Eur. J. Immunol. 1990.20: 244-2449 HLA expression and NK lysis of tumor cells 2445

Jose PeiiaoA, Corona Alonsoo, Rafael SolanaoA, Rafael SerranoO, Julia Carracedon and Rafael RamirezO

Department of Biochemistryo, Faculty of Medicine and Immunology UnitA, Reina Sofia Hospital, University of Cbrdoba, C6rdoba

Natural killer susceptibility is independent of HLA class I antigen expression on cell lines obtained from human solid tumors*

The susceptibility to natural killer (NK) cell-mediated cytotoxicity of 20 cell lines obtained from human solid tumors and their class I histocompatibility antigen (HLA) levels were studied in an attempt to determine whether major histocom- patibility complex (MHC) products expressed on cells derived from human solid tumors influence NK susceptibility. The effect of interferon-y (IFN-y) treatment on these elements was also analyzed.The MHC class I (HLA-ABC, HLA-A and HLA-B) antigen levels and degree of NK lysis were very heterogeneous and no correlation was found on comparison. After treatment with IFN-y a marked decrease in NK susceptibility was observed in all the cell lines, including the control line K-562. However, the level of HLA class I expression was not modified in any of the lines with the exception of the K-562, which increased. In some cell lines the expression of HLA class I-like antigens, CDla, b and c, was also measured before and after IFN treatment; however, no correlation was found between CD1 levels and NK susceptibility. Consequently, from our results it is possible to conclude that HLA class I antigens do not play a decisive role in NK susceptibility of cell lines derived from human solid tumors and to suggest that molecules which are not HLA class I antigens but IFN-y inducible may confer NK resistance to these lines.

1 Introduction

The structures involved in NK cytotoxicity are still un- known at the molecular level. Several molecules have been reported to play a role in the interaction between the target and effector cells [l-31. However, their precise nature and functions have not yet been clearly defined.

NK-mediated lysis is not restricted by MHC products, contrary to that which occurs with theTcR a@ T cells which recognize the antigens only when they are presented by these molecules. Several authors have suggested that MHC class I antigens participate in the targetleffector NK cell interaction, thus reducing the susceptibility of the target cells to NK-mediated cytolysis. Nevertheless, there is still a great deal of controversy, as the influence of these antigens has been demonstrated only for some systems. In murine models, a relationship between low NK susceptibility and a high level of MHC class I antigens has been reported in EBV-transformed lymphoblastoid cell lines [4] and YAC-1 [5] and RBL-5 [6] lymphomas, but no relationship was found with B16 melanoma [7], lung carcinoma [8], hepa- toma [9] and embryonal carcinoma [lo]. In human target cells contradictory results have also been reported. A correlation was found in EBV-transformed B cell lines [ l l , 121 and other lymphoblastoid-derived cell lines [13], whereas no correlation was found on human small cell lung carcinomas [ 141.

[I 84981 * This work was supported by grants from the CICYTand FIS

(Spain). Present address: Department of Immunology, London Hospital Medical College, London, GB.

Correspondence: Jose Peiia, Faculty of Medicine, Department of Biochemistry, University of Cbrdoba, 14004 Cbrdoba, Spain

IFN-y treatment has been found to reduce NK susceptibil- ity of different target cells [15-201; however, the mechan- isms and molecular structures involved are unknown. The decreased susceptibility observed has been linked to an increased MHC antigen expression in some experimental models [5,21], although this cannot be applied when other cell lines are used as targets [22, 231.

To analyze the afore-mentioned discrepancies further the aim of this study was to investigate the influence of HLA class I antigen expression on NK lysability of target cells obtained from human solid tumors. The NK susceptibility before and after treatment with IFNy of cell lines express- ing different levels of HLA antigens was examined. Fur- thermore, in an attempt to determine whether the expres- sion of HLA class I-like (CD1) antigens on target cells have any effect on NK lysis, the levels of CDla, b and c were also considered.

2 Materials and methods

2.1 Cell l i e s

Twenty cell lines derived from human tumors, obtained and characterized in our laboratory [23,24], were used.Ten cell lines were derived from primitive brain tumors (seven gliomas and three meningosarcomas); five were brain metastases of different origins (one uterus fibrosarcoma, one melanoma, one colonic adenocarcinoma and two lung adenocarcinomas). Another five cell lines were obtained from pleural metastasis of four lung adenocarcinomas and one breast adenocarcinoma. The HLA class I-negative K-562 cell line was used as the conventional NK target cell. Cell lines were grown in WMI 1640 supplemented with 10% heat-inactivated FCS, 2 mM L-glutamine, 100 U/ml penicillin and 50 pg/ml gentamycin (all reagents from Flow Labs., Baar, Switzerland) at 37°C in 5% CO2. Cell lines

0 VCH Verlagsgesellschaft mbH, D-6940 Weinheim, 1990 0014-2980/90/1111-2445$3.50+ .25/0

2446 J. PeAa, C. Alonso, R. Solana et al. Eur. J. Immunol. 1990. 20: 2445-2449

RS-1, RS-3, RS-6, CAC and CAP2 wee treated for 48 h with 100 UI/ml of IFNy (Sigma, Poole, GB).

then collected, washed extensively with PBS and resus- pended in culture medium at a concentration of 5 x lo6 cell/ml.

2.2 mAb and indirect immunofluorescence

The following mAb were used: W6/32 (anti-HLA-ABC), IOT6 (anti-CDla), IOT6b (anti-CDlb), M241 (anti- CDlc), BBM-1 (anti-P2-microglobulin), 108-2C5 (anti- HLA-A), Joan-1 (anti-HLA-B), L243 (anti-HLA-DR), Tu-22 (anti-HLA-DQ) and B7.21 (anti-HLA-DP). Cells (5 x lo5) were incubated in 50 p1 PBS containing 0.02% sodium azide and 0.2% BSA (PBS/BSA/azide) for 1 h at 4°C with each mAb at the optimal dilution. Cells were then washed twice and resuspended in 50 pl of the same buffer. The suspension was treated with goat anti-mouse IgG conjugated to fluorescein isothiocyanate (FITC-GAM) at a final dilution of 1/15 and incubated for another 20 min at 20°C. Cells were washed twice in PBS/BSA/azide and examined with a cytofluorimeter (Epics Profile, Coulter, Harpenden, GB).The results are reported as percentage of positive cells and as the mean of fluorescence channel (MFC). Units of MFC are arbitrary (0 to 100) assigned to a logarithmic scale of fluorescence channels between 1 to 1024.

2.3 NK assay

Target cells (4 x loh) were incubated with 250 yCi (= 9.25 MBq) Nal5lCr04 (1 mCi/ml, CEA-OFUS, Saclay, France) for 1 h in 0.5 ml of culture medium at 37°C in 5% C02. After washing four times, target cells were resus- pended at a concentration of 5 x 104 celVml. The cytotox- icity assay was performed in triplicate in 96-well round- bottom microtiter plates (Greiner, Nurtingen, FRG) by adding effector cells to 5 x lo3 target cells producing Em ratios of 100/1,50/1 and 254 in a final volume of 200 pl. For spontaneous release control, samples of target cells were resuspended in culture medium alone. After a 4-h incuba- tion period at 37°C in 5% COl, 100 yl of SN was removed by centrifugation of the plate at 500 x g for 5 min and counted in a gamma counter (Ultrogamma LKB, Bromma, Sweden). Maximum release of W r from the target cells was obtained by mixing 5 x lo3 labeled target cells obtained by mixing 5 x lo3 labeled target cell in 0.1 ml of 10% NP40 (Sigma) in culture medium. Background release was usually < 10% of total uptake. Specific 51Cr release percentages were calculated with the following equation:

Exp. release - background release Total release - background release Specific release = x loo.

Effector cells were prepared from Ficoll-Hypaque 2.4 Binding assay (Pharmacia, Uppsala, Sweden)-buoyant PBMC, depleted of monocytes by incubation in 75-cm2 plastic culture flasks (Costar Europe, Badkoevedorp,The Netherlands) for 1 h at 37°C in 5% COa. The resulting nonadherent cells were

Cell lines were tested for their capacity to form conjugates with PBL. Briefly, 100 pl of lymphocytes (1 x 106/ml) and 100 yl of tumor cells (1 x 10Vml) were mixed, centrifuged

Table 1. HLA class I antigen expression on tumor cell l ined

HLA-A HLA-B HLA-ABC Name Origin (Yo) MFC (%) MFC (Yo) MFC

Brain primitive tumors RS-1 Astrocytoma IV RS-2 Astrocytoma IV RS-3 Astrocytoma 11 RS-4 Astrocytoma IV RS-5 Astrocytoma 111 RS-6 Astrocytoma IV RS-7 Mengingosarcoma MS-1 Mengingosarcoma MS-I Mengingosarcoma MS-3 Mengingosarcoma Brain metastasis CAC Colon CAP- 1 Lung CAP-2 Lung MELA Melanoma FISAR Unterus fibros. Pleural exudatcs MP- 1 Breast PP- 1 Lung PP-2 Lung LP-132 Lung LP-305 Lung

K-562 Eryt hrole ukemia

2 0 3 0 3 0

2 0 2 0 2 0 7 - 0 3 0 2 0

34 11 41 13 37 10

2 0 3 0 2 0 2 0 2 0 2 0 3 0 3 0 2 0

s7 1s 48 16 61 16

3 0 3 0 2 n

2 0 3 0 4 n

43 12 45 10 51 14 7 1 2n 62 22 SY 18 78 26 71 22 71 24 83 42 7') 45 82 44 2 0 2 0 3 n

3 0 3 0 4 0 42 24 37 26 3s 24 25 16 22 15 27 12 21 10 2s 9 23 10 1s 8 23 8 10 9

2 0 3 0 2 0 a) %: Percentage of positive cells; MFC: mean fluorescence channel.

Eur. J. Immunol. 1990. 20: 2445-2449 HLA expression and NK lysis of tumor cells 2447

at 120 x g for 5 min and incubated at 37°C for 1 h. The percentage of lymphocytes bound to target cells was determined by counting at least 200 lymphocytes after gently resuspending them.

3 Results

3.1 NK susceptibility is independent of the level of HLA class I expression on solid tumor cell lines

HLA-ABC, HLA-A and HLA-B antigen levels on the tumor cell surface are shown in Table 1 and ranged from nondetectable to high levels. No significant differences between HLA-A and HLA-B expression were observed. Interestingly, cell lines derived from primitive brain tumors did not express HLA class I antigens, with the exception of the RS-5 and MS-2 lines, whereas those derived from metastasis and pleural exudate expressed high levels with only one exception in each group. NK susceptibility of these cell lines is also very heterogeneous as shown in Fig. 1. When the level of HLA class I antigen expression on the target cells is compared to their NK lysis no relationship is found. Cell lines expressing high levels of HLA antigens can be either NK resistant or NK susceptible. Likewise, those expressing nondetectable levels can also be resistant or susceptible to NK cytotoxicity.

Cel I type MFC

K562 0 RS-1 0 RS-2 0 RS-3 0 RS-4 0 RS-6 0 RS-7 0 MS-1 0 MS-2 0 RS-5 10 MS-3 16 FlSAR 0 CAC 14 CAP 1 18

CAP 2 24 MELA 44

M P l 0 PP 1 24 PP 2 12 LP 132 10 LP 305 9

Fig. 1 also shows the capacity of the target cells to bind NK effector cells which was analyzed to determine the influence of HLA expression in the initial steps of NK lysis. These results suggest that the capacity of tumor cells to bind to effector cells is not related to the levels of HLA class I expression.When the conjugate formation and NK lysis are compared, a correlation between both parameters is found, indicating that in these cell lines binding is a major event determining NK programming for lysis.

3.2 NK susceptibility decrease in the brain tumor cell lines after IFN-y treatment without detectable changes in HLA expression

Cell lines were treated with IFN-y and HLA class I antigen expression was studied before and after treatment. As shown in Fig. 2, IFN-y does not induce any changes in HLA antigen expression in cell lines from solid tumors although, as has been previously reported, a marked increase is observed in K-562 cells. Cell lines treated with IFN-y had a decreased susceptibility to NK lysis (Fig. 3), indicating that in these tumor cell lines the reduced susceptibility to NK is not related to HLA antigen expression. However, treat- ment of K-562 cells with IFN-y decreases NK susceptibility and induces HLA class I antigen expression.

3.3 HLA class I-like (CD1) expression is not involved in NK target susceptibility

Considering the possibility that NK susceptibility could be related to the expression of HLA class I-like antigens such

Class I (MFC)

60 -

40 -

0 20 40 60

~b NK lysis E B ~b Conjugates

Figure 1. NK susceptibility and HLA-ABC expression of human tumor cell lines. PBL was used as effector cells; the Emrat io was

0 K562 RS-1 US-3 RS-7 CAC CAP 2

100/1. The HLA-ABC level was measured by reactivity with the mAb W6/32 and expressed as the mean fluorescence channel (MFC).The percentage of Enconjugate formation at 37°C for 1 h is shown. The human cell lines used were derived from primitive brain tumor, and brain and pleural metastasis. K-562 cells were used as conventional targets.

Tumor cell lines

Figure 2. HLA-ABC expression represented as MFC of brain primitive (RS-1, RS-3 and RS-7) and metastatic (CAC and CAP-2) tumor cell lines before (B) and after 48-h treatment (a) with 100 UI/ml of IFN-y.

2448 J. Peiia, C. Alonso, R. Solana et al. Eur. J. Immunol. 1990. 20: 2445-2449

Table 2. Expression of CDla, CDlb and CDlc on cell lines before and after IFN-y treatment

Cell line CDla CDlb CDlc (%) MFC (%) MFC (%) MFC

K-562 8 5 2 4 2 4 K-562 + IFN 9 6 4 5 2 4 RS-2 4 5 . 5 5 4 5 RS-2 + IFN 7 6 3 4 4 5 RS-5 2 3 4 5 2 3 RS-5 + IFN 9 7 7 6 5 5 RS-7 5 4 9 6 6 6 RS-7 + IFN 18 8 8 6 12 8 CAC 5 5 2 3 2 2 CAC + IFN 2 3 6 5 4 5 CAP-2 7 5 4 5 4 5 CAP-2 + IFN 2 4 3 4 2 4 Thymocytes 68 25 65 28 72 25

% NK lyslr

"1

K562 RS-1 RS-3 RS-7 CAC CAP 2 HLA lnductlon

Figure 3. NK lysis (%) brain primitive (RS-1, RS-3 and RS-7) and metastatic (CAC and CAP-2) tumor cell lines before (m) and after

) with 100 UI/ml of IFN-y.

as CD1, the expression of these molecules was studied in some target cells before and after IFN treatment. Results in Table 2 show that the expression of these antigens is very low and not inducible with the doses and times used. Thus, we can conclude that at least in this experimental model, the levels of CD1 antigens do not significantly participate in the NK resistance induced after IFN-y treatment.

4 Discussion

Our results show no relationship between HLA class I antigen expression and NK susceptibility of target cells

derived from solid tumors. Whereas this data agrees with observations made by some authors [7-9,13-151, it is not in accord with others [4-6,10, 111. A possible explanation for the discrepant results could be related to the nature of the tumors used. Thus, when target cells from hematopoietic origin are used, a significant influence of the MHC class I products on NK susceptibility is found while in those experiments performed on cell lines derived from solid tumors, such as ours, this relationship is not generally found.

To determine whether HLA antigens affect the adhesion capacity of cells involved in NK activity, the conjugate formation between the tumor and effector cells was also studied.The results show that this capacity correlates to the NK susceptibility of the tumor, indicating that in these cell lines the target/effector cell binding is a major event determining NK programming for lysis. However, no relationship between HLA class I antigen expression on tumor cells and their capacity to form conjugates with the NK effector cells was found. This suggests that the capacity of tumor cells to bind to effector cells is not related to HLA class I expression.

Most probably, the fact that there is an inverse correlation between HLA and NK susceptibility when target cells of hematopoietic origin are used could be due either to a coordinated genetic regulation between MHC class I expression and the NK target structures or to a physical interaction between these molecules on the hematopoietic cells. However, the association between high HLA antigen levels and increased resistance to NK lysis cannot be generalized as this only occurs when certain target cells are used. This supports previous findings that NK cells interact with target cells using several surface molecules and their corresponding ligands [25] which can be differentially expressed on tumor cells of different origin. Our results suggest a relationship between NK susceptibility and the degree of tumor cell differentiation which is independent of HLA class I antigen expression, i.e. grade I11 and IV showing susceptibility and grade I1 being resistant.This is in agreement with previous results, although a greater num- ber of cell lines should be considered to clarify this point further.

IFN-y treatment of different target cells has been shown to induce a partial resistance to NK lysis [5,6,15-231. It is well known that IFN-y induces an increased expression of MHC class I antigens on a numbr of cells [26,27]; thus it has been assumed that this increased expression may play a role on NK resistance. However, we cannot confirm this hypothesis as IFN-y failed to induce changes in the levels of HLA antigen expression on the cell lines studied, although a significant reduction of NK susceptibility was observed in all cases. These results indicate that the decrease in NK susceptibility after treatment with IFN-y is not related to HLA changes and suggest that this treatment could inter- fere with NK susceptibility through a mechanism which is not HLA dependent.Thus, structures other than the HLA class1 antigens and which are induced by IFN-y are probably involved in NK resistance, as has been previously suggested [28-301.

Other molecules, such as CD1, Qa or TL, have been reported to be recognized by non-MHC-restricted cyto-

Eur. J. Immunol. 1990. 20: 2445-2449 HLA expression and NK lysis of tumor cells 2449

toxic cells [31-331. Thus, in an attempt to determine whether these structures could be related to NK resistance, the expression of these antigens (CDla, b and c) on target cells was measured before and after IFN-y treatment. Our results do not show a clear correlation between CDl expression and NK susceptibility, indicating that in the cell lines tested, these molecules do not play a major role in the determination of NK resistance or susceptibility.

Hence, from our results, it is possible to conclude that given that no influence of HLA class I antigen expression on NK susceptibility on cell lines was found, an unknown mole- cule, other than HLA class I antigens and which is I F N y inducible, participates in the reduction of NK susceptibility of the target cells. In any case further studies, including an extensive panel of other structures, need to be carried out to characterize this molecule.

We thank Drs. R. Vilella and J. Vives for providing mAb Joan-I and 108-2C5, and Prof. C. Milstein for mAb M241.

Received April 22, 1990; in revised form July 16, 1990.

5 References

1

2

3

4

5

6

7

8

9

10

Lanier, L. L., Phillips, J. H., Hackett, J., Tutt, M. and Kumar, V., J. Immunol. 1986. 137: 2735. Savary, C. A. and Lotzova, E., in Lotzova, E. and Herberman, R. B. (Eds.), Immunology of Natural Killer cells, CRC Press, Boca Raton 1986, p. 46. Anderson, P., Caligiuri, M., Ritz, J. and Schlossman, S. F., Nature 1989. 341: 159. Ohlen, C., Bejarano, M. T., Gronberg, A., Torsteinsdottir, S., Fransksson, L., Ljunggren, H. G., Klein, E., Klein, G. and Karre, K., J. Immunol. 1989. 142: 3336. Piontek, G. E.,Taniguchi, K., Ljunggren, H., Gronberg, A., Kiessling, R., Klein, G. and Karre, K., J. Immunol. 1985.135: 4281. Karre, K., Ljunggren, H. G., Piontek, G. and Kiessling, R., Nature 1986. 319: 657. Gorelik, E., Gunji,Y. and Herberman, R., J. Immunol. 1988. 140: 2096. Dennert, G., Landou, C., Lord, E. M., Bahler, D. W. and Frelinger, J. G., J. Immunol. 1988. 140: 2472. Nishimura, M., Stroynowski, I., Hood, L. and Ostrand- Rosenberg, S., J. Immunol. 1988. 141: 4403. Stern, P., Gidlund, M., Om, A. and Wigzell, H., Nature 1980. 285: 341.

11 Harel-Bellan, A., Quillet, A., Marchiol, C., Demars. R. ,Tursz, T. and Fradelizi, D., Proc. Natl. Acad. Sci. USA 1986. 83: 5688.

12 Shimizu, Y. and DeMars, R., Eur. J. Immunol. 1989. 19: 447.

13 Storkus,W. J., Howell, D. N., Salter, R. D., Dawson, J. R. and Cresswell, P., J. Immunol. 1987. 138: 1657.

14 Storkus,W. J., Alexander, J., Payne, J. A., Cresswell. P. and Dawson, J. R., J. Immunol. 1989. 143: 3853.

15 Storkus,W. J., Alexander, J., Payne, A. J., Dawson, J. R. and Cresswell, P., Proc. Natl. Acad. Sci. USA 1989. 86: 2361.

16 Starn,N. J.,Kast,W. M.,Voordouw,A. C.,Pastoors, L. B.,Van der Hoeven, F. A., Melief, C. J. M. and Ploegh, H. L.. J. Immunol. 1989. 142: 4113.

17 Trinchieri, G. and Santoli, D., J. Exp. Med. 1978. 147: 1314. 18 Wallach, D., J. Interferon Res. 1982. 2: 239. 19 Gronberg, A., Kiessling, R., Masucci, G., Guevara, L. A..

Eriksson, E. and Klein, G., Int. J. Cancer 1983. 32: 609. 20 Giuliani-Bonmassar, A., Graziani, L., Frati, L. and Bonmas-

sar, E., Int. J. Tissue React. 1984. 6: 35. 21 Karre, K., in Herberman, R. B. and Callewaert, D. (Eds.).

Mechanisms of Cytotoxicity by Natural Killer cells, Academic Press, Orlando 1985, p. 81.

22 Powell, J., Stone, J., Chan, W. C.,Yan, Z., Leatherbury, A.. Sell, K. W., Wiktor-Jedrzejczak, W. and Ahmed-Ansari, A.. Cell. Immunol. 1989. 118: 250.

23 Alonso, C., Serrano, R., Solana, R. and Peiia, J., Int. Arch. Allergy Appl. Immunol. 1989. 88: 71.

24 Serrano, R., Alonso, C., Solana, R., Monserrat, J., Muiioz, J. and Peiia, J., Immunol. Lett. 1989. 20: 311.

25 Herberman, R. B., Reynolds, C.W. and Ortaldo, J. R., Annu. Rev. Imrnunol. 1986. 4: 651.

26 Alonso, C., Navarrete, C., Solana, R.,Torres, A., Peiia, J. and Festenstein, H., J. Immunogenet. 1986. 13: 255.

27 Taniguchi, K., Peterson, M., Hoglund, P., Kiessling, R., Klein, G. and Karre, K., Proc. Natl. Acad. Sci. USA 1987.84: 3405.

28 Shchez-Madrid, F., Krensky, A. M.,Ware, C. F., Robbins, E., Strominger, J. L., Burakoff, S. J. and Springer,T., Proc. Natl. Acad. Sci. USA 1982. 79: 7489.

29 Chambers,W. H.,Vujanovic, N. L., De Leo, A. B., Olszowy, M. W., Herberman, R. B., Hiserodt, J. C., J. Exp. Med. 1989. 169: 1373.

30 Harris, D.T., Jaso-Friedrnann, L., Devlin, R. B., Koren, H. S.. Evans, D. L., J. Immunol. 1989. 143: 727.

31 Porcelli, S., Brenner, M. B., Greenstein, L., Balk, S. P., Terhorst, C. and Bleicher, B., Nature 1989. 341: 447.

32 Vidovic, D., Rogric, M., McKune, K., Guerder, S., McKay, C. and DembiC, Z., Nature 1989. 340: 646.

33 Koide, J., Rivas, A. and Engleman, E. G., J. Immunol. 1989. 142: 4161.