Cell cycle controlling the silencing and functioning of mammalian activators

Brief Communication 1695

Cell cycle controlling the silencing and functioning of mammalian activators Alan C. Mullen*§, Anne S. Hutchins*§, Alejandro V. Villarino*, Hubert W. Lee*, Frances A. High*, Nezih Cereb‡, Soo Y. Yang‡, Xianxin Hua† and Steven L. Reiner* Naı¨ve CD4ⴙ helper T (TH) cells respond to stimulation by terminally differentiating into two mature classes, TH1 cells, which express interferon ␥ (IFN-␥), and TH2 cells, which express interleukin 4 (IL-4) [1]. The transcriptional activators T-bet [2, 3] and Gata-3 [4, 5] mediate commitment to the TH1 and TH2 fates, respectively, including chromatin remodeling of signature genes. The cytokine IL-12 fosters growth of committed TH1 cells [3], while IL-4 fosters growth of committed TH2 cells [6]. IL-12 and IL-4 also play critical roles in commitment by promoting transcriptional silencing of Gata-3 [7] and T-bet [3], respectively. We now show that both T-bet and Gata-3 are induced in a cell cycle-independent manner in bipotent progenitor cells. In contrast, both lineage-restricted gene induction by the activator proteins and heritable silencing of the transcription of each activator, the hallmarks of terminal differentiation, are cell cycle dependent. We found that cells that cannot cycle remain uncommitted and bipotent in response to the most polarizing signals for maturation. These results provide mechanistic insight into a mammalian model of terminal differentiation by illustrating that cell cycle-coupled epigenetic effects, as originally described in yeast [8, 9], may represent an evolutionarily conserved strategy for organizing signaling and cell fate. Addresses: * Department of Medicine and † Department of Cancer Biology, Abramson Family Cancer Research Institute, University of Pennsylvania, Philadelphia, Pennsylvania 19104. ‡ Histogenetics, Inc., and Center for Genetic Polymorphism, Hawthorne, New York 10532 Correspondence: Steven L. Reiner E-mail: [email protected] §These

two authors contributed equally to this work.

Received: 23 July 2001 Revised: 4 September 2001 Accepted: 18 September 2001 Published: 30 October 2001 Current Biology 2001, 11:1695–1699 0960-9822/01/$ – see front matter  2001 Elsevier Science Ltd. All rights reserved.

Results and discussion Heritable silencing of lineage-determining activators during the cell cycle

Consistent with prior observations [2, 4] (see Figure 1a for diagram of experimental system), we found that the transcriptional activators T-bet and Gata-3 are reciprocally expressed in terminally differentiated TH1 and TH2 cells, respectively (Figure 1b). In newly differentiating cells, IL-12 and IL-4 can inhibit transcription of Gata-3 [7] and T-bet [3], respectively. Thus, we found that naı¨ve cells stimulated with mitogens (hereafter, simply “stimulated”) in the presence of recombinant (r) IL-12 plus antiIL-4 antibody (TH1 conditions) exhibited repression of Gata-3, and cells stimulated in rIL-4 plus anti-IL-12 antibody (TH2 conditions) exhibited repression of T-bet (Figure 1c). In contrast, cells stimulated in anti-IL-12 antibody plus anti-IL-4 antibody expressed both activators (Figure 1d). The negative regulation of activator expression by IL-4 and IL-12 seems to play a causal role in cell fate, since retroviral transduction of each activator under nonpermissive conditions was sufficient to restore the prohibited fate (Figure 1g) [2, 3, 7]. Furthermore, we found that activator expression did not require entry into the cell cycle (Figure 1d), and even appeared to be augmented by G1 cell cycle arrest, using the drug mimosine. The preceding results suggest that, by default (in the absence of IL-4 and IL-12), progenitor cells are bipotent, expressing both T-bet and Gata-3, prior to entry into S phase of the cell cycle. We, therefore, examined whether progenitors could be bipotent under prohibitive cytokine conditions. While activator expression was repressed by specific cytokines in proliferating cells, the same signals failed to repress activator transcription in cells that were arrested in G1 using mimosine (Figure 1e,f). We also tested if G1 arrest would prevent terminal TH commitment. Naı¨ve cells were stimulated in prohibitive cytokines and either allowed to proliferate or arrested in G1. After 3 days, cells were released from arrest to determine whether they could assume the previously prohibited fate. Cells that were initially allowed to proliferate behaved as terminally differentiated TH cells, unable to adopt the prohibited fate efficiently (Figure 1h,i), while cells initially arrested in G1 behaved as undifferentiated cells, subsequently able to adopt either fate (Figure 1h,i; data not shown). The inability of arrested cells to terminally commit probably was not due to lack of appropriate stimulation, since virtually all cells expressed markers of activa-

1696 Current Biology Vol 11 No 21

Figure 1

Helper T cells as a model system for terminal differentiation. (a) Schematic of established and hypothetical regulation of gene expression. Naı¨ve (progenitor) cells, which have not undergone chromatin remodeling of IFN-␥ or IL-4 loci [14, 22], may rapidly express T-bet and Gata-3, activators of the TH1 and TH2 fate, respectively. Committed TH1 (blue) and TH2 (red) cells have undergone chromatin remodeling of IFN-␥ and IL-4 loci, respectively, mediated by T-bet [3] and Gata-3 [5]. Differentiation also involves heritable induction of T-bet or Gata-3, along with heritable silencing of the opposing activator [2, 4]. Experimental TH1 conditions inhibit development of TH2 fate because IL-12 favors growth of TH1 cells [3] and suppresses transcription of Gata-3 [7]. Experimental TH2 conditions inhibit TH1 fate because IL-4 favors growth of TH2 cells [6] and suppresses transcription of T-bet [3]. The present study tests the hypotheses (“?”) that S phase of the cell cycle plays a role in both (1) heritable silencing of activator loci and (2) the ability of activators to mediate heritable induction of signature cytokine loci. (b) T-bet (top) and Gata-3 (middle) mRNA levels were determined by RT-PCR in terminally differentiated TH1 and TH2 clones. Competitive amplification of HPRT was performed

with an internal standard (bottom panel, upper band). (c) Naı¨ve cells were stimulated in experimental TH1 (left) or TH2 (right) conditions for 3 days. (d) Naı¨ve cells were stimulated in the presence of both anti-IL-4 and anti-IL-12 antibodies for 2 days in the absence (⫺) or presence (⫹) of G1 arrest, using mimosine throughout the article. (e,f) Cells were stimulated in TH1 conditions or TH2 conditions for 3 days in the absence (⫺) or presence (⫹) of mimosine. (g) Stat6⫺/⫺ (left) or Stat4⫺/⫺ (right) cells were stimulated for 24 hr in TH1 or TH2 conditions prior to infection with bicistronic Gata-3- or T-bet-GFP retrovirus (RV). After 2 more days, GFP and cytokine expression were determined by flow cytometry. Percentages above each column indicate frequency of untransduced (left portion) and transduced (right portion) cells expressing the indicated cytokine (y axis). (h) Cells were stimulated for 3 days in TH1 conditions with (right) or without (left) G1 arrest. Cells were then washed (Release) and re-stimulated in TH2 conditions for 4 days, in the absence of cell cycle inhibitors. Percentage of IL-4-positive cells is indicated. (i) Cells were treated as in (h), except initial stimulation was in TH2 conditions and cells were then released into TH1 conditions.

tion, indicating successful receptor-mediated signal transduction [10] (data not shown).

mRNA within a cell division correlated with the ability to adopt the prohibited fate (Figure 2b). We do not yet know if the loss of activator with cell division contributes to pericentric repositioning of cytokine loci, another correlate of terminal commitment [12]. Together, these results suggest that the cytokine-driven silencing of activators and terminal commitment are regulated by the cell division cycle.

Terminal commitment, or the inability to adopt the previously prohibited fate, has been correlated with repeated stimulation [11] and extent of cell division [12]. We, therefore, examined activator expression as a function of cell division number in prohibitive culture conditions. Cells were loaded with carboxyfluorescein diacetate succinimidyl ester (CFSE) prior to stimulation, such that fluorescence intensity diminishes 2-fold with each cell division and can be tracked and sorted by flow cytometry (see Supplementary material). We found that mRNA levels of the prohibited activator remained detectable in undivided cells and became undetectable by the fifth division (Figure 2a). Moreover, we found that the levels of activator

The aforementioned results do not formally exclude the possibility that silencing is simply time dependent and that cell division is required merely to dilute residual activator mRNA from the progenitor cell. To address this, cells were stimulated and arrested either in early S phase, using hydroxyurea, or in G2/M, using nocodazole, prior to release (Figure 3a). Cells that were arrested in early

Brief Communication 1697

Figure 2

Cell division regulating repression of the activators Gata-3 and T-bet. (a) Cells were labeled with CFSE to resolve cell divisions (see Supplementary material) and stimulated in TH1 (left) or TH2 conditions (right). After 4 days, asynchronously dividing CD4⫹ cells were sorted by division number and mRNA levels among individual cell divisions were determined by RT-PCR for Gata-3 (TH1 conditions, left) and T-bet (TH2 conditions, right). (b) CFSE-labeled cells were stimulated and sorted as described in (a). Individual divisions were then washed extensively. Cells initially stimulated in TH1 conditions were restimulated in TH2 conditions plus anti-IFN-␥ antibody (10 ␮g/ml), while cells initially stimulated in TH2 conditions were restimulated in TH1 conditions. After 4–5 days, cells switched into TH2 conditions were analyzed for IL-4 expression (three separate experiments) and cells switched into TH1 conditions were analyzed for IFN-␥ expression (six separate experiments) by flow cytometry (see Supplementary material for sample data). The initial cell division number yielding the highest frequency of cytokine expression after switching was set at 100% (maximum), and the frequencies obtained from the other initial cell divisions were expressed as percentages of the maximum (% Max). In some experiments, divisions 0 and 1 were combined to obtain sufficient cell numbers.

S phase remained less terminally committed than cells arrested in G2/M, as judged by their relative ability to adopt the prohibited fate following release. These findings prompted us to characterize the transcription of an activator during the cell cycle using two separate signals that lead to silencing. In cultures containing rIL-4, silencing of T-bet was defective in cells arrested in G1 or S phase, whereas silencing appeared more efficient in those that progressed to G2/M (Figure 3b). In addition, TGF␤ signaling, which can inhibit cytokine expression during differentiation [13], was sufficient to silence T-bet (Figure 3c), but this also appeared dependent on passage through S phase of the cell cycle. Thus, the heritable silencing of lineage-defining activators seems to depend on passage through S phase, a finding reminiscent of silencing of the mating type loci in yeast, which occurs sometime in S phase and before G2/M, between the hydroxyurea and nocodazole arrest points [8]. Progenitors are, therefore, likely to remain bipotent as they enter their first cell cycle, regardless of prohibitive cytokine signaling.

Figure 3

Terminal TH commitment and silencing of activators during passage through S phase of the cell cycle. (a) Cells were stimulated in TH1 (left) or TH2 (right) conditions and simultaneously arrested either in S phase, using hydroxyurea (HU), or G2/M, using nocodazole (Noc). After 3 days, cells were washed (Release) and restimulated in the opposite conditions for 4 more days, in the absence of all cell cycle inhibitors. Cells were then analyzed for expression of CD4 (x axis) and either IL-4 or IFN-␥ (y axes) by flow cytometry. (b) Cells were stimulated for 3 days in rIL-4 plus anti-IL-12 antibody and either allowed to proliferate (far left lane) or arrested at the indicated stages of the cell cycle using mimosine (Mim), hydroxyurea (HU), aphidicolin (Aph), nocodazole (Noc), or taxol (Tax). Levels of T-bet (top) and HPRT (bottom) mRNA were determined by RT-PCR. (c) Cells were stimulated for 3 days in anti-IL-4 and anti-IL-12 antibodies, prior to RT-PCR as in (b). Human rTGF␤ (200 pM) and cell cycle inhibitors were used where indicated.

Cell cycle enabling activator function

It has been reported that progression into S phase plays an important role in initiating expression of IFN-␥ and IL-4 in newly differentiating TH cells [14, 15], although some studies have suggested that a requirement for S phase may not be absolute [16, 17]. In yeast, there is a precedent for an activator to require cooperativity from the cell cycle in order to induce a silenced gene [9]. We, therefore, tested whether the specific activators of TH fate are dependent on progression into the cell cycle in order to execute their function. To mark the expression of activator, we employed bicistronic retroviral vectors, with T-bet and GFP encoded in the same transcript. Cells were stimulated in conditions prohibitive for endogenous T-bet expression and then transduced with T-bet- or control-GFP retrovirus. After completion of the viral life cycle, cells were arrested with cell cycle inhibitors. Proliferating cells, as well as those that were arrested either in

1698 Current Biology Vol 11 No 21

Figure 4

IL-4-expressing cells (Figure 4b). Virtually all differentiation, however, was prevented in butyrate-treated or untreated cells that were arrested in G1 (Figure 4b). Thus, both the activators and the posttranslational modifications of histones that potentiate activation in a chromatin context seem to require downstream cooperativity from the cell cycle to confer TH fate, perhaps due to the opening of replication forks, which occurs between the mimosine and hydroxyurea arrest points [18]. Although there may be alternative pathways of TH differentiation [16, 17], the present results suggest that some activator-dependent functions cannot occur without entry into S phase. Summary

Activators of TH differentiation functioning in a cell cycle-dependent manner. (a) Stat4⫺/⫺ cells were stimulated in the presence of rIL-4 for 24 hr prior to infection with T-bet- (top row) or control- (bottom row) GFP retrovirus (RV) and continued culture in rIL-4. 24 hr after retroviral infection, some cells were treated with mimosine (Mim) or hydroxyurea (HU), as indicated. All groups of cells were analyzed for GFP (x axis) and IFN-␥ (y axis) expression 48 hr after retroviral infection. Frequency of IFN-␥-expressing cells in the untransduced (GFP-negative, left) and transduced (GFP-positive, right) populations is indicated above each column of cells. (b) Cells were stimulated for 3 days with no drug addition or in the presence of mimosine, sodium butyrate, or both, prior to analysis of IFN-␥ (x axis) and IL-4 (y axis) expression. Percentages of cytokine-expressing cells are indicated. In both (a) and (b), only CD4⫹ gated events are displayed.

S phase (Figure 4a) or in G2/M (data not shown), exhibited robust IFN-␥ expression, mediated specifically by T-bet. In contrast, the action of T-bet in specifying the IFN␥-expressing, TH1 fate was defective in cells that were arrested in G1 (Figure 4a). In light of the apparent requirement to traverse the G1/S boundary in order for T-bet to confer TH1 fate, we performed additional controls to demonstrate that G1 arrest did not prevent IFN-␥ production in naı¨ve CD8⫹ T cells (Supplementary material) or committed TH1 cells [3], and that the inhibitors were arresting cells at the appropriate stage of the cell cycle (Supplementary material; data not shown). Similarly, we found that ectopic Gata-3 was unable to mediate TH2 differentiation without progression into S phase (unpublished data). We also tested whether inhibiting histone deacetylases, an intervention known to potentiate TH differentiation [14], could bypass the requirement for the cell cycle. Naı¨ve TH cells stimulated for 3 days in the presence of sodium butyrate had increased frequency of IFN-␥- and

Our results support a model of helper T cell commitment in which lineage-determining activators are induced prior to entry into the cell cycle, yet dependent on entry into S phase of the cell cycle in order to execute a program of differentiation. Similarly, transcription of activator loci, which begins prior to commitment to the cell cycle, apparently can become heritably silenced only during progression through S phase of the cell cycle. We can only speculate about the reason for these associations, but coupling both the silencing and the functioning of activators to the cell cycle may ensure that any progenitor cell can initiate a program that confers diverse fates to its progeny, while remaining responsive to polarizing signals of maturation. Although previously described for telomeric chromatin in yeast [9] and in vitro models of gene remodeling using cell extracts [19], to our knowledge, the induction of cytokine genes by activators is one of the first examples of cell cycle-dependent derepression in metazoan cells. The induction of cytokine genes in TH cells and telomeric genes in yeast is further linked by their stochastic nature [3, 9, 14], even though the point in S phase when induction occurs is somewhat different. Additionally, our analysis of activator expression shows, for the first time in metazoa, that silencing of some loci may require coupling to the cell cycle, another phenomenon previously described in yeast [8]. It is still uncertain what aspect of S phase is essential for this form of silencing [20, 21], just as it is unclear why particular chromatin contexts cannot be activated in noncycling cells. These questions about cell cycle-coupled epigenetic effects, potentially prevalent controls in metazoan differentiation, may now also be explored in a mammalian model system. Supplementary material Supplementary material, including the Materials and methods section and additional data, is available at http://images.cellpress.com/supmat/ supmatin.htm.

Acknowledgements We are grateful to G. Koretzky, M. Bartolomei, C. Vakoc, and M. Harris for critical comments; J.D. Engel for the Gata-3 cDNA; W. Pear and L. Xu for advice and reagents; and W. DeMuth for cell sorting. This work was supported by the National Institutes of Health (AI42370 to S.L.R., EY07131 to A.C.M., and AI10662 to A.V.V.).

Brief Communication 1699

References 1. Glimcher LH, Murphy KM: Lineage commitment in the immune system: the T helper lymphocyte grows up. Genes Dev 2000, 14:1693-1711. 2. Szabo SJ, Kim ST, Costa GL, Zhang X, Fathman CG, Glimcher LH: A novel transcription factor, T-bet, directs Th1 lineage commitment. Cell 2000, 100:655-669. 3. Mullen AC, High FA, Hutchins AS, Lee HW, Villarino AV, Livingston DM, et al.: Role of T-bet in commitment of TH1 cells before IL12-dependent selection. Science 2001, 292:1907-1910. 4. Zheng W, Flavell RA: The transcription factor GATA-3 is necessary and sufficient for Th2 cytokine gene expression in CD4 T cells. Cell 1997, 89:587-596. 5. Ouyang W, Lohning M, Gao Z, Assenmacher M, Ranganath S, Radbruch A, et al.: Stat6-independent GATA-3 autoactivation directs IL-4-independent Th2 development and commitment. Immunity 2000, 12:27-37. 6. Kurt-Jones EA, Hamberg S, Ohara J, Paul WE, Abbas AK: Heterogeneity of helper/inducer T lymphocytes. I. Lymphokine production and lymphokine responsiveness. J Exp Med 1987, 166:1774-1787. 7. Ouyang W, Ranganath SH, Weindel K, Bhattacharya D, Murphy TL, Sha WC, et al.: Inhibition of Th1 development mediated by GATA-3 through an IL-4-independent mechanism. Immunity 1998, 9:745-755. 8. Miller AM, Nasmyth KA: Role of DNA replication in the repression of silent mating type loci in yeast. Nature 1984, 312:247-251. 9. Aparicio OM, Gottschling DE: Overcoming telomeric silencing: a trans-activator competes to establish gene expression in a cell cycle-dependent way. Genes Dev 1994, 8:1133-1146. 10. Doyle AM, Mullen AC, Villarino AV, Hutchins AS, High FA, Lee HW, et al.: Induction of Cytotoxic T Lymphocyte Antigen 4 (CTLA-4) restricts clonal expansion of helper T cells. J Exp Med 2001, 194:893-902. 11. Murphy E, Shibuya K, Hosken N, Openshaw P, Maino V, Davis K, et al.: Reversibility of T helper 1 and 2 populations is lost after long-term stimulation. J Exp Med 1996, 183:901-913. 12. Grogan JL, Mohrs M, Harmon B, Lacy DA, Sedat JW, Locksley RM: Early transcription and silencing of cytokine genes underlie polarization of T helper cell subsets. Immunity 2001, 14:205215. 13. Gorelik L, Flavell RA: Abrogation of TGF␤ signaling in T cells leads to spontaneous T cell differentiation and autoimmune disease. Immunity 2000, 12:171-181. 14. Bird JJ, Brown DR, Mullen AC, Moskowitz NH, Mahowald MA, Sider JR, et al.: Helper T cell differentiation is controlled by the cell cycle. Immunity 1998, 9:229-237. 15. Richter A, Lohning M, Radbruch A: Instruction for cytokine expression in T helper lymphocytes in relation to proliferation and cell cycle progression. J Exp Med 1999, 190:1439-1450. 16. Laouar Y, Crispe IN: Functional flexibility in T cells: independent regulation of CD4⫹ T cell proliferation and effector function in vivo. Immunity 2000, 13:291-301. 17. Ben-Sasson SZ, Gerstel R, Hu-Li J, Paul WE: Cell division is not a “clock” measuring acquisition of competence to produce IFN-␥ or IL-4. J Immunol 2001, 166:112-120. 18. Krude T: Mimosine arrests proliferating human cells before onset of DNA replication in a dose-dependent manner. Exp Cell Res 1999, 247:148-159. 19. Barton MC, Emerson BM: Regulated expression of the betaglobin gene locus in synthetic nuclei. Genes Dev 1994, 8:2453-2465. 20. Li Y-C, Cheng Z-H, Gartenberg MR: Establishment of transcriptional silencing in the absence of DNA replication. Science 2001, 291:650-653. 21. Kirchmaier AL, Rine J: DNA replication-independent silencing in S. cerevisiae. Science 2001, 291:646-650. 22. Agarwal S, Rao A: Modulation of chromatin structure regulates cytokine gene expression during T cell differentiation. Immunity 1998, 9:765-775.