Prevalence and molecular characteristics of Staphylococcus aureus, including methicillin resistant strains, isolated from bulk can milk and raw milk products in pastoral communities of South-West Uganda

Staphylococcus aureus strains, especially those resistant to methicillin (methicillin resistant S. aureus, MRSA) have been regarded as zoonotic agents and there is growing genuine concern about the likely transmission of MRSA between animals and humans from close interaction or from handling and/or consuming MRSA infected animal products [1–5]. In settings where humans depend on animals and their products for food and livelihood, such as in pastoral Africa, contact is intimate and multi-drug resistant microorganisms have become an emerging veterinary and zoonotic issue of public health concern [6–8]. Additionally, S. aureus is known to be the third most reported cause of food-borne diseases in the world [4], culpable for production of staphylococcal enterotoxins that cause food poisoning. Moreover, contaminated milk and milk products have frequently been implicated in staphylococcal food poisoning in different industrialized countries [9]. The study of S. aureus prevalence, antimicrobial resistance and clonal lineages in humans, animals and food in African settings has great relevance, taking into consideration the high diversity of ethnicities, cultures and food habits that determine the lifestyle of the people.

The treatment of bacterial infections in Uganda’s health and veterinary care settings is largely empirical with no laboratory confirmation to guide therapy, and antibiotics for both human and veterinary use are readily available over the counter. This continuous use and misuse of the antibiotics has resulted into a surge in multidrug microorganisms, now a growing problem on farms, in health care settings and the community [7, 10, 11]. In Uganda, pastoralist communities are vulnerable to zoonotic infections, including MRSA, not only because of the weak veterinary/public health delivery systems but also due to their cherished way of life (culture) dependent on animals and their products. There is limited data about prevalence and spread of S. aureus strains in rural pastoral farm environments in Uganda. Previous studies have reported on bacterial carriage of milk in peri and urban farmers in and around the capital, Kampala [7], while another reported on burden of S. aureus infection on dairy farms in the neighboring district of Kiboga [6]. However these studies neither assessed resistance to methicillin nor further characterized the isolates for resistance and virulence markers as well as genetic relatedness and spread of the strains in the study environments.

In Uganda, 92% of the milk is marketed unregulated and in raw form [12]. This practice increases the risk of spread of staphylococcal food poisoning due to poor storage and transportation conditions. S. aureus is among the leading causes of foodborne bacterial intoxications worldwide [13, 14], with the most notable virulence factors being the heat-stable enterotoxins that cause the sporadic food-poisoning syndrome or foodborne outbreaks [15]. Staphylococcal enterotoxins (SEs) cause toxic shock-like syndromes and have also been implicated in several allergic and autoimmune diseases, in addition to food poisoning; therefore SEs function not only as potent gastrointestinal toxins but also as superantigens that stimulate non-specific T-cell proliferation [16].

The aim of the present study was to establish the prevalence and distribution of methicillin-susceptible S. aureus (MSSA) and MRSA in bulk can milk and raw milk products samples at household level. Furthermore, we tested the susceptibility of the isolates to nine drugs and analyzed the genetic relatedness of the isolates by PFGE and spa typing. Additionally, resistance and virulence genes (SCCmec and PVL) as well as frequency of genes encoding three of the classical staphylococcal enterotoxins (SEs): SEA, SEB and SEC; and five additional enterotoxins: SEG, SEH, SEI, SEL, and SEM were assayed.

Pastoralists within Uganda’s cattle corridor are afflicted by a variety of ecosystem-based human/livestock health problems, especially those zoonotic in nature, with potential to spill over to neighboring communities and affecting the tourism industry. Unfortunately, not much has been done to estimate the burden of many such potentially zoonotic infections in these marginalized communities. S. aureus, including MRSA, has recently been found to be a zoonotic problem in different settings [2, 3, 28]. We set out to study the prevalence of methicillin resistance, virulence factors, and genetic relatedness of S. aureus strains isolated from bulk can milk and milk products from pastoral households around L. Mburo national Park.

S. aureus was isolated in 20.3% of bulk can milk samples and 12.1% of sour milk samples. Generally, there is a paucity of data on S. aureus from milk and raw milk products in the region. In urban and peri-urban Kampala, a recent study of 97 milk samples yielded 85 pure cultures, but only 1 of 58 (2%) Gram-positive bacteria was identified as S. aureus while 20 of 58 (34%) were coagulase negative Staphylococci [7]. Another study carried out in dairy farms in rural Kiboga (another cattle corridor district) in Uganda, however, found that the most prevalent bacterial pathogens associated with sub-clinical mastitis in dairy cattle were Staphylococci such as Coagulase negative Staphylococci and S. aureus in 64.42 and 16.56% of the cases respectively [6], comparable to our finding of 19.6% S. aureus in fresh bulk can milk samples in the current study. These findings support the expectation that S. aureus from infected udders may contaminate bulk milk and, subsequently, raw milk products as has been observed elsewhere [29].

SCCmec PCR showed that 23/41 (56.1%) of the isolates carried the mecA gene implicated in resistance to methicillin. Elsewhere in Africa, however, low level MRSA carriage in raw milk has been reported in Algeria where only two of five isolates were screened by the disc diffusion test, but there was no testing for the mecA gene [30]. In the current study, only two isolates had zones of clearance less than 22 mm around cefoxitin discs as recommended by EUCAST, to be considered MRSA. The two isolates also had MIC of >2 mg/L on oxacillin strips, hence resistant. The discrepancy between the phenotypic and genotypic results seen in this study may be due to a lack of expression of mecA. Routine oxacillin tests often fail to detect very heterogeneous MRSA populations, which consequently are considered methicillin-susceptible S. aureus (MSSA) because of their usual susceptibility to most non-β-lactam antistaphylococcal antibiotics. Therefore, several parameters have been recommended to improve results and these include increasing the inoculum, growth at a low temperature, an oxacillin screen test with NaCl, or protracted incubation [31]. More recent methods for detection of MRSA include the oxacillin E-test (AB BIODISK, Solna, Sweden) for determination of MICs, and the automated Vitek 2 system (bioMe’rieux, La Balme les Grottes, France). Phenotypic results will therefore vary between laboratories depending on the system used. Detection of the mecA gene, therefore, would be a good way of standardizing identification of MRSA across laboratories. The results are similar to those in another study comparing methods for the detection of MRSA on isolates from foods of animal origin in Italy by Corrente et al. [31]. In that study, none of the six MRSA strains identified by mecA PCR were detected by the cefoxitin test (sensitivity of 0%), and they concluded that analysis for MRSA in isolates from food of animal origin is better done with the mecA gene-specific PCR rather than conventional phenotypic assays [31]. More studies in human derived strains also render support to low sensitivity of the cefoxitin test. For example, a recent study of low level methicillin resistance in mecA positive S. aureus strains showed zones of clearance of 28 mm diameter by cefoxitin testing [32]. In the current study, 17 of the 23 mecA positive strains also had zones of clearance of ≤28 mm (Table 1). The observation of a high proportion of MRSA in the study sample is very worrisome because studies have shown that MRSA readily jumps from animals to exposed veterinarians, farm works and other farm animals [2, 5, 28]. This, to the best of our knowledge, is the first report of MRSA isolated from milk and raw milk products under communally grazing pastoral households in Uganda. In this setting where animal – human interaction is intimate due to the pastoralists’ cherished way of life, it creates an urgent public health concern.

Genotype analysis was done by SCCmec typing and PFGE. Of the 23 mecA positive (MRSA) isolates, 21 harboured the SCCmec type V. The other SCCmec types found can be seen in Table 2. It has recently been noted that community-associated MRSA (CA-MRSA) isolates, which carry SCCmec type IV or V, are now prevalent and exceed methicillin-susceptible S. aureus (MSSA) in skin and soft tissue infections, and such isolates were reported to cause severe, often necrotizing, soft tissue infections and pneumonia [33, 34]. Their presence in milk creates risk of spread of these genotypes in the communities. Additionally, assaying for the virulence marker PVL showed that only four isolates were positive for this gene. A previous study of 50 S. aureus isolates from 14 dairy cow herds in Italy identified the PVL gene in more than 50% of the isolates [35], while another in Brazil reported that none of the 84 milk derived S. aureus isolates harbored the PVL gene [36].

According to criteria proposed by Tenover et al., [22], strains that differ by three bands or fewer by PFGE are likely to be closely related. The authors emphasized that these criteria apply to epidemiologically related isolates. By these criteria, isolates in the current study group into ten clusters of two to seven isolates each, and only eight unique strains. Since these were not outbreak strains, spa typing was used to determine the clonal lieneages of the isolates. The most common lineages were t7753 and t1398 with four isolates each. In all, there were 13 different lineages, but most of the isolates (17/41, 41.5%) could not be assigned to any lineage and were designated as unknown (Table 2). Only one lineage in the current study, t645, had been isolated from one mastitic cow and four milkmen in peri-urban Kampala [7]. The most common spa types associated with clone ST398 (t011, t034, t108, t567, t571, t588, t753, t753, t779, t898, t899, t943, t1184, t1197, t1254, t1255, t1451, t1456, t1457, t2123, t2287, t2329, t2330, t2383, t2582, t2748, t2971, t2974, t3013, t3014, t3053, t3146, or t3208) according to the recent work [37] were not found among our isolates. However, we detected other spa types that have been reported to be associated with LA-MRSA strains other than ST398 (Table 2).

This, to the best of our knowledge, is the first report of SEs in milk from Uganda. While literature elsewhere suggests that the classical SEA is the most frequently observed in enterotoxigenic strains of S. aureus [38], it was found in only two (4.9%) of the strains in the current study. This is probably because of the fact that different foods and strains carry different enterotoxins, as observed elsewhere [39]. On the other hand, the more recently discovered seg gene was observed in 21 (51.2%) of 41 isolates, and in 85.7% of the cases, it was associated with sei, which was present in 25 (61%) of the isolates, similar to reports in other studies in Brazil [40] and France [41]. These genes are known to be frequently found together because they are within the same cluster, in a 3.2 kb DNA fragment as reported elsewhere [42]. In East Africa, most milk is produced by small holders and the bulk of it (86% in Kenya and 92% in Uganda) is traded through unregulated channels as unpasteurized milk or milk products [12]. Raw and sour milk are consumed in these settings for such beliefs as enhanced nutritional quality and better taste. The fact that over 90% of the isolates in the current study carried at least one gene encoding for enterotoxins shows a high risk of spread of foodborne diseases in this setting, with milk as a vehicle for dispersion. There is a need for farmer education and health promotion messages in the communities about the need for pasteurization of milk as this will kill S. aureus, therefore reducing enterotoxin production.

Our results show a high proportion of MRSA in raw milk and its products in the cattle corridor of Uganda, threatening treatment options for S. aureus in the near future in this setting, especially if the strains cross the species barrier into humans. Isolates carrying SCCmec type V are wide spread in this setting. The fact that over 90% of the isolates in the current study carried at least one gene encoding for enterotoxins shows a high risk of spread of foodborne diseases. A limitation of this study is that the enterotoxin genes were not studied for expression, therefore we cannot confirm their significance in cases of food poisoning. Additionally, the initial processing and culture technique for the samples was not good enough whereby only 82 of the 356 samples yielded suspected S. aureus.