Divergence of Binding, Signaling, and Biological Responses to Recombinant Human Hybrid IFN

This information is current as Renqiu Hu, Joseph Bekisz, Mark Hayes, Susette Audet, Judy of September 24, 2021. Beeler, Emanuel Petricoin and Kathryn Zoon J Immunol 1999; 163:854-860; ; http://www.jimmunol.org/content/163/2/854 Downloaded from

References This article cites 29 articles, 10 of which you can access for free at: http://www.jimmunol.org/content/163/2/854.full#ref-list-1

Why The JI? Submit online. http://www.jimmunol.org/

• Rapid Reviews! 30 days* from submission to initial decision

• No Triage! Every submission reviewed by practicing scientists

• Fast Publication! 4 weeks from acceptance to publication

*average by guest on September 24, 2021

Subscription Information about subscribing to The Journal of Immunology is online at: http://jimmunol.org/subscription Permissions Submit copyright permission requests at: http://www.aai.org/About/Publications/JI/copyright.html Email Alerts Receive free email-alerts when new articles cite this article. Sign up at: http://jimmunol.org/alerts

The Journal of Immunology is published twice each month by The American Association of Immunologists, Inc., 1451 Rockville Pike, Suite 650, Rockville, MD 20852 Copyright © 1999 by The American Association of Immunologists All rights reserved. Print ISSN: 0022-1767 Online ISSN: 1550-6606. Divergence of Binding, Signaling, and Biological Responses to Recombinant Human Hybrid IFN1

Renqiu Hu,2* Joseph Bekisz,* Mark Hayes,* Susette Audet,† Judy Beeler,† Emanuel Petricoin,* and Kathryn Zoon*

Three human IFN-␣ hybrids, HY-1 [IFN-␣21a(1-75)/␣2c(76-165)], HY-2 [IFN-␣21a(1-95)/␣2c(96-165)], and HY-3 [IFN-␣2c(1- 95)/␣21a(96-166)], were constructed, cloned, and expressed. The hybrids had comparable specific antiviral activities on Madin- Darby bovine kidney (MDBK)3 cells but exhibited very different antiproliferative and binding properties on human Daudi and WISH cells and primary human lymphocytes. Our data suggest that a portion of the N-terminal region of the molecule is important for interaction with components involved in binding of IFN-␣2b while the C-terminal portion of IFN is critical for antiproliferative activity. A domain affecting the antiproliferative activity was found within the C-terminal region from amino acid residues 75–166. The signal transduction properties of HY-2 and HY-3 were evaluated by EMSA and RNase protection assays. Downloaded from Both HY-2 and HY-3 induced activation of STAT1 and 2. However, HY-2 exhibited essentially no antiproliferative effects at concentrations that activated STAT1 and 2. Additionally, at concentrations where no antiproliferative activity was seen, HY-2 induced a variety of IFN-responsive to the same degree as HY-3. RNase protection assays also indicate that, at concentra- tions where no antiproliferative activity was seen for HY-2, this construct retained the ability to induce a variety of IFN-inducible genes. These data suggest that the antiproliferative response may not be solely directed by the activation of the STAT1 and STAT2 pathway in the cells tested. The Journal of Immunology, 1999, 163: 854–860. http://www.jimmunol.org/

nterferons are that have been shown to affect a To date, our laboratory has purified and characterized 22 IFN-␣ wide variety of cellular functions, including inhibition of cel- components produced by Sendai virus-induced human lympho- I lular proliferation, antiviral activity, immune regulatory func- blastoid cells (12). One of these species, component o, was found tions, and activation of multiple cellular genes (1). Antiviral and to be noteworthy for its high antiproliferative activity and its poor antiproliferative activities are clinically important properties of ability to compete with the IFN-␣2b binding site (13). Based on IFN-␣s. Several types of IFN-␣2 are currently approved for a va- our partial amino acid sequence data, component o was indistin- riety of diseases including hepatitis b and c, hairy cell leukemia, guishable from IFN-␣21a. Therefore, rIFN-␣21a was cloned and by guest on September 24, 2021 and chronic myelogenous leukemia (2). expressed. It was determined to behave functionally like compo- Recombinant techniques are useful methods for the production nent o in that it exhibits a high antiproliferative sp. act. and com- and modification of IFN-␣ . The first IFN hybrid, IFN- petes poorly for the IFN-␣2b binding site (13, 14). ␣1/␣2 was constructed using recombinant technology in 1981 (3). The goals of our studies are to determine the specific region(s) Subsequently, a number of hybrids have been constructed and have of IFN-␣21a responsible for the enhanced antiproliferative activity been informative for examining the activity of IFNs (4–11). These and to further understand the signaling mechanism it uses to elicit hybrid constructs have resulted in novel IFNs that either combine this response. Three IFNA2 and IFNA21 chimeras (HY-1, HY-2, different biological properties from the parental proteins or have and HY-3) were constructed using PCR technology and cloned significantly different biological activity from both the parents (3). into a pQE-30 expression vector. The hybrid proteins were ex- Therefore, IFN hybrids have provided a powerful tool for studying pressed in Escherichia coli. These chimeras were selected based the structure-function relationship of these molecules. The engi- on the 1) amino acid sequences of IFN-␣2c and IFN-␣21a, 2) the neered IFN-␣ proteins may have important new therapeutic appli- biological activities and binding properties of these IFN-␣s, and 3) cations and may provide greater insights into understanding of the information derived from the three-dimensional structures of hu- clinical activities of existing IFN-␣s. man IFN-␣ consensus and human IFN-␤ (15, 16). The three-di- mensional model of a human IFN-␣ consensus sequence reveals *Division of Biology, Office of Therapeutics Research and Review, and that the two domains 29–35 and 123–140 are in close spatial prox- †Division of Viral Products, Office of Vaccines Research and Review, Center for imity and may constitute a receptor recognition domain. In con- Biologics Evaluation and Research, Food and Drug Administration, Rockville, MD trast, the residues 78–95 are distant from this region on the mol- 20892 ecule and may influence species specificity and differential Received for publication December 7, 1999. Accepted for publication April 26, 1999. biological activities of the human IFN-␣s and IFN-␤ (14). 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 The human IFN hybrid proteins were purified using nickel-ni- with 18 U.S.C. Section 1734 solely to indicate this fact. trilotriacetic acid. Agarose and mAb affinity chromatography. The 1 The sequences presented in this article have been submitted to GenBank under antiviral, antiproliferative, binding, signal transduction, and RNase accession numbers HY-1 AF085803, HY-2 AF085804, and HY-3 AF085805. protection properties of the IFN-␣ hybrids were analyzed and com- 2 Address correspondence and reprint requests to Dr. Renqui Hu, Center for Biologics pared. Our data suggest that a portion of the N-terminal region of Evaluation and Research, Food and Drug Administration, 1401 Rockville Pike, Rock- ville, MD 20892. E-mail address: [email protected] the molecule is important for interaction with receptor compo- ␣ 3 Abbreviations used in this paper: MDBK, Madin-Darby bovine kidney; GRR, ␥ nents, while the C-terminal region of IFN- is critical for antipro- response region; ISRE, IFN-stimulated response element. liferative activity. The EMSA and RNase protection assays of the

Copyright © 1999 by The American Association of Immunologists 0022-1767/99/$02.00 The Journal of Immunology 855 hybrid IFN-␣s suggest that the antiproliferative response may not Expression and purification be directed by STAT1 and 2 activation alone in either a nonhe- IFNA2, IFNA21, and hybrids HY-2, HY-3 plasmid DNAs were individ- matopoietic cell line, a hematopoietic cell line, or in a primary ually transformed into E. coli strain SG13009[pREP4] (Qiagen), and HY-1 human cell type. plasmid DNA was transformed into E. coli strain DH5␣F’IQ (Life Tech- nologies, Gaithersburg, MD). Bacteria were grown in Luria-Bertani broth containing 100 ␮g/ml ampicillin (HY-1) or 100 ␮g/ml ampicillin and 25 Materials and Methods ␮g/ml kanamycin (HY-2, HY-3, IFNA2, and IFNA21) in a 37°C shaker IFN and radiolabeled IFN-␣2b incubator overnight. The cultures were diluted 1:50 in Luria-Bertani broth containing the appropriate antibiotic(s) and incubated at 37°C with shaking ␣ Recombinant human IFN- 2b (obtained from Schering, Kenilworth, NJ) to an A600 of 0.8–0.9. expression was induced by 2 mM isopropyl- 8 has an antiviral sp. act. of 2 ϫ 10 IU/mg protein. IFN-␣2c (cDNA was 1-thio-␤-D-galactopyranoside. The bacteria were then incubated at 30°C cloned into the vector pBluescript; Stratagene, La Jolla, CA) and IFN-␣21a for 4–5 h, after which cells were harvested and sonicated. The cell lysate (cDNA from Genentech, South San Francisco, CA) proteins were ex- was clarified by centrifugation at 10,000 ϫ g for 30 min at 40°C. IFN pressed and purified from E. coli strains transformed with pQE30 as de- purification was performed by nickel-nitrilotriacetic acid. Agarose resin scribed below. IFN-␣2b was labeled with 125I-Bolton-Hunter reagent (Am- metal-affinity chromatography (20, 21) and 4F2 mAb affinity chromatog- ersham, Arlington Heights, IL) as previously described (13). raphy (12). The mAb 4F2 recognizes amino acid residues 113–149 of IFN-␣ consensus (Dr. Bruce Altrock, unpublished observations). Cell and cell culture Protein determination Human Daudi cells were obtained from Dr. P. Grimley (Department of Purified recombinant protein concentrations were determined using the Pathology, Uniformed Services University of the Health Sciences, Be- coomassie plus protein assay (Pierce, Rockford, IL). Purity of the rIFN-␣s thesda, MD). Cells were grown in suspension using RPMI 1640 with 10% were assessed by SDS-PAGE and HPLC analysis (data not shown). Downloaded from FCS, 2 mM glutamine, and 0.2% gentamicin. WISH cells (American Type Culture Collection (ATCC), Manassas, VA) were grown as monolayer cul- Antiproliferative assay tures using Eagle’s minimal essential medium supplemented with 10% ␣ FCS and gentamicin (50 ␮g/ml). The cultures were incubated at 37°C in a The antiproliferative activities of the IFN- constructs were analyzed using humidified atmosphere containing 5% CO . All cultures were determined human Daudi and WISH cells and primary human lymphocytes. The assays 2 on Daudi cells were performed as previously described (13). The assays on to be free of mycoplasma. We obtained primary human lymphocytes from ␣ normal donors by centrifugal elutriation after Ficoll-Hypaque sedimenta- WISH cells were performed by incubating the cells with various IFN- sat tion (lymphocyte separation medium package insert, Organon Teknika, the indicated concentrations for 72 h at 37°C. Fifty microliters of 2 mg/ml http://www.jimmunol.org/ MTT was added into each well and incubated for4hat37°C. Then, 10% Durham, NC). The resultant cells were resuspended in RPMI 1640 media ␮ supplemented with 10% FCS and fungizone (250 ng/ml amphotericin B, SDS, 0.01 N HCl (250 l) was added to each well and incubated overnight at 37°C. The OD of each well was determined, and the percentage of 100 U/ml penicillin, and 100 ␮g/ml streptomycin). 570 growth inhibition was calculated by comparing the control cultures (no IFN) with the IFN-treated cultures. Primary human lymphocytes were Construction of pQE30/A21 treated with PHA (Promega, Madison, WI) at 1 ␮g/ml overnight. The Ј resultant PHA blasts were then treated for 72 h with the various IFNs at the A pair of oligonucleotides, 5 -TCCGGATCCTGTGATCTGCCTCAGA concentrations indicated (see Fig. 2). Percent inhibition was calculated C-3Ј (sense) and 5Ј-GAGCTCGCATGCTCATCATTCCTTCCTCCTTA Ј from direct cell counts that were done by Coulter counter analysis (Coulter, ATCT-3 (antisense), with BamHI and SphI restriction sites were synthe- Palo Alto, CA). sized based on the cDNA coding region for mature human IFN-␣21a by guest on September 24, 2021 protein and were used as primers. Human IFNA21 cDNA was used as the Antiviral assay template. PCR was used to amplify an entire coding region for mature human IFN-␣21a protein. The PCR reactions were conducted by standard Antiviral activity was determined as previously described using MDBK procedures (17), and the resulting products were cleaved with restriction cells (ATCC) and WISH cells (ATCC) (12). All IFN units are expressed endonucleases BamHI and SphI and cloned into the E. coli expression with reference to the National Institutes of Health human lymphoblastoid vector pQE30 (purchased from Qiagen, Chatsworth, CA). The final con- IFN standard Ga 23-901-532. struct was confirmed by DNA sequencing (18). The Edmonston strain of measles virus (low passage, human embryonic kidney 7, VERO 5) was plaque purified and used to infect 1 ϫ 106 primary ␣ human lymphocytes (in triplicate) following no treatment or pretreatment Construction of IFN- hybrids DNA with 100 ng/ml of parental or hybrid IFN for 24 h before infection as Hybrid IFN cDNAs were constructed by PCR technology (19). PCR prim- previously described (22). Cells were infected with measles virus at 0.1– ers for the construction of HY-1 [IFN-␣ 21a(1-75)/IFN-␣ 2c(76-165)], 1.6 multiplicity of infection or mock infected with virus-free medium and HY-2 [IFN-␣ 21a(1-95)/IFN-␣ 2c(96-165)], and HY-3 [IFN-␣ 2c(1-95)/ harvested 72 h postinfection. Cell-associated virus and supernatants were IFN-␣ 21a(96-166)] include: primer 1, 5Ј-TCC GGA TCC TGT GAT CTG titrated on VERO cell monolayers. Measles virus cytopathic effect was CCT CAG ACC-3Ј (for HY-1, 2, and 3); primer 2, 5Ј-AGC AGA TGA evaluated microscopically after 6 days and confirmed by staining with GTC CTT TGT GCT GAA GAG-3Ј (for HY-1); primer 3, 5Ј-CTC TTC crystal violet. The results reported are a comparison of the mean virus in AGC ACA AAG GAC TCA TCT GCT-3Ј (for HY-1); primer 4, 5Ј-GAG supernatants harvested from the untreated group with supernatant virus in CTC GCA TGC TCA TCA TTC CTT ACT TCT TAA ACT-3Ј (for HY-1 each treated group. and 2); primer 5, 5Ј-CAC GCA GGC CTC GAG GTC ATT CAG-3Ј (for EMSA HY-2 and 3); primer 6, 5Ј-CTG AAT GAC CTC GAG GCC TGC GTG-3Ј (for HY-2 and 3); and primer 7, 5Ј-GAG CTC GCA TGC TCA TCA TTC Cells were treated with the various IFNs at the concentrations and times CTT CCT CCT TAA TCT-3Ј (for HY-3). Primers 2, 3, 5, and 6 were used indicated (see Fig. 4) and centrifuged at 1500 ϫ g at 40°C for 10 min. After as inside primers for all three hybrids, and primers 1 and 4 (HY-1 and being washed once in ice-cold PBS, the cells were recentrifuged at the HY-2) or 1 and 7 (HY-3) were used as outside primers. The BamHI re- above conditions. The cell pellet was then resuspended in lysis buffer (20 Ј striction site is on the 5 outside primer (primer 1), and the SphI restriction mM HEPES, pH 7.5, 1 mM EDTA, 1 mM MgCl2, 150 mM NaCl, 1% site is on the 3Ј outside primers (primers 4 and 7). The DNA sequences of Triton X-100, 3 mM sodium orthovanadate, 10 mM B-glycerophosphate, IFNA2 and IFNA21 are identical in regions recognized by primer 1, hence, 20 mM DTT, and 2 mM PMSF) followed by centrifugation for 5 min at this primer was used in the construction of both IFNA2 and IFNA21 frag- 14,000 ϫ g. EMSA were then performed. The clarified whole-cell lysates ments. Plasmid DNAs Bluescript/A2 and pQE30/A21 were used as the containing 10 ␮g of total protein were diluted 1:1 in a binding buffer templates for the first PCR reactions for all hybrids. Purified DNA frag- containing 20% glycerol, 2 mM DTT, 150 mM KCl, 2 ␮g polydeoxyi- ments from the first PCRs were mixed as templates, and primers 1 and 4 nosine-deoxycytosine, 0.5 mM EDTA, and 0.5 mM MgCl2 and incubated (for HY-1 and 2) or primers 1 and 7 (for HY-3) were used for the secondary for 15 min on ice. One nanogram of 32P-end labeled probe was then added PCR reactions. to the reaction and allowed to sit an additional 5 min on ice. For the Confirmatory DNA sequencing on the final constructs was performed supershift experiments, rabbit anti-STAT1 or anti-STAT2 serum or normal using a DNA sequencing (Amersham) as previously described (18). The rabbit serum was then added at this point at a 1:200 final dilution (in following are the GenBank accession codes assigned to the hybrids: HY-1 binding buffer) and was allowed to incubate for an additional 30 min on AF085803, HY-2 AF085804, and HY-3 AF085805. ice. The samples were then subjected to electrophoresis on a native 6%, 856 IFN-␣ HYBRIDS EXHIBIT DIFFERENT BIOLOGICAL RESPONSES

FIGURE 1. The amino acid sequences of IFN-␣2c, IFN-␣21, and IFN-␣ hybrids HY-1, HY-2, and HY-3. The amino acid leader sequence for all of these Downloaded from sequences is MRGSHHHHHHGS.

0.25ϫ Tris-borate-EDTA gel for 4 h. The probes used consisted of the ␥ IFN-␣21a on MDBK cells (2.0 ϫ 108 IU/mg to 3.7 ϫ 108 IU/mg). response region (GRR) found within the promoter of the Fc␥RI (5Ј- However, the antiviral sp. act. of HY-1 and HY-2 were, at least, Ј AGCATGTTTCAAGGATTTGAGATGTATTTCCCAGAAAAG-3 ) and 7-fold lower than that of HY-3, IFN-␣2c, or IFN-␣21a on WISH the IFN-stimulated response element (ISRE) of the IFN-stimulated gene 15 http://www.jimmunol.org/ (5Ј-GATCCATGCCTCGGGAAAGGGAAACCGAAACTGAAGCC-3Ј). cells. The ability of the IFN-␣ hybrids and the parental IFN-␣sto RNase protection assay inhibit the growth of Daudi, WISH, and primary human lympho- This assay was performed as described previously (23). cytes was also examined. The inhibition curves are shown in Fig. 2. The concentrations of IFN-␣s that inhibited Daudi and WISH Binding assay cell growth by 50% are shown in the Table I. Hybrid HY-3 always IFN binding assays were performed as previously described on human exhibited a higher antiproliferative sp. act. compared with the Daudi (13) and WISH (24) cells. other hybrids and IFN-␣2c and IFN-␣21a on all three cell popu-

lations. In comparison, hybrid HY-2 had a lower antiproliferative by guest on September 24, 2021 Results sp. act. than either of the other hybrids or either of the parental ␣ Purification and characterization of the IFN- s and hybrids IFN-␣s on the same cell populations. HY-2 displayed 10,000-fold Human IFN-␣2c, IFN-␣21a, and three IFN hybrids, HY-1, HY-2, lower antiproliferative sp. act. compared with HY-3 on Daudi cells and HY-3, were expressed in E. coli using a pQE30 expression and Ͼ1,000-fold less sp. act. on WISH and primary human cells system. The sequences of the IFN-␣s deduced from their cDNA compared with HY-3. The hybrid HY-1 had a 2- to 8-fold greater sequences are shown in Fig. 1. Initial purification of the IFNs on antiproliferative sp. act. than HY-2 on Daudi and WISH cells. [The a Ni-NTA-Agarose column yielded partially purified material, relative ability of HY-2 and HY-3 to induce an antiviral state in a with antiviral sp. act. ranging from 3 ϫ 106 to 4.5 ϫ 106 IU/mg primary cell was determined by examining fresh human lympho- protein on MDBK cells. The IFN-␣s were further purified by 4F2 cytes primed with PHA]. The PHA blasts were treated with con- or NK2 mAb affinity chromatography (12). After the two purifi- centrations of HY-2 and HY-3 known to induce STAT1 and cation steps, each IFN-␣ hybrid appeared as a single band with an STAT2 and subsequently infected with measles virus. Fig. 3 shows apparent m.w. of 20,000 on reducing SDS-PAGE. The antiviral sp. that for two separate normal donors at concentrations of HY-2 in act. of the purified IFN-␣s are shown in Table I; they ranged from which no antiproliferative effects were seen and STAT1 and 2 ϫ 108 to 3.7 ϫ 108 IU/mg protein on MDBK cells and 0.1 ϫ 108 STAT2 activation was readily apparent, there was suppression of to 1.9 ϫ 108 IU/mg protein on WISH cells. The antiviral sp. act. release of measles virus. These data indicated that the antiviral of the three hybrids were similar to each other and to IFN-␣2c and pathway was operational. However, there was no effect on measles

Table I. Summary of antiproliferative, antiviral, and binding activities of hybrid and parent IFNs

Binding Activity (competes with Antiproliferative Activity (ng/ml)a Specific Antiviral Activity (IU/mg) 125I-IFN-␣2b) (ng/ml)b

Daudi Wish Ratio wish/ Ratio MDBK/ IFN cells cells Daudi MDBK cells Wish cells Wish Daudi cells Wish cells

IFN-␣2c 0.005 80 1.6 ϫ 104 3.3 ϫ 108 1.9 ϫ 108 1.7 35 25 IFN-␣21a 0.0008 95 1.2 ϫ 105 3.7 ϫ 108 0.7 ϫ 108 5.3 Ͼ400 Ͼ200 HY-1 0.5 110 2.2 ϫ 102 2.0 ϫ 108 0.1 ϫ 108 20.0 Ͼ400 Ͼ200 HY-2 1.0 Ͼ1000 Ͼ1 ϫ 103 3.0 ϫ 108 0.1 ϫ 108 30.0 Ͼ400 Ͼ200 HY-3 0.0001 8 8 ϫ 104 2.0 ϫ 108 0.7 ϫ 108 2.9 30 30

a Concentration of IFN species that inhibits cell growth by 50%. b Concentration of IFN species that inhibits binding of 125I-IFN-␣2b by 50%. The Journal of Immunology 857 Downloaded from FIGURE 3. Antiviral studies of IFN-␣2c, IFN-␣21a, and the IFN-␣ hy- brids using primary human lymphocytes. Comparison of no IFN (Ⅺ,No IFN) vs treatment with parental (o, ␣2C and s, ␣21a) and hybrid (p, HY2 and t, HY3) IFN on measles virus replication in primary human lympho- cytes. A total of 1 ϫ 106 cells were resuspended in RPMI 1640 Ϯ 105 pg/ml IFN 24 h prior and during infection with measles virus at multiplicity http://www.jimmunol.org/ of infection of 0.1–1.6 TCID50/ml. Measles virus-infected and control cells were harvested at 72 h postinfection. Supernatants were titrated on VERO cell monolayers. All experiments were performed in triplicate, and the results are expressed as percent of control; 100% for donors 1 and 2 were

4.95 and 5.7 TCID50/ml log10, respectively.

virus replication when titers of treated and untreated cell-associ- ated virus were compared. by guest on September 24, 2021 To examine the cell-surface binding properties of the hybrid molecules, competitive binding analysis was performed with IFN- ␣2c, IFN-␣21a, the hybrids, and IFN-␣2b. Fig. 4 shows the ability of IFN-␣2c, IFN-␣21a, and the three hybrids (HY-1, HY-2, HY-3) to compete for IFN-␣2b binding sites on Daudi and WISH cells, respectively. IFN-␣2c and the hybrid HY-3 inhibited the binding of 125I-IFN-␣2b to Daudi and WISH cells very effectively, while hybrids HY-1 and HY-2 like IFN-␣21a, poorly inhibited 125I-IFN- ␣2b binding on both cell lines. These data suggest that the amino terminal region of the IFN-␣ molecule is important in binding to its receptor.

IFN signaling pathways and transcription analysis Because HY-3 has the highest antiproliferative activity and HY-2 has the lowest on Daudi and WISH cells and primary lymphocytes (Fig. 2), these two constructs were examined to explore potential signaling differences that could explain the dissimilarity seen in their antiproliferative activity. One signaling pathway that is acti- vated by IFN-␣ is the JAK/STAT signaling pathway. Both STAT1

(ƒ). A, Daudi cells. B, WISH cells. Cells were incubated in the presence of indicated concentrations of IFN for 3 days at 37°C. Results are expressed as a percentage of growth relative to a control culture to which medium without IFN was added. Experiments were performed in triplicate (Daudi) and duplicate (WISH). C, Primary human lymphocytes. Antiproliferative effects of the parental IFN constructs IFN-␣2c (Ⅺ) and IFN-␣21a (ƒ) and chimeras. Incubation with IFNs is same as for (ء) the HY- 2 (E) and HY-3 Daudi and WISH cells. Here direct cell counts were performed by analysis FIGURE 2. Inhibition of cell proliferation by IFN hybrids HY-1 (‚), on a Coulter counter. The experiment was performed in triplicate and the HY-2 (E), HY-3(*), as well as parent IFNs IFN-␣2c (Ⅺ) and IFN-␣21a results of the mean are represented. 858 IFN-␣ HYBRIDS EXHIBIT DIFFERENT BIOLOGICAL RESPONSES

FIGURE 5. HY-2 and HY-3 both activate STAT1- and STAT2 -con- taining DNA binding complexes. EMSA of extracts from HY-2- and HY- 3-treated Daudi, Wish, and primary human lymphocytes are shown. Nu- clear extracts (10 ␮g/lane) were prepared from 2 ϫ 107 cells treated for 30 min with either the HY-2 or HY-3 IFN construct at concentrations chosen from the antiproliferative profiles of each of the cells. For the Daudi cells, this was 0.1 pg/ml for HY-3 and 10 pg/ml for HY-2; for the Wish cells, this was 1 ng/ml for HY-3 and 10 ng/ml for HY-2; and for the primary human lymphocytes, this was 1 ng/ml for HY-3 and 100 ng/ml for HY-2. The extracts were incubated with radiolabed GRR oligonucleotide previously Downloaded from shown to specifically bind IFN-␣-activated STAT1 (A) or the ISRE oli- gonuleotide shown to specifically bind the IFN-␣ activated STAT2-con- taining ISGF3 complex (B).

complex seen in Fig. 5 contained STAT1 or STAT2 specifically, http://www.jimmunol.org/ supershift experiments using anti-STAT1 and anti-STAT2 Abs were performed. The ability of the anti-STAT1 Ab to shift the complex is seen in Fig. 6, indicating that this complex does, in fact, contain STAT1. When anti-STAT2 Ab is used, the complex in- duced by either of the IFNs tested is knocked out, which can occur when the Ab recognizes the DNA-binding domain as an epitope. Normal rabbit serum was used as a negative control for all of the supershift experiments and did not effect the mobility of the IFN-

inducible complex. The results indicated that, in all three cell by guest on September 24, 2021 types, HY-2 induced the optimal activation of STAT1 and STAT2 at concentrations at which essentially no antiproliferative activity was observed. These data suggest another pathway may also be activated in addition to the activation of STAT1 and STAT2 to elicit the antiproliferative activity.

FIGURE 4. Competitive binding of IFN-␣ hybrids, IFN-␣2c and IFN- ␣21a on Daudi (A) and WISH (B) cells. The concentration of 125I-IFN-␣2b used in these experiments was 0.25 nM. Hybrids HY-1(‚), HY-2(E), HY- -and parent IFNs IFN-␣2c(Ⅺ) and IFN-␣21a(ƒ) were used as com (ء)3 petitors. The experiments were performed in duplicate.

and 2 are normally activated after IFN receptor occupancy. There- fore, to understand the mechanism governing this activity, we compared the ability of HY-2 and HY-3 to induce STAT1 and 2 activation by EMSA. Two 32P-labeled double-stranded oligonu- cleotide probes were used. One contains a GRR element that spe- FIGURE 6. Supershift analysis of EMSA of HY-2 and HY-3 IFNs. cifically binds STAT1, while the second contains an ISRE that EMSA of whole-cell lysates from HY-2- and HY-3-treated primary human binds STAT2. The data are shown in Fig. 5. A and B show the lymphocytes are shown and experiments performed as in Fig. 4 except that results using Daudi and WISH cells and primary human lympho- rabbit anti-STAT1 (top), anti-STAT2 (bottom), and normal rabbit serum cytes that were treated for 30 min with a concentration of IFN that Abs were added to the reaction and allowed to incubate on ice for 30 min showed a clear antiviral effect for HY-3 and very little or no an- before the addition of either radiolabeled GRR oligonucleotide (top)or tiproliferative effect for HY-2. To validate that the IFN-inducible radiolabeled ISRE oligonucleotide (bottom). The Journal of Immunology 859

activation of STAT1 and STAT2 alone. The gel shift data show that STAT1 and 2 activation and IFN-dependent gene transcription is induced by HY-2 at concentrations where HY-2 shows no an- tiproliferative activity. The antiviral activity was comparable be- tween the two hybrids at concentrations in which no antiprolifera- tive activity was seen. HY-3 was seen to induce STAT1 and STAT2 activation to a greater extent than HY-2 when the cells were treated with equivalent concentrations (data not shown). FIGURE 7. RNase protection assays using IFN-␣2c, IFN-␣21a, and However, the kinetic profiles of STAT1 and STAT2 activation by IFN-␣ hybrids. The parental and hybrid IFN constructs activate an IFN- HY-2 and HY-3 at the various concentrations used in the antipro- ␣-inducible gene to the same extent at concentrations where antiprolifera- liferative assay were identical so that the differences seen between tive activities of HY-2 are not observed. An RNase protection assay was the two hybrids cannot be explained by STAT activation off rates performed using a radiolabeled riboprobe for the ISG15 gene (top)asa (data not shown). Conversely, HY-2 exhibits little or no antipro- marker for the potential of each of the IFNs to induce transcriptional ac- liferative effects at concentrations that induce STAT1 and STAT2 ␮ tivation of an IFN-regulated gene. Then, 15 g of total cellular RNA was activation more robust than that seen with low concentrations of isolated from cells treated for 16 h with the various IFNs at concentrations HY-3 (data not shown). of 50 pg/ml, 5 ng/ml, and 100 ng/ml for the Daudi cells, Wish cells, and the primary PHA-treated human lymphocytes, respectively. A probe for Comparing the sequences of HY-1 and HY-2, there are seven GAPDH was included as a housekeeping gene control (bottom). amino acid differences in the 75–95 region (Fig. 1). Comparisons of published structural and homology models of IFN-␤ and -␣ (16, Downloaded from 26, 27), show that the amino acid region 75–95 forms the C helix, RNase protection assays were used to assess the ability of the which is one of the most exposed helices. The amino acids in this parental and hybrid IFNs to induce several IFN-responsive genes. region may be easily accessible to react with their appropriate No differences were seen among the parental and hybrid IFNs in receptors (28) and thus may be important for biological activity. Di their ability to induce the ISG15 gene in Daudi or WISH cells, and Marco et al. (29) conducted site-directed mutagenesis in the C the induction was comparable for HY-2 and HY-3 in PBLs (Fig. helix of the IFN molecule. They have reported that the mutated http://www.jimmunol.org/ 7). The concentrations of HY-2 and HY-3 used were identical with amino acids (position 84, 86, 87, 90) of helix C of IFN-␣ faces those used in the EMSA experiments and were at levels where no toward one of the subunits of the receptor and are part of one of antiproliferative activity was seen for HY-2 but good antiprolif- two binding sites of IFN-␣. Mutations markedly decreased the erative activity was noted for the other IFNs. In addition, the abil- biological activity on murine cells compared with the unmutated ity of the IFNs to activate IFN-induced gene expression was also protein. Korn et al. (15) and Fish (30) have proposed that there identical for the ISG54, IRF1, GBP, and IP10 genes (data not may be sufficient complexity in the type 1 IFN receptor to account shown). These results indicate that at concentrations where no an- for the differential sensitivities between IFN-␣s and IFN-␤ that

tiproliferative activity was seen for HY-2, this construct retained may be associated with a residue difference in the region 78–95. by guest on September 24, 2021 the ability to induce a variety of IFN-inducible genes. The IFN sequence 78–95 has been implicated as influencing spe- cies specificity among the murine and human IFN-␣s and affecting Discussion the differential specificity of action between human IFN-␣s and ␤ One prominent effect of the IFN-␣s is their ability to inhibit cell IFN- . It is intriguing to speculate that these residue positions may growth, which has also been suggested to be of major importance constitute specific recognition sites associated with an accessory ␣ in determining antitumor action. Grander et al. (25) indicated that component of the IFN- receptor. Our studies suggest that region IFNs can exert profound anticellular effects in vitro, caused by 75–95 is very important for the antiproliferative activity of human ␣ either cell cycle arrest and/or induction of apoptosis. Our studies IFN- , with maximal activity observed when the construct has the ␣ ␣ show human IFN-␣2c, IFN-␣21a, and the hybrids HY-1, HY-2, 2c sequence from 75–95 and the 21a sequence from 96–166. and HY-3 exhibit a broad range (1,000–10,000) of differences in Our competitive binding experiments show that the hybrids or ␣ ␣ their antiproliferative sp. act. on Daudi and WISH cells and human IFN- with the 2c domain at the N terminus (residues 1–95; like ␣ 125 ␣ lymphocytes (see Table I and Fig. 2). HY-3 exhibited the highest IFN- 2c and HY-3) compete well with I-IFN- 2b binding. In ␣ ␣ antiproliferative sp. act. The concentration of HY-3, which inhibits contrast, hybrids or IFN- with the IFN- 21a domain at the N ␣ cell growth by 50%, is 0.0001ng/ml on Daudi cells and 8.0 ng/ml terminus (residues 1–95; IFN- 21a, HY-1, HY-2) show reduced 125 ␣ on WISH cells. However, these IFN-␣s exhibited comparable an- ability to compete with I-IFN- 2b binding. These data suggest tiviral sp. act. on MDBK cells and sp. act. within 7-fold of each that the N terminus is involved in IFN receptor binding interaction. other on WISH cells. Similarly, in primary human lymphocytes, Several reports have proposed two binding sites for IFN-␣, one the antiviral activity of HY-2 and HY-3 are essentially the same which is located in the N-terminal region and would determine the for the concentrations examined (see Fig. 6). Thus, the biological binding to high-affinity receptors. The second site would be in the effects of HY-2 and HY-3 provide evidence for a divergent mech- C-terminal region of the IFN molecule and would influence low- anism by which IFN-␣s induce their antiproliferative and antiviral affinity binding to cells (3, 31, 32). Extensive evidence suggests effects. Therefore, the IFN-␣s may turn on several signaling path- that this receptor system is complex, possibly consisting of either ways to different degrees (5). Recently, it has been reported that multiple receptors or a multisubunit receptor (13, 33). the antiproliferative effects of IFN-␣ are mediated by components In conclusion, our data demonstrates that a portion of the N- of TCR signaling pathways and that activation of the JAK-STAT terminal region of the molecule may be important for the receptor pathway alone is insufficient for the induction of the antiprolifera- binding activities. Furthermore, a domain critical for antiprolifera- tive activity of IFN-␣ (22). Our observations not only support that tive activity appears to exist within the C-terminal region from data but expand it beyond a cell line and a hematopoietic-based amino acid residues 75–166. In our studies the IFN region cell system. In addition our EMSA and RNase protection results [␣2c(75-95)/␣21a(96-166)] appears to be responsible for the suggest that the antiproliferative response may not be directed by higher antiproliferative sp. act. seen with HY-3. In addition, our 860 IFN-␣ HYBRIDS EXHIBIT DIFFERENT BIOLOGICAL RESPONSES data define IFN domains that dissociate anitiviral and antiprolif- 13. Hu, R., Y. Gan, J. Liu, D. Miller, and K. C. Zoon. 1993. Evidence for multiple ␣ erative activities. The hybrids (HY-1, HY-2, and HY-3) had sim- binding sites for several components of human lymphoblastoid - . J. Biol. Chem. 268:12591. ilar antiviral sp. act. on MDBK cells and primary human lympho- 14. Hu, R., and K. C. Zoon. 1994. HuIFNA21 gene expression and properties of cytes but different antiproliferative and binding properties on recombinant IFNA21. J. Interferon Res. 14:S98. Daudi, WISH, and primary human cells. Analysis of gel shift gene 15. Korn, A. P., D. R. Rose, and E. N. Fish. 1994. Three-dimensional model of a human interferon-␣ consensus sequence. J. Interferon Res.12:1. induction and biological responses suggest that the antiprolifera- 16. Mitsui, Y. 1997. Elucidation of the basic three-dimensional structure of type I tive response requires critical signals in addition to those that ac- and its functional and evolutionary implications. J. Interferon Cyto- tivate STAT1 and STAT2. kine Res. 17:319. 17. Saiki, R. K., and R. Higuchi. 1990. PCR Protocols: A Guide to Methods and Applications. Academic Press, New York, NY. Acknowledgments 18. Sanger, F., S. Nicklen, and A. R. Coulson. 1977. DNA sequencing with chain- terminating inhibitors. Proc. Natl. Acad. Sci. USA 74:5463. We thank Joan C. Enterline for performance of antiviral assays, Valerie 19. Horton, R. M., H. D. Hunt, S. N. Ho, J. K. Pullen, and L. R. Pease. 1989. Calvert for elutriation of lymphocytes, and Raj Puri and Barbara Rellahan Engineering hybrid genes without the use of restriction enzymes. Gene 77:61. for review of the manuscript. 20. 1997. The QIA expressionist: the high level expression and protein purification system. Qiagen, Santa Clarita, CA. 21. Janknecht, R., G. DeMartynoff, J. Lou, R. A. Hipskind, A. Nordhei, and References H. G. Stunnenberg. 1991. Rapid and efficient purification of native, histidine- tagged protein expressed by recombinant vaccinia virus. Proc. Natl. Acad. Sci. 1. Hayes, M. P., and K. C. Zoon. 1994. Production and action of interferons: new USA 88:8972. insights into molecular mechanisms of gene regulation and expression. In 22. Petricoin, E. F., S. Ito, B. L. Williams, S. Audet, L. F. Stancato, A. Gamero, Progress in Drug Research, Vol. 43. Ernst Jucker, ed. Birkhauser Verlag, Berlin, K. Clouse, P. Grimley, A. Weiss, J. Beeler, D. S. Finbloom, E. W. Shores,

p. 239. R. Abraham, A. C. Larner. 1997. Antiproliferative action of interferon-␣ requires Downloaded from 2. Weiss, K. 1998. Safety profile of interferon-␣ therapy. Semin. Oncol. 25 (Suppl. components of T-cell receptor signalling. Nature 390:629. 1):9. 23. Larner, A. C., M. David, G. M. Feldman, K. Igarashi, R. H. Hackett, D. S. Webb, 3. Streuli, M., A. Hall, W. Boll, W. E. Stewart, S. Nagata, and C. Weissmann. 1981. S. M. Sweitzer, E. F. Petricoin III, and D. S. Finbloom. 1993. Tyrosine phos- Target-cell specificity of two species of human interferon-␣ produced in E. coli phorylation of DNA binding proteins by multiple cytokines. Science 261:1730. and of hybrid molecules derived from them. Proc. Natl. Acad. Sci. USA 78:2848. 24. Zoon, K. C., D. Zur Nedden, and H. Arnheiter. 1982. Specific binding of human ␣ 4. Horisberger, M. A., and S. DiMarco. 1995. Interferon- hybrids. Pharmacol. ␣ interferon to a high affinity cell surface binding site on bovine kidney cells. Ther. 66:507. J. Biol. Chem. 257:4695. 5. Rehberg, E., B. Kelder, E. G. Hoal, and S. Pestka. 1982. Specific molecular 25. Grander, D., O. Sangfelt, and S. Erickson. 1997. How does interferon exert its http://www.jimmunol.org/ activities of recombinant and hybrid leukocyte interferons. J. Biol. Chem. 257: cell growth inhibitory effect? Eur. J. Haematol. 59:129. 11497. 26. Walter, M. R. 1997. Three-dimensional models for interferon-␣ subtypes IFN- 6. Meister, A., G. Uze, K. E. Morgensen, I. Gresser, M. G. Yovey, M. Gruetter, and conI, IFN-␣8 and IFN-␣1 derived from the crystal structure of IFN-␣2b. Semin. F. Meyer. 1986. Biological activities and receptor binding of two human recom- Oncol. 24:S52. binant interferons and their hybrids. J. Gen. Virol. 67:1633. 27. Viscomi, G. C. 1997. Structure-activity of type I interferons. Biotherapy 10:59. 7. Fidler, I. J., R. Heicappell, I. Saiki, M. G. Grutter, M. A. Horisberger, and J. Nuesch. 1987. Direct antiproliferative effects of recombinant human interfer- 28. Murgolo, N. J., W. T. Windsor, A. Hruza, P. Reichert, A. Tsarbopoulos, ␣ S. Baldwin, E. Huang, B. Pramanik, S. Ealick, and P. P. Trotta. 1993. A homol- on- B/D hybrids on human tumor cell lines. Cancer. Res. 47:2020. ␣ 8. Sperber, S. J., S. B. Hunger, B. Schwartz, and S. Pestka. 1993. Anti-rhinoviral ogy model of human interferon- 2. Proteins Struct. Funct. Genet. 17:62. activity of recombinant and hybrid species of interferon ␣. Antiviral Res. 22:121. 29. DiMarco, S., M. G. Grutter, J. P. Priestle, and M. A. Horisberger. 1994. Muta- ␣ 9. Mitsui, Y., T. Senda, T. Shimazu, S. Matsuda, and J. Utsumi. 1993. Structural, tional analysis of the structure-function relationship in interferon- . Biochem. functional and evolutionary implications of the three-dimensional crystal struc- Biophys. Res. Commun. 202:1445. ture of murine interferon-␤. Pharmacol. Ther. 58:93. 30. Fish, E. N. 1992. Definition of receptor binding domains in interferon-␣. J. In- by guest on September 24, 2021 10. Muller, U., U. Steinhoff, L. F. L. Reis, S. Hemmi, J. Pavlovic, R. M. Zinkernagel, terferon Res. 12:257. and M. Aguet. 1994. Functional role of type I and type II interferons in antiviral 31. Fish, E. N., K. Banerjee, and N. Stebbing. 1989. The role of three domains in the defense. Science 264:1918. biological activity of human interferon-␣. J. Interferon Res. 9:97. 11. Zav’Yalov, V. P., and G. A. Zav’yalov. 1997. Interferons ␣/␤ and their receptors: 32. Kontsek, P., L. Borecky, E. Kontsekova, I. Macikova, A. Kolcunova, M. Novak, place in the hierarchy of cytokines. APMIS 105:161. and V. Krchnak. 1991. Mapping of two immunodominant structures on human 12. Zoon, K. C., D. Miller, J. Bekisz, D. ZurNedden, J. C. Enterline, N. Nguyen, and interferon ␣2c and their role in binding to cells. Mol. Immunol. 28:1289. R. Hu. 1992. Purification and characterization of multiple components of human 33. Colamonici, O., and L. M. Pfeffer. 1991. Structure of the human interferon ␣ lymphoblastoid interferon-␣. J. Biol. Chem. 267:15210. receptor. Pharmacol. Ther. 52:227.