natural killer activity of lymphoid cells isolated from human ascitic ovarian tumors

Int. J . Cancer: 25, 573-582 (1980)

NATURAL KILLER ACTIVITY OF LYMPHOID CELLS ISOLATED FROM HUMAN ASCITIC OVARIAN TUMORS

Albert0 MANTOVANI~, Paola ALLAVENA I , Cristiana SESSA I , Giorgio BOLIS’ and Costantino MANGIONI~

Ginecologica, Universita di Milano, Via Commenda, 12-20100 Milan, Italy. lstituto di Ricerche Farmacologiche “Mario Negri”, Via Eritrea, 62-20157 Milan; and Clinica Ostetrica e

Lymphocytes and tumor cells were isolated from the carcinomatous ascites of 24 patients with epithelial ovarian tumors by stepwise application of density and velocity sedimentation on discontinuous Ficoll-lsopaque gradients and fetal bovine serum. Tumor-associated lymphocytes showed a lower percentage of cells with receptors for sheep erythrocytes (E) or for complement than did peripheral blood lymphocytes from the same patients. NK activity was measured, ”Cr-labelled K562 cells being used as targets in a 20-h assay. Tumor-associated lymphocytes showed significant NK activity. Cy- totoxicity levels were lower than for peripheral blood effector cells from the same patients, and these in turn showed significantly lower cytotoxic capacity than peripheral blood lymphocytes from 64 control subjects. Similar results were obtained when lysis was measured after 4 h of incubation. Tumor-associated lymphocytes forming E rosettes were at least as effective as the unseparated population. When tumor-associated lymphocytes were mixed with normal effector cells, in three of six preparations with low NK activity tested, significant inhibition of normal lymphocyte NK activity was observed. Adherent macrophages from carcinomatous ascites, which contained lymphocytes that had suppressive activity, showed no inhibitory activity. Interferon (IF) boosted the NK activity against K562 of tumor-associated lymphocytes. Purified ovarian carcinoma cells were relatively resistant to lysis by normal lymphocytes. However, they inhibited lysis of K 562 cells in cold target competition assays, though less efficiently than K 562 itself, and were consistently lysed when effector cells were stimulated with IF. It is therefore suggested that ovarian carcinoma cells express NK-relevant recognition structures, but are relatively resistant to cytolysis by unstimulated effector cells.

Non-adherent lymphoid cells from various animal species can express significant levels of natural killer (NK)3 activity in vitro (recently reviewed by Herber- man et al., 1979). NK activity has been thought to play a role in vivo as one line of natural resistance against infection and neoplasia (Kiessling et al., 197%; Herberman et al., 1979). The significance of NK cells in the control of established malignancy remains t o be determined. Studies on NK activity of lymphocytes isolated from experimental and human tumors are limited and have yielded variable results. In two studies, lymphocytes isolated from selected solid human neoplasms had no NK activity (Vose et al., 1977a; Totterman et al., 1978), but in two murine tumors some NK cytotoxicity was detected (Becker and Klein, 1976; Moore and Moore, 1979). Freshly isolated human tumor cells were found to be relatively resistant to the NK activity of unstimulated effector cells (Herberman et al., 1979; Zarling et al., 1979; Vose and Moore, 1980) although some lysis of leukemia cells was observed in one study (Rosenberg et al., 1972); however, when the expression on human tumor cells of NK recognition struc-

tures was studied in cold target competition assays, conflicting data were obtained (Ortaldo et al., 1977; Vose and Moore, 1980).

The present investigation was designed to elucidate the NK activity of lymphocytes isolated from human ascitic ovarian tumors and the susceptibility to NK cells of purified ovarian carcinoma cells, ascitic ovarian neoplasms providing an easily accessible source of viable tumor cells, lymphocytes and macrophages (Mantovani et al., 1979b; 1980c) in suspension.

MATERIAL AND METHODS

Human subjects Twenty-four patients with histologically con-

firmed ascitic ovarian epithelial tumors admitted to the Department of Oncology, Clinica Ostetrica e Ginecologica, Universita di Milano, formed the case-list for this study. All patients were classified as stage I11 and IV. The presence of tumor cells in ascitic fluids was checked by independent pathologists.

The control population consisted of 19 age- matched patients admitted to the same clinic for non-malignant, non-infectious gynecological disor- ders (benign ovarian cysts, myomas, etc.) and 45 normal healthy adult volunteers. In each experiment a t least three control subjects were tested concomitantly with ovarian cancer patients.

Heparinized venous blood (10-40 ml) was obtained by venipuncture and carcinomatous ascites was collected by paracentesis or at laparotorny. Peritoneal exudate (PE) was collected from 15 patients undergoing surgery for non-malignant, non- infectious gynecological diseases (benign ovarian cysts, myomas, etc.) as previously described (Man- tovani et al., 1 9 8 0 ~ ) .

Peripheral blood lymphocytes (PBL) Blood was diluted 1:4 with phosphate-buffered

saline (PBS, Eurobio, Paris, France) and 40 ml on 10 ml Ficoll-Hypaque (Eurobio, Paris, France) were centrifuged at 400 g for 20 min at room temperature.

3Abbreviations: PB, peripheral blood; PE, peritoneal exudate; PBL, peripheral blood lymphocytes; PEL, peritoneal exudate lymphocytes; TAL, tumor-associated lymphocytes; E, sheep erythrocytes; Zy, complement-coated zymosan; RFC, rosette-forming cells; RCI, relative cytotoxicity index; CV, coefficient of variation; IF, interferon; PBS, phosphate-buffered saline; C, complement; C.parvurn, Corynebacterium parvum; NK, natural killer.

Received: January 21, 1980.

574 MANTOVANI m AL.

Mononuclear cells were collected at the interface and washed with PBS, then 10-30 X lo6 cells were incubated in 10-20 ml of RPMI 1640 medium sup- plemented with 20% fetal bovine serum (FBS, GIB- CO-Biocult, Paisley, Scotland) for 45 min at 37°C in plastic Petri dishes (3003, Falcon, Oxnard, Calif., USA). Non-adherent cells were collected, centrifuged at 400 g for 5 min and finally resuspended in RPMI 1640 medium with 10% FBS and 50 &ml gentamicin. Non-adherent cell preparations w e v only partially depleted of monocytes as 2-4% of these cells were mononuclear phagocytes, as assessed by morphology, avid uptake of neutral red and staining for non-specific esterase, as compared to 10- 30 96 of the original mononuclear cell suspensions (Mantovani et al., 1980b).

Peritoneal exudate lymphocytes ( P E L ) Macrophages can be isolated in relatively large

amounts from peritoneal exudates (median 3 X 10 , range 0.5 - 31 X 10' for 28 donors; Mantovani et al., 1980a) but we have not been able to isolate lymphocytes from this anatomical site, except from two donors. Peritoneal exudates were centrifuged at 400 g for 20 min on Ficoll-Isopaque. Macrophages were removed by adherence on plastic as described above for PBL (Mantovani et al., 1980a,b,c). The procedure was repeated, until contamination with monounclear phagocytes was less than 2-4% as assessed by morphology, avid uptake of neutral red and non-specific esterase staining.

Tumor-associated lymphocytes (TAL) and ovarian carcinoma cells

Enriched preparations of ovarian tumor cells and tumor-associated lymphocytes were obtained by stepwise application of density and velocity sedimentation on discontinuous Ficoll-Hypaque or FBS gradients (Vose et al., 1977a, b). Carcinomatous ascites was centrifuged at 400 g for 5 min. Cells were resuspended in PBS and 20 ml of the cell suspension (1-5 X 10h cells/ml) were layered on 10 ml Ficoll-Hypa- que. The tubes (25330, Corning, N.Y.) were then centrifuged at 400 g for 20 min at room temperature and the mononuclear cell layer was aspirated. To remove macrophages from ovarian ascites cell suspensions, mononuclear cells in serum-free medium (Mantovani et al., 1979b, 1980c) were incubated for 45 rnin at 37°C in plastic Petri dishes. The procedure was repeated 2-3 times until contamination with mononuclear phagocytes, assessed by morphology and avid uptake of neutral red, was less than 5 % . After washing with 50 ml PBS, 10 ml of the mononuclear cell suspension (-2X 106/rnl) were layered on a discontinuous gradient consisting of 5 ml of 100% Ficoll-Hypaque, 10 ml of 50% Ficoll-Hypa- que (i.e. diluted 1:2 with PBS) and 10 ml of 25% Ficoll-Hypaque. The tubes (25330-Corning, N.Y.) were held stationary for 2 h at 37°C. Lymphocyte- enriched preparations were collected at the PBS - 25 % interface, and tumor cells were concentrated at the 50% - 100% interface. This procedure was adopted in the first seven patients studied. More recently, the method was simplified. Mononuclear cell suspensions were layered on discontinuous gradients consisting of 10 ml of 100 % and 15 ml of 75 % Ficoll-

Hypaque. The tubes were centrifuged at 500 X g for 10 min at room temperature. Lymphocyte-enriched suspensions were harvested at the 75 %-loo% interface and tumor cells were found on top of the 75 % fraction. The procedure was repeated once or twice if separation was not satisfactory. When separation of lymphocytes from tumor cells on discontinuous Ficoll-Hypaque gradients was not satisfactory, as judged by morphological examination of Wright- stained smears (see below), cell suspensions (10-20 X 10h/10 ml PBS) were layered onto 10 ml FCS and incubated at room temperature for 60-90 min; tumor cells sedimented at the bottom of the discontinuous gradient whereas cells with lymphoid morphology remained in the upper part of the gradient. The various phases of the separation procedure were moni- tored by morphological examination of Wright- stained smears. Cell suspensions from the unseparated carcinomatous ascites and occasionally from the purified lymphocytes or tumor cells were stained with Papanicolau's stain and examined by independent pathologists. Lymphoid cell and tumor-cell suspensions were accepted for use when at least 95% (usually more than 98%) of the cells had the typical unequivocal morphology of the respective cell type. Lymphocyte preparations were also studied (see below) for the presence of sheep erythrocyte and complement receptors, which were not found on tumor cells. Less than 5 % of the lymphoid cell suspensions belonged to the rnonocyte-macrophage lineage, as assessed by morphology, rapid uptake of neutral red and non-specific esterase staining, and work in progress shows that TAL have a significant blastogenic response to phytohaemmaglutinin. Primary ovarian carcinoma cultures with typical morphology and lack of contact inhibition could be established from purified tumor cells (Mantovani et al., 1980c), but attempts to cultivate tumor cells from lymphocyte preparations consistently failed. The separation procedure described here was not effective in four patients, in spite of the presence in the carcinomatous ascites of cells with lymphoid morphology. Viability of enriched fractions always exceeded 90 % and lymphocyte recovery ranged from SO to 0.5 X loh cells for different patients. When normal PBL were submitted to the same separation procedure as used for ovarian tumors, NK activity was not significantly affected.

Lymphocyte subpopulations To determine the number of E-rosette-forming

cells (E-RFC), 2 X 10' lymphoid cells were incubated for 5 min at 37°C with 6 X lo7 sheep erythrocytes (E) in 0.5 ml RPMI 1640 with 40% absorbed FBS. The cells were centrifuged at 80 X g for 5 min and held at 6"-8"C for 18h. After gentle suspension, cells binding at least 3E were scored as E-RFC. Complement (C) receptor-bearing cells were evaluated using C-coated zymosan particles (Zy) as previously described (Huber and Wigzell, 1975). Briefly, lymphocytes (2 X 10') were mixed with Zy and centrifuged at 700 rpm at room temperature. After 30 min incubation in an ice bath, cells binding at least three particles were scored as Zy-RFC. To separate E-RFC-enriched lymphocytes, after interaction with E for 2h at 4°C lymphocytes were separated by cen-

NK CELLS IN HUMAN TUMORS 575

trifugation at 400 g for 20 min on Ficoll-Hypaque (West et al., 1977). E-FRC, collected at the bottom of the gradient, were washed with 50 ml PBS and finally disaggregated by vigorous resuspension with a Pasteur pipette. No attempt was made to lyse E as these procedures may alter NK activity and, in preliminary tests, we had confirmed that E had no effect on cytolysis (West et al., 1977).

Tumor-associated macrophages Macrophages were isolated from ascitic ovarian

tumors by adherence on microexudate-coated plastic and EDTA as previously described (Mantovani, 1980; Mantovani et al . , 1979a,b, 1980~).

Interferon (IF) Partially purified human fibroblast IF (Lot No. 45-

10-9) was obtained from HEM, Rockville, Md., USA. Lymphoid cells (1-5 X 106/ml) were cultured overnight in the presence of 1,000 units/ml IF in growth medium. Control lymphocytes were cultured alone. Lymphoid cells were washed with 50 ml PBS before their cytotoxic capacity was tested.

Targei cells Tumor cell lines used in the present study were:

the KS62 line (originally obtained from Dr. G. Bon- nard, National Cancer Institute, Bethesda, Md.) derived from a human myeloid leukemia (Lozzio and Lozzio, 1973); the murine SV40-transformed mKSATU5 (TU5) kidney line (Kit et al., 1969) originally obtained from Dr. J . Dean, Litton Bionetics, Kensington, Md.; the Raji cell line (Pulvertaft, 1965) obtained through the courtesy of Dr. V. Rosso di San Secondo, Centro Trafusionale e di Im- munologia dei Trapiati, Milan. Cell lines were main- tained in vitro in RPMI 1640 medium with 20 % FBS and 50 pg/ml gentamicin (growth medium). K562 cells were used as standard targets to measure NK activity of lymphocytes. TU5 and Raji cells are not lysed by normal lymphocytes (Mantovani et a[., 1979a; Ortaldo et al., 1977) and they were used as controls in cold target competition assays. In addi- tion, purified ovarian carcinoma cells, freshly isolated from nine carcinomatous ascitic fluids as described above, were used as targets. Tumor cells (1-5 X lo6 in 1 ml growth medium) were incubated with 20-50 pCi of slCr (sodium chromate, Radiochemical Centre, Amersham, Bucks., England) at 37°C for 45 min. Labelled tumor cells were washed twice with 50 ml of medium before use in cytolysis assays. Cytotoxicity assay

51Cr-labelled tumor cells (2X 1 04) were cultured with as many attacker to target (A:T) ratios as possible (usually 50:1, 25:1, 12:l and 6:l) in 0.2 ml growth medium in round-bottomed wells of microplates (Sterlin, Teddington, Middlesex, England). When NK activity was measured with K562 cells, the routinely employed incubation time was 20 h , but in a number of experiments lysis was also measured after 4 h of incubation. Isotope release was calculated as (A/B)X100, where A is the isotope in the supernatant and B is the total in incorporated radioactivity released by incubation with 1 % sodium dodecyl sulfate. Specific lysis was calculated by sub-

7 p < o o 1 ,

..... i ..

.. .. I 2.

#s rz

I 20 -!-

1 ..... : lo] -? * * rz

20 : lo] - p<o 01 - l0I p < o o 1 O J - - ~~ - -

Z Y E Z Y E Z Y Pecipheral blood Peripheral blood A S C I ~ I C I u m o r s

Controls O v a r l a " cancer patients

FIGURE 1 - E and Zy-RFC in human ovarian tumors.

tracting spontaneous isotope release of tumor cells alone. Spontaneous W r release from K562 cells was 0.5-1.5% per hour of incubation. With the nine freshly isolated ovarian carcinoma preparations used in the present study, median spontaneous release after 4 h of culture was 27% (range 9-33%, nine tumor-cell preparations) and after 20 h 33 % (range 31-35, five tumor cell preparations).

Cold target inhibition assay The assay (Herberman et al., 1976) was performed

to establish whether ovarian cancer cells expressed NK-relevant recognition structures. K562 cells (2X104) were cultured with normal peripheral blood lymphocytes at an A:T ratio of 25:l in the presence of different numbers of unlabelled inhibitors; the routinely employed inhibitor:target (1:T) cell ratio ranged from 0.3:l to 10:1. Cells were cultured either in 1 ml growth medium in flat-bottomed 16-mm wells (3524 Costar, Cambridge, Mass.) for 4-6 h on a rocking platform (6 cycles/min) or in 0.2 ml in microplates as in the standard cytolysis assay described above.

Since comparable results were obtained with the two methods, no distinction will be made between them in the presentation of data. The percentage of inhibition was calculated as (l-A/B)XlOO, where A is the cytotoxicity in samples with cold inhibitors and B is the cytotoxicity of effector cells alone.

Statistical analysis Three replicates per experimental group were ern-

ployed throughout and results of cytotoxicity tests were calculated as mean f SD; 4-6% increases in isotope release above baseline were usually statisti- cally significant at p<O.O5 (Student's t-test). Lym- phoid cell preparations from different subject populations or from different anatomical sites from the same patients were analysed by the Mann-Whitney U-test.

In each NK experiment, ovarian cancer lymphocytes were tested along with at least three control subjects. A semilog plot of the specific cytotoxicity data against the number of effector cells per sample was obtained; the number of effector cells required for 33 % specific lysis was graphically determined and arbitrarily defined as one lytic unit 33 (LU 33).

576 MANI'OVANI K l AL.

8 0 ,

7 0 -

60-

50.

2 LO.

= 30- c oi

20~

10-

01

I

!.

1

-@

P r

& I '

* * ; - - Controls O v P I Controls O v P t C ~ n l r o l ~ O v Pt

A 1 ~ 5 0 I A 1 : 2 5 1 A T;12 1

_ _ _ - - _ _ _ _ _ - - _ _ _ _ PS PE P B Ascllrs PB P E PB Ascilrs P B PE PB AscileS P B PE P B Asc)tes

Controls O v Pt A 1:6 1

Expression of NK cytotoxicity results relative to normal subjects tested concomitantly in the same experiment has been reported to consistently reduce the day-to-day coefficient of variation (CV, defined as the ratio SD/mean) of the assay (Pross and Baines, 1976; Santoli et al., 1978) as also observed with blastogenesis tests (Dean et al., 1977). This was coii- firmed in preliminary experiments in this laboratory (see also "Results"). Therefore, for each lymphoid cell preparation, a relative cytotoxicity index (RCI) was calculated as: xnLU33ALU33, where xnLU33 is the average LU33 value for control subjects tested concomitantly and tLU33 is the same value for the test subject. Thus, a n RCI of 0.2 indicates that five times more test lymphocytes are required to obtain 33 % lysis when compared to control effector cells.

RESULTS

Figure 1 shows the percentages of E-RFC and Zy- RFC in lymphoid cells isolated from the peripheral blood and carcinomatous ascites of patients with ovarian epithelial tumors. T A L had a lower percentage of E-RFC and Zy-RFC (p<O.Ol) than PBL from the same patients. The latter, in turn, showed significantly (p<0.02) less E-RFC than control subjects.

Tumor-associated lymphoid cells had significant cytolytic activity against K.562, cytotoxicity levels varying widely in different patients (Fig. 2 ) . NK activity values of T A L and PBL from ovarian cancer patients were significantly lower than for control PBL (p t0 .01 , Fig. 2). It should be noted that when normal PBL were submitted to the same preparation procedure as used to obtain T A L from ovarian tumors, NK activity was not significantly affected. The obtain a more appropriate control for ovarian TAL, we tried to isolate lymphocytes from peritoneal washings of patients with benign gynecological diseases. From most (13) peritoneal exudates we could not obtain enough lymphocytes to test NK activity; lymphoid cells from two exudates (Fig. 2 ) showed very little cytolytic activity, specific lysis being 9.5 and 7 % at an A:T ratio of SO: 1.

I n agreement with previous reports on NK and blastogenesis tests (Pross and Baines, 1976; Santoli et al., 1978; Dean et al., 1977) calculation of cytotoxicity as RCI reduced the day-to-day C V of the assay.

FIG~JRE 2 - NK activity in ovarian tumors. Results are presented as percentage specific lysis. Horizontal bars indicate the medians. Ov. Pt. = ovarian cancer patients.

The C V for RCI in four normal subjects repeatedly tested over a period of 4 months (see below, Fig. 4) was 0.27 k 0.02 (mean + SD). Accordingly, differences between normal PBL, ovarian cancer PBL and T A L were observed better when data expressed as RCT (Fig. 3). From Figure 3 it can also be seen that T A L were significantly (p<0.05) less cytotoxic than PBL from the same patients. In fact, in 1.5 o u t of 20 experiments in which PBL and T A L from the same patient were tested in parallel in the same experiment. PBL were more efficient killers. whereas the

2

1.6

1.4

1.2

1 - V

LT 0.8

0.6

0.4

0.2

0

. .. . . 7 - - - - - - - - - - - r-::

r-PcO.0l-i r P < O . O l T

7 p<o.os 1 . 3:. . ..... ..

0 Ul N .

.-0. . - 2 .. v .:.. c Y.. c ... 0 ...

.

. .M.

P B P B A sc i tes Controls Ovarian cancer patients

FIGURE 3 - NK activity in ovarian carcinoma. Results are presented as RCI.


A L y m p h o c y t e s f r o m norma l donors . Tumor -as soc ia t ed l y m p h o c y t e s

I ,*

0 1 I’ /*/---‘

.. , I

0 10 2 0 30 5 0 100 Day a f t e r f i r s t a s s a y

FIGURE 4 - NK activity of PBL from S normal donors (A) and TAL from 7 ovarian cancer patients (0) tested at different times over a period of 4 months.

reverse was observed in only four subjects. In one donor no difference was detected between PBL and TAL.

NK activity of T A L varied widely among different donors, but when ascitic tumor samples were obtained at different times from the same patients, donors with low, intermediate or high reactivity consistently had similar cytotoxicity (Fig. 4).

NK activity was routinely measured in a 20-h test. This incubation time was chosen because a shorter assay (4-6 h) would have been impractical for application after a long separation procedure (see “Mate-

A

4 h

0 0-0 0

6 : l 12:l 25:l 50 : 1 A:T ratio

1

rial and Methods”; Tumor-associated lymphocytes). However, in a first limited series of experiments lysis was also measured after 4 h of incubation. A s shown by the representative experiment presented in Fi- gure 5 , results at 4 h were qualitatively similar to those at 20 h.

There was some concern that the significant, though defective, NK activity of ovarian T A L was in fact due to effector cells of different lineage from NK, e.g. macrophages. Macrophages, with natural cytotoxicity against susceptible targets (Mantovani et al., 1979a,b; 1980a,b,c) were less than 5 % (usually less than 2 % ) of the TAL, and in agreement with previous data (Mantovani et al., 1979a,b), they had little cytotoxicity against K562 cells (see below, Table I). Moreover (Fig. 6), cells with receptors for E isolated from ascitic ovarian tumors were at least as effective as the unseparated TAL, as previously shown by others (Ortaldo et al., 1977). E-RFC from normal PBL had significant cytotoxicity.

Defective NK activity in mice after treatment with Corynebacterium parvum (C. pnrvum), hydrocortisone or irradiation has been related to the presence of suppressor activity (Hochman et al., 1979; Hoch- man and Cudkowicz, 1979; Savary and Lotzova, 1978). Therefore, in a series of experiments, lymphocytes from ovarian tumors were mixed with normal lymphocytes and NK activity was measured (Table I). We calculated the inhibitory activity of T A L relative to normal lymphocytes alone, thus possibly underestimating suppressive activity because we chose not to take into account the possible contribution t o cytotoxicity of the NK TAL. Table I

100.

ul ul >r

.- - U Z 50. .- U Q, a ul

s

0.

6

2 0 h

...I..*::* /.----O

0

o-------o 1 1

6 :l 12:l 25:l 50:l A : T r a t i o

/.----O 0‘ 0

o-------o

6 :l 1211 25:l 50:l A : T r a t i o

FIGURE 5 - NK activity of normal PBL (3 donors, 0) . ovarian cancer PBL (1 donor, *) and TAL (2 patients, 0), after 4 (a ) or 20 h ( b ) of culture with KS62 target cells.

578 MANTOVANI ET Al..

v)

v)

2 20.

100,

1 ;!Normal lymphocytes

ln ul > .-

-

I Tumor- associated lymphocytes

U .- T so- a u * ul -

#

6 : l 12: 1 25:l 50: 1 A : T ratio

FIGURE 6 - N K activity of E-RFC from carcinomatous ascites. Continuous and dotted lines indicate unseparated lymphoid cells and E-RFC, respectively.

A:T i 5O:l @ A:T. 25:l '"1

1 2 3 4 5 6 7 8 9 1 2 3 L 5 6 7 8 9 Ovar i an t u m o r cells

A:T; 25:l 0 A:Tz 5O:l

I

' 1 0 f

1

I I

f

f 1 2 3 4 5 6 7 8 9 1 2 3 k 5 6 7 8 9

Ovarian tumor c e l l s

FIGURE 7 - Lysis of ovarian carconoma cells from normal peripehral blood lymphocytes after 4 (a ) or 20 h ( h ) of culture. Numbers indicate carcinoma cells from different donors.

TABLE I

EFFECT OF TUMOR-ASSOCIATED LYMPHOCYTES ON NK ACTIVITY OF NORMAL LYMPHOCYTES

Patient Normal PBL Tumor-associared Percentage specific Percentage NO. (A:T ratio) ' ~ ~ $ ' ~ ~ ~ ~ ~ lysis inhihition'

11 50:l - 40.2 (1.8) - 25:l - 36.1 (0.8) - 12: 1 - 27.6 (0.4) - 6:1 - 18.5 (0.7) -

- 50:l 17.3 (3.0) -

- 12:1 8.2 (3.3) - - 6: I 6.5 (4.1) -

25:l 25:1 15.5 (2.7)4 58 12:l 12: I 11.9 (2.4)' 57

- 25:1 9.6 (4.8) -

6:l 6: 1 15.2 (0.3)' 21 - 25:l 4.2 (2.3)* -

25:l 25:1* 38.9 (2.3)2 -

12 50:l - 75.0 (0.1) - 25:l - 74.0 (0.1) - - 50: 1 41.4 (3.3) - - 25: 1 33.2 (3.9) -

25:l 25:l 63.4 (0.5)' 14

12:l - 69.0 (1.6) - - 25:l 42.8 (2.4) -

12:1 12:l 55.8 (0.9) 19 14 50:l - 49.2 (2.3) -

25:l - 41.7 (4.5) - 12:l - 29.5 (7.1) - 6:l - 20.8 (2.7) -

- 25:l 17.4 (1.0) - 12:l 8.5 (0.3) -

24 25:l - 74.4 (1.2) -

- 12:l 37.3 (4.4) ~

-

6: 1 4.9 (2.2) ~

2S:l 25: 1 37.4 (0.7) -

-

12:l 12:l 27.3 (2.3) - 6:l 6:l 19.5 (4.0) -

2 S0:l - 63.6 (9.0) - 25:l - 45.2 (2.0) - 12:l - 37.2 (5.0) - 6:l - 29.3 (4.0) -

25:1 27.6 (0.6) - - 12:l 16.7 (4.3) -

25:l 25:l 70.6 (3.1) - 6:l 6:1 31.0 (7.1) -

23 50:l - 76.5 (2.8) - 25:1 - 68.7 (3.4) - 12:l - 68.2 (3.2) - 6:l - 59.5 (2.7) - - 25:l 37.7 (2.4) - - 12:1 21.6 (0.7) -

6:l 15.9 (1.1) -

6:l 6:1 67.2 (4.4) -

-

- 6: 1 11.8 (4.7) -

12:1 12:l 53.5 (7.9) -

-

i Inhibition was calculated as a percenta e of normal lymphocytes alone. - * Tumor-associated macrophages. - 'Significantly below lysis of normal lymphocytes alone, p<O.O5. - ' Significantly below lysis of normal lymphocytes alone, p<O.Ol.

N K CELLS IN HUMAN TUMORS

TABLE 11

EFFECT O F INTERERON ON THE NK ACTIVITY OF TUMOR-ASSOCIATED LYMPHOCYTES AGAINST K562 CELLS

579

Percentage specific lysis

6 : l 12:1 25:l 50: 1 Effector cells IF

Normal PBL - + + -

+ Ovarian cancer PBL - + Tumor-associated - lymphocytes + Tumor-associated - lymphocytes +

36.6 (1.6) 39.8 (0.4) 8.9 (1.0) 14.6 (2.9)’ 26.7 (2.0) 31.1 (2.4)’ - -

15.4 (3.7) 32.3 (3.3)*

13.5 (0.6)* 3.9 (0.8)

53.6 (2.6)

13.8 (0.9) 28.3 (5.2)7 40.5 (3.0) 55.3 (4.1)2 31.7 (3.8) 49.8 (5.2)2 38.0 (2.6) 52.3 (3.8)2 9.1 (1.2) 17.2 (2.6)*

59.4 (0.4)2 64.4 (1.7) 69.1 (3.9) 22.0 (2.9) 34.0 (1.2)2 50.5 (3.4) 66.0 (1..5)? 40.0 (7.9) 63.8 (1.6); 54.7 (5.0) 61.2 (3.3)’ 17.1 (1.7) 34.0 (3.0)‘

31.5 (1.0) 48.4 (l.S)* 60.2 (5.8) 67.0 (0.5)’

25.3 (1.2) 55.2 (3.0)’

presents six experiments in which TAL with depressed NK activity were tested. With three of six preparations evaluated significant inhibition of the cytolytic activity of normal PBL was observed, suppression values ranging from 14 to 58%. Experi- ments presented in Table I were performed using a ratio of 1:l between presumed inhibitor cells and normal lymphocytes, but increasing the ratio did not increase the levels of suppression nor the number of positive results (data not shown). It is of interest that tumor-associated adherent macrophages from the donor with the highest levels of inhibitory activity (patient No. 11) had little cytotoxicity and no inhibitory activity on cytotoxicity (Table I). From one of the patients with significant, though low, suppressive capacity (No. 24) TAL obtained on two separate occasions had similar inhibitory effects (results not presented). TAL from three ovarian cancer patients with relatively high NK activity (in the same range as controls, Fig. 3) had no inhibitory activity, nor was suppression observed when PBL from normal donors or ovarian cancer patients were tested.

In vitro or in vivo exposure to IF can boost NK activity in rodents and humans (Trinchieri et al., 1978; Trinchieri and Santoli, 1978; Herberman et al., 1979). Therefore it was of interest to investigate whether IF could stimulate the depressed NK activity of TAL from ovarian neoplasms. Preincubation with IF markedly enhanced the cytolytic capacity of TAL against K562 (Table 11).

In an effort to elucidate further the immunobiolo- gy of NK cells in human ovarian carcinoma, the susceptibility of freshly isolated ovarian tumor cells to normal NK activity was studied. The interaction of ovarian cancer lymphocytes with autologous and heterologous ovarian carcinoma tumor cells is cur- rently being investigated. As shown in Figure 7 , ovarian carcinoma cells were relatively resistant to NK activity, some lysis being observed only occasionally, and most frequently at the highest (50:l) A:T ratio and longest incubation time (20 h). However, in cold target competition assays (Fig. S), ovarian carcinoma cells consistently inhibited lysis of K562, though

less efficiently than K562 itself. Conversely, Raji and TU5, used as negative controls, had no consis- tent effect on lysis of K562. Inhibition of lysis of K562 by ovarian carcinoma cells was never detected at I:T ratio<2.5:1 which, when related to the number of effector lymphocytes present in the assay (25: l), corresponded to a 10 % contamination of effector cells with tumor cells.

Thus ovarian tumor cells, though relatively resistant to lysis, appeared to express the relevant recognition structures for NK cells. To further test this hypothesis, the capacity of normal lymphocytes stimulated by IF to lyse ovarian carcinoma was evaluated. IF-boosted effector cells consistently kil-

TABLE I11

EFFECT OF INTERFERON ON CYTOLYSIS OF OVARIAN CARCINOMA CELLS

Ovarian

(No.) 25:l’ 50: 1

Percentage specific lysis f SD tumor c:lls l:g:a;$’ IF

5 4 - NT 2.7 (0.3) + NT 4.6 (0.6)’ 20 - NT 13.0 (3.1)3 + NT 15.2 (1,3)3

7 4 - 3.9 (1.113 4.7 (1.213 + 21.4 (1.3)3.4 34.4 (5.1)3.4 20 - 6.4 (4.2)’ 8.5 (1.5)’ + 27.3 (3.7)3,4 35.7 (5.5)3,4 4 - 10.5 (4.1)’ 12.6 (1.4)’

20 - 16.7 (1.0)’ 26.4 (8.1)3 + 18.6 (2.4)3.4 27.8 (1.5)3.4

+ 18.9 (l.l)3 32.3 (0.5)’.‘

+ 0 0 20 - 0 0 + 7.7 (3.8)3.4 13.8 (l .8)3.4

9 4 - 0 0

I Effector cells were normal PBL preincubated overnight alone o r with IF. Cytotoxicity was measured after 4 or 20 h incubation. -’ A:T ratio. -

pt0.05 versus spontaneous release. - p<0.05 versus lymphocytes cultured without IF.

580 MAN'I'OVAI

led ovarian tumor cells (Table 111). In one experiment with carcinoma No. 9 , IF, when left throughout the assay at a concentration of 1,000 units/ml, did not protect target cells from lysis by IF-stimulated effector cells (results not presented), as observed with some cell lines (Trinchieri and Santoli, 1978).

C

2 .- 20-

s o

n

c 10- .- L .-

301

U .- t 1 44 :: 10

\ K 5 6 2 \ 2 f- ?

0 0 None 2.5 :1 5 : l 10: 1 Ratio of inhibitor: labeled K 5 6 2

cel ls

'Z 5 0

c 40 ;1 r

R a j i 6 T U 5 rn K 5 6 2 . Ovarian tumor cells

-----*- I+!

a i l o.&i i . z j i 2.511 511 i o l i Ratio 01 inhibitor to labeled K 5 6 2 c e l l s

FIGURE 8 - Cold target competition assays with ovarian carcinoma cells. In panel A, results of a typical experiment are presented as percentage specific lysis, after 4 h of culture with rocking. In panel B data are presented as percentage inhibition of lysis of K562. Data with K562 (positive control), TU5 and Raji (negative controls) are means of five experiments.

DISCUSSION

Human ascitic ovarian carcinomas provide an easily accessible, unique source of tumor cells, macrophages (Mantovani et al., 1979b, 1980c) and lymphocytes in suspension. Lymphoid cells can be purified with relative ease from carcinomatous as-

cites. T A L showed a lower proportion of E-RFC and Zy-RFC than PBL from the same patients or normal PBL; thus a larger percentage of T A L lacked these membrane markers. Similar data on E-RFC were reported for some human solid tumors (Vose et al., 19770; Hiiyry and Ttitterman, 1978).

T A L from ovarian carcinomatous ascites showed significant NK activity against K562 cells. Unlike T A L from ascitic ovarian tumors, lymphoid cells from a series of disaggregated human solid tumors were reported to have no significant NK activity (Vose el a / . , 1 9 7 7 ~ ; Totterman et al., 1978). Murine peritoneal exudate cells have low NK activity (Her- berman et d., 1975; Kiessling et al., 1 9 7 5 ~ ) . We have not been able to isolate lymphoid cells from human peritoneal exudates, except in the case of two subjects, whose peritoneal lymphocytes had little cytoci- dal capacity. However the peritoneal cavity might represent a n accessible anatomical site for NK cells, which would explain the significant NK activity in ovarian carcinomatous ascites in contrast to other human solid tumors (Vose et a/., 1 9 7 7 ~ ; Tijtterman et al., 1978). An alternative or complementary ex- planation for the NK effector capacity of T A L from ascitic ovarian tumors is that the disaggregation procedures applied to human solid neoplasms (Vose et al., 1 9 7 7 ~ ; Hayry and Totterman, 1978) may affect the viability or functional status of tumor-associated NK cells. In fact, lymphoid cells with significant NK activity have been recovered from some rodent solid neoplasms (Becker and Klein, 1976; Moore and Moore, 1979).

Several lines of evidence suggest that tumor-associated NK cells from ovarian carcinomatous ascites share many characteristics with normal peripheral blood NK cells. ( 1 ) Although NK activity was routinely measured in a 20-h assay, significant cytolysis was also detected after 4 h with TAL, as expected for NK cells. (2) It is unlikely that macrophages, which have natural cytotoxicity (Manto- vani el a / . , 1979a,b, 1980a), had any role in the NK activity of TAL. In fact, in agreement with previous data (Mantovani et a/. , lY79a, b ) K562 cells were relatively resistant to the cytotoxic activity of macrophages, which anyway were less than 5 % (usually less than 2%) of the TAL. (3) E-RFC separated from ascitic tumors were at least as effective as, o r more so than, the unseparated TAL, and it has been shown that at least some of the NK cells have low affinity receptors for E (West et al., 1977; Timonen et al., 1979; Herberman et al., 1979).

Although T A L from ascitic ovarian tumors had significant levels of N K activity, these were much lower than normal PBL or even PBL from the same cancer patients. The mechanismfs) responsible for the impaired N K capacity of T A L remain(s) t o be fully elucidated. A smaller number of relevant effector cells may be present in T A L , or suppressor cells may inhibit NK activity.

In three of six patients with defective NK effector capacity, T A L inhibited the expression of NK activity by normal lymphocytes. Tumor cells contaminating T A L preparations have been thought to account for depressed NK activity in one rat sarcoma model (Moore and Moore, 1979). In the present study,


ovarian tumor cells were always less than 5 % of the TAL and usually less than 2%. Purified ovarian carcinoma cells were able to inhibit NK activity in competition assays, but this effect was only observed at tumor-cell concentrations exceeding 10 % of effector lymphocytes. Thus, it appears extremely unlikely that contaminating tumor cells have any role in the defective NK activity of TAL and in their suppressive activity, unless it is assumed that lymphocyte- associated tumor cells are a population with much greater affinity for NK cells than the whole tumor population. Moreover, it is possible that Moore and Moore (1979) have overestimated the inhibitory capacity of tumor cells as they used a tissue culture line as a reference standard.

Following treatment with C. parvurn, carrageenan or irradiation, cells able to inhibit NK activity have been reported in mice (Savary and Lotzova, 1978; Hochman et al . , 1978; Hochman and Cudkowicz, 1979). The nature of the supposedly suppressor cells has not been elucidated, but they appeared not to adhere to plastic, although carbonyl iron removed the suppressive activity (Hochman and Cudkowicz, 1979). Interestingly enough, in the present investigation adherent tumor-associated macrophages from a patient with inhibitory TAL had no suppressive activity, and a similar lack of suppressive capacity on NK activity was noted with tumor-associated macrophages from other subjects (results not presented). Thus the nature of the supposedly suppressor cell(s) and the relevance of this phenomenon remain to be defined, although it is tempting to speculate that the defective NK capacity of TAL may in part be related to their suppressive activity. It should be noted that in the present investigation, as well as in previous studies in rodents, there is no unequivocal evidence that inhibition of NK activity is in fact due to a suppression of NK cell function, as a protection of target cells from lysis cannot be for- mally excluded.

Freshly isolated ovarian carcinoma cells were relatively resistant to lysis by normal unstimulated lymphocytes, but some cytotoxicity was occasionally observed, particularly when spontaneous release values permitted testing after 20 h of culture with effector cells. Resistance of freshly obtained tumor cells has previously been observed with human leukemia cells (Zarling et al., 1979) and lung carcinoma cells (Vose and Moore, 1980) and with murine tumors (Herberman et al . , 1979). However, ovarian tumor cells inhibited NK activity in target competition assay, though less efficiently than unlabelled K562 cells. Moreover, lysis of ovarian carcinoma cells was observed after stimulation of effector cells with IF. Thus ovarian carcinoma cells may express the recognition structures for NK cells but be relatively resistant to cytolysis. A dissociation between susceptibility to lysis and inhibitory capacity in cold target

competition assays has been reported with the murine YAC lymphoma. After a few hours in culture this tumor could already better express NK- relevant recognition structures but enhancement of cytolysis required longer in vitro culture (Becker el al. , 1978).

Previous studies on the expression of NK-relevant recognition structures on human tumors using cold target competition assays have given conflicting results. Ortaldo et al. (1977) showed that unseparated, freshly disaggregated human neoplasms inhibited NK activity, but the effect might in fact have been related to contaminating cells from the stroma or the lymphoreticular infiltrate, or to dead cells which frequently are a major component of suspensions from human solid neoplasms. At variance with these data, Vose and Moore (1980) recently observed that purified lung carcinoma cells did not express NK recognition structures. The apparent discrepancy between results reported here, indicating that human ovarian carcinomas bear NK recognition structures, and Vose and Moore’s finding may in part be related to the differences between the tumors studied or to the disaggregation procedures somehow selecting tumor-cell populations not expressing recognition structures.

Ovarian carcinoma cells were relatively resistant to cytolysis in their interaction with NK cells, and TAL from ovarian tumors showed defective NK capacity. It is therefore unlikely that this effector mechanism plays a significant role in the control of advanced ovarian carcinoma. On the other hand, IF could boost NK activity of TAL and ovarian carcinoma cells were susceptible to lysis by normal PBL stimulated by IF. C. parvurn, an IF inducer in human lymphocyte cultures (Sugiyama and Epstein, 1978; Mantovani el al . , 1980b), has been reported to cause regression of ascitic ovarian tumors (Webb et al . , 1978) and this was confirmed by our group in a preliminary series of six patients resistant to conven- tional chemotherapy. Studies are in progress to elucidate the role of NK cells, along with other populations (e.g., macrophages), in the regression of human ovarian carcinomatous ascites induced by C. parvurn.

ACKNOWLEDGEMENTS

This work was supported by the Consiglio Nazionale delle Ricerche, Rome, Italy (Contract No. 79.00643.96, No. 79.02417.65 and No. 78.02166.04) and by Grant 1 R 0 1 CA 26824 from National Cancer Institute, USA. We thank Drs. C. Belloni, M. D’Incalci, C. Bortolozzi, P. Gritti and P. Molteni for invaluable help in the collection of biological samples. Drs. C. Bordignon, M. Masse- rini, and L. Silliprandi microscopically examined biological samples.

REFERENCES

BECKER, S. , KIESSLING, R., LEE, M., and KLEIN. G., Modula- tion of sensitivity to natural killer cell lysis after in vitro explan- tation of a mouse lymphoma. J . nat. Cancer Inst.. 61, 1495- 1498 (1978).

BECKER, S., and KLEIN, E., Decreased ‘‘natural killer” effect in tumor-bearing mice and its relation to the immunity against oncornavirus-determined cell surface antigens. Europ. J . Im- munol., 6 , 892-898 (1976).

582 MANTOVANI ET AL

DEAN, J.H., CONNOR, R. , HERBERMAN, R.B., SII.VA, J . , McCoy, J.L., and OLDHAM, R.K., The relative proliferation index as a more sensitive parameter of evaluating lymphop- roliferative responses of cancer patients to mitogens and al- loantigens. Inr. J. Cancer, 20,359-370 (1977). HAYRY, P., and TOTTERMAN, T.H., Cytological and functional analysis of inflammatory infiltrates in human malignant tumors. I. Composition of the inflammatory infiltrates. Europ. J. Immunol., 8, 866-871 (1978). HERBERMAN, R.B., D J E ~ J , J.Y., KAY, H.D. , ORTALDO, J.R.. RICCARDI, C.. BONNARD, G.D., HOI.DEN, H.T., FAGNANI, R., SANTONI, A , , and PUCCETI, P . , Natural killer cells: characteristics and regulation of activity. Immunol. Rev., 44, 43-70 (1 979). HERBERMAN, R.B., NUNN, M.E., and HOLDEN, H.T., Cytotox- icity inhibition assay for analysis of specificity of cell-mediated T r release cytotoxicity. In: B.R. Bloom and J.R. David (ed.), In vitro methods in cell-mediated and tumor immunity, pp, 489-495, Academic Press, New York (1976). HERBERMAN, R.B., NUNN, M.E., and LAVRIN, D.H., Natural cytotoxic reactivity of mouse lymphoid cells against syngeneic and allogeneic tumors. 1. Distribution of reactivity and specificity. In/. J. Cancer, 16, 216-229 (1975). HOCHMAN, P.S., and CiJDKOWICZ, G., Suppression of natural cytotoxicity by spleen cells of hydrocortisone-treated mice. J . Immunol., 123, 968-976 ( 1979). HOCHMAN, P . S . , CUDKOWICZ, G. , and DAUSSET, J . , Decline of natural killer cell activity in sublethally irradiated mice. J . nut. Cancer Inst., 61, 265-268 (1978). HlJBtR, Ch., and WIGZELL, H. , A simple rosette assay for demonstration of complement receptor sites using complement-coated zymosan beads. Europ. J. Immunol., 5 , 432-435 (1975). KIESSLING, R., KLEIN, E., and WIGZELL, H., “Natural” killer cells in the mouse. I. Cytotoxic cells with specificity for mouse Moloney leukemia cells. Specificity and distribution according to genotype. Eurup. J. Immunol.. 5, 112.117 (19751). KIESSIJNG, R., PETRANYI, G., KLEIN, G., and WIGZELL, H., Genetic variation of in vitro cytolytic activity and in vivo rejec- tion potential of non-immunized semi-syngeneic mice against a mouse lymphoma line. int. J. Cancer, 15, 933-940 (19756). KIT, S. , KURIMURA, T., and DLJBBS, D.R.. Transplantable mouse tumor line induced by injection of SV40-transformed mouse kidney cells. Int. J. Cancer, 4, 384-392 (1969). LOZZIO, C.B., and LOZZIO, B.B., Cytotoxicity of a factor isolated from human spleen. J. nut. Cancer Inst., 50, 535-538 (1 973). MANTOVANI, A. , Separation of mononuclear phagocytes by adherence on microexudate-coated plastic. In: H.B. Hers- cowitz, H.T. Holden, J.A. Bellantiand A. Ghaffar (ed.), Man- uul of macrophage metfiodoiogy, Marcel Dekker, New Yark (1980) in press.

N., B O R U I ~ ~ N O N , C., SESSA, C., and MANGIONI, C. , Natural cytotoxicity on tumour cells of human macrophages obtained from diverse anatomical sites. Clin. exp. Imrnunol., 39,778-784 ( 1 980a). MA~TOVANI, A., DEAN, J.H., JERRELLS, T.R., and HERBER- MAN, R.B., Augmentation of tumoricidal activity of human monocytes and macrophages by lymphokines. Int. J. Cancer, in press (1 980b). MRWOVANI, A , , JERRELLS, T.R., D ~ A N , J.H., and HtRBER- MAN, R.B., Cytolytic and cytostatic activity on tumor cells of circulating human monocytes. In / . J. Cancer, 23,18-27 (1979~). MANTOVANI, A. , PERI, G., POLENTARIJITI, N . , BOLE, G. , MANGIONI, C., and SPREAFICO, F. . Effects on in vitro tumor growth of macrophages isolated from human ascitic ovarian tumors. Inr. J . Cancer, 23, 1.57-164 (19796).

Z., VECCHJ, A., BOLIS, G., and MANGIONI, C., Cytotoxic activity on tumor cells of peripheral blood monocytes and tumor- associated macrophages in patients with ascitic ovarian tumors. J. nuf. Cancer Inst., in press (1980~).

MANTOVANI, A, , BAR SHAVIT, z., PERI, G. , POI FNTARLlTTl,

MA~TOVANI, A . , POLtNTARUTII, N., PERI, G. , BAR SHAVIT,

MOORE, K., and MOORE, M., Systemic and in situ natural killer activity in tumour-bearing rats. Brit. J. Cancer, 39, 636-647 (1979). ORI-ALDO, J.R., OLDHAM, R.K., CANNON, G.C., and H E R B ~ K - M A N , R.B., Specificity of natural cytotoxic reactivity of normal human lymphocytes against a myeloid leukemia cell line. J . nut. Cancer Inst., 59, 77-82 (1977).

PROSS, H.F., and BAINES. M.G., Spontaneous human lymphocyte-mediated cytotoxicity against tumour target cells. I. The effect of malignant disease. Int. J. Cancer, 18, 593-604 (1976).

Pui.vnwAw, J.V., A study of malignant tumours in Nigeria by short-term tissue culture. J . din. Path., IS, 261-273 (1965).

ROSENBERG, E.B., HERRERMAN, R.B. , LEV~NE, P.H., HAITER- MAN, R.H. , McCoy, J .L. , and WUNDERLICH, J.R., Lympho- cyte cytotoxicity reactions to leukemia-associated antigens in identical twins. In / . J . Cancer, 9, 648-658 (1972).

SAVARY, C.A., and LOTZOVA, E., Suppression of natural killer cell cytotoxicity by splenocytes from Corynebacferium yarvurn- injected, bone marrow-tolerant, and infant mice. J . Immunol., 120, 239-243 (1978). SAPVTOLI, D., TRINCHIERI, G., MORE~TA, L., ZMIJEWSKJ, C.M., and KOPROWSKI, H. , Spontaneous cell-mediated cytotoxicity in humans. Distribution and characterization of effector cells. Clin. exp. Immunol., 33, 309-318 (1978).

SUGIYAMA, M., and EPSTEIN, L.B., Effect of Corynehacterium parvum on human T-lymphocyte interferon production and T- lymphocyte proliferation in vitro. Cancer Res., 38, 4467-4473 (1978). TIMONEN, T. , RANKI, A , , SAKSELA, E . , and HAYKY, P. , Hu- man natural cell-mediated cytotoxicity against fetal fibroblasts. 111. morphological and functional characterization of the effector cells. Cell. Immunol., 48, 121-132 (1979).

TOTTERMAN, T.H. , HAYRY, P . , SAKSELA, E., TIMONEN, T., and EKLUND, B., Cytological and functional analysis of inflammatory infiltrates in human malignant tumors. 11. Functional in- vestigations of the infiltrating inflammatory cells. Europ. J. Immunol., 8, 872-875 (1978). T K I N C H I ~ R I , G., and SANTOI I , D., Anti-viral activity induced by culturing lymphocytes with tumor-derived or virus-transformed cells. Enhancement of human natural killer cell activity by interferon and antagonistic inhibition of susceptibility of target cells to lysis. J . exp. Med., 147, 1314-1333 (1978).

TKINCHltKI, G., SANTO1.1, D., DEE, R.R.. and KNOWLESS, B.B., Anti-viral activity induced by culturing lymphocytes with tumor-derived or virus-transformed cells. Identification of the antiviral activity as interferon and characterization of the human effector lymphocyte subpopulation. J . exp. Med., 147, 1299- 13 13 (1978). VOSE, B.M., and MOORE. M., Natural cytotoxicity in man: low susceptibility of freshly isolated tumour cells to lysis. Submitted for publication (1980). VOSE, B.M., VANKY, F., ARGOV, S . , and KLEIN, E., Natural cytotoxicity in man: activity of lymph node and tumor-infiltrating lymphocytes. Europ. J. Immunol., 7,753-757 (19770).

VOSE, B.M., VANKY, F., and KLEIN, E., Human tumour-lymphocyte interaction in vitro. V. Comparison of the reactivity of tumour infiltrating blood and lymph-node lymphocytes with autologous tumour cells. In[ . J. Cancer, 24, 895-902 (19776).

WtBB, H.E., OATEN, S.W., and PIKE, C.P., Treatment of malignant ascitic and pleural effusions with Corynebacterium parvum. Brit. med. J . , 1, 338-340 (1978).

WEST, W.H., CANNON, G.H., KAY, H.D., HONNARD, G.D., and HERBERMAN, R.B., Natural cytotoxic reactivity of human lymphocytes againat a rnyeloid ccll line: characterization of effector cells. J . Immunol., 118, 355-361 (1977).

ZARLING, J.M., ESKRA, L., B O H D ~ N , E.C., HORW~ZEWICZ, J., and CARTER, A.W.. Activation of human natural killer cells cytotoxic for human leukemia cells hy purified interferon. J . Immunol., 123, 63-70 (1979).

natural killer activity of lymphoid cells isolated from human ascitic ovarian tumors

Documents