Mce

MMa

b

A

a

ARRA

KTOMGP

0d

Veterinary Parasitology 183 (2012) 305– 316

Contents lists available at ScienceDirect

Veterinary Parasitology

jo u rn al hom epa ge : www.elsev ier .com/ locate /vetpar

icrohabitat use, not temperature, regulates intensity of Gyrodactylusichlidarum long-term infection on farmed tilapia—Are parasitesvading competition or immunity?

iguel Rubio-Godoya,∗, Germán Munoz-Córdovab, Mario Garduno-Lugob,artha Salazar-Ulloab, Gabriel Mercado-Vidala

Instituto de Ecología, A.C., km 2.5 ant. carretera a Coatepec, Xalapa, Veracruz 91070, MexicoCentro de Ensenanza, Investigación y Extensión en Ganadería Tropical (CEIEGT), Facultad de Medicina Veterinaria y Zootecnia, Universidad Nacionalutónoma de México, km 5.5 Carretera Federal Martínez de la Torre-Tlapacoyan, AP 103, Martínez de la Torre, Veracruz 93600, Mexico

r t i c l e i n f o

rticle history:eceived 2 March 2011eceived in revised form 13 July 2011ccepted 21 July 2011

eywords:ilapiareochromisonogenea

yrodactylusarasite ecology

a b s t r a c t

Gyrodactylids (Monogenea) are ectoparasites of fish, some of which negatively affect com-mercially valuable fishes. Temperature strongly regulates population dynamics of theseviviparous flatworms in farmed and wild fish populations, with most gyrodactylid speciesshowing positive temperature-abundance associations. In agreement with epidemiologi-cal theory, numerous laboratory studies demonstrate that these parasites cannot persist inconfined fish populations without periodic introduction of susceptible hosts. Extinction ofgyrodactylid populations is due to host immunity, which develops in several fish species.In this one-year study, we followed populations of the recognized pathogen Gyrodactyluscichlidarum infecting four genetic groups of confined tilapia (wild type Nile tilapia Ore-ochromis niloticus niloticus, red O. n. niloticus, Mozambique tilapia O. mossambicus and a redsynthetic population called Pargo-UNAM) kept under farming conditions and subject tonatural environmental fluctuations. Based on the antecedents given, we postulated the fol-lowing three hypotheses: (1) parasite abundance will be regulated by water temperature;(2) parasites will induce host mortality, particularly during periods of rapid infrapopulationgrowth; and (3) gyrodactylid populations will eventually become extinct on confined fishhosts. We disproved the three hypotheses: (1) parasite numbers fluctuated independentlyof temperature but were associated to changes in microhabitat use; (2) although gyro-dactylid populations exhibited considerable growth, no evidence was found of negativeeffects on the hosts; and (3) infections persisted for one year on confined fish. Microhabitatuse changed over time, with most worms apparently migrating anteriorly from the caudalfin and ending on the pectoral fins. Gyrodactylid populations followed similar trajecto-ries in all fish, aggregating and dispersing repeatedly. Several instances were found whereincreased parasite dispersion coincided with increased intensity of infection; as well as the

opposite, where increased aggregation coincided with parasite population declines. Threealternative explanations could account for these observations: that parasites (1) experi-ence differential mortality on different anatomical regions of the fish; (2) migrate to avoid

etition; and (3) migrate to escape localized immune responses induced

intraspecific comp by infection. Our data do not allow us to demonstrate which of these alternatives is correct,so we discuss the merits of each. We provide circumstantial evidence in support of the third

∗ Corresponding author. Tel.: +52 228 842 1800x6208; fax: +52 228 818 7809.E-mail addresses: miguel.rubio@inecol.edu.mx, mrubiogodoy@yahoo.com (M. Rubio-Godoy).

304-4017/$ – see front matter © 2011 Elsevier B.V. All rights reserved.oi:10.1016/j.vetpar.2011.07.040

306 M. Rubio-Godoy et al. / Veterinary Parasitology 183 (2012) 305– 316

explanation, because as shown in other fish host–gyrodactylid interactions where immuneresponses have been characterized, in this study worms progressively moved away fromfins with high mucus cell density to those with low density – what would be anticipated ifimmune defenses occur and reach the fish surface through mucus.

1. Introduction

The negative impact that farm-origin fish parasitescan have on wild fish stocks has become apparent inthe past decades, mainly because of two ectoparasiticpathogens that severely affect aquaculture and endangerthe survival of wild salmonid stocks: the monogeneanflatworm Gyrodactylus salaris Malmberg, 1957 (see Bakkeet al., 2007) and the copepod sea louse Lepeophtheirussalmonis Krøyer, 1837 (see Costello, 2009). Gyrodactylidreproduction is characterized by a short generation timeand completion of the life-cycle in situ without the needfor transmission; these characteristics make these para-sites both pathogenic and valuable epidemiological models(Bakke et al., 2007). Major factors affecting the seasonalprevalence and abundance of gyrodactylid worms includeabiotic factors like water temperature and chemistry, aswell as host-related factors like behaviour, sex, age, resis-tance and mortality (Bakke et al., 2002). Temperature isprobably the most important macroenvironmental factordriving gyrodactylid population dynamics, with severalfreshwater Gyrodactylus species exhibiting positive cor-relations of parasite numbers and water temperature inboth the laboratory (e.g., Scott and Nokes, 1984; Jansenand Bakke, 1991) and the field (e.g., Appleby and Mo,1997; Mo, 1997; Anttila et al., 2008; Winger et al., 2008)– although field studies usually show peak parasite inten-sities in spring and sharp declines during summer; anda few gyrodactylid species exhibit negative temperature-abundance associations (Kirby, 1981; Dávidová et al., 2005;Bakke et al., 2007).

Laboratory-based experiments conducted at constanttemperature have demonstrated that although gyro-dactylid population dynamics are unstable and oscillate(Scott and Anderson, 1984; Bakke et al., 1991), infectionsgenerally cannot persist in confined fish populations with-out periodic introduction of susceptible hosts (e.g., Scottand Anderson, 1984; Boeger et al., 2005; Pie et al., 2006).Parasite extinctions in confined fish groups agree withepidemiological theory stressing the need for pathogenpersistence of new, naïve individuals within a host pop-ulation (Altizer et al., 2006). New susceptible individualsare required for parasite survival due to the developmentof host immunity, which severely limits infection. Indeed,graphs illustrating population dynamics of gyrodactylidson experimentally infected three-spined sticklebacks, Gas-terosteus aculeatus L. (see Lester and Adams, 1974) andguppies, Poecilia reticulata Peters (see Scott, 1985), areparadigmatic of the effect of host immunity on parasites:

in naïve fish, worm burdens rise rapidly to a peak inducingsignificant host mortality, but then infection levels declineand surviving hosts are refractory to subsequent parasitechallenges. Gyrodactylus infections have been shown to

© 2011 Elsevier B.V. All rights reserved.

induce protective responses in several fish species, includ-ing several salmonid fishes (Buchmann and Bresciani,1998; Bakke et al., 2002; Buchmann et al., 2004; Rubio-Godoy, 2007).

Some gyrodactylids exhibit marked site specificity(topographical specialization), although microhabitat usevaries between species (Bakke et al., 2007). For instance,Gyrodactylus bullatarudis Turnbull, 1956 is most commonlyfound on the head and mouth of guppies, whereas Gyro-dactylus turnbulli Harris, 1986 preferentially occupies thecaudal region. Gyrodactylid microhabitat use is believedto be determined by a variety of parasite and host fac-tors, including access to resources, intra- and interspecificcompetition, parasite age, mating, transmission, and hostimmunity (Buchmann and Bresciani, 1998; Pie et al., 2006;Bakke et al., 2007). A few studies documenting the dynam-ics of gyrodactylid spatial specialization over the courseof an infection suggest that changes in microhabitat usemay be related to parasite abundance. Thus, increases inparasite infrapopulations have been suggested to resultin worms occupying sites that favour transmission, likefins (Harris, 1988; Mo, 1997). Gyrodactylid transmissionoccurs primarily through direct host contact (Soleng et al.,1999), although active swimming and persistence on deadhosts, the substrate or the water surface also contributeeffectively to dispersal (Bakke et al., 1992; Cable et al.,2002). High worm burdens may also force occupancy ofless favourable anatomical sites on the host. For example, G.salaris preferentially occurs on the fins of Atlantic salmon,Salmo salar L., but colonizes the body at high parasite inten-sity (>1000 worms/host) (Jensen and Johnsen, 1992). Hostimmunity may also result in shifts in microhabitat use,as shown for Gyrodactylus derjavini [sic G. derjavinoidesMalmberg, Collins, Cunningham et Jalali, 2007] on rain-bow trout, Oncorhynchus mykiss Walbaum (see Buchmannand Bresciani, 1998). In this case, as infection progressedand host immunity was activated, worms moved awayfrom epithelial regions rich in mucus cells, through whichimmune effectors are secreted.

The present work describes the changes over a one-year period in parasite intensity and the dynamics ofmicrohabitat use of Gyrodactylus cichlidarum Paperna,1968 infecting four genetic groups of farmed tilapia,exposed directly to water from a river and thus subjectedto natural fluctuations in temperature and other waterphysicochemical variables. Tilapia (Oreochromis spp.) arecommercially important fish: global production of culturedtilapia approaches 3 million tonnes per annum, secondonly to carp (common carp, grass carp, silver carp, bighead

carp and crucian carp) production (F.A.O., 2008). G. cich-lidarum is a recognized pathogen of these tropical cichlidfishes (Paperna, 1996; García-Vásquez et al., 2010). Basedon the antecedents given, we hypothesized the following:

ary Para

(adgorac

2

eN

2

tnLmlGFOSasafmatdt(

eiti

TM

M. Rubio-Godoy et al. / Veterin

1) parasite abundance will be regulated by water temper-ture; (2) parasites will induce host mortality, particularlyuring periods of rapid infrapopulation growth; and (3)yrodactylid populations will eventually become extinctn confined fish hosts. Testing these hypotheses would beelevant for the management of farmed tilapia in Mexicond in other tropical countries where these cichlids areultivated.

. Materials and methods

All procedures described here were approved by thethics committee of the Veterinary School, Universidadacional Autónoma de México, Mexico City.

.1. Fish and farming conditions

Four genetic groups of tilapia were studied: wildype (grey) and selected red Nile tilapia, Oreochromisiloticus niloticus L. (WNT and RNT, respectively) (Garduno-ugo et al., 2004); red Mozambique tilapia, Oreochromisossambicus Peters (MOT); and the red synthetic popu-

ation called Pargo-UNAM (UNAM) (Munoz-Córdova andarduno-Lugo, 2003), whose genetic composition is: 50%lorida red tilapia, 25% Rocky Mountain tilapia and 25% red. n. niloticus. The experiment was conducted at the Escuelaecundaria Técnica Pesquera (ESTP) no. 90, in Nautla, Ver-cruz State, Mexico (20◦12′49′′N, 96◦46′36′′W). The studyite is located on the banks of the River Nautla, ca. 3 kmway from its mouth into the Gulf of Mexico. To obtainry, 20 mature male and 30 female fish were placed in 5-

diameter concrete tanks with 50 cm deep water columnnd 20% daily water change. Once fry were detected in theanks, they were collected and passed through a 2.0 mmiameter plastic mesh to ensure all fish were <11 mm inotal length, a prerequisite for the sexual reversal techniqueHiott and Phelps, 1993).

The sex reversal period lasted 28 days. Fry (1500 fish of

ach genetic group in separate 750-l tanks) were sexuallynverted to male, a common practice in tilapia aquacul-ure to increase yield and avoid uncontrolled procreationn production tanks. For the sexual reversal procedure,

able 1ean size and body condition factor of four genetic groups of tilapia (Oreochromi

Month Standard length (cm) ± SE

Wild typeO. n.niloticus

Red O. n.niloticus

Wild typeO. mossam-bicus

Pargo-UNAMa

September 3.2 ± 1.94 3.5 ± 4.68 3.9 ± 0.73 3.8 ± 1.9October 6.8 ± 0.55 7.0 ± 1.41 6.5 ± 0.75 7.6 ± 2.8November 10.3 ± 1.37 9.8 ± 2.12 8.5 ± 0.96 10.3 ± 2.9December 12.4 ± 0.55 12.3 ± 3.48 11.4 ± 0.63 13.4 ± 2.9January 15.1 ± 1.68 14.1 ± 1.74 13.6 ± 2.28 14.1 ± 3.6February 15.6 ± 1.24 14.8 ± 3.87 14.3 ± 1.50 15.5 ± 2.1March 16.5 ± 1.14 16.1 ± 6.00 15.4 ± 1.07 16.3 ± 3.3April 19.6 ± 2.64 20.1 ± 4.00 18.3 ± 0.75 19.5 ± 4.5May 21.7 ± 1.93 20.3 ± 4.10 20.0 ± 3.26 22.2 ± 1.7June 23.7 ± 1.38 22.6 ± 1.38 20.7 ± 3.78 23.8 ± 1.0July 25.5 ± 2.54 24.0 ± 1.63 23.5 ± 3.26 25.0 ± 4.9August 27.2 ± 1.74 25.2 ± 5.63 23.8 ± 4.75 23.9 ± 4.2

a Pargo-UNAM is a synthetic tilapia hybrid, whose genetic composition is: 50%

sitology 183 (2012) 305– 316 307

fluoximesterone was added to feed at a dose of 10 mg/kgfeed (Phelps et al., 1992). Parasitological examinationsstarted in September 2006, when the sex-reversal periodended and ca. one month-old fish (standard length ca.3–4 cm; Table 1) were transferred to experimental tanks.During the remainder of the study, the four genetic groupswere kept in separate quadruplicate, randomly allocated,750-l plastic tanks, all of which received water directlyfrom the River Nautla. During the first three months (nurs-ery stage), fish received commercial feed with 45% crudeprotein and were kept at high density: initially, 200 frywere placed in the tanks, and five animals were removedfrom each tank every month for parasite sampling. Dur-ing the remainder of the study (growth-out stage), fishreceived commercial feed containing 32% crude proteinand stocking density was low: 65 fish were placed ini-tially in the 750-l tanks, and after 9 monthly samples offive tilapias/tank, only 20 fish survived per tank.

2.2. Physico-chemical water variables

The only physico-chemical variable manipulated wasoxygenation, as tanks were provided with 24-h aeration;other variables (temperature, dissolved ammonia, pH andsalinity) were monitored at regular intervals, but weresubject to natural variation. Temperature and concentra-tion of dissolved oxygen were determined with a portableoxymeter (YSI Incorporated, Yellow Springs, OH, USA);salinity was measured with an SR1 refractometer (AquaticEcosystems, Apoka, FL, USA); ammonia concentration andpH were determined with commercial test kits (Rolf C.Hagen Inc., Montreal, Canada).

2.3. Parasitological examination of fish

G. cichlidarum burdens on the surface of individual fishwere determined at monthly intervals from September2006 to August 2007. Most monthly samples had n = 20,

except: all groups in December (n = 10); WNT in Septem-ber (n = 18); RNT in November (n = 19), March (n = 19) andJuly (n = 12); MOT in February (n = 19); and UNAM in July(n = 19) and August (n = 18). At the end of the experi-

s spp.) farmed between September 2006 and August 2007.

Condition factor K ± SE

Wild typeO. n.niloticus

Red O. n.niloticus

Wild typeO. mossam-bicus

Pargo-UNAMa

4 3.2 ± 0.11 3.3 ± 0.07 3.0 ± 0.07 3.5 ± 0.061 2.9 ± 0.06 3.4 ± 0.10 3.3 ± 0.15 3.5 ± 0.106 3.0 ± 0.12 3.3 ± 0.05 2.9 ± 0.17 3.6 ± 0.091 3.1 ± 0.04 3.5 ± 0.06 3.6 ± 0.08 3.8 ± 0.109 2.9 ± 0.04 3.2 ± 0.11 3.0 ± 0.06 3.7 ± 0.140 2.8 ± 0.05 3.3 ± 0.06 3.0 ± 0.15 3.5 ± 0.204 2.8 ± 0.05 3.3 ± 0.08 3.0 ± 0.06 3.5 ± 0.056 2.9 ± 0.04 3.1 ± 0.09 2.9 ± 0.10 3.5 ± 0.074 3.0 ± 0.25 3.1 ± 0.05 2.8 ± 0.17 3.5 ± 0.097 2.8 ± 0.04 3.0 ± 0.05 3.0 ± 0.12 3.3 ± 0.163 3.0 ± 0.07 3.2 ± 0.06 3.0 ± 0.08 3.4 ± 0.074 2.9 ± 0.06 3.1 ± 0.05 2.8 ± 0.06 3.5 ± 0.07

Florida red tilapia, 25% Rocky Mountain tilapia and 25% red O. n. niloticus.

ary Para

308 M. Rubio-Godoy et al. / Veterin

ment, total sample sizes were: WNT n = 228, RNT n = 220,MOT n = 229, UNAM n = 227. Fish were killed by pithing,measured, weighed and immediately placed in a dish con-taining water. The whole surface of hosts was examinedunder a dissection microscope and the number of G. cich-lidarum was recorded for each of the following anatomicalregions: head, eye, anterior trunk (region between the headand a horizontal line located at the anal protuberance),posterior trunk, and the five types of fins (caudal, dorsal,anal, pelvic and pectoral). Parasite specimens were fixedin 96% (v/v) ethanol for subsequent taxonomic identifica-tion. Samples representative of the four fish groups andall months were identified as G. cichlidarum through mor-phological measurements of the attachment hooks of thehaptor, as described by García-Vásquez et al. (2007). Thegills of all fish were fixed in 10% (v/v) formalin and laterinspected under a dissection microscope. The coefficient ofbody condition K was calculated as proposed by Lemly andEsch (1984).

2.4. Mucus cell density

The density of mucus cells on tilapia fins was deter-mined by adapting the method of Buchmann and Bresciani(1998). In brief, the five types of fin were excised from20 naïve, uninfected fish (10 WNT, 10 UNAM) and fixedovernight in phosphate-buffered 4% formaldehyde, rinsedin distilled water, stained in 1% Alcian blue (Sigma Chemi-cal Co.) with 3% acetic acid, rinsed in distilled water, andfinally mounted in 50:50 glycerine–formaldehyde jelly.The density of superficial mucus cells was determinedunder a compound microscope (1000× magnification) bycounting cells present in 0.64 mm2 fields (five to six ineach fin) from the edge of the fin to the fin base. Fieldswere photographed with a digital camera coupled tothe microscope, and stained mucus cells were countedlater.

2.5. Statistical analyses

2.5.1. Parasitological dataUse of parasitological terms follows that of Bush et al.

(1997). Parasite mean intensities and their 95% confidenceintervals were calculated for all monthly samples with thesoftware Quantitative Parasitology 3.0 (QP) (Rózsa et al.,2000). Monthly mean intensities were compared with QPby calculating bootstrap t-tests, with 2000 replications;comparisons included intraspecific contrasts from monthto month, and monthly comparisons between fish geneticgroups. General Linear Models (GLM) were used in Minitab15.0 to assess whether the use of quadruplicate tanks hadan effect on parasite intensity; the significance level was setat P ≤ 0.05. Spearman’s rank correlations were calculatedwith Minitab to analyze the association between parasiteintensity and host body condition, as well as between par-asite intensity and host size (length).

2.5.2. Microhabitat useAnalyses were performed using the software R 2.6.0 (R

Development Core Team, 2007); the significance level wasset at P ≤ 0.05. Parasite microhabitat use on each individual

sitology 183 (2012) 305– 316

fish was described by three parameters: (1) distributionpattern; (2) parasite aggregation; and (3) migration overthe fish surface.

Parasite distribution was described every month forindividual hosts, calculating the proportion of the wormpopulation found on each anatomical region of the fish(except head and eyes, which only occasionally har-boured a few parasites). Parasite distribution on individualfish was then assigned to one of three defined dis-tribution patterns (caudal, intermediate, pectoral) usingKolmogorov–Smirnov (KS) tests; details of the distributionpatterns are presented in Section 3.

To describe parasite aggregation on the fish surface,Topographical Specialization Indices (TSI) were calculatedfor individual hosts, as described by Pie et al. (2006). Inbrief, the relative number of parasites present in all avail-able anatomical regions of the fish is calculated for eachcensus day, and the TSI are the variance among the result-ing scores. If parasites are present in all available bodystructures, the variance (TSI) will tend to be small; con-versely, high variance (TSI) values indicate that parasitesare aggregated. Mean monthly TSI values were comparedby ANOVA and significant differences were detected posthoc based on Tukey’s Honestly Significant Difference (HSD)test.

Over time, worms tended to migrate from the caudalfin to the pectoral fins (see Section 3 for details). We there-fore quantified in individual fish the inverse logit functionof the difference between parasite intensity in the caudaland pectoral fins, and used this parameter (called migra-tion) as a numerical indication of directionality of parasitemovements on the host surface.

General Linear Models (GLM) with gamma error struc-ture and log link function were used to examine parasiteintensity as a function of month of sampling, parasiteaggregation (TSI), parasite migration and the interac-tion TSI × migration. GLM procedures initially includedinteractions between all factors available including KSdistribution, temperature and water physico-chemicalvariables measured, and were subjected to stepwise sim-plification to obtain minimal significant models.

2.5.3. Mucus cell densityDensity of mucus cells on naïve fish fins was not

normally distributed and was thus analyzed with QP inthe same manner as intensity of infection, by calculatingmeans ± 95% confidence intervals and contrasting groupswith bootstrap t-tests.

3. Results

Experimental fish probably acquired G. cichlidaruminfections from their parents, as fry in the sex-reversaltanks harboured parasites when inspected at the begin-ning of the study (data not shown). Some host mortalitywas observed in all fish groups during the sex-reversalperiod, but this was not quantified nor did we attempt

to establish whether mortality was due to gyrodactylidinfection. No gyrodactylids were found on the gills of anyexperimental fish, so only parasites found on the skin werequantified. There were no significant differences in para-

M. Rubio-Godoy et al. / Veterinary Parasitology 183 (2012) 305– 316 309

Fig. 1. Monthly mean intensities (bars) and aggregation (discontinuous lines) of Gyrodactylus cichlidarum on tilapia, Oreochromis spp. Increases of theT on indivL tstrap tl ent mea

ssPodiosiptww

TCS

N

opographical Specialization Indices (TSI) reflect aggregation of parasitesetters group months with intensities that do not differ significantly in booine of the TSI curve connect consecutive samples with significantly differ

ite intensity between quadruplicate tanks throughout thetudy period (GLM: WNT, P = 0.832; RNT, P = 0.195; MOT,

= 0.697; UNAM, P = 0.398); therefore, monthly samplesf each fish type were pooled. Prevalence of G. cichli-arum infection was 85–100% for all tilapia genetic groupsn September and October, and 100% from Novembernwards; we thus report mean parasite intensities (Fig. 1hows intraspecific contrasts of monthly intensities ofnfection for each fish group over the study period; Table 2

resents monthly interspecific contrasts). With the excep-ion of RNT, no significant changes in intensity of infectionere detected between November and December, a periodhen aquacultural procedures (change from the nursery to

able 2omparison of monthly mean intensities of Gyrodactylus cichlidarum infection oneptember 2006 and August 2007.

Month Mean intensity of infection (worms/fish)

Wild type O. n. niloticus Red O. n. nil

September 19.5a 35.7b

October 5.5a 9.4b

November 8.4ab 12.8a

December 7.0a 23.8b

January 28.8a 56.9b

February 9.5a 20.0b

March 10.8a 26.9b

April 11.5a 23.0b

May 17.6a 22.9a

June 53.4a 42.5a

July 16.5a 14.8a

August 15.5a 38.0b

Overall mean ± SE 17.4 ± 1.55 27.5 ± 1.85

ote: Letters group fish groups with monthly mean intensities that do not differ s

idual hosts, decreases indicate dispersion. Most monthly samples n = 20.-tests with 2000 replications (P < 0.01). Solid arrows on the discontinuousn TSI.

the growth-out stage) resulted in changes in the diet andstocking density of fish, from high to low protein contentand density, respectively (Fig. 1 and Table 2). In January,when the lowest mean water temperature was recorded(Table 3), significant increases in intensity of infection wereobserved in all groups, leading to a distinct peak. A fur-ther peak in parasite intensity was observed in June, whenthe second highest mean temperature was registered. Junesamples included the overall highest mean intensities and

highest individual burdens recorded in three of the fourfish groups studied (Table 2). A third intensity peak wasrecorded in August for RNT and UNAM. Throughout thestudy, UNAM had the highest parasite burdens (overall

four different genetic types of farmed tilapia (Oreochromis spp.) between

oticus Wild type O. mossambicus Pargo-UNAM

5.1c 76.4d

5.7a 17.8c

7.3b 42.4c

6.2a 30.0b

26.5a 104.8c

6.0a 51.5c

8.3a 116.5c

4.9c 22.9b

21.5a 93.8b

34.1a 275.5b

31.3b 40.9b

32.9ab 140.6c

16.1 ± 1.28 86.5 ± 11.50

ignificantly in bootstrap t-tests with 2000 replications (P < 0.05).

310 M. Rubio-Godoy et al. / Veterinary Parasitology 183 (2012) 305– 316

Table 3Mean monthly water physico-chemical quality values ± SE in a culture of four genetic groups of tilapia (Oreochromis spp.) farmed between September 2006and August 2007.

Month Temperature(◦C) Oxygen (mg/l) pH Salinity (ppt) NH4 (mg/l)

September 27.0 ± 0.35 3.7 ± 0.14 7.5 ± 0.08 0.5 ± 0.28 0.3 ± 0.01October 25.7 ± 1.11 3.0 ± 0.24 7.6 ± 0.09 1.0 ± 0.41 0.4 ± 0.11November 20.7 ± 1.45 3.9 ± 0.34 7.6 ± 0.09 1.0 ± 0.41 0.5 ± 0.05December 21.4 ± 1.08 3.9 ± 0.28 7.5 ± 0.01 1.3 ± 0.75 0.5 ± 0.24January 19.6 ± 1.06 4.7 ± 0.21 7.5 ± 0.01 3.0 ± 0.41 0.3 ± 0.05February 21.7 ± 1.18 3.9 ± 0.19 7.5 ± 0.01 1.3 ± 0.48 0.4 ± 0.01March 24.5 ± 1.21 4.0 ± 0.16 7.6 ± 0.10 2.8 ± 1.48 0.3 ± 0.05April 26.7 ± 1.45 3.9 ± 0.16 7.6 ± 0.10 3.0 ± 1.47 0.5 ± 0.11May 28.6 ± 0.46 4.8 ± 0.03 7.8 ± 0.15 4.0 ± 1.82 0.5 ± 0.01June 29.0 ± 0.07 4.7 ± 0.10 7.6 ± 0.12 2.8 ± 1.55 0.7 ± 0.01

July 28.1 ± 1.34 4.7 ± 0.07

August 29.7 ± 0.43 4.3 ± 0.20

mean intensity ± 1SE 86.5 ± 11.50 worms/host), RNT inter-mediate burdens (27.5 ± 1.85 worms/host), and WNT andMOT similarly low burdens (17.4 ± 1.55 and 16.1 ± 1.28worms/host, respectively) (Table 2). No adverse effect ofinfection was detected on the hosts: no significant asso-ciations between body condition (Table 1) and parasiteintensity of individual fish were observed in any geneticgroup during any of the monthly samples (all P val-ues > 0.05); and no significant increases of host mortalitynor decreases of fish growth rate were observed dur-ing periods of parasite population increase (Salazar-Ulloa,BSc thesis, 2008). No significant associations were foundeither between parasite intensity and host size (standardlength) (all P values > 0.05). Except for pH, water abioticparameters fluctuated considerably throughout the study(Table 3); water temperature ranged from ca. 20 ◦C inJanuary to ca. 30 ◦C in August. Variance in parasite inten-sity was not significantly associated to variance of abioticfactors in any of the four genetic groups of tilapia (allP values > 0.05).

Microhabitat selection followed the same trend in allgenetic groups: at the beginning of the experiment, mostworms were found on the caudal fin, whereas towards theend of the study the majority occurred on the pectoralfins (Fig. 2). The distribution pattern of worms found inSeptember was defined as caudal distribution, as the high-est proportion of worms on all tilapia groups occurred onthe caudal fin (means of all groups: 49.7% caudal fin; 17.1%dorsal fin; 13.2% anal fin; 9.1% pectoral fin; 5.2% pelvic fin).Pectoral distribution was defined as that found in August,when most parasites occurred on the pectoral fin (61.6%pectoral fin; 16.2% dorsal fin; 13.8% pelvic fin; 4.9% analfin; 2.8% caudal fin). Parasite distribution found in Octoberwas defined as intermediate, since worms were generallydistributed throughout the whole body and this sample hadleast variance between the percentages of each anatomicalregion of all 12 monthly samples (27.7% dorsal fin; 18.5%pectoral fin; 17.9% caudal fin; 14.2% anal fin; 14.1% pelvicfin). KS tests demonstrated that over the course of one year,parasite populations shifted their preference for attach-ment from the caudal fin to the pectoral fins (Fig. S1). KS

analyses also showed that during most of the year, consid-erable proportions of worms fluctuated between pectoraland intermediate distributions (mainly occurring on dorsaland pelvic fins; Fig. 2).

7.6 ± 0.12 2.8 ± 2.43 0.6 ± 0.107.5 ± 0.10 1.0 ± 0.45 0.4 ± 0.01

Parasite aggregation on the host surface – described byTopographical Specialization Indices (TSI; Fig. 1) – likewisefluctuated over time in all fish groups. Several instanceswere found where changes in parasite aggregation coin-cided with variations in intensity of infection (Fig. 1).In particular, nine cases were documented of decreasedworm aggregation (i.e., increased worm dispersion illus-trated by decreased TSI values) being associated withincreased parasite burdens; and seven of the converse,when increased parasite aggregation (increased TSI) coin-cided with decreased intensity of infection. Correlationsbetween changes in worm aggregation and parasite inten-sity coincided in most cases with peaks of worm intensity;in RNT, MOT and UNAM, several significant decreases in TSIcoincided with significant increases in parasite intensity,and vice versa (Fig. 1).

The effect of parasite migration on intensity of infec-tion was significant in all tilapia groups (Table 4). Variationin intensity was significantly associated to month of sam-pling in MOT and UNAM; and to both parasite aggregation(TSI) and its interaction with migration in MOT and RNT.In all cases, positive correlations were observed betweenparasite intensity and migration, although this associationwas not significant for UNAM (Table 4). Significant negativecorrelations between migration and parasite aggregationwere observed in three out of four cases, except for WNT.As illustrated in the graphic representation of the combinedeffects of parasite migration and aggregation on intensityof infection (Fig. 3), in WNT, worms tended to be moreaggregated at higher parasite burdens. In contrast, in RNTand MOT, worms tended to be less aggregated as parasiteintensity increased. The same was true for UNAM but onlyto a certain extent, because GLM predicted a distinct peakcorresponding to the highest parasite intensities, whereworms tend to be highly aggregated and not to migrate(Fig. 3). The intensity peak in the UNAM graph results inthe marked inflection of the aggregation curve (Fig. 3) andis probably the cause for the non-significant associationfound between parasite intensity and migration (Table 4).

Tilapia fins can be separated in three categories, basedon the mean density ± 1SE of superficial mucus cells

determined in naïve fish: (1) caudal fins, which havesignificantly more mucus cells than all other fin types(35.4 ± 1.61 cells/field); (2) dorsal (17.8 ± 2.05 cells/field),anal (23.0 ± 1.92 cells/field) and pelvic fins (20.4 ± 2.04

M. Rubio-Godoy et al. / Veterinary Parasitology 183 (2012) 305– 316 311

ylus cich

cdsdwU

4

wttic

TSo

NMM

Fig. 2. Microhabitat use of Gyrodact

ells/field), which have similar, intermediate mucus cellensities; and (3) pectoral fins, which have the lowest den-ity (13.0 ± 1.29 cells/field) (bootstrap t-tests, all significantifferences had P values < 0.001). No significant differencesere found in mucus cell densities between WNT andNAM (P values > 0.05).

. Discussion

This one-year study on the infection of farmed tilapiaith G. cichlidarum disproved the three hypotheses we pos-

ulated: (1) parasite numbers fluctuated independently ofemperature but were associated to changes in microhab-tat use; (2) although gyrodactylid populations exhibitedonsiderable growth, no evidence was found of negative

able 4ummary of the statistical analyses of the effect of different factors on the variatiof farmed tilapia (Oreochromis spp.).

GLM factor Tilapia genetic group

WNT R

Migration 0.003** (3.052) 1Aggregation (TSI) 0.757 (0.310) 0Migration: TSI 0.813 (−0.236) 0Month 0.062 (1.875) 0Intensity: migration correlation (coefficient) 6.426e−05*** (0.261) 2TSI: migration correlation (coefficient) 0.475 (−0.048) 0

ote: Figures shown are significance probabilities, some of which are shown as

odels (GLM) and coefficients for correlations. Abbreviations: WNT, wild type NileOT, red Mozambique tilapia, Oreochromis mossambicus; UNAM, Pargo-UNAM; T* P < 0.01.

** P < 0.001.*** P < 0.0001

lidarum on tilapia, Oreochromis spp.

effects on the hosts; and (3) even though fish were con-fined, infections persisted for one year.

No significant effect of temperature nor any of the otherwater physicochemical parameters measured, was foundon G. cichlidarum population dynamics. In fact, peaks in par-asite intensity were observed simultaneously in the fourtilapia groups during both the coldest month (January) andthe second hottest month (June) recorded. Cold negativelyaffects fish immunity (Tatner, 1996; Rubio-Godoy, 2010)and this could explain the increase in infection levels in Jan-uary; and warm summer temperatures favour gyrodactylid

reproduction (Bakke et al., 2007), which would accountfor the peak observed in June. However, our results are incontrast to studies showing that temperature is the mainenvironmental factor underlying the seasonal patterns of

n in intensity of Gyrodactylus cichlidarum infection in four genetic groups

NT MOT UNAM

.02e−14*** (8.318) 9.93e−16*** (8.652) 0.004** (2.893)

.006*** (2.794) 0.015* (2.450) 0.122 (1.550)

.006*** (−2.804) 0.003** (−3.005) 0.091 (−1.697)

.348 (0.941) 2.89e−13*** (7.765) 0.000*** (3.872)

.541e−08*** (0.365) 5.8e−11*** (0.415) 0.106 (0.108)

.019* (−0.157) 7.313e−05*** (−0.259) 0.003** (−0.196)

exponential (e) exact values; brackets show t-values for General Linear tilapia, Oreochromis niloticus niloticus; RNT, red Nile tilapia, O. n. niloticus;SI, Topographical Specialization Indices.

312 M. Rubio-Godoy et al. / Veterinary Parasitology 183 (2012) 305– 316

Fig. 3. Effect of parasite migration and aggregation on the intensity of Gyrodactylus cichlidarum on tilapia, Oreochromis spp. Worm migration was calculatedas the inverse logit function of the difference between parasite intensity in the caudal and pectoral fins of individual fish. General Linear Models (GLM)with gamma error and log link function were used to model parasite intensity as a function of month of sampling, parasite aggregation (TopographicalSpecialization Index), caudal or pectoral distribution, parasite migration and the interaction TSI × migration. GLM initially included the interactions between

variable

all factors available including temperature and water physico-chemical

Significance codes: ***P < 0.0001; **P < 0.001.

abundance and transmission of gyrodactylids (e.g., Lesterand Adams, 1974; Scott and Nokes, 1984; Bakke et al., 2002)and controls the seasonality of infection in natural pop-ulations of several salmonid fishes in temperate regions,including Atlantic salmon, S. salar (see Appleby and Mo,1997; Anttila et al., 2008), brown trout, Salmo trutta L. (seeMo, 1997) and arctic charr, Salvelinus alpinus L. (see Wingeret al., 2008). Whether the lack of association between

temperature and parasite population growth documentedhere applies to other tropical fish–gyrodactylid interac-tions remains to be studied. To our knowledge, no studiesare available on the effect of temperature on the biology of

s measured, and were simplified to obtain minimal significant models.

G. cichlidarum; thus, we do not know whether optimal tem-peratures for this species lie within the observed range ofroughly 20–30 ◦C, nor whether temperatures outside thisrange would exert a more pronounced effect on parasitepopulation dynamics. Nonetheless, it is likely that adequatetemperatures for G. cichlidarum lie within the range weobserved, as these are within the temperature range of itshosts: O. n. niloticus lives between 14 ◦C and 33 ◦C, and O.

mossambicus between 17 ◦C and 35 ◦C (Froese and Pauly,2010).

G. cichlidarum is a recognized pathogen of tilapia (Tilapiaand Oreochromis spp.) and several cases of infection result-

ary Para

i(waidtnmc2sdiissfifit

dtSeyo2cicsfifvdaf(it(it1ded11vahadsicfi

M. Rubio-Godoy et al. / Veterin

ng in high host mortality have been recorded worldwidePaperna, 1996; García-Vásquez et al., 2010). Nonetheless,e did not find any significant associations between par-

site intensity and host condition; nor did we observencreased fish mortality or decreased host growth rateuring periods of parasite population increase. However,his does not necessarily indicate that G. cichlidarum isot pathogenic, because this parasite is known to pri-arily affect tilapia fry – infected juvenile fish stocks

an experience significant mortality (García-Vásquez et al.,010). We observed limited mortality of fish during theex-reversal period, but we did neither quantify it noretermine whether it was a consequence of gyrodactylid

nfection. Nonetheless, it is important to consider thatn this study, the stocking density of fish during theex-reversal procedure (1500 fry in 750-l tanks) was con-iderably lower than that usually employed in commercialsh farms, and this probably reduced both the stress ofsh and the probabilities of host contact and parasiteransmission.

Contrary to our expectation based on various reportsocumenting parasite extinctions in confined host popula-ions (Lester and Adams, 1974; Scott and Anderson, 1984;cott, 1985; Boeger et al., 2005; Pie et al., 2006; Bakket al., 2007), gyrodactylid populations persisted for oneear on farmed tilapia. One reason for this may be thatf the 400+ Gyrodactylus described species (Harris et al.,004), some host–parasite associations exhibit a stableoexistence, with low host mortality and pathogenic-ty. This might be the case for chub (Leuciscus) Squaliusephalus (L.), and Gyrodactylus lomi Ergens et Gelnar, 1988,ince laboratory infections of isolated fish have persistedor up to 10 months (King et al., 2008). Escape frommmunity through transmission between hosts may be aurther mechanism enabling long-term gyrodactylid sur-ival. Previous studies have shown that host immunityoes not permanently damage gyrodactylids, if parasitesre able to escape from it. Thus, gyrodactylids that detachedrom immune sticklebacks were able to infect new hostsLester and Adams, 1974), and 60% of G. salaris exper-mentally transferred to naïve salmon after six days onhe innately resistant S. trutta were able to reproduceBakke et al., 2002). Similarly, fish can develop immunitynvolving both innate and acquired components againsthe ectoparasitic ciliate Ichthyophthirius multifiliis Fouquet,876 (see Dickerson, 2006), resulting in parasites aban-oning immune hosts within 2 h of invasion. However,scaped “Ich” theronts are able to colonize and produceisease in naïve fish. As suggested previously (Bakke et al.,996; Buchmann and Bresciani, 1998; Richards and Chubb,998; Buchmann et al., 2004), evasion of immune indi-iduals may be the reason why gyrodactylids can survivet low intensities of infection on populations of immuneosts. This probably applies to natural fish populationsnd was observed in previous records of long-term gyro-actylid infection of confined fish. In particular, G. salarisurvived up to 280 days on laboratory populations of shoal-

ng arctic charr, S. alpinus (see Bakke et al., 1996). Inontrast, fish which had been infected as part of a shoalor 115 days, lost their parasites within 20–30 days whensolated, suggesting that the fish were immune to the

sitology 183 (2012) 305– 316 313

parasite at the time of isolation, but that worms couldsurvive this when their hosts were shoaling. A furtherexample of persistent, low intensity infection on smallpopulations of confined hosts is that of G. turnbulli on P.reticulata, where six fish maintained infections with <10parasites/host for up to 210 days (Richards and Chubb,1998). Both of these studies document long-term per-sistence of low-level gyrodactylid infections, followingrelatively high host mortality during episodes of par-asite population growth. Our results differ from theseantecedents, because we registered several peaks of para-site population growth with no concurrent host mortality,and relatively high intensities of infection throughout theyear.

In the present study, gyrodactylids periodically aggre-gated and dispersed on the fish and modified their useof host microhabitats over the course of one year, byapparently performing a gradual and progressive migrationtowards the anterior of hosts. Given the monthly intervalbetween samples, we cannot determine whether changesin the relative abundance of parasites on the differentregions of the fish surface effectively resulted from wormsmigrating or from differential parasite survival on theseregions. Nonetheless, changes in microhabitat use wereassociated to several significant changes in parasite inten-sity: thus, various instances were found where increasedworm burdens coincided with increased parasite disper-sion; as well as the converse, where decreased parasiteloads were associated with decreased worm disper-sion. Statistical analyses confirmed significant associationsbetween changes in parasite intensity and microhabitatuse in all tilapia groups. In particular, worm migration wasthe most significant factor of statistical models describingvariation in intensity of infection, and parasite aggrega-tion was also a significant component. A significant positivecorrelation between migration and intensity of infectionwas found in three out of four cases; this, however,does not imply causality, as our analyses cannot deter-mine whether worms migrated first and then increasedin numbers, or whether increasing parasite intensitiesdrove worm migration. Nonetheless, two previously pro-posed alternative explanations (Richards and Chubb, 1996;Buchmann and Bresciani, 1998; Buchmann et al., 2004;van Oosterhout et al., 2008) can account for the signifi-cant association found between gyrodactylid intensity andmicrohabitat use: that parasites are migrating to evadeintraspecific competition or to escape localized immuneresponses.

Microhabitat saturation and intraspecific competitionare thought to be uncommon among the Monogenea, par-ticularly among gill-infecting species (Euzet and Combes,1998; Morand et al., 1999). Indeed, detailed analyses ofmicrohabitat use of gill-inhabiting monogeneans belong-ing to the genus Dactylogyrus Diesing, 1850 infecting roach,Rutilus rutilus L., and crucian carp, Carassius carassius L.,have shown that conspecific parasites are usually aggre-gated (Simková et al., 2001; Bagge et al., 2005). However,

the degree of intraspecific aggregation decreased withincreasing worm burdens, suggesting density-dependentshifts in microhabitat use. Furthermore, microhabitat satu-ration may partially account for the changes of distribution

ary Para

314 M. Rubio-Godoy et al. / Veterin

on the gill arches observed at different infection intensi-ties of several monogeneans, like Haliotrema spp. Johnstonet Tiegs, 1922 infecting damselfish, Dascyllus aruanus L.(see Lo and Morand, 2000); Polylabris mamaevi Ogawaet Egusa, 1980 and Tetrancistrum nebulosi Young, 1967,infecting mottled spinefoot, Siganus fuscescens (Houttuyn,1782) (see Yang et al., 2006); and Discocotyle sagittata(Leuckart, 1842) Diesing, 1850 infecting rainbow trout,O. mykiss (see Rubio-Godoy, 2008). Density-dependentchanges in microhabitat use have also been recorded inseveral gyrodactylids (Harris, 1988; Jensen and Johnsen,1992; Mo, 1997), where topographical specialization mightarise either to evade intraspecific competition or to boosttransmission. Studies of the dynamics of topographicalspecialization of G. anisopharynx Popazoglo et Boeger,2000 infecting three species of armored catfishes of thegenus Corydoras (Gill, 1858) demonstrated that parasitesmigrated on the host surface to avoid competition, ratherthan facilitate transmission (Pie et al., 2006). Our resultsshow that peaks in parasite abundance coincided withincreased worm dispersion; as well as the opposite, whendecreased parasite burdens were associated to augmentedworm aggregation. This would suggest that migratingaway from saturated microhabitats enables gyrodactylidsto escape competition, gain access to resources (includingfood and space) and thrive; or the opposite phenom-ena leading to infrapopulation declines. To test thishypothesis, detailed studies are needed of the temporal-ity and directionality of parasite migration, and of thereproductive output of worms located in different micro-habitats.

A further factor potentially driving the migration ofworms observed in this study is host immunity. Gyro-dactylid infection has been shown to induce immunedefenses in virtually every fish species examined (Lesterand Adams, 1974; Scott, 1985; Buchmann et al., 2004;Buchmann and Bresciani, 2006; Bakke et al., 2007; Cableand van Oosterhout, 2007), and host responses not onlylimit worm proliferation, but have also been shown to influ-ence parasite microhabitat use (Buchmann and Bresciani,1998). Although host responses against gyrodactylids havenot been fully characterized, these probably involve bothspecific and non-specific immune components (Buchmannet al., 2004; Cable and van Oosterhout, 2007). In salmonidfishes, infection with Gyrodactylus spp. induces expres-sion of a series of immune-response genes (Lindenstrømet al., 2004; Matejusová et al., 2006; Collins et al., 2007).Innate defense mechanisms are probably also involved inprotection against these parasites – particularly comple-ment, which has been shown to kill gyrodactylids in vitro(Buchmann, 1998; Harris et al., 1998). Mucus cells arelikely to play a pivotal role in host responses againstgyrodactylids, because complement and other immunecomponents reach the fish surface via mucus (Buchmannet al., 2004). Thus, the action of mucus cells can be thoughtof as a two-edge sword: initially, attachment sites withhigh mucus cell density potentially provide a rich source of

food for epidermal cell- and mucus-feeding gyrodactylids,facilitating parasite growth and reproduction. However,once the immune defense mechanisms are in motion,mucus cell-rich surfaces become inhospitable, as they are

sitology 183 (2012) 305– 316

flooded with protective effectors. Development of local-ized, superficial host immunity influences gyrodactylidmicrohabitat selection, as shown in O. mykiss infected withG. derjavini [sic G. derjavinoides Malmberg, Collins, Cun-ningham et Jalali, 2007] (see Buchmann and Bresciani,1998). Initially, G. derjavini occurred most frequently onmucus cell-rich areas of rainbow trout, but moved awayfrom these over time; and six weeks after infection, sur-viving worms were only found in mucus cell-poor areas,such as the corneal surface. Conceivably, once the wholesurface of fish becomes immune, parasite populationsbecome extinct. Host immunity accounts for the inability ofgyrodactylids to persist on confined populations of experi-mentally infected fish, without the continual introductionof naïve, susceptible hosts.

Although no studies show whether tilapia developimmunity against gyrodactylids and we measured noimmune parameters, we present circumstantial evidencethat this might be the case. First, host immunity couldexplain why in the present study, parasite populationsdid not grow unchecked and did not kill the fish. Fur-ther, the changes in microhabitat use we observed in thefour fish types might be related to immunity – if tilapiaare indeed able to mount immune responses against gyro-dactylids, and immune effectors reach the fish surface viamucus. Provided that density of mucus cells is related to theeffectiveness of putative immune reactions, moving awayfrom mucus cell-rich areas would allow parasites to escapehost defenses. This might explain the observation that G.cichlidarum rapidly migrated away from the caudal fin,which has the highest density of mucus cells; persisted formonths by switching between fins of intermediate mucuscell density; and ultimately were most frequently foundon the pectoral fins, which have the lowest mucus celldensity overall. Nonetheless, further study is needed todemonstrate whether tilapia are able to develop immunityagainst gyrodactylids; whether expression/deployment ofimmune effectors changes over the course of infection anddiffers between distinct anatomical structures of the fish;and whether host defenses influence parasite microhabitatuse.

Acknowledgements

We thank Daniel Aguirre-Fey (INECOL, Mexico) forinspection of fish gills; Lamberto Aragón (INECOL, Mexico)for staining and mounting fins; Roger Guevara (INECOL,Mexico) for help with statistical analysis of microhab-itat use; Adriana García-Vásquez and Andrew P. Shinn(Institute of Aquaculture, University of Stirling, UK) forthe taxonomic identification of parasites; Trevor Williamsand Carlos Montana (INECOL, Mexico) for comments onprevious versions of the manuscript. This study was sup-ported by funds awarded to MRG and GMC by CONACYT(grants FOMIX 32679 and 37487, and CB 58050) and byINECOL and CEIEGT institutional grants. Sponsors had no

role in the study design, in the collection, analysis andinterpretation of data; in the writing of the manuscript;nor in the decision to submit the manuscript forpublication.

ary Para

A

cd

R

A

A

A

B

B

B

B

B

B

B

B

B

B

B

B

C

C

C

C

D

D

M. Rubio-Godoy et al. / Veterin

ppendix A. Supplementary data

Supplementary data associated with this arti-le can be found, in the online version, atoi:10.1016/j.vetpar.2011.07.040.

eferences

ltizer, S., Dobson, A., Hosseini, P., Hudson, P., Pascual, M., Rohani, P.,2006. Seasonality and the dynamics of infectious diseases. Ecol. Lett.9, 467–484.

nttila, P., Romakkaniemi, A., Kuusela, J., Koski, P., 2008. Epidemiologyof Gyrodactylus salaris (Monogenea) in the River Tornionjoki, a Balticwild salmon river. J. Fish Dis. 31, 373–382.

ppleby, C., Mo, T.A., 1997. Population dynamics of Gyrodactylus salaris(Monogenea) infecting Atlantic salmon, Salmo salar, parr in the riverBatnfjordselva, Norway. J. Parasitol. 83, 23–30.

agge, A.M., Sasal, P., Valtonen, E.T., Karvonen, A., 2005. Infracommunitylevel aggregation in the monogenean communities of crucian carp(Carassius carassius). Parasitology 131, 367–372.

akke, T.A., Cable, J., Harris, P.D., 2007. The biology of gyrodactylid mono-geneans: the “Russian-doll killers”. Adv. Parasitol. 64, 161–376.

akke, T.A., Harris, P.D., Cable, J., 2002. Host specificity dynamics: obser-vations on gyrodactylid monogeneans. Int. J. Parasitol. 32, 281–308.

akke, T.A., Harris, P.D., Jansen, P.A., Hansen, L.P., 1992. Host specificityand dispersal strategy in gyrodactylid monogeneans, with particularreference to Gyrodactylus salaris (Platyhelminthes, Monogenea). Dis.Aquat. Org. 13, 63–74.

akke, T.A., Jansen, P.A., Harris, P.D., 1996. Differences in susceptibilityof anadromous and resident stocks of Arctic charr to infections ofGyrodactylus salaris under experimental conditions. J. Fish Biol. 49,341–351.

akke, T.A., Jansen, P.A., Kennedy, C.R., 1991. The host specificity ofGyrodactylus salaris Malmberg (Platyhelminthes, Monogenea): sus-ceptibility of Oncorhynchus mykiss (Walbaum) under experimentalconditions. J. Fish Biol. 39, 45–57.

oeger, W.A., Kritsky, D.C., Pie, M.R., Engers, K.B., 2005. Mode of trans-mission, host switching, and escape from the red queen by viviparousgyrodactylids (Monogenoidea). J. Parasitol. 91, 1000–1007.

uchmann, K., 1998. Binding and lethal effect of complement fromOncorhynchus mykiss on Gyrodactylus derjavini (Platyhelminthes:Monogenea). Dis. Aquat. Org. 32, 195–200.

uchmann, K., Bresciani, J., 1998. Microenvironment of Gyrodactylus der-javini on rainbow trout Oncorhynchus mykiss: association betweenmucous cell density in skin and site selection. Parasitol. Res. 84, 17–24.

uchmann, K., Bresciani, J., 2006. Monogenea (Phylum Platyhelminthes).In: Woo, P.T.K. (Ed.), Fish Diseases and Disorders, Volume 1: Proto-zoan and Metazoan Infections. CAB International, Wallingford, pp.297–344.

uchmann, K., Lindenstrøm, T., Bresciani, J., 2004. Interactive associa-tions between fish hosts and monogeneans. In: Wiegertjes, G.F., Flik,G. (Eds.), Host–Parasite Interactions. Garland Science/BIOS ScientificPublishers, Oxford, pp. 161–184.

ush, A.O., Lafferty, K.D., Lotz, J.M., Shostak, A.W., 1997. Parasitology meetsecology on its own terms: Margolis et al. revisited. J. Parasitol. 83,575–583.

able, J., Scott, E.C.G., Tinsley, R.C., Harris, P.D., 2002. Behavior favoringtransmission in the viviparous monogenean Gyrodactylus turnbulli. J.Parasitol. 88, 183–184.

able, J., van Oosterhout, C., 2007. The role of innate and acquiredresistance in two natural populations of guppies (Poecilia reticulata)infected with the ectoparasite Gyrodactylus turnbulli. Biol. J. Linn. Soc.90, 647–655.

ollins, C.A., Olstad, K., Sterud, E., Jones, C.S., Noble, L.R., Mo, T.A., Cun-ningham, C.O., 2007. Isolation of a novel fish thymidylate kinasegene, upregulated in Atlantic salmon (Salmo salar L.) following infec-tion with the monogenean parasite Gyrodactylus salaris. Fish ShellfishImmunol. 23, 793–807.

ostello, M.J., 2009. How sea lice from salmon farms may cause wildsalmonid declines in Europe and North America and be a threat tofishes elsewhere. Proc. R. Soc. B 276, 3385–3394.

ávidová, M., Jarkovsky, J., Matejusová, I., Gelnar, M., 2005. Seasonaloccurrence and metrical variability of Gyrodactylus rhodei Zitnan 1964(Monogenea, Gyrodactylidae). Parasitol. Res. 95, 398–405.

ickerson, H.W., 2006. Ichthyophthirius multifiliis and Cryptocaryon irritans(Phylum Ciliophora). In: Woo, P.T.K. (Ed.), Fish Diseases and Disorders,

sitology 183 (2012) 305– 316 315

Volume 1: Protozoan and Metazoan Infections. CAB International,Wallingford, pp. 116–153.

Euzet, L., Combes, C., 1998. The selection of habitats among the Monoge-nea. Int. J. Parasitol. 28, 1645–1652.

F.A.O., 2008. In: Department, F.a.A. (Ed.), FishStat Plus. Food and Agricul-ture Organization of the United Nations, Rome.

Froese, R., Pauly, D., 2010. In: Froese, R., Pauly, D. (Eds.), FishBase. WorldWide Web electronic publication, www.fishbase.org.

García-Vásquez, A., Hansen, H., Christison, K.W., Rubio-Godoy, M., Bron,J.E., Shinn, A.P., 2010. Gyrodactylids (Gyrodactylidae, Monogenea)infecting Oreochromis niloticus niloticus (L.) and O. mossambicus(Peters) (Cichlidae): a pan-global survey. Acta Parasitol. 55, 215–229.

García-Vásquez, A., Hansen, H., Shinn, A.P., 2007. A revised descriptionof Gyrodactylus cichlidarum Paperna, 1968 (Gyrodactylidae) from theNile tilapia, Oreochromis niloticus niloticus (Cichlidae), and its syn-onymy with G. niloticus Cone, Arthur et Bondad-Reantaso, 1995. FoliaParasitol. 54, 129–140.

Garduno-Lugo, M., Munoz-Córdova, G., Olvera-Novoa, M.A., 2004. Massselection for red color in Oreochromis niloticus (Linnaeus, 1758).Aquacult. Res. 35, 340–344.

Harris, P.D., 1988. Changes in the site specificity of Gyrodactylus turn-bulli Harris, 1986 (Monogenea) during individual infections of guppies(Poecilia reticulata Peters). Can. J. Zool. 66, 2854–2857.

Harris, P.D., Shinn, A.P., Cable, J., Bakke, T.A., 2004. Nominal species of thegenus Gyrodactylus von Nordmann 1832 (Monogenea: Gyrodactyli-dae), with a list of principal host species. Syst. Parasitol. 59, 1–27.

Harris, P.D., Soleng, A., Bakke, T.A., 1998. Killing of Gyrodactylus salaris(Platyhelminthes, Monogenea) mediated by host complement. Para-sitology 117, 137–143.

Hiott, A.E., Phelps, R.P., 1993. Effects of initial age and size on sex reversal ofOreochromis niloticus fry using methyltestosterone. Aquaculture 112,301–308.

Jansen, P.A., Bakke, T.A., 1991. Temperature-dependent reproduction andsurvival of Gyrodactylus salaris Malmberg, 1957 (Platyhelminthes:Monogenea) on Atlantic salmon (Salmo salar L.). Parasitology 102,105–112.

Jensen, A.J., Johnsen, B.O., 1992. Site specificity of Gyrodactylus salarisMalmberg, 1957 (Monogenea) on Atlantic salmon (Salmo salar L.) inthe river Lakselva, northern Norway. Can. J. Zool. 70, 264–267.

King, T.A., Harris, P.D., Cable, J., 2008. Long-term Gyrodactylus lomi infec-tion on isolated juvenile chub, Leuciscus cephalus. J. Parasitol. 94,1426–1427.

Kirby, J.M., 1981. Seasonal occurrence of the ectoparasite Gyrodactylusatratuli on spotfin shiners. Trans. Am. Fish Soc. 110, 462–464.

Lemly, A.D., Esch, G.W., 1984. Effects of the trematode Uvulifer ambloplitison juvenile bluegill sunfish, Lepomis macrochirus: ecological implica-tions. J. Parasitol. 70, 475–492.

Lester, R.J., Adams, J.R., 1974. Gyrodactylus alexanderi: reproduction, mor-tality, and effect on its host Gasterosteus aculeatus. Can. J. Zool. 52,827–833.

Lindenstrøm, T., Secombes, C.J., Buchmann, K., 2004. Expression ofimmune response genes in rainbow trout skin induced by Gyrodactylusderjavini infections. Vet. Immunol. Immunopathol. 97, 137–148.

Lo, C.M., Morand, S., 2000. Spatial distribution and coexistence of mono-genean gill parasites inhabiting two damselfishes from Moorea islandin French Polynesia. J. Helminthol. 74, 329–336.

Matejusová, I., Felix, B., Sorsa-Leslie, T., Gilbey, J., Noble, L.R., Jones, C.S.,Cunningham, C.O., 2006. Gene expression profiles of some immunerelevant genes from skin of susceptible and responding Atlanticsalmon (Salmo salar L.) infected with Gyrodactylus salaris (Monogenea)revealed by suppressive subtractive hybridisation. Int. J. Parasitol. 36,1175–1183.

Mo, T.A., 1997. Seasonal occurrence of Gyrodactylus derjavini (Monoge-nea) on brown trout, Salmo trutta, and Atlantic salmon, S. salar, in theSandvikselva river, Norway. J. Parasitol. 83, 1025–1029.

Morand, S., Poulin, R., Rohde, K., Hayward, C., 1999. Aggregation andspecies coexistence of ectoparasites of marine fishes. Int. J. Parasitol.29, 663–672.

Munoz-Córdova, G., Garduno-Lugo, M., 2003. Mejoramiento genético entilapia: sistemas de cruzamiento y mecanismos genéticos en la deter-minación de color (Veracruz, Sistema de Investigación del Golfo deMéxico del Consejo Nacional de Ciencia y Tecnología). p. 84.

Paperna, I., 1996. Parasites, infections and diseases of fishes in Africa. Anupdate. CIFA Technical Paper No. 31. FAO, Rome, 220 pp.

Phelps, R.P., Cole, W., Katz, T., 1992. Effect of fluoxymesterone on sex ratioand growth of Nile tilapia, Oreochromis niloticus (L.). Aquacult. Res. 23,405–410.

Pie, M.R., Engers, K.B., Boeger, W.A., 2006. Density-dependent topo-graphical specialization in Gyrodactylus anisopharynx (Monogenoidea,

ary Para

316 M. Rubio-Godoy et al. / Veterin

Gyrodactylidae): boosting transmission or evading competition? J.Parasitol. 92, 459–463.

R Development Core Team, 2007. R: A Language and Environment forStatistical Computing. R Foundation for Statistical Computing, Vienna.

Richards, G.R., Chubb, J.C., 1996. Host responses to initial and challengeinfections, following treatment of Gyrodactylus bullatarudis and G.turnbulli (Monogenea) on the guppy (Poecilia reticulata). Parasitol. Res.82, 242–247.

Richards, G.R., Chubb, J.C., 1998. Longer-term population dynamics ofGyrodactylus bullatarudis and G. turnbulli (Monogenea) on adult gup-pies (Poecilia reticulata) in 50-I experimental arenas. Parasitol. Res. 84,753–756.

Rózsa, L., Reiczigel, J., Majoros, G., 2000. Quantifying parasites in samplesof hosts. J. Parasitol. 86, 228–232.

Rubio-Godoy, M., 2007. Fish host–monogenean parasite interactions,with special reference to Polyopisthocotylea. In: Terrazas, L.I. (Ed.),Advances in the Immunobiology of Parasitic Diseases. Research Sign-post, Trivandrum, pp. 91–109.

Rubio-Godoy, M., 2008. Microhabitat selection of Discocotyle sagittata(Monogenea) in farmed rainbow trout. Folia Parasitol. 55, 270–276.

Rubio-Godoy, M., 2010. Inmunología de los peces óseos. Revisión/Teleostfish immunology. Review. Rev. Mex. Tec. Pecu. 1, 43–57.

Scott, M.E., 1985. Dynamics of challenge infections of Gyrodactylus bul-latarudis Turnbull (Monogenea) on guppies, Poecilia reticulata (Peters).J. Fish Dis. 8, 495–503.

sitology 183 (2012) 305– 316

Scott, M.E., Anderson, R.M., 1984. The population dynamics of Gyrodactylusbullatarudis (Monogenea) within laboratory populations of the fishhost Poecilia reticulata. Parasitology 89, 159–194.

Scott, M.E., Nokes, D.J., 1984. Temperature-dependent reproduction andsurvival of Gyrodactylus bullatarudis (Monogenea) on guppies (Poeciliareticulata). Parasitology 89, 221–227.

Simková, A., Gelnar, M., Sasal, P., 2001. Aggregation of congeneric parasites(Monogenea: Dactylogyrus) among gill microhabitats within one hostspecies (Rutilus rutilus L.). Parasitology 123, 599–607.

Soleng, A., Jansen, P.A., Bakke, T.A., 1999. Transmission of the monogeneanGyrodactylus salaris. Folia Parasitol. 46, 179–184.

Tatner, M.F., 1996. Natural changes in the immune system of fish. In:Iwama, G., Nakanishi, T. (Eds.), The Fish Immune System: Organ-ism, Pathogen and Environment. Academic Press Ltd., London, pp.255–287.

van Oosterhout, C., Potter, R., Wright, H., Cable, J., 2008. Gyro-scope: anindividual-based computer model to forecast gyrodactylid infectionson fish hosts. Int. J. Parasitol. 38, 541–548.

Winger, A.C., Kanck, M., Kristoffersen, R., Knudsen, R., 2008. Seasonaldynamics and persistence of Gyrodactylus salaris in two riverine

anadromous Arctic charr populations. Environ. Biol. Fish 83, 117–123.

Yang, T., Liu, J., Gibson, D.I., Dong, A., 2006. Spatial distributions of twospecies of monogeneans on the gills of Siganus fuscenscens (Hout-tuyn) and their seasonal dynamics in caged versus wild-caught hosts.J. Parasitol. 92, 933–940.