Microbiological and clinical characteristics of Streptococcus gallolyticus subsp. pasteurianus infection in China

Background Infections by Streptococcus gallolyticus subsp. pasteurianus (SGSP) is often underestimated. Herein, the epidemiological features and resistant characteristics of SGSP in mainland China are characterized to enable a better understanding of its role in clinical infections. Methods In the present work, 45 SGSP isolates were collected from the samples of bloodstream, urine, aseptic body fluid, and fetal membrane/placenta from patients in 8 tertiary general hospitals of 6 cities/provinces in China from 2011 to 2017. The identification of all isolates was performed using traditional biochemical methods, 16S rRNA and gyrB sequencing, followed by the characterization of their antibiotic resistance profiling and involved genes. Results Among 34 non-pregnancy-related patients, 4 (4/34,11.8%) patients had gastrointestinal cancer, 10 (10/34, 29.4%) patients had diabetes, and one patient had infective endocarditis. Moreover, 11 cases of pregnant women were associated with intrauterine infection (9/11, 81.2%) and urinary tract infection (1/11, 9.1%), respectively. Except one, all other SGSP isolates were correctly identified by the BD Phoenix automated system. We found that all SGSP isolates were phenotypically susceptible to penicillin, ampicillin, cefotaxime, meropenem, and vancomycin. Forty strains (40/45, 88.9%) were both erythromycin and clindamycin-resistant, belonging to the cMLSB phenotype, and the majority of them carried erm(B) gene (39/40, 97.5%). Although the cMLSB/erm(B) constituted the most frequently identified phenotype/genotype combination (25/40, 62.5%) among all erythromycin-resistant cMLSB isolates, erm(B)/erm(A), erm(B)/mef(A/E), and erm(B)/erm(T) was detected in 7, 4, and 3 isolates, respectively. Furthermore, 43 strains (43/45, 95.6%) were tetracycline-resistant, and out of these, 39 strains (39/45, 86.7%) carried tet(L), 27(27/45, 60.0%) strains carried tet(O), and 7 (7/45, 15.6%) strains carried tet(M), alone or combined, respectively. All erythromycin-resistant isolates were also resistant to tetracycline. Conclusions It is important to study and draw attention on SGSP, an underreported opportunistic pathogen targeting immunodeficient populations, notably elderly subjects, pregnant women and neonates.


Background
Streptococcus gallolyticus subsp. pasteurianus (SGSP), formerly known as S. bovis biotype II/2 [1] and is one member of Group D streptococci, is a cause and a potential pathogen of bacteremia and infective endocarditis (IE), as well as urinary tract infection (UTI), in elderly and immunodeficient people [2][3][4], septicemia and meningitis in newborns, and as well as intrauterine infection in pregnant woman [5][6][7][8]. This species is also associated with gastrointestinal malignancy [3,9]. It colonizes the digestive and female genital tract and therefore can lead to UTI and neonatal invasive infection, resembling what happens with group B Streptococcus (S. agalactiae, GBS). However, frequently occurring erroneous identification of SGSP might lead to an underestimation of the real incidence of infections caused by the species [6,10].
Additionally, the susceptibility of SGSP strains to β-lactam and vancomycin has remained relatively stable over the past years, while variable resistance rates were observed against clindamycin, erythromycin, tetracycline and levofloxacin [4,11,12].
Considering gradually increased clinical infections caused by SGSP [10,11], the clarification of its clinical features and antibiotic resistance is highly desired and should be valuable for its prevention and treatment. Unfortunately, epidemiological studies on SGSP isolates circulating in mainland China have not been conducted yet. To this end, we retrospectively analyzed SGSP isolates collected from 8 tertiary teaching hospitals in 6 cities/provinces in China from 2011 to 2017, and wanted to properly group these strains into species/subspecies level using traditional biochemical methods and 16S rRNA as well as gyrB sequencing to obtain their phenotypic and genotypic antibiotic resistance traits. The clinical and antibiotic resistance features of these SGSP isolates would help to understand the infections caused by the species circulating in China and for decision making in the context of empiric therapy.

Sample sources
Forty-seven non-duplicate isolates that were originally identified as SGSP in line with the new taxonomy criteria [13,14] were recovered from 8

Phylogenetic analysis of 16S rRNA gene
DNAs were extracted from SGSP stains and subjected to PCR amplification and sequencing using a commercial DNA purification kit (Promega) according to the manufacturer's instructions. The 16S rRNA genes from all SGSP strains were amplified with two universal primers (27F and 1492R), and the amplification of the DNA gyrase subunit B (gyrB) gene was performed using primers gyrB F 5′-GAAGTDGTIAARATYACBAAYCG-3′ and gyrB R5′-ACATCDGCATCRGTCAT-3′ as described elsewhere [15]. The sequencing of the 16S rRNA and gyrB was conducted by Ruibiotech (Beijing, China). The consequent comparison of the respective 16S rRNA and gyrB sequences against those in GenBank was performed using online BLASTn (www.ncbi.nlm.nih.gov/blast). A sequence similarity of 99 and 96% was used as a "cut-off" value for SGSP species identification [16,17]. In addition, the phylogenetic tree was generated based on the 16S rRNA gene using the neighbour-joining algorithms using MEGA version 10.0.5 and iTOL v4 (https://itol.embl.de). To this end, the sequences were aligned with reference sequences of SGSP type strain AJ297216.1 that is available in the GenBank database.
Bacterial identification using BD Phoenix automated microbiology system All 47 isolates were examined by the department of clinical microbiology of CAGH for further confirmation based on the new taxonomy. BD Phoenix 100 Automated Microbiology System STREP (SMIC/ID) panel (Becton Dickinson, Sparks, MD, USA) was used as the identification method. The misidentified isolates by BD Phoenix 100 system, including S. infantarius (now designated as S. infantarius subspecies infantarius) (1 isolate) and Enterococcus faecalis (1 isolate), were excluded for further analysis. Finally, 45 SGSP isolates were included in the subsequent study and the clinical data of the patients are shown in Table 1. Their detailed geographic distribution was shown in Fig. 1.

Phenotypical and genotypical features of antibiotic resistance
Susceptibility tests against penicillin, cefotaxime, vancomycin, meropenem, erythromycin, clindamycin, and tetracycline were performed using STREP (SMIC/ID) panel. The interpretive criteria for antibiotic susceptibility test (AST) were according to the Clinical and Laboratory Standards Institute (CLSI, 2017) for Streptococcus spp. viridans group.
Strains that showed resistant to erythromycin and tetracycline were also analyzed by PCR using conditions as described previously [1,18], to examine the presence of antibiotic resistance genes, which are commonly found among the isolates of S. bovis group, including erm(A), erm(B), erm(T), mef(A/E), tet(K), tet(L), tet(M) and tet(O).

Statistical analysis
In the study, MIC50 and MIC90 are defined as the MICs of a given agent that inhibits the growth of 50 and 90% of the isolates, respectively. MIC data of each antibiotic were recorded and analyzed by WHONET 5.6 software, and MIC50 and MIC90 were also calculated. Furthermore, the distribution of SGSP, as well as ages and infection types, was determined by using GraphPad Prism version 8.0.1.
Furthermore, in 34 non-pregnant patients, some underlying conditions had a higher prevalence: 10 patients (29.4%) had diabetes, 4 patients (11.8%) had Strain identification and phylogenetic analysis of the 16S rRNA gene All 45 isolates were positive for Streptococcus Lancefield antigen D grouping sera as examined by latex agglutination test.
Initial identification by the automated Phoenix system revealed that all isolates belonged to S. bovis biotype II. Nucleotide sequencing of 16S rRNA amplicons classified all S. bovis biotype II isolates as SGSP. Furthermore, gyrB sequencing also identified the isolates as S. pasteurianus. The phylogenetic analysis of the 16S rRNA gene (1422 bp) was performed by the neighbour-joining method between the 45 SGSP strains and the reference strain of SGSP species (Fig. 4) [19].

Antibiotic-resistant phenotypes and genotypes
Antimicrobial susceptibility results are shown in Table 2.
All the isolates were phenotypically susceptible to  Table 3. Moreover, in 43 (95.6%) out of 45 were tetracycline-resistant strains, 29 strains harbored tet(L), 27 strains tet(O) and 7 strains tet(M), singly or combined. One isolate harbored triple resistance genes of tet(O)/tet(L)/   Interestingly, all erythromycin-resistant isolates were also resistant to tetracycline, and both tetracycline-sensitive isolates were also sensitive to erythromycin.

Literature review
To better understand the features of SGSP infections worldwide, we searched MEDLINE database (https:// www.ncbi.nlm.nih.gov/pubmed) for the studies reporting the clinical infections caused by SGSP. Five reports including185 clinical SGSP isolates were included for comparison, and the details were summarized in Table 4 [4,11,14,20,21].

Discussion
This work is, to the best of our knowledge, the first comprehensive study on infective SGSP isolates in mainland China. The complexity of S. bovis taxonomy and relatively limited infection reports constrain clinical studies of SGSP, which is thus considered as an underreported opportunistic pathogen [10,14]. In the present study, we found that 22 out of 34 non-pregnancy-related subjects (64.7%) were elderly subjects with ages over 65 years. Interestingly, a study conducted in southern-central Israel reported that 75% bacteremia by S. bovis was over 65 years [20]. This, together with our data, suggests that elderly people are prone to SGSP infection. Moreover, we also observed that among 15 SGSP isolates (15/ 34, 44.1% non-pregnant subjects) recovered from UTI patients, 10 had diabetes. This observation was in line with a retrospective study in Italy, which reported that among 63.6% of patients (14/22) with UTI caused by SGSP, diabetes was the most common underlying disease (7/22, 31.8%) [4]. Another study in Spain also found that most S. bovis group isolates (72%) causing UTI were SGSP [2]. These observations thus collectively hint that SGSP can be taken as a potential pathogen in UTI, especially in those with diabetes [2]. Additionally, it should be noted that 62.5% patients (10/16) with SGSP bacteremia were male, demonstrating a correlation between gender and SGSP isolation from the urinary tract, as suggested in another two previous studies [4]. Together, the elderly, pregnant women and the immunodeficient population are the main people who are under the risk of SGSP infection. Bacteremia caused by SGSP was shown to be associated with malignancy of various parts of the digestive tract, including gastric, pancreatic, hepatobiliary and colorectal cancers [11,[22][23][24][25]. In our study, 1 case of malignant bile duct tumor, 1 case of endometrial carcinoma/colon cancer, 1 case of pancreatic tumor and 1 case of liver cancer were identified in non-pregnant-related infections, respectively. It was reported that all S. bovis strains from bile were likely associated with biliary tract malignancy [20], and SGSP was more frequently identified in the bacteremia with a biliary source (15/27 cases, 55.6%,) than S. infantarius (20/46, 43.5%) and S. gallolyticus subsp. gallolyticus (SGSG, 2/112, 1.8%) [23]. It should be noted, in colorectal carcinoma, a lower risk was noticed for SGSP, compared with SGSG at an odds ratio of 7.26 [26]. As a consequence, considering the association between S. bovis subspecies and specific pathogenesis, it is thus mandatory for every S. viridans organism isolated from the bloodstream to be identified into a species/subspecies level in order to distinguish SGSP from other S. bovis group members [4,10,22]. Additionally, SGSP bacteremia was observed to be less associated with IE than SGSG (8~29% in SGSP vs 43~100% in SGSG), too [22,26]. Only one case (1/16, 6.25%) of SGSP bacteremia was diagnosed with IE in the present study. Another observation was that the hematological diseases, including aplastic anemia (with hematopoietic stem cell transplantation) and acute lymphoblastic leukemia, were detected in two patients in our study, and this has been rarely documented previously [22]. The underlying mechanism remains elusive.
SGSP colonizes asymptomatically in the gastrointestinal and genitourinary tracts in pregnant women, and thus might potentially cause neonatal meningitis and bacteremia [7,8,27,28]. There is a very high one-year mortality rate of 58.7% in SGSP bacteremia [22]. The current study involves 9 cases (9/45, 20%) of intrauterine infections, and 2 cases (2/45, 4.4%) of bacteremia in pregnant women and neonates, hinting that SGSP is an important pathogen of pregnancy-related infection [8,[28][29][30]. One case of late-onset meningitis in a preterm male infant is detected in this study. Furthermore, our previous report found one case of intrauterine infection and post-partum bacteremia that was attributed to SGSP providing evidence of a possible portal of entry in cases of maternal or neonatal infection [8]. This potential infective pathway might be confirmed because more similar cases exist in this work. Altogether, we support the hypothesis that SGSP, which is different from other subspecies of S.bovis group, is a potential pathogen of maternal-fetal infection similarly to GBS [8].
Phenotypic variations always limit a correct identification of S. bovis species by the use of conventional microbiology and biochemical methods. However, in this study, Phoenix100 system identified most SGSP strains into subspecies level correctly, except the KT445 strain collected from fetal membrane in a 32year female, and that was misidentified as S. bovis I (Strep. group D) by Phoenix100, but confirmed as SGSP (S. pasteurianus strain CIP 107122) using 16S rRNA gene. Therefore, it is tempting to conclude that the classical biochemical methods are suitable and sufficient to fulfill clinical purposes. Treatment of SGSP infection, especially in meningitis, often includes intravenous penicillin, ampicillin and cefotaxime administration [7]. Considering all SGSP isolates were susceptible to penicillin [2], cefotaxime, vancomycin, meropenem, and chloramphenicol, the narrowest spectrum antibiotic penicillin should be considered as the drug of choice. In line with our results, this antibiotic choice should be recommended in mainland China. Furthermore, SGSP resistance rates varied for clindamycin, erythromycin, tetracycline and levofloxacin [11]. In our study, most SGSP isolates (40/45, 88.9%) simultaneously exhibited resistance to macrolides and clindamycin, dramatically higher than that of 31.8% in Italy [4] and 37.5% in Spain [14]. The resistance of SGSP isolates was due to the presence of either erm(B) and erm(T) genes or to a lesser extent mef(A/E) gene [4,31,32]. All erythromycin-resistant isolates in this study also displayed resistance to clindamycin, with the cMLS B resistance phenotype caused mainly by the erm(B) gene. While efflux-encoding mef(A/E) genes were only detectable in 5 isolates, singly (1 case) or combined with erm(B) (4 cases), which is different from previous reports in which erm(T) was found to be responsible for most macrolide resistance [12,31]. Among the reported SGSP strains in Italy, 68.2% (15/22) were tetracycline-resistant, and most of them harbored either tet(O) (10 cases) or tet(M) (4 cases) [4]. In the current study, however, 93.3% strains were tetracycline-resistant, most carried tet(L) gene, and less carried tet(O) and tet(M) genes, singly or combined, while no isolate carried tet(K) gene. This discrepancy might be explained by geographic and/or species differences. Furthermore, all erythromycin-resistant isolates were also resistant to tetracycline, similar to a previous study [4] . Taken together, these findings demonstrated that antibiotic resistance was widespread among SGSP clinical isolates, thus representing a serious problem particularly when the emerging infection rates are considered, especially in those allergic to β-lactam antibiotics.

Conclusions
In summary, this study on infective SGSP isolates circulating in mainland China underscores the clinical importance of this microorganism and provides valuable information about clinical features and epidemiological characteristics of SGSP. It is important to draw attention to this underreported opportunistic pathogen targeting immunodeficient populations, notably elderly subjects, pregnant women and neonates.  Z171100001017118) for specimen's collection and analysis. The funding bodies had no role in the design of the study and collection, analysis, and interpretation of data and in writing the manuscript.
Availability of data and materials I can confirm I have included a statement regarding data and material availability in the declaration section of my manuscript. All the data and material involved in the current study are available from the corresponding author on reasonable request.

Ethics approval and consent to participate
The institutional review boards at the Henan provincial people's hospital approved the study protocol. The written informed consent from participants was exempted, because it focused only on the epidemical features of SGSP strains, and the privacy of involved subjects was not affected.

Consent for publication
Not applicable.