Emergence of Erythromycin- and Clindamycin-Resistant Streptococcus pyogenes emm 90 Strains in Hawaii

Abstract

We identified 12 erythromycin- and clindamycin-resistant emm 90 group A streptococcus (GAS) isolates during a retrospective invasive disease survey in Hawaii. A comparison with 20 type-matched isolates showed all resistant isolates to be emm 90.4b with the constitutive or inducible macrolide-lincosamide-streptogramin B resistance phenotype (cMLS_B or iMLS_B). All isolates had the same pulsed-field gel electrophoresis (PFGE) pattern, suggesting clonal spread.

The erythromycin resistance rate (∼3%) has been low among group A streptococcus (GAS) isolates in Hawaii, where impetigo, acute rheumatic fever, and clusters of necrotizing fasciitis have been reported (5-7, 8-10; G. Erdem, G. Matsuura, G. Wheelen, C. Mizumoto, D. Esaki, and P. V. Effler, presented at the XVIIth Lancefield International Symposium on Streptococci and Streptococcal Diseases, Porto Heli, Greece, 2008). During a retrospective survey of 250 GAS cultures collected from patients with invasive GAS disease identified between the years 2005 and 2007, the reviews of the clinical information incidentally showed erythromycin and clindamycin resistance in five patients. emm typing of these isolates showed a single emm type: emm 90.4b. In addition to these five isolates, 15 additional isolates from individual patients with invasive disease were emm type 90.4b. The high percentage of invasive isolates (8%) belonging to this infrequent emm type and associated drug resistance prompted us to characterize the associated drug resistance patterns. We wanted to compare the resistance patterns among the isolates identified during this survey and other emm type 90 isolates identified from patients with noninvasive disease during our previous surveys. Of the 20 invasive GAS isolates, 12 were from blood cultures. Eight isolates were identified from wound cultures of inpatients with severe skin and soft tissue infections. Twelve additional emm type 90 pharyngeal isolates from our archived collection of 1,660 GAS isolates (collected from 1997 to 2007) were studied.

(Preliminary data were presented in part at the American Society of Tropical Medicine and Hygiene 58th Annual Meeting in November 2009.)

emm sequencing (http://www.cdc.gov/ncidod/biotech/strep/strepblast.htm) was done for all isolates. Etest methodology (AB Biodisk, Piscataway, NJ) was used to determine MICs for erythromycin and clindamycin (2, 3). The erythromycin-clindamycin double disk test was applied to resistant isolates. The “blunting” in the inhibition zone around the clindamycin disk next to the erythromycin disk was interpreted to be an inducible macrolide-lincosamide-streptogramin B resistance phenotype (iMLS_B), and susceptibility to clindamycin with no blunting indicated the M phenotype. The mef(A), erm(B), erm(A) (20), and erm(TR) resistance genes and prtF1 were detected by PCR amplification (11, 19, 23, 24, 26, 27). Pulsed-field gel electrophoresis (PFGE) was performed with a Chef-DR III apparatus (Bio-Rad) after SmaI digestion.

All (20 isolates) invasive GAS emm 90 strains were of the allelic variation type 90.4b. Of the 12 noninvasive pharyngitis isolates, three emm 90 subtypes were identified: emm type 90.0 (one isolate from the year 1997), emm type 90.2 (6 isolates from 2004, 2005, and 2006), and emm type 90.4b (5 isolates: one isolate from 2005 and 4 isolates from 2007). Five of the invasive isolates had the constitutive MLS_B (cMLS_B) phenotype, with MICs showing resistance to both clindamycin (MIC, 3 to 4 μg/ml) and erythromycin (MIC, 12 to 64 μg/ml). Seven of them showed the iMLS_B phenotype (Table 1). All of the resistant strains were isolated after July 2006. Of the 12 noninvasive isolates, only three of the emm 90.4 isolates showed resistance (erythromycin MIC, 12 to 24 μg/ml) with the iMLS_B phenotype, and they were identified in the year 2007. Other emm 90 subtypes did not show resistance.

TABLE 1.

Isolation years, sources, emm typing, drug MICs, phenotypes, and resistance genes identified in invasive GAS isolates listed according to original culture and identification dates

Isolate	emm type	Yr	MIC (μg/ml) of druga		Drug resistance phenotype	Resistance genes	Source
			EM	CM
1	90.4b	2005	0.19	0.125			Blood
2	90.4b	2006	24	0.125	iMLSB	erm(A), erm(TR)	Blood
3	90.4b	2006	0.25	0.19			Blood
4	90.4b	2006	32	0.19	iMLSB	erm(A), erm(TR)	Blood
5	90.4b	2006	0.38	0.25			Blood
6	90.4b	2006	0.25	0.125			Wound
7	90.4b	2006	48	0.094	iMLSB	erm(A), erm(TR)	Wound
8	90.4b	2006	24	0.19	iMLSB	erm(A), erm(TR)	Wound
9	90.4b	2006	24	0.125	iMLSB	erm(A), erm(TR)	Wound
10	90.4b	2006	24	4	cMLSB	erm(A), erm(TR)	Wound
11	90.4b	2006	12	0.125	iMLSB	erm(a), erm(TR)	Wound
12	90.4b	2006	0.125	0.125			Blood
13	90.4b	2006	24	0.047	iMLSB	erm(A), erm(TR)	Foot
14	90.4b	2006	0.19	0.047			Blood
15	90.4b	2006	0.125	0.125			Blood
16	90.4b	2006	0.25	0.25			Wound
17	90.4b	2006	48	3	cMLSB	erm(A), erm(TR)	Blood
18	90.4b	2007	24	3	cMLSB	erm(A), erm(TR)	Blood
19	90.4b	2007	48	3	cMLSB	erm(A), erm(TR)	Blood
20	90.4b	2007	64	3	cMLSB	erm(A), erm(TR)	Blood

^aEM, erythromycin; CM, clindamycin.

The erythromycin resistance rates increased each subsequent year (0% in 2005, 56% in 2006, and 86% in 2007). All of these MLS_B isolates had the erm(A) and erm(TR) genes and tested negative for the presence of mef(A) and erm(B) genes. All noninvasive isolates tested positive for the presence of the prtF1 gene. Four PFGE patterns were identified among emm 90.4 isolates, and all drug-resistant isolates (cMLS_B and iMLS_B phenotypes) had one identical PFGE pattern. Among the emm type 90.4b isolates, there was an approximately 250-kb-size band difference between the drug-resistant and nonresistant isolates.

Erythromycin or clindamycin is recommended as an alternative treatment of streptococcal pharyngitis in case of allergy to beta-lactams, and clindamycin could be used in combination with penicillin for treatment of severe streptococcal infections (12, 31). Resistance of Sreptococcus pyogenes to macrolides and lincosamides has been reported, and overenthusiastic use of several macrolides may further increase this problem (14, 15, 19, 27, 28). A few emm types, such as emm 4, 6, 12, 58, 75, 77, 89 and 114, have been associated with resistance (1, 4, 14, 17, 19, 24, 26, 28). The emm type 90 isolates have not been associated with invasive GAS disease or with macrolide resistance (1, 4, 22, 25, 27, 30). This emm type appeared to be common after 2006 among invasive GAS strains, although it has been infrequently identified during our previous studies (5-7, 8-10; Erdem et al., XVIIth Lancefield International Symposium on Streptococci and Streptococcal Diseases). Considering that we have previously showed a wide variety of rather infrequent continental U.S. emm types and low levels of resistance, the emergence of these clonal, drug-resistant and invasive strains in Hawaii is concerning (29).

Macrolide resistance has been linked to invasiveness with possible introduction of drug resistance and virulence genes via lateral gene transfer (11). This may explain the presence of drug-resistant and invasive strains in our study. Facinelli et al. described the drug-resistant intracellular streptococci escaping the effects of macrolides (11). They found a high proportion of drug-resistant isolates to be PrtF1 positive, suggesting PrtF1 as a virulence factor (11). In contrast, almost none of the invasive and drug-resistant strains in our study had this gene. This finding may be more in-line with PrtF1 association with persistent infection or carriage (11, 21).

We do not know whether our strains have a higher invasion efficiency, since we have not done similar internalization and cell invasion assays. However, we have shown isolates harboring the rarely reported, constitutively expressed erm(TR) gene sharing the same PFGE pattern with the inducible resistant phenotypes after 2006 (17, 18, 20). Expression of the erm genes can be constitutive and inducible, and the selective pressure of clindamycin may lead to selection of constitutively resistant derivatives with alterations in the attenuator of the erm(TR) gene (12, 15). Point mutations in the erm regulatory region leading to constitutive methylase expression and high-level resistance due to the destabilization of the stem-loop structure have also been reported (17). In pneumococci, constitutive expression of resistance was found to be due to deletions, duplications, or point mutations in the attenuator sequence, leading to the derepressed production of the methylase (15, 16).

The single PFGE band difference between most of the drug-resistant and -sensitive isolates was an interesting finding of our study. Macrolide resistance is more likely to occur from a horizontal gene transfer than a spontaneous chromosomal mutation when observed in a rapid spread of a single emm type (28). In one study, the acquisition of the resistance phenotype was associated with the appearance of a new ∼30- to 40-kb band in the PFGE pattern of a recipient macrolide-susceptible strain, suggesting a transposable element (13). Further studies need to be done on the 250-kb genomic fragment of our drug-resistant isolates to understand whether emm 90.4b developed drug resistance from a genomic insertion. Clonal spread of erythromycin-resistant strains belonging to a single emm type has been reported (14, 19, 24, 26). Our study shows similar concerns of clonal emergence in Hawaii with this multidrug-resistant emm type.