Dose-Dependent Glucocorticoid Regulation of Transcription Factors in Vocal Fold Fibroblasts and Macrophages

Abstract

Objective.

Variable outcomes of glucocorticoid (GC) therapy for laryngeal disease are putatively due to diverse interactions of the GC receptor (GR) with cell signaling pathways, limited consideration regarding concentration-dependent effects, and inconsistent selection of GCs. In the current study, we evaluated concentration-dependent effects of three frequently administered GCs on transcription factors with an emphasis on phosphorylation of GR at Ser203 and Ser211 regulating nuclear translocation of GR. This study provides foundational data regarding the diverse functions of GCs to optimize therapeutic approaches.

Study design:

In vitro

Methods:

Human vocal fold fibroblasts and THP1-derived macrophages were treated with different concentrations of dexamethasone, methylprednisolone, and triamcinolone in combination with IFN-γ, TNF-α, or IL4. Phosphorylated STAT1, NF-κB family molecules, and phosphorylated STAT6 were analyzed by Western blotting. Ser211-phosphorylated GR (S211-pGR) levels relative to GAPDH and Ser203-phosphorylated GR (S203-pGR) were also analyzed.

Results:

GCs differentially altered phosphorylated STAT1 and NF-κB family molecules in different cell types under IFN-γ and TNF-α stimuli. GCs did not alter phosphorylated STAT6 in IL4-treated macrophages. The three GCs were nearly equivalent. A lower concentration of dexamethasone increased S211-pGR/GAPDH ratios relative to increased S211-pGR/S203-pGR ratios regardless of cell type and treatment.

Conclusion:

The three GCs employed in two cell lines had nearly equivalent effects on transcription factor regulation. Relatively high levels of Ser203-phosphorylation at low GC concentrations may be related to concentration-dependent differential effects of GCs in the two cell lines.

Lay Summary.

Steroid-based therapies are associated with variable outcomes for laryngeal disease. This study provides incremental data related to potential mechanisms of this variability including divergent effects on the glucocorticoid receptor phosphorylation.

INTRODUCTION

Annual healthcare costs associated with dysphonia are estimated at $13 billion.¹ Glucocorticoids (GCs) are broadly administered for laryngeal disease in various formulations.^2–4 Affordability and access likely contribute to pervasive GC use. However, reported clinical outcomes are disparate across laryngeal disease and GCs.⁵ For example, triamcinolone acetonide injection had a relatively unclear therapeutic effect on vocal scar,⁶ compared to dexamethasone.⁷ Risks of GC injection on muscular and mucosal atrophy have also been reported,^8,9 and more recent data suggest a pro-fibrotic response to GCs in the larynx.¹⁰ In that regard, GCs have divergent effects on fibrosis across organs^11,12,13,14 In vitro data from our group confirmed both anti- and pro-fibrotic effects of GCs in human vocal fold (VF) fibroblasts and macrophages.^10,15 Despite their primary anti-inflammatory role, GCs increased several inflammatory genes in macrophages.¹⁵ We hypothesize complex functions of GCs underlie disparate clinical outcomes for VF disease.⁵ Ideally, elucidating this complexity will optimize GC therapies by enhanced favorable effects of GCs while limiting the unfavorable effects.

The GC receptor (GR) is activated via GC binding and interacts with numerous proteins and pathways.¹⁶ Activated GR regulates proteins both in the nucleus and cytoplasm. In the cytoplasm, GR binds to and inhibits transcription factors (TFs), such as nuclear factor-κB (NF-κB), a primary mediator of inflammation.¹⁷ In the nucleus as a TF, GR binds to GR-responsive elements (GREs) to induce transcription. Alternatively, recruitment of GR to GREs is accompanied with other TFs and co-activators to support transcription. Negative regulation is mediated through GR binding to negative gene regulatory sequences called a negative GRE (nGRE).¹⁸ Due to diverse GR functions, GCs elicit complex responses across cell types and cellular contexts in which signaling pathways are differentially activated.

In injured tissue, fibroblasts and macrophages participate in inflammation and subsequent tissue repair. In response to inflammatory stimuli, VF fibroblasts and macrophages secrete inflammatory cytokines such as tumor necrosis factor-α (TNF-α), interleukin-1β (IL1β), and interferon-γ (IFN-γ); an inflammatory macrophage phenotype (M1) is induced.^19,20 C-X-C motif chemokine ligand 10 and 11 (CXCL10, CXCL11), which induces leukocyte infiltration, are also upregulated by inflammatory stimuli. Following the inflammatory phase, fibroblasts synthesize and reorganize extracellular matrix (ECM). Macrophages typically shift into an anti-inflammatory/pro-fibrotic phenotype (M2) and support ECM formation via paracrine signaling and expression of ECM-crosslinking enzymes such as lysyl oxidase (LOX) and transglutaminase-2 (TGM2).^20,21 However, excess accumulation of ECM leads to fibrosis.^19,22

We recently reported differential effects of GCs on fibrotic and inflammatory gene expression in human VF fibroblasts and macrophages dependent on cell type, cytokine stimuli, and GC concentration.¹⁵ Inflammatory genes such as TNF, IL1B, and PTGS2, were downregulated by GCs in IFN-γ- and TNF-α-stimulated cells. Accordingly, CXCL10 and CXCL11 expression were inhibited by GCs in TNF-α-treated VF fibroblasts. However, IFN-γ-treated VF fibroblasts were partially resistant to inhibition, and GCs upregulated CXCL10 and CXCL11 in IFN-γ-treated macrophages. Regarding fibrotic genes, LOX and TGM2 expression stimulated by interleukin-4 (IL4), a potent inducer of the M2 phenotype, was inhibited by GCs in macrophages. In contrast, connective tissue growth factor (CTGF), a mediator of fibrosis,²³ was upregulated by GCs in macrophages and VF fibroblasts under TNF-α, IFN-γ, and IL4 stimulation. Interestingly, higher concentrations of GCs were required to upregulate CTGF, as well as a major transcription target of GR (glucocorticoid-induced leucine zipper; GILZ) relative to concentrations required to inhibit inflammatory genes (CXCL10, CXCL11, TNF, PTGS2, IL1B) and other fibrotic genes (LOX, TGM2).¹⁵

STAT1 and NF-κB are major TFs stimulated by IFN-γ and TNF-α.^24,25 IFN-γ binding to IFN-γ receptor mediates activation of Janus kinase 1 and 2 (JAK1 and 2) and subsequent phosphorylation of STAT1 at Tyr701. Phosphorylation induces STAT1-target genes including CXCL10 and CXCL11.²⁵ NF-κB signaling is canonically stimulated through the activation of the TNF-α receptor.²⁶ Downstream, p50 NF-κB is truncated from a p105 precursor protein,²⁷ and dimers of NF-κB family molecules (predominantly p65/p50 and c-Rel/p50 dimers) are released from the inhibitor of NF-κB (IκB), resulting in induction of various inflammatory genes.²⁸ In addition, due to the supportive functions of NF-κB to the JAK/STAT1 pathway, IFN-γ and TNF-α synergistically induce inflammatory responses.^29,30 The primary TF activated by IL4 is STAT6.^31,32 IL4 induces cross phosphorylation between IL4 receptor-α and associated proteins. In turn, STAT6 is phosphorylated at Tyr641 and enters the nucleus to induce transcription, including an M2 macrophage marker (CCL17). ^33,34 Given the diversity of GR function across signaling pathways related to both cell type and exogenous stimuli, investigation regarding cell type- and context-dependent effects of GCs is critical to optimize GC-based therapeutic approaches.

Phosphorylation is critical for the localization of GR within cells. Ser211 and Ser203 are particularly relevant phosphorylation sites for GR cellular distribution. Ser211 is phosphorylated by ligand activation of GR,^35,36 leading to conformational changes and nuclear accumulation of GR.³⁶ In contrast, Ser203 phosphorylation restricts nuclear translocation of GR, and GR phosphorylated at both Ser203 and Ser211 is predominantly distributed in the cytoplasm.^23,35 Nuclear concentration is important for dimerization of GR, a requirement for GRE-dependent transcription.^18,37,38 A lower GR concentration is adequate to bind to nGREs, which accepts GR monomers. We hypothesize the Ser211- and Ser203-phosphorylation ratio is critical to nuclear concentration of GR, and downstream, transcriptional regulation. The ratio of Ser203/Ser211 phosphorylation has not been extensively investigated, particularly in response to GC concentration.

The current study seeks to provide foundational data regarding differential effects of GCs on TFs across cell types and cytokine stimuli as well as GC and GC concentration. Human VF fibroblasts and macrophages were treated with different concentrations of three GCs commonly administered to the larynx under IFN-γ, TNF-α, and IL4 stimuli. Altered TFs stimulated by these conditions were analyzed. In addition, we interrogated relationships between GC concentration-dependent regulation of Ser203/Ser211 phosphorylation and transcription. GC concentrations required to increase Ser211-phosphorylated GR (S211-pGR) levels relative to a housekeeping gene and Ser203-phosphorylated GR (S203-pGR) were compared to previously reported GC concentrations required to inhibit and promote inflammatory and fibrotic gene expression.¹⁵

MATERIALS AND METHODS

Detailed information is available in the Supporting Information.

Cell culture.

An immortalized human VF fibroblast cell line, HVOX,³⁹ and macrophages derived from THP-1 were treated with 10 ng/mL TNF-α, 100 ng/mL IFN-γ, or 40 ng/mL IL-4 +/− various concentrations of GCs for 24 hours.

Western blotting.

Cell lysates were subjected to Western blotting as described previously.⁴⁰ Primary/secondary antibodies listed in Table S1 were employed. Intensities of Western blots were quantified using ImageJ.

Data analysis.

All experiments were repeated in technical triplicate, at least. Determination of EC₅₀ and EC₉₀, and illustration of concentration-response curves were performed for each technical replicate using R drc.⁴¹

RESULTS

GCs differentially altered transcriptional factors under cytokine stimuli.

STAT1 and NF-κB signaling are typically activated by IFN-γ and TNF-α and have synergistic effects across cell types.^25,28 Upon activation, Tyr701-phosphorylated STAT1 (Y701-pSTAT1) and NF-κB increase.⁴² Western blotting was employed to interrogate the effects of GCs on phosphorylation of STAT1 and p65 and p50 of NF-κB in IFN-γ- and TNF-α-treated human VF fibroblasts and macrophages (Figures 1 and 2). Because genetic/epigenetic backgrounds can impact the response to GCs, a single cell line for each cell type was employed to simplify concentration-dependent effects of GCs. In independently performed and technically triplicated experiments, IFN-γ increased Y701-pSTAT1 and total STAT1 (tSTAT1) as well as p50 and p65 NF-κB in HVOX fibroblasts (Figure 1A–C, Supplemental Figure S1A). Although tSTAT1 decreased in response to all three GCs, Y701-STAT1 appeared stable. GCs had no clear effects on p65 or p50 NF-κB levels. Similar to IFN-γ, TNF-α increased Y701-pSTAT1, tSTAT1, and p65 and p50 NF-κB in VF fibroblasts (Figure 1I–K, Supplemental Figure S1B). However, these molecules decreased in response to GCs in TNF-α-stimulated fibroblasts, in contrast to IFN-γ-stimulated cells. IFN-γ increased Y701-pSTAT1 and tSTAT1 in macrophages, but had no clear effects on p65 and p50 NF-κB; the three GCs decreased tSTAT1 (Figure 2A–C, Supplemental Figure S2A). GCs did not clearly alter p65 and p50 NF-κB in IFN-γ-treated macrophages. TNF-α did not increase Y701-pSTAT1 or p65 NF-κB (Figure 2I–K, Supplemental Figure S2B), but increased p50 NF-κB in macrophages. This response was attenuated by GCs.

Figure 1. Concentration-dependent effects of GCs on TFs in IFN-γ- and TNF-α-treated human VF fibroblasts.

HVOX fibroblasts were cultured +/− 100 ng/mL IFN-γ (A-H) and +/− 10 ng/mL TNF-α (I-P) with 0-1,000 nM dexamethasone (Dex; A, D-I, L-P), 0-10,000 nM methylprednisolone (Met; B, J), and 0-10,000 nM triamcinolone (Tri; C, K) for 24 hours. Representative Western blots for TFs and GAPDH (A-C, I-K). Plots of S211-pGR/GAPDH, S211-pGR/tGR, S203-pGR/GAPDH, and S203-pGR/tGR ratios in dexamethasone-treated cells (D-H). Protein levels in the control, without either cytokine or GCs, were determined as ‘1’ (black dashed lines), and relative levels are plotted as color dots. Smoothed curves (D-F) and concentration-dependent sigmoid curves (H) are illustrated as colored dashed lines and solid lines. Data from independently-performed technical replicates are respectively shown in different colors.

Figure 2. Concentration-dependent effects of GCs on TFs in IFN-γ- and TNF-α-treated human macrophages.

THP1-derived macrophages cultured +/− 100 ng/mL IFN-γ (A-H) and +/− 10 ng/mL TNF-α (I-P) with 0-1,000 nM dexamethasone (Dex; A, D-I, L-P), 0-10,000 nM methylprednisolone (Met; B, J), and 0-10,000 nM triamcinolone (Tri; C, K) for 24 hours. Representative Western blots for TFs and GAPDH (A-C, I-K). Plots of S211-pGR/GAPDH, S211-pGR/tGR, S203-pGR/GAPDH, and S203-pGR/tGR ratios in dexamethasone-treated cells (D-H). Protein levels in the control, without either cytokine or GCs, were determined as ‘1’ (black dashed lines), and relative levels are plotted as color dots. Smoothed curves (D-F) and concentration-dependent sigmoid curves (H) are illustrated as colored dashed lines and solid lines. Data from independently-performed technical replicates are respectively shown in different colors.

Phosphorylation of STAT6 in IL4-treated macrophages was investigated. Since IL4 did not stimulate fibrotic genes (ACTA2, Col1A1, TGM2, LOX) in VF fibroblast,¹⁵ unlike in pulmonary and hepatic fibroblasts,^43,44 VF fibroblasts were not treated with IL4. IL4 increased Tyr641-phosphorylated STAT6 (Y641-pSTAT6) in macrophages; this response was not altered by GCs (Figure 3A–C, Supplemental Figure S3). Overall, the three GCs were nearly equivalent in their effect on TFs.

Figure 3. Concentration-dependent effects of GCs on TFs in IL4-treated human macrophages.

THP1-derived macrophages were collected after treatment with 0-1,000 nM dexamethasone (Dex; A, D-I), 0-10,000 nM methylprednisolone (Met; B), and 0-10,000 nM triamcinolone (Tri; C) +/− 40 ng/mL IL4 for 24 hours. Representative Western blots for TFs and GAPDH (A-C). Plots of S211-pGR/GAPDH, S211-pGR/tGR, S203-pGR/GAPDH, and S203-pGR/tGR ratios in dexamethasone-treated cells (D-H). The protein levels in the control, without either IL4 or GCs, were determined as ‘1’ (black dashed lines), and relative levels are plotted as color dots. Smoothed curves (D-F) and concentration-dependent sigmoid curves (H) are illustrated as colored dashed lines and solid lines. Data from independently-performed technical replicates are respectively shown in different colors.

Dexamethasone differentially altered phosphorylation of GR at Ser211 and Ser203.

The ratio of S211-pGR normalized to GAPDH was elevated with increased concentrations of dexamethasone up to 1~10nM, and decreased by higher concentrations of dexamethasone regardless of cell types or cytokine stimuli in triplicate experiments (Figure 1A, 1D, 1I, 1L, 2A, 2D, 2I, 2L and 3A, 3D, and Table 1). Due to decreased total GR (tGR) levels with increased concentrations of dexamethasone, the S211-pGR/tGR ratio increased in a concentration-dependent manner (Figure 1E, 2M, 2E, 2M, and 3E). Effects of dexamethasone on S203-pGR/GAPDH and S203-pGR/tGR ratios were unclear at low concentrations in VF fibroblasts and macrophages regardless of cytokine stimulation (Figure 1F, 1G, 1N, 1O, 2F, 2G, 2N, 2O, 3F, 3G). However, higher concentrations of dexamethasone decreased S203-pGR/GAPDH and Ser203-pGR/tGR ratios. Altered S211-pGR/S203-pGR ratios fit to concentration-responsive sigmoid curves and EC₉₀ to increase S211-pGR/S203-pGR ratios ranged from 64-103 nM (Figure 1H, 1P, 2H, 2P, 3H and Table 1).

Table 1:

Dexamethasone concentrations required to increase S211-pGR/S203-pGR and S211-pGR/GAPDH ratios.

	VF fibroblast		Macrophage
	IFN-γ	TNF-α	IFN-γ	TNF-α	IL4
S2211/GAPDH ratio
Peaked concentration (nM)*	1~10	3~10	3~10	3~10	1~10
S211-pGR/S203-pGR ratio
EC50 (nM)**	6.50±4.15	3.11±4.1.76	5.09±2.62	3.89±1.73	3.87±2.17
EC90 (nM)**	78.76±40.69	64.60±33.91	100.13±49.31	103.42±40.17	85.70±33.81

^*Ranges of concentrations in triplicate independent experiments

^**Mean ± SD from triplicate independent experiments

DISCUSSION

Investigation of multiple mechanisms underlying the diverse functions of GCs is critical to optimize GC-based therapies. Our prior work using HVOX and THP-1 cells elucidated different GC functions dependent on cell type, cytokine stimuli, and GC concentrations.¹⁵ The current study sought to provide further insight regarding molecular mechanisms related to these differential GC functions. TFs and TF phosphorylation were differentially altered in the current study, likely reflective of the diversity of GC effects on gene expression. These data suggest increased GC concentrations were required to increase the S211-pGR/S203-pGR ratio compared to the S211-pGR/GAPDH ratio, regardless of cell type or cytokine stimuli. Although obtained in only two cell lines and need to be confirmed with biological replicates, these data likely imply a possible mechanism underlying dose dependent GC-mediated gene regulation.

STAT1 and NF-κB are major TFs, stimulated by IFN-γ and TNF-α, and synergistically stimulate inflammatory gene expression.²⁵ Activated GR can affect STAT1 and NF-κB signaling pathways at multiple levels, including regulation of degradation, sequestration from the nucleus, and/or occupation of response elements in the genome.^18,45 Regarding degradation, activated GR induces STAT1 degradation mediated by suppressor of cytokine signaling 1.^46–48 GR prevents NF-κB from translocating to the nucleus and leads to proteasomal degradation of NF-κB.^18,45 In TNF-α-stimulated VF fibroblasts, GCs decreased p65 NF-κB and Y701-pSTAT1, suggesting degradation was mediated through activated GR. However, GCs had no clear effect on those molecules in IFN-γ-stimulated fibroblasts as well as IFN-γ- and TNF-α-stimulated macrophages. GCs also decreased p50 NF-κB, which supports p65 to induce gene expression,^49,50 in TNF-α-stimulated macrophages suggesting an alternative pathway to suppress NF-κB-dependent gene expression.

Divergent data regarding anti-inflammatory mechanisms of GCs have been recently reported. In Crohn’s disease, for example, GCs increased IκBα, an endogenous inhibitor of NF-κB, in vascular endothelial cells, but not in mononuclear cells.⁵¹ Dexamethasone supports STAT1-dependent gene expression in human peripheral blood monocytes by recruiting PU.1 TF to IFN-γ-activated sites.⁵² In the current study, GCs had no clear effect on Y701-pSTAT1 in either fibroblasts or macrophages under IFN-γ stimulation. GCs had unclear effects on, or even enhanced, transcription of a STAT1 target gene (CXCL11) in VF fibroblasts and macrophages under IFN-γ stimulation.¹⁵ Sustained Y701-pSTAT1 in IFN-γ-treated cells may be partially related to failure to inhibit transcription via STAT1. PU.1 recruitment may be mechanistic for GC-induced CXCL11 upregulation in macrophages, as previously observed in human peripheral blood monocytes.⁵²

GCs had contrasting effects on CCL17 and ECM cross-linking enzymes expression in IL4-treated macrophages.¹⁵ IL4 receptor signaling induces CCL17 through the JAK/STAT6 pathway^33,34 and induces TGM2 and LOX through p38 MAPK and PI3K/AKT activation.^53,54 These mechanistic differences in gene regulation likely underlie contrasting effects of GCs. However, interactions between GCs and IL4 signaling are poorly described. GCs did not clearly alter phosphorylation of STAT6 in macrophages, suggesting GR supported STAT6-dependent transcription without affecting upstream events in the JAK/STAT6 pathway. GR may mediate recruitment of STAT6 or a co-activator to corresponding gene promoter sites; further investigation is required to clarify this mechanism. GCs reportedly inhibit p38 MAPK and PI3K/AKT signaling pathways activated by non-IL4 stimulation in various cells.^53,54

Phosphorylation at Ser203 and Ser211, respectively, induce cytoplasmic and nuclear localization of GR.^18,35,55 Dogma suggests elevated GR concentration in the nucleus increases GR dimerization, which is critical for GR binding to GREs to drive transcription.^18,37,38 Data from the current study appear consistent with this hypothesis; the EC₉₀ for CTGF and GILZ gene upregulation (58-99 nM)¹⁵ were comparable to the EC₉₀ to increase the S211-pGR/S203-pGR ratio (64-104 nM) rather than the S211-pGR/GAPDH ratio (1-10 nM). In contrast, dimerization is not required for GR binding to negative gene regulatory sites in the genome (nGREs, GRE half sites, and κBRE),¹⁸ implying that a lower level of nuclear concentration is adequate for GR to inhibit transcription than to induce GRE-dependent transcription. Additionally, nuclear translocation is not necessary for cytoplasmic inhibitory mechanisms of GR.^18,37 Concentrations of dexamethasone at the peaks of S211-pGR/GAPDH ratio (1-10 nM) were similar to the IC₉₀ to downregulate inflammatory and ECM cross-linking enzyme genes (0.3-12 nM).¹⁵ Based on these findings, a reasonable hypothesis might be that low GC concentrations partially activate GR with limited nuclear translocation contributing to reduced GR-mediated sequestration of other TFs in the cytoplasm and occupation of negative gene regulatory sites in the nucleus.

Dexamethasone, methylprednisolone, and triamcinolone have no mineral corticoid activity.⁵⁶ These compounds were developed to induce GR activation. Accordingly, these GCs had nearly equivalent effects on TFs and gene expression in the current and previous studies, except disparities between EC₉₀ and IC₉₀ for gene regulation are different across the three GCs.¹⁵ Penetration into cells, diffusion speed, and inactivation rates must also be considered, and these three commonly employed GCs have different in vivo half-lives.⁵⁶ Additional pharmacokinetic studies are required to detail GC-dependent outcomes.

In spite of these mechanistic data, elucidating the numerous activities associated with GC signaling is challenging. The current study is not without significant limitations. Protein and phosphorylation levels of several major TFs were investigated, but the current study stopped short of interrogating the influences of GC signaling on multiple alternative process such as distribution of TFs and upstream molecules as well as regulation of degradation and modification of chromatin accessibility. In addition, with regard to the model, clinical translation could be enhanced by investigating the effects of GCs on cells already stimulated by cytokines; GC therapy is typically implemented after inflammation is observed clinically. Temporal dynamics were also not included in the current study; cells likely differentially respond to duration of GC treatment. Similarly, genetic/epigenetic background, one of the most likely factors altering outcomes of GC treatment, was also not investigated in the current study. The most significant limitation, however, is the lack of biological replicates. Responses to GCs observed in HVOX and THP-1 cells might be specific in these cell lines and further analyses using multiple cell lines with different genetic/epigenetic background are required to confirm the current findings and/or to characterize genetic/epigenetic differences that significantly alter responses to GCs.

CONCLUSION

TFs were differentially affected by GCs across cell types and cytokine stimuli. These data increased the knowledge regarding the complexity of GC function in regulating cell signaling. However, higher concentrations of dexamethasone were required to increase S211-pGR/S203-pGR ratios than S211-pGR/GAPDH ratios in the two cell lines employed in this study. This finding is likely related to the disparity in dexamethasone concentrations required for gene regulation in these cell lines.

Supplementary Material

supinfo1

Supplemental Figure S1. Concentration-dependent effects of GCs on transcription factors in IFN-γ- and TNF-α-treated human VF fibroblasts. HVOX fibroblasts were cultured +/−100 ng/mL IFN-γ (A) and +/−10 ng/mL TNF-α (B) with 0-1,000 nM dexamethasone (Dex), 0-10,000 nM methylprednisolone (Met), and 0-10,000 nM triamcinolone (Tri) for 24 hours. Western blots for transcription factors and GAPDH were quantified. Protein levels in the control, without cytokines or GCs, were determined as ‘1’ (black dashed lines), unless blots were undetectable. Protein levels relative to the control or cells treated only with cytokines are plotted as color dots. Smoothed curves are illustrated as colored dashed lines. Data from independently-performed technical replicates are shown in different colors. ND: Not determined, due to undetectable protein levels.

Supplemental Figure S2. Concentration-dependent effects of GCs on transcription factors in IFN-γ- and TNF-α-treated human macrophages. THP1-derived macrophages were cultured +/−100 ng/mL IFN-γ (A) and +/−10 ng/mL TNF-α (B) with 0-1,000 nM dexamethasone (Dex), 0-10,000 nM methylprednisolone (Met), and 0-10,000 nM triamcinolone (Tri) for 24 hours. Western blots for transcription factors and GAPDH were quantified. Protein levels in the control, without cytokines or GCs, were determined as ‘1’ (black dashed lines), unless blots were undetectable. Protein levels relative to the control or cells treated only with cytokines are plotted as color dots. Smoothed curves are illustrated as colored dashed lines. Data from independently-performed technical replicates are shown in different colors. ND: Not determined, due to undetectable protein levels.

Supplemental Figure 3. Concentration-dependent effects of GCs on transcription factors in IL4-treated human macrophages. THP1-derived macrophages were collected after treatment with 0-1,000 nM dexamethasone (Dex), 0-10,000 nM methylprednisolone (Met), and 0-10,000 nM triamcinolone (Tri) +/−40 ng/mL IL4 for 24 hours. Western blots for transcription factors and GAPDH were quantified. Protein levels in the control, without IL4 or GCs, were determined as ‘1’ (black dashed lines), unless the blots were undetectable. Protein levels relative to the control or cells treated only with cytokines are plotted as color dots. Smoothed curves are illustrated as colored dashed lines. Data from independently-performed technical replicates are shown in different colors. ND: Not determined, due to undetectable protein levels.