Low Level of Colistin Resistance and mcr Genes Presence in Salmonella spp.: Evaluation of Isolates Collected between 2000 and 2020 from Animals and Environment

Salmonellosis is one of the most important zoonoses in Europe and the world. Human infection may evolve in severe clinical diseases, with the need for hospitalization and antimicrobial treatment. Colistin is now considered an important antimicrobial to treat infections from multidrug- resistant Gram-negative bacteria, but the spreading of mobile colistin-resistance (mcr) genes has limited this option. We aimed to evaluate colistin minimum inhibitory concentration and the presence of mcr (mcr-1 to mcr-9) genes in 236 Salmonella isolates previously collected from different animals and the environment between 2000 and 2020. Overall, 17.79% of isolates were resistant to colistin; no differences were observed in relation to years of isolation (2000–2005, 2009–2014, and 2015–2020), Salmonella enterica subspecies (enterica, salamae, diarizonae, and houtenae), origin of samples (domestic animals, wildlife, and environment), or animal category (birds, mammals, and reptiles); only recently isolated strains from houseflies showed the most resistance. Few isolates (5.93%) scored positive for mcr genes, in particular for mcr-1, mcr-2, mcr-4, mcr-6, and mcr-8; furthermore, only 2.54% of isolates were mcr-positive and colistin-resistant. Detected resistance to colistin was equally distributed among all examined Salmonella isolates and not always related to the presence of mcr genes.


Introduction
Colistin is an old antibiotic, discovered in 1947, and belonging to the class of polymyxins. The target of this antimicrobial is the outer membrane lipopolysaccharides (LPS) of Gram-negative bacteria [1]. In human medicine, colistin use has been largely abandoned since the mid-1970s as a consequence of some adverse events, primarily nephrotoxicity and neurotoxicity, combined with the discovery of new antibiotics [2]. In contrast, colistin was largely used in veterinary medicine worldwide to treat infections, mainly from Enterobacteriaceae, in terrestrial and aquatic animals. Furthermore, it was used for decades as a growth promoter in farm animals in many countries; this practice is still allowed in some territories [3]. The increase in antimicrobial resistance led, by the mid-1990s, to the rediscovery of colistin for human therapy as a last-resort antibiotic in multidrug-resistant Gram-negative infections [2,4]. Consequently, policies for the reduction in use of colistin in animals were advanced in many countries, although the resistance to this antimicrobial was not so diffuse [1,5]. Resistance to colistin remained a rare event for years, probably because it is encoded chromosomally and consequently transferred only vertically [2,5].
In 2015, the first plasmid gene able to confer resistance to colistin, mcr-1, was detected in Escherichia coli in China [6]. Subsequently, different variants of this gene, from mcr-2 to mcr-10, and numerous subvariants have been reported [7,8]. Surprisingly, retrospective investigations highlighted the presence of mcr genes in strains isolated from the 1980s, but particularly in bacteria cultured in the last twenty years [9][10][11]. However, before 2015, colistin resistance was not considered a serious threat and resistance levels were generally low [12]. Starting in 2015, reports of colistin-resistant and mcr-positive bacteria increased; now, these events are reported worldwide in many Gram-negative bacteria [12,13].
Despite E. coli remaining the main bacterial species involved in this phenomenon, colistin resistance and mcr-mediated resistance were also reported in Salmonella [14]. Phenotypic resistance to colistin in Salmonella is variable in relation to serotypes, year of isolation, source (humans, animals, food, and environment), and geographic area. However, percentages of resistant strains are generally low, ranging between about 1% and 12% [15][16][17][18], with some exceptions (40%) [19]. The gene mcr-1 was the most frequently detected in Salmonella, but mcr-2, mcr-3, mcr-4, mcr-5, and mcr-9 were also detected [20]. Domestic animals were identified as the main reservoir of mcr-positive salmonellae and S. ser Typhimurium as the most frequent serovar carrying these genes [14,20]. However, detection rate of mcr-positive Salmonella strains is generally low [17,18,21], with some exceptions [22]. As reported for other bacteria, the direct relation between the presence of mcr genes and colistin resistance in Salmonella has not been fully clarified; these genes have also been found in susceptible strains [15], and, frequently, only a small number of resistant strains carry these genes [16][17][18], suggesting other mechanisms of resistance [2]. In S. enterica, different chromosomally encoded mechanisms were identified as conferring resistance to colistin. In particular, the addition of 4-amino-4-deoxy-L-arabinose (L-Ara4N) and phosphoethanolamine (PEtN) to the phosphate groups of lipid A causes the alteration of the LPS structure; this event alters the primary target of colistin. The arnBCADTEF and pmrCAB operons are responsible for this process; they are up-regulated by the two-component regulatory systems (TCSs) PmrA/PmrB and PhoP/PhoQ. Overactivation of TCSs by external stimuli or by mutations may lead to the hyper-expression of arnBCADTEF and pmrCAB. Mutations in mgrB, a suppressor gene of the PhoP/PhoQ system, may produce the same result. Deacylation of lipid A, mediated by pagL, rpoN, and ipxM genes, may be another mechanism mediating colistin resistance, although less common in S. enterica than in other bacteria, such as Acinetobacter baumannii. Finally, it has been proposed that alteration of efflux pumps (sapABCDF operon) and porins by cytoplasmic molecules (ydeI/ompD system) can increase bacterial resistance to polymyxins [2,13,23,24].
The aim of this study was to investigate the colistin resistance and presence of mcr genes (mcr-1 to mcr-9) in Salmonella strains isolated from animals or related environments, including strains cultured from atypical samples (reptiles and wild animals) and belonging to serotypes that are not frequently detected or investigated, collected between 2000 and 2020.

Discussion
Salmonella is one of the most important food-borne zoonotic pathogens worldwide. In Europe, salmonellosis is the second most prevalent zoonosis considering the total number of confirmed cases and the most prevalent considering the total number of outbreaks. In 2019, in EU, Salmonella was responsible for 87,923 confirmed human cases, 16,628 hospitalizations, and 140 deaths [25]. In Europe, antimicrobial resistance, mainly multidrug resistance, was generally lower in Salmonella than in other Enterobacteriaceae, with differences related to serotype and geographic territory [26]. Colistin was recently reintroduced in human therapy and it is now considered a critically important antimicrobial [27]. For this reason, the discovery of mobile genes conferring colistin resistance caused many concerns. These genes were more commonly observed among E. coli strains, but they were also detected in other Gram-negative bacteria, such as Salmonella [28].
The number of studies on phenotypic colistin resistance in Salmonella is limited. Past works were usually based on the disk diffusion test, a method not suitable for colistin resistance detection [1]; on the other hand, recent investigations usually focused on mcr-positive strains. In our study, the percentage of colistin-resistant strains was 17.79%, which is slightly higher than the data reported in recent literature. Recent surveys registered percentages of colistin-resistant Salmonella strains ranging between 1% and 12% [15,17,18,29,30]. These differences may be due to many factors: source, geographic area, and serotype. It is interesting to note that no differences were observed in relation to years of isolation, suggesting a basal level of diffused resistance. Unfortunately, our study suffers from the heterogeneity of samples from which Salmonella isolates were cultured; thus, it was not possible to draw a conclusion in terms of monitoring over time. However, this limitation may also represent a strength of the study, showing a condition largely spread among the Salmonella genus and not restricted to strains coming from a particular matrix. To underline this hypothesis, no differences were observed among strains from wild animals, domestic animals, or environment, or from different animal categories: birds, mammals, and reptiles. Salmonella isolates from houseflies (Musca domestica) were statistically the most resistant strains. Considering the behavior of houseflies, the obtained results may suggest that these insects might be involved in the diffusion of colistin-resistant salmonellae.
Most of the available studies in the literature focused on a subset of serovars, the ones mainly involved in human and animal salmonellosis, such as Typhimurium and Enteritidis [14,16,20]. Few data are available on colistin resistance in Salmonella belonging to uncommon serovars or subspecies [31]. In our investigation, more than 70 different serovars, belonging to four subspecies, were included. In relation to phenotypic colistin resistance, no statistical differences were observed among the subspecies. Furthermore, no statistical differences emerged among the six serovars most frequently detected in human infections in EU between 2015 and 2019 [25]. These findings support the hypothesis of a diffused resistance among tested salmonellae. Moreover, our results show that the less investigated serovars may also be involved in the phenomenon of colistin resistance.
Regarding mobile colistin-resistance genes, few isolates scored positive (5.93%) for one or more genes. This finding is in agreement with other studies reporting a percentage of mcr-positive Salmonella strains ranging between 0.6% and 3.3% [15,[17][18][19]29]. The slightly higher positivity detected in our survey may be related to the extended set of investigated genes. Our data confirm the low circulation of these genes in the Salmonella genus compared to other Enterobacteriaceae, especially E. coli. E. coli seems to be the primary species having mcr genes [9,28,32], also confirmed by the high prevalence of mcr-positive E. coli strains (44.6%) isolated from wild boar from the same geographic area [33].
The low detection rate of mcr genes compared to found phenotypic resistance suggests the low involvement of these genes in colistin resistance in Salmonella, as also supposed by other authors [17,29]. Detected resistance may be related to other chromosome-related mechanisms. From this point of view, the TCSs (pmrA/pmrB and phoP/phoQ) play a key role in colistin resistance in Salmonella. Mutations occurring in these genes or in their regulatory genes, such as mgrB for PhoP/Q, lead to the modification of lipid A and, consequently, the alteration of the colistin target [2,13,24]. Up-regulation of TCSs may also be the consequence of environmental stimuli, including low Mg 2+ concentrations [2]. For this reason, cation-adjusted Mueller-Hinton medium was employed in MIC tests, as recommended by CLSI [34]. Other resistance mechanisms reported in Salmonella are related to efflux pumps and porins, but they seem to play a secondary role in colistin resistance in these bacteria [1].
Only 6/14 mcr-positive isolates were resistant to colistin. All mcr-positive and colistinsusceptible strains were from reptiles sampled in 2002. The gene mcr-1 was detected for the first time in E. coli in 2015 [6]. However, studies performed since 2015 have also shown the presence of this gene in previously obtained isolates [9], including Salmonella strains [14,20]. Currently, the exact relation between the presence of mcr genes and phenotypic resistance to colistin is controversial and far from being fully understood. In most studies, only colistinresistant strains were screened for mcr and this aspect may cause a lack of information. Recent investigations also showed the abundant presence of these genes in susceptible bacteria [35]; furthermore, some mcr variants are unable to confer colistin resistance [31]. The obtained results highlighted the importance of carrying out both phenotypic and genotypic analyses to investigate the antimicrobial resistance.
The present research, which focused on the resistance to colistin in Salmonella strains cultured from different sources, suffers from some limitations. It would be useful to investigate if colistin-resistant and mcr-positive strains are also resistant to other antibiotics and whether they harbor other resistant genes, particularly at the plasmid level. Furthermore, we investigated only the involvement of mcr genes in colistin resistance. As reported above, other chromosomally related mechanisms induce resistance to colistin; the study of these systems may explain some resistances that were phenotypically detected.

Salmonella Strains Included in the Study
This investigation was conducted on 236 Salmonella spp. strains preserved in the biobank of the Laboratory of Infectious Disease of the Department of Veterinary Sciences, University of Pisa, Italy. All strains were collected during routine investigations or specific research. Most of the isolates came from apparently healthy animals. The strains were selected based on their pulsotype profile and/or taking into consideration epidemiological information. All strains were isolated from feces or organs of various domestic or wild animals (mammals, birds, reptiles, and arthropods), with the exception of 11 strains coming from environmental or feed samples collected at animal farms. Supplementary Table S1 shows information for each strain related to the type of sample, isolation year, species, subspecies, and serotype.

Evaluation of Minimum Inhibitory Concentration for Colistin
The broth microdilution method was employed for evaluation of minimum inhibitory concentrations (MICs) of colistin. The protocol described by CLSI was adopted [34]; cationadjusted Mueller-Hinton (MH) broth (Oxoid, Milan, Italy) was used and colistin sulfate (CARLO ERBA Reagents, Cornaredo, Italy) was diluted 2-fold from 256 to 0.5 µg/mL. Isolates showing an MIC > 2 µg/mL were recorded as resistant [36].

Molecular Detection of mcr Genes
DNA was extracted from overnight agar cultures with a commercial kit, DNA Plus Kits (Zymo Research, Irvine, CA, USA), following the manufacturer's guidelines.
Two multiplex PCR assays were employed: the first for the detection of genes mcr-1, mcr-2, mcr-3, mcr-4, and mcr-5; and the second to detect mcr-6, mcr-7, mcr-8, and mcr-9 genes. The assays were executed using primers and protocols published elsewhere [37,38] and are shown in Table 7. PCR reactions were carried out in a volume of 50 µL including 25 µL EconoTaq PLUS GREEN 2X Master Mixes (Lucigen Corporation, Middleton, WI, USA), 0.5 µL of each primer (10 µM concentration), 3 µL of template DNA, and PCR-grade water to reach the final volume.

Statistical Analyses
Chi-square (X 2 ) and Fisher (F) tests were used to evaluate the obtained data. In particular, statistical tests were used to compare distribution of resistant strains and positivity to mcr genes in relation to: year of isolation, source of isolates, Salmonella subspecies, and serovars. The statistical significance threshold was set at a p-value ≤ 0.05.

Conclusions
In the present study, we focused on the resistance to colistin in Salmonella strains isolated and collected over the last twenty years. No relevant differences emerged among the investigated isolates, suggesting a broad resistance in strains from different sources belonging to different serovars and isolated in different years. Phenotypic resistance was not always associated with mcr genes, confirming that this rising issue marginally involves Salmonella strains. The results of the resistance of the studied isolates to other antimicrobials have been reported in previous surveys. However, the present investigation may be improved by further data about other possible genetic mechanisms responsible for colistin resistance, as suggested by some authors who assumed that chromosomally related mechanisms are involved [2,13,24]. Programs of antibiotic-resistance monitoring should include Salmonella of uncommon serovars and isolated from atypical hosts to obtain more real and valuable information.