Molecular Characterization of Methicillin-Sensitive Staphylococcus aureus Isolates from Bacteremic Patients in a Norwegian University Hospital

Abstract

Staphylococcus aureus bacteremia is common in both nosocomial and community settings, and the pathogenicity of the microbe depends upon a large repertoire of virulence factors. S. aureus bacteremia isolates (n = 126) were characterized using DNA microarrays. Clonal complexes 5, 8, 15, 30, and 45 accounted for 74.6% of the isolates. We identified geographical differences in dominating clones and toxin gene profiles. One isolate was methicillin resistant. Potential associations between age and genotype were detected.

TEXT

Staphylococcus aureus bacteremia (SAB) is often associated with increased morbidity and poor outcome (1, 2). The pathogenicity of S. aureus depends upon a large repertoire of virulence factors influenced by environmental conditions and host interactions (3). Associations between specific virulence factors and severe invasive infection still remain unclear. Our recent study genotyping SAB isolates (4) demonstrated a knowledge gap regarding molecular epidemiology of methicillin-sensitive S. aureus (MSSA) isolates despite the increasing incidence of SAB (5). Here we conducted an extensive characterization of MSSA isolates detected in blood cultures using DNA microarrays in order to determine clonal complex (CC) affiliation and carriage of clinically relevant virulence and antibiotic resistance genes.

Isolates (collected in routine diagnostics) from 126 consecutive SAB patients hospitalized in 2006–2007 at Akershus University Hospital, Lørenskog, Norway, were genotyped as described earlier (6, 7). The complete microarray hybridization data are provided in the supplemental material. Clinical parameters are listed in Table 1. The median age of patients was 65 years (range, 0 to 98), and 54.8% were men. Community-acquired (CA) and hospital-acquired (HA) SABs were defined by a positive blood culture obtained earlier or later than 48 h after hospital admission, respectively. Statistical analysis involved Fisher's exact test, with a P value of <0.05 considered significant.

Table 1.

Distribution of S. aureus clonal complexes in patients with bacteremia in relation to clinical parameters

S. aureus CC	No. (%) of patients with CC	No. of patients with condition or characteristic
		HA infection	CA infection	Diabetes mellitus	Intravascular deviceb	Deep-seated abscessc	Endocarditisd	Osteomyelitis/arthritis	All-cause hospital mortality	Under median age 65 yrs	Above median age 65 yrs
CC45	31 (24.6)	13	18	3	3	4	3	4	5	21e	10
CC15	24 (19.0)	14	10	3	7	0	1	1	7	11	13
CC30	19 (15.1)	6	13	2	4	2	0	2	5	7	12
CC5	12 (9.5)	7	5	2	3	1	0	2	2	2	10f
CC8	8 (6.3)	4	4	0	3	0	0	0	1	4	4
Others (15 CCs)a	32 (25.4)	14	18	5	11	4	1	3	8	17	15
Total no. (%)	126 (100.0)	58 (46.0)	68 (54.0)	15 (11.9)	31 (24.6)	11 (8.7)	5 (4.0)	12 (9.5)	28 (22.2)	62 (49.2)	64 (50.8)

^aOthers (n): CC1 (4), CC6 (1), CC7 (3), CC9 (1), CC12 (4), CC20 (3), CC22 (3), CC25 (3), CC50 (1), CC59 (2), CC97 (2), CC101 (2), CC121 (1), CC188 (1), and CC707 (1).

^bIncluding both long-term arterial and venous catheters.

^cWell-defined abscesses in musculature and inner organs.

^dDefined according to the modified Duke criteria (8).

^e67.7%, versus 43.2% for all other CCs; P = 0.023.

^f83.3%, versus 47.4% for all other CCs; P = 0.030.

Twenty different CCs were detected. Five CCs (CC5, -8, -15, -30, and -45) dominated and accounted for 74.6% of the isolates (Table 1). This genetic diversity supports the perception that virtually any S. aureus lineage may cause invasive infection, although some clones potentially are more virulent than others (9, 10). Differences in CC distribution and frequencies were observed in comparing both various geographical origins and sampling sites. CC15 strains were more prevalent (19% versus 2.2% to 13.3%) and CC30 less prevalent (15% versus 19% to 35%) than for SAB isolates from other geographic areas (9, 11, 12) and for nasal carrier isolates and deep surgical site infections (SSI) from the same catchment area (13). The most abundant SAB lineage, CC45, was more common among nasal carriers than in invasive strains in the Netherlands (10, 14) but had a prevalence similar to those of locally found carrier and deep SSI isolates (13). A scarcity of recombination contributes to the highly clonal structure of S. aureus, and the lineage preferences may be due to differences within S. aureus distribution geographically or in relation to host characteristics. Local frequency similarities may reflect that a substantial proportion of health care-related SABs are endogenous infections (15).

Overall, the SAB isolates showed low antibiotic resistance except that nearly 80% were penicillin resistant (blaZ). CC30 strains were significantly more likely to contain the blaZ gene (Table 2). Genes encoding resistance to chloramphenicol (cat), tetracyclines (tetK and tetM), fusidic acid (far1), macrolide-lincosamide-streptogramins (ermA, ermC, msrA, mpbBM, and vga), and aminoglycosides (aacA-aphD, aadD, and aphA) were rare, each occurring in one to four isolates. No clones harbored genes conferring resistance toward vancomycin (vanA, vanB, and vanZ), linezolid (cfr), or mupirocin (mupR). The same susceptibility pattern was found among other MSSA isolates originating with the same geographic area (13), demonstrating that these isolates do not represent a reservoir of various antibiotic resistance genes. Only one isolate carried mecA, encoding methicillin resistance (CC5-MRSA-IV), which is consistent with the low incidence of methicillin-resistant S. aureus (MRSA) found in blood cultures nationwide (16).

Table 2.

Frequencies of selected genes within the five dominant S. aureus clonal complexes in bacteremia patients

S. aureus CC	No. (%) of patients with CC	Frequency of gene(s) (n)
		agrI	agrII	agrIII	agrIV	cap5	cap8	mecA	blaZ	fosB	lukF-PV, lukS-PV	tst	seag	seb	sec, sel	sed	egc	seh	sej, ser	sek, seq	eta, etb, etd	sak	chp	scn	bbp	cna	ebh	sdrD
CC45	31 (24.6)	30	0	0	1	0	31	0	23	0b	0	0d	0	1b	15j	0	31	0	1	0	0	31	31	31	29	31	31	31
CC15	24 (19.0)	0	24	0	0	0	24	0	22	24	0	0e	0	0	0	0	0	0	0	0	0	0	24	24	22	0c	24	23
CC30	19 (15.1)	0	0	19	0	0	19	0	19a	19	1	17f	12h	0	0	0	18	3	0	0	0	17	15	17	18	16c	19	19
CC5	12 (9.5)	0	12	0	0	12	0	1	6	12	0	0	8i	0	0	3	12	0	2	1	0	12	8c	12	10	0	12	11
CC8	8 (6.3)	8	0	0	0	8	0	0	5	8	0	1	2	0	1	3	0	0	3	1	0	8	3	7c	8	0	8	6
Other (15 CCs)	32 (25.4)	20	5	5	2	12	19	0	23	15c	2	2	12	4b	1	0	12	4	0	5	4	28	13c	30	25	16k	27	26
Total	126 (100)	58	41	24	3	33	93	1	98	78	3	20	34	5	17	6	73	7	6	7	4	96	94	121	112	63	121	116

^a100%, versus 73.8% for all other CCs; P = 0.007.

^bThree isolates with ambiguous results not included.

^cOne isolate with ambiguous results not included.

^d0%, versus 21.1% for all other CCs; P = 0.003.

^e0%, versus 19.6% for all other CCs; P = 0.013.

^f89.5%, versus 2.8% for all other CCs; P < 0.0005.

^gsea analyzed together with seaN315.

^h63.2%, versus 20.6% for all other CCs; P < 0.0005.

ⁱ66.7%, versus 22.8% for all other CCs; P = 0.003.

^j48.4%, versus 1.1% for all other CCs; P < 0.0005.

^kTwo isolates with ambiguous results not included.

The toxin gene profiles were highly diverse among the five dominant CCs (Table 2). Panton-Valentine leukocidin (PVL)-encoding genes (lukF-PV and lukS-PV) were detected in three isolates (2.4%), belonging to CC1, CC22, and CC30. The tst gene, encoding toxic shock syndrome toxin 1, found in 20 (15.9%) isolates, was significantly more common in CC30 (Table 2). CC45 and CC15 strains were less likely to carry tst (Table 2). All main CCs, except CC15, carried various S. aureus enterotoxin (SE) and enterotoxin-like (SEl) genes. The highest frequency rates were seen for the enterotoxin gene cluster, egc (seg, sei, sem, sen, seo, and seu), with 57.9%, sea, with 27.0%, and sec and sel, with 13.5% (Table 2). The sea gene was significantly more prevalent in CC5 and -30 (Table 2). CC45 was more likely to contain the sec and sel genes (Table 2). The overall picture is similar with previous reports, but with some geographical differences within the lineages. CC45 isolates are found to contain both tst and sea and to lack egc, while CC15 can harbor tst and egc (11, 17–19). The findings support the hypothesis that the toxins are distributed mainly according to clonal lineages and that transmission is vertical. The absence of toxin genes in C15 may be attributed to the restriction modification system which prevents acquisition of mobile gene elements carrying superantigens by horizontal transfer (17, 20). Geographical toxin profile similarities may indicate that these lineages are evolutionarily old and share conserved genomic structures (21). The assessment of virulence determinants in this study are based only on gene carriage and not quantitative expression of genes or the presence of single nucleotide polymorphisms which may influence the functions of gene products.

Associations between S. aureus genotype and patient age have previously been described only to a small extent. Sangvik et al. found a decrease in prevalence of a specific genotype with increasing age of colonized hosts (22). Interestingly, the age preferences among CC5 and CC45 isolates found in this study (Table 1) may indicate a host and microbe matching, where factors in both parties are relevant for interactions and bacterial pathogenesis. Persistent S. aureus nasal carriers that tested positive for their original resident strain after inoculation with a mixture of different S. aureus strains demonstrated the importance of a good match between host and bacterial factors (23). No significant differences between HA and CA infections were revealed. Analysis of potential associations between bacterial characteristics and clinical manifestations was limited by a lack of comorbidity data and host factors likely to affect outcome.

In summary, our study confirms that a diversity of MSSA clones is responsible for SAB. We identified geographical frequency differences between dominating CCs and between toxin gene profiles within identical bacterial lineages. Interestingly, potential associations between age and genotype were detected. Further investigations to determine the relative role of bacterial, host, and environmental factors in the pathogenesis and outcome of SAB are warranted.