Nonspecific\'Blocking of Human Ovarian Carcinoma-Associated Cellular Cytotoxicity In Vitro

'Nonspecific' Blocking of Human Ovarian Carcinoma-Associated Cellular Cytotoxicity In Vitro E. SAKSELA, K. PENTTINEN & S. PYRHONEN Third Department of Pathology and Laboratory of Viral Immunopathology, Department of Virology, University of Helsinki; and Departments of Gynecology and Obstetrics, University Central Hospital; Helsinki, Finland

Saksela, E., Penttinen, K. & Pyrhonen, S. 'Nonspecific' Blocking of Human Ovarian Carcinoma-Associated Cellular Cytotoxicity In Vitro. Scand. J. Immunol. 3, 78 L788, 1974. The blocking ability of heat-aggregated human polyclonal IgG and complexes formed by preservative-free tetanus toxoid (TT) and human anti-tetanus serum was investigated on the cell-mediated tumor-specific cytotoxicity reaction of ovarian carcinoma patients against cultured ovarian carcinoma cells. Aggregated IgG effectively blocked the cytotoxic reaction when the purified effector cells were preincubated with various concentrations of this material, whereas no blocking activity was detected when target cells were preincubated and washed before addition of cytotoxic lymphocytes. Optimal concentration was about 1.6 ^g/ml, and higher or lower concentrations were less effective. Anti-TT/TT immunocomplexes behaved similarly and could effectively block at the effector cell but not at the target cell level. The results suggest that the effector cells in human ovarian carcinoma-associated cellular cytotoxicity are Fc-receptor-carrying cells and that their cytotoxic activity can be 'nonspecifically' blocked by saturating these sites with aggregated IgG or unrelated immunocomplexes. E. Saksela, M.D., Third Department of Pathology, University of Helsinki, Helsinki 29, Finland

Patients with malignant neoplasms have circulating lymphocytes that are capable of specifically reducing the number of in vitro-cultured tumor cells derived either from the patient's own tumor or from a histologically similar neoplasm (8). This cell-mediated cytotoxicity can be blocked by serum from patients with progressively growing neoplasms of the same histological type (9). The blocking factors are presumably antigen-antibody complexes (22, 1), and also free antigen has been shown to block the effector cells (2). The effector cells involved in the inhibition of cultured human bladder carcinoma cells can be removed by passage through glass-bead columns coated with antigen-antibody complexes, presumably due to attachment through Fc receptors (16), indicating a possible thymus-independent origin of these cells. The cytotoxic

00290

lymphoid cells in syngeneic animal systems reacting in microcytotoxicity assays against tumorassociated or virus-induced antigens have also been shown to be non-T cells (13). In the present study we have analyzed the role of Fc receptors on the effector cells by preincubation in media containing heat-aggregated IgG or unrelated antigen-antibody complexes known to contain activated immunoglobulin Fc sites, . „ , ^ ^ ^..^ MATERIALS AND METHODS Patients. The patients had various histological types of ovarian carcinoma of surface epithelium derivation. They were inpatients from the Department of Obstetrics and Gynecology, University Central Hospital, Helsinki, and the blood samples were obtained at various times after radical operation.

782

£. Saksela, K. Penttinen & S. Pyrhonen

Ttitnor cells. All tumor cell lines described were obtained from ascitic fluids of ovarian carcinoma patients which were cytologically proved to contain abundant carcinoma cells. The samples were centrifuged, the cells separated, and cultures started in BME (Basal Medium Eagle, Gibco, Chagrin Falls, Ohio, USA) supplemented with 10% calf serum (Flow Laboratories, Irvine, Scotland) and 10 jag streptomycin and 100 IU penicillin per milliliter. During the primary culture phase, 50% autologous ascitic fluid was used in the medium. Subcultures were made once or twice a week by trypsinization, and between 10 and 16 passages (subcultures in 1:2 ratio) were obtained before cessation of growth and degeneration of the cells. No continuous lines have been obtained so far. The cells were large and polygonal, had pleomorphic nuclear structure, and grew without definite pattern, often piling up and showing faulty contact inhibition. By these morphological criteria they were tumor cells and distinguishable from macrophages and mesothelial cell elements. As control target cells, two different established lines from human malignant neoplasms 4T and 396T were used. The former was established from a giant-cell synovial sarcoma (20) and the latter from human osteosarcoma by Dr. Jan Ponten (Wallenberg Laboratory, Uppsala, Sweden), who kindly placed a seeder culture at our disposal. Effector cells, Heparinized blood was drawn (10-ml plastic syringes, Jintan, Tokyo, Japan) from the patients and control persons (usually patients with other types of malignant neoplasms from the same ward) and purified by colloidal iron and magnetism, followed by gradient centrifugation in Ficoll-Isopaque (5). May-Griinwald-Giemsa (MGG)-stained cytocentrifuge preparations showed the population to be more than 95% pure mononuclear cells. Heat-aggregated IgG, Aggregated IgG was prepared by heating native human polyclonal IgG (2 to 3 mg/ml in phosphate-buffered saline (PBS)) for 10 min in a water bath at 63°C, followed by rapid cooling in ice water (7). Antigen—antibody complexes. Soluble tetanus toxoid (TT)-human anti-TT complexes at se-

lected Ag/Ab ratio levels were prepared by mixing equal volumes (usually 0.2 ml) of undiluted antiserum with various doses of antigen in 2-ml conical centrifuge tubes (24). After incubation for 2 hr in a water bath a 37°C and for 2 to 3 days in a refrigerator at 4 to 8°C, the mixtures were centrifuged at room temperature for 10 min at 4,600 g. The supernatants ('soluble complexes') were used for experiments. Cytotoxicity tests. The tests were performed in plastic microcytotoxicity plates (Falcon 3034, Falcon Plastics, Los Angeles, Calif., USA), following the method of Takasugi & Klein (23). Two hundred tumor cells in the growth medium described above were plated in each well, resulting in about 50 to 100 attached cells during overnight incubation at 37°C in humified 5% CO2 air atmosphere. Purified mononuclear cells, also suspended in the same medium, were added to each well in 20-jal volumes. The target to effector cell ratio was 1:250 (2.5 X lO** cells/well). The incubation was continued for 48 hr, after which the plates were washed three times in PBS to remove nonadherent cells, fixed in methanol, air-dried, and stained with MGG. Six to twelve wells of each effector-target cell combination were prepared, with the control effector cells always on the same plate. The remaining cells in each well were counted and the average ± standard error determined. Inhibition was expressed as percentage from the formula: % Cx = Control lymphocytes — Patient lymphocytes (C) (P) Control lymphocytes (C)

-XIOO.

Blocking tests. Blocking tests were performed either on the effector cells or on target cells. In the former case effector cells were incubated in the blocking material for 30 min at 37°C, washed once in PBS, and suspended in the growth medium before being added on the target cells. When blocking was tested on the target cells, these were first overlaid by 20 jal of the, PBS-diluted blocking material and washed once with PBS before addition of the effector cells suspended in growth medium. The

Blocking of Cellular Cytotoxicity

783

Table I. Effect of heat-aggregated IgG (AggrIgG) and non-aggregated IgG on the cell-mediated in vitro cytotoxicity of ovarian carcinoma patients Target cells *

Lymphocyte donor

Blocking combination •*

Surviving cells/well

+ SE

Percent reduc-

Ptt

tion t

Percent block-

Pof difference in block-

Ptt

ing t t t

ing t t OCLa OCLa OCLa OCLa OCLa OCLa OCLa OCLa OCLa

OvCaBa OvCaBa OvCaBa OvCaBa OvCaBa OvCaBa OvCaBa OvCaBa Control

None Autologous serum AggrIgG 1:10 IgG 1:10 AggrIgG 1:100 IgG 1:100 AggrIgG 1:1000 IgG 1:1000 None

OCMu OCMu OCMu OCMu OCMu OCMu OCMu OCMu OCMu

OvCaBa OvCaBa OvCaBa OvCaBa OvCaBa OvCaBa OvCaBa OvCaBa Control

None Autologous serum AggrIgG 1:10 IgG 1:10 AggrIgG 1:100 IgG 1:100 AggrIgG 1:1000 IgG 1:1000 None

393T 393T

OvCaBa Control

None None

31 + 4 32 + 3 32 + 5 31 + 3 47 + 5 32 + 2 43 ± 5 26 + 1 47 + 6 41 + 2 38 + 2 48 + 3 40 + 1 53 + 10 38 + 8 49±5 29 + 8 64 + 7 30 + 2 33±4

34

0.025

36 40 25 38 17 40 23 55 -

9

_ 6

33 33 34 0 31 9 45 -

6 0 100 6 75 -31 0.005

1 1 1 J 1 1

ns KJ,\J J

0 002 5

_ -13 30 -4 52 -13

ns 0.05 1 ns 1

34

0.05 \

-50 ns

ns ns ns 0.05 ns 0.05 ns

0.1

\

ns

J

ns

J

0.025 ns 0.0

_ -

* OCLa and OCMu are cell cultures derived from ovarian carcinoma ascitic fluid; 393T is a human osteosarcoma cell line. *• Starting concentration of IgG (before aggregation) was 160 f),g/ml. t The percentage reduction of number of cells/well when patient lymphocytes were added in comparison to wells with control lymphocytes. t t Significance of differences is calculated by Student's / test, ns = not significant. t t t The percentage abrogation of the reduction by patient lymphocytes alone by the blocking agent tested.

abrogation of the lymphocyte-mediated inhibition was calculated from the formula

pared with that of the corresponding non-aggregated IgG. Effector cells were incubated for 30 min at 37°C with various dilutions of these ^ _,, (C - P) - (C - P blocked) preparations in PBS, washed once, and suspend% Bl = J^ i—b ^ X 100. ed in growth medium before being added on C-P the target cells. The results are shown in Table Statistical analysis. The statistical significance I. At high concentrations (16 fig/ml, final conof differences was determined by Student's / centration) neither showed blocking activity, test. but at higher dilutions heat-aggregated IgG effectively blocked the cell-mediated inhibition of mononuclear cells from ovarian carcinoma paRESULTS tients against homologous carcinoma cells. No Effect of heat-aggregated IgG blocking activity could be detected in the In the first series of experiments the effect starting material IgG, and the differences were of heat-aggregated IgG (AggrIgG) was com- statistically highly significant (P ^ 0.005).

784

£. Saksela, K. Penttinen & S. Pyrhotteii

Table II. Effect of various dilutions of AggrIgG on the cell-mediated in vitro cytotoxicity of ovarian carcinoma patients Target cells *

Lymphocyte donor

Blocking combination **

Surviving cells/well

± SE 37±4

Percent reduction t

Ptt

Ca ov Ru Ca ov Ru Ca ov Ru Ca ov Ru Control

None Autologous serum AggrIgG 1:10 AggrIgG 1:100 None

50 36 24 13 -

0.005

48 + 5 56±4 65 + 6 75 + 4

OCHa OCHa OCHa

Ca ov Ru Ca ov Ru Control

None AggrIgG 1:1000 None

43 + 8 50 + 4 68 + 8

36 27 -

0.05

4T 4T

Ca ov Ku Control

None None

OCSl OCSl OCSl OCSl OCSl OCSl

Ca ov Ku Ca ov Ku Ca ov Ku Ca ov Ku Ca ov Ku Control

None Autologous serum AggrIgG 1:10 AggrIgG 1:100 AggrIgG 1:1000 None

4T 4T

Ca ov Ku Control

None None

41 + 2 38 + 3 48 + 3 62 + 4 50 + 2 65 + 2 156 + 21 138+15

-t-13 37 42 27 5 24 -

-fl3

Ptt

ing ttt

OCHa OCHa OCHa OCHa OCHa

135 + 25 119 + 16

Percent block-

29 51 73 -

ns 0.001

26 -

ns 0.005 0.0025

ns

_ _ -13 27 88 36 -

ns ns 0.001 0.005

ns

—

* OCHa and OCSl are cell cultures derived from ovarian carcinoma ascitic fluids; 4T is a human cell line derived from giant-cell synovial sarcoma. ** Starting concentration of AggrIgG was 160

t-i'tt For rest of footnotes, see Table I.

Because there was some indication of a zone effect dependent on the concentration of heataggregated IgG in the blocking suspension, further experiments with various dilutions of AggrIgG were carried out. The results in Table II show that, in the system used, the optimal blocking effect could be obtained with around 1.6 ^g/ml AggrIgG in PBS, and both higher and lower dilutions were less efficient. The blocking mechanism was further analyzed by comparing the blocking effect of AggrIgG when administered on the effector cells as above and when added on target cells followed by a PBS wash before addition of the effector cells. In these experiments a single concentration of 1.6 ;ug/ml of AggrIgG was used. The results are shown in Table III. In four of five experiments a significant blocking of the cell-mediated tumor cell inhibition was

obtained when AggrIgG was added on the effector cells, whereas AggrIgG was ineffective in all experiments when added on the target cells.

Effect of antigen—antibody complexes Complexes of purified, preservative-free tetanus toxoid and anti-tetanus serum of high specific activity were prepared as described above by combining equal volumes of serum and various dilutions of TT (1,400 Lf*/ml), starting from 1:8. This material was further diluted 1:4 in the effector cell incubation mixture, since higher concentrations of TT could be toxic to cultured cells, as shown in preliminary experiments. * Lf = : limit of flocculation.

Blocking of Cellular Cyiotoxicit'y 785 Table III. Effect of AggrIgG on effector cells and taiget cells in cell-mediated cytotoxicity of ovarian carcinoma patients Target cells *

Lymphocyte donor

Blocking combination **

Surviving . cells/well

Percent reduction t

+ SE

Pii

Percent block-

P-fi

Pol

difference in block-

ing ttt

ing t t OCMii OCMa OCMa OCMii OCMa 4T 4T

Ca ov To Ca ov To Ca ov To Ca ov To Control Ca ov To Control

None Autologous serum AggrIgG on lymph. AggrIgG on targets None None None

16 + 1 40 + 7 49 + 7 15 + 2 48 + 4 67 + 5 53 + 6

OCKo OCKo OCKo OCKo OCKo 4T 4T

Ca ov Ap Ca ov Ap Ca ov Ap Ca ov Ap Control Ca ov Ap Control

None Autologous serum AggrIgG on lymph. AggrIgG on targets None None None

39 ± 5 62 + 8 43 + 3 33 + 6 60 + 5 66 + 8 53 + 6 30 + 6 38 + 4 66 + 4 23 + 3 55 + 6 31 + 3 32 + 4

OCMu OCMu OCMu OCMu OCMu 4T

Ca ov Ku Ca ov Ku Ca ov Ku Ca ov Ku Control Ca ov Ku Control

None Autologous serum AggrIgG on lymph. AggrIgG on targets None None None

OCMu OCMu OCMu OCMu OCMu 4T 4T

Ca ov To Ca ov To Ca ov To Ca ov To Control Ca ov To Control

None Autologous serum AggrIgG on lymph. AggrIgG on targets None None None

OCKo OCKo OCKo OCKo OCKo 4T 4T

CaovKu Ca ov Ku Ca ov Ku CaovKu Control Ca ov Ku Control

None Autologous serum AggrIgG on lymph. AggrIgG on targets None None None

OCKo OCKo OCKo OCKo OCKo 4T 4T

Ca ov To Ca ov To Ca.ovTo. Ca ov To Control Ca ov To Control

4T

• None •

Autologous serum AggrIgG on lymph. AggrIgG on targets None None None

23 + 5 36 + 4 62 + 5 18 + 5 48 + 7 37 + 6 32+4 42 + 7 74+3 60 + 2 36 + 6 79 + 6 31 + 3 32 + 4 . 42 + 8 • 43 + 2 65 + 4 39 + 6 70 + 6 37 + 6 32 + 4

.

67

75 103

+2 69 _

+ 26 35

_

0.0005

17

0.01

+3

-3 ... _ _ 110 19 -29 _

28 45 -

0.01 0.005

"1

ns

J "•"'^^

0.05 ns ns

I

no

+ 24 -

-

45

_ 32

0.01

31 58 .-

•

3 52 25

ns

144 0.0025 -28. ns

+ 20

_

-

0.01 52 156 -20 _ _ -

0.05 0.0005

6

86

24 54

49

0.005 0.05

+29 63 10 47

- 1 6 ns —

3 -

44 10 —

ns

Y •

> 0.0005

0.005 ,

-r

40 39 7

\

J

-

1 i. 0.005"

. 0.01 ,- 0 82 -10 ... _

ns

0.05

1

ns

J 0.005

* OCMa, OCMu, and OCKo are cell cultures derived from carcinoma SiScitic fluid oi ovarian carcinoma patients; 4T is a human cell line from giant-cell synovial sarcoma. ** Concentration of AggrIgG is 0.16 ^ t - t t t For rest of footnotes, see Table I.

786

E. Saksela, K. Pentt'inen & S. Pyrhonen

Table IV. Effect of human antitetanus-tetanus toxoid complexes on the cell-mediated cytotoxicity of ovarian carcinoma patients Target cells *

Lymphocyte donor

Blocking combination **

OCKa OCKa OCKa OCKa OCKa OCKa OCKa OCKa 393T 393T

Ca ov Ba Ca ov Ba Ca ov Ba Ca ov Ba Ca ov Ba Ca ov Ba Ca ov Ba Control Ca ov Ba Control

None Autologous serum Anti-TT TT 5 X Ag excess Ag/Ab eq. 5 X Ab excess None None None

OCKa OCKa OCKa OCKa OCKa OCKa OCKa OCKa 39 3T 393T

Ca ov Sa Ca ov Sa Ca ov Sa Ca ov Sa Ca ov Sa Ca ov Sa Ca ov Sa Ca ov Sa Ca ov Sa Control

None Autologous serum Anti-TT

TT 5 X Ag excess Ag/Ab eq. 5 X Ab excess Healthy None None

Surviving cells/well + SE

25 + 25 + 28 + 28 + 48 + 43 +

2 2 3 4 3 5

37±4

43 + 2 17 + 3 15 + 2

Percent reduction t

Ptt

42 42 35 35 -fl2 0 14 _ -fl3 -

0.005

13 + 5 47 + 5 18 + 6 12 + 2 32 + 5 21 + 7 30 + 10 47 + 3 15 + 2 16+3

Ptt

ing ttt _ 0 17 17 127 100 67 -

ns ns 0.0025 0.01 • 0.025

ns _

72 0 62 74 32 55 36 _

6 -

Percent block-

100 14 -3 56 24 50

0.0025 ns ns 0.01 ns 0.1

ns -

* OCKa is a cell culture derived from ovarian carcinoma ascitic fluid; 393T is a human osteosarcoma cell line. ** The anti-TT (tetanus toxoid) serum was always used in 1:2 final dilution. The starting material of tetanus toxoid contained 1,400 limit of flocculation (Lf)/ml. Final dilutions of this were 1:32 in 5 X Ag excess, 1:160 in Ag/Ab equivalence, and 1:800 in 5 X Ab excess. TT alone was tested at 1:32 final dilution. t - t t t For rest of footnotes, see Table I.

DISCUSSION When the patients' purified mononuclear cells were preincubated with these complexes, an effective blocking of their tumor cell inhibitory effect could be obtained (Table IV). This was directly proportional to the amount of antigen present, as indicated by the results when mixtures with various antigen concentrations were used, as seen in the table. When experiments were performed by using complexes with a constant amount of antigen (7 Lf/ml), the blocking obtained with complexes in 5 X Ag or 5 X Ab excess (data not shown) was similar and effective. When solutions of tetanus-antitetanus complexes were first added on the target cells and washed before the addition of effector cells, no blocking effect could be seen.

The results show that effective blocking of the tumor-specific cell-mediated in vitro inhibition of cultured tumor cells can be obtained by heat-aggtegated human IgG and immune complexes unrelated to the tumor. Heat-aggregated IgG is known to have a high affinity for Fc receptors occurring on the surface of B lymphocytes but not on non-activated T lymphocytes (3, 10). Fc portions with high affinity are likewise present in antigen-antibody complexes. Since blocking is obtained with preincubation of the effector cells and not of the target cells with AggrIgG or immune complexes, the findings strongly suggest that this effect is mediated through Fc receptors of the effector cells, which could, according to this criterion, possibly be-

Blocking of Cellular Cytotoxicity

long to non-T populations, although activated T-cells have also been shown to carry Fc receptors (26). This finding is in agreement with the recent reports of O'Toole et al. (16) and Lamon et al. (13) indicating that non-T cells are effective in mediating tumor-specific iri vitro inhibition in certain human as well as syhgeneic animal tumor systems. The effector cell(s) involved in mediating human tumor-specific inhibition in microcytotoxicity assays has certain controversial characteristics. In the human melanoma system (25) and in the present ovarian carcinoma system (Saksela & Virolainen. Unpublished results) the effector cells seem to carry receptors for sheep erythrocytes (SRBC), as T cells do, since they can be greatly enriched by collecting the SRBC rosette-forming cells with gradient sedimentation. Therefore, since the fractionated cells are also blocked AggrIgG, the effector cells involved would seem to be mononuclear cells carrying both Fc receptors and SRBC receptors. However, different classes of lymphocytes may be involved in different stages of tumor progression (11, 12), which makes it difficult to compare results obtained in different systems and with few markers. Whether the surface characteristics assigned to the 'effector cell' reflect properties of a unique cell type or of different classes of lymphocytes operative at different stages remains to be tested by simultaneous characterization with several markers. In many respects the present system resembles antibody-mediated target cell lysis (14, 19). It has been shown that also in the latter can effector cells adsorb antigen-antibody complexes (18), are effector cells removed by anti-IgGcontaining columns (21), and can unrelated antigen-antibody complexes block the Fc-mediated binding of the antibody responsible for the specific cytotoxicity against antigen-coated target cells (15). In human tumor patients great specificity has been observed in the serum-mediated blocking of cellular cytotoxicity. Serum from melanoma patients can block lymphocytes of other melanoma patients tested against melanoma cells but not the inhibitory action of lymphocytes from colon carcinoma patient against colon carcinoma

787

cells, which again are specifically blocked by colon carcinoma serum. Also, some of the blocking autologous sera in the present experiments (Tables I to V), when tested for specificity, blocked only in the ovarian carcinoma system. This effect apparently cannot, then, be mediated through the Fc receptor pathway, which seems to represent another, unspecific way of blocking. The specific effector cell blocking could be mediated through binding of free antigen to specific receptor sites or to free Ab sites of possible arming complexes, as demonstrated on mouse mononuclear cells made specifically cytotoxic by Fc-mediated attachment of antigen-antibody complexes in antibody excess (6). A third mechanism for blocking, also specific, can be demonstrated at the target cell level, where the antibody part of tumor antigen-antibody complexes apparently results in binding and coating of the antigenic sites. The strongest evidence for this is the finding that blocking activity can be removed from the serum by absorption with tumor cells of the same type but not with other cells (22). If the activated Fc part of IgG can exert similar blocking effects both in vitro and in vivo, it may play a role in tumor immunology. Practically it may explain the occasional blocking effect of control sera in cytotoxicity tests, since in some cases aggregation of IgG and activation of Fc have been shown to occur already at the temperature used for routine inactivation of complement (17). Conceivably, if Fc-mediated blocking of tumor-specific cellular immunity can also operate in vivo, certain chronic infectious diseases where antigen-antibody complexes are formed continuously (4) might disturb the immunologic balance and facilitate the escape of tumor growth from immunologic control. ACKNOWLEDGEMENTS The technical assistance of Mrs. Maija-Liisa Mantyla is gratefully acknowledged. The investigation was financially supported by grants from the Finnish Cancer Society, Sigrid Juselius Foundation, and the National Research Council for Medical Sciences, Finland.

788

E. Saksela, K. Penttinen &S:. Pyrhonen

The anti-tetanus toxoid antiserum was a generous gift from Dr. Odd Wager (Aurora Hospital, Helsinki, Finland) who also kindly prepared the complexes. Preservative-free tetanus toxoid was provided by Dr. Lauri Jannes, Central Laboratory of Public Health, Helsinki, Finland.

REFERENCES 1. Baldwin, R. W., Price, M. R. & Robins, R. A. Blocking of lymphocyte-mediated cytotoxicity for rat hepatoma cells by tumor-specific antigen-antibody complexes. Nature (Lond.) 238, 185, 1972. 2. Baldwin, R. W., Price, M. R. & Robins, R. A. Inhibition of hepatoma-immune lymph-node cell cytotoxicity by tumor-bearer serum and solubilized hepatoma antigens. Int. J. Cancer. 11, 527, 1973. 3. Basten, A., Miller, J. F. A. P., Sprent, J. & Pye, J. Receptors for antigen-antibody complexes used to separate T cells from B cells. Nature (Lond.) 235, 178, 1972. 4. Baumal, R. & Broder, I. Studies into the occurrence of soluble antigen-antibody complexes in disease. III. R. A. and other human diseases. Clin. exp. Immunol. 3, 558, 1968. 5. Boyum, A. Isolation of mononuclear cells and granulocytes from human blood. Scand. J. clin. Lab. Invest. 21, Suppl. 97, 77, 1968. 6. Greenberg, A. H. & Shen, L. A class of specific cytotoxic cells demonstrated in vitro by arming with antigen-antibody complexes. Nature [New Biol.] (Lond.) 245, 282, 1973. 7. Hannestad, K. Monoclonal and polyclonal gammaM rheumatoid factors with anti-di and anti-trinitrophenyl activity. Clin. exp. Immunol. 4, 555, 1969. 8. Hellstrom, I., Hellstrom, K. E., Sjogren, H. O. & Wagner, G. A. Demonstration of cell-mediated immunity to human neoplasms of various histological types. Int. ]. Cancer. 7, 1, 1971. 9. Hellstrom, I., Sjogren, H. O., Wagner, G. & Hellstrom, K. E. Blocking of cell-mediated tumor immunity by sera from patients with growing neoplasms. Int. f. Cancer. 7, 226, 1971. 10. Jondal, M., Wigzell, H. & Ainti, F. Human lymphocyte subpopulations: classification according to surface workers and/or functional characteristics. Transplant. Rev. 16, 163, 1973. 11. Lamon, E. W., Andersson, B., Wigzell, H., Eenyo, E. M. & Klein, E. The immune response to primary moloney sarcoma virus tumors in BALB/c mice: cellular and humoral activity of longterm regressors. Int. J. Cancer. 13, 91, 1974. 12. Lamon, E. W., Skurzak, H. M., Klein, E. & Wigzell, H. In vitro cytotoxicity. by a nonthymusprocessed lymphocyte population with specificity

for a virally determined tumor cell surface antigen. / . exp. Med. 136, 1072, 1972. 13. Lamon, E. W., Wigzell, H., Andersson, B. & Klein, E. Antitumour activity in vitro dependent on immune B lymphocytes. Nature [New Biol.} (Lond.) 244, 209, 1973. 14. MacLennan, L C. M. Antibody in the induction and inhibition of lymphocyte cytotoxicity. Transplant. Rev. 13, 61, 1972. 15. MacLennan, I. C. M. Competition for receptors for immunoglobulin on cytotoxic lymphocytes. Clin. exp. Immunol. 10, 275, 1972. 16. O'Toole, C , Perlmann, P., Wigzell, H., Unsgaard, B. & Zetterlund, C. G. Lymphocyte cytotoxicity in bladder cancer. No requirement for thymusderived effector cells? Lancet 1, 1085, 1973. 17. Penttinen, K., Wager, O., Rasiinen, J. A., Myllyla, G. & Haapanen, E. Platelet aggregation and CryoIgM in the study of hepatitis and immune complex states. Clin. exp. Immunol. 15, 409, 1973. 18. Perlmann, P., Perlmann, H. & Biberfeld, P. Specifically cytotoxic lymphocytes produced by preincubation with antibody-complexed target cells. /. Immunol. 108, 558, 1972. 19. Perlmann, P., Perlmann, H. & Wigzell, H. Lymphocyte mediated cytotoxicity in vitro. Induction and inhibition by humoral antibody and nature of effector cells. Transplant. Rev. 13, 91, 1972. 20. Ponten, J. & Saksela, E. Two established in vitro cell lines form human mesenchymal tumours. Int. } . Cancer 2, 434, 1967. 21. Schirrmacher, V. & Golstein, P. Cytotoxic immune cells with specificity for defined soluble antigens. I. Assay with antigen-coated target cells. Cell. Immunol. 9, 198, 1973. 22. Sjogren, H. O., Hellstrom, I., Bansal, S. C. & Hellstrom, K. E. Suggestive evidence that the 'blocking antibodies' of tumor-bearing individuals may be antigen-antibody complexes. Proc. nat. Acad. Sci. (Wash.) 68, 1372, 1971. 23.Takasugi, M. & Klein, E. A micro-assay for cellmediated tumor immunity. Transplantation 9, 219, 1970. 24. Wasastjerna, C , Klemola, H., Rasanen, J. A. & Wager, O. IgG cryoglobulinaemia. Case report and immunological studies of a patient with recurrent ulcerative stomatitis and high content of cold precipitable immunoglobulin in serum. Scand. J. Haemat. 4, 473, 1967. 25. Wybran, J., Hellstrom, L, Hellstrom, K. E. & Fudenbergh, H. H. Cytotoxicity of human rosetteforming blood lymphocytes on cultivated human tumor cells. Int. J. Cancer 13. In press, 1974. 26. Yoshida, T. O. & Andersson, B. Evidence for a receptor recognizing antigen complexed immunoglobulin on the surface of activated mouse thymus lymphocytes. Scand. J. Immunol. 1, 401, 1972. Received 25 April 1974 Received in revised form 2 August 1974