Given the multiple reports describing the effects of retinoids on cytokine production by murine splenocytes and T cells, we initially examined the effects of various concentrations of ATRA and 9-cis-RA on the production of Th1- and Th2-associated cytokines by anti-CD3-stimulated human T cells and PBMC in vitro. Culture supernatants were examined 48 hrs after stimulation for cytokine levels. The results in Figure 1 demonstrate a representative dose response curve of ATRA and 9-cis-RA on the expression of IL-4, IL-5 and IFN-γ by human T cells and PBMC. These data reveal that the Th2 cytokines, IL-4 and IL-5, are induced in T cells and PBMC in a dose-dependent manner, while the production of the Th1 cytokine, IFN-γ, is inhibited in response to increasing concentrations of ATRA and 9-cis-RA. These findings were highly reproducible in greater than 95% of the PBMC and T cells donors examined (n > 20). Based on the above curves, ATRA and 9-cis-RA were utilized in subsequent experiments at the 10^-7 M (100 nM) concentration, a dose range that we and others 34 have shown to be optimal for human T cell stimulation.

Besides these Th1 and Th2 cytokines, additional cytokines were also examined including IL-12p70, IL-13 and TNF-α. Protein levels of IL-12p70, TNF-α and IFN-γ in the culture supernatants were significantly decreased, while IL-4, IL-5 and IL-13 protein levels were significantly increased by ATRA treatment in all donors tested (Table 1). ATRA and 9-cis-RA were equipotent in their effect on the majority of these cytokines. However, 9-cis-RA was significantly more effective than ATRA in inhibiting TNF-α. ATRA and 9-cis-RA failed to significantly alter IL-10 and IL-15 cytokine production by stimulated human PBMCs. While not shown, IL-18 levels were not consistently affected by ATRA or 9-cis-RA.

As PBMCs are a heterogeneous population of cytokine producing cell populations, we also examined the influence of ATRA on Th1- and Th2-associated cytokine production from purified CD4⁺, CD8⁺ and non-fractionated T cells. Similar to the above results, our data revealed that IL-4 and IL-5 synthesis was increased and IFN-γ expression was inhibited in the supernatants of all stimulated T cell populations examined (data not shown). The ATRA-induced changes in the ratio of these two cytokines were greatest in PBMCs followed by CD4⁺ T cells, T cells and then by CD8⁺ T cells.

ATRA also modulated the expression of several other inflammatory cytokines and chemokines by human PBMCs (Table 2). Many of these factors have not been previously shown to be influenced by retinoids. Interestingly, the expression of IL-8 and MCP-1 were both augmented by ATRA treatment, while the expression of RANTES, MIP-1β, G-CSF, IL-1β and IL-6 were significantly inhibited by ATRA treatment of anti-CD3 mAb-stimulated human PBMCs. While MIP-1α was not significantly different between the control and ATRA-treated groups, there was a trend towards downregulation (p = 0.09). IL-17, GM-CSF and IFN-α failed to demonstrate any significant differences between control and experimental cultures. The relevance of exogenous or endogenous retinoids and their nuclear receptors in the expression and production of various cytokines and chemokines remains to be defined.

We next sought to determine whether the RA-induced decrease in IFN-γ and increase in IL-4 production may be due to a shift in the frequency of cytokine producing T cells and/or the quantity of cytokine being produced by individual cytokine-producing cell. Using intracellular flow cytometric analysis of several donors, we observed a decrease in the percentage of T cells staining for IFN-γ in cultures treated with either ATRA (5.2%) or 9-cis-RA (6.3%) when compared to control T cells (15.8%). Similar studies were performed for IL-4 and IL-5; however, consistently poor intracellular staining for these cytokines was observed using several human T cell donors in our hands. So to more accurately measure the effects of RA on the frequency of Th2 cytokine production, ELISPOT analysis was performed. The results in Table 3 demonstrate the representative results of anti-CD3 mAb-activated T cells derived from single donor of two examined where a range of an approximate 3-fold increase in the frequency of IL-4- and an approximate 2-fold increase in the frequency of IL-5-producing T cells in ATRA-treated T cell cultures compared to vehicle-treated cells. In addition, a 1.5-fold decrease in IFN-γ-producing T cells was also observed between cultures treated with ATRA compared to a vehicle control. These data suggest that treatment with RA results in an increase in the frequency of T cells producing Th2 cytokines, possibly through the direct induction of type 2 cytokine or transcription factor RNA.

We believe that inhibition of accessory cell cytokine production by ATRA or 9-cis-RA may account for the greater type 2 polarization observed in PBMCs versus non-fractionated T cells or the CD4⁺ and CD8⁺ T cell populations. We therefore examined whether ATRA-induced alterations in IL-4, IL-5 and IL-13 production persisted under strong Th1- or Th2-polarizing conditions. ATRA-induced changes in IL-5 and IL-13 were not further altered upon the addition of exogenous IL-12, IL-4, or neutralizing anti-IL-4 and anti-IL-12 antibody (Figure 2). The biological activity of such manipulation is evident from corresponding increase or decrease in IFN-γ production. Due to extreme variations in the degree of stimulation by IL-4 and IL-12 in the human donors examined, only a representative experiment with one donor is shown in Figure 2. Three individual donors were examined in this experimental series and each donor demonstrated similar effects. Moreover, we have found that the addition of other potential type 1-inducing cytokines including IFN-γ or IL-18 in the presence of neutralizing anti-IL-4 antibody also failed to alter the effect of ATRA on IL-5 or IL-13 production within the cultures (data not shown). Moreover, similar to primary human T cells and these polarization cultures, we have also found that ATRA increases the proliferation and IL-4 production by human and murine Th2 clones with little to no major effects or inhibition on Th1 clones (data not shown).

To assess whether this same requirement was necessary for RA using human cells, purified human T cells were stimulated with anti-CD3 mAb and ATRA in the presence or absence of CD28 costimulation. The results in Figure 4 demonstrate that even in the absence of CD28 costimulation, ATRA increases IL-5 production by primary human T cells. Upon stimulation with anti-CD28mAb, there is even a greater level of IL-5 production by T cells stimulated with ATRA. In addition, CD28-costimulation was previous found to be necessary to observe the IFN-γ-inhibitory effects of ATRA using a murine model 35. Similarly, we have found that ATRA inhibited IFN-γ production in the presence of anti-CD28 mAb (Figure 3). Thus, ATRA is capable of inducing Th2 cytokine expression in the absence of costimulatory signals such as CD28.

As type 2 differentiation is thought to involve two temporally separate signals, a differentiation and proliferation signal 36, we next examined the time course of ATRA-induced Th2 polarization of CD4⁺ T cells in response to anti-CD3 and -CD28 mAb. As shown in Figure 4, the effects of ATRA were evident as early as 12 hrs after the initiation of the cultures. However, these differences did not reach statistical significance for IL-5 levels until 24 and 48 hr of culture. Similar trends for IL-4 expression were observed as shown here for IL-5 (data not shown).

We next examined the effects of ATRA and 9-cis-RA on the level of mRNA of each Th1- or Th2-associated cytokine. As shown in Figure 5, the relative copy number of IL-4 and IL-5 mRNA was significantly increased by ATRA and 9-cis-RA utilizing PBMC derived from 5 different donors. Although similar levels of IL-4 protein were induced by both retinoids, 9-cis-RA was found to be at least two-fold more effective at inducing IL-4 message compared to ATRA. In addition, the relative copy number of IFN-γ mRNA was modestly but significantly decreased by both 9-cis-RA and ATRA. The results in Figure 5 were performed in the presence of low dose IL-2 to assist in cell activation and the maintenance of the cells. Similar results were observed in the absence of IL-2 or in the presence of anti-CD28 mAb (data not shown). Moreover, ATRA treatment of human T cells over various time intervals enhanced the expression of the Th2 cytokine, IL-4 and the Th2 transcription factors, GATA-3, c-MAF and STAT6 at 6 and 24 hours post stimulation and decreased the expression of the Th1 transcription factor, T-bet, at these same time points (Figure 6). The specific effects on the expression of Th2 transcription factors and subsequently cytokine production strongly support the ability of RA to facilitate the polarization of human T cells to a Th2 phenotype.

ATRA, 9-cis-RA were purchased from Sigma, St. Louis, MO. Retinoids were dissolved at various concentrations in 100% ETOH, overlayered with argon gas and stored at -80°C in the dark until used. Recombinant human IFN-γ, IL-4, IL-12 and IL-18 were obtained from R & D Systems (Minneapolis, MN). Recombinant IFN-γ was obtained from Biosource Int. (Camarillo, CA). Neutralizing monoclonal antibodies (mAbs) to human IFN-γ (clone 25718.111) IL-4 (clone 34019.111), and IL-12 (clone 24910.1) were also obtained from R & D Systems. Based upon manufacturer's testing lot-specific testing, 1 :g of anti-IFN-γ, anti-IL-4, and anti-IL-12 will neutralize 5 ng/ml, 300 pg/ml and 700 pg/ml of rhIFN-γ, rhIL-4, and rhIL-12, respectively.

Whole blood was acquired from healthy human volunteers between the ages of 21–55 years who provided informed consent. PBMC were isolated by Ficoll Paque (Amersham Pharmacia Biotech, Piscataway, NJ) density gradient centrifugation followed by treatment with ammonium chloride (ACK) lysis solution (Biofluids, Gaithersburg, MD) to eliminate the remaining erythrocytes. The isolated cells were subsequently washed 2 times in PBS and resuspended in RPMI 1640 (Biofluids) supplemented with 10% heat-inactivated FBS (Sigma), 2% heat-inactivated pooled human AB serum (Sigma), 50 μM mercaptoethanol (Gibco BRL Gaithersburg, MD), 1 mM sodium pyruvate (Biofluids), 2 mM glutamine, 1 × non-essential amino acid solution (Biofluids), 1 mg/ml gentamicin (Biowhittaker, Walkersville, MD), 100 U/ml penicillin (Biofluids), 100 :g/ml streptomycin (Biofluids), and 20 mM HEPES buffer (Biofluids). T cells, CD4⁺ T cells and CD8⁺ T cells were isolated by negative selection using enrichment columns according to manufacturer's instructions (R & D Systems). All of these cells were typically > 93% pure as assessed by flow cytometric analysis. The contaminating cell population was largely CD8⁺ cells. Most likely, these cells were NK cells based on their size and granularity.

PBMC (2.5 × 10 ⁶ cells/ml) were activated with 200 ng/ml of immobilized anti-CD3ε OKT-3, Ortho, Raritan, NJ) ± 0.001 to 1 μM of various retinoids or ETOH vehicle control for 48 h. IL-2 (Teceleukin, Hoffman LaRoche, Nutley, NJ) at 10 U/ml was added where indicated. Where indicated, 1 μg/ml of neutralizing anti-cytokine mAb was added to the cultures. Alternatively, T cells (1.0 × 10 ⁶ cells/ml) were activated with 200 ng/ml of immobilized anti-CD3 and 0.1 to 1 μg of soluble anti-CD28 (clone 28.2, Pharmingen) or IL-2 at 10 U/ml was added where indicated to provide co-activation or costimulatory signals. PBMCs or T cells were harvested at various times after incubation at 37° and 5% CO₂. The 48 hr time interval was selected for many of the studies shown based on the optimal and reproducible cytokine expression in response to anti-CD3 mAb. Non-adherent cells were decanted from the flasks and centrifuged to obtain supernatants. The flasks were then treated with Enzyme-Free cell disassociation solution (Specialty Media, Phillipsburg, NJ) to remove the adherent cells (a typical result of cell activation and the anti-CD3 coated flasks) and were gently scraped to remove and harvest the cells. Viable cells from the decanted cells and cell removal mixture were isolated by Ficoll Paque density gradient centrifugation (as described above). Cell viability was not significantly affected by this enzyme treatment process (viability >95%).

It should be noted that we have also utilized the serum free medium AIMV in these various cultures and observed similar effects to serum containing medium (data not shown).

ELISA (Biosource) and Bio-Plex Human Cytokine 17-Plex (Biorad, Hercules, CA) were utilized to examine the following human-specific cytokines: IFN-α, IFN-γ, IL-1α, IL-1β, IL-4, IL-5, IL-6, IL-7, IL-8, IL-10, IL-12 p70, IL-13, IL-15, IL-17, G-CSF, GM-CSF, MIP-1∀, MIP-1∃, RANTES, MCP-1 and TNF-α. The IL-18 ELISA was obtained from R & D Systems (Minneapolis, MN). All of the ELISAs and multiplex assays were performed according to the manufacturer's instructions. The results are expressed as pg/ml or ng/ml and all assays were run in duplicate with at least three separate experiments being examined.

Cytoplasmic RNA was extracted and purified using a commercially available kit (RNAeasy, Qiagen, Valencia, CA). Purified RNA was electrophoresed on a 1% agarose gel to assess the integrity of the purified RNA. One :g of RNA was reverse transcribed into cDNA using a commercial available kit (Applied Biosystems, Foster City, CA). One hundred pg RNA equivalent of this cDNA was used for PCR amplification. PCR reactions were performed in special optical tubes in a 96 well microtiter plate format on an ABI PRISM 7700 Sequence Detector System (PE Applied Biosystems) using pre-developed FAM- and TAMRA-labeled internal oligonucleotide probes and primers for IFN-γ, IL-4, IL-5, IL-10, IL-12p30, IL-12p40, IL-15, and TNF-α (PE Applied Biosystems). Each reagent also contains VIC- and TAMRA-labeled internal oligonucleotide probes and primers specific for the 18S RNA ribosomal subunit. Amplification conditions were as follows 25°C for two min; 95°C for 10 min; 40 cycles of 95°C 15 s and 60°C for 1 min. Fluorescence signals measured during amplification were processed post-amplification and were regarded as positive if the fluorescence intensity was ten fold greater than the standard deviation of the baseline fluorescence. This level is defined as the threshold cycle (Ct). The Ct value for 18S ribosomal subunit was subtracted from the Ct value for each cytokine message to normalize for RNA content. This value is defined as ΔCT. To evaluate the effects of retinoids, ΔCT_treatment was subtracted from ΔCt_control. This value is defined as ΔΔCT. The relative folds increase or decrease was then calculated as 2 ^-ΔΔCT.

The PCR was also set up using SYBR green Master Mix (Applied Biosystems), 1 μl cDNA and gene-specific primers at a final concentration of 0.3 μM. Thermal cycling was carried out on the Applied Biosystems GeneAmp 7700 Sequence Detector and SYBR green dye intensity was analyzed using GeneAmp 7700 SDS software. Primers for human GATA3, STAT6, C-MAF, and T-bet genes and glyceraldehyde-3-phosphate dehydrogenase (GAPDH) as control were designed using ABI prism software (PE Applied Biosystems). Primers are available upon request. Similar amplification procedures and data computation were followed as described above. No PCR products were generated from genomic versus cDNA template.

Mouse anti-human CD3-Cy-Chrome^®, anti-human IL-4-phycoerytrin (PE), anti-human IFN-(-fluorescein isothiocyanate (FITC) anti-human TNF-∀-FITC, and the appropriately labeled isotype control mAbs were obtained from Pharmingen. PBMCs were cultured for 36 hrs as described above after which the cells were treated with 2 :M monensin (Sigma) and subsequently harvested as described above. Cells (0.25 × 10⁶) were suspended in 50 :L of staining buffer (1% FCS, 1% goat serum, 2.5 :g of mouse IgG/50 μL) in round-bottom 96-well plates at 4°C for 15 min. 5 :L of the appropriate dilution of each antibody was then added. Cells were then incubated for 30–40 min at 4°C. Cells were pelleted and medium was aspirated carefully. Cells were then washed with 100 :L of PBS/FBS buffer twice. Cells were fixed and permeabilized with Cytofix/Cytoperm^® solution (Pharmingen). Various combinations of labeled anti-human cytokines were used to stain cells. Cytochrome-conjugated mouse IgG₁ mAb and PE mouse-IgG were used as isotype controls at the same concentrations as the anti-cytokine antibody. Additional controls in which the labeled mAbs and 10 fold saturating recombinant cytokine proteins were pre-incubated for 30 min at room temperature before staining (IL-4). Alternatively, cells were pretreated with unlabeled mAb (IFN-γ). Three-color cytofluorometry was performed using a FacScan (Becton Dickinson, San Diego, CA). A minimum of 10,000 CD3⁺ cells were analyzed in these experiments. Data are expressed as the % of CD3⁺ cells expressing the marker of interest or the mean channel number (MCN) of the marker's fluorescent intensity.

The ELISPOT assays (BD Biosciences) used to quantify IFN-γ-, IL-4, and IL-5-producing T cells were performed according to the manufacturer's instructions. Briefly, T cells were prepared at different cell densities ranging from 1 × 10⁵, 1 × 10⁶, and 2.5 × 10⁶ cells per ml and 100 :l of the suspensions were added to each well of mouse anti-human cytokine antibody-coated BD ELISPOT plates. The cells were stimulated using ATRA or 9-cis-RA and/or soluble anti-CD3 and CD28 antibodies described above under cell culture. The plates were then incubated at 37°C in a 5% CO2 humidified incubator for 24 hr after which the cell suspension were aspirated, the wells were washed with various combinations of deionized water and wash buffer, and subsequently developed using the proper antibody, conjugate and substrate pairs defined by the manufacturer. The plates were air-dried overnight at room temperature and the plates were stored in a sealed bag in the dark until analyzed. Spots were then enumerated using an ImmunoSpot^® Series 2 Analyzer (Cellular Technology Limited, Cleveland, Ohio) and the supporting ImmunoSpot^® Software. Spots were counted by an automated system using a defined set of parameters for size, intensity, and gradient. The background (the mean numbers of spots in wells without stimulation) was subtracted from each well on each cytokine plate. A response was considered positive if the average number of cytokine-producing cells (CPC) per triplicate wells exceeded background +/- 2SD. The data are shown as the average number of CPC per 10⁶ cells.

Data were analyzed for equality of variance using Fischer's F test (Statview 5.0 for Macintosh, Abacus Concepts, Berkeley, CA). If the variance was heterogeneous, the appropriate transformation of the data was performed. A two-tailed paired T test was then used to determine statistical significance. A P < 0.05 was considered statistically significant for all analysis.