Acta Tropica 92 (2004) 193–203

Susceptibility of TNF-�-deficient mice toTrypanosomacongolenseis not due to a defective antibody response

Jan Naessensa,∗, Hiroshi Kitanib, Eiichi Momotanib, Kenji Sekikawac,Joseph M. Nthalea, Fuad Iraqia

a International Livestock Research Institute, P.O. Box 30709, Nairobi 00100, Kenyab National Institute of Animal Health (NIAH), Tsukuba, Ibaraki, Japan

c National Institute of Agrobiological Sciences (NIAS), 3-1-5 Kannondai, Tsukuba, Ibaraki 305-8602, Japan

Received 4 September 2002; received in revised form 6 April 2004; accepted 11 May 2004Available online 25 August 2004

Abstract

C57BL/6 mice deficient in one or two copies of the gene for tumor necrosis factor alpha (TNF-�) were more susceptibleto Trypanosoma congolenseinfection than their resistant, wild-type counterparts. The number of TNF-� genes was correlatedwith the capacity to control parasitaemia and with survival time. Absence of TNF-� resulted in a diminished capacity toform germinal centres in lymph nodes and spleen. Since germinal centres are involved in antibody production and affinitymaturation, the susceptibility of the TNF-�-deficient mice could have been due to this secondary defect. Despite the lack oft me antigensw the twom ento©

K

v

f

edgicalpoly-and

0

he germinal centres, the antibody responses to internal and exposed trypanosome antigens and to non-trypanosoere not significantly different. Also the relative avidities measured in infected sera did not significantly differ betweenouse strains. These data suggest that the role of TNF-� in control ofT. congolensewas not due to its role in the developmf an antibody response.2004 Elsevier B.V. All rights reserved.

eywords:TNF-�; Trypanosomosis; Trypanosoma congolense; Disease resistance; Antibody response; Germinal centres

Abbreviations:d.p.i., days post-infection; PCV, percent red cellolume

∗ Corresponding author. Tel.: +254 2 630743x4406;ax: +1 707 885 7781.

E-mail address:j.naessens@cgiar.org (J. Naessens).

1. Introduction

Murine infections withTrypanosoma bruceiandTrypanosoma congolenseparasites are characterizby very high parasitaemias and serious patholocomplications such as anaemia, splenomegaly,clonal B cell-activation and immune suppression,

001-706X/$ – see front matter © 2004 Elsevier B.V. All rights reserved.doi:10.1016/j.actatropica.2004.05.015

194 J. Naessens et al. / Acta Tropica 92 (2004) 193–203

are usually lethal to the host (Sileghem et al., 1994;Naessens et al., 2001, 2002). Increased parasitaemiaand earlier death rates after B cell-depletion (Campbellet al., 1977), but not after T cell-depletion (Campbellet al., 1978; Clayton et al., 1979; Bakhiet et al., 1990)suggested a major role for antibodies in parasitaemiacontrol. The capacity of variant antigen-specific anti-bodies, but not antibodies to internal parasite antigens,to control trypanosomes by complement-mediated ly-sis (Crowe et al., 1984), agglutination (Russo et al.,1994), immobilization (Wei et al., 1990) or increasedmacrophage uptake (Ngaira et al., 1983; Takayanagiet al., 1987) suggested a role for antibodies to surface-exposed epitopes in trypanosome clearance. Duringthe first wave of parasitaemia, the production of vari-ant antigen-specific antibodies is correlated with areduction in parasitaemia. However, the capacity ofbloodstream forms to continuously express new variantsurface glycoproteins ensures survival of adequatenumbers of parasites to give rise to new waves of par-asitaemia and disease.

A number of investigations have suggested an im-portant role for tumor necrosis factor� (TNF-�) in thepathogenesis of trypanosomosis. Differential expres-sion of TNF-� has been reported in resistant and sus-ceptible strains of mice (Uzonna et al., 1998). TNF-�was demonstrated to have trypanolytic activity in vitro(Lucas et al., 1994; Magez et al., 1997) and to reduceparasite numbers ofTrypanosoma b. brucei(Magez etal., 1993) andTrypanosoma cruzi(Silva et al., 1995;L etg este ,1 ,1f2a er-ec er ofm2 estt ofp con-t ayb ii

Tumor necrosis factor alpha could contribute to try-panosome resistance through any of its known func-tions, its role in induction of acute phase proteins,role in suppression of haemopoiesis or its central rolein inflammatory processes. Another way that TNF-�could interfere with resistance is through suppress-ing antibody production. Although TNF-� is not di-rectly involved in adaptive immunity, its central rolein the cytokine cascade may influence immune re-sponses during particular situations (Le Hir et al., 1996;Pasparakis et al., 1996; Wedlock et al., 1999). Lackof TNF-� during foetal development or during an im-mune response may induce secondary changes to theimmune system, which in turn could lead to defectiveimmune responses. The greater susceptibility of TNF-�-deficient mice may therefore be due to defectiveresponses caused by such developmental alterations,rather than from a primary lack of TNF-�. Previousfindings (Pasparakis et al., 1996; Taniguchi et al., 1997)showed that TNF-�-deficient mice were unable to de-velop germinal centres in response to T cell-dependentantigens in the peripheral lymphoid organs and thismay be due to lack of competent follicular dendriticcells (Korner et al., 1997). Since germinal centres are amajor site for antibody production and affinity matura-tion during immune responses (Przylepa et al., 1998),the role of TNF-� in trypanoresistance could thus sim-ply be attributed to an inefficient antibody responseresulting from a lack of germinal centre formation.

This paper describes the effect of TNF-� gene abla-ta es ina

2

2

ereg ta yh es-t -� tedi erica icew L/6

ima et al., 1997) in rodent models. In addition, murinrypanosomosis resistance QTL mappings inT. con-olenseinfections suggest that the locus with largffect encompasses the TNF-� gene (Kemp et al., 1996997). TNF-� gene-knockout mice (Taniguchi et al.997) were highly susceptible toT. congolensein-

ection, died earlier than wild-type mice (Iraqi et al.,001) and developed higher parasitaemias (Kitani etl., 2002). In contrast, survival rates were not diffnt in TNF-�-deficient mice infected withT. b. bru-ei, despite higher parasitaemias of the same ordagnitude as inT. congolenseinfections (Kitani et al.,002; Magez et al., 1999). These observations sugg

hat TNF-� plays an important role in the controlarasitaemia in murine trypanosomosis, but its

ribution to control of pathogenesis and survival me different betweenT. congolenseandT. b. bruce

nfections.

ion in mice on resistance to aT. congolenseinfectionnd attempts to correlate these effects to changcquired immune responses.

. Materials and methods

.1. Mice

Tumor necrosis factor alpha-deficient mice wenerated by gene-targeting technology (Taniguchi el., 1997). In brief, the TNF-� gene was disrupted bomologous recombination in embryonic stem cells

ablished from the CBA× C57BL/6 F1 mice. TNFgene-knock-out embryonic stem cells were injec

nto C57BL/6 blastocysts, and the resulting chimanimals were mated to C57BL/6 mice. The mutant mere backcrossed for six generations with C57B

J. Naessens et al. / Acta Tropica 92 (2004) 193–203 195

mice. Homozygous and hemizygous TNF-�-deficientmice and wild-type mice were obtained from thecrosses of hemizygous mice and maintained for 8–12weeks of age. After confirmation of genotype by PCRamplification (Iraqi et al., 2001), equal numbers ofmales and females of each genotype were used for theexperiment, to avoid sex-linked differences.

In a first experiment nine wild-type, 10 hemizy-gous mice and 12 homozygous mice were comparedthroughout infection for survival and parasitaemia. Ina second experiment only wild-type and TNF-�-knock-out mice were infected and six mice (three male andthree female) from each strain sacrificed at differenttime points and blood collected from the heart for an-tibody analysis. A third experiment compared indi-vidual wild-type and TNF-�-knock-out mice. Fifty-six mice of each type were infected, and at eachtime point eight mice were sacrificed and blood col-lected. Blood from eight uninfected mice was taken ascontrol.

2.2. Parasite challenge and serum collection

Trypanosoma congolenseclone IL 1180 (Nantulyaet al., 1982) was grown in sublethally irradiatedSprague-Dawley rats, and trypanosomes were isolatedfrom infected rat blood by anion exchange column(Lanham, 1968). Mice were infected intra-peritoneallywith 1 × 104 parasites in 200�l of phosphate bufferedsaline (PBS, pH 8.0) containing 1.5% glucose. Un-i nge,b ter-v aredf s oft singa

sumo byt

erek acp wascr ughs plingp thes ones .

2.3. Histopathological procedures

Peripheral lymph nodes and spleen were fixed with10% buffered formalin, and paraffin sections were ex-amined with hematoxylin-eosin staining for germinalcentre formation.

2.4. Antibody titres

The relative titres of trypanosome-binding anti-bodies were measured by an ELISA method. Wholelysate from trypanosomes ofT. congolensecloneIL1180 were prepared as previously described (Buzaet al., 1997). Recombinant trypanosome immunoglob-ulin heavy chain binding protein HSP70 (Boulange andAuthie, 1994) was a gift of Dr. Alain Boulange (ILRI).Wells of a 96-well Polysorp plate (Nunc, Roskilde,Denmark) were coated overnight at 4◦C with recom-binant HSP70 or with 5�g/ml of the parasite lysate incarbonate buffer, pH 9.6, rinsed 3 times in PBS andblocked for 30 min at 37◦C with PBS containing 1%Casein and 0.05% Tween20. The blocking buffer wasdiscarded and the plates were subsequently incubatedfor 1 h at 37◦C with sera or plasma from trypanosome-infected mice diluted at 1:50 in PBS containing 1%skimmed milk and 0.1% Tween20. The plates werebriefly washed 5 times and incubated for 10 min at roomtemperture (RT) in PBS containing 0.1% Tween20.Trypanosome-binding IgG and IgM antibodies weredetected by incubating plates for 30 min at 37◦C withg RPc di-l ilka anda erei b-s ISAp ki,F

2

oft ys-t CNcm ounto ody

nfected mice were used as controls. After challelood samples were taken by tail snip, at 3–7 day inals up to 60 days, and wet blood smears were prepor determination of parasitaemia. Actual numberrypanosomes per ml of blood were enumerated uhaemocytometer.The average survival time was calculated as the

f the survival times of the individual mice, dividedhe number of mice.

On different sampling days, groups of six mice willed in a CO2 jar, blood was collected by cardiuncture and allowed to clot after which serumollected, aliquoted and stored at−80◦C until use. Toeduce individual variations and to obtain large enoamples to do several experiments for each samoint, equal volumes of sera from three males (andame for three females) were pooled together asample for monitoring antibody titres and affinities

oat anti-mouse IgG- or goat anti-mouse IgM-Honjugates (Sigma Chemical Co, St. Louis, MO)uted at 1:800 in PBS containing 1% skimmed mnd 0.1%Tween20. After making five brief washesn additional 10-min incubation wash, the plates w

ncubated for a further 30 min at RT with K-Blue sutrate and the optical density was read in an ELlate reader (Titertek Multiscan MCC340, Helsininland) at 620 nm.

.5. Relative antibody avidities

To determine the relative binding aviditiesrypanosome-binding IgG antibodies, the ELISA sem described above was modified to include a S-

haotropic ion elution step (Goldblatt, 1997). Thisethod is based on the observation that the amf chaotropic ion necessary to denature the antib

196 J. Naessens et al. / Acta Tropica 92 (2004) 193–203

and reduce its binding is proportional to its affinity. Inthis way, relative avidities of different antibodies to thesame antigen can be compared.

Briefly, for each sample under investigation, sixreplicate wells were coated and blocked as above. Thewells were subsequently incubated for 1 h at 37◦C witha 1:50 dilution of mouse sera or plasma. Plate wellswere washed 5 times and 1× 10 min in PBS contain-ing 0.1% Tween20 and incubated for 15 min at 37◦Cwith increasing molar solutions of SCN−: 0, 0.25, 0.5,1, 2 and 3 M to elute trypanosome-bound antibodies.The wells were rinsed 3 times with PBS and the re-maining amount of binding antibody was quantified byincubating wells for 30 min at 37◦C with goat anti-mouse IgG-HRP conjugate (Sigma Chemical Co, St.Louis, MO). Wells were washed 5 times and incubatedonce for 10 min, before incubation with K-Blue sub-strate for 30 min at RT and read at 620 nm as describedabove.

The relative avidities were calculated as describedpreviously (Goldblatt, 1997). The relative avidity in-dex was defined as the molarity of SCN- needed to re-duce binding of specific antibody by 50%. An estimateof the ratio of antibody eluted by different molaritiesof SCN− was calculated using the following formula:(mean O.D. of wells without SCN−−mean O.D. ofwells containing SCN−)/mean O.D. of wells withoutSCN−.

2.6. Flow cytometry analysis

pesw ecti de-s ro tobp , in9 hedt u-c ds ernBo es.A icea lsw Eu-r ean

fluorescence was measured using CellQuest software(Becton Dickinson).

2.7. Statistical analyses

Statistical analyses of variance of parasitaemia andantibody titres were performed by Genstat with termsfor mouse strain, day, and strain× day. Parasite num-bers were transformed to logarithms before analysis,because of skewed variation. Only data of antibodyresponses after 1 week were included in the analysisof variance, as before that time no parasites were de-tectable in blood and no significant responses were ex-pected. Differences were considered significant whenP < 0.05.

3. Results

3.1. Parasitaemia

Parasitaemia was compared between three groupsof infected C57BL/6 mice: mice homozygous for thedeficient TNF-� gene, hemizygous mice with one de-ficient TNF-� gene and wild-type mice. Parasitaemiaswere significantly different between the three groups(P < 0.001, for parasitaemias between 14 and 35 dayspost infection (d.p.i.)). The lowest parasitaemias wererecorded in wild-type mice and the highest in the TNF-�-deficient mice, while parasitaemias in the hemizy-gous mice were intermediate (Fig. 1). The highest dif-f

3

rt ea( lyl0 andt s-t entm em-i

3

test inal

Antibody titres to trypanosome surface epitoere measure by flow cytometry using an indir

mmunofluorescnce-binding assay as previouslycribed (Williams et al., 1996). Twenty five microlitef 25-fold dilutions of mouse sera were allowedind to 25�l of 5 × 105 T. congolenseIL-1180 try-anosomes, DEAE-purified from infected rat blood6-well plates for 10 min on ice. Cells were was

wice with 200�l ice-cold PBS containing 1.5% glose in a cooled centrifuge at 4◦C. The fluoresceinateecond step antibody to mouse IgG or IgM (Southiotechnology Associates, AL, USA) was added (25�lf a 30-fold dilution) to the resuspended trypanosomfter 10 min on ice, trypanosomes were washed twnd fixed in PBS containing 2 % formaldehyde. Celere analysed on FACScan (Becton Dickinson,

opean HQ, Erembodegem-Aalst, Belgium) and m

erences were noted after 3 weeks of infection.

.2. Survival time

The average survival time was 52± 18 days fohe wild-type mice, 37± 18 for the hemizygous micnd 34± 8 for the homozygous, TNF-�-deficient miceFig. 2). The wild-type mice survived significantonger than mice with deficient TNF-� genes (P <.02), but the difference between the hemizygous

he homozygous TNF-�-deficient mice was not statiically significant. However, no homozygous deficiice survived beyond day 45, unlike some of the h

zygous mice.

.3. Germinal centres

One important histological observation that relao the immunological responses is the lack of germ

J. Naessens et al. / Acta Tropica 92 (2004) 193–203 197

Fig. 1. Mean parasitaemia inT. congolense-infected C57BL/6 mice. Broken line with empty circle: mice deficient for both TNF-� genes (TNF-�−/−); broken line and + symbol: hemizygous mice deficient for one TNF-� gene (TNF-� +/−); solid line and solid circle: wild-type (TNF-�+/+)mice.

centres in TNF-�-deficient mice (Taniguchi et al.,1997). Since germinal centres are crucial in thegeneration of antibody responses in lymphoid tis-sues, their development was carefully monitored dur-ing a T. congolenseinfection. Sections from spleenand five lymph nodes from a male and a femalemouse and from wild-type and TNF-�-deficient strainswere examined at different days post-infection. Un-like infected wild-type animals, germinal centreswere not detected in infected TNF-� knock-out mice(Fig. 3).

Fig. 2. Survival rate ofT. congolense-infected wild-type mice (full line, TNF-� +/+) and mice deficient for one (hyphenated line, TNF-� +/−)or two copies (broken line, TNF-�−/−) of the TNF-� gene.

3.4. Antibody levels

Serum was collected from wild-type and TNF-�-deficient mice, sacrificed at different times after infec-tion and samples from three mice pooled. The total IgMresponses to trypanosome antigens did not statisticallydiffer between the wild-type and the TNF-�-deficientmice (Fig. 4A). Similarly, the IgM response to non-trypanosome antigens�-galactose and LPS, which hasoccur during trypanosome infections (Hudson et al.,1976) and which could play a role in the first-line of

198 J. Naessens et al. / Acta Tropica 92 (2004) 193–203

Fig. 3. Lack of germinal centre formation in TNF-�-deficient mice. Spleen sections from wild-type (left) and TNF-�-deficient (right) mice 3days after infection withT. congolense, stained with hematoxylin and eosin. Arrow indicates a germinal centre in a B cell-follicle.

defence (Buza et al., 1997), were also similar in bothmouse strains and no significant differences were found(Fig. 4C and D). Differences in the IgG responses werenot statistically significant (P> 0.05) up to day 24 d.p.i.However after that time significantly lower IgG titreswere observed in TNF-�-deficient mice (Fig. 4B).

In a repeat experiment the first parasitaemic wavepeaked at day 11 post-infection, but now antibody titreswere measured in samples from individual mice(Fig. 5). The IgM and IgG antibody responses to apurified, internal antigen (HSP70) were compared be-tween wild type and TNF-�-knock-out mice (Fig. 5Aand B). The anti-HSP70 IgM titres increased duringthe first peak of parasitaemia, but no statisticallysignificant differences were observed between wildtype and TNF-�-knock-out mice (P > 0.05). Nodifferences were seen between the anti-HSP70 IgGtitres of the two mouse breeds during the first waveof parasitaemia and up to 21 d.p.i., but on 28 d.p.i. theIgG titre in wild type breed was significantly higherthan in the TNF-�-knock-out mice.

The same sera were tested by flow cytometry for an-tibodies to surface-exposed epitopes (Fig. 5C and D).Titres of both trypanosome-binding IgM and IgG anti-bodies increased from day 7 till day 21 in both mousestrains. No statistically significant difference could bedetected (P < 0.05), except for day 28 post-infection,when a statistically higher titre was observed in theIgG, but not IgM, response in wild type mice.

3

e-b opic

ion elution ELISA (Fig. 6). The relative avidities in-creased slowly during infection, but there were no sig-nificant differences in the relative avidities of anti-trypanosome antibodies between wild-type and TNF-�-deficient mice.

4. Discussion

This experiment confirmed a previous report (Iraqiet al., 2001) that trypanoresistant C57BL/6 mice weremore susceptible toT. congolenseinfections when de-ficient for the TNF-� gene. The capacity of the miceto control parasitaemia correlated with the number ofTNF-� gene copies present, confirming the previousobservation that TNF-� plays a role in the control ofparasitaemia in infections ofT. congolense(Kitani etal., 2002) and the related parasiteT. b. brucei(Magezet al., 1999). The data also suggested a positivecorrelation between survival times and number of TNF-� gene copies. Wild-type mice survived longest, fol-lowed by hemizygous and homozygous deficient mice.Although the mean survival times of homo- and hem-izygous deficient mice were not significantly different,no homozygous mice survived beyond day 45, whilethe last hemizygous survivor reached day 68. How-ever, no correlation between time of death and para-sitaemia was found within mouse strains. Several TNF-�-knock-out mice survived with higher parasitaemiastm ii ns( g

.5. Antibody affinities

The average, relative affinities of trypanosominding antibodies were measured using a chaotr

han mice that had succumbed earlier. TNF-�-deficientice also had a reduced capacity to controlT. bruce

nfections but did not die earlier from these infectioKitani et al., 2002; Magez et al., 1999), suggestin

J. Naessens et al. / Acta Tropica 92 (2004) 193–203 199

Fig. 4. Mean relative antibody titres and standard deviations fromT. congolense-infected mice. IgM (A) and IgG (B) titres to a trypanosomelysate and IgM titres to non-trypanosome antigens�-galactosidase (C) and LPS (D). Solid symbols and full lines represent optical densities ofwild-type mice; empty symbols and broken lines represent those of TNF-�-deficient mice.

that factors other than parasitaemia determine survivaltime. Studies in trypanosome-infected cattle suggestedthat survival and productivity were not correlated withparasitaemia, but there was a correlation with the de-gree of anaemia (Naessens et al., 2001, 2003).

Although a previous paper described no germinalcentre formation in spleens ofT. congolense-infectedmice (Morrison et al., 1981), in our experiments severalgerminal centres were observed in spleens and lymph

nodes of C57BL/6 wild-type mice. This discrepancymay be due to the fact that susceptible C3H/He micewere used in the previous experiment and these miceare expected to show more severe pathology than themore resistant C57BL/6 mice. As expected from pre-vious studies in TNF-�-knock-out mice (Pasparakis etal., 1996; Taniguchi et al., 1997), these mice showed asevere lack of germinal centre development in our ex-periments, not only in spleen, but also in lymph nodes.

200 J. Naessens et al. / Acta Tropica 92 (2004) 193–203

Fig. 5. Mean relative antibody titres and standard deviations fromT. congolense-infected mice. IgM (A) and IgG (B) titres to an internaltrypanosome antigen, HSP70. IgM (C) and IgG (D) titres to trypanosome surface epitopes. Solid symbols and full lines represent opticaldensities of wild-type mice; empty symbols and broken lines represent those of TNF-�-deficient mice.

Antibody isotype titres and relative antibody avidi-ties to parasite antigens in a trypanosome lysate did notsignificantly differ between the two mouse strains. IgMtitres to trypanosome as well as to non-trypanosomeantigens varied strongly between the pooled samplesbut this variation was similar in the two mouse strains.A separate experiment confirmed that no significantdifferences could be detected between IgM titres from

wild type and TNF-�-knock-out mice to a recombinantinternal trypanosome antigen and to surface-exposedepitopes. In those two experiments, IgG titres to inter-nal and exposed antigens were also similar during thefirst parasitaemic wave, although our results showed adifference much later in infection. The responses dur-ing the first peak of parasitaemia proved that TNF-�-deficient mice could mount IgM and IgG responses

J. Naessens et al. / Acta Tropica 92 (2004) 193–203 201

Fig. 6. Relative avidities of antibodies to trypanosome antigens inwild-type (closed circles) and TNF-�-deficient mice (open circles) inexperiment 1 (top) and 2 (bottom). There is no significant differencebetween the two strains.

to the same degree as wild type mice, and suggestedthat they had no inherent defect that prevented induc-tion and secretion of antibody during infection. It isinteresting to note that this was observed in both the Tcell-dependent responses to internal antigens and theresponse to surface-exposed epitopes, which is mainlya T cell-independent response (Reinitz and Mansfield,1988, 1990). These observations corroborate and ex-tend the findings ofMagez et al. (1999), who found nodifferences in IgM and IgG titres toT. bruceiantigens.However, in our experiments a statistically significantdifference in IgG titres to internal and exposed antigenswas observed by day 28 post-infection, which is longafter the first peak of parasitaemia that occurred aroundday 11 post-infection. The data thus suggest that thehigher trypanosome load in the TNF-� knock-out miceduring the first parasitaemic wave is not a consequenceof a lower antibody response. The lower IgG titres ob-served in knock-out mice during the later stages of theinfection, after 3 weeks post-infection, may be due tothe more severe pathology (f.e. immune suppression)resulting from the higher parasitaemic load, rather thanfrom the lack of TNF-�. Previous reports on differ-

ent mouse models also suggest that parasitaemia maynot be correlated to the size of the antibody response.Analysis of the resistance trait toT. b. rhodesiensein more susceptible C3HeB mice and more resistantB10.BR mice, revealed no link between parasitaemiaand change in hematocrit with antibody production orsurvival time (Seed and Sechelski, 1989). Comparingthe capacity of C3H/He and C57BL/6 mice to resistT. b. bruceiinfection, it was suggested that the reduc-tion in the ability to secrete Ig in susceptible mice laterin infection was a consequence rather than the causeof differences in parasitaemic waves (Newson et al.,1990).

Antibody avidities did not differ between the twomouse strains, suggesting that although TNF-� is im-portant in the development of germinal centres, it doesnot affect the relative antibody avidity. A large frac-tion of the antibody response to surface-exposed anti-gens in trypanosome-infected mice is produced in aT cell-independent way (Reinitz and Mansfield, 1990)and may not even occur in the germinal centres. Dur-ing the first 3 weeks of infection, the avidity increasedslightly, but significantly, in both wild-type and TNF-�-deficient mice. Mice deficient for lymphotoxin-�(or TNF-�) do not form spleen germinal centres andin addition lack lymph nodes and Peyer’s patches(Matsumoto et al., 1996). Despite the lack of thesecentres for affinity maturation, high affinity antibodieswere produced upon repeated immunizations as mea-sured by typical somatic mutations (Matsumoto et al.,1 fora -a ss.

pos-s cu-l trol,s ssen-t f ac andn nti-b bilityt

roleo in apb ,1t ical

996). Other sites may compensate as a locationntibody production and affinity maturation and LT�nd TNF-� do not seem to be involved in this proce

Of course, these experiments do not rule out theibility that differences exist in antibodies with partiar specificities that may play a role in parasite conuch as antibodies to flagellar pocket antigens or eial receptors. But there is little reason why lack oytokine would affect certain antibody specificitiesot others. It remains unlikely that differences in aody responses account for the diminished capa

o control parasite growth in TNF-� knock-out mice.An alternative mechanism that could explain the

f TNF-� in trypanosome control was suggestedrevious report. TNF-� was shown to directly lyseT.. bruceiat certain developmental stages (Magez et al.997). However, the concentrations of TNF-� needed

o lyse trypanosomes were higher than physiolog

202 J. Naessens et al. / Acta Tropica 92 (2004) 193–203

levels. Furthermore, in vitro culture studies in our lab-oratory (Kitani et al., 2002) suggested that the growthof T. b. bruceiandT. congolenseclones could not beinhibited by recombinant TNF-� even at maximal con-centrations that could be attained in the culture systems(3000–15,000 units TNF-�/ml medium). So, while thisdirect lytic mechanism may contribute to the control ofparasitaemia in certain trypanosome species, it is pos-sible that TNF-� may be indirectly involved in the con-trol of parasitaemia through the host’s defence mecha-nisms. Since TNF-� is also a crucial mediator of innateand inflammatory responses in the host during infec-tions, it may indirectly influence trypanosome growththrough modified secretion of acute phase proteins orother innate factors.

Acknowledgements

The authors would like to thank Dr. John Rowlandsand Sonal Nagda from the Biometrics Group ILRI, forhelp with statistical analysis, Mr. Robert King fromILRI, for production of mice from different strains, Ms.Peris Boit and Mr. Tindih Shelton Hesborne from JomoKenyatta University of Agriculture and Technology, foradditional technical help. They also appreciate the in-tellectual contributions of Dr. Yoshio Nakamura andDr. John Gibson. This study was conducted under theCollaborative Research Project between JIRCAS andI er2

R

B 990.alin.

B igenus toogy

B 5+ Bnon--

C an.

Campbell, G.H., Esser, K.M., Phillips, M., 1978.Trypanosoma bru-cei rhodesienseinfection in congenitally athymic (nude) mice.Infect. Immun. 20, 714–720.

Clayton, C.E., Ogilvie, B.M., Askonas, B.A., 1979.Trypanosomabrucei infection in nude mice. B lymphocyte function is sup-pressed in the absence of lymphocytes. Parasite Immunol. 1,39–48.

Crowe, J.S., Lamont, A.G., Barry, J.D., Vickerman, K., 1984. Cyto-toxicity of monoclonal antibodies toTrypanosoma brucei. Trans.Roy. Soc. Trop. Med. Hyg. 78, 508–513.

Goldblatt, D., 1997. Simple solid phase assays of avidity. In: John-stone, A.P., Turner, M.W. (Eds.), Immunochemistry 2. A PracticalApproach. Oxford University Press, Oxford (pp. 31-51).

Hudson, K.M., Byner, C., Freeman, J., Terry, R.J., 1976. Immuno-suppression, high IgM levels and evasion of the immune responsein murine trypanosomiasis. Nature 264, 256–258.

Iraqi, F., Sekikawa, K., Rowlands, J., Teale, A., 2001. Susceptibilityof TNF-� genetically deficient mice toTrypanosoma congolenseinfection. Parasite Immunol. 23, 445–451.

Kemp, S.J., Darvasi, A., Soller, M., Teale, A.J., 1996. Genetic controlof resistance to trypanosomosis. Vet. Immunol. Immunopathol.54, 239–243.

Kemp, S.J., Iraqi, F., Darvasi, A., Soller, M., Teale, A.J., 1997. Lo-calization of genes controlling resistance to trypanosomosis inmice. Nat. Genet. 16, 194–196.

Kitani, H., Black, S.J., Nakamura, Y., Naessens, J., Murphy, N.B.,Yokomizo, Y., Gibson, J., Iraqi, F., 2002. Recombinant tumornecrosis factor alpha does not inhibit the growth of Africantrypanosomes in axenic cultures. Infect. Immun. 70, 2210–2214.

Korner, H., Cook, M., Riminton, D.S., Lemckert, F.A., Hoek, R.M.,Ledermann, B., Kontgen, F., Fazekas de St Groth, B., Sedgwick,J.D., 1997. Distinct roles for lymphotoxin-� and tumor necro-sis factor in organogenesis and spatial organization of lymphoidtissue. Eur. J. Immunol. 27, 2600–2609.

L lood218,

Lfer-sesed.

L P.,or in.

L , J.P.,ping63,

M aet-ctiveith

M eren,ptake

LRI. This study comprises ILRI publication numb00218.

eferences

akhiet, M., Olsson, T., Van der Meide, P., Kristensson, K., 1Depletion of CD8+ T cells suppresses growth ofTrypanosombrucei bruceiand interferon-gamma production in mice. CExp. Immunol. 81, 195–199.

oulange, A., Authie, E., 1994. A 69 kDa immunodominant antof Trypanosoma (Nannomonas) congolense is homologoimmunoglobulin heavy chain binding protein (BiP). Parasitol109, 163–173.

uza, J., Sileghem, M., Gwakisa, P., Naessens, J., 1997. CDlymphocytes are the main source of antibodies reactive withparasite antigens inTrypanosomacongolense-infected cattle. Immunology 92, 226–233.

ampbell, G.H., Esser, K.M., Weinbaum, F.I., 1977.Trypanosomrhodesienseinfection in B-cell-deficient mice. Infect. Immu18, 434–438.

anham, S.M., 1968. Separation of Trypanosomes from the bof infected rats and mice by anion-exchangers. Nature1273–1274.

e Hir, M., Bluethmann, H., Kosco-Vilbois, M.H., Muller, M., diPadova, F., Moore, M., Ryffel, B., Eugster, H.P., 1996. Difentiation of follicular dendritic cells and full antibody responrequire tumor necrosis factor receptor-1 signalling. J. Exp. M183, 2367–2372.

ima, E.C., Garcia, I., Vicentelli, M.H., Vassalli, P., Minoprio,1997. Evidence for a protective role of tumor necrosis factthe acute phase ofTrypanosoma cruziinfection in mice. InfectImmun. 65, 457–465.

ucas, R., Magez, S., De Leys, R., Fransen, L., ScheerlinckRampelberg, M., Sablon, E., De Baetselier, P., 1994. Mapthe lectin-like activity of tumor necrosis factor. Science 2814–817.

agez, S., Lucas, R., Darji, A., Songa, E.B., Hamers, R., De Bselier, P., 1993. Murine tumour necrosis factor plays a proterole during the initial phase of the experimental infection wTrypanosoma brucei brucei. Parasite Immunol. 15, 635–641.

agez, S., Geuskens, M., Beschin, A., del Favero, H., VerschuH., Lucas, R., Pays, E., de Baetselier, P., 1997. Specific u

J. Naessens et al. / Acta Tropica 92 (2004) 193–203 203

of tumor necrosis factor-alpha is involved in growth control ofTrypanosoma brucei. J. Cell Biol. 137, 715–727.

Magez, S., Radwanska, M., Beschin, A., Sekikawa, K., de Baetse-lier, P., 1999. Tumor necrosis factor alpha is a key mediator inthe regulation of experimentalTrypanosoma bruceiinfections.Infect. Immun. 67, 3128–3132.

Matsumoto, M., Lo, S.F., Carruthers, C.J., Min, J., Mariathasan, S.,Huang, G., Plas, D.R., Martin, S.M., Geha, R.S., Nahm, M.H.,Chaplin, D.D., 1996. Affinity maturation without germinal cen-tres in lymphotoxin-alpha-deficient mice. Nature 382, 462–466.

Morrison, W.I., Murray, M., Bovell, D.L., 1981. Response ofthe murine lymphoid system to a chronic infection withTry-panosoma congolense. I. The spleen. Lab. Invest. 45, 547–557.

Naessens, J., Grab, D.J., Sileghem, M., 2001. Identifying the mech-anisms of trypanotolerance in cattle, In: Black, S.J., Seed, J.R.(Eds.), World Class Parasites, Vol.1. Kluwer, Boston, Dordrecht,London, pp. 97–111, The African trypanosomes.

Naessens, J., Teale, A.J., Sileghem, M., 2002. Identification of mech-anisms of natural resistance to African trypanosomiasis in cattle.Vet. Immunol. Immunopathol. 87, 187–194.

Naessens, J., Leak, S.J.A., Kennedy, D.J., Kemp, S.J., Teale, A.J.,2003. Responses of bovine chimaeras combining trypanosomo-sis resistant and susceptible genotypes to experimental infectionwith Trypanosoma congolense. Vet. Parasitol. 111, 125–142.

Nantulya, V.M., Musoke, A.J., Rurangirwa, F.R., Barbet, A.F.,Ngaira, J.M., Katende, J.M., 1982. Immune depression in Africantrypanosomosis: the role of antigenic competition. Clin. Exp. Im-munol. 47, 234–242.

Newson, J., Mahan, S.M., Black, S.J., 1990. Synthesis and secretionof immunoglobulin by spleen cells from resistant and suscepti-ble mice infected withTrypanosoma brucei bruceiGUTat 3.1.Parasite Immunol. 12, 125–139.

Ngaira, J.M., Nantulya, V.M., Musoke, A.J., Hirumi, K., 1983.Phagocytosis of antibody-sensitizedtrypanosoma brucei in vitroby bovine peripheral blood monocytes. Immunol. 49, 393–400.

P 96.

l-rs,Exp.

P elop-biol.

R of Bfrican

trypanosome variable surface glycoproteins. J. Immunol. 141,620–626.

Reinitz, D.M., Mansfield, J.M., 1990. T-cell-independent and T-cell-dependent B-cell responses to exposed variant surface glycopro-tein epitopes in trypanosome-infected mice. Infect. Immun. 58,2337–2342.

Russo, D.C.W., Williams, D.J.L., Grab, D.J., 1994. Mechanisms forthe elimination of potentially lytic complement-fixing variablesurface glycoprotein antibody-complexes inTrypansoma brucei.Parasitol. Res. 80, 487–492.

Seed, J.R., Sechelski, J.B., 1989. African trypanosomes: inheri-tance of factors involved in resistance. Exp. Parasitol. 69, 1–8.

Sileghem, M., Darji, A., De Baetselier, P., Flynn, J.N., Naessens,J., 1994. African Trypanosomosis. In: Kierszenbaum, F. (Ed.),Parasitic infections and the immune system. Academic Press,San Diego, CA, pp. 1–51.

Silva, J.S., Vespa, G.N., Cardoso, M.A., Aliberti, J.C., Cunha, F.Q.,1995. Tumor necrosis factor alpha mediates resistance toTry-panosoma cruziinfection in mice by inducing nitric oxide pro-duction in infected gamma interferon-activated macrophages. In-fect. Immun. 63, 4862–4867.

Takayanagi, T., Kawaguchi, H., Yabu, Y., Itoh, M., Appawy, M.A.,1987. Contribution of the complement system to antibody-mediated binding ofTrypanosoma gambienseto macrophages.J. Parasitol. 73, 333–341.

Taniguchi, T., Takata, M., Ikeda, A., Momotani, E., Sekikawa, K.,1997. Failure of germinal center formation and impairment ofresponse to endotoxin in tumor necrosis factor alpha-deficientmice. Lab. Invest. 77, 647–658.

Uzonna, J.E., Kaushik, R.S., Gordon, J.R., Tabel, H., 1998. Ex-perimental murineTrypanosoma congolenseinfections. I. Ad-ministration of anti-IFN-gamma antibodies alters trypanosome-susceptible mice to a resistant-like phenotype. J. Immunol. 161,5507–5515.

W ar-dju-nol.

W eanti-

W ns,nti-unol.

asparakis, M., Alexopoulou, L., Episkopou, V., Kollias, G., 19Immune and inflammatory responses in TNF�-deficient mice: acritical requirement for TNF� in the formation of primary B celfollicles, follicular dendritic cell networks and germinal centeand in the maturation of the humoral immune response. J.Med. 184, 1397–1411.

rzylepa, J., Himes, C., Kelsoe, G., 1998. Lymphocyte devment and selection in germinal centers. Curr. Top. MicroImmunol. 229, 85–104.

einitz, D.M., Mansfield, J.M., 1988. Independent regulationcell responses to surface and subsurface epitopes of A

edlock, D.N., Goh, L.P., McCarthy, A.R., Midwinter, R.G., Plane, N.A., Buddle, B.M., 1999. Physiological effects and avanticity of recombinant brushtail possum TNF-alpha. ImmuCell. Biol. 77, 28–33.

ei, G., Qualtiere, L., Tabel, H., 1990.Trypanosoma congolens:complement-independent immobilization by a monoclonalbody. Exp. Parasitol. 70, 483–485.

illiams, D.J.L., Taylor, K., Newson, J., Gichuki, B., NaesseJ., 1996. The role of anti-variable surface glycoprotein abody responses in bovine trypanotolerance. Parasite Imm18, 209–218.