Modulation of the tumorigenicity of human adenovirus-12-transformed cells by interferon

Cell, Vol. 43, 263-267,

November

1985, Copyright

0092-8674/85/l

0 1985 by MIT

10263-05

$02,00/O

Modulation of the Tumorigenicity of Human Adenovirus-12-Transformed Cells by Interferon Hiroaki Hayashi,’ Kenichi Tanaka,’ Francis George Khoury; and Gilbert Jay’ Laboratory of Molecular Virology National Cancer Institute Bethesda, Maryland 20205 t Department of Medical Microbiology University of Manitoba Winnipeg, Manitoba R3E OW3, Canada

Jay,t

l

Summary Human adenovirus-124ransformed cells express greatly reduced levels of the major histocompatibility complex class I antigens and an? highly tumorigenic in syngeneic hosts. The finding that expression of a transfected class I gene is sufficient to abrogate their tumorigenicity underscores the importance of defining the conditions that will lead to derepression of endogenous class I genes in these cells. Brief treatment of AdlP-transformed cells with interferon results in the rapid but transient expression of class I antigens, and these interferon-treated cells have significantly reduced tumorigenicity in immunocompetent hosts. We have further demonstrated that subcutaneous administration of interferon, subsequent to the introduction of a tumorigenic dose of AdlS-transformed cells, results in complete protection against this tumor. The ability of interferon to’induce” class I gene expression may be an important modality in the treatment of a variety of spontaneous tumors that exhibit greatly reduced levels of class I antigens on their cell surface. Introduction The major histocompatibility complex (MHC) class I antigens are expressed in virtually all cell types and play a role in the immune presentation of cells bearing “foreign” antigens (Zinkernagel and Doherty, 1979). The absence of class I molecules from the surface of tumor cells would likely afford protection against immune recognition, allowing the cells to escape destruction. Indeed, many natural tumors of epithelial derivation, including embryonal carcinomas (Jacob, 1977) choriocarcinomas (Jones and Bodmer, 1980; Tanaka and Chang, 1982) cervical carcinomas (Sanderson and Beverley, 1983), mammary carcinomas (Fleming et al., 1981; Natali et al., 1983) eccrine porocarcinemas of the skin (Holden et al., 1983, 1984) small-cell carcinomas of the lung (Doyle et al., 1985), neuroblastomas (Trowsdale et al., 1980; Lampson et al., 1983) colorectal carcinomas (Pavers et al., 1982; Umpleby et al., 1985) and melanomas (Nanni et al., 1983) either do not express ordisplay greatly reduced levels of class I antigens on their cell surface. Tumors of lymphoid origin have also been found in certain cases to have a similar phenotype (Seigler et al., 1971; Elkins et al., 1984). Such an alter-

ation could represent an essential requirement of malignant cells. Evidence that suppression of class I antigens in tumor cells is a critical determinant for their malignant state of growth comes from recent experiments which show that expression of a transfected class I gene can abrogate the tumorigenicity of different malignant cell lines (Hui et al., 1984; Tanaka et al., 1985; Wallich et al., 1985). These studies demonstrate that cell transformation alone is often insufficient to induce malignancy and suggest that immunoselection of those transformed cells which have lost the expression of their class I antigens is responsible for their escape from host surveillance. We have attempted to evaluate the efficacy of interferon to induce in vitro the expression of endogenous class I genes and to subvert in vivo the tumorigenicity of cells transformed by human adenovirus-12. Results The tumorigenicity of Adl2-transformed mouse cells is due in part to their not expressing significant levels of class I antigens (Tanaka et al., 1985). Southern blot analysis has revealed that for the two distinct cell lines studied, the class I genes are neither deleted nor rearranged (data not shown). In an attempt to ‘derepress” the endogenous class I genes in these Ad12 tumors, we have made use of the murine a- and /3-interferons (IFN-a and IFN-fi). C3ATl is an Adl2-transformed line derived from C3H cells of embryonic origin (Maeta and Hamada, 1979). It expresses class I transcripts at a level less than 5% that observed in nontransformed cells and is highly tumorigenic in. both syngeneic and allogeneic hosts (Tanaka et al., 1985). Using a rat monoclonal antibody that detects all murine class I antigens nondiscriminately, we studied the effect of varying doses of murine IFN-a and -fi on these C3ATl cells (Figure 1A). Incubation with 400 IUlml (curve c) caused a significant increase over control cells (curve b) in total surface class I antigens; incubation with 1000 IUlml (curve d) resulted in maximal stimulation. Such doses of IFN were neither cytotoxic nor cytopathic, as determined by cell counts and by propidium iodide staining (Herzenberg and Herzenberg, 1980). The induction process was gradual and required about 8 hr to reach a plateau (Figure lB, curve a). However, this effect was transient. Upon removal of interferon, the amount of surface class I antigen rapidly decreased with a half-life of about 20 hr (Figure lB, curve b). Similar results were observed when murine y-interferon (IFN-r) was used instead of IFN-a and IFN-8 (data not shown). Since the stimulatory effects of IFN-a and IFN-fl, and of IFN-y, were neither additive nor synergistic, only IFN-a and IFN-/I were used in subsequent functional studies. The stimulation of class I antigen expression on the cell surface was reflected at the RNA level, as determined by

Cdl 294

B

‘0

0 FLUORESCENCE

20 HOURS

Northern blot analysis (Figure 2). Poly(A)+ RNA from control C3ATi cells (lanes 1) and from CBATI cells treated with either 400 IU/ml (lanes 2) or 2000 IU/ml (lanes 3) of IFN-a and -/3 were compared in parallel with RNA from L cells (lanes 4). Hybridization with both the Ad12 ElA DNA probes (A) and the Ad12 ElB DNA probes(B) revealed that interferon did not alter the expression of either of the two genes, which together were responsible for the transformation of the C3ATl cells. Hybridization with either of the two synthetic oligonucleotide probes (Kress et al., 1983) one specific for the H-2K transcript (C) and the other for the H-2L transcript (D), showed that the expression of both class I genes was greatly stimulated by treatment with interferon. However, while the H-2K transcript was elevated more than 20-fold and reached a level higher than that observed for L cells (C), the H-2L transcript was much less elevated and did not approach the level seen in L cells (D; exposure was 4 times as long as in C). In the interferontreated cells, the accumulation of the H-2L transcript was estimated to be less than one-fifth that of the H-2K transcript. Together, these data suggest that IFN-(r and -b stimulate class I gene expression in C3ATl cells without affecting the expression of the transforming functions. Although the induction of class I antigens by interferon was transient, it was of interest to determine whether this expression has any biological effect. C3ATl cells were treated for 24 hr with IFN-a and IFN-/3, and were injected subcutaneously into C3H/HeJ mice. Propidium iodide staining confirmed that the treatment had no detectable cytostatic effect. At a tumor-cell dose of 5 x lo5 per animal, the average tumor size was found to differ significantly between animals receiving untreated C3ATl cells and those receiving interferon-treated C3ATl cells (Figure 38). The marked reduction in the average tumor size observed for the treated cells was the result of only four animals responding to the tumor cell challenge as compared to eight for the control group, and also of an overall

I 40

Figure 1. Effect face Expression

of Interferon on the Cell Surof Class I Antigens

Single C3ATl cell suspensions derived from different monolayer cultures by treatment with trypsin and EDNA were incubated with the rat monoclonal antibody (43-3-9-8) specific for the mouse class I antigens, stained with fluoresceinconjugated rabbit anti-rat IgG, and analyzed by flow microfluorometry in the Becton Dickinson FACS Analyzer. A: (a) untreated cells incubated with a nonspecific rat monoclonal antibody, (b) untreated cells incubated with the 453-9-9 rat monoclonal antibody, (c)cells treated with 400 IUlml interferon and incubated with the 43-3-9-9 antibody, (d) cells treated with 1000 IlJ/ml interferon and incubated with the 43-3-9-9 antibody. B: (a) mean fluorescence of the cell population using the 43-3-9-9 antibody after different lengths of treatment with interferon, (b) mean fluorescence of cells treated with interferon for 24 hr and chased for different times in the absence of interferon. The mean fluorescence represents the value above that observed for the untreated cells.

delay in the appearance of tumors when compared to the control animals. The effect of in vitro treatment with interferon is further demonstrated by an analysis of the survival curves derived from the same experiment (Figure 3D). Whereas 50% of the untreated group of animals survived to about 45 days, for 50% of the interferon-treated group survival was prolonged to 90 days. When the tumor cell dose was increased to 5 x 106, treatment with interferon had almost no effect. There was a slight delay in the appearance of tumors in animals challenged with the treated cells (Figure 3A), but no detectable difference in overall survival between the interferontreated group and the control group (Figure 3C). This observation confirms the suspicion that when the dose of tumor cells is sufficiently high, the immune system is incapable of combating the challenging cells despite their expression of high levels of class I antigens. However, when the tumor-cell dose was sufficiently low, the effect of interferon was significant. The less than complete loss of tumorigenicity was expected because of the transient nature of the interferon effect on class I gene expression. The positive results observed with the in vitro studies prompted us to use the Ad12 tumor model for evaluating the efficacy of interferon to subvert tumorigenesis in vivo. Recognizing that C3ATl cells at a dose of 1.5 x lo6 required from 10 to 15 days to induce palpable tumors in syngeneic C3H/HeJ mice, we attempted to give single daily injections of IFN-(r and IFN-fi for a total of 10 days, beginning one day after inoculation with tumor cells. Two doses of interferon were used in these studies. Although a dose of 2000 IU per animal of about 30 gm was chosen because 1000 IUlml was sufficient to induce maximal expression of class I antigens in tissue culture, a higher dose of 8000 IU per animal was also selected in accordance with the recommended clinical dose presently being used on cancer patients. Neither dose was lethal under the protocol used in our studies.

Interferon

Effect

on Tumorigenicity

Figure 2. Effect of Interferon State Levels of Class I mRNA

on the

Steady

Total poly(A)’ RNA obtained from untreated C3ATl cells (lanes l), C3ATI cells treated with IFN at 400 IlJlml (lanes 2) or 2000 IUlml (lanes 3) and untreated L cells (lanes 4) were fractionated by electrophoresis in a 1.0% agarose gel and transferred to a nitrocellulose membrane. The same RNA blot was hybridized successively, after stripping with boiling water, to one of several 52P-labeled DNA probes: nicktranslation probes derived from a subclone of either the Ad12 EIA gene (A) or the Ad12 ElB gene (B), and end-labeled oligonucleotide probes specific for either the H-2K transcript (C) or the H-2L transcript (D). The top arrow indicates the position of the 28s ribosomal RNA and the bottom arrow that of 18s ribosomal RNA.

Figure 3. Average Tumor Size and Survival Rates of Mice Injected with C3ATl Cells Treated with Interferon C3HIHeJ mice in groups of ten, were each given a subcutaneous dose of 5 x lo6 cells (A and C) or 5 x lo5 cells (B and D) in the thigh. The cells were untreated (curves a) or were treated with 1000 IUlml of interferon for 24 hr before injection (curves b). At regular intervals thereafter, the size of the tumor mass in each animal was measured (A and B) and individual deaths resulting from tumor load were recorded (C and D).

DAYS AFTER INJECTION

DAYS AFTER INJECTION

When compared to control animals injected with phosphate-buffered saline (Figure 4A, curve a), mice injected with 2000 IU IFN-(r and IFN-8 showed significantly reduced average tumor size (Figure 4A, curve b). While 9 of 10 control mice developed tumors (Figure 48, curve a), only 4 of 10 mice that were given 10 injections of interferon at 2000 IU each responded with progressive tumor growth (Figure 48, curve b). More dramatically, mice that were given 10 injections of 8000 IU each were fully protected against growth of the C3ATl tumor. Tumor growth progressed slowly at first, and then regressed totally (Figure 4A, curve c). Since the rejection occurred well after the cessation of interferon treatment, it was concluded that an immune response had been induced in these mice, which appeared cured (Figure 48, curve c). Discussion It has long been suspected that the class I antigens play a critical role in the immune presentation and recognition

of tumor cells (Zinkernagel and Doherty, 1979). Direct evidence in support of this assumption came from recent studies showing that the expression of transfected class I genes can reverse the tumorigenicity of different transformed cells which do not display surface class I antigen (Hui et al., 1984; Tanaka et al., 1985; Wallich et al., 1985). Given that a large variety of malignant tumors of different tissue derivations express greatly reduced levels of class I antigens, the identification and characterization of “natural” immunomodulators that can derepress class I gene expression may have significant therapeutic value. Although there is great interest in interferon as an antitumor agent, the results so far obtained have been ambiguous. The fact that interferon can induce class I gene expression in different tissue culture cells (Vignaux and Gresser, 1977; Fellous et al., 1979; Wivel and Pitha, 1982) suggests that in tumor systems in which there is overall suppression of surface class I antigens, treatment with interferon may have a prophylactic effect. The Ad12 tumor system was selected for our studies be-

Cl?ll 266

Figure 4. Effect of Interferon in Animals Receiving a Tumorigenic Dose of CBATI Cells

DAYS AFTER INJECTION

DAYS AFTER INJECTION

cause cells transformed by this virus not only have reduced levels of surface class I antigens (Schrier et al., 1983; Bernards et al., 1983) but also are highly tumorigenie in syngeneic animals (Tanaka et al., 1985). Because these two properties of Ad12 transformation are highly reproducible and are unique among tissue culture systems, we believe that results derived from this experimental model would prove more meaningful than results obtained using cell lines whose derivation and representativeness are unknown. We have demonstrated that interferon can stimulate class I gene expression in an Ad12 tumor which shows overall suppression of all surface class I antigens, and that mice treated with interferon can effectively reject a tumorigenic dose of this same tumor. We cannot yet distinguish whether this antitumor activity of interferon is the result of more effective presentation of tumor cells to cytotoxic T cells, the stimulation of natural killer cells, the activation of macrophages, the acquisition of an antiproliferative effect, or a combination of the above (Stewart, 1980). Because our previous studies showed that expression of a transfected class I gene led to tumor rejection (Tanaka et al., 1985) we suspect that interferon induction of the MHC locus is central to its antitumor effect. Our results suggest that this tumor model may be useful for optimizing the conditions for the clinical use of interferon as an antitumor agent. For example, it should be possible to determine an effective interferon dose by monitoring the in vivo induction of class I antigens in the target tumor cells and the subsequent infiltration of immune effector cells. We have already observed that interferon treatment has no effect at high tumor loads. This finding is consistent with the speculation that successful use of interferon in cancer treatment requires the prior removal of the bulk of the locally growing tumor mass. Accordingly, these results suggest that interferon may prove to be an important modality in the treatment of a variety of spontaneous tumors that exhibit greatly reduced levels of class I antigens on their cell surface. Experimental

Groups of ten C3HIHeJ mice were injected with a fixed daily dose of interferon for a total of ten doses beginning one day after the inoculation with tumor cells. The various groups were given phosphate-buffered saline (curves a), or 2000 IU interferon (curves b), or 8000 IU interferon (curves c) per day. The size of the tumor mass in each animal was measured at regular intervals (A) and individual deaths resulting from tumor load were recorded (6).

Procedures

Cell Cultures All cells were maintained at 37°C in Dulbecco’s modified Eagles’ medium containing 10% fetal calf serum. The interferon preparations (IFN-a and IFN-p, 1.3 x 10’ IRUlmg) were obtained from Lee BioMolecular Research Laboratories, Inc. (San Diego, California). The

C3ATl cell line was Hamada, 1979).

Detection

kindly-

of Cell Surface

provided

by C. Hamada

(Maeta

and

Antigens

For the flow microfluorometry assay, single cell suspensions obtained from monolayer cultures by treatment with trypsin and EDTA were incubated with the rat monoclonal antibody (43-3-9-8) specific to mouse class I antigens and stained with fluorescein-conjugated rabbit anti-rat IgG. The stained cells were analyzed in the Becton Dickinson FACS Analyzer. The monoclonal antibody was kindly provided by Dr. Keiko Ozato (National Institutes of Health).

RNA Blot Analysis Poly(A)’ RNA was isolated as previously described (Tanaka et al., 1983). Aliquots of RNA were adjusted to 50% formamide, 20 mM morpholinepropane-sulfonic acid (MOPS, pH 7.0), 5 mM sodium acetate, 1 mM EDTA, 2.2 M formaldehyde, heated at 60% for 10 min, and subjected to electrophoresis on a 1.0% agarose gel containing 2.2 M formaldehyde. The running buffer was 20 mM MOPS (pH 7.0) 5 mM sodium acetate, 1 mM EDNA; and electrophoresis was performed at 35-40 mA for 4 hr at 4%. The RNA was transferred to a nitrocellulose filter as described (Thomas, 1980). The conditions for hybridization have been described (Tanaka et al., 1983).

Tumor

Induction

C3HIHeJ (H-2k haplotype) mice, in groups of ten, were each given a single subcutaneous injection on the right thigh of a fixed dose of the control or of the appropriately treated C3ATl cells. The progressive increase in the size of the tumor mass, assessed by measuring the diameter in millimeters, was determined at regular intervals for each animal. Deaths of individual mice resulting from tumor growths were recorded.

We thank C. Hamada for kindly providing the AdlP-transformed cell lines, K. Ozato for the rat monoclonal antibody against murine class I antigens, and A. J. van der Eb for the Ad12 genomic clones. 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. Received

July 18, 1985

References Bernards, R., Schrier, P I., Houweling, A., Bos, J. L., Van der Eb, A. J., Zijlstras, M., and Melief, C. J. M. (1983). Tumorigenicity of cells transformed by adenovirus type 12 by evasion of T-cell immunity. Nature 305, 776-779. Doyle, A., Martin, W. J., Funa, K., Gazdar, A., Carney, D., Martin, S. E., Linnoila, I., Cuttitta, F., Mulshine, J., Bunt-r, P., and Minna, J. (1985). Markedly decreased expression of class I histocompatibility antigens, protein, and mRNA in human small-cell lung carcinoma. J. Exp. Med. 161, 1135-1151.

Interferon 267

Effect

on Tumorigenicity

Elkins, W. L., Pickard, A., and Pierson, G. R. (1984). Deficient expression of class I HLA in some cases of acute leukemia. Cancer Immunol. Immunother. 78, 91-100.

W. F. (1982). Lack of expression of HLA-ABC antigens in choriocarcinoma and other human tumor cell lines. Natl. Cancer Inst. Monogr. 60, 175-180.

Fellous, M., Kamoun, M., Gresser, I., and Bono, R. (1979). Enhanced expression of HLA antigens and /I,-microglobulin on interferon-treated human lymphoid cells. Eur. J. Immunol. 9, 446-449.

Trowsdale, J., Travers, P., Bodmer, W. F., and Patillo, R. A. (1980). Expression of HLA-A, B, and -C and 6,-microglobulin on human neuroblastoma cell lines. J. Immunol. 730, 2471-2478.

Fleming, K. A., f&Michael, A., Morton, J. A., Woods, J., and McGee, J. 0. D. (1981). Distribution of HLA class I antigens in normal human tissue and in mammary cancer. J. Clin. Pathol. 34, 779-784.

Umpleby, H. C., Heinemann, D., Symes, M. O., and Wiliamson, R. C. N. (1985). Expression of histocompatibility antigens and characterization of mononuclear cell infiltrate in normal and neoplastic colorectal tissues of humans. J. Natl. Cancer Inst. 74, 1161-1168.

Herzenberg, L. A., and Herzenberg, L. A. (1980). In Handbook of Experimental Immunology, 2, D. M. Weir, ed. (Oxford: Blackwell Scientific Publications), pp. 22.1-22.21. Holden, C. A., Sanderson, A. R., and MacDonald, D. M. (1983). Absence of human leukocyte antigen molecules in skin tumor and some cutaneous appendages: evidence using monoclonal antibodies. J. Am. Acad. Dermatol. 9, 867-871. Holden, C. A., Shaw, M., McKee, P H., Sanderson, A. R., and MacDonald, D. M. (1984). Loss of membrane /I,-microglobulin in eccrine porocarcinoma: its association with the histopathologic and clinical criteria of malignancy. Arch. Dermatol. 720, 732-735. Hui, K., Grosveld, F, and Festenstein, H. (1984). Rejection of transplantable AKR leukaemia cells following MHC DNA-mediated cell transformation. Nature 377, 750-752. Jacob, F. (1977). Mouse munol. Rev. 33, 3-32.

teratocarcinoma

Jones, E. A., and Bodmer, gens on choriocarcinoma

and embryonic

antigens.

Im-

W. F. (1980). Lack of expression of HLA anticell lines, Tissue Antigens 76, 195-202.

Kress, M., Liu. W.-Y., Jay, E., Khoury, G., and Jay, G. (1983). Comparison of class I (H-2) gene sequences: derivation of unique probes for members of this multigenic family. J. Biol. Chem. 258, 13929-13936. Lampson, L. A., Fisher, C. A., and Whelan, J. l? (1983). Striking paucity of HLA-A, 8, C and /3,-microglobulin on human neuroblastoma cell lines. J. Immunol. 730, 2471-2478. Maeta, Y., and Hamada, C. (1979). Susceptibility S(+) and S(-) mouse cells to cell-mediated Microbial. Immunol. 23, 1085-1095.

of Adl2-transformed immunity in vitro.

Nanni, l?, Colombo, M., DeGiovanni, C., Lollini, P., Nicoletti, G., Parmiani, G., and Prodi, G. (1983). Impaired H-2 expression in BL6 melanoma variants. J. Immunogenet. 70, 361-370. Natali, f? G., Giacomini, l?, Bigotti, A., Lami, K., Nicotra, M. R., Ng, A. K., and Ferrone, S. (1983). Heterogeneity in the expression of HLA and tumor-associated antigens by surgically removed and cultured breast carcinoma cells. Cancer Res. 43, 660-668. Sanderson, A. R., and Beverley, microglobulin and immunoselection munol. Today 4, 211-213.

l? C. L. (1983). Interferon, in the pathway to malignancy.

&Im-

Schrier, l? I., Bernards, R., Vaessen, R. T M. J., Houweling, A., and Van der Eb, A. J. (1983). Expression of class I major histocompatibility antigens switched off by highly oncogenic adenovirus 12 in transformed rat cells. Nature 305, 771-775. Seigler, H. D., Kremer, W. B., Metzgar, R. S., Ward, F. E., Hwang, A. T., and Amos, D. 8. (197l). HLA antigenic loss in malignant transformation, .i J. Natl. Cancer Inst. 46, 577-584. Stewart, Verlag).

W. E., II. (1980). The Interferon

System.

(New York: Springer-

Tanaka, K., and Chang, K. S. S. (1982). Modulation of natural killer sensitivity of murine trophoblast cells by tumor promoter and interferon. Int. J. Cancer 29, 315-321. Tanaka, K., Appella, E., and Jay, G. (1963). Developmental activation of the H-2K gene is correlated with an increase in DNA methylation. Cell 35, 457-465. Tanaka, K., Isselbacher, K. J., Khoury, G., and Jay, G. (1985). Reversal of oncogenesis by the expression of a major histocompatibility complex class I gene. Science 228, 26-30. Thomas, P S. (1980). fragments transferred 5201-5205.

Hybridization of denatured RNA and small DNA to nitrocellulose. Proc. Natl. Acad. Sci. USA 77,

Travers,

J. L., Trowsdale,

P. J., Arklie,

J., Patillo, Ft. A., and Bodmer,

Vignaux, F., and Gresser, I. (1977). Differential the expression of H-2K, H-2D, and la antigens J. Immunol. 778, 721-723.

effects of interferon on on mouse lymphocytes.

Wallich, R., Bulbuc, N., Hammerling, G. J., Katzav, S., Segal, S., and Feldman, M. (1985). Abrogation of metastatic properties of tumor cells by de novo expression of H-2K antigens following H-2 gene transfection. Nature 375, 301-305. Wivel, N. A., and Pitha, P M. (1982). Interferon treatment of murine Meth-A sarcoma cells: effects on the malignant phenotype and expression of tumor-specific and H-2 antigens. Int. J. Cancer 30, 649-654. Zinkernagel, R. M., and Doherty, P. C. (1979). MHC-restricted cytotoxic T cells: studies on the biological role of polymorphic major transplantation antigens determining T-cell restriction-specificity, function, and responsiveness. Adv. Immunol. 27, 51-77.