates
Shenzhen People's Hospital is a 2500-bed tertiary teaching hospital in Shenzhen, China. In the current study, a total of 99 non-duplicate GBS clinical isolates were recovered from 41 neonates and 58 pregnant women In 2012, the number of beds in neonatal intensive care unit (NICU) increased from 60 to 100, indicating the significant increase of inpatients. Neonatal GBS isolated from blood (n = 26), umbilical secretion (n = 3), pharynx swabs (n = 11) and endotracheal aspirate (n = 1); Maternal GBS isolated from cervical secretion (n = 53), vaginal secretion (n = 2), urine (n = 2) and blood (n = 1). All isolates were identified as GBS by Vitek 2 system (bio-Merieux), since 2014, by Vitek MS system (bioMerieux), along with phenotypic and manual biochemical tests.

Serotyping and antimicrobial susceptibility testing
Capsular serotyping was carried out by a latex agglutination assay with the ImmuLex™ Strep-B kit (Statens Serum Institute, Copenhagen, Denmark), including 10 antibodies of GBS capsular polysaccharide (CPS Ia, Ib, II to IX), according to the manufacturer's instructions.
All GBS isolates were tested for susceptibility to erythromycin, clindamycin, tetracycline, levofloxacin, penicillin, and vancomycin by using the E-test method according to the Clinical and Laboratory Standards Institute (CLSI) guidelines. Inducible clindamycin resistance (iMLS B ) was tested by disk diffusion (D-zone test) with erythromycin and clindamycin [19].

MLST
MLST was performed by sequencing seven housekeeping genes (adhP, pheS, atr, glnA, sdhA, glcK and tkt) of GBS as described previously [28]. Alleles and sequence types (STs) were identified by using the S. agalactiae MLST database (http://pubmlst.org/sagalactiae/). The new alleles and STs were deposited at the S. agalactiae MLST database. The eBURSTv3 software (http://eburst. mlst.net) was used to display the relationships between STs, analyze clonal complexes (CCs), and create a minimum spanning tree. CCs were defined at the single locus-variant (SLV) level.

Investigation of medical records
Medical records of GBS infected/colonized neonates and their mothers, as well as GBS infected/colonized pregnant women were investigated. The invasive infection was confirmed by the positive culture of GBS isolated from a normally sterile site (e.g. blood and cerebrospinal fluid) or an infected site (e.g. umbilicus) in our cases, along with infection-related symptoms and signs, as well as laboratory examinations. The sole positive culture of GBS isolated from a non-sterile or non-infected site (e.g. pharynx, trachea, urinary tract, cervix, and vagina), without consistent clinical findings, was defined as GBS colonization. Infection occurring in infants less than 7 days old, and that occurring in infants 7 to 89 days old, was defined as EOD and LOD [3,5], respectively.

Additional files
Additional file 1: Table S1. The details of serotypes and genotypic characteristics of GBS recovered from neonates and pregnant women, respectively. The CPS types, sequence types, clonal complexes, and predominant CPS type-STs were listed. (DOCX 15 kb) Additional file 2: Table S2. The details of Clinical and biological characteristics of GBS neonatal infections. The sex, age, birth weight, gestational age, infections confirmed by culture, clinical signs of infection, targeted and empirical antibiotic treatment, and outcomes were listed.