Generation of macrophage migration inhibitory activity by ...

Proc. Natl. Acad. Sci. USA Vol. 74, No. 4, pp. 1570-1574, April 1977

Cell Biology

Generation of macrophage migration inhibitory activity by plasminogen activators (simian virus 40-transformed 3T3 cells/urokinase/dense cell-surface coat/plasmin)

RICHARD 0. ROBLIN*, M. ELIZABETH HAMMOND, NORMAN D. BENSKY, ANN M. DVORAK, HAROLD F. DVORAK, AND PAUL H. BLACK Infectious Disease Unit and Department of Pathology, Massachusetts General Hospital and Departments of Microbiology and Molecular Genetics, Pathology, and Medicine, Harvard Medical School, Boston, Massachusetts 02114

Communicated by Jerome Gross, October 18, 1976

extended by Poste (5). He also demonstrated that the SV3T3 MIF-like activity was inhibited by diisopropylfluorophosphate (Dip-F), and suggested that the active molecule might be a serine protease (5). After comparing the properties of the SV3T3 MIF-like substance and plasminogen activator, a serine protease known to be released by SV3T3 cells (6, 7) and many human tumor cell lines (8, 9), Poste concluded that they were probably different proteins. In this paper, we show the partially purified preparations of plasminogen activator from SV3T3 cells, as well as a known plasminogen activator, human urokinase (UK), can generate MIF-like activity when incubated with medium containing 15% guinea pig serum (GPS). We suggest that the action of plasminogen activator on a serum component(s) generates the molecule(s) which inhibits the migration of GPPE cells.

ABSTRACT Medium from cultures of simian virus 40transformed mouse 3T3 cells (SV3T3) inhibits the migration in vitro of peritoneal exudate cells (macrophages) from guinea pigs while medium from untransformed 3T3 cultures does not [Hammond, M. E., Robbin, R. D., Dvorak, A. M., Selvaggio, S. S., Black, P. H. & Dvorak, H. F. (1974) Science 185,955-9571. The present paper describes the generation of migration inhibitory factor (MIF)like activity for peritoneal exudate cells from guinea pigs after incubation of a serum-free harvest fluid from SV3T3 cells with guinea pig serum. Inhibited macrophages lose a densely staining material from the cell surface coat compared with uninhibited guinea pig peritoneal exudate cells. The factor in SV3T3 harvest fluids which generates the migration inhibitory activity appears to be plasminogen activator, i.e., a serine protease, because (i) plasminogen activator activity and the factor which generates MIF-like activity copurify, and cochromatograph on Sephadex G-200 columns, and (ii) plasminogen activator activity and capacity to generate MIF-like activity are simultaneously lost upon treatment with [3H]diisopropylfluorophosphate. In addition, a purified preparation of a known plasminogen activator, human urokinase, can also generate MW-like activity upon reaction with guinea pig serum. Because transformation of 3T3 cells by SV40 increases their plasminogen activator secretion, enhanced secretion of plasminogen activator by SV3T3 cells may explain why formation of MIF-like activit is observed in SV3T3 but not 3T3 cultures. These results revearla biochemical pathway whereby a product secreted by virus-transformed cells affects one function of a cell central to the host's immunological defense system.

MATERIALS AND METHODS Cells. Balb/c 3T3 (clone A31) and Balb/c SV3T3 (clone T2) cells were grown in Eagle's minimal essential medium containing four times the usual amount of essential amino acids and vitamins (MEM X 4)/fetal calf serum (FCS), 10%/penicillin, 250 units/ml/and streptomycin, 250 Atg/ml (standard medium). The cell lines were repeatedly found to be negative for mycoplasma by culture assays in the laboratory of L. Hayflick and by [3H]thymidine autoradiography. GPPE cells were obtained from Hartley guinea pigs (normal males), 3-4 days after intraperitoneal injection of 25 ml of sterile light mineral oil (Exxon Corp., Freehold, N.J.), by lavage with 150 ml of Hanks' balanced salt solution. GPPE cells were collected in a separatory funnel and washed three times in Hanks' solution by centrifugation (300 X g for 10 min). GPPE cells were counted, resuspended at 5 X 107 eirs per ml in Eagle's minimal essential medium containing 100 units/ml penicillin and 100,gg/ml streptomycin, and placed in capillary tubes sealed with sterile paraffin. Only exudates containing 70% or more of large, oil-containing macrophages were used. In all cases, viability of the GPPE cells before packing in capillary tubes was 97% or better. Concentration and Purification of Plasminogen Activator. SV3T3 cells, seeded at 4 X 106 cells per 150 mm dish, were grown for 3 days in standard medium. Cultures containing 2 to 4 X 107 cells were then incubated overnight in 20 ml of standard medium without FCS to prepare a serum free harvest fluid (HF). Plasminogen activator was partially purified from 2.14 liters of SV3T3 HF by a modification of the technique described by Unkeless et al. (10). The HF was acidified to pH 3.0 with 1 M HCl, 31.3 g ammonium sulfate was added per 100 ml of acidified HF, and, after 30 rmin to 2 hr at 40, the solution was centrifuged at 15,000 X g, for 20 min at 4°. The pellet was dissolved in 1% of the original HF volume in 0.05 M glycine-

When sensitized lymphocytes are grown in vitro in the presence of specific antigen, they produce a substance known as migration inhibitory factor (MIF), which inhibits the migration of guinea pig peritoneal exudate (GPPE) cells (primarily macrophages) in vitro (1). We have previously reported that culture fluids from simian virus 40 (SV40)-transformed 3T3 cell cultures, but not from untransformed 3T3 cell cultures, also inhibit the migration of GPPE cells (2). In addition, such inhibited GPPE cultures lost a densely staining material from the cell-surface coat (DSM), as also occurs when GPPE cultures are treated with lymphocyte MIF (3). MIF-like activities have previously been shown to be produced by several different nonlymphoid cell lines in vitro (4). Recently, our finding that MIF-like activity was produced in SV3T3 cell cultures has been confirmed and considerably Abbreviations: MIF, migration inhibition factor; GPPE, guinea pig peritoneal exudate;- SV40, simian virus 40; DSM, densely staining material from the cell-surface coat; Dip-F, diisopropylfluorophosphate; UK, human urokinase; MEM X 4, Eagle's minimal essential medium containing four times the usual amount of essential amino acids and vitamins; GPS, guinea pig serum; FCS, fetal calf serum; HF, serum-free harvest fluid; CTA, Committee on Thrombolytic Agents. * To whom reprint requests should be addressed at (present address): Basic Research Program, Frederick Cancer Research Center, Frederick, Md. 21701.

1570

Cell Biology: Roblin et al.

Proc. Natl. Acad. Sci. USA 74 (1977)

1571

Table 1. Requirements for generation of MIF-like activity Experiment number 1

Medium/treatment

Control/incubated SV3T3 HF/incubated 3T3 HF/incubated Control/unincubated SV3T3 HF/unincubated 3T3 HF/unincubated

2

3

4

Average migration area*

% Inhibition


% Inhibition


% Inhibition

Average migration

1980 1249

0

37t

2197 1693 2434 2775 2597 2773

0 23 0 0 7 1

3059 1180 2232 2203 1698 2660

0 62t 28 0 23 0

3118 1324 2743 2829 2600 2796

20.73 2255 2046 1808

0 0 10 20

area*t

% Inhibition

0 58t 12 0 8 2

Aliquots of unconcentrated HF from 8 ml-cultures of SV3T3 cells (1 to 2 X 107 cells, 100 mm dish), 3T3 cells (2 to 4 X 106 cells, 100 mm dish) or plates without cells (control HF) contained 15% GPS and were incubated 16-18 hr at 37°. After overnight storage of the remaining HF (SV3T3, 3T3, and control) at 00, aliquots of these HF were combined with GPS (final concentration 15% vol/vol, also stored overnight at 00) to make the unincubated samples. Incubated and unincubated media were added to capillary tube cultures of GPPE cells for assay of MIF activity. * Average of two chambers each containing two capillary tubes; experiment 4, average of four chambers each containing two capillary tubes. t Glucose concentrations were determined by glucose oxidase technique on HF after overnight incubation with or without cells: control (no cells) 0.98 mg/ml; SV3T3 cells = 0.19 mg/ml; 3T3 cells = 0.69 mg/ml. Average migration area in SV3T3 HF medium is less than average migration in control medium; statistically significant (P < 0.01) by Student's t-test.

HCO buffer, pH 3.3, dialyzed against 500 volumes of glycine. HCI buffer, and stored at 4°. These plasminogen activator preparations were further concentrated 10- to 30-fold by negative pressure dialysis (Schleicher and Schuell, collodion bag no. 100) prior to Sephadex G-200 chromatography. Assay of Plasminogen Activator. Plasminogen activator activity was assayed on 125I-labeled fibrin-coated plastic petri dishes (7, 11) in the presence of partially purified FCS plasminogen. The plasminogen was isolated from FCS by one cycle of affinity chromatography on lysine-Sepharose 4B (12) and was determined to be 80% plasminogen by sodium dodecyl sulfate/polyacrylamide gel electrophoresis. Assay mixtures contained 0.1 ml of the G-200 column fraction, FCS plasminogen fractiQn (10-20 ,ug of protein as bovine serum albumin) and 0.1 M Tris-HC1 buffer, pH 8.1, in a total volume of 1.0 ml. Plates were incubated at 370 for 4 hr and 0.2 ml-aliquots of the supernatant removed for determination of the amount of released 12'I radioactivity. Generation and Assay of Macrophage Migration Inhibition Factor (MIF)Like Activity. Crude or partially purified SV3T3 HF or UK was incubated overnight (16-18 hr) at 37° in plastic tubes iiVMEM X 4 medium -containing 15% freshly thawed GPS. This medium was then assayed for MIF-like activity by adding it to chambers containing capillary tubes of pelleted GPPE cells. Chambers were incubated for 18-20 hr at 37° in 5% C02/95% air atmosphere, after which the area of migration image was traced and measured by planimetry. Trypan blue dye exclusion tests were performed on all test chambers to assess cell viability. Migration inhibition tests were considered valid only if 95% or more of the GPPE cells were viable. Because of the variability in migration areas in replicate assays, only >20% inhibition of macrophage migration is judged to be significant. GPPE cells in chamber cultures were fixed in situ and prepared for electron microscopic examination of DSM as previously described (3). UK. Purified UK ("Winkase" Sterling Winthrop) (35,000 CTA units/mg protein) was obtained through the courtesy of William P. Blackmore (Sterling Winthrop) and Joseph C. Fratantoni (Blood Disease Branch, National Heart and Lung Institute). (CTA refers to the Committee on Thrombolytic

Agents.) This preparation showed several peaks on Coomassie brilliant blue-stained sodium dodecyl sulfate (7.5%)/polyacrylamide gels with estimated molecular weights of 90,000 (minor component), 51,000 and 42,000 (major components), and 33,000 (minor). The 51,000, 42,000, and 33,000 molecular weight components were labeled after reaction with [3H]Dip-F (13), and this suggested that the 42,000 and 33,000 molecular weight components are breakdown products of UK which retain the ability to react with [3H]Dip-F. Preparation of Guinea Pig Serum. Blood was collected by cardiac puncture from a group of ether-anesthetized Hartley guinea pigs (males) into sterile glass tubes, allowed to clot for 45 min to 2 hr at room temperature, then centrifuged at 300 X g for 10 min at 4°. Aliquots of the resulting serum were stored at -20° until use. Serum was thawed immediately before use and was never refrozen. RESULTS Requirements for Generation of MIF-Like Activity. Study of the conditions required for the generation of MIF-like activity in SV3T3 cultures (Table 1) revealed that three factors were necessary. These were (i) a cellular factor which was secreted by the cells when they were cultured in serum-free medium, (ii) serum component(s) supplied by GPS, and (iii) overnight incubation at 37°. Thus, as shown by the data in Table 1, combination of an SV3T3 HF and guinea pig serum (GPS) plus overnight incubation yielded significant migration inhibition, in contrast to results obtained with HF from 3T3 cells or with unincubated samples. The morphological aspects of migration inhibition in GPPE cell cultures treated with HF-incubated medium resembled those previously described (2, 3) for cells whose migration was inhibited by medium from SV3T3 cultures. As shown in Fig. 1A, inhibited cells in HF-incubated medium lost the DSM detectable with the osmium potassium ferrocyanide technique (3). Control cells, cultured in SV3T3 HF-medium which had not been incubated, did not show migration inhibition and had a readily detectable DSM (Fig. 1B). Thus, a morphological correlate of migration inhibition in other systems (2, 3), i.e., the

1572


Proc. Natl. Acad. Sci. USA 74 (1977) Table 2. Generation of MIF-like activity by UK Experiment number

1 Volume of UK added (mll )

Incubated 0 0.04 (1600 0.06 (2400 0.08 (3200 0.10 (4000 Boiled UK,


2 % Inhibition 0


CTA units) CTA units) CTA units) CTA units) 0.04 ml

2694 1696 N.D. N.D. N.D. 2459

N.D. N.D. N.D. 9

2204 1445 1380 1019 1641 N.D.

0 0.04 (1600 CTA units) 0.10 (4000 CTA units) Boiled UK, 0.04 ml

2607 2913 N.D. 2426

0 0 N.D. 7

1432 1583 2036 N.D.

37t

3

4

bitiont


% Inhibition


% Inhibitiont

0 35 37 54 25 N.D.

2590 2108 2029 2237 2002 N.D.

0 19 22 14 23 N.D.

2673 1634 1540 1594 1884 N.D.

0 39 43 41 30 N.D.

0 0 0 N.D.

1805 2358 2237 N.D.

0 0 0 N.D.

1279 1553 1503 N.D.

0 0 0 N.D.

% Inhi-

Unincubated

Aliquots of UK (40 CTA units/ul in 0.12 M sodium phosphate buffer, pH 7) were added to MEM X 4 medium containing 15% GPS and antibiotics, and incubated overnight for 16-18 hr at 37°. After overnight storage separately of the remaining MEM X 4, GPS, and UK at 00, 0.04 or 0.10 ml of stored UK solution was added to medium containing 15% GPS to prepare the unincubated samples. Incubated and unincubated media were then added to chambers containing capillary-tube cultures of GPPE cells for assay of MIF activity. N.D., not determined. * Average of two chambers each containing two capillary tubes. t Average migration area in medium containing UK is less than average migration in control medium; statistically significant (P < 0.01) by Student's t-test.

loss of cell surface DSM, is also observed when GPPE cell migration is inhibited under the conditions described here. Generation of MIF-Like Activity by Purified UK. Because SV3T3 cells are known to secrete plasminogen activator into serum-free harvest fluids (5-7), we thought that plasminogen activator might be involved in the generation of MIF-like activity by hIF from SV3T3 cells. To obtain additional support

for this hypothesis, we determined whether UK, a known plasminogen activator, could also generate MIF-like activity when incubated in medium containing GPS. As shown in Table 2, overnight incubation of UK in MEM X 4 medium containing 15% GPS produced a medium which inhibited the migration of GPPE cells. Boiling the UK solution for 30 min destroyed its ability to hydrolyze '251-fibrin when assayed using 15% GPS as the source of plasminogen (data not shown) and also eliminated its ability to generate MIF-like

FIG. 1. Electron micrographs of GPPE cells migrating (A) in the presence of incubated SV3T3 HF medium (migration inhibited) and (B) in the presence of unincubated SV3T3 HF medium (migration uninhibited). Magnification: X22,800. Note the abundant dense surface material in (B) and its absence in (A).

activity (Table 2, experiment 1). The generation of MIF-like activity required a minimum concentration of 0.04 ml (1600 units) of UK. Increasing the dose of UK above this minimum concentration led to the generation of little further inhibitory activity (Table 2, experiments 2-4). Cochromatography of SV3T3 Cell Plasmihogep Activator and Ability to Generate MIF-Like Activity. The plasminogen activator activity present in HF from SV3T3 cultures was precipitated with ammonium sulfate and further concentrated by negative pressure dialysis as described in Materials and Methods. An aliquot of this concentrated SV3T3 HF was reacted with [3H]Dip-F to label the plasminogen activator (13), which then served as a marker for chromatography experiments. Examination of this [3H]Dip-F-labeled preparation by electrophoresis on sodium dodecyl sulfate/5% polyacrylamide gels revealed one major radioactive component (estimated molecular weight of 55,000) and three additional minor components (data not shown). Assay of slices from a duplicate gel showed a small peak of plasminogen activator activity which coelectrophoresed with the major [3H]Di>-F-labeled peak (data not shown). Thus, the major [3H]Dip-F-labeled component of the concentrated SV3T3 HF had the molecular weight and functional characteristics expected for a major species of mouse cell plasminogen activator (13). Aliquots of partially purified, untreated SV3T3 HF and [3H]Dip-F-labeled SV3T3 HF were mixed and cochromatographed on a Sephadex C-200 column. As shown in Fig. 2, the plasminogen activator activity cochromatographed with the [3H]Dip-F label. When aliquots of the column fractions were assayed for their ability to generate MIF-like activity, an excellent correlation was observed between the amount of plasminogen activator in the fraction (as judged by the amount of [3H]Dip-F label), the amount of plasminogen activator activity, and the ability to generate migration-inhibitory activity. Without overnight incubation, fraction 168 (which contained maximal plasminogen activator activity) failed to yield migration-inhibitory activity. These results establish that the factor


Cell Biology: Roblin et A Table 3. Effect of [3H]Dip-F treatment on plasminogen-activator activity and capacity to generate MIF-like activity

w z 0. 0 0

0

Treatment*

% 125I-labeled fibrin (cpm released)

Average migration areat

% Migration inhibition

6.15 17.6 3.3

2658 1625

0

39t

2659

0

1573

0 ous

UcF c

0~o cJ

_

ii +

[3H]Dip-F

Mock Buffer control Plasminogen alone MEM x 4 medium

5.1

* Pooled G-200 column fractions (167 and 169, Fig. 2) were dialyzed against 0.1 M Tris-sulfate, pH 7.2 and 0.5 ml-aliquots were combined with either [3HJDip-F (New England Nuclear) (0.1 ml, 100 MCi, 0.021 mg in propylene glycol) or an equal volume of propylene glycol (mock-treatment). After incubation at room temperature for 7 hr, the reaction mixtures were dialyzed against 1 liter of 0.05 M glycine-HCl buffer, pH 3.3. Aliquots of these dialyzed activator preparations (0.2 ml of [3H]Dip-F treated, 0.1 ml of mock-treated) were assayed for plasminogen activator. Aliquots (0.4 ml) were incubated overnight in MEM X 4 containing 15% GPS (2.6 ml) and assayed for MIF-like activity as described under Materials and Methods. t Average of three test chambers each containing two capillary tubes. I Average migration area in mock-treated sample is less than average migration area in buffer control; statistically significant (P < 0.01) by Student's t-test.

in SV3T3 HF which generates migration inhibition has the same molecular size as SV3T3 cell plasminogen activator. Dip-F Treatment Inactivates Plasminogen Activator Activity and Capacity to Generate MIF-Like Activity. If SV3T3 cell plasminogen activator is responsible for generation of the MIF-like activity of SV3T3 cultures, then inactivation of the plasminogen activator activity would also be expected to inactivate the capacity to generate MIF-like activity. To test this hypothesis, we combined fractions from the Sephadex G-200 column plasminogen activator peak region (Fig. 2 upper), treated one aliquot with [3H]Dip-F in propylene glycol, while the other aliquot was mock-treated. As shown in Table 3, treatment with [3H]Dip-F inactivated about 95% of the plasminogen-activator activity and completely eliminated the ability to generate MIF-like activity. This experiment has been repeated three times, on either crude SV3T3 HF material or on Sephadex G-200 column fractions containing SV3T3 cell plasminogen activator. In each case, [3H]Dip-F treatment led to approximately 95% inhibition of plasminogen activator activity and completely abolished the capacity to generate MIF-like activity when incubated overnight with medium containing 15% GPS.

DISCUSSION Our results indicate that (i) migration-inhibitory activity for GPPE (macrophage) cell cultures can be generated by the interaction of a cellular component secreted by SV3T3 cells and component(s) present in guinea pig serum; (ii) the ability of crude and partially purified SV3T3 HF to generate MIF-like activity and the plasminogen activator activity of the HF are both abolished by [3H]Dip-F treatment; (iii) the major [3H]Dip-F reactive component of partially purified SV3T3 HF has the molecular weight expected of mouse cell plasminogen activator; and (iv) the plasminogen activator activity and the ability to generate MIF-like activity are both concentrated by

zN

C

A 0.IL 0.

U.

a EQ

FRACTION NUMBER

FIG. 2. Sephadex G-200 chromatography of partially purified SV3T3 HF. Plasminogen activator activity from 2.14 liters of SV3T3 HF was partially purified and concentrated to 1 ml as described under Materials and Methods. An 0.4 ml-aliquot of this material was labeled by reaction with [3H]Dip-F (13). This [3H]Dip-F-labeled material was combined with 0.5 ml of untreated, concentrated SV3T3 HF and 1 Ml of Na35SO4 (New England Nuclear, 1 gCi/ml) and the mixture was chromatographed on a 52 X 2.5 cm diameter column of Sephadex G-200 in 0.05 M glycine-HCl buffer, pH 3.3. Fractions of 1 ml were collected and 0.1-ml aliquots were assayed for plasminogen activator activity (upper panel, total radioactivity: *, 4.9 X 104 cpm/plate; 0, 6.7 X 104 cpm/plate; 0, 6.9 X 104 cpm/plate). Fractions 130-190, assayed in the absence of plasminogen with 125I, released a maximum of 500 cpm/0.2 ml. Aliquots (0.3 ml) of several different fractions were combined with 0.45 ml of GPS and 2.25 ml of MEM X 4, incubated overnight at 370, and assayed for MIF-like activity in three replicate chambers, each containing two capillary tubes (upper panel, dark bars). Inhibition by material in fractions 168 and 180 was statistically significant (P < 0.01) by the Student's t-test. Fraction 168 (0.3 ml), assayed at the same time for MIF-like activity without overnight incubation, showed zero inhibition of cell migration. 3H cpm were determined by assaying 0.1 ml-aliquots in 10 ml Bray's scintillation solution (lower panel, 0).

the low pH/ammonium sulfate precipitation procedure and cochromatograph on a Sephadex G-200 column. The simplest interpretation of these results is that the SV3T3 cell-plasminogen activator generates MIF-like activity upon interaction with a component(s) of guinea pig serum. Our finding that UK, a known plasminogen activator, also generates MIF-like activity upon overnight incubation with GPS (Table 2) is consistent with this interpretation. Our previous results (2) and data reported here (Table 1) show that untransformed 3T3 cell cultures and HF prepared from such cultures, generally do not yield MIF-like activity under conditions where SV3T3 cell culture medium and SV3T3 HF do so. We suggest that the quantity of plasminogen activator in the HF is the determining factor because the HF from SV3T3 cells assayed in Table 1, experiment 4, contained eight times more plasminogen activator activity than the 3T3 cell HF in that experiment (data not shown). The observation that generation of MIF-like activity by urokinase is not reproducibly observed until at least 1600 CTA units of UK are used (unpublished data and Table 2) demonstrates that a certain threshold dose of plasminogen activator is required for generation of the MIF-like activity and is consistent with this expla-

1574


nation of the lack of formation of MIF-like activity by 3T3 cell cultures and HF from 3T3 cells. Our results indicate that plasminogen activator activity produced by SV3T3 cells can initiate the formation of molecule(s) which inhibit the migration of GPPE cells. The requirement for overnight incubation to generate MIF-like activity indicates that SV3T3 cell plasminogen activator is probably not the migration inhibitory factor (MIF) itself. In addition, the fact that unincubated SV3T3 HF is not inhibitory (Table 1) indicates that migration inhibition is not due solely to depletion of glucose or other nutrients from the medium (14), despite the fact that SV3T3 HF contains less residual glucose than 3T3 HF (Table 1, footnote). By adding extra glucose to the SV3T3 HF (Table 1, experiment 4) to a final concentration of 1 mg/ml before overnight incubation with GPS, we did not significantly reduce the migration inhibitory activity observed in the incubated sample (data not shown). Addition of SV3T3 cell plasminogen activator or UK to medium containing 15% GPS produces a burst of fibrinolytic activity which decreases progressively over the next several hours, and indicates formation and inactivation of plasmin under such conditions. Thus, formation of the MIF-like activity could be a consequence of direct action of plasminogen activator upon some serum molecule, or generation of the MIF-like activity could be a secondary consequence of the formation of plasmin by plasminogen activator. It is noteworthy, however, that little or no fibrinolytic activity can be detected after overnight incubation of plasminogen activator in 15% GPS medium, at a time when maximal migration inhibitory activity is observed. This suggests that free-plasmin activity is not responsible for migration inhibition, but leaves open the possibility that plasmin formation is an intermediate step in the process of generating the migration-inhibitory molecule(s). Poste (5) has compared the size and function of SV3T3 cell MIF-like activity and plasminogen activator. Our work cannot be directly compared with that of Poste because he was studying the properties of the active molecule with MIF-like activity while we have provided evidence for the role of plasminogen activator in the initial generation of the active molecule. Plasminogen activator does not appear to be the MIF produced by guinea pig lymphocytes in response to concanavalin A or sensitizing antigen, because antigen activity is not inhibited by Dip-F and does not require overnight incubation for activity (15). Plasminogen activator may play a role in MIF produced by human lymphocytes incubated with purified protein derivative, since Havemann et al. have shown that this activity is inhibited by incubation with Dip-F (16). On the assumption that several factors will be found to play a role in as complex a phenomenon as macrophage migration in vitro, it is perhaps not surprising that there appear to be at least two different types of factors responsible for migration inhibition. Finally, our results suggest a possible explanation for the


observation by Cohen et al. (17) that sera from patients with a variety of lymphoproliferative diseases contain a migrationinhibitory activity for GPPE cells. If the tumor cells involved in such malignancies resemble other human tumor cells in secreting high levels of plasminogen activator (8, 9), then this plasminogen activator secretion might generate MIF-like activity as we have shown here for SV3T3 cell plasminogen activator and UK. Alternatively, the malignant lymphoid cells might be producing an MIF which is independent of plasminogen activator. In any event, the mechanism(s) responsible for the migration inhibitory activity in the sera of certain cancer patients, and the possible effects of such activity on macrophage function and consequent tumor growth appear worthy of further investigation. We thank Michael Fitzsimmons, Sara P. O'Donnell, Salvatore S.

Selvaggio, Jane Goodwin, and Ellen Morgan for excellent technical assistance and Dr. Sidney Rieder for the glucose determinations. This work was supported by U.S. Public Health Service Grants CA-15889, CA-10126, CA-16881, and CA-19141. During much of this work R.O.R. was a Faculty Research Associate of the American Cancer Society (PRA-75). M.E.H. is a Research Scholar Award recipient from the American Cancer Society. 1. David, J. R. (1971) Fed. Proc. 30, 1730-1735. 2. Hammond, M. E., Roblin, R. O., Dvorak, A. M., Selvaggio, S. S., Black, P. H. & Dvorak, H. F. (1974) Science 185,955-957. 3. Dvorak, A. M., Hammond, M. E., Dvorak, H. F. & Karnovsky, M. J. (1972) Lab. Invest. 27,561-574. 4. Papageorgiou, P. S., Henley, W. L. & Glade, P. R. (1972) J. Immunol. 108, 494-504. 5. Poste, G. (1975) Cancer Res. 35, 2558-2566. 6. Ossowski, L., Unkeless, J. C., Tobia, A., Quigley, J. P., Rifkin, D. B. & Reich, E. (1973) J. Exp. Med. 137, 112-116. 7. Chou, I. N., Black, P. H. & Roblin, R. 0. (1974) Nature 250, 739-741. 8. Rifkin, D. B., Loeb, J. N., Moore, G. & Reich, E. (1974) J. Exp. Med. 139, 1317-1328. 9. Laug, W. E., Jones, P. A. & Benedict, W. F. (1975) J. Natl. Cancer Inst. 54, 173-179. 10. Unkeless, J., Dano, K., Kellerman, G. M. & Reich, E. (1974) J.

Biol. Chem. 249,4295-4305. 11. Unkeless, J. C., Tobia, A., Ossowski, L., Quigley, J. P., Rifkin, D. B. & Reich, E. (1973) J. Exp. Med. 137, 85-111. 12. Deutsch, D. G. & Mertz, E. T. (1970) Science 170, 1095-1096. 13. Unkeless, J. C., Gordon, S. & Reich, E. (1974) J. Exp. Med. 139, 834-850. 14. Taylor, M. M., Burman, C. J. & Fantes, K. H. (1975) Cell. Immunol. 19, 41-57. 15. David, J. R. & Becker, E. L. (1974) Eur. J. Immunol. 4, 287289. 16. Havemann, K., Burger, S., Schmidt, W., Sodomann, C. P. & Stein, P. (1973) in Proceedings of the Seventh Leulocyte Conference, ed. Daguillard, F. (Academic Press, New York), pp. 315-325. 17. Cohen, S., Fisher, B., Yoshida, T. & Bettigole, R. (1974) N. Engi. J. Med. 290,882-886.