Production of migration-inhibitory factor by a human T-lymphoblast cell line

CELLULAR

IMMUNOLOGY

78, 305-3 13 (1983)

Production of Migration-Inhibitory Factor by a Human T-Lymphoblast Cell Line’ H. G. REMOLD, A. MEDNIS, T. KAWAGUCHI,N. BERSCH,AND D. W. GOLDE The Department of Medicine, Harvard Medical School and the Department of Rheumatology/ Immunology, Brigham and Women’s Hospital, Boston, Massachusetts 02115, and the Division of Hematology/Oncology, Department of Medicine. UCLA School of Medicine, Los Angeles. California 90024 Received January 6. 1983; accepted March 9, 1983

Migration-inhibitory-factor (MIF) activity was detected in culture supematants of the human T-lymphoblast cell line MO after stimulation with phytohemagglutinin and phorbol myristate acetate. MIF activity was not detected in unstimulated cultures reconstituted with phytohemagglutinin and phorbol my&ate acetate. Conditioned medium from the cell line MO was fmctionated by Sephadex G- 100 gel filtration. MIF-containing Sephadex fractions corresponding to a M, of 60,000 to 70,000 were. further fractionated by isoelectrofocusing resulting in a sharp peak of activity with a pI of 4.6 to 5.2. This MIF speciesconstitutes a major form secreted by MO cells; it adheres to Con A-Sepharose, is trypsin-resistant, and is denser than pure protein as determined by CsCl density gradient centrifugation. These are the same physicochemical characteristics previously established for second-day pHS-MIF from peripheral blood mononuclear cells (W. Y. Weiser et al., J. Immunol. 126, 1958, 1981). In contrast, Sephadex fractions corresponding to larger molecules (M, 70000-90000) contain at least two additional MIF species. These larger MIF forms have a pI of 3.0 to 3.5 and of 4.6 to 5.2 and lack affinity to Con ASepharose.Thus, the MO T-cell line produces large quantities of at least three different species of human MIF.

INTRODUCTION It is now generally accepted that lymphokines play a vital role in modulating cellular immune responses. Some of these mediators which are produced by lymphocytes after stimulation by antigen or mitogen are directed toward the macrophage. These lymphokines include migration-inhibitory factor (MIF),* a lymphokine which is thought to concentrate macrophages at inflammatory sites (2, 3) and a closely related lymphokine, macrophage-activating factor (4, 5). Human MIF has been partially purified and characterized using behavior on gel filtration, isoelectrofocusing, and sensitivity to proteolytic enzymes and neuraminidase. Three MIF specieshave been identified. Supernatants from a 24-hr incubation of concanavalin A (Con A)-stimulated mononuclear cells contain a single MIF species ’ This study was supported by NIH Grants CA-30388, CA-32737, AI-121 10-07, and HL-2253402. ’ Abbreviations used: MIF, migration-inhibitory factor, TPA, phorbol my&tic acetate; PBS, phosphatebuffered saline; MEM-PS, minimum essential medium containing 100 U/ml penicillin and 100 &ml streptomycin; Con A, concanavalin A; percent I, percentage inhibition; PHA, phytohemagglutin; HBSS, Hanks’ balanced salt solution. 305 0008-8749/83 $3.00 Copynght 0 1983 by Academic Press. Inc. All rights of reproduction in any foorm reserved

306

REMOLD ET AL.

with a pZ of 5 (first-day MIF), whereas two MIF species, a pH 3 MIF and a pH 5 MIF, were detected in supernatants from a further 24-hr incubation of the same cells (second-day pH 3 MIF and pH 5 MIF) (1). It was anticipated that this pronounced polymorphism would complicate the biochemical analysis of MIF. Moreover it has been difficult to find a suitable cellular source producing sufficient amounts of MIFrich supernatants to make large-scale biochemical analysis feasible. The present report describes the production of MIF on mitogenic stimulation of the cell line MO, a T-lymphoblast line derived from a patient with a T-cell variant of hairy cell leukemia (6). MO cells were previously described to produce other lymphokines including gamma interferon and colony-stimulating factor (7, 8). MIF activity secreted by stimulated MO cells consists of at least three different molecular specieswhich can be separatedfrom each other by gel filtration and isoelectrofocusing. MATERIALS

AND METHODS

Production of MO MZF. MO cells were maintained in Iscove’s medium (9) with 20% fetal calf serum, 2 mM glutamine, 100 U/ml penicillin, and 100 &ml streptomycin. For the production of MO MIF the cells were cultured at 1 X 106/ml in serum-free Iscove’s medium with 3% PHA (Burroughs Wellcome, Greenville, N.C.) and 5 rig/ml phorbol myristic acetate (TPA) for 96 hr at 37°C. TPA was added 2 hr prior to the addition of PHA. Control supematants were produced by incubation of the MO cells in the absence of PHA and TPA and reconstitution of PHA and TPA after removal of the cells from the supematant by centrifugation. After removal of the cells, the media were concentrated by vacuum dialysis to l/100 the original volume and stored at -70°C. It should be noted that MO cells also produced substantial amounts of MIF when incubated with PHA alone in the absence of TPA. Production of radiolabeled MO MZF. To produce radiolabeled MO MIF, lo6 MO cells were cultured in 1 ml serum-free and leucine-free Iscove’s medium with 3% PHA and 5 pg/ml TPA substituted with L-[4,5-3H]leucine at 250 ~1 (1 mCi/ml; 1 Ci = 3.7 X 10” Bq; Amersham/Searle, Arlington Heights, Ill.) for 96 hr at 37°C. The cells were removed by centrifugation and the supematants processed as described above. Sephadex G-100 fractionation. Aliquots of 2.5 to 3 ml were dialyzed against phosphate-buffered saline (PBS) for 3 hr and filtered over a Sephadex G-100 column (2.5 X 100 cm). The column effluent was pooled into fraction 1 containing the void volume, fraction 2 containing molecules with an apparent it& of 70,000 to 90,000, fraction 3 including molecules with an apparent 1M,of 55,000 to 70,000, fraction 4 with molecules with an apparent it& of 40,000 to 55,000, and fraction 5 containing molecules with an apparent M, of 25,000 to 40,000. Zsoelectrofocusing. The fractions 2 and 3 containing MIF activity were dialyzed separately against distilled water and subjected to isoelectrofocusing in a 0.4 to 47.0% sucrosegradient containing 0.5% pH 2.5 to 4.0 and 1%4.0 to 6.0 Ampholines (LKB; Bromma, Sweden) (10) utilizing an LKB 8 102 Ampholine electrofocusing column. Prefocusing for 2 1 hr at 6 W before application of the sample established a pH gradient from 2.0 to 6.2. The sample was injected into the center of the column and isoelectrofocused for 22 hr at 6 W. Four-milliliter fractions of the column effluent were collected into tubes containing bovine serum albumin, resulting in a final concentration of 75 pg/ml bovine serum albumin. Six fractions were then pooled and di-

PRODUCTION

OF MIGRATION-INHIBITORY

FACTOR

BY MO

307

alyzed against 20 mM EDTA, 20 mM NaCl, 5.3 mJ4 MgQ, and 0.5 mil4 dithiothreitol and vacuum concentrated to 4.0 ml for MIF assay. Con A-Sepharose afinity chromatography. MIF-containing Sephadex G- 100 fractions (0.5 ml) were dialyzed for 24 hr against PBS at 4°C and applied to a Con ASepharosecolumn (6 X 1 cm) equilibrated with PBS. The column was washed with 40 ml of PBS; the combined effluent was called the nonadherent fraction. Adherent components were then eluted with 40 ml of 0.2 M a-methyl-Dmannoside in PBS. The nonadherent fraction and the fraction which adhered and was eluted were vacuum concentrated to 2.5 ml for the MIF assay. Zsopycnic centrijkgation. Isopycnic centrifugation of MO MIF was performed as described ( 11). In brief 1.5 ml concentrated MO MIF from the isoelectrofocusing step was added to 0.02 M Tris-HCI buffer, pH 8.0, 0.3 mM EDTA, 20 mM NaCl, and 5.3 mM MgC12to a volume of 5 ml and the concentration of CsCl adjusted to 30.0% (w/w). After centrifugation for 48 hr at 40,000 rpm in a SW-50.1 rotor (Beckman Instruments, Palo Alto, Calif.) at 4°C seven fractions were collected ranging from p25 1.244 to 1.485 g/ml each covering a density increment of approximately 0.02 g/ml. The fractions were dialyzed against minimum essential medium containing 100 U/ml penicillin and 100 &ml streptomycin (MEM-PS) and assayed for MIF activity. Treatment of MO MZF with trypsin. Concentrated MO MIF containing isoelectrofocusing fractions (0.6-1.0 ml) were diluted to 2.0 ml with Hanks’ balanced salt solution (HBSS) and incubated with and without 16 to 20 U/ml TPCK-trypsin (80 to 200 U/mg, Sigma Chemical Co., St. Louis, MO.) for 1 hr at 37°C (12) and pH 8.0. To terminate the reaction diisopropylphosphofluoridate was added to IO-’ M and the mixture incubated for 15 min at room temperature, dialyzed against MEMPS for 3 hr at 4°C and assayedfor MIF activity. MZF assay. The fractions to be tested were dialyzed against MEM, adjusted to 15%with normal guinea pig serum, and tested in a capillary-tube assaywith normal guinea pig peritoneal exudate cells as target cells (11). MIF activity was expressedas percentage of inhibition of cell migration as percent I = 100 -

average migration in MIF-containing fractions x 100. average migration in control fractions

Inhibition of migration greater than 20% was considered to be significant (11). RESULTS Production of MIF by MO Cells

MO cells produce MIF when stimulated with PHA and TPA. The highest MIF titers are obtained when the MO cells are incubated at a concentration of 1 X IO61 ml in serum-free Iscove’s medium in the presence of PHA and TPA as described under Materials and Methods. The quantity of MIF activity secreted by MO cells is 10times that secretedby human peripheral blood mononuclear cells. MO MIF activity is still detected when 0.7 ~1 pooled active Sephadex G- 100 fraction/ml medium is assayed.By contrast, human peripheral lymphocyte MIF is active only when 60-70 ~1 of Sephadex G- 100 fraction/ml medium are tested (eight experiments). Cells incubated in the absenceof PHA and TPA did not produce MIF. Likewise, supernatants

308

REMOLD ET AL. TABLE 1 MIF Activity of Media from Stimulated MO Cells Fractionated on Sephadex G-100 MIF Activity (Percent I) Fractions

Experiment

Microliters

1

300 150

2

(O.OfO.9)

(0.09L.27)

0 0

23 0

50 36

4 0

6 8

300 150

8 5

38 26

32 13

9 0

16 14

3

300 150

15 6

42 31

41 24

15 22

2 9

4

300 150

27 13

43 15

45 24

5 0

3 9

(0.27255)

(0.55L.74)

DFractions 1-5 were concentrated to 0.0 1 of the volume of the starting supematant. Each fraction of 300 and 150 pl was analyzed for MIF activity. The numbers represent percentage I. The experiments were analyzed by the two-way analysis of variance. MIF activity in fractions 2 and 3 was statistically significant (P < 0.001). bNumbers in parenthesesindicate Kd range.

from cells incubated in the absenceof mitogen and reconstituted with PHA and TPA after removal of the cells showed no MIF activity (experiment not shown). Fractionation of Mo MIF by Sephadex G-100 Gel Filtration Culture media from MO cells incubated with PHA and TPA were fractionated on Sephadex G- 100. MIF activity was consistently found in fractions 2 and 3 which contained molecules with an apparent M, of 55,000 to 90,000 (& of 0 to 0.27). No activity was found in fractions 4 and 5 which contain molecules of the size of ovalbumin (M, 45,000) and chymotrypsinogen (Mr 25,000). Isoelectrofocusing Fraction 3 from Sephadex G-100 was subjected to isoelectrofocusing. In seven experiments maximum MIF activity was consistently found in fractions containing molecules with a pZ of 4.9 to 5.3. Three typical experiments are shown in Fig. 1. Recovery and Purification of MO MIF from Fraction 3 from Sephadex G-100 Gel Filtration and Isoelectrofocusing Figure 2 shows the activity of different concentrations of MIF on the ordinate of samples from crude MO-cell supematant, from Sephadex G- 100 gel filtration fraction 3 and from the pooled active isoelectrofocusing fractions plotted against their protein content on the abscissa. It can be seen that an amount of crude MIF supematant yielding 50% I contains an average of 90 &ml protein. After Sephadex G-100 gel filtration, 1 ml of effluent yielding the same amount of MIF activity contains 16 pg/ ml protein (five- to sixfold purification) with a yield of 60%. Following isoelectrofocusing 1 ml containing the same MIF activity contained 2 pg protein giving a

PRODUCTION OF MIGRATION-INHIBITORY

40

FACTOR BY MO

309

6

4

20

0

Expeviment2

P -6

P 140s i= B .2Qs 2

-2

40

6

4

20

2 0 0

I 153

I

I

mo

I

4ocl

FIG. 1. Isoelectrofocusing of MO MIF from Sephadex G-100 fraction 3 (& range 0.09 to 0.27). Three experiments are shown. The column effluent in milliliters is indicated on the abscissa,the pH of the fractions (solid line) and the MIF activity (percent I) are indicated on the ordinate. Material from Sephadex G-100 fractionation of 300 ml of MO cell-conditioned media was isoelectrofocused in a pH gradient from 2.0 to 6.1. Eleven fractions as indicated by the bars were collected, concentrated to 4.0 ml, 0.4 ml of which was tested for MIF activity. One peak of MIF activity was recovered at a pH of 4.9 to 5.3. The experiments were analyzed by the two-way analysis of variance. In 10 experiments MIF-activity was statistically sign&ant (P -C0.001) in fractions 2, 8 and 9.

FIG. 2. Specific activity of MO MIF after partial purification by Sephadex G-100 gel filtration and isoelectrofocusing. The activity of decreasing dilutions of MO MIF in crude supematants (crude SN), in Sephadex G- 100 fraction 3 (G- 100), and after IEF-fractionation (IEF) is seen on the ordinate. The abscissa shows the protein content per milliliter.

310

REMOLD ET AL.

further eightfold purification with a yield of 46%. We have thus obtained an overall 45-fold purification and a yield of 28%. Fractionation

of MO MIF by Con A Ajinity

Chromatography

To determine whether MO MIF is glycosylated, Sephadex fraction 3 was further purified by isoelectrofocusing and applied to a Con A-Sepharose column. No MIF activity was detected in the nonadherent fraction in three experiments (Table 2). When adherent proteins were eluted from the Con A column with a-methyl-D-mannoside, 70 to 80% of the MIF activity was recovered. Therefore, MO MIF from Sephadex fraction 3 adheres to Con A-Sepharose and can be eluted with cy-methylD-mannoside suggesting that the protein is glycosylated. Isopycnic Centrifugation

To verify its glycoprotein nature, MO MIF was subjected to isopycnic centrifugation in a CsCl density gradient ranging from 1.164 to 1.517 g/ml. Two experiments are shown in Fig. 3. MIF activity was recovered from fractions with p25of 1.340 to 1.400 g/ml which includes the densities of typical glycoproteins. Thus, MO MIF is denser than albumin, indicating it is a glycoprotein. The Efect of Trypsin on MO MIF

MO MIF from Sephadex G- 100 fraction 3 most closely resembles the pH 5 MIF which is produced by human mononuclear cells on the second day of culture. A TABLE 2 MIF Activity of MO-Media Fractions from Sephadex G- 100 Chromatogmphed on Con A-Sepharose’ MIF Activity (Percent I) Nonadherent fraction

Adherent fraction

Sephadex fraction 2 (M, 70,000 to 90,000) Expt. 1 2 3

47 37 42

9 18 9

Sephadex fraction 3 (M, 55,000 to 70,000) Expt. I 2 3

10 0 13

47 24 26

a Sephadex fractions 2 and 3 were chromatographed on Con A-Sepharose column. After application of the sample, the effluent was collected (nonadherent fraction) and adherent material was then eluted using 0.2 M cY-methyl-D-mannoside(adherent fraction). An additional fraction eluted with 50% ethylene glycol in 0.2 M cu-methyl-D-mannosidewas inactive in three experiments (not shown). All the experiments within each group were analyzed together by the two-way analysis of variance: There was statistically significant MIF activity in the nonadherent fraction from Sephadex fraction 2 compared ta the adherent fraction (P < 0.01) and in the adherent fraction from Sephadex fraction 3 when compared to the nonadherent fraction (P < 0.05).

PRODUCTION OF MIGRATION-INHIBITORY

I

I I 1234567

I

I

I

FACTOR BY MO

I

311

I

FRACT/OrV NUMBER 3. Isopycnic centrifugation of MO MIF recovered from Sephadex G-100 in fraction 3 (Kd of 0.09 to 0.27). Two experiments are shown. Fraction 3 derived from 200 ml crude conditioned medium was concentrated to 2 ml and centrifuged in a CsCl density gradient as described under Materials and Methods. The bars show the MIF activity (percent I) of 0.3 ml from each fraction. The buoyant density (pZS)is indicated on the abscissa.The experiments were analyzed by two-way analysis of variance. MIF activity in fractions 3 and 4 was statistically significant (P < 0.001). FIG.

further criterion allowing differentiation between the MIF speciesproduced by human blood mononuclear cells is trypsin sensitivity. Human firstday MIF is trypsin sensitive, whereas second-day pH 5 MIF is trypsin resistant (12). When MO MIF was subjected to trypsin treatment, no alteration of MIF activity was detectable. In three experiments MIF preparations with activities of 58, 37, and 66% inhibition showed activities of 57, 37, and 70% inhibition after treatment with 16 U/ml of TPCKtrypsin for 1 hr at 37°C. Properties of MO MIF from Sephadex G-100 Fraction 2 (Kd 0.0 to 0.09) In further experiments, Sephadex fraction 2 was chromatographed on Con ASepharose.MIF activity was consistently found in the nonadherent fractions, whereas the cY-methyl-Dmannoside eluate contained no activity (Table 2). MO MIF from the Sephadex G-100 fraction 2 containing molecules with an apparent M, of 70,000 to 90,000 amounts to approximately one-half of the total MIF activity (Table 1). When this MIF fraction was subjected to isoelectrofocusing, a peak of MIF activity was found at pH 4.9 to 5.3, cofractionating with MIF activity from Sephadex G-100 fraction 3. In addition, another peak of MIF activity was detected at pH 3.0 to 3.5 (Fig. 4) which was not present when Sephadex fraction 3 was subjected to isoeiectrofocusing (Fig. 1).

312

REMOLD ET AL. Expefhent

f 6

2

80

6

60

FIG. 4. Isoelectrofocusing of MO MIF recovered from Sephadex G-100 gel filtration in fraction 2 (& range of 0.0 to 0.09). The column effluent is indicated on the abscissa.The eleven pooled fractions are represented by bars. MIF activity and the pH of the fractions (solid line) are indicated on the ordinates. Material from 300 ml MO cell-conditioned media was fractionated on Sephadex G-100 and the fractions were concentrated to 3.0 ml. Fraction 2 was isoelectrofocusedand assayedfor MIF activity. Note that two peaks with MIF activity can be discerned, one at a pI of 3.5 to 4.5 and the other at pI of 4.8 to 5.2. All experiments were analyzed by two-way analysis of variance. MIF activity was statistically significant in fractions 5, 8, and 9 in all experiments; in Experiment 1, fraction 7, and in Experiment 3 fractions 5 and 8 were statistically significant (P -z 0.001).

DISCUSSION Thesestudies demonstrate that the human T-cell line MO produces upon stimulation with PHA and TPA significant quantities of at least three different MO MIF species. One speciesis found in Sephadex G-100 fraction 3, has a pI of 4.6-5.2, and is bound to Con A. A second species is in fraction 2, has a pI of 4.6-5.2, and is apparently not bound to Con A. A third form is also found in fraction 2, has a p1 of 3.0, and is not bound to Con A.

PRODUCTION OF MIGRATION-INHIBITORY

FACTOR BY MO

313

Another lymphokine produced by MO cells similar to MIF is gamma-interferon (IFN-y). The MO cells were found to produce an IFNy specieswhich chromatographs on Ultrogel columns with molecules having a i&f, of 40,000, is retained on Con ASepharosecolumn, and is acid and heat labile (8). As of now we cannot separate the pH 5 MO MIF from MO IFN-y on the basis of its isoelectric point and its behavior on Con A-Sepharose columns. On the other hand, we found that Mo-1FN-y activity is never detected in Sephadex G- 100 fractions of an approximate molecular weight greater than 70,000. On the other hand, neutrophil migration-inhibition factor, erythroid-potentiating factors, and colony-stimulating factor, also products of this cell line, are different entities (4, 7, 14, 15). There have been a number of reports describing the production of lymphokines by human B- and T-cell lines ( 16, 17). Analysis of the supematants of several of these cell lines in our laboratory has been disappointing. MIF activity from the supematants of several of these cell lines was either unstable after several purification steps or was associatedwith molecules with a M, smaller than 15,000. The MO cell line has thus gained importance because it produces human MIF with close similarities to a classical MIF species from mononuclear cells. REFERENCES I. 2. 3. 4. 5.

Weiser, W. Y., Greineder, D. K., Remold, H. G., and David, J. R., J. Immunol. 126, 1958, 198I. Bloom, B. R., and Bennett, B., Science 153, 80, 1966. David, J. R., Proc. Nat. Acad. Sci. USA 56, 72, 1966. Piessens,W. F., Churchill, W. H., and David, J. R., J. Immunol. 114, 293, 1975. Churchill, W. H., and Cameron, D., In “Biochemical Characterization of Lymphokines” (A. L. de Week et al.. Eds.), p. 137. Academic Press,New York, 1980. 6. Saxon, A., Stevens, R. H., Quart, S. G., and Golde, D. W., J. Immunol. 120, 777, 1978. 7. Golde, D. W., Bersch, N., Quan, S. G., and Lusis, A. J., Proc. Nat. Acad. Sci. USA 77, 593, 1980. 8. Nathan, I., Groopman, J. E., Quan, S. G., Bersch, N., and Golde, D. W., Nature (London) 292, 842, 1981. 9. Iscove, N. N., and Melchers, F., J. Exp. Med. 147, 923, 1978. 10. Remold, H. G., and David, J. R., J. Zmmunol. 107, 1090, 1971. 11. Remold, H. G., Katz, A. B., Haber, E., and David, J. R., Cell Immunol. 1, 113, 1970. 12. Remold, H. G., and Mednis, A. D., J. Immunol. 122, 1920, 1979. 13. Remold, H. G., and Mednis, A. D., J. Immunol. 118, 2015, 1977. 14. Lusis, A. J., Quon, D. H., and Golde, D. G., Blood 57, 13, 1981. 15. Weisbart, R. H., Lusis, A. J., Chart, G., Billing, R., Ashman, R. F., and Golde, D. G., J. Immunol. 128,457,

1982.

16. Yoshida, T., Kuratsuji, D., Takada, A., Takada, Y., Minowada, Y., and Cohen, S., J. Zmmunol. 117, 548, 1976. 17. Schook, L., Otx, U., Lazary, S., and de Week, A. L., In “Biochemical Characterization of Lymphokines, Proceedings of the Second International Lymphokine Workshop” (A. L. de Week et al., Eds.), p. 67. Academic Press,New York, 1980.