Int. J . Cancer: 23, 344-352 (1979)

CHARACTERISTICS OF A LOW-MOLECULAR-WEIGHT FACTOR EXTRACTED FROM MOUSE TUMORS THAT AFFECTS I N VZTRO PROPERTIES OF MACROPHAGES Hou T. CHEUNG, Walter D. CANTAROW and G . SUNDHARADAS Immunohiology Research Center and the Department of Medical Microhiolo~y,University of Wisconsin, Madison, Wisconsin 53706, U S A

A dialysable low-molecular-weight factor capable of affecting in vitro properties of macrophages was extracted from four different mouse tumors. This factor not only modulates closely related properties of peritoneal macrophages such as spreading and migration but also inhibits lipopolysaccharideinduced tumoricidal activity of these cells. It can be extracted not only from tumor tissues but also from tumor cells grown in vitro. The appearance of this factor i s unique t o tumors and it is not present in detectable quantities in normal tissues. The factor from one of the tumors, Lewis lung carcinoma, was purified extensively and the partially purified factor retains all the above effects on macrophages. It i s not sensitive t o pronase o r a mixture of bovine spleen phosphodiesterase II, E. coli alkaline phosphatase and pancreatic ribonuclease. The factor i s lipid-like in character and i t is soluble in both organic solvents and aqueous media. It has ionizable group(s) and is anionic a t neutral p H but non-ionic under acidic conditions.

Both specific anti-tumor immunity and nonspecific anti-tumor resistance involving macrophages are part of the host defense mechanisms against the spread of tumor (Hellstrom and Hellstrom, 1969; Currie, 1976). Macrophages in the activated state have the capacity to distinguish tumor cells from normal cells and to mediate their cytotoxicity in a non-specific manner. The importance of this cytotoxicity as a mechanism of surveillance has been receiving increasing support (Currie, 1976; Hibbs, 1976). Recent findings suggest that tumors have the capacity to subvert host defense mechanisms and thus escape destruction. In fact, there appears to exist a n inverse correlation between the extent of inetastasis and the degree of infiltration of the tumor mass by inflammatory cells (Alexander et a/., 1976; Evans, 1973 ; Berg, 1959). Macrophages isolated from regressing Moloney sarcomas are more cytotoxic than those recovered from progressing sarcomas (Russell and Mclntosh, 1977). It has also been observed that macrophages derived from cancer patients exhibit a depressed chemotactic response in vitro whereas surgical removal of the neoplasm results in reversal to a normal response (Boetcher and Leonard, 1974; Snyderman et a / . , 1975). I n the mouse, malignant cells impair locally induced inflammation and suppress the accumulation of leukocytes at the inflammatory site (Fauve et a/., 1974; Snyderman er a/., 1976). Also, malignant cells have been found to repulse macrophages in vitro (Fauve ct a/., 1974).

All the above effects may be mediated by soluble factors released from the tumor cells. In fact, Fauve et a / . (1974) have made the observation that malignant mouse cells in culture produce into the medium a factor of molecular weight between 1,000 and 10,000 daltons with anti-inflammatory activity. Tumor-cell culture supernatants are also known to contain an activity capable of affecting the spreading and migration of macrophages (Fauve a n d Hevin, 1977; Rabatic e t a / . , 1977; Razvorotnev et a/.,1977). Snyderman a n d Pike (1976) have established that mouse neoplasms possess a factor of molecular weight between 6,000 and 10,000 daltons capable of inhibiting macrophage chemotaxis in vitro. North et a / . (1 976) obtained results suggesting that tumors produced a low-molecular-weight factor which suppressed macrophage-mediated resistance to Listeria infection. We now present results showing that a factor extracted from different mouse tumors not only affects closely related macrophage properties such as spreadyg and migration but also inhibits lipopolysaccharide-induced tumoricidal activity of macrophages. This factor is lipid-like in character and is soluble in both organic solvents and aqueous media. It is small in size with a molecular weight of less than a few thousand daltons. I t has ionizable group(s) and is anionic a t neutral PH but non-ionic under acid conditions. I t is not sensitive to pronase or a mixture of bovine spleen phosphodiesterase TI, E. coli alkaline phosphatase and pancreatic ribonuclease. This factor has been extensively purified and the partially purified factor retains all the above effects on macrophages. MATERIAL AND METHODS

Mouse tumors The following mouse tumors were used. mKS-A TU-5 (mKSA), an SV40-induced fibrosarcoma derived from BALB/c was a gift from Dr. Joyce Zarling of this research center. Meth IA, a niethylcholanthrene-induced fibrosarcoma derived from BALB/c a n d Lewis lung carcinoma (LLC), a spontaneous lung carcinoma derived from C57BL/IO, were obtained from the National Cancer Institute, To whom communications and reprint requests should be addressed. Received: November 24, 1978.

A TUMOR FACTOR THAT AFFECTS MACROPHAGES

Bethesda, Md. B16, a spontaneous melanoma derived from C57BL/10, was obtained from Jackson Laboratories, Bar Harbor, Me. RAW8, a lymphosarcoma derived from BALB/c, BW5147, a leukemia derived from AKR/J and EL4, a leukemia derived from C57BL/6, were obtained from the Salk Institute. P815, a mastocytoma derived from DBA/2, was a gift from Dr. Sugamura of this research center. mKSA, Meth 1A, LLC and B16 were maintained in the corresponding syngeneic hosts subcutaneously whereas the four other tumors were maintained in culture in Dulbecco's modified eagle medium (DMEM) containing 10% fetal calf serum (FCS). Culturing of tumor cell5 in vitro

In vivo grown tumors were minced and further disrupted by treatment with collagenase (Leffert and Paul, 1972) in phosphate-buffered saline (PBS). The cells were washed twice in PBS and then cultured in DMEM containing 10% FCS in Falcon tissue culture dishes to confluency in a humidified atmosphere of 5 % CO, in air. Preparation of dialysates of tissue extracts

Dialysates from in vivo grown tumor and normal tissues were prepared as follows. In all cases 1 g of tissue was homogenized in 5ml of PBS and the homogenate was sonicated twice for 30 sec each. I n the case of cultured cells, the cell pellet was suspended in 5 vol of PBS and sonicated. The homogenate was centrifuged at 2,OOOg for 10 min and the supernatant was further centrifuged at 100,OOOg for 1 h. The supernatant was dialysed against 2 vol of PBS for 48 h at 5" C using a dialysis membrane that allows the passage of molecules of less than 3,500 (approximately) daltons only. The dialysates were sterilized by filtration and stored at 5" C or -20" C. Peritoneal exudate cells and peritoneal macrophages

Peritoneal exudate cells were obtained from male C57BL/6 mice, 6 to 10 weeks of age, which had been injected with 0.2 ml of complete Freund's adjuvant (CFA) intraperitoneally 36 h previously. The mice were killed by cervical dislocation and the peritoneal cells were collected by lavaging the peritoneal cavity with 8 ml of PBS. The cells were washed once with PBS and suspended in DMEM containing 10% FCS at a density of 4 x lo7 cells per ml. The peritoneal exudate cells contained approximately 51 % macrophages, 7 % lymphocytes and 42 % polymorphonuclear leukocytes as determined by differential counting after staining with hemotoxylin. Peritoneal exudate cells containing approximately 91 % macrophages were obtained by injecting mice intraperitoneally with 2 ml of Brewer's thioglycollate medium (DIFCO) 6 days previously. Purified macrophages (,99%) were obtained by the fractionation of the thioglycolfate-induced peritoneal exudate cells (Sembala and Asherson, 1970). Assay for spreading of macrophages

Peritoneal exudate cells, 5.2 i( loe in 2 nil DMEM containing 10% FCS, were incubated for 2 h in a 3 5 ~ 1 0 m mFalcon tissue culture dish. I n some experiments tissue culture plates (FB-16-34 TC,

345

Linbro Chemical Co.) were used in place of tissue culture dishes and in such cases 1.5 x lo6 peritoneal exudate cells were used per well. Non-adherent cells were removed and the adherent monolayer was washed once with 1.0 ml PBS. The non-adherent fraction contained on an average 2.5 x lo6 cells. The macrophage monolayers (>98 % macrophages) thus obtained were incubated at 37" C in medium containing 10% FCS in a humidified atmosphere of 5 % CO, in air for 24 h by which time the cells had begun t o spread. The non-adherent cells were removed and the adherent cells incubated with medium containing 10% FCS and various dialysates for 24 h. At this time the cells were examined. The spread cells with elongated processes and the round cells with minimal t o no elongated processes were scored. Assay for the migration of peritoneal exudate cells

The migration assay was carried out as described by Gorczynski et al. (1 973). Approximately 2 X 10' cells in 0.05 ml were drawn into sterile 15 x 100-mm glass capillary tubes and sealed at one end with a 60:40 mixture of paraffin and vaseline. The capillaries were centrifuged (200g for 6 min) and cut at the cell-fluid interface. Two stubs were placed in each well of a leukocyte migration plate (Catalogue No. 132, York Scientific Inc., Ogdensburg, N. Y. 13669) and secured with vaseline. The wells were filled with approximately 0.6 ml of DMEM containing 10% FCS and the tumor or normal tissue dialysale. The wells were covered with sterile 22 x22-mm cover slips and the plates were incubated at 37" C in a humidified atmosphere of 5 % CO, in air. The migration areas were projected onto a paper and measured with a planimeter. Results were expressed as arbitrary units of migration area. Activation of macrophages and their tumoricidal activity CFA-induced peritoneal exudate cells, 1.5 x loR, in 0.5 ml of DMEM containing 10% FCS were incubated in wells of tissue culture plates (FB-16-34 TC, Linbro Chemical Co.) for 24 h and the nonadherent cells were removed. The adherent cells (>98 % macrophages) were activated by culturing with 1 nil of medium containing 10% FCS and 20 pg per ml lipopolysaccharide (LPS) (E. coli 0127 :B8, Difco Laboratories) for 48 h. The various dialysates, whose effects were to be tested, were added along with the LPS. To the activated macrophage monolayer 5 x lo4 chromate (51Cr)-labelled tumor target cells in 0.1 ml were added and incubated for 12-24 h depending upon the target cells used. A constant aliquot (0.1-0.4 ml) of the supernatant fluid was then aspirated from each well and the amount of 51Crradioactivity released into the supernatant was determined. The maximum release of radioactivity was determined by lysing 5 x lo4 labelled cells with 0.25 % Triton X-100. Spontaneous release of radioactivity was determined by incubating 5 x lo4labelled cells alone. All incubations were at 37" C in a humidified atmosphere of 5 % CO, in air. In order to label the tumor target cells, lo7 cells were incubated in 0.5 ml medium containing 10%FCS'and 250 pCi of [51Cr]sodi~mchromate for 1 h, washed three

346

CHEUNG ET AL.

times with PBS and suspended in medium containing 10 FCS. The percentage cytotoxicity was calculated

as cpm experimental release -cpm spontaneous release x 100. cpm maximum release - cpni spontaneous release Study of the effect of pronase on the tumor factor Pronase (Calbiochem) 5 mg was conjugated to 1 ml of Sepharose 6B (Pharmacia Fine Chemicals) as follows. Sepharose 6B was activated with cyanogen bromide (50mg/ml), washed on a sintered glass filter with water followed by 0.1 M NaHCO8. The activated Sepharose and the pronase were mixed in a small volume and incubated overnight at 4" C. The mixture was treated with 50 mM ethanolamine for 15 min and the conjugated Sepharose was washed extensively with PBS. The conjugate retained the activity of pronase as judged by its action on casein

yellow substrate. The LLC tumor dialysate, 0.5 ml, was incubated with an equal volume pronasesepharose for 1 h at 37" C. The mixture was centrifuged and the supernatant was sterilized and assayed for tumor factor activity. In a control experiment the pronase-sepharose was incubated with PBS and the supernatant obtained did not show any activity similar to that of the tumor factor. Also, the activity of the bound enzyme was not inhibited by the tumor factor preparation when tested on casein yellow substrate. Study of the effect of phosphodies:;rase, alkaline phosphatase and ribonuclease on the tumor factor The LLC tumor dialysate, 0.5 ml, was incubated with a mixture of 0.9 units of bovine spleen phosphodiesterase 11 (Sigma Chemical Co.), 5Opg E. coli alkaline phosphatase (Sigma Chemical Co.) and 0.1 mg pancreatic ribonuclease (Calbiochem) or PBS (control) at 37" C for 1 h. This was dialysed against an equal volume of PBS and the dialysate was tested for the tumor factor activity. When the enzyme was incubated alone (control), the dialysate did not show any activity similar to that of the tumor factor. Also, the ability of the enzyme mixture to degrade yeast RNA was not inhibited by the tumor dialysate.

- -++0

L-- I

I

I

5

Concentration

(Yo)

10

20

FIGURE2 - Effect of LLC dialysate on the spreading of peritoneal macrophages. Concentration of dialysate is shown as a percentage, vjv. Data shown are from one typical experiment.

RESULTS

FIGURE1 - Effect of tumor dialysate on peritoneal macrophage spreading. Macrophage monolayers were treated with PBS 25%, v/v (a) or LLC dialysate, 25%, v/v (6).After 24 h of treatment the cells were fixed, stained with hemotoxylin and photographed.

Eflect of a tumor factor on macrophage spreading Homogenates Of four different LLC, mKSA* Meth l A and B16, were dialysed against PBS and the dialysates were tested for their effect on the spreading of macrophages. Liver, kidney, heart

A

IUMUK

- 1

^---

tACIUK

--_.1mAl AFFECTS MACROPHAGES

TABLE I EFFECT OF LLC DIALYSATE ON THE RELEASE OF LACTATE DEHYDROGENASE BY PERITONEAL MACROPHAGES IN CULTURE Addition

PBS, 25 %

LLC dialysate, 25 %

LDH in the cells

LDH in the supernatant

%LDH in the supernatant

3.62 3.29

0.41 0.38

11.4 10.3

Monolayers of peritoneal macrophages were prepared as described under “Methods” and cultured for 2 4 h with and without LLC dialysate. The supernatants were removed and assayed for lactate dehydrogenase (LDH) activity. The cells in each monolayer were lysed with 0.1% Triton X-100 and the lysates were assayed for LDH activity. LDH activity was determined as described in Worthington Enzyme Manual, Worthington Biochemical Corporation. LDH activity is expressed in units; one unit of enzyme is that amount which, when present in 1 ml, gives a change in A,,. of I .O in 2 min under the assay conditions.

and embryonic fibroblasts cultured to confluency were used as normal tissue controls. Monolayers of mouse peritoneal macrophages (>98 % macrophages) induced with CFA were cultured for 24 h, at which time they had begun to spread. The adherent spread macrophages were treated with the different dialysates and incubated for an additional 16-24 h. The results obtained with dialysates derived from LLC tumor and PBS control are shown in Figure 1. As shown in Figure 1b the LLC dialysate causes the cells to round up. Similar results were also observed with dialysates derived from the three other tumors. In contrast, in the presence of dialysates derived from the control tissues, the cells had a spread morphology similar to that of the control cells (Fig. la). Thus the dialysates derived from tumors but not normal tissues contain a factor capable of affecting macrophages and reversing their spreading; in the presence of this factor the spread macrophages revert to a round morphology. The reversal of macrophage spreading by tumor factor is concentration-dependent as shown in Figure 2. The cells treated with the different dialysates and control cells were examined for viability by the eosin exclusion test and the percentage of viable cells was found to be very similar (>95 %) in all cases. Furthermore, if the cells were washed to remove the tumor dialysate and further cultured overnight in the absence of tumor dialysate, they reverted from the round to a spread morphology. Thus the effect of the tumor dialysate is reversible. In order to test whether the tumor dialysate had any sublethal effect on the macrophages, these cells were cultured in the presence and absence of the tumor dialysate and their release of the cytoplasmic enzyme lactic dehydrogenase into the medium was studied. As shown in Table I, the tumor dialysate did not cause any increase in the release of this enzyme. The above experiments were repeated many (>25) times and the results were reproducible though the effect of the tumor dialysates on spreading varied slightly between preparations of peritoneal macrophages. Also, these experiments were repeated by using thioglycollate-induced peritoneal macrophages and very similar results were obtained.

341

Enhancement of peritoneal cell migration caused by the tumor factor Since spreading of cells and their movement are related properties (Allison, 1973; Cheung et al., 1978) the effects of the dialysates derived from the tumor tissues and the control normal tissues on the niigration of CFA-induced peritoneal cells out of capillary tubes were also examined. Figure 3 shows the effects of various concentrations of the dialysates derived from LLC and liver. The LLC dialysate enhanced the migration (expressed in migration area) of peritoneal cells; at the highest concentration (25 %) tested, the enhancement in migration area is approximately 300 %. This result is reproducible but the migration enhancement varied among peritoneal exudate cell preparations between 200 % and 300 %. Similar results were also obtained with dialysates derived from mKSA, B16 and Meth 1A tumors. In contrast, the dialysate derived from liver (Fig. 3) had no measurable effect on the migration as compared to the medium control. Dialysates derived from the

+ + 0 1

I ~

5

10

I

20

Concent rat ion ( % I

FIGURE 3 - Effect of tumor dialysate on the migration of CFA-induced peritoneal exudate cells. Migration areas in the presence of dialysates derived from the tumor, LLC ( 0 ) and liver (0)are shown. Concentration of dialysate is expressed as a percentage, v/v. Each point is the mean of four determinations+sD.

other normal tissues, kidney, heart and embryonic fibroblasts grown to confluency also had no significant effect on the migration of peritoneal cells. The CFA-induced peritoneal exudate cells used in the migration assay were largely macrophages and polymorphonuclear leukocytes with a few lymphocytes. To establish the nature of the cells which were affected in the migration assay, the cells that migrated out of the capillary were fixed, stained with Wright’s stain and examined. The cells in the periphery of the migration area and those identifiable cells in the more

348

CHEUNG ET AL. TABLE 11 EPI-ECTS OF TUMOR A N D LIVER IIIALYSATES ON T H E LPS-INDUCED TUMORICIDAL ACTIVITY OF MACROPHAGES



Addition

Percentage cytotoxicity

PBS, 25 % LLC dialysate, 25 % rnKSA dialysate, 25 Meth IA dialysate, 25 ”/, B I6 dialysate, 25 % Liver dialysate, 25 ~~~~

~~

~

35.01 1.6 -4.1 &5.3 1.8*2.2 11.5313.5

4.0 $6.2 43.0 t I .6

~~~

’ EL4 target cells were used. Concentrations are in percentage, v/v.

The data shown are the means of triplicatesfso.

crowded interior areas were mostly ( > 95 %) niacrophages as judged by the morphology of their nuclei and the nuc1eus:cytoplasm ratios. In addition, the tumor factor enhanced the migration of thioglycollate-induced peritoneal exudate cells containing approximately 90 % macrophages. Also, the migration assay was repeated using purified niacrophages. Macrophages from the thioglycollate-induced peritoneal exudate cells were purified as described under ‘‘ Methods” and the migration of these purified macrophages (>99 % macrophages as judged by hemotoxylin staining and the non-specific esterase content) was also enhanced by the tumor dialysate. This demonstrates the ability of the tumor factor to act directly on macrophages and affect their migration. The cells in the migration chambers were examined for viability by the eosin exclusion test and the percentage of viable cells was found to be very similar (:.-90%) in both the presence and absence of the various dialysates. The enhanced migration of peritoneal cells observed in the presence of tumor dialysates is sensitive to sodium azide. Sodium azide a t 1 mM inhibited the migration enhancement by approximately 50% without affecting viability of the cells. Ejfect of the macrophage-modulating tumor factor on the tumoricidal activity of macrophages induced by lipopolysaccaride Since the tumor factor affected macrophages and modulated their properties, its effect o n LPS-induced tumoricidal activity of macrophages was also tested, Peritoneal macrophages could be activated by treatment with LPS for 48 h and rendered cytotoxic for syngeneic EL4 tumor cells (Table 11). The effects of the dialysates derived from all four tumors, LLC, mKSA, Meth 1A and B16 and normal liver (control) were tested by adding the dialysates a t the time of addition of LPS. As shown in Table 11, alf the tumor dialysates significantly inhibited the cytotoxicity. However, no inhibition was observed with the liver dialysate. Similar results were also obtained when the tumor-cell lines, mKSA, P815, RAW8 and BW5147, which are allogeneic to the macrophages were used as targets (data not shown). In these experiments, the tumor dialysate was present during the entire period of activation of the macrophages and during

the cytotoxicity phase. But, inhibition of cytotoxicity by the tumor dialysate was observed even if it was added 24 h after the addition of LPS. Furtherniorc, inhibition of cytotoxicity by the tumor dialysatc is concentration-dependent, as shown in Figure 4. Presence of macrophage-modulating factor in crrltiirecl tumor cells The above results show that the dialysates derived from tumor tissues possess a factor capable of modulating in vitro properties of macrophages. But these results d o not show whether this factor is a product of malignant cells o r whether its occurrence in tumor tissue is due to events such as proteolysis taking place in and around the tumor tissue. In order to differentiate between these possibilities, the Lewis lung carcinoma and mKSA tumor tissues were disrupted by treatment with collagenase and the disrupted cells were cultured in vitro. After five passages the cells were harvested, washed, suspended in PBS, sonicated, and dialysates of the homogenates were prepared. The dialysates derived from both Lewis lung carcinoma and inKSA cells reversed the spreading of peritoneal macrophages and enhanced their migration (Table 111) suggesting that the macrophage-modulating factor is a product of malignant cells. Some mouse tumors are known to contain lactate dehydrogenase (LDH) elevating virus (Riley, 1968) and it was important to test whether the presence of the tumor factor in the tumor cells was related to the presence of this virus. Since this virus is known to disappear from cells cultured in vitro, the LLC and mKSA tumor cells were cultured in vitru for five passages and tested for the presence of LDH elevat-

c

C 0,

2 10

a“

i

0’1

t

5

10

20

Concentration ( %) FIGURE 4 -Effect of LLC dialysate on the LPS-induced tumoricidal activity of macrophages EL4 tumor cells were used as targets. The LLC dialysate was added at the time of addition of LPS and the concentration of the dialysate is expressed as a percentage, v/v. The data shown are the means of triplicates t s D .

A TUMOR FACTOR THAT AFFECTS MACROPHAGES

TABLE I l l EFFECT OF DlALYSATES FROM CULTURED TUMOR CELLS ON THE MIGRATION OF PERITONEAL CELLS ' Migration a r e a i s u in the presence of Source of dialysate

LLC cells mKSA cells

PBS (control)

5.41 1.7 3.9+1.8

Dialysate

14.9*2.7 11.0&2.4

LLC and mKSA cells were cultured in vifro for five passages and dialysates $ their homogenates were prepared as described under Methods The effect of these dialysates on the migration of CFAinduced peritoneal cells were tested at a concentration of 25%. vjv. The data shown are means of triplicatesfso.

.

'I

ing virus as well as the tumor factor. Testing for the virus was carried out for us by Dr. Vernon Riley (Pacific Northwest Foundation, Seattle, Wash.) by injecting 2.5 x lo5 tumor cells intraperitoneally into female BDF mice and following their serum LDH levels (Riley, 1968). Both LLC and mKSA tumor cells were found to be free of the virus but the same cells contained the tumor factor (Table 111). Thus the appearance of the tumor factor is not due to any presence of LDH elevating virus but is a property of the tumor cells. Characteristics of the tumor factor

The activities affecting the different properties of macrophages described above dialyzed out of a dialysis membrane which allows the passage of molecules of less than 3,500 daltons. The factors extracted from all four tumors behaved similarly in this regard. In order to gain further insight into the nature of the tumor factor the following additional studies were carried out with the LLC factor. The LLC tumor factor was not affected by heating at loo" C for 30 min or by exposure at PH 2 or 11 for 30min. It was not inactivated to any measurable

349

degree upon treatment with pronase or a mixture of bovine spleen phosphodiesterase 11, E. coli alkaline phosphatase and pancreatic ribonuclease (see " Methods "). It has ionizable group(s) and is negatively charged at PH 7.4 but not at PH 2.4. Thus it binds to anion exchange resin (AGI-X2, BioRad) in the presence of ten-fold diluted PBS at PH 7.4 and the bound factor is eluted by 10-times diluted PBS at PH 2.4. It does not bind to cellulose phosphate at a PH of 1.5 or 3.0, suggesting that it does not have a positive charge under acid conditions. The solubility of the tumor factor in organic solvents was also investigated. It could not be extracted from PBS (PH 7.4) with chloroform whereas it could be readily extracted with this solvent after the PH was lowered to 1.5 by the addition of hydrochloric acid. In addition, after a solution of the factor in PBS was lyophilized, it could be extracted from the residue with chloroform though the extraction was not rapid. These extracts contained the activity capable of affecting all the different in vitro properties of macrophages. These results indicate that the tumor factor is lipid-like with one or more anionic groups. Purtiul purification of the furnorfactor

Partial purification of the tumor factor was carried out by a combination of chloroform extraction and molecular sieving. Absorbance at 220 nm was used for monitoring the fractions during purification. A volume of 700ml of LLC dialysate (dialyzed out of a membrane that allows the passage of molecules of less than 3,500 daltons) was lyophilized and the residue was extracted with chloroform. The chloroform extract was evaporated and the resulting residue extracted with 6 ml of PBS. This gave a 14-fold purification from the dialysate based units. The extract was then fractionated by on molecular sieving using Sephadex G-25. As shown in Figure 5, the activity that inhibits macrophage spreading and the activity that causes the enhance-

0.E

-

:

R

15

I

0.6

0

6

N L

-I0

a 0)

0.4

-P c

0 5?

4

H

0

$ 0.2t

Fraction number

FIGURE5 - Molecular sieving using Sephadex (3-25. Column (2.6 x 89 cm) was eluted with PBS at a flow rate of 10 ml/h. Fractions of 3 ml were collected. Macrophage spreading inhibiting activity and macrophage migration enhancing activity after 50-fold dilution of the fractions are shown. Blue dextran peak was at fraction 66. Peaks of prostaglandins E,, Ez and B, and [3H,]0 were as indicated.

350

CHEUNG ET AL.

ment of macrophage migration appear together in fractions 135-150, the peak of activity being around fraction 142. The activity inhibiting macrophagemediated cytotoxicity also appears in the same range of fractions with the peak activity around fraction 142. Prostaglandins El, E, and B1 were used as markers in molecular sieving using the same column. Peaks of both PGE, and PGEz appeared at fraction 133 whereas the peak of PGB, appeared at fraction 147. Molecular sieving resulted in a purification of approximately 10 fold giving a total purification of approximately 140 fold from the dialysate. The absorption spectrum of the material present in fraction 142, which has the peak activity, is shown in Figure 6. The absorption at 260 or 280 nm is very small compared to the absorption at 220nm. This is also true for the other fractions containing the activity. DISCUSSION

The results presented here show that all the four mouse tumors tested possess a low molecular weight factor which affects peritoneal macrophages in vitro, resulting in the modulation of their properties. This factor reverses the spreading of macrophages without affecting their viability. The tumor factor also causes an enhancement of macrophage migration; this enhancement is sensitive to sodium azide and thus energy-dependent. This effect on migration is observed not only with unfractionated peritoneal exudate cells but also with macrophages purified from the peritoneal exudate cells. Our results (Table 11, Fig. 4) also show that the tumor factor has the capacity to inhibit the LPSinduced tumoricidal activity of macrophages. It is not clear whether the tumor factor causes this by interfering with the activation of macrophages or by inhibiting the cytotoxic activity of activated macrophages. Since activated macrophages are thought to play an important role in surveillance against the spread of tumor, this inhibition may be relevant to the tumor-mediated subversion of macrophage function in vivo. It is not known whether the modulation of the different properties of macrophages and the inhibition of LPS-induced macrophage-mediated cytotoxicity are brought about by the same molecule or by different molecules. But the finding that the partially purified factor exhibits all the effects is consistent with the notion that all these effects may be brought about by a single factor. The macrophage-modulating tumor factor has a molecular weight very likely less than a few thousand daltons as it dialyzes through a dialysis membrane which allows the passage of molecules of less than 3,500 daltons. The chemical nature of this factor is not known. However, it appears to be lipid-like as it is soluble in organic solvents. It can be extracted from aqueous medium by chloroform at low PH (1.5) but not at high PH (7.4). It has a negative charge at PH 7.4 but not at PH 2.4, suggesting that it has ionizable group@) such as carboxyl. It is not sensitive to pronase or a mixture of RNA-degrading enzymes nor is its activity lost upon boiling or

incubation in acid or basic solution. In molecular sieving by Sephadex (3-25 (Fig. 5) it is eluted in a region where prostaglandins are eluted. Prostaglandins are also lipids with an anionic group (carboxyl) that can be extracted by organic solvents from aqueous medium at low PH but not at high PH. Tumor cells are known to produce more prostaglandins of the E type than normal cells, the predominant one being PGEz (Jaffe, 1974; Plescia et ul., 1976). However, as shown in Figure 6, the peak of macrophage-modulating activity in molecular sieving does not coincide with the peak of either PGEL,PGE, or PGB,. In addition, PGA2, PCBz and PGF,, appear to differ from the tumor factor in their effects on the spreading, adhesion and migration of macrophages (Cantarow et al., 1978). Thus it can be concluded that the macrophage-modulating tumor factor is different from PGAz, PGB,, PGB,, PGE1, PGE,, and PGF,, and its chemical nature remains to be established.

a, 0

c

0

-? u) 0

22

U

I

I

I

I

220

260

300

340

Wavelength, nrn

FIGURE6 - Absorption spectrum of the material present in fraction 142 of molecular sieving described in Figure 5 .

Tumor tissues are known to contain host cells (Evans, 1972; Eccles and Alexander, 1974) and also degradative events might be occurring in the vicinity of tumors. Hence it could be argued that the macrophage-modulating factor is not a product of tumor cells. But, the fact that this factor could be extracted from mKSA cells and LLC cells after five passages in vitro indicates that the factor is produced by the tumor cells themselves. In order that the tumor factor may affect macrophages in vivo it should be produced extracellularly and appear in the circulation. In vitro studies show that the factor is in fact secreted into the culture medium by the tumor cells (unpublished observation). Some mouse tumors are known to contain LDHelevating virus (Riley, 1968) and it can be argued that the appearance of the tumor factor in the tumor cells is related to the presence of this virus. But this

A TUMOR FACTOR THAT AFFECTS MACROPHAGES

possibility is ruled out by the finding that LLC and mKSA tumor cells cultured in vitro for five passages do not contain the virus but contain normal levels of the tumor factor. The appearance of the macrophage-modulating factor may be unique to tumor cells as it is present in all tumors examined but not in normal tissues such as heart, kidney and liver or in embryonic fibroblasts grown to confluency. As described earlier, tumor cells have been shown to produce a lowmolecular-weight factor with anti-inflammatory activity (Fauve et ul., 1974). Tumor-cell culture supernatants are also known to contain an activity capable of affecting the spreading and migration of macrophages (Fauve and Hevin, 1977; RabatiC et al., 1977; Razvorotnev et al., 1977). Low-molecularweight factors capable of inhibiting macrophage chemotaxis in vitro (Snyderman and Pike, 1976) and suppressing macrophage-mediated resistance to Listeriu infection (North et al., 1976) are also known to be produced by tumors. It is not known at this

351

time whether any or all of these factors have identity with the macrophage-modulating tumor factor we have described. ACKNOWLEDGEMENTS

We thank Dr. Vernon Riley, Pacific Northwest Foundation, Seattle, Washington for testing the tumor cells for the presence of LDH virus. We also thank Prof. Fritz H. Bach for his support and Ms. C. Smith for assistance with the preparation of the manuscript. This is paper number 155 from the Immunobiology Research Center. This work was aided by grants DRG1283 from the Damon Runyon Memorial Fund for Cancer Research, Inc. and a grant from the University of Wisconsin Graduate School. H.T.C. is a postdoctoral trainee supported by training grant T32CA09106 from the National Cancer Institute, DHEW. W.D.C. was supported by postdoctoral fellowship F32CA05596 from the National Cancer Institute, DHEW.

CARACTERISTIQUES D’ UN FACTEUR D E FAIBLE POIDS MOLECULAIRE EXTRAIT D E TUMEU R S MURINES QUI I NFLUE SUR LES PROPRIETES I N VITRO DES MACROPHAGES Un facteur dialysable, de faible poids mokculaire, capable d’influer sur les proprietes in vitro des macrophages, a extrait de quatre turn:urs murines differentes. Ce facteur module les proprietes etroitement apparentees des macrophages peiitoneaux telles que I’etalement et la migration; de plus, il inhibe I’activite tumoricide de ces cellules induite par le lipopolyssccharide. I1 peut &re extrait non seulement de tissus tumoraux, mais aussi de cellules tumorales cultiv6es in vitro. I1 ne semble present qu: d m 3 les tumsurs; les tissu; normaux n’en contiennent pas de quantites decelables. Le facteur provenant de l’une dzj tumzurs, le carcinome pulmmaire de Lewis, a BtB purifie dans une large mesure; le facteur partiellement purifie conserve tJus les effits dicrits ci-dessus sur les macrophages. 11 n’est pas sensible a la pronase, ni A un melange de phosphodiesterase spleniqu: bovine 11, de phosphatase alcaline d’E. coli et de ribonuclkase pancrkatique. 11 est de type lipidique et est soluble dans des sdvants organiques aussi bien que dans un milieu aqueux. 11 comporte un o u plusieurs groupes ionisables, et est anionique a PH neutre mais non ionique en milieu acide.

REFERENCES

ALEXA NDE R,ECCLES, ~., S. A., and GAUCI,C. L. L.,The significance of macrophages in human and experimental tumors. Ann. N . Y. Acad. Sci., 276, 124-133 (1976). ALLISON, A. C., The role of microfilaments and microtubules in cell movement, endocytosis, and exocytosis. In: Locomotion of tissue cells, Ciba Foundation Symposium, Vol. 14, pp.109148, Associated Scientific Publishers, Amsterdam (1973). BERG,J. W., Inflammation and prognosis in breast cancer. A search for host resistance. Cancer, 12,714-720 (1959). BOETCHER,D., and LEONARD, E. J., Abnormal monocyte chemotactic response in cancer patients. J. nat. Cancer Inst., 52, 1091-1099 (1974). CANTAROW, W. D., CHEUNG, H. T., and SUNDHARADAS, G., Effects of prostaglandins on the spreading, adhesion and migration of mouse peritoneal macrophages. Prostaglandins, 16, 39-46 (1978). CHEUNG, H. T., CANTAROW, W. D., and SUNDHARADAS, G., Colchicine and cytochalasin B (CB) effects on random movement, spreading and adhesion of mouse macrophages. Exp. Cell Res., 111,95-103 (1978). CURRIE,G., Immunological aspects of host resistance to the development and growth of cancer. Biochem. Biophys. Acta, 458, 135-165 (1976). ECCLES,S. A., and ALEXANDER, P., Macrophage content of tumors in relation to metastatic spread and host immune reaction. Nature (Lond.), 250,667-669 (1974). EVANS, R., Macrophages in syngeneic animal tumors. Transplantation, 14, 468-473 (1972).

EVANS,R.,Macrophages and the tumor bearing host. Brit. J. Cancer, 28, Suppl. 1, 19-25 (1973). FAUVE,R. M., and HEWN,M. B., Inflammation and host resistance against tumors. 11. Antagonism between bradykinin and a fraction isolated from the supernatant of cultured malignant cells on the spreading of macrophages. Ann. Immunol. (Inst. Pasteur), 128C, 1079-1083 (1977).

J. A., and FAUVE,R. M., HCVIN,B., JACOB,H., GAILLARD, JACOB, F., Antiinflammatoryeffects of murine malignant cells. Proc. nat. Acad. Sci. (Wash.), 71,4052-4056 (1974). GORCZYNSKI, R . M., MILLER,R. G., and PHILLIPS,R. A,, Reconstitution of T cell depleted spleen cell populations by factors derived from T cells. I. Conditions for the production ofactive Tcellsupernatants. J. Immunol., 110,968-983 (1973). K. E., and HELLSTROM, I., Cellular immunity HELLSTROM, against tumor antigens. Advanc, Cancer Res., 12, 167-223 (1969). HIBBS,H. B., JR., Role of activated macrophages in nonspecific resistance t o neoplasia. J. reticuloendothel. Soc., 20, 223-231 (1976). JAPFE,B. M., Prostaglandins and cancer: an update. Prostaglandins, 6, 453-461 (1974). LEFFERT, H. L., and PAUL,D., Studies on primary cultures of differentiated fetal liver cells. J . Cell Biol., 52, 559-568 (1972). D. P., and TUTTLE, R. L., Subversion NORTH,R. J., KIRSTEIN, of host defense mechanisms by murine tumors. I. A circulat-

352

CHEUNG ET AL.

ing factor that suppresses macrophage mediated resistance to infection. J . exp. Med., 143,559-573 (1976). PLESCIA, 0.J., GRINWICH, K., and PLESCIA, A. M., Subversive activity of syngeneic tumor cells as an escape mechanism from immune surveillance and the role of prostaglandins. Ann. N . Y . Acad. Sci., 216, 455-465 (1976). RABATI~., S., JURIN,M., and DEKARIS, D., In virro suppression of macrophage spreading caused by supernatants of tumour, thymus and lymph node cells. Folia Biologica (Prahaj, 23, 317-326 (1977). RAZVOROTNEV, V. A., LAVROVSKY, V.A., and VIKSLEV, V. Kh., Secretion by tumor and nontumor cells of substances modifying the macrophage migration and the cytopathogenic action of immune lymphocytes in vitro. Zh. Microbiol. Epidimiol. Microbiol., 6,48-51 (1977). RILEY,V.. Lactate dehydrogenase in the normal and malignant state in mice and the influence of a benign enzymeelevating viius. Merh. Cancer Res., 4, 493-618 (1968).

RUSSELL, S. W., and MCINTOSH, A . T., Macrophages isolated from regressing Moloney sarcomas al e more cytotoxic than those recovered from progressing sarcomas. Nature (Lond.), 268, 69-71 (1977). SEMBALA, M., and ASHERSON, G. L., The rapid purification of peritoneal exudate macrophages by Ficoll (polysucrose) density gradient centrifugation. Immunology, 19, 677-681 (1970). SNYDERMAN, R., and PIKE,M. C., An inhibitor of macrophage chemotuxis produced by neoplasms. Science, 192, 370-372 (1976). SNYDERMAN, P., PIKE,M. C., BLAYLOCK, B. L., and WEINSTEIN, P., Effects of neoplasms o n inflammation: depression of macrophage accumulation after tumor implantation. J. Immunol., 116, 585-589 (1976). SNYDERMAN, R., PIKE,M. C., MEADOWS, L., HEMSTRPET, G., and WELLS,S., Depression of monocyte chemotaxis by neoplasms. Clin. Res., 23,297a (1975).

Characteristics of a low-molecular-weight factor extracted from mouse tumors that affects in vitro properties of macrophages.

Int. J . Cancer: 23, 344-352 (1979) CHARACTERISTICS OF A LOW-MOLECULAR-WEIGHT FACTOR EXTRACTED FROM MOUSE TUMORS THAT AFFECTS I N VZTRO PROPERTIES OF...
832KB Sizes 0 Downloads 0 Views