FormalPara Key Summary Points

Why carry out this study?

The implementation of anti-pneumococcal conjugate vaccines has resulted in a significant decrease of the number of pneumococcal infections caused by vaccine serotypes; however, an increase in non-vaccine serotype 19A infections was observed in several countries using 7- and GSK 10-valent conjugate vaccines.

Unexpectedly, in Poland we noticed an increase of invasive infections caused by serotype 19A Streptococcus pneumoniae prior to the compulsory antipneumococcal vaccination of children.

The aim of the study was to characterize serotype 19A S. pneumoniae invasive isolates of GPSC1/CC320 circulating in Poland before the introduction of PCV into the Polish NIP and to compare them to isolates from other countries which implemented pneumococcal conjugated vaccines much earlier than Poland.

What was learned from the study?

The most common among Polish invasive 19A pneumococci were isolates of highly resistant GPSC1/CC320 clone, rarely identified elsewhere in Europe before and even after 7-valent vaccine introduction.

The majority of Polish 19A pneumococci of GPSC1/CC320 belong to a group genetically separate from other known worldwide representatives of that clone.

This particular epidemiological situation may be a consequence of our special conditions associated with a late vaccine introduction, long-lasting only partial coverage by vaccination via private market (i.e., lack of population-based vaccination), and a relatively strong selective pressure caused by antibiotic consumption.

Introduction

Streptococcus pneumoniae (pneumococcus) is an important cause of morbidity and mortality all over the world. It is also a common bacterium of the human commensal flora, as approximately 10% of healthy adults and up to 27–65% of healthy children may carry pneumococcus in their nasopharynx [1, 2]. S. pneumoniae isolates cause non-invasive diseases such as pneumonia, otitis media, and sinusitis; these are milder but more common presentations. They may also be responsible for more severe infections such as bacteremic pneumonia, meningitis, and sepsis. These are known as invasive pneumococcal disease (IPD) and have a high case fatality ratio. Several risk factors are associated with pneumococcal infections, but young children and adults over 65 are the most numerous among vulnerable groups [3].

The polysaccharide capsule constitutes one of the most important pneumococcal virulence factors. On the basis of its structure and antigenic differences, over 100 serotypes have been distinguished so far [4]. This division is crucial since all currently used pneumococcal vaccines are based on serotype-specific capsular polysaccharide antigens. The distribution of serotypes and their variability over time differ around the world. These phenomena occur naturally, but over the past 2 decades, changes in serotype circulation have been greatly influenced by the widespread children’s vaccination [3, 5]. The introduction of the seven-valent pneumococcal conjugate vaccine (PCV7) into the national immunization programs (NIPs) has significantly reduced the number of pneumococcal infections, especially invasive ones, caused by vaccine serotypes (VT) [6]. Since some VT pneumococci were frequently resistant to antibiotics, vaccination also resulted in decreased antibiotic resistance levels [7]. The spectacular reduction of pneumococcal infections was observed not only in vaccinated children but also in older age groups owing to herd immunity [3]. However, infections caused by isolates of non-vaccine serotypes (NVT) started spreading shortly thereafter [8]. That phenomenon called serotype replacement occurred as a result of NVT filling the gap left following the reduction of vaccine serotypes following the introduction of the vaccine. Additionally, new clones have also spread as a result of capsule/serotype switching, e.g., 19A isolates with resistance to penicillin and with the genetic background of PCV7 serotype 4 were identified, which were no longer influenced by the PCV7 vaccine [9]. Therefore, vaccines targeting a larger number of serotypes were introduced, including the GSK 10-valent pneumococcal conjugate vaccine (GSK PCV10) and the 13-valent pneumococcal conjugate vaccine (PCV13).

In many countries using PCV7 and GSK PCV10, serotype 19A pneumococci have been of particular importance among NVT isolates and in a relatively short period 19A strains have become the main cause of IPD burden, with the majority additionally being multidrug resistant (MDR) [5, 8, 10, 11]. In parallel, more and more differentiation was observed among pneumococci of the 19A serotype, manifested by a growing number of clonal complexes (CCs) detected worldwide [5, 12]. In many countries MDR CC320 of the 19A serotype has become the most prevalent [5]. The introduction and extensive use of whole-genome sequencing (WGS) in microbiological research has allowed the definition of Global Pneumococcal Sequence Clusters (GPSCs) for more accurate monitoring of pneumococcal clones [13]. On the basis of that system, CC320 corresponds to GPSC1. The introduction of PCV13 reduced the incidence of 19A IPD, including GPSC1/CC320 [5, 14, 15].

Compared to the majority of European countries, PCV was introduced to the NIP in Poland very late, only in 2017, for children born after December 31, 2016, using a 2 + 1 dose schedule (at 2 and 4 months of age plus a booster at 13–15 months). The selection of the vaccine for the NIP is carried out through tenders and so far, with a single exception, GSK PCV10 has been selected, and beside the private market PCV13 has been available for specific risk groups. Previously, PCVs (first PCV7, then GSK PCV10 and PCV13) were available free of charge only to children from selected risk groups and commercially on the private market. Although in several countries the dissemination of 19A isolates was mainly related to the PCV7 and GSK PCV10 usage, some authors reported the 19A rise before the PCV introduction or in countries with a very low vaccine coverage [16,17,18]. Also in Poland, the number and proportion of 19A invasive infections increased before introduction of PCV into NIP, and among 19A pneumococci, those belonging to GPSC1/CC320 turned out to be the most common (https://koroun.nil.gov.pl/dane-epidemiologiczne/) [19].

The aim of this study was to characterize serotype 19A S. pneumoniae invasive isolates belonging to GPSC1/CC320 (n = 97) and circulating in Poland during 2009–2016, i.e., before the introduction of PCV into the Polish NIP, and to compare them to isolates from other countries which implemented pneumococcal conjugated vaccines much earlier than Poland.

Methods

Bacterial Isolates and Their Identification

In Poland, cases of IPD are identified through two independent surveillance systems, a hospital-based surveillance system run by the National Institute of Public Health NIH—National Research Institute (NIPH NIH—NRI) and a laboratory confirmed surveillance system in the National Reference Centre for Bacterial Meningitis (NRCBM), described elsewhere [20]. All pneumococcal isolates collected during laboratory surveillance of invasive pneumococcal disease in Poland conducted by the NRCBM between 1997 and 2016 were identified on the basis of typical colony morphology, Gram stain, susceptibility to optochin (bioMerieux, Marcy l’Etoile, France), and bile solubility [21].

The study was conducted as continuous surveillance of the NRCBM and in accordance with the World Health Medical Association 1966 Declaration of Helsinki and the EU rules of Good Clinical Practice. The NRCBM was established by the Ministry of Health in 1997 to monitor invasive infections in Poland, including invasive pneumococcal disease. Currently the NRCBM acts under the Act of 5 December 2008 on preventing and combating infections and infectious diseases in humans (Dz. U. 2008 Nr 234 poz. 1570) and therefore institutional review board approval is not required.

Serotyping

Depending on the year, serotypes of all S. pneumoniae received between 1997 and 2016 were determined either by the Quellung reaction test (in the Statens Serum Institut, Copenhagen, Denmark or in the National Reference Center for Streptococci in Aachen, Germany), a Pneumotest-Latex kit (Statens Serum Institut, Copenhagen, Denmark), or by PCR [22, 23]. As the NRCBM implemented serotyping with the Quellung reaction test in 2020, all isolates initially identified as 19A were re-serotyped using this method [24].

Antimicrobial Susceptibility Testing

Antimicrobial susceptibility was determined using the broth microdilution method “in house” according to the EUCAST guidelines [25]. Only susceptibility to ceftaroline was tested using MIC Test Strips (Liofilchem, Roseto degli Abruzzi, Italy). Twelve antimicrobial groups were tested: penicillins (penicillin, ampicillin, amoxicillin), cephalosporins (cefuroxime, cefotaxime, cefepime, ceftaroline), carbapenems (meropenem), glycopeptides (vancomycin), fluoroquinolones (levofloxacin, moxifloxacin), tetracyclines (tetracycline, doxycycline), phenicols (chloramphenicol), folate pathway antagonists combinations (co-trimoxazole), macrolides (erythromycin), lincosamides (clindamycin), oxazolidinones (linezolid), and rifamycins (rifampicin). The double disc test was additionally used to establish erythromycin and clindamycin resistance phenotypes, including efflux-mediated resistance (M-phenotype), constitutive MLSB (cMLSB), and inducible MLSB (iMLSB) phenotypes. The results were interpreted according to the EUCAST guidelines [25]. When available, meningitis breakpoints were used. The ATCC 49619 strain of S. pneumoniae was used as a quality control. Multiple drug-resistant (MDR) isolate was defined as resistant to at least one agent from three or more antimicrobial groups, respectively [26, 27].

DNA isolation, WGS, and Sequence Assembly

Genomic DNA was purified using the SDS/phenol method [28]. DNA quality was assessed by measuring the ratio of absorbance at 260 nm and 280 nm and by electrophoresis in 0.8% agarose gels (Agarose BASICA LE, Prona). Short-read bacterial genome sequencing was performed using the MiSeq instrument (Illumina, San Diego, CA). The template concentration was determined using Qubit fluorimeter (Thermo Fisher Scientific, Waltham, MA). DNA libraries were constructed using the Nextera XT (Illumina, San Diego, CA) and NEB Ultra II FS kits (New England Biolabs, Ipswich, MA), followed by Paired End 300 cycle sequencing, targeting at least 50 × genome coverage.

Sequence quality metrics were assessed using the FASTQC software v.0.11.8 (https://www.bioinformatics.babraham.ac.uk/projects/fastqc/). Raw sequencing reads were trimmed for quality and residual library adaptors were removed using the fastp software [29]. Cleaned Illumina reads were assembled into contigs using SPAdes v.3.13.1 [30].

Data Analysis

Sequence types (STs), ribosomal sequence types (rSTs), and core genome sequence types (cgSTs) were determined with the use of the BIGSdb database (https://pubmlst.org/software/bigsdb/; date last accessed 1 April 2023) [31]. Isolates differing by up to one allele from ST320 belonged to the clonal complex CC320. Global Pneumococcal Sequence Clusters (GPSCs) were defined using PopPUNK included in the Pathogenwatch v. 20.0.15 online platform [13, 32].

Serotype 19A capsule subtype analysis was performed in silico [33]. Geneious Prime 2023.0.4 (https://www.geneious.com) was used and to analyze antimicrobial determinants such as mutations in gyrA, gyrB, parC, parE (fluoroquinolone resistance), folA, folP (co-trimoxazole resistance) as well as structure and localization of Tn2010 carrying erm(B) and tet(M) [AB426620] [34]. The deduced amino acid sequences of gyrA, gyrB, parC, parE, folA, and folP were compared with their counterparts in the R6 strain [35]. Local BLAST searches for genes associated with virulence in pneumococci [36], using genes from the 3641/15 isolate (PubMLST id 112,577) as queries [37], were done in Geneious.

Polish GPSC1/CC320 pneumococci were clustered with GPSC1/CC320 19A isolates from other countries with genomic sequences available in the PubMLST Pneumococcal Genome Library (date last accessed 3 April 2023) [31]. Phylogenetic trees were constructed using a BIGSdb plugin Grape Tree (https://pypi.org/project/grapetree/).

The Fisher exact test or the chi-square test were used, depending of sample size, to compare difference in proportions; a p value ≤ 0.05 was considered as significant.

Results

Polish 19A Pneumococci and Characteristics of Patients Infected with GPSC1/CC320 Isolates

During the 1997–2016 study period, the NRCBM collected 4237 S. pneumoniae isolates responsible for invasive infections in Poland. Among them, 200 (4.7%) isolates belonged to serotype 19A, with the first isolate of this serotype registered in 2000. The MLST analysis revealed that CC320 isolates were most prevalent among 19A pneumococci, with 97 isolates accounting for 48.5% of the group. CC320 corresponds to GPSC1 [13], and this assignation was further confirmed by the cgMLST analysis (see below). The first GPSC1/CC320 isolate was obtained by the NRCBM in 2009. The remaining 19A serotype isolates belonged to CC230 (67 isolates, 33.5%), CC199 (26 isolates, 13.0%), CC193 (three isolates, 1.5%), CC63 (two isolates, 1.0%), and CC81 (one isolate, 0.5%); four isolates were not assigned to any known CC (Fig. 1). The distribution of the STs and CCs of 19A S. pneumoniae responsible for invasive infections among age groups of Polish patients is presented on Fig. 2. The number and percentage of 19A isolates increased significantly from 1.3% (4/299) to 4.9% (13/266, p < 0.024) between 2009 and 2010, and that higher percentage of 19A was maintained throughout the subsequent years until 2016 (6.3%, 42/672, Fig. 1). Between the periods 2009–2012 and 2013–2016, the number of GPSC1/CC320 also increased (Fig. 1) from 43.3% among all 19A isolates (26/60) to 52.2% (71/136); this increase, however, was not statistically significant. Pneumococci of this complex were isolated at hospitals located in 15 from 16 Polish regions (1–17 isolates/region). Most of the GPSC1/CC320 pneumococci were isolated from blood (74.2%, 72/97), followed by cerebrospinal fluid (24.7%, 24/97) and pleural fluid (1.0%, 1/97). Meningitis, bacteremic meningitis, or meningitis with sepsis were diagnosed in 32.0% (31/97) of patients, 25.8% (25/97) had bacteremic or septic pneumonia, 23.7% (23/97) sepsis, and the remaining 18.6% (18/97) other or unspecified manifestations of IPD. Of the patients, 32.0% (31/97) were under 5 years of age and 28.9% (28/97) were over 65 years of age. General sex ratio (M/F) was 1.9 (63:34), whereas in children under 5 years of age it was 2.1 (21:10), and in the age group > 65 years old it was 1.2 (15:13). Among 31 children under 5 years of age one was vaccinated with four doses of GSK PCV10; 20 children were not vaccinated and for 10 children data concerning vaccination was not reported. There were 21 fatal cases among patients with infections caused by GPSC1/CC320 pneumococci; of these, 17 (81.0%) were diagnosed with sepsis alone or in combination with meningitis and pneumonia. The case fatality ratio (CFR) was 23.6% (21/89) or 21.6% (21/97), taking into account cases with known outcomes only or all cases, respectively. The CFR for all other invasive pneumococci collected in the NRCBM between 2013 and 2016 CFR was 32.8% and 23.4%, respectively. Differences in CFRs between GPSC1/CC320 and other pneumococci were not significant. The highest CFR among GPSC1/CC320 cases with known outcome was observed in the age group 40–64 years old (41.7%, 10/24), followed by patients older than 65 years (30.8%, 8/26). In children under five the CFR was 10.0% (3/30).

Fig. 1
figure 1

Distribution of the clonal complexes (CC) among invasive S. pneumoniae of serotype 19A and percentage of 19A isolates among all pneumococci responsible for invasive infection in Poland, 1997–2016 (in the case when isolates were not assigned to any known CC, the sequence type (ST) was provided)

Full size image
Fig. 2
figure 2

Distribution of the sequence types (STs) and clonal complexes (CCs) of S. pneumoniae of serotype 19A responsible for invasive infections among age groups of Polish patients, 1997–2016

Full size image

Sequence Types, Ribosomal Sequence Types, and the cps Locus

Three STs were identified among the GPSC1/CC320 isolates. The most numerous ST320 was represented by 75.2% (73/97) of isolates, followed by ST4768 (23.7%, 23/97) and ST15047 (1.0%, 1/97). ST4768 and ST15047 are single locus variants (SLVs) of ST320, differing by the xpt alleles. Twelve ribosomal sequence types (rSTs) were recognized with the most frequent being rST589 (78 isolates), followed by rST572 (8 isolates), rST119912 (2 isolates), and nine other rSTs represented by single isolates (Supplementary Material). All isolates carried the cps locus specific for 19A serotype in agreement with the results of serotyping obtained using the Quellung test, and belonged to the 19A-III cps subtype [33].

Antimicrobial Susceptibility and Determinants of Resistance

Antimicrobial susceptibility data are provided in Table 1 and Supplementary Material. All 97 isolates were resistant to penicillin using the meningitis breakpoints (MIC > 0.06 mg/L) and 87 isolates were resistant to this antibiotic using also the breakpoint for indications other than meningitis (MIC > 2.0 mg/L). The remaining 10 isolates had MIC = 2.0 mg/L. Resistance to cefotaxime using meningitis breakpoints also characterized all isolates, including eight isolates with cefotaxime MIC > 2.0 mg/L. All isolates of our study had the same allele 13 of pbp1a gene and 90 isolates (92.8%) had the pbp profile (pbp1a-pbp2b-pbp2x) 13-11-16. Another variant of pbp2b (allele 14) was found in three isolates from the same region of the country (the Podkarpackie Voivodeship). The pbp2x gene was the most diverse, with four variants, including two new ones, differing by single nucleotide substitutions from the main allele 16 (Supplementary Material). All five isolates with the allele 26 of pbp2x demonstrated the highest MIC of cefotaxime (≥ 8.0 mg/L) among all GPSC1/CC320 pneumococci. These isolates harbored several substitutions in the deduced amino acid sequences of PBP2x, including the M339F and M400T changes, typically associated with increased levels of β-lactam resistance [38].

Table 1 Susceptibility data of serotype 19A Streptococcus pneumoniae isolates of Global Pneumococcal Sequence Clusters/clonal complexes GPSC1/CC320 (n = 97) responsible for invasive infections in Poland, 2009–2016
Full size table

All isolates were resistant to erythromycin; of these, 96 represented MLSB phenotype and a single isolate demonstrated inducible MLSB phenotype. The presence of both mef(E) and erm(B) genes was characteristic for the whole group. Ninety-six isolates were tetracycline resistant but retained doxycycline susceptibility. While tet(M) was present in all isolates, a single tetracycline-susceptible isolate had a 2-bp deletion at the positions 1004–1005 nt of the gene, resulting in a frameshift and TetM presumably shorter by 300 amino acid residues. In all isolates erm(B), mef(E), and tet(M) were localized within the Tn2010-type element [34], inserted between counterparts of the spr1764 and spr1772 genes in the R6 genome. In the case of 73 isolates, the transposon sequence was identical to the original Tn2010 (Li 2011) with the exception of a silent SNP within the tet(M) gene. Thirteen isolates harbored one or two additional SNPs in various parts of the transposon. Three isolates, including the isolate with iMLSB phenotype, had a 42-bp deletion in the 3′ terminal part of ORF20, located upstream the leader peptide ORF of erm(B), three isolates had a 77-bp sequence replaced by a 4-bp fragment in ORFP3, two isolates had a 184-bp deletion at the beginning of ORF15, and one isolate had a 7-bp deletion in ORF21. In the case of a single isolate, the presence of 23 SNPs in ORF16 suggested a local recombination event.

All isolates demonstrated an intermediate susceptibility to levofloxacin and susceptibility to moxifloxacin. While several differences in deduced amino acid sequences of the gyrA, gyrB, parC, and parE genes in comparison to R6 were observed (details in Supplementary Material), none of the isolates harbored changes typically associated with fluoroquinolone resistance, such as substitutions in the quinolone-resistance determining regions (QRDRs) of GyrA and ParC or in the EGDSA and P(I/L)RGK motifs of GyrB and ParE [39]. All but two isolates had the I460V mutation in ParE, which was previously reported as a change involved in reducing susceptibility to fluoroquinolones [40]. All isolates were fully resistant to co-trimoxazole and had the I100L substitution in FolA and the 65EI66 duplication in FolP amino acid sequences (details in Supplementary Material) [41, 42]. In summary, all Polish GPSC1/CC320 isolates were resistant to penicillins, cephalosporins, carbapenems, macrolides, and trimethoprim-sulfamethoxazole, and were MDR. Compounds retaining activity to all isolates of this CC included moxifloxacin, doxycycline, rifampicin, linezolid, and vancomycin.

GPSC1/CC320 Virulome

The distribution and variability of investigated virulence-associated factors and regulators are summarized in Table 2 and details are provided in Supplementary Material. Genes of certain factors, such as LPxTG proteins ZmpC, ZmpD, PsrP, choline-binding proteins (CBPs) CbpF, CbpG, CbpI, and the non-canonical surface-exposed protein NanC were not found among Polish GPSC1/CC320 isolates. Some genes, such as cbpC, cbpJ, and pcpA were absent in a few (1–3) isolates. Genes bgaA, spGH101, nanA, pfbA, prtA, hysA, pce, piaA, ppiA, psaA, ppmA, phtE, gap, pavA, nanB, and pppA were present in the whole group and showed 100% sequence identity among isolates. Several other genes demonstrated the presence of various changes such as SNPs, indels, larger deletions, and duplications; some of these changes, in particular in pclA, zmpA, lytC, pcpA, sp_0667, pbp1b, and fic-like resulted in variants of genes encoding truncated forms of proteins or presumably a formation of pseudogenes in some isolates. In four isolates an additional copy of the lytA gene was detected and it was associated with the presence of 38.1–41.5 kb phages in three isolates and a 38.4 kb prophage in a single isolate (details in Supplementary Material). The variability of certain genes, such as pavB, zmpB, lytB, pcpA, pspA, pspC, phtA, phtB, ply, mgrA, and psaR could not be analyzed in some isolates because of problems with Illumina assemblies, especially in regions containing repeats, and in consequence splitting of a gene into two contigs or appearance of N calls in the sequence.

Table 2 Distribution of pneumococcal virulence-associated factor and regulator genes among Polish GPSC1/CC320 isolates
Full size table

All isolates had complete loci for both PI-1 and PI-2 pili types. All PI-1 loci were 99.9–100% identical and belonged to clade 1, characteristic for Spain9V-ST156 and Taiwan19F-ST236 international clones [43]. However, only nine isolates carried the wild-type regulator gene rlrA and the remaining isolates had the C1117T point mutation, not reported before, resulting in the appearance of the TAG stop codon and truncation of RlrA by 137 amino acid residues. One of these isolates additionally contained a 2-bp novel deletion at the position 1468–1469 bp in the rrgA gene, encoding the pilus tip adhesin, resulting in a frameshift and truncated RrgA. Another isolate contained several SNPs in rrgA and rrgB, the latter responsible for pilus backbone biosynthesis. This isolate had also the C529T mutation in the sortase gene srtD resulting in appearance of a stop codon and a truncated form of SrtD. In the PI-2 locus, all isolates had a characteristic stop codon in the pilus tip protein gene pitA [44]. Ninety-one isolates harbored the identical PI-2 locus. Three isolates harbored a new single-nucleotide deletion at the position 428 of sipA, leading to shortening of SipA (a presumable peptidase involved in pilus assembly) by 27 amino acid residues. One isolate had a novel extensive 513-bp in-frame deletion in pitB, encoding pilus backbone protein. In another isolate, a novel single-nucleotide deletion at the position 49 of pitB was found, presumably resulting in a pseudogene.

Clonal Relationships Among Polish GPSC1/CC320 and Their Relationships with Other Pneumococci of this Complex

The cgMLST approach was used to study the relationships among isolates from our collection (Fig. 3). Isolates belonging to STs 4768 and 15074 formed a cluster separate from ST320 isolates. Another cluster, containing nine isolates, was observed among ST320 pneumococci; except one, all of them were from one region (Kujawsko-Pomorskie), including five from one city, and were isolated between 2010 and 2016.

Fig. 3
figure 3

Phylogenetic tree of 97 Polish GPSC1/CC320 S. pneumoniae isolates of serotype 19A, constructed based on cgSTs

Full size image

The clustering with other 300 GPSC1/CC320 pneumococci of 19A, circulating globally, with genomic sequences available in the PubMLST Pneumococcal Genome Library (accessed 3 April 2023), revealed a separate cluster containing predominantly Polish isolates from the whole study period and, at the same time, grouping the vast majority (89 isolates) of our collection. The remaining eight Polish isolates belonged to another relatively large cluster, together with isolates from other countries (Fig. 4).

Fig. 4
figure 4

Phylogenetic tree of 97 Polish GPSC1/CC320 S. pneumoniae isolates of serotype 19A clustered with 300 isolates of GPSC1/CC320 and serotype 19A from other countries with genomic sequences available in the PubMLST Pneumococcal Genome Library (date last accessed 3 April 2023). Tree was constructed based on cgSTs

Full size image

Discussion

Results of our study showed that the number of invasive infections caused by 19A pneumococci already increased in Poland before the introduction of PCV in the NIP. In the majority of countries such an increase was observed only after PCV7 or GSK PCV10 introduction. For example, in the USA, the first country to introduce mass vaccination of children with PCV7 in 2000, an increase of 19A was detected in eight pediatric hospitals, from 24% of NVT in 2003 to 46% in 2007–2008 [45]. The rise of 19A infections has also been observed in other countries which have introduced PCV7 and GSK PCV10 in their NIPs, and moreover this pneumococcal serotype is one of the most commonly associated with invasive infections [5, 8, 46,47,48]. However, in some cases, the increase in 19A infections was not always associated with vaccination pressure and such exceptions were previously described. In South Korea, 19A IPD in children under 5 years of age increased from 0% in 1991–1994 to 26% in 2001–2003, with CC320 responsible for this rise [16]. In Israel, a 63.1% increase of 19A was found in acute otitis media in Bedouin children from 1999–2001 to 2004–2006 but it was not related with an expansion of a single clone [17]. Instead, natural epidemiological fluctuations, antibiotic prescribing policy, and the emergence of new epidemic clones were suggested as possible reasons for such an increase [16, 17]. In contrast to other European countries, in Poland PCV7 was available commercially on a voluntary basis and compulsory vaccination with GSK PCV10 was introduced in 2017, for all children born after December 31, 2016. In most Western European countries, such vaccination with PCV7 was implemented between 2006 and 2008, and shortly after (2009–2011) it was replaced with the higher-valent vaccines, GSK PCV10 or PCV13 [49].

In our study, GPSC1/CC320 isolates were most common among Polish invasive 19A isolates before the introduction of PCV, accounting for almost half of the cases caused by this serotype. All GPSC1/CC320 pneumococci of our study were MDR, but according to some authors such pneumococci should even be considered extensively drug-resistant [26, 27]. The determinants of resistance, in particular the main pbp profile 13-11-16, mutations in folA and folP specifying co-trimoxazole resistance and the presence of Tn2010 harboring erm(B), mef(E), and tet(M) were shared with other representatives of GPSC1/CC320 from Japan characterized recently in detail [50]. Further, the spread of GPSC1/CC320 pneumococci in Poland may lead to treatment challenges and failures. It was reported that patients with 19A bacteremia require 1.4 times longer antimicrobial treatment than patients with invasive infections caused by pneumococci of other serotypes [51].

ST320 is a double-locus variant of ST236, characteristic of the previously defined Taiwan19F-14 clone, which likely originated in the late 1980s by acquiring a number of antimicrobial resistance determinants and became distributed worldwide [26, 27, 52,53,54]. Subsequently, one of at least four independent capsule switching events gave rise to the highly successful ST320 clade [55]. GPSC1/CC320 includes isolates of various serotypes with the highest prevalence of serotypes 19A and 19F. In the PubMLST database for S. pneumoniae (date accessed 1 February 2023), among 793 isolates of ST320 with known serotypes, serotypes 19A and 19F characterized 687 and 102 isolates, respectively. GPSC1/CC320 isolates became widely distributed in the Americas, including Canada, the USA after the PCV7 introduction, and Brazil after the GSK PCV10 introduction [5, 10, 52,53,54, 56]. In 2007, over 82% of the 19A isolates tested in the USA belonged to GPSC1/CC320 which represented the fastest growing CC associated with 19A serotype in the country [52, 53, 57]. In 2008–2009, CC320 pneumococci also predominated in 10 Asian countries, including countries lacking PCV7 or with its low coverage [18]. Also in Bulgaria, CC320 was the most common in the post-GSK PCV10 period [48]. By contrast, isolates of CC320 were found extremely rarely before the PCV era and were also infrequent after PCV introduction in most of Europe, where 19A pneumococci were generally more heterogeneous. In Italy, Germany, Czechia, and Norway in the post-PCV7 period, the increase of 19A isolates was related mostly to the spread of unrelated CC199. Pneumococci of CC230 and CC63 were also quite often found in Europe at that time [5]. In Germany, the first three CC320 infections were observed in 2008–2009 in children who entered the country from the USA, UK, and India [58]. We do not know how and when the first GPSC1/CC320 isolates appeared in Poland, but the majority of them form a clearly distinct cluster from other 19A pneumococci of this GPSC/CC, based on the cgMLST data available in the PubMLST Pneumococcal Genome Library (Fig. 4) [31]. Only three isolates from the USA dating back to 2008 and 2009 and single isolates from Germany, China, and Cambodia were also included in this distinct cluster. The remaining eight Polish isolates belonged to another relatively large cluster, dominated by isolates from the USA, but containing also isolates from several other countries such as Iceland, Germany, Belarus, Israel, China, India, Cambodia, South Africa, Mexico, and Peru. The two oldest isolates in this cluster were obtained in 2004 in the USA (Fig. 4).

Such country-specific clusters within GPSC1/CC320 were found also in Ireland and Japan [50, 59]. In Ireland, the unique GPSC1/CC320 sub-clade harboring a rare variant of the galE gene, whose product is involved in the biosynthesis of capsular polysaccharide, was responsible for vaccine failures, being more likely associated with the IPD in vaccinated children than other 19A genotypes [59]. The cluster observed in our study included a subcluster grouping 23 representatives of ST4768, which were found in eight Polish regions, but the majority of them (57%) were responsible for infections in the Małopolska region. ST4768 pneumococci are found outside our country only sporadically, as single isolates in the USA and Brazil (PubMLST database, date accessed 1 February 2022), and in Sweden [27].

Among Polish GPSC1/CC320 the instances of recombination events were observed, as mentioned before, in pbp genes and in Tn2010 as well as in several virulence-associated genes such as cbpJ, gnd, lytC, rrgA, spuA, stkP, and strA (Supplementary Material). Certain recombinational hotspots have been observed in main pneumococcal CCs, encompassing especially genes associated with antimicrobial resistance (in particular pbp genes), mobile genetic elements, and genes influencing the antigenic properties of the bacterial cell (e.g., pspA) [55, 60, 61]. Single-nucleotide substitutions, which may have arisen by either mutation or recombination, as well as deletions and insertions of various extent were an additional source of variability of these genes observed in our study. Importantly, several such changes resulting in stop codons or frameshift mutations rendered genes associated with virulence either inactive or presumably producing truncated proteins. Such a phenomenon is believed to represent a response to selective pressure from the host immune system [61]. Especially striking examples of such mutations were observed in our study for pili gene clusters. In the case of PI-1, the majority of isolates had a stop codon within the rlrA gene, responsible for a biphasic expression of pili genes [62]. Changes were also observed in other genes of the PI-1 operon and in the PI-2 operon. Pili promote adhesion to host tissues and colonization, and have a role in the pathogenesis [63, 64] but also stimulate an immune response [65].

As emphasized above, in most countries the number of 19A infections increased only after PCV was introduced, and moreover, isolates of GPSC1/CC320, the most common serotype 19A CC in Poland, were rare in Europe both before and after mass vaccination of children. In our opinion, this dissimilarity may be a consequence of the very late introduction of pneumococcal vaccination to the Polish NIP as described above, long-term low vaccine coverage, parallel use of different vaccines in the country, and high antibiotic consumption. Long-term partial coverage by vaccination available free of charge only for some risk groups such as prematurely born and immunodeficient children, but otherwise only at the patient’s own expense, allowed pneumococci access to a wide genetic pool and to gradually adapt to vaccine selective pressure. For example, in the years 2010, 2011, and 2012, 139,039, 159,166, and 173,248 children under the age of 15, respectively, completed their primary or reinforcing vaccination, with an estimated coverage rate of approximately 15% [66]. The selective pressure caused by high antibiotic consumption should also be considered as a factor beyond the dissemination of GPSC1/CC320 in the country since it is a known main driver of the emergence and spread of resistance not only in the pneumococcal population [67]. Of note, between 1998 and 2015 Poland belonged to 10 European countries with the highest antibiotic consumption at the community level [68].

Our study has a certain limitation. The Polish laboratory surveillance system of IPD is voluntary and during the research period, especially after 2008, the performance of the system has clearly improved. This may have affected the number of 19A isolates submitted to the NRCBM, but most likely affected all other serotypes in the same manner; therefore, we do believe that our results reflect some real epidemiological trends among Polish pneumococci.

Conclusions

Before introduction of PCV in the Polish NIP we noticed an unexpected increase of serotype 19A in invasive pneumococcal infections, with the most common being representatives of highly drug-resistant GPSC1/CC320 clone, rarely identified in Europe both before and even after PCV introduction. That dissimilarity may be a consequence of the late introduction of pneumococcal vaccination to the Polish NIP, long-lasting only partial coverage by vaccination via private market, parallel use of different vaccines in the country, and a relatively strong selective pressure caused by high antibiotic consumption. On the basis of the experience of other countries after the introduction of PCV7 and GSK PCV10, and taking into account that Poland mainly uses GSK PCV10 in its NIP, the circulation of 19A strains, in particular the virulent and highly resistant GPSC1/CC320 isolates, should be constantly monitored.