Linezolid- and Multidrug-Resistant Enterococci in Raw Commercial Dog Food, Europe, 2019-2020.

We describe enterococci in raw-frozen dog food commercialized in Europe as a source of genes encoding resistance to the antibiotic drug linezolid and of strains and plasmids enriched in antibiotic-resistance and virulence genes in hospitalized patients. Whole-genome sequencing was fundamental to linking isolates from dog food to human cases across Europe.

Raw meat–based diets are increasingly popular for feeding dogs, but the extent of antimicrobial-resistant bacteria in raw dog food is rarely addressed globally (). The Centers for Disease Control and Prevention does not recommend feeding raw diets to pets because of frequent contamination with Salmonella and Listeria (https://www.cdc.gov/healthypets/publications/pet-food-safety.html), but awareness about this issue is not as evident in Europe. Eating raw meat has been considered a risk factor for carriage of clinically relevant ampicillin-resistant (AmpR) Enterococcus faecium and optrA-positive linezolid-resistant E. faecalis in dogs (,), but data for commercial pet food are not available. We evaluated multidrug-resistant (MDR) Enterococcus in raw-frozen dog food commercialized in countries in Europe; we focused on transferable linezolid resistance (LinR) genes because linezolid is a last-resort drug to treat gram-positive infections ().

We purchased 14 raw-frozen dog food samples from the 2 commercially available brands in Portugal in specialized stores (September 2019–January 2020). Brand A (produced in Europe) is available in specialized stores, brand B (produced in the United Kingdom) in specialized stores and online; both are commercialized across different countries in Europe. We enriched samples (25 g) in buffered peptone water (1:10), then in brain–heart infusion broth with or without different antibiotic drugs (ampicillin [16 μg/mL], vancomycin [6 μg/mL], chloramphenicol [16 μg/mL]), and plated them onto Slanetz-Bartley agar with and without the same drug concentrations. We identified isolates with different morphologies per plate by PCR. We performed antibiotic susceptibility testing by disk diffusion using European Committee on Antimicrobial Susceptibility Testing (EUCAST) () or Clinical and Laboratory Standards Institute () guidelines. We used broth microdilution for linezolid and Etest for ampicillin. We searched acquired LinR genes (optrA/poxtA/cfrA-E) and typed representative isolates by multilocus sequence typing (n = 20; https://www.pubmlst.org) and whole-genome sequencing (LinR E. faecalis [n = 6] and AmpR/LinR E. faecium [n = 5]) using the Hi Seq 2500 Sequencing System (Illumina, https://www.illumina.com). We deposited assemblies (SPAdes version 3.11.1; https://cab.spbu.ru/software/spades) in GenBank (Bioproject PRJNA663240) and characterized them using in silico tools (http://www.genomicepidemiology.org) and in-house databases ().

All samples carried enterococci resistant to erythromycin, streptomycin, chloramphenicol, and tetracycline; 93% resistant to ampicillin, ciprofloxacin, and quinupristin/dalfopristin; 79% resistant to gentamicin; and 50% resistant to linezolid. We detected acquired LinR genes among 20 MDR isolates from 64% of samples from both brands and with different types of ingredients (Table): optrA (4 E. faecalis, 1 E. faecium), poxtA (2 E. faecium), optrA+poxtA (8 E. faecalis, 3 E. faecium) or optrA+cfrD (2 E. faecalis). Of those, 15 expressed LinR (MIC = 8 mg/L), whereas 5 were susceptible (MIC = 4 mg/L) (Table).

Table

Characterization of Enterococcus isolates obtained from raw dog food samples, Porto, Portugal, 2019–2020*

Species	cgMLST†	MLST‡	Sample (brand)§	Antimicrobial drug resistance profile#	Antibiotic resistance genotype	MIC LIN,mg/L	Transferof LinRgenes
E. faecalis	CT1206	ST40	Duck (B)	ERY, TET, CHL, LIN	optrA, fexA, cat, erm(B), Isa(A), tet(M), dfr(G)	8	–
	CT1207	ST674	Salmon (A)	CIP, ERY, TET, STR, CHL, LIN	optrA, cfrD, fexA, cat, ant(6)-Ia, aph(3′)-III, erm(B), Isa(A), tet(M), tet(L), dfr(G)	8	++
	CT1205	ST1008	Turkey (A)¶	ERY, TET, GEN, STR, CHL	optrA, poxtA, fexB, cat, aac(6')-aph(2”), ant(6)-Ia, ant(9)-Ia, aph(3′)-III, erm(B), lnu(B), Isa(A), Isa(E), tet(M), tet(L), dfr(G)	4	–
	CT1205	ST1008	Turkey (A)¶	ERY, TET, STR, CHL	optrA, poxtA, fexB, cat, aac(6')-aph(2”), ant(6)-Ia, ant(9)-Ia, aph(3′)-III, erm(B), lnu(B), Isa(A), Isa(E), tet(M), tet(L), dfr(G)	4	–
	CT1209	ST1008	Chicken + lamb (A)	ERY, TET, STR, CHL, LIN	optrA, poxtA, fexB, cat, aac(6')-aph(2”), ant(6)-Ia, ant(9)-Ia, aph(3′)-III, erm(B), lnu(B), Isa(A), Isa(E), tet(M), tet(L), dfr(G)	8	–
	CT1208	ST1009	Turkey + goose (B)	ERY, CHL, LIN	optrA, cfrD, fexA, cat, erm(B), Isa(A), dfr(G)	8	–
E. faecium	CT106	ST80	Salmon (A)	AMP (>256 mg/L), CIP, ERY, TET, GEN, STR, QD	aac(6')-aph(2”), ant(6)-Ia, aph(3′)-III, erm(B), msr(C), tet(M), tet(L), dfr(G)	ND	NA
	CT284	ST25	Beef (A)	AMP (32 mg/L), CIP, ERY, TET, GEN, STR, QD, CHL	poxtA, fexB, aac(6')-aph(2”), ant(6)-Ia, ant(9)-Ia, aph(3′)-III, erm(A), erm(B), msr(C), Inu(B), Isa(E), tet(M), tet(L), dfr(G)	4	–
	CT374	ST264	Beef (A)	AMP (32 mg/L), CIP, TET, STR, QD	cat, ant(6)-Ia, Inu(G), tet(M), tet(L), dfr(G)	ND	NA
	CT272	ST1091	Duck (B)	AMP (>256 mg/L), CIP, ERY, TET, STR, QD	ant(9)-Ia, erm(A), erm(B), msr(C), tet(M), tet(L), dfr(G)	ND	NA
	CT3399	ST1263	Deer (B)	AMP, ERY, TET, STR, QD, CHL	poxtA, fexB, cat, ant(6)-Ia, ant(9)-Ia, aph(3′)-III, erm(A), msr(C), Inu(B), Isa(E), tet(L), dfr(G)	4	+

*AMP, ampicillin; cgMLST, core-genome MLST; CIP, ciprofloxacin; CHL, chloramphenicol; CT, complex type; ERY, erythromycin; GEN, high-level resistance to gentamicin; LIN, linezolid; LinR, linezolid-resistant; MLST, multilocus sequence typing; NA, not applicable; ND, not done: QD, quinupristin/dalfopristin; STR, high-level resistance to streptomycin; ST, sequence type: +, positive (transfer frequency of 10−8); ++, positive (transfer frequency of 10−7); –, negative. †The E. faecalis CT1205-CT1209 and the E. faecium CT3399 were identified in this study by submitting them to the cgMLST database (https://www.cgMLST.org) through Ridom SeqSphere+ version 7.2 software (https://www.ridom.de/seqsphere). ‡The novel E. faecalis ST1008–ST1009 were submitted to the MLST database (https://www.pubmlst.org). §Brand A is produced in Europe; Brand B is produced in the United Kingdom. ¶These 2 samples correspond to 2 different batches and were acquired at different times (October 2019 and January 2020). #QD resistance was tested only against E. faecium isolates. Successful transfer of ampicillin resistance is underlined (AMP) and all transconjugants exhibited high values of ampicillin resistance (16–256 mg/L).and.

The E. faecium isolates (n = 39) were mostly MDR (70%), expressing resistance to tetracycline (85%), quinupristin/dalfopristin (72%), erythromycin (64%), ciprofloxacin (59%), streptomycin (57%), ampicillin (56%), gentamicin (23%), chloramphenicol (21%), or linezolid (10%). We compared selected dog food AmpR E. faecium genomes with 7,660 available GenBank E. faecium genomes by complex types (CTs) through core-genome multilocus sequence typing (Ridom SeqSphere+ version 7.2, https://www.ridom.de/seqsphere). Those data (Figure) and data from single-nucleotide polymorphisms (Appendix Figure 1) showed different clusters grouping related isolates obtained from dog food and hospitalized patients (sequence type [ST] 80/CT106; ST264/CT374) or from pet food and livestock or wastewaters (ST1091/CT284; ST1263/CT3399) in different countries. Dog food E. faecium was enriched in acquired antibiotic-resistant and virulence genes as strains from different sources (Appendix Figure 1). ST80 E. faecium from brand A was phylogenetically related to other strains from Germany and Netherlands; ST1091 and ST1263 from brand B were phylogenetically related to UK strains (Figure). By filter-mating (), we found that 3 (ST25, ST80, ST1263) of 5 AmpR E. faecium isolates transferred a chromosomal genetic platform containing pbp5 to GE1 E. faecium strain (Table). Following our previous description of a large transferable pbp5-containing platform in a clinical isolate (), we partly identified highly similar genetic platforms carrying different adaptive features including virulence genes (e.g., sgrA) in ST80 and ST1263 dog food AmpR E. faecium (Appendix Figure 2). ST1263 E. faecium was able to transfer poxtA by conjugation (Table).

Figure

Minimum-spanning tree based on the core-genome multilocus sequence typing (cgMLST) data from Enterococcus faecium isolates (n = 15) from different sources in Europe. The tree is based on cgMLST (1,423 genes) analyses made with Ridom SeqSphere+ version 7.2 software (https://www.ridom.de/seqsphere). Each circle represents 1 allele profile. The numbers on the connecting lines represent the number of cgMLST allelic differences between 2 isolates. Sequence types are shown in colored circles (see key); numbers in circles are isolate identifications. Gray shading around nodes indicates clusters of closely related isolates (<20). CK, chicken; DE, Denmark; DF, dog food; HP, hospitalized patient; PT, Portugal; ST, sequence type; SW, swine; UK, United Kingdom; WW, wastewater.

The E. faecalis isolates (n = 52) recovered were mostly MDR (75%), resistant to chloramphenicol (83%), tetracycline (79%), erythromycin (75%), streptomycin (63%), gentamicin (31%), linezolid (21%), or ciprofloxacin (10%). ST40, ST674, ST1008, and ST1009 sequences corresponded to novel complex types carrying antimicrobial resistance (aac(6')-aph(2″)/ant(6)-Ia/aph3″-III/erm(B)/tet(M),tet(L),dfr(G)) and virulence (ace/gelE/elrA) genes linked to clinically relevant MDR lineages (Table) (,). ST674 E. faecalis carried optrA on a pheromone-responsive plasmid (pAPT110) identical to others from non–clonally related E. faecalis in hospitalized patients in Spain and China (Appendix Figure 3). Similarly to pAPT110 in this study transferring optrA in high rates (Table), pEF10748 (China) is an optrA highly transferable plasmid with a complete sex-pheromone response module ().

In conclusion, the diversity and rate of E. faecium and E. faecalis with linezolid-resistance genes (optrA/poxtA/cfrD) we identified were unexpectedly high. Our data suggest that raw dog food could be a sentinel of emerging antimicrobial resistance traits because this type of food may accumulate raw ingredients of different origins, namely from animals associated with intensive farming, adding a new concern to the global health burden of antimicrobial resistance.

Appendix

Additional information about drug-resistant enterococci in raw commercial dog food, Europe, 2019–2020.