- Email: [email protected]
Morphologic and phenotypic analysis of canine natural killer cells: Evidence for T-cell lineage
CELLULAR
IMMUNOLOGY
Morphologic
95,207-2 17 (1985)
and Phenotypic Analysis of Canine Natural Killer Cells: Evidence for T-Cell Lineage’
THOMASP.LOUGHRAN,JR.,'H.JOACHIMDEEG,ANDRAINERSTORB Fred Hutchinson Cancer Research Center, Seattle, and Division of Oncology, Department of Medicine, University of Washington School of Medicine, Seattle, Washington Received January 16, 1985; accepted June 14, 1985
Caninenatural killer (NK) activity and antibody-dependent cell-mediated cytotoxicity were studied utilizing a canine thyroid adenocarcinoma cell line and a lymphoblastoid cell line (CT45s) respectively, as cell targets. Fractionation of peripheral blood mononuclear cells by Percoll discontinuous-gradient centrifugation resulted in a six- to sevenfold enrichment in large granular lymphocytes (LGL) in parallel with a twofold increase in NK activity (%specific lysis) in lowdensity fractions. Further enrichment in LGL (78 + 6%) and NK activity (threefold increase)was obtained by lytic treatment of low-density fractions 2 and 3 with monoclonal antibody WIG4. By means of cytolytic treatment with additional monoclonal antibodies the phenotype of canine NK cells was determined as Dly-1+, Dly-6+, lAl+, E-l l+, DT-2-, WIG4-. Some NK cells were also Ia+. NK activity was relatively radioresistantwith 40% specificlysis even alter irradiation with 40 Gy. Among the populations examined, the highest NK activity was found in peripheral blood mononuclear cells, followed by splenic mononuclear cells and bone marrow mononuclear cells. These results indicate that canine NK cells have the morphology of LGL, are relatively radioresistant, and express cell surface antigens suggesting a T-cell lineage. 0 1985 Academic Press. Inc.
INTRODUCTION Natural killer (NK) cells are the mononuclear cell population responsible for spontaneous in vitro cytotoxicity against a variety of cell targets. The in vivo function of NK cells is not entirely certain; however, an involvement in anti-viral and tumor surveillance (1,2), and hematopoietic (3,4) and B-cell (5-9) regulation has been postulated. It has also been suggestedthat NK cells may mediate resistanceto hemopoietic grafts in mice (lo- 13). Recently, Krakowka has established a canine NK assayusing a canine thyroid adenocarcinoma cell line as target (14). We were interested in further characterizing NK cells in the dog, with the aim of applying that knowledge to the canine model of marrow transplantation. We describe here the morphologic and phenotypic characteristics of canine NK cells obtained by enrichment on Percoll discontinuous gradients and analyzed with monoclonal antibodies. ’ This investigation was supported by Grant Nos. CA 31787, CA 18221, CA 30924, and CA 18105 awarded by The National Cancer Institute, DHHS, and Department of Defense Contract, No. DNA OOl83-C-0294. Dr. Loughran is a Fellow of the Leukemia Society of America. * To whom requests for reprints should be addressedat: Division of Oncology, Fred Hutchinson Cancer ResearchCenter, 1124 Columbia Street, Seattle, Wash. 98 104. 207 0008-8749185$3.00 Copyright 0 1985 by Academic Press, Inc. All rights of reproduction in any form reserved.
208
LOUGHRAN,
DEEG, AND STORB
MATERIALS
AND METHODS
Dogs. Beagles,0.5 to 2 years old, served as normal donors of blood, bone marrow, and spleen cells. NK assay. Peripheral blood, bone marrow, or splenic mononuclear cells served as effector cells. Peripheral blood mononuclear cells were separatedusing Ficoll-Hypaque density gradient centrifugation at 9OOgfor 25 min, as previously described ( 15). Cells from a canine thyroid adenocarcinoma (CTAC) cell line served as target for the NK assay (16). Target cell preparation involved incubation of CTAC with 300 &i of Na”CrO4 (New England Nuclear, Boston, Mass.) at 37°C for 90 min, followed by 1 wash with cold medium [RPM& GIBCO, Grand Island, N.Y.; containing 10% fetal calf serum (GIBCO) and 1%nonessential amino acids and 1%penicillin-streptomycin (GIBCO)] and incubation for 30 min at 4°C as previously described (14). Target solutions were then adjusted to 1 X lo6 labeled targets/ml. Targets were added to 96well U-bottom plates (Costar, Cambridge, Mass.) at a concentration of 1 X 1O5labeled targets/well. One-tenth milliliter of effector cell suspensionswas added at concentrations yielding effecter/target ratios of 1O/1, 20/ 1, 50/ 1, 80/l, and lOO/1, in triplicate wells. Plates were then centrifuged at 2008 for 2 min, incubated at 37°C for either 4, 8, or 18 hr, and centrifuged again at 1OOgfor 2 min. From each well 0.1 cc of supematant was harvested and radioactivity was quantitated in a gamma scintillation counter. Percentage cytotoxicity (%specific lysis) was calculated by a standard formula, using the mean value of triplicate cultures: %Specific lysis =
experimental release - spontaneous release x 100. maximum release - spontaneous release
Spontaneous releasewas determined in wells containing target and medium alone. In preliminary experiments three methods for determining maximum releasewere compared: (i) freeze-thawing, (ii) detergent lysis, and (iii) using 90% of the value for total incorporation. Since there was no difference among the results obtained with these methods, in all subsequent experiments the maximum release was derived by using 90% of the value for total incorporation. ADCC assay. A lymphoblastoid cell line (CT-45s) described previously ( 17) served as the cell target for the ADCC assay. CT45-S cells were incubated wth 200 &i of Na5’CrOh and 0.1 ml of heat-inactivated normal dog serum for 1 hr, followed by three washeswith cold medium (Waymouths MB 752-l) Irvine Scientific, Santa Ana, Calif., containing 20% fetal calf serum, 1% nonessential amino acids, and 1% penicillinstreptomycin). The procedure for the remaining portion of the ADCC assaywas identical to that described for the NK assay with the exception of antibody coating of CT45-S target. An antibody to CT45-S was prepared by immunizing rabbits with CT45-S target cells (unpublished results). Antibody (0.044 ml) was added to wells containing target cells at dilutions of either l/100 or l/1000. As a control, 0.044 ml of heat inactivated normal rabbit or dog serum was added to wells containing target cells and no antibody (uncoated targets). Methods of NK-cell enrichment. Percoll discontinuous-gradient centrifugation was performed as described by Timonen and Saksela(18), with the exception of the following modifications: Peripheral blood mononuclear cells were suspendedin 9 ml of 50%Percoll (Pharmacia, Upsala, Sweden)in a 50-ml centrifuge tube (Costar). Gradients of 37.5-508 were then established by adding 6 ml of Percoll in successiveincrements
CANINE
NK CELLS
209
of 2.5%; 3 cc of PBS (GIBCO) was placed on top. The tube was then centrifuged at 55Ogfor 30 min. Seven fractions of 6 ml each were collected and washed twice with medium (Waymouths). In some experiments, fractions 2 and 3 were further treated with monoclonal antibody WIG4 (described below), adjusted to original cell number, and used in NK assays. Morphology. The morphology of peripheral blood mononuclear cells and that of cells obtained in the various Percoll gradients were examined using cytospin preparations. Cells were added to slides at a concentration of 2 X 105/ml and were spun at 700 r-pm for 7 min. Slides were then stained with Wright-Giemsa. Similar cytospin preparations were performed on effector-target conjugates, as previously described (19). For electron microscopy examination, cells were fixed in glutaraldehyde and examined as described (20). In vitro irradiation. For theseexperiments, peripheral blood mononuclear cells were irradiated with a ‘37Cssource 2 hr prior to use in the NK assay.Total exposures were 20,40, or 80 Gy. Phenotyping. Phenotyping was carried out with a panel of monoclonal antibodies: Antibody Dly- 1 (IgGZb)identifies a 2 13,000-Da surface protein found on all canine lymphohematopoietic cells (2 1). Antibody Dly-6 (IgM) identifies a 60,000-Da protein found on the majority of canine lymphocytes but not on monocytes (2 1). Antibody DT-2 (IgG2,) identifies a 72,000-Da protein found on a subset of canine T lymphocytes, recognizing 60-70% of T-cells in peripheral blood (22). This antibody recognizes a canine T-lymphocyte subpopulation which is functionally equivalent to CD4 + (T4) human lymphocytes (i.e., helper subset). Antibody E- 11 (IgG3) is directed at a subsetof canine T lymphocytes complementary to that recognized by DT-2 (23). In addition E- 11 recognizesa fraction of lymphocytes without T- or B-cell markers (23). E-l 1+ canine T lymphocytes are functionally equivalent to CD8 + (T8) human lymphocytes (i.e., T-cytotoxic/suppressor subset). Antibody 1Al (IgM) identifies a bimolecular complex of 80,000 and 28,000 Da and is directed at canine T lymphocytes. In addition it is reactive with a proportion of lymphocytes without other T- or B-cell markers (24). Antibody WIG4 (IgGJ is reactive with canine monocytes/macrophages and approximately 15% of peripheral blood lymphocytes (unpublished observations). Although it is of the IgG, isotype it activates complement and is cytolytic in vitro. Antibody 7.2 (IgG& identifies a bimolecular complex of 29,000 and 34,000 Da representing a framework determinant of la-like antigens found on human cells (25), and is cross-reactive with a similar complex on canine cells (26). These antibodies were used in conjunction with complement for negative selection of peripheral blood mononuclear cells to be used in NK assays.All antibodies were used in the form of ascites fluid, at dilutions of 1:1000. The peripheral blood mononuclear cells that were to be treated were first adjusted to a concentration of 10 X lo6 cells/ml. Under intermittent shaking, cells were incubated for 30 min at 22°C with medium or monoclonal antibody at a volume ratio of 1:1, followed by incubation for 90 min with the same volume of complement (rabbit serum screened for HLA-DR typing, Pel FreezeRogers,Ark.), or medium. Cells were then washedonce with medium and counted for viability (Trypan blue exclusion). Effecters, adjusted to the original volume and not reconstituted to original cell number, were plated in the NK assay. Efficiency of lytic treatment was evaluated by the number of viable cells remaining.
210
LOUGHRAN,
DEEG, AND STORB
For Dly-1 this averaged 10%; for Dly-6 35%; for DT-2 60%; for E-l 1 86%; for IA1 44%; for WIG4 48%; and for 7.2 50%. RESULTS NK assay.Results utilizing peripheral blood mononuclear cells as effectersare shown in Fig. 1. Only minimal NK activity was observed with 4-hr assays.Eight- and 18-hr assaysproduced specific lysis of greater than 50%. NK activity peaked at an effector/ target ratio of 80/l. Spontaneous release for CTAC ranged from 5-17%, mean 12% at 4 hr, 9-l 8%, mean 12% at 8 hr, and 1l-30%, mean 20% at 18 hr. ADCC assay. Forty percent specific lysis was observed in a 4-hr assay utilizing peripheral blood mononuclear cells at an effecter/target ratio of 100/l. Optimum cytotoxicity was achieved using an antibody concentration of I/ 1000. There was no killing when uncoated targets were used. Enrichment of NK activity. Table 1 shows the cell yields resulting from Percoll discontinuous-gradient centrifugation of peripheral blood mononuclear cells as well as the percent of LGL in each fraction. The total number of cells recovered averaged 65% of the total input. Highest NK activity occurred in the low-density fractions, as shown in Fig. 2. The fractions with the highest NK activity corresponded to those with the highest proportion of LGL. Further enrichment of LGL and NK activity was obtained by lytic treatment of cells from fractions 2 and 3 with monoclonal antibody WIG4 plus complement, as shown in Fig. 2 (bar graph). Cells enriched in the highdensity fractions had the morphology of typical lymphocytes and had little, if any,
.
..
i3toctw
- Target
I
I
10
20
I
so
I
60
... I
100
Ratio
FIG. 1. NK assay results utilizing peripheral blood mononuclear cells.at various etTector/target ratios, showing mean ?&specificlysis after 4-hr (dotted line, n = 12), 8-hr (dashedline, n = 16),and IS-hr incubation (solid line, n = 22). Mean Sbspecificlysis for ADCC assay utilizing peripheral blond mononuclear cells is shown for target coated with no antibody (dotted line), and with I/ 100(dashedline), and l/ 1000(solid line) dilutions of antibody (n = 12). Standard errors of mean are shown in brackets.
211
CANINE NK CELLS TABLE 1
Cell Yields from Percoll Discontinuous-Gradient Centrifugation of Peripheral Blood Mononuclear Cells” Fraction
%LGL (M + SEM)
Unfractionated 1 2 3 4 5 6 7 2 + 3 (WIG4 + C’)*
6& 1.2 31 1- 4.4 44 f 1.7 41 f 5.7 21 + 4.2 13 -f: 3.0 10 + 1.8 6+ 1.0 78 + 6.2
%Input (M + SEM) 9.8 k 8.5 f 6.5 + 5.3 + 7.3 f 7.0 f 15.6 f -
1.7 0.9 0.7 0.1 0.9 1.2 1.6
’ LGL, large granular lymphocytes; M, mean of 12 experiments. ’ Fractions 2 and 3 combined and then treated with monoclonal antibody WIG4 plus complement; results represent the means + SEM of five experiments.
NK activity. Separation of peripheral blood mononuclear cells by nylon-wool columns into nonadherent population did not deplete NK activity when compared to unseparated peripheral blood mononuclear cells (data not shown). Morphology of NK cells. Figure 3 shows the binding of effector cells to CTAC target
20
70
50
d R 10 w--l
90 IO
FIG. 2. Resultsof 8-hr NK assaysusing various cell fractions of Percoll discontinuous gradientsof peripheral blood mononuclear cells as effecters at an effecter/target ratio of 20/ 1. Mean %specific lysis is shown using scale on left (solid line, n = 7) while mean WLGL in each fraction is shown using scale on right (dashed line, n = 12). Further enrichment in NK activity (open bar) and %LGL (solid bar) was obtained by lytic treatment of low-density fractions 2 and 3 with monoclonal antibody WIG4 plus complement (n = 5). Standard errors of mean for Ispecific lysis are shown in brackets, and for %LGL in Table 1.
212
LOUGHRAN,
DEEG, AND STORB
FIG. 3. Binding of LGL effector cells (arrow heads) to target cells (arrows) (Wright-Giemsa, original magnification X400).
cells, illustrating the characteristic LGL morphology of the effector cell with eccentric nucleus, relatively large cytoplasmic/nuclear ratio, and azurophilic granules. Figure 4 reveals the ultrastructural features of the canine LGL, showing multiple characteristic electron-dense granules. Tissue distribution of NK activity. As shown in Fig. 5, among the three populations examined, the highest NK activity was found in peripheral blood mononuclear cells, followed by spleen and bone marrow. Radiation sensitivity. Figure 6 depicts NK activity of peripheral blood mononuclear cells after in vitro radiation exposure. NK activity was relatively radioresistant, as demonstrated by 40% specific lysis remaining even after irradiation with 40 Gy. Phenotype of NK cells. Results are summarized in Figure 7. Lytic treatment of peripheral blood mononuclear cells with complement and monoclonal antibodies Dly- 1, Dly-6, 1Al, and E- 11, respectively, depleted NK activity, such that the remaining cytotoxicity ranged from only lo-34% of the complement controls. In most cases, this was also true of monoclonal reagent 7.2, although in some dogs lytic treatment with this antibody had no significant effect on cytotoxicity. As a result, cytotoxic activity remaining after such lytic treatment with 7.2 averaged50% of the complement controls. In contrast, treatment with complement and monoclonal antibodies DT-2 or WIG4, respectively, failed to reduce NK activity. Since both antibodies strongly activate complement, this suggeststhat canine NK cells do not express the antigens recognized by theseantibodies. None of the antibodies when added to the assayblocked NK activity in the absenceof complement treatment (data not shown).
CANINE NK CELLS
213
FIG. 4. Ultrastructural features of LGL, demonstrating typical electron-dense granules (arrow heads) (original magnification X8000).
DISCUSSION In the mouse, rat, and man it has been shown that NK function resides in a subpopulation of large lymphocytes characterizedby the presenceof prominent azurophilic granules and consequently termed LGL (27-29). Several investigators have shown that mononuclear cells possessingNK function can be distinguished by phenotyping with specific monoclonal antibodies (30-33). We have shown in the present study that canine NK activity resides in a subpopulation of peripheral blood mononuclear cells that have the morphology of LGL and are nonadherent, low-density cells, which is similar to findings in other species(27-29). Ultrastructural features were similar to those observed in human LGL (34). The phenotype of canine NK cells was Dly-l+, Dly-6+, lAl+, E-l l+, DT-2-, and WIG4-. Our studies confirm the results reported by Krakowka in defining the suitability of CTAC as a cell target in the NK assay(14). Eight- and 18-hr assaysresulted in easily demonstrated killing, with acceptable values for spontaneous release. In agreement with results in other species(35), cells recognized as having NK function (LGL morphology) also functioned as effector cells in ADCC. Furthermore, enrichment of ADCC activity also occurred in low-density fractions, suggesting that canine LGL are responsible for both NK and ADCC activity.
214
LOUGHRAN,
10
20
DEEG, AND STORB
50
80
100
EMCtCf-TarpstFMiO
FIG. 5. Mean Zspecific lysis (brackets indicate standard error of the mean) in 18-hr NK assayby canine mononuclear cells of different origin at various effecter/target ratios. Peripheral blood mononuclear cells are represented by solid line (n = 12), spleen cells by dashed line (n = 5), and marrow cells by dotted line (n = 4).
Canine NK activity was studied in peripheral blood, splenic, and marrow mononuclear cells, with peripheral blood mononuclear cells having highest cytotoxicity and marrow mononuclear cells least. Splenic mononuclear cells appeared to be intermediate; however, these results were obtained using frozen cells as effecters since fresh
10
20
50
80
100
Effecter-TmwtFtath FIG. 6. Effect of in vitro irradiation on 18-hr NK assaysat various effecter/target ratios utilizing peripheral blood mononuclear cells as effecters(n = 6). Mean Ispecific lysis with unirradiated cells (control) is depicted by open circles, 20 Gy by solid circles, 40 Gy by open squares,and 80 Gy by solid squares.Standard errors of mean ranged from 9.8 to 16.7%for mm-radiated cells, from 10.1 to 15.3%at 20 Gy, 10.2 to 16.2%at 40 Gy, and 9.9 to 17.7%at 80 Gy.
215
CANINE NK CELLS
r
ldcmelml
Antibodiis
FIG. 7. NK activity (I 8-hr assay)remaining after lytic treatment of peripheral blood mononuclear cells with monoclonal antibodies plus complement, (expressedas % of complement controls). Results represent mean values of combined experiments performed at ratios of 50/ 1 or 80/I (n = 12 for antibodies Dly- I and Dly-6; n = 8 for lA1, E-l 1 and DT-2; n = 22 for 7.2; and n = 4 for WIG4).
cells were not available. Freezing of canine peripheral blood mononuclear cells resulted in a decreaseof approximately 33% in NK activity (data not shown). Therefore, cytotoxicity might have been greater if freshly isolated splenic mononuclear cells had been used as effecters. The ontogeny of human NK cells is uncertain, with T-cell (36) myeloid (37, 38), and separatelineage (39,40) having all been postulated. Our results suggestthat canine NK cells share phenotypic characteristics of canine T cells. Experiments utilizing lytic treatment of peripheral blood mononuclear cells with complement and a panel of monoclonal reagents indicated that canine NK cells, as defined in vitro, express cell surface antigens defined by monoclonal antibodies Dly- 1, Dly-6, 1A 1, and E- 11, and are negative for DT-2 and WIG4. The presenceof E- 11, a marker which identifies a canine T-lymphocyte subset functionally equivalent to CD8 + (T8) human lymphocytes [i.e., suppressor/cytotoxic T-cell subset (23)] on canine NK cells is similar to human NK cells, a subset of which coexpressessuppressor cell and NK-cell markers (30, 33, 41, 42). In contrast, NK cells did not expressthe antigen recognized by DT2 which identifies a canine T-lymphocyte subpopulation functionally equivalent to CD4 + (T4) human lymphocytes (i.e., T-helper subset)and which is otherwise expressed on a subset of T cells complementary to the E-l l+ subset. The fact that WIG4 (normally expressed on monocytes) did not recognize NK cells suggeststhat the canine NK lineage is different from that of the monocyte. This is further supported by the observation that elimination of NK activity could be accomplished using lytic treatment with Dly-6, an antibody not reactive with monocytes. Together with the E-l 1 and 1A 1 results, this would seemto be in agreement with previous findings in other species suggesting a common lineage for NK cells and cytotoxic T cells (20, 43). Also in agreement with this, Ringler and Krakowka have recently shown that canine NK cells expressthe canine T-lymphocyte marker, Thy-l (44).
216
LOUGHRAN,
DEEG,
AND
STORB
Our data also suggestthat canine NK cells expressIa antigens, although these results were variable. Several investigators have reported that a subset of human NK cells expressesIa antigens (30, 41), others have not confirmed ths finding (32). Recently, human NK cells generated in mixed lymphocyte cultures were shown to expressclass II antigens (45). The finding of classII antigen expression on NK cells may be relevant in regardsto their putative role in the resistanceto hemopoietic grafts (lo- 13). Gambel et al. have shown that pretreatment of mice with anti-class II monoclonal antibodies before the infusion of haploidentical donor marrow allows for the development of lymphohemopoietic chimerism (46). We have shown recently that treatment of dogs with anti-Ia antibody in vivo prior to transplantation facilitates engraftment of marrow from donors differing for both DLA haplotypes (47). Sincethe distribution of Ia antigens on canine lymphocytes is different than that of human lymphocytes (27), it remains to be determined whether this effect is due to inactivation of activated canine NK cells expressingIa antigens or to another host population that also expressesIa antigens. It is of note, however, that, similar to the inconsistent elimination of NK activity in vitro, the in vivo administration of anti-class II antibody (7.2) did not uniformly allow for sustained engrafiment (47). We are presently investigating whether there is a correlation between in vivo and in vitro results. Furthermore, the observation that NK activity was radioresistant may have implications in this model of histoincompatible marrow transplantation, since it has been shown in mice that genetic resistance to hemopoietic grafts is relatively radioresistant (10, 48). Furthermore, we have shown in our canine model that resistance to marrow engraftment can be overcome by highdose fractionated total-body irradiation (49), a scheme of irradiation similar to that which produces NK-cell deficiency in mice (50). Further experiments utilizing the phenotypic data presented here may be helpful in elucidating the role of NK cells in marrow graft resistance in the dog. ACKNOWLEDGMENTS We thank Susan DeRose and Liz Caldwell for excellent technical assistanceand Dr. P. Martin, Dr. W. Ladiges, and Dr. D. R. Krawiec for providing antibodies 7.2, WIG4, and lA1, respectively. We appreciate Jan Freedle’s help in preparing the manuscript.
REFERENCES 1. Herberman, R. B., Djeu, J. Y., Kay, H. D., Ortaldo, J. R., Riccardi, C., Bonnard, G. D., Holden, H. T., Fagnani, R., Santoni, A., and Puccetti, P., Immunol. Rev. 44,43, 1979. 2. Herberman, R. B., and Ortaldo, J. R., Science (Washington, D.C.) 214,24, 1981. 3. Hansson, M., Beran, M., Andersson, B., and Kiessling, R., J. Immunol. 129, 126, 1982. 4. Mangan, K. F., Hartnett, M. E., Matis, S. A., Winklestein, A., and Abo, T., Blood 63, 260, 1982. 5. Tilden, A. B., Abo, T., and Balch, C. M., J. Zmmunol. 130, 1171, 1983. 6. Arai, S., Yamamoto, H., Itoh, K., and Kumagai, K., J. Immunol. 131,651, 1983. 7. Abruzzo, L. V., and Rowley, D. A., Science (Washington. D.C.) 222,581, 1983. 8. Brieva, J. A., Targan, S., and Stevens, R. H., J. Immunol. 132,611, 1984. 9. James, K., and Ritchie, W. S., Immunol. Today 5, 193, 1984. IO. Kiessling, R., Hochman, P. S., Hailer, O., Shearer, G. M., Wigzell, H., and Cudkowicz, G., Eur. J. Immnol. I, 655, 1977. 1I. Miller, S. C., J. Immunol. 131, 92, 1983. 12. Warner, J. F., and Dennert, G., Nature (London) 300, 31, 1982. 13. Lotzova, E., Pollack, S. B., and Savary, C. A., In “NK Cells and Other Natural Effector Cells” (R. B. Herberman, Ed.), pp. 1535-l 540. Academic Press,New York, 1982.
CANINE NK CELLS
217
14. Krakowka, S., Amer. .I. Vet. Res. 44, 635, 1983. 15. Deeg, H. J., Storb, R., Gerhard-Miller, L., Shulman, H. M., Weiden, P. L., and Thomas, E. D., Trunsplantation 29, 230, 1980. 16. Kasza, L., Amer. J. Vet. Res. 25, 1178, 1964. 17. Krakowka, S., Olsen, R., and Cockerell, G., In Vitro 13, 119, 1977. 18. Timonen, T., and Saksela,E., J. Immunol. Methods 36,285, 1980. 19. Timonen, T., Ortaldo, J. R., and Herberman, R. B., J. Immunol. 128,25 14, 1982. 20. Brooks, C. Cl., Kuribayashi, K., Sale, Cl. E., and Henney, C. S., J. Immunol. 128,2327, 1982. 21. Wulff, J. C., Durkopp, N., Aprile, J., Tsoi, M.-S., Deeg, H. J., and Storb, R., Exp. Hematol. 10, 535, 1983. 22. Wulff, J. C., Deeg, H. J., and Storb, R., Transplantation 33, 616, 1982. 23. Szer, J., Deeg, H. J., Sevems, E., and Storb, R., Transplantation 39, 187, 1985. 24. Krawiec, D. R., and Muscoplat, C. C., Amer. J. Vet. Res. 45,491, 1984. 25. Hansen, J. A., Martin, P., and Nowinski, R. C., Immunogenetics 10,247, 1980. 26. Deeg, H. J., Wulff, J. C., DeRose, S., Sale, G. E., Braun, M., Brown, M. A., Springmeyer, S. C., Martin, P. J., and Storb, R., Immunogenetics 16,445, 1982. 27. Kumagai, K., Itoh, K., Suzuki, R., Hinuma, S., and Saitoh, F., J. Immunol. 129, 388, 1982. 28. Reynolds, C. W., Timonen, T., and Herberman, R. B., J. Immunol. 127,282, 1981. 29. Timonen, T., Ortaldo, J. R., and Herberman, R. B., J. Exp. Med. 153,569, 1981. 30. Ortaldo, J. R., Sharrow, S. O., Timonen, T., and Herberman, R. B., J. Immunol. 127, 2401, 1981. 31. Abo, T., and Balch, C. M., J. Immunol. 127, 1024, 1981. 32. Perussia,B., Starr, S., Abraham, S., Fanning, V., and Trinchieri, G., J. Immunol. 130, 2 133, 1983. 33. Lanier, L. L., Le, A. M., Phillips, J. H., Warner, N. L., and Babcock, G. F., J. Immunol. 131, 1789, 1983. 34. Grossi, C. E., Cadoni, A., Zicca, A., Leprini, A., and Ferrarini, M., Blood 59, 277, 1982. 35. Bradley, T. P., and Bonavida, B., J. Immunol. 129, 2260, 1982. 36. Grossman,Z., and Herberman, R. B., In “NK Cells and Other Natural Effector Cells” (R. B. Herberman, Ed.), pp. 229-238. Academic Press,New York, 1982. 37. Kay, H. D., and Horowitz, D. A., J. Clin. Invest. 66, 847, 1980, 38. Zarling, J. M., and Kung, P. C., Nature [London) 288,394, 1980. 39. Ferrarini, M., and Grossi, C. E., In “NK Cells and Other Natural Effector Cells” (R. B. Herberman, Ed.), pp. 257-264. Academic Press,New York, 1982. 40. Ortaldo, J. R., In “NK Cells and Other Natural Effector Cells” (R. B. Herberman, Ed.), pp. 265-271. Academic Press,New York, 1982. 41. Abo, T., Cooper, M. D., and Balch, C. M., J. Immunol. 129, 1752, 1982. 42. Perussia, B., Fanning, V., and Trinchieri, G., J. Immunol. 131,223, 1983. 43. Fast, L., Beatty, P., Hansen, J. A., and Newman, W., J. Immunol. 131,2404, 1983. 44. Ringler, S. S., and Krakowka, S., Vet. Immunol. Immunopathol., 1985, in press. 45. Phillips, J. H., Le, A. M., and Lanier, L. L., J. Exp. Med. 159, 993, 1984. 46. Gambcl, P., Francescutti, L. H., and Wegmann, T. G., Transplantation 38, 152, 1984. 47. Deeg, H. J., Storb, R., Szer, J., Appelbaum, F. R., Hackman, R. C., and Thomas, E. D., Transplant. Proc. 17,493, 1985.
48. Rauchwerger, J. M., Gallagher, M. T., Monie, H. J., and Trentin, J. J., Transplantation 23, 158, 1977. 49. Deeg,H. J., Storb, R., Shulman, H. M., Weiden, P. L., Graham, T. C., and Thomas, E. D., Trunsp~unt~tion 33,443, 1982. 50. Parkinson, D. R., Brightman, R. P., and Waksal, S. D., J. Immunol. 126, 1460, 1981.
IMMUNOLOGY
Morphologic
95,207-2 17 (1985)
and Phenotypic Analysis of Canine Natural Killer Cells: Evidence for T-Cell Lineage’
THOMASP.LOUGHRAN,JR.,'H.JOACHIMDEEG,ANDRAINERSTORB Fred Hutchinson Cancer Research Center, Seattle, and Division of Oncology, Department of Medicine, University of Washington School of Medicine, Seattle, Washington Received January 16, 1985; accepted June 14, 1985
Caninenatural killer (NK) activity and antibody-dependent cell-mediated cytotoxicity were studied utilizing a canine thyroid adenocarcinoma cell line and a lymphoblastoid cell line (CT45s) respectively, as cell targets. Fractionation of peripheral blood mononuclear cells by Percoll discontinuous-gradient centrifugation resulted in a six- to sevenfold enrichment in large granular lymphocytes (LGL) in parallel with a twofold increase in NK activity (%specific lysis) in lowdensity fractions. Further enrichment in LGL (78 + 6%) and NK activity (threefold increase)was obtained by lytic treatment of low-density fractions 2 and 3 with monoclonal antibody WIG4. By means of cytolytic treatment with additional monoclonal antibodies the phenotype of canine NK cells was determined as Dly-1+, Dly-6+, lAl+, E-l l+, DT-2-, WIG4-. Some NK cells were also Ia+. NK activity was relatively radioresistantwith 40% specificlysis even alter irradiation with 40 Gy. Among the populations examined, the highest NK activity was found in peripheral blood mononuclear cells, followed by splenic mononuclear cells and bone marrow mononuclear cells. These results indicate that canine NK cells have the morphology of LGL, are relatively radioresistant, and express cell surface antigens suggesting a T-cell lineage. 0 1985 Academic Press. Inc.
INTRODUCTION Natural killer (NK) cells are the mononuclear cell population responsible for spontaneous in vitro cytotoxicity against a variety of cell targets. The in vivo function of NK cells is not entirely certain; however, an involvement in anti-viral and tumor surveillance (1,2), and hematopoietic (3,4) and B-cell (5-9) regulation has been postulated. It has also been suggestedthat NK cells may mediate resistanceto hemopoietic grafts in mice (lo- 13). Recently, Krakowka has established a canine NK assayusing a canine thyroid adenocarcinoma cell line as target (14). We were interested in further characterizing NK cells in the dog, with the aim of applying that knowledge to the canine model of marrow transplantation. We describe here the morphologic and phenotypic characteristics of canine NK cells obtained by enrichment on Percoll discontinuous gradients and analyzed with monoclonal antibodies. ’ This investigation was supported by Grant Nos. CA 31787, CA 18221, CA 30924, and CA 18105 awarded by The National Cancer Institute, DHHS, and Department of Defense Contract, No. DNA OOl83-C-0294. Dr. Loughran is a Fellow of the Leukemia Society of America. * To whom requests for reprints should be addressedat: Division of Oncology, Fred Hutchinson Cancer ResearchCenter, 1124 Columbia Street, Seattle, Wash. 98 104. 207 0008-8749185$3.00 Copyright 0 1985 by Academic Press, Inc. All rights of reproduction in any form reserved.
208
LOUGHRAN,
DEEG, AND STORB
MATERIALS
AND METHODS
Dogs. Beagles,0.5 to 2 years old, served as normal donors of blood, bone marrow, and spleen cells. NK assay. Peripheral blood, bone marrow, or splenic mononuclear cells served as effector cells. Peripheral blood mononuclear cells were separatedusing Ficoll-Hypaque density gradient centrifugation at 9OOgfor 25 min, as previously described ( 15). Cells from a canine thyroid adenocarcinoma (CTAC) cell line served as target for the NK assay (16). Target cell preparation involved incubation of CTAC with 300 &i of Na”CrO4 (New England Nuclear, Boston, Mass.) at 37°C for 90 min, followed by 1 wash with cold medium [RPM& GIBCO, Grand Island, N.Y.; containing 10% fetal calf serum (GIBCO) and 1%nonessential amino acids and 1%penicillin-streptomycin (GIBCO)] and incubation for 30 min at 4°C as previously described (14). Target solutions were then adjusted to 1 X lo6 labeled targets/ml. Targets were added to 96well U-bottom plates (Costar, Cambridge, Mass.) at a concentration of 1 X 1O5labeled targets/well. One-tenth milliliter of effector cell suspensionswas added at concentrations yielding effecter/target ratios of 1O/1, 20/ 1, 50/ 1, 80/l, and lOO/1, in triplicate wells. Plates were then centrifuged at 2008 for 2 min, incubated at 37°C for either 4, 8, or 18 hr, and centrifuged again at 1OOgfor 2 min. From each well 0.1 cc of supematant was harvested and radioactivity was quantitated in a gamma scintillation counter. Percentage cytotoxicity (%specific lysis) was calculated by a standard formula, using the mean value of triplicate cultures: %Specific lysis =
experimental release - spontaneous release x 100. maximum release - spontaneous release
Spontaneous releasewas determined in wells containing target and medium alone. In preliminary experiments three methods for determining maximum releasewere compared: (i) freeze-thawing, (ii) detergent lysis, and (iii) using 90% of the value for total incorporation. Since there was no difference among the results obtained with these methods, in all subsequent experiments the maximum release was derived by using 90% of the value for total incorporation. ADCC assay. A lymphoblastoid cell line (CT-45s) described previously ( 17) served as the cell target for the ADCC assay. CT45-S cells were incubated wth 200 &i of Na5’CrOh and 0.1 ml of heat-inactivated normal dog serum for 1 hr, followed by three washeswith cold medium (Waymouths MB 752-l) Irvine Scientific, Santa Ana, Calif., containing 20% fetal calf serum, 1% nonessential amino acids, and 1% penicillinstreptomycin). The procedure for the remaining portion of the ADCC assaywas identical to that described for the NK assay with the exception of antibody coating of CT45-S target. An antibody to CT45-S was prepared by immunizing rabbits with CT45-S target cells (unpublished results). Antibody (0.044 ml) was added to wells containing target cells at dilutions of either l/100 or l/1000. As a control, 0.044 ml of heat inactivated normal rabbit or dog serum was added to wells containing target cells and no antibody (uncoated targets). Methods of NK-cell enrichment. Percoll discontinuous-gradient centrifugation was performed as described by Timonen and Saksela(18), with the exception of the following modifications: Peripheral blood mononuclear cells were suspendedin 9 ml of 50%Percoll (Pharmacia, Upsala, Sweden)in a 50-ml centrifuge tube (Costar). Gradients of 37.5-508 were then established by adding 6 ml of Percoll in successiveincrements
CANINE
NK CELLS
209
of 2.5%; 3 cc of PBS (GIBCO) was placed on top. The tube was then centrifuged at 55Ogfor 30 min. Seven fractions of 6 ml each were collected and washed twice with medium (Waymouths). In some experiments, fractions 2 and 3 were further treated with monoclonal antibody WIG4 (described below), adjusted to original cell number, and used in NK assays. Morphology. The morphology of peripheral blood mononuclear cells and that of cells obtained in the various Percoll gradients were examined using cytospin preparations. Cells were added to slides at a concentration of 2 X 105/ml and were spun at 700 r-pm for 7 min. Slides were then stained with Wright-Giemsa. Similar cytospin preparations were performed on effector-target conjugates, as previously described (19). For electron microscopy examination, cells were fixed in glutaraldehyde and examined as described (20). In vitro irradiation. For theseexperiments, peripheral blood mononuclear cells were irradiated with a ‘37Cssource 2 hr prior to use in the NK assay.Total exposures were 20,40, or 80 Gy. Phenotyping. Phenotyping was carried out with a panel of monoclonal antibodies: Antibody Dly- 1 (IgGZb)identifies a 2 13,000-Da surface protein found on all canine lymphohematopoietic cells (2 1). Antibody Dly-6 (IgM) identifies a 60,000-Da protein found on the majority of canine lymphocytes but not on monocytes (2 1). Antibody DT-2 (IgG2,) identifies a 72,000-Da protein found on a subset of canine T lymphocytes, recognizing 60-70% of T-cells in peripheral blood (22). This antibody recognizes a canine T-lymphocyte subpopulation which is functionally equivalent to CD4 + (T4) human lymphocytes (i.e., helper subset). Antibody E- 11 (IgG3) is directed at a subsetof canine T lymphocytes complementary to that recognized by DT-2 (23). In addition E- 11 recognizesa fraction of lymphocytes without T- or B-cell markers (23). E-l 1+ canine T lymphocytes are functionally equivalent to CD8 + (T8) human lymphocytes (i.e., T-cytotoxic/suppressor subset). Antibody 1Al (IgM) identifies a bimolecular complex of 80,000 and 28,000 Da and is directed at canine T lymphocytes. In addition it is reactive with a proportion of lymphocytes without other T- or B-cell markers (24). Antibody WIG4 (IgGJ is reactive with canine monocytes/macrophages and approximately 15% of peripheral blood lymphocytes (unpublished observations). Although it is of the IgG, isotype it activates complement and is cytolytic in vitro. Antibody 7.2 (IgG& identifies a bimolecular complex of 29,000 and 34,000 Da representing a framework determinant of la-like antigens found on human cells (25), and is cross-reactive with a similar complex on canine cells (26). These antibodies were used in conjunction with complement for negative selection of peripheral blood mononuclear cells to be used in NK assays.All antibodies were used in the form of ascites fluid, at dilutions of 1:1000. The peripheral blood mononuclear cells that were to be treated were first adjusted to a concentration of 10 X lo6 cells/ml. Under intermittent shaking, cells were incubated for 30 min at 22°C with medium or monoclonal antibody at a volume ratio of 1:1, followed by incubation for 90 min with the same volume of complement (rabbit serum screened for HLA-DR typing, Pel FreezeRogers,Ark.), or medium. Cells were then washedonce with medium and counted for viability (Trypan blue exclusion). Effecters, adjusted to the original volume and not reconstituted to original cell number, were plated in the NK assay. Efficiency of lytic treatment was evaluated by the number of viable cells remaining.
210
LOUGHRAN,
DEEG, AND STORB
For Dly-1 this averaged 10%; for Dly-6 35%; for DT-2 60%; for E-l 1 86%; for IA1 44%; for WIG4 48%; and for 7.2 50%. RESULTS NK assay.Results utilizing peripheral blood mononuclear cells as effectersare shown in Fig. 1. Only minimal NK activity was observed with 4-hr assays.Eight- and 18-hr assaysproduced specific lysis of greater than 50%. NK activity peaked at an effector/ target ratio of 80/l. Spontaneous release for CTAC ranged from 5-17%, mean 12% at 4 hr, 9-l 8%, mean 12% at 8 hr, and 1l-30%, mean 20% at 18 hr. ADCC assay. Forty percent specific lysis was observed in a 4-hr assay utilizing peripheral blood mononuclear cells at an effecter/target ratio of 100/l. Optimum cytotoxicity was achieved using an antibody concentration of I/ 1000. There was no killing when uncoated targets were used. Enrichment of NK activity. Table 1 shows the cell yields resulting from Percoll discontinuous-gradient centrifugation of peripheral blood mononuclear cells as well as the percent of LGL in each fraction. The total number of cells recovered averaged 65% of the total input. Highest NK activity occurred in the low-density fractions, as shown in Fig. 2. The fractions with the highest NK activity corresponded to those with the highest proportion of LGL. Further enrichment of LGL and NK activity was obtained by lytic treatment of cells from fractions 2 and 3 with monoclonal antibody WIG4 plus complement, as shown in Fig. 2 (bar graph). Cells enriched in the highdensity fractions had the morphology of typical lymphocytes and had little, if any,
.
..
i3toctw
- Target
I
I
10
20
I
so
I
60
... I
100
Ratio
FIG. 1. NK assay results utilizing peripheral blood mononuclear cells.at various etTector/target ratios, showing mean ?&specificlysis after 4-hr (dotted line, n = 12), 8-hr (dashedline, n = 16),and IS-hr incubation (solid line, n = 22). Mean Sbspecificlysis for ADCC assay utilizing peripheral blond mononuclear cells is shown for target coated with no antibody (dotted line), and with I/ 100(dashedline), and l/ 1000(solid line) dilutions of antibody (n = 12). Standard errors of mean are shown in brackets.
211
CANINE NK CELLS TABLE 1
Cell Yields from Percoll Discontinuous-Gradient Centrifugation of Peripheral Blood Mononuclear Cells” Fraction
%LGL (M + SEM)
Unfractionated 1 2 3 4 5 6 7 2 + 3 (WIG4 + C’)*
6& 1.2 31 1- 4.4 44 f 1.7 41 f 5.7 21 + 4.2 13 -f: 3.0 10 + 1.8 6+ 1.0 78 + 6.2
%Input (M + SEM) 9.8 k 8.5 f 6.5 + 5.3 + 7.3 f 7.0 f 15.6 f -
1.7 0.9 0.7 0.1 0.9 1.2 1.6
’ LGL, large granular lymphocytes; M, mean of 12 experiments. ’ Fractions 2 and 3 combined and then treated with monoclonal antibody WIG4 plus complement; results represent the means + SEM of five experiments.
NK activity. Separation of peripheral blood mononuclear cells by nylon-wool columns into nonadherent population did not deplete NK activity when compared to unseparated peripheral blood mononuclear cells (data not shown). Morphology of NK cells. Figure 3 shows the binding of effector cells to CTAC target
20
70
50
d R 10 w--l
90 IO
FIG. 2. Resultsof 8-hr NK assaysusing various cell fractions of Percoll discontinuous gradientsof peripheral blood mononuclear cells as effecters at an effecter/target ratio of 20/ 1. Mean %specific lysis is shown using scale on left (solid line, n = 7) while mean WLGL in each fraction is shown using scale on right (dashed line, n = 12). Further enrichment in NK activity (open bar) and %LGL (solid bar) was obtained by lytic treatment of low-density fractions 2 and 3 with monoclonal antibody WIG4 plus complement (n = 5). Standard errors of mean for Ispecific lysis are shown in brackets, and for %LGL in Table 1.
212
LOUGHRAN,
DEEG, AND STORB
FIG. 3. Binding of LGL effector cells (arrow heads) to target cells (arrows) (Wright-Giemsa, original magnification X400).
cells, illustrating the characteristic LGL morphology of the effector cell with eccentric nucleus, relatively large cytoplasmic/nuclear ratio, and azurophilic granules. Figure 4 reveals the ultrastructural features of the canine LGL, showing multiple characteristic electron-dense granules. Tissue distribution of NK activity. As shown in Fig. 5, among the three populations examined, the highest NK activity was found in peripheral blood mononuclear cells, followed by spleen and bone marrow. Radiation sensitivity. Figure 6 depicts NK activity of peripheral blood mononuclear cells after in vitro radiation exposure. NK activity was relatively radioresistant, as demonstrated by 40% specific lysis remaining even after irradiation with 40 Gy. Phenotype of NK cells. Results are summarized in Figure 7. Lytic treatment of peripheral blood mononuclear cells with complement and monoclonal antibodies Dly- 1, Dly-6, 1Al, and E- 11, respectively, depleted NK activity, such that the remaining cytotoxicity ranged from only lo-34% of the complement controls. In most cases, this was also true of monoclonal reagent 7.2, although in some dogs lytic treatment with this antibody had no significant effect on cytotoxicity. As a result, cytotoxic activity remaining after such lytic treatment with 7.2 averaged50% of the complement controls. In contrast, treatment with complement and monoclonal antibodies DT-2 or WIG4, respectively, failed to reduce NK activity. Since both antibodies strongly activate complement, this suggeststhat canine NK cells do not express the antigens recognized by theseantibodies. None of the antibodies when added to the assayblocked NK activity in the absenceof complement treatment (data not shown).
CANINE NK CELLS
213
FIG. 4. Ultrastructural features of LGL, demonstrating typical electron-dense granules (arrow heads) (original magnification X8000).
DISCUSSION In the mouse, rat, and man it has been shown that NK function resides in a subpopulation of large lymphocytes characterizedby the presenceof prominent azurophilic granules and consequently termed LGL (27-29). Several investigators have shown that mononuclear cells possessingNK function can be distinguished by phenotyping with specific monoclonal antibodies (30-33). We have shown in the present study that canine NK activity resides in a subpopulation of peripheral blood mononuclear cells that have the morphology of LGL and are nonadherent, low-density cells, which is similar to findings in other species(27-29). Ultrastructural features were similar to those observed in human LGL (34). The phenotype of canine NK cells was Dly-l+, Dly-6+, lAl+, E-l l+, DT-2-, and WIG4-. Our studies confirm the results reported by Krakowka in defining the suitability of CTAC as a cell target in the NK assay(14). Eight- and 18-hr assaysresulted in easily demonstrated killing, with acceptable values for spontaneous release. In agreement with results in other species(35), cells recognized as having NK function (LGL morphology) also functioned as effector cells in ADCC. Furthermore, enrichment of ADCC activity also occurred in low-density fractions, suggesting that canine LGL are responsible for both NK and ADCC activity.
214
LOUGHRAN,
10
20
DEEG, AND STORB
50
80
100
EMCtCf-TarpstFMiO
FIG. 5. Mean Zspecific lysis (brackets indicate standard error of the mean) in 18-hr NK assayby canine mononuclear cells of different origin at various effecter/target ratios. Peripheral blood mononuclear cells are represented by solid line (n = 12), spleen cells by dashed line (n = 5), and marrow cells by dotted line (n = 4).
Canine NK activity was studied in peripheral blood, splenic, and marrow mononuclear cells, with peripheral blood mononuclear cells having highest cytotoxicity and marrow mononuclear cells least. Splenic mononuclear cells appeared to be intermediate; however, these results were obtained using frozen cells as effecters since fresh
10
20
50
80
100
Effecter-TmwtFtath FIG. 6. Effect of in vitro irradiation on 18-hr NK assaysat various effecter/target ratios utilizing peripheral blood mononuclear cells as effecters(n = 6). Mean Ispecific lysis with unirradiated cells (control) is depicted by open circles, 20 Gy by solid circles, 40 Gy by open squares,and 80 Gy by solid squares.Standard errors of mean ranged from 9.8 to 16.7%for mm-radiated cells, from 10.1 to 15.3%at 20 Gy, 10.2 to 16.2%at 40 Gy, and 9.9 to 17.7%at 80 Gy.
215
CANINE NK CELLS
r
ldcmelml
Antibodiis
FIG. 7. NK activity (I 8-hr assay)remaining after lytic treatment of peripheral blood mononuclear cells with monoclonal antibodies plus complement, (expressedas % of complement controls). Results represent mean values of combined experiments performed at ratios of 50/ 1 or 80/I (n = 12 for antibodies Dly- I and Dly-6; n = 8 for lA1, E-l 1 and DT-2; n = 22 for 7.2; and n = 4 for WIG4).
cells were not available. Freezing of canine peripheral blood mononuclear cells resulted in a decreaseof approximately 33% in NK activity (data not shown). Therefore, cytotoxicity might have been greater if freshly isolated splenic mononuclear cells had been used as effecters. The ontogeny of human NK cells is uncertain, with T-cell (36) myeloid (37, 38), and separatelineage (39,40) having all been postulated. Our results suggestthat canine NK cells share phenotypic characteristics of canine T cells. Experiments utilizing lytic treatment of peripheral blood mononuclear cells with complement and a panel of monoclonal reagents indicated that canine NK cells, as defined in vitro, express cell surface antigens defined by monoclonal antibodies Dly- 1, Dly-6, 1A 1, and E- 11, and are negative for DT-2 and WIG4. The presenceof E- 11, a marker which identifies a canine T-lymphocyte subset functionally equivalent to CD8 + (T8) human lymphocytes [i.e., suppressor/cytotoxic T-cell subset (23)] on canine NK cells is similar to human NK cells, a subset of which coexpressessuppressor cell and NK-cell markers (30, 33, 41, 42). In contrast, NK cells did not expressthe antigen recognized by DT2 which identifies a canine T-lymphocyte subpopulation functionally equivalent to CD4 + (T4) human lymphocytes (i.e., T-helper subset)and which is otherwise expressed on a subset of T cells complementary to the E-l l+ subset. The fact that WIG4 (normally expressed on monocytes) did not recognize NK cells suggeststhat the canine NK lineage is different from that of the monocyte. This is further supported by the observation that elimination of NK activity could be accomplished using lytic treatment with Dly-6, an antibody not reactive with monocytes. Together with the E-l 1 and 1A 1 results, this would seemto be in agreement with previous findings in other species suggesting a common lineage for NK cells and cytotoxic T cells (20, 43). Also in agreement with this, Ringler and Krakowka have recently shown that canine NK cells expressthe canine T-lymphocyte marker, Thy-l (44).
216
LOUGHRAN,
DEEG,
AND
STORB
Our data also suggestthat canine NK cells expressIa antigens, although these results were variable. Several investigators have reported that a subset of human NK cells expressesIa antigens (30, 41), others have not confirmed ths finding (32). Recently, human NK cells generated in mixed lymphocyte cultures were shown to expressclass II antigens (45). The finding of classII antigen expression on NK cells may be relevant in regardsto their putative role in the resistanceto hemopoietic grafts (lo- 13). Gambel et al. have shown that pretreatment of mice with anti-class II monoclonal antibodies before the infusion of haploidentical donor marrow allows for the development of lymphohemopoietic chimerism (46). We have shown recently that treatment of dogs with anti-Ia antibody in vivo prior to transplantation facilitates engraftment of marrow from donors differing for both DLA haplotypes (47). Sincethe distribution of Ia antigens on canine lymphocytes is different than that of human lymphocytes (27), it remains to be determined whether this effect is due to inactivation of activated canine NK cells expressingIa antigens or to another host population that also expressesIa antigens. It is of note, however, that, similar to the inconsistent elimination of NK activity in vitro, the in vivo administration of anti-class II antibody (7.2) did not uniformly allow for sustained engrafiment (47). We are presently investigating whether there is a correlation between in vivo and in vitro results. Furthermore, the observation that NK activity was radioresistant may have implications in this model of histoincompatible marrow transplantation, since it has been shown in mice that genetic resistance to hemopoietic grafts is relatively radioresistant (10, 48). Furthermore, we have shown in our canine model that resistance to marrow engraftment can be overcome by highdose fractionated total-body irradiation (49), a scheme of irradiation similar to that which produces NK-cell deficiency in mice (50). Further experiments utilizing the phenotypic data presented here may be helpful in elucidating the role of NK cells in marrow graft resistance in the dog. ACKNOWLEDGMENTS We thank Susan DeRose and Liz Caldwell for excellent technical assistanceand Dr. P. Martin, Dr. W. Ladiges, and Dr. D. R. Krawiec for providing antibodies 7.2, WIG4, and lA1, respectively. We appreciate Jan Freedle’s help in preparing the manuscript.
REFERENCES 1. Herberman, R. B., Djeu, J. Y., Kay, H. D., Ortaldo, J. R., Riccardi, C., Bonnard, G. D., Holden, H. T., Fagnani, R., Santoni, A., and Puccetti, P., Immunol. Rev. 44,43, 1979. 2. Herberman, R. B., and Ortaldo, J. R., Science (Washington, D.C.) 214,24, 1981. 3. Hansson, M., Beran, M., Andersson, B., and Kiessling, R., J. Immunol. 129, 126, 1982. 4. Mangan, K. F., Hartnett, M. E., Matis, S. A., Winklestein, A., and Abo, T., Blood 63, 260, 1982. 5. Tilden, A. B., Abo, T., and Balch, C. M., J. Zmmunol. 130, 1171, 1983. 6. Arai, S., Yamamoto, H., Itoh, K., and Kumagai, K., J. Immunol. 131,651, 1983. 7. Abruzzo, L. V., and Rowley, D. A., Science (Washington. D.C.) 222,581, 1983. 8. Brieva, J. A., Targan, S., and Stevens, R. H., J. Immunol. 132,611, 1984. 9. James, K., and Ritchie, W. S., Immunol. Today 5, 193, 1984. IO. Kiessling, R., Hochman, P. S., Hailer, O., Shearer, G. M., Wigzell, H., and Cudkowicz, G., Eur. J. Immnol. I, 655, 1977. 1I. Miller, S. C., J. Immunol. 131, 92, 1983. 12. Warner, J. F., and Dennert, G., Nature (London) 300, 31, 1982. 13. Lotzova, E., Pollack, S. B., and Savary, C. A., In “NK Cells and Other Natural Effector Cells” (R. B. Herberman, Ed.), pp. 1535-l 540. Academic Press,New York, 1982.
CANINE NK CELLS
217
14. Krakowka, S., Amer. .I. Vet. Res. 44, 635, 1983. 15. Deeg, H. J., Storb, R., Gerhard-Miller, L., Shulman, H. M., Weiden, P. L., and Thomas, E. D., Trunsplantation 29, 230, 1980. 16. Kasza, L., Amer. J. Vet. Res. 25, 1178, 1964. 17. Krakowka, S., Olsen, R., and Cockerell, G., In Vitro 13, 119, 1977. 18. Timonen, T., and Saksela,E., J. Immunol. Methods 36,285, 1980. 19. Timonen, T., Ortaldo, J. R., and Herberman, R. B., J. Immunol. 128,25 14, 1982. 20. Brooks, C. Cl., Kuribayashi, K., Sale, Cl. E., and Henney, C. S., J. Immunol. 128,2327, 1982. 21. Wulff, J. C., Durkopp, N., Aprile, J., Tsoi, M.-S., Deeg, H. J., and Storb, R., Exp. Hematol. 10, 535, 1983. 22. Wulff, J. C., Deeg, H. J., and Storb, R., Transplantation 33, 616, 1982. 23. Szer, J., Deeg, H. J., Sevems, E., and Storb, R., Transplantation 39, 187, 1985. 24. Krawiec, D. R., and Muscoplat, C. C., Amer. J. Vet. Res. 45,491, 1984. 25. Hansen, J. A., Martin, P., and Nowinski, R. C., Immunogenetics 10,247, 1980. 26. Deeg, H. J., Wulff, J. C., DeRose, S., Sale, G. E., Braun, M., Brown, M. A., Springmeyer, S. C., Martin, P. J., and Storb, R., Immunogenetics 16,445, 1982. 27. Kumagai, K., Itoh, K., Suzuki, R., Hinuma, S., and Saitoh, F., J. Immunol. 129, 388, 1982. 28. Reynolds, C. W., Timonen, T., and Herberman, R. B., J. Immunol. 127,282, 1981. 29. Timonen, T., Ortaldo, J. R., and Herberman, R. B., J. Exp. Med. 153,569, 1981. 30. Ortaldo, J. R., Sharrow, S. O., Timonen, T., and Herberman, R. B., J. Immunol. 127, 2401, 1981. 31. Abo, T., and Balch, C. M., J. Immunol. 127, 1024, 1981. 32. Perussia,B., Starr, S., Abraham, S., Fanning, V., and Trinchieri, G., J. Immunol. 130, 2 133, 1983. 33. Lanier, L. L., Le, A. M., Phillips, J. H., Warner, N. L., and Babcock, G. F., J. Immunol. 131, 1789, 1983. 34. Grossi, C. E., Cadoni, A., Zicca, A., Leprini, A., and Ferrarini, M., Blood 59, 277, 1982. 35. Bradley, T. P., and Bonavida, B., J. Immunol. 129, 2260, 1982. 36. Grossman,Z., and Herberman, R. B., In “NK Cells and Other Natural Effector Cells” (R. B. Herberman, Ed.), pp. 229-238. Academic Press,New York, 1982. 37. Kay, H. D., and Horowitz, D. A., J. Clin. Invest. 66, 847, 1980, 38. Zarling, J. M., and Kung, P. C., Nature [London) 288,394, 1980. 39. Ferrarini, M., and Grossi, C. E., In “NK Cells and Other Natural Effector Cells” (R. B. Herberman, Ed.), pp. 257-264. Academic Press,New York, 1982. 40. Ortaldo, J. R., In “NK Cells and Other Natural Effector Cells” (R. B. Herberman, Ed.), pp. 265-271. Academic Press,New York, 1982. 41. Abo, T., Cooper, M. D., and Balch, C. M., J. Immunol. 129, 1752, 1982. 42. Perussia, B., Fanning, V., and Trinchieri, G., J. Immunol. 131,223, 1983. 43. Fast, L., Beatty, P., Hansen, J. A., and Newman, W., J. Immunol. 131,2404, 1983. 44. Ringler, S. S., and Krakowka, S., Vet. Immunol. Immunopathol., 1985, in press. 45. Phillips, J. H., Le, A. M., and Lanier, L. L., J. Exp. Med. 159, 993, 1984. 46. Gambcl, P., Francescutti, L. H., and Wegmann, T. G., Transplantation 38, 152, 1984. 47. Deeg, H. J., Storb, R., Szer, J., Appelbaum, F. R., Hackman, R. C., and Thomas, E. D., Transplant. Proc. 17,493, 1985.
48. Rauchwerger, J. M., Gallagher, M. T., Monie, H. J., and Trentin, J. J., Transplantation 23, 158, 1977. 49. Deeg,H. J., Storb, R., Shulman, H. M., Weiden, P. L., Graham, T. C., and Thomas, E. D., Trunsp~unt~tion 33,443, 1982. 50. Parkinson, D. R., Brightman, R. P., and Waksal, S. D., J. Immunol. 126, 1460, 1981.