DISEASES OF AQUATIC ORGANISMSDis Aquat Org

Vol. 59: 141150, 2004 Published May 5

INTRODUCTION

The plerocercoid larvae of the pseudophyllideancestode Schistocephalus solidus are frequent parasitesin the body cavity of the 3-spined stickleback through-out its geographical range (Arme & Owen 1967). Theinfection is under investigation as a model of host

parasite interaction in studies on behaviour (Milinski1984, 1985, 1990, verli et al. 2001), reproduction(Meakins 1974, McPhail & Peacock 1983, Candolin &Voigt 2001) and evolutionary ecology (Barber & Ruxton1998, Aeschlimann et al. 2000, Arnott et al. 2000, Binzet al. 2000, Reimchen & Nosil 2001, Christen & Milinski2003, Kurtz et al. 2004). Yet little is known about

Inter-Research 2004 www.int-res.com*Corresponding author. Email: kalbe@mpil-ploen.mpg.de

Modulation of granulocyte responses in three-spined sticklebacks Gasterosteus aculeatus infected

with the tapeworm Schistocephalus solidus

J. P. Scharsack1, 2, M. Kalbe1,*, R. Derner1, J. Kurtz1, M. Milinski1

1Max Planck Institute of Limnology, Department of Evolutionary Ecology, August-Thienemann-Strasse 2, 24306 Pln, Germany

2Present address: Biology and Immunology Group, Wageningen University, PO Box 338, 6700 AH Wageningen, The Netherlands

ABSTRACT: Leukocytes isolated from the head kidney and peripheral blood of 3-spined sticklebacksGasterosteus aculeatus L. were analysed by means of flow cytometry during infection with the tape-worm Schistocephalus solidus (Mller, 1776). Although parasites increased their body weight contin-uously throughout the observation period (98 d), proportions of granulocytes increased in blood andhead kidney only up to Day 63 post-infection (p.i.). Thereafter, declining proportions of granulocyteswere observed in both organs. Thus the relative decrease in granulocyte number was not correlatedto a decline in the parasitic load of the fish. To investigate a possible modulatory impact of S. soliduson granulocyte function, head kidney leukocytes were isolated at times before Day 63 p.i. and testedin vitro for their capacity to produce reactive oxygen species (ROS). Head kidney leukocytes from S.solidus-infected fish, analysed immediately after isolation (ex vivo, Day 40 p.i.), exhibited a higherROS production when stimulated with phorbol myristate acetate (PMA), than leukocytes from nave,sham-treated control fish and fish that had resisted or cleared the infection (exposed but notinfected). The latter showed an increased spontaneous ROS production that was not correlated to thenumbers of granulocytes present in the head kidney isolates. In infected sticklebacks, spontaneousand PMA-induced ROS production was significantly correlated with numbers of granulocytes pres-ent in the head kidney isolates, suggesting that elevated ROS production was due to higher numbersof responding cells rather than an increased capacity of single cells. In vitro, after cultivation for 4 din the presence of pokeweed mitogen (PWM) or extracts from S. solidus, head kidney leukocytes fromcontrol fish showed an increased ROS production and phagocytic activity compared with non-stimulated control cultures. In contrast, head kidney leukocytes from infected fish isolated on Days 48and 44 p.i., failed to respond to S. solidus antigens in vitro. During S. solidus infection, granulocytemobilisation resulted in elevated numbers of these cells in head kidneys, but the lack of an in vitroresponse to S. solidus antigens indicates an in vivo priming of granulocytes by the parasite. Theseobservations may reflect the ability of S. solidus to impair the hosts immune response once the para-site is developing in the body cavity of G. aculeatus.

KEY WORDS: Gasterosteus aculeatus Schistocephalus solidus Granulocytes Reactive oxygenspecies Phagocytosis

Resale or republication not permitted without written consent of the publisher

Dis Aquat Org 59: 141150, 2004

immune mechanisms that may play a role in S. solidusinfection of the stickleback.

Mobilisation and activation of granulocytes has beenconsidered a significant part of the immune responseof cyprinids (Hoole & Arme 1983, Taylor & Hoole 1993,1995, Nie & Hoole 2000) and rainbow trout (Sharp et al.1992) to helminth parasites. Increased production ofoxygen intermediates like nitric oxide (NO) and reac-tive oxygen species (ROS) occur (Whyte et al. 1990,Secombes & Chappell 1996), indicating the functionalimportance of granulocyte and macrophage responsesto helminths. Indeed, in vitro larvacidal activity ofmacrophages from rainbow trout to diplostomules ofthe eye fluke Diplostomum spathaceum has beendemonstrated (Whyte et al. 1989). It is not yet clearwhether large helminth parasites of fishes are killed bycellular responses in vivo (Secombes & Chappell 1996).The fate of Schistocephalus solidus plerocercoids infish species other than the natural host, the 3-spinedstickleback Gasterosteus aculeatus, do however indi-cate that the plerocercoids can be harmed by fishimmune defence. Growth of plerocercoids was muchslower in 9-spined sticklebacks Pungitius pungitius, inwhich degenerative changes in the plerocercoidstegument were detected (Orr et al. 1969). Heterotrans-plants of S. solidus from G. aculeatus to other species offishes (Cottus gobio, Nemacheilus barbatula, Phoxinusphoxinus, Salmo trutta, Coregonus clupeoides, Percafluviatilis, Rutilus rutilus, Esox lucius), including P.pungitius, died within 2 to 10 d of transfer, while homo-transplants between G. aculeatus survived (Brten1966). Whether the parasites were actively destroyedby the immune system was not analysed. Howeverthese observations indicate that S. solidus plerocer-coids may be able to suppress an effective immuneresponse in G. aculeatus, but not in other fish species.

Information on the cellular immune response ofstickleback against Schistocephalus solidus is still lim-ited. Although encystment of S. solidus plerocercoidsin the body cavity of infected sticklebacks was notobserved (Arme & Owen 1967), total white blood cellcounts increased during S. solidus infection (Barber etal. 2001). Thus leukocytes are mobilised during S.solidus infection of stickleback, but the parasite mayhave evolved strategies of immuno-suppression orimmuno-avoidance.

The present study investigated cellular responses ofsticklebacks infected with Schistocephalus solidus.Leukocytes isolated from the peripheral blood and thehead kidney of stickleback were analysed by means offlow cytometry for the frequency of leukocyte subsetsand production of reactive oxygen species (ROS), andphagocytosis activity of head kidney leukocytes wereanalysed following in vitro stimulation with S. solidusextracts.

MATERIALS AND METHODS

Propagation of Gasterosteus aculeatus and Schisto-cephalus solidus. Both fish and tapeworms were labo-ratory-raised offspring of individuals originating fromthe Neustdter Binnenwasser (northern Germany), abrackish lagoon of the Baltic Sea. Sticklebacks wereraised and kept in groups of 10 to 20 fish in 20 l tankswith a continuous supply of aerated tap water at 18Cand a day:night rhythm of 16:8 h. During the first fewweeks after hatching, fish were fed daily with Artemiasp. naupliae, and later 3 to 5 times per week withfrozen mosquito larvae and occasionally frozen clado-cerans ad libitum. Mature S. solidus plerocercoidswere removed from the fish body-cavity asepticallyand allowed to reproduce in an in vitro system modi-fied from Smyth (1954), as described in Wedekind et al.(1998). Eggs released by the worms were washed andstored in tap water at 18C in the dark. After 20 d incu-bation, hatching of coracidia was stimulated by light(Dubinina 1980).

Infection experiments. As first intermediate hosts,laboratory-bred copepods Macrocyclops albidus wereeach individually incubated with 6 Schistocephalussolidus coracidia, and screened for intensity of infec-tion after 2 wk. Subsequently, individual sticklebackswere exposed to 3 copepods containing together a totalof 4 procercoids in tanks with 500 ml of water for 48 h.Ingestion of the intermediate hosts was confirmed bysubsequent filtration of the water and screening forremaining copepods. To follow up the development ofS. solidus plerocercoids associated with expectedchanges in frequencies of leukocyte subpopulations,reflecting an immune response, 105 sticklebacks wereexposed to infected copepods and kept in groups of 15in separate tanks. All fish from individual tanks weresacrificed and sampled at set time points (see Fig. 2).Infection was monitored by measuring the freshweight of S. solidus plerocercoids. Head kidney andperipheral blood leukocytes were isolated frominfected fish. From the same stock, 20 fish that werenot exposed to copepods were analysed as a Day 0control. For in vitro analysis of leukocytes from S.solidus-infected sticklebacks, 36 sticklebacks wereexposed to infected copepods and 18 fish wereexposed to non-infected copepods (sham-exposed).Exposed fish were kept in groups of 6 in separate tanksuntil sampling. Head kidney leukocytes were collectedat time points before the peak of granulocyte frequen-cies (see Fig. 2) from 12 fish exposed to infected cope-pods and 6 sham-exposed fish at 40, 44 and 48 d.

Preparation of Schistocephalus solidus extract. Ple-rocercoids of S. solidus, with an individual weight of 15to 200 mg, were collected from laboratory-infectedsticklebacks. Plerocercoids were washed with phos-

142

Scharsack et al.: Granulocyte responses in tapeworm-infected sticklebacks

phate-buffered saline (PBS) and frozen at 20C in PBS(1 g wet weight ml1 PBS, pH 7.4). After thawing, theworms were chopped up with a scalpel and sonicatedfor 60 s on ice. The resulting suspension was cen-trifuged at 20 800 g and 4C for 15 min. Thesupernatant was removed and 0.45 m-filtered. Subse-quently the protein content was determined colori-metrically using bovine serum albumin as a standard,and aliquots were stored at 20C.

Culture media. Culture media for cell separationand cultivation and phosphate-buffered saline (PBS)were diluted with 10% (v/v) distilled water to adjusttheir osmotic pressure according to stickleback serumosmolarity. For blood collection, diluted Leibovitz 15(L-15) medium (Sigma Aldrich) with 50 000 IU l1

sodium heparin (Sigma-Aldrich) was used (hepar-inised medium). Washing procedures were conductedwith diluted L-15 (wash medium). For cultivationexperiments diluted L-15 was supplemented with100 000 IU l1 penicillin, 100 mg l1 streptomycin,4 mmol l1 L-glutamine, 7% (v/v) foetal bovine serum(FBS; all chemicals: Sigma Aldrich) and 3% (v/v) carpserum (culture medium). Carp serum was used as asource of piscine serum because of the limitationsinherent in the preparation of sufficient amounts ofstickleback serum. Carp serum was a pool of sera from10 individual fish. It was heat-inactivated for 30 min at56C, 0.2 m-filtered and stored at 22C until use.

Leukocyte isolation. Media and cells were kept onice, and washing procedures were performed at 4C.Fish were anaesthetised by a blow on the head andkilled by incision of the brain. Blood was collected intowells of 24-well plates prefilled with 0.5 mlheparinised medium, after cutting off the tail of thefish. Peripheral blood leukocytes (PBL) were separatedfrom erythrocytes by centrifugation (30 min, 750 g)over Lymphoprep (1.077 g ml1, Nycomed) discontinu-ous density-gradient, as described by Miller & McKin-ney (1994). Cell suspensions from head kidneys wereprepared by forcing the tissues through a 40 m nylonscreen (BD-Falcon, USA). Isolated PBL and head kid-ney leukocytes (HKL) were washed once with washmedium (10 min, 550 g) and once with culturemedium, and resuspended in cell-culture medium.Numbers of viable cells (exclusion of propidium-iodide-positive cells) were enumerated by means offlow cytometry.

Flow cytometry analysis. Suspensions of freshly iso-lated and of cultured PBL or HKL were analysed byflow cytometry (PAS III, Partec, Germany). All sampleswere supplemented with propidium iodide (2 mg l1,Sigma Aldrich) to detect dead cells, and tubes werekept on ice until measurement. Forward- and side-scatter values (FSC/SSC characteristics) of at least10 000 events were acquired in linear mode; fluores-

cence intensities at wavelengths of 530 nm and 650 nmwere acquired at log-scale. All flow-cytometry datawere analysed with the software WinMDI, Version 2.8(Trotter 1998: http://facs.scripps.edu/software.html).Cellular debris with low FSC characteristics and deadcells (propidium-iodide-positive) were excluded fromfurther evaluation. Different cellular subsets wereidentified according to their characteristic FSC/SSCprofiles. For the assessment of phagocytosis activityafter in vitro stimulation, plates with cultured cellswere placed on ice (15 min) and briefly shaken, andthe whole content of each well was transferred to indi-vidual flow-cytometer tubes. Phagocytic active cellswere green-fluorescence-positive after ingestion offluorescent latex beads. For the adjustment of cellnumbers for in vitro assays, total cell counts in suspen-sions of freshly isolated cells were determined with thestandard cell dilution assay (SCDA, Pechhold et al.1994) in a modified form: 2 105 green fluorescentstandard particles (4 m, Polyscience, USA) wereadded to each tube. Standard particles (green-fluorescence-positive) could be easily discriminatedfrom viable cells (propidium-iodide-negative, green-fluorescence-negative). Absolute numbers of cells inindividual samples were calculated according to N(vital cells) = events (vital cells) number (standardbeads)/events (standard beads).

Leukocyte cultivation. For cell-culture experiments,HKL were incubated in 96-well flat-bottomedmicrotitre plates (2 105 cells well1 in a final volume of175 l). All setups were made in duplicate at least.Mitogen stimulation of cultured leukocytes was in-duced with PWM (pokeweed mitogen, 2 mg l1). Para-site extracts were added at final concentrations of 10and 100 mg l1 (protein fraction). The cultures wereincubated for 4 d at 18C in a water-vapour-saturatedatmosphere with 2% CO2.

Production of reactive oxygen species by head kid-ney leukocytes. Generation of reactive oxygen species(ROS) by HKL was measured by the nitro blue tetra-zolium salt (NBT, Sigma-Aldrich) reduction assay. Afirst set of HKL was collected at Day 40 p.i. and testedimmediately after isolation to assess the ex vivo ROSproduction. Therefore 4 105 cells per well of 96-wellflat-bottomed microtitre plates were incubated in175 l culture medium containing the reactive re-agents. A second set of HKL was isolated at Day 48 p.i.and tested for ROS production after in vitro stimulationwith Schistocephalus solidus antigens, as describedabove. Culture supernatants were removed from eachwell and replaced by 175 l HKL medium containingthe respective reactive components. Both HKL setupswere tested following the same protocol for the NBT-reduction assay. Receptor-independent ROS produc-tion was induced by adding 0.15 mg l1 phorbol myris-

143

Dis Aquat Org 59: 141150, 2004

tate acetate (PMA, Sigma-Aldrich). The indicator NBTwas added at 1 g l1, and wells without PMA served todetermine spontaneous ROS generation. After incuba-tion for 2 h at 22C, the supernatants were removedand the cells were fixed by adding 125 l of 100%methanol. Each well was washed 2 times with 125 l of70% (v/v) methanol and the fixed cells were air-driedfor 20 h. The reduced NBT (formazan) was dissolved in125 l 2 mol l1 KOH and 125 l DMSO per well. Opti-cal densities were recorded with a spectrophotometerat 650 nm.

Phagocytic activity. Phagocytic activity of culturedHKL was tested as described by Chilmonczyk &Monge (1999). HKL isolated at Day 44 p.i. were cul-tured as described above for 3 d in vitro and co-incubated with 2.5 106 green fluorescent latex parti-cles (1 m, Polyscience, USA) for an additional 18 h.Harvested cells were analysed by flow cytometry for

the presence and fraction of green fluorescent cells(phagocytosis-positive).

Statistics. To determine the significance of differ-ences between independent groups, data were com-pared by ANOVA, and by ANOVA-on-ranks if theequal variance test failed. Results were confirmedusing Students t-test, and a Mann-Whitney rank-sumtest, if the normality test failed. Differences betweendependent groups were tested with a paired t-test.Correlation between data sets was tested by regres-sion analysis. All statistical tests were done using Jan-del, Sigma Stat 2.0 software.

RESULTS

Lymphocyte and granulocyte frequencies in bloodand head kidney of sticklebacks

In isolates of stickleback peripheral blood and headkidney leukocytes, 2 distinct cell populations wereidentified by means of flow cytometry according totheir forward-/side scatter characteristics (FSC/SSCprofiles; Fig. 1). A population of small cells with low-complexity values (FSC/SSC low) was characterised aslymphocytes by parallel microscopic analysis (R1,Fig. 1). The majority of large cells with high complex-ity (FSC/SSC high), microscopically showed character-istics of granulocytes, such as increased numbers ofgranules and polymorphic nuclei, and were gated inRegion 2 (R2, Fig. 1).

To follow potential changes in frequencies of lym-phocytes (R1-cells) and granulocytes (R2-cells) duringSchistocephalus solidus infection, leukocytes from thehead kidney and blood were analysed by flow cytome-try. During the course of infection, increasing propor-tions of granulocytes and decreasing proportions oflymphocytes were initially observed in the head kid-ney and blood (Fig. 2a,b). Interestingly, this progres-sion was reversed after 63 d p.i. and the proportion ofgranulocytes started to decrease. Development of S.solidus plerocercoids did not appear to be affected bythe mobilisation of granulocytes, as parasites weregrowing continuously throughout the observationperiod up to 98 d p.i. (Fig. 2c).

Respiratory burst-activity of freshly isolated headkidney leukocytes

To analyze whether modulation by Schistocephalussolidus plerocercoids could be the cause for the drop ingranulocyte mobilisation, a second infection experi-ment was conducted. Head kidney leukocytes werefunctionally analysed at time points before the peak

144

R1

R2

a)

40950

040

95

Cell size (FSC)

R1

R2

b)

40950

040

95

Cell size (FSC)

Cel

l com

plex

ity (

SS

C)

Fig. 1. Density plots showing flow-cytometry characteristics ofstickleback leukocytes. In leukocytes isolated from (a) theperipheral blood and (b) the head kidney, 2 distinct cell pop-ulations were detected according to their FSC/SSC (forward-/side-scatter) profiles. Cells gated in R1 were small lympho-

cytes, cells gated in R2 were granulocytes

Scharsack et al.: Granulocyte responses in tapeworm infected sticklebacks

granulocyte activity seen in the first infection experi-ment. When analysed with the NBT-reduction assayimmediately after isolation at Day 40 p.i., head kidneyleukocytes from infected and from sticklebacks thathad cleared the infection (exposed but not infected)exhibited a significantly higher production of reactiveoxygen species (ROS) when stimulated with phorbolmyristate acetate (PMA) than sham-exposed controlfish (exposed to non-infected copepods; Fig. 3).Exposed but non-infected sticklebacks showed an

increased spontaneous (not PMA-induced) ROS pro-duction (Fig. 3) compared with infected and sham-exposed fish.

In infected fish, but not in sham-exposed fish, andexposed but not infected fish, proportions of granulo-cytes (R2 cells) present in head kidney isolates weresignificantly correlated to the amount of ROS producedin PMA (r = 0.819; p < 0.05)-stimulated cultures andcultures without PMA (r = 0.794; p < 0.05). This indi-cates that higher numbers of granulocytes in head kid-ney isolates from infected fish, rather than elevatedcapacities for ROS production of single cells, causedthe increased PMA response.

Respiratory burst-activity of head kidney leukocytesafter in vitro stimulation

For further characterisation of the influence of Schis-tocephalus solidus on head kidney leukocytes, cellswere stimulated in culture with S. solidus antigens andsubsequently tested for production of ROS. Head kid-ney leukocytes were isolated 48 d p.i. and incubated inmedium alone, with poke weed mitogen (PWM) andwith S. solidus extracts, and tested in the NBT-

145

20 11 13 12 11 11 11 10

**(* (* (*(* (*

(*(*

* ** *

**

**

*

*

No. of fish testedLymphocytes Granulocytes

a) Head kidney leukocytes

%

0

20

40

60

80

100

0 28 42 56 63 70 77 98

Infection time [days]

b) Peripheral blood leukocytes

%

0

20

40

60

80

100

c) S. solidus wet weight

0

50

100

150

200

250

mg

Fig. 2. Gasterosteus aculeatus infected with Schistocephalussolidus. Changes in leukocyte subpopulations in sticklebacks.(a) Head kidney leukocytes and (b) peripheral blood leuko-cytes were analysed by flow cytometry during (c) course ofinfection. Proportions of lymphocytes and granulocytes weredetermined according to the regions shown in Fig. 1. Day 0measurement represents non-infected fish (n = 20); at subse-quent time points, data are mean and SD of indicated num-ber of infected fish out of 15 fish exposed to the parasite. Notedecreasing proportions of granulocytes (R2 cells) after 63 dp.i. although S. solidus were still present and growing. Dataon parasite growth (c) were collected by Hberli et al. (2002).": Significantly different from Day 0; p < 0.05; (": significantly

different from Day 63; p < 0.05

ROS production

sham-exposed

*

*

**

exposednot infected

exposedinfected

OD

0.0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

PMA 0 mg/l PMA 0.15mg/l

* significantly different from sham-exposed fish (P < 0.05)** significantly different from sham-exposed and infected fish (P < 0.05)

Fig. 3. Respiratory-burst activity of freshly isolated head kid-ney leukocytes from sticklebacks infected and not infectedwith Schistocephalus solidus. Production of reactive oxygenspecies (ROS) was measured immediately after cell isolation(40 d p.i.). Optical density (OD) values are mean SD oftriplicate cultures from control (n = 6), exposed but non-infected (n = 5) and infected (n = 7) fish. Note that headkidney leukocytes from S. solidus-infected and -exposed butnot infected sticklebacks exhibited a significantly higherphorbol myristate acetate (PMA) response than sham-exposed controls. In exposed but non-infected fish, spontaneous (not PMA-induced) ROS production was sig-

nificantly increased

Dis Aquat Org 59: 141150, 2004

reduction assay. After 4 d cultivation in medium alone(Ctr; Fig. 4), head kidney leukocytes from sham-exposed and S. solidus-infected sticklebacks exhibiteda significant ROS response to stimulation with phorbolmyristate acetate (PMA). When cultured in mediumwith PWM or S. solidus extract (SSE; Fig. 4), supple-mentation with PMA did not result in an elevated ROSproduction in either setup. Interestingly, leukocytesfrom sham-exposed fish exhibited a significantlyhigher spontaneous (not PMA-induced) ROS produc-tion in cultures stimulated with PWM and extracts of S.solidus than the corresponding control (Fig. 4), indicat-ing an in vitro response to these stimuli (Fig. 4). In con-trast, spontaneous ROS production of head kidneyleukocytes from infected fish in cultures stimulatedwith S. solidus antigen did not exceed the valuesrecorded for control cultures. Thus head kidney leuko-cytes from S. solidus-infected fish in this assay did notrespond to stimulation with S. solidus antigens. Data of

head kidney leukocytes from exposed but not infectedfish (n = 2) did not show significant differences be-tween treatment groups (data not shown).

Phagocytosis activity of stickleback head kidneyleukocytes after in vitro stimulation

Comparable observations were made when headkidney leukocytes (isolated at Day 44 p.i.) wereanalysed for phagocytic activity after 4 d in vitro stim-ulation (Fig. 5). Head kidney granulocytes from sham-exposed fish significantly increased their phagocyticactivity in cultures stimulated with PWM and extractsof S. solidus. Granulocytes from infected fish wereresponding to PWM stimulation, but not to stimulationwith S. solidus extracts with increased phagocyticactivity. Data of head kidney leukocytes from exposedbut not infected fish (n = 2) are not shown, as signifi-cant differences between treatment groups were notfound.

Taken together, Schistocephalus solidus-primedgranulocytes ex vivo showed increased capacities forROS production compared with sham-treated controls(Fig. 3). Cultivation of head kidney leukocytes with S.solidus extract triggered in vitro ROS production and

146

Sham-exposed

P

Scharsack et al.: Granulocyte responses in tapeworm infected sticklebacks

phagocytic activity of cells from sham treated controlfish, but leukocytes from infected fish did not respondto direct contact with S. solidus extracts (Figs. 4 & 5).

DISCUSSION

The present study investigated responses of leuko-cytes from the stickleback to the pseudophyllideancestode Schistocephalus solidus. During the initialphase of infection, proportions of granulocytes inleukocyte isolates from the head kidney and bloodincreased, while the proportions of lymphocytesdecreased (Fig. 2a,b). In another study by Barber et al.(2001), elevated counts of total white blood cells weredetected in S. solidus-infected G. aculeatus; however,their study did not distinguish between lymphocytesand granulocytes. Mobilisation of granulocytes is acommon feature of the immune responses of fishes tohelminth parasites (Taylor & Hoole 1989, Sharp et al.1992, Secombes & Chappell 1996). The initial increasein granulocyte proportions in the blood and head kid-ney of infected stickleback described here could bedue to the mobilisation of granulocytes against S.solidus rather than to changes in lymphocyte frequen-cies. Whether the changes in the lymphocyte andgranulocyte populations are associated with aleukopenia or leukocytosis remains to be verified, astotal leukocyte counts in the blood were not done dur-ing the present study. Production of reactive oxygenspecies (ROS) by head kidney leukocytes from S.solidus-infected stickleback in response to stimulationwith phorbol myristate acetate (PMA) showed anincrease when analysed immediately after isolation(40 d p.i., Fig. 3). This finding is corroborated byanother experiment, in which head kidney leukocytesfrom sticklebacks infected with S. solidus showed ahigher respiratory-burst reaction upon phagocytosisof zymosan particles in vitro (Kurtz et al. 2004). In thepresent study, ROS production was positively cor-related with proportions of granulocytes present inthe head kidney isolates of infected fish, indicatingthat higher numbers of granulocytes in head kidneyisolates rather than elevated capacities for ROS pro-duction of single cells caused increased PMAresponse. Stickleback that had cleared the infection(exposed but not infected) exhibited an increasedspontaneous (not PMA-induced) ROS production(Fig. 3) that was not correlated to the proportions ofgranulocytes in the head kidney isolates, suggestingthat single cells had an elevated capacity for ROS pro-duction. This was not observed in sticklebacks show-ing development of S. solidus plerocercoids, indicat-ing a differential regulation of granulocyte activationin the groups tested, perhaps suppression by the par-

asite. Granulocyte mobilisation and activation obvi-ously did not affect the growth of S. solidus plerocer-coids, which increased their body weight continuouslythroughout the observation period (Fig. 2c). In roachinfected with the cestode Ligula intestinalis, parasitessurvive apparently unharmed, despite the presence ofa host cellular immune-response, as long as the host isalive (Hoole & Arme 1982). In the present study, noneof the infected or sham-exposed fish died during theexperiment. Interestingly, after 63 d p.i., the propor-tions of granulocytes started to decrease in the bloodand head kidney, while lymphocytes were increasing(Fig. 2). This might reflect the ability of S. solidus toimpair the cellular response of its host. To furtherinvestigate this assumption, head kidney leukocytesfrom infected and sham-treated stickleback were iso-lated before peak granulocyte mobilisation and func-tionally tested in vitro.

When analysing the ROS production of head kidneyleukocytes from infected and sham-exposed fish aftercultivation in medium alone (Ctr, Fig. 4), leukocytesfrom both groups responded to receptor-independentstimulation with PMA, indicating that responsivenessof cells from both groups was still present after theincubation period. In corresponding cultures incu-bated with PWM and Schistocephalus solidus extract,a further increase in ROS production was not inducedby the addition of PMA. This might be due to the factthat, after incubation with PWM and S. solidus extract,leukocytes were already triggered for maximum ROSproduction and were no longer responsive to receptor-independent PMA stimulation. However, head kidneyleukocytes from sham-exposed fish showed a signifi-cantly higher spontaneous ROS production followingin vitro stimulation with PWM and S. solidus extractsthan medium controls (Fig. 4). In contrast, head kidneyleukocytes from S. solidus-infected stickleback did notshow elevated ROS production in the presence of S.solidus extracts. Comparable observations were madewhen analysing the phagocytic activity of head kidneyleukocytes from infected and sham-treated stickle-back. While leukocytes from sham-exposed fishshowed increased phagocytosis after stimulation withPWM and S. solidus extracts, leukocytes from infectedfish responded to PWM but not to S. solidus extract. Asleukocytes from infected fish showed a significantresponse to PWM, the lack of response to S. solidusextracts was not due to a general anergy of these cells.Nie & Hoole (2000) used a polarisation assay to testgranulocytes responses to Bothriocephalus acheilo-gnathi. In unstimulated controls at 10C, 98 to 100% ofgranulocytes remained spherical. Addition of parasiteextracts and increase in temperature resulted in theformation of lamellipodia at the anterior cell-pole andsmall tails at the posterior cell-pole. Head kidney gran-

147

Dis Aquat Org 59: 141150, 2004

ulocytes from carp, naturally and laboratory-infectedwith the cestode B. acheilognathi, exhibited a reducedpolarisation response to extracts of the parasite com-pared with granulocytes from nave carp (Nie & Hoole2000). This suppression of polarisation was reduced byaddition of carp serum to the in vitro system (Nie &Hoole 2000). Thus, also in the system used by Nie &Hoole (2000), a general anergy of leukocytes frominfected carp can be excluded. In both systems, cyto-toxic effects induced by the parasite extracts areunlikely, as leukocytes from nave fish clearly re-sponded.

The reduced in vitro responsiveness of leukocytesfrom infected Gasterosteus aculeatus seems to be partof a complex system. A specific in vivo priming ofleukocytes by Schistocephalus solidus plerocercoidsseems to be a prerequisite for the in vitro non-responsiveness to S. solidus extracts. For mammalsinfected with helminth parasites, it has been suggestedthat granulocyte activity is capable of killing larval, butnot adult, parasite stages (Meusen & Balic 2000). In ourexperiments, the ongoing growth of S. solidus through-out the observation period of 98 d indicates that oncethe plerocercoid is developing in the body cavity of G.aculeatus, the immune response of the host seems nolonger able to repress the parasite completely. In arecent experiment (Kurtz et al. 2004), S. solidus tape-worms reached a relatively smaller size 11 wk p.i. intheir stickleback hosts, which possessed an optimalallelic diversity at their major histocompatibility com-plex (MHC) genes. In stickleback with optimal MHCdiversity, a reduced impact of S. solidus and other par-asites (Wegner et al. 2003a,b) was observed. Duringongoing S. solidus infection, sticklebacks may be moredependent on mechanisms of acquired immunity anddown-regulate functions of the innate immune system.However, mechanisms of immuno-suppression andimmuno-avoidance of helminth parasites in fishes arebarely understood (Secombes & Chappell 1996). Hoole& Arme (1983) suggested that in gudgeon Gobio gobioinfected with Ligula intestinalis, protective host pro-teins are absorbed onto the surface of the parasite. Inmammals infected with helminth parasites, pseudo-cytokines produced by the parasites are suggestedto play a role in immune evasion and exploitation(reviewed by Damian 1997). In fishes, factors producedby helminths can influence the immune response. Forexample, the proliferative response of splenic lympho-cytes from roach to mitogens was suppressed in thepresence of L. intestinalis extracts (Taylor & Hoole1994), and inhibition of temperature-induced polarisa-tion of roach leukocytes occurred with L. intestinalis-derived extracts (Taylor & Hoole 1993).

Brten (1966) observed an ongoing growth of Schis-tocephalus solidus plerocercoid homotransplants from

infected to nave Gasterosteus aculeatus, but not inheterotransplants to nave Pungitius pungitius. Het-erotransplants of S. solidus plerocercoids to other fishspecies died within 2 to 10 d (Brten 1966). Thusmechanisms of immuno-avoidance described in thepresent study are likely to be a specific feature of theS. solidusG. aculeatus hostparasite relationship. Ina few sticklebacks in the present study, no S. solidusplerocercoids were found after exposure to the para-site. All copepods used for the exposure were micro-scopically checked for the presence of S. solidus pro-cercoids and were definitely ingested, since noremaining copepods were found in the tanks. Thus,exposed but not infested sticklebacks were able toeffectively prevent infective S. solidus procercoidsfrom entering the body cavity or eliminate alreadyestablished plerocercoids. Interestingly, elevatedspontaneous ROS production was observed exclu-sively with leukocytes from exposed but not infectedfish (Fig. 3), indicating that production of ROS isindeed part of a successful immune response againstS. solidus. Infected fish showed a significantly lowerspontaneous ROS production, but a high responseafter PMA stimulation. This might indicate that, onceS. solidus plerocercoids are growing in the bodycavity of G. aculeatus, granulocytes are unable todevelop their activity against the parasite. Investiga-tions on peritoneal leukocytes isolated during S.solidus infection might provide interesting comple-ments to the present work. However, access to leuko-cytes in sufficient numbers for functional assays islimited in the stickleback. The use of leukocytes fromthe pronephros is a compromise; however, functionalresponses of pronephros leukocytes from infected andnon-infected stickleback are differential, reflecting anin vivo priming by the parasite.

Schistocephalus solidus appears to be very welladapted to the immunological features of its specificintermediate host Gasterosteus aculeatus. The initialphase of infection seems to be decisive for the devel-opment of a parasitosis. It is likely that mobilisationand activation of granulocytes can be effective againstprocercoid stages of S. solidus during the first weeks ofinfection, but once plerocercoid stages are present inthe body cavity, granulocytes are unable to developtheir activity against S. solidus.

Acknowledgements. We would like to thank P. B. Aeschli-mann, M. A. Hberli and M. Michaud for their support in theinfection experiments. Furthermore, we thank N. Karstens forher help in flow-cytometer trouble-shooting. We are gratefulto G. F. Wiegertjes, Cell Biology and Immunology Group,Wageningen University, The Netherlands, for critical readingof the manuscript.

148

Scharsack et al.: Granulocyte responses in tapeworm infected sticklebacks

LITERATURE CITED

Aeschlimann P, Hberli M, Milinski M (2000) Threat-sensitivefeeding strategy of immature sticklebacks (Gasterosteusaculeatus) in response to recent experimental infectionwith the cestode Schistocephalus solidus. Behav EcolSociobiol 49:17

Arme C, Owen RW (1967) Infections of the three-spined stick-leback, Gasterosteus aculeatus L., with the plerocercoidlarvae of Schistocephalus solidus (Mller, 1776), with spe-cial reference to pathological effects. Parasitology 57:30114

Arnott SA, Barber I, Huntingford FA (2000) Parasite-associ-ated growth enhancement in a fishcestode system. ProcR Soc Lond B 267:65763

Barber I, Ruxton GD (1998) Temporal prey distribution affectsthe competitive ability of parasitized sticklebacks. AnimBehav 56:14771483

Barber I, Arnott SA, Braithwaite VA, Andrew J, HuntingfordFA (2001) Indirect fitness consequences of mate choice insticklebacks: offspring of brighter males grow slowly butresist parasitic infections. Proc R Soc Lond B 268:7176

Binz T, Reusch TBH, Wedekind C, Schrer L, Sternberg JM,Milinski M (2000) Isolation and characterization ofmicrosatellite loci from the tapeworm Schistocephalussolidus. Mol Ecol 9:19261927

Brten T (1966) Host specificity in Schistocephalus solidus.Parasitology 56:657664

Candolin U, Voigt HR (2001) No effect of a parasite on repro-duction in stickleback males: a laboratory artefact? Para-sitology 122:457464

Chilmonczyk S, Monge D (1999) Flow cytometry as a tool forassessment of the fish cellular immune response topathogens. Fish Shellfish Immunol 9:319333

Christen M, Milinski M (2003) The consequence of self-fertil-ization and out-crossing of the cestode Schistocephalussolidus in its second intermediate host. Parasitology 126:369378

Damian RT (1997) Parasite immune evasion and exploitation:reflections and projections. Parasitology 115(Suppl): S169S175

Dubinina MN (1980) Tapeworms (Cestoda, Ligulidae) of thefauna of the USSR. Amerind Publishing, New Delhi

Hberli MA, Aeschlimann PB, Michaud M, Milinski M (2002)Growth pattern of infection and sex bias to prevalence ofSchistocephalus solidus larvae in three-spined stickle-backs (Gasterosteus aculeatus). In: Hberli MA, Odourcues and sexual selection in the three-spined stickleback(Gasterosteus aculeatus). PhD thesis, University of Berne

Hoole D, Arme C (1982) Ultrastructural studies on the cellularresponse of roach, Rutilus rutilus L., to the plerocercoidlarva of the pseudophyllidean cestode, Ligula intestinalis.J Fish Dis 5:131144

Hoole D, Arme C (1983) Ultrastructural studies on the cellularresponse of fish hosts following experimental infectionwith the plerocercoid of Ligula intestinalis (Cestoda:Pseudophyllidae). Parasitology 87:139149

Kurtz J, Kalbe M, Aeschlimann PB, Hberli MA, Wegner KM,Reusch TBH, Milinski M (2004) Major histocompatibilitycomplex diversity influences parasite resistance andinnate immunity in sticklebacks. Proc R Soc Lond B271:197204

McPhail JD, Peacock SD (1983) Some effects of the cestodeSchistocephalus solidus on reproduction in the three-spined stickleback Gasterosteus aculeatus, evolutionaryaspects of hostparasite interaction. Can J Zool 61:901908

Meakins RH (1974) A quantitative approach to the effects ofthe plerocercoid of Schistocephalus solidus Mller 1776on the ovarian maturation of the three-spined sticklebackGasterosteus aculeatus L. Z Parasitenkd 44:7379

Meusen EN, Balic A (2000) Do eosinophils have a role in thekilling of helminth parasites? Parasitol Today 16:95101

Milinski M (1984) Parasites determine a predators optimalfeeding strategy. Behav Ecol Sociobiol 15:3537

Milinski M (1985) Risk of predation of parasitized stickle-backs (Gasterosteus aculeatus) under competition forfood. Behaviour 93:203216

Milinski M (1990) Parasites and host decision making. In:Barnard CJ, Behnke JM (eds) Parasitism and host behav-iour. Taylor & Francis, London, p 95116

Miller NW, McKinney EC (1994) In vitro culture of fish leuko-cytes. In: Hochachka PW, Mommsen TP (eds) Biochem-istry and molecular biology of fishes. Elsevier Science,Amsterdam, p 341353

Nie P, Hoole D (2000) Effects of Bothriocephalus acheilog-nathi on the polarization response of pronephric leuco-cytes of carp, Cyprinus carpio. J Helminthol 74:253257

Orr TSC, Hopkins CA, Charles GH (1969) Host specificity andrejection of Schistocephalus solidus. Parasitology 59:683690

verli O, Pall M, Borg B, Jobling M, Winberg S (2001) Effectsof Schistocephalus solidus infection on brain monoaminer-gic activity in female three-spined sticklebacks Gasteros-teus aculeatus. Proc R Soc Lond B 268:14111415

Pechhold K, Pohl T, Kabelitz D (1994) Rapid quantification oflymphocyte subsets in heterogeneous cell populations byflow cytometry. Cytometry 16:152159

Reimchen TE, Nosil P (2001) Ecological causes of sex-biasedparasitism in threespine stickleback. Biol J Linn Soc 73:5163

Secombes C J, Chappell LH (1996) Fish immune responses toexperimental and natural infection with helminth para-sites. Annu Rev Fish Dis 6:167177

Sharp GJ, Pike AW, Secombes CJ (1992) Sequential develop-ment of the immune response in rainbow trout(Oncorhynchus mykiss [Walbaum, 1792]) to experimentalplerocercoid infections of Diphyllobothrium dendriticum(Nitzsch, 1824). Parasitology 104:169178

Smyth JD (1954) Studies on tapeworm physiology. VII. Fertil-ization of Schistocephalus solidus in vitro. Exp Parasitol 3:6471

Taylor MJ, Hoole D (1989) Ligula intestinalis (L.) (Cestoda:Pseudophyllidea): plerocercoid-induced changes in thespleen and pronephros of roach, Rutilus rutilus (L.) andgudgeon, Gobio gobio (L.). J Fish Biol 34:583596

Taylor MJ, Hoole D (1993) Ligula intestinalis (L.) (Cestoda:Pseudophyllidea): polarization of cyprinid leucocytes asan indicator of host- and parasite-derived chemoattrac-tants. Parasitology 107:433440

Taylor MJ, Hoole D (1994) Modulation of fish lymphocyte pro-liferation by extracts and isolated proteinase inhibitors ofLigula intestinalis (Cestoda). Fish Shellfish Immunol 4:221230

Taylor MJ, Hoole D (1995) The chemiluminescence of cyprinidleucocytes in response to zymosan and extracts of Ligulaintestinalis (Cestoda). Fish Shellfish Immunol 5:191198

Wedekind C, Strahm D, Schrer L (1998) Evidence for strate-gic egg production in a hermaphroditic cestode. Parasitol-ogy 117:373382

Wegner KM, Kalbe M, Kurtz J, Reusch TBH, Milinski M(2003a) Parasite selection for immunogenetic optimality.Science 30:1343

Wegner KM, Reusch TBH, Kalbe M (2003b) Multiple para-

149

Dis Aquat Org 59: 141150, 2004

sites are driving major histocompatibility complex poly-morphism in the wild. J Evol Biol 16:233241

Whyte SK, Chappell LH, Secombes CJ (1989) Cytotoxic reac-tions of rainbow trout, Salmo gaidneri Richardson, macro-phages for larvae of the eye fluke Diplostomum

spathaceum (Digenea). J Fish Biol 35:333345Whyte SK, Chappell LH, Secombes CJ (1990) Protection of

rainbow trout (Salmo gaidneri) against Diplostomumspathaceum (Digenea): the role of specific antibody andactivated macrophages. J Fish Dis 13:281291

150

Editorial responsibility: Wolfgang Krting,Hannover, Germany

Submitted: June 30, 2003; Accepted: November 10, 2003Proofs received from author(s): April 21, 2004