effects on in vitro tumor growth of macrophages isolated from human ascitic ovarian tumors

Znt. J . Cancer: 23, 157-164 (1979)

EFFECTS ON I N VITRO TUMOR GROWTH OF MACROPHAGES ISOLATED FROM HUMAN ASCITIC OVARIAN TUMORS

Albert0 MANTOVANI l, Giuseppe PERI l, Nadia POLENTARUTTI l, Giorgio BOLIS 2, Costantino MANGIONI and Federico SPREAFICO

Ostetrica e Ginecologica dell’ Universita di Milano, Via Commenda 12, 20100 Milan, Italy Istituto di Ricerche Farmacologiche ‘‘ Mario Negri ”, Via Eritrea 62, 20157 Milan; and Prima Clinica

Macrophages were isolated from 22 human ascitic ovarian epithelial tumors and their growth- inhibitory capacity was tested using as targets the following in vitro tumor cell liner: murine TLXP lymphoma and FS6 sarcoma; human myeloid K562 leukemia and human E cell line derived from an ovarian carcinoma. Macrophage preparations were heterogeneous in their interaction w i th tumor target cells, and assay conditions, such as the type of target cell, incubation time, and attacker t o target cell (A:T) rat io critically affected the evaluation of the cytotoxic potential of tumor-associated macrophages. A t an A:T rat io of 7:l no cytostatic activity on TLXP and K562 cells was ever observed, but in the presence of specific antibody 8 out of 12 macrophage preparations tested showed significant antibody-dependent cytotoxicity on TLXP lymphoma cells. Macrophage preparations f rom two patients significantly inhibited growth of the FS6 sarcoma and a cytostatic activity on E cells was observed in five additional patients. Significant stimulation of the proliferative capacity of a t least one of the target cell lines was observed in 11 subjects at an A:T rat io of 7:l. I n 12 patients, macrophage cytostatic activity on E cells was also tested a t an A:T rat io of 35:l; eight out of 12 preparations showed significant cytotoxicity under these conditions. When the same subject was repeatedly tested a t short intervals the same pattern of inhibition or stimulation of tumor growth was observed.

Macrophages are a major component of the lymphoreticular infiltrate of experimental and human tumors (Evans, 1972, 1973, 1976; Pross and Kerbel, 1976; Gauci and Alexander, 1975; Wood and Gollahon, 1977). In rodent tumors, there appears to be a correlation between the macrophage content of tumors, on the one hand, and tumor immuno- genicity and metastasizing capacity, on the other (Eccles and Alexander, 1974).

Macrophages isolated from experimental neo- plasms can express specific or usually non-specific cytotoxicity (lysis or stasis) on tumor cells in vitro (Evans, 1976). The in vivo significance of the in vitro cytotoxic activity of tumor-associated macrophages has not been firmly established (Evans, 1976), but there is evidence in experimental systems consistent with the hypothesis that tumor-associated macrophages may play a role in the regulation of tumor growth and metastasis (Birbeck and Carter, 1972; Eccles and Alexander, 1974; Russell and McIntosh, 1977; Wood and Gillespie, 1975; Zarling and Tevethia, 1973; Mantovani, 1978).

No data concerning the isolation and cytotoxic activity of human tumor-associated macrophages

is available and, in general, little is known about the effects of human mononuclear phagocytes on tumor cells (Holtermann et al., 1974). The present study was designed to isolate pure macrophage preparations from human ascitic ovarian tumors and to characterize their in vitro effects on tumor growth.

MATERIAL AND METHODS

Patients Ascitic fluid (200-3,000 ml) was obtained by

paracentesis or at laparatomy from 22 patients with histologically proven ovarian epithelial tumors (stages 111 and IV). The presence of tumor cells in the ascitic fluid was checked by an independent pathologist. Twelve patients had never received either chemotherapy or radiotherapy and the remaining 10 subjects had not been given chemotherapy during the 4 weeks before the assay.

Macrophages Ascitic fluid was centrifuged a t 4OOg for Smin,

the pellet was resuspended in 0.14 M NH,CI to lyse contaminating red cells and the cells were centrifuged at 400s for 5 min. The cells were resuspended in basal medium Eagle (BME) and 20 ml of the cell suspension, at a concentration of approximately 1 x lo6 cells/ml, were stratified on 10ml Ficoll- Hypaque (MSL, Eurobio, Paris, France). The tubes (cat. 25,330 Corning, N. Y., USA) were then centrifuged at 400 g for 20 min at room temperature and the mononuclear cell layer was aspirated. This step was introduced in the separation procedure after preliminary experiments had shown that, in some ascitic fluids, polymorphonuclear leukocytes represented up to 90 % of the adherent population. After washing twice with 50ml BME, the number of mononuclear phagocytes was estimated by neutral red uptake and approximately 3.5 x lo6 macrophages in 1 ml BME were seeded in the wells of Costar trays (3524 Costar, Cambridge, Mass., USA). These culture conditions were selected after preliminary experiments had shown that in the presence of serum, which provides optimal conditions for adherence of human circulating monocytes (Koller et al., 1973; Mantovani et al., 1977a), a large number of non-phagocytic (possibly tumor) cells adhered to

Received: October 2, 1978. Abbreviations used: BME, basal medium Eagle; A:T,

attacker to target cell; ADCC, antibody-dependent cellular cytotoxicity; NK, natural killer.

158 MANTOVANI ET AL.

FIGURE 1 - Mononuclear cells were incubated for 30 min in serum-free BME and non-adherent cells were removed with jets of medium. Giemsa, phase contrast, x 200.

plastic. After 1 h of incubation at 37" C the wells were thoroughly washed with jets of medium. Virtually all adherent cells were mononuclear phagocytes as assessed by the following criteria: (1) cells adhered and rapidly spread out on plastic or glass surfaces in serum-free BME and had the typical morphology of mature macrophages (Fig. 1) (Evans, 1972; 1973); (2) adherent cells could take up neutral red (0.03% in BME); (3) trypsin (0.1 or 0.25% of 30 min in BME) did not detach adherent mononuclear cells as shown for murine tumor- associated macrophages (Evans, 1972, 1973); (4) when incubated for 60 min in growth medium with 1 :20 diluted latex particles (3102-65 Difco, Detroit, Mich., USA) more than 90% of the adherent cells were phagocytic; (5) in the presence of yeast particles, coated with complement as described by Huber and Wigzell (1975), at a ratio of 100 particles per macrophage, binding and phagocytosis were observed after 60min of incubation; (6) in the presence of antibody-coated sheep erythrocytes prepared as previously described (Mantovani et al., 1976), binding and phagocytosis of red cells were observed over a 60-min incubation period in growth medium. The number of macrophages actually adhering to the

wells was routinely evaluated by direct counting of Giemsa-stained preparations with the aid of an eyegrid : by this method the calculated percentage (mean$sE) of phagocytic adherent cells in the ascitic fluid in 27 experiments was 35210 .4% (range 4-62 %) of the total mononuclear cells.

Monocytes Monocytes were obtained from heparinized

peripheral venous blood of 10 normal healthy adult volunteers, as previously described (Mantovani et al., 1977~). Briefly, mononuclear cells were separated on Ficoll-Hypaque as described above and washed twice with 50 ml of BME. Monocytes (3.5 x lo5 per well in 1 ml of RPMI 1640 medium with 20% fetal bovine serum) were separated by adherence for 1 h at 37" C followed by washing with BME.

Target cells Tumor cell lines used as targets were: murine

TLX9 lymphoma and FS6 sarcoma of C57BL/6 origin; human K562 myeloid leukemia, a gift from Dr. G. Bonnard, National Cancer Institute, Bethesda, Md. (Lozzio and Lozzio, 1973) and E carcinoma line, originally derived from an ovarian tumor, obtained through the courtesy of Drs. L. Curatolo and L. Morasca of this Institute. All cell lines were maintained in RPMI 1640 medium (Eurobio, Paris, France) supplemented with 10 % fetal bovine serum and 50 pg/ml gentamycin (growth medium). TLX9 and K562 cells grew as suspension cultures whereas the FS6 and E lines adhered to plastic and were collected by exposure for 5 min to 0.25% trypsin and 0.02% EDTA. Previous data indicated that these tumor cells are susceptible to the non-specific growth-inhibitory activity of murine macrophages (Mantovani, 1977, 1978; Mantovani et al., 19776; Kurland et a/., 1978).

Growth inhibition assay Cytotoxicity was determined essentially as pre-

viously described (Mantovani, 1977; Mantovani et al., 1977a, b). Briefly, tumor cells were seeded on macrophage monolayers at a concentration of 5 x lo4 cells in 1 ml growth medium, resulting in an attacker to target cell (A:T) ratio of approximately 7:1, unless otherwise stated. The samples were incubated for 48 h in air with 5 % CO, at 37" C, unless otherwise specified, and tumor growth was checked under an inverted microscope. At the end of incubation, the medium was removed from wells containing adherent FS6 and E target cells and the samples were incubated for an additional 4-6 h with 1 ml growth medium containing 0.5pCi [3H]- thymidine (specific activity 1.9 Ci/mrnol, Schwarz Mann., Orangeburg, N. Y., USA). When TLX9 and K562 cells were used as targets, tumor cells were vigorously resuspended, transferred to plastic tubes and washed as previously described (Mantovani, 1977; Mantovani et al., 1977a, b) prior to pulsing with [3H]thymidine as described above. The percentage of inhibition or stimulation of [3H]thymidine uptake was calculated as: (1 -A/B) x 100, where A is the isotope uptake by tumor cells with macrophages and B is the isotope incorporated by tumor cells alone. Samples containing macrophages alone were

CYTOTOXICLTY OF HUMAN TUMOR MACROPHAGES 159

TABLE 1

EFFECTS ON I N VITRO TUMOR-PROLIFERATIVE CAPACITY OF HUMAN MACROPHAGES ISOLATED FROM ASCITIC OVARIAN TUMORS

Target ;ells

TLX9 K562 FS6 E Patient Macrophages

A.G.

Av.G.

B.V.

C.A.

C.G.

C.M.

D'A.M.

De G.A.

E.M.

F.C.

F.I.

F.M.

F.R.

G.C.

G.G.

G.M.

G.R.

G.S.

Lo G.P.

M.B.

M.G.

P.R.

NT

NT

142.2 rt 5.4 150.1 1 6 133.614.2 120.6 f 7

NT

NT

118.2+6 11 1.3 17 .8 64.5 1 1.5

95+3.9 NT

NT

25.5f5.4 25.011.3

NT

63.4h0.5 141.1k18.6 2, '

5.9f2.3 66.4f6.1 3*

NT

NT

34.6113 148.4It10.5 39

NT

NT

NT

NT

NT

NT

NT

32.7f6.2 42.6~k13.8 36.413.4 49.3rt0.4

NT

NT

10.912.2 9.2k0.6

NT

NT

24.2k5.1 17.710.2

NT

NT

NT

31.6f3.8 41.4rt1.8

NT

NT

10612 93.7*7.2

NT

NT

NT

NT

NT

NT

NT

63.712.2 60.6f0.6 35.21-2.3 50.5 12.7 9.9f0.1

12.810.4 NT

7.210.4 10.611.2 23.612.4

27k0.8 NT

156.7*4.3 42f3.2 39

48.2f4.2 98.4k4.2 3,

8It0.1 20.4f0.6 39

NT

199.1 f6.8 269.211.7 39

5.62~0.4 56.2 &0.06 3,

NT

25.8 5 1.8 14.6f0.2

NT

9.9k0.1 20.2It0.8 3,

67.2%2.9 57.9% 1.1

NT

NT

19.6It0.3 7.210.3 x, 2.3 10.03 2.810.03

24.7%2.2 28.8f1 7.1 1 0 . 3 4.410.06

12.4%0.6 10.8 k0.2 7.510.3 6.0 f 0.3 7.6k1.4 6.7f0.7

10.5 10.1 24.810.2 3, 11.5 It0.5 13.910.4

NT

NT

7.1 10.5 3.8k0.1

12.410.2 7.410.4 37

6.1 10.1 6.8k0.3

NT

9.610.5 lO.O&O.l

NT

38.9f 10.3 55.4+3 12.4 50.6 2.9k0.2 7.1 1 0 . 5 5.5f0.9 6.9 10.9 7.250.3 7.4k0.3 7.4f0.4

1 Results are presented as cpmfsa ( x loa). The A:T ratio was 7 : l . - p i 0 . 0 5 . - * p<O.Ol. - ' Not tested. - Microscopically evident inhibition or stimulation.

included in all experiments, but since their [3H]- Inhibition of upake of DNA precursors, used here thymidine uptake was negligible (tlOO cpm), this as a measure of cytostasis can be mediated by soluble value was not taken into account in the calculations. thymidine-like compounds and thus does not Results presented are mean +sE of three replicates necessarily imply actual growth inhibition (Evans per group and statistical significance was assessed by and Booth, 1976). However, in the assay used in Student's t-test. these studies, cells were washed prior to pulsing with


[3H]thymidine and, although cell counts were not routinely made, in the presence of cytotoxicity values >50 %, inhibition of tumor-cell proliferation was evident under an inverted microscope. Moreover, direct tumor-cell counts performed in a number of experiments gave cytotoxicity data consistent with the isotope uptake results. Thus, under these conditions, [3H]thymidine uptake can be taken as a measure of tumor-cell proliferative capacity.

ADCC assay

Antibody-dependent cellular cytatoxicity (ADCC) was assayed as previously described for human monocyte-derived macrophages (Mantovani et al., 1977~). Briefly, TLX9 lymphoma cells ( 5 ~ 1 0 ~ in 1 ml growth medium) were cultivated for 24 h on macrophage preparations in the presence of 1 :20,000 diluted rabbit anti-TLX9 serum. At the end of the incubation, lymphoma cells were counted, washed and pulsed with 0.5 pCi [3H]thymidine as described (Mantovani et al., 1977~). The percentage antibody- dependent inhibition of [3H]thymidine uptake was calculated as: (1 -A/B) x 100, where A is the isotope uptake with antibody and B is the isotope incorporated in control cultures. The same formula was employed to calculate growth inhibition in terms of cell counts, A and B indicating in this case the number of lymphoma cells with and without antiserum, respectively. The rabbit anti-TLX9 serum used in these experiments did not affect growth and DNA synthesis cf lymphoma cells in the absence of effector cells.

RESULTS

All patients except nine were tested with at least two tumor target cells at an A:T ratio of 7: l (Table I). Macrophages isolated from human ascitic ovarian tumors never showed cytostatic activity on TLX9 or K562 leukemia cells. Two patients (G.R. and F. C.) had macrophages which significantly inhibited FS6 sarcoma proliferative capacity (73 and

TABLE 11

ANTIBODY-DEPENDENT CYTOSTATIC ACTlVlTY OF HUMAN TUMOR-ASSOCIATED MACROPHAGES

ON TLX9 LYMPHOMA CELLS

Patient % inhibition of ['Hlthymidine uptake

B.V. C.A. D'A.M. De G.A. F.C. F.M. F.R. G.C. G.G. G.R. Lo G.P. M.B.

Percentage inhibition of tumor growth as determined by cell counts is shown in parenthesis. - * p <O.OS. - * p <0.01.

TABLE 111

CYTOSTATIC ACTIVITY OF HUMAN TUMOR- ASSOCIATED MACROPHAGES ON E TARGET CELLS

TESTED AT DIFFERENT A:T RATIOS

% inhibition (-) or stimulation (+) of IPH]thymidine uptake

7:1 ' 35:l ' Patient

B.V. C.A. C.G. C.M. D'A.M. De G.A. F.M. G.C. G.S. Lo G.P. M.B. M.G.

f 1 7 1 . 4 -28+3

0 - 1 2 1 4 - 13 i 10,

- 4 6 i 2 +10+10

81.3 -771.3 +9&12

4 ~ t 2

+135+5

-54+1 -39+3

-32hI a -37+7 -19&24 --45*3 -36r t l '

0 -79*3 - 3 9 1 2

4 i 3

0

'The A:T ratios of 7:l and 35:l were obtained using 5 x lo4 and 10' E cells respectively as targets. - * p <0.05. - a p <0.01.

43%) and five (A.G., C.A., Lo G. P., F.M. and F.R.) showed a significant cytostatic activity on E target cells, cytotoxicity values ranging from 77 to 28% (Table I). Significant stimulation of the proliferative activity of at least one of the target cells tested was observed in 11 patients (Table 1). As regards the effects of the same macrophage preparation on different target cells, in some patients different results were obtained with the various target cells. In fact, four macrophage preparations (C.A., Lo G.P., F.M., F.R.) which inhibited E-cell proliferation, in the same experiments enhanced growth of the FS6, K562 or TLX9 tumor.

Similarly patient G . R. who had cytostatic activity on FS6 cells, stimulated the proliferative capacity of TLX9 lymphoma cells. Overall, of all the cell lines tested, E cells appeared to be the most sensitive to the cytostatic activity of human macrophages isolated from ascitic ovarian tumors and the least susceptible to the growth-enhancing capacity of these effector cells; in only one subject (De G.A.) was a stimulatory effect seen on E carcinoma cells (Table I).

As a positive control that, under these experimental conditions, human tumor-associated macrophages could express cytotoxicity on tumor cells, their growth-inhibitory capacity on antibody- coated TLX9 lymphoma cells was evaluated. As shown in Table 11, eight out of 12 macrophage preparations significantly inhibited [3H]thymidine uptake by TLX9 lymphoma cells in the presence of specific antibody, cytotoxicity values ranging from 95 to 40%. Similar results were obtained when lymphoma cells were counted microscopically. Moreover, significant ADCC was also observed with two macrophage preparations which, in preliminary experiments, were not tested for direct cytostatic activity on other tumor cell lines.

CYTOTOXICITY OF HUMAl

The cytotoxicity tests presented in Table I were made using an A:T ratio of 7:1, at which rodent tumor-associated or peritoneal macrophages can reportedly express cytotoxic activity on tumor cells (Keller, 1973, 1976; Kurland et a[., 1978; Russell and McIntosh, 1977). When sufficient numbers of macrophages were available, a higher (35:l) A:T ratio was also employed by reducing the number of E target cells to lo4. As shown in Table 111, eight out of 12 macrophage preparations tested at an A:T of 35:l significantly inhibited the proliferative capacity of E target cells. Macrophages obtained from patient De G.A., which at an A:T ratio of 7 : l more than doubled [3H]thymidine uptake by E tumor cells, had no significant effect at the ratio of 35: l . Three macrophage preparations had no effect on tumor growth at A :T ratios of both 7 and 35 :1.

In a few cases the cytotoxic activity of human macrophages could be evaluated on two occasions in the same patient at short intervals. Table IV presents the results obtained with E target cells; macrophages obtained from the same subject constantly gave the same pattern of activity (inhibition or stimulation) on tumor cells. Similar data were obtained with different target cells (results not presented) or when ADCC effector capacity (Table V) was measured.

In four patients, E target cells were incubated for different times (24 or 48 h) in vitro with macrophages prior to measuring [3H]thymidine uptake. As shown in Table VI, macrophages from patients C.G., G.S. and M.G. significantly enhanced tumor-cell proliferative capacity at 24 h and had no effect at 48 h, whereas macrophages from patient Lo G.P. had no significant effect at 24 h of culture and markedly inhibited tumor-cell proliferation at 48 h. Keller (1976), working with rat macrophages, ieported a

TABLE IV

GROWTH-INHIBITORY OR ST'IMULATORY CAPACITY OF MACROPHAGES OBTAINED AT DIFFERENT

TIMES FROM THE SAME PATIENT

% inhibition (-1 or Patient , ~ ~ $ t $ i ~ ~ r ~ A:T stimulation (+) of

[JH]thymidine uptake _ _ ~ ~

A.G. 6/14/1978 '

7/6/1978 AvG . 6/5/1978

61611 978 C. A. 411 211978

4/18/1978

De G.A. 3/9/1978 4/4/1978

F.M. 3/30/1978

4/3/1978

7: l - 6 3 5 2 ' 7:l -2411 ' 7:l +20&3 7:1 1 1 7 1 5

35:l - 1 7 5 4 7:l -36zt2

35:l -39zt6 ' 7:l -28+7 7 : l t 135rt5 ' 7:l - t 1 7 0 1 8

35:1 - 4 9 1 3 ' 7:l -22h3

35:l -4553 7:l -46rt2 '

lp<O.Ol.- *p<O.O5.

\I TUMOR MACROPHAGES 161

TABLE V

ADCC OF MACROPHAGES OBTAINED AT DIFFERENT TIMES FROM THE SAME PATIENT

%inhibition of Patient Day Of testing ['Hlthymidine uptake

B.V. 3/1/1978 23 3/8/1978 6 4 2

C.A. 4/12/1978 75 4/18/1978 48

D'A.M. 3/3/1978 11 3/20/1978 19

De G.A. 3/29/1978 95 4/4/1978 88

F.R. 4/4/1978 30 4/6/1978 0

'p<O.OS.- *p<O.OI.

similar pattern of relative resistance of some target cells to macrophage cytostasis during early phases of the interaction, or even enhancement of growth.

The cytotoxic activity of most rodent tumor- associated macrophages is rapidly lost on culture (Evans, 1976). Macrophages from one patient (C.A.) which inhibited growth of E target cells at A:T ratios of 7 and 35 :1 (Table 111), were cultured in wells in the absence of tumor cells for 24 or 48 h prior to testing their cytotoxicity at an A:T ratio of 35:l . After 24 h of culture, macrophages significantly inhibited growth of E target cells whereas after 48 h macrophages were not significantly cytostatic. No gross loss of macrophages during the 48-h culture period could be seen microscopically and minor reductions in the number of macrophages are unlikely to account for the loss of cytotoxic activity as this macrophage preparation also showed cytostatic capacity at an A:T ratio of 7 : l (Table 111).

We could not obtain " normal " peritoneal macrophages from non-tumor-bearing patients undergoing minor abdominal surgery (the only appropriate control for macrophages obtained from ascitic ovarian tumors) in sufficient numbers for use in the growth-inhibition assay employed in the present investigation. Therefore, by way of comparison, the growth-inhibitory activity of circulating monocytes obtained from 10 normal adult donors was tested at an A:T ratio of 7 : l . As shown in Table VII, TLX9 and K562 cells were not significantly affected by monocytes, only one donor (P.G.) showing significant inhibitory activity on K 562 cells. All preparations inhibited the proliferative capacity of the murine FS6 sarcoma and four out of 10 monocyte preparations (BSZ, MA, PG and V. A.) showed significant cytostatic activity on E ovarian cells.

DISCUSSION

The results presented here show that human macrophages isolated from ascitic ovarian tumors are heterogeneous in their interaction with tumor target cells and that in vitro assay conditions, such as the target cell used, the A:T ratio and the incubation


TABLE VI

TIME COURSE OF THE EFFECTS OF MACROPHAGES ON I N VITRO TUMOR GROWTH

Incubation %,inhibition (-) or Patient A :T ratio time ' stimulation ( f ) of

(hours) ['Hlthymidine uptake

C.G. 7:l 24 + 4 8 1 6 48 0

48 0

48 + 8 5 3

48 0

Lo G.P. 7:1 24 +8&3 48 -77+3

48 -79f3 7:1 24 +67&6

48 + 4 f 2 35:1 24 +48*4

48 + 4 f 3

35:1 24 + 4 9 1 7 a

G.S. 7:1 24 +go513

35:l 24 +50i-2 a

35: l 24 0

M.G.

I Tumor cells were incubated with macrophages for 24 or 48 h prior to pulsing with [aHlthymidine. - 1p<0.05. - 'p<O.Ol.

time, critically affect evaluation of the cytotoxic potential of human macrophages essentially as shown in experimental systems (Keller, 1973, 1976). At an A:T ratio of 7:1 the tumor-cell lines used as targets showed differing susceptibilities to niacro- phage activity in a 48-h growth-inhibition assay. Tumor-associated macrophages never inhibited growth of TLX9 and K562 cells, but, at an A:T ratio of 7:1, most macrophage preparations (eight out of 12 tested) significantly inhibited growth of TLX9 lymphoma cells in the presence of specific rabbit antiserum. This indicates that, under these experimental conditions, most human tumor-associated macrophages have the potential to express cytostatic activity on these tumor cells. Moreover, these data

confirm that, in contrast to circulating monocytes, human mature macrophages can show ADCC on tumor cells, as previously described using in v i m matured monocyte-derived macrophages (Mantovani et al., 1977a). Macrophage monolayers did not exert cytotoxicity on K562 cells, thus excluding a significant contamination with natural killer (NK) cells. Since NK and K cells appear to be closely related, if not identical, cell populations (Herberman and Holden, 1978), this observation further indicates that the capacity of macrophage preparations of mediating ADCC was not due to contamination with non-adherent K cells. Two macrophage preparations out of 16 were cytostatic on FS6 sarcoma cells and four on E ovarian carcinoma cells. The importance of the target cell employed is stressed by the observation that macrophages cytostatic on E or FS6 target cells enhanced the proliferative capacity of other tumor cells in the same experiment. Of the tumor cell lines used in this study, E cells appear to be the most sensitive to the cytostatic activity of human macrophages isolated from ascitic ovarian tumors and the least susceptible to their growth-enhancing capacity. Similarly, rat proteose-peptone-induced macrophages reportedly have different cytostatic activity on different target cells and, as shown here with human macrophages and E cells, they tend to show lower levels of growth inhibition, or even enhance cell proliferation, in the early phases of the interaction with some target cells (Keller, 1976). Inhibition of E-cell proliferation was better observed at an A:T ratio of 35:1, eight out of 12 macrophage preparations being significantly cytostatic under these conditions. It has been noted in experimental systems that the cytotoxic activity of macrophages is critically affected by the A :T ratio employed (Keller, 1973, 1976; Kurland ef af., 1978) and cytotoxic mact ophages from strongly immunogenic rodent tumors can express cytotoxicity at A:T ratios of less than 1O:l (Russell and McIntosh, 1977; Holden ef al., 1976; Mantovani, 1978). In a still limited series of patients (four) the cytotoxic activity of macrophages isolated from ascitic ovarian tumors has been confirmed in a long-term cytolysis assay based on the release of [3H]thymidine from prelabelled target cells.

TABLE V11

GROWTH-INHIBITORY ACTIVITY OF CIRCULATING HUMAN MONOCYTES FROM NORMAL DONORS

%inhibition (-) or stimulation (+) of ['Hlthymidine uptake Donor

TLX9 K562 FS6 E

B.C. B.S.Z. C.M. G.R. H.C. M.A. P.G. P.N. T.S. V.A.

+513 b NTi'

-1oi-2 1 2 f 2 NT

- 1 5 1 4 -8+2 - 9 t 2

NT - 5 1-1

--3&9 -5f3 -5 5 4 - 8 k 4

- 2 1 2 -53513

- 3 5 1 -25rt5

-4+12

0

-69~1r2 ', --65.1:6 ', -7651 *, -7653 ', '' -70-t1 ', --55:t:6 *, -65 1.3 ', -63:t:l ', -661t2 ', -61 - t 2 ', '

+ I 6 110 -46+6 l,

0 - - 2 0 1 9

t 7*9 - 3 1 1 1 - 5 8 t 2 ' 9

- 1 3 1 4 + I 3 1 1 1 -47&10 ' 9 ''

p <0.05. - a p <O 01. - Microscopically evident cytotoxicity. - ' Not tested

CYTOTOXICITY OF HUMAN TUMOR MACROPHAGES 163

At an A:T ratio of 7:1, macrophage preparations ( I 1 out of 22) stimulated the proliferative capacity of at least one of the tumor lines, including, in one subject, E target cells. Rodent macrophages have previously been reported to stimulate in vitro tumor growth. Peritoneal macrophages from experimental animals reportedly enhanced growth of lymphoid tumors growing poorly in tissue culture (Nathan and Terry, 1975; Hewlett et al., 1977; Namba and Hanaoka, 1972) and also, under certain circum- stances, of in vitro cell lines (Keller, 1973, 1976; Krahenbuhl et al., 1976; Olivotto and Bomford, 1974). Cytotoxic macrophages infiltrating chemically- induced rodent sarcomas, lose their cytotoxic activity after culture and can promote growth of tumor cells (Evans, 1976). Moreover, macrophages tested immediately after isolation from a weakly immunogenic metastasizing murine sarcoma were found to stimulate in vitro tumor-proliferative capacity (Mantovani, 1978). The in vivo significance of the tumor-growth-enhancing effect of human and rodent tumor-associated macrophages and its possible relevance in the context of an immunostimulation theory of tumor development (Prehn, 1977) remain to be elucidated.

Attempts to obtain “ normal ” human peritoneal macrophages from patients undergoing minor abdominal surgery have so far failed to give enough cells for use in the cytotoxicity test used here. By way of comparison, the growth-inhibitory activity of monocytes obtained from 10 normal laboratory donors was investigated. Circulating human monocytes from normal healthy donors have recently been shown to express “natural” cytolytic and cytostatic activity on tumor cells (Mantovani et al., 1978) and a growth-inhibitory effect on FS6 (10 subjects) and E cells (four subjects) was in fact

observed. On the other hand, the murine TLX9 and RBL5 (unpublished data) lymphomas and the human K562 myeloid leukemia were resistant to monocyte-mediated cytotoxicity. However, circulating monocytes differ from mature macrophages, which were the predominant adherent cell type isolated from ascitic tumors, in many respects including proliferative activity, adherence, phagocytosis (Territo and Cline, 1976) and cytotoxicity against antibody-coated lymphoma cells (Mantovani et al., 1977a). Thus they cannot be considered an appropriate normal control for the tumor-associated macrophages investigated in the present study. It therefore remains to be ascertained whether the effects on in vitvo tumor growth of macrophages isolated from ovarian ascitic tumors are disease- related, possibly a consequence of cell-mediated immunity to ovarian tumor-associated antigens (Faiferman et al., 1977), or are expressed by normal human peritoneal macrophages.

Human macrophages can be easily isolated from ascitic ovarian tumors and their properties studied in vitro. They should thus provide a useful tool for investigating the immunobiology of human tumor- associated macrophages, a cell population which according to animal data, appears to play a role in the regulation of tumor growth and metastasis (Birbeck and Carter, 1972; Eccles and Alexander, 1974; Russell and McIntosh, 1977; Wood and Gillespie, 1975; Zarling and Tevethia, 1973 ; Mantovani, 1978).

ACKNOWLEDGEMENTS

This work was supported by CNR, Rome, Italy. We thank Mr. W. Luini for technical assistance and Mr. E. Erba for preparing photographs.

EFFETS DES MACROPHAGES ISOLES A PARTIR DE TUMEURS OVARIENNES ASCITIQUES HUMAINES SUR LA CROISSANCE TUMORALE I N VfTRO

Des macrophages ont ete isoles ti partii de 22 tumeurs Bpitheliales ovariennes ascitiques et leur capacite d’inhibition de la croissance a ete testee sur les lignbes de cellules tumorales ci-apres, cultivees in vitro: lymphome murin TLX9 et saicome murin FS6, leucemie myelolde humaine K562 et lignee de cellules E humaines provenant d’un carcinome ovarien. Les preparations de macrophages etaient hCt6rogknes dans leurs interactions avec les cellules tumorales et les conditions d’essai, telles que le type de cellule cible, la durke de l’incubation et le rapport attaquants/cibles (A:T), avaient une influence determinante sur l’evaluation du potentiel cytotoxique des maciophages associes L la tumeur. A un rapport A:T de 7 : I , on n’a jamais observ6 d’activitk cytostatique sur les cellules TLX9 et K562, mais en presence de I’anticorps specifique, 8 des 12 preparations de macrophages testees ont fait preuve d’une nette cytotoxicit6 dependant de I’anticorps sur les cellules de lymphome TLX9. Les preparations de macrophages de deux patients inhibaient significativement la croissance du sarcome FS6 et I’on a observe une activite cytostatique sur les cellules E chez cinq autres patients. A un rappoit A:T de 7:1, on a constate une stimulation significative de la capacite prolifkiative d’au moins une des lignees de cellules cibles chez 1 I sujets. L’activite cytostatique des macrophages de 12 patients sur les cellules E a aussi et6 testte a un rapport A:T de 35: l ; dans ces conditions, 8 des 12 preparations ont fait preuve d’une cytotoxicite significative. Lorsque le mdme sujet Ctait teste B plusieurs reprises, a de courts intervalles, on observait le mdme type d’inhibition ou de stimulation de la croissance turnorale.

REFERENCES

BIREIECK, M. S. C., and CARTER, R. L., Observations on the ultrastructure of two hamster lymphomas with particular reference to infiltrating macrophages. Int. J. Cancer, 9, 249- 257 (1972). ECCLES, S. A,, and ALEXANDER, P., Macrophage content of tumours in relation to metastatic spread and host immune reaction. Nature (Lond.), 250, 667-669 (1974). EVANS, R., Macrophages in syngeneic animal tumours. Transplantation, 14, 468-473 (1972).

EVANS, R., Preparation of pure cultures of tumor macrophages. J. nut. Cancer Inst. , 50,271-273 (1973). EVANS, R., Tumor macrophages in host immunity to malig- nancies. In: M. A. Finck (ed.), The macrophage in neoplash, pp. 27-42, Academic Press, New York (1976). EVANS, R., and BOOTH, G., Inhibition of l*sIUdR incorpora- tion by supernatants from macrophage and lymphocyte cultures: a cautionary note. Cell Immunol., 26, 120-124 (1 976).


FAIPERMAN, I., GLEICHER, N., COHEN, C. J., and KOFPLER, D., Leucocyte migration in ovarian carcinoma: comparison of inhibitory activity of tumor extracts. J. nut. Cancer Insr., 59, 1593-1597 (1977). GAUCI, C. L., and ALEXANDER, P., The macrophage content of some human tumours. Cancer Lett., 1, 29-39 (1975).

HERBERMAN, R. B., and HOLDEN, H. T., Natural cell- mediated immunity. Advanc Cancer Res., in press (1978).

HEWLETT, G., OPITZ, H.-G., SCHLUMBERGER, H. D., and LEMKE, H., Growth regulation of a murine lymphoma line by a 2-mercaptoethanol or macrophage-activated serum factor. Europ. J. Immunol., 7, 781-785 (1977).

HOLDEN, H. T., HASKILL, J. S., KIRCHNER, H., and HERBER- MAN, R. B., Two functionally distinct anti-tumor effector cells isolated from primary murine sarcoma virus-induced tumors. J. Immunol., 117, 440-446 (1976).

HOLTERMANN, 0. A,, DJERASSI, I., LISAFELD, B. A., ELIAS, E. G., PAPERMASTER, 8. W., and KLEIN, E., In vitro destruc- tion of tumor cells by human monocytes. Proc. Soc. exp. Eiol. ( N . Y . ) , 147, 456-459 (1974).

HUBER, Ch., and WIOZELL, H., A simple rosette assay for demonstration of complement receptor sites using complement-coated zymosan beads. Europ. J. Immunol., 5, 432-435 (1975). KELLER, R., Cytostatic elimination of syngeneic rat tumor cells in virro by nonspecifically activated macrophages. J. exp. Med., 138, 625-644 (1973). KELLER, R., Susceptibility of normal and transformed cell lines to cytostatic and cytocidal effects exerted by macrophages. J. nat. Cancer Inst., 56, 369-374 (1976).

KOLLER, C. A., KING, G. W., HURTUBISE, P. E., SAGONE, A. L., and LOBUGLIO, A. F., Characterization of glass adherent human mononuclear cells. J. Immunol., 111, 1610- 1612 (1973). KRAHENBUHL, J. L., LAMBERT, L. H. JR., and REMINGTON, J. S., Effect of Corynebacterium parvum treatment and Toxoplasma gandii infection on macrophage-mediated cytoslasis of tumour target cells. Immunology, 31, 837-846 (1976). KURLAND, J. I., TRAGANOS, F., DARZYNKIEWICZ, Z., and MOORE, M. A. S., Macrophage-mediated cytostasis of neoplastic hemopoietic cells: cytofluorometric analysis of the reversible cell cycle block. Cefl. Immunoi., 36, 318-330 (1978).

LOZZIO, C. B., and LOZZIO, B. B., Cytotoxicity of a factor isolated from human spleen. J. nat. Cancer Insr., SO, 535-538 (1973). MANTOVANI, A., “ In vitro ” and “ i n vivo ” cytotoxicity of adriamycin and daunomycin for murine macrophages. Cancer Res., 37. 815-820 (1977). MANTOVANI, A., Effects on in vitro tumor growth of murine macrophages isolated from sarcoma lines differing in im- munogenicity and metastasizing capacity. Int. J. Cancer, in press (1978).

MANTOVANI, A., CAPRIOLI, V., GRITTI, P., and SPREAFICO, F., Human mature macrophages mediate antibody-dependent cellular cytotoxicity on tumor cells. Transplantation, 24, 291 - 293 (1977a). MANTOVANI, A,, EVANS, R., and ALEXANDER, P., Non- specific cytotoxicity of spleen cells in mice bearing trans- planted chemically induced fibrosarcomas. Erif. J . Cancer.

MANTOVANI, A., JERRELLS, T. R., DEAN, J. H., and HERBER- MAN, R. B., Cytolytic and cytostatic activity on tumor cells of circulating human monocytes. I n f . J . Cancer, 23, 18-27 (1979). MANTOVANI, A., TAGLIABUE, A., VECCHI, A,, and SPREAFICO, F., Effects of adriamycin and oaunomycin on spleen populations in normal and tumor allografted mice. Europ. J . Cancer,

NAMBA, Y., and HANAOKA, M., lmmunocytology of cultured IgM-forming cells of mouse. I. Requirement of phagocytic cell factor for- the growth of IgM-forming tumor cells in tissue culture. J . Immunol., 109, 1193-1200 (1972). NATHAN, C. F., and TERRY, W. D., Differential stimulation of murine lymphoma growth in vitro by normal and BCG- activated macrophages. J . exp. Med., 142, 887-902 (1975).

OLIVOTTO, M., and BOMPORD. R., ‘‘ In vitro ” inhibition of tumour growth and DNA synthesis by pzritoneal and lung macrop$ges from mice injected with Corynebacterium parvum . Inr. J. Cancer, 13, 478-488 (1974).

PREHN, R. T., Immunostimulation of the lymphodependent phase of neoplastic growth. J . nat. Cancer Inst., 59, 1043- 1049 (1977). PROSS, H. F., and KERBEL, R. S., An assessment ofintratumor phagocytic and surface marker-bearing cells in a series of autochthonous and early passaged chemically induced murine sarcomas. J . nut. Cancer Inst., 57, 1157-1167 (1976).

RUSSELL, S. W., and MCINTOSH, A. T., Macrophages isolated from regressing Moloney sarcomas are more cytotoxic than those recovered from progressing sarcomas. Nature (Lond.), 268, 69-71 (1977). TERRITO, M., and CLINE, M. J., Macrophages and their disorders in man. In: D. S. Helson (ed.), Immunobiulugy qf fhe macrophage, pp. 593-616, Academic Press, New York (1976). WOOD, G. W., and CiILLEsPIE, G. Y . , Studies on the role of macrophages in regulation of growth and metastasis of murine chemically induced fibrosarcomas. Int. J. Cnncer, 16,

WOOD, G . W., and GOLLAHON, K. A,, Detectionand quantita-

36, 35-40 (19776).

12, 381-387 (1976).

1022-1029 (1975).

tion of macrophage infiltration into primary human tumois with the use of cell-surface markers. J. nat. Cunrer Inst. , 59, 108 1-1087 (1 977). ZARLING, J. M., and TEVETHIA, S. S., Transplantation immunity to Simian virus 40-transformed cells in tumol- bearing mice. 11. Evidence for macrophage participation at the effector level of tumor rejection. J . nut. Cancer Insr., 50, 149-157 (1973).

effects on in vitro tumor growth of macrophages isolated from human ascitic ovarian tumors

Documents