Immunobiol., vol. 183, pp. 23-39 (1991) 1 Department

of Molecular Genetics and Virology, 2 Department of Chemical Immunology, Department of Membrane Research, The Weizmann Institute of Science, Rehovot and 4 Department of Biological Chemistry, Institute of Life Science, The Hebrew University of Jerusalem, Jerusalem, Israel 3

Resistance to NK Cell-Mediated Cytotoxicity (in K-562 Cells) does not Correlate with Class I MHC Antigen Levels ZVI REITER 1, YORAM REITER2, ZVI FISHELSON2, MEIR SHINITZKy3, ABRAHAM KESSLER 3, AVRAHAM LOYTER\ OFER NUSSBAUM\ and MENACHEM RUBINSTEIN! Received August 13, 1990· Accepted in Revised Form April 2, 1991

Abstract Natural Killer (NK) cells probably function as an early line of defense against virus-infected cells and tumor cells. In all cases, the killing by NK cell-mediated cytotoxicity (NK-CMC) is not MHC-restricted and the factors which determine the sensitivity to NK-CMC have not yet been identified. A positive correlation between resistance to NK-CMC and the level of class I MHC antigen (MHC I) expression on target cells has been reported in many studies, and in some cases a functional linkage between the two has been claimed. Several other studies have shown that there is no such correlation. By employing several experimental systems, we demonstrate here a lack of correlation between the level of MHC I and the sensitivity of K-562 cells to NK-CMC. Transfer of MHC I to MHC I-negative cells via vesicles had no effect on their resistance to NK-CMC. In addition, a decrease in resistance to NK-CMC and increase of MHC I levels was observed following target-cell membrane modulation by both application of cholesterol and hydrostatic pressure. Finally, no correlation between sensitivity to NK-CMC and MHC I expression was found in three sublines of K-562 cells. Since NK-CMC is a multistage process, it is concluded that components other than class I MH C antigens have a more prominent role in modulating the sensitivity of target cells to NK-CMC.

Introduction

Cell surface class I MHC antigens (MHC I) playa key role in cell-to-cell interactions that regulate the immune response. Specifically, killing of virus-infected cells by cytotoxic T lymphocytes (CTL) is MHC-restricted and there is a correlation between the level of class I MHC expression on the virus-infected cell and it's sensitivity to CTL (1). NK cells are probably involved in protection against primary and metastatic tumor cells as well as killing of virus-infected cells. Unlike CTL, NK cells are not MHCAbbreviations: NK = Natural Killer; NKCF = NK Cytotoxic Factor; NK-CMC = NKCell Mediated Cytotoxicity; IFN = Interferon; MHC I = MHC class I antigens; P2m = f3r microglobulin

24 . Z. REITER et al.

restricted and the elements which confer specificity to the killing by NK cells have not yet been identified (2). Despite the lack of MHC restriction in NK-cell mediated cytotoxicity (NK-CMC), many studies suggested a possible correlation between resistance of a target cell to NK-CMC and the level of MHC I on cell membranes (3-7). This correlation was further established by studies with interferons (IFNs) which were shown to elevate MHC I levels on target cell surface and concomitantly increase their resistance to NK-CMC (8, 9). We have recently demonstrated a lack of such a correlation in one cell type. In this case IFN increased the resistance of Daudi cells to NK-CMC without induction of cell surface MHC I (10). In the present study, the correlation between MHC I levels and resistance to NK-CMC was further investigated in cell types routinely used for measuring NK-CMC. Several approaches were taken, including transfer of MHC I via fusion with membrane vesicles and modulation of the MHC I level by chemical or physical methods. In all cases, no correlation was found between MHC I level and resistance to NK-CMC.

Materials and Methods Target cells

K-562, human leukemic cells (ATCC CCL 243) were used as NK target cells and as a source of membrane proteins. Two sublines of the parental K-562 line were used in this study: a complement-resistant subline which was obtained by selection (K -562/CR; 11) and a complement-sensitive subline which appeared spontaneously during culture (K-562/S; 12). All cells were grown and assayed at 37 DC in a humidified 5°/c, CO 2 incubator in RPM I 1640 medium (GIBCO, Detroit, MI, USA) supplemented with fetal calf serum (10 %), glutamine (2 mM), penicillin (100 units/ml) and streptomycin (100 f,lg/ml) (RPMI 10 %). NKCF generation and assay of NKCF activity were performed in RPMI 1640 medium supplemented similarly but with 1 % FCS (RPMI-1 %). NK-target cells (2x 106/ml) were cultured in RPMI-10% in the presence or absence of IFN-y (1000 units/ml). Effector cells

NK effector cells were isolated from freshly collected peripheral blood of healthy donors by density gradient sedimentation of Ficoll-Paque (13). Adherent cells were removed by adsorption to nylon wool (14). The non-adherent cells were used as effector cells. Effector cells (5 x 106/ml) were cultured in RPMI-10 % in the presence (priming) or absence of IFN-aC (1000 units/ml) for 18 hat 37 DC. Cytokines and other reagents

Homogeneous E. coli IFN-aC (specific activity 109 units/mg) and homogeneous IFN-y from recombinant Chinese hamster ovary cells (specific activity 5 x 107 units/mg) were from InterPharm Laboratories Ltd, Nes-Ziona, Israel. Sodium [51 Cr]-chromate (specific activity 5-40 Ci/mmol) was from the Nuclear Research Center, Israel. Anti HLA-A, B, C, murine monoclonal antibody W6/32 was from Serotec, England. Fluorescein isothiocyanate (FITC)labeled goat anti-mouse IgG was from Bio-Makor, Israel, FITC-labeled rabbit anti-human ~z­ micro globulin (~2m) antibody was from DAKO-Immunoglobulins, Denmark. Octadecylrhodamine Bl chloride (R I8 ) was from Molecular Probes, USA. Ficoll-Paque was from Pharmacia, Sweden. Triton X-lOO (ocylphenoxy polyethoxyethanol), dithiothreitol (DTT),

MHC Class I Antigens and Resistance to NK Lysis . 25 phenylmethane sulfonyl fluoride (PMSF) and N a -2-hydroxyethyl piperazine-N-2-ethanesulfonic acid (hepes) were from Sigma, St. Louis, MO, USA. Macroporous polystyrene beads Bio-beads Bio-beads SM-2 (20/50 mesh) were from Bio-Rad, Richmond, CA, USA. Preparation and solubilization of cell membranes Cell membranes were prepared by a modification of the method described by MAEDA et al. (15). Plasma membrane-enriched fractions were prepared from at least 10 8 control or IFN-ytreated K-562 cells. All steps were carried out at 4 dc. Cells were washed twice with cold PBS, spun (150 x g, 10 min) and the pellet resuspended in a hypotonic Disruption Buffer (10 mM hepes, pH 7.4, 3 mM MgCb, 0.1 mM PMSF, 5 ml) for 15 min. Cells were then disrupted in a cold Dounce homogenizer by 35 strokes. The disrupted cell suspension was sedimented (500 x g, 5 min) and the supernatant was collected. The pellet still containing some intact cells was subjected once more to Disruption Buffer treatment, followed by homogenization and the supernatant was then collected. The supernatants were pooled, layered on top of a 42°/') sucrose cushion in Disruption Buffer and sedimented (100,000 x g, 60 min). The interface containing mainly plasma membranes was collected with a Pasteur pipette, suspended in Disruption Buffer (5 volumes) and sedimented (100,000 x g, 60 min). The pellet was resuspended in Disruption Buffer (1 ml, containing 15 % v/v glycerol) and kept at -70 DC until further processed. For solubilization of membrane proteins, the pellet was suspended by gently shaking for 30 min in Triton X-I00 (2 %) or CHAPS (15 mM) in Resolution Buffer (NaCI 0.1 M, hepes 50 mM, pH 7.4, PMSF 0.1 mM, 0.5 ml). The solubilized preparation of plasma membranes was sedimented (100,000 x g, 60 min) and the clear supernatant was collected. Solubilized plasma membranes were used immediately or stored at -70 dc. The amount of plasma membrane proteins obtained from 5 x 10 8 cells was 150-350 !Ag. The solubilized plasma membrane proteins were purified 13-fold from intact cells as determined by comparing the specific radioactivity of the proteins following surface radio labeling and sucrose gradient centrifugation. 125

1 surface labeling of cells

K-562 cells (5 x 10 8 cells) were washed twice and suspended phosphate-buffered saline (PBS; 0.5 ml). Sodium [125 I]iodide (100 !ACi), glucose (5 mg/ml, 50 !AI), glucose oxidase (0.5 mg/ ml, 25 !AI) and lactoperoxidase (1 mg/ml, 25 !AI) were added and the cells were kept for 10 min at room temperature. Tyrosine (1 mM, 100 !AI) was then added and after 2 min the cells were washed 3 times with PBS. Cell viability was 85-95 % following this treatment. About 11-18 % of the input radioactivity was found to be associated with the cells. Preparation and testing of fusogeneic vesicles Viral-plasma membrane (VPM) vesicles were prepared from solubilized plasma membranes and Sendai virus proteins at various proportions as previously described (16-20). The final yield of the VPM vesicles was 30--40 % based on protein content. Reconstituted Sendai virus envelopes (RSVE) were prepared from Sendai virus proteins alone and used as control vesicles. The precipitated co-reconstituted VPM vesicles and RSVE were resuspended in a Resolution Buffer to a protein concentration of 0.1--4 mg/ml and stored at -70 dc. The fusogenic vesicles were tested by their hemagglutination titer and hemolytic activity as determined by a modification of the method described by HOSAKA et al. (21). RSVE or VPM vesicles were fluorescently labeled with octydecylrhodamine Bj chloride as previously described (22). The extent of fluorescence (excitation at 560 nm, emission at 590 nm) was measured before and after solubilization with Triton X-I00 (1 %). The percentage of fluorescence dequenching was then calculated as previously described (22, 23). Fusion of vesicles with acceptor cells Acceptor cells were washed twice with cold PBS and incubated for 60 min at 4 DC with occasional shaking with various amounts of VPM vesicles or RSVE in Fusion Buffer (0.14 M

26 . Z.

REITER

et al.

NaCl, 20 mM Tris-HCI, pH 7.1, 3 mM KCI, 0.8 mM MgS0 4 , 200 [-II). CaClz (50 mM) was then added to the cells (final concentration 5 mM), the cells were incubated at 37°C for 30 min, washed twice with PBS, resuspended and kept in RPMI-10% at 37°C until used. 51Cr-release cytotoxicity assay This assay was performed as described (24). Briefly, target cells were labeled with sodium rsICr]-chromate (0.4 [-ICi/ml) for 1.5 hat 3rc in a humidified incubator with 5 % CO2 • The cells were washed twice with PBS, treated with reconstituted vesicles, washed again with RPMI-10%, suspended to a density of 2x10 5 cells/ml and dispensed into 96-well conical bottom microtiter plates (10 4 cells/well). Various amounts of effector cells were then added (in triplicates, final volume 150 [-II/well), the plates were spun (150 X g, 3 min) and incubated for 4 h at 37°C. Aliquots of the supernatant were then collected and their radioactivity measured. Total cpm was obtained by adding Triton X-100 (1 %) to wells with target cells alone. Spontaneous release (6--18 % of the total cpm) was also determined in wells containing target cells alone. Data are expressed as percent specific release according to the formula: % specific release = 100x (experimental cpm - spontaneous cpm)/(total cpm - spontaneous cpm). Each data point represents an average of a triplicate.

Conjugate formation assay Effector and target cells (8 x 104 and 4 x 104 cells, respectively) were mixed in RPMI-10 % (100 [-II) in 96-well plates. The plates were spun (150 x g, 5 min) and then incubated for 30 min at 3rC. Trypan blue (1 %, 50 [-II) was added followed by glutaraldehyde (3 %, 50 [-II). The number of conjugates and the viability of the bound target cells were determined by counting them under a light microscope. Each sample was assayed in duplicate and data are expressed as percent binding according to the formula: % Binding = 100 x (conjugated effector cells)/(total effector cells).

Preparation of crude NKCF IFN-a-primed or non-primed effector cells (5 x 106 /ml) were co-cultured with various amounts of intact target cells or their isolated membranes in RPMI-1 % for 24h at 37°C in a humidified incubator with 5 % COz. Following incubation, the supernatants were harvested and either used immediately or kept frozen at -70°C until used. Supernatants from cultures of either target cells or effector cells alone were used as controls.

Assay of NKCF 5lCr-Iabeled K-562 cells (2 x 104 cells/well, 100 [-II RPMI-1 %) were seeded in 96-well plates. Supernatants containing NKCF (100 [-II) were added and the cultures were incubated for 20-24 h. Aliquots of the supernatants were then collected for radioactivity measurement. Total, spontaneous and specific cpm were calculated as described in the 5lCr-release assay for NK-CMC.

Measurement of MHC I and 132m antigens on cell surface The entire procedure was performed at 4°C. Target cells (2 x 106 cells/tube) were washed with cold PBS containing 0.02 % NaN 3 and then incubated with anti-HLA-A, B, C antibody (W6/32, diluted 1: 100 in PBS + 0.02 % NaN3 ; 45 min; 100 [-II). Excess antibody was removed by washing with cold PBS and the cell suspension was further incubated with FITC-Iabeled goat anti-mouse IgG (diluted 1:20 in PBS+0.02% NaN 3 ; 45 min; 100 [-II). The cells were washed with PBS, fixed with formaldehyde (3 %; 18 h), washed again with PBS, passed through glass wool and then analyzed by fluorescence activator cell sorter (FACS) on a FACS II (Becton-Dickinson, Mount View, CA, USA). The level of ~zm was measured similarly by directly staining the cells with FITC-Iabeled anti-~2m antibody (1:20 in PBS, 45 min) followed by formaldehyde fixation and fluorescence analysis. At least 10,000 cells were analyzed in each

MHC Class I Antigens and Resistance to NK Lysis . 27 sample for fluorescence intensity and for size at different gains as detected with a photomultiplier tube set at 500-650 nm. Cholesterol treatment

Rigidification of the cell membrane lipid layer was carried out by incorporation of cholesterol as previously described (25). Briefly, cholesterol was dissolved in ethanol (10 mg/ ml) and dispersed in a rapidly stirring solution of polyvinylpyrrolidone (PVP, MW 40,000; 3.5 % w/v), bovine serum albumin (1 %) and glucose (0.5 %) in PBS, pH 7.4, to a final concentration of 200 ltg/m!. Cells (3 x 106/ml) were incubated in this medium at room temperature with a gentle shaking for 1.5 h. The cells were then washed three times with PBS and resuspended in RPMI -10 %. Viability of the treated cells was 90-98 % as assessed by trypan-blue exclusion. Hydrostatic pressure treatment

This treatment was done as described (26) with some modifications. Cells were dispersed in PBS (10 8 /ml) in a capped polypropylene tube (1.5 ml) filled to the top. A short needle inserted through the cap served as a vent for pressure equalization. Both the tube and the needle were filled with PBS without air bubbles to avoid cell rupture. Tubes were placed in a 40 ml pressure bomb (Aminco), filled with PBS and sealed. Pressure was applied gradually to reach the maximum of 800 atmospheres after 5 min and after additional 15 min the pressure was gradually released during 5 min. The cells were then sedimented (150 x g, 5 min) and resuspended in RPMI-10% (10 6 cells/ml). Protein determination

Protein concentration was determined by a system based on fluorescamine (27) using crystalline bovine serum albumin as a standard. The fluorescamine method allows accurate (± 10 %) determination of protein concentration even at the range of 1-2 ltg/m!. Statistics

51Cr-release assay and conjugate formation assay were expressed as the arithmetic mean ± standard error. Results were compared by Student's t-test.

Results 1. Transfer of MHC I via reconstituted membrane vesicles does not increase resistance to NK-CMC Isolated membrane vesicles from IFN -y-treated target cells were fused with non treated target cells with the aid of Sendai virus envelope glycoproteins. Isolated membrane vesicles from non treated target cells were used as a control. Initially it was demonstrated that reconstituted viral-plasma membrane (VPM) vesicles which were prepared at various ratios of virusto-cell components retained most of the expected Sendai virus derived hemagglutination and hemolytic activites (Table 1). VPM vesicles were then prepared at a 1:4 ratio of virus to plasma membrane proteins, respectively. This ratio was found to be optimal in terms of fusion efficiency as determined independently by measuring the ability to transfer 125I-Iabeled and fluorescently labeled isolated plasma membranes to acceptor cells (Table I). VPM vesicles bearing MHC I were then prepared from membranes of IFN-treated K-562 cells. These VPM vesicles were fused with

28 . Z. REITER et al. Table I. Construction of optimal VPM vesicles (1) MembranelViral protein ratio

Hemagglutination units

Hemolysis %

Fluorescence dequenching % (2)

Transfer efficiency %(3)

1 3 4 10 RSVE

6.8 N.D. 4.2 2.3 8.3

6.3 N.D. 4.1 1.9 11.6

52.1 57.6 64.3 31.3

N.D. N.D. 22 6 34

(1). Vesicles were prepared from viral and cell membrane proteins at the indicated ratios. Reconstituted Sendai virus envelopes (100 % viral proteins, RSVE) were used as controls. In all experiments a constant amount (2.5 mg) of viral protein was used. All values represent an average of triplicate or duplicate results. (2). Octadecyl rhodamine Bl chloride-labeled RSVE or VPM vesicles were used for determining their fusion efficiency with K-562 cells as measured by fluorescence dequenching, taking triton-X-I00-solubilized vesicles as 100 % dequenching. (3). VPM vesicles made from 125I-labeled plasma membranes of K-562 cells were fused with K562 cells. Input cpm in each experiment was taken as 100 %.

class I negative K-562 cells which were then stained with either anti HLAA, B, C monoclonal antibody or anti B2m antibody. It was found that the resulting cells indeed acquired MHC I as determined by F ACS analysis. The level of MHC I increased with the amount of VPM vesicles in a dosedependent manner (Fig. 1). In a control experiment, VPM vesicles made from membranes of non treated K-562 cells were fused with the same type of cells and no increase in MHC I was obtained (Fig. 1). In parallel, following membrane fusion, both types of the fused cells and the parental cells were tested for their sensitivity to NK-CMC. Both fused cell types remained highly sensitive to NK-CMC, as much as their parental cells (Fig. 2). A

5 M

B

4

I

0 )(

IJ)

3 2

Q)

()

..... 10

100

1000

10

100

1000

RELATIVE FLUORESCENCE INTENSITY

Figure 1. Insertion of MHC I to K-562 cells by membrane fusion. FACS analysis of K-562 cells following labeling with anti -HLA-A, B, C antibodies (A) or anti-132m antibodies (B).

Control (MHC I-negative) cells = ----; cells fused with VPM vesicles from control K-562 cells = - -..; cells fused with VPM vesicles (2 mg/ml = - ; 5 mg/ml = ... .. ) from IFN-y-treated K562 cells.

MHC Class I Antigens and Resistance to NK Lysis . 29 60

50

l

.. CD 1/1

II

CD

40

~ ~

.,

CJ

30

u ;:

20

'uCD

Q.

II)

10

0

10: 1

50: 1

25: 1

Effector:Target

ratio

Figure 2. Insertion of MHC I to K-562 cells does not affect the sensitivity to NK-CMC. Specific 51Cr release from K-562 cells was determined at different E:T ratios. Control K-562 cells = 0; IFN-y-treated (1000 units/ml, 18 h) K-562 cells = .; K-562 cells fused with VPM vesicles from IFN-y-treated K-562 cells = 0; K-562 cells fused with VPM vesicles from non treated K-562 cells = 6. The curves represent an arithmetic mean (standard error was 2.5-6.3 %) of triplicate measureme~ts with effector cells from 7 different donors.

2. Loading of cellular membranes with cholesterol exerts opposite effects on resistance to NK-CMC and on expression of MHC

In order to further study the possible correlation between resistance to NK-CMC and surface levels of MHC I, control and IFN-y-treated (MHC 5

B

4

M

I

A

0 )(

3

~

- 2 If)



()

...... 10

100

1000

10

100

1000

RELATIVE FLUORESCENCE INTENSITY

Figure 3. Cholesterol treatment lowers the level of MHC I in K-562 cells. FACS analysis of K562 cells following labeling with anti-HLA-A, B, C (A) or anti-b2m (B) antibodies. Control (MHC I-negative) cell = ---; IFN-y-treated K-562 cells = -'-'; IFN-y and cholesterol-treated K-562 cells = - .

30 . Z. REITER et al.

ror-------~------._------_r------_,

50 GI

::

40

GI

~

u

~

20

GI

Q. C/)

10

0~----~10~:~1----~2-5L : 1------5~0~:-1--~--~

Effector:Target

ratio

Figure 4. Cholesterol treatment lowers the sensitivity to NK-CMC. Specific 51Cr release from K-562 cells was determined at different E:T ratios, with IFN-y-primed effector cells. Control K-562 cells = 0 ; IFN-y (1000 units/ ml, 16 h)-treated cells = e; cholesterol-treated cells = . ; cells treated with IFN-y and cholesterol = .... The curves represent an arithmetic mean (standard error was 0.8-4.6 %) of experiments with IFN-a primed-PBMC from 3 different donors.

I-positive) K-562 cells were loaded with cholesterol. Cholesterol loading significantly reduced the population of IFN-y treated cells which expressed high levels of MHC I, while no effect of cholesterol was observed in the control MHC I-negative cells (Fig. 3). These cells were then tested for their sensitivity to NK-CMC and it was found that cholesterol pre-treatment reduced the sensitivity of both control and IFN-y-treated cells to IFN-aprimed NK-CMC (Fig. 4). No such reduction in sensitivity to NK-CMC was observed with control cells that were treated with the medium used for incorporation of cholesterol. As expected, the same pattern of response was Table 2. The effect of IFN-y and cholesterol on conjugate formation and on sensitivity of K-562 cells to NK-CMC Treatment IFN-y

+ +

cholesterol

+

+

Conjugated cells

Dead conjugated cells

(%)

(%)

17.3 ± 1.6

14.6 ± 2.1

19.4±2.1

16.8 ± 2.4 19.6 ± 1.7

6.8 ± 1.1

5.9 ± 2.3 5.2 ± 1.8

MHC Class I Antigens and Resistance to NK Lysis' 31

~r--------------------------------'

l

30

III III CII

III

..

!

_u

20

II)

~

U III a.

(/J

10

o

100

50

10

NKCF (I! L)

Figure 5. Cholesterol treatment does not change the ability of target cells to induce NKCF release. Specific 51Cr release from K-562 cells treated with supernatants of co-cultures of effector cells and K-562 cells at 100:1 ratio for 24 h. Co-culturing was done with control K-562 cells = . ; IFN-y-treated K-562 cells = 0; cholesterol-treated K-562 cells =~; or IFN-y- and cholesterol-treated K-562 cells = D. The histograms represent an arithmetic mean (standard error was 1.1-3.6%) of experiments with PBMC from 3 different donors.

30r--------------------------------,

III

til

....

~

20

Q)

_0

-

10

.~

u

Q)

a.

(/J

o

100

50

10

NKCF (I! L)

Figure 6. Cholesterol treatment reduces the sensitivity of cells to NKCF. Specific SICr release from K-562 cells incubated (20 h) with various amounts of NKCF. Control K-562 cells = II1II; cholesterol-treated K-562 cells = §iI. The histograms represent an arithmetic mean (standard error was 1.8-4.7 %) of three experiments.

32 . Z.

REITER

et al.

7o

4

)(

3

B

A

5

2 Q)

U

10

100

1000

10

100

1000

RELATIVE FLUORESCENCE INTENSITY

Figure 7. Hydrostatic pressure reduces the level of MHC I on K-562 cells. FACS analysis of K-562 cells following labeling with anti-HLA-A, B, C (A) or anti-~2m (B) antibodies. Control cells = ---; IFN-y-treated cells = - ; IFN-y- and pressure-treated cells = _._ ..

obtained with non-primed NK effector cells as with IFN-a-primed cells, except that the overall cytotoxicity was lower in the first case (not shown). The mechanism by which cholesterol increases resistance to NK -CM C was then studied. It was found that cholesterol had no effect on the level of conjugate formation between K-562 and NK cells. However, the percent of dead target cells in the conjugates was significantly reduced (Table 2). The combined protection following treatments with both IFN-y and cholesterol was higher than each of the treatments alone. Unlike IFN-y, cholesterol did not affect the ability of target cells to induce the release of NKCF from NK

60

~ e...-

50

CII III IG CII

40

~

...

0

30

It)

~

·u CII

D-

20 10

C/)

0

1 0 : 1

25 : 1

Effector:Target

50: 1

ratio

Figure 8. Hydrostatic pressure increases resistance of K-562 cells to NK-CMC. Specific 51Cr release from K-562 cells was determined at different E:T ratios. Control K-562 cells = 0; IFN-y-treated cells = .; pressurized cells = .; IFN-y-treated and pressurized cells = ~. The curves represent an arithmetic mean (standard error was 4.6-7.2 %) of two experiments (each performed in triplicate) with IFN-y-primed NK effector cells from 5 different donors.

MHC Class I Antigens and Resistance to NK Lysis . 33

cells (Fig. 5). However in contrast to IFN-y, cholesterol treatment rendered target cells more resistant to killing by NKCF (Fig. 6).

3. Hydrostatic pressure increases resistance to NK-CMC and reduces the level of surface MHC I K-562 cells that were subjected to hydrostatic pressure of 800 atmospheres for 15 min were 75-85 % viable but the population of cells which exhibited the highest level of MHC-I was significantly reduced (Fig. 7). However such a treatment significantly increased the resistance of the cells to NK-CMC (Fig. 8). The protective effects of a combined treatment with

40

~

A

l!Q)

1/1

30

111 Q)

... ... -u

Qi 20

'"

~u

10

Q)

Q. rJ)

0

10: 1

25 : 1

Effector:Target

50 : 1

ratio

60

~ l!-

B

50

Q)

1/1

111

Q)

... _u...

40

Qi

'"

u

30 20

~

'0 Q)

Q. rJ)

10 0

10 : 1

25: 1

Effector:Target

50 : 1

ratio

Figure 9. The sensitivity of K-562 sublines to NK-CMC. Specific SICr release from various K562 sublines incubated either with effector cells (A) or with IFN-y-primed effector cells (B) at different E:T ratios. K-562 parental cells = 0; K-562/CR cells = .; K-562/S cells = •. The curves represent an arithmetic mean (standard error was 2.3-4.4 %) of experiments with NK effector cells from 6 different donors.

34 . Z.

REITER

et al.

IFN-y and hydrostatic pressure was larger than that of each treatment alone. The same result was observed when non primed effector cells were used, except that the cytotoxicity was lower (not shown). Although the spontaneous release of SlCr slightly increased following pressurization from 5-12 % to 15-25 %, the effects of both IFN-y and pressurization on resistance to NK-CMC could be determined with confidence. A high level of surface MHC I was present in K-562 cells only following IFN-y treatment. Consequently, the effect of hydrostatic pressure treatment on MHC I level was measurable only in IFN-y-treated cells (Fig. 7).

4. Lack of correlation between sensitivity to NK-CMC and MHC I levels in K-562 sublines The sensitivity to NK-CMC of K-562/CR and K-562/S cells was compared with that of the parental cells. K-562/CR cells were significantly more resistant to both spontaneous and IFN-y-primed NK-CMC as compared with their parental K-562 cells (Fig. 9 A, B). K-562/S cells were found to be almost completely resistant in both tests (Fig. 9 A, B). The parental cells and the two sublines exhibited a similar low level of MHC I antigens and following IFN-y treatment the level of MHC I increased to the same extent in all sublines as demonstrated by F ACS analysis (not shown). Both K-562 and K-562/S cells became more resistant to NK-CMC following IFN treatment, however the resistance of K-562/CR cells did not increase in spite of the elevation in their MHC-I level. The phase of the killing process which was associated with the differences in K-562/CR and K-562/ S phenotypes was then identified. No difference was observed between the level of conjugate formation of K-562/CR, K-562/S or K-562 parental cells 30.-----------------------------,

Q) II)

:

20

~

...

~o It)

10

u

:;::

'u Q)

Q.

en

o

100

50

10

NKCF (11 L) Figure 10. The sensitivity of K-562 sublines to NKCF. Specific SICr release from various K562 sub lines incubated (20 h) with different amounts of NKCF. K-562 cells = . ; K-562/S cells = D; K-562/CR cells = §l. The histograms represent an arithmetic mean (stand ard error was 1.4-4.3 %) of three experiments.

MHC Class I Antigens and Resistance to NK Lysis . 35

and NK effector cells. In addition, the stimulatory capacity of K-562/CR cells as inducers of NKCF release was similar to that of the parental cells. However, it was found that K-562/CR cells were significantly more resistant than both the parental and the K-562/S cells to killing by NKCF (Fig. 10). Therefore, it was concluded that the increased resistance of K562/CR cells to NK-CMC was not related to their MHC-I level.

Discussion Class I MHC antigens (MHC I) which are present on all cell types have a central role in cell-to-cell interactions in the immune system. Many MHCI-negative cells were found to be highly sensitive to NK-CMC and it was therefore suggested that a lack of MHC I is a signal for attack by NK cells (the «missing self theory», 7). The possible role of MHC I as signals which confer resistance to NK-CMC was suggested on the basis of many observations (3, 4, 8, 9, 28-36). Recently, transfection of cells with MHC I genes resulted in an NK-CMC resistant phenotype (34-36). The correlation between MHC I expression and resistance to NK-CMC as reported in all of these studies was primarily circumstantial. However, due to it's extensive nature a functional linkage between these two phenomena was proposed. However in some other studies, including several transfection experiments, no correlation was found between the level of MHC I on target cell surface and resistance to NK-CMC (37-42). We have shown recently that IFN-aC protected Daudi cells from both IFN-primed and non primed NK cells without inducing MHC I on cell surface. Therefore, we concluded that in other cell types these two phenomena could represent independent responses to IFN (10). Further support for an independent regulation of MHC I and resistance to NK-CMC is provided by the present membrane fusion experiments. Membrane fusion was proven as an efficient mean for transfer of various membranal components from cell-to-cell. Optimal conditions for such a transfer with the aid of Sendai virus proteins were worked out and it was found that transfer of membranal components from resistant to sensitive cells did not confer resistance to the later ones. Despite the fact that the acceptor cells acquired cell surface MHC I, no increase in resistance to NKCMC was obtained. Additional experimental approaches were investigated as well. The fluidity of the lipid bilayer of the cell membrane is affected by the lipid composition and can be modulated by incubation of the cells with various fatty acids and steroids (43, 44). Changes in membrane fluidity were reported to modulate the sensitivity to NK-CMC and T cell mediated cytotoxicity (45-49). Such changes also affected the level of MHC I in mouse spleen cells (50). Indeed, following cholesterol treatment of IFNtreated target cells, the number of cells exhibiting a high level of MHC I was reduced, while resistance to NK-CMC was increased. An increase in

36 . Z. REITER et al.

resistance was also observed when cells were treated with cholesterol alone. It is therefore likely that the IFN-induced protection was independent from modulation of the MHC I level. Alternatively, it is possible that an IFNinduced increase in the MHC-I level above a certain threshold exerted resistance to NK-CMC while further increase, reversible by cholesterol, had no additional effect on NK-CMC. Application of hydrostatic pressure on whole cells is a gentle method for exposure and (at higher pressure) release of membranal and even transmembranal proteins (51). As shown here the number of cells exhibiting a high level of MHC I was significantly reduced by exposing the cells to a pressure of 800 Atmospheres and the antigens were probably released into the medium (51). Following such treatment, the cells became more resistant rather than more sensitive to NK-CMC, despite the loss of MHC 1. Indeed such a treatment affects not only MHC I but many other cell surface structures, some of which may be involved in the various stages of NK-CMC. Further evidence for an independent regulation of MHC I levels and resistance to NK-CMC came from studies of the K-562 sublines which exhibit marked differences in resistance to NK-CMC in spite of the fact that they normally do not carry MHC I and following IFN induction exhibit the same elevated level of MHC 1. The data presented in this work further demonstrates a lack of exclusive functional linkage between the level of MHC class I antigens and regulation of target cell sensitivity to NK-CMC. It is possible that cholesterol treatment and hydrostatic pressure evoked NK-protective systems independent from MHC I, which could overcome the loss of MHC 1. If this is the case, one can at least conclude that in addition to MHC I there are other, more critical systems which provide protection from NK-CMC. The mechanisms by which IFN, cholesterol and hydrostatic pressure induce resistance to NK-CMC remains elusive but it appears that structures other than MHC I are responsible for target cell refractoriness to NK-CMC. Acknowledgments The technical assistance of Mrs. R. EISENSTADT is greatly acknowledged. MENACHEM RUBINSTEIN has the Maurice and Edna Weiss chair in Interferon Research. ZVI FISHELSON is an incumbent of the Barecha Foundation Career development Chair.

References 1. BERKE, G. 1989. Functions and mechanisms of lysis induced by cytotoxic T-lymphocytes and Natural Killer cells. In: Fundamental Immunology, 2nd ed., Chapter 27 (W. E. PAUL, ed.). Raven Press Ltd., New York. 735 pp. 2. Trinchieri, G. 1989. Biology of Natural Killer cells. In: Adv. in Immunology, Vol. 47 (F. J. DIXON, ed.). Academic Press Inc., London-New York. 187 pp. 3. STORKUS, W. J., D. N. HOWELL, R. D. SALTER, J. R. DAWSON, and P. CRESSWELL. 1987. NK susceptibility varies inversely with target cell class I HLA antigen expression. J. Immunol. 138: 1657.

MHC Class I Antigens and Resistance to NK Lysis . 37 4. SUGAWARA, S., T. ABO, H. ITHO, and K. KUMAGAI. 1989. Analysis of mechanisms by which NK cells acquire increased cytotoxicity against class I MHC-eliminated targets. Cell. Immunol. 119: 304. S. OHLEN, c., M. T. BEJARANO, A. GRONBERG, S. TORSTEINSDOTTIR, L. FRANKSSON, H. G. LJUNGGREN, E. KLEIN, G. KLEIN, and K. KARRE. 1989. Studies of sub lines selected for loss of HLA expression from an EBV -transformed lymphoblastoid cell line. Changes in sensitivity to cytotoxic T cells activated by allostimulation and Natural Killer cells activated by IFN or IL-2. J. Immunol. 142: 3336. 6. LJUNGGREN, H. G., S. PAABO, M. COCHET, G. KLING, P. KOURILSKY, and K. KARRE. 1989. Molecular analysis of H-2-deficient lymphoma lines. Distinct defects in biosynthesis and association of MHC class I heavy chains and ~2-microglobulin observed in cells with increased sensitivity to NK cell lysis. J. Immunol. 142: 2911. 7. LJUNGGREN, H. G., and K. KARRE. 1990. In search of the «missing selL,: MHC molecules and NK cell recognition. Immunol. Today 11: 237. 8. GRONBERG, A., R. KIESSLING, and W. FIERS. 1985. Interferon-y is a strong modulator of NK susceptibility and expression of ~2-microglobulin but not of transferrin receptors of KS62 cells. Cell. Immunol. 95: 19S. 9. YEOMAN, H., and R. A. ROBINS. 1988. The effect of interferon-gamma treatment of rat tumor cells on their susceptibility to Natural Killer cell, macrophage and cytotoxic T-cell killing. Immunology 63: 291. 10. REITER, Z., and M. RUBINSTEIN. 1990. Interleukin-l a and tumor necrosis factor-a protect cells against Natural Killer cell-mediated cytotoxicity and Natural Killer cytotoxic factor. Cell. Immunol. 125: 326. 11. FISHELSON, Z., and Y. REITER. 1989. Monoclonal antibody-C3b conjugates: killing of KS62 cells and selection of a stabile complement resistant variant. In: Mechanism of action and therapeutic applications of biological in cancer and immuno-deficiency disorders. Alan R. Liss Inc., New York. 273 pp. 12. REITER, Y., and Z. FISHELSON. Complement membrane attack complexes induce transient synthesis of a high molecular weight protein (L-CIP) in human leukemic cells. Submitted. 13. BOYUM, A. 1968. Isolation of mononuclear cells and granulocytes from human blood. Scand. J. Lab. Invest. 21: 77 (Suppl. 97). 14. LEUNG, K. H., and H. S. KOREN. 1982. Regulation of human Natural Killing. II. Protective effect of interferon on NK cells from suppression by PGE2. J. Immunol. 129: 1742. IS. MAEDA, T., K. BALAKRISHNAN, and S. Q. MEHDI. 1983. A simple and rapid method for the preparation of plasma membranes. Biochim. Biophys. Acta 731: lIS. 16. ]AKOBOVITS, A., N. SHARON, and I. ZAN-BAR. 1982. Acquisition of responsiveness by non responding lymphocytes upon insertion of appropriate membrane components. J. Exp. Med. 156: 1274. 17. PRUJANSKY-]AKOBOVITS, A., D. J. VOLSKI, A. LOYTER, and N. SHARON. 1980. Alteration of lymphocytic surface properties by insertion of foreign functional components of plasma membranes. Proc. Nat!. Acad. Sci. USA 77: 7247. 18. JAKOBOVITS, A., M. ROSENBERG, and N. SHARON. 1981. Human B lymphocytes from rosettes after insertion of T lymphocyte membrane constituents. Eur. J. Immunol. 11: 440. 19. JAKOBOVITS, A., A. FRENKEL, N. SHARON, and 1. R. COHEN. 1981. Inserted H-2 gene membrane products mediate immune response phenotype of antigen presenting cell. Nature 291: 666. 20. VAINSTEIN, A., M. HERSHKOVITZ, S. ISRAEL, S. RABIN, and A. LOYTER. 1984. A new method for reconstitution of highly fusogenic Sendai virus envelopes. Biochim. Biophys. Acta 773: 181. 21. HOSAKA, Y., H. KITANO, and S. IKEGUCHI. 1966. Studies on the pleomorphism of HVJ virions. Virology 29: 20S. 22. CITOVSKY, V., R. BLUMENTHAL, and A. LOYTER. 1985. Fusion of Sendai virions with phosphatidyl choline-cholesterol liposomes reflects the viral activity required for fusion with biological membranes. FEBS Lett. 193: 13S.

38 . Z. REITER et al. 23. CHEJANOVSKY, N., and A. LOYTER. 1985. Fusion between Sendai virus envelopes and biological membranes. J. bioI. Chern. 260: 7911. 24. WELSH, R. M. 1978. Cytotoxic cells induced during lymphocytic choriomeningitis virus infection of mice. I. Characterization of Natural Killer cell induction. J. Exp. Med. 148: 163. 25. SHINITZKY, M., Y. SKORNICK, and N. HARAN-GHERA. 1979. Effective tumor immunization induced by cells of elevated membrane-lipid microviscosity. Proc. Natl. Acad. Sci. USA 76: 5313. 26. RICHERT, L., A. OR, and M. SHINITZKY. 1986. Promotion of tumor antigenicity in EL-4 leukemic cells by hydrostatic pressure. Cancer Immunol. Immunother. 22: 119. 27. STEIN, S., and J. MOSCHERA. 1981. High performance liquid chromatography and picomole level detection of pep tides and proteins. In: Methods in Enzymology 79 (S. P. COLOWICK and N. O. KAPLAN, eds.). Academic Press, New York. 7 pp. 28. BECKER, S., R. KIESSLING, N. LEE, and G. KLEIN. 1978. Modulation of sensitivity to Natural Killer cell lysis after in vitro explantation of mouse lymphoma. J. Nat!. Cancer Inst. 61: 1495. 29. STERN, P., M. GIDLUND, A. ORN, and H. WIGZELL. 1980. Natural Killer cells mediate lysis of embryonal carcinoma cells lacking MHC. Nature 285: 341. 30. HANSSON, M., R. KIESSLING, B. ANDERSSON, and R. M. WELSH. 1980. Effect of interferon and interferon inducers on the NK sensitivity of normal mouse thymocytes. J. Immunol. 125: 2225. 31. WALLACH, D. 1982. Regulation of susceptibility to Natural Killer cell's cytotoxicity and regulation of HLA synthesis; differing efficacies of alpha, beta and gamma interferons. J. Interferon Res. 2: 329. 32. LOBO, P. I., and C. E. SPENCER. 1989. Use of anti-HLA antibodies to mask major histocompatibility complex gene products on tumor cells can enhance susceptibility of these cells to lysis by Natural Killer cells. J. Clin. Invest. 83: 278. 33. HAREL-BELLAN, A., A. QUILLET, C. MARCHIOL, R. DEMARS, T. TURsz, and D. FRADELIZI. 1986. Natural Killer susceptibility of human cells may be regulated by genes in the HLA region on chromosome 6. Proc. Nat!. Acad. Sci. USA 83: 5688. 34. STORKUS, W. J., J. ALEXANDER, J. A. PAYNE, J. R. DAWSON, and P. CRESSWELL. 1989. Reversal of Natural Killing susceptibility in target cells expressing transfected class I HLA genes. Proc. Nat!. Acad. Sci. USA 86: 2361. 35. QUILLET, A., F. PRESSE, C. MARCHIOL-FoURNIGAULT, A. HAREL-BELLAN, M. BENBUNAN, H. PLOEGH, and D. FRADELIZI. 1988. Increased resistance to non-MHC-restricted cytotoxicity related to HLA A, B expression. Direct demonstration using ~2-microglobu­ lin-transfected Daudi cells. J. Immunol. 141: 17. 36. SHIMIZA, Y., and R. DEMARS. 1989. Demonstration by class I gene transfer that reduced susceptibility of human cells to Natural Killer cell-mediated lysis is inversely correlated with HLA class I antigen expression. Eur. J. Immunol. 19: 447. 37. LEIDEN, J. M., B. A. KARPINSKI, L. GOTTECHALK, and J. KORNBLUTH. 1989. Susceptibility to Natural Killer cell-mediated cytolysis is independent of the level of target cell class I HLA expression. J. Immunol. 142: 2140. 38. NISHIMURA, M. I., I. STROYNOWSKI, L. HOOD, and S. OSTRAND-RoSENBERG. 1988. H2Kb antigen expression has no effect on Natural Killer susceptibility and tumorigenicity of a murine hepatoma. J. Immunol. 141: 4403. 39. GOPAS, J., S. SEGAL, G. HAMMERLING, M. BAR-ELI, and B. RAGER-ZISMAN. 1988. Influence of H-2K transfection on susceptibility of fibrosarcoma tumor cells to natural killer (NK) cells. Immunol. Lett. 17: 261. 40. KARRE, K., H. G. LJUNGGREN, G. PIONTEK, and R. KIESSLING. 1986. Selective rejection of H-2-deficient lymphoma variants suggests alternative immune defence strategy. Nature 319: 675. 41. CHERVENAK, R., and R. M. WOLCOTT. 1988. Target cell expression of MHC antigens is not (always) a turn-off signal to Natural Killer cells. J. Immunol. 140: 3712.

MHC Class I Antigens and Resistance to NK Lysis . 39 42. GRONBERG, A., R. KIESSLING, E. ERIKSSON, and M. HANSSON. 1981. Variants from MLVinduced lymphoma selected for decreased sensitivity to NK lysis. J. Immunol. 127: 1734. 43. MANDEL, G., S. SHIMIZU, R. GILL, and W. GLARK. 1978. Alteration of the fatty acid compositions of membrane phospholipids in mouse lymphoid cells. J. Immunol. 120: 1631. 44. KLAUSNER, R. D., A. M. KLEINFELD, R. L. HOOVER, and M. J. KARNOVSKY. 1980. Lipid domains in membranes. Evidence derived from structural perturbations induced by free fatty acids and lifetime heterogeneity analysis. J. BioI. Chern. 255: 1286. 45. BERKE, G., R. TzuR, and M. IN BAR. 1978. Changes in fluorescence polarization of a membrane probe during lymphocyte-target cell interaction. J. Immunol. 120: 1378. 46. DABROWSKI, M. P., W. E. PELL, and A. E. R. THOMSON. 1980. Plasma membrane cholesterol regulates human lymphocyte cytotoxic function. Eur. J. Immunol. 10: 821. 47. HEINIGER, H. J., and J. D. MARSHALL. 1982. Cholesterol synthesis in polyclonally activated cytotoxic lymphocytes and its requirement for differentiation and proliferation. Proc. Natl. Acad. Sci. USA 79: 3823. 48. ROOZEMOND, R. L., and B. BONAVIDA. 1985. Effect of altered membrane fluidity on NK cell-mediated cytotoxicity. I. Selective inhibition of the recognition or post recognition events in the cytotoxic pathway of NK cells. J. Immunol. 134: 2209. 49. ROOZEMOND, R. L., M. MEVISSEN, D. C. URLI, and B. BONAVIDA. 1987. Effect of altered membrane structure on NK cell-mediated cytotoxicity. III. Decreased susceptibility to Natural killer cytotoxic factor (NKCF) and suppression of NKCF release by membrane rigidification. J. Immunol. 139: 1739. SO. MULLER, C. P., D. A. STEPHANY, M. SHINITZKY, and J. R. WUNDERLICH. 1983. Changes in cell surface expression of MHC and thy-1.2 determinants following treatment with lipid modulating agents. J. Immunol. 131: 1356. 51. DECKMANN, M., R. HAIMOVITZ, and M. SHINITZKY. 1985. Selective release of integral proteins from human erythrocyte membranes by hydrostatic pressure. Biochim. Biophys. Acta 821: 334. Dr. M. RUBINSTEIN, Department of Molecular Genetics and Virology, The Weizmann Institute of Science, Rehovot 76100, Israel