Plague is a bacterial zoonotic disease, which infects both human beings and animals especially rodents, caused by a gram-negative bacterium known as Yersinia pestis. Other animals which are occasionally infected and can serve as carriers and consequently act as a source of infection to humans include domestic and wild dogs, cats, other small carnivores and camels [1]. The disease is mainly transmitted from rodents to other mammals by the bite of an infectious flea. The bacterium infects human beings through the bite of an infected flea or by direct contact with contaminated materials from an infected animal [2]. Transmission to animals such as cats and other small carnivores can also occur through bites by fleas escaping from infected rodents caught by such predators or through consumption of such rodent carcases [3].

In Zambia, there have been three outbreaks of human plague in three different zones that are far apart and a number of cases were recorded. The outbreak in North-Western region occurred in December 1993 [4], while in Southern zone, the outbreak occurred in December 1996 [5]. The recorded outbreak in Eastern zone occurred in January 2001, where 436 human cases of bubonic plague with 11 deaths were reported and in 2007, 32 cases were involved [6–8]. All these outbreaks were attributed to the possible movement of rodents from burrows to the nearby villages as a result of heavy rains which flooded these burrows. Plague has been known to occur following a period of total silence as has been observed in India, which experienced a large outbreak in 1994 after 30 years [9], in Algeria, cases were reported in 1940 after a 50 year period of quiescence [10–12] and in Libya, after 25 years, only to have cases in 2009 [13, 14]. The purpose of this study was to assess the role of these hosts, found or living close to humans, as indicators of plague occurrence in the area using molecular and serological tools in a previously plague outbreak area.

The detection of Y. pestis DNA plasminogen activator gene in 5 (2.5%) rodents and shrews from the villages which were involved in the 2007 outbreak of plague can be interpreted to indicate that Y. pestis is still present in the area. These animals may have been infected with the bacterium either through flea bites or from soil contamination as the bacterium can persist in the soil for substantial periods of time [20, 21]. The fact that Mastomys natalensis had the highest infection rate as compared to the house rats and Crocidura spp. is consistent with observations by Kilonzo et al. [22]. All the four villages, where the pla gene of Y. pestis was detected, were involved in the 2007 outbreaks of plague [7]. However, neither of the villages was affected by the 2001 outbreak of the disease. The sequencing results also indicated that Y. pestis circulating in the area was similar to the bacterium which has caused the disease in other parts of the world (Reference sequence accession: CP009936.1; CP010294.1).

The negative PCR results on Xenopsylla spp. fleas collected from rodents and shrews suggests that chances of transmitting the bacterium from rodents and shrews to other mammals including humans through flea bite were minimal or non-existent during the time of this study. This observation is probably attributable to the fact that specimen collections were done at a time of disease quiescence, when the bacterium may be lower in the reservoirs. Clover et al. [23] similarly reported that when there is no plague disease outbreak, most fleas become negative for Y. pestis DNA.

Despite the presence of Y. pestis DNA in rodents and shrews, there was no indication of the outbreak of disease probably due to the partly very low population of vector fleas (Xenopsylla spp.) whose individual rat-flea index ranged between 0.0 and 0.4 among the rodents. Such indices are too low to institute a risk of transmitting the disease outbreak. However, the study indicated that rodents and shrews are suitable hosts/carriers of Y. pestis in the area and may serve as sources of plague outbreaks if and when climatic conditions become favourable to the main hosts and vectors [24].

The ELISA results further revealed that Y. pestis antibodies were present in rodents and shrews and further proved the role of these animals as the natural reservoirs of Y. pestis once survived plague infection in the study area. Detection of such antibodies in 35/199 (17.6%) (95% CI 12.57–23.6%) animals and 25/199 (12.6%) (95% CI 8.30–17.98%) when anti-mouse antibody and anti-rat antibody respectively were used as secondary antibodies, suggests that anti-mouse immunoglobulin is the most cross-reactive to the heterologous antibodies in the rodent population. These results are consistence with those obtained by Nakamura et al. [15]. In a similar study conducted in California, Smith et al. [25] detected Y. pestis antibodies in 18% of rodents by ELISA tests. By using similar tests (ELISA), Kilonzo et al. revealed that 11% rodents and 2% domestic dogs in Northern Tanzania had anti-plague antibodies [26].

The current studies further suggest that M. natalensis and Crocidura spp. were the highest anti-plague positive animals when anti-mouse antibody was used as secondary antibodies, suggesting their strong involvement as plague reservoirs in the area. These observations are consistence with those reported by [22, 27] that M. natalesis in Tanzania were the most plague-positive rodents for both IgG and IgM (10.6 and 9.2% respectively). In Mbulu district, Tanzania, it was observed that 60% of Crocidura spp. were positive for Y. pestis antibodies when tested using ELISA [27].

The detection of specific antibodies against Fra1 antigen of Y. pestis in domestic dogs indicates that the animals had been exposed to the bacteria within their movement range. According to literature, however, anti-plague antibodies in small carnivores remain detectable up to 2 years [28]. It is, therefore, likely that the animals came into contact with the infected materials (e.g. tissues of rodents, shrews and hares) which domestic dogs can easily kill and or feed on. The current results are consistent with those reported by Kilonzo et al. [26]. Similarly, in other studies, dogs were reportedly sero-positive for the anti-plague antibodies in Zimbabwe [29] and Democratic Republic of Congo [30]. Since dogs are very mobile it is most likely that they may have been infected either at the villages of their residence or in neighbouring villages. Dogs are generally considered to be suitable sentinel animals and hence, the presence of plague IgG antibodies in such animals is an indication that the latter had been in contact with Y. pestis either recently or recent past. The relationship of the dog to man and the dog’s rapid serologic response to the plague bacillus suggest that these animals could serve as useful amplifiers and/or sentinel for the detection of plague in areas frequented by man [30, 31].

The observation that 29/37 (78.3%) of rodent fleas collected were hosted by M. natalensis and that all the fleas collected from rodents were Xenopsylla spp. suggests that M. natalensis are the most favourite hosts of efficient plague vectors in the area. Such results are similar to those reported by Kilonzo et al. [26], that M. natalensis had the highest flea infestation rate (18%) followed by Rattus rattus (12%) in a plague endemic area in Northern Tanzania. The SFI in the area was very low (0.19), a fact which probably explains why there have not been any reported human plague cases in the area since the last plague outbreak. Indeed, it has been well documented that a rat-flea index of greater than one increases potential risk for plague infection of humans and an increase in the flea population in a plague endemic area may result in outbreaks of the disease [26, 32]. The rise in the flea population may be caused by climatic conditions such as temperature, rainfall and relative humidity. These have direct effect on development, survival, behaviour and reproduction of fleas and their populations [24]. An SFI of at least 0.5–1 is considered sufficient to maintain plague in a locality [33] and an index greater than one is reported to represent a potentially dangerous situation with respect to plague risk [34]. In the current study, however, the SFI was too low to results in the disease outbreak but gives useful information for the surveillance of plague and serves as an indicator of potential plague transmission [17, 32]. Elsewhere, a rat-flea index ranging from 1.3 to 3.7 was reported in a plague endemic focus when the disease was prevalent [35].

The detection of plasminogen activator gene (Pla) of Y. pestis in rodents and shrews, and specific antibodies against Y. pestis F1 antigen in domestic dogs, rodents and shrews indicates that Y. pestis is still present in the study area and was similar to other Y. pestis bacterium elsewhere. These animals could also act as the sentinel to alert the community to the condition of Y. pestis in the area.