Prevalence, Molecular Identification, and Risk Factors for Cryptosporidium Infection in Edible Marine Fish: A Survey Across Sea Areas Surrounding France

Abstract

Cryptosporidium, a zoonotic pathogen, is able to infect a wide range of hosts including wild and domestic animals, and humans. Although it is well known that some parasites are both fish pathogens and recognized agents of zoonosis with a public health impact, little information is available concerning the prevalence of Cryptosporidium in wild aquatic environments. To evaluate the prevalence of Cryptosporidium spp. in commercially important edible marine fish in different European seas (English channel, North sea, Bay of Biscay, Celtic sea and Mediterranean sea), 1,853 specimens were collected as part of two surveys. Nested PCR followed by sequence analysis at the 18S rRNA gene locus was used to identify Cryptosporidium spp. The overall prevalence of Cryptosporidium spp. in sampled fish reached 2.3% (35 out of 1,508) in a first campaign and 3.2% (11 out of 345) in a second campaign. Sequence and phylogenetic analysis of positive samples identified Cryptosporidium parvum (n = 10) and seven genotypes which exhibited between 7.3 and 10.1% genetic distance from C. molnari, with the exception of one genotype which exhibited only 0.5-0.7% genetic distance from C. molnari. Among 31 analyzed fish species, 11 (35.5%) were identified as potential hosts for Cryptosporidium. A higher prevalence of Cryptosporidium spp. was observed in larger fish, in fish collected during the spring-summer period, and in those caught in the North East Atlantic. Pollachius virens (saithe) was the most frequently Cryptosporidium positive species. In fish infected by other parasites, the risk of being Cryptosporidium positive increased 10-fold (OR: 9.95, CI: 2.32-40.01.04, P = 0.0002). Four gp60 subtypes were detected among the C. parvum positive samples: IIaA13G1R1, IIaA15G2R1, IIaA17G2R1, and IIaA18G3R1. These C. parvum subtypes have been previously detected in terrestrial mammals and may constitute an additional source of infection for other animals and in particular for humans. Microscopical examination of histological sections confirmed the presence of round bodies suggestive of the development of C. parvum within digestive glands. We report herein the first epidemiological and molecular data concerning the detection of Cryptosporidium in edible marine fish in European seas surrounding France broadening its host range and uncovering potential novel infection routes.

Keywords: 18S rRNA gene; Cryptosporidium; European seas; edible marine fish; gp60; molecular epidemiology; novel genotypes; phylogeny.

FIGURE 1

Multiple correspondence analysis. The cluster analysis explained 41% of the total variation. Variables which are closest to each other and distant from the center on the scatterplot are the most likely related. Even if these variables are not grouped as perfect clusters, most Cryptosporidium positive fishes are in the zone encircled (dot lines) that we called cluster 1. In general, this cluster contained larger fishes mainly of size and weight groupings (calibers) 4 and 5, from the order Gadiformes, caught in the North East Atlantic. PELGAS (PELagiques GAScone) 1, campaign 2011; PELGAS 2, campaign 2012; PELMED (PELagiques MEDiterranée) 1, campaign 2011; PELMED 2, campaign 2012; EVHOE (EValuation Halieutique de l’Ouest de l’Europe) 1, campaign 2011; EVHOE 2, campaign 2012; IBTS (International Bottom Trawl Survey), campaign 2012; MGBF, Retailers. Season 0, fall/winter; Season 1, spring/summer. Order 1, Gadiformes; Order 2, Perciformes; Order 3, Clupeiformes; Order 4, Lophiiformes; Order 5, Pleuronectiformes; Order 6, Salmoniformes; Order 7, others. Fishes were classified according to weight and size and five groups were defined ranging from 1-smallest fishes to 5-largest fishes (Supplementary Figure S1) using a hierarchical cluster analysis with the R stats package.

FIGURE 2

Phylogenetic tree showing the evolutional relationships of Cryptosporidium piscine isolates inferred by ML analysis of 18S rRNA gene sequences. Percentage support (>50%) from 1,000 pseudoreplicates from ML and distance analyses and posterior probabilities from Bayesian analysis are indicated at the left of the supported node. Red texts correspond to the sequences from this study.

FIGURE 3

Global distribution of different types of Cryptosporidium sequences identified at the 18S rRNA gene locus according to fish species found as hosts in both surveys (n = 46). Eleven new species of fish were identified as potential hosts for Cryptosporidium. Cryptofish 1 was the most frequently identified novel genotype. Cryptosporidium genotypes had less host diversity when compared to C. parvum which was found in seven different fish species.

FIGURE 4

Distribution of different types of Cryptosporidium sequences identified at the 18S rRNA gene locus according to the anatomical location. (A) First survey (n = 35). C. parvum was identified in the intestine of 7 fishes and in the stomach of one fish, while the C. molnari-like genotype (#Cryptofish 6) was found in the stomach of one fish. The various novel Cryptosporidium genotypes were detected in the stomach of 20 fish, in the intestine of one fish and in both the intestine and stomach of 5 fish. (B) Second survey (n = 11). C. parvum was identified in the bowel of one fish and the stomach of one fish. The 3 novel Cryptosporidium genotypes were identified simultaneously in the stomach and bowel of one fish, in the stomach only of four fishes, and in the intestine only of four fishes.

FIGURE 5

Stained sections of the intestinal tract of fishes. (A,B) Presence of round bodies suggestive of the developmental stages of C. parvum observed in the apical position (arrows) within the intestinal epithelial cells. Bars = 15 μm (A) and 5 μm (B).