Anti-viral cytotoxic cells in the channel catfish (Ictalurus punctatus)

Developmentaland ComparativeImmunology,Vol. 20, No. 2, pp. 115-127,1996 Copyright0 1996ElsevierScienceLtd. All rights werved Printed in Great Britain 0145-305x/% s15.OO+o.OO

0145-305x(95)00043-7

ANTI-VIRAL

CYTOTOXIC CELLS IN THE CHANNEL (lctalurus punctatus)

CATFISH

Robert J. Hogan, Tor 6. Stuge, L. William Clem, Norman W. Miller, V. Gregory Chinchar* Department

of Microbiology,

University of Mississippi Medical Center, Jackson,

and

MS, 39216, U.S.A.

(Submitted August 1995; Accepted November 1995)

qlAbstract-Cytotoxlc cells isolated from the head kidney and peripheral blood of the channel catfish appear to represent distinct subpopulations of effector cells. Previous studies showed that the former lyse xenogeneic natural killer (NK) cell targets, whereas the latter preferentially lyse allogeneic cells. Here we extend these studies and present data suggesting a third class of cytotoxic effecters responsible for killing virusinfected allogeneic and autologous cells. Peripheral blood leukocytes (,PBLs) freshly isolated from unimmunixed catfish lyse uninfected allogeneic target cells as well as virus-infected allogeneic and autologous cells. Cell depletion and unlabeled (“cold”) target inhibition studies discriminated between putative effector classes and supported the view that at least two populations of cytotoxic cells are present within peripheral blood leukocytes. One population lyses allogeneic targets, whereas a second population kills channel catfish virus (CCV)infected cells. In addition, inhibitor studies demonstrated that early virus gene products are sufIicient to render infected cells susceptible to lysis. These results suggest that channel catfish possess distinct populations of NK-like, PBL-derived cytotoxic cells capable of lysing allogeneic and virusinfected target cells. Copyright 0 1996 Elsevier Science Ltd qlKeywords-Channel catfish; Channel cat&h virus; Natural killer cells; NK cells; Allogeneic target cells; Viral immunology; Cell-mediated immunity; Cytotoxic cells; Ictaluruspunctatus.

Address correspondence

to V. G. Chinchar.

Nomenclature CCV CTL NK PBLs mAb MO1 NCC ACG HK ADCC LCL

channel catfish virus cytotoxic T lymphocyte natural killer cell peripheral blood leukocytes monoclonal antibody multiplicity of infection nonspecific cytotoxic cell acycloguanosine head kidney antibody-dependent cell-mediated cytotoxicity lymphoid cell line

Introduction In mammalian

systems, cell-mediated immunity plays a major role in resolving

viral infections. NK cells lyse virusinfected cells in a non-MHC restricted fashion and are thought to provide a first line of defense by slowing virus replication before antigen-specific responses are generated. Subsequently, following a period of clonal expansion, antigenspecific, MHC class I-restricted, CD8+ cytotoxic T lymphocytes (CTLs) and virus-specific antibodies act to resolve the current infection and protect against reinfection (l-3). The mechanisms by which CTLs recognize and kill virus-infected targets have been, for the most part, successfully elucidated. Through the T cell receptor, CDS’ cytotoxic cells bind endogeneously processed viral antigens associated with 115

116 MHC class I molecules on the target cell surface. Target cell recognition initiates CTL-mediated cytolysis through necrotic and/or apoptotic mechanisms (l-3). In contrast, the molecular mechanisms by which NK cells recognize target cells and mediate lysis have not been conclusively identified. Recent studies have associated the surface expression of HLA class I molecules with resistance to NK-m~iated lysis and suggest that the absence or alteration of class I molecules on the target cell surface triggers NK-mediated lysis (4). Since virus infection may downregulate surface expression of MHC class I molecules or replace cellular peptides within the class I antigen binding site with viral products, virus-infected cells are also not recognized as “self” and are lysed by NK effecters ($6). Compared to ma~a~an systems, very little is known about teleost ~~-media~d i~unity. Cytotoxic effector cells have been demonstrated in several fish species including rainbow trout (7,8), carp (9,10), damselfish (1 l), tilapia (12) and channel catfish (13-16). In the most extensively studied system, channel catfish head kidney (HK)-derived non-specific cytotoxic cells (NCCs) were shown to lyse a variety of murine and human natural killer cell targets as well as some parasites (14,17). In contrast, recent studies demonstrated that catfish peripheral blood leukocytes contain cytotoxic effector cells that preferentially lyse allogeneic target cells and are not inhibited by a monoclonal antibody targeted against catfish NCCs (15,16). These results suggest that HK-derived NCCs and PBL-derived effecters may represent distinct subpopulations of naturally occurring catfish cytotoxic cells. Despite compelling evidence that effector cells from fish lyse xenogeneic and allogeneic targets, there is little evidence that teleosts possess effector cells capable of killing virus-infected cells. In this study, PBL-mediated cytotoxicity against channel catfish virus

R.J. Hogan et al. (CCV)-infected allogeneic and autologous cell lines was demonstrated. Furthermore, the results of unlabeled (“cold”) target inhibition studies and cell depletion assays suggest that two populations of PBL-derived, cytotoxic effecters exist: one that lyses allogeneic cells, and another that lyses virus-infected targets. Based on these observations and comparison to m~malian systems, it is likely that catfish PBL-derived cytotoxic cells represent subpopulations of NK-like cells distinct from NCCs.

Materials and Methods

Animals Channel catfish weighing approximately l-2 kg were obtained from ConAgra (Isola, MS) and maintained in a recirculating filtration system as previously described (18).

Cell Lines Long-term catfish lymphoid cell lines were established spontaneously (line 28s) or following treatment of PBLs with phorbol ester and calcium ionophore [Iines F13G3.1 and 64M2 (19,20)]. A cloned catfish B cell line (lG8) was generated following stimulation of antibody-selected B cells with lipopolysaccharide (21). Cell lines 1GS and 64M2 supported CCV replication and were used as targets as described below. Catfish lymphoid cell lines (LCLs) were maintained in AL medium (22) containing 5% pooled catfish serum (AL5).

Propagation and titration of CCV (Auburn 1, clone A, American Type Culture Collection) were performed using channel cat&h ovary (CCO) cells

117

Catfish cytotoxic cells

(23). CC0 cells, obtained from Dr Larry Hanson (College of Veterinary Medicine Mississippi State University) were maintained in Eagle’s minimum essential medium containing Hank’s salts and supplemented with 5% fetal bovine serum. When required, CCV was inactivated by ultraviolet (UV) irradiation at 254 nm using a Bio-Rad GS Gene-Linker at a setting of 150 mJ. This level of irradiation was sufficient to reduce the virus titer by greater than 99.9% (VG Chinchar and EG Chinchar, unpublished observation).

Efector

Cell Preparation

Catfish were anesthetized with MS-222 (tricaine-methanesulfonate, Crescent Research Chemicals) and bled from the caudal sinus into heparinized Vacutainer tubes (Becton-Dickinson). PBLs were isolated by layering diluted blood over either Lymphoprep (Accurate Chemical Co.) or 51% Percoll and centrifuging at 400 x g for 25 min. PBLs were removed from the interface, washed once, and with enumerated a hemocytometer (20,24).

Cytotoxic Assays Target cells (2 x lo6 cells in 100 uL AL5) were mock-infected (i.e. uninfected) or infected with CCV at multiplicities of infection (MOI) ranging from 10-20 PFU/cell and simultaneously labeled with 100 pCi Na$?r]O, (100-500 mCi/ mg Cr, Amersham) for 1.5-2 h at 28°C in a humidified atmosphere containing 5% C0~95% air. After labeling, target cells were washed in catfish RPMI-1640 (21) and resuspended in AL5. Where indicated, a lo-fold excess of unlabeled (“cold”) inhibitors was added to radiolabeled targets to assess the specificity of lysis. PBLs, isolated from individual fish, were added to triplicate wells of 96-well

round bottom plates at various effector to target (E:T) ratios. Subsequently, 5 x lo4 “0-labeled target cells were added to each well and the plates centrifuged at 200 x g for 1 min. Cultures were incubated at 28°C for 4 h, and, following a second centrifugation (200 x g, 5 mm), 100 uL of cell-free tissue culture medium was harvested and assayed for ‘iCr release by gamma counting (15,16). Percent specific release was calculated using the equation: % Specific release = Experimental cpm - Spontaneous cpm x 1oo Maximum cpm - Spontaneous cpm Reproducibility among triplicate samples was high and calculated values of specific lysis showed standard deviations of less than 5%. Results were depicted graphically by plotting percent specific lysis versus the effector-to-target ratio, and the data analyzed by multiple comparison test (Student-Newman-Keuls) using the SigmaStat program (Jandel Scientific Software).

[35S] Methionine Labeling To monitor virus infection and to ensure that cells used in the cytotoxic assay were productively-infected with CCV, replicate cultures were pulse labeled in vivo with 30 uCi/mL [35S]methionine (1000 Ci/mmol, Amersham) from 4-6 h post-infection. Subsequently, the labeling medium was removed, the cells lysed in sample buffer containing SDS and 2mercaptoethanol, and the samples denatured by boiling. Protein samples, representing equal numbers of cells, were separated by electrophoresis on 8% SDS-polyacrylamide gels and radiolabeled polypeptides visualized by autoradiography (23). Because approximately 2 h were required to prepare CCVinfected, “Cr-labeled target cells, pulselabeling replicate cultures from 4 to 6 h

118

after infection provides a reasonable approximation of viral biosynthetic activity in 51Cr-labeled targets immediately prior to assay.

R. J. Hogan et al.

and the synthesis of late CCV genes (27, unpublished observations).

Results Removal of Alloreactive Effector Cells To remove effector cells capable of reacting with allogeneic targets, a depletion assay was performed. lG8 cells (4 x 10s cells) were incubated for 20 min on ice with 600 uL of hybridoma tissue culture supematant containing murine monoclonal antibodies (mAb) lH12 (25) and 13CS (26) which bind 75% and > 98% of lG8 cells, respectively (data not shown). Antibody-coated lG8 cells were washed three times with catfish RPM1 to remove unbound antibody and mixed with 4 x lo7 freshly isolated PI3L.s. To facilitate conjugate formation between PBL effecters and allogeneic targets, the PBLflG8 mixture was centrifuged at 200 x g for 8 min, the supernatant decanted, and the culture incubated on ice in 200 pL catfish RPMI. After 30 min, 400 ~.LL of catfish RPM1 containing 4 x lo8 goat anti-mouse immunoglobulin conjugated Dynabeads (Dynal) was added and the PBL/l GSlDynabead suspension gently mixed. Following a 20 min incubation on ice, 4 mL cat&h RPM1 was added and PBL/lGS/ Dynabead conjugates were removed magnetically. Unconjugated PBLs were decanted, counted, and used as effecters in subsequent cytotoxic assays.

PBL-derived Eflector Cells Lyse Allogeneic, but not Autologous Targets Previous studies demonst~ted that catfish PBLs efficiently lysed allogeneic target cells (15,16). To confirm this finding and to determine whether autologous cells were also killed by cytotoxic effecters, lysis of “Cr-labeled allogeneic and autologous targets by catfish PBL-derived cytotoxic effecters was measured. In a representative marked lysis of two experiment, independently-derived allogeneic catfish l~phoid cell lines (lG8 and FI3L-3.1) was observed, whereas an autologous cell line (28s) was not lysed (Fig. IA). Addition of either the same (homologous) or different (heterologous) unlabeled allogeneic targets markedly reduced PBLmediated lysis of “Cr-labeled allogeneic target cells whereas the presence of a lofold excess of unlabeled autologous cells had no effect on killing (Fig. 1B). The inability of a IO-fold excess of unlabeled 28s cells to inhibit killing of radiolabeled allogeneic targets indicated that inhibition by either heterologo~ or homologous “cold” targets was not due to the physical blocking of “0-labeled targets by unlabeled inhibitors. These results indicate that PBL effecters distinguished between allogeneic and autologous cells, but not between cells with presumably different haplotypes.

Treatment of Infected Cells with Acycloguanosirte To block late virus gene expression, the

caffish lymphoid cell line lG8 was infected with CCV and maintained in media containing 1 pg/mL acycloguanosine. This concentration of drug has been shown to block viral DNA replication

PBL-mediated Lysis of Allogeneic Cells is Enhanced after CCV-infection To determine whether PBL-mediated lysis of allogeneic targets is increased following infection with 0.3, IG8 cells, a cloned catfish B cell line which supports

119

Catfish cytotoxic ceils

0

5

10

15

20

25

Effector:Target Ratio

100 90 60 .;

70

$

60

B

50

Y g

40

s

30

m

No inhlbltors

0

10X 1G8

@SSS iOXF13L31, 10x 28s

20

CCV replication, were either mock- or WV-infected and simultaneously labeled with Na2[51Cr]04 for 1.5 h. Lysis of mock- and virus-infected cells was measured in 4 h “Cr-release assays using PBLs from unimmunized fish as effecters. SDS-polyacrylamide gel analysis of replicate cultures pulse-labeled with [%]methionine contirmed that productive CCV infections occurred (data not shown). A representative %r-release assay showed that lysis of CCV-infected lG8 cells (CCV lG8) was greater than that of mock-infected lG8 cells suggesting that CCV infection increased the susceptibility of allogeneic cells to PBLmediated lysis (Pig. 2). In multiple repetitions of this experiment, percent specific lysis of virus-infected cells was 1.58 + 0.43 fold higher than in mockinfected lG8 cells. In addition, the enhanced lysis of virus-infected lG8 cells

10 0

1G5

F13L 3.1 Labeled

28s

I-*

Targets

Figure 1. Interaction of catfish PBL effecters with catfish allogeneic and autologous target cells. Panel A: Effector PBLs isolated from catfish 28 were incubated with “0-labeled allogensic cells (lG8 and F13L-3.1) or an autologous cell line derived from the same fish (28s) at the indicated effector:target ratios, and 51Cr release was assayed at 4 h. Panel 8: PBLs from fish 28 were incubated with “Cr-labeled lG8, F13L-3.1, and 28s cells at an effector to target ratio of 20: 1 in the presence or absence of the indicated unlabelled (“cold”) inhibitors at a ratio of ten “cold” targets to one 5’Cr-labeled target cell. Identities of “0-labeled targets and unlabelled inhibitors are as indicated. The asterisk (“) preceding the cell designation, in this and all other figures, indicates “Crlabeled cells. Except where noted, multiple comparison analysis using the StudentNewman-Keuls test demonstrated that the percent specific lysis of aiiotargets was significantly different from the lysis of autologous cells or allotargets in the presence of cold targets at P < 0.05. All SD < 3%. In panel B there was no significant difference between the killing of radiolabeled F13L-3.1 cells in the presence or absence of “cold” 28s Inhibitors; there were also no significant differences in the killing of radiolabeled 28s cells In the presence of a lo-fold excess of either lG8, F13L3.1, or 28s cells.

t t ~C

1

‘1G8+CC”lG8(101, ‘CC” lG8 ‘CC”lG8+1G8(101) ‘CC”lG*+CC”lG8(101),

_) 0

5

10

15

20

E”ector:T*rgat

30

25

35

40

45

Ft.tio

Flguro 2. Lysis of mock and CCV-infected lG8 cells by PBL effecters. lG8 cells, either mockinfected or infected with CCV at a MOI of 10 PFUlcell, were “0-labeled for 1.5 h and used as targets in a standard 4 h cytotoxic assay. 5’Cr-labeled, mock-infected lG8 cells are denoted by llGB; lCCV lG8 indicates “Cr-labeled, CCV-infected 168 cells. Where indicated, unlabeled mock- or CCV-infected lG8 cells were present in a lo-fold excess. With the exception of the lowest three curves, all values for percent specific lysis were statistlcally significant at P < 0.05. All SD < 5%. These results are representative of 13 independent experiments done with eight dlfferentfish.

120

was not simply a consequence of an increase in membrane permeability due to virus-infection since spontaneous release values were essentially identical in mock- and CCV-infected cells. To determine whether the same population of effector cells lyses both uninfected and CCV-infected allotargets, a IO-fold excess of unlabeled 1G8 or (XV-infected 1G8 cells was added to “Cr-labeled targets and cytolysis was monito~d in a 4 h %-release assay. If the same population of PBL effecters was responsible for lysis of both uninfected and CCV-infected lG8 cells, then a lo-fold excess of unlabeled lG8 cells should inhibit the lysis of both mockand WV-infected 51Cr-labeled target cells equally well. As shown in Figure 2, addition of a IO-fold excess of unlabeled lG8, or CCV-infected fG8 cells, reduced lysis of “Cr-labeled mock-inf~ted targets to background levels. In contrast, addition of a lo-fold excess of unlabeled, mock-infected lG8 cells only partially inhibited lysis of “Cr-labeled CCV lG8 cells, whereas greater inhibition of 51Crrelease was achieved with the addition of a IO-fold excess of unlabeled, CCV-infected targets. Taken together, these results indicate that catfish possess effector cells capable of efficiently killing uninfected and virus-infected allogeneic targets and suggest that PBLs contain at least two populations of functionally distinct cytotoxic cells-one able to kill allogeneic targets, the other capable of killing virusinfected cells. Furthermore, the results of cold target inhibition studies suggest that lysis of WV-infected allogeneic cells appears to be mediated by both alloreactive and virus-reactive cytotoxic cells.

repletion of AZZore~c~iveerector Cells does not Abrogate PBLmediated Anti-viral Cytotoxicity To strengtben the contention that two populations of cytotoxic effector cells, one

R. J.

Hogan et al.

targeted to allogeneic cells and a second to virus-infected cells, were present in the PBL fraction, catfish PBLs were depleted of alloreactive effector cells and the remaining cells assayed for cytotoxic activity against both mock-infected and CCV-infected allogeneic targets. To achieve this, lG8 cells were fitst treated with two different catfish lymphoid cellspecific murine mAbs (lH12 and 13C5) and subsequently incubated with freshly isolated PBLs. Following incubation with goat anti-mouse immunoglobulin-conjugated Dynabeads, conjugates composed of alloreactive effecters bound to lG8 cells were removed magnetically. Unfractionated PBLs and the alloreactivedepleted population were subsequently tested for their abilities to lyse mock- and CCV-infected lG8 cells in a standard 4 h “Cr-release assay (Fig. 3). Comparison of

60

Figure 3. Cytotoxic activity of PBLs depleted of alloreactive effecters. Catfish PBL effectore were isolated and either used as effecters without further treatment (unsorted) or following depletion of alloreactive effector cells. Lysis of “Cr-labeled mock- or CCVinfected 1G8 target cells was monitored in standard 4 h “Cr-release assays. Effector and target identities are shown in the insert; ARE, alloreactive effector cells. At E:T ratios > 5, values for percent specific lysis of mock- and CC%infected targets by unsorted and depleted PBLs were statisticafty sfgniffcant at P < 0.05. All SD < 5%. These results are representative of four experiments done with two different fish.

Catfish cytotoxic cells

cytotoxicity against mock- and CCVinfected lG8 cells showed that, consistent with previous observations, untreated PBLs lysed virus-infected allotargets better than mock-infected targets (i.e. 65% specific lysis of virus-infected cells compared to 45% lysis of mock-infected targets at an E:T ratio of 4011). However, following depletion of putative alloreactive effector cells, killing of mockinfected targets was almost completely inhibited, whereas lysis of virus-infected lG8 cells was only slightly reduced. These results support the view that two populations of effecters may exist among catfish PBLs and suggest that alloreactive effector cells can be selectively removed without impairing lysis of virus-infected targets.

Early Virus Gene Expression is Required for Enhanced Lysis of CCV-infected Allogeneic Cells We next determined whether full virus for gene expression was required enhanced lysis of CCV-infected allogeneic cells. To block late virus gene expression, CCV-infected lG8 cells were incubated in the presence of 1 pg/mL acycloguanosine (ACG), an inhibitor of virus-specific DNA synthesis and late gene expression (27). In other experiments designed to determine whether binding of virions to target cells was sufficient to render a cell sensitive to cytolysis, lG8 cells were infected with W-inactivated CCV. Winactivation should not prevent binding of virions to target cells, but should block productive infection and virus-specific gene expression. Lysis of mock- and CCV-infected cells treated as described above was measured in 4 h “Cr-release assays. Cells infected with W-inactivated CCV showed only a marginal increase in lysis over mock-infected allogeneic cells, whereas virus-infected cells treated with ACG showed slightly higher levels of lysis than productively-infected cells (Fig. 4). To ascertain the effectiveness of acyclo-

121

-Bt t

‘1G8 ‘CCV IGB ‘CCV 1GB + KG

7

5,

Flgure 4. Effects of ACG treatment and infection with UV-inactivated CCV on PBLmediated cytotoxicity. lG8cells were infected with wild-type CCV and “Cr-labeled in the absence or presence of 1 ug/mL ACG for 1.5 h. Subsequently the cells were washed, resuspended in AL5 containing 1 ug/mL ACG, and used as targets. In addition, lG8 cells were also simultaneously “Cr-labeled and either mock-infected or infected with UVinactivated CCV at a MOI equivalent to 10 PFUlcell. With the exception of lCCV lG8 and ‘CCV lG8+ ACG at an E:T ratio of 40, all values for percent specific lysis are significant at P < 0.05. All SD < 2%. These experiments are representative of three experiments performed with U&inactivated CCV using three different fish, and four experiments with ACG using four different fish.

guanosine treatment and W-inactivation in blocking virus gene expression, infection was monitored in replicate cultures by labeling with [35S]methionine. Analysis of viral gene products confirmed that ACG inhibited late viral protein synthesis, and that W-inactivation effectively blocked all CCV gene expression (Fig. 5). These results indicate that binding of W-inactivated virions to target cells did not result in the enhanced lysis seen following productive infection and underscore the importance of active virus replication for cell killing. Furthermore, our findings indicate that expression of early viral gene products was sufficient to render target cells susceptible to lysis.

122

R. J. Hogan et al.

CCV-infected 168

97.4 kDa 66.2 kDa

55.0 kDa

42.7 kJIa 40.0 kDa

31.0 kDa 1

2

3

4

Flgure 5. Analysis of CCV protein synthesis in target cells. 1Gfl cells were either mockinfected (lane 1) or cultured with CCV at a MOI of 10 PFU/cell under the following conditions: wild type CCV (lane 2); wild type CCV in the continuous presence of 1 pg/mL ACG (lane 3); or UV-inactivated CCV @ne 4). Cells were labeled with 30 pCi/mL [ Slmethionine from 4 to 6 h post-infection, lysed, and radiolabeled proteins analyzed on an 6% SDS-polyacrylamide gel. Molecular weight markers (in kDa) are shown to the left of the figure. A representative late CCV protein is indicated by the large arrowhead, and a major early CCV protein is shown by the small arrowhead (23).

PBL Efectors Eficiently Lyse CCV-infected Autologous Cells To support the hypothesis that catfish possess cytotoxic cells capable of killing cytolysis was cells, virus-infected monitored in an autologous system. Here, because killing of mock-infected cells is minimal, it is much easier to demonstrate lysis of virus-infected targets. Three independent catfish lymphoid cell lines, permissive for CCV infection, were generated and used as targets in 4 h “Cr-

Flguro 6. PBL-mediated lysis of CC&infected autologous cells. PBL effecters, isolated from fish 64, were incubated with 51Cr-labeled mock- or CCV-infected autologous cells, and specific lysis determined at 4 h. Where indicated, a IO-fold excess of unlabeled CCVinfected targets was added to radiolabeled CCV-infected cells. At an E: T ratio of 40, all values for percent specific lysis are significant at P < 0.05; at lower ratios there is no difference between lysis of ‘64M2 and ‘CCV 64M2 + CCV 64M2 at P < 0.05. All SD < 3%. These results are representative of three experiment performed with three different fish.

release assays with autologous PBLs as effecters. Representative data using uninfected and CCV-infected 64M2 cells as targets indicated that mock-infected cells were not lysed when autologous PBLs were used as effecters, whereas CCVinfected cells were readily killed (Fig. 6). Moreover, “Cr-release from radiolabeled, CCV-infected targets could be reduced nearly 50% by a IO-fold excess of unlabeled, virus-infected autologous cells. These results clearly indicate that under conditions where uninfected targets are spared, virus-infected autologous cells were recognized and readily lysed by PBL-derived effecters.

Discussion In this report we demonstrate catfish PBL-derived, cytotoxic discriminated between autologous

that cells and

Catfish cytotoxic cells

allogeneic targets, and between mockand virus-infected cells. Allogeneic cell lines were efficiently lysed by PBLs from unimmunized catfish, whereas autologous targets were not killed. Moreover, lysis of individual allotargets was not cell line specific as indicated by the ability of heterologous unlabeled targets to inhibit lysis. In addition, a second population of cytotoxic effecters which recognized and killed virus-infected autologous and alloTaken geneic cells was identified. together, these data suggest that catfish PBLs contained distinct populations of cytotoxic effecters that recognized and killed allotargets and virus-infected cells. The two populations of PBL-derived cytotoxic cells identified here appear to be distinct from previously described catfish NCCs by several criteria. In the first place, PBL-derived cytotoxic effecters do not react with and are not inhibited by mAb 5C6 (16), an antibody which recognizes a function-associated molecule present on the surface of catfish NCCs and mammalian NK cells and blocks their cytotoxicity (28-30). Secondly, the cytotoxic potential of NCCs is inhibited by catfish serum (31), whereas PBL-derived cytotoxic cells are not blocked (data not shown). Finally, catfish NCCs differ from PBL-derived effecters in target cell preference, i.e. NCCs preferentially lyse xenogeneic cells (e.g. mouse and human NK targets) and parasitic protozoans, whereas PBL effecters lyse allogeneic and virus-infected cells (14-17). Taken together these data suggest that channel catfish contain at least three populations of cytotoxic effector cells: head kidneyderived NCCs, and two populations of PBL-derived cytotoxic cells-one which efficiently kills allogeneic targets (15,16), and a second which appears to kill virusinfected cells. Currently the identities of catfish PBLderived cytotoxic cells are not known. For the reasons outlined below, we postulate that the cytolytic activity of catfish PBLs is most Iikely due to different subpopu-

123

lations of NK-like effecters. In the first place, it is unlikely that lysis of allotargets or (XV-infected cells is due to antibodydependent cell-mediated cytotoxicity (ADCC). Pooled catish serum (which was included in cytotoxic assays to maintain cell integrity and lower the level of spontaneous “Cr-release) failed in plaque reduction assays to block CCV infectivity indicating that pooled (nonimmune) catfish serum lacks neutralizing antiCCV activity. Moreover, although “Crrelease assays were generally performed in the presence of catfish serum, several conducted in the absence of serum showed no reduction in cytolytic activity, thus supporting the assertion that ADCC is not involved in target cell lysis. Finally, it is unlikely that ADCC plays a role in the killing of allogeneic cells because catfish, unless intentionally immunized with allogeneic cells, do not contain anti-all0 antibodies (32). Furthermore, it is unlikely that lysis of CCV-infected and allo-targets is mediated by an authentic MHC-restricted, antigenspec&c cytotoxic T cell response. First, because the catfish studied were from an outbred population, it is extremely rare to find fish which were matched at the MHC locus (32,33). Thus, antigen-specific, MHC-restricted killing by cytotoxic T cells would be unlikely when effector and target cells are derived from different outbred animals. Secondly, in most mammalian systems cytolysis of virus-infected (34-36) or allogeneic (37) cells is demonstrable only after stimulation and expansion of CTL precursor cells in vitro. This was clearly not the case with catfish PBL effecters since lysis of CCV-infected and allo-targets occurred directly with freshly isolated effecters and did not require expansion in vitro. Thus, our results with CCV, a fish herpesvirus, are consistent with those in mammalian systems where NK cells efficiently lyse herpesvirus-infected cells in an MHC-unrestricted manner without the requirement for in vitro stimulation or expansion

124

(3g-40). Furthermore in both teleost (the present data) and mammalian (38,39) systems, susceptibility of herpesvirusinfected cells to lysis by cytotoxic cells requires only early virus gene expression and does not occur when target cells are infected with UV-inactivated virus. In view of the above observations, it is likely that the cytolytic activity directed and virus-infected against allogeneic targets is most probably due to NK-like cells and not to cytotoxic T cells or ADCC. Further characterization of catfish NK-like effector cells will require the development of molecular and specific for immunological reagents different subpopulations of catfish cytotoxic cells. The work reported here is one of a few studies demonstrating lysis of virusinfected teleost cells. An earlier study by Moody et al. (41) reported enhanced lysis of infectious pancreatic necrosis virusinfected targets by kidney cells from rainbow trout (Uncorhynchus mykiss) and Atlantic salmon (Salmo salur). However, although they were able to detect lysis of virus-infected cells, they were unable to demonstrate expression of viral proteins in infected cells within the period of their assay system. Recently Yoshinaga et al. (42) showed that virusinfected RTG-2 cells were lysed by rainbow trout HK effecters, but observed marked variation in the cytotoxic activity of individual rainbow trout against both allogeneic and virus-infected targets. Here, using permissive autologous and allogeneic cell lines, we showed that PBLs, isolated from catfish that had not been previously immunized or intentionally-infected with CCV, mediated high levels of lysis of CCV-infected targets. Al~ough some variation in the extent of killing was observed between different fish, fish which failed to lyse mock- or CCV-infected allogeneic cells were not detected. Cold target inhibition studies demonstrated that lysis of CCV-infected allotargets was only partially inhibited by

I?.J. Hogan et al.

a W-fold excess of uninfected allogeneic cells suggesting that CCV-infected allogeneic cells are recognized by two distinct populations of cytotoxic cells-one targeted to allogeneic cells, and another directed against virus-infected cells. Partial inhibition by a lo-fold excess of uninfected cells is explicable if one assmnes that only the alloreactive population of cytotoxic effecters is inhibited, leaving vim-specific effecters capable of killing. Alternatively, differences in killing may reflect differences in the activation state or differentiation level of putative NK cells rather than the activity of different cell populations. At present, we are unable to distinguish among these alternative hypotheses. Although a number of models have been suggested, the precise mechanisms by which NK cells recognize target cells have not yet been elucidated. The ability of both homologous and heterologous unlabeled allotargets to inhibit killing of “0-labeled allogeneic catfish cells is consistent with the “missing self” hypothesis i.e. failure to recognize or detect self MHC Class I molecules triggers lysis (4). However, the inability of uninfected allogeneic cells to block completely the lysis of “Cr-labeled CCV-infected allogeneic cells does not fit this model. Vimsinfected cells are thought to be susceptible to NK-mediated lysis either because virus+nfection down-regulates cell surface expression of MHC (i.e. “missing self”) or because viral peptides replace self peptides within class I molecules [i.e. “altered self” (6,7)]. If recognition of altered or missing self MHC were the sole determinant of lysis, then a IO-fold excess of unlabeled, uninfected allogeneic cells should inhibit lysis of labeled, virusinfected target cells by about 90%. That only a partial inhibition of lysis occurred, suggests that recognition of virus-infected cells may involve factors other than the absence or alteration of class I molecules. Interestingly, our findings with catish NK-like cells are similar to those seen in

125

Catfish cytotoxic cells

mammalian NK systems in which individual NK clones showed different patterns of cytotoxi~ity toward uninfected, virus-infected, and tumor cell targets (43). In addition, the results of cold target inhibition studies suggested that NK cells targeted to virus-infected cells may be different than those directed against tumor cells (40). Thus NK mediated lysis of virus-infected cells may occur via a “multiple choice” model in which MHC is only one of several features controlling cell killing (39). Finally, two unique features of cytolysis mediated by channel catfish PBLs deserve further note. The ability to generate lymphoid cell lines from individual channel catfish and the permissiveness of these lines for CCV replication permits, for the first time in teleosts, cytotoxic experiments to be performed in an autologous environment. Moreover, an autologous system not only increases the chances of detecting killing of virusinfected cells (because lysis of mockinfected targets is reduced), but also provides an opportunity to screen for authentic cytotoxic T cells. Unfortunately, one impediment to full development of an autologous cytotoxic assay system has been the initial resistance of

newly generated LCLs to CCV infection. Pre~~na~ evidence suggested that newly es~b~sh~ LCLs secrete factors (e.g. interferon-like or tumor necrosis factorlike molecules) that inhibit CCV replication. However, since catfish lymphoid cell lines appear to become progressively more permissive with time in culture, this phenomenon does not limit the long-term usefulness of catfish LCLs as targets. A second key feature of the above cytotoxic system is its reproducibility, Because catfish PBLs, rather than HK leukocytes, are employed as effector cells, repeat expe~ments can be performed easily with a single fish reducing fish-tofish variability in the cytotoxic response. Thus catfish PBL effecters and autologous target cells may prove to be an excellent system in which to demonstrate the existence of authentic, antigenspecific, MHC-restricted cytotoxic cells in fish.

~c~u~~e~ge~e~?~~s work was supported by grants from the U.S. Department of Agriculture (92-37208-8187 and 94-372040853) and the National Institutes of Health

(R37-AI-19530). We thank Ornar Logue for technical assistance.

References 1, Janeway, C. A.; Travers, P. ~~0~010~: The immune system in health and disease. New York: Garland; 1994. 2. Whittou. J. L.: Oldstone. M. B. A. Virusinduoed‘immu& response ‘interactions: Principles of immunity and immunopathology. In: Fields, B. N.; Knipe, D. M., Eds. Virology. New York: Raven Press; 1990. 3. Zinkernanel, R. M. Immunity to viruses. In: Paul, W. E., Ed. Fundamentalimmunolog, 3rd edn. New York: Raven Press: 1993. 4. Ljunggrcn, H-G.; Karre, K.. In search of the n&sing self MHC molecules and NK cell reconnition. Immunol. Todav 11~237-244: 1990. 5. ?&&at& M. S.; Lusso, P.; &cone, E.; l&ret@ A.; Moretta, L.; Long, E. 0. Recognition of virus-infected cells by natural killer cell clones is controlled by polymorphic cell elements. J. Exp Med. 178961-969; 1993.

6. Malnati, M. S.; Pens&, M.; Parker, K. C.; Biddison, W. E.; C&one, E.; Moretta, A.; Long, E. 0. Peptide specificity in the recognition of MHC Class I by natural biller cell clones. Science 267~101~1018; 1995. 7. Greenlee, A. R.; Brown, R. A.; Ristlow, S. S. Nonspecific cytotoxic cells of rainbow trout (0ncorhynchu.s mykir~) kill Yac-1 targets by both necrotic and anontotic mechanisms, Dev. Comp. Immunol. 15:153-1&Q;1991. 8. Havden. B. J.: Laux. D. C. Cell-mediated lvsis of mu&e ‘target cells’ by non-immune sahnonid lymphoid preparations. Dev. Comp. Immunol. 9:627-639; 1985. 9. I-I&ma, S.; Abo, T.; Kumaga, IL; Hata, M. The potent activity of fresh-water fish kidney o&s in ~~-~ng: I. Characterization and s~es~s~bution of cytotoxicity. Dev. Comp. Immunol. 4~653-666; 1980.

126 10. Suzumura, I%.;Kurata, 0.; Okamoto, N.; Ikeda,

Y. Characteristics of natural killer-like cells in carp. Fish Path. 29:199-203; 1994. 11. McKinney, E. C.; S&male, M. C. Damselfish with neurofibromatosis exhibit cytotoxicity toward tumor targets. Dev. Comp. Immunol. 4305-313; 1994. 12. Faisal, M.; Ahmed, I. I.; Peters, G.; Cooper, E. L. Natural cytotoxicity of tilapia leucoytes. Dis. Aquatic Org. 7:17-22; 1989. 13. Graves, S. S.; Evans, D. L.; Cobb, D.; Dawe, D. L. Nonspecific cytotoxic cells in fish (IEtafzuus punc#utuff: I. Optimmn ~~~~~ for target cell lysis. Dev. Comp. Immunol. 8:293-302; 1984. 14. Evans, D. L.; Jaso-Friedmann, L. Nonspecific cytotoxic cells as effecters of immunity in fish. Ann. Rev. Fish Dis. 2:109-121; 1992. 15. Yoshida, S. H.; Stuge, T. B.; Miller, N. W.; Clem, L. W. Phylogeny of lymphocyte heterogeneity: Cytotoxic activity of channel catfish peripheral blood leukocytes directed against allogeneic targets. Dev. Comp. Immunol. 1971-77; 1995. 16. Stuge, T. B.; Miller, N. W.; Clem. L. W. Channel cat&h cytotoxic effector cells from peripheral blood and pronephroi are different. Fish Shellfish Immunoi. 5~469471; 1995. 17. Evans, D. L.; Graves, S. S.; Cobb, D.; Dawe, D. L. Nonspecific cytotoxic cells in fish (Ic~uZ~~ punctarus): II. Parameters of target cell lysis and specificity. Dev. Comp. Immunol. 8:303-312; 1984. 18. Miller, N. W.; Clem, L. W. Microsystem for in vitro primary and secondary immunization of channel catfish (Ictulunrs punctatus) leukocytes with hapten-carher conjugates. J. Immunol. Methods. 72367-379: 1984. 19. Lin, G. L.; Ellsaeskr, C. F.; Clem, L. W.; Miller, N. W. Phorbol ester/calcium ionophore activate fish leukocytes and induce long term cultures. Dev. Comp. Immunol. l&153-163; 1992. 20. Miller, N. W.; Chinchar, V. G.; Clem, L. W. Development of leukocyte cell lines from the channel catfish (Zetuiutk punctatus). J. Tissue Culture Methods 16:117-123; 1994. 21. Miller, N. W.; Rycyzyn, M. A.; Wilson, M. R.; Warr. G. W.: Naftel. J. P.; Clem, L. W. Development and cha;acteriz&on of channel catfish long term B cell lines. J. Immunol. 1522180-2189; 1994. 22. Luft, J. C.; Clem, L. W.; Bly, J, E. A serum-free medium for channel catfish in vitro immune responses. Fish Shellfish Immunol. 1:131-139; 1991. 23. Chinchar, V. G.; Rycyzyn, M.; Clem, L. W.; Miller, N. W. Productive infection of ~ntinuo~ lines of channel catfish leukocytes by channel cattish virus. Virology 193:989-992; 1993. 24. Miller, N. W.; McKinney, E. C. In vitro culture of fish leukocytes. In: Hochachka, P. W.; Mommsen, T. P.; Eds. Biochemistry and molecular biology of fishes. New York: Elsevier; 1994:341-353.

R. J. Hogan et at. 25. Lobb, C. J.; Olson, M. 0. J. I~~o~ob~n heavy H chain isotypes in a teleost fish. J. hnmunol. 141:12361245; 1988. 26, Bly, J. E.; Miller, N. W.; Clem, L. W. A monoclonal antibody specific for neutrophils in normal and stressed channel catfish. Dev. Comp. Imnlunol. 1421 l-221; 1990. 27. Buck, C. C.; Loh, P. C. In vitro effect of acyclovir and other antiherpetic compounds on the replication of channel catfish virus. Antiviral Res. 5:269-280; 1985. 28. Evans, D. L.; Jaso-Friedmann, L.; Smith, E. E.; St. John, A.; Koren, H. S.; Harris, D. T. Identification of a putative antigen receptor on fish nonspecific cytotoxic cells with monoclonal antibodies. J. Immunol. 141:324-332; 1988. 29. Harris, D. T.; Jaso-Friedmann, L.; Devlin, R. B.; Koren, H. S.; Evans, D. L. Identification of an evolutionarily conserved function associated molecule on human natural killer cells. Proc. Natl. Acad. Sci. USA 88:3009-3013; 1991. 30. Harris, D. T.; Camenisch, T. D.; JasoFriedmann, L.; Evans, D. L. Expression of an evolutionarily conserved function associated molecule on sheep, horse, and cattle NK cells. Vet. Immunol. r~unonathol. 38~273-282: 1993. 31. Evans, D. L.; Hogan, K. T.; Graves, S. S.; Car&on, R. L.; Floyd, E.; Dawe, D. L. Nonspecific cytotoxic cells in fish (Zctaium ptmctutur): III. Biophysical and biochemical properties affecting cytolysis. Dev. Comp. Immunol. 8:599-610; 1984. 32. Vallejo, A. N.; Miller, N. W.; Clem, L. W. Phylogeny of immune recognition: Role of alloantigens in antigen presentation in channel catfish immune responses. Immunology 74: 165168; 1991. 33. Vallejo, A. N.; Miller, N. W.; Clem, L. W. Antigen processing and presentation in teleost immune responses. Ann. Rev. Fish Dis. 273-89; 1992. 34. Torpey, D. J.; Lindsey, M. D.; Rinaldo, C. R. HLA-restricted lysis of herpes simplex virusinfected monocytes and macrophages mediated by CD4+ and CD8+ lymphocytes. I. Immunol. 1421325-1332; 1989. 35. Demkowicz, W. E.; Ennis, F. A. Vaccinia virusspecific CD8+ cytotoxic T lymphocytes in humans. J. Virol. 67~1538-1544: 1993. 36. Allen, G.; Yeargan, M.; Costa, ‘R. R.; Cross, R. Major histocompatibility complex class Irestricted cytotoxic T-lymphocyte responses in horses infected with equine herpesvirus 1. J. Virol. 69&&i-612; 1995. 37 Cerottini, J-C.; Bnmner, K. T. Cell-mediated cytotoxicity, allograft rejection, and tumor i~u~ty. A&. Immunol. 18:67-132; 1974. 38. Fitzgerald-Bocarsly, P.; Howell, D. M.; Pettera, L.; Tehrani, S.; Lopez, C. Immediate-early gene expression is sufficient for induction of natural killer cell-mediated lysis of herpes simplex virus type-l infected fibroblasts. Virology 65:31513160; 1991.

Catfish cytotoxic cells

39. Kaufman, D. S.; Schoon, R. A.; Leibson, P. J.

Role for major hi&compatibility complex class I in regulating natural killer cell-mediated killing of virus-infected cells. Proc. Natl. Acac. Sci. USA 8983378341; 1992. 40. Lopez-Guerrero, J. A.; Alarcon, B.; Fresno, M. Mechanism of recognition of heroes simplex virus type l-infect& cells by natural killer cells. J. Gen. Virol. 6928592868: 1988. 41. Moody, C. E.; Serreze, D. V.; Reno, P. W. Non-specific cytotoxicity activity of teleost

127

leukocytes. Dev. Comp. Immunol. 9:51&k 1985. 42. Yoshinaga, K.; Okamoto, N.; Kurata, 0.; Ikeda, Y. Individual variations of natural killer activity of rainbow trout leukocytes against IPN virus-infected and uninfected RTG-2 cells. Fish Pathology 29:1* 1994. 43. Mason, P. D.; Sissons, J. G. P.; Borysiewicz, L. K. Heterogeneity amongst natural killer cells revealed by limiting dilution culture: Selectivity against virus-infected and tumor cell targets. Immunology 80:625--632; 1993.