Role of Interferon and 2’,5’-Oligoadenylate Synthetase in Erythroid Differentiation of Friend Leukemia Cells STUDIES WITH INTERFERON-SENSITIVE AND -RESISTANT VARIANTS*

It has been suggested that the interferon (IFN)-induced 2',5'-oligoadenylate (2-5A) synthetase, which polymerizes ATP into a series of 2',5'-linked oligomers with the general formula pppA(2'p5'A)n, plays a general role in cell growth and terminal differentiation. For instance, an increase in 2-5A synthetase activity has been described during dimethyl sulfoxide (Me2SO)-induced erythroid differentiation of Friend leukemia cells (FLC). 2-5A synthetase has been measured in two Friend leukemia cell sublines by various techniques including a radioimmunoassay of its products which would detect 10(-16) mol of 2-5A cores. Although cells of both sublines fully differentiate (as measured by benzidine staining), only in one subline was there an increase in 2-5A synthetase activity upon treatment with Me2SO. Hexamethylenebisacetamide, another potent agent of differentiation in this system, did not increase 2-5A synthetase activity in either of these two sublines. An IFN-resistant FLC variant differentiated normally upon treatment with Me2SO or hexamethylenebisacetamide while it was noninducible for 2-5A synthetase activity by exogenous IFN or by the inducers themselves. A similar situation has been observed with regard to the level of phosphorylation of the IFN-induced Mr = 67,000 protein band. In addition, treatment of IFN-sensitive and resistant FLC sublines with mouse alpha beta IFN antiserum did not affect differentiation. Even though we have duplicated previous findings on the increase of 2-5A synthetase activity in Me2SO-induced FLC, the lack of any such increase with other inducers or other sublines indicates that there is no causal relationship between the enzyme activation and FLC differentiation.

2-5A synthetase has been measured in two Friend leukemia cell sublines by various techniques including a radioimmunoassay of its products which would detect 10"' mol of 2-5A cores. Although cells of both sublines fully differentiate (as measured by benzidine staining), only in one subline was there an increase in 2-5A synthetase activity upon treatment with Me2S0. Hexamethylenebisacetamide, another potent agent of differentiation in this system, did not increase 2-5A synthetase activity in either of these two sublines. An IFNresistant FLC variant differentiated normally upon treatment with M e 8 0 or hexamethylenebisacetamide while it was noninducible for 2-5A synthetase activity by exogenous IFN or by the inducers themselves. A similar situation has been observed with regard to the level of phosphorylation of the IFN-induced M, = 67,000 protein band. In addition, treatment of IFNsensitive and resistant FLC sublines with mouse &IFN antiserum did not affect differentiation. Even though we have duplicated previous findings on the increase of 2-5A synthetase activity in MeaSO-induced FLC, the lack of any such increase with other inducers or other sublines indicates that there is no causal relationship between the enzyme activation and FLC differentiation.
Besides their antiviral activity, interferon($ can serve as pleiotropic effector molecules in a variety of systems, including cells differentiating in vitro . Extensive studies have revealed the major role played by two * Collaborative work has been aided by the Centre National de la Recherche Scientifique. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "aduertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. independent 1FN'-induced enzymatic systems, namely the protein kinase and 2-5A synthetase pathways, both activated by double-stranded RNAs (for a review, see Lebleu and Content, 1982). The protein kinase phosphorylates the cy subunit of initiation factor eIF-2 and a M , = 67,000 protein which may represent the kinase itself (for a review, see Sen, 1981). 2-5A synthetase polymerizes ATP into a series of 2',5'-linked oligomers with the general formula pppA(2'p5'A),. These in turn bind specifically and activate an endoribonuclease (RNase L) which degrades RNAs (for a review, see Lengyel, 1981).
Since 2-5.4 synthetase activity varies extensively in different cell lines and within a given cell line according to growth conditions, a wider role for this pathway in the regulation of cell growth and differentiation has been proposed (Stark et al., 1979;Krishnan and Baglioni, 1980;Creasey et al., 1980; and others). In particular, increased levels of 2-5A synthetase activity (Kimchi, 1981) and the production of P-type IFN (Friedman-Einat et al., 1982) have been reported in growing and differentiating FLC exposed to Me2S0. It has also been reported that rabbit reticulocytes contain high levels of protein kinase and 2-5A synthetase (Farrell et al., 1977;Hovanessian and Kerr, 1978). Similarly, low doses of IFN enhance the expression of differentiation markers upon induction of FLC with Me2S0 (Lieberman et al., 1975;Luftig et al., 1977;Dolei et al., 1980;. In this report, we describe experiments carried out in this context on two FLC sublines and on an IFN-resistant variant isolated from one of them . This variant does not exhibit any 2-5A synthetase activity even when exposed to 10,000 units/ml of aPIFN (Affabris et aL., 1983).
Poly(r1) .poly(rC) was from P-L Biochemicals. Benzidine dihydrochloride was from Ciba. Me2S0 and polyethylene glycol 6000 were from Merck. HMBA was a generous gift of C. Delfini, Istituto Superiore di Sanita, Rome, Italy. Bacterial alkaline phosphatase (EC 3.1.3.1) type IIIR was purchased from Sigma. All other chemicals and solvents were of reagent grade.
Cells-Two sublines, both derived from the 745A clone and ob-3262 2-5A Synthetase and Differentiation in Friend Cells tained from C. Friend, Mount Sinai Medical Center, New Yo&, were independently passaged for several years in the laboratories of L.
Viruses-Vesicular stomatitis virus (Indiana strain) stocks, obtained by infecting L929 cell monolayers with low multiplicity of infection (0.01-0.1 plaque-forming unit/cell) were titrated by plaque assay on the same cells. Titers ranged between and lo9 plaqueforming units/ml. Newcastle Disease virus, strain F, stocks were obtained by infecting 10-11-day-old embryonated eggs in the allantoic cavity. The virus was harvested after 3 days of incubation at 35 "C and titrated by hemagglutination assay. The titers ranged between 300 and 600 hemagglutinating units/ml.
Interferon-Mouse aBIFN was prepared on C243-3 cells according to Cachard and De Maeyer-Guignard (1981). Briefly, confluent monolayers were treated for 24 h with 1 mM Na butyrate in MEM + 2% NBS (v/v) in the presence of 150 units/ml aPIFN ("priming"). Cells were then infected with Newcastle disease virus (10 hemagglutinating units/106 cells). The unattached virus was removed after 1 h of adsorption. The monolayers were washed with isotonic buffered saline and reincubated with MEM containing 0.5% (v/v) NBS, and 5 mM theophylline for 16-18 h. Clarified supernatant fluids were brought to pH 2.0 with 6 N HCl, kept at 4 "C for 5 days, neutralizing to pH 7.0, and stored a t -80 "C. Partial purification by (NH4)zS04 precipitation between 30 and 80% saturation was followed by extensive dialysis against isotonic buffered saline. Specific activities of IFN preparations averaged about lo6 units/mg of protein. Amounts of IFN are given throughout this paper in laboratory units, i.e. the amount of IFN reducing by 50% plaque production by vesicular stomatitis virus. This unit equals 4 research reference units of the National Institutes of Health Standard Preparation, code G-002-904-511, whose titer was 12,000 IU/ml when reconstituted. Details about this standard preparation are reported in Research Reagents Note No. 15 (World Health Organization Standard, 1979).
Induction of FLC Differentiation-Friend cells were seeded a t lo5/ ml in the presence of 1.5% (v/v) MezSO or 5 mM HMBA. Cells were counted daily by using a Coulter Counter or a hemocytometer. Cell mortality, evaluated by the trypan blue dye exclusion method, never exceeded 2%. The degree of differentiation was determined by spectrophotometric evaluation, a t 414 nm, of hemoglobin content, according to Kabat et al. (1975), or by the percentage of benzidine-positive cells (B+ cells) according to Orkin et al. (1975). Cell extracts were prepared from the same cell cultures and kept frozen at -70 "C.
Radioimmunological Assay of 2-5A Actiuity-Cell extracts and incubations were performed as described above, except for the absence of I3H)ATP. Radioimmunoassay of 2-5A products was performed as described by Cailla et al. (1982a). Briefly, dilutions of the samples containing 2-5A oligomers were incubated with alkaline phosphatase for 30 min at 37 "C in 100 mM Tris-HC1, pH 9.3,lO mM MgClz buffer. Samples were boiled 3 min at 100 "C to inactivate the phosphatase and were diluted 3-fold with a citrate buffer (100 mM, pH 6.3). 50 pl of samples were mixed with 50 g1 of the Iz5I-labeled iododisuccinyl-AZ'p5'A tyrosine methyl ester probe, and 100 p1 of monoclonal antibody, both prepared according to Cailla et al. (1982b). After 18 h of incubation at 4 "C, the antigen-antibody complexes were precipi-tated with 300 gl of polyethylene glycol 6000 in the presence of human plasma as carrier. Samples were centrifuged 10 min at 3000 X g and the radioactivity of the precipitate was measured in a y-counter. The 2-5A concentration of each sample was obtained by comparison with a reference curve. The results were expressed as nanomoles of 2-5A/ lo6 cells for the indicated period of time. None of the samples that were not treated with phosphatase displaced the radioactive probe, indicating that only phosphorylated 2-5A was synthesized in the cellfree extracts. Likewise, phosphatase-treated samples did not displace the radioactive probe if they were digested with snake venom phosphodiesterase.

2-5A Synthetase Activity in Several Growing and Differentiating FLC Sublines Treated with
Me2SO-Cells of two FLC sublines ( 7 4 5 A~ and 745AH) and of one IFN-resistant variant (745AR 1Cll) were seeded at lo5 cells/ml in medium supplemented with Me2S0 and were grown for 6 days under stationary conditions. As shown in Fig. lA, the three growth curves were similar. Under these conditions, erythroid differentiation, as measured by hemoglobin-containing (benzidine-positive) cells, attained nearly 100% in all three lines (Fig. 1B). In accordance with published data (Kimchi,.l981), 2-5A synthetase activity increased in parallel with the accumulation of benzidine-positive cells in the 745AH subline. On the contrary, neither the wild type 745.4~ subline nor the IFNresistant variant derived therefrom, exhibited any detectable enzyme activity (Fig. IC). It was verified that 7 4 5 A~ cells responded to exogenous mouse a@IFN with the expected elevation in 2-5A synthetase activity (as shown by Affabris et al., 1983) while 7 4 5 A~ l c l l cells did not respond to such treatment (data not shown), as previously described (Affabris et al., 1983). As 2-5A synthetase levels have been reported to be related to growth conditions (Krishnan and Baglioni, 1980;Creasey et al., 1980), two additional protocols were also followed. 745AR cells seeded a t 5 X 105/ml in Me2SO-rich medium were counted daily and diluted in Me2SO-containing medium to the initial concentration for 8 days. Alternatively, cells of the same subline were seeded at 105/ml in the same medium, counted 3 days later (when cell saturation density had not yet been attained), diluted to 2 X 105/ml in MenSO-rich medium, and grown for 8 days, essentially as described by Friedman-Einat et al. (1982). Although the percentage of benzidinepositive cells attained nearly 100% in all these conditions, no increase in 2-54 synthetase activity could be detected (data not shown). Likewise, no increase in 2-5A synthetase was observed when cells were diluted in Me2SO-free medium (data not shown).
Since even very low levels of 2-5A synthetase activity could suffice to activate RNase L and therefore be biologically meaningful, these negative data have been re-assessed with a radioimmunological assay of 2-5A cores (Cailla et al., 1982a). The cell-free products of 2-5A synthetase were dephosphorylated with alkaline phosphatase and quantified using monoclonal antibodies specifically detecting as low as mol of 2-5A cores (Cailla et aL, 1982b). Fig. 2 shows the linear conversion of ATP into 2-5A expected for an enzymatic assay. The use of appropriate dilutions allowed detection of both the induction of 2-5A synthetase activity upon exogenous IFN treatment of 745A~ cells (Fig. !&I) and a low but measurable basal 2-5A synthetase activity in cell-free extracts prepared from untreated cells (Fig. 2B). Similar low levels of 2-5A synthetase activity were detected in extracts of IFN-resistant variants, regardless of stimulation by IFN (data not shown). As described in Table I, however, no modulation of 2-5A synthetase activit,y was detectable by this sensitive assay procedure in extracts prepared from Me2SO-treated cells, even in the conditions described by Friedman-Einat et al. (1982).   ylenebisacetamide-Since Me2S0 has been shown to increase 2-5A synthetase activity in several cell lines (Besanqon et al., 1981),' the effects of HMBA, another potent inducer of differentiation in the FLC system (Reuben et al., 1980), have been investigated. As shown in Table 11, nearly 100% of either 745A~ or 745A~ cells became benzidine-positive after 6 days of exposure to 5 mM HMBA. 2-5A synthetase activity remained undetectable in fully differentiated 745AE and 745AR l C l l cells after HMBA treatment, thus confirming the data obtained with MezSO for these two sublines. Surprisingly, HMBA did not stimulate 2-5A synthetase activity in 745AH cells (Table II), at variance with the data obtained with M e 3 0 (see above). HMBA even diminished the basal level of 2-5A synthetase activity which could be measured with the more sensitive radioimmunoassay of 2-5-4 products ( Table  11). The latter observations could not be accounted for by a trivial inhibitory effect, since HMBA did not affect 2-5A synthetase activity whether it was added to intact IFN-treated cells (Table 11)  After 3 days, cells were diluted to 2 X 105/ml with medium containing the same inducer and grown for an additional 4 days. Cell extracts were prepared, as described under "Experimental Procedures," 3, 5, and 7 days after cell seeding. Percentages of B' cells were 49, 85, and 89 for Me2SO-treated cells and 80, 97, and 99 for HMBA-treated cells, when measured 3, 5, and 7 days after seeding, respectively. The patterns of phosphorylation of the cell extracts, in the absence (-) or presence (+) of poly(r1) .poly(rC), were analyzed by polyacrylamide gel electrophoresis as described in the text. Assay mixture applied to each lane were as follows: a, 745A~ (3 days); b, 7 4 5 A~ + Me2S0 (3 days); c, 745& + HMBA (3 days); d, 745& + npIFN (3 days); e, 745A~ (5 days); 1,745A~ + Me2S0 (7 days); g, 745& + HMBA (5 days); h, 7458 (7 days); i, 745& + Me2S0 (7 days); 1, 7 4 5 A~ + HMBA (7 days); rn, L929 + nBIFN (1 day). The indicated molecular weights correspond to those of bovine serum albumin (68,000) and chymotrypsinogen (25,000). FLC-If IFN were released during FLC differentiation, as recently proposed (Friedman-Einat et al., 1982), it should increase the double-stranded RNA-activated protein kinase activity. Extracts were thus prepared from FLC induced to differentiate with either HMBA or Me,SO, for increasing periods of time and incubate with y3'P-labeled ATP with or without double-stranded RNA. None of these conditions led to an increased level of phosphorylation of the M, = 67,000 protein band in 745AR and 745AR IC11 sublines (Fig. 3) and in 745AH up to day 6 (data not shown). On the contrary, extracts prepared from exogenous IFN-treated FLC did exhibit the characteristic double-stranded RNA-dependent increase of the M, = 67,000 band phosphorylation, although less pronounced than for IFN-treated L929 cells taken as a positive reference (Fig. 3).

2-5A
Effects of IFN Antiserum on FLC Erythroid Differentiation-Neutralization of IFN effects with a specific antiserum should provide an alternative approach to the elucidation of the role of IFN-induced enzymes in erythroid cell differentiation, if indeed any increase in their activity resulted from IFN release. FLC were treated with an antiserum raised against an unfractionated preparation of mouse nPIFN at doses shown in preliminary experiments to completely inhibit exogenous IFN-induced 2-5A synthetase activity. As described in Fig. 4, such a treatment did not significantly affect FLC growth (A) or differentiation induced by MezSO as measured by the increase in benzidine-positive cells (B) or by spectrophotometric determination of hemoglobin content (C). IFN was included in parallel experiments as a control of antiserum efficacy, as described in Fig. 4.

DISCUSSION
The terminal differentiation of one FLC subline is accompanied by an increase of the IFN-induced 2-5A synthetase, as originally described by Kimchi (1981). However, the data presented in this paper do not support a general role of 2-5A synthetase induction in the differentiation of erythroid cells, as: 1) an increase in 2-5A synthetase activity following Me2S0 treatment is not observed in a closely related but independently passaged FLC subline; and 2) HMBA does not modulate 2-5A synthetase activity even in the FLC subline ('i45A~) which shows an increase of 2-5A synthetase activity following Me2S0 treatment.
These data are in keeping with the absence of any detectable increase in 2-5A synthetase activity in five of five of the IFNresistant variants isolated in this laboratory (Affabris et al., 1983)3 while all of them respond to inducers of differentiation The observed differences in the ability of Me2S0 and HMBA to increase 2-5A synthetase activity in at least ;ne FLC subline is not surprising. It has been shown that differentiation inducers such as MezSO or Na butyrate increase 2-5A synthetase activity in several unrelated lines such as C243 murine fibroblasts (Besanqon et al., 1981) or HeLa cells? which are not known to express markers specific to a given differentiation lineage.
Since subnanomolar amounts of 2-5A suffice to activate RNase L (Lengyel, 1981), conventional assays for measuring 2-5A synthetase activity might not have been sensitive enough to detect low amounts of 2-5A products still of possible biological significance. Indeed, a recently developed sensitive radioimmunoassay of 2-5A (Cailla et al., 1982a) did allow US to measure previously undetectable basal levels of 2-5A synthetase activity in all three sublines. Yet, Me2S0 was unable to modulate these levels in 745AR and the IFN-resistant variant cells. We cannot exclude, however, a role of the 2-5A synthetase recently detected in the nuclei (Nilsen et al., 1982;St. Laurent et al., 1983) in the differentiation process of FLC, as experimental procedures conventionally used to assay 2-5A synthetase do not lead to a major disruption of the nuclei. Likewise, a localized activation of the 2-5A system (as hypothesized by Nilsen and Baglioni, 1979) in the intact differentiating FLC could have escaped our detection and yet play a significant biological role. 2-5A synthetase induction could merely reflect the excretion of IFN, as reported by Friedman-Einat et al. (1982).
There are examples of excretion of minute amounts of IFN undetectable by its biological assay. As we could not detect any IFN level in the supernatants of differentiating FLC,4 cells were exposed to a potent mouse aPIFN antiserum which would neutralize any IFN produced under these circumstances. As the differentiation process was not affected by this treatment, we can exclude a role of ''spontaneously'' produced nPIFN in the course of MezSO-induced hemoglobinization. In this respect, it is not surprising that protein kinase is not induced by Me2S0 or HMBA under these experimental conditions in 745AR sublines. Likewise, IFN antiserum suppresses the increased level of 2-5A synthetase activity accompanying mouse myeloid leukemia cell differentiation without preventing differentiation itself (Sokawa et al., 1981). This does not exclude, however, a role of IFN and of the 2-5A system in the regulation of growth and differentiation in other systems such as rat liver regenerating after partial hepatectomy (Etienne-Smekens et al., 1981) or lymphocyte activation (Kimchi, 1981).
Considering the presence of an elevated 2-5A synthetase activity reported in mammalian reticulocytes (Hovanessian and Kerr, 1978) and the data reported in this paper, we might hypothesize that any such increase takes place very late in the process of erythroid cell differentiation, e.g. a t a stage preceding closely the extrusion of the nucleus. Differences in FLC sublines would then easily explain why only some of them (see the data reported here for 745& and those reported by Kimchi (1981)) do reach this stage while others do not (see data reported here for 745AR and for HMBA-induced 745AH). Whatever the case, the 2-5A system per se (i.e. independently from exogenous IFN addition) would not play any role either in the commitment of FLC to terminal differentiation or in the hemoglobinization process which is fully expressed in this in vitro system (Cioi! et al., 1978).