Genotypic and Phenotypic Characterization of Staphylococcus aureus Isolates from the Respiratory Tract in Mechanically-Ventilated Patients

Abstract

Staphylococcus aureus is a commensal and frequent colonizer of the upper respiratory tract. When mechanical ventilation disrupts natural defenses, S. aureus is frequently isolated from the lower airways, but distinguishing between colonization and infection is difficult. The objectives of this study were (1) to investigate the bacterial genome sequence in consecutive isolates in order to identify changes related to the pathological adaptation to the lower respiratory tract and (2) to explore the relationship between specific phenotypic and genotypic features with the patient's study group, persistence of the clinical isolate and clinical outcome. A set of 94 clinical isolates were selected and corresponded to 34 patients that were classified as having pneumonia (10), tracheobronchitis (11) and bronchial colonization (13). Clinical strains were phenotypically characterized by conventional identification and susceptibility testing methods. Isolates underwent whole genome sequencing using Illumina HiSeq4000. Genotypic characterization was performed with an in-house pipeline (BacterialTyper). Genomic variation arising within-host was determined by comparing mapped sequences and de novo assemblies. Virulence factors important in staphylococcal colonization and infection were characterized using previously established functional assays. (1) Toxin production was assessed using a THP-1 cytotoxicity assay, which reports on the gross cytotoxicity of individual isolates. In addition, we investigated the expression of the major virulence factor, alpha-toxin (Hla) by Western blot. (2) Adhesion to the important extracellular matrix molecule, fibronectin, was determined using a standardized microtitre plate assay. Finally, invasion experiments using THP-1 and A539 cell lines and selected clinical strains were also performed. Repeated isolation of S. aureus from endotracheal aspirate usually reflects persistence of the same strain. Within-host variation is detectable in this setting, but it shows no evidence of pathological adaptation related to virulence, resistance or niche adaptations. Cytotoxicity was variable among isolates with 14 strains showing no cytotoxicity, with these latter presenting an unaltered Fn binding capacity. No changes on cytotoxicity were reported when comparing study groups. Fn binding capacity was reported for almost all strains, with the exception of two strains that presented the lowest values. Strains isolated from patients with pneumonia presented a lower capacity of adhesion in comparison to those isolated during tracheobronchitis (p = 0.002). Hla was detected in 71 strains (75.5%), with most of the producer strains in pneumonia and bronchial colonization group (p = 0.06). In our cohort, Hla expression (presence or absence) in sequential isolates was usually preserved (70%) although in seven cases the expression varied over time. No relationship was found between low cytotoxicity and intracellular persistence in invasion experiments. In our study population, persistent S. aureus isolation from airways in ventilated patients does not reflect pathological adaptation. There is an important diversity of sequence types. Cytotoxicity is variable among strains, but no association with study groups was found, whereas isolates from patients with pneumonia had lower adhesion capability. Favorable clinical outcome correlated with increased bacterial adhesion in vitro. Most of the strains isolated from the lower airways were Hla producers and no correlation with an adverse outcome was reported. The identification of microbial factors that contribute to virulence is relevant to optimize patient management during lower respiratory tract infections.

Keywords: Staphylococcus aureus; adhesion; mechanical ventilation; persistence; pneumonia; toxicity.

Figure 1

Summary of genomic characterization results for all samples analyzed in this study. From left to right: (1) Sequence similarity clustering tree; (2) bacterial isolate numeric identifiers (culture number in Supplementary Table S2) Red color corresponds to samples containing Enterococcus faecalis DNA; (3) patient, individual patient numeric identifiers (ID patient in Supplementary Table S2), color coded by status (red, no change; green, change) diverging sequence type (ST) assignment between isolates from the lower respiratory tract (tracheal aspirate and sputum) of the same patient; patients 15, 146 and 181, are graphically labeled. (4) Origin, isolation source with the following color code: dark blue, lower respiratory tract (LRT) including tracheal aspirate, pleural effusion, protected brush specimen and sputum; light blue, nasal swab and light green, blood culture (including one central venous catheter); (5) order, corresponding to order of isolation for serial LRT samples and (6) ST (strain type) multilocus sequence typing (MLST) assignment results. Asterisks (*) indicate ST could not be assigned. (7) Spa typing protein results. Code corresponds to Ridom Spa Server nomenclature (https://spaserver2.ridom.de/, accessed on 1 August 2020). Asterisks (*) indicate Spa type could not be assigned; hash (#) corresponds to an unknown spa type, see Supplementary Table S2 for details. (8) Genotyping results for several genes of interest, color coded by presence (cyan), absence (salmon) or ambiguous (green). The first block corresponds to some resistance and virulence genes (mecA, fnbA, fnbB, pvl, chp, scn and sak) and the following block contains hemolysin encoding genes (hla, hlb and hld), including Western blot (WB) results for Hla (Hla WB). (9) MGE, mobile genetic elements. Grey scale quantitative heatmap showing the number of phage insertions (Phage) and genomic islands (GI) detected for each sample. Data from 91 isolates and 34 individuals is shown.

Figure 2

Maximum likelihood trees of the populations found from each individual. Circles represent distinct genotypes, and are shaded according to the time of isolation of the genotype: white is earliest, darkest grey is latest. Branches between nodes are shaded according to the types of variants found: intergenic (grey), synonymous (green), nonsynonymous (orange) and nonsense/premature codon stop (red). SNPs and indels are represented by solid and dashed lines, respectively. Small black circles represent hypothetical intermediate genotypes. Data from 25 individuals is shown.

Figure 2

Maximum likelihood trees of the populations found from each individual. Circles represent distinct genotypes, and are shaded according to the time of isolation of the genotype: white is earliest, darkest grey is latest. Branches between nodes are shaded according to the types of variants found: intergenic (grey), synonymous (green), nonsynonymous (orange) and nonsense/premature codon stop (red). SNPs and indels are represented by solid and dashed lines, respectively. Small black circles represent hypothetical intermediate genotypes. Data from 25 individuals is shown.

Figure 3

Cytotoxicity profile for the 48 clinical strains. (A). Distribution of cytotoxicity results. (B). Cytotoxicity according to the study group considered. (C). Cytotoxicity according to persistent isolation in the respiratory tract. (D). Cytotoxicity according to the clinical outcome, defined as development of respiratory complications and/or mortality related. Mann–Whitney U test and one-way ANOVA test were used. No statistical differences were found.

Figure 4

Adhesion profile for the 48 clinical strains. (A). Distribution of adhesion results. (B). Adhesion according to the study group considered. (C). Adhesion according to persistent isolation in the respiratory tract. (D). Adhesion according to the clinical outcome, defined as development of respiratory complications and/or mortality related. Mann–Whitney U test and one-way ANOVA test were used. Statistical differences were found when comparing values according to the study group and the clinical outcome.

Figure 5

Hla expression. (A). Image with examples of positive (pos) and negative (neg) Hla WB results, including clinical isolates and control strains (USA300, NE1354 Tn::hla). (B). Number of isolates (one per patient) according to the study group and Hla expression (positive/ negative). (C). Number of isolates (one per patient) according to the clinical outcome and Hla expression. (D). Cytotoxicity and (E). Adhesion values according to Hla expression. Mann–Whitney U test and Fisher’s exact test were used. No statistical differences were found.

Figure 6

Invasion assays in murine alveolar macrophage (MH-S) and A549 cell lines. (A). Time course analysis for 4 S. aureus clinical isolates in infected MH-S. Log 10 c.f.u./well are the average of three independent experiments. (B). Adhesion and internalization bacterial counts after infection of the A549 cell line with selected clinical strains. c.f.u./mL are the average of three independent experiments.