Occurrence,Phenotypic and Molecular Characteristics of Vancomycin-Resistant Enterococci Isolated from Retail Raw Milk in Egypt

Introduction

Enterococci are Gram-positive bacteria that exist ubiquitously in the gastrointestinal tracts of animals and humans, as well as in water and soil (Sadowsky and Whitman, 2011). Although generally recognized as predominantly nonpathogenic bacteria, they occasionally are opportunistic and cause human diseases (Ramos et al., 2020). Over the past few years, enterococci have become one of the primary causes of nosocomial infections (Ahmed and Baptiste, 2018).

Within the genus Enterococcus, the two species, Enterococcus faecium, and Enterococcus faecalis are the most frequently reported enterococcal species causing hospital-acquired infections, such as bacteremia and urinary tract infections (Faron et al., 2016). On the other hand, Enterococcus casseliflavus, and Enterococcus gallinarum, which were infrequently recovered from human clinical specimens, have become increasingly implicated in health-care-associated infections (Monticelli et al., 2018; Ramos et al., 2020). They are opportunistic pathogens and are usually inhabitant in animal and human intestinal tracts, causing different types of invasive infections in the hospitalized patients, such as bacteremia, endophthalmitis, endocarditis, and peritonitis (Monticelli et al., 2018).

The glycopeptide antibiotic vancomycin was formerly nominated as a “last resort” for the treatment of severe infections caused by multidrug-resistant Enterococcus spp. (Muhlberg et al., 2020). However, its clinical efficiency is now being questioned with the emergence of resistant strains. To date, nine types of glycopeptide resistance genes have been identified in enterococci: vanC is an intrinsic property of E. casseliflavus, E. gallinarum, and E. flavescens, and eight are acquired (vanA, vanB, vanD, vanE, vanG, vanL, vanM, and vanN) (Ahmed and Baptiste, 2018). The surge in the detection of vancomycin resistance among enterococci in different environments worldwide has been associated with the conjugative transfer of mobile genetic elements and the acquisition of the van gene clusters (Rios et al., 2020).

Of note, the World Health Organization categorized vancomycin-resistant E. faecium as high-priority pathogens for which new and effective therapeutic options are imperatively needed (World Health Organization, 2017). After the discovery of linezolid, it was regarded as a proposed therapeutic option for the treatment of vancomycin-resistant enterococci (VRE) infections. However, it shortly turned out that some strains of enterococci can acquire resistance to both vancomycin and linezolid (Ramos et al., 2020).

The ubiquitous nature and exceptional resilience of enterococci to adverse environmental conditions explain why they are frequently found in foods as contaminants. Enterococci can contaminate milk during the milking process, and this relies on the sanitary condition of the environment, milking personnel, and utensils used for milking (Gelsomino et al., 2002). Of serious public health concern, on a global basis, is the increasing consumption of raw milk.

Unpasteurized milk for human consumption is currently legalized for sale in retail stores and at farms in many countries worldwide. For instance, based on the proposed health benefits of raw milk, about 3% of the Americans consume unpasteurized milk from animal sources (Liu et al., 2020). In Egypt, the milk industry is still primarily traditional as small family farms comprise most of the dairy producers and unpasteurized raw milk is widely consumed. Loose milk accounts for almost 85% of all milk sold in Egypt, which is sold in plastic bags on the street or at dairy stores (Oxford Business Group, 2012). Consequently, milk is prone to contamination with foodborne pathogens, particularly ubiquitous Enterococcus spp., during production and distribution.

Despite the expanding efforts to include a One Health perspective in action plans to determine the prevalence and spread of antimicrobial resistance in human, animal, and environmental sectors around the globe (McEwen and Collignon, 2018), data about antibiotic-resistant bacteria such as VRE from the African continent remain scarce.

Further, the growing demand for consumption of unpasteurized milk in developing and developed countries shed light on the imperative need for elucidating its role as a reservoir for virulent and multidrug-resistant opportunistic pathogens. In this study, we aimed at assessing the occurrence and molecular characteristics of VRE isolated from 100 retail raw milk samples collected from shops in Egypt.

Materials and Methods

Results and Discussion

To date, little information is available regarding the molecular characteristics of VRE isolated from retail raw milk. In Africa, and particularly in Egypt, it is still unclear to what extent the overuse of antibiotics in animals contributed to the evolution of antimicrobial resistance in enterococci as surveillance systems are inadequate. This study examined, for the first time, the occurrence and molecular characteristics of VRE from retail raw milk.

We recovered 31 presumptive enterococcal isolates from 24 samples. All enterococcal isolates were screened by genus-specific primers, of which 24 isolates and one from each sample were identified as enterococci, with an occurrence rate of 24% (24%, 24/100). E. faecium (29.1%, 7/24), E. faecalis (12.5%, 3/24), E. casseliflavus (16.6%, 4/24), and E. gallinarum (4.1%, 1/24) were determined by using multiplex PCR. Other strains were identified as Enterococcus spp. (37.5%, 9/24).

Interestingly, the presence of enterococci in raw milk could not be eliminated by the pasteurization process as enterococci can survive pasteurization temperatures (McAuley et al., 2012). Accordingly, our results underscore the importance of the application of hygienic measures during milk production and shed light on the unsafety of consumption of raw milk from retail shops.

Analysis of the antimicrobial susceptibility of the 24 enterococcal strains revealed that the genus Enterococcus was resistant to clindamycin (100%, 24/24), linezolid (91.6%, 22/24), teicoplanin (91.6%, 22/24), erythromycin (87.5%, 21/24), and tetracycline (29.1, 7/14). None of the enterococcal isolates was resistant to ampicillin, chloramphenicol, ciprofloxacin, streptomycin, gentamycin, or quinupristin-dalfopristin. According to the CLSI guidelines (CLSI, 2017), 22 isolates (91.6%, 22/24) were classified as VRE (MIC ≥32), and 2 isolates (8.3%, 2/24) were classified as intermediate resistant to vancomycin (≤16).

Phenotypic characteristics of isolated strains are presented in Table 1. Visualization of the antibiotic resistance patterns of isolated strains using a heatmap (Fig. 1) revealed four clusters. Cluster 1 included 13 strains (54.2%, 13/24) that were resistant to erythromycin, clindamycin, vancomycin, teicoplanin, and linezolid; Cluster 2 included 5 strains (20.8%, 5/24) that showed additional resistance to tetracycline than group 1; Cluster 3 included 2 strains (8.3%, 2/24) that were only resistant to clindamycin, vancomycin, teicoplanin, and linezolid; and Cluster 4 included 4 strains (16.6%, 4/24) that showed miscellaneous resistance to different antibiotics.

OPEN IN VIEWER

FIG. 1.

Heatmap showing the antibiotic-resistant patterns of isolated strains. Resistance to antibiotics was visualized by using the ComplexHeatmap (Gu et al., 2016). The black and white colors of the boxes indicate resistance and susceptibility to antibiotics, respectively. The dendrogram on the left reflects the hierarchical clustering of antibiotic resistance in the isolated strains. The numbers on the dendrogram (1–4) indicate the numbers of clusters. The dendrogram on the top reflects the hierarchical clustering of screened antibiotics. CLI, clindamycin; ERY, erythromycin; LNZ, linezolid; STR, streptomycin; TEC, teicoplanin; TET, tetracycline; VAN, vancomycin.

Of note, some reports from Egypt highlighted the emergence of multidrug resistance and extensive drug resistance among enterococcal clinical isolates (Hashem et al., 2015; Said and Abdelmegeed, 2019). However, there are little published data from animals and food of animal origin in Egypt, leading to a lack of awareness and indiscriminate use of antibiotics in veterinary practices. The high occurrence of VRE (22%) in the analyzed samples contrasted with other studies that could not detect VRE in raw milk and raw milk cheese in Ireland and Italy (Kagkli et al., 2007; Silvetti et al., 2019).

However, a higher occurrence was reported from raw milk in Botswana (72–84%) (Chingwaru et al., 2003) and cheese in Italy (>50% of samples) (Giraffa et al., 2000). In Egypt, the absence of proper antibiotic stewardship and legislations in place to manage the sale and use of antibiotics for animals and humans hinder the prudent use of antibiotics. Thus, it is not surprising to detect a high occurrence of VRE in retail raw milk samples from Egypt.

The PCR-based detection identified the genetic determinant vanB in 7 strains (29.1%, 7/24), including E. faecium (3 strains), E. faecalis (1 strain), E. casseliflavus (2 strains), and Enterococcus spp. (1 strain) (Table 1). VanB bacteria are resistant to variable levels of vancomycin concentrations and are susceptible to teicoplanin (Faron et al., 2016). They were frequently detected in clinical isolates in a few countries, such as Sweden, Australia, and Germany (Ahmed and Baptiste, 2018).

Our findings, together with recent results from humans (Said and Abdelmegeed, 2019) and animals (Osman et al., 2019) from Egypt, suggest further dissemination of this genotype that may contribute to the dominating vanA genotype in increasing the burden of enterococcal infection. A striking finding in our study is that 6 out of 7 VRE strains (VRM-8, VRM-15, VRM-20, VRM-28, VRM-33, and VRM-34) with vanB genotype showed the VanA phenotype, which is characterized by resistance to both vancomycin and teicoplanin.

Further, resistance to teicoplanin was detected in 17 VRE strains that did not carry any of the glycopeptide resistance genes screened in this study, including E. faecium (4 strains), E. faecalis (2 strains), E. casseliflavus (2 strains), E. gallinarum (1 strain), and Enterococcus spp. (8 strains). Consistent with our results, teicoplanin resistance was detected in the majority of VRE strains isolated from different food products such as cheese, meat preparations, and processed meat (Giraffa et al., 2000; Guerrero-Ramos et al., 2016; Sabenca et al., 2020).

A possible explanation for our result is that the isolated strains carried other vancomycin resistance genes that were not screened in this study, such as vanM gene, which confers resistance to both vancomycin and teicoplanin or vanZ gene, which can be integrated into the van operon and confer resistance to both vancomycin and teicoplanin (Faron et al., 2016). Also, alterations of the vanS_B and vanR_B genes in response to exposure to vancomycin can lead to the emergence of teicoplanin resistance in VanB-type Enterococcus spp.

The vanC genes are intrinsically present in E. gallinarum and E. casseliflavus (Ahmed and Baptiste, 2018). They confer low levels of resistance against vancomycin but not to teicoplanin (Cattoir and Leclercq, 2013). In this study, 5 strains (20.8%, 5/24) showed vanC genotypes (Table 1). E. gallinarum strain VRM-29 showed a typical VanC phenotype, as it was resistant to a low level of vancomycin (≤16) and susceptible to teicoplanin. Three strains of E. casseliflavus showed the vanC genotype/VanA phenotype.

Such co-resistance to both vancomycin and teicoplanin was rarely reported in E. casseliflavus strains (Guerrero-Ramos et al., 2016). Another phenotype–genotype discrepancy was detected in E. casseliflavus strains, VRM-4, and VRM-28, which showed vanC-vanB genotype/VanB phenotype. To the best of our knowledge, such discrepancy had never been described in E. casseliflavus strains.

The high MIC of vancomycin against our E. casseliflavus strains was unexpected in food isolates, as it was rarely reported in E. casseliflavus from clinical settings where vancomycin is regularly used. This may be attributed to the acquisition of other van genes (Monticelli et al., 2018) or the induction of vanC genes in response to exposure to vancomycin (Panesso et al., 2005) during the isolation process.

Of serious concern and a growing health problem, on a global basis, is the emergence of resistance to linezolid in enterococci, particularly in VRE. Here, we report for the first time a high occurrence of co-resistance to vancomycin, teicoplanin, and linezolid in nonclinical enterococcal isolates (83.3%, 20/24) (Fig. 1). Lower rates of linezolid resistance were observed among VRE strains recovered from hospitalized patients in the United States (11.6%) and Germany (1.4%) (Dilworth et al., 2019; Heininger et al., 2020).

Considering that linezolid is not used for the treatment of food animals, the reason for the high occurrence of linezolid resistance in our VRE strains is unclear. It is worth mentioning that when an isolate is resistant to an antimicrobial agent due to a selection process, exposure to an antimicrobial of a different class favors the selection of a new resistant variant, thus accumulating various resistance phenotypes in a process coined as “genetic capitalism” (Baquero, 2004). In Africa and particularly in Egypt, there is a misuse of antibiotics as they can be purchased over the counter. Therefore, the risk of the emergence of co-resistance to several classes of antibiotics in foodborne bacteria is likely to occur.

Of note, testing for linezolid resistance was not considered in several previous studies, particularly those that reported high incidences of VRE similar to our finding (Chingwaru et al., 2003; Guerrero-Ramos et al., 2016; Sabenca et al., 2020), leading to an underestimation of the prevalence of resistance against this antibiotic in food isolates. Thus, it was difficult to compare our results with previous reports from other countries.

However, in agreement with our results, a previous study reported a high incidence of linezolid-resistant coagulase-negative staphylococci (79.5%) from Turkey in Egypt (Moawad et al., 2019). The high occurrence of linezolid resistance reported here prompts increased attention of researchers to routinely perform linezolid susceptibility in food isolates.

Tetracycline resistance in enterococci is generally associated with the existence of tet(M), tet(O), tet(S), tet(K), and tet(L). PCR amplification of tetracycline-resistant genes revealed the presence of tet(M) (8.3%, 2/24), or both tet(M) and tet(L) (8.3%, 2/24) (Table 1). Interestingly, tetracycline-resistant genes can be disseminated to other bacterial species in the human gastrointestinal tract through conjugative transposons such as Tn916-family transposons.

The integrase (IntTn916) and excisionase (XisTn916) encoding genes that are responsible for the transposition of Tn916 were detected in two tetracycline-resistant enterococcal strains (8.3%, 2/24). On the other hand, erm(B) gene, which confers cross-resistance to macrolide, lincosamide, and streptogramin B antibiotics, was detected in one isolate (41.6%, 1/24).

All Enterococcus spp. are intrinsically resistant to a low level of aminoglycosides. Fortunately, in this study, none of the isolated strains showed high-level aminoglycoside resistance according to CLSI criteria (CLSI, 2017). Consistent with our findings, high-level resistance to aminoglycosides was found to be absent or rare in VRE strains isolated from animals and food of animal origin (Sabenca et al., 2020; Wist et al., 2020).

Although enterococci are commensal in the human gut, they can act as opportunistic pathogens, especially in elderly patients and in other immunocompromised patients who have been hospitalized for long periods. The enrichment of virulence genes in E. faecalis and E. faecium such as agg (aggregative pheromone-inducing adherence to extra-matrix protein), esp (enterococcal surface protein), hyl (hyaluronidase), and gelE (gelatinase) was found to contribute to the intestinal colonization, invasion of host tissue, and translocation through epithelial cells.

Also, these genes help E. faecalis and E. faecium evade the host's immune response. In this study, the virulence genes gelE and esp were detected in 16.6% and 12.5% of isolates, respectively. The virulence genes in strains that carried antibiotic resistance genes are shown in Table 1. Of note, asa1, cylA, and hyl genes could not be detected in the isolated strains.

Conclusions

The high occurrence of co-resistance to vancomycin, teicoplanin, and linezolid reported in this study is alarming and underscores the importance of retail raw milk as a potential reservoir of multidrug-resistant enterococcal species that could transfer their antibiotic resistance genes to other more potentially pathogenic bacteria, such as methicillin-resistant Staphylococcus aureus. The reason for the high occurrence of resistance to the most potent antibiotics used for the treatment of enterococcal infection remains unclear, but it could be a consequence of co-selection due to the overuse of other antibiotics.

The similarities between the phenotype–genotype discrepancies described here in VRE isolates from food of animal origin and those reported from humans suggest that both human and animal enterococcal strains have similar evolutionary pathways. This study provides the first detailed analysis about the ecology of antibiotic resistance and virulence in a variety of VRE isolated from retail raw milk, declares potential food safety risks from consumption and improper handling of raw milk regarding clinically important bacteria, and promotes necessary legislation for forbidding the selling and consumption of retail raw milk.

Ethics Statement

This study did not involve any animal experiments.