Expression cloning of an interferon-inducible 17-kDa membrane protein implicated in the control of cell growth.

Interferon-inducible membrane proteins of approximately 17 kDa have been suggested to play a role in the antiproliferative activity of interferons based on (1) their pattern of induction in interferon-sensitive and -resistant cell lines and (2) the ability of a membrane fraction enriched in 17-kDa proteins to inhibit cell growth. To gain insight into the nature of the proteins that mediate the antiproliferative activity of interferons, a monoclonal antibody, 13A5, was generated that reacted specifically with a 17-kDa interferon-inducible cell surface protein. The expression pattern of this 17-kDa protein by human cell lines correlated with sensitivity to the antiproliferative activity of interferons. To obtain information regarding the structure of this protein, the 13A5 antibody was used to screen COS cells transfected with a human cDNA expression library. Sequence analysis of a full-length cDNA clone revealed identity with the 9-27 cDNA, previously isolated on the basis of its interferon inducibility by differential screening. In addition, the 17-kDa protein encoded by the 9-27 gene was shown to be identical to the Leu-13 antigen. Leu-13 was previously identified as a 16-kDa interferon-inducible protein in leukocytes and endothelial cells and is a component of a multimeric complex involved in the transduction of antiproliferative and homotypic adhesion signals. These results suggest a novel level of cellular regulation by interferons involving a membrane protein, encoded by the interferon-inducible 9-27 gene, which associates with other proteins at the cell surface, forming a complex relaying growth inhibitory and aggregation signals.

Interferon-inducible membrane proteins of approximately 17 kDa have been suggested to play a role in the antiproliferative activity of interferons based on (1) their pattern of induction in interferon-sensitive and -resistant cell lines and (2) the ability of a membrane fraction enriched in 17-kDa proteins to inhibit cell growth. To gain insight into the nature of the proteins that mediate the antiproliferative activity of interferons, a monoclonal antibody, 13A5, was generated that reacted specifically with a 17-kDa interferon-inducible cell surface protein. The expression pattern of this 17-kDa protein by human cell lines correlated with sensitivity to the antiproliferative activity of interferons. To obtain information regarding the structure of this protein, the 13A5 antibody was used to screen COS cells transfected with a human cDNA expression library. Sequence analysis of a full-length cDNA clone revealed identity with the 9 -27 cDNA, previously isolated on the basis of its interferon inducibility by differential screening. In addition, the 17-kDa protein encoded by the 9 -27 gene was shown to be identical to the Leu-13 antigen. Leu-13 was previously identified as a 16-kDa interferoninducible protein in leukocytes and endothelial cells and is a component of a multimeric complex involved in the transduction of antiproliferative and homotypic adhesion signals. These results suggest a novel level of cellular regulation by interferons involving a membrane protein, encoded by the interferon-inducible 9 -27 gene, which associates with other proteins at the cell surface, forming a complex relaying growth inhibitory and aggregation signals.
Interferons (IFN) 1 are multifunctional cytokines that play a critical role in the defense against viral or parasitic infections. These cytokines also exhibit antiproliferative and differentiat-ing activities, prompting an evaluation of their potential as antitumor agents. While the results of animal testing have been encouraging, clinical trials in humans have shown IFNs to be effective for only a small subset of cancer types (1). This dramatic host difference in susceptibility to the antitumoral activity of IFNs underscores the need to understand at the molecular level the mechanism by which IFNs exert this activity. Exposure to IFNs leads to a modulation in the levels of an estimated 50 -100 cellular proteins, which are thought to collectively mediate IFNs pleiotropic effects (2). Identifying the role and contribution of each protein is a major challenge, but it will provide the basis for designing better therapeutic strategies based on IFNs.
In addition to soluble cytokines and growth factors, cell-cell interactions involving specific membrane proteins are likely to play an important role in the control of cell growth and differentiation. Indeed, it has been shown that the antiproliferative activity of IFN could be transferred from cell to cell (3). Moreover, it has been suggested that IFN-induced cell surface proteins of approximately 17 kDa could be involved in their antiproliferative activity, since their expression pattern in various cell lines correlates with the sensitivity to the inhibition of cell growth induced by IFNs (4 -6), and a membrane fraction enriched for 17-kDa proteins inhibits cell growth (7).
The purpose of the present study was to investigate the possible contribution of IFN-inducible cell surface proteins to IFNs antiproliferative activity. Our approach was based on the generation of monoclonal antibodies (mAbs) directed against IFN-inducible cell surface proteins. The mAbs could then be used to analyze the expression and the function of their target and eventually to clone the corresponding cDNA.
Here we report the characterization of a mAb directed against a 17-kDa membrane antigen inducible by IFNs. The expression pattern of this 17-kDa protein by human cell lines correlated with sensitivity to the antiproliferative activity of IFNs. Expression cloning revealed that the 17-kDa protein is encoded by the 9 -27 gene, which is a member of an IFNinducible gene family isolated by differential screening and whose function was unknown (5,6). We also demonstrated that the product of the 9 -27 gene is a protein previously identified as Leu-13, a leukocyte antigen that is part of a membrane complex of proteins involved in the transduction of antiproliferative and homotypic adhesion signals (4, 5, 8 -12).

EXPERIMENTAL PROCEDURES
Cell Lines and Interferons-Daudi, Raji, and Namalwa are human lymphoblastoid B cells (Burkitt lymphoma). DIF8 and Daudi-R are two different IFN-resistant cell lines derived from Daudi cells (13,14). Jurkat and H9 are human lymphoblastoid T cells (derived from an acute lymphoblastic leukemia and from a lymphoid leukemia, respectively). Reh and RS4;11 are human lymphoblastoid null cells (acute lymphocytic leukemia). K562 are human leukemic cells derived from a * This work was supported by grants of the National Fund for Scientific Research (Belgium) (to G. D. and M. G. W.) and by grants of the Fund for Medical Scientific Research (Belgium), the Fonds Cancérologique de la CGER, and of the Actions de Recherche Concertées-Université Libre de Bruxelles (to G. A. H.). 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.
Production of mAbs-BALB/c mice were immunized with IFN-␣treated Daudi cells after a procedure of immunosuppression. The immunosuppression and immunization schedule were adapted from a procedure described by Matthew and Sandrock (15). Four tolerization steps were repeated at 2-week intervals and were followed by three immunizations with IFN-␣-treated cells. One tolerization consisted in the injection of 10 7 untreated Daudi cells at day 0 followed by three cyclophosphamide injections (100 g/g) at days 0, 1, and 2. One immunization consisted in the injection of 10 7 IFN-␣-treated Daudi cells. The fusion of spleen cells and Sp2/0 myeloma cells was performed at day 4 after the last intravenous boosting as described previously (16).
Hybridoma supernatants were screened for their differential reactivity to IFN-␣-treated or untreated Daudi cells by flow cytometry. Positively selected hybridomas were repeatedly cloned by limiting dilution in the presence of 100 units/ml of recombinant murine interleukin-6 to maintain Ig secretion. Antibody isotype was determined using the mouse mAb isotyping kit from Life Technologies. IgM mAbs were purified by ammonium sulfate precipitation of ascites fluid followed by exclusion chromatography on A-0,5 resin (Bio-Rad 200 -400 mesh), and their concentration was quantified by enzyme-linked immunosorbent assay.
Additional murine mAbs used in this study include the following: anti-Leu-13, an IgG 1 originally characterized by Chen et al. (10); anti-TAPA-1 antibody 5A6 (17), an IgG 1 (kind gift of Dr. S. Levy, Stanford University School of Medicine, Stanford CA); anti-CD3 mAb OKT3, an IgG 2a (obtained from the American Type Culture Collection); and two isotype-specific control antibodies that were nonreactive against human cell lines, A4 (an IgG 1 that is an anti-idiotypic antibody) and M104E (an IgM directed against B1355S dextran, kindly provided by Dr. R. Ward (Roswell Park Cancer Institute, Buffalo, NY)).
Fluorocytometric Analysis-Expression of membrane antigens was measured by indirect immunofluorescence analysis on a FACSCAN flow cytometer (Becton Dickinson). Briefly, 10 6 cells were washed 3 times with PBS and incubated in PBS, 0.5% bovine serum albumin, 0.01% NaN 3 with saturating amounts of mAb for 1 h on ice; washed with PBS; and stained in 100 l of 10 g/ml goat anti-mouse-Ig-fluorescein isothiocyanate (Sigma) for 1 h on ice. In each experiment, an isotype-matched negative control mAb was substituted for the primary mAb to determine background levels of fluorescence.
cDNA Library Construction-Total RNA from HeLa cells treated for 24 h with 1000 units/ml IFN-␣ was extracted by the guanidium thiocyanate method followed by ultracentrifugation through a CsCl cushion (19). Poly(A) ϩ RNA was isolated by oligo(dT) cellulose chromatography as described by Aviv and Leder (20). An oligo(dT) primed library was prepared from 500 ng of poly(A) ϩ RNA using the primer-adapter procedure described previously (21). A library of approximately 3 ϫ 10 5 independent clones was screened.
COS Cells Transfection and Screening by Panning-Plasmid DNA was prepared using the polyethylene glycol purification method and was transfected into COS cells using DEAE-dextran as described previously (22). Briefly, 2 ϫ 10 6 cells in a 10-cm dish were transfected with 12 g of DNA in 4 ml of Dulbecco's modified Eagle's medium containing 800 g of DEAE-dextran (Pharmacia Biotech Inc.), 200 M chloroquine diphosphate, and 5% FCS. After 5 h at 37°C, the transfection medium was removed, and the cells were treated for 1 min at 4°C with 10% Me 2 SO in PBS. Cells were then rinsed twice with PBS and overlaid with Dulbecco's modified Eagle's medium, 1% FCS for 72 h. Transfected COS cells that expressed the specific membrane protein were first enriched by panning (23), exactly as described previously (24). Episomal DNA was recovered from the selected cells by Hirt extraction (25) and used to transform MC1061 bacteria by electroporation. Plasmid DNA recovered from the transformants was polyethylene glycol-purified and used to transfect COS cells for the next round of panning. Six panning cycles were performed. Thereafter, plasmid DNA was prepared from bacterial pools, then from subpools, and finally from individual clones.
Synthesis of 1-8u cDNA-The 1-8u cDNA was obtained by reverse transcription polymerase chain reaction using mRNA isolated from IFN-␣-treated HeLa cells and primers 1-8-5, CGGGATCCCGACCGC-CGCTGGTC, and 1-8-3, CGGAATTCATGCCTCCTGATCTATC, according to the manufacturer protocol (Perkin Elmer). After restriction with BamHI and EcoRI, the cDNA was inserted into the BamHI-EcoRIrestricted pcD␤A vector, a derivative of pcDNAI/Amp (Invitrogen) in which the SV40 t intron sequence has been replaced by the sequence of a human ␤-globin intron. 10 constructs were sequenced and found to contain the 1-8u cDNA sequence. The HindIII-EcoRI 1-8u-containing fragment was subcloned into the HindIII-EcoRI-restricted pse1 plasmid.
Adhesion Assays-100 l of K562, Jurkat, or U937 cells at 5 ϫ 10 4 cells/ml were distributed in flat bottom 96-microwell plates with or without 1000 units/ml IFN-␣. Each well received an additional 100 l of medium containing 1 g of one of the following mAbs: 13A5, a control IgM, anti-Leu-13, or a control IgG1. Cells were incubated for 24 h at 37°C, and adhesion was assessed microscopically.
Lymphocyte Proliferation Assays-The effect of anti-Leu-13 and 13A5 mAb on anti-CD3-driven proliferation of human lymphocytes was determined essentially as described previously (4,10). Briefly, human peripheral blood mononuclear cells (PBMC) were isolated from normal donor buffy coat leukocyte concentrates (American Red Cross, Buffalo, NY chapter) by Ficoll-Hypaque centrifugation (4,5). PBMC were resuspended at a final concentration of 2.5 ϫ 10 6 cells/ml in flat bottom 96-well microtiter plates in RPMI 1640 containing 10% FCS, 100 units/ml penicillin, 100 g/ml streptomycin, and 2 mM glutamine. Lymphocyte proliferation was induced by culturing PBMC in the presence of 0.5 g/ml of anti-CD3 murine mAb. Anti-Leu-13, 13A5, and isotypematched control antibodies were included in anti-CD3-activated cultures at a final concentration of 10 g/ml. Each sample was set up in triplicate. At 72 h, cultures were pulsed with 0.6 M [ 3 H]thymidine (6.7 Ci/mmol; DuPont NEN), and [ 3 H]thymidine incorporation was measured in a ␤ scintillation counter (Beckman Instruments).

Generation of a Monoclonal Antibody Reactive with a 17-kDa Membrane Protein Induced by IFN-␣ on Daudi Cells-Daudi
cells treated with IFN-␣ were used as a source of IFN-induced antigens since they are very sensitive to the antiproliferative activity of type I IFNs (13). To increase the frequency of mAbs recognizing IFN-␣-induced antigens, mice were first injected with untreated Daudi cells and submitted to a cyclophosphamide regiment in an attempt to selectively kill lymphocytes stimulated by constitutively expressed Daudi antigens. Tolerization was followed by immunization with IFN-␣-treated Daudi cells. Spleenocytes were fused to Sp2/0 myeloma cells, and the resulting hybridomas were screened by indirect immunofluorescence for secretion of mAbs reacting differentially with IFN-␣-treated and untreated Daudi cells.
One mAb termed 13A5, an IgM, reacted with a protein that was highly inducible by IFN-␣ on Daudi cells. The antigen recognized by this mAb was characterized by immunoprecipitation of 125 I-labeled surface proteins from IFN-␣-treated or untreated Daudi cells. SDS-polyacrylamide gel electrophoresis analysis of the immunoprecipitation products revealed that the molecular mass of the protein recognized by 13A5 was approximately 17 kDa and that its expression on Daudi cells was strongly induced upon IFN-␣ treatment (Fig. 1, A and B). The antigen recognized by 13A5 had the same mobility as a predominant IFN-inducible protein that can be detected in whole 125 I-labeled IFN-␣-treated Daudi cell lysates (Fig. 1, A and C). An IFN-inducible 17-kDa protein on the surface of Daudi cells has been previously described and implicated in the antiproliferative activity of IFN (5,7).
Expression of the 17-kDa Protein in Response to Interferons-Established human cell lines differ widely in their sensitivity to the antiproliferative activity of IFNs. Expression of the 17-kDa protein in response to type I and II IFNs was assayed by indirect immunofluorescence on a panel of IFN-sensitive and -resistant cell lines. Representatives of epithelial, fibroblastic, promyelomonocytic and lymphoblastoid (B, T, and null) origin were included (Table I). DIF8 and Daudi R are mutant cell lines derived from Daudi cells that were selected for resistance to the antiproliferative effect of IFN-␣ (13,14). U937, Raji and Namalwa cell growth is also insensitive to the inhibitory effects of IFN-␣.
The results of the 17-kDa protein expression analysis are presented in Table I. Fig. 2 shows the fluorescence histograms of four representative cell lines on a logarithmic scale. We observed that IFN-␣ treatment increased the expression of the 13A5 antigen on most of the cell lines tested. As much as a 20-fold increase was observed on Daudi, Jurkat, Reh, and FS-7F cells. Interestingly, the level of induction was very weak on several resistant cell lines (e.g. on DaudiR or U937) or undetectable on DIF8. The pattern of expression appeared to correlate closely with the ability of IFN-␣ to inhibit cell growth. IFN-␥ also significantly affected the expression of the 17-kDa protein although to a lesser extent than did IFN-␣. Induction of the 17-kDa protein by IFN-␥ was observed in most cell lines tested, except in those derived from B or T cells. This result is consistent with our previous observation that genes that are predominantly induced by type I IFN are not or very poorly inducible by IFN-␥ in cell lines of lymphoblastoid origin (26). We concluded that the 17-kDa protein is inducible by both types of IFNs on sensitive cells.
Some of the cell lines tested expressed a basal level of the 17-kDa protein, which was particularly prominent on K562 cells. Confirming the fluorocytometric analysis, immunoprecipitation of 125 I-labeled surface proteins with 13A5 showed the high basal level of the 17-kDa protein on K562 cells (Fig. 3).
Overexposure of the autoradiogram allowed the detection of a 28-kDa component, similar to the one observed in the proteins immunoprecipitated by 13A5 from IFN-␣-treated Daudi cells (compare Figs. 1B and 3). Immunoprecipitated proteins of 17 and 28 kDa were also detected under nonreducing conditions (Fig. 3B). These data suggest that the 17-kDa protein is noncovalently associated in the cell membrane with a protein of 28 kDa. Expression Cloning of the cDNA Coding for the 17-kDa Protein-We constructed a directional cDNA expression library in the pse1 vector (21) using mRNA extracted from HeLa cells that had been incubated for 24 h in the presence of IFN-␣. Screening of transfected cells reactive with the 13A5 mAb was undertaken using the "panning" method previously described for the cloning of various eukaryotic membrane proteins (17, 23, 24, 27). Six successive amplification and enrichment cycles  Three additional screening cycles were performed by transfecting plasmid extracted from counted colony pools and by detecting the positive pools by indirect immunofluorescence. This procedure lead to the isolation of two identical clones containing a cDNA insert of 650 base pairs. COS cells transfected with the cDNA from one of these clones was strongly reactive with the 13A5 mAb, as determined by flow cytometric analysis (Fig. 4A).
The cDNA sequence was obtained by the dideoxy termination method. Comparison with the EMBL data bank showed that the cDNA encoding the 17-kDa protein was nearly identical to the 9 -27 cDNA. The two sequences diverge at positions 280 (A3 T), 391 (C3 T), 499 (G3 A), 624 (C3 T), and 643 (G3 A) of the previously cloned 9 -27 cDNA. Only the replacement at position 624 results in a modification in the peptide sequence (Ser 103 3 Leu). These divergences are probably due to allelic variations. The 9 -27 cDNA was first isolated on the basis of its IFN inducibility (28) and is part of a family containing at least two other members, 1-8u and 1-8d, both IFN-inducible as well. The isolated cDNA was full-length, containing the entire open reading frame of 9 -27 preceded by a 5Ј-untranslated region corresponding to the second start site of transcription reported for the 9 -27 gene (29).
The 9 -27 Gene Encodes Leu-13-Leu-13 is a 16-kDa leukocyte surface antigen whose expression is stimulated on endothelial cells and B lymphocytes after exposure to both types of IFNs (4,5,10,30). Previous immunoprecipitation studies have established that Leu-13 is noncovalently associated with polypeptides of 28, 48, and 90 kDa (10,12). Taken together, these observations prompted us to determine if the 9 -27 gene, or the closely related 1-8u or 1-8d genes, encoded the Leu-13 antigen. In the experiment shown in Fig. 4A, COS cells transfected with the 9 -27 cDNA were first reacted with 13A5 and anti-Leu-13 mAbs and then were counterstained with a fluorescein-conjugated antibody and analyzed by flow cytometry. Fig. 4A shows that both the 13A5 and anti-Leu-13 mAbs positively reacted with COS cells transfected by the 9 -27 encoding plasmid but not with the cells transfected with the control plasmid. We conclude that 13A5 and anti-Leu-13 each recognize an epitope in the 9 -27 protein.
The 9 -27, 1-8u, and 1-8d coding sequences diverge at the amino and carboxyl termini. Thus, it was theoretically possible that 13A5 and anti-Leu-13 could recognize an epitope common to both 9 -27 and 1-8 proteins. An reverse transcription po- lymerase chain reaction approach using primers designed to amplify the 1-8u and 1-8d mRNAs resulted in the obtention of the 1-8u but not of the 1-8d cDNA, which is less abundant and less homologous. As shown in Fig. 4B, neither anti-Leu-13 nor 13A5 reacted with COS cells transfected with the 1-8u cDNA. These results were confirmed by immunoprecipitation experiments performed on in vitro synthesized 9 -27 and 1-8u proteins. Both mAbs were able to precipitate 9 -27 but not 1-8u products (data not shown). Therefore, the epitope recognized by 13A5 and anti-Leu-13 on the 9 -27 protein is not present in the 1-8u protein. Thus, we conclude that the leukocyte antigen, Leu-13, is composed of only one polypeptide encoded by the IFN-inducible 9 -27 gene.
Proteins Coprecipitated with the 17-kDa Protein-We next investigated whether the 28-kDa protein coprecipitated with the 17-kDa protein by 13A5 could correspond to TAPA-1. TAPA-1 is a 26-kDa protein that can associate with Leu-13 (8,9,11,12). 125 I-labeled K562 cells were lysed in a 2% Nonidet P-40 buffer, as described previously in Fig. 3, or in buffer containing 1% CHAPS, a mild nonionic detergent, to preserve most of the noncovalent protein interactions. Fig. 5A shows that the protein whose interaction with 17-kDa protein is preserved in the Nonidet P-40 lysate did not correspond to TAPA-1. However, when the cells were lysed in the milder CHAPS buffer, several proteins were precipitated with either 5A6, the anti-TAPA-1 mAb, or 13A5, the anti-9 -27-encoded 17-kDa protein mAb. Using the 5A6 mAb, TAPA-1 was detected by Western blot among the proteins precipitated with the 13A5 mAb in CHAPS lysate (Fig. 5B), in agreement with previously reported results obtained using anti-Leu-13 mAb (8,9,11,12). The 28-kDa protein coprecipitated by 13A5, but not by 5A6, from Nonidet P-40 lysate appeared in both precipitation products when CHAPS lysate was used. Thus 9 -27/Leu-13 is associated with several proteins at the cell surface, including TAPA-1, and the strength of these interactions can be probed using different detergents.
Effect of 13A5 and anti-Leu-13 mAbs on Homotypic Adhesion and Proliferation-Previous studies using anti-Leu-13 mAb have established that this mAb can induce the homotypic adhesion of cells expressing Leu-13 (4, 5, 10) and can inhibit the anti-CD3-driven proliferation of PBMC (4, 12, 17). Fig. 6A shows the results of an homotypic adhesion assay on U937, Jurkat, and K562 cells using 13A5, anti-Leu-13, and control mAbs. Interestingly, the 13A5 mAb did not induce cell adhesion. In contrast, the anti-Leu-13 mAb triggered adhesion of cells expressing 9 -27/Leu-13, either constitutively (K562 or Jurkat), or in response to IFN treatment (U937). Similarly, the 13A5 mAb had no significant effect on the anti-CD3-driven proliferation of PBMC as measured by incorporation of [ 3 H]thymidine, while anti-Leu-13 mAb inhibited proliferation by more than 50% (Fig. 6B). These results suggest that 13A5 and anti-Leu-13 mAbs reacted with distinct epitopes on the 9 -27/Leu-13 molecule and that the latter mAb could mimic an interaction with a physiological ligand, inducing a biological response. DISCUSSION Modulations of gene expression in cells exposed to IFNs play an essential role in the ability of these cytokines to affect vital  3 and 4). These cell extracts were immunoprecipitated by 5A6, the anti-TAPA-1 antibody (lanes 1  and 3), or by 13A5 (lanes 2 and 4). Arrows, from higher to lower molecular mass, correspond to p28, TAPA-1, 9 -27. Molecular masses are given in kDa. B, the same immunoprecipitates shown in A were separated by SDS-polyacrylamide gel electrophoresis under nonreducing conditions, transferred to a nitrocellulose sheet, and the presence of TAPA-1 was revealed using the 5A6 antibody, a horseradish peroxidaseconjugated anti-mouse antibody, and ECL chemiluminescence. Upper bands correspond to the antibodies used in the immunoprecipitation.
FIG. 6. Effect of 13A5 and anti-Leu-13 mAbs on homotypic adhesion and proliferation. A, induction of homotypic adhesion by mAbs. K562, Jurkat, and U937 cells were cultured for 24 h with or without 500 units/ml IFN-␣ (ϩifn or Ϫifn) in the presence of the indicated mAbs at a final concentration of 5 g/ml. Control IgG1 and control IgM had no effect on cell-to-cell adhesion. B, anti-Leu-13 inhibits anti-CD3-driven proliferation of human PBMC. Human PBMC were cultured in the absence or presence of 0.5 g/ml of anti-CD3 mAb. Anti-Leu-13, 13A5, or isotype-matched control antibodies were included in cultures at a concentration of 10 g/ml. biological processes. Several proteins whose synthesis is induced by IFNs have been implicated in the antiviral activity shared by all IFNs (for review, see Refs. 2 and 31-33). The inhibition of cell growth by IFNs is also likely to result from an action on multiple pathways affecting different steps at which cell growth can be regulated. Interestingly, two doublestranded RNA-dependent antiviral pathways induced by IFNs have been recently implicated in their antiproliferative activity as well (34 -36). IFN treatment also directly affects the function of two genes known to be involved in the control of the cell-cycle, the protooncogene c-myc and the tumor-suppressor retinoblastoma gene, resulting in arrest at the G 0 /G 1 phase of the cell cycle (37)(38)(39)(40).
It has been shown that the antiproliferative activity of IFN could be transferred from cell to cell, not by diffusion of a soluble mediator but through direct contact between cells (3). To explore the role of membrane proteins in the antiproliferative activity of IFNs, we generated mAbs directed against IFNinducible membrane antigens. We obtained an antibody that reacted specifically with a 17-kDa protein. This 17-kDa protein was induced by type I and II IFNs on a wide range of cell lines sensitive to the antiproliferative effect of IFNs, but not significantly on resistant cell lines. Screening of an expression library yielded a full-length cDNA clone encoding the 17-kDa protein, and sequence analysis revealed identity with the 9 -27 cDNA.
The 9 -27 gene is induced by both type I and II IFNs, and the mRNA level can increase as much as a 100-fold upon induction by IFNs. Induction by IFNs is transcriptionally controlled by a single IFN-stimulated response element present in the promoter of the gene (41). The protein encoded by the 9 -27 gene has a calculated molecular mass of 13.9 kDa, in close agreement with the observed 16 -17 kDa (7,10). In addition, we demonstrated that 9 -27 was coding for the leukocyte antigen Leu-13. In most cell types studied thus far, Leu-13 was shown to be part of a membrane complex containing several distinct proteins. One of these proteins is the 26-kDa TAPA-1 protein, the target of an antiproliferative antibody (12). TAPA-1 is strongly related at the sequence level to two other surface proteins involved in the regulation of cell growth, the ME491 melanoma-associated antigen, and CD37, another leukocyte antigen (17). Leu-13 and TAPA-1 are involved in a mechanism controlling cellular adhesion since anti-Leu-13 and anti-TAPA-1 mAbs each triggers a general homotypic adhesion phenomenon. Interestingly, homotypic adhesion induced by anti-Leu-13 or anti-TAPA-1 is not dependent on adhesion pathways involving known adhesion molecule including the leukocyte function-associated antigen-1, the intercellular adhesion molecule-1, CD44, or VLA-4 (4,5,12).
Along with their involvement in the regulation of cell aggregation, Leu-13 and TAPA-1 are likely to also play a role in the control of cell growth. Indeed, anti-Leu-13 and anti-TAPA-1 mAbs can directly inhibit cell growth, although this effect is observed in a more restricted subset of cell lines (4,12,17). Furthermore, anti-Leu-13 mAbs were shown to potentiate the antigrowth effect of IFN-␣ on leukemic B cells (4). Insensitivity to the growth inhibitory effect of anti-Leu-13 and anti-TAPA-1 mAbs in some cell lines can probably be accounted for by the alterations in the control of cell growth that occur when a cell line is established for in vitro growth. We also attempted to establish cell lines stably expressing 9 -27/Leu13 under the control of a constitutive promoter. Although some clones did express heterogeneous levels of the protein early in the selection procedure as revealed by staining, expression rapidly declined to undetectable levels (data not shown). This result suggested that stable expression of 9 -27/Leu13 might be hin-dering cell growth.
Evidence that antibodies against 9 -27/Leu-13 and TAPA-1 can induce a biological response suggests that the multimeric cell surface complex containing 9 -27/Leu-13, TAPA-1, and other molecules is a receptor for an as yet unidentified ligand. Indeed, activation of a signal transduction pathway by extracellular signals often requires ligand-induced dimerization or oligomerization of the corresponding receptor. Kinase(s) associated with the receptor are brought together, resulting in their reciprocal phosphorylation and the subsequent activation of further downstream components of the signal transduction pathway (42). Therefore, antibodies against cell surface receptors that are activated by dimerization can sometimes function as an agonist because they artificially induce dimerization of the receptor, whereas Fab fragments of such antibodies have no activity (42).
Taken together, these results show that the 9 -27 gene is coding for a cell surface protein that associates with other membrane proteins, forming a multimeric complex that relays growth inhibitory and cell adhesion signals.
Interferons seem to exert their inhibition on cell growth by acting at many different levels, (1) directly affecting the function of protein such as c-myc and Rb that are intimately involved in cell-cycle control, (2) increasing the level of enzymes such as the double-stranded RNA-dependent protein kinase and the 2-5A synthetases that inhibit cell anabolism, and (3) inducing cell surface proteins such as 9 -27/Leu-13 that relay other growth inhibitory signals. While the activation of a single pathway might be sufficient to inhibit the growth of a given cell, activation of multiple pathways by IFNs allows for both more efficiency and flexibility in the control of cell growth. Indeed, in a physiological setting, IFNs are acting on a population of cells at distinct stages and in different programs of differentiation, cells that are continuously exposed to various other stimuli and have to maintain the ability to perform other essential functions, hence the requirement for efficiency and flexibility, i.e. multiple pathways.