We investigated Trichinella sp. outbreaks on pig farms in Poland in 2013 and 2014, identifying infected pigs and rats. The prior epidemiological investigation of outbreak 1 substantiated the suspicion that contaminated meat may have been the source of infection [40]. Here, genetic data affirm the likelihood that a single infection source gave rise to all infections (documented in half of the herd and in 17 synanthropic rats). In the second outbreak, a source of infection had not previously been established. However, the epidemiological inquiry indicated that illegal movements of pigs between these farms had probably occurred. Molecular tests previously identified T. spiralis as the parasite involved in both outbreaks, but it was not possible to define whether the two outbreaks had the same origin.
Attempts to distinguish among European isolates of T. spiralis have been limited by low genetic variability [24, 41]. Recently, the development of new molecular tools enabled characterization of the genetic structure of individual larvae of the same species, making distinction of genetic differences between two isolates possible [25, 26]. Here, we confirmed for the first time that two T. spiralis isolates, from two separate outbreaks, were notably more similar to each other than would be expected for randomly chosen larval cohorts, supporting the hypothesis that they were somehow related. Despite the limited genetic variability shown by T. spiralis in Europe, the suspicion of a link between the two outbreaks 40 km apart was confirmed by demonstrable genetic affinities (from pigs and a rat).
We reached this conclusion despite low overall genetic variability, consistent with prior regional estimates. Across all investigated loci, we observed globally an average Na of only 1.42 alleles/locus and an average Ho of 0.10. These low values are comparable to prior estimates from 41 wild boar larval cohorts (genotyped using 5 polymorphic loci) from Extremadura (Spain) (Na = 1.59 and Ho = 0.17) [25] and from 13 cohorts (using 7 loci) collected from throughout Europe (Na = 1.27 and Ho 0 = 0.09) [23]. By contrast, Asian T. spiralis cohorts harbor more genetic variability [22, 23]. Recently, Li et al. [42] recognized 7 to 19 alleles/locus studying 16 new microsatellite markers of 12 Chinese isolates. Why Asia and Europe differ so markedly in their polymorphism remains to be determined, but the practical effect has been to severely limit the application of molecular epidemiology to discriminate among T. spiralis isolates in Europe. These tools have adjudicated epidemiological hypotheses for T. britovi, which harbors more variation and therefore more discernable geographic structure [26]. Asia may have been a source of European T. spiralis spread or may simply have maintained larger effective populations for longer [22, 43].
Previous studies have shown that individual hosts generally harbor larval cohorts that can be distinguished, to greater or lesser extents, from larval cohorts of other individual hosts [23]. Here, pairwise Fst among larval cohorts from ten wild boars ranged from 0.045 to 0.934 (average 0.52). The seemingly large differences between cohorts reflects the biology of cohort formation: each host ingests a small number of infectious larvae that then have limited opportunities to outcross because the larvae from a particular meal reach adulthood at the same time. The result is frequent inbreeding, which minimizes the genetic variability of each the larval cohort in any host. Only rarely will a host eat sequential independent T. spiralis-infected meals, producing genetically distinct adults maturing at the same time in the same gut, enabling outcrossing. These rare occurrences can increase genetic diversity locally and limit allelic loss through genetic drift. From a molecular epidemiological perspective, these differences provide opportunities to distinguish infections, even when allelic diversity is limited.
Prior epidemiological investigations of these outbreaks raised suspicion that pigs had been moved between the investigated farms; one farm likely introduced T. spiralis to the other. This study sustains this suggestion by identifying notably little genetic differentiation among the larvae derived from the five hosts on the two farms studied. The larvae sampled on farms account for 7% of the total genetic variation (mean Fst = 0.07, 95% CI 0–0.21). When wild boars were considered, six times as much genetic variation was observed (mean Fst = 0.417), indicating that the larvae of outbreak cohorts share notably recent common ancestors.
The multivariate analysis, based on the Fst matrix, confirms the genetic similarity of the larvae collected from the two outbreaks. The five larval cohorts from the four domestic pigs and from the rat trapped at the pig farm define a cluster distinct from all the larva cohorts from all sampled wild boar except for wild boar WB6; as shown in Fig. 1, nine out of ten larval cohorts from wild boar are scattered in the 3D space, supporting their genetic independence.
The UPGMA phylogenetic tree of MLGs of all individual larvae divided almost all individuals from wild boar from individuals isolated from outbreak hosts, creating a ‘domestic’ subclade within a ‘wild’ clade (Fig. 2). The ‘wild’ clade is characterized by larger genetic distances between MLGs, suggesting the circulation of a wider subset of genotypes among wildlife. The ‘domestic’ subclade includes mostly uniform genotypes from domestic pigs and the synanthropic rat, larvae from the WB6 wild boar, and several individual larvae from other wild boar, suggesting that the genotypes of the domestic infections are likely a subset of wildlife genotypes. In addition, these data suggest that the four animals sampled on farm 1 harbored larvae from a single source. These results support the preliminary suspicion that contaminated feed on farm 1 was likely a shared source of infection for the half of the pig herd and rats.
Analyzing multilocus microsatellite genotypes using the Bayesian algorithm in STRUCTURE confirmed epidemiological assessment despite the limited variability characterizing these isolates. This is similar to what has previously been reported in T. spiralis in Spain and T. britovi on two Mediterranean islands [23, 26]. Isolates derived from populations generally lacking in genetic variability (as is true for T. spiralis in Europe) may become distinct through successive passages, since inbred parasites lose alleles through genetic drift. Here, some larval cohorts are comprised of individuals assigned to a single cluster (each individual has Q > 0.7) but differing from one another (i.e. WB5, WB7, WB10, and domestic isolates F1d and F2). This suggests that these cohorts likely have two lineages that have undergone generations of inbreeding. Other cohorts harbor admixed individuals showing a plethora of different patterns of membership (0.3 < Q > 0.7), indicating that they may have undergone more recent admixture. This pattern occurred primarily in cohorts from wild boar, indicating significant gene flow among wildlife isolates. More hosts likely contribute to the sylvatic cycle, since wild hosts will eat from multiple infected carcasses over their lifetime. The admixed patterns in the wild samples described here likely reflect the history of sequential infections in the lifespan of wild animals.
However, the evolutionary forces driving genetic diversity depend on the amount of genetic variability in a population. In cases of species as homogeneous as T. spiralis in Europe, which has a small effective population size, the admixture level is limited by the number of genotypes circulating in the environment [44]. Of the parasites sampled from wild boar, those of WB6 most closely resembled those from the two farms; this is notable, given how far away WB6 was hunted from these farms. Low overall genetic variability of T. spiralis in Poland increases the likelihood that geographically disparate isolates share genetic similarities.
In Trichinella spp., transmission of larvae from wild to domestic hosts typically occurs as a single, isolated event. The limited number and relatedness of larvae ingested in a single meal often results in further inbreeding in the domestic host. This inbreeding leads to clonal expansions that differ from subpopulations in other hosts. We have demonstrated that epidemiologically related infections can be recognized as especially similar even when genetic variability is low (with fewer than two alleles per locus and an expected heterozygosity of only 10%). This was only achieved through the use of microsatellite loci, which proved variable, despite the low genetic variability among European T. spiralis subpopulations. Although outbreak tracing would be further aided by greater polymorphism or the occurrence of private alleles, we conclude that these markers can help traceback efforts even under conditions of limited overall genetic variability.