Lymphocytes: Early Induction of Cyclin D3 IL

The Journal of Immunology

All-trans Retinoic Acid Stimulates IL-2-Mediated Proliferation of Human T Lymphocytes: Early Induction of Cyclin D31

Nikolai Engedal,* Tone Gjevik,* Rune Blomhoff,† and Heidi Kiil Blomhoff2*

Vitamin A is established as an important immune regulator, but the mechanisms whereby vitamin A regulates T cell biology are poorly deﬁned. In this study, we show that an active metabolite of vitamin A, all-trans retinoic acid (RA), potently stimulates T cell proliferation by modulating IL-2-mediated signaling downstream of IL-2R and independent of the induction of IL-2. Thus, at concentrations as low as 0.1 nM, RA enhanced the division of normal human T lymphocytes that were simultaneously stimulated with anti-CD3 mAbs and saturating concentrations of IL-2. At the optimal concentration of RA (50 nM), a 3-fold increase in T cell proliferation was observed. The induced proliferation was preceded by increased phosphorylation of the retinoblastoma protein and enhanced G1- to S-phase progression. Interestingly, the promitogenic effect of RA was found to be particularly directed toward increased expression of cyclin D3 at both the mRNA and protein level. Furthermore, the stimulatory effect of RA on cyclin D3 expression as well as on cell proliferation was completely abolished in the presence of the JAK inhibitor AG-490 or blocking IL-2R␣ mAbs, and RA also enhanced cyclin D3 expression and T cell proliferation in the presence of IL-2 alone. Finally, we showed that the proliferative effect of RA was mimicked by agonists of the retinoic acid receptor (RAR) and completely inhibited by a RAR-selective antagonist. In conclusion, our results indicate that RA, via RAR, stimulates IL-2-induced signaling in a JAK-dependent manner to enhance cyclin D3 expression and thereby promote T cell proliferation. The Journal of Immu- nology, 2006, 177: 2851–2861.

itamin A is required for normal immune function (1), In the latter case, progression into S phase can be achieved by and a deficiency in this vitamin is associated with in- addition of exogenous IL-2. V creased morbidity and mortality from infections (2). The A key event in the regulation of S-phase entry is the phosphor- immune regulatory activities of vitamin A are mediated by its ylation of the retinoblastoma protein (pRB) (14). During G1, pRB acidic derivatives, most notably by all-trans retinoic acid (RA)3 is phosphorylated at multiple sites by various cyclin-dependent (3). We and others have shown previously that RA can potentiate kinases (CDKs), leading to the release of E2F transcription factors the proliferation of T lymphocytes in vitro (4–7), suggesting that and the subsequent induction of S-phase genes (15). CDKs are vitamin A may enhance immune function by directly modulating positively regulated by D-, E-, and A-type cyclins, and negatively mitogenic signals in T cells. regulated by CDK inhibitors (CKIs) such as p27Kip1 and p21Cip1, Resting T lymphocytes require two sequential signals to prolif- and thus G1- to S-phase transition is largely dictated by the erate. First, antigenic stimulation of the TCR/CD3 complex drives relative expression levels of cyclins and CKIs (16). Down-reg- the cells from their quiescent state (G0) into the G1 phase of the ulation of p27Kip1 protein levels has been proposed to be the cell cycle (8, 9). Second, the binding of IL-2 to its high affinity decisive effect of IL-2 on the cell cycle progression of mature receptor leads to G - to S-phase transition and thus proliferation 1 murine T cells (8, 17). (10). Strong TCR/CD3 signaling leads to expression of both IL-2R Physiologic IL-2 signaling occurs through the high affinity IL- and IL-2, thus creating an autocrine/paracrine proliferative loop 2R, which consists of three subunits: IL-2R␣, ␤, and ␥. Resting T (11, 12). In contrast, weak TCR/CD3 signaling leads to expression cells display low to intermediate levels of the ␤- and ␥-chains on of IL-2R, but not IL-2, and the cells are arrested in G (8, 11, 13). 1 their surface, but the ␣-chain is virtually absent (18, 19). Upon T cell activation, the expression of IL-2R␣ is rapidly and dramati- cally induced, whereas the levels of IL-2R␤ and ␥ are more mod- *Department of Biochemistry, Institute of Basic Medical Sciences, University of Oslo, Oslo, Norway; and †Department of Nutrition, Institute of Basic Medical Sci- erately increased at later time points (20–22). Upon IL-2 binding, ences, University of Oslo, Oslo, Norway the ␤- and ␥-chain-associated kinases JAK1 and JAK3 are acti- Received for publication October 20, 2005. Accepted for publication June 9, 2006. vated through reciprocal phosphorylation (23–25). JAK-mediated The costs of publication of this article were defrayed in part by the payment of page phosphorylation of tyrosine residues on IL-2R␤ leads to the re- charges. This article must therefore be hereby marked advertisement in accordance cruitment of the transcription factor STAT5 and the adaptor pro- with 18 U.S.C. Section 1734 solely to indicate this fact. tein Shc. STAT5, which is directly phosphorylated and thereby 1 This work was supported by the Norwegian Cancer Society, the Norwegian Re- search Council, Freia Research Foundation, Jahre Research Foundation, and the Blix activated by JAKs (23), seems to play a major role in IL-2-medi- Family Legacy. ated T cell proliferation (26, 27). Furthermore, Shc recruitment to 2 Address correspondence and reprint requests to Dr. Heidi Kiil Blomhoff, Depart- IL-2R␤ results in the activation of PI3K-protein kinase B (PKB)/ ment of Biochemistry, Institute of Basic Medical Sciences, University of Oslo, P.O. Akt and Ras-Raf-MAPK/ERK signal pathways, both of which are Box 1112 Blindern, N-0317 Oslo, Norway. E-mail address: h.k.blomhoff@ basalmed.uio.no thought to be involved in IL-2-mediated T cell proliferation (24, 3 Abbreviations used in this paper: RA, all-trans retinoic acid; CDK, cyclin-depen- 25). Additionally, IL-2R ligation leads to the activation of STAT3, dent kinase; CKI, CDK inhibitor; PI, propidium iodide; PKB, protein kinase B; pRB, which may also contribute to mitogenic signaling (24). retinoblastoma protein; RAR, retinoic acid receptor; RXR, retinoid X receptor; TTNPB, 4-((E)-2-(5,6,7,8-tetrahydro-5,5,8,8-tetramethyl-2-naphthalenyl)-1-prope- We have shown previously that when isolated peripheral blood nyl)benzoic acid. T lymphocytes are suboptimally stimulated, RA potently augments

␮ G1- to S-phase progression and proliferation by enhancing the pro- T cells were cultured in triplicates in 96-well plates, pulsed with 1.25 Ci duction of IL-2 (6, 7). However, in our previous studies, we had of [3H]thymidine (TRA120; Amersham Biosciences) for 18 h, transferred observed that RA also had a significant, albeit less potent effect on to Unifilter-96 GF/C filters (PerkinElmer) with a Packard FilterMate cell harvester, and counted on a Topcount liquid scintillation counter (Packard the proliferation of T cells under IL-2R-saturating conditions (6). Instrument) using 25 ␮l of MicroScint mixture (Packard Instrument) Moreover, the amount of exogenous IL-2 needed to mimic the per well. effect of RA was higher than the level of IL-2 that was actually produced by RA (6). These observations led us to hypothesize that in addition to modulating TCR/CD3-initiated signals and thereby Analysis of CD69, CD25, CD122, and CD132 cell surface increasing IL-2 production, RA may also stimulate IL-2-induced expression mitogenic signaling per se. In the present study, we tested this T lymphocytes (6 ϫ 105 cells) were washed with PBS-0.5% FBS and hypothesis by assessing the effect of RA on highly purified human resuspended in 50 ␮l of PBS-0.5% FBS into which 5 ␮l of anti-CD69 FITC peripheral blood T lymphocytes that were simultaneously stimu- (555530; BD Pharmingen), 10 ␮l of anti-CD25 FITC (P3191; Diatec), 5 ␮l of anti-CD122 PE (Mik-␤3, 554525; BD Pharmingen), 5 ␮l of anti-CD132 lated with anti-CD3 mAbs and high concentrations of IL-2. We PE (AG184, 555900; BD Pharmingen), or 5 ␮l of FITC- or PE-conjugated found that T cell proliferation was potently enhanced by physio- isotype-matched (IgG1) control Abs (P2011, Diatec, and 349043, BD Bio- logic concentrations of RA, and that this event was closely linked sciences, respectively) was mixed. After 15 min at room temperature in the to an early induction of cyclin D3 at both the RNA and protein dark, cells were washed twice with PBS-0.5% FBS, and 10,000 cells were level. Furthermore, we found that the effect of RA was mediated analyzed by flow cytometry using a FACSCalibur flow cytometer and CellQuest software (BD Biosciences). by the retinoic acid receptor (RAR) and that it was completely dependent on IL-2-induced JAK activity. Western and Northern blot analyses Materials and Methods For Western blot analyses, T cells were lysed in radioimmunoprecipitation Reagents assay (RIPA) buffer, as previously described (29). Protein concentrations RA, 4-((E)-2-(5,6,7,8-tetrahydro-5,5,8,8-tetramethyl-2-naphthalenyl)-1- were determined by Bradford assay (Bio-Rad) of duplicate samples and propenyl)benzoic acid (TTNPB), Am580, propidium iodide (PI), and BrdU equalized with RIPA buffer. Equal amounts of protein were loaded onto were from Sigma-Aldrich. Ro 41-5253 was provided by M. Klaus (Hoff- SDS-polyacrylamide gels, and Western blot analyses were performed, as previously described (29). Northern blot analyses were performed, as pre- mann-LaRoche, Basel, Switzerland). Retinoids were dissolved in ethanol ϫ 7 or DMSO, flushed with argon, and stored in lightproof containers at viously described (30). RNA was isolated from 1 10 T cells using the Ϫ20°C. Experiments with retinoids were performed in subdued light. RA RNeasy Mini Kit (Qiagen). The probes were as follows: for cyclin D3, a stability was checked weekly on the basis of UV spectra. Human rIL-2, 1.4-kb EcoRI/BamHI fragment from pEF1-HisC-cyclin D3 (see Ref. 45), anti-IL-2R␣ mAb (MAB223), and isotype-matched control mAb and for c-myc, a 1.4-kb ClaI/EcoRI fragment from pMC41-3RC (31). (MAB002) were from R&D Systems. AG-490 was from Calbiochem. OKT-3 (anti-CD3␧ mAb) was purified from OKT-3-producing hybridoma cell culture supernatants. Anti-pRB Ab (G3-254) was from BD Pharmin- Statistical analyses gen. Abs against p27Kip1 (C-19), p21Cip1 (C-19), STAT5 (C-17), c-Myc SPSS12.0.1 for Windows was used to perform paired samples t test (9E10), cyclin D2 (C-17), E (HE12), and A (C-19) were from Santa Cruz analyses. Biotechnology. Anti-cyclin D3 mAb (DSC-22) was from MBL. Abs against phospho-STAT3 (Tyr705, 9135), phospho-STAT3 (Ser727, 9134), STAT3 (9132), phospho-STAT5 (Tyr694, 9351), phospho-Akt (Ser473, 9271), phospho-p44/42-MAPK (9101), and p44/42-MAPK (9102) were Results from Cell Signaling Technology. RA potentiates T cell proliferation induced by OKT-3 and IL-2 T lymphocyte isolation, culture, and treatment We wished to determine whether RA is able to modulate IL-2- PBLs were isolated from healthy human blood donors (Ullevaal Hospital induced mitogenic signaling in T lymphocytes. Previously, we had Blood Bank), as previously described (6). Next, CD4ϩ T lymphocytes shown that RA stimulates IL-2 production in isolated human T were positively selected by use of magnetic CD4 MicroBeads (Miltenyi cells (6, 7). To avoid the contribution to T cell proliferation caused Biotec), as described by the manufacturer. More than 98% of the isolated by RA-induced IL-2 production in the present study, we added cells were positive for CD3 and CD4, as measured by flow cytometry using FITC- or PE-conjugated Abs against CD3 (SK7) and CD4 (SK3) (BD IL-2R-saturating concentrations of human rIL-2 (4 ng/ml) (32) to- Biosciences). T cells were cultured in standard 96- or 24-well flat-bottom gether with anti-CD3 mAbs (OKT-3) at the beginning of the cul- microtiter plates (BD Biosciences; Falcon 3072 and 3047) at 1.5 ϫ 106 tures. In initial experiments performed in the presence of 10% cells/ml in RPMI 1640 supplemented with 2 mM glutamine, 125 U/ml FBS, we found that RA treatment resulted in a slight increase in T ␮ penicillin, and 125 g/ml streptomycin at 37°C in a humidified incubator cell proliferation (data not shown and Fig. 1A). FBS is known to with 5% CO2. In some experiments, the culture medium was in addition supplemented with heat-inactivated FBS to final concentrations of 2.5 or contain the parent vitamin A compound retinol (33, 34), which 10% (v/v). OKT-3, IL-2, and RA were added simultaneously at the start of may be intracellularly metabolized to RA and thereby mask the the cultures. Unless otherwise stated, the final concentrations of OKT-3, effect of exogenously added RA. The retinol content in a given ␮ Ϸ IL-2, and RA were 2.5 g/ml, 4 ng/ml ( 250 pM), and 50 nM, batch of FBS will vary depending on, for example, conditions of respectively. storage and handling. Because the heat-inactivated FBS used in the Determination of retinol concentration in FBS present study was found to contain ϳ0.6 ␮M retinol (data not shown), we decided to compare the effect of RA in the absence or The retinol concentration in heat-inactivated FBS was determined by HPLC analysis, as previously described (28). The FBS did not contain presence of FBS. Cell divisions were tracked by CFSE staining detectable amounts of RA. after 4 days of culture, as previously described (7). Only a slight effect of RA was observed in the presence of 2.5 or 10% FBS (Fig. Cell proliferation assays 1A). In the absence of FBS, however, RA potently enhanced T cell To determine cell divisions, freshly isolated T lymphocytes were stained proliferation induced by OKT-3/IL-2. Thus, under serum-free con- with CFSE (Molecular Probes), as previously described (7), except that the ditions, a Ͼ3-fold increase in the percentage of divided cells was final concentration of CFSE was 0.15 ␮M. Cell cycle distribution was analyzed by incubating cells with BrdU (10 ␮M) for 1 h, followed by observed in the presence of RA (Fig. 1A). Serum-free conditions staining with a FITC-conjugated anti-BrdU Ab (556028; BD Pharmingen) were therefore used in the subsequent experiments to study the and PI, as previously described (29). For determination of DNA synthesis, effect of RA on T cell proliferation. The Journal of Immunology 2853

FIGURE 1. RA potentiates OKT-3/IL-2-induced T cell proliferation in a dose- and time-dependent manner. A, CFSE-stained T cells were either cultured in medium alone, or stimulated with OKT-3 and IL-2 with or without RA (100 nM) for 4 days, in the absence or presence of 2.5 or 10% FBS. Percentages of divided cells were measured, as described in Materials and Methods. Average percentages of divided cells Ϯ SEMs from four independent experiments are shown. e, u, and f, Represent cells cultured in medium alone, OKT3/IL-2, or OKT-3/IL-2/RA, respectively. B, CFSE-stained T cells were treated as indicated, and after 4 days percentages of divided cells were determined, as described in A. Average percentages of divided cells Ϯ SEMs from three independent experiments are shown. C, T cells were treated as indicated (RA concentrations: 0.01, 0.1, 1, 10, 50, and 100 nM), and after 48 h the cells were pulsed with [3H]thymidine for 18 h, as described in Materials and Methods. Average cpm values Ϯ SDs of triplicates from one representative experiment of three are shown. D, CFSE-stained T cells were treated with OKT-3, IL-2, and RA (100 nM), as indicated, and analyzed for percentages of divided cells after 3, 4, or 5 days. Gatings were set to accommodate for loss of CFSE ﬂuorescence by dying cells, as previously described (7). Percentages of divided cells are indicated. One representative experiment of eight is shown. E, CFSE-stained T cells from eight blood donors were treated and analyzed p Ͻ 0.01; paired samples t test). e, u, and f, Represent ,ءء) for cell divisions, as described in D. Average percentages of divided cells Ϯ SEMs are shown cells cultured in medium alone, OKT3/IL-2, or OKT-3/IL-2/RA, respectively. F, In the same experiments as those described in E, cell death was determined .(p Ͻ 0.01; paired samples t test, n ϭ 8 ,ءء ;p Ͻ 0.05 ,ء) by PI staining. Average percentages of dead cells Ϯ SEMs are shown

RA potentiates T cell proliferation in a dose- and time- RA were noted at very low concentrations. In fact, in these exper- dependent manner iments, effects of RA were observed already at 10 pmol (Fig. 1C). Using the CFSE-staining technique, we tested the dose depen- Next, we examined the kinetics of the RA effect. CFSE staining dency of the effect of RA on OKT-3/IL-2-induced T cell prolifer- was used to track cell divisions at various time points after T cell ation. RA enhanced T cell proliferation at doses as low as 0.1 nM, stimulation. As shown in the dot blots from one representative whereas optimal effects were seen at 10–100 nM (Fig. 1B). At 1 experiment (Fig. 1D), T cells cultured in medium alone did not ␮M, the effect of RA was essentially lost (Fig. 1B), probably due undergo cell division. No cell division was observed in stimulated to the increased cell death observed at this concentration of RA T cells at day 1 or 2 (data not shown), but at day 3 RA had induced (data not shown). T cell division was not stimulated by RA alone one round of cell division in a proportion of T cells (Fig. 1D). At (Fig. 1B). The dose-dependent effect of RA was conﬁrmed by mea- day 4, a larger proportion of T cells had divided once and a com- surement of DNA synthesis; as shown in Fig. 1C, optimal uptake parable proportion of the cells had undergone two cell divisions in of [3H]thymidine was obtained at 100 nM RA, and again effects of the presence of RA. Thus, 42% of RA-treated T cells had divided, 2854 RA STIMULATES IL-2-MEDIATED T CELL PROLIFERATION

FIGURE 2. RA stimulates the cell cycle machinery and S-phase entry in OKT-3/IL-2-treated T cells. A, T cells were either cultured in medium alone (M), or stimulated with OKT-3 and IL-2 in the absence (OI) or presence (OIR) of RA. After 24, 48, or 72 h, whole cell extracts were prepared, and Western blot analyses, using Abs against the indicated proteins, were performed, as described in Materials and Methods. Hypo- and hyperphosphorylated pRB proteins are indicated as pRB and ppRB, respectively. One representative experiment of three is shown. B, T cells were cultured in medium alone, or stimulated with OKT-3, IL-2, and RA, FIGURE 3. AG-490 inhibits T cell proliferation, cyclin D3 expression, as indicated. At days 1 or 2, cells were pulsed for 1 h with BrdU and and late IL-2R␣, ␤, and ␥ expression, but not the early induction of IL- analyzed for BrdU incorporation and DNA content, as described in Mate- 2R␣. A, T cells were treated with various concentrations of AG-490 as rials and Methods. Dead cells were excluded from the dot plots by gating indicated and either cultured in medium alone, or stimulated with OKT-3 for living cells on forward/side scatter plots (data not shown). Percentages and IL-2 in the absence or presence of RA. After 48 h, cells were pulsed of cells in S phase are indicated. One representative experiment of three is with [3H]thymidine for 18 h, as described in Materials and Methods.Av- shown. C, The same CFSE-stained T cells that had been analyzed for erage cpm values Ϯ SDs of triplicates from one representative experiment percentages of divided cells in Fig. 1E were now analyzed for number of of three are shown. e, u, and f, Represent cells cultured in medium alone, discernible cell divisions from dot plots, such as the example shown in Fig. OKT3/IL-2, or OKT-3/IL-2/RA, respectively. B, T cells were either cul- 1D. Average numbers of cell divisions Ϯ SEMs at days 3, 4, and 5 for T tured in medium alone (M), or stimulated with OKT-3 and IL-2 in the cells from the eight different blood donors are shown. absence (OI) or presence (OIR) of RA, as well as in the absence or presence of AG-490 (20 ␮M), as indicated. After 72 h, whole cell extracts were prepared, and Western blot analyses, using Abs against the indicated proteins, were performed. After immunoprobing with cyclin D3- or phospho- compared with only 10% in the absence of RA. After 5 days, 52% STAT3-specific Abs, each membrane was stripped and reprobed with Abs of RA-treated cells had divided, among which a substantial pro- recognizing total ERK (tot-ERK) or total STAT3 (tot-STAT3) proteins, portion of cells had undergone three cell divisions. In comparison, respectively, to control for equal loading. One reproducible experiment of two is shown. C and D, T cells were treated with OKT-3, IL-2, RA, and only 21% of the cells had divided in the absence of RA (Fig. 1D). AG-490 (20 ␮M), as indicated, and analyzed for cell surface expression of These data suggest that RA stimulates T cell proliferation by ac- CD25 (C) after 24 or 72 h (bold lines), or CD122 and CD132 (D) after 72 h celerating the T cell division process at a relatively early stage. (bold lines). In each case, FITC (C)- or PE-conjugated (D) isotype-matched Results from eight different blood donors (Fig. 1E) confirmed the control Abs were included as negative controls (dotted lines). CD25 pos- RA effects observed in Fig. 1D. The time-dependent effect of RA itively stained cells (C) and induced expression of CD122 or CD132 rel- was also confirmed when measuring DNA synthesis; the effect of ative to untreated cells (D) are indicated (%). The values in parentheses RA was notable at day 2, and at days 3 and 4 RA potently en- indicate mean fluorescence intensities of the gated cells. One representative hanced DNA synthesis (data not shown). experiment of three is shown. The Journal of Immunology 2855

more evident, but still relatively modest effect observed at day 3 (Fig. 2A). The expression of p21Cip1 was unaffected by RA (data not shown). In conclusion, the earliest and most striking effect of RA on the cell cycle machinery was enhanced expression of cyclin D3, followed by increased pRB phosphorylation. To confirm that the observed effect of RA on the cell cycle machinery was due to RA stimulating S-phase entry rather than enhancing the rate of cell division among dividing cells, we directly determined percentages of cells entering S phase after T cell stimulation by use of BrdU incorporation and PI staining (29). As shown in Fig. 2B, the first T cells entered S phase at day 2 after stimulation with OKT-3/IL-2. In the presence of RA, the percentage of cells in S phase was clearly increased (from 4.8 to 15.5%; Fig. 2B), demonstrating that RA indeed stimulated S-phase entry. That RA stimulated S-phase entry rather than enhancing the rate of cell division among dividing cells was further confirmed by ex- amining the number of cell divisions that had occurred over time. When the same cells as those described in Fig. 1E were analyzed FIGURE 4. A blocking anti-IL-2R␣ mAb abolishes OKT-3/IL-2-in- for number of cell divisions, we found that the rate of division was duced DNA synthesis and cyclin D3 expression. A, T cells were either not significantly enhanced by RA after day 3, i.e., after the first cell cultured in medium alone (med), or stimulated with OKT-3 and IL-2 (1 division had taken place (Fig. 2C). Thus, we concluded that RA ng/ml) in the absence (OI) or presence (OIR) of RA, as well as in the exerts its effects on T cell proliferation primarily by stimulating the absence or presence of anti-human IL-2R␣ mAb (␣-IL-2R␣-Ab; 10 ␮g/ cell cycle machinery and S-phase entry during the first cell ml), as indicated. After 48 h, the cells were pulsed with [3H]thymidine for 18 h, as described in Materials and Methods. Average cpm values Ϯ SEMs division. from three independent experiments are shown. B, T cells were either cultured in medium alone (M), or stimulated with OKT-3 and IL-2 (1 ng/ml) RA-mediated enhancement of T cell proliferation and cyclin D3 in the absence (OI) or presence (OIR) of RA, as well as in the absence or presence of anti-human IL-2R␣ mAb (␣IL-2R␣-Ab; 10 ␮g/ml) or an iso- expression is dependent on IL-2-induced signaling type-matched (IgG1) control mAb against an irrelevant protein (keyhole By using IL-2R-saturating concentrations of IL-2, we had ruled out limpet hemocyanin, which is not expressed in human cells) (irr-Ab; 10 the possibility that RA stimulated the cell cycle machinery and T ␮g/ml), as indicated. After 48 h, whole cell extracts were prepared, and cell proliferation by increasing IL-2 production. Still, we could not Western blot analyses, using Abs against cyclin D3, were performed. One exclude the possibility that RA mediated its effect by modulating reproducible experiment of two is shown. TCR/CD3-, and not IL-2-induced signaling. To elucidate which of these two signaling pathways was affected by RA, we used a specific inhibitor of JAK, AG-490 (35). By inhibiting the activity of Next, we assessed whether the enhanced cell division observed JAK-3, AG-490 efficiently blocks mitogenic signaling from IL-2R in the presence of RA could be a result of decreased cell death. As (35). First, we assessed the effects of various concentrations of shown in Fig. 1F, RA did not affect T cell viability at day 3 and AG-490 on OKT-3/IL-2-induced T cell proliferation. AG-490 in- caused only a minor reduction in cell death at days 4 and 5. Thus, hibited T cell proliferation in a dose-dependent manner, both in the the effect of RA on T cell division could not be explained by absence and presence of RA (Fig. 3A), without affecting cell via- inhibition of cell death. bility (data not shown). Moreover, the potency of RA to enhance T cell proliferation was gradually diminished with increasing RA stimulates the cell cycle machinery and S-phase entry in concentrations of AG-490. Thus, whereas RA induced a 4-fold T cells increase in [3H]thymidine uptake in the absence of the JAK Multiple phosphorylation of the pRB protein is a major require- inhibitor, only 2.9-, 1.8-, and 1.2-fold increases were observed in ␮ ment for G1- to S-phase transition (14). Hyperphosphorylated the presence of 2.5, 10, or 25 M AG-490, respectively (Fig. 3A). forms of pRB can be detected by their slower migration in SDS- Next, we assessed the effect of JAK inhibition on the expression polyacrylamide gels. As shown in Fig. 2A, pRB phosphorylation of cyclin D3. Strikingly, as shown in Fig. 3B, OKT-3/IL-2-induced was strongly enhanced by RA; the effect of RA was first notable at cyclin D3 expression was almost totally blocked by AG-490, both day 2, and at day 3 pRB was predominantly observed in its hy- in the absence and presence of RA. To verify that AG-490 inhib- perphosphorylated forms in the presence of RA. ited JAK activity, we assessed the effect of the inhibitor on STAT3 Phosphorylation of pRB is mediated by CDKs, whose activities tyrosine phosphorylation, which is known to be mediated by JAKs are positively and negatively regulated by cyclins and CKIs, re- (36). As shown in Fig. 3B, OKT3/IL-2-induced tyrosine phosphor- spectively. Each cyclin family comprises several members. T cells ylation of STAT3 was indeed markedly reduced by AG-490. express cyclin D2, D3, E1, E2, and A2 (8, 9). As shown in Fig. 2A, TCR/CD3-induced signaling leads to a rapid up-regulation of the protein levels of all these cyclins were clearly enhanced upon IL-2R␣ on the T cell surface (22). Thus, if AG-490 only blocks OKT-3/IL-2 treatment. Whereas cyclin D2 expression was unaf- signaling from IL-2R and not from the TCR/CD3 complex, the fected by RA, cyclin D3 protein levels were increased by RA early expression of IL-2R␣ should not be inhibited by AG-490. already at day 1, and at days 2 and 3 a strong up-regulation was Indeed, AG-490 did not affect the cell surface expression of IL- observed (Fig. 2A). Cyclin E levels were unaffected by RA, 2R␣ at day 1 (Fig. 3C). Moreover, IL-2R-␣ expression was unaf- whereas cyclin A expression was considerably up-regulated by RA fected by RA. We also examined expression of the activation at days 2 and 3. p27Kip1 levels were markedly, and in a time- marker CD69, whose early expression is mediated by TCR/CD3- dependent manner reduced by OKT3/IL-2 treatment (Fig. 2A). RA induced signaling (37). Like IL-2R␣, early CD69 expression was augmented this effect, with a minor effect noted at day 2, and a not inhibited by AG-490 and was also unaffected by RA (data not 2856 RA STIMULATES IL-2-MEDIATED T CELL PROLIFERATION shown). Once IL-2R␣ is expressed, its levels can be further up- 2R␣ mAb completely inhibited the effect of RA on OKT-3/IL-2- regulated by IL-2-induced signals (10). Thus, at day 3, it is ex- induced DNA synthesis. Moreover, OKT-3/IL-2-induced DNA syn- pected that the expression of IL-2R␣ is at least partly regulated by thesis was by itself almost fully repressed, indicating that T cell IL-2-induced signaling. Indeed, we found that the cell surface ex- proliferation under our conditions was almost entirely dependent on pression level of IL-2R␣ was partly inhibited by AG-490 at day 3 signaling from IL-2R. Next, we tested the effect of the IL-2R␣-block- (Fig. 3C). Moreover, although RA did not increase the percentage ing mAb on cyclin D3 expression. As shown in Fig. 4B, the Ab of IL-2R␣-positive cells, the average intensity of the staining was against IL-2R␣ totally repressed OKT-3/IL-2-induced cyclin D3 ex- doubled, and this effect was totally blocked by AG-490 (Fig. 3C). pression, both in the absence and presence of RA. Thus, we concluded Cell surface expression levels of IL-2R␤ and ␥ are enhanced upon that the effect of RA on cyclin D3 expression and T cell proliferation T cell activation and may be regulated by IL-2-induced signaling was indeed strictly dependent on IL-2-induced signaling. (20–22). Thus, we examined whether RA affected ␤- and ␥-chain expression in our system. Strikingly, the expression of both IL- RA stimulates signaling induced by IL-2 alone 2R␤ and ␥ was substantially increased by RA at day 3, and in both Resting T cells express IL-2Rs with intermediate IL-2 affinity cases this effect was completely abolished by AG-490 (Fig. 3D). (consisting of IL-2R␤␥ heterodimers) (18, 19). Therefore, high These results indicate that RA stimulates T cell proliferation and concentrations of exogenously added IL-2 can induce some pro- enhances the expression of cyclin D3 and IL-2R␣, ␤, and ␥,by liferative IL-2R-mediated signaling in the absence of T cell-acti- modulating IL-2-induced signaling. vating stimuli (18, 38). Thus, we tested whether RA could stimu- Although the JAK inhibitor AG-490 efficiently repressed IL-2- late the proliferation of T cells that were stimulated with IL-2 induced signaling, while leaving TCR-induced signaling unaf- alone. Indeed, as assessed by CFSE-staining experiments, RA en- fected, we could not formally exclude the possibility that AG-490 hanced T cell division in the presence of IL-2 in a dose-dependent was repressing the activity of other bioactive molecules than IL-2, manner (Fig. 5A). Furthermore, RA clearly potentiated the ability which were mediating the effects of RA. Therefore, to directly test of IL-2 to induce DNA synthesis (Fig. 5B), S-phase entry (Fig. for the dependency of RA on IL-2R-induced signaling, we used a 5C), and cyclin D3 expression (Fig. 5D), as well as the cell surface specific mAb against IL-2R␣, which blocks signaling downstream expression of IL-2R␥ (Fig. 5E). These results demonstrate that RA of the high affinity IL-2R. As shown in Fig. 4A, the blocking IL- is able to modulate IL-2-induced signaling also in the absence of

FIGURE 5. RA stimulates T cell proliferation and the expression of cyclin D3 and IL-2R␥ induced by IL-2 alone. A, CFSE-stained T cells were treated as indicated, and after 4 days, percentages of divided cells were determined, as described in Materials and Methods. Average percentages of divided cells Ϯ SEMs from three independent experiments are shown. B, T cells were either cultured in medium alone (medium) or treated with IL-2 and/or RA, as indicated. After 48 h, the cells were pulsed with [3H]thymidine for 18 h, as described in Materials and Methods. Average cpm values Ϯ SEMs from four independent experiments are shown. C, T cells were either cultured in medium alone (medium) or treated with IL-2 and RA, as indicated. At day 1 or 3, cells were pulsed for 1 h with BrdU and analyzed for BrdU incorporation and DNA content, as described in Materials and Methods. Dead cells were excluded from the dot plots by gating for living cells on forward/side scatter plots (data not shown). Percentages of cells in S phase are indicated. One reproducible experiment of two is shown. D, T cells were either cultured in medium alone (medium; Med) or treated with IL-2 and RA, as indicated. After 48 h, whole cell extracts were prepared, and Western blot analyses, using Abs against cyclin D3, were performed. One reproducible experiment of two is shown. E, T cells were either cultured in medium alone (medium) or treated with IL-2 and/or RA, as indicated, and analyzed for cell surface expression of CD132 (IL-2R␥) after 7 days (bold lines). In each case, PE-conjugated isotype-matched control Abs were included as negative controls (dotted lines). Induced expression of CD132 relative to untreated cells (medium) is indicated (%). The Journal of Immunology 2857

TCR/CD3-initiated signals, and importantly that RA can enhance As shown in Fig. 7A, STAT5 was strongly phosphorylated after 1 cyclin D3 expression and S-phase entry when acting in concert day of OKT-3/IL-2 treatment, and the level of phospho-STAT5 with IL-2-mediated signals alone. proteins remained high at days 2 and 3. However, at neither of these time points was STAT5 phosphorylation enhanced by RA. RA-mediated enhancement of IL-2R expression is a relatively STAT3 requires both tyrosine and serine phosphorylation for full late event activation (39). As shown in Fig. 7A, both tyrosine and serine Having established that the effects of RA on OKT-3/IL-2-induced phosphorylation of STAT3 was strongly induced by OKT-3/IL-2. cyclin D3 expression and T cell proliferation most likely were However, RA did not affect either phosphorylation events at any of mediated via regulation of IL-2-induced signaling, we next studied the time points examined. the potential mechanism(s) involved. Because we had observed We also examined STAT5 phosphorylation at earlier time that RA increased the expression of all three subunits of the IL-2R points. Elevated levels of phospho-STAT5 proteins were observed at day 3 (Fig. 3, C and D), we ﬁrst examined the kinetics of this RA already after 15 min of OKT-3/IL-2 treatment and peaked between effect. As shown in Fig. 6, none of the receptor subunits were 2 and 24 h (Fig. 7B). However, RA did not increase phospho- up-regulated by RA at day 1, and only the expression of IL-2R␤ STAT5 levels at any of the time points examined (Fig. 7B). In was increased at day 2. At day 3, the cell surface expression levels separate experiments, and in accordance with a previous study of both IL-2R␤ and ␥ were strongly increased (Fig. 6, B and C). (26), the early phosphorylation of STAT5 was found to be induced The percentage of cells staining positive for IL-2R␣ was not en- by IL-2 and not by OKT-3 (data not shown). hanced by RA (Fig. 6A). However, the average intensity of posi- RA does not affect OKT-3/IL-2-induced phosphorylation of tively stained cells was substantially increased at day 3, but not at PKB/Akt or MAPK/ERK, but enhances the expression of c-Myc day 1 or 2 (data not shown and Fig. 3C). These results indicate that RA-mediated enhancement of IL-2R expression is a relatively late IL-2R ligation activates the PI3K-PKB/Akt and Raf-MAPK/ERK event, making it unlikely that increased IL-2R expression can ex- pathways, both of which have been reported to up-regulate the plain the early and strong effects of RA on cyclin D3 expression.

RA does not affect OKT-3/IL-2-induced phosphorylation of STAT5 or STAT3 JAK-mediated tyrosine phosphorylation of STAT5 is a proximal IL-2-induced signaling event that is involved in IL-2-mediated T cell proliferation (23, 26, 27). We therefore examined whether RA-induced enhancement of cyclin D3 expression and T cell proliferation could be mediated by increased STAT5 phosphorylation.

FIGURE 7. RA enhances cyclin D3 mRNA levels and c-Myc protein and mRNA levels, but does not affect OKT3/IL-2-induced phosphorylation of STAT5, STAT3, PKB/Akt, or MAPK/ERK. A, T cells were either cultured in medium alone (M), or stimulated with OKT-3 and IL-2 in the absence (OI) or presence (OIR) of RA. After 24, 48, or 72 h, whole cell extracts were prepared, and Western blot analyses, using Abs against the indicated proteins, were performed. After immunoprobing with phospho- STAT-specific Abs (pSTAT), each membrane was stripped and reprobed with Abs recognizing total STAT5 (tot-STAT5) or total STAT3 (tot- STAT3) proteins. B, T cells were either harvested immediately after isolation (C), or stimulated with OKT3 and IL-2 in the absence (OI) or presence (OIR) or RA. At the indicated time points, whole cell extracts were prepared, and Western blot analyses, using Abs against the indicated proteins, were performed. One representative experiment of four is shown. C, T cells were either harvested immediately after isolation (C), or stimulated with OKT-3 and IL-2 in the absence (OI) or presence (OIR) of RA. After 5 h, total cellular RNA was isolated and subjected to Northern blot analysis FIGURE 6. Effect of RA on the cell surface expression levels of IL- of c-myc mRNA. The ethidium bromide-stained 28S and 18S rRNA bands 2R␣, ␤, and ␥. T cells were either cultured in medium alone, or stimulated are shown as loading control. One representative experiment of three is with OKT-3 and IL-2 in the absence or presence of RA. After 24, 48, or shown. D, T cells were either cultured in medium alone (M), or stimulated 72 h, the cells were analyzed for cell surface expression of CD25 (A), with OKT-3 and IL-2 in the absence (OI) or presence (OIR) of RA. After CD122 (B), or CD132 (C). Average percentages of gated cells Ϯ SEMs 36 h, total cellular RNA was isolated and subjected to Northern blot anal- from three independent experiments are shown. Gatings were set, as de- ysis of cyclin D3 mRNA. The ethidium bromide-stained 28S and 18S scribed in Figs. 3 and 5. e, u, and f, Represent cells cultured in medium rRNA bands are shown as loading control. One representative experiment alone, OKT3/IL-2, or OKT-3/IL-2/RA, respectively. of three is shown. 2858 RA STIMULATES IL-2-MEDIATED T CELL PROLIFERATION expression of D-type cyclins (40, 41). We therefore asked whether RA under our experimental conditions may enhance cyclin D3 expression by stimulating the activities of Akt and/or ERK. Phos- pho-Akt protein levels were enhanced within2hofOKT-3/IL-2 treatment and remained up-regulated for at least 22 h (Fig. 7B). However, at neither of the time points examined was Akt phosphorylation significantly affected by RA (Fig. 7B). OKT-3/IL-2 induced a strong and sustained phosphorylation of ERK; elevated levels of phospho-ERK proteins were observed already after 15 min of stimulation and peaked between 2 and 24 h (Fig. 7B). However, RA did not significantly alter phospho-ERK levels at any of the time points examined (Fig. 7B). Together, these results suggest that RA-mediated induction of cyclin D3 neither involves the PI3K-PKB/Akt nor the Raf-MAPK/ERK pathway. In support of this conclusion, we found that neither wortmannin (a PI3K inhibitor) nor U0126 (a MEK inhibitor) could diminish the effect of RA on cyclin D3 expression (data not shown). Furthermore, we found that these inhibitors did not affect the ability of RA to stimulate OKT3/IL-2-induced DNA synthesis (data not shown), suggesting that the potentiating effect of RA on IL-2-induced T cell FIGURE 8. RAR-selective agonists, TTNPB and Am580, mimic, and proliferation is independent of both Akt and ERK. RAR-selective antagonist, Ro 41-5253, abolishes the effect of RA on T cell proliferation. A, T cells were either cultured in medium alone (Med), or c-Myc is a transcription factor that is often implicated in the stimulated with OKT-3 and IL-2 (OI) in the absence or presence of: RA promotion of cell proliferation. The expression of c-Myc is up- (100 nM), TTNPB (100 nM), or Am580 (100 nM), as indicated. After 48 h, regulated by IL-2 through a JAK3-dependent pathway (10). It was the cells were pulsed with [3H]thymidine for 18 h, as described in Mate- recently reported that c-Myc can enhance the expression of cyclin rials and Methods. Average cpm values Ϯ SEMs from six independent D3 (42). We therefore examined whether RA could enhance c- experiments are shown. B, T cells were either cultured in medium alone Myc expression. Protein levels of c-Myc were consistently in- (Med), or stimulated with OKT-3 and IL-2 in the absence (OI) or presence creased within2hofOKT3/IL-2 treatment and remained up-reg- of RA (0.5 nM) (OIR) and/or Ro 41-5253 (Ro) (250 nM), as indicated. ulated for at least 22 h (Fig. 7B and data not shown). In the After 48 h, the cells were pulsed with [3H]thymidine for 18 h, as described presence of RA, c-Myc expression was even further enhanced, the in Materials and Methods. Average cpm values Ϯ SEMs from three in- effect of RA being most prominent at 5 h (Fig. 7B). RA also independent experiments are shown. creased the mRNA levels of c-myc at 5 h, as assessed by Northern blot analysis (Fig. 7C). These results indicate that an early induction of c-Myc could at least in part mediate the effect of RA on effect of RA on OKT3/IL-2-induced T cell proliferation was me- cyclin D3 expression. diated via RAR.

RA up-regulates cyclin D3 mRNA levels Discussion It is evident that cyclin D3 protein expression can be regulated at In the present study, we demonstrate for the ﬁrst time that vitamin multiple levels. These include regulation of gene transcription A can stimulate IL-2-mediated proliferation of peripheral blood T (43), mRNA stability (44), mRNA translation (30), and protein cells. Physiological concentrations of RA potently augmented stability (45). To elucidate how RA enhances cyclin D3 protein DNA synthesis and promoted the division of freshly isolated hu- levels under our experimental conditions, we determined cyclin D3 man T lymphocytes that were simultaneously stimulated with anti- mRNA levels by Northern blot analysis. As shown in Fig. 7D, CD3 mAbs and saturating concentrations of IL-2. OKT3/IL-2 treatment potently induced cyclin D3 mRNA expres- Of interest is that the effect of RA on T cell proliferation was sion, and in the presence of RA the mRNA levels were markedly observed in the absence of serum. In fact, the addition of FBS further increased. largely abrogated the RA effect. One explanation for this could be that FBS-contained retinol is metabolized by the T cells, producing The effect of RA on OKT3/IL-2-induced T cell proliferation is enough RA to stimulate T cell proliferation under our experimental mediated via RAR conditions. Indeed, we found that OKT-3/IL-2-induced T cell pro- Most of the biological effects of RA are believed to be exerted via liferation was very sensitive to RA, responding to RA concentra- binding to RAR, which, when heterodimerized with the retinoid X tions as low as 10 pmol. Moreover, the FBS that was used in the receptor (RXR) functions as a ligand-dependent transcription fac- present study contained ϳ0.6 ␮M retinol (data not shown), a con- tor capable of regulating several hundred genes (46). In some centration that is comparable to that found by others (33, 34). cases, however, RA may also act via RAR-independent mecha- Accordingly, when studying the effects of RA or retinol on T cells, nisms (47, 48). We assessed whether RAR was involved in RA- serum has frequently been omitted from the culture medium (52– mediated potentiation of OKT3/IL-2-induced T cell proliferation 54). An alternative explanation as to why the RA effect was di- by the use of two different RAR-selective agonists, TTNPB (49) minished in the presence of FBS could be that serum contains and Am580 (50), and the RAR-selective antagonist Ro 41-5253 factors (other than retinol) that replace and/or inhibit the effect of (51). TTNPB and Am580 almost completely mimicked the effect RA on IL-2-mediated signaling. Indeed, serum has been reported of RA on OKT3/IL-2-induced thymidine incorporation (Fig. 8A). to contain factors that can either stimulate (55) or inhibit (56) Furthermore, the effect of RA was completely abolished in the IL-2-induced T cell proliferation. Irrespectively of how serum af- presence of Ro 41-5253 (Fig. 8B). Essentially, the same results fects IL-2-induced signaling, the fact that we in the present study were obtained when cell proliferation was determined by the use of found RA to enhance IL-2-mediated proliferation in the absence of CFSE staining (data not shown). We therefore concluded that the serum suggests that RA stimulates IL-2-induced signaling directly The Journal of Immunology 2859 and not by modulating the effects of factors that are present in activation of CDK2 through cyclin D3-CDK4/6 complex binding serum. It should also be noted, however, that the effect of RA was and thereby sequestration of p21Cip1 and p27Kip1 (16). Down-reg- not exclusively observed in the absence of serum. Indeed, we have ulation of p27Kip1 expression has been reported to be the critical found that RA signiﬁcantly enhances OKT-3/IL-2-induced T cell IL-2-mediated event for CDK2 activation and thereby G1-toS- proliferation in the presence of 0.5 or 1% FBS (data not shown). phase transition in murine T cells (8, 17), and interestingly, we did We argue that the increased T cell proliferation observed with observe that RA modestly reduced p27Kip1 protein levels. How- RA in the present study is due to RA stimulating IL-2-, rather than ever, the effect of RA on p27Kip1 was observed at days 2 and 3

TCR/CD3-initiated signaling. First, because we used IL-2R-satu- only, and could thus well be a result, rather than the cause of G1- rating concentrations of IL-2, a possible contribution from RA- to S-phase transition. RA did not alter p21Cip1 protein levels, and mediated enhancement of IL-2 production was excluded. Second, the expression of cyclin D2 and E was also unaffected by RA. in the presence of the JAK inhibitor AG-490, which blocked IL- Therefore, it is tempting to speculate that RA allows CDK2 acti- 2-induced signaling (as demonstrated by inhibition of STAT3 ty- vation and pRB hyperphosphorylation predominantly by cyclin rosine phosphorylation), but not TCR/CD3-induced signaling (as D3-CDK4/6 complex-mediated sequestration of p21Cip1 and demonstrated by the intact induction of early IL-2R-␣ expression), p27Kip1. RA also increased the expression of cyclin A. However, RA-mediated enhancement of cyclin D3 expression and T cell pro- because cyclin A gene transcription has been shown to be a down- liferation was abolished. Third, the effect of RA on cyclin D3 stream regulatory target of both cyclin D (61) and pRB (62), the expression and DNA synthesis was also abolished in the presence induced expression of cyclin A could well be a consequence of of blocking anti-IL-2R␣ mAbs. Fourth, we showed that RA in- RA-mediated up-regulation of cyclin D3 and enhancement of pRB creased cyclin D3 expression and T cell proliferation in the pres- phosphorylation. Interestingly, previous studies indicate that ence of IL-2 alone. Finally, our proposal that RA stimulates IL- whereas the expression of cyclin D2 (9, 63, 64), E (8, 9, 17), and 2-mediated proliferation of peripheral blood T cells is supported A (8) is strongly induced by TCR/CD3-initiated signals and only by previous studies showing that RA can potentiate IL-2-induced modestly induced by IL-2-initiated signals, cyclin D3 expression DNA synthesis in resting and PHA-preactivated human thymo- seems to be primarily induced by IL-2 signaling (9, 63, 64). In line cytes (57) as well as in Con A-preactivated murine splenocytes and with this, we found that cyclin D3 expression was almost com- in the murine IL-2-dependent T cell line HT-2 (58). Interestingly, pletely abolished, both in the absence and presence of RA, when RA has also been reported to enhance IL-2-induced cytolytic ac- IL-2 signaling was repressed by AG-490 or blocking anti-IL-2R␣ tivity of lymphokine-activated killer cells (59). mAbs. Moreover, RA augmented cyclin D3 expression also in the We found that RA substantially up-regulated the cell surface presence of IL-2 alone, indicating that IL-2-induced signaling may expression of all three subunits of the high affinity IL-2R. In a be sufficient as well as required for the effect of RA on cyclin D3 previous study, it was proposed that RA stimulates the prolifera- expression. tion of murine T splenocytes through an early and transient po- We believe that the present study provides some important clues tentiation of mitogen-induced up-regulation of IL-2R␣ expression with respect to the molecular mechanism(s) involved in RA-me- (5), and another study suggested that RA enhances the prolifera- diated stimulation of cyclin D3 protein expression and T cell pro- tion of IL-2-maintained human thymocyte lymphoblasts by aug- liferation. First, because RAR agonists mimicked and a RAR an- menting the expression of both IL-2R␣ and ␤ (60). However, the tagonist abolished the effect of RA on T cell proliferation, we data obtained in the current study indicate that the effects of RA on believe that RA acts primarily through RAR-dependent transcrip- cyclin D3 expression and T cell proliferation are not mediated by tional regulation. Second, considering the strong induction of cy- an up-regulation of IL-2Rs. First, RA-mediated enhancement of clin D3 mRNA by RA, it seems likely that RA predominantly IL-2R expression was a relatively late event; RA did not increase increases cyclin D3 protein expression by enhancing its mRNA cell surface expression of the ␣- and ␥-chains before day 3, and levels. Because, to our knowledge, there does not exist any RAR- RA-induced enhancement of IL-2R␤ expression was first noted at responsive elements in the cyclin D3 promoter, we believe that the day 2. In comparison, RA-mediated up-regulation of cyclin D3 effect of RA on cyclin D3 mRNA expression is indirect. RAR/ protein levels could be observed already at day 1. Second, if the RXR-dependent, de novo synthesized proteins could act to mod- effect of RA on T cell proliferation was mediated by induction of ulate a single, upstream IL-2-induced signaling event or alterna- IL-2Rs, RA would be expected to enhance proximal IL-2 signal- tively cooperate with IL-2 via multiple mechanisms. The data ing, e.g., the phosphorylation of STAT proteins. However, the obtained with AG-490 indicate that the effect of RA is dependent phosphorylation statuses of both STAT5 and STAT3 were unaf- on JAK activity. However, because the levels of phospho-STAT5 fected by RA, even at day 3. This suggests that proximal IL-2 and phospho-STAT3 proteins were not increased by RA, we con- signaling was already saturated in the absence of RA. Finally, we cluded that RA mediates its effects downstream or independently found that RA-mediated induction of IL-2R␣, ␤, and ␥ was totally of proximal JAK-STAT signaling. Furthermore, we could exclude abolished when IL-2 signaling was blocked by AG-490. We there- the involvement of two other major IL-2-induced signaling path- fore believe that RA-induced up-regulation of IL-2Rs is a result, ways, namely the PI3K-PKB/Akt and the Raf-MAPK/ERK path- rather than the cause of RA-mediated stimulation of IL-2-induced ways. Thus, at present, we do not have any indications that RA signals. regulates a single, upstream IL-2-induced signaling event. Rather,

IL-2 initiates T cell proliferation by promoting G1- to S-phase it is possible that RA via RAR/RXR acts further downstream, e.g., transition (10). Thus, we examined potential effects of RA on es- at the level of gene transcription, to modulate the expression of sential components of the cell cycle machinery that regulate G1-to IL-2-responsive genes. Interestingly, we found that the expression S-phase transition in T cells. Consistent with the potent enhancing of c-myc, a candidate RAR/RXR target gene (46) and a known effect of RA on DNA synthesis noted at day 3, the phosphorylation target of IL-2-induced signaling (10), was up-regulated by RA at of pRB was strongly enhanced by RA at days 2 and 3. The earliest both the RNA and protein level within 5 h after stimulation with and most striking effect of RA was the induction of cyclin D3, OKT-3/IL-2. Because c-Myc has been shown to have the capacity which was observed already at day 1, and was very strong at days to induce the expression of cyclin D3 (42), RA-mediated up-reg- 2 and 3. Cyclin D3 may contribute to cell cycle progression in at ulation of c-Myc could provide a mechanistic explanation for the least two ways: 1) direct activation of CDK4/6, and 2) indirect observed induction of cyclin D3 by RA. We have put strong effort 2860 RA STIMULATES IL-2-MEDIATED T CELL PROLIFERATION into assessing this possibility by the use of siRNA against c-myc 19. Ishii, N., T. Takeshita, Y. Kimura, K. Tada, M. Kondo, M. Nakamura, and ␥ RNA. However, it has proven to be very difficult to obtain specific K. Sugamura. 1994. Expression of the IL-2 receptor chain on various popula- tions in human peripheral blood. Int. Immunol. 6: 1273–1277. knockdown of c-myc under serum-free conditions in primary T 20. Page, T. H., and M. J. Dallman. 1991. Molecular cloning of cDNAs for the rat cells, and thus we have not been able to verify our hypothesis. interleukin 2 receptor ␣ and ␤ chain genes: differentially regulated gene activity In conclusion, we have demonstrated that RA potently stimu- in response to mitogenic stimulation. Eur. J. Immunol. 21: 2133–2138. 21. Nakarai, T., M. J. Robertson, M. Streuli, Z. Wu, T. L. Ciardelli, K. A. Smith, and lates IL-2-induced proliferation of freshly isolated T lymphocytes, J. Ritz. 1994. Interleukin 2 receptor ␥ chain expression on resting and activated and have linked the effect of RA to an early induction of cyclin D3. lymphoid cells. J. Exp. Med. 180: 241–251. 22. Lin, J. X., and W. J. Leonard. 1997. Signaling from the IL-2 receptor to the The elucidation of the molecular mechanism(s) involved in RA- nucleus. Cytokine Growth Factor Rev. 8: 313–332. mediated potentiation of IL-2-induced T cell proliferation may be 23. Lin, J. X., and W. J. Leonard. 2000. The role of Stat5a and Stat5b in signaling by particularly important from a physiological point of view, consid- IL-2 family cytokines. Oncogene 19: 2566–2576. 24. Ellery, J. M., and P. J. Nicholls. 2002. Alternate signalling pathways from the ering the recently described role of Ag-presenting dendritic cells in interleukin-2 receptor. Cytokine Growth Factor Rev. 13: 27–40. providing T cells with both IL-2 (65, 66) and RA (67) during 25. Benczik, M., and S. L. Gaffen. 2004. The interleukin (IL)-2 family cytokines: dendritic cell-T cell interaction. IL-2 is a promising therapeutic survival and proliferation signaling pathways in T lymphocytes. Immunol. Invest. 33: 109–142. agent for the treatment of HIV/AIDS (68) and several types of 26. Moriggl, R., D. J. Topham, S. Teglund, V. Sexl, C. McKay, D. Wang, cancers (69). The results obtained in the present study imply that A. Hoffmeyer, J. van Deursen, M. Y. Sangster, K. D. Bunting, et al. 1999. Stat5 RA may be exploitable as an enhancer of IL-2-mediated is required for IL-2-induced cell cycle progression of peripheral T cells. Immunity 10: 249–259. immunotherapy. 27. Lord, J. D., B. C. McIntosh, P. D. Greenberg, and B. H. Nelson. 2000. The IL-2 receptor promotes lymphocyte proliferation and induction of the c-myc, bcl-2, and bcl-x genes through the trans-activation domain of Stat5. J. Immunol. 164: Acknowledgments 2533–2541. We thank Nasser Bastani for preparation and analyses of RA and for anal- 28. Austenaa, L. M., H. Carlsen, A. Ertesvag, G. Alexander, H. K. Blomhoff, and yses of the retinol content in FBS. R. Blomhoff. 2004. Vitamin A status significantly alters nuclear factor-␬B activity assessed by in vivo imaging. FASEB J. 18: 1255–1257. 29. Naderi, S., J. Y. Wang, T. T. Chen, K. B. Gutzkow, and H. K. Blomhoff. 2005. Disclosures cAMP-mediated inhibition of DNA replication and S phase progression: involve- Cip1 The authors have no financial conflict of interest. ment of Rb, p21 , and PCNA. Mol. Biol. Cell 16: 1527–1542. 30. Gutzkow, K. B., H. U. Lahne, S. Naderi, K. M. Torgersen, B. Skalhegg, M. Koketsu, Y. Uehara, and H. K. Blomhoff. 2003. Cyclic AMP inhibits trans- References lation of cyclin D3 in T lymphocytes at the level of elongation by inducing eEF2-phosphorylation. Cell. Signal. 15: 871–881. 1. Stephensen, C. B. 2001. Vitamin A, infection, and immune function. Annu. Rev. 31. Dalla-Favera, R., F. Wong-Staal, and R. C. Gallo. 1982. Onc gene amplification Nutr. 21: 167–192. in promyelocytic leukemia cell line HL-60 and primary leukemic cells of the 2. Semba, R. D. 1999. Vitamin A and immunity to viral, bacterial and protozoan same patient. Nature 299: 61–63. infections. Proc. Nutr. Soc. 58: 719–727. 32. Robb, R. J., A. Munck, and K. A. Smith. 1981. T cell growth factor receptors: 3. Hayes, C. E., F. E. Nashold, F. Enrique Gomez, and K. A. Hoag. 1999. Retinoids quantitation, specificity, and biological relevance. J. Exp. Med. 154: 1455–1474. and immunity. In Retinoids: The Biochemical and Molecular Basis of Vitamin A 33. Fuchs, E., and H. Green. 1981. Regulation of terminal differentiation of cultured and Retinoid Action, Vol. 139. H. Nau and W. S. Blaner, eds. Springer-Verlag, human keratinocytes by vitamin A. Cell 25: 617–625. Berlin, pp. 589–610. 34. Ishida, S., T. Kinoshita, N. Sugawara, T. Yamashita, and K. Koike. 2003. Serum 4. Valone, F. H., and D. G. Payan. 1985. Potentiation of mitogen-induced human inhibitors for human mast cell growth: possible role of retinol. Allergy 58: T-lymphocyte activation by retinoic acid. Cancer Res. 45: 4128–4131. 1044–1052. 5. Jiang, X. L., M. P. Everson, and E. W. Lamon. 1993. A mechanism of retinoid potentiation of murine T-cell responses: early up-regulation of interleukin-2 re- 35. Kirken, R. A., R. A. Erwin, D. Taub, W. J. Murphy, F. Behbod, L. Wang, ceptors. Int. J. Immunopharmacol. 15: 309–317. F. Pericle, and W. L. Farrar. 1999. Tyrphostin AG-490 inhibits cytokine-mediated JAK3/STAT5a/b signal transduction and cellular proliferation of antigen- 6. Ertesvag, A., N. Engedal, S. Naderi, and H. K. Blomhoff. 2002. Retinoic acid activated human T cells. J. Leukocyte Biol. 65: 891–899. stimulates the cell cycle machinery in normal T cells: involvement of retinoic acid receptor-mediated IL-2 secretion. J. Immunol. 169: 5555–5563. 36. Witthuhn, B. A., M. D. Williams, H. Kerawalla, and F. M. Uckun. 1999. Dif- 7. Engedal, N., A. Ertesvag, and H. K. Blomhoff. 2004. Survival of activated human ferential substrate recognition capabilities of Janus family protein tyrosine ki- T lymphocytes is promoted by retinoic acid via induction of IL-2. Int. Immunol. nases within the interleukin 2 receptor (IL2R) system: Jak3 as a potential mo- 16: 443–453. lecular target for treatment of leukemias with a hyperactive Jak-Stat signaling machinery. Leuk. Lymphoma 32: 289–297. 8. Firpo, E. J., A. Koff, M. J. Solomon, and J. M. Roberts. 1994. Inactivation of a Cdk2 inhibitor during interleukin 2-induced proliferation of human T lympho- 37. Ziegler, S. F., F. Ramsdell, and M. R. Alderson. 1994. The activation antigen cytes. Mol. Cell. Biol. 14: 4889–4901. CD69. Stem Cells 12: 456–465. 9. Lea, N. C., S. J. Orr, K. Stoeber, G. H. Williams, E. W. Lam, M. A. Ibrahim, 38. Mookerjee, B. K., and J. L. Pauly. 1990. Mitogenic effect of interleukin-2 on 3 unstimulated human T cells: an editorial review. J. Clin. Lab. Anal. 4: 138–149. G. J. Mufti, and N. S. Thomas. 2003. Commitment point during G0 G1 that controls entry into the cell cycle. Mol. Cell. Biol. 23: 2351–2361. 39. Wen, Z., Z. Zhong, and J. E. Darnell, Jr. 1995. Maximal activation of transcrip- 10. Nelson, B. H., and D. M. Willerford. 1998. Biology of the interleukin-2 receptor. tion by Stat1 and Stat3 requires both tyrosine and serine phosphorylation. Cell 82: Adv. Immunol. 70: 1–81. 241–250. 11. Modiano, J. F., J. Domenico, A. Szepesi, N. Terada, J. J. Lucas, and 40. Brennan, P., J. W. Babbage, B. M. Burgering, B. Groner, K. Reif, and E. W. Gelfand. 1995. Symmetry of the activation of cyclin-dependent kinases in D. A. Cantrell. 1997. Phosphatidylinositol 3-kinase couples the interleukin-2 re- mitogen and growth factor-stimulated T lymphocytes. Ann. NY Acad. Sci. 766: ceptor to the cell cycle regulator E2F. Immunity 7: 679–689. 134–148. 41. Sears, R. C., and J. R. Nevins. 2002. Signaling networks that link cell prolifer- 12. Malek, T. R., G. Ortega, J. P. Jakway, C. Chan, and E. M. Shevach. 1984. The ation and cell fate. J. Biol. Chem. 277: 11617–11620. murine IL 2 receptor. II. Monoclonal anti-IL-2 receptor antibodies as specific 42. Yu, Q., M. A. Ciemerych, and P. Sicinski. 2005. Ras and Myc can drive onco- inhibitors of T cell function in vitro. J. Immunol. 133: 1976–1982. genic cell proliferation through individual D-cyclins. Oncogene 24: 7114–7119. 13. Stingl, L. A., A. Sinska, U. Landesmann, and J. S. Smolen. 1987. Induction of 43. Wang, Z., Y. Zhang, J. Lu, S. Sun, and K. Ravid. 1999. Mpl ligand enhances the interleukin 2 receptiveness and proliferation in resting peripheral T cells by transcription of the cyclin D3 gene: a potential role for Sp1 transcription factor. monoclonal anti-CD3 (T3) antibodies does not require the presence of macro- Blood 93: 4208–4221. phages. Clin. Exp. Immunol. 68: 146–155. 44. Garcia-Gras, E. A., P. Chi, and E. A. Thompson. 2000. Glucocorticoid-mediated 14. Planas-Silva, M. D., and R. A. Weinberg. 1997. The restriction point and control destabilization of cyclin D3 mRNA involves RNA-protein interactions in the of cell proliferation. Curr. Opin. Cell Biol. 9: 768–772. 3Ј-untranslated region of the mRNA. J. Biol. Chem. 275: 22001–22008. 15. Harbour, J. W., and D. C. Dean. 2000. The Rb/E2F pathway: expanding roles and 45. Naderi, S., K. B. Gutzkow, H. U. Lahne, S. Lefdal, W. J. Ryves, A. J. Harwood, emerging paradigms. Genes Dev. 14: 2393–2409. and H. K. Blomhoff. 2004. cAMP-induced degradation of cyclin D3 through 16. Sherr, C. J., and J. M. Roberts. 1999. CDK inhibitors: positive and negative association with GSK-3␤. J. Cell Sci. 117: 3769–3783. regulators of G1-phase progression. Genes Dev. 13: 1501–1512. 46. Balmer, J. E., and R. Blomhoff. 2002. Gene expression regulation by retinoic 17. Nourse, J., E. Firpo, W. M. Flanagan, S. Coats, K. Polyak, M. H. Lee, acid. J. Lipid Res. 43: 1773–1808. J. Massague, G. R. Crabtree, and J. M. Roberts. 1994. Interleukin-2-mediated 47. Konta, T., Q. Xu, A. Furusu, K. Nakayama, and M. Kitamura. 2001. Selective elimination of the p27Kip1 cyclin-dependent kinase inhibitor prevented by rapa- roles of retinoic acid receptor and retinoid X receptor in the suppression of ap- mycin. Nature 372: 570–573. optosis by all-trans-retinoic acid. J. Biol. Chem. 276: 12697–12701. 18. Bich-Thuy, L. T., M. Dukovich, N. J. Peffer, A. S. Fauci, J. H. Kehrl, and 48. Takahashi, N., and T. R. Breitman. 1994. Retinoylation of proteins in mammalian W. C. Greene. 1987. Direct activation of human resting T cells by IL-2: the role cells. In Vitamin A in Health and Disease. R. Blomhoff, ed. Marcel Decker, New of an IL-2 receptor distinct from the Tac protein. J. Immunol. 139: 1550–1556. York, pp. 257–273. The Journal of Immunology 2861

49. Boehm, M. F., M. R. McClurg, C. Pathirana, D. Mangelsdorf, S. K. White, 60. Sidell, N., B. Chang, and L. Bhatti. 1993. Up-regulation by retinoic acid of J. Hebert, D. Winn, M. E. Goldman, and R. A. Heyman. 1994. Synthesis of high interleukin-2-receptor mRNA in human T lymphocytes. Cell. Immunol. 146: specific activity [3H]-9-cis-retinoic acid and its application for identifying retin- 28–37. oids with unusual binding properties. J. Med. Chem. 37: 408–414. 61. Schulze, A., K. Zerfass-Thome, J. Berges, S. Middendorp, P. Jansen-Durr, and 50. Kagechika, H. 2002. Novel synthetic retinoids and separation of the pleiotropic B. Henglein. 1996. Anchorage-dependent transcription of the cyclin A gene. Mol. retinoidal activities. Curr. Med. Chem. 9: 591–608. Cell. Biol. 16: 4632–4638. 51. Apfel, C., F. Bauer, M. Crettaz, L. Forni, M. Kamber, F. Kaufmann, P. LeMotte, 62. Knudsen, K. E., A. F. Fribourg, M. W. Strobeck, J. M. Blanchard, and W. Pirson, and M. Klaus. 1992. A retinoic acid receptor ␣ antagonist selectively E. S. Knudsen. 1999. Cyclin A is a functional target of retinoblastoma tumor counteracts retinoic acid effects. Proc. Natl. Acad. Sci. USA 89: 7129–7133. suppressor protein-mediated cell cycle arrest. J. Biol. Chem. 274: 27632–27641. 52. Allende, L. M., A. Corell, A. Madrono, R. Gongora, C. Rodriguez-Gallego, 63. Lucas, J. J., A. Szepesi, J. F. Modiano, J. Domenico, and E. W. Gelfand. 1995. A. Lopez-Goyanes, M. Rosal, and A. Arnaiz-Villena. 1997. Retinol (vitamin A) Regulation of synthesis and activity of the PLSTIRE protein (cyclin-dependent is a cofactor in CD3-induced human T-lymphocyte activation. Immunology 90: kinase 6 (cdk6)), a major cyclin D-associated cdk4 homologue in normal human 388–396. T lymphocytes. J. Immunol. 154: 6275–6284. 53. Friedman, A., O. Halevy, M. Schrift, Y. Arazi, and D. Sklan. 1993. Retinoic acid 64. Malek, T. R., A. Yu, P. Scibelli, M. G. Lichtenheld, and E. K. Codias. 2001. promotes proliferation and induces expression of retinoic acid receptor-␣ gene in Broad programming by IL-2 receptor signaling for extended growth to multiple murine T lymphocytes. Cell. Immunol. 152: 240–248. cytokines and functional maturation of antigen-activated T cells. J. Immunol. 54. Garbe, A., J. Buck, and U. Hammerling. 1992. Retinoids are important cofactors 166: 1675–1683. in T cell activation. J. Exp. Med. 176: 109–117. 65. Creusot, R. J., and N. A. Mitchison. 2004. How DCs control cross-regulation 55. Clevenger, C. V., D. H. Russell, P. M. Appasamy, and M. B. Prystowsky. 1990. between lymphocytes. Trends Immunol. 25: 126–131. Regulation of interleukin 2-driven T-lymphocyte proliferation by prolactin. Proc. Natl. Acad. Sci. USA 87: 6460–6464. 66. Granucci, F., I. Zanoni, S. Feau, and P. Ricciardi-Castagnoli. 2003. Dendritic cell 56. Kucharz, E. J., and J. S. Goodwin. 1988. Serum inhibitors of interleukin-2. Life regulation of immune responses: a new role for interleukin 2 at the intersection Sci. 42: 1485–1491. of innate and adaptive immunity. EMBO J. 22: 2546–2551. 57. Sidell, N., and F. Ramsdell. 1988. Retinoic acid up-regulates interleukin-2 re- 67. Iwata, M., A. Hirakiyama, Y. Eshima, H. Kagechika, C. Kato, and S. Y. Song. ceptors on activated human thymocytes. Cell. Immunol. 115: 299–309. 2004. Retinoic acid imprints gut-homing specificity on T cells. Immunity 21: 58. Jiang, X. L., D. L. Dillehay, M. P. Everson, A. B. Tilden, and E. W. Lamon. 1992. 527–538. Potentiation of IL-2-induced T-cell proliferation by retinoids. Int. J. Immunop- 68. Smith, K. A. 2001. Low-dose daily interleukin-2 immunotherapy: accelerating harmacol. 14: 195–204. immune restoration and expanding HIV-specific T-cell immunity without toxic- 59. Lin, T. H., and T. M. Chu. 1990. Enhancement of murine lymphokine-activated ity. AIDS 2(Suppl. 15): S28–S35. killer cell activity by retinoic acid. Cancer Res. 50: 3013–3018. 69. Atkins, M. B. 2002. Interleukin-2: clinical applications. Semin. Oncol. 29: 12–17.