The colonization of indwelling medical devices by biofilm-forming bacteria is one of the major causes of healthcare-associated infections (Percival et al., 2015). Staphylococcus epidermidis, a biofilm-forming commensal bacterium that inhabits human skin and mucosae, is considered one of most important causes of medical devices-related infections, being particularly associated with the use of intravascular catheters (Mack et al., 2013). Although S. epidermidis biofilms are classically associated with the development of chronic infections (Costerton et al., 1999), the release of cells from the biofilm has been associated with onset of acute infections such as embolic events of endocarditis (Pitz et al., 2011), bacteremia, or even septicemia (Cole et al., 2016). Bloodstream infections caused by S. epidermidis are typically indolent and difficult to eradicate significantly increasing patient's morbidity (Kleinschmidt et al., 2015) and mortality among immunocompromised (Khashu et al., 2006) and immunosuppressed patients (Bender and Hughes, 1980). In addition, the costs associated with the diagnosis and treatment of these secondary infections is estimated to be approximately $20,000 per occurrence (Kilgore and Brossette, 2008). Henceforth, it is imperative to redefine strategies for the management of the pathologic events associated with biofilm disassembly. Since bloodstream infections are one of the most frequent complications caused by S. epidermidis biofilm disassembly (Cole et al., 2016), a comprehensive analysis of the interplay between S. epidermidis biofilm-released cells (BRC) and hosts' blood components would be invaluable. Herein, as the first step toward the understanding of this interaction, we have characterized, using RNA sequencing (RNAseq) technology, the transcriptome of S. epidermidis BRC upon interaction with whole human blood, polymorphonuclear, or mononuclear leukocytes and plasma.

The transcriptomic profile of S. epidermidis BRC under the conditions described in this study was deposited in Gene Expression Omnibus (GEO) database, at NCBI, under the accession number GSE79948. Raw and trimmed/filtered datasets are available and can be downloaded as fastq and text files, respectively.

Bacteremia is one of the major clinical complications associated with the release of cells from S. epidermidis biofilms formed on intravascular catheters. These infections are associated with increased hospitalization periods, healthcare costs and patient morbidity, or even mortality (Kleinschmidt et al., 2015). Our goal was thus to gather the first insights into the interaction between S. epidermidis BRC and human blood, in order to better understand the role of these cells in the pathogenesis of S. epidermidis biofilm-related infections. To this end, we analyzed the transcriptome of BRC upon interaction with human blood or its cellular and soluble components, using an ex vivo model previously used for other microorganisms (Fradin et al., 2003; Mereghetti et al., 2008; Malachowa et al., 2011; França et al., 2014).

The transcriptome of S. epidermidis BRC changed significantly after incubation with whole human blood or its cellular or soluble components. Nevertheless, before analyzing the alterations observed, RNAseq results were confirmed using qPCR. As can be seen in Supplementary Figure 1, a strong correlation between both methods was observed. Thereafter, we compared the transcriptome of BRC incubated with human blood, PMNs, MNs, or plasma with the one of BRC incubated with TSB, a rich medium frequently used in in vitro assays. This comparison allowed the identification the genes whose transcription was affected only by the presence of biological factors. Within the genes found differentially transcribed (fold change >3 and P < 0.05, Kal's test with FDR) only 21–25% of the genes were found in BRC incubated with whole human blood or its cellular and soluble components and these were the genes considered for further analysis (Figure 1).

KEGG pathways analysis showed that the major alterations occurring in the presence of human blood circulating factors were associated with basic pathways involved in the biosynthesis or metabolism of amino acids and in the import or export of substances (Figure 2). In addition, KEGG pathways such as biosynthesis of secondary metabolites and phenylalanine, tyrosine, tryptophan, and folate biosynthesis were also found significantly enriched (P < 0.05, Hypergeometric test with FDR) but only in BRC incubated with both PMN and MN leukocytes (Figure 2).

As it would be expected in a low free iron environment such as human blood/plasma, iron transferrin receptors, iron ABC transporter permeases, and iron ABC transporter ATP binding proteins were found significantly enriched. Interestingly, biotin metabolism was also found significantly enriched (P < 0.05, Hypergeometric test with FDR) in all the conditions tested. In fact, the complete biotin operon, which is composed by the genes bioA, bioD, bioF, and bioW, was found highly transcribed, with fold change values ranging from 13 to 110. The same observation was made in S. aureus after short-term incubation with either blood or plasma (Malachowa et al., 2011). Biotin is an important cofactor involved in prokaryotic central pathways being particularly important during infection, as bacteria have a high demand for micronutrients (Streit and Entcheva, 2003). Due to the essential nature of biotin metabolism, the target of biotin metabolism-associated proteins has been under special attention, being considered a promising strategy to combat drug-resistant pathogens including S. aureus (Soares da Costa et al., 2012; Pendini et al., 2013; Paparella et al., 2014). No enrichment was found among genes with decreased transcription. Interestingly, in all the conditions tested, ~40% of the genes with down-regulated transcription encoded hypothetical proteins or pseudogenes.

Analyzing the effect of each of the blood components independently, using hierarchical cluster analysis, we found two major clades: one that separates BRC incubated with whole human blood from the remaining conditions and a second clade that separates BRC incubated with plasma from those incubated with either PMN or MN cells (Figure 2). Interestingly, this suggested that both PMN and MN leukocytes, despite their involvement in innate immunity and phagocytic capability, had little influence on S. epidermidis BRC gene expression profile. Also, these results indicated that the majority of the alterations observed in whole human blood occurred in plasma, as recently observed in a small panel of S. epidermidis genes (França and Cerca, 2016). We cannot, however, discard the possibility that a different effect on S. epidermidis gene expression profile could be seen if these cells were circulating in human blood, as crosstalk with and production of signaling molecules by other cells may have an important influence in S. epidermidis gene expression.

Overall, this study enabled us to identify the pathways involved in the adaptation of the bacterium to the stressful environment encountered in human blood consequently contributing to its survival and persistence. Hence, these results may be helpful in the selection of potential targets for future studies aiming to develop preventive and/or therapeutic strategies for S. epidermidis biofilm-based infections. One interesting target may be biotin metabolism-associated proteins, which are already being tackled in other important human pathogens.

GP, MV, and NC conceived the study and participated in its design and coordination. AF performed the experiments, collected, analyzed, and deposited the data. AF prepared the draft and GP, MV, and NC proofread the final draft. All authors have read and approved the manuscript.

This study was funded by the Portuguese Foundation for Science and Technology (FCT) by the project with the reference FCOMP-01-012014-FEDER-041246 (EXPL/BIA-MIC/0101/2013), the strategic funding of UID/BIO/04469/2013 unit, COMPETE 2020 (POCI-01-0145-FEDER-006684), BioTecNorte operation (NORTE-01-0145-FEDER-000004) funded by European Regional Development Fund under the scope of Norte2020 - Programa Operacional Regional do Norte. NC is an Investigador FCT. AF is supported by the FCT fellowship SFRH/BPD/99961/2014. The funders had no role in study design, data collection and interpretation, or decision to submit the work for publication.