Natural cytotoxicity towards allogeneic tumour targets in Xenopus mediated by diverse splenocyte populations

Natural cytotoxicity towards allogeneic tumour targets in Xenopus mediated by diverse splenocyte populations

\ Pergamon Developmental and Comparative Immunology\ Vol[ 11\ No[ 1\ pp[ 106Ð129\ 0887 Þ 0887 Elsevier Science Ltd[ All rights reserved Printed in G...

554KB Sizes 0 Downloads 6 Views

\

Pergamon

Developmental and Comparative Immunology\ Vol[ 11\ No[ 1\ pp[ 106Ð129\ 0887 Þ 0887 Elsevier Science Ltd[ All rights reserved Printed in Great Britain 9034Ð294X:87 ,08[99¦9[99

PII] S9034Ð294X"87#99992Ð1

NATURAL CYTOTOXICITY TOWARDS ALLOGENEIC TUMOUR TARGETS IN Xenopus MEDIATED BY DIVERSE SPLENOCYTE POPULATIONS Trudy L. Horton, Pamela Ritchie, Martin D. Watson and John D. Horton Department of Biological Sciences, University of Durham, Durham, U.K. "Received October 0886^ Accepted December 0886#

Abstract*We have recently demonstrated NK!like activity in the spleen of the clawed frog\ Xenopus laevis[ This paper investigates the cellu! lar basis of this natural cytotoxicity[ Signi_cant levels of cytotoxicity towards B2B6 allogeneic thymus tumour targets\ that express neither class Ia nor class II MHC proteins\ occurred after splenocytes from either control or early!thy! mectomized "Tx# year!old Xenopus were cultured for 37 hours[ Killing by Tx cells required their culture in growth factor!rich medium "GFM# obtained from concanavalin A!stimulated cells[ Immunomagnetic cell sorting revealed that cyto! toxic effectors in both control and Tx frogs were found in the B cell!depleted population\ but never in the B cell!enriched fraction[ Splenocytes from control Xenopus\ depleted of T cells by magnetic sorting and following culture in GFM\ also developed natural cytotoxicity towards allo! tumour cells[ Magnetic cell sorting also revealed that puri_ed "CD4¦# T cells cultured for 37 hours in GFM also became able to lyse the allogeneic tumour targets[ Cytotoxicity mediated by T cells resided not only in the CD4¦\ CD7¦ population\ but also in the CD4¦\ CD7− "putative CD3¦# T cell subset[ Ontogenetic studies revealed that splenocytes from 5Ð6 week!old "stage 45Ð46# con! trol larvae\ even after 37 hr culture in GFM\ were unable to spontaneously lyse the allotumour targets\ whereas cultured splenocytes from 5

Address correspondence to Dr[ John D[ Horton\ Department of Biological Sciences\ University of Durham\ South Road\ Durham DH0 2LE\ U[K[ Tel] 9080 263 2248^ Fax] 9080 263 1306[

month old froglets were effective killers[ Thy! mocytes from larvae or adults routinely failed to kill tumour cells[ The work highlights the need to use Tx Xenopus to further explore non!T!cell! mediated\ NK!like cytotoxicity at the amphibian level of evolution[ Þ 0887 Elsevier Science Ltd[ All rights reserved Keywords*Magnetic cell sorting^ Natural cytotoxicity^ NK!like cells^ Spleen^ T cells^ Thy! mectomy^ Tumour cells^ Xenopus laevis[

Nomenclature Tx Con A NK CTL mAb NCC GFM

early!thymectomized^ concanavalin A^ natural killer^ cytotoxic T lymphocyte^ monoclonal antibody^ nonspeci_c cytotoxic cell^ growth factor!rich medium from Con A! stimulated splenocytes[

0[ Introduction Natural killer "NK# cells provide innate\ non antigen!speci_c defence against target cells that have lost their normal expression of major histocompatibility complex "MHC# class I proteins as a consequence of viral infection or metastasis "0#[ Such aberrant targets cannot be dealt with by cytotoxic T lymphocytes "CTL#\ whose clonally!distributed antigen receptors are MHC restricted[ The realisation that NK

217

218

cells provide crucial immune defence\ tog! ether with increasing awareness that these cells are important regulators of adaptive immunity "1#\ has lead to a wealth of research attempting to de_ne the molecular basis of antigen receptors and target ligands involved in the activation and inhi! bition of mammalian NK cells "2Ð5#[ Since NK cells are believed to be the evolutionary forerunners of CTL "6#\ stud! ies on lower vertebrates are likely to reveal novel animal models to explore crucial issues concerning NK cell biology[ Indeed avian studies\ that identi_ed cytoplasmic CD2 proteins in chicken NK cells "7#\ sup! port the notion "8# that T cells and NK cells derive from a common progenitor[ Work on _sh indicates evolutionary con! servation of NK cell antigen receptors] thus cat_sh non!speci_c cytotoxic cells "NCC# and human NK cell killing are blocked by monoclonal antibodies "mAbs# raised against a vimentin!like molecule on cat_sh NCC "09#[ Despite extensive research on the immune system of Xenopus "00# candidate NK cells have yet to be identi_ed in this amphibian\ although NK!like activities are suspected in other anuran species "01#[ We have recently begun to try and address this issue through the use of early!thy! mectomized Xenopus "02\ 03#[ Early thy! mectomy can be routinely performed on days 4Ð5 of larval life\ when the thymus is rudimentary "04\ 05#[ Despite displaying crippled thymus!dependent immune func! tions "04\ 05\ 06# and loss of bona _de T cells from spleen\ blood and intestine "07Ð 19#\ Tx Xenopus generally survive sur! prisingly well[ Such survival may re~ect enhanced NK activity following thymec! tomy\ since splenocyte e}ectors taken directly from some Tx adult frogs\ but not from control siblings\ show natural lytic activity towards allogeneic thymus tumour cells "as shown by use of 5 hour 40Cr release assay# "02#[ Xenopus thymus tumour cell lines are appropriate targets for candidate NK cells\ since they are de_cient in MHC antigen expression "10#[

T. L. Horton et al.

Magnetic cell sorting of mAb!labelled T and B cells is used here to examine the cellular basis of Xenopus splenocyte cyto! toxicity towards allothymus tumour cells[ Flow cytometry is employed to con_rm the purity of spleen subpopulations[ The experiments reveal that a non!T:non!B leu! cocyte population in both control and Tx frogs\ when cultured for 37 hours in GFM "active supernatant derived from ConA! stimulated cultures#\ can lyse allotumour targets[ Under the same culture conditions\ CD7¦ and putative CD3¦ T cell subsets "from control adult spleen# are shown to display signi_cant allocytotoxicity[ Unsep! arated splenocytes from control larvae cul! tured in GFM "and thymocytes from all stages of development# are unable to lyse allotumour targets[

1[ Materials and Methods 1[0[ Animals and Operations Outbred X[ laevis were reared in the lab! oratory at 1221>C as described elsewhere "04#[ Thymectomy of larvae by micro! cautery was carried out at 4Ð5 days of age\ when the thymus is at a rudimentary stage of di}erentiation "04\ 05\ 07#[ Absence of thymus in Tx larvae was con_rmed by stereo!microscopy "the thymus is normally readily apparent below the transparent lar! val skin#\ whereas thymic absence after metamorphosis was checked by dissection when the spleen was removed for analysis[ Controls were non!operated siblings[ The cytotoxicity assays were on 01 month old adult frogs\ except in the ontogenetic study\ which utilised stage 45Ð46 "5Ð6 week old# larvae "just prior to metamorphosis# and 5 month old froglets[

1[1[ Ma`netic Cell Sortin` Splenocyte suspensions were prepared from individual adults as described else! where "11#\ except that cells were kept throughout the separation procedure in

Allotumour killing in Xenopus

amphibian strength Hanks| balanced salt solution "HBSS# supplemented with 0) foetal calf serum "FCS#[ For T cell sorting\ cells were initially incubated on ice with mAb 1B0 "IgG0 anti!Xenopus CD4\ "12##[ B cells were obtained by initial incubation with a 49]49 mix of mAb 7E3 "IgG anti! Xenopus IgM "13## and a cocktail of 2 anti! Xenopus Ig light chain mAbs "14#[ CD7¦ populations were obtained using mAb AM11 "IgM anti!Xenopus CD7 "15##[ After thorough washing\ approximately 4×095 mAb!incubated cells were resus! pended in 79 ml HBSS\ and 19 ml of MACS microbeads conjugated with either goat anti!mouse IgG or rat anti!mouse IgM "Miltenyi Biotec# added prior to 04 min incubation at 3>C[ The mixture of mAb! labelled and unlabelled cells was sub! sequently passed over a prewashed Mini! MACS separation column placed in a separation unit "magnet#[ The e/uent from the column contained cells depleted of those lymphocytes that had bound the mAb and magnetic beads[ The column was then removed from the magnet and the retained mAb!stained cells eluted "{enriched| population#[ Production of 2 subsets of splenocytes i[e[ candidate cyto!

219

toxic T cells "CD4¦\ CD7¦#\ putative CD3¦ T cells "CD4¦\ CD7−# and non!T cells "CD4−# was achieved by sequential MACS separations as described in the leg! end to Fig[ 3[ After assessing the purity of separated populations by ~ow cytometry\ sorted cells were established in vitro[

1[2[ Cell Culture of Splenocytes and Thymocytes 1[2[0[ Experimental Protocol[ Splenocyte e}ectors from control and Tx frogs were only tested directly ex vivo for their capacity to lyse tumour targets in the experiment shown in Fig[ 0[ Apart from this\ the ability of splenocytes and thy! mocytes to lyse tumour targets was assessed only after these e}ectors had been pre!cultured for 37 hrs in tumour medium\ with or without growth factor sup! plementation[ Cells from individual ani! mals were always cultured separately\ any pooling of cells "required with larval lym! phocytes to give su.cient e}ector num! bers* see Fig[ 5# being carried out at the time of the chromium!release assay[ ðThis precludes mixed lymphocyte reactions

Figure 1. Cytotoxicity towards B3B7 allothymus tumour targets displayed by unsorted splenocytes from adult control and early-thymectomized (Tx) Xenopus. Splenocytes were tested directly ex vivo (0 hrs) and then again after 48 hrs in tumour culture medium alone (med) or in culture medium supplemented with 25% growth factor-rich medium (GFM) generated from Con A-stimulated cells. Data show mean specific 51Cr release2S.E. from 5×104 targets at various effector:target (E:T) ratios following 6 hr assay. Nine separate experiments were carried out with spleens from individual control frogs, 5 with individual Tx frogs.

220

from occurring\ which might promote cytotoxicity[Ł Cells were cultured at 16>C in 4) CO1\ adult cells in 0 ml aliquots in 13!well plates "½2×095 leucocytes:ml#\ larval cells in 199 ml aliquots in ~at!bottom 85!well plates "½0×095:ml#[ 1[2[1[ Tumour Medium[ 399 ml serum!free medium "as below# was diluted to amphib! ian strength with distilled water and sup! plemented with 7) supernatant from the Xenopus A5 kidney cell line "16#\ 1) decomplemented FCS "First Link# and 199 mg:ml kanamycin "Gibco#[ Serum!free medium contains 499 ml Iscove|s medium\ 4 ml 099× non!essential amino acids\ 49 mg:ml penicillin:streptomycin "all from Gibco#\ 4 mg:ml insulin "Sigma#\ 49 mmol[ mercaptoethanol "Sigma# and 0[4 ml 09) Primatone "Roche#[ 1[2[2[ Growth Factor Supplementation[ This involved addition\ at the start of culture\ of 14) growth factor!rich medium "GFM#[ The latter was derived from X[ laevis splen! ocytes stimulated for 13 hrs with an opti! mal mitogenic dose of concanavalin A "ConA#\ as described in detail elsewhere "11#[ ConA was removed from the GFM by adsorption with a!methyl mannoside and the supernatant _ltered through a 9[1 mm _lter prior to use[ Cells cultured without exogenous growth factor received instead a control volume of tumour medium[

1[3[ In Vitro Cytotoxicity Assay 1[3[0[ Tar`ets[ The Xenopus thymus tumour cell line B2B6 was the principle tar! get used[ This cell line is derived from thy! mus tumours found in the partially inbred\ MHC!homozygous Xenopus family ff\ and expresses neither classical class I "class Ia# nor class II MHC proteins "10#[ Tumour cells are maintained in tumour cell medium\ ideally at 0Ð1×095:ml to main! tain ×84) viable cells[ In some experi! ments 2!day Con A!induced allogeneic

T. L. Horton et al.

splenic lymphoblasts were also used as tar! gets[ 1[3[1[ 40Cr Release Assay[ Tumour targets "0×095:ml# in log phase growth "and allo! geneic lymphoblast targets# were incubated overnight "05 hours# with 099 mCi:ml Na1ð40CrŁ93 "199Ð499 mCi:mg Cr\ Amer! sham#[ Following three washes in tumour medium\ radiolabelled targets were incu! bated for 29 min "½0×095 cells in 1 ml medium# to remove initial spontaneous release of 40Cr\ then adjusted to 4×094 tar! gets:ml for assay[ E}ector splenocytes or thymocytes\ either ex vivo or after 37 hour culture\ were washed twice prior to serial dilution in tumour medium in 85 well U! bottom plates "Greiner#[ To each well\ con! taining 099 ml e}ectors\ was added 099 ml 40 Cr!labelled targets "4×093 cells#[ E}ec! tor]target ratios ranging from 39]0 to 4]0 were established[ Plates containing targets and e}ectors were centrifuged at 299 ` for 1 min and then incubated for 5 hours at 16>C in 4) CO1[ ðThe 40Cr!release assays were carried out at 16>C because both splenocyte e}ectors and tumour cell targets are routinely cultured in vitro at this tem! perature[ Additionally\ 16>C promotes accelerated immune reactivity compared with 12>C*the temperature that Xenopus are maintained in vivo[Ł After incubation\ the contents of each well were gently pip! etted and plates then centrifuged "299 ` for 09 min#\ prior to careful removal of 099 ml supernatant[ The latter was then prepared for scintillation counting by addition of 2 ml Ecosint scintillation solution "National Diagnostics#\ and radioactivity detected using a Packard Tri!carb analyser[ The percentage of speci_c 40Cr release was calculated from the formula "E!C:M! C×099#\ where Ecpm 40Cr for exper! imental supernatant\ Ccpm for mini! mum control "that is\ spontaneous release using 099 ml target cells cultured in 099 ml medium alone#\ and M the maximum release value "using 099 ml target cells ¦099 ml water\ freeze!thawed three times#[ Maximum 40Cr release values for 4×093

Allotumour killing in Xenopus

B2B6 targets were extremely reproducible and cpm were routinely in the range 00\999Ð04\999[ Spontaneous release for B2B6 targets never exceeded 09) of maximum release[ For ConA blasts\ spon! taneous release was around 19) of total[

1[4[ Flow Cytometry Splenocytes were routinely assessed for expression of CD4\ CD7 or IgM prior to cell culture using primary murine mAbs 1B0\ AM11 and 7E3 respectively "see {mag! netic sorting| Section 1[1 for details on these mAbs#[ The control mouse mAb was CT2\ an anti!chicken CD2 of IgG0 isotype "17#[ Single colour ~ow cytometric analysis employed FITC!conjugated rabbit anti! mouse Ig "DAKO# to detect mAb binding[ The protocol for staining unsorted splen! ocytes followed procedures given in detail elsewhere "07#[ To determine purity of enriched populations produced by mag! netic sorting\ an aliquot of separated cells was stained with secondary FITC!con! jugated anti!mouse antibody alone[ Sur! face antigen expression on depleted populations was con_rmed by incubation of these cells both with primary mAb and secondary FITC!conjugated antibody[ Gates were set by forward and side light scatter to delineate lymphoid cells and exclude dead cells from analysis\ the latter being con_rmed by propidium iodide staining[ Five to 09\999 gated cells were analysed on a Coulter XL ~ow cytometer[

2[ Results 2[0[ Cytotoxicity towards Tumour Cells induced by In Vitro Culture of Splenocytes from Control and Thymectomized Xenopus The capacity of unsorted splenocytes from either control or Tx year!old X[ laevis to lyse B2B6 targets is shown in Fig[ 0[ In these experiments both control and Tx splenocytes tested directly ex vivo "immedi!

221

ately following spleen removal# failed to lyse the tumour targets[ Following 37 hours in vitro\ control splenocytes become able to lyse B2B6 cells "½24) 40Cr release at E]T ratio of 39]0\ with a gradual decline in amount of lysis as E]T ratios declined#[ Supplementation with GFM had no appar! ent e}ect on the level of cytotoxicity achieved by control splenocytes[ In contrast\ 37 hours in vitro culture of Tx splenocytes in the absence of GFM failed to generate cytolytic cells "only ½09) speci_c 40Cr release from tumour targets at E]T ratio of 39]0#[ However\ splenocytes from the same Tx animals cultured for 37 hours in the presence of GFM were now able to kill targets e}ectively\ achieving ½49) speci_c 40Cr release at E]T ratio of 39]0[

2[1[ Effect of T and B Cell! Enrichment:Depletion on Tumour Cytotoxicity Mediated by Cultured Splenocytes from Control and Tx Xenopus 2[1[0[ Ma`netic Cell Sortin`[ This was employed to produce CD4¦ "T cell enriched# and CD4− "T cell depleted# populations from control spleens and also IgM¦:Ig light chain¦ "B cell enriched# and Ig− "B cell depleted# populations from con! trol and thymectomized spleens[ It should be noted that CD4 is a pan T cell marker in Xenopus but\ unlike the situation in some mammals\ is not constitutively expressed on Xenopus B cells "12#[ Figure 1 shows the outcome of 2 experiments demonstrating the e}ectiveness of the MiniMACS sep! aration system to procure populations depleted of T and B cells[ In experiment 0 "Fig[ 1 top# 59) T cells and 19) B cells were found in the unseparated control spleen[ Depletion of CD4¦ cells proved extremely e}ective and the depleted popu! lation showed a corresponding B cell increase[ In experiment 1 "Fig[ 1 middle# the control spleen displayed 41) T cells and 29) B cells prior to sorting with anti!

222

T. L. Horton et al.

Figure 2. Representative flow cytometric analysis of splenocytes from adult control and earlythymectomized (Tx) Xenopus to illustrate efficiency of MACS magnetic bead separation to deplete T or B lymphocytes. In experiment 1 (top), levels of surface CD5 and IgM expression are monitored on control cells before separation and again following depletion of CD5+ T cells by magnetic sorting. Experiment 2 (middle) employs a different control spleen to show purity of an Ig-depleted population. In experiment 3 (bottom) CD5 and IgM expression in unseparated and Ig-depleted splenocyte populations from a Tx frog are shown. Percentages of positive stained cells relate to the marker set to exclude 98% cells stained with control mAb (CT3).

Ig mAbs[ B cells are absent in the Ig! depleted population "T cells displaying proportional increase#[ In experiment 2 "Fig[ 1 bottom#\ prior to sorting the Tx frog spleen failed to display normal CD4 expression\ although some 7) CD4¦lo cells "CD4 ~uorescence being of a lower order of magnitude than seen in control cells# were recorded[ Depletion of Ig¦ cells is still successful with the Tx spleen\ despite a high proportion "35)# of B lymphocytes prior to separation[ The T and B cell

enriched populations yielded purities of ×84) "data not shown here\ but see Fig[ 3 demonstrating purity of CD4¦ and CD7¦ enriched populations#[ 2[1[1[ Cytotoxicity Assays[ The ability of the above T and B enriched:depleted splen! ocyte populations to lyse B2B6 targets after 37 hours culture is shown in Fig[ 2[ T cell sortin`] Medium!cultured control cells\ depleted of CD4¦ lymphocytes\ were unable to lyse B2B6 targets\ whereas after

Allotumour killing in Xenopus

223

Figure 3. Cytotoxicity towards B3B7 allothymus tumour targets displayed by sorted splenocyte subsets from adult control and early-thymectomized (Tx) Xenopus. The subsets were produced by MACS magnetic bead separation as illustrated in Fig. 2. Various cell populations were cultured for 48 hrs in medium alone or in GFM and then set up with targets at various effector:target (E:T) ratios in 6 hr 51Cr release assay. Top 2 graphs show mean data for T cell-depleted splenocytes (CD5−) and purified T cells (CD5+) from 4 control animals. Mean data for B cell-depleted splenocytes (Ig−) and purified B cells (Ig+) from 3 control frogs are shown in middle 2 graphs, and from 3 Tx frogs in bottom 2 graphs. Mean specific 51Cr release2S.E. from 5×104 targets is shown.

culture in GFM\ they became e}ective kil! lers "speci_c 40Cr release ½29) at E]T ratio 19]0#[ Enriched CD4¦ cells cultured in medium alone failed to display cyto! toxicity\ but displayed maximally ½19) killing after growth factor supplement! ation[ CD4− and CD4¦ cells cultured for 37 hours in GFM were unable to lyse allo! geneic lymphoblast targets\ there being ³1) killing even at E]T ratio of 19]0 "data

not shown#[ B cell sortin`] Both medium! and GFM!cultured control cells\ depleted of Ig¦ lymphocytes proved to be e}ective killers\ causing ½29) speci_c 40Cr release at E]T ratio 19]0[ Ig− splenocytes from Tx frogs were able to lyse B2B6 targets "½34) killing#\ but only if cultured in GFM[ Pur! i_ed B cells from either control or Tx spleen were unable to lyse target cells\ even after culture in GFM[

224

2[2[ Identi_cation of Three Subsets of Splenic Cytotoxic Cells in Control Xenopus Successive rounds of magnetic sorting of control splenocytes\ the _rst following incubation with anti!CD7 mAb "to provide CD7¦ and CD7− populations#\ and the second following incubation of the CD7− cells with anti!CD4 mAb\ yielded three mAb!de_ned subpopulations\ namely CD7¦"CD4¦#\ CD7−"CD4¦# and CD7− "CD4−# cells "see ~ow cytometric analysis shown in Fig[ 3#[ Each population was cul! tured for 37 hours in medium\ or with added GFM\ prior to cytotoxicity assay with B2B6 targets[ Figure 4 reveals that all three populations cultured in medium alone were ine}ective killers of the allo! tumour cells[ In contrast\ after culture in

T. L. Horton et al.

GFM\ each population was able to kill the tumour targets[ Maximal 40Cr release "at E]T ratio of 19]0# mediated by the CD7¦ T cell and non!T cell subsets was ½19)\ whereas the CD7− T cell subset achieved ½24) killing under the same conditions[

2[3[ Onto`eny of Splenocyte Cytotoxicity towards B2B6 Cells Figure 5 shows the outcome of cyto! toxicity assays "E]T ratios of 19]0# on unsorted splenocytes and thymocytes taken from 5Ð6 week old control larvae\ and also 5 and 01 month old control adults[ Each larval data point involved the pooling of 01Ð04 individual splenocyte:thymocyte cultures to provide su.cient cells for the assay[ Larval splenocytes routinely failed

Figure 4. Representative flow cytometric analysis to show production of CD8+ and CD8− T cells and a T cell-depleted population through 2 successive rounds of magnetic cell sorting. Splenocytes from a control adult were initially incubated with mAb AM22 to label CD8+ cells (top left histogram). The first magnetic sorting (MACS 1) of these mAb-labelled cells produced a CD8− fraction and a highly purified CD8+ population (middle 2 histograms). The CD8− fraction was then incubated with mAb 2B1 to label CD5+ cells and a second sorting (MACS 2) yielded a CD5−/CD8− splenocyte population and a purified CD5+/CD8− subset (bottom 2 histograms). Top right histogram illustrates unseparated splenocytes stained with both 2B1 and AM22 mAbs. Results of the MACS sorting experiments reveal that the 2 peaks in this top right histogram represent bright-staining CD8+ cells (30%) and a duller CD5+ population (41%). [Dual colour analysis (18) reveals that CD8 cells routinely coexpress CD5, but this coexpression is not evident in single colour analysis.] Percentages of positive-stained cells relate to a marker set to exclude 98% unseparated cells stained with CT3.

Allotumour killing in Xenopus

225

Figure 5. Cytotoxicity towards B3B7 allotumour targets mediated by the 3 control splenocyte populations produced by magnetic cell sorting as illustrated in Fig. 4. Each population was cultured for 48 hrs in medium alone or in GFM and then set up with targets at various effector:target (E:T) ratios in 6 hr 51Cr release assay. Mean data from 3 animals show specific 51Cr release2S.E. from 5×104 targets.

Figure 6. Ontogeny of cytotoxicity towards B3B7 allotumour cells. Unsorted splenocytes or thymocytes from 6–7 week old larvae, 6 and 12 month old adults were maintained in vitro for 48 hrs in GFM, prior to testing for cytotoxicity towards tumour targets in 6 hr 51Cr release assay at effector:target ratios of 20:1. Specific 51Cr release is shown for splenocytes (hatched box) and thymocytes (solid box) from two individual 12 month-old adults, and three individual 6 month-old frogs. Two separate experiments with larval lymphocytes were carried out, each involving splenocyte or thymocyte cultures established separately from 12–15 individual larvae, cells being pooled only at the time of chromium-release assay.

to kill the tumour targets\ whether cultured for 37 hours in GFM "as shown in Fig[ 5# or in medium alone "data not shown#[ In contrast\ 5 month old froglets "some 2[4 months post!metamorphosis# contained splenocytes which after culture in GFM were almost as e}ective tumour cell killers as splenocytes from year old adults[ At no stage of development did medium!cultured "data not shown# or GFM!cultured "Fig[ 5# thymocytes have the capacity to kill B2B6 targets[

3[ Discussion The likelihood that NK!like activity exists in Xenopus stems from our initial observation "02# that splenocytes from some unimmunised\ T cell depleted "Tx# frogs spontaneously lyse MHC!de_cient allothymus tumour targets[ That work also indicated that the capacity of splenocytes to lyse allogeneic tumour cells is rapidly induced "within a few days# in control Xen! opus and elevated in Tx frogs by either in

226

vitro cell culture or in vivo allotumour cell injection[ In contrast\ the mammalian ery! thro!leukaemic "NK target# cell line K451\ allogeneic splenic lymphoblasts and allo! geneic erythrocyte targets coated with Xen! opus IgY "the amphibian IgG equivalent# all proved resistant to lysis by Xenopus splenocytes taken from tumour cell! injected frogs "02#[ Our current work has concentrated on probing the cellular basis of cytotoxicity towards tumour cells mediated by 37 hour! cultured splenocytes[ This protocol was adopted in view of our initial experiments on cells tested directly ex vivo "Fig[ 0#\ which in contrast to our earlier _ndings "02#\ revealed insigni_cant levels of killing with splenocytes from Tx frogs[ Such vari! ation in natural cytotoxicity displayed by splenocytes from separate batches of frogs may well relate to di}erences in natural antigenic challenge in vivo and associated cytokine modulation[ Thus it is known that antigen administration can in~uence the ability of Tx splenocytes to kill tumour targets "02#[ For in vitro culture to promote cytotoxicity in Tx splenocytes\ growth factor!supplemented medium "GFM# is required[ GFM is likely to contain an array of Xenopus T cell!derived cytokines\ including IL!1!like material that has been partially characterised "18#[ Although this cytokine!induced cytotoxicity is remi! niscent of lymphokine!activated killer "LAK# activity\ the question of whether culture in GFM induces splenocytes from Tx frogs to kill a wide array of target types was not examined[ It would be interesting to observe whether prolonged incubation of Xenopus splenocytes in GFM "×37 hours# results in increasing promiscuity of target cell killing\ typical of mammalian LAK cells "30#[ LAK activity was _rst sus! pected in Xenopus splenocytes from frogs injected with the iridovirus FV2\ which after restimulation in vitro in the presence of T cell growth factor!rich supernatant\ exhibited non!speci_c cytotoxicity "29#[ The potential of GFM to stimulate NK! like activity in Tx splenocytes is consistent

T. L. Horton et al.

with previous proliferative studies\ where supernatants from activated T cells were shown to induce enhanced tritiated thy! midine incorporation in splenocytes from Tx Xenopus\ especially in the Ig− sorted population "11#[ In humans a NK cell pro! liferation!inducing factor distinct from IL! 1 has recently been identi_ed in culture supernatants from ConA!activated PBL "20#[ In this paper we have set out to inves! tigate which splenocyte populations in Tx and control Xenopus are able to display natural cytotoxicity towards allotumour targets[ The initial goal of this work is to identify alloreactive NK cells "21# at the amphibian level of evolution[ These studies will be expanded to include other potential sources of NK!like cells in Xenopus\ such as peripheral blood and intestinal lym! phocytes\ the latter being a major source of candidate NK cells in birds "7#[ Het! erogeneous populations of cytotoxic cells have recently been described in _sh[ For example\ separate peripheral blood leu! kocyte "PBL# populations from unim! munised cat_sh are spontaneously cytotoxic to either allogeneic "22# or vir! ally!infected "23# targets[ Cytotoxicity mediated by these two PBL populations is not inhibited by anti!NCC mAb "09\ 23#\ whereas the xenoreactive NCC cells extracted from cat_sh kidney are inhibited by this mAb[ Distinct populations of allo! reactive and xenoreactive non!speci_c cytotoxic cells have also been observed in tumour!bearing damsel_sh "24#[ Although killing by _sh cytotoxic cells frequently dis! plays lack of antigen speci_city "25#\ the unavailability of either T cell!depleted _sh or appropriate anti!_sh T cell antibodies precludes documentation as to whether such killing represents diverse NK!like populations rather than T cells[ Anti!Xenopus T cell mAbs have been used here in conjunction with magnetic sorting to show that control splenocytes\ depleted of T cells and then cultured for 37 hours\ behave like unsorted Tx splen! ocytes\ in that they become allocytotoxic

Allotumour killing in Xenopus

e}ectors\ but only after culture in GFM[ Sorting with anti!Ig mAbs revealed that puri_ed B cells from both control and Tx Xenopus\ even after culture in GFM\ rou! tinely failed to kill allotumour targets\ whereas B cell!depleted splenocytes from these same frogs became e}ective killers[ These experiments have therefore ident! i_ed that non!T:non!B splenic leucocytes in unimmunised control and Tx Xenopus can display natural lytic activity towards allotumour targets and represent NK can! didates[ Magnetic cell sorting has also identi_ed cytotoxicity against tumour cells "but not allogeneic lymphoblast targets# mediated by CD4¦ T cells from control spleen\ when these have been cultured in GFM[ ðThe failure of Xenopus lymphocytes to display natural cytotoxicity towards MHC! expressing allogeneic lymphoblast targets "see also "02## points to the possibility that MHC expression inhibits this type of allo! cytotoxicity[Ł This _nding that T cells from unimmunised\ outbred X[ laevis can lyse allogeneic B2B6 tumour cells "from ff strain X[ laevis# is of interest\ since in vitro cyto! toxicity of allogeneic splenic lymphoblasts by CTL requires in vivo immunisation with allogeneic lymphocytes or skin grafts\ and restimulation of splenocytes in vitro for 4 days "26#[ Furthermore\ CTL activity was detected only when cells bearing the same MHC haplotype were used for in vivo and in vitro stimulation "26#[ CTL activity towards minor alloantigens in Xenopus is even more di.cult to detect\ requiring pro! longed in vivo priming prior to in vitro restimulation "27\ 28#[ Interestingly\ some minor H antigen!reactive cell lines derived in this manner in the presence of Xenopus! derived T cell growth factor supernatant\ could also kill allogeneic lymphoblast tar! gets that were MHC disparate to the immunising cells "28#[ Our _nding that cul! tured T cells can kill allotumour cells is consistent with the recent _nding from early thymectomy experiments "39# that T cells play an important role in the rejection of ff strain thymus tumour cells implanted

227

into minor histocompatibility antigen!dis! parate ff Xenopus hosts[ The question of which T cell subset in unimmunised control Xenopus is capable of lysing allothymus tumour targets fol! lowing cell culture was investigated through sequential magnetic sorting\ _rst of CD7¦\ then CD4¦ cells[ The CD7¦ lym! phocytes "all of which coexpress CD4 "07##\ following 37 hour culture in GFM\ became e}ective killers[ It has been known for some time that murine alloantigenic CTL cultured continuously in T cell growth fac! tor become NK!like\ such LAK cells being able to kill an array of tumour target cells "30#[ IL!1 treatment over 4 days can raise CTL derived from human PBL to a state of activation where killing of tumour targets seems to become independent of antigen stimulation through the T cell receptor "0\ 31#[ Furthermore\ it is known that a close developmental relationship exists between mammalian T cells and NK cells "8\ 32# The residing of NK!like activity within Xenopus CTL kept in GFM is therefore not totally unexpected[ More surprising was the allocytotoxic capacity of the CD7−\ CD4¦ "putative CD3¦# T cell population cultured in GFM[ ðIt should be noted that there is no anti!CD3 reagent yet generated to identify bona _de Xenopus CD3¦ T cells[Ł This population achieved higher levels of tumour cell lysis than the enriched CD7¦ T cell "and T cell!depleted# subpopulations[ In mammals CD3¦ CTL clones exist that can kill targets relatively quickly\ probably through binding to the TNF receptor family member Fas "33\ 34#[ The nature of tumour target antigens recognised by Xenopus cytotoxic cells remains to be resolved[ Although class Ia and class II MHC proteins are not expre! ssed on the B2B6 cells "10#\ allocytotoxicity may be directed against minor his! tocompatibility antigens\ perhaps coded for by non!classical class Ib genes "35# since MHC class Ib mRNA is expressed in Xen! opus tumour cells "10#[ In this respect\ class Ib CD0 gene products are believed to be recognised by murine NK0¦T cells "36#[

228

T. L. Horton et al.

On the other hand\ the ability of Xenopus cytolytic cells to rapidly kill tumour cells\ even from minor H antigen disparate donors "10# may relate to expressison of tumour!speci_c antigens identi_ed on these target cells "39\ 37#[ Our ontogenetic studies reveal that splenocytes taken from 5Ð6 week!old X[ laevis larvae "just prior to metamorphosisŁ\ following 1 days culture with or without GFM\ were unable to lyse B2B6 allothymus tumour cells\ in contrast to the e}ective cytotoxicity manifested by splenocytes from 5 month and year!old froglets[ This failure of larval cells in vitro to destroy tumour targets from MHC!disparate Xen! opus di}ers from the in vivo _nding "10# that larvae are able to reject such tumour cells[ On the other hand\ ff tumour cells transplanted to ff Xenopus larvae are not rejected\ the ability to eliminate such minor H antigen!disparate tumours becoming e}ective only 3Ð4 weeks after the com! pletion of metamorphosis "39\ 37#[ The lar! val lack of rejection of tumour cells from MHC!compatible donors is due to incom! plete e}ector function\ rather than an absence of tumour cell recognition "37#[ Our in vitro cytotoxicity _ndings with larval cells reported here suggest that in!

complete e}ector function prior to meta! morphosis extends even to tumour cells derived from MHC!disparate donors[ Such immunologic immaturity may not be evident when the total larval immune sys! tem is investigated in vivo\ but becomes apparent only when components of allo! cytotoxicity are measured in vitro[ In summary\ in vitro studies with Xenopus splenocytes reveal that a non!T: non!B population and also two T cell sub! sets all display natural cytotoxicity towards allothymus tumour cells following 37 hours culture in T cell!derived growth factor!rich medium[ Since Tx Xenopus do not develop bona _de T cells "07Ð19#\ they provide a useful phylogenetic model for generation and screening of monoclonal antibodies against the non!T cytotoxic population\ that is likely to include can! didate natural killer cells[

Acknowled`ements*The authors thank Dr[ Louis DuPasquier for the generous gift of Xen! opus B2B6 thymus tumour cells[ The animal hus! bandry skills of Paul Loftus are gratefully acknowledged[ Work funded by Research Grants from The Leverhulme Trust "JDH# and The University of Durham Research Com! mittee "JDH + MDW#[

References 0[ Valiante\ N[ M[ and Parham\ P[ NK cells and CTL] opposite sides of the same coin[ In Molec! ular Basis of NK Cell Reco`nition and Function\ ed L[ Moretta[ Karger\ Basel\ 0885 pp[ 035Ð052[ 1[ Kos\ F[ J[ and Engleman\ E[ G[ Immune regu! lation] a critical link between NK cells and CTLs[ Immunol[ Today\ 0885\ 06\ 063Ð065[ 2[ Yokoyama\ W[ M[ Natural killer cell receptors[ Current Opinion Immunol[\ 0884\ 6\ 009Ð019[ 3[ Raulet\ D[ H[ Recognition events that inhibit and activate natural killer cells[ Current Opinion Immunol[\ 0885\ 7\ 261Ð266[ 4[ Lanier\ L[ L[ Natural killer cell receptors and MHC class I interactions[ Current Opinion Immu! nol[\ 0886\ 8\ 015Ð020[ 5[ Long\ E[ O[ and Wagtmann\ N[ Natural killer cell receptors[ Current Opinion Immunol[\ 0886\ 8\ 233Ð249[ 6[ Janeway\ C[ A[ Natural killer cells] a primitive immune system[ Nature\ 0878\ 230\ 097[

7[ Gobel\ T[ W[ F[\ Chen\ C[ H[\ Shrimpf\ J[\ Grossi\ C[ E[\ Bernot\ A[\ Bucy\ R[ P[\ Au}ray\ C[ and Cooper\ M[ D[ Characterization of avian natural killer cells and their intracellular CD2 protein complex[ Eur[ J[ Immunol[\ 0883\ 13\ 0574Ð0580[ 8[ Brooks\ C[ G[\ Georgiou\ A[ and Jordan\ R[ K[ The majority of immature fetal thymocytes can be induced to proliferate to IL!1 and di}erentiate into cells indistinguishable from mature natural killer cells[ J[ Immunol[\ 0882\ 01\ 5534Ð5545[ 09[ Harris\ D[ T[\ Kapur\ R[\ Frye\ C[\ Acevedo\ A[\ Camenisch\ T[\ Jaso!Friedmann\ L[ and Evans\ D[ L[ A species!conserved NK cell antigen recep! tor is a novel vimentin!like molecule[ Devel[ Comp[ Immunol[\ 0881\ 05\ 284Ð392[ 00[ Du Pasquier\ L[\ Schwager\ J[ and Flajnik\ M[ F[ The immune system of Xenopus[ Ann[ Rev[ Immunol[\ 0878\ 6\ 140Ð164[ 01[ Ghoneum\ M[\ Cooper\ E[ L[ and Sadek\ I[ Varia! bility of natural killer cell activity in anuran

Allotumour killing in Xenopus

02[

03[ 04[

05[ 06[

07[

08[

19[

10[

11[

12[

13[

14[ 15[

16[

amphibians[ Devel[ Comp[ Immunol[\ 0889\ 03\ 248Ð254[ Horton\ T[ L[\ Ritchie\ P[\ Watson\ M[ D[ and Horton\ J[ D[ NK!like activity against allogeneic tumour cells demonstrated in the spleen of con! trol and thymectomized Xenopus[ Immunol[ Cell Biol[\ 0885\ 63\ 254Ð262[ Horton\ T[ L[\ Ritchie\ P[\ Watson\ M[ D[ and Horton\ J[ D[ NK cell evolution] studies on Xeno! pus[ Biochem[ Soc[ Trans[\ 0886\ 14\ 152S[ Horton\ J[ D[ and Manning\ M[ J[ Response to skin allografts in Xenopus laevis following thy! mectomy at early stages of lymphoid organ matu! ration[ Transplantation\ 0861\ 03\ 030Ð043[ Horton\ J[ D[ Amphibians[ In Immunolo`y] A Comparative Approach\ ed R[J[ Turner[ John Wiley\ Chichester\ 0883\ pp[ 090Ð025[ Nagata\ S[ and Cohen\ N[ Speci_c in vivo and non!speci_c in vitro alloreactivities of adult frogs "Xenopus laevis# that were thymectomized during early larval life[ Eur[ J[ Immunol[\ 0872\ 02\ 430Ð 434[ Gravenor\ I[\ Horton\ T[ L[\ Ritchie\ P[\ Flint\ E[ and Horton\ J[ D[ Ontogeny and thymus!depen! dence of T cell surface antigens in Xenopus] ~ow cytometric studies on monoclonal antibody! stained thymus and spleen[ Devel[ Comp[ Immunol[\ 0884\ 08\ 496Ð412[ Horton\ J[ D[^ Horton\ T[ L[^ Ritchie\ P[ Immune system of Xenopus] T cell biology[ In The Biolo`y of Xenopus\ ed H[ R[ Kobel and R[ C[ Tinsley[ Oxford University Press\ Oxford\ 0885\ pp[ 168Ð 188[ Horton\ J[\ Horton\ T[\ Ritchie\ P[\ Gravenor\ I[\ Gartland\ L[ and Cooper\ M[ Use of monoclonal antibodies to demonstrate absence of T cell devel! opment in early!thymectomized Xenopus[ Devel[ Comp[ Immunol[\ 0883\ 07\ S84[ Robert\ J[\ Guiet\ C[ and Du Pasquier\ L[ Lymph! oid tumours of Xenopus laevis with di}erent capacities for growth in larvae and adults[ Devel[ Immunol[\ 0883\ 2\ 186Ð296[ Turner\ S[ L[\ Horton\ T[ L[\ Ritchie\ P[ and Horton\ J[ D[ Splenocyte response to T cell! derived cytokines in thymectomized Xenopus[ Devel[ Comp[ Immunol[\ 0880\ 04\ 208Ð217[ Jurgens\ J[ B[\ Gartland\ L[ A[\ Du Pasquier\ L[\ Horton\ J[ D[\ Gobel\ T[ W[ F[ and Cooper\ M[ D[ Identi_cation of a candidate CD4 homologue in the amphibian Xenopus laevis[ J[Immunol[\ 0884\ 044\ 3107Ð3112[ Langeberg\ L[\ Ruben\ L[ N[\ Clothier\ R[ H[ and Shiigi\ S[ The characterization of the toad splenocytes which bind mouse anti!human IL!1 receptor antibody[ Immunol[ Letters\ 0876\ 05\ 32Ð37[ Hsu\ E[\ Lefkovits\ I[\ Flajnik\ M[ and Du Pasqu! ier\ L[ Light chain heterogeneity in the amphibian Xenopus[ Molec[ Immunol[\ 0880\ 17\ 874Ð883[ Flajnik\ M[ F[\ Ferrone\ S[\ Cohen\ N[ and Du Pasquier\ L[ Evolution of the MHC] antigenicity and unusual tissue distribution of Xenopus "frog# class II molecules[ Molec[ Immunol[\ 0889\ 16\ 340Ð351[ Ra}erty\ K[ A[ Mass culture of amphibian cells] methods and observations concerning stability of

229

17[

18[

29[

20[

21[

22[

23[

24[

25[

26[

27[ 28[

39[

30[

cell type[ In Biolo`y of Amphibian Tumors\ ed[ M[ Mizell[ Springer!Verlag\ New York\ 0858\ pp[ 41Ð 70[ Chen\ C[ H[\ Ager\ L[ L[ Jr\ Gartland\ G[ L[ and Cooper\ M[ D[ Identi_cation of a T2:T cell receptor complex in chickens[ J[ Exp[ Med[\ 0875\ 053\ 264Ð279[ Haynes\ L[ and Cohen\ N[ Further charac! terization of an interleukin!1 like cytokine pro! duced by Xenopus laevis T lymphocytes[ Devel[ Immunol[\ 0882\ 2\ 120Ð127[ Watkins\ D[ T cell function in Xenopus] studies on T cell ontogeny and cytotoxicity using an IL! 1 like growth factor[ Ph[D[ Thesis\ University of Rochester\ New York\ 0874[ Saxena\ Q[ B[\ Saxena\ R[ K[\ Goldfarb\ R[ H[\ Herberman\ R[ B[ and Whiteside\ T[ L[ Identi! _cation and partial puri_cation of a human natu! ral killer cell proliferation!inducing factor[ J[ Biosciences\ 0885\ 10\ 344Ð358[ George\ T[\ Yu\ Y[ Y[ L[\ Liu\ J[\ Davenport\ C[\ Lemieux\ S[\ Stoneman\ E[\ Mathew\ P[ A[\ Kumar\ V[ and Bennett\ M[ Allorecognition by murine natural killer cells] lysis of T!lym! phoblasts and rejection of bone!marrow grafts[ Immunol[ Rev[\ 0886\ 044\ 18Ð39[ Yoshida\ S[ H[\ Stuge\ T[ B[\ Miller\ N[ W[ and Clem\ L[ W[ Phylogeny of lymphocyte het! erogeneity] cytotoxic activity of channel cat_sh peripheral blood leukocytes directed against allo! geneic targets[ Devel[ Comp[ Immunol[\ 0884\ 08\ 60Ð66[ Hogan\ R[ J[\ Stuge\ T[ B[\ Clem\ L[ W[\ Miller\ N[ W[ and Chinchar\ V[ G[ Anti!viral cytotoxic cells in the channel cat_sh "Ictalurus punctatus#[ Devel[ Comp[ Immunol[\ 0885\ 19\ 004Ð016[ McKinney\ E[ C[ and Schmale\ M[ C[ Damsel_sh with neuro_bromatosis exhibit cytotoxicity towards retrovirus infected cells[ Devel[ Comp[ Immunol[\ 0886\ 10\ 176Ð187[ Stuge\ T[ B[^ Yoshida\ S[ H[^ Chinchar\ V[ G[^ Miller\ N[ W[^ Clem\ L[ W[ Cytotoxic activity generated from channel cat_sh peripheral blood leukocytes in mixed leukocyte cultures[ In press\ 0886[ Bernard\ C[ C[ A[\ Bordmann\ G[\ Blomberg\ B[ and Du Pasquier\ L[ Immunogenetic studies on the cell!mediated cytotoxicity in the clawed toad Xenopus laevis[ Immuno`enetics\ 0868\ 8\ 332Ð 343[ Lallone\ R[ L[ and Horton\ J[ D[ In vitro cor! relates of in vivo skin graft rejection in Xenopus[ Transplantation\ 0874\ 28\ 112Ð115[ Watkins\ D[\ Harding\ F[ and Cohen\ N[ In vitro proliferative and cytotoxic responses against Xen! opus minor histocompatibility antigens[ Trans! plantation\ 0877\ 34\ 388Ð490[ Robert\ J[\ Guiet\ C[\ Cohen\ N[ and Du Pasquier\ L[ E}ects of thymectomy and tolerance induction on tumor immunity in adult Xenopus laevis[ Int[ J[ Cancer\ 0886\ 69\ 229Ð223[ Brooks\ C[ G[\ Urdal\ D[ L[ and Henney\ C[ S[ Lymphokine!driven {di}erentiation| of cytotoxic T!cell clones into cells with NK!like speci_city] correlations with display of membrane macro! molecules[ Immun[ Rev[\ 0872\ 61\ 32Ð61[

230 31[ Dianzani\ C[\ Zarcone\ D[\ Pistoia\ V[\ Grossi\ C[ E[\ Pileri\ A[\ Massaia\ M[ and Ferrarini\ M[ CD7¦CD00b¦ peripheral blood lymphocytes contain lymphokine!activated killer cell precur! sors[ Eur[ J[ Immunol[\ 0878\ 08\ 0926Ð0933[ 32[ Lanier\ L[ L[\ Spits\ H[ and Phillips\ J[ H[ The developmental relationship between NK cells and T cells[ Immunol[ Today\ 0881\ 02\ 281Ð284[ 33[ Aposov\ S[ D[\ Redegeld\ F[ and Sitkovsky\ M[ Cell!mediated cytotoxicity] contact and secreted factors[ Current Opinion Immunol[\ 0882\ 4\ 393Ð 309[ 34[ Squier\ M[ K[ T[ and Cohen\ J[ J[ Cell!mediated cytotoxic mechanisms[ Current Opinion Immunol\ 0883\ 5\ 336Ð341[

T. L. Horton et al. 35[ Flajnik\ M[ F[\ Kasahara\ M[\ Shum\ B[ P[\ Salter!Cid\ L[\ Taylor\ E[ and Du Pasquier\ L[ A novel type of class I gene organization in ver! tebrates] a large family of non!MHC!linked class I genes is expressed at the RNA level in the amphibian Xenopus[ EMBO Journal\ 0882\ 01\ 3274Ð3285[ 36[ Bendelac\ A[\ Rivera\ M[ N[\ Park\ S[ H[ and Roark\ J[ H[ Mouse CD0!speci_c NK0 T cells] Development\ speci_city\ and function[ Ann[ Rev[ Immunol[\ 0886\ 04\ 424Ð451[ 37[ Robert\ J[\ Guiet\ C[ and Du Pasquier\ L[ Ontogeny of the alloimmune response against a transplanted tumor in Xenopus laevis[ Dif! ferentiation\ 0884\ 48\ 024Ð033[