Introduction

Cystic echinococcosis (CE) caused by the metacestode of the dog worm Echinococcus granulosus is a zoonotic disease of worldwide importance that is widespread on the island of Sardinia, Western Mediterranean (Euzeby 1991; Garippa et al. 2004). Livestock raising occurs extensively in Sardinia, wherein some 3 million sheep are present, representing two thirds of all sheep in Italy. These are exclusively of the Sarda breed, which are good dairy sheep and supply some 100 to 170 l of milk annually during a 180-day lactation period, providing the raw material for the manufacture of cheese, a major industry in Sardinia. Other factors that favor the transmission of CE in Sardinia are the high number of dogs (150,000), prevalence of illegal slaughtering, and other socioeconomic and cultural conditions that contribute to transmission leading to a high rate of infection particularly in sheep (Scala et al. 2004a). Recent data on CE in sheep in Sardinia has revealed a prevalence of 76%, with a fertility rate of 6.9%, while the prevalence of CE in cattle has been recorded as 19.6% (Scala et al. 2004a,b). The contribution of infections in cattle to transmission of CE in Sardinia is likely to be minimal, however, because the cysts in bovines have only 0.76% fertility, which has been suggested to relate to these parasites possibly being of the G1 strain of E. granulosus (Scala et al. 2004b). Less information is available about the epidemiology of CE in pigs in Sardinia, because the infection can be found only in animals raised free in the fields, which are often illegally butchered on the farms. The prevalence of CE in such farm-raised pigs has been determined to be 20% (Garippa et al. 2004).

Given the substantial level of transmission of CE in Sardinia, it is not surprising that CE constitutes one of the major public health problems on the island, where human CE continues to have an average annual incidence of 9.8%°°° (Ecca et al. 1998). Marked differences are apparent in the prevalence of human CE among the various provinces in Sardinia as well as fluctuations in annual incidence (8%°°° to 15%°°° between 1969 and 1990, see Ecca et al. 1998).

Considerable information is available about the epidemiology of different genetic strains of E. granulosus around the world (reviewed by Thompson and McManus 2002). Ten genotypes of E. granulosus, which exhibit a diversity of morphology, development and host range confirmed by genetic studies (Eckert and Thompson 1997; Thompson and McManus 2002; Lavikainen et al. 2003) have been described. Several of these strains have been confirmed to occur in the general area of the Mediterranean. In addition to the G1 or common sheep strain, the equine strain (G4) has been found in Spain, Italy, Lebanon, and Syria. The camel strain (G6) has been found in North Africa and the Middle East and the swine strain (G7) in Spain, the Slovak Republic, and Poland (Thompson and McManus 2002). However, no information is available about the role of different genotypes of E. granulosus in Sardinia. Here we examine CE cysts from sheep, cattle, and pigs in Sardinia to characterize the genetic variants of the parasite and to correlate the results with prevalence and fertility data on the cysts and with the epidemiology of the disease in the different species.

Material and methods

Liver, lung, and spleen samples were obtained from 770 sheep, 229 cattle, and 277 pigs slaughtered in Sardinia between January 2003 and April 2005. By examining the animals’ teeth, we estimated their age, and their origin within Sardinia was obtained from slaughterhouse records. The aforementioned organs of slaughtered livestock were examined for hydatid cysts by visual inspection and palpation. When cysts were found, they were examined in the laboratory to determine number, location, type, and fertility. Fertility was assessed by microscopic observation of the germinal layer. The viability of protoscoleces was determined by examination at 10× without staining, observing flame cell movements and using a vital dye (eosin 0.1%). In cases of nonfertile cysts, the presence of nonviable protoscoleces and degenerative modifications (calcification or caseation) were also determined. From each cyst, laminar layers and protoscoleces were removed and stored at −20°C. DNA was extracted from 91 samples of hydatid material obtained from sheep (63), cattle (14), and pigs (14) using a commercial kit (Roche DNA template extraction kit). The protocol established by Dinkel et al. (2004) was performed on all DNA samples to discriminate the G1 strain of E. granulosus from the G5 and G6/G7 strains with four different polymerase chain reactions (PCR). PCR was carried out for the amplification of target sequence of the mitochondrial 12S ribosomal RNA gene in a 25-μl volume containing 10 mM Tris–HCl (pH 8.3), 50 mM KCl, 2 mM MgCl2, 250 μM of each deoxyribonucleic triphosphate (dNTP), 2 U of Taq polymerase (Roche) and 25 pmol of each primer. PCR assay specific for E. granulosus G1 (g1 PCR) was performed with primer (E.g.ss1for. 5 GTA TTT TGT AAA GTT GTT CTA 3 and E.g.ss1rev. 5 CTA AAT CAC ATC ATC TTA CAA T 3). The samples that were negative in this PCR protocol were then tested with a second PCR assay, which amplifies a 254-bp fragment of E. ortleppi (G5) and E. granulosus G6/G7, with the primer pair E.g.cs1for. (5 ATT TTT AAA ATG TTC GTC CTG 3) and E.g.cs1rev. (5 CTA AAT AAT ATC ATA TTA CAA C 3). Seminested PCRs specific for G6/G7 (g6/g7 PCR; primer pair E.g.camel.for. 5 ATG GTC CAC CTA TTA TTT CA 3 and E.g.cs1rev.) and for E. ortleppi (g5 PCR; primer pair E.g.cattle.for. 5 ATG GTC CAC CTA TTA TTT TG 3 and E.g.cs1rev.) were used to discriminate between E. ortleppi and E. granulosus G6/G7 in a second step, each amplifying a different fragment of 171 bp, as described by the authors (Dinkel et al., 2004). Amplification was carried out for 40 cycles as follows: denaturation for 30 s at 94°C, annealing for 1 min at 57°C and elongation for 40 s at 72°C. After amplification, 10 μl of the amplification products were detected and photographed on a 1.5% ethidium bromide stained agarose gel.

The Eg9 PCR-restriction fragment length polymorphism (RFLP) protocol was performed according to the method described by of González et al. (2002) using the primers PEg9F1, 5’-ATG GCA TGG GTA GCA CGG AGA G-3’ and PEg9R1, 5’-GGT TTG GGA ATG GCG ATG TTG A-3’. Eg9-PCR was performed in a total volume of 25 μl containing PCR buffer (PCR buffer I, Perkin-Elmer), 0.4% glycerol, 200 mM of each dNTP (Amersham-Pharmacia), 0.5 mM PEg9F1, 0.5 mM PEg9R1, and 2.5 U Taq polymerase (Perkin-Elmer). The PCR conditions were: 94°C for 5 min (initial denaturation) followed by 35 cycles at 94°C for 1 min, 60°C for 30 s, 72°C for 30 s, and 72°C for 10 min (final extension). The amplification products (8 μl) obtained from the Eg9-PCR were digested (15 μl final volume) for 4 h with 10 units of restriction site endonuclease RsaI (Roche) according to the manufacturer’s recommendations. The digestion products were fractionated in 2.5% agarose gels, visualized under UV light after ethidium bromide staining and then photographed. Sequencing reactions were undertaken on PCR products as described by Bowles and McManus (1993a,b) for NADH and COI mitochondrial genes. Genomic DNA was isolated as described previously and mitochondrial DNA regions were amplified from 10–20 ng of template by PCR. ND1 (∼500 bp) was amplified with primers JB11 (5-AGATTCGTAAGGGGCCTAATA-3) and JB12 (5-ACCACTAACTAATTCACTTTC-3) and COI with primers JB3 (5-TTT TTT GGG CAT CCT GAG GTT TAT-3) and JB4.5 (5-TAA AGA AAG AAC ATA ATG AAA ATG-3) (Bowles and McManus 1993a,b; Bowles et al. 1992). PCRs (25 μl) were performed in 10mM Tris–HCl (pH8.4); 50 mM KCl; 2 mM MgCl2; 250 μM of each dNTP; 25 pmol of each primer and 2 U Taq polymerase (Roche) under the following cycling conditions: 94°C, 5 min (initial denaturation), followed by 30 cycles of 94°C, 30 s (denaturation); 55°C, 30 s (annealing); 72°C, 30 s (extension), followed 72°C for 5 min (final extension). Thirty-five cycles of denaturation (94°C for 30 s), annealing (55°C for 30 s), and extension (72°C for 30 s) were carried out. For each set of PCR reactions, negative controls (no DNA) were included. DNA from PCR products was purified with a commercial kit (MinElute PCR Purification Kit, Qiagen SA). Sequencing was performed on the ABI Prism 310 (Applied Biosystems) with the ABI Prism Big Dye Terminator Kit (Applied Biosystems) using the corresponding PCR primers. Nucleotide sequence analysis was undertaken using the National Center for Biotechnology Information BLAST programs and databases. Multiple sequence alignments were made with the ClustalW method with Bioedit software and compared with GenBank sequences.

Results and discussion

A total number of 4,070 hydatid cysts were found in the examined sheep, with a prevalence of infected animals of 75.3% (580/770). The percentage of sheep with fertile cysts was 10.3% (80/770). CE was found in 62.2% of the livers and in 53.7% of the examined lungs (χ 2=11.26; P=0.0007). A greater percentage of viable hydatids was found in the lungs (8.7%) than in the liver (4.4%) (χ 2=11.54; P=0.0006). Cysts were found in both the liver and lungs in 41.2% of infected sheep, while the percentage of animals that had a large number of cysts (more than 10) was 16.4% of the total examined (21.8% of the positives). Abundance (number of hydatids/animals sampled) was 5.28, while average intensity (number of hydatids/positive animals) was 7.01. In relation to the morphology of the recovered cysts, 57.4% of them were calcified, while fertile, acephalocysts and caseous hydatids were found in 13.1, 25.1, and 4.4% of the total number of cysts examined respectively.

Hydatid cysts were found in 41.5% of the cattle investigated, although only 2.6% of the animals harbored fertile hydatids of the 1,189 cysts examined. CE was found in 33.2% of cases in the liver and in 31.8% of cases in the lungs (χ 2=0.09; P=0.76). A greater number of viable hydatids were found in the lungs (52) than in the liver (2) (χ 2=65.15; P<0.00001). More than 10 cysts were found in 16.2% of the cattle examined, and 23.6% of the investigated animals had hydatid cysts in the liver and in the lungs at the same time (38.9% of the positives). Abundance was 5.19, while average intensity was 12.5. Caseous cysts amounted to 58% of the recovered cysts, while fertile, acephalocysts and calcified hydatids were found respectively in 4.5, 20.5, and 17% of the total. Hydatid cysts were found in 9.4% of pigs examined (26/277) during home inspection visits. The cysts were found to be viable in 69.2% of the infected animals, representing 6.1% of the total number of livers examined and 6.4% of the lungs (χ 2=0,03; P=0.86). The G1 PCR selectively amplified the G1 genotype of E. granulosus with a specific band of 254 bp, leading us to discriminate 89 on 91 DNA samples while the G5/G6/G7 PCR was positive on two pig samples with a band of 254 bp. The seminested G6/G7 PCR was performed on these two samples to discriminate between E. granulosus G6/G7 and G5 showing a specific product of 171 bp. It was not possible to discriminate between E. granulosus genotypes G6 and G7 by using these methodologies. The same result was obtained using PCR-RFLP on all the samples (91): 89 isolates were identified as the E. granulosus G1 genotype, whereas two isolates were identified as E. granulosus G7 genotype. As in the PCR protocols, PCR-RFLP detected two strains among the E. granulosus pig isolates, which yielded one band (G1) or two bands with the RsaI digestion of the Eg9-PCR amplification product (G7). Sequencing of mitochondrial genes showed that all the sheep and cattle examined were infected by the G1 or common sheep strain of E. granulosus. The DNA extracted from hydatids harbored by pigs corresponded to two different strains: 11 were G1 (sheep strain) while two belong to the pig strain (G7). The mitochondrial ND1 and COI fragment sequences showed a great uniformity, and the only differences found within each genotype correspond to punctual base substitutions in a few isolates (DQ154008; DQ062857). The G7 sequence shows an identity of 99% (Identities=460/462, Gaps=2/462) with Kedra et al. (1999) and Bowles and McManus (1993a,b) isolates (DQ023703, DQ062858). The results of genotyping are summarized in Table 1. Despite there having been three separate hydatid control projects carried out in Sardinia over the last 30 years, CE continues to be highly endemic in Sardinia in several livestock species (Garippa et al. 2004). This study has determined the strains of E. granulosus that infect sheep, cattle, and pigs in Sardinia and represents one of the first investigations of this kind in Italy. The PCR and PCR-RFLP methods used for the first screening of the samples allowed the discrimination of a limited number of genotypes. Hence, sequencing was found to be essential to characterize individual strains or to possibly identify any new genetic variants of the parasite. DNA samples were analyzed from 75 hydatid cysts representing sheep (63), cattle (14) and pigs (14). With the exception of two samples taken from pigs, all samples were determined to be the G1 or common sheep strain. This genotype seems to be the most important strain in the Mediterranean region due to its common occurrence in a wide range of intermediate hosts and also because it is the main parasite genotype known to be responsible for infections in humans (Eckert and Thompson 1997). The low level of viability of the hydatids found in sheep appear to be related to the farm management practiced in Sardinia over the last 10 years and not as a consequence of the parasite genotype found on the island (Scala et al. 2004a; Garippa et al. 2004). Many farmers use anthelmintic treatments against sheep intestinal parasites, and in 47% of these cases the anthelmintic is a benzimidazole compound. Several papers have demonstrated these compounds to have activity against the metacestoda of E. granulosus (Garippa et al. 2004; Scala et al. 2004a). In cattle, the G1 strain of E. granulosus was found to be frequently infertile (2.6% fertile cysts). Similar results have been reported frequently for the G1 strain in cattle, with higher cyst fertility rates in cattle more commonly being associated with infections involving the G5 strain (Euzeby 1991; Thompson and McManus 2002). However, in some situations high levels of cyst fertility have been described for G1 strain parasites in cattle, such as in Tunisia (fertility rate of 47.6%) and Kazakhstan (Torgerson et al. 2003; M’rad et al. 2005). In Sardinia, cattle appear to play a small role in the transmission of E. granulosus infections in dogs and consequently would be unlikely to play a significant role in transmission leading to human infections. However, the G1 strain could constitute a significant economic problem because the liver of infected animals must be destroyed when they are found to be infected with hydatid cysts. In contrast, pigs infected with the G1 strain of E. granulosus were found to have highly fertile cysts (69.2%). These animals potentially play a significant role in the transmission of E. granulosus to dogs in Sardinia, particularly because of the common practice whereby infected pigs are slaughtered informally and without any inspection for hydatid or other infections. The pig strain (G7) was detected in two samples of those examined from Sardinia. This is the first identification in Italy of the presence of this genotype. The G7 parasite has been identified previously in pigs from Spain, Poland, Slovak Republic, and Ukraine (Pawtowski and Stefaniak 2003; Kedra et al. 2000; González et al. 2002; Turceková et al. 2003; Pawtowski and Stefaniak 2003). The G7 strain differs morphologically, developmentally, epidemiologically, and genetically from other strains of E. granulosus, particularly from the G1–G2–G3 group (Eckert and Thompson 1997). The G7 strain is known to be infective to humans (Kedra et al. 1999). On the basis of the relative abundance of G1 strain parasites and the presence of fertile cysts in sheep and pigs, sheep and pigs are the intermediate hosts most likely to play the most important role in the transmission of G1 strain parasites to dogs leading to infections in humans. The relative role of G7 parasites would be expected to be less than that of the G1 strain in Sardinia, however the contribution of these two strains to infections in humans will require specific future investigation of the genotype of parasites isolated from human cases in Sardinia.

Table 1 Samples of Echinococcus granulosus hydatid cysts from Sardinia
Full size table