Synthesis and Secretion of Multiple Forms of &Interferon/B-cell ...

THEJOURNAL OF BIOLOGICAL CHEMISTRY 0 1988 by The American Society for Biochemistry and Molecular Biology, Inc.

Vol. 263, No. 16, Issue of June 5,pp. 7760-7766,1988 Printed in O.S.A.

Synthesis and Secretion of Multiple Forms of &Interferon/B-cell Differentiation Factor2/Hepatocyte-~timulating Factor by Human Fibroblasts andMonocytes* (Received for publication, October 2,1987)

Lester T. May, John GhrayebS, Uma Santhanam, Stephen B. Tatter, Zev Sthoeger, David C. Helfgott, Nicholas Chiorazzis, Gerd Grieningerll, and Pravinkumar B. SehgalII From The Rockefeller University, New York, New York 10021,SCentocor, Maluern, Pennsylvania 19355, and 7The Lindsley F. Kimball Research Institute of The New York BEood Center, New York, New York 10021

The cDNA for human &-interferon (IFN-B,)/B-cell kDa which co-migrate in polyacrylamide gels with differentiation factor 2/hepatocyte-~timulatingfactor those obtained from FS-4 cells. Our observations help was expressed in Escherichia coli to yield a fusion relate previous descriptions of multiple forms of heto proteins derived protein which contains the182 carboxyl-terminal patocyte-stimulating factor specific amino acids of IFN-& fused to a 34-amino acid pro- from the IFN-B, gene. karyotic leader peptide(rIFN-8,). When added to cultures of human hepatoma cell line Hep3B2, rIFN-/3, as well as preparations of natural IFN-0, enhance secreThe “acute phase” alterations in the composition of blood tion of positive acute phase reactants such as al-antichymotrypsin, complement C3, fibrinogen, and al-acid were, in essence, observed 2500 years ago (1-4). Hippocratic glycoprotein and inhibit secretion of albumin, confirm- medicine (approximately 400 B. C.) wasbased on the principle ing that a protein derived from the IFN-8, gene can that disease was due to the inability of the “four humors” in have hepatocyte-stimulating factor activity. We have the body to mix properly (1-4). The empirical observation prepared a rabbit polyclonal antiserum to theE. coli- was that blood withdrawn from an acutely ill person separated derived human IFN-B2 fusion protein. This polyclonal into a crustainflammatoria or a grayish-white fibrin layer as antiserum inhibits the hepatocyte-stimulating and B- it clotted. The scientific basis for the separation of blood from cell differentiation activities of appropriate IFN-j3, an acutely ill person into a crustainflammatoria or buffy coat preparations.The anti-rIFN-B, antiserumhas been was described by Fahraeusin 1921 (1) and concerns the used in immunoprecipitation experiments and in West- enhancedsedimentation of erythrocytes. It is understood ern blots to help define the secretory proteins derived today that the increase in erythrocyte sedimentation rate from the IFN-82 gene in fibroblasts and monocytes. during acute infections reflects, in large part, a change in “Uninduced” human FS-4 fibroblasts as well as those plasma protein composition (particularly an increase in the induced with interleukin-la, tumor necrosis factor, or fibrinogen level) (2,5). During the period from approximately bacterial lipopolysaccharide secrete at least five forms 1930 to 1950, investigators helped define specific protein of IFN-& of apparent molecular mass in the range from alterations in the sera of host animals duringthe acute phase 23 to 30 kDa which can be resolved by polyacrylamide of different microbial infections (Refs. 6-10; reviewed in Refs. gel electrophoresisunderdenaturingandreducing 11-13). The systemic response to tissue injury caused by conditions. The threehigher molecular mass forms are infection or trauma is referred to as theacute phase response not observed when FS-4 cells are induced in the pres- (8,13-15) and includes alterations in the plasma concentraence of tunicamycin, suggesting that these forms are tions of a large number of proteins secreted by the liver. The N-glycosylated (gp28, gp29, and gp30). Although se- plasma levels of proteins such as C-reactive protein, fibrinocretion of the two lower molecular mass forms is re- gen, a,-antichymotrypsin, a,-acid glycoprotein, complement sistant to tunicamycin, they are labeled by [3H]gluc~- C3, haptoglobin, hemopexin, serum amyloid A protein, cerusamine (gp23-gp25). The inclusion of cycloheximide loplasmin, and many more are increased in response to acute during the [S5S]methioninelabeling of induced FS-4 tissue injury (positive acute phase reactants),whereas plasma cells results in the preferential synthesis and secretion of the 29-kDa triplet.Human monocytes induced with levels of proteins such as albumin, transthyretin,retinoldecreased (negative acute bacterial lipopolysaccharide also secrete several dis- binding protein, and transcortin are tinct forms of IFN-Bz in the size range from 2 3 to 30 phase reactants) (13-15). The possibility that a hormone-like messenger released at the site of infection or injury may be involvedin eliciting the * This work was supported by United States Public Health Service alterations in hepatic gene expression observed during the Research Grants AI-16262 (to P. B. S.), CA-44365 (to P. B. S.), and acute phase has drawn increasing attention (14,16). In recent AI-10811 (to N. C.), a grant-in-aid from the American Heart Association (to G. G.), a contract from the National Foundation for Cancer years, numerous investigators have observedthat interleukinResearch (to P. B. S.), and anEstablished Investigatorship from the l a (IL-la)’ and tumor necrosis factor (TNF), mediators of American Heart Association (to P. B. S.). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. 5 Present address: North Shore University Hospital/Cornell University Medical College, Manhasset, NY 11030. 11 To whom correspondence should be addressed The Rockefeller University, 1230 York Ave., New York, NY 10021.

The abbreviations used are: IL-la, interleukin-la;BSF-2, B-cell differentiation factor 2; HSF, hepatocyte-stimulating factor; IFN, interferon; rIFN-&, recombinant &interferon; LPS, bacterial lipopolysaccharide; TNF, tumor necrosis factor; SDS-PAGE, sodium dodecyl sulfate-polyacrylamide gel electrophoresis; FPLC; fast protein liquid chromatography; DTT, dithiothreitol; MEM, minimal essential medium; NP-40, Nonidet P-40.

7760

Multiple Forms of IFN-P2/BSF-2/HSF

7761

some of the non-hepatic acute phase reactions suchas fever A 1 2 3 4 5 B 1 2 and increased vascular permeability (17, 18), induce only a 92* limited alteration in liver gene expression in hepatocyte cell 66* cultures (reviewed in Refs. 19-21). The major alterations in 4* hepatic gene expression that constitute the acute phase response are attributed to the hepatocyte-stimulatingfactor, a protein of molecular mass approximately 25-30 kDa (estimated bygel filtration) which is usually isolated from the culture medium of bacterial lipopolysaccharide (LPS)-acti14. vated human monocytes (Refs. 22 and 23; reviewed in Ref. 20). The recent observations of Gauldie et al. (20) that antiIFN-8 antisera which react with human IFN-8, (24-33) neuFIG.2. Immunoprecipitation( A )and Western blot( B )analtralize the ability of several different preparations of natural yses of IFN-8, proteins using a rabbit antiserum to E. colihuman hepatocyte-stimulating factor to elicit the synthesis derived rIFN-B2.A, FS-4 cell cultures in T-175 flasks were induced of acute phase proteins in primary rat hepatocyte cultures with IL-ln (1 ng/ml) for 4 h, washed 10 times with phosphateand in human hepatomaG2 cells as well as their observation buffered saline, and then labeled for 2 h with [:%]methionine (100 that a preparation of Escherichia coli-derived IFN-P2/BSF-2 pCi/ml) in methionine-free Dulbecco's medium supplemented with 0.001% (w/v) bovine serum albumin (10 ml/flask). One-ml aliquots (Ref. 34; reviewed in Ref. 29) elicits the major acute phase were lyophilized and immunoprecipitated with preimmune serum response in hepatocytes led to the conclusion that the hepa- (lane I ) ; immune serum, bleed 2 (lanes 2 and 3 ) ;or immune serum, tocyte-stimulating factor, the major mediator of the hepatic bleed 3 (lanes 4 and 5). The immunoprecipitates were analyzed by acute phase response, is derived from the IFN-P, gene. SDS-PAGE (17.5%)and autoradiography. R , FS-4 cell cultures in TInterest in the IFN-P, gene originated with the observation 175 flasks were induced with IL-la as described for A except that the that a poly(1) poly(C)-inducible 1.3-kilobase mRNA prepa- induced cells were allowed to secrete unlabeled IFN-j3, into the culture medium for 14-16 h. A 1-ml aliquot was lyophilized and used for the ration could yield a translation product (in Xempus laevk Western blot analysis (lane 2) using anti-rIFN-8, immune serum, oocytes) which had antiviral activity of the IFN-P serology bleed 3. The E. coli-derivedrIFN-B, was included in the Western blot (24, 25). In contrast to the IFN-8, gene, whichisfree of analysis as a control (lane I ). introns andresides onhuman chromosome 9,the IFN-B, gene contains four introns (26,27) and has been localized to human (also see below). The detection of this heterogeneity is in chromosome7(28). Proteins variouslydescribed as B-cell contrast to theprevious descriptions of a single IFN-B,/BSFstimulatory factor 2, hybridoma/plasmacytoma growth factor, 2 protein of 21 kDa (34) or that of two proteins of 23 and 26 26-kDa protein, and interleukin-6 are also derived from the kDa (35,361. same gene (Refs. 34-37; reviewed in Ref. 29). The amino acid The anti-rIFN-& antiserumcould also be used in Western sequence deduced from the cDNAclonereveals an open blot analyses of rIFN-P, (Fig. 2B, lane 1 ) as well as in those reading frame of 212 residues with two potential N-glycosyl- of natural IFN-P, polypeptides secretedby IL-la-induced FSation sites (26, 30, 31, 34). 4 cells (Fig. 2B, lane 2 ) at dilutions in the range from 1:250 We have expressed human IFN-P, cDNA in E. coli and to 1:10,000. Heterogeneity of IFN-Pz polypeptides secretedby have confirmed that rIFN-8, has hepatocyte-stimulating ac- FS-4 cells isalsodetected in Western blots. In additional tivity. Furthermore, a rabbit polyclonal antiserum prepared experiments, we have observed that E. coli-derived rIFN-P, against the E. coli-derived protein rIFN-P2 neutralized not blocks the immunoreactivity in Western blots of all of the only the hepatocyte-stimulating activity but also the B-cell species observed in FS-4 culture medium. This antiserum differentiation activityof appropriate preparationsof IFN-8,. does not immunoprecipitate IFN-8, or recognize rIFN-P, in a We have used this polyclonal antiserum in immunoprecipi- Western blot (not shown). tation and Western blot experiments to detect and characterHepatocyte-stimulating Activity of the rIFN-8, Preparation ize a complex series of at least five proteins in.the size range and Its Inhibition by the Anti-rIFN-P, Antiserum-The hefrom 23 to 30 kDa derived from theIFN-8, gene and secreted patocyte-stimulating activity of the rIFN-B, preparation illusby human fibroblasts and monocytes induced by various cytrated in Fig. lA (fraction 23) was investigated in Hep3B2 tokines andby LPS. These experiments provide a description cultures. Fig. 3 summarizesrocketimmunoelectrophoretic of the synthesis and secretion of a complex set of proteins analyses which show that rIFN-8, increased the accumulation that mediate themajor hepatic acute phaseresponse in man. of fibrinogen and complement C3 in the culture medium of Hep3B2 cells while decreasing the accumulation of albumin MATERIALS AND METHODS~ (compare lanes 1 and 2 in each case), The inclusion of antirIFN-& immune serum (lanes3 and 5)in the culture medium, RESULTS but not that of control rabbit serum(lanes 4 and 6),inhibited Rabbit Polyclonal Antiserum to rIFN-P,-The ability of the increase in accumulation of complement C3 in response rabbit antiserum to rIFN-8, to immunoprecipitate [:"SS]me- to rIFN-8,. Anti-rIFN-P, also inhibited theincrease in fibrinthionine-labeled IFN-P, proteins from IL-la-induced FS-4 ogen accumulation in response to rIFN-P2 (not shown). Fig. 4 cell culture medium is illustrated inFig. 2A, which shows that summarizes a [:"SS]methionine labeling/immunoprecipitation the postimmune, but not the preimmune, rabbit sera immu- analysis showing the increase in synthesis of a,-antichymonoprecipitate a series of polypeptides in the size range from trypsin and al-acid glycoprotein and thedecrease in synthesis 23 to 30 kDa from the culture medium of induced FS-4 cells of albumin in response to a purified preparation of natural IFN-P, and in response to rIFN-P, (fraction 23 of Fig. L4). Portions of this paper (including "Materials and Methods" and (LPS by itself does not alter plasmaprotein synthesis or Fig. 1)are presented in miniprint at the end of this paper. Miniprint is easily read with the aid of a standard magnifying glass. Full size secretion by hepatocytes (22, 23).) We have observed that photocopies are included in the microfilm edition of the Journal that immunoaffinity-purified or fast protein liquid chromatograis available from Waverly Press. phy-purified preparations of rIFN-& as well as additional e

Multiple Forms of IFN-P,/BSF-B/HSF

7762 Fib

AI b =*

c3

TABLE I

cc

Inhibitory effect of anti-rIFN-& immune serum on differentiation of CESS cells into I& secretors Plaque-forming cells/culture" Additions to CESS cells Exp. 1

Exp. 2

additions No 76 (8) 144 (9) TNFand CHXb-induced FS-4 cell 124 (7) 359 (35) supernatant +Preimmune serum (1:lOO) (30) 178 325 (52) 1 2 1 2 1 2 3 4 5 6 +Preimmune serum (1:lOOO) 174 (9) 391 (18) FIG. 3. Hepatocyte-stimulating activity of rIFN-p,. Cultures +Immune serum (1:lOO) 74 (13) 192 (25) of human Hep3B2 cells in 24-well plates were treated for 72 h with +Immune serum (1:lOOO) 138 (13) 398 (42) rIFN-8, (fraction 23 in Fig. lA dialyzed against phosphate-buffered "CESS cells (lo') were cultured in 0.2 ml of medium containing saline; 300 ng/ml, 0.5 ml/well) in the presence or absenceof a 1:lOO the various additions for 60 h, a t which time the number of IgC dilution of anti-rIFN-8, immune serum or control rabbit serum. The secreting cells per culturewas determined using the reverse hemolytic accumulation of fibrinogen (Fib), albumin (Alb), and complement C3plaque assay. Each variable was assayeda t least intriplicate; numbers ( C 3 )in the culturemedium was then assayed by rocket immunoelecindicate mean (standard errorof mean). trophoresis. Each experiment was performed in triplicate; the entire * CHX, cycloheximide. experiment was repeated three times. The figure illustrates representative results from one experiment. Lane I , untreated control; lane 2, rIFN-B,; lane 3, rIFN-8, with immune rabbit serum; lane 4, rIFN& with control rabbit serum; lane 5, immune rabbit serumalone; lane 6, control rabbit serum alone.

1

FIG.4. Hepatocyte-stimulatingactivity

of naturaland

rIFN-p,. Cultures of Hep3B2 cells in T-75 flaskswere treated with a purified preparation of natural IFN-& proteins (DEAE-Sephacel purification; 20 ng/ml) or rIFN-& (fraction 23 in Fig. lA dialyzed against phosphate-buffered saline;200 ng/ml) for 24 h. The cells were then washed, and the cultureswere labeled with ["S]methionine (100 pCi/ml, 3 ml/flask) in methionine-free Dulbecco's medium for another 24 h. Appropriate aliquots of the culture medium were then immunoprecipitated using antisera specific for human a,-antichymotrypsin(ACT),a,-acid glycoprotein (AGP), or albumin (Alb), followed by SDS-PAGE and autoradiography. Duplicate aliquots of culture medium were analyzed for a,-antichymotrypsin and a,-acid glycoprotein, and single aliquots for albumin. Croup 1 , untreated Hep3B2 control; group 2, natural IFN-8, added; group 3, rIFN-& added.

2

3

4

5

6

7

8

FIG. 5. Multiple forms of IFN-B, proteins induced by TNF or IL-lain humanfibroblasts. FS-4 cells in T-175 flasks (one per group) were induced with T N F (10 ng/ml) or IL-la (1ng/ml) for 4 h in the presence or absence of tunicamycin, washed extensively, and then labeled with [:"S]methionine for 2 h as described for Fig. 2 . 4 ; and 1-ml aliquots of the culture medium were then immunoprecipitated using anti-rIFN-& serum. Lanes 1 and 2, uninduced controls; lane 3, T N F alone; lane 4, T N F with tunicamycin (0.8 pg/ml); lane 5, T N F with tunicamycin (1.6 pg/ml); lane 6, IL-la alone; lane 7, IL-la with tunicamycin (0.8 pg/ml), lane 8, IL-la with tunicamycin (1.6 pglml). Insets show laser densitometric scans of the respective areas of the original autoradiograms (drawn to different scales; the left inset is at a scale enlarged five times compared to theright inset). A clear separation was observed between peaks 2 and 3 in shorter exposures of the autoradiogram inlanes 3 and 6.

mediumfrom TNF- and cycloheximide-induced FS-4 cells preparations of natural IFN-@, also elicit the enhanced syn- which contain IFN-@,. In additional experiments, we have thesis of a,-macroglobulin and a-thiol proteinase inhibitor in observed that the anti-rIFN-@, antiserum in dilution the range Hep3B2 cells. Furthermore, the increases elicited by natural from 1:32 to 1:256 inhibits the proliferation of the murine IFN-8, preparations can also be inhibited by the anti-rIFN- plasmacytoma cell line T1165 in response to natural IFN-@, 8, antiserum. present in the culturemedium of TNF-, cycloheximide-, and actinomycin D-induced FS-4 cells. The anti-rIFN-@, antiGel-purified, electrophoretically homogeneous rIFN-@, preparations also enhanced a,-antichymotrypsin secretion by serum also neutralizes the ability of electrophoretically hoHep3B2 cells (not shown). The specific activity of gel-purified mogeneous monocyte-derived IFN-/3, species to enhance prorIFN-8, in this assay is estimated be greater to thanlo6units/ liferation of human B-cells recently transformed by Epsteinmg of protein. Barr virus (51). Inhibition of B-cell Differentiation by Anti-rIFN-@, AntiMultiple Forms of IFN-@,Secreted by Human Fibroblastsserum-The rabbit polyclonal anti-rIFN-@, antiserum inhib- FS-4 cell cultures induced with I L - l a for 4 h were labeled ited the abilityof culture medium from TNF- andcyclohexi- with [:'%]methionine for 2 h, and the secreted, labeled promide-induced FS-4 cells (40, 41, 44) to enhance the differen- teins were immunoprecipitated using the anti-rIFN-@, serum tiation of B-lymphoblastoid CESS cells into immunoglobulin- (Fig. 5). Fig. 5 (lanes 1 and 2) shows that control"uninduced" secreting cells. In a series of four experiments, two of which FS-4 cultures do synthesize detectable levels of IFN-@, proare illustrated in Table I, immune serum (at a dilution of teins immunoprecipitable with anti-rIFN-@,antibodies. This 1:lOO) inhibited by 40-50% the increase in the numbers of observation is consistent with the detection of IFN-8, mRNA immunoglobulin-secreting CESS cells in response to culture in uninduced FS-4 cell cultures (40, 41). Fig. 5 (lanes 3-8)

Multiple Forms of IFN-P2/BSF-2/HSF

7763

represents immunoprecipitatesof TNF- or IL-la-induced FS1 2 3 4 cell culture media using the anti-rIFN-P, antiserum.Fig. 5 (lanes 3 and 6) shows a complexset of IFN-P2 formsof which five can be resolved clearly in these samples (also see Figs. 6 and 9). Thus, the major forms of IFN-P2 secretedby cytokineinduced FS-4 cells consist of a triplet a t approximately 29 kDa and at least a doublet a t approximately 25 kDa. The synthesis and secretion of the upper triplet are inhibited in the presence of tunicamycin (compare lanes 4 and 5 and 7 and 8 with 3 and 6 in Fig. 5);butthe lower doubletis tunicamycin-resistant, suggesting that the three higher molecular mass forms contain N- or asparaginyl-linked carbohydrate (gp28, gp29, and gp30), whereas thetwo lower molecular mass forms do not. Amore extensivesampling of ourimmunoprecipitation kD experimentsissummarizedin Fig. 6. Fig. 6A shows that multiple IFN-8, species are observed following immunopre4 31 cipitation of the IL-la-induced FS-4 cell culture medium even when the protein samples are boiled for up to 10 min under reducing conditions priorto gel electrophoresis andwhen they 4 21 are run through a different gel system (urea-polyacrylamide) than used in Fig. 5 (SDS-polyacrylamide). A variety of different induction conditions lead to the increased secretion of multiple forms of IFN-P2. Fig. 6B illustrates the IFN-P, proteins detected in FS-4 cell cultures induced with TNF, IL-la, or LPS. In each instance, the inclusion of tunicamycin inhibits the productionof the upper triplet(Fig. 6 C ) . The conclusion that the tunicamycin-sensitive28-30-kDa M G G IFN-8, species are glycosylated in a manner different from the tunicamycin-resistant23-25-kDa species is confirmed by FIG. 7. IFN-& proteins secreted by induced FS-4 cells lathe observation that the 28-30-kDa proteins bind to a con- beled with ["H]glucosamine.SDS-PAGE is shown of radioactively canavalin A-Sepharose column (and are eluted off with a- labeled IFN-B, proteins immunoprecipitated by the anti-rIFN-B, imA

1

2

3

4

mune serumfrom the medium of FS-4 cell cultures that were induced with IL-la for 14 h in the continuous presence of ~-[6-'H]glucosamine ( G ; lanes 2 and 3 ) or ["S]methionine ( M ; lane 1 ) . The fluorogram was exposed for 30 h.

5

-31 -21

methyl mannoside), whereas the 23-25-kDa species fail to bind such a column.3 However, the 23-25-kDa secreted IFNP2 species are notdevoid of carbohydrate because they canbe -31 labeled using ["H]glucosamine as the radioactive precursor --- -21 (Fig. 7). The labeling of the faster mobility species of IFN-8, by glucosamine together with their tunicamycin resistance C l 2 3 4 5 6 7 8 suggests that these glycoproteins do not contain N-linked sugars. Whether this ["H]glucosamine label is actually incor-31 porated into the 23-25 kDa IFN-P, glycoproteins as glucosa"--mi= mine or its rapidly isomerized form (galactosamine) remains -21 to be investigated. The possibility that the observed heterogeneity of IFN-8, FIG.6. Multiple forms of IFN-& proteins secreted by human fibroblasts under different experimental conditions. A, proteins may be due to degradation of the secreted proteins IFN-6, proteins prepared as described for Fig. 2A were immunopre- was tested in several pulse-chase experiments in which ILcipitated with preimmune ( l a n e 1 ) or anti-rIFN-0, immune(lanes 2la-induced FS-4 cells were labeled with ["'S]methionine for 5) serum, solubilized in sample buffer, and then heated for 1 min different lengths of time and then chasedwithunlabeled ( l a n e I ) , 0 min ( l a n e 2), 1 min ( l a n e 3 ) , 5 min ( l a n e 4 ) , and 10 min methionine. Multiple forms of IFN-P, were detectable in the ( l a n e 5). Electrophoresis was carried out through a urea-polyacrylculture medium at the earliest time that such proteins could amide gel. B, IFN-j3, proteins labeled with ["S]methionine as described for Fig. 2.4 following induction of FS-4 cell cultures with IL- be seen (10-min pulse),and thebasic pattern of polypeptides la and cycloheximide (lanes 1 and 2), IL-la alone ( l a n e 4 ) , LPS persisted throughout the pulse-chase experiment (5-min pulse alone (lane 5),o r T N Falone (lane 6) were immunoprecipitated using followed by a 5-h chase),suggesting that thevarious observed the anti-rIFN-j3, immune serum followed by electrophoresis through forms of IFN-P, represent distinct primary secretion products a urea-polyacrylamide gel. Lune 3 illustrates proteins secreted by derived from this gene (data not shown). C, IFN-P, proteins labeled with ["SI uninducedcontrolcultures. Alterations in the Relative Composition of IFN-8, Forms methionine as described for Fig. 24 following induction of FS-4 cell cultures with IL-la (lanes 6-8),T N F (lanes 2 and 3 ) ,or LPS (lanes Secreted by Human Fibroblasts-The relative proportions of 4 and 5 ) in the presence (lanes 2,4, and 6 ) or absence (lanes 1 , 3 , 5 , the different IFN-P, forms synthesized and secreted by human 7,and 8 ) of tunicamycin (1 pg/ml) were immunoprecipitated using fibroblasts can be altered experimentally. Fig. 8 illustrates an the anti-rIFN-j3, immune serum followed by SDS-PAGE and auto- experiment in which FS-4 cells induced with IL-1 or T N F radiography. Lane 1 illustrates proteins secreted by uninduced conB

1

2

3

"

-"

4

5

6

-

c

trols, and lane 8 by cultures treated with IL-la in the presence of cycloheximide.

U. Santhanam et al., manuscript in preparation.

7764

Multiple F o r m s of IFN-&/BSF-B/HSF 1

2

3

4

5

6

A 0

-

4

I

,

,

,

,

31 21

,

inating LPS (endotoxin). This possibility has been confirmed in additional experiments in which monocytes were cultured in media low in LPS. Under these experimental conditions, there is no detectable synthesis andsecretion of IFN-& species. The overall pattern of IFN-8, forms secretedby induced monocytes is similar to that obtained from equivalently induced fibroblasts (compare lanes 1 and 2 in Fig. 9).

B DISCUSSION

We have expressed human IFN-B2/J3SF-2/HSF as a fusion rabbit polyclonal proteinin E. coli and haveprepareda antiserum to this E. coli-derived protein. In this paper, we report that this anti-rIFN-B2 antiserum inhibits the hepatocyte-stimulating andB-cell differentiation activities of appropriate IFN-& preparations. Using this antiserum in immunoprecipitationandimmunoblottingexperiments, we ob-21 served acomplex series of IFN-8, forms in thesize range from FIG.8. Relative composition of IFN-8, forms secreted by 23 to 30 kDa secreted by appropriately induced human fibrohuman fibroblasts can be altered experimentally. FS-4 cell blasts and monocytes. The secretion of the tripletof proteins cultures in T-175flasks were labeled with [RsS]methioninein the at approximately 29 kDa was inhibited by tunicamycin, sugabsence of any added inducer ( l a n e I ) or in the presence of IL-la gesting that these IFN-8, forms are N-glycosylated (gp28, alone ( l a n e Z),IL-la and cycloheximide ( l a n e 3 ) , TNF alone ( l a n e gp29, and gp30); whereas that of the doublet of proteins a t 25 4 ) , TNF and cycloheximide ( l a n e 5),or cycloheximide alone ( l a n e 6 ) . One-ml aliquots of medium were withdrawn from each culture at 1 h kDa was resistant to tunicamycin, although they could be ( A ) , 5 h ( B ) ,or 24 h (C)after the beginning of the induction and labeled with [3H]glucosamine (-23-25). Furthermore, we labeling. IFN-& forms synthesized and secreted by the FS-4 cells have observed that the relative composition of IFN-8, forms were characterized by immunoprecipitation with the anti-rIFN-& secreted by human fibroblasts can be experimentally shifted immune serum followedby SDS-PAGE and autogradiography. in favor of the triplet 29 at kDa by partially inhibitingcellular protein synthesis (such as mightoccur in injured or infected 1 2 3 4 tissues). Thus, the synthesis and secretion of multiple IFN,T2 forms do not appear to be passive processes but are likely to be an actively regulated seriesof post-translational events. *31 The immunological reagents developed by us have led to rapid progress in purification of individual IFN-8, forms prelimi-21 nary to an investigation of the precise chemical structures and biological activities of each of the secreted forms of IFNFIG.9. Multiple forms of IFN-8, secreted by activated human monocytes. Adherent human monocytes (7 X lO‘/group) were induced for 4 h with IL-la ( l a n e 2) or LPS ( l a n e 3 ) or were left uninduced ( l a n e 4 ) and were then labeled for 2 h with [3sSS)methionine. IFN-8, proteins secreted into the culture medium were analyzed by immunoprecipitation using the anti-rIFN-& immune serum and SDS-PAGE and autoradiography. For comparison,lane I shows the immunoprecipitation of IFN-P, proteins in IL-la-induced FS-4 culture medium.

82-

The discrepant observations on the antiviral activity of recombinant IFN-P, species produced by different investigators (26, 34, 35) may be resolved through the determination of the biological properties of the different forms of IFN-8,. Revel and colleagues (26, 32, 33, 52) have expressed IFN-& cDNA and genomic DNA clones in mammalian cell expression systems toyield recombinant IFN-8, with high antiviral activity and low B-cell activity. Hiranoet al. (34)4 have were labeled with [3sS]methionine in thecontinuous presence expressed IFN-&/BSF-2 cDNA in transient expression sysof cycloheximide. It is clear that the inclusion of cyclohexi- tems as well as in E. coli in a manner that yields what they of call the mature21-kDa protein andhave observed no antiviral mide during the experiment greatly enhances the secretion the IFN-8, tripleta t 29 kDa. Fig. 8C (lanes5 and 6)illustrates activity, but have observed high B-cell activity. Their E. coli particularly clear examples of the resolution of the three preparation is alsoactive in the hepatocyte stimulation assay individual components a t 29 kDa. Whereas the biochemical (20). However, our rIFN-P2fusion protein preparation elicits basis for the observed shift in the composition of IFN-8, an antiviralstate (53), stimulateshepatocytes, and hasB-cell forms is not understood, it is likely to include not only the activity. T o be able to reconcile these different observations, ability of cycloheximide to inhibit the rate of protein synthesis it will be necessary to establish the identity of the mature21(approximately 40-fold inhibition at 50 pg/ml), but also its kDa form of Hirano et al. (34) within the series of IFN-& ability to greatly enhance (“superinduce”) the IFN-8, mRNA forms that we detect. It will also be important to determine content in FS-4 cells (approximately 300-fold increase; Refs. the biological properties of multimers of IFN-8, which we 40-42). have detected in the 50-70-kDa range;’ Thus, the precise Multiple Forms of IFN-& Secreted by Human Monocyteschemical structure of recombinant IFN-8, expressed by difMost investigators have prepared hepatocyte-stimulating fac- ferent investigators using differentexpression systems needs tor by inducing human monocyteswith LPS (22, 23). We to beelucidated in order to determine whether biological have investigatedIFN-j3, forms secretedby uninduced human observations reported by the different laboratories refer to monocytes and following induction of monocytes with LPS the same or different forms of IFN-P2. or IL-la (Fig. 9). It is clear thatinduced monocytes secrete Baumann etal. (54) haveobserved differences in theability multiple forms of IFN-B, (Fig. 9, lanes 2 and 3). Uninduced of hepatocyte-stimulating preparationsof different gel filtramonocytes also produce significant amounts of IFN-& (Fig. 9, lane 4 ) presumably because the culture media used in this ‘Hirano, T., Matsuda, T., Hosoi, K., Okano, A., Matsui, H., and experiment already contained sufficient amounts of contam- Kishimoto, T. (1988)Immunol. Lett. 17,41-45.

Multiple Forms of IFN-P2/BSF-2/HSF

7765

tion properties to enhance secretion of different plasma pro-

11. Dubos, R. J. (1976) The Professor, the Institute, and DNA, The Rockefeller

b i n s from hepatocytes (see Fig. 4 in Ref. 54). In Reuber H35 rat hepatoma cells, the 5°-70-kDa HSF produced by COL016 cells stimulates production of cysteine proteinase inhibitor and a,-macroglobulin without significant effects on secretion of complement C3, haptoglobin, and a,-acid glycoProtein. On the other hand, the 25-30-kDa HSF stimulates complement C3,haptoglobin, and al-acid glycoprotein secretion andhas littleeffect on secretion of a,-macroglobulin and cysteine proteinase inhibitor. Whereas it is not clear whether the properties of the 50-70-kDa material studiedby Baumann et al. (54) can be assigned to products of the IFN-P, gene, we have, in preliminary experiment^,^ observed that theindividual separated forms of IFN-P, have different quantitative effects on hepatocyte plasmaproteinsynthesis. The outstanding physiological significance of IFN-& as an endogenous cellular regulatory mOleCUle Stems from the fact that itpossesses hepatocyte-stimulating activity and appears to be the major mediator of the acute phase response. The induction Of IFN-& by Other acute phase cytokines such as TNF and IL-la in a variety of different tissues ensures that IFN-& is in any cascade Of events during the acute phase. The ability of platelet-derived growth factor to induce IFN-B, in stromal fibroblasts also ensures the production of IFN-6, in and around sites of thrombosis in traumatized tissues. The Of products such as Lps toxin) to directly induce IFN-0, secretion not only from monocytes but also from the ubiquitous stromal fibroblast ensures the production of IFN-P, by tissuesharboring microbial infections. The ability of viruses such as Sendai virus to greatly enhance IFN-P, secretion' ensures the production of 1FN-b during viral infections. From our perspective, it aPpears that the elementalproperty of polypeptides derived from the IFN-P, gene is likely to be hormone-like communication between damaged tissues and thehepatocyte. We suggest that this mediator be designated "hippocratin" in honor ofthe signal achievement ofthe Greek physicians of antiquity who described the acute phase alterations in the properties of the blood.

12. McCarty, M. (1982) Ann. N. Y. Acad. Sci. 389,1-10 13. Kushner, I. (1982) Ann. N. Y.Acad. Sci. 3 8 9 , 39-48 14. Koj, A. (1974) in Structure and Function of Plasma Proteins (Allison, A. C., ed) Vol. 1, pp. 73-131, PlenumPress, New York 15. Gauldie, J., Lamontagne, L., and Stadnyk, A. (1985) Inf[ammtion 4,126151 16. Kampschmidt, R. F., Upchurch, H. F., and Pulliam, L. A. (1982) Ann. N . Y. Acad. Sci. 389,338-353 17. Beutler, B., and Cerami, A. (1986) Nature 320,584-588 18. Dinarello, C. (1984) Rev. Infect. &. 6 5 1 - 9 5 19. Evans, E., Courtois, G. M., Killian, P. L., Fuller, G. M., and Crabtree, G. R. (1987) J. Biol. Chem. 262,10850-10854 20. Gauldie, J., Richards, C., Hamish, D., Lansdorp, P., and Baumann, H. (,l987) Proc. Natl. Acad. Sei. U. S. A. 8 4 , 7251-7255 21. KO], A,, Kurdowska, A,, Magielska-Zero, D., Rokita, H.,Sipe, J. D., Dayer, J. M., Demczuk, S., and Gauldie, J. (1987) Bimhem. Int. 14,553-560 22. Ritchie, D. G., Levy, B. A., Adams, M. A., and Fuller, G. M. (1982) Proc. Natl. Acad. Sei. U. S. A. 79,1530-1534 23. Ritchie, D.G., andFuller, G. M. (1983) Ann. N . Y. Acad. Sei. 4 0 8 , 490502 24. Weissenbach, J., Chernajovsky, Y., Zeevi, M., Shulman, L., Soreq, H., Nir, U., Wallach, D., Perricaudet, M., Tiollais, P., and Revel, M. (1980) Proc. Natl. Acad. Sci. U. S. A. 7 7 , 7152-7156 25. Sehgal, P. B., and Sagar, A. D. (1980) Nature 288,95-97 26. Zilherstein, A., Ruggieri, R., Kom, J. N., and Revel, M. (1986) EMBO J . 6 , 2529-2537 27. Yasukawa, K., Hirano, T., Watanabe,Y., Muratani, K., Matsuda, T., Nakai, S., and Kishimoto, K. (1987) EMBO J. 6,2939-2945 28. Sehgal, P. B., Zilberstein, A,, Ruggieri, R., May, L. T., Ferguson-Smith, A,, Slate, D. L., Revel, M., and Ruddle, F. H. (1986) Proc. Natl. Acad. Sci. U. S. A. 8 3 , 5219-5222. 29. Sehgal, P. B., May, L. T., Tamm, I., and VilEek, J. (1987) Science 2 3 6 , 731-732 30. Haegeman, G., Content, J., Volckaert, G., Derynck, R., Tavemier, J., and Fiers, W. (1986) Eur. J. Biochern. 159,625-632 31. May, L. T., Helfgott, D.C., and Sehgal, P. B. (1986) Proc. NatLAcad. Sci. U. S. A. 83,8957-8961 32. Zilherstein, A., Ruggieri, R., and Revel, M. (1985) Seronn Symp. Publ. Raven Press 24.73-83 33. Revel, M., Rwgieri, R., and Zilberstein, A. (1986) in The Biology of the Interferon System 1985 (Schellekens, H., and Stewart, W. E., 11, e&) pp. 119-124, Elsevier/North-Holland Biomedical Press, Amsterdam 34. Hirano, T., Yasukawa, K., Harada, H., Taga, T., Watanabe, Y.,Matsuda,

Ackmwkdgments-We thank Dr.Igor Tamm for his enthusiastic support and numerous helpful discussions. We also thank Robert Clarick for excellent technical assistance. REFERENCES 1. FHhraeus, R. (1921) Acta. Med. Scand. 65, 1-228 2. Wintrobe, M. M. (1974) Clinical Hematology, pp. 125-128, Lea and Febiger,

Philadelphia 3. Adams, F. (1929) The Genuine Works of Hippocrates, William Wood & Co.,

New York

4. Jones, W. H. S. (1923) Hippocrates (an English Translation of the Origiml Greek, from Nature of Man) Vol. 4, pp. 1-41, Harvard University Press,

Cambridge

5. Bedell, S. E., andBush, B. T. (1985) Am. J . Med. 78,1001-1009 6. Tillett, W. S., and Francis, T., Jr. (1930) J . Exp. Med. 52, 561-571 7. Abernethy, T. J., and Avery, 0.T. (1941) J. Exp. Med. 7 3 , 173-182 8. MacLeod, C. M., and Avery, 0. T. (1941) J. Exp. Med. 7 3 , 183-190 9. MacLeod, C. M., and Avery, 0.T. (1941) J . Exp. Med. 73,191-200 10. Anderson, H.C., and McCarty, M. (1950) Am. J. Med. 8,445-455 5 Sehgal, p, B,,~ ~ lD. fc., santhanam, ~ ~ ~ u., ~ Tatter, , S. B., Clarick, R. H., Ghrayeb, J., and May, L. T. J. Exp. Med., in press.

University Press, New York

T., Kashiwamura, S., Nakajima, K., Koyama, K., Iwamatsu, A., Tsunasawa, S., Sakiyama, F., Matsui, H., Takahara, Y., Taniguchi, T., and Kishimoto, T. (1986) Nature 324,73-76 35. Poupart, P., Vandenabeele, P., Cayphas, S., Van Snick, J., Haegeman, G., Kruys, V., Fiers, W., andContent, J. (1987) EMBO J. 6,1219-1224 36. Content, J., De Wit, L., Pierard, D., Derynck, R., De Clercq, E., and Fiers, W. (1982) Proc. Natl. Acad. Sci. U. S. A. 79,2768-2772 37. Van Damme, J. M., Opdenakker, J. R., Simpson, M. R., Rubira, S., Cawhas, A., Vink, A,, Billiau, A,, and Van Snick, J. (1987) J. Exp. Med. 1 6 6 ,

914-922 38. Wong-Staal, F., Chanda, P. K., and Ghrayeh, J. (1987) AIDS Res. Hum. Retroviruses 3 , 33-39 39. Franchini, G., Robert-Guroff, M., Wong-Staal, F., Ghrayeb, J., Kato, I., Chang, T.W., and Chang, N. T. (1986) Proc. Natl. Acad. Sci. U. S. A. 83,5282-5285 40. Kohase, M., Henriksen-DeStefano, D., May, L. T., VilEek, J., and Sehgal, P. B. (1986) Cell 45,659-666 41. Kohase, M., May, L. T., Tamm, I., VilEek, J., and Sehgal, P. B. (1987) Mol. Cell. Biol. 7 , 273-280 42. Walther, Z., May, L. T., and Sehgal, P. B. (1987) J. Immunol. 1 4 0 , 974917 43. Sehgal, P. B., Walther, Z., and Tamm, I. (1987) Proc. NatL Acad. Sci. U. S. A. 84,3663-3667 44. Helfgott, D. C., May, L. T., Sthoeger, Z., Tamm, I., and Sehgal, P. B. (1987) J. Exp. Med. 1 6 6 , 1300-1309 45. Boyum, A. (1968) Scand, J. Lab. Clin. Invest. 2 1 , Suppl. 97, 77-89 46. Brodersen, M.Burns, P., and P. (1974) Proc. Soc. Exp. Biol. Med. 1 4 4 , 941-944 47. Grieninger, G., Pindyck, J., Hertzberg, K. M., and Mosesson, M. W. (1979) Ann. Clin. Lab. Sci. 9, 511-517 48. Gronwicz, E., Coutinho, A., and Melchers, F. (1976) Eur. J. Immunol. 6 , 588-590 49. Laemmli, U. K. (1970) Nature 227,680-685 50. Towbin, H., Staehelin, T., and Gordan, J. (1979) Proc. Natl. A c d . Sci. U. S. A. 76,4350-4355 51. Tosato, G., Seamon, K. B., Goldman, N. D., Sehgal, P. B., May, L. T., Washington,G. C., Jones, K.D., andPike, S. E. (1988) Science 2 3 9 , 502-504 52. Revel, M., Zilherstein, A., Ruggieri, R. M., Ruhinstein, M., and Chen, L. (1987) J . Interferon Res. 7 , 529-536 53. SehgaL P. B., and May. L. T. (1987) J . Interferon Res. 7,521-527 54. Baumann, H., Onorato, V., Gauldie, J., and Jahreis, G. P. (1987) J. Biol. Chem. 262,9756-9768

Multiple Forms of IFN-P2/BSF-2/HSF

-

"

"-

-.

""-

" "

\

\

I-

e-

rn

rn

-