Differential Effects of Recombinant Human Leukocyte Interferons on Cell Surface Antigen Expression John W

[CANCER RESEARCH 46, 4984-4990, October 1986] Differential Effects of Recombinant Human Leukocyte Interferons on Cell Surface Antigen Expression John W. Greiner,1Paul B. Fisher, Sidney Pestka, and Jeffrey Schlom

Laboratory of Tumor Immunology and Biology, NJH/National Cancer Institute, Bethesda, Maryland 20892 [J. W. G., J. SJ; Department of Microbiology, Comprehensive Cancer Center/Institute of Cancer Research, Columbia University, College of Physicians and Surgeons, New York, New York 10032 [P. B. F.]; and Roche Institute of Molecular Biology, Nutley, New Jersey 07110 fS. P.]

ABSTRACT interferons, several studies have found significant differences with respect to their biological actions. Interferons are generally Human leukocyte (a) interferon (IFN-a) is composed of a multigene regarded as having antiviral, antiproliferative, differentiation- family within which at least eight different species have been expressed modulatory, and immunomodulatory activities. Studies have in Escherichia coli, isolated, and shown to exert a wide range of biological activities on different human target cells. In this study we utilized eight shown substantial quantitative differences among these species species of IFN-a (A, B, C, D, F, I, J, and K) and investigated then- of recombinant leukocyte interferon in their varied biological respective capabilities to alter the proliferation of a human breast carci and biochemical effects on a range of target cells (9-17). Other noma cell line (MCF-7). The antigens studied were all constitutively studies have demonstrated that the individual species of leu expressed on the MCF-7 cell surface: the M, 180,000 carcinoembryonic kocyte interferon can elicit different antigrowth effects on the antigen; a high molecular weight (>10*) glycoprotein, termed tumor- same target cell (9, 11, 15, 16-19). It was also shown that a associated glycoprotein 72; and a major HLA histocompatibility antigen. single leukocyte interferon species was an effective antiviral The level of expression of each antigen was measured by the binding of agent yet was completely inactive in boosting human natural monoclonal antibodies HI.I, B72J, and W6/32, respectively. A high killer activity (10). Therefore, as originally observed with the degree of diversity was found among the various IFN-a species with respect to their ability to enhance antigen expression and inhibit MCF-7 various natural leukocyte interferon species (15) there seem to cell growth. The two most potent species, IFN-aA and IFN-aB, were exist quantitative and qualitative differences with respect to found to increase the expression of tumor antigens as well as the HLA their abilities to regulate a variety of biological properties of determinant by 2-5-fold. In contrast, IFN-aD and IIVÂ«.) were virtually target cells. inactive in altering antigen expression but did inhibit the growth of MCF- Considerable attention has been focused on the ability of the 7 cells. The remaining IFN-a species, -Â«('.-<Â»!â€¢',-a!,and -aK, exerted an interferons to modulate surface antigens on a variety of human intermediate range of activities for both antigen enhancement and inhi cells (reviewed in Ref. 18). Partially purified as well as recom bition of MCF-7 cell growth. The relative ability of each species of 11N- binant leukocyte interferon can enhance the expression of class a to inhibit MCF-7 cell growth appeared to be independent of their I histocompatibility antigens on both normal and transformed effectiveness in augmenting antigen expression. IFN-aD and IFN-oJ, human cells (18, 20-26). Our laboratories have reported that the two species that failed to alter tumor antigen expression, did, however, IFN-aA can increase the binding of MAbs to the surface of seem to interact with the interferon receptor since they inhibited MCF-7 cell growth and competed with other IFN-a species for the increase in human breast and colon carcinoma cells (27, 28). We have carcinoembryonic antigen, tumor-associated glycoprotein 72, or HLA shown that such an increase is a result of the enhanced expres expression. A comparison of the concentrations of each IFN-a necessary sion of tumor antigens, such as the M, 180,000 CEA and the high molecular weight (>106) mucin, TAG-72, which react with to enhance antigen expression revealed that the surface HLA determinant was approximately 10-fold more sensitive to enhancement than was the MAbs B1.1 and B72.3, respectively (18, 24, 26). The anti-CEA tumor antigen, carcinoembryonic antigen. The individual members of the and TAG-72 MAbs are currently being used in several areas in IFN-a family thus differ extensively in their ability to alter the level of the management of human cancers, including (a) serum assays antigen expression on the surface of MCF-7 breast carcinoma cells. The for the detection of antigen; (/>) immunohistochemical assays differential response of these cells to the IFN-a species, which share a for the detection of occult tumor cells in pleural effusions, high degree of sequence homology and bind to the same cell membrane ascites, and fine needle aspirate biopies (29-31); and (c) the receptor, suggest that these biologically related compounds may differ in detection of in situ occult tumors by radio labe led MAbs (32, the biochemical and molecular signals induced distal to binding to the surface interferon receptor. 33). The use of recombinant interferon to augment tumor antigen expression and thereby enhance detection by MAbs for diagnosis and treatment in each of the above situations merit INTRODUCTION consideration. The present study was carried out to evaluate Human IFN-a2 consists of a family of individual species with the different species of IFN-a for their abilities to alter cell proliferation and modulate the level of cell surface antigen amino acid residues that have been shown to differ by as much expression. The antigens monitored were the M, 180,000 CEA, as 20% (1-8). At least eight different species of leukocyte a high molecular weight (>106) mucin termed TAG-72, and interferons have been expressed in Escherichia coli and purified HLA, all of which are constitutively expressed by the human and their biological activities have been compared (6, 9, 10). breast carcinoma cell line, MCF-7. The levels of expression of Besides the difference in amino acid composition of these these antigens were measured by the binding of MAbs Bl.l, Received 4/10/86; revised 6/27/86; accepted 7/1/86. B72.3, and W6/32, respectively (Table 1). 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. 1To whom requests for reprints should be addressed, at Laboratory of Tumor MATERIALS AND METHODS Immunology and Biology, NIH/NCI, Bldg. 10, Room 8B07, Bethesda, MD 20892. Recombinant Human Leukocyte Interferons. The isolation, expres 'The abbreviations used are: IFN-a, (alpha) interferon; MAIÂ»,monoclonal sion, and purification of eight different clones of human leukocyte antibodies; CEA, carcinoembryonic antigen; TAG-72, tumor-associated glycopro interferon have been described (2, 5-7, 11). The interferons were tein 72; USA, bovine serum albumin; RIA, radioimmunoassay; Ali,,,, molar concentration or molecules per cell required to induce a 50% increase in HLA or prepared and purified as described (8, 9, 11, 34). Unless otherwise tumor antigen expression. noted, the specific activity of the preparations on MDBK cells was 1- 4984

5 x 10s antiviral units/mg protein with respect to the human IFN-a M!of 2 N NaOH and the amount of monoclonal antibody bound to the reference standard. Preparations with specific activities ^\ x IO8units/ cell surface (i.e., I25l-labeled radioactivity) was determined in a gamma ml were at least 90% homogeneous as determined by sodium dodecyl counter. Control MCF-7 cells received either a control MAb, the mouse sulfate-polyacrylamide gel electrophoresis (34). IFN-aB represented myeloma IgG (termed MOPC-21), or RPMI 1640 containing 1% BSA. about 75% of the total protein. IFN-aF was a crude bacterial extract. The background radioactivity (<300 cpm/well) was subtracted from The interferons were diluted with RPMI 1640 containing 25 mM 4-(2- the wells that received the monoclonal antibodies of interest. Finally, hydroxyethyl)-l-piperazineethenesulfonic acid and 1% bovine serum cells from 6-10 wells were trypsinized and counted and the total cpm albumin and stored at -70Â°C. Periodically an aliquot of each IFN-a bound was normalized for 5 x IO4cells. species was rechecked for antiviral activity which remained virtually Solid Phase RIA. The immunoreactivities of MAbs Bl.l, B72.3, and unchanged with storage at -70*C. Prior to use an aliquot of each was W6/32 were also determined in whole cell extracts with a solid phase thawed, diluted, and added to the growth medium at the indicated RIA. Untreated and interferon-treated MCF-7 cells were routinely antiviral titers. scraped from T-75 flasks and pelleted by centrifugation at 1000 x g. Hybridoma Methodology. The details of the generation and charac The medium was removed and the cell pellet was resuspended and terization of the MAbs B1.1 and B72.3 have been reported (26-28). homogenized for 2-3 rain on ice in 10 mM Tris-HCl (pH 7.2)-0.2 HIM Briefly, B1.1 is an IgG2a that recognizes a M, 180,000 CEA, whereas CaCl2 (10 g, wet weight/100 ml). The homogenate was further disrupted B72.3 is an IgGl that reacts with a high molecular weight (>106) with a cell bomb (Parr Instrument Co., Moline, IL) for 5 min at 1000 glycoprotein antigen complex termed TAG-72. Bl.l has been shown lb/iir. The disrupted homogenate was sonicated on ice for 1-2 min to react primarily with human breast and colon carcinoma cells, while and centrifuged at 10,000 x g for 10 min. The protein concentration some reactivity has been found with melanomas and selected normal of the supernatant (i.e., cell extract) was determined by the method of tissues. B72.3 has a more restrictive range of reactivity that includes Lowry et al. (40). The cell extracts were diluted to l mg/ml in Dulbec co's phosphate-buffered saline containing calcium and magnesium (pH primarily human breast, colon, and ovarian carcinoma tissues (29, 35, 36). The W6/32 monoclonal antibody is IgG2a that recognizes the 7.2); then 50 pi were added to each well of a 96-well microtiter plate ABC loci of the HLA major histocompatibility complex (37) and was and dried. The solid phase RIA was carried out essentially by the same purchased from Cappell Laboratories (Malvern, PA). procedure outline for the surface antigen expression (40). The radiola- Cells and Growth Inhibition Assay. The MCF-7 cell line was gener beled goat anti-mouse IgG was added at a concentration of 75,000 ated from a pleural effusion of a primary breast cancer patient and was cpm/25 n\ of phosphate-buffered saline containing 1% BSA. After the 1-h incubation the plates were washed and the bound cpm were deter obtained from the Breast Cancer Task Force (National Cancer Institute, NIH). The cell line was routinely subcultured in Dulbecco's modified mined by cutting the wells and measuring the radioactivity in a gamma Eagle's medium containing insulin (5 /jg/nil). 1 mM sodium pyruvate, counter. Ix nonessential amino acids, 10% heat-inactivated fetal bovine serum, and gentamicin (50 ftg/ml). All determinations of antigen expression RESULTS were performed on MCF-7 cells between passages 125 and 140. The MCF-7 cells were routinely tested for Mycoplasma and were con Fig. 1 shows the results of a representative experiment in sistently negative. which the relative effect of the various species of IFN-a on the To assess the antiproliferative actions of the recombinant leukocyte growth of MCF-7 cells were compared. IFN-aA, -aB, -aD, and interferons, the MCF-7 cells were seeded at a density of 5 x 10s cells/ -aJ were chosen because the IFN-aA and -aB clones were T 25 flask in complete growth medium. The cells attached and grew among the most potent while the IFN-aD and -aJ clones were for 48 h at which time the complete growth medium was removed and the least effectiveleukocyte interferons in inhibiting the growth medium with and without the appropriate titers of the different recom of MCF-7 cells (see Table 2). All four IFN-a species exhibited binant interferons was added. After 5 additional days the cells were a dose-dependent relationship showing demonstrable inhibition trypsinized, resuspended in serum-free RPMI 1640, and counted with a hemocytometer. Each control and interferon-treated group consisted of MCF-7 cell growth with 10 antiviral units of interferon. IFN-aB was substantially more effective than the other three of at least four T 25 flasks. Determination of Cell Surface Antigen Expression. The MCF-7 cells IFN-a species in inhibiting MCF-7 cell growth at low (<100 were maintained in T-75 flasks. Subconfluent cells were harvested with units) antiviral titers. For example, 10 units of IFN-aB reduced 0.1% trypsin-0.05 mM EDTA; then 5 x 10* MCF-7 cells in 0.1 ml MCF-7 cell growth by more than 40%, whereas the same complete medium with and without interferon were seeded in each well antiviral titer of IFN-aA, -aD, or aJ resulted in a reduction of of a 96 well microtiter plate. The cells were incubated for 24-36 h in a cell growth by approximately 10%.At higher antiviral titers the humidified atmosphere of 95%:5% air:CO2. Due to the effect of cell IFN-aA clone was as effectiveas the IFN-aB clone in suppress cycle on surface antigen expression (38), all assays were performed with ing MCF-7 cell growth. Incubation with either of these two subconfluent, exponentially growing cells. A complete description of IFN-a species at antiviral titers greater than 1000 units reduced this radioimmunoassay to detect antigen expression on live cells has MCF-7 cell growth by 50-55%. In contrast, the IFN-aD and been described (39). Briefly, the growth medium was removed and 100 -aJ species were clearly less effective in reducing MCF-7 cell n\ of RPMI containing 10% BSA and 0.08% sodium azide were added for 1 h. This "blocking" medium was aspirated and each purified growth. Maximum growth inhibition with IFN-aD and IFN- aJ was obtained with 500-1000 antiviral units and any addi monoclonal antibody diluted in 50 n\ of RPMI 1640 containing 1% BSA was added for 1 additional h. Routinely, an antibody dilution tional increase in dose did not result in a further significant curve which ranged from 500 to 0.3 ng of each antibody was performed reduction of growth. Table 2 summarizes the antiviral titers of for each experiment. The unbound antibody was removed and 75,000 each IFN-a species required to achieve a 25% reduction in cpm of I2sl-labeled purified goat anti-mouse IgG was added in 50 n\ to MCF-7 cell growth. There was a greater than 100-fold range in each well for 1 h. After extensive washing the cells were lysed with 100 the antigrowth capabilities of these eight IFN-a species. As

Table 1 Characteristics of monoclonal antibodies AntibodyW6/32 43,000 HLA-ABC loci human cells, except RBC Bl.lB72.3AntigenM,M, 180,000 carcinoembryonic an IgG2aIgGlReactivityAllHumanmelanomasHuman breast and colon carcinomas, some 26-2826, tigen M, >10* glycoprotein complex breast, ovarian, and colon carcinomas; no reactivity to 27, 29, 35 (TAG-72)IsotypeIgG2a normal human cellsRef.37 4985

1000 10.000 100 1.0 0.1 001 INTERFERONCONC. (antiviral unte/mil ngB 11/well Fig. l. Effects of different IFN-a species on the in vitro growth kinetics of the MCF-7 cells. An in vitro growth inhibition assay was used to assess the effects of IFN-aA, -aB, -aD, and -oj on MCF-7 cell growth. The assays were carried out as outlined in "Materials and Methods." MCF-7 cells were transferred to T-25 10r flasks (5 x 10* cells/flask) and grown in complete medium for 2 days. The complete medium was then changed to medium containing various antiviral titers of the different IFN-os. The cells were grown for an additional 5 days at which time cells from each group were harvested and counted with a hemocytometer. The mean number of cells per group was determined from 4 flasks. The range of cell number for the untreated MCF-7 flasks was 4.4-6.1 x id" cells/flask. The data are presented as the means of triplicate experiments.

Table 2 Comparison of the antiproliferative properties of the different recombinant leukocyte interfÃ©ronsona human breast carcinoma cell MCF-7 cell growth inhibition IFN-, (I.'0" (antiviral units/ml) A (25-60)* B 2(0.5-12) CD 60(40-110) 250(180-410) 500 5 .5 005 33 (25-80) FI ng W6/32/W6II 92(80-125) J 75 (60-95) Fig. 2. Effects of IFN-aA, -aB, and -aD on the expression of CEA (A) and K38 58 (45-70) HLA (B) on the MCF-7 cell surface. MCF-7 (5 x IP/well) were added to 96- well mit n niter plates in growth medium alone (â€¢)or containing 1000 antiviral * MCF-7 (5 x IO9cells) cells were added in complete medium to T-25 flasks. units of IFN-aA (A), -aB (Â»),or -aD (T). After a 24-30-h treatment the binding After 48 h 10-2000 antiviral units of each recombinant species of leukocyte of MAbs B1.1 and W6/32 was measured as a function of antibody concentration. interferon were added. After 5 days in the interferon-containing growth medium, Data presented are the means of 3-5 determinations for each species of IFN-a. the cells were removed by trypsinization and counted with a hemocytometer. I?Â« represents the approximate antiviral titer for each leukocyte interferon which resulted in a 25% inhibition of cell growth. -aB. It should be noted that the increase in HLA expression * The values are shown as the mean of 3-5 experiments for each leukocyte (i.e., W6/32 binding) (4-5-fold) in the cell extracts from inter- interferon species with the appropriate range in parentheses. The values for UN aB were estimated by using 1-1000 units for testing. feron-treated MCF-7 cells was much greater than that detected on the cell surface (Table 3). Other IFN-a species, in particular shown, IFN-aB, the most potent species, inhibited MCF-7 cell -aC, -aF, -Â«I,and-aK, were also found to augment the level of growth by 25% with the addition of only 2 units/ml. For expression of CEA, TAG-72, and HLA on the MCF-7 cell comparison, it requires 125-fold more units of IFN-aD (250 surface. For example, treatment with 1000 units of IFN-aK units) to evoke the same degree of growth inhibition. The overall enhanced B1.1, B72.3, and W6/32 binding to the MCF-7 cell order of potency of these eight IFN-a species for the inhibition surface by 1.7-, 1.7-, and 2.4-fold, respectively. Of the eight MCF-7 growth wasB>A = F>K = OJ>IÂ»D. Removal IFN-a species tested, IFN-aD and -oJ, which were among the of any of these IFN-a species from the MCF-7 growth medium least effectivefor inhibiting MCF-7 cell growth, were relatively reversed their antiproliferative effects and resulted in the re ineffectivein enhancing HLA expression and even less effective sumption of normal MCF-7 growth kinetics. in altering the levelof tumor antigen expression. Whereas other As demonstrated previously (18, 24), MCF-7 cells express IFN-a species enhanced cell surface HLA expression at least two distinct tumor antigens, CEA and TAG-72, as well as the 2-fold, IFN-aD and -aJ at antiviral titers of greater than 1000 class I major histocompatibility HLA complex on their surface. units/ml could boost HLA expression by only 30%(Fig. 2B; Fig. 2 shows the MAb dilution curves for the reactivity of Bl.l Table 2). Similarly, analysis of whole cell extracts from the and W6/32 to the surface of MCF-7 cells following treatment same IFN-aD- or -aJ-treated cells also revealed a modest 30- with IFN-aA, -aB, or -aD. A 30-h incubation of MCF-7 cells 60% enhancement of W6/32 binding. As noted above, these in the presence of 1000 units IFN-aA or -aB resulted in two IFN-a species (-aD and -aJ) were found to be inactive for approximately a 2-fold enhancement of Bl.l binding to the increasing the expression of CEA or TAG-72. Analysis of the MCF-7 cell surface (Fig. 2A). Similarly, incubation with either reactivity of both MAbs Bl.l and B72.3 to the MCF-7 cell IFN-a species also increased the cell surface expression of the surface (Fig. 2; Table 3) and whole cell extracts (Table 3) high molecular weight TAG-72 antigen by 2-2.5-fold (Table revealed little or no change in the levelof Bl.l or B72.3 binding 3). An analysis of the whole cell extracts revealed an increase following incubation with 2000-5000 antiviral units of either of the two tumor antigens as well as HLA expression following IFN-aD or IFN-aJ. treatment with 10 or 1000 antiviral units of either IFN-aA or The molecular weights and specific activities (i.e., 1-5 x 10* 4986

Table 3 Comparison of the effects of leukocyte interferons on antigen reactivity to the surface and whole cell extracts of human breast carcinoma cells reactivity"CEA5,8607,280 reactivity*CEA6,9059,240(1.3)cell extract Concentration (units/ml)Control,Interferon species NoneA untreated

10 (1.2)c 1,000B 10,840(1.9)6,420(1.1)3,010(2.0)NT*1,510(1.0)9,440(2.2)7,410(1.7) 14,490(2.1)NT 1,860(1.8)NTNT 10,560(5.1)NTNT

10 100 9,220(1.6) 7,800(1.7) NT11,650(1.7)NTNT 1,000C 9,840(1.7)9,240 2,550(1.7)2,110(1.4)1,410(1.0)10,410(2.4)8,400 1,580(1.5)NTNTNT9,140(4.4)NTNT

1,000D (1.6)5,860 (2.0)4,480(1.0)

500 (0.8) 1,000 6,040(1.0) 1,550(1.0) 5,550 (1.3) NT NT2,870(1.4) 2,000 5,640(1.0) 1,600(1.1) 5,710(1.3) NT NT 5,000F 5,190(1.1)8,410(1.4)10,210(1.7)6,170(1.1)1,330(0.9)1,940(1.3)3,140(2.1)1,660(1.1)5,410(1.3)10,880 6,620(1.0)NTNTNT1,220(1.2)NTNTNT 2,740(1.3)NTNTNT3,340(1.6)NT

1,000I (2.6)9,990(2.3)5,070(1.2)

1,000J

1,000 2,000K 6,040(1.0)10,080(1.7)TAG-721,4801,725(1.2)1,550(1.1)2,440(1.7)HLA4,2756,660(1.6)5,840(1.4)10,140(2.4)Whole7,000(1.0)NTTAG-721,0501,420(1.4)1,220(1.2)NTHLA2,0706,810(3.3)

1,000Surface "MCF-7 cells were plated in 96-well mictotiter plates (5x10* cells/well) in complete medium with or without the indicated concentrations of the various leukocyte interferons. A titration curve was determined for the binding of each monoclonal antibody from 500 to 0.03 ng/well. Maximum binding occurred with the addition of 20 ng W6/32,100 ng B1.1, and 500 ng B72.3 to their respective surface antigens. Data represent the mean of the different experiments in which the standard error was less than 15%. * Whole cell extract of untreated and interferem-Ireated MCF-7 cells as described in "Materials and Methods." The binding of each monoclonal was determined as stated previously but with 5 fig of extract protein dried to each well of a 96-well plate. ' Numbers in parentheses, increase in MAh binding as determined by cpm of the inter fenm -treated cells/cpm of the untreated cells. ' NT, not tested. antiviral units/mg protein) are known for each purified species for HLA and CEA were greater than 350- to 600-fold higher of IFN-a. Data on the antigrowth and antigen augmentation of than the values for IFN-aA and -aB. The overall ranking of the the various IFN-a species have been presented in terms of eight different IFN-a species for enhancement of both HLA antiviral units of interferon. The data are shown as AE5oin and CEA expression was A = B>I>F>C = K >Â» J, D. Table 4. Table 4 permits a more reliable comparison among This ranking represents the relative potencies of the IFN-a the various IFN-a species for their enhancement of surface species for the enhancement of both the HLA antigen and the antigen expression. The AEjo for HLA expression requires tumor antigen, CEA. substantially less IFN-a than for a similar increase in the The combination of weak antiproliferative and relatively in expression of CEA. For example, 10 times more interferon effective tumor antigen-enhancing activities by IFN-aD, in molecules per cell are required to induce a 50% increase in B1.1 particular, might be explained by inefficient binding of this (CEA expression) binding with IFN-aA or IFN-aB than to IFN-a species to the MCF-7 surface interferon receptor. This elicit the same increase in W6/32 binding. In addition, there is hypothesis was tested by preincubating MCF-7 cells in the a wide range of potencies within the different species of IFN-a presence of IFN-aD (1000 units/ml) followed by the addition with respect to their abilities to increase HLA and CEA expres of IFN-aA (1000 units/ml). As shown previously, 1000 units sion. IFN-aA and -aB were the most effective in increasing IFN-a A/ml resulted in a significant increase in B1.1 binding to HLA expression by 50% following the binding of 1.2 x 10"and CEA (approximately 2-fold) on the surface of MCF-7 cells. A 3.6 x IO4 molecules/cell, respectively. In contrast, approxi 2-h preincubation with IFN-aD or -aJ completely blocked the mately 10-fold more of each of these interferons was required IFN-aA-induced increase in B1.1 binding (Fig. 3, IFN-aD data to increase CEA expression by 50% (i.e., 110,000 and 460,000 shown). These findings indicate that IFN-aD can interact with molecules/cell, respectively). IFN-aD and -aJ were virtually the MCF-7 cell membrane and block the enhancement of tumor inactive in enhancing the expression of CEA so that their AEsoS antigen expression induced by IFN-aA. It was also observed

Table 4 Comparison of the molar amounts of the purified interferons required to increase surface HLA and CEA expression on human breast carcinoma cells AI;M|)ULeukocyte HLA expression ( (AEM)Concentrationexpression Antiviral IFNpg/ml18.9 of of mole IFNpg/ml177 of of mole interferon titer cules/cell titer cules/cell species(units/ml)A (x IO"*)1.2 (units/ml)94 (x IO-4)11.1 10 B 7 58.3 3.046.4 3.6 90 750 38.5 46.0 C 90 900 55.8 860 8.6 x IO3 443.2 533.8 D >5000 >3x 10* >\0> >103 >5000 >3x IO3 >103 >103 F 46 511 26.5 31.9 1250 1.4 x 10* 725 867.0 I 18 150 7.8 9.4 540 4.5 x IO3 233.7 281.7 J >3000 >2 x 10* >103 >103 >5000 >3x 10s >1(P >103 K 141Concentration1175PM0.9 59.5No. 71.9Antiviral 480CEA 4X IO3PM9.2 202.6No. 245.3 * A 50% increase in the expression of HLA or CEA as measured by the binding of MAbs W6/32 and Bl.l was considered significant. Therefore, the antiviral titer of each leukocyte interferon required to boost each antigen expression by 50% is termed the Al-,â€ž.Calculationswere performed as described previously (11,13, 14). 4987

The results of the present and previous studies (21, 23, 24) 14 suggest that the mechanisms by which interferon enhances lu2

10 1.0 0.01 face HLA may be a result of an increased synthesis of this ng B1.I/WELL surface antigen (20,21,23). The augmentation of tumor antigen Fig. 3. Inhibition by IFN-aD of the IFN-aA-induced BI.I binding to the expression by IFN-aA is different from the enhancement of surface of MCF-7 cells. MCF-7 cells were incubated for 36 h in the presence of HLA in several aspects, (a) MAb B1.1 reacts with 40-60% of 1500 antiviral units of IFN-aA/ml and the level of B1.1 binding to the cell surface was measured by the live cell RIA outlined in "Materials and Methods." In the the MCF-7 cells, but following IFN-aA treatment more than same experiments, MCF-7 cells were pretreated with 1500 units IFN-aD/ml for 90% of the cell population is CEA positive. This is a result of 2 h after which the IFN-aD-containing medium was removed and the cells were recruitment of antigen-negative cells to become antigen posi washed with complete MCF-7 growth medium. After a washing, the cells were treated for 36 h with growth medium containing 1500 units II NHA and B1.1 tive, resulting in a more homogeneous cell population with binding was measured as described previously. Control cells received RPMI 1640 respect to the expression of that particular tumor antigen, (b) containing 1%BSA. Data represent the mean of each treated and untreated group Analysis of single cell subclones of the MCF-7 cell line indicate from at least three separate experiments. that three different classes of cells exist (41): cells expressing that the inhibition of the H N <,.\ mediated increase in B1.1 either high or low levels of constitutive tumor antigen expres sion which are enhanced following IFN-aA treatment; cells that binding by IFN-aD could be abrogated by substantially increas ing the titer of the IFN-aA to >25,000 units/ml which suggests express a constitutive level of antigen but do not respond with increased expression by IFN-aA treatment; and cells that are the competitive nature of the IFN-aD inhibition. negative for a particular tumor antigen either before or after interferon treatment. These results suggest that the parental DISCUSSION MCF-7 cell line consists of a heterogeneous population of cells Treatment of the breast carcinoma cell line, MCF-7, with which vary in their capacity to respond to IFN-aA with in eight different species of IFN-a which display a high degree of creased tumor antigen expression, (c) The data presented in amino acid sequences homology with complete identity at 85 dicate that IFN-aD and -aJ can induce a modest increase in of their 165 amino acids (for review see Ref. 8), results in both HLA expression, whereas both species are unable to alter tumor quantitative and qualitative differences in antiproliferative ac antigen expression. Because of the previous lack of availability tivity and enhancement in the cell surface expression of tumor of homogeneous IFN-a preparations of known specific activity, and HLA antigens. Six of eight IFN-a species were virtually it was possible only to evaluate IFN-a enhancement of tumor equipotent for boosting the level of HLA expression as mea antigen and HLA antigen expression using antiviral titers. sured by the binding of MAb W6/32 (2-3-fold enhancement). However, with the advent of recombinant DNA techniques and In contrast, there seems to be a broader range of activities improved biochemical purification procedures homogeneous among the different IFN-a species with respect to their abilities preparations of various IFN-a species are available (6, 8) and to enhance tumor antigen expression. IFN-aA, -aB, and -ol permit an evaluation of the functional and molecular activities induced the most dramatic increases in CEA and TAG-72 of these biological agents. Such data can demonstrate in abso expression on the surface of the MCF-7 cells. Treatment with lute terms whether tumor antigens expressed by human carci IFN-aC, -aF, or -aK evoked an intermediate elevation in B1.1 noma cells are more susceptible to enhancement by the IFN-as and B72.3 binding to their respective tumor antigens. An inter than the normal HLA surface antigen. A comparison of IFN- esting observation was that although at high antiviral titers aA and -aB for increasing the expression of the two tumor IFN-aD and -aJ could induce a small increase in HLA expres antigens, CEA and TAG-72, and HLA indicated that the AE50 sion, these species were incapable of evoking any change in the for expression of the normal cell surface determinant could be expression of either tumor antigen. As stated previously, other achieved by the administration of 10-fold less IFN-a than investigators have reported a wide diversity among the IFN-a required for the same level of enhancement of either tumor species for eliciting a particular biological function (i.e.. anti antigen. Therefore, these observations coupled with the disso viral, antigrowth, augmentation of the natural killer activity, ciation of tumor antigen-enhancing activities from that of the augmenting tumor antigen expression, and inducing alterations HLA antigen-enhancing activities for IFN-aD and -aJ suggest in cellular differentiation). In particular, Ortaldo et al. (10) that multiple, if not different, mechanisms are responsible for reported that the J species of IFN-a lacked any ability to boost the interferon-mediated increase in tumor-associated than those natural killer activity. It would be attractive to postulate that, modulating the expression of normal cell surface antigens. since IFN-aJ fails to alter natural killer activity and cell surface The lack of antigen-enhancing ability of the IFN-aD and -aJ tumor antigen expression yet exhibits effective antiviral and species may be a result of: (a) preferential inactivation of these antigrowth activity, it represents an interferon species that is IFN-a species by the MCF-7 target cell; (b) the inability of the functionally inactive as an immunomodulatory compound. cell surface interferon receptor to recognize these compounds More studies with a variety of human target cells are necessary leading to ineffective binding to the receptor; and/or (c) their to substantiate these findings. Furthermore, the IFN-aD was binding to a component of the interferon receptor which cannot as ineffective as the IFN-aJ for enhancing tumor antigen elicit the required transmembrane signals to ultimately evoke expression, but in contrast to IFN-aJ, this species was a potent alterations in cell surface tumor antigen expression. The data inducer of natural killer activity (10). in this paper demonstrating that both IFN-aD and -aJ can 4988

inhibit the growth of MCF-7 cells as well as reduce the antigen or low expressors of the antigen. The ability to alter the antigen enhancement induced by IFN-aA or -aB suggest that the -aD phenotype of these tumor cells with a pretreatment with recom and -aJ species can competitively interact with the interferon binant human interferon would result in an accurate discrimi receptor. The third possibility, which requires further experi nation between tumor and nontumorous tissue. Furthermore, mentation, is suggested by the observations that various IFN-a our data suggest that normal cells which are negative for a preparations differ in their ability to induce a number of inter- distinct MAb-defined tumor antigen (i.e., TAG-72) remain so feron-related properties, including antiviral, antiproliferative, even after treatment with high levels of recombinant interferon differentiation-modulating, and antigen-enhancing effects. Be (24, 26). These findings suggest that interferon treatment in cause of the differences in the structures of IFN-aD and -aJ vivo could enhance the signal (i.e., surface antigen expression) versus active IFN-a species (reviewed in Ref. 8), it is possible and thus be considered an adjuvant for the detection and that the primary amino acid sequence or the tertiary structures treatment of occult tumor cells by therapeutically conjugated of these molecules which allows for binding and/or activation MAbs. Testing these hypotheses is integral to the development of only a portion of the cell membrane IFN-a//3 receptor, i.e., of combination therapies which include a biological response the region which regulates antiviral (IFN-aD and -aJ), antipro modifier and monoclonal antibody. liferative (IFN-aD and -aJ), natural killer cell (IFN-aD), and HLA expression (IFN-aD and -aJ), but not the receptor domain ACKNOWLEDGMENTS required for modulating tumor antigen expression may be re sponsible for the observed varied effects. Different IFN-a spe We wish to thank Garland Davis and Robin Riley for their expert cies, as a result of structural dissimilarities, could then exhibit technical assistance and Bruce Kelder for the interferon assays. a high, intermediate, or low affinity for specific binding domains within the interferon receptor which are required to elicit the REFERENCES appropriate biological and biochemical responses after inter 1. Nagata, S.. Manu-i, N., and Weissmann, C. The structure of one of the eight feron binding. These observations provide additional support or more distinct chromosomal genes for human interferon-a. Nature (Lond.), for the hypothesis that the interferon receptor contains multiple 287:401-408, 1980. trigger sites that regulate specific subsets of biochemical path 2. Maeda, S., McCandliss, R., Gross, M., Sloma, A., Familletti, P. C., Tabor, J. M., Evinger, M., Levy, W. P., and Pestka, S. Construction and ick-ntifi ways distal to receptor occupancy (8). cation of bacterial plasmids containing nucleotide sequence for human leu The discovery that the multiple species of IFN-a have distinct kocyte interferon. Proc. Nati. Acad. Sci. USA, 77: 7010-7013, 1980. functional properties suggests that preferential expression of 3. Maeda, S., McCandliss, R., Gross, M., Sloma, A., Familletti, P. C., Tabor, the different IFN-a species might occur in response to different J. M., Evinger, M., Levy, W. P., and Pestka, S. Construction and identifi cation of bacterial plasmids containing nucleotide sequence for human leu environmental stimuli. Evidence has been presented indicating kocyte interferon. Proc. Nati. Acad. Sci. USA, 78:4648, 1981. that the complementary DNA clones of IFN-aA and -aD occur 4. Goeddel, D. V., Leung, D. W., Dull, T. J., Gross, M., Lawn, R. M., McCandliss, R., Seeburg, P. H., Ullrich, A., Yelverton, E., and Gray, P. W. more frequently in the library prepared from the Sendai-induced The structure of eight distinct cloned human leukocyte interferon cDNAs. m>clohlast oid cell line KG-1 (7). Furthermore, the major IFN- Nature (Lond.), 290: 20-26, 1981. 5. Allen, G., and Fantes, K. H. A family of structural genes for human lym a species present in the lymphoblastoid Namalwa cell line phoblastoid (leukocyte type) interferon. Nature (Lond.), 287:408-411,1980. closely resembles IFN-aF (42). An attempt to correlate any 6. Pestka, S. The human interferonsâ€”from protein purification and sequence functional differences reported for the various IFN-a species to cloning and expression in bacteria: before, between and beyond. Arch. Biochem. Biophys., 221: 1-37, 1983. with their known structural differences failed to reveal any 7. Goeddel, D. V., Yelverton, E., Ullrich, A., Heyneker, H. L., Miozzari, G., obvious relationships. A more detailed understanding of the Holmes, W., Seeburg, P. H., Dull, T., May, L., Stebbing, N., Crea, R., relationship between the structure and the function of specific Maeda, S., McCandliss, R., Sloma, A., Tabor, J. M., Gross, M., Familletti, P. C., and Pestka, S. Human leukocyte interferon produced by E. coti is IFN-a species may lead to the development of an IFN-a mole biologically active. Nature (Lond.), 287:411 -416, 1980. cule that exerts significant enhancement of tumor antigen 8. Langer, J., and Pestka, S. Structure of interferons. Pharmacol. 1her.. 27; 371-401, 1985. expression on a population of human tumor cells in vivo while 9. Ortaldo, J. R., Mason, A., Rehberg, E., Moschera, J., Kelder, B., Pestka, S., lacking some of the clinical side effects such as pyrogenicity. and Herberntan, R. B. Effect of recombinant and hybrid recombinant human Such a natural or designed species of IFN-a would be useful as leukocyte interferons on cytotoxic activity of natural killer cells. J. Biol. Chem., 258:15011-1 SOI5, 1983. an adjuvant for the detection and/or therapy of human carci 10. Ortaldo, J. R., Herberman, R. B., Harvey, C., Osheroff, P., Pan, Y. E., nomas by therapeutically labeled monoclonal antibodies. Kelder, B., and Pestka, S. A species of human a interferons that lacks the Some select MAbs, such as the anti-CEAs and the anti-TAG- ability to boost human natural killer activity. Proc. Nati. Acad. Sci. USA, 81:4926-4929, 1984. 72s, are currently being evaluated in different experimental 11. Rehberg, E., Kelder, B., Hoal, E. G., and Pestka, S. Specific molecular models as well as available clinical material to determine their activities of recombinant and hybrid leukocyte interferons. J. Biol. Chem., 257:11497-11502,1982. usefulness in the diagnosis and management of human carci 12. Kramer, M. J., Dennin, R., Kramer, C., Jones, G., Connell, E., Rolon, N., nomas. These MAbs have been tagged with different radio- Gruarin, A., Kale, R., and Trown, P. W. Cell and virus sensitivity studies nuclides and used to localize occult tumor lesions /// \itii (32, with recombinant alpha interferons. J. Interferon Res., 3:425-435, 1983. 13. Lee, S. H., Kelley, S., ChiÃ¹, H., and Stebbing, N. Stimulation of natural 33). They also have been used to detect antigen in serum (43) killer cell activity and inhibition of proliferation of various leukemic cells by as well as carcinoma cells in pleural effusions, ascites, and fine purified human leukocyte interferon subtypes. Cancer Res., 42: 1312-1316, needle aspirate biopsies (29-31, 37). In addition to the above 1982. 14. Harada, H., Shioiri-Nakano, K., Magumi, M., and Kawai, T. Distinction of immunodiagnostic uses, the potential of MAbs to deliver ther two subtypes of human leukocyte interferon (IFN-a) on B cell activation. J. apeutic effectors to the tumor site is being investigated. The use Immunol., 131: 238-243, 1983. 15. Evinger, M., Rubinstein, M., and Pestka, S. Antiproliferative and antiviral of recombinant interferon to augment tumor antigen expression activities of human leukocyte interferons. Arch. Biochem. Biophys., 210: in each of these situations merits consideration. For example, 319-329, 1981. a major question in the use of MAbs in the diagnosis of 16. Fish, E. N., Banerjee, K., and Stebbing, N. Human leukocyte interferon subtypes have different antiproliferative and antiviral activities on human exfoliated cells (i.e.. fine needle aspirates) is the extent of cells. Biochem. Biophys. Res. Commun., 112: 537-546, 1983. heterogeneity of tumor antigen expression. The heterogeneity 17. Fisher, P. B., and Grant, S. Effects of interferon on differentiation of normal and tumor cells. Pharmacol. Ther., 27:143-166, 1985. would allow human carcinoma cells to go undetected in im 18. Greiner, J. W., Schlom, J., Pestka, S., Giacomini, P., Kusama, M., Ferrane, munodiagnostic procedures if the cells analyzed were negative S., and Fisher, P. B. Modulation of tumor-associated antigen expression and 4989

shedding by recombinam human leukocyte and fibroblast interferons. Phar- 31. Johnston, W. W., Szpak, C. A., Lottich, S. C., Thor, A., and Schlom, J. Use macol. I hLT..in press, 1986. of a monoclonal antibody (B72.3) as a novel immunohistochemical adjuvant 19. Balkwill, F., Goldstein, I... and Stebbing, N. Differential action of six human for the diagnosis of carcinomas in tine needle aspiration biopsies. Human interferons against two human carcinomas growing in nude mice. Int. J. Path., 17:501-513,1986. Cancer, 35:613-617,1985. 32. Colcher, I)., Zalutsky, M., Kaplan, W., Kufe, D., Austin, F., and Schlom, J. 20. Basham, T. Y., Bourgeade, M. F., Creasey, A. A., and Merigan, T. C. Radiolocalization of human mammary tumors in athymic mice by a mono Interferon increases III A synthesis in melanoma cell: interferon-resistant clonal antibody. Cancer Res., 43: 736-742,1983. and sensitive cell lines. Proc. Nati. Vail. Sci. USA, 79:3265-3269, 1982. 33. Colcher, I).. Keenan, A. M., Larson, S. M., and Schlom, J. Prolonged binding 21. Heron, I., IInk laml. M . and Berg, K. Enhanced expression of /3j-microglob- of a radiolabeled monoclonal antibody (B72.3) used for the in situ radioim- ulin and HLA antigen on human lymphoid cells by interferon. Proc. Nati. munodetection of human colon carcinoma xenografts. Cancer Res., 44:5744- Acad. Sci. USA, 75:6215-6219,1978. 5751, 1985. 22. Imai, K., Ng, v K., Glassy, M. C, and Ferrane, S. Differential effect of 34. Staehelin, T., Hobbs, D. S., Rung, H.-F., Lai, C.-Y., and Pestka, S. Purifi interferon on the expression of tumor-associated antigens and histocompat- cation and characterization of recombinant human leukocyte interferon ibility antigens on human melanoma cells: relationship to susceptibility to (IFLrA) with monoclonal antibodies. J. Biol. (hem.. 256:9750-9754,1981. immune lysis mediated by monoclonal antibodies. J. Immunol., 127: 505- 35. Ã‘utÃ•,M.,Teramoto, Y. A., Mariani-Costantini, R., Horan Hand, P., Colcher, 508,1981. D., and Schlom, J. A monoclonal antibody (B72.3) defines patterns of 23. Fellous, M . Kamoun, M., Gresser, I., and Bono, R. Enhanced expression of distribution of a novel tumor-associated antigen in human mammary carci HLA antigens and fo-microglobulin in interferon-treated human lymphoid noma cell populations. Int. J. Cancer, 29:539-545, 1982. cells. EUT.J. Immunol., 9:446-449, 1979. 36. Friedman, E., Thor, A., Horan Hand, P., and Schlom, J. Surface expression of tumor-associated antigens in primary cultured human colonie epithelial 24. Greiner, J. W., Horan Hand, P., Noguchi, P., Fisher, P. H.. Pestka, S., and Schlom, J. Enhanced expression of surface tumor-associated antigens on cells from carcinomas, benign tumors and normal tissues. Cancer Res., 45: human breast and colon tumor cells after recombinant human leukocyte ,. - 5648-5655,1985. interferon treatment. Cancer Res., 44:3208-3214, 1984. 37. Barnstable, C. J., Bodmer, W. F., Brown, G., Galfre, G., Milstein, C., 25. <.ucomim. P., Aguzzi, A., Pestka, S., Fisher, P. B., and Ferrane, S. Modu Williams, A. F., and Ziegler, A. Production of monoclonal antibodies to group A erythrocytes, HLA and other human cell surface antigens: new tools lation by recombinam DNA leukocyte (a) and fibroblast (ÃŸ)interferons of for genetic analysis. Cell, 14:9-20, 1978. the expression and shedding of HLA and tumor associated antigens by human melanoma cells. J. Immunol., 133:1649-1655,1984. 38. Kufe, D. W., Nadler, L., Sargent, L., Shapiro, H., Hand, P., Austin, F., Colcher, D., and Schlom, J. Biological behavior of human breast carcinoma 26. Greiner, J. W., Horan Hand, P., Pestka, S., Noguchi, P., Fisher, P. B., associated antigens expressed during cellular proliferation. Cancer Res., 43: Colcher, I>. and Schlom, J. Detection and enhancement (by recombinant 851-857, 1983. interferon) of carcinoma cell surface antigens, using monoclonal antibodies. 39. Greiner, J. W., Horan Hand, P., Wunderlich, D., and Colcher, D. Radioim- IK Cancer Cells, I/The Transformed Phenotype, pp. 265-291. New York: munoassay for detection of changes in cell surface tumor antigen expression Cold Spring Harbor Press, 1984. induced by interferon. Methods Enzymol., 119:682-685, 1986. 27. Colcher, I>.. Horan Hand, P., Nuti, M., and Schlom, J. A spectrum of 40. Lowry, O. H., Rosebrough, N. J., Fair, A. L., and Randall, R. J. Protein monoclonal antibodies reactive with human mammary tumor cells. Proc. measurement with the Folin phenol reagent. J. Biol. (hem., 193: 265-275, Nati. Acad. Sci. USA, 78:3199-3203, 1981. 1951. 28. Colcher, I>..Horan Hand, P., Nuti, M., and Schlom, J. Differential binding 41. Greiner, J. W., Tobi, M., Fisher, P. B., Langer, J. A., and Pestka, S. to human mammary and nonmammary tumors of monoclonal antibodies Differential responsiveness of cloned mammary carcinoma cell populations reactive with carcinoembryonic antigen. Cancer Invest., /: 127-138, 1983. to the human recombinant leukocyte interferon enhancement of tumor anti 29. Johnston, W. W., Szpak, C. A., Lottich, C., Thor, A., and Schlom, J. Use of gen expression. Int. J. Cancer, 26: 159-166, 1985. a monoclonal antibody (B72.3) as an immunocytochemical adjunct to diag 42. Hiscott, J., Cantei!, K.. and Weissman, C. Differential expression of human nosis of adenocarcinoma in human effusions. Cancer Res., 45: 1894-1900, interferon genes. NucÃ.Acids Res., 12:3727-3746,1984. 1985. 43. Paterson, A. J., Schlom, J., Sears, H. F., Bennett, J., and Colcher, D. A 30. Nuti, M., Mottolese, M., Viora, M., Donnorso, R. P., Schlom, J., and Natali, radioimmunoassay for the detection of a human tumor-associated glycopro- P. G. Use of monoclonal antibodies to human breast tumor associated antigen tein (TAG-72) using monoclonal antibody B72.3. Int. J. Cancer, 37: 659- in Tine needle aspirate cytology. Int. J. Cancer, 37:493-498, 1986. 666, 1986.

4990

Downloaded from cancerres.aacrjournals.org on September 23, 2021. © 1986 American Association for Cancer Research. Differential Effects of Recombinant Human Leukocyte Interferons on Cell Surface Antigen Expression

John W. Greiner, Paul B. Fisher, Sidney Pestka, et al.

Cancer Res 1986;46:4984-4990.

Updated version Access the most recent version of this article at: http://cancerres.aacrjournals.org/content/46/10/4984

E-mail alerts Sign up to receive free email-alerts related to this article or journal.

Reprints and To order reprints of this article or to subscribe to the journal, contact the AACR Publications Subscriptions Department at [email protected].

Permissions To request permission to re-use all or part of this article, use this link http://cancerres.aacrjournals.org/content/46/10/4984. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site.