Genomic and Epidemiological Features of Two Dominant Methicillin-Susceptible Staphylococcus aureus Clones from a Neonatal Intensive Care Unit Surveillance Effort

Abstract

Methicillin-susceptible Staphylococcus aureus (MSSA) is a more prevalent neonatal intensive care unit (NICU) pathogen than methicillin-resistant S. aureus (MRSA). However, the introduction and spread of MSSA, the role of systematic decolonization, and optimal infection prevention and control strategies remain incompletely understood. We previously screened infants hospitalized in a university-affiliated level III to IV NICU twice monthly over 18 months for S. aureus colonization and identified several prevalent staphylococcal protein A (spa) types. Here, we performed whole-genome sequencing (WGS) and phylogenetic comparisons of 140 isolates from predominant spa types t279, t1451, and t571 to examine possible transmission routes and identify genomic and epidemiologic features associated with the spread of dominant clones. We identified two major MSSA clones: sequence type 398 (ST398), common in the local community, and ST1898, not previously encountered in the region. ST398 NICU isolates formed distinct clusters with closely related community isolates from previously published data sets, suggesting multiple sources of acquisition, such as family members or staff, including residents of the local community. In contrast, ST1898 isolates were nearly identical, pointing to clonal expansion within the NICU. Almost all ST1898 isolates harbored plasmids encoding mupirocin resistance (mupA), suggesting an association between the proliferation of this clone and decolonization efforts with mupirocin. Comparative genomics indicated genotype-specific pathways of introduction and spread of MSSA via community-associated (ST398) or health care-associated (ST1898) sources and the potential role of mupirocin resistance in dissemination of ST1898. Future surveillance efforts could benefit from routine genotyping to inform clone-specific infection prevention strategies. IMPORTANCE Methicillin-susceptible Staphylococcus aureus (MSSA) is a significant pathogen in neonates. However, surveillance efforts in neonatal intensive care units (NICUs) have focused primarily on methicillin-resistant S. aureus (MRSA), limiting our understanding of colonizing and infectious MSSA clones which are prevalent in the NICU. Here, we identify two dominant colonizing MSSA clones during an 18-month surveillance effort in a level III to IV NICU, ST398 and ST1898. Using genomic surveillance and phylogenetic analysis, coupled with epidemiological investigation, we found that these two sequence types had distinct modes of spread, namely the suggested exchange with community reservoirs for ST398 and the contribution of antibiotic resistance to dissemination of ST1898 in the health care setting. This study highlights the additional benefits of whole-genome surveillance for colonizing pathogens, beyond routine species identification and genotyping, to inform targeted infection prevention strategies.

Keywords: MSSA; NICU; genomic surveillance; methicillin-susceptible Staphylococcus aureus; mupirocin resistance; neonatal intensive care unit (NICU).

FIG 1

ST398 and ST1898 epidemiologic curves during an 18-month NICU surveillance effort. (A) Number of newly identified infants with ST398 (harboring spa types 1451 or 571) (red) and total number of ST398 cases per month (blue). There were multiple months with no new cases of ST398, and no ST398 isolates were identified beyond April 2018. (B) Number of newly identified infants with ST1898 (purple) and total number of ST1898 (harboring spa t279) cases per month (blue). In every month of the surveillance period, except January and June 2017, one to eight new cases of ST1898 were identified.

FIG 2

Phylogenetic tree of NICU and local community MSSA ST398. MSSA isolates from our NICU which were identified as ST398 using WGS (n = 31) were mapped against a previously published ST398 reference genome (GenBank CP003045) isolated from within the local community, along with publicly available short reads from 271 ST398 isolates also collected from our local household study (25). Concatenated SNPs covering the core chromosome of the reference genome totaled 5,338 bp, with pairwise distances between isolates ranging from 0 to 298 SNPs (median, 78 SNPs). Maximum-likelihood phylogenetic reconstruction was performed using RAxML with 100 bootstrap replicates (48). (A) Unrooted phylogenetic tree of all NICU (blue branches) and community (black branches) MSSA ST398 isolates reveals that many NICU isolates cluster with community isolates, suggesting possible introduction of these isolates from the community rather than the NICU. (B) The MSSA ST398 phylogeny, rooted at the reference isolate. The inner ring denotes isolate spa type, and colors along the outer ring represent individual infants to whom each NICU isolate belongs. The largest NICU cluster, comprising 11 isolates from 5 patients, is marked cluster 1 and was used for targeted transmission analysis (Fig. 3).

FIG 3

Contact tracing for MSSA ST398 spa type 1451 phylogenetic cluster. Temporal and spatial distribution of a subset of infants harboring MSSA ST398, identified as cluster 1 (Fig. 2). Whole-genome sequencing (WGS) revealed that 11 ST398 isolates from the 5 infants shown, all spa type 1451, were phylogenetically more closely related than other isolates (median pairwise distance, 6 SNPs). Rooms within the NICU are coded by color, with similar shades corresponding to different units. ST398 cultures (circles), non-ST398 MSSA cultures (diamonds), and negative surveillance swabs (Xs) are shown for each infant throughout the study period. Figure S1 shows the spatiotemporal distributions for all infants who harbored MSSA ST398 throughout the 18-month surveillance.

FIG 4

Phylogenetic reconstruction of NICU clonal complex CC15. WGS was performed on 106 CC15 isolates collected from our NICU, belonging to sequence types ST1898 and ST15. (A) The 106 CC15 isolates clustered by both spa type (inner ring) and sequence type (ST, middle ring). Screening for the high-level mupirocin resistance gene mupA (also referred to as ileS-2) revealed that this gene was only present in CC15-t279 isolates, with one CC15-t279 isolate containing a mupA variant allele and two CC15-t279 isolates missing the mupA gene. (B) We next generated a phylogenetic tree specifically for the 94 CC15-t279 isolates. Serial isolates from the same infant (colored dots on branch ends; no dot indicates singleton isolates) did not necessarily cluster together, and some isolates from vastly separated collection dates (outer ring) were closely related, highlighting the close relationships between CC15-t279/ST1898 isolates and indicative of local proliferation within the NICU.