Pathway Development Via Retinoid X Receptor Vitamin a Enhances In

Home , Retinoic acid, Retinoic acid receptor, Retinoid X receptor

Vitamin A Enhances in Vitro Th2 Development Via Retinoid X Receptor Pathway

This information is current as Charles B. Stephensen, Reuven Rasooly, Xiaowen Jiang, of September 24, 2021. Michael A. Ceddia, Casey T. Weaver, Roshantha A. S. Chandraratna and R. Patterson Bucy J Immunol 2002; 168:4495-4503; ; doi: 10.4049/jimmunol.168.9.4495

http://www.jimmunol.org/content/168/9/4495 Downloaded from

References This article cites 40 articles, 24 of which you can access for free at: http://www.jimmunol.org/content/168/9/4495.full#ref-list-1 http://www.jimmunol.org/

Why The JI? Submit online.

• Rapid Reviews! 30 days* from submission to initial decision

• No Triage! Every submission reviewed by practicing scientists

• Fast Publication! 4 weeks from acceptance to publication by guest on September 24, 2021

*average

Subscription Information about subscribing to The Journal of Immunology is online at: http://jimmunol.org/subscription Permissions Submit copyright permission requests at: http://www.aai.org/About/Publications/JI/copyright.html Email Alerts Receive free email-alerts when new articles cite this article. Sign up at: http://jimmunol.org/alerts

The Journal of Immunology is published twice each month by The American Association of Immunologists, Inc., 1451 Rockville Pike, Suite 650, Rockville, MD 20852 Copyright © 2002 by The American Association of Immunologists All rights reserved. Print ISSN: 0022-1767 Online ISSN: 1550-6606. Vitamin A Enhances in Vitro Th2 Development Via Retinoid X Receptor Pathway1

Charles B. Stephensen,2* Reuven Rasooly,* Xiaowen Jiang,* Michael A. Ceddia,3* Casey T. Weaver,† Roshantha A. S. Chandraratna,‡ and R. Patterson Bucy†

Vitamin A deﬁciency diminishes Th2-mediated Ab responses, and high-level dietary vitamin A or treatment with the vitamin A metabolite retinoic acid (RA) enhances such responses. To identify a potential mechanism(s) underlying this in vivo activity of vitamin A, we examined the effects of all-trans and 9-cis RA on development of Th1 and Th2 cell populations using in vitro stimulation of Ag-naive Th0 cells from the DO11.10 TCR-transgenic mouse. Treatment with 9-cis, but not with all-trans RA, at primary stimulation strongly enhanced Th2 development. IL-4-neutralizing Ab blocked this activity, but IL-12- and IFN-␥- neutralizing Ab did not. Because 9-cis RA regulates gene transcription via either RA receptors or retinoid X receptors (RXRs), we tested the Th2-enhancing activities of the RXR- and RA receptor-selective agonists AGN194204 and 4-((E)-2-(5,6,7,8-tetra- Downloaded from hydro-5,5,8,8-tetramethyl-2-naphthalenyl)-1-propenyl)benzoic acid (TTNPB). AGN194204 strongly enhanced Th2 development, (whereas TTNPB did not. This RXR agonist also enhanced Th2 development when puriﬁed, naive Th0 cells (L-selectinhigh/CD4؉ were stimulated with CD3 and CD28 Abs in the absence of APCs. During primary antigenic stimulation of naive Th0 cells from DO11.10 mice, AGN194204 increased IL-4 and IL-5 production, decreased IFN-␥ production, increased mRNA in responding T cells for genes involved in Th2 development (IL-4, GATA-3, and c-maf), and decreased mRNA for genes involved in Th1 development (IFN-␥, T-bet, and IL-12R). These data show that stimulation of the RXR pathway enhances Th2 development, perhaps http://www.jimmunol.org/ by affecting the relative expression of pertinent transcription factors, cytokines, and cytokine receptors. The Journal of Immu- nology, 2002, 168: 4495–4503.

itamin A deficiency increases mortality from common increases IL-5, IL-10, and IL-4 production (11Ð13). Thus, vitamin childhood infections (1, 2). The specific mechanisms un- A deficiency biases the immune response in a Th1 direction, V derlying this increased risk of death have not been whereas high-level dietary vitamins may bias the response in a Th2 clearly defined but presumably involve impairment of specific and direction. RA appears to be the metabolite of vitamin A that is

nonspecific host defense mechanisms (3). Although data from hu- most potent in restoring impaired Ab responses (14). Although it by guest on September 24, 2021 mans are limited, rodent studies indicate that vitamin A deficiency is known that exogenous RA can down-regulate IFN-␥ transcrip- alters the Ab response to T cell-dependent Ags (4, 5). Vitamin A tion (15), little else is known about how RA modulates Th1/Th2 deficiency decreases the IgA, IgG1, and IgE responses but in- balance. creases the IgG2a response to viral infection (6Ð8). The underly- Vitamin A and other fat-soluble nutrients are precursors for ing pattern of cytokine production by APCs and T cells is consis- compounds that act as ligands for nuclear receptors that regulate tent with these effects on Ab production: vitamin A deficiency gene transcription in response to changes in nutritional status (16). increases constitutive IL-12 production by macrophages (9), and These receptors regulate key metabolic processes, such as energy during secondary in vitro stimulation of lymphocytes with Ag, metabolism, but are also found in cells of the immune system vitamin A deficiency increases IFN-␥ production, but decreases where they appear to modulate immune function based on envi- IL-4 and IL-5 production (10). Conversely, supplemental treat- ronmental (in this case, nutritional) signals. The vitamin A deriv- ment with vitamin A or retinoic acid (RA)4 decreases IFN-␥ and atives all-trans and 9-cis RA regulate gene transcription by binding to the RA receptors (RARs) ␣, ␤,or␥ or to the retinoid X receptors (RXRs) ␣, ␤,or␥ (17, 18). Both all-trans and 9-cis RA *U.S. Department of Agriculture Western Human Nutrition Research Center and Nutrition Department, University of California, Davis, CA 95616; †Department of bind to the RARs, whereas 9-cis RA also acts via the RXRs. Do- Pathology, School of Medicine, University of Alabama, Birmingham, AL 35243; and cosahexaenoic acid also binds to RXRs and may be a physiolog- ‡Departments of Chemistry and Biology, Retinoid Research, Allergan, Irvine, CA 92623 ically important ligand in tissues in which concentrations are suf- ficient (Ն10 ␮mol/L) to trigger transcriptional regulation (19). Received for publication December 18, 2001. Accepted for publication March 1, 2002. RAR and RXR belong to a family of nuclear receptors that also The costs of publication of this article were defrayed in part by the payment of page includes the vitamin D receptor (VDR), thyroid hormone receptor, charges. This article must therefore be hereby marked advertisement in accordance and the peroxisome proliferation/activation receptor (PPAR), with 18 U.S.C. Section 1734 solely to indicate this fact. which binds specific fatty acids. In brief, the structure of these 1 This work was supported by the U.S. Department of Agriculture, National Research receptors includes a 5Ј “A/B” domain, a DNA-binding “C” do- Initiative Grant No. 97-35200-4229, and program funds from the Western Human Nutrition Research Center. main, a hinge region (“D” domain), the ligand-binding “E” do- Ј 2 Address correspondence and reprint requests to Dr. Charles B. Stephensen, Nutri- main, and a 3 “F” domain of uncertain function. Receptors bind to tion Department, University of California, 3243 Meyer Hall, One Shields Avenue, specific DNA response elements in the regulatory regions of genes Davis, CA 95616. E-mail address: [email protected] 3 Current address: IAMS, 6571 State Route 503 North, Lewisburg, OH 45338. 4 Abbreviations used in this paper: RA, retinoic acid; RAR, RA receptor; RXR, ret- tion receptor; TTNPB, 4-((E)-2-(5,6, 7,8-tetrahydro-5,5,8,8-tetramethyl-2-naphthale- inoid X receptor; VDR, vitamin D receptor; PPAR, peroxisome proliferation/activa- nyl)-1-propenyl)benzoic acid.

for which transcription is regulated by these receptors. Transcrip- gies; containing penicillin, streptomycin, and fungizone), 5 ml nonessential tional activity is regulated via specific receptor sequences that in- amino acid mix (Life Technologies), 5 ml sodium pyruvate (Life Technol- teract with coactivator proteins to affect transcription by RNA ogies), and 0.25 ml of 100 mM 2-ME (Life Technologies). Cells were centrifuged at 800 ϫ g at 4¡C for 10 min. Contaminating RBCs were then polymerase II. Sequences in the ligand binding domain are also lysed by adding red cell lysis buffer. Cells were again centrifuged and responsible for formation of heterodimers among these receptors, resuspended in Russ-10 medium, and viable cells were counted using with RXR being one of the heterodimer partners. Unlike other trypan blue and a hemocytometer. Serum-free medium, made as described receptors, RXR can also form homodimers that can positively reg- above with 1% HL-1 serum replacement (BioWhittaker, Walkersville, MD) instead of 10% serum, was used in some experiments. In some ex- ulate transcription. In these experiments, we used all-trans RA periments, CD4ϩ cells were positively selected using Ab-coated magnetic (which binds to RAR), 9-cis RA (which binds to both RAR and beads (CD4 Dynabead; Dynal Biotech, Oslo, Norway) and a magnetic RXR), the RAR-selective retinoid 4-((E)-2-(5,6,7,8-tetrahydro- particle concentrator (MPC-2; Dynal Biotech). CD4ϩ cells were then de- tached from the magnetic beads using an Ab reagent (mouse CD4 Detach- 5,5,8,8-tetramethyl-2-naphthalenyl)-1-propenyl)benzoic acid (TT- ϩ NPB), and the RXR-selective retinoid AGN194204 (20, 21) to a-bead; Dynal Biotech). Positively selected CD4 cells, unselected lymph node cells, and a remixture of the selected and unselected cells were com- demonstrate that stimulation of the RXR pathway enhances Th2 pared for their ability to produce IL-4 and IFN-␥ after antigenic stimulation development. Use of these latter compounds is particularly impor- (as described in Stimulation protocol) under Th2 conditions (1000 U/ml, 10 tant because they are stable, whereas RA isomers can interconvert ng/ml IL-4; BD PharMingen), nonselective conditions (no additional cytokine or Ab), and Th1 conditions (5 ␮g/ml anti-IL-4 plus 10 U/ml, 0.2 under physiologic conditions (18). ϩ ␣␤ ng/ml IL-12), and no qualitative differences were seen. Purity of CD4 The DO11.10 -TCR transgenic mouse (22) is a useful model cells was assessed by flow cytometry and was Ͼ97% in our hands. Spleno- for examining early events in the development of Th1 and Th2 cytes were collected from BALB/c mice in essentially the same fashion and memory cells from Ag-naive Th0 cells. The DO11.10 TCR rec- were then irradiated with 2500 rad before use as APCs. Downloaded from ognizes amino acids 323Ð339 of the protein OVA (OVA323Ð339), Stimulation protocol and the TCR can be identified using the clonotypic mAb KJ1-26 ϩ (23). The OVA synthetic peptide stimulates proliferation For primary antigenic stimulation of naive DO11.10 CD4 T cells, splenic 323Ð339 APCs (5 ϫ 106/ml) in Russ-10 medium containing 6 ␮g/ml OVA and cytokine production when it is presented in the MHC class II 323Ð339 d (purchased from the Peptide Synthesis Core Facility of the Comprehensive context I-A . We have used this system to examine the impact of Cancer Center, University of Alabama, Birmingham, AL) were placed in

retinoid treatment during primary in vitro antigenic stimulation on wells of 24-well (1.0 ml) or 48-well (0.5 ml) plates (Nunc, Rochester, NY). http://www.jimmunol.org/ the subsequent development of a Th1 or Th2 pattern of cytokine Ab, cytokine, or retinoid treatments were added, followed immediately by ϫ 5 ϩ ϫ 5 production. lymph node cells (4 10 /ml) or CD4 cells (2 10 /ml) (1.0 ml for 24-well plate, 0.5 ml for 48-well plate). Cultures were placed at 37¡Cin5%

CO2. Retinoids were diluted in DMSO. The final concentration of DMSO Materials and Methods in the cell culture medium was 0.1% (v/v), which did not affect cytokine Mice production by these cultures (data not shown). Three days after primary stimulation, cultures were expanded into six-well plates by adding 2 or 3 DO11.10 mice were bred in our facility. BALB/c mice were purchased vol of Russ-10 medium. For secondary stimulation on day 7, cells were from Charles River Breeding Laboratories (Wilmington, MA). Peripheral collected by centrifugation at 800 ϫ g at 4¡C for 10 min, washed twice in blood from DO11.10 mice was used to screen lymphocytes by two-color Russ 10, and counted using a System 9000 Cell Counter (Baker Instru- flow cytometric analysis using anti-CD4 Ab and the TCR clonotype-spe- ments, Fort Collins, CO). Concentration was then adjusted to 2 ϫ 105/ml by guest on September 24, 2021 cific mAb KJ1-26 (23) to identify transgenic (double-positive) animals and cells were restimulated as just described. For tertiary stimulation, on when heterozygous breeders were used and when mice were back-crossed day 14, cells restimulated on day 7 were collected and restimulated in an onto the BALB/c background. identical fashion. Supernatants were collected 24 or 48 h after secondary or Retinoids tertiary stimulation by centrifuging the plates at 800 ϫ g for 10 min at 4¡C and aspirating supernatants with a micropipette. All-trans RA, 9-cis RA, and the RAR agonist TTNPB were purchased In experiments to examine the effect of retinoid treatment on mature Th1 from Sigma-Aldrich (St. Louis, MO). The RXR agonist AGN194204 was and Th2 populations, DO11.10 lymph node cells were isolated and stim-

provided by Allergan (Irvine, CA). Retinoids were diluted aseptically in ulated with OVA323Ð339 and splenic APCs (primary stimulation) under cell culture grade DMSO (Sigma-Aldrich). The concentration of 9-cis and Th1-differentiating conditions (IL-12 plus anti-IL-4), Th2-differentiating all-trans RA were conﬁrmed by HPLC analysis as described (24). Stock conditions (IL-4), or without exogenous treatment. One week after primary solutions at 1 or 10 mM were frozen at Ϫ70¡C and were discarded after stimulation, cells were washed and equal numbers of cells were restimu- three freeze-thaw cycles. The retinol concentration in FBS was measured lated (secondary stimulation) without any exogenous treatment. One week by the same method. after secondary stimulation, cells were again washed and equal numbers

were restimulated with OVA323Ð339 (tertiary stimulation) in the presence of Abs and cytokines 10Ϫ7 M all-trans or 9-cis RA. Abs used in cell culture were purchased from BD PharMingen (San Diego, In some experiments, naive Th0 cells were stimulated with plate-bound ␮ CA) and included neutralizing rat mAbs for murine IL-4 (clone BVD40- anti-CD3 (1 g/ml, clone 145-2C11; BD PharMingen) and soluble anti- ␮ 1D11, IgG2b isotype), IL-12 (p40/p70, clone C17.8, IgG2a isotype), and CD28 (5 g/ml, clone 37.51; BD PharMingen) in the presence of IL-4 (10 IFN-␥ (clone R4-6A2). Isotype control Abs for IgG1 (clone 11010D), ng/ml). Similar results were found with anti-CD3 concentrations of 1 and IgG2a (clone 11020D), and IgG2b (clone R35-38) were also used. Recom- 10 ng/ml and anti-CD28 concentrations of 5 and 10 ng/ml. Cells were binant purified IL-4 and IL-12 were also purchased from BD PharMingen. expanded after 3 or 4 days. After 6 or 7 days, cells were washed twice, Abs for FACS analysis included R-PE-labeled anti-CD4 (clone L3T4; BD counted, and treated with PMA (5 ng/ml final concentration; Sigma-Al- PharMingen), FITC-labeled KJ1-26 (Caltag Laboratories, Burlingame, drich) and ionomycin (500 ng/ml final concentration; Sigma-Aldrich) to CA), and biotin-labeled CD62 ligand (mouse IgG2a clone MEL-14, used stimulate cytokine production. Supernatants were collected after overnight with the biotin-labeled IgG2a isotype control; BD PharMingen). Cy- incubation for measurement of cytokines by ELISA. Chrome-labeled streptavidin (BD PharMingen) was used as a second-step Cytokine ELISAs reagent. IL-4 and IFN-␥ were measured in cell culture supernatants by capture Cells ELISA using Abs, purified standards, and protocols suggested by the man- ␥ Lymph node cells from male DO11.10 mice were prepared aseptically ufacturer (BD PharMingen). The limit of detection of the IL-4 and IFN- from the popliteal, femoral, inguinal, dorsal, axillary, brachial, subman- ELISAs were 0.03 and 0.06 ng/ml, respectively. dibular, and mesenteric lymph nodes, which were disrupted between FACS frosted glass slides in Russ-10 cell culture medium (made by combining 450 ml of RPMI 1640 medium without glutamine (Life Technologies, At various times after stimulation, KJ1-26ϩ/CD4ϩ cells were isolated us- Grand Island, NY), 50 ml FBS (Life Technologies), 5 ml 200 mM glu- ing a MoFlo (Cytomations, Fort Collins, CO) high-speed cell sorter. Cells tamine (Life Technologies), 5 ml antibiotic-antimycotic (Life Technolo- in the lymphocyte gate by forward and side scatter were also gated for The Journal of Immunology 4497 viability using propidium iodide staining. Target cells were from 4 to 10% To confirm that each primer pair correctly amplified the sequence of of total cells, depending on the time point. We routinely obtained purities interest, initial PCR products from T cell cDNA were run on an agarose gel from 96 to 99%. (3% SeaKem 3:1), stained with 0.5 ␮g/ml ethidium bromide, and viewed by UV transillumination to confirm that a single product of the predicted mRNA analysis size was produced. PCR products were cloned into the PCR-II-TOPO plasmid vector (Invitrogen, San Diego, CA) and used to transform Escherichia Total RNA was prepared using the Trizol reagent as described by the ϩ coli competent cells (Life Technologies). White colonies were picked and manufacturer (Life Technologies). RNA was prepared from CD4 cells grown, and plasmid DNA was isolated and analyzed for the presence of the (prepared using Dynabeads, as described in Stimulation protocol)0,4,24, receptor sequence by digestion with EcoRI restriction enzyme that flanked and 48 h after primary antigenic stimulation in two experiments and from the insert. Inserts were sequenced completely in one direction using M13 CD4ϩ ϩ /KJ1-26 cells (purified by FACS, as described above) 0, 24, 48, and forward and reverse primers. 72 h after stimulation in two additional experiments. Analysis of gene To confirm specificity of the reaction product during each run, the melt- expression revealed essentially identical patterns using both cell isolation ing profile of each sample was analyzed using the LightCycler. The melting methods, and results from the four experiments were analyzed together profile was determined by holding the reaction at 55¡C for 10 s and then (Table I). heating slowly to 94¡C with a linear rate of 0.2¡C/s while the fluorescence After RNA isolation, first-strand cDNA was synthesized from 0.5 or 1.0 emitted was measured. Melting curve analysis demonstrated that each of ␮g of total RNA using Moloney murine leukemia virus reverse transcrip- ␮ the primer pairs described amplified a single product. Rare samples that tase (Superscript II; Life Technologies) in a 25- l reaction with oligo-dT demonstrated a significant second peak in the melting profile were not used (Life Technologies) as a primer. RNA was then degraded by treatment with in analysis. RNaseH (Life Technologies). Quantitative PCR analysis was performed using 1 ␮l of cDNA with 0.25 ␮M primers in a 20-␮l reaction using a LightCycler rapid thermal cycler system (Roche Diagnostics, Indianapolis, Statistical analysis

IN) and DNA Master Mix reagents (Roche Diagnostics) with SYBR Green Downloaded from I dye (Molecular Probes, Eugene, OR). The amplification program began Statistical analysis was performed with the SigmaStat program (Jandel Sci- with a 30-s denaturation step at 94¡C followed by 40 cycles with a 94¡C entific, San Rafael, CA). A p value of 0.05 was used to determine statistical denaturation for 1 s, annealing for5sat56¡C, and extension at 72¡C for significance unless otherwise indicated. All variables were compared 17 s. Upstream and downstream primers (respectively) for genes of interest among the treatment groups using either one-way or two-way ANOVA. were as follows: IL-4,5Ј-CATCGGCATTTTGAACGAGGTCA-3Ј,5Ј-C Two-way ANOVA was used when data from experiments done at different TTATCGATGAATCCAGGCATCG-3Ј; GATA-3,5Ј-TACTTGCGTTTT times were analyzed together (i.e., data were analyzed by “experiment” and TCGCAGGA-3Ј,5Ј-GCGTCCTTCATGCCTTTCTT-3Ј; c-maf,5Ј-GAG by “treatment group”). Pair-wise multiple comparisons among the treat- CAGTTGGTGACCATGTCG-3Ј;5Ј-TTGCTCACCAGCTTCTCGTATTT ment groups were routinely made (in conjunction with the ANOVA pro- http://www.jimmunol.org/ C-3Ј; STAT-6,5Ј-TCTGCTGCAGTGCCAGTCC-3Ј,5Ј-GGCTTTGTTTC cedure) by the Student-Newman-Keuls method or Bonferroni’s t test. TGGGCTTCTC-3Ј; IL-4R,5Ј-CGTGCCCTTATTTACTTTCGGACT-3Ј, When data were missing from one cell of a two-way ANOVA, the general Ј Ј ␥ Ј linear model was used to perform the ANOVA. This occurred with the 24- 5 -GGTTGGCTTCTGGTGGTATTCC-3 ; IFN- ,5-CATTGAAAGCCT ␥ AGAAAGTCTG-3Ј,5Ј-CTCATGAATGCATCCTTTTTCG-3Ј; T-bet,5Ј-C and 48-h IL-4 and IFN- data in Table I, where one experiment lacked the TGCCCGAACTACAGTCACGA-3Ј,5Ј-ACCCTCTCTTGTCTCCTCGCA IL-4 treatment, and with T-bet at 24 h in the IL-4 treatment, where RNA 3Ј; Fyn,5Ј-CTTTGGATTGGCTCGGTTGAT-3Ј,5Ј-GGAAGCCCTGC was not available for this assay. Preplanned, two-group comparisons were AAGTACTCGA-3Ј; and IL-12,5Ј-CACAACCTGAGCTCTGCGAAA-3Ј, also made using the Student t test. 5Ј-CATGCCATCAGGAGATTATCCG-3Ј. ␤-Actin primers were purchased from Promega (Madison, WI). Results

The cDNA copy number for each gene of interest was determined using by guest on September 24, 2021 a four-point standard curve (of plasmid DNA). Standard curves were run DO11.10 T cells have naive phenotype before antigenic with each set of samples. Correlation coefficients (r2) for standard curves stimulation were typically Ն0.98. The precision of target gene-␤-actin ratios from the ϩ ϩ same cDNA sample (within run) were quite good. For example, the aver- Before in vitro antigenic stimulation (day 0), CD4 /KJ1-26 cells age coefficient of variation for the 12 IL-4-␤-actin ratios from Table I was from DO11.10 mice in our colony had high-level L-selectin ex- 9.2 Ϯ 7.5% (range, 0Ð26%). Furthermore, the day-to-day reproducibility pression, consistent with a naive phenotype (data not shown). (precision) was also good. For example, the correlation coefficient for IL- Seven days after tertiary stimulation, L-selectin expression was 4-␤-actin ratios from nine samples (measured in duplicate) from a single experiment was r2 ϭ 0.87 (r ϭ 0.93) (quantitative PCR was performed on low—similar to that of the isotype control Ab (data not shown)— different days using two different batches of cDNA, prepared on different which is consistent with a memory cell phenotype, as has been days). shown previously for these mice (25).

Table I. mRNA expression in puriﬁed Th0 cells from DO11.10 mice following primary stimulation in vitro with OVA323Ð339 peptide and treatment with RXR agonista

24 h 48 h 72 h

Gene AGN IL-4 DMSO SE AGN IL-4 DMSO SE AGN IL-4 DMSO SE

Th2 genes IL-4 0.037 0.046 0.046 0.003 0.080b 0.16c 0.056d 0.004 0.15b 0.22c 0.05d 0.003 GATA-3 0.19b 0.34c 0.15d 0.009 0.30b 0.38c 0.22d 0.013 0.54b 0.53b 0.34c 0.017 c-maf 0.033b 0.035b 0.016c 0.003 0.018 0.020 0.011 0.002 0.018b 0.045c 0.038c 0.003 STAT-6 0.38 0.48 0.29 0.035 0.30 0.30 0.29 0.028 0.19 0.23 0.21 0.015 IL-4R 0.64 1.11 0.54 0.155 0.73 0.69 0.52 0.095 0.23b 0.66c 0.40b 0.043 Th1 genes IFN-␥ 0.042 0.048 0.045 0.005 0.019b 0.033c 0.044c 0.004 0.0073b 0.026c 0.019d 0.0002 T-bet 0.081 0.099 0.068 0.007 0.084b 0.18c 0.18c 0.003 0.063b 0.17c 0.17c 0.002 IL-12R ND in most cases 0.015b 0.040c 0.063d 0.004 0.049b 0.20c 0.31c 0.028 Fyn 0.36b 0.26c 0.31d 0.009 0.23b 0.21b 0.26c 0.008 0.15 0.15 0.21 0.009

a AGN, 1 ϫ 10Ϫ7 M AGN194204; IL-4, 10 ng/ml; or vehicle (DMSO, 0.1% DMSO). SE, pooled SE. The copy number for each gene is expressed relative to ␤-actin and was measured in duplicate or triplicate by real-time PCR in two (72 h) to four (24 and 48 h) independent experiments. b,c,d When a significant difference among treatment means (AGN, IL-4, DMSO) at a given time was identified by two-way ANOVA (using a p value of 0.05/9 ϭ 0.0056 to correct for multiple comparisons; n ϭ 9 genes), all means were compared with Bonferroni’s t test and differences ( p Ͻ 0.05) were identified by different superscripts. Bold type indicates means that differ from the vehicle control (DMSO). 4498 RXR AGONISTS ENHANCE IN VITRO Th2 DEVELOPMENT

9-cis, but not all-trans, RA promotes Th2 development Because conditions at primary stimulation of Ag-naive Th0 cells influence subsequent Th1/Th2 development, we treated primary cultures with all-trans and 9-cis RA and monitored IL-4 and IFN-␥ production at secondary and tertiary stimulation. As shown in Fig. 1, 9-cis, but not all-trans, RA increased IL-4 production at both secondary and tertiary stimulation ( p Ͻ 0.01), with 10Ϫ7 M9-cis RA having the greatest effect, stimulating a 10-fold increase over the DMSO control culture. In three additional experiments (one using serum-free medium), 10Ϫ7 9-cis RA significantly increased IL-4 production at secondary or tertiary stimulation in all cases, whereas 10Ϫ7 all-trans RA increased IL-4 production in just one FIGURE 2. The RXR agonist AGN194204 enhances Th2 development experiment. of naive DO11.10 T cells, whereas the RAR agonist TTNPB does not. At primary antigenic stimulation, lymph node cells from male DO11.10 mice RXR-selective agonist AGN194204 promotes Th2 development were treated with the indicated concentrations of the synthetic retinoids AGN194204 and TTNPB. Cells were restimulated 7 days later without Because 9-cis RA can bind to both RARs and RXRs, we used the retinoid treatment (secondary stimulation). The means Ϯ SD of triplicate synthetic RXR agonist AGN194204 to determine whether treat- cultures are shown. Asterisks indicate significant differences in IL-4 or

ment with an RXR-selective ligand would reproduce the Th2-en- IFN-␥ concentration between the AGN194204 and TTNPB treatments at Downloaded from hancing effect of 9-cis RA. Parallel cultures were treated with the the indicated concentration (p Ͻ 0.05 by Student’s t test). RAR-selective agonist TTNPB as a control for stimulation via RAR. As shown in Fig. 2, treatment of primary cultures with AGN194204 significantly enhanced Th2 development at concen- Ϫ Ϫ trations of 10 6 through 10 10 mol/L. In three of three additional If this were true, then neutralizing these cytokines during primary Ϫ experiments, treatment of primary cultures with 10 7 M stimulation would block the Th2-enhancing effect of 9-cis RA. We http://www.jimmunol.org/ AGN194204 also strongly enhanced IL-4 production at secondary conducted three experiments to test this hypothesis and found that stimulation ( p Ͻ 0.05; data not shown). On average, the IL-4 when neutralizing Abs and 9-cis RA were added at primary stim- concentration was 21-fold greater in the AGN194204 cultures than ulation, significant increases in IL-4 (and reductions in IFN-␥) in the DMSO control cultures (range, 3.6Ð56). were still seen in the 9-cis RA cultures at both secondary and As expected, treatment of primary cultures with the RAR ago- tertiary stimulation (Fig. 3) when either IFN-␥-(p Ͻ 0.001) or nist TTNPB had little effect on IL-4 production at secondary stim- IL-12-neutralizing Ab ( p Ͻ 0.001) were used. However, neutral- ulation (Fig. 2). In two of three additional experiments, the IL-4 ization of IL-4 blocked the Th2-enhancing activity of 9-cis RA Ϫ concentration was significantly higher after treatment with 10 7 M (Fig. 3). In two additional experiments, essentially the same results

TTNPB, compared with DMSO control cultures ( p Ͻ 0.05; data were reproduced for the RXR agonist AGN194204 (data not by guest on September 24, 2021 not shown), although the increases were small. On average, IL-4 shown). Ϫ concentrations were 1.8-fold greater in the 10 7 M TTNPB cul- To control for the possible nonspecific effect of Abs on cyto- tures than in the DMSO cultures (range, 1.3Ð2.3). IL-4 concentra- kine, isotype-control Abs for the IFN-␥ (IgG1), IL-12 (IgG2a), and tions in the AGN194204 cultures were always significantly greater IL-4 (IgG2b)-neutralizing Abs were included in several experi- than in the TTNPB cultures ( p Ͻ 0.05; data not shown). ments. These Abs did not significantly alter the Th2-enhancing effect of either AGN194204 or 9-cis RA. For example, mean IL-4 IL-4 is required for Th2 enhancement by RXR agonists, but concentrations at secondary stimulation for IgG1, IgG2a, IgG2b, ␥ IL-12 and IFN- are not and no Ab cultures with 10Ϫ7 M9-cis RA treatment were 1.88 Ϯ Because RA treatment can decrease IL-12 (9) and IFN-␥ produc- 0.11, 1.99 Ϯ 0.11, 2.14 Ϯ 0.13, and 1.94 Ϯ 0.07 ng/ml, respec- tion (26), it is possible that 9-cis RA and the RXR agonist tively, whereas IL-4 concentrations for the DMSO control cultures AGN194204 enhance Th2 development by decreasing production that received the same Abs were 0.20 Ϯ 0.01, 0.30 Ϯ 0.00, 0.19 Ϯ of these Th1-enhancing cytokines by APCs or bystander cells (26). 0.03, and 0.15 Ϯ 0.03 ng/ml, respectively.

FIGURE 1. 9-cis, but not all-trans, RA treatment at primary stimulation increases IL-4 production at secondary and tertiary stimulation. At primary antigenic stimulation, lymph node cells from DO11.10 mice were treated with the indicated concentrations of all-trans or 9-cis RA or with a vehicle control (0.1% DMSO). Cells were restimulated 7 days later without RA treatment (secondary stimulation). Cells from the secondary stimulation were again restimulated without RA 14 days after primary stimulation (tertiary stimulation). The means Ϯ SD of triplicate cultures are shown. The Journal of Immunology 4499

treated cultures were from 18% higher to 48% lower than in the DMSO control cultures.

RXR agonists diminish IL-12-induced Th1 development Under Th1-polarizing conditions (i.e., treatment of primary cultures with IL-12 and IL-4-neutralizing Ab), 9-cis RA treatment consistently decreased IFN-␥ production. In six experiments, 9-cis RA treatment decreased IFN-␥ production from 100 Ϯ 0.3 (mean Ϯ SE) to 50 Ϯ 0.3 ng/ml, with the percent decrease ranging from 18 to 85%. The mean decrease was 50% ( p Ͻ 0.001 by two-way ANOVA). The RXR agonist AGN194204 had a similar effect (data not shown).

Th2 enhancement by RXR agonist does not require APCs Although irradiated cells are unlikely to respond to retinoid treatment, it is possible that APCs, or bystander cells from the spleno- cyte population, may be required for RXR-mediated enhancement FIGURE 3. IL-4 neutralization blocks the Th2-enhancing effect of 9-cis of Th2 development. Thus, we wished to determine whether the

RA, but neutralization of IL-12 and IFN-␥ does not block this effect. At Downloaded from RXR agonist acts directly on naive Th0 cells in the absence of primary antigenic stimulation, lymph node cells from DO11.10 mice were treated with the indicated neutralizing Abs and 1 ϫ 10Ϫ7 M9-cis RA. Cells APCs. In addition, we wished to determine whether Th2 enhance- were restimulated 7 days later without RA treatment (secondary stimula- ment is limited to BALB/c mice (the background strain for tion). Cells from the secondary stimulation were again restimulated without DO11.10 mice), which are prone to Th2 development, or whether RA 14 days after primary stimulation (tertiary stimulation). Data from it could be reproduced in C57BL/6 mice, which are not (27). To either three (secondary stimulation) or two experiments (tertiary stimula- address these two questions, we purified naive Th0 cells (CD4ϩ, tion) are combined in this figure and mean values Ϯ SE (triplicate cultures L-selectinhigh) from C57BL/6 mice by flow cytometry and stimu- http://www.jimmunol.org/ per experiment) are reported. lated proliferation using anti-CD3 and anti-CD28 Ab in the presence of IL-4. We conducted two experiments, which gave essentially identical results. As shown in Fig. 5, treatment with To determine whether Th2 enhancement by RXR agonists was AGN194204 enhanced Th2 development compared with the ve- limited to low (endogenous) levels of IL-4 or if further enhance- hicle control, indicating that naive Th0 cells respond directly to ment could be seen under Th2-polarizing conditions, we examined stimulation with an RXR agonist. the effect of 9-cis RA and AGN194204 on Th2 development using different levels of IL-4 during primary stimulation. As shown in All-trans and 9-cis RA decrease IFN-␥ production by mature

␮ by guest on September 24, 2021 Fig. 4, blocking IL-4 activity with either 5 or 0.5 g/ml of neu- Th1 cells tralizing Ab largely blocked the ability of these compounds to enhance IL-4 production (although small, statistically significant In addition to examining the effects of RA treatment at primary increases were seen). However, both compounds substantially de- stimulation on subsequent Th1 and Th2 development, we also ex- ␥ creased IFN-␥ production in the presence of IL-4-neutralizing Ab amined the effect of RA treatment on IFN- and IL-4 production ( p Ͻ 0.001), with the effect of AGN194204 being greater than that by Th1 and Th2 populations produced by previous treatment with ␥ of 9-cis RA ( p Ͻ 0.05). When IL-4 was added to the primary IL-12 and IL-4, respectively. IFN- production by Th1 cultures Ͻ cultures at 1 ng/ml or 10 ng/ml, both compounds significantly in- differed significantly among the three treatment groups ( p ␥ creased IL-4 production at secondary stimulation ( p Ͻ 0.001), 0.001) (Fig. 6). IFN- production by the all-trans RA-treated cul- Ͻ with the effect of AGN194204 consistently being greater than that tures was 19% lower than in control cultures ( p 0.05) and was of 9-cis RA ( p Ͻ 0.05) (Fig. 4). The 9-cis RA did not consistently increase IL-4 production under strong Th2-enhancing conditions (10 ng/ml IL-4). In six additional experiments, IL-4 levels in 9-cis-

FIGURE 5. The RXR agonist AGN194204 enhances Th2 development of naive C57BL/6 T cells. At primary stimulation with anti-CD3 and anti- CD28 Ab, CD4ϩ/L-selectinhigh lymph node cells from male C57BL/6 mice FIGURE 4. IL-4 is required for the Th2-enhancing effects of 9-cis RA were treated with 1 ϫ 10Ϫ7 M AGN194204 or with the vehicle control and the RXR agonist AGN194204. During primary antigenic stimulation, (0.1% DMSO). Cells were washed, counted, and restimulated 6 days later lymph node cells from DO11.10 mice were treated with 1 ϫ 10Ϫ7 M9-cis without retinoid treatment using PMA and ionomycin. Supernatants were RA or AGN194204 and with anti-IL-4 Ab or IL-4, as indicated. Cells were collected after overnight incubation for measurement of IL-4 and IFN-␥. The restimulated 7 days later without RA treatment (secondary stimulation). means Ϯ SD of triplicate cultures are shown. The concentrations of both cy- The means Ϯ SD of triplicate cultures are shown. tokines differed significantly between the treatment groups (p Ͻ 0.05). 4500 RXR AGONISTS ENHANCE IN VITRO Th2 DEVELOPMENT Downloaded from FIGURE 6. RA treatment of Th1 cultures decreases IFN-␥ production but has little effect on IL-4 production by a Th2 culture. At primary anti- FIGURE 7. The RXR agonist AGN194204 increases production of IL-4 genic stimulation, lymph node cells from DO11.10 mice were treated with (p Ͻ 0.05) and IL-5 (p Ͻ 0.05), decreases production of IFN-␥ (p Ͻ 0.05), IL-4 to promote Th2 development or with IL-12 plus IL-4-neutralizing Ab and does not affect production of IL-10 relative to a vehicle control (0.1% to promote Th1 development or were not treated with cytokines or Abs. DMSO) during primary antigenic stimulation of lymph node cells from Cells were restimulated 7 days later (secondary stimulation). At tertiary DO11.10 mice. (The p values indicate comparison of the two treatments at Ϫ7 http://www.jimmunol.org/ stimulation (14 days after primary), cultures were treated with 1 ϫ 10 M all time points by two-way ANOVA.) Cells were treated with 1 ϫ 10Ϫ7 M all-trans,9-cis RA, or 0.1% DMSO (the diluent for RA). The means Ϯ SD AGN194204, 0.1% DMSO, or 10 ng/ml IL-4. Values are the means Ϯ SE of triplicate cultures are shown. for individual concentrations from three independent experiments. Means at a given time point with different superscripts differ significantly from one another (p Ͻ 0.05) by two-way ANOVA. 40% lower in the 9-cis cultures ( p Ͻ 0.05). No significant differences were seen in IL-4 production by Th2 cultures ( p ϭ 0.08) (Fig. 6). Similar results were seen when RA treatments were added ulation were 1.5 Ϯ 0.5 ϫ 108 copies/50 ng total RNA (mean Ϯ SE, at secondary stimulation (data not shown). n ϭ 4 experiments). ␤-Actin mRNA levels increased over time. Levels did not differ by retinoid or IL-4 treatment at 24 and 48 h, by guest on September 24, 2021 RXR agonist increases IL-4 and IL-5 and decreases IFN-␥ but at 72 h, ␤-actin levels were higher in the IL-4 treatment than production by naive Th0 cells in the DMSO treatment ( p Ͻ 0.05). The ␤-actin levels 24 h after IL-4 and IFN-␥ are produced by developing Th0 cells and enhance Ag stimulation in the DMSO, AGN194204, and IL-4 treatments Th2 and Th1 development, respectively. To test the hypothesis that were 2.7 Ϯ 0.3, 2.7 Ϯ 0.3, and 2.9 Ϯ 0.3 ϫ 108 copies/50 ng total an RXR agonist would enhance production of the former and di- RNA, respectively. At 48 h, the corresponding values were 4.6 Ϯ minish production of the latter during primary antigenic stimula- 0.4, 5.2 Ϯ 0.4, and 5.3 Ϯ 0.5 ϫ 108 copies/50 ng total RNA, tion of naive Th0 cells, DO11.10 cultures were treated with respectively. At 72 h, these values had increased to 7.7 Ϯ 0.2, AGN194204 exactly as described for experiments examining Th1/ 8.2 Ϯ 0.2, and 9.0 Ϯ 0.2 ϫ 108 copies/50 ng total RNA, respec- Th2 phenotype at secondary stimulation. The RXR agonist tively (n ϭ 2 experiments). AGN194204 significantly increased IL-4 concentrations relative to Treatment with the RXR agonist AGN194204 increased IL-4 the vehicle control ( p ϭ 0.014 by two-way ANOVA); differences and decreased IFN-␥ mRNA levels 48 and 72 h after primary Ag were statistically significant 48 and 72 h after stimulation ( p Ͻ stimulation (Fig. 8, C and E; Table I), consistent with the changes 0.05) (Fig. 7A). The RXR agonist also decreased IFN-␥ concen- in IL-4 and IFN-␥ protein levels seen in supernatants at these same trations relative to the vehicle ( p Ͻ 0.05) and IL-4 treatments ( p Ͻ time points (Fig. 7). Because the balance in expression of these 0.05), which did not differ from one another (Fig. 7B). When in- genes is important in Th1/Th2 development, we compared the dividual time points were considered separately, the only signifi- mean IL-4-IFN-␥ ratios (Fig. 8A) in the AGN194204, IL-4, and cant difference was that the IFN-␥ concentration in the IL-4 treat- vehicle treatments at 48 h and found that they were 20, 12, and 3.1, ment was greater than in the AGN194204 treatment ( p Ͻ 0.05). In respectively (SE ϭ 0.85; n ϭ 4; p Ͻ 0.001; all comparisons among addition, IL-5 concentrations were increased over both the vehicle means significant at p Ͻ 0.05). At 72 h, these ratios were 20, 11, control ( p Ͻ 0.05) and IL-4 treatment groups ( p Ͻ 0.05) by treat- and 2.3 (SE ϭ 0.67; p ϭ 0.007; all means significantly different) ment with the RXR agonist at both 48 ( p Ͻ 0.05) and 72 h ( p Ͻ (Fig. 8A). This analysis reveals that treatment with the RXR ago- 0.05) (Fig. 7C). IL-10 concentrations were not affected by treat- nist shifted the IL-4-IFN-␥ ratio more strongly in favor of Th2 ment with the RXR agonist (Fig. 7D). development than did treatment with IL-4. RXR agonist treatment also enhanced expression of the Th2 RXR agonist increases Th2 and decreases Th1 gene expression transcription factor GATA-3 within 24 h of antigenic stimulation. by naive Th0 cells GATA-3 determines which Th0 cells progress to become Th2 Because RXR is a nuclear receptor that modulates gene expres- memory cells (28) by enhancing IL-4 and diminishing IFN-␥ ex- sion, we examined mRNA levels by real-time PCR 24, 48, and pression (29). GATA-3 also strongly enhances IL-5 production 72 h after antigenic stimulation and treatment with DMSO (vehicle (30). Treatment with the RXR agonist increased GATA-3 expres- control), AGN194204, or IL-4. ␤-Actin mRNA levels before stim- sion at all time points (Fig. 8F; Table I). Treatment with IL-4 had The Journal of Immunology 4501

tios at 48 h and found that they were 8.5, 5.7, and 3.2 (SE ϭ 0.14; p Ͻ 0.001; all means signiﬁcantly different) in the AGN194204, IL-4, and vehicle control groups, respectively. Results at 72 h were similar, with the mean ratios being 10, 5.4, and 4.7 (SE ϭ 0.06; p Ͻ 0.001; all means signiﬁcantly different). Treatment with the RXR agonist had an even greater effect than IL-4 in shifting the GATA-3:T-bet ratio in favor of Th2 development. Finally, we had also hypothesized that an RXR-mediated decrease in mRNA for the protein tyrosine kinase Fyn might occur because RA treatment of embryonic stem cells decreases Fyn mRNA expression (36). This decrease could enhance Th2 development because Fyn mediates inhibition of Th2 cytokine expression during Th0 development (37). However, treatment with the RXR agonist did not produce a consistent change in Fyn mRNA expression (Table I).

Discussion We have demonstrated that in vitro treatment of naive DO11.10 Downloaded from Th0 cell cultures with the vitamin A metabolite 9-cis RA during primary antigenic stimulation strongly enhances subsequent development of IL-4-secreting Th2 memory cells. The level of enhancement is comparable to that seen with IL-4 treatment, the classic mechanism of inducing Th2 development (38). This novel activity

of 9-cis RA appears to be mediated via the RXR family of nuclear http://www.jimmunol.org/ retinoid receptors, because treatment with the RXR-selective agonist AGN194204 has even more potent Th2-enhancing activity than does 9-cis RA (which binds to both RARs and RXRs). In FIGURE 8. mRNA ratios for Th1 and Th2 pathway genes in CD4ϩ/ ϩ addition, the RXR pathway in T cells seems to be responding KJ1-26 Th0 cells purified by FACS after antigenic stimulation with ϫ Ϫ7 directly to this treatment, because enhanced Th2 development oc- OVA323Ð339 and irradiated splenic APCs. Cells were treated with 1 10 M AGN194204, 0.1% DMSO, or 10 ng/ml IL-4. Values are means Ϯ curred in purified, naive Th0 cells stimulated with anti-CD3 and pooled SE determined by two-way ANOVA for duplicate measurements. anti-CD28 Ab, as well as in DO11.10 cultures using splenic APCs Statistical analysis and mean values for each gene (from three to four plus peptide Ag. This effect was also seen in two genetic back- independent experiments) are given in Table I. grounds: BALB/c and C57BL/6. These data provide strong evi- by guest on September 24, 2021 dence that vitamin A-mediated stimulation of the RXR pathway acts directly on naive Th0 cells to enhance Th2 development. a similar effect (Table I). Initially, IL-4 treatment caused a greater Our finding that RXR agonists enhance Th2 development in increase in GATA-3 (relative to the vehicle control) than did vitro is consistent with previous work demonstrating that vitamin AGN194204 (2.26- vs 1.24-fold at 24 h), but by 72 h the ratios Adeficiency diminishes Th2 responses, whereas treatment with were essentially the same (1.55 vs 1.57). Another Th2-specific high-level dietary vitamin A or RA enhances such responses. An- transcription factor is c-maf, which induces IL-4 expression (31) imal studies have shown that vitamin A deficiency diminishes and enhances Th2 development by other pathways (32). Like polyclonal and Ag-specific production of Th2 cytokines (6, 10), GATA-3, c-maf expression was enhanced by both AGN194204 diminishes Th2-mediated Ab responses (4, 8), increases Ag-spe- and IL-4 24 h after stimulation, although not at 48 and 72 h cific production of IFN-␥ (10, 26), increases IL-12 production by (Table I). unstimulated lymph node cells (9), and increases anti-viral (Th1- The expression of all Th2-enhancing genes was not increased by mediated) IgG2a responses (7). In addition, in vitro treatment with AGN194204 and IL-4 treatment. The expression of STAT-6, RA decreases production of IL-12 by APCs and IFN-␥ by T cells which also plays a key role in Th2 development by participating in and NK cells (9, 39). Conversely, supplemental vitamin A in- signal transduction from the IL-4R to IL-4 responsive genes (33), creases the IgA response and production of IL-5 and IL-10 while was not altered by AGN194204 or IL-4 treatment (Table I). In decreasing the IgG response and production of IFN-␥ (11, 12). addition, we measured mRNA levels for the ␣-chain of IL-4R Similarly, RA treatment of mice with experimental allergic en- because the promoter region for this gene contains a possible RA cephalomyelitis (which is mediated by Th1 cells) decreases the response element (our unpublished observation). However, mRNA severity of disease, whereas in vitro treatment of lymph node levels for IL-4R were not altered by treatment with the RXR ag- cells from these mice with all-trans RA decreases IFN-␥ and onist (Table I). increases IL-4 production (13). Because Th1 and Th2 development genes can be cross-regula- Whereas data in the literature suggest that vitamin A deficiency tory (34), we also examined expression of Th1-enhancing genes. diminishes Th2 responses by creating a lymph node environment Although expression of such genes was not altered at 24 h, we conducive to Th1 development (9), our findings suggest that RXR found that expression of both T-bet, a transcription factor that agonists may, in addition, directly enhance Th2 development by plays a key role in Th1 development (35), and IL-12R were de- acting on naive Th0 cells. Addition of IL-12-neutralizing Ab to creased by RXR agonist treatment 48 and 72 h after stimulation cultures during primary stimulation did not diminish the ability of (Fig. 8, B and D; Table I). Because the relative levels of GATA-3 either 9-cis RA or AGN194204 to enhance Th2 development, as and T-bet expression may be important in determining eventual would be expected if this Th2-enhancing activity were due to mod- Th1/Th2 development, we compared the mean GATA-3:T-bet ra- ulation of IL-12 production by APCs. Similarly, 9-cis RA could 4502 RXR AGONISTS ENHANCE IN VITRO Th2 DEVELOPMENT decrease IFN-␥ production by NK or T cells, thus indirectly en- variety of important physiological stimuli. Although not yet widely hancingTh2 development. Again, addition of IFN-␥-neutralizing studied in immunology, nuclear receptors are a family of mole- Ab did not alter the ability of 9-cis RA and AGN194204 to en- cules that modulate immune function, as well as other physiologic hance Th2 development, supporting an IFN-␥-independent mech- processes, in response to a variety of signals, including diet (e.g., anism. Thus 9-cis RA and AGN194204 enhance Th2 development vitamins act via RAR, RXR, and VDR; fatty acids act via PPAR; by pathways not requiring the presence of extracellular IL-12 or energy and iodine modulate thyroid hormone receptor activity; and IFN-␥. In addition, we found that the RXR agonist AGN194204 cholesterol-derived oxysterols bind to liver X receptor), gender stimulated Th2 development in purified, naive Th0 cells when no and reproduction (sex hormones act via the estrogen, androgen, APCs were present. This provides strong evidence that the RXR and progesterone receptors), stress (cortisol acts via the glucocor- agonist acts directly on Th0 cells. ticoid receptor), and exposure to xenobiotics (some of which bind Although the presence of IFN-␥ and IL-12 are not required to to the pregnane X receptor) (16). Their role in the regulation of mediate the Th2-enhancing effect of 9-cis RA and AGN194204, immune function deserves further study. both compounds require the presence of IL-4 to enhance Th2 de- In summary, we have found that RXR agonists strongly enhance velopment. This point was clearly shown in experiments demon- Th2 development by acting directly on naive Th0 cells. This ac- strating that treatment of primary cultures with IL-4-neutralizing tivity appears to be independent of the previously demonstrated Ab blocks the ability of these compounds to increase IL-4 produc- ability of RA to decrease the production of the Th1-enhancing tion by secondary and tertiary cultures. This requirement for IL-4 cytokine IL-12 by APCs and IFN-␥ by T cells. Our data suggest in primary cultures suggests that these retinoids could enhance that RXR agonists may enhance production of the Th2 transcrip- production of IL-4. However, treatment of mature Th1, Th2, or tion factors GATA-3 and c-maf, although this hypothesis requires Downloaded from uncommitted Th cells with 9-cis RA did not enhance IL-4 produc- confirmation. Thus, RXR agonists may play a significant role in tion (Fig. 6), and treatment of naive Th0 cells with 9-cis RA in the both diminishing Th1 development and enhancing Th2 develop- presence of 10 ng/ml IL-4 seemed to slightly (but significantly) ment by distinct pathways. This work also suggests that the use of decrease IL-4 concentrations at secondary stimulation (data not RXR-selective compounds might provide a means of modulating shown). These observations do not support a direct role for the the Th cell response in treating autoimmune or chronic inflamma-

RXR pathway in enhancing IL-4 production. However, treatment tory diseases, which may be Th1-mediated, or in promoting Th2 http://www.jimmunol.org/ of primary Th0 cultures with AGN194204 did increase IL-4 pro- responses, which are beneficial in protecting against infectious dis- duction as well as decrease IFN-␥ production by these cultures. eases caused by extracellular pathogens. Both effects were seen at 48 h, in addition to a significant increase in IL-5. These findings could be due to a direct effect of the RXR Acknowledgments pathway on increasing IL-4 production and subsequent enhance- We thank Leslie Hayden for HPLC analysis of RA stock solutions and ment of Th2 development, but a direct effect of RXR agonists on retinol analysis of FBS. We also thank Carol Oxford for performing the cell other genes in the Th1 or Th2 development pathway could also sorting. indirectly enhance IL-4 production in primary cultures. by guest on September 24, 2021 Results of our mRNA expression studies using purified Th0 References cells isolated from primary stimulation cultures support the hy- 1. Glasziou, P. P., and D. E. Mackerras. 1993. Vitamin A supplementation in in- pothesis that the RXR pathway indirectly increases IL-4 produc- fectious diseases: a meta-analysis. BMJ 306:366. tion during primary stimulation by enhancing GATA-3 and c-maf 2. Fawzi, W. W., T. C. Chalmers, M. G. Herrera, and F. Mosteller. 1993. Vitamin A supplementation and child mortality: a meta-analysis. J. Am. Med. Assoc. 269: transcription. Treatment with AGN194204 first increases mRNA 898. levels for the GATA-3 and c-maf genes (24 h), followed by in- 3. Ross, A. C., and C. B. Stephensen. 1996. Vitamin A and retinoids in antiviral creased IL-4 mRNA and protein (48 h), increased IL-5 protein (48 responses. FASEB J. 10:979. 4. Smith, S. M., and C. E. Hayes. 1987. Contrasting impairments in IgM and IgG h), decreased IFN-␥ mRNA and protein (48 h), and decreased responses of vitamin A-deficient mice. Proc. Natl. Acad. Sci. USA 84:5878. mRNA levels for the Th1-enhancing genes T-bet and IL-12R (48 5. Pasatiempo, A. M., M. Kinoshita, D. T. Foulke, and A. C. Ross. 1994. The Ab response of vitamin A-deficient rats to pneumococcal polysaccharide is enhanced h). This scenario is plausible because GATA-3 and c-maf both through coimmunization with lipopolysaccharide. J. Infect. Dis. 169:441. enhance IL-4 transcription, whereas GATA-3 also enhances IL-5 6. Wiedermann, U., L. A. Hanson, H. Kahu, and U. I. Dahlgren. 1993. Aberrant and diminishes IFN-␥ transcription. The decreased mRNA levels T-cell function in vitro and impaired T-cell dependent Ab response in vivo in vitamin A-deficient rats. Immunology 80:581. for T-bet and IL-12R could be a direct effect of the RXR agonist 7. Stephensen, C. B., Z. Moldoveanu, and N. N. Gangopadhyay. 1996. Vitamin A or an indirect effect of IL-4 on the expression of these Th1 pathway deficiency diminishes the salivary immunoglobulin A response and enhances the genes. Because the promoter regions of GATA-3 and c-maf are not serum immunoglobulin G response to influenza A virus infection in BALB/c mice. J. Nutr. 126:94. well characterized, further work is clearly needed to determine 8. Carman, J. A., S. M. Smith, and C. E. Hayes. 1989. Characterization of a helper how the RXR pathway acts to enhance Th2 development. Inter- T lymphocyte defect in vitamin A deficient mice. J. Immunol. 142:388. estingly, a recent study found that stimulation of naive Th0 cells 9. Cantorna, M. T., F. E. Nashold, and C. E. Hayes. 1995. Vitamin A deficiency results in a priming environment conducive for T1 cell development. Eur. J. Im- with anti-CD3 and anti-CD28 plus vitamin D also enhanced munol. 25:1673. GATA-3 and c-maf RNA levels (and subsequent Th2 develop- 10. Cantorna, M. T., F. E. Nashold, and C. E. Hayes. 1994. In vitamin A deficiency ment) compared with a vehicle control (40). Because vitamin D multiple mechanisms establish a regulatory T helper cell imbalance with excess Th1 and insufficient Th2 function. J. Immunol. 152:1515. also binds to a nuclear receptor (VDR) and can form heterodimers 11. Nikawa, T., M. Ikemoto, M. Kano, K. Tokuoka, K. Hirasaka, S. Uehara, with RXR, it is possible that the same pathway may be stimulated K. Takatsu, K. Rokutan, and K. Kishi. 2001. Impaired vitamin A-mediated mu- cosal IgA response in IL5 receptor knockout mice. Biochem. Biophys. Acta 285: by two nutrients: vitamins A and D. 546. The observations that treatment of T cells with agonists for 12. Cui, D., Z. Moldoveanu, and C. B. Stephensen. 2000. High-level dietary vitamin RAR, RXR, VDR (discussed above), and PPAR-␥ (41) modulate A enhances T-helper type 2 cytokine production and secretory immunoglobulin a response to influenza A virus infection in BALB/c mice. J. Nutr. 130:1132. T cell activity emphasize the important role of nutrients in regu- 13. Racke, M. K., D. Burnett, S. H. Pak, P. S. Albert, B. Cannella, C. S. Raine, lating immune function via nuclear receptors. Nuclear receptors D. E. McFarlin, and D. E. Scott. 1995. Retinoid treatment of experimental aller- represent an important second level of regulation of the immune gic encephalomyelitis: IL-4 production correlates with improved disease course. J. Immunol. 154:450. response (after the first-level regulation by direct exposure to 14. Chun, T. Y., J. A. Carman, and C. E. Hayes. 1992. Retinoid repletion of vitamin pathogens and the regulatory molecules that they induce) by a A-deficient mice restores IgG responses. J. Nutr. 122:1062. The Journal of Immunology 4503

15. Cippitelli, M., J. Ye, V. Viggiano, A. Sica, P. Ghosh, A. Gulino, A. Santoni, and 29. Ferber, I. A., H. J. Lee, F. Zonin, V. Heath, A. Mui, N. Arai, and A. O’Garra. H. A. Young. 1996. Retinoic acid-induced transcriptional modulation of the hu- 1999. GATA-3 significantly downregulates IFN-␥ production from developing man interferon-␥ promoter. J. Biol. Chem. 271:26783. Th1 cells in addition to inducing IL-4 and IL-5 levels. Clin. Immunol. 91:134. 16. Chawla, A., J. J. Repa, R. M. Evans, and D. J. Mangelsdorf. 2001. Nuclear 30. Zhang, D. H., L. Y. Yang, and A. Ray. 1998. Differential responsiveness of the receptors and lipid physiology: opening the X-files. Science 294:1866. IL-5 and IL-4 genes to transcription factor GATA-3. J. Immunol. 161:3817. 17. Mangelsdorf, D., K. Umesono, and R. Evans. 1994. The retinoid receptors. In The 31. Ho, I. C., M. R. Hodge, J. W. Rooney, and L. H. Glimcher. 1996. The proto- Retinoids: Biology, Chemistry and Medicine, 2nd ed. M. Sporn, A. Roberts, and oncogene C-MAF is responsible for tissue-specific expression of interleukin-4. D. Goodman, eds. Raven Press, New York, p. 319. Cell 85:973. 18. Pfahl, M., and F. Chytil. 1996. Regulation of metabolism by retinoic acid and its 32. Ho, I. C., D. Lo, and L. H. Glimcher. 1998. c-maf promotes T helper cell type 2 nuclear receptors. Annu. Rev. Nutr. 16:257. (Th2) and attenuates Th1 differentiation by both interleukin 4-dependent and 19. de Urquiza, A. M., S. Y. Liu, M. Sjoberg, R. H. Zetterstrom, W. Griffiths, independent mechanisms. J. Exp. Med. 188:1859. J. Sjovall, and T. Perlmann. 2000. Docosahexaenoic acid, a ligand for the retinoid X receptor in mouse brain. Science 290:2140. 33. Kaplan, M. H., U. Schindler, S. T. Smiley, and M. J. Grusby. 1996. Stat6 is 20. Johnson, B. S., R. A. Chandraratna, R. A. Heyman, E. A. Allegretto, L. Mueller, required for mediating responses to IL4 and for development of Th2 cells. Im- and S. J. Collins. 1999. Retinoid X receptor (RXR) agonist-induced activation of munity 4:313. dominant-negative RXR-retinoic acid receptor ␣403 heterodimers is develop- 34. Rengarajan, J., S. J. Szabo, and L. H. Glimcher. 2000. Transcriptional regulation mentally regulated during myeloid differentiation. Mol. Cell. Biol. 19:3372. of Th1/Th2 polarization. Immunol. Today 21:479. 21. Vuligonda, V., S. M. Thacher, and R. A. S. Chandraratna. 2001. Enantioselective 35. Szabo, S. J., S. T. Kim, G. L. Costa, X. Zhang, C. G. Fathman, and syntheses of potent retinoid X receptor ligands: differential biological activities of L. H. Glimcher. 2000. A novel transcription factor, T-bet, directs Th1 lineage individual antipodes. J. Med. Chem. 44:2298. commitment. Cell 100:655. 22. Murphy, K. M., A. B. Heimberger, and D. Y. Loh. 1990. Induction by antigen of 36. Forrester, L. M., A. Nagy, M. Sam, A. Watt, L. Stevenson, A. Bernstein, ϩ ϩ intrathymic apoptosis of CD4 CD8 TCRlo thymocytes in vivo. Science 250: A. L. Joyner, and W. Wurst. 1996. An induction gene trap screen in embryonic 1720. stem cells: identification of genes that respond to retinoic acid in vitro. Proc. Natl. 23. Marrack, P., R. Shimonkevitz, C. Hannum, K. Haskins, and J. Kappler. 1983. The Acad. Sci. USA 93:1677. major histocompatibility complex-restricted antigen receptor on T cells. IV. An

37. Tamura, T., O. Igarshi, A. Hino, H. Yamane, S. Aizawa, T. Kato, and Downloaded from anti-idiotypic Ab predicts both antigen and I-specificity. J. Exp. Med. 158:1635. H. Nariuchi. 2001. Impairment in the expression and activity of Fyn during dif- 24. Ohata, M., M. Lin, M. A. Satre, and H. Tsukamoto. 1997. Diminished retinoic ferentiation of naive CD4ϩ T cells into the Th2 subset. J. Immunol. 167:1962. acid signaling in activated hepatic stellate cells in liver fibrosis. Am. J. Physiol. 38. Hsieh, C. S., A. B. Heimberger, J. S. Gold, A. O’Garra, and K. M. Murphy. 1992. 272:G589. Differential regulation of T helper phenotype development by interleukins 4 and 25. Rogers, W. O., C. T. Weaver, L. A. Kraus, J. Li, L. Li, and R. P. Bucy. 1997. 10 in an ␣␤ T-cell-receptor transgenic system. Proc. Natl. Acad. Sci. USA 89: Visualization of antigen-specific T cell activation and cytokine expression in 6065. vivo. J. Immunol. 158:649. 26. Carman, J. A., C. E. Hayes. 1991. Abnormal regulation of IFN-␥ secretion in 39. Frankenburg, S., X. Wang, and Y. Milner. 1998. Vitamin A inhibits cytokines produced by type 1 lymphocytes in vitro. Cell. Immunol. 185:75. vitamin A deficiency. J. Immunol. 147:1247. http://www.jimmunol.org/ 27. Bix, M., Z. E. Wang, B. Thiel, N. J. Schork, and R. M. Locksley. 1998. Genetic 40. Boonstra, A., F. J. Barrat, C. Crain, V. L. Heath, H. F. J. Savelkoul, and A. ϩ regulation of commitment to interleukin 4 production by a CD4ϩ T cell-intrinsic O’Garra. 2001. 1␣,25-Dihydroxyvitamin D3 has a direct effect on naive CD4 T mechanism. J. Exp. Med. 188:2289. cells to enhance the development of Th2 cells. J. Immunol. 167:4974. 28. Farrar, J. D., W. Ouyang, M. Lohning, M. Assenmacher, A. Radbruch, 41. Harris, S. G., and R. P. Phipps. 2001. The nuclear receptor PPAR␥ is expressed O. Kanagawa, and K. H. Murphy. 2001. An instructive component in T helper by mouse T lymphocytes and PPAR␥ agonists induce apoptosis. Eur. J. Immunol. cell type 2 (Th2) development mediated by GATA-3. J. Exp. Med. 193:643. 31:1098. by guest on September 24, 2021