Comparison of anisakid infection levels between two species of Atlantic mackerel ( Scomber colias and S . scombrus ) off the Atlantic Portuguese coast

1 Faculty of Sciences of University of Porto, Biology Department, Animal Pathology Laboratory, Rua do Campo Alegre, s/n, FC4, 4169-007 Porto, Portugal. (MJS) (Corresponding author) E-mail: mjsantos@fc.up.pt. ORCID-iD: http://orcid.org/0000-0001-6655-491X (RC) E-mail: sr.ricardocastro@gmail.com. ORCID-iD: http://orcid.org/0000-0002-4381-3605 (FC) E-mail: fcavaleiro@fc.up.pt. ORCID-iD: http://orcid.org/0000-0001-9978-3401 (LR) E-mail: luisfiliperangel@sapo.pt. ORCID-iD: http://orcid.org/0000-0002-8503-7763 2 CIIMAR/CIMAR Interdisciplinary Centre of Marine and Environmental Research, University of Porto, Terminal de Cruzeiros do Porto de Leixões, Av. General Norton de Matos s/n, 4450-208 Matosinhos, Portugal. 3 Rostock University, Faculty of Agricultural and Environmental Sciences, Aquaculture und Sea Ranching, Justus-von-Liebig-Weg 6, 18059 Rostock, Germany. (HP) E-mail: harry.palm@uni-rostock.de. ORCID-iD: http://orcid.org/0000-0003-2918-3253


INTRODUCTION
Portugal has a long tradition of fishing and is the leading country in fish consumption in the European Union (EUMOFA 2016), with 55.3 kg of sea food per capita per year in 2014. For this reason, zoonotic infections such as anisakiasis are a potential risk and a case of major public health concern in this country. Fish dishes, such as grilled (slightly cooked) fish, are one of the most popular food specialties in Portugal, not only among the local people but also among the many tourists who visit the country each year. Atlantic mackerels Scomber colias Gmelin, 1789 and Scomber scombrus L., 1758 are pelagic-neritic scombrids of great commercial interest that are consumed lightly grilled. According to the Portuguese Institution for Statistical Data, 29543 t of S. colias and 588 t of S. scombrus were landed in 2014 (DGRM 2015), with S. colias being one of the top four fish sold in the country.
Anisakid nematodes are the most abundant parasites of marine fishes worldwide (Mattiucci and Nascetti 2008). They are parasites of zoonotic potential, having the ability to infect humans through the consumption of raw or lightly cooked fish, causing the emerging infection anisakiasis, and thus being a food safety concern (MacCarthy and Moore 2000). Associated symptoms of anisakiasis include severe gastric or intestinal disease or, in the mild version, allergies (Audicana et al. 2002(Audicana et al. , 2003. In Portugal, no reports of severe disease due to Anisakis infections have been recorded so far. However, in a survey conducted among the population in a fishery town located in the south of the country, it was found that 8% of the people were allergic to anisakids (Nunes et al. 2003). In Spain, France, Italy, Germany, the Netherlands and Japan, severe episodes of anisakiasis have already been recorded (Arizono et al. 2012, Mattiucci et al. 2013, Ubeira et al. 2000. Both S. colias and S. scombrus are common hosts of anisakids in different geographic areas, particularly in Europe (Abollo et al. 2001, Mattiuci and Nascetti 2008, Pontes et al. 2005. There are also several records from African countries (Abattouy et al. 2011, Farjallah et al. 2008, Kijewska et al. 2009). So far, no updated information is available for Atlantic mackerels fished off continental Portugal, and no attempt has been made to identify the anisakids isolated from S. scombrus to the species level (Rego et al. 1985).
The main purpose of this study was to determine and compare the infection levels (prevalence and intensity) of anisakid nematodes in S. colias and S. scombrus, to correlate parasite and host data, and to evaluate whether the fish species can represent a danger when consumed raw or undercooked, taking into account (i) the infecting species and (ii) the recorded infection levels.

Host sampling
Several samples of Atlantic chub mackerel, S. colias (n=40 in total), and Atlantic mackerel, S. scombrus (n=42 in total), were collected from October to December 2009 (n=19 and 21, respectively) and from January to June 2010 (n=21 for both species). The Atlantic mackerels were fished in the FAO area Atlantic Northeast 27, subarea Portuguese waters IX, by trawling fisheries operated by local boats. The fish were all purchased at the harbour fish market and then freshly dissected or frozen for subsequent parasitological analyses. Host identification was done according to Collette (1986), based on the external and internal features. Fish length was also evaluated [mean±standard deviation (range)]: 32.6±3.6 (25.4-39.2) cm for the Atlantic chub mackerel; and 31. 9±2.3 (21.8-36.4) cm for the Atlantic mackerel (see Table 1). As some parasites might accumulate in older, i.e. larger fishes, and thus bias our sample, a set of hosts with similar lengths were used for the analysis. Also, fish weight (g), Fulton's condition factor (K=weight/length 3 g cm -3 ), hepatosomatic index (HI=liver weight (g) / total weight (g) × 100), gonadosomatic index (GI=gonadal weight (g) / total weight (g) × 100) and sex ratio were recorded for each fish species (see Table 1).

Parasitological survey
During the parasitological survey for anisakids, the following infection sites were analysed under the primer hospedador intermediario principal de A. pegreffii en comparación com lo de A. simplex (s.s.). Los mayores niveles de infección de A. simplex (s.s.) (más infecciosos para los humanos) en S. scombrus sugieren que su consumo ligeramente cocinados, como en el pescado a la parrilla (tan popular en Portugal), podría ser más problemático en relación con el desarrollo de la anisakiasis en los seres humanos, en comparación con el consumo de S. colias y por lo tanto ser de posible preocupación de salud pública. stereo-microscope (magnification 30×): digestive tract, gonads, heart and muscle. The muscle portion examined was cut off from the ventral part of the fish, and also from the belly flaps, as this is considered one of the most infected muscular regions in the fish according to Mehrdana et al. (2014). The isolated worms were cleaned in saline solution (0.9% NaCl) and fixed and preserved in 70% ethanol. For morphological identification purposes, anisakids were cleared and mounted in glycerine, following Moravec (1998). Since around 700 worms were sampled from each fish species, a subsample of worms found around the viscera (and morphologically identified as Anisakis simplex s.l., type I) was identified to the species level using molecular tools. Moreover, around 19 anisakids were taken from about 19 specimens of each species of scombrids (i.e. one worm/fish at random). A minimum sample size of 15 worms was chosen as recommended for calculating reliable prevalence levels (Jovani and Tella 2006). Specimens of Hysterothylacium, occasionally found in both fish hosts, were also characterized using the same molecular methods.
Twenty µl of each DNA sample and 2.5 µl of each primer per sample was sent to the laboratory (GATC Biotec AG, European Custom Sequencing Centre, Cologne, Germany) for sequencing. In total, 38 worms were sequenced. The species were identified by comparison with sequences previously deposited in Gen-Bank, using the BLASTn algorithm and the BioEdit software version 7.1.3.0 (Hall 1999) to previously aligned sequences forward and reverse provided by the sequence laboratory.

Infection levels of anisakids
Prevalence and intensity [mean±SD (minimummaximum)] were calculated for each identified taxon according to Bush et al. (1997) and considering each of the two species of host.
Binary Pearson correlations among anisakid intensity and host total weight, total length, condition factor, hepatosomatic index and gonadosomatic index were conducted. Moreover, the anisakid intensities were analysed in smaller versus larger fish for host total weight, total length, condition factor, hepatosomatic index, gonadosomatic index and sex (male versus female), using the Mann-Whitney's U test. In the later analysis, the two classes of fish were separated by the median value for each parameter. In the case of detecting significant differences, the anisakid intensities [mean±SD (minimum-maximum)] for the smaller and bigger hosts were determined. Non-parametric tests were chosen for the statistical analysis because the study of the normality using the Kolmogorov-Smirnov's test for some variables, such as anisakid intensity, showed that they did not follow a normal distribution (needed in parametric tests), as for instance. Occurrence of A. simplex s.s. and A. pegreffii was compared between the two species of host using the chi-squared test.
All statistical analyses were conducted in SPSS for Windows, version 23 (level of significance: P<0.05).

Anisakid infection levels
Both Atlantic mackerels recorded very high infection levels -a prevalence of anisakids of 85.0% and 83.3% for S. colias and S. scombrus, respectively. In total, 1312 anisakid worms were sampled from both fish, 737 from S. colias and 575 from S. scombrus. Mean intensity was higher for S. colias, with 21.7 worms/host, than for S. scombrus, with the 16.4 worms/host. These infection values were recorded for worms mainly found around the viscera, because the muscular tissue was found infected only once, with a single anisakid recorded in S. scombrus (prevalence of 2.4%) and none in S. colias (prevalence of 0%).
Binary correlations between anisakid intensity and host total weight, total length, condition factor, hepatosomatic index and gonadosomatic index were found to lack statistical significance.
The comparison of anisakid intensities among bigger and smaller hosts, separated by the median value, for host total weight, total length, condition factor, hepatosomatic index, gonadosomatic index and sex (males versus females) were in most cases non-significant, with probability values higher then 0.306, except for the gonadosomatic index, which was significant and recorded a probability value of 0.001. The intensities [mean±SD (range)] for the immature and mature hosts, according to the gonadosomatic index weight, were 6.0±8.6 (1-34) and 27.5±56.8 (2-245) worms per fish, respectively.

Anisakid molecular identification
Among the 36 Anisakis worms sequenced, several sequences were deposited in GenBank (accession numbers KF923929 and KF923930 for A. simplex (s.s.) and KF923927 for A. pegreffii from S. scombrus; and KF923928 for A. simplex (s.s.), and KF914636 and KF923926 for A. pegreffii from S. colias). The distribution for each Anisakis species for both fish hosts is given in Table 2. A. simplex (s.s.) was significantly more frequent in S. scombrus than in S. colias, whereas the opposite trend was reported for A. pegreffii (χ 2 =9.03, P<0.01).

DISCUSSION
In this study, infection levels for anisakids were evaluated considering two species of Atlantic mackerel, S. colias and S. scombrus, sampled at the continental Portuguese coast. The high infection levels (83%-85% prevalence) recorded for both fish are greater than the 10% reported for S. scombrus off continental Portugal by Rego et al. (1985), and the 69.5% reported for S. colias off the Madeira Islands, Portugal, by Costa et al. (2003). However, they are lower than the 100% reported for S. colias off the Azores Islands, Portugal, by Shukhgalter (2004). High prevalence values were also recorded in other geographic regions: 87% for S. colias in El Rincon, Argentina (Cremonte and Sardella 1997) and 92% for S. scombrus and 100% for S. colias in the Adriatic Sea off Croatia (Mladineo 2003, Mladineo andPoljak 2014). In contrast, in Moroccan waters, prevalence of anisakids in S. colias varied from 57% to 67.9% (Abattouy et al. 2011). High prevalence values were also recorded for anisakids of other fish species from the Portuguese mainland, such as the blue whiting (Micromesistius poutassou, 77.7%), the black scabbardfish (Aphanopus carbo, 100%) (Cruz et al. 2007(Cruz et al. , 2009 and the blackspot seabream (Pagellus bogaraveo, 100%) (Hermida et al. 2012). Moreover, Costa et al. (2004) also reported a high prevalence (89.6%) for the blackspot seabream off Madeira Islands, Portugal.
In our samples, not only the overall prevalence but also the mean intensity of anisakids was high, with 21.7 worms per host for S. colias and 16.4 worms per host for S. scombrus. These values were similar to the ones recorded from the Azores Islands (Hyeres Bank), with 17.1 worms per host for S. colias (Shukhgalter 2004).
The correlation analysis of anisakid intensity and host total weight, total length, condition factor, hepatosomatic index and gonadosomatic index for S. colias and S. scombrus was found not to be significant. Moreover, the comparison analysis of anisakid intensity among smaller and bigger fish according to host total weight, total length, condition factor, hepatosomatic index, gonadosomatic index and sex (males versus females) showed mostly the same behaviour: the results were non-significant, except for the gonadosomatic index in S. scombrus. In the latter, more mature fish, with higher index values, recorded significantly higher intensities of anisakids than fish with lighter gonads, with mean values of 6.0 and 27.5, respectively. These findings can be interpreted as a conspicuous high preference for the first intermediate host by the more mature hosts. The lack of significant values for correlation between intensity of anisakids and host data was also recorded by Mladineo and Poljak (2014) for S. colias and Sardina pilchardus in the Adriatic Sea, and by Costa et al. (2004) for Pagellus bogaraveo off Madeira Island. By contrast, Cruz et al. (2009)  recorded a significant increase in anisakid intensity according to the host length for Aphanopus carbo off Sesimbra (mainland) and Madeira Islands, Portugal. Additionally, Chou et al. (2011) found significantly higher intensity of Anisakis larvae in larger and older spotted mackerel (Scomber australasicus) off the Taiwanese coast. Furthermore, Mladineo and Poljak (2014) recorded significant values for host length and abundance of Anisakis from anchovies (Engraulis encrasicolus), European hake (Merluccius merluccius) and whiting (Merlangius merlangus) in the Adriatic Sea. The correlation significance between either intensity or abundance of Anisakis and host length, seems to be dependent on the fish species, and for S. colias the present work is at least the second time that it has been measured (Mladineo and Poljak 2014), but with no significance. In our study, the edible portion of the fishes, the muscle, recorded low or absent infection values, with 2.4% for S. scombrus and 0% for S. colias. Nevertheless, a potential risk of infection should not be discarded, because anisakid larvae can migrate from the viscera to the muscle after death of the host (Smith andWootten 1975, Hauck 1977). These results suggest that fish caught from Portuguese northern Atlantic waters are highly susceptible to carry infections with anisakid nematodes and thus represent a potential human health problem. Taking into account these infection values, we advise that Atlantic mackerels caught off Portugal be ingested with care, and only after being properly frozen or cooked. The European Union (2011) recommends that for the safe consumption of fishery products, raw or undercooked, a previous period of freezing and storage at a core temperature of -20°C or below for not less than 24 h, or of -35°C or below for not less than 15 hours should be applied. In addition, the recommendation of fish evisceration and low temperatures on board just after the catch should be followed more closely. These procedures will avoid as much as possible the migration of the worm to the muscle after the host dies.
The present study identifies for the first time the anisakid species found in mainland Portuguese Atlantic mackerels, S. scombrus, and recognizes the presence of mixed infections with both A. simplex s.s. and A. pegreffii in the two scombrid fish species. For S. colias off Madeira Island, Pontes et al. (2005) recorded six different Anisakis species for S. colias, namely A. simplex s.s., A. pegreffii, A. nascetti, A. typica, A. ziphidarum and A. physeteris. The detection of A. simplex (s. s.) and A. pegreffii mixed infections have also been recorded for S. scombrus in Galicia from the northwestern coast of Spain (Abollo et al. 2001), and for S. colias and S. scombrus in the Alboran Sea off the southern coast of Spain (Abollo et al. 2003). These findings corroborate the statement that the Portuguese and Spanish coasts are a sympatric area for both parasite species (Abollo et al. 2001, Marques et al. 2006. Taking into account the distribution of the different Anisakis species in our fish sample, we further hypothesize that this pattern may be related to differences in the anisakid life cycle hosts, since the two fish mainly occur at different depths. A. pegreffii larvae are probably found in deeper waters, as are their main first host, than A. simplex (s.s.) larvae in less deeper waters, because A. pegreffii larvae were more frequent in S. colias (a deeper fish living at 50-300 m) than in S. scombrus (living at 0-200 m depth). However, this subject is controversial. According to Mattiucci et al. (1997), A. simplex (s.s.) is found mainly in benthic or demersal fishes, while A. pegreffii is found mainly in pelagic fishes, based on the marine mammal final host distribution. Abollo et al. (2001) confirmed the occurrence of A. simplex (s.s.) in benthic and demersal fishes but could not confirm the occurrence of A. pegreffii in pelagic fishes, having analysed a range of different fishes and cephalopods. Kuhn et al. (2013) found that both Anisakis species have similar proportions of fish hosts that live in pelagic, benthopelagic and demersal environments. However, after a second analysis of this author's data, we may see that A. pegreffii occurs in a higher number of reef-associated fish species than A. simplex (s.s.), thus supporting our hypothesis. Also, the high abundance of A. pegreffii in Trachurus trachurus (Abattouy et al. 2013), a benthopelagic fish reaching depths of 0 to 1050 m (Lloris and Moreno 1995), supports our hypotheses. The geographic distribution of a parasite will be the conjunction of the distributions of its different hosts belonging to its life cycle. And a parasite with high host specificity will have a smaller distribution, probably limited to that of its hosts, than a non-specific parasite, which is expected to have a broad distribution.
We are aware that further data on the distribution within the Anisakis simplex complex at the invertebrate host level is needed in order to prove our hypothesis. Smith (1983) proposed that euphausiids were the major intermediate host of A. simplex s.l., in the northeast Atlantic and northern North Sea, and perhaps universally. Recently, Gregori et al. (2015) detected the occurrence of larvae of A. simplex (s.s.) and A. pegreffii in the euphausiid Nyctiphanes couchii, justifying the co-occurrence of both species in the same fish species, and its sympatry in the northeast Atlantic. However, other invertebrate species were also indicated as host candidates for Anisakis species, such as copepods (eg. Acartia tonsa, and Oitona similis) (Koie 2001). The low infection levels recorded for these first intermediate hosts to date in the ocean do not allow us to fully understand what species are more important for the life cycle of each Anisakis species, but further research studies will certainly solve this problem in the short term.
According to earlier studies, the two recorded Anisakis species have a different potential zoonotic level in humans: A. simplex (s.s.) is more prone to migrate into the muscle than the other species, and also withstands better the gut juices of the human stomach, with high survival rates (Arizono et al. 2012). These differences in parasite behaviour explain why more human infections have been reported due to A. simplex (s.s.) than to A. pegreffii in Japan, where both species co-occur and the ingestion of raw fish is particularly common (Umehara et al. 2007). Moreover, the average number of Anisakis muscle larvae in Scomber japonicus can be 12 times higher for A. simplex s.s. than for A. pegreffii, (Suzuki et al. 2010). Taking into account that our Atlantic mackerels have significantly different infections with these two Anisakis species and likewise a similar anisakid load, we may conclude that it is less safe to eat S. scombrus, since it harbours more A. simplex (s.s.) than S. colias, unless it is appropriately frozen or well cooked.