First report of highly pathogenic Echinococcus granulosus genotype G1 in dogs in a European urban environment

Background Echinococcus granulosus and E. multilocularis are tapeworm parasites of major medical and veterinary importance, causing cystic and alveolar echinococcosis, respectively. Both diseases are listed among the most severe parasitic diseases in humans, representing 2 of the 17 neglected diseases prioritised by the World Health Organisation. However, little is known about the role of urban animals in transmission of both parasite species. Findings A sensitive non-invasive genetic method was used to monitor E. granulosus and E. multilocularis infection among dog faecal samples collected from an urban area in Estonia in 2012–13. Out of 181 dog faecal samples analysed, 2.2% tested positive for E. granulosus, determined by sequencing as genotype G1. None of the samples tested positive for E. multilocularis. Conclusions We report contamination of an urban environment with highly pathogenic E. granulosus G1 disseminated by dogs, and a potential risk to human health.


Urban dogs, Zoonoses
Findings Background Tapeworms of the genus Echinococcus are important parasites of mammals, causing life-threatening diseases called echinococcoses. Human echinococcoses are zoonotic diseases that in Europe are caused by two parasite species: E. granulosus, the causative agent of cystic echinococcosis or hydatid disease; and E. multilocularis, which causes alveolar echinococcosis [1]. In Estonia, as well as in the other Baltic States, the number of human cases of echinococcosis is increasing [2]. Human infections are most commonly associated with E. granulosus sensu stricto and, in particular, with its genotype G1 [3,4]. While E. granulosus s.s. in Europe mainly uses domestic dogs as definitive and domestic ungulates as intermediate hosts, the typical transmission cycle of E. multilocularis is wildlife-based, predominantly involving red foxes (Vulpes vulpes) as definitive and small rodents as intermediate hosts. Humans are considered as accidental intermediate hosts to both species, and they can be infected by ingesting parasite eggs via direct contact with a definitive host or through contact with contaminated water, soil or food. Due to their close contact with humans, dogs present a considerable risk factor in the spread of echinococcoses to humans [5].
Following recent increases in abundance, red foxes have started to occur regularly in European cities, including urban areas in Estonia [6]. Mirroring the range shift of its definitive host, E. multilocularis has also colonised the urban environment. The parasite has been recorded in foxes in a number of European cities [7,8] and recently also in Tartu, Estonia [9], prompting considerable concern for public health. To date the only report of E. granulosus infection in urban areas within the European Union occurred several decades ago when the parasite infection was detected in urban dogs in Rome [10]. However, the problem seems more serious than previously thought: in other European countries, including Montenegro and Serbia, the infection rate of urban dogs was found to be very high, up to 65% [11,12], and a cat with E. granulosus G1 cysts was recently found in Saint-Petersburg, Russian Federation [13].
The aim of this study was to monitor urban dog faecal samples for contamination with E. granulosus and E. multilocularis eggs using non-invasive genetics.

Non-invasive sample collection
Dog faecal samples were collected from January to March in 2012 and 2013 from the streets and green areas of Tartu, Estonia. We surveyed 14 transects, each approximately 4 km long (more than 850 km in total), incorporating all major districts in the city ( Figure 1). Each transect was searched weekly during the study period. We collected 102 dog scat samples that were deep frozen at −80°C for at least one week to avoid the risk of Echinococcus infection [5], since both E. multilocularis [9,14] and E. granulosus [15,16] have recently been found in Estonia and have elsewhere been found to infect dogs [17][18][19].

Molecular analysis
Samples were processed and PCR-based molecular analysis was carried out as described in Laurimaa et al. [9]. Essentially, scat samples of approximately 250 mg were placed into tubes, heated for 15 minutes and placed back at −80°C. The heating and cooling procedure helps to break the parasite egg shells, enabling efficient DNA extraction. DNA was extracted using the QIAamp DNA Stool Mini Kit (Qiagen) according to the manufacturer's instructions. Primers F1/RVu and Dog1f/HW1r were used to amplify short sequences of mitochondrial DNA with total lengths of 76 bp and 56 bp to distinguish between fox and dog, respectively [9].
All samples were further analysed for E. granulosus and E. multilocularis using primer pairs specific to E. multilocularis (EMfor1, EMrev1) and E. granulosus (EGfor1, EGrev1), amplifying 120 bp and 149 bp sequences, respectively [9]. PCR was performed twice for each sample in a total volume of 20 μl using the touchdown protocol described in Laurimaa et al. [9]. All positive samples for E. granulosus were sequenced with the same primers as used for PCR. PCR product purification and sequencing procedures followed Saarma et al. [20].

Results
DNA was successfully PCR-amplified from 90 samples out of 102 (success rate 88%). The PCR-based genetic analysis and sequencing verified that all samples belonged to dogs (data not shown). We also included 91 dog samples previously identified and analysed for E. multilocularis [9], but not for E. granulosus. Thus, in total a further 181 dog samples were analysed to determine infection with E. granulosus, and 90 samples were analysed for E. multilocularis. While none of the analysed dog samples were positive for E. multilocularis, we detected E. granulosus in four dog faecal samples (2.2%; 95% binomial confidence interval 0.6-5.6%; Figure 1; Figure 2).
Sequence analysis demonstrated that E. granulosus found in urban dogs in Tartu belonged to genotype G1 (the 'sheep strain') and was 100% homologous to E. granulosus genotype 1 sequence AF297617 (pos. 9984-10133; small-subunit rDNA), whereas in comparison with the other Echinococcus species, the identity was ≤95%; Figure 3).

Discussion
This study provides the first record of E. granulosus genotype G1 in dogs in a European urban environment. Moreover, while E. multilocularis has recently been reported in urban areas of the European Union, including Estonia [7][8][9], this is the first such record of E. granulosus for approximately 20 years. In rural areas of Estonia, both E. multilocularis and E. granulosus sensu lato are present: the former has been described in red foxes [14]; the latter in wolf (Canis lupus) and moose (Alces alces) [15,16]. Moks et al. [15,16] demonstrated that the sylvatic cycle of E. granulosus in Estonia involves moose (genotypes G8 and G10) as the primary intermediate host, and wolf (G10) as the definitive host. Roe deer (Capreolus capreolus) and wild boar (Sus scrofa) have also been examined for the presence of E. granulosus, and published results have reported no cysts [16]. Furthermore, E. granulosus has recently been  Alignment of E. granulosus mitochondrial small-subunit rDNA sequences. 'Dog_Tartu_genotype 1' is the E. granulosus genotype G1 sequence obtained in this study, all others are homologous sequences from various E. granulosus genotypes and Echinococcus species from GenBank. The small-subunit rDNA fragment corresponds to positions 9984-10133 in AF297617. Note that the E. granulosus sequence obtained in this study shows 100% identity with the E. granulosus genotype G1 (AF297617). For other Echinococcus species the identity is lower (≤95%).
reported in cows and pigs, but not in sheep or other farm animals, as reviewed in Marcinkute et al. [2].
Dogs usually become infected with E. granulosus genotype G1 after eating the discarded offal of wild or domestic ungulates [17,19]. Since feeding the viscera of wild and domesticated animals to dogs is commonly practiced, dogs may carry the parasite into urban environments following hunting trips in rural areas. Free-roaming stray dogs that move between rural and urban areas can also be a source of contamination. Since the genotype (G1) we detected differs from those recorded from wild mammals in Estonia (G8 and G10), an origin in the sylvatic cycle seems unlikely. The genotype G1 is most probably transmitted via the domestic cycle, and further sampling of production animals is necessary to determine the transmission path of E. granulosus G1 in more detail.
We expected that dogs included in this study might also harbour E. multilocularis infection, since about 30% of foxes in rural areas [14] and 7.1% of the analysed fox scats from the same urban area (Tartu) were found to be infected [9]. Moreover, 1.6% of raccoon dogs (Nyctereutes procyonoides) harboured the adult tapeworms in their small intestines (Laurimaa et al., unpublished). Given the relatively high densities of foxes and raccoon dogs in Estonia, both species seem to represent important definitive host species for E. multilocularis (and for other zoonotic pathogens in the country) [21]. However, none of the dogs analysed in Tartu were identified as E. multilocularis-positive, though such cases have been reported in Slovakia, where hunting rodents and feeding on raw viscera were found to be the main risk factors of E. multilocularis infection in dogs [17].
Decisions about the treatment and control of echinococcosis should rely on the accurate identification of Echinococcus species and genotypes. Morphological methods often fall short in such identifications due to specific limitations. Recent advances in the development of molecular methods have laid a solid basis for accurate detection of Echinococcus species and genotypes [9,[22][23][24][25][26]. Nonetheless, it is highly desirable to have a sensitive and low-cost molecular diagnostic method that also allows Echinococcus parasites and their host species to be detected from degraded samples (e.g. faecal). The molecular diagnostics we used in Laurimaa et al. [9] and in this study is highly sensitive, being able to detect infection of a scat sample with a single Echinococcus egg. It is especially suitable for analyzing faecal samples since it uses relatively short sequences of mitochondrial DNA, allowing species-specific signals to be detected from samples where DNA may be highly degraded. As the method is based on regular PCR, it is also relatively low-cost.