The first case of plasmid-mediated resistance to colistin due to the mcr-1 gene was reported in China in late 2015 (1). Soon after, the same gene was detected by a number of retrospective analyses (2). Resistance to colistin is generally believed to be related to its use in animal husbandry, but epidemiological data on the prevalence and transmission of mcr-1 along the food chain are still lacking, as the majority of studies focus on clinical isolates of Escherichia coli (2). To contribute to closing of the gap, we focused on one of the main foodborne agents, Salmonella, leveraging the cross-species nature of this pathogen, which often moves between animals and humans through foodstuff, thus experiencing the selective factors posed by the different environments, including antimicrobials.

Here we present the outcome of the screening for colistin resistance of 4,473 isolates of Salmonella enterica systematically collected from human, animal, food, and environmental sources in the Emilia-Romagna region of northern Italy between January 2012 and December 2015.

The region studied, covering a population of 4.5 million, performs a systematic laboratory surveillance of clinical isolates of S. enterica from regional hospitals and outpatient samples. The surveillance system also includes isolates from active and passive monitoring of animals and foodstuffs, covering 13, 10, and 18% of Italian swine, dairy, and poultry production, respectively.

Of the 4,473 isolates tested for colistin resistance, 3,294 were from humans (139 serotypes), 1,143 were from veterinary sources (food products, food-producing animals and pet and nonpet animals), and 36 were environmental (seawater). The five most frequent serotypes among human isolates, covering ~74% of the total, were S. enterica serotype 1,4,[5],12:i:− (n = 1,581), S. enterica serotype Typhimurium (n = 456), S. enterica serotype Enteritidis (n = 280), S. enterica serotype Napoli (n = 115), and S. enterica serotype Derby (n = 86). The isolates were screened in a 96-well liquid format with cation-adjusted Mueller-Hinton broth containing colistin at the epidemiological cutoff value (2 mg/liter) indicated by the European Committee on Antimicrobial Susceptibility Testing (EUCAST) (3). The screening Z′ score was 0.79, which is indicative of a robust assay (4). Colistin resistance was detected in 269 isolates (6%), 217 from humans, 51 from veterinary sources, and 1 from the environment, that were further screened for the presence of the mcr-1 and mcr-2 genes by PCR with previously reported primers (1, 5). The colistin MICs for PCR-positive isolates were also determined by standard broth microdilution, and the isolates were also tested for susceptibility to other antimicrobial classes (Fig. 1). mcr-2 was not detected in the 269 colistin-resistant isolates, whereas mcr-1 was found in 25, 10 from humans and 15 from veterinary sources (Table 1). Sequencing of the PCR amplicons evidenced 100% similarity to the mcr-1 gene sequence reported by Liu et al. (1). This confirmed the presence of mcr-1-like genes; at the same time, the existence of mutations outside the region sequenced was not excluded. The prevalence of colistin-resistant isolates differed between sources (Table 1), and the proportion of mcr-1 carriers among colistin-resistant isolates was higher in those from food-producing animals and food sources than in those from humans: 32.4, 30.8, and 4.6%, respectively. A total of 56.3% of the colistin-resistant isolates from swine and 15.4% of those from poultry were mcr-1 positive, while mcr-1 was not detected in pet, nonpet, or environmental isolates. The only food source associated with mcr-1 was pork (Table 1). The larger proportion of samples positive for Salmonella serotype 1,4,[5],12:i:− (17/25) (Fig. 1) reflects its high prevalence among humans (48%) and pigs (33.2%). mcr-1-positive isolates of human origin were concentrated in 2015 (70%, n = 7/10), while veterinary isolates increased from 2013 (20%, n = 3/15) to 2014 (53.3%, n = 8/15) and decreased in 2015 (26.7%, n = 4/15) (Table 1). Clonal diversity among mcr-1-positive isolates, assessed by XbaI pulsed-field gel electrophoresis (PFGE) (www.cdc.gov/pulsenet/), suggests possible horizontal gene transfer (Fig. 1). The greater prevalence of mcr-1 in the swine-pork industry than in humans and poultry (Table 1) is consistent with the high frequency among mcr-1-positive veterinary isolates of S. Derby (Fig. 1), which is highly prevalent in pigs and has been suggested to be adapted to this species (6). Considering that Italy is the European Member State with the second largest use of polymyxins in veterinary medicine (7), these data are suggestive of gene flow from pigs to humans along the food chain. Notably, all mcr-1-positive isolates were susceptible to broad-spectrum cephalosporins; therefore, they were neither extended-spectrum β-lactamase nor AmpC producers.

Open in a separate window

FIG 1

Phylogenetic relatedness and antimicrobial resistance (AMR) profiles of mcr-1-positive Salmonella isolates collected from different sources between 2012 and 2015 in the Emilia-Romagna region of Italy. Clustering of PFGE profiles was generated by the unweighted-pair group method using averages based on the Dice similarity index (optimization, 1%; band matching tolerance, 1%; BioNumerics software, Applied Maths). MICs were determined by the EUCAST method (4). Colistin resistance was defined as a MIC of >2 mg/liter according to EUCAST breakpoints (4). Isolates were tested for antimicrobial susceptibility in a liquid format with cation-adjusted Mueller-Hinton broth with the addition of each single molecule at the epidemiological breakpoint according to EUCAST breakpoints (4). Tested antimicrobials were: AMP, ampicillin; FFN, florfenicol; CTX, cefotaxime; CIP, ciprofloxacin; CHL, chloramphenicol; MER, meropenem; TET, tetracycline; STR, streptomycin.

TABLE 1

Results of screening for colistin resistance and presence of the mcr-1 gene in 4,473 Salmonella isolates collected from different sources between 2012 and 2015 in Emilia-Romagna, Italy

Source	Yr of isolation	No. of isolates		No. of mcr-1-positive isolates (% of those tested)c
		Testeda	Growing on screening platesb
Farmed animals
Poultry	2012–2015	243	13	2 (0.8)
Swine	2012–2015	222	16	9 (4.1)
Bovines	2013–2015	30	3	0 (0)
Mussels	2013–2015	21	1	0 (0)
Horses	2014–2015	19	1	0 (0)
Goats	2014	1	0	0 (0)
Feed	2013–2015	28	0	0 (0)
Pets
Cats	2013–2015	6	1	0 (0)
Dogs	2012, 2014	2	0	0 (0)
Nonpets
Mammals	2012–2015	48	0	0 (0)
Birds	2013–2015	39	3	0 (0)
Reptiles	2013–2015	7	0	0 (0)
Humans	2012–2015	3,294	217	10 (0.3)
Food
Pork	2013–2015	223	6	4 (1.8)
Poultry meat	2013–2015	93	1	0 (0)
Beef	2013–2015	7	0	0 (0)
Eggs	2013–2014	2	0	0 (0)
Other	2012–2015	152	6	0 (0)
Environment (seawater)	2013–2015	36	1	0 (0)
Total		4,473	269	25 (0.6)

Open in a separate window

^aIsolates were collected within the laboratory surveillance system for human enteropathogens of the Emilia-Romagna region of Italy and the routine veterinary microbiological activity on animals and foodstuff of the Istituto Zooprofilattico Sperimentale della Lombardia e dell'Emilia-Romagna.

^bIsolates were phenotypically screened for colistin susceptibility in a 96-well liquid format with cation-adjusted Mueller-Hinton broth containing colistin at the epidemiologic cutoff value (2 mg/liter), according to EUCAST breakpoints (4).

^cIsolates displaying growth in the liquid assay of colistin susceptibility underwent mcr-1 detection by PCR and sequencing (1).

Colistin is one of the last-resort antimicrobials used to treat deadly infections by members of the family Enterobacteriaceae. The discovery of the mcr-1 gene on a mobile element has shed alarming light on the future availability of colistin. Furthermore, a new colistin resistance gene, mcr-2, was recently discovered in Belgium (5) and a variant of mcr-1 (mcr-1.2) was recently discovered in Italy (8). Our high-coverage, territory-focused data demonstrate the presence of mcr-1 in Salmonella isolates at least since 2012 in Italy and indicate that the swine industry is a focal point in the food chain.