Emergence and Dissemination of a Community-Associated Methicillin-Resistant Panton-Valentine Leucocidin-Positive Staphylococcus aureus Clone Sharing the Sequence Type 5 Lineage with the Most Prevalent Nosocomial Clone in the Same Region of Argentina

Abstract

Epidemiological surveillance for community-associated methicillin-resistant Staphylococcus aureus revealed prevalences of 33% and 13% in pediatric and adult patients, respectively, in Cordoba, Argentina, in 2005. This study describes for the first time the emergence and dissemination of the sequence type 5 (ST5) lineage as the most prevalent clone (89%) (pulsed-field gel electrophoresis type I-ST5-staphylococcal cassette chromosome type IVa-spa type 311) harboring the Panton-Valentine leukocidin and enterotoxin A genes.

Methicillin-resistant Staphylococcus aureus (MRSA) is a traditionally established nosocomial pathogen. Furthermore, virulent and nonmultiresistant community-associated MRSA (CA-MRSA) strains have been reported in many countries since 1993.

CA-MRSA strains are commonly associated with skin and soft tissue infections, although invasive infections such as osteomyelitis and necrotizing pneumonia have been described (9). These strains frequently harbor staphylococcal cassette chromosome (SCCmec) type IV or V and the Panton-Valentine leukocidin (PVL) genes, along with a particular repertoire of virulence genes (4). Moreover, the simultaneous assessment of the genetic background, virulence gene profiles, and accessory gene regulator locus (agr) increases the discriminatory power of genetic investigations into S. aureus pathogenesis (4, 6).

In Argentina, MRSA is one of the most important pathogens causing health care-associated infections, with health care-associated MRSA (HA-MRSA) showing 48% prevalence in 2005 (15), whereas no data about CA-MRSA prevalence are available. The Cordobes/Chilean clone was reported as the most prevalent among HA-MRSA isolates (53% in 2001) in Cordoba, the second most populated city in Argentina (18). Additionally, one sporadic isolate that was susceptible to gentamicin, nonmultiresistant, and characterized as pulsed-field gel electrophoresis (PFGE) type I-sequence type 5 (ST5)-SCCmec type IVa was recovered in 2001 (18). Four other isolates with PFGE type I1 were also detected in another study carried out in 2004 (C. Sola and J. L. Bocco, unpublished data).

In this work, the prevalence of CA-MRSA infections was investigated during a surveillance period in 2005 in Cordoba. The molecular genetic characteristics and virulence gene contents of CA-MRSA strains, as well as the patients' clinical features, were analyzed and compared to those for HA-MRSA strains isolated during the same time period.

This prospective study of laboratory-based surveillance for MRSA infections in 14 hospitals (H1 to H14; 1,878 beds) in Cordoba, Argentina, was conducted in two steps: (i) single patient HA- and CA-MRSA isolates were collected during April to June 2005, and (ii) single patient CA-MRSA isolates were recovered during January to June 2006. Three tertiary-care community hospitals (H11 to H13) and a primary-care pediatric hospital (H14; 49 beds) were included in this study in addition to those previously reported (18). Since no HA-MRSA was recovered in H14 in 2002 and since H14 had a high prevalence of CA-MRSA in 2005, all S. aureus strains, both MRSA and methicillin-susceptible S. aureus (MSSA), were analyzed to determine the genetic relationship between them during the 2006 period.

All isolates were identified by standard microbiologic procedures and were characterized by antibiotic susceptibility (2) and PFGE type (18). The mecA and pvl genes were determined for all strains by PCR (19). Representative isolates of CA-MRSA (all PFGE subtypes) and HA-MRSA (more prevalent subtypes) were characterized by multilocus sequence typing (MLST) and SCCmec and spaA typing, as previously described (18). The sequences obtained by spaA typing were compared to those held on the SpaServer (http://spaserver.ridom.de) (5). Representative isolates of the major PFGE types of MSSA recovered in H14 were also characterized by MLST and spaA typing. The agr group (1 to 4) (14) and virulence gene contents (4, 7, 14) were detected by PCR (Table 1).

TABLE 1.

Characteristics of CA-MRSA and HA-MRSA clones in Cordoba, Argentina

Organisma	PFGE type/n (%)	PFGE subtype/n (%)	ST/CC	RIDOM spa type	spaA repeats	SCCmec type	pvlb	Virulence gene profilec	agr type	Non-β-lactam drug resistance (%)d
CA-MRSA	I/42 (89)	I1/34 (81)	5/5	311	TJMBDMGMK	IVa	+	sea-egc-lukDE	2	ERY (10), CLIi (10)
		I1/1 (2)	5/5	2121	TJMBDMMK	IVa	+	sea-egc-lukDE	2
		I1/1 (2)	5/5	1125	TJMBDMGMMK	IVc	+	sea-egc-lukDE	2
		I2/2 (2)	5/5	1452	TJMBBDMGMK	IVa	−	egc-lukDE	2
		I3/1 (5)	5/5	002	TJMBMDMGMK	IVa	−	egc-lukDE	2
		I4/3 (7)	5/5	311	TJMBDMGMK	IVa	+	sea-egc-lukDE	2
	C/3 (6)	C1/1 (33)	100/5	002	TJMBMDMGMK	IVNv	−	egc-lukDE	2	GEN (100)
		C7/1 (33)	100/5	002	TJMBMDMGMK	IVNv	−	egc-lukDE	2
		C9/1 (33)	100/5	002	TJMBMDMGMK	IVNv	−	egc-lukDE	2
	R/1 (4)	R1/1 (100)	917/8	148	UJGFGMDMGGM	IVNv	−	egc-lukDE	1
	S/1 (4)	S1/1 (100)	918/singleton	1194	ZMOKJBT2T2T2M	IVa	−	sea-sek-egc-lukDE	3
HA-MRSA	A/86 (64)	A1/9 (10)	5/5	149	TO2MEMDMGMGMK	I	−	egc-lukDE	2	GEN (99), CIP (100), ERY (97), CLIc (97), RIF (14), CMP (20)
		A3/13 (15)	5/5	149	TO2MEMDMGMGMK	I	−	egc-lukDE	2
		A23/8 (9)	5/5	149	TO2MEMDMGMGMK	I	−	egc-lukDE	2
	B/24 (18)	B1/6 (25)	239/8	037	WGKAOMQ	IIIA	−	lukDE-bsa	1	GEN (100), CIP (100), ERY (100), CLI (100), RIF (96), CMP (92), SXT (100), MIN (17)
		B2/9 (38)	239/8	037	WGKAOMQ	IIIA	−	lukDE-bsa	1
		B5/5 (21)	239/8	037	WGKAOMQ	IIIA	−	lukDE-bsa	1
	C/20 (15)	C1/6 (43)	100/5	002	TJMBMDMGMK	IVNv	−	egc-lukDE	2	GEN (100), CIP (30), ERY (40), CLIc (15), CLIi (25), RIF (55), CMP (10)
		C15/5 (21)	100/5	002	TJMBMDMGMK	IVNv	−	egc-lukDE	2
CA-MSSA	I/4 (40)	I5/1 (25)	5/5	311	TJMBDMGMK		+	sea-egc-lukDE	2
		I6/1 (25)	5/5	002	TJMBMDMGMK		−	egc-lukDE	2
		I7/1 (25)	5/5	002	TJMBMDMGMK		−	egc-lukDE	2
		I8/1 (25)	5/5	002	TJMBMDMGMK		−	egc-lukDE	2
	M/1 (10)	M1/1 (100)	5/5	002	TJMBMDMGMK		+	egc-lukDE	2	ERY, CLIi
	U/2 (20)	U1/1 (50)	8/8	NDe	ND		+	ND	ND	GEN (50), ERY (50), CLIi (50)
		U2/1 (50)	8/8	ND	ND		−	ND	ND
	N/1 (10)	N1/1 (100)	30/30	012	WGKAKAOMQQ		+	ND	3	ERY, CLIi
	T/1 (10)	T1/1 (100)	ND	ND	ND		−	ND	ND	ERY, CLIi
	V/1 (10)	V1/1 (100)	ND	ND	ND		−	ND	ND	ERY, CLIc

^aCA-MRSA, n = 47 (22 in 2005 and 25 in 2006); HA-MRSA, n = 134 (2005 period); CA-MSSA, n = 10 (2006 period).

^blukS-PV-lukF-PV.

^cThe following genes were analyzed, and those detected are indicated: for enterotoxins, cluster egc (seg, sei, sem, sen, and seo), sea, seb, sec, sed, see, seh, sej, and sek; for toxic shock syndrome toxin 1, tst-1; for exfoliative toxin, eta and etb; for leukocidin, lukED and the class F leukocidin lukM; for adhesin, cna and bpp; and for bacteriocin, bsa.

^dResistance to the non-β-lactams gentamicin (GEN), ciprofloxacin (CIP), erythromycin (ERY), clindamycin (CLIc and CLIi; constitutive and inducible resistance to macrolides-lincosamides-streptogramin B, respectively), rifampin (RIF), chloramphenicol (CHL), trimethoprim-sulfamethoxazole (SXT), and minocycline (MIN) was determined. The percentage of strains resistant to these antibiotics within each PFGE type is indicated when more than one isolate with this PFGE type was detected.

^eND, not determined.

A total of 376 isolates of S. aureus were collected during the 2005 period. The proportions for HA-MRSA and CA-MRSA infections were 57% (134 MRSA/235 HA S. aureus isolates) and 16% (22 MRSA/141 CA S. aureus isolates), respectively. The prevalence of CA-MRSA differed significantly between pediatric and adult patients (33% [7 CA-MRSA/21 CA S. aureus isolates] versus 13% [15 CA-MRSA/120 CA S. aureus isolates]; P = 0.02). Twenty-five episodes of CA-MRSA infections were detected during the 2006 period. In H14, 11 CA-MRSA and 10 CA-MSSA isolates were recovered (52% prevalence). These results revealed the emergence of CA-MRSA in the community; moreover, this is the first report of the prevalence of CA-MRSA in Argentina.

Different definitions for distinguishing CA-MRSA and HA-MRSA as well as differences in the settings and populations analyzed could explain why the prevalence rates of CA-MRSA reported so far vary widely among different studies and countries. The high prevalence (31% in 2005 and 52% in 2006) observed in H14 could reflect the true prevalence of CA-MRSA in this region, because the patients in this hospital were pediatric outpatients affected by nonsevere infections. In fact, patients with mild S. aureus skin diseases are often treated empirically and are more frequently treated in primary or intermediate-care institutions (such as H14) than in tertiary-care community hospitals (such as H11 to H13). In conclusion, the results from H14 during 2005 and 2006 strongly suggest that the current prevalence of CA-MRSA in Argentina may be higher than was detected in all hospitals in 2005 and thus may call for closer systematic surveillance.

The CA-MRSA strains have been shown to be continent specific for some STs, although the results of a recent study suggest intercontinental exchanges of these PVL-positive clones and detection of new ones since 2003. Among the latter, ST5 was described only for sporadic isolates from Europe and the United States (1, 10, 11, 16, 20). Nevertheless, in this study, among all the CA-MRSA isolates, a predominant (89% [91% in 2005 and 88% in 2006]) major PFGE type, PFGE type I, was detected (Table 1), belonging to the ST5 lineage. All isolates of this PFGE type were characterized as agr type 2 and carried the IVa SCCmec type, except for one isolate (SCCmec type IVc). Most of them had spa type 311 (88%) and harbor pvl and sea genes (94%) (Table 1). Therefore, the emergence of the ST5 lineage in 2001 (18) and its dissemination as the most prevalent clone among CA-MRSA is described for the first time in this study. Muller-Premru and coworkers reported isolates of the ST5 lineage harboring pvl genes but associated with SCCmec type I-spa type 002 instead of SCCmec type IVa-spa type 311 (12). The acquisition of pvl genes by this lineage is of extreme epidemiological importance for two reasons: (i) a high epidemic potential of the New York/Japan/ST5-II and Pediatric/ST5-IV pandemic clones was demonstrated in hospital settings, and (ii) the lineage has the ability to acquire vancomycin resistance (18).

In addition, two STs (ST917/CC8 and ST918 [singletons]) were described in this study for the first time, along with the already-characterized ST100 (three isolates), among CA-MRSA strains (Table 1).

The distributions of CA-MRSA and HA-MRSA with respect to patients' clinical features, age, and sex are shown in the Table 2. In agreement with several reports (13), the results of this work showed that most CA-MRSA isolates were from skin and soft tissue infections and occurred with higher frequency among children and young adults (Table 2). Invasive and severe infections (three cases), such as sepsis with pulmonary involvement and bone and joint infections, were also detected. This supports the high virulence of the new CA-MRSA clone detected in Argentina. The presence of PVL appears to be associated with increased severity (9), and the sea genes were linked to severe and invasive S. aureus infections (3). Thus, the presence of both pvl and sea genes, as observed for the clone described here, seems to be a genetic signature of highly toxic/superantigenic strains involved in CA-MRSA infections.

TABLE 2.

Demographic and clinical characteristics of patients with CA-MRSA versus HA-MRSA infections during April through June 2005^a

Characteristic	No. (%) in group		P valueb
	HA-MRSA (n = 134)	CA-MRSA (n = 22)
Gender
Female	48 (36)	11 (50)
Male	86 (64)	11 (50)	0.20
Age group (yr)
Children (0-14)	16 (12)	7 (32)	0.02
Young adults (<30)	8 (6)	13 (59)	<0.001
Children and young adults	24 (18)	20 (91)	<0.001
Adults (>30)	93 (69)	2 (9)	<0.001
Elderly (>69)	17 (13)
Infection type
Skin and soft tissue	37 (28)	20 (90)	<0.001
Abscesses		9
Cellulitis		6
Folliculitis		1
Foruncles		1
Wound		3
Surgical site	37
Bloodstream	37 (28)
Sepsis	32
Catheter associated	5
Bone and joint	15 (11)	1 (5)	0.3
Osteomyelitis	6	1c
Arthritis	2
Prosthetic joint	7
Respiratory tract	23 (17)	1 (5)	0.1
Pneumonia	5
Ventilator-associated pneumonia	18
Acute middle otitis		1
Urinary tract	9 (8)
Other	13 (10)

^aCA-MRSA was defined as MRSA that was isolated from patients who acquired it outside the hospital setting or within 48 h of admission and had no direct or indirect exposure to the health care system in the previous year. HA-MRSA infection was defined according to published criteria, including infection diagnosed in patients with risk factors for hospital acquisition during the previous 6 months (17, 18).

^bP values are based on the chi-square test or Fisher's exact test if 20% of the expected values were smaller than 5 for comparisons of CA-MRSA versus HA-MRSA for each of the categorical variables (males, females, age group, and infection site). A P value of <0.05 was considered statistically significant with chi-square distribution. Epi Info version 6.0.4 software (Centers for Disease Control and Prevention, Atlanta, GA) was used.

^cThe patient affected by CA-MRSA osteomyelitis also had metastatic pulmonary disease and severe sepsis syndrome.

It is of significance that CA-MRSA strains have been reported as the cause of HA-MRSA infections (8). In our study, four patients from different hospitals with surgical site infections were considered to have CA-MRSA infections according to the molecular characterization. Hence, the systematic follow-up of infections caused by CA-MRSA and HA-MRSA in this region is necessary to determine whether this new virulent CA-MRSA clone could also emerge as a cause of HA-MRSA infections. The eventual spread of this CA-MRSA clone in hospitals and in the community could be a serious public health challenge in the coming years.

On the other hand, among the HA-MRSA strains, the Cordobes/Chilean clone is still the most prevalent (64% in 2005) (18) since 1999, and it behaves at present like an endemic clone. Although the total number of HA S. aureus infections did not vary significantly (data not shown), a significant increase in the prevalence of HA-MRSA infections was observed in 2005 (43% in 2001 versus 57% in 2005) (P < 0.0001). Thus, this clone not only gained this dominant behavior at the expense of the other epidemic strains (e.g., the Brazilian clone) but also enhanced the burden of HA-MRSA infections.

The ST30 and ST5 lineages were the most prevalent among invasive MSSA strains with the capacity to acquire pvl genes in nosocomial and community settings in this region during 1999 to 2002 (19). In this study, most strains of MRSA and MSSA belonged to the same PFGE type I and had similar molecular characteristics (Table 1 and Fig. 1A and B). Comparative analysis of virulence gene content and genetic background among those local strains along with the sequential emergence of different genotypes of PFGE type I (Fig. 1C) allowed us to delineate the possible evolutionary pathway within clonal complex 5 (CC5) (Fig. 1D). Therefore, the new epidemic CA-MRSA clone likely arose from insertion of SCCmec type IVa into a successful pvl- and sea-positive MSSA clone (Fig. 1B and D). The type diversity of the SCCmec elements (types IVa, IVc, IVNv, and I) identified in ST5/ST100-CC5 indicates that their integration occurred independently in different time periods. This observation argues in favor of the existence of a successful MSSA clone having a genomic background (PFGE type M-ST5-spa type 002) with the capacity to acquire pvl genes, adapted to the local ecological niche, and more receptive to different SCCmec types. These characteristics could confer a selective advantage for its dissemination in Argentina. According to these results, it is interesting to hypothesize that both the prevalent CA-MRSA and HA-MRSA clones could have evolved from the same successful MSSA ancestor (Fig. 1D).

FIG. 1.

(A) SmaI restriction patterns from representative CA-MRSA and CA-MSSA strains of each subtype and representative HA-MRSA strains of more prevalent subtypes were digitized, presented schematically (middle), and analyzed to calculate the Dice coefficient of correlation and to generate a dendrogram (left) by the unweighted-pair group method using average linkage clustering. Isolates differing by up to six fragments were considered to be subtypes of a given clonal type, and the similarity cutoff was 80%. Genotypes are denoted as subtype (PFGE)-ST (by MLST)-MRSA/MSSA-SCCmec type (right). Gs and Gr are representative strains of the A1 subtype showing susceptibility and resistance to gentamicin, respectively. Nc-8325 refers to the NCTC 8325 control strain. (B) PFGE DNA patterns of representative subtypes of the type I major clone, comparing MRSA (I1, I2, and I4) and MSSA (I5 and I6) strains isolated in Cordoba, Argentina, during 2006. White arrows indicate a single band of difference between I6 (CA-MSSA) and I2 (CA-MRSA) as well as between I5 (CA-MSSA) and I1 (CA-MRSA). The SCCmec IVa element was detected by PCR within the 190-kb band found in the CA-MRSA strains but not in the CA-MSSA-derived 150-kb band (data not shown). Subtypes are indicated at the top by numbers. L, DNA molecular size markers in kb (lambda DNA ladder; Promega). Nc, NCTC 8325 control strain. (C) Annual distribution of infections caused by CA-MRSA strains belonging to the type I clonal family from 2001 to 2006. n, total number of strains characterized as PFGE type I recovered in each year, as indicated. Each genotype was named as subtype (PFGE)-ST (by MLST)-SCCmec type-spa type. The more prevalent genotype among CA-MRSA strains is shaded in gray. (D). Proposed genetic events leading to the emergence of different CA-MRSA and HA-MRSA clones within the CC5 lineage in Cordoba, Argentina. Different HA-MRSA and CA-MRSA clones are indicated by boxes. Shaded and white boxes represent MSSA and MRSA, respectively. Full genotypes are given for the ST, SCCmec type, spaA repeats/RIDOM spa type, PFGE type or subtypes (I1 to I6), and presence of pvl and sea genes. Proposed evolutionary pathways are indicated by filled arrows. Strains diversification through the horizontal transfer of pvl and sea genes and SCCmec allotypes as well as minor changes detected in spa type (differs only by deletion or duplication of one repeat) and in the PFGE type/subtype are indicated next to the arrows. Dashed boxes indicate PVL-positive CA-MRSA and CA-MSSA clones derived from the ancestral MSSA PVL-negative clone (double-lined box). The new CA-MRSA clone that emerged and disseminated in this region is indicated by a double-dashed box. Dashed arrows represent the evolutionary pathway of MRSA clones already proposed (18).

Our results reveal that PVL-producing MRSA strains must now be considered as a possible cause of CA infections in Argentina when choosing therapeutic options. Most CA-MRSA strains were resistant only to β-lactam antibiotics and were susceptible to mupirocin and fusidic acid (only 10% resistance to clindamycin was detected).

In conclusion, the emergence of CA-MRSA in Cordoba, Argentina, was due to the expansion, for the first time, of a virulent clone belonging to the ST5 lineage associated with an SCCmec IVa element and thus differing from the ST30-IVc clone disseminated in the bordering country of Uruguay since 2003.