Journal of Leukocyte Biology Volume 57, June 1995 943

Elevation of rainbow trout Oncorhynchus mykiss macrophagerespiratory burst activity with macrophage-derived supernatants

Seon I. Jang, Laura J. Hardie, and Christopher J. Secombes

Department ofZoology, University ofAberdeen, Tillydrone Avenue, Aberdeen� United Kingdom

Abstract: A variety of supernatants were prepared bystimulating rainbow trout Oncorhynchus mykiss head kid-ney macrophages with lipopolysaccharide (LPS), tumornecrosis factor a (TNF-ct), or a leucocyte-derived macro-phage-activating factor (1-MAF), individually and in com-bination. If generated using a 12-h stimulation period,such supernatants were found to elevate significantly therespiratory burst activity of target macrophages; that is,they contained a macrophage-derived MAF (m-MAF), butsupernatants generated using a shorter incubation pe-nod showed no significant activity. Combinations ofthese treatments were particularly effective in generatingm-MAF--containing supernatants. The elevation of respi-ratory burst activity by supernatants generated usingcombined treatments could be partially inhibited by priortreatment of the target macrophages with anti-TNF-czreceptor 1 (TNFR1) monoclonal antibodies (mAbs). Simi-larly, treatment of macrophages with combinations of1-MAF and m-MAF generated supernatants with potentm-MAF activity and this activity was partially inhibited byprior treatment of the target cells with anti-TNFR1 mAb.In addition, the presence of anti-transforming growthfactor 13i (TGF-�1) serum while generating these lattersupernatants resulted in significantly increased m-MAFactivity. Such data suggest that fish leukocytes secrete avariety of potent macrophage-activating (TNF-a) and -de-activating (TGF-f3) factors. J. Leukoc. Biol. 57: 943-947;1995.

Key Words: macrophages . MAF . respiratory burst ‘ rain-bow trout ‘ TNF-ct

INTRODUCTION

It is well established that fish macrophages can be acti-vated in vivo and in vitro for enhanced bactericidal activ-ity [i, 2]. Although many factors are known to induce thisphenomenon, the mechanism(s) of in vivo activation hasstill to be elucidated. On the other hand, it is clear that invitro this can be brought about by one or more mole-cules, termed macrophage-activating factors (MAFs), re-leased from head kidney and peripheral blood leukocytes[3]. Several lines of evidence suggest that fish T cellsrelease factors with MAF activity. MAF is released follow-ing stimulation of leucocytes with a T cell mitogen (con-canavalin A) [3], and its release in response to specificantigen can be enhanced by prior immunization [4], thehallmark of lymphocyte responses. MAF release following

Con A stimulation is temperature sensitive [5], as areother knOwn activities of fish T cells [6], and removal ofsurface immunoglobulin-negative cells by panning pre-vents MAY release [7]. In mammals, the predominant

MAF present in supernatants from mitogen-stimulatedleukocytes is interferon-y (IFN-’y) [8]. Similarly, the MAFactivity in supernatants from mitogen-stimulated troutleukocytes cofractionates with IFN activity, and both ac-tivities share similar temperature and pH sensitivities,suggesting that a cytokine akin to IFN-y is released fromfish T cells [9].

Cytokines from other leucocyte types can also activatemammalian macrophages, including macrophage prod-ucts themselves [iO]. In particular, tumor necrosis factora (TNFa) is a potent autocrine signal [1 1], able to elevatemacrophage respiratory burst [12] and microbicidal activ-ity [i3], and to synergise with IFN-’y for the induction ofmicrobicidal [14, 15] and tumoricidal activity [16, i7].Recently, it has been shown that rainbow trout lympho-cytes, macrophages, and neutrophils can respond to re-combinant human TNF-cz [i8, 19] and that theseresponses can be inhibited by prior treatment of themacrophages with monoclonal antibodies (mAbs) to the55-kDa TNF-a receptor (TNFR1) [19]. Such findingsstrongly suggest that this TNFR1 has been conservedwithin the vertebrates. In addition, responsiveness toMAF activity in trout leukocyte supernatants can be par-tially inhibited by treatment of test macrophages with theanti-TNFRi mAb [19]. This implies that either TNF-a waspresent in such supernatants as a component of the MAFactivity or that it was released from MAF-stimulatedmacrophages.

To investigate these observations further, the presentstudy was undertaken to confirm whether trout macro-phages can be triggered to release factors with MAF activ-ity, following incubation with a variety of stimulatorysignals. In addition, the ability of anti-TNFR1 mAb toinhibit the activities of such supernatants was studied toelucidate, indirectly, whether fish macrophage-derivedMAY (m-MAF) activity involves a TNF-a-like molecule.

Abbreviations: Con A, concanavalin A; FCS, fetal calf serum; HBSS,

Hanks’ balanced salt solution; IFN-y, interferon-i 1-MAF, leukocyte-de-

rived MAF; LPS, lipopolysaccharide; mAb, monoclonal antibody; MAF,

macrophage-activating factor; m-MAF, macrophage-derived MAF; PMA,

phorbol myristate acetate; TCF-f31 ,transforming growth factor �i; TNF-a,

tumor necrosis factor a; TNFR1, TNF.a receptor.Reprint requests: Christopher J. Secombes, Department of Zoology,

University ofAberdeen, Tillydrone Avenue, Aberdeen, AB9 2TN, United

Kingdom.

Seon I. Jang’s present address: Department of Aquaculture, National

Fisheries University of Pusan, Republic of Korea.Received May 2, 1994; acceptedJanuary 25, 1995.

LPS (�tg/ml)

944 Journal of Leukocyte Biology Volume 57, June 1995

MATERIALS AND METHODS

Fish

Rainbow trout Oncorhynchus mykiss, weighing 300-500 g, were obtainedfrom College Mill Trout Farm, Almond Bank, Perthshire. They werekept at 14*C and fed twice daily on a commercial pelleted diet. Fishwere acclimatized to the aquarium system for at least 2 weeks prior to

use.

Isolation of head kidney macrophages

Phagocyte-ennched head kidney cell suspensions were obtained as de-scribed previously [20]. Briefly, head kidney leukocyte suspensions pre-

pared in Leibovitz medium (L15, Gibco) were separated on a 34%/51%Percoll density gradient, and the phagocyte-enriched fraction at theinterface was collected. These cells were washed twice in L15, adjustedto I x io� viable cells/ml L15 plus 0.1% fetal calf serum (FCS, Gibco),and either 100 p1 was added to wells of a 96-well tissue culture plate(Nunc) or I ml was added to wells of a 24-well plate (Nunc), in triplicateper treatment. Following an adherence period (3 h at 18’C), the cellswere washed and incubated in L15 medium plus 5% FCS overnight at18C. Macrophages were washed vigorously with L15 medium threetimes prior to use, leaving approximately 20% of the cell number in thewells and giving a macrophage purity of 90-95%.

Production of macrophage supernatants

Three types of treatment were used to stimulate head kidney macro-phages; Escherichia coli (serotype 01 1 1 :B4) lipopolysaccharide (LPS,

Sigma), recombinant human tumor necrosis factor a (TNF-a, BritishBiotechnology) and supernatants from mitogen-stimulated trout head

kidney .leukocytes. The trout leukocyte supernatants could increasemacrophage respiratocy burst activity and thus were deemed to containleukocyte-derived MAF (1-MAF) activity [3].

The concentrations of LPS that were stimulatory for trout macro-phages were first determined on macrophage monolayers in 96-wellplates from five fish. After washing the cells three times with L15medium, 100 �il of medium containing 0.1-100 jig LPS/ml L15 plus 5%FCS was added to triplicate wells per concentration, and their respira-

tory burst activity was analyzed (as described below) following a 24- or48-h incubation at 18C. Subsequently, LPS was used only at a concen-

tration of 50 pg/ml to stimulate macrophages. Concentrations of TNF-aand 1-MAY supernatants deemed to have optimal stimulatory effects ontrout macrophages have been determined previously [19] and were 25IU/ml and a dilution of 1:4, respectively.

In experiments to investigate whether trout macrophages could re-lease factors with MAY activity, macrophages in 24-well plates, from fourto five fish, were stimulated with MAF, LPS, and ThF-a alone, or invarious combinations (at the above concentrations), in triplicate per

treatment. Macrophages incubated with medium were used to generatecontrol supernatants. The macrophages were then cultured for 3, 6, and12 h at 18’C before being washed five times with Li5 medium warmed

to 18’C and cultured for a further 24 h in the absence of exogenousstimuli. Finally, the supernatants from these macrophages were col-lected after centrifugation of the culture medium at 400g for 30 mm at4’C and stored at -70C prior to testing for m-MAF activity.

In the same manner, generated m-MAF supernatants deemed to bethe most stimulatory for “target” macrophages (see below) were diluted1:4 and used in combination with l-MAF supernatants to stimulate

macrophages for 12 h (i.e., to give a potential mixture of fish lympho-cyte and macrophage cytokines). The macrophages were then washedand supernatants harvested 24 h later as described above. Finally, in oneset of such experiments a chicken antiporcine transforming growthfactor I3� (TGF-�31) gamma globulin preparation (British Biotechnology)at 2 �tg/ml was added together with the above supernatants to generatem-MAF, because TGF4I1 is a potent macrophage-deactivating factor that

could be present in the 1-MAF supernatants. Indeed, studies have shownthat fish macrophages can respond to natural bovine TGF-�1 and thatthis responsiveness is inhibited by preincubation of the TCF431 with theabove antibody [21]. The anti-TCF-�1 gamma globulin preparation had

no effect on macrophage respiratory burst activity in the absence of

TGF-�31.

Detection of MAF activity

m-MAF activity was assessed by incubating test supernatants with targetmacrophages in a 96-well plate, prior to determination of their respira-tory burst activity. Supernatants were diluted from 1:2 to 1:32 in L15medium plus 5% FCS and 100-�tl aliquots added to washed macrophages

(as for LPS above) in triplicate for 24 h at 18C. In some experiments,two anti-TNFRI mAbs, 5R2 and 5R16 (Celltech), were preincubated

with the target macrophages for 1 h, at 6 and 0.45 pg/mI, respectively,before addition of the macrophage supernatants. These mAbs havetheir activity ablated in the presence of soluble TNFR1 (personal obser-vation) and are able to neutralize the responsiveness of trout macro-

phages to human TNF-a at the above concentrations [19]. Other mAbsused alongside 5R2 and 5R16 have no effect on trout macrophagerespiratory burst activity [19].

Macrophage respiratory burst activity was assessed via the reduction

of ferricytochrome c by released superoxide anion (02), following stimu-

lation of the cells with phorbol myristate acetate (PMA, Sigma) [20].Macrophages were washed twice in phenol red-free Hanks’ balancedsalt solution (HBSS, Gibco) before 100 �tl of 2 mg/ml fen-icytochromec plus 1 �tg/ml PMA in phenol red-free HBSS was added to triplicatewells per treatment. Macrophages incubated with PMA and superoxide

dismutase (Sigma), at 300 IU/ml, were used as the blank. Optical den-sity values were taken at 550 nm after 30 mm, on a multiscan spectro-photometer (MDC), and converted to nmole 02 according to Pick [22].Data were analyzed by one-way analysis of variance and Student’s t-test.

RESULTS

As can be seen in Figure 1, incubation of head kidneymacrophages with LPS had a significant impact on theirrespiratory burst activity. Analysis of variance demon-strated a significant dose effect (P < .01) after both 24 and48 h of culture in the presence of LPS, with maximalincreases seen using 50 �tg/ml for 24 h. At both incuba-tion times higher concentrations of LPS (100 jig/ml) gavesignificantly lower (P < .05) respiratory burst activity rela-tive to levels seen using 50 j.tg/ml, although the nanomo-les 02 produced were still significantly (P < .05) above

background levels in the absence of LPS.Incubation of target macrophages for 24 h with I :4

diluted macrophage supernatants from control cultureshad a small but significant (P < .05) impact on respiratoryburst activity compared with target macrophages incu-bated with medium alone (Figs. 2-5). Incubation of targetmacrophages for 24 h with 1:4 diluted supernatants frommacrophages stimulated for 3 and 6 h with LPS, TNF-aand 1-MAF alone and in various combinations had nosignificant impact on respiratory burst activity relative tocontrol supernatants (Fig. 2). However, incubation with1:4 diluted supernatants prepared by stimulating macro-

Fig. 1. Respiratory burst activity of rainbow trout head kidney macro-

phages incubated with varying concentrations of E. coli LPS for 24 and48 h prior to stimulation with PMA for 30 mm. Data are means + SE for

five fish.

E� � U.0 �0 (1)0 LL� LLCJ) LL�SC1)

�0 � <� �C c�C <� <a. <,�o.� g� � � _J.� � �� Q� + +

0.

U)

I-MAF + TNF-cz + LPS

1:8

Dilution

fang et al. Elevation of trout macrophage respiratory burst activity 945

0

0a

�00

0.

0U)0

0EC

03h�6h

U 12 h

Fig. 2. Respiratory burst activity of rainbow trout head kidney (target)

macrophages incubated for 24 h with medium alone or 1:4 diluted

macrophage supernatants prior to stimulation with PMA for 30 mm. The

macrophage supernatants were prepared by incubating head kidney

macrophages for 3, 6, or 12 h with medium (control supernatant) or

various combinations of 1:4 diluted 1-MAF, 25 lU/mI TNF-a, and 50

pg/mI LPS. Following this incubation period the macrophages were

washed and the supernatants harvested 24 h later. The data are means +

SE for four fish. *J� < #{149}#{216}5,**�P < #{216}�, and ***� < .001 compared with the

respiratory burst activity of macrophages incubated with control super-

natants.

phages for 12 h did significantly increase (P < .05) therespiratory burst activity of target macrophages. In gen-eral, supernatants from macrophages given combinations

of the different treatments had the greatest impact onrespiratory burst activity relative to control supernatants,as evidenced by higher respiratory burst activity. In addi-tion, with the exception of l-MAF versus 1-MAF + LPS, allpaired comparisons between single and combined treat-

ments were significantly (P < .05) different. Furthermore,dilution of the supernatants revealed that those frommacrophages given combined treatments retained theirability to elevate significantly respiratory burst activity oftarget macrophages beyond that of supernatants frommacrophages given a single treatment (Fig. 3). However,the use of all three treatments to stimulate macrophagesdid not generate supernatants with more m-MAF activitycompared with those from macrophages given two treat-ments (Figs. 2 and 3).

Prior exposure of target macrophages with anti-TNFR1

mAb significantly reduced the ability of supernatantsfrom macrophages stimulated for 12 h with the combinedtreatments to elevate macrophage respiratory burst activ-ity (Fig. 4). Anti-TNFR1 exposure did not result in signifi-cant reductions in activity using supernatants frommacrophages given single treatments, despite a similartrend, due to the low or insignificant activity of suchsupernatants.

Supernatants from macrophages stimulated with com-binations of trout 1-MAF- and m-MAF-containing super-natants also possessed m-MAF activity. Thus, they wereable to increase significantly the production of 02 bytarget macrophages above that of macrophages incubatedwith control supernatants (Fig. 5). Indeed, these super-natants were more active than supernatants from macro-phages stimulated with 1-MAF alone used in the sameexperiment (P < .05), and as above only the supernatants

generated using the combined supernatants were signifi-cantly affected by prior treatment of the target macro-phages with anti-TNFR1 mAb. The addition ofanti-TGF-�1 serum when generating these supernatantsalso had a significant impact (P < .05) on their ability to

elevate macrophage respiratory burst activity (Fig. 5). The

8

0

�000

�000.4

‘C”0U)00

EC

6

2

0

g Medium alone

U Control supernatant

� I-MAF (1:4)

�J LPS(50�tg/mI)

U l-MAF + TNF-a

i;� I-MAF + LPS

Fig. 3. Respiratory burst activity of rainbow trout head kidney

(target) macrophages incubated for 24 h with medium alone

or varying dilutions ofmacrophage supernatants prior to stimu-

lation with PMA for 30 mm. The macrophage supernatants

were prepared by incubating head kidney macrophages for 12

h with medium (control supernatant) or various combinations

of 1:4 diluted 1-MAF, 25 lU/mI TNF-a, and 50 �ig/ml LPS.

Following this incubation period the macrophages were

washed and the supernatants harvested 24 h later. The data are

means + SE for five fish. *� < 05, **� < � and ***� < .001

compared with the respiratory burst activity of macrophages

incubated with control supernatants.

E0 � U.Q � g� LL.u a-C’) LL.t?CI)

� � � � .� � _�CU < � LA. 0 _j 0 < LL �<a.

� o� -� -I-.0. + +U)

0 Cells aloneDCells + 5R16I Cells + 5R2

10

7.5

5.

2.5�

E0 � u-0 +� +� +�� �CU <C U.� � � <Li. <LI� <LI-

�< �< �<� Q� � .J.� �� � �

0.(I)

946 Journal of Leukocyte Biology Volume 57, June 1995

0

�00C)

0.

0U)00EC

0 Cells aloneD Cells + 5R16

U Cells + 5R2

Fig. 4. Respiratocy burst activity of rainbow trout head kidney (target)

macrophages incubated for 24 h with medium alone or 1:4 diluted

macrophage supernatants prior to stimulation with PMA for 30 mm. The

macrophage supernatants were prepared by incubating head kidney

macrophages for 12 h with medium (control supernatant) or various

combinations of 1:4 diluted 1-MAY, 25 lU/mi TNF-a, and 50 �tg/ml LPS.

Following this incubation period the macrophages were washed and the

supernatants harvested 24 h later. In some cases the target macrophages

were preincubated with anti-TNFR1 mAb for I h at 6 �tg/ml(5R2) or 0.45

�ig/ml (5R16) before addition ofthe macrophage supernatants. The data

are means + SE for four fIsh. *J� < 05 and **� < .01 compared with the

respiratory burst activity of macrophages incubated with the respective

supernatants without anti-TNFR1 treatment.

presence of anti-TGF-�1 serum produced supernatantswith significantly more m-MAF activity (P < .05) than therespective supernatants generated without i

serum [e.g., 1-MAF + m-MAF (b) in Fig. 5].

DISCUSSION

It seems remarkably easy to induce rainbow trout macro-phages to release factors that have an autocrine effect ontheir respiratory burst activity. Although it is true thatother cells were present in the cultures in addition to themacrophages, the small percentage of contaminating leu-kocytes maximally represents 2 x io� lymphocytes perwell, and this has been shown previously to be insufficientto generate bioassay-detectable 1-MAF supernatants [7].Thus, macrophages do appear to be the source of thesefactors. A range of treatments were used to stimulate therelease of these factors, including bacterial products(LPS), a recombinant cytokine (TNF-a) known to cross-re-act with trout leukocytes [18, 19], and a trout leukocytesupernatant known to “activate” trout macrophages [3,21]. Although these treatments were effective when usedindividually, the most active supernatants were generatedusing these treatments in pairs. However, combining allthree treatments did not elevate the m-MAF activity of thegenerated supernatants further, although this could havebeen due to a limit on the maximum response possible bythe fixed number of target cells present per well. Theincrease in respiratory burst activity induced by the latter

m-MAF supernatants was at least as high as those seenwith 1-MAF, which result in increased bactericidal activityagainst the fish pathogen Aeromonas salmonicida [3].

The activity in the macrophage supernatants generatedwith combined treatments was significantly inhibited bytreatment of the target macrophages with an anti-TNFR1mAb. Thus, part of their MAF activity is potentially dueto released TNF-ct. It has been shown previously that

trout leukocytes can respond to human TNF-a [18, 19]

and that the anti-TNFR1 mAb used in the present studycould inhibit these responses [19]. Furthermore, additionof anti-TNFR1 mAb has no effect on the increased respi-ratory burst activity induced directly by Con A or LPS(personal observation), ruling out the possibility that cellsurface binding of any antibody could decrease the activ-ity of these cells. Such studies suggest that at least oneTNFR is conserved on trout leukocytes, which itself im-plies that a TNF-ct-like molecule may be released fromtrout leukocytes, although more definitive proof is clearlywarranted. That human TNF-cZ was used to generatesome of these supernatants in the present study is alsocause for caution, because it is not possible to excludeentirely the potential carryover of some human TNF-cz in

the macrophage supernatants. However, supernatantsgenerated using TNF-cz alone were clearly not very stimu-latory, nor inhibitable by anti-TNFR1 mAb, and some

combinations of treatments lacking TNF-a (e.g., 1-MAF-containing supernatants plus LPS) were as stimula-tory as those generated with TNF-cx.

TNF-cx is well known to synergize with other cytokines[23, 24], and with respect to effects on macrophages it can

Fig. 5. Respiratory burst activity of rainbow trout head kidney (target)macrophages incubated for 24 h with medium alone or I :4 diluted

macrophage supernatants prior to stimulation with PMA for 30 mm. The

macrophage supernatants were prepared by incubating head kidney

macrophages for 12 h with medium (control supernatant) or 1:4 diluted

1-MAF plus various m-MAF supernatants (a, b, c) diluted 1:4. The m-MAF

supernatants were prepared by stimulating head kidney macrophages for

12 h with 1:4 diluted l-MAF plus 25 IU/ml TNF.a (a), “a” plus 50 �tg/ml

LPS (b), or “b’ plus 2 jig/ml rabbit anti-TGF-�31 serum (c) In all cases,following the 12-h incubation period the macrophages were washed and

the supernatants harvested 24 h later. In some cases the target macro-

phages were preincubated with anti-TNFR1 mAb for 1 h at 6 �tg/ml(5R2)

or 0.45 �tg/ml (5R16) before addition of the macrophage supernatants.

The data are means + SE for five fish. *� < 05 and **� < .01 compared

with the respiratory burst activity of target macrophages incubated with

the respective supernatants without anti-TNFRI treatment.

Jang et al. Elevation of trout macrophage respiratory burst activity 947

synergize with IFN-y to increase microbicidal [14, 15] andtumoricidal activity [16, 17]. Human TNF-a has also beenshown to synergize with trout 1-MAY-containing super-natants to enhance macrophage respiratory burst activity[18]. Because these leukocyte supernatants contain IFNactivity that coelutes with MAF activity, it has been sug-gested previously that their MAF activity may be mediatedvia an IFN-y-like molecule [9] and hence the synergy withhuman TNF-a. Thus, in the present study the ability ofthese two types of trout supernatant, potentially contain-ing trout molecules functionally equivalent to TNF-cx andIFN-y, to stimulate macrophages when used in combina-tion was also investigated. Supernatants generated in thismanner certainly had a far larger stimulatory effect ontarget macrophages than supernatants generated bystimulation with 1-MAF alone, and this effect was signifi-cantly inhibited by treatment with anti-TNFR1 mAb aswith the supernatants generated using combined treat-ments discussed earlier. Whether these supernatants wereacting synergistically in the present study is not known,because the macrophage supernatants were not testedalone.

Finally, the presence of anti-TGF-f�i serum during thegeneration of macrophage supernatants significantly in-creased their MAF activity. It has been shown previouslythat leukocyte MAF supernatants often have inhibitoryeffects on macrophage activity when used at suboptimalconcentrations [7, 9], possibly due to the presence ofsuppressive factors. More recendy, it has been shown thattrout macrophages can respond to natural bovine TGF-�1[21]. Coincubation of macrophages with leukocyte MAFsupernatants plus TGF-�1 resulted in a significant de-crease in macrophage respiratory burst activity, and incu-bation of activated macrophages with TGF-�31 deactivatedthem. In mammals and birds the mature TGF431 peptideis 99-100% identical [25] but has not been isolated fromother vertebrate groups. Similarly, the mature TGF-�2peptide is �95% identical across amphibians, birds, andmammals. Even TGF-f�5, to date a TGF-� unique to am-phibians, is 75% identical to TGF-�1. So TGF43s are veryconserved and consequently polyclonal anti-TGF-�31 serashow considerable cross-species reactivity. In the presentstudy, the addition of anti-TGF-�1 serum eliminated asmall but significant suppressive effect of the super-natants on macrophage activity, and experiments are on-going to characterize trout TGF-�.

ACKNOWLEDGMENTS

This work was supported by a grant from AFRC (no.AG1/551). Many thanks to Celltech Ltd, Slough, Berks,for the generous gift of the anti-TNFR1 mAb.

REFERENCES

1. Secombes, C.J., Fletcher, T.C. (1992) The role of phagocytes in theprotective mechanisms offish. Annu. Rev. Fish Dis. 2, 53-71.

2. Secombes, C.J. (1994) Macrophage activation in fish. Modulators FishImmune Responses 1, 49-57.

3. Craham, S., Secombes, C.J. (1988) The production ofa macrophage

activating factor from rainbow trout Salmo gairdneri leucocytes.Immunology 65, 293-297.

4. Marsden, M.J., Cox, D., Secombes, C.J. (1994) Antigen-inducedrelease of macrophage activating factor from rainbow trout On-corhynchus mykiss leucocytes. Vet. Immunol. Immunopathol. 42,199-208.

5. Hardie, L.J., fletcher, T.C., Secombes, C.J. (1994) Effect of tempera-ture on macrophage activation and the production of macrophageactivating factor by rainbow trout (Oncorhynchus mykiss) leucocytes.Dev. Comp. Immunol. 18, 57-66.

6. Clem, L.W., Miller, N.W., Bly, J.E. (1991) Evolution of lymphocytesubpopulations, their interactions, and temperature sensitivities. InPhylogenesis oflmmune Functions (C. Warr and N. Cohen, eds) CRC

Press, Boca Raton, FL, 191-213.7. Craham, S., Secombes, C.J. (1990) Cellular requirements for lym-

phokine secretion by rainbow trout Salmo gairdneri leucocytes. Dcv.

Comp. ImmunoL 14, 59-68.8. Trinchieri, C., Perussia, B. (1985) Immune interferon: a pleiotropic

lymphokine with multiple effects. Immunol. Today 6, 131-136.9. Craham, S., Secombes, C.J. (1990) Do fish lymphocytes secrete

interferon-’y.J. Fish BioL 36, 563-573.10. Adams, D.O., Hamilton, T.A. (1992) Molecular basis of macrophage

activation: Diversity and its origins. In The Mac.rciphage (C.E. LewisandJ.O.’D. McCee, eds) IRL Press, New York, 75-114.

11. Manogue, K.R., van Denter, S.J.H., Cerami, A. (1991) Tumournecrosis factor alpha or cachectin. In The Cytokine Handbook (A.Thomson, ed) Academic Press, London, 241-256.

12. Sheehan, K.C.F., Schreiber, R.D. (1992) The synergy and antago-nism ofinterferon-yand TNF. In TumorNecrosis Faaors: The Moleculesand TheirEmergingRole in Medicine(B. Beutler, ed) Raven Press, NewYork, 145-178.

13. Theodos, C.M., Povinelli, L, Molma, R., Sherry, B., Titus, R.C.

(1991) Role of tumor necrosis factor in macrophage leishmanicidalactivity in vitro and resistance to cutaneous leishmaniasis in vivo.Inftct. Immun. 59, 2839-2842.

14. Liew, F.Y., Millott, S. (1990) Tumor necrosis factor-alpha synergizeswith IFN-gamma in mediating killing of Leishmania major throughthe induction ofnitric oxide.J. ImmunoL 145, 4306-4310.

15. Bogdan, C., Moll, H., Solbach, W., Rollinghoff, M. (1990) Tumornecrosis factor-alpha in combination with interferon-gamma, but notwith interleukin 4 activates murine macrophages for elimination ofLeishmania major amastigotes. Eur.J. Immunol. 20, 1131-1135.

16. Chen, L., Suzuki, Y., Wheelock, E.F. (1987) Interferon-y synergizeswith tumor necrosis factor and with interleukin 1 and requires the

presence of both monokines to induce antitumor cytotoxic activity

in macrophages.j ImmunoL 139, 4096-4101.17. Sibley, L.D., Adams, LB., Fukutomi, Y., Krahenbuhl, J.L. (1991)

Tumor necrosis factor-a triggers antitoxoplasmal activity of IFN-yprimed macrophages.j ImmunoL 147, 2340-2345.

18. Hardie, L.J., Chappell, L.H., Secombes, C.J. (1994) Human tumor

necrosis factor a influences rainbow trout Oncorhynchus mykiss len-cocyte responses. Vet. ImmunoL ImmunopathoL 40, 73-84.

19. Jang, 5.1., Mulero, V., Hardie, L.J., Secombes, C.J. (1995) Inhibitionofrainbow trout phagocyte responsiveness to human tumor necrosisfactor a(hTNFa) with monoclonal antibodies to the hTNFa 55 kDareceptor. Fish Shellfish ImmunoL5, 61-69.

20. Secombes, C.J. (1990) Isolation ofsalmonid macrophages and analy-sis of their killing activity. In Techniques in Fish Immunology, VoL 1(J.S. Stolen, T.C. fletcher, D.P. Anderson, B.S. Robertson, and W.B.van Muiswinkel, eds) SOS Publications, Fair Haven, NJ, 137-154.

21. Jang, 5.1., Hardie, L.J., Secombes, C.J. (1994) The effects of trans-forming growth factor � on rainbow trout Oncorhynchus mykissmacrophage respiratory burst activity. Dcv. Comp. ImmunoL 18,315-323.

22. Pick, E. (1986) Microassays for superoxide and hydrogen peroxideproduction and nitroblue tetrazolium reduction using an enzymeimmunoassay microplate reader. Methods EnzymoL 132, 407-421.

23. Vilcek, J., Le, J. (1994) Immunology of cytokines: an introduction.In The Cytokine Handbook 2nd edition (A. Thomson, ed) Academic

Press, London, 1-19.24. Munoz-Fernandez, M.A., Fernandez, M.A., Fresno, M. (1992) Syner-

gism between tumor necrosis factor.a and interferon-y on macro-

phage activation for the killing of intracellular Ttypanosoma cruzithrough a nitric oxide-dependent mechanism. Eur.J. ImmunoL 22,301-307.

25. Roberts, A.B., Sporn, M.B. (1990) The transforming growth factor-fls. In Peptide Growth Factors and Their Receptors Vol. I (M.B. Spornand A.B. Roberts, eds) Springer-Verlag, Berlin, 419-472.