Glucocorticoid Dose Dependency on Gene Expression in Vocal Fold Fibroblasts and Macrophages

Abstract

Objective:

Glucocorticoids (GCs) modulate multiple cellular activities including inflammatory and fibrotic responses. Outcomes of GC treatment for laryngeal disease vary, affording opportunity to optimize treatment. In the current study, three clinically employed glucocorticoids were evaluated to identify optimal in vitro concentrations at which GCs mediate favorable anti-inflammatory and fibrotic effects in multiple cell types. We hypothesize a therapeutic window will emerge as a foundation for optimized therapeutic strategies for patients with laryngeal disease.

Study design:

In vitro

Methods:

Human vocal fold fibroblasts and human macrophages derived from THP-1 monocytes were treated with 0.03–1,000 nM dexamethasone, 0.3–10,000 nM methylprednisolone, and 0.3–10,000 nM triamcinolone in combination with interferon-γ, tumor necrosis factor-α, or interleukin-4. Real-time polymerase chain reaction was performed to analyze inflammatory (CXCL10, CXCl11, PTGS2, TNF, IL1B) and fibrotic (CCN2, LOX, TGM2) genes, and TSC22D3, a target gene of GC signaling. EC₅₀ and IC₅₀ to alter inflammatory and fibrotic gene expression was calculated.

Results:

Interferon-γ and tumor necrosis factor-α increased inflammatory gene expression in both cell types; this response was reduced by GCs. Interleukin-4 increased LOX and TGM2 expression in macrophages; this response was also reduced by GCs. GCs induced TSC22D3 and CCN2 expression independent of cytokine treatment. EC₅₀ for each GC to upregulate CCN2 was higher than the IC₅₀ to downregulate other genes.

Conclusion:

Lower concentrations of GCs repressed inflammatory gene expression and only moderately induced genes involved in fibrosis. These data warrant consideration as a foundation for optimized clinical care paradigms to reduce inflammation and mitigate fibrosis.

INTRODUCTION

Healthcare costs associated with voice disorders are estimated at nearly 13 billion dollars annually in the United States.¹ Glucocorticoids (GCs) are affordable and widely administered in various formulations across the spectrum of laryngeal disease.^2–4 Due to their potent anti-inflammatory actions, GCs are primarily used to reduce laryngeal inflammation.⁵ However, clinical outcomes differ across types of laryngeal diseases and GCs.⁶ Less favorable outcomes have been reported for triamcinolone acetonide injection for Reinke’s edema and vocal scar relative to other benign VF lesions.⁷ These data are inconsistent outcomes of dexamethasone injection which had a clear therapeutic effect on scarred vocal folds.⁸ Multiple factors including occupational vocal demands and laryngopharyngeal reflux have been posited to affect the therapeutic effects of GCs.⁹ GCs, however, vary substantively with regard to critical properties (e.g., potency, half-life, etc.), potentially underlying variability in efficacy. Based on the diversity of GCs employed across published reports, it appears these drugs are considered interchangeable with equivalent effectiveness. In addition, the efficacy of GCs for fibrosis across tissues is controversial. For example, GCs inhibited fibrosis in mouse liver¹² and airway.¹³ However, fibrosis of the murine lung was exacerbated by methylprednisolone.¹⁴ Similarly, dexamethasone induced renal fibrosis¹⁵ and ~50% of keloid scars are GC resistant.¹⁶ These disparate outcomes across tissues may be related to differential effects of GCs based on cell type and GC.^6,17 In addition, our recent work suggested both anti- and pro-fibrotic responses in a single cell type.¹⁷ The inherent complexity of GCs and cell-level responses, putatively, are the source of variability in clinical outcomes for fibrosis of the vocal folds.⁶ Broadly, we hypothesize optimization of glucocorticoid therapy will substantively improve clinical outcomes.

The GC receptor (GR) interactome consists of numerous proteins involved in various pathways.¹⁸ Following activation via GC binding, GR enters the nucleus to regulate transcription, both positively and negatively, through interactions with co-activators and co-repressors.¹⁹ GR activates transcription by binding to GR-responsive elements (GREs) with other transcription factors and co-activators. In contrast, GR represses gene transcription by binding negative gene regulatory sequences called a negative GRE (nGRE), or by binding to proteins such as nuclear factor-κB (NF-κB) to repress NF-κB-mediated transcription. NF-κB is a major transcription factor activated by inflammatory stimuli to modulate pro-fibrotic pathways such as SMAD, Wnt, and Hippo.^19–23 Specific pathways, however, are preferentially affected by GR. We recently reported increased dexamethasone concentrations were required to synergistically enhance the fibrotic effects of transforming growth factor-β (TGF-β) relative to inhibition of inflammatory genes in human VF fibroblasts.¹⁷ Dose-dependent differential effects of GCs have also been reported in neuronal cells.²⁴ Beyond these preliminary data, little information exists to direct optimal clinical care paradigms involving GCs.

Fibroblasts primarily maintain extracellular (ECM) matrix through synthesis and orientation of proteins.²⁵ Following injury, fibroblasts differentiate into myofibroblasts with contractile properties critical for wound closure; hyperactivation of myoblasts, however, is associated with fibrosis.^26,27 In addition to ECM management, fibroblasts/myofibroblasts also secrete inflammatory cytokines and enzymes under inflammatory stimuli such as tumor necrosis factor-α (TNF-α), interleukin-1β (IL1β), and/or interferon-γ (IFN-γ).²⁷ Macrophages are large phagocytic cells mobilized in response to injury to participate in both inflammation and ECM reorganization.²⁸ Macrophages alter phenotypes under different circumstances; TNF-α, IFN-γ, lipopolysaccaride, and other inflammatory stimuli leads to a conventional M1 macrophage phenotype. Macrophages are alternatively activated to secrete anti-inflammatory cytokines and support ECM formation.²⁸ IL4 and IL13, potent inducers of the alternatively activated M2 phenotype, contribute to a Th2 immune response.²⁸ In addition, IL4 and IL13 upregulate ECM cross-linking enzymes including TGM2 and LOX (encoding transglutaminase-2 and lysyl oxidase) in human macrophages.^29,30 Given the role of both macrophages and fibroblasts in the development and maintenance of fibrosis, therapeutic strategies that favorably impact multiple cell types are likely to yield favorable outcomes.

In the current study, we hypothesized different concentrations of three commonly employed GCs-dexamethasone, methylprednisolone, and triamcinolone-differentially affect inflammatory and fibrotic gene expression in macrophages and fibroblasts in combination with IFN-γ, TNF-α, and IL4. We compared potencies of GCs to alter gene expression by analyzing EC₅₀ and EC₉₀ for genes induced by GCs, and IC₅₀, and IC₉₀ for genes repressed by GCs as indicators of drug potency related to concentration. EC_x and IC_x are drug concentrations that mediate effects with x% efficiency, and useful to determine a therapeutic window to elicit favorable effects. Ultimately, we seek to provide foundational, pre-clinical data to support treatment approaches to optimize GC therapy.

MATERIALS AND METHODS

Detailed information is available in the Supplemental Information.

Cell culture.

An immortalized human VF fibroblast cell line, HVOX,³¹ and macrophages derived from THP-1 (ATCC, Manassas, VA) were treated with 10 ng/mL TNF-α, 100 ng/mL IFN-γ, or 40 ng/mL IL4 +/- various concentrations of GCs for 24 hours.

Quantitative Real-Time Polymerase Chain Reaction (qPCR).

Total RNA were isolated and reverse transcribed with commercially available kits. The TaqMan Gene Expression kit (Life Technologies) and StepOne Plus (Applied Biosystems) were employed for quantitative analyses.

Data analysis.

All experiments were repeated in triplicate, at least. Determination of IC₅₀, EC₅₀, IC₉₀, EC₉₀, and illustration of concentration-response curves were performed using the R drc package.³²

RESULTS

Gene regulation differed based on GC concentration in human VF fibroblasts.

Pro-inflammatory genes (CXCL10, CXLC11, TNF, and PTGS2) were quantified as well as the pro-fibrotic gene CCN2 in the presence of IFN-γ or TNF-α +/- dexamethasone, methylprednisolone, or triamcinolone (Supplemental Figure S1 and S2; representative plots of qPCR data). To determine the concertation of GCs (EC₅₀) required for transcriptional activation, expression of TSC22D3 (GILZ), a major transcriptional target of GR,³³ was also quantified in HVOX fibroblasts treated with IFN-γ or TNF-α over a range of GC concentrations. IFN-γ and TNF-α stimulation had little effect on TSC22D3 expression. A dose dependent increase in TSC22D3 expression was observed in HVOX fibroblasts in response to all GCs. The EC₅₀ of dexamethasone, methylprednisolone, and triamcinolone was approximately 7, 32, and 40 nM in IFN-γ-treated fibroblasts and 10, 37, and 60 nM in TNF-α-treated fibroblasts (Table 1). This disparity in response is consistent with data regarding potencies of the three GCs estimated from their ability to suppress eosinophils and inhibit inflammation.³⁴ The EC₅₀ for GC-induction of CCN2 expression was similar to concentrations required to upregulate TSC22D3 (Table 1). These data suggest a hierarchy of GC potency for gene induction: dexamethasone >> methylprednisolone > triamcinolone (rank order).

Table 1:

IC50/EC50 and IC90/EC90 of GCs' effects on gene expression in human VF fibroblasts

			Dexamethasone			Methylprednisolone			Triamcinolone
	Genes	Major role	Down- or up-regulation	IC50/EC50 (nM)*	IC90/EC90 (nM)*	Down- or up-regulation	IC50/EC50 (nM)*	IC90/EC90 (nM)*	Down- or up-regulation	IC50/EC50 (nM)*	IC90/EC90 (nM)*
IFN-γ-treated VF fibroblast
	TSC22D3	Target of GR	Up	7.36±2.18	58.84±10.22	Up	32.09±3.22	421.11±91.86	Up	40.07±15.20	486.25±80.03
	CXCL10	Chemoattractant of leukocytes	Down	0.85±0.37	3.00±1.47	Down	2.08±0.44	10.17±1.63	Down	5.29±2.17	22.34±5.48
	CXCL11	Chemoattractant of leukocytes	NA	NA	NA	NA	NA	NA	NA	NA	NA
	TNF	Inflammatory cytokine	NA	NA	NA	NA	NA	NA	NA	NA	NA
	PTGS2	Inflammatory enzyme	Down	0.47±0.19	1.85±1.19	Down	1.82±0.52	10.64±2.12	Down	4.25±0.26	38.23±5.96
	CCN2	Fibrotic matricellular protein	Up	8.19±2.68	88.57±23.20	Up	30.21±9.06	493.54±119.63	Up	32.47±9.04	372.35±57.22
TNFα-treated VF fibroblast
	TSC22D3	Target of GR	Up	9.72±2.59	65.71±11.58	Up	37.18±10.18	445.71±94.58	Up	59.72±14.25	593.12±129.96
	CXCL10	Chemoattractant of leukocytes	Down	0.45±0.23	3.74±1.71	Down	1.80±0.73	7.36±2.36	Down	4.48±1.02	31.59±10.71
	CXCL11	Chemoattractant of leukocytes	Down	0.36±0.15	2.69±1.30	Down	1.19±1.05	3.22±0.95	Down	3.48±0.56	17.91±4.72
	TNF	Inflammatory cytokine	Down	0.43±0.12	1.74±0.78	Down	2.75±0.77	10.45±3.79	Down	4.83±1.46	17.36±4.78
	PTGS2	Inflammatory enzyme	Down	0.39±0.19	1.70±1.01	Down	2.16±1.12	8.67±2.52	Down	5.46±2.12	19.30±5.65
	CCN2	Fibrotic matricellular protein	Up	7.73±2.01	72.88±13.73	Up	34.34±8.45	481.02±131.35	Up	37.67±8.37	354.73±60.99

^*Data are shown as mean ± SD from triplicate independent experiments

NA: Not applicable.

We next examined the concentration of GCs required to repress expression (IC₅₀) of pro-inflammatory genes CXCL10, CXCL11 and PTGS2. A dose dependent decrease in IFN-γ and TNF-α-mediated CXCL10 expression was observed in response to all three GCs in HVOX fibroblasts. Although GCs had a tendency to suppress CXCL11 expression in IFN-γ-treated fibroblasts, clear concentration-dependent curves were not obtained. Similarly, concentration-dependent curves were not obtained for TNF expression in IFN-γ-treated fibroblasts (Supplemental Figure- S1D, J, P). In contrast, CXCL11 upregulation by TNF-α was inhibited by GCs in a concentration dependent manner. Upregulation of TNF by TNF-α was also inhibited by the three GCs (Supplemental Figure- S2D, J, P). Interestingly, 6-times less GC (IC₅₀) was required to inhibit expression of the inflammatory genes (1, 3, and 6 nM, respectively) compared to the EC₅₀ required to induce TSC22D3 and CCN2 expression (Table 1). This finding was confirmed by IC₉₀ of dexamethasone, methylprednisolone, and triamcinolone (20–50, 45–160, and 10–20 times lower than EC₉₀ to enhance CCN2 expression; Table 1). Less GC was required to repress IFN-γ and TNF-α stimulated expression of CXCL10, CXCL11 and PTGS2 compared to GC-dependent induction of TSC22D3 and CNN2 in HVOX fibroblasts.

HVOX fibroblasts were also treated with IL4; IL4 upregulates fibrotic genes including smooth muscle actin and collagens in pulmonary and hepatic fibroblasts.^35,36 IL4 did not alter ACTA2, COL1A1, LOX, or TGM2 expression (data not shown); additional GC experiments were not performed on IL4-treated fibroblasts.

Gene regulation differed based on GC concentration in human macrophages.

Supplemental Figures S3-S5 show representative plots of qPCR results obtained from THP1-derived macrophages treated with IFN-γ, TNF-α, or IL4 in combination with GCs. Similar to HVOX fibroblasts, dexamethasone, methylprednisolone, and triamcinolone increased TSC22D3 in a concentration dependent manner in macrophages treated with IFN-γ, TNF-α, or IL4 (Supplemental Figure- S3A, G, M, S4A, G, M and S5A, F, K). EC₅₀ of dexamethasone, methylprednisolone, and triamcinolone were approximately 9, 26, and 91 nM in IFN-γ-treated cells; 14, 37, and 59 nM in TNF-α-treated cells; and 9, 36, and 71 nM in IL4-treated cells (Table 2). CXCL10 and CXCL11 were markedly upregulated in IFN-γ-treated macrophages and moderately upregulated in TNF-α-treated macrophages (Supplemental Figures- S3B, C, H, I, N, O and S4B, C, H, I, N, O). GCs further upregulated these genes only in IFN-γ-treated macrophages; the EC₅₀ of dexamethasone, methylprednisolone, and triamcinolone was 4, 12, and 31 nM for CXCL10 upregulation, and 3, 9, and 19 nM for CXCL11 (Table 2).

Table 2:

IC50/EC50 and IC90/EC90 of GCs' effects on gene expression in human macrophages

			Dexamethasone			Methylprednisolone			Triamcinolone
	Genes	Major role	Down- or up-regulation	IC50/EC50 (nM)*	IC90/EC90 (nM)*	Down- or up-regulation	IC50/EC50 (nM)*	IC90/EC90 (nM)*	Down- or up-regulation	IC50/EC50 (nM)*	IC90/EC90 (nM)*
IFN-γ-treated macrophage
	TSC22D3	Target of GR	Up	9.38±3.21	61.16±17.08	Up	25.62±11.82	422.48±122.07	Up	90.67±20.14	803.48±160.56
	CXCL10	Chemoattractant of leukocytes	Up	3.60±0.98	34.39±9.12	Up	12.36±2.57	108.13±28.53	Up	31.39±10.72	278.00±61.22
	CXCL11	Chemoattractant of leukocytes	Up	2.70±0.44	20.49±3.89	Up	9.36±2.47	65.82±6.75	Up	19.21±4.18	180.09±74.91
	TNF	Inflammatory cytokine	Down	0.33±0.09	3.69±1.39	Down	5.06±1.32	33.74±11.59	Down	10.45±3.92	49.74±20.59
	IL1B	Inflammatory cytokine	Down	0.11±0.04	0.30±0.23	Down	2.06±0.37	8.89±1.03	Down	3.20±0.86	11.13±1.53
	CCN2	Fibrotic matricellular protein	Up	17.59±4.37	89.46±24.37	Up	100.29±23.14	561.31±64.25	Up	144.69±19.14	880.33±171.57
TNFα-treated macrophage
	TSC22D3	Target of GR	Up	13.60±3.22	65.74±20.04	Up	37.28±4.06	509.32±98.72	Up	58.74±14.25	510.95±102.67
	CXCL10	Chemoattractant of leukocytes	NA	NA	NA	NA	NA	NA	NA	NA	NA
	CXCL11	Chemoattractant of leukocytes	NA	NA	NA	NA	NA	NA	NA	NA	NA
	TNF	Inflammatory cytokine	Down	1.78±0.58	11.53±3.41	Down	6.64±2.33	40.93±5.94	Down	15.82±3.77	100.01±18.74
	IL1B	Inflammatory cytokine	Down	0.55±0.29	1.65±0.74	Down	2.02±0.71	6.56±1.86	Down	5.03±1.07	19.74±4.54
	CCN2	Fibrotic matricellular protein	Up	25.52±8.70	98.65±24.23	Up	122.46±19.68	728.36±124.57	Up	136.13±23.34	795.01±161.04
IL4-treated macrophage
	TSC22D3	Target of GR	Up	9.37±1.68	64.33±7.51	Up	36.23±9.26	628.07±129.68	Up	70.76±13.40	506.51±93.65
	CCL17	Chemoattractant of Th2 cells	Up	13.78±3.02	78.04±18.97	Up	36.26±5.21	409.64±71.97	Up	60.22±10.34	531.67±85.41
	TGM2	Enzyme of ECM-crosslinking	Down	0.22±0.12	2.81±1.36	Down	1.20±0.39	13.39±3.96	Down	3.71±1.31	26.57±8.53
	LOX	Enzyme of ECM-crosslinking	Down	1.75±0.32	6.66±2.13	Down	4.07±1.27	36.18±11.61	Down	8.51±1.87	74.99±18.27
	CCN2	Fibrotic matricellular protein	Up	25.22±6.25	94.00±21.79	Up	117.42±26.63	640.98±97.26	Up	121.96±19.63	821.43±122.86

^*Data are shown as mean ± SD from triplicate independent experiments

NA: Not applicable.

TNF and IL1B expression in IFN-γ and TNF-α treated macrophages was inhibited by the three GCs (Supplemental Figure- S3D, E, J, K, P, Q and S4D, E, J, K, P, Q). IC₅₀ of dexamethasone, methylprednisolone, and triamcinolone to inhibit TNF and IL1B expression was less than 2, 7, and 16 nM (Table 2). GCs upregulated CCN2 in both IFN-γ and TNF-α treated macrophages (Supplemental Figures- S3F, L, R and S4F, L, R) and the EC₅₀ of dexamethasone, methylprednisolone, and triamcinolone was approximately 17–26, 100–122, and 136–145 nM (Table 2). In IL4-treated macrophages, GCs increased TSC22D3, CCL17, and CCN2 expression in a concentration-dependent manner (Supplemental Figure- S5A, B, E, F, G, J, K, L, O). Despite CCL17 upregulation, TGM2, a M2 marker, was downregulated by GCs (Supplemental Figure- S5C, H, M). Moreover, LOX expression was also downregulated by GCs (Supplemental Figure- S5D, I, N). EC₅₀ of dexamethasone, methylprednisolone, and triamcinolone to upregulate CCN2 in IL4-treated cells were approximately 25, 117, and 122 nM (Table 2). IC₅₀ of dexamethasone, methylprednisolone, and triamcinolone to inhibit expression of ECM cross-linking enzymes, TGM2 and LOX, was lower than 2, 5, and 9 nM. Collectively, with cytokine stimulation, human VF fibroblasts and macrophages, the EC₉₀ of dexamethasone, methylprednisolone and triamcinolone to upregulate CCN2 was 72–99, 481–729, 354–881 nM, respectivley (Tables 1, 2). The EC₉₀ was >6, >11, and >3 times higher than the IC₉₀ to downregulate CXCL10, CXCL11, PTGS2, TNF, IL1B, TGM2, and LOX.

DISCUSSION

In spite of relatively little data to direct clinical practice, steroids are ubiquitous in laryngology. The current study provides foundational in vitro data regarding dosing, a critical initial step to optimized therapeutic strategies. Inflammatory and fibrotic genes tended to be downregulated by GCs and the concentrations required to downregulate these genes were lower than concentrations required to upregulate mediators of fibrosis.^15,37 Limiting fibrotic expression via low-concentration GCs likely limits the fibrotic response with potential to improve clinical outcomes.

Although GCs increased CXCL10 and CXCL11 in IFN-γ-treated macrophages, this response occurred at concentrations higher than required for inhibition of other genes. Upregulation of TSC22D3 also required higher concentrations of GCs than downregulation of inflammatory genes.TSC22D3 has multiple GREs on its promoter region and is a major target of GR. GC-induced leucine zipper (GILZ), encoded by TSC22D3, has multiple functions including inhibition of inflammation, cooperation in TGF-β/SMAD signaling, and mediation of M2 shift in macrophages.^38,39 Basal levels of GILZ are likely an intrinsic regulator of inflammation and GCs do not necessarily require GILZ upregulation to inhibit inflammatory responses in macrophages,³⁸ consistent with our data.

Peak plasma concentrations of free dexamethasone and methylprednisolone after oral administration of 1.5 mg dexamethasone and 4 mg methylprednisolone is approximately 8.9 nM and 21.1 nM, respectively.^34,40,41 These concentrations are lower than the EC₉₀ for CCN2 upregulation in the current study. Oral GCs appear suitable for GC-induced downregulation of inflammatory genes. Conversely, concentrations of injectable dexamethasone, methylprednisolone, and triamcinolone are 4–10 mg/mL (10–25 mM), 20–80 mg/mL (53–214 mM), and 40 mg/mL (100 mM) and higher than the EC₉₀ for CCN2 upregulation.⁴² These data suggest local injection of high-concentration GCs may elicit unfavorable responses. Further analyses of in-vivo pharmacokinetics are required to confirm optimal GC concentrations.

Dexamethasone, methylprednisolone, and triamcinolone exhibited similar effects on gene regulation and similar clinical outcomes are expected assuming altered dosages based on potency. However, regarding disparities in EC₉₀ to upregulate CCN2 and IC₉₀ to downregulate inflammatory mediators and ECM crosslinking enzymes, methylprednisolone had the greatest disparity in effective concentrations. This finding likely suggests methylprednisolone has a broader therapeutic window to avoid pro-fibrotic responses.

GC effects on gene expression varied depending on cytokine and cell type. In TNF-α treated VF fibroblasts, CXCL10, CXCL11, TNF, and PTGS2 decreased in response to GCs. In contrast, VF fibroblasts treated with IFN-γ were partially resistant to GCs; no clear inhibitory effects on CXCL11 or TNF expression were observed. CXCL10 and CXCL11 expression in macrophages was enhanced by GCs in the presence of IFN-γ. STAT1 and NF-κB are major transcription factors stimulated by IFN-γ and TNF-α and synergistically stimulate inflammatory gene expression, including CXCL10 and CXCL11.⁴³ Activated GR can affect STAT1 and NF-κB signaling pathways at multiple levels, such as upregulation of endogenous antagonists and agonists, binding to response elements in the genome, recruitment of co-activators and repressors, and inhibition of nuclear translocation.^19,44–47 Anti-inflammatory mechanisms of GCs are not necessarily common across cell types and can vary based on context. In Crohn’s disease, for example, GCs increased IκBα, an endogenous inhibitor of NF-κB, in vascular endothelial cells, but not in mononuclear cells.⁴⁴ Although dexamethasone inhibited STAT1 activation via toll-like receptor signaling in peritoneal macrophages by inducing suppressor of cytokine signaling 1 (SOCS1),⁴⁵ dexamethasone supports STAT1-dependent gene expression in human peripheral blood monocytes by recruiting PU.1 transcription factor to IFN-γ-activated sites.⁴⁸ To further optimize GC treatment, experimentation inclusive of all cell types and contexts related to the diseased tissue phenotype must be considered. Furthermore, in vitro experimentation is contrived and limited by the lack of in vivo biological complexity. The current study investigated a single time point in two cell types; further experimentation is indicated prior to shifts in clinical care paradigms. This work, however, is a critical initial step in that regard.

GC-induced CCN2 upregulation has been observed in various models.^23,49–51 For example, renal fibrosis induced by high dose GCs is dependent on CTGF.¹⁵ In the current study, GCs increased CCN2 expression in VF fibroblasts and macrophages regardless of cytokine stimuli. YAP/TEAD signaling is important for GC-induced CCN2 expression.^21,23 GR forms a complex with TEAD to promote TEAD-dependent gene expression. However, indirect activation of YAP/TEAD signaling through upregulation of fibronectin has also been proposed.^23,52,53 Collectively, transcriptional regulation by GCs appears to be specific to cell type and context.

CONCLUSION

GC-induced downregulation of inflammatory molecules and ECM-crosslinking enzymes was elicited at lower concentrations of GCs compared to GC-mediated upregulation of pro-fibrotic genes in both human VF fibroblasts and macrophages. These data suggest restricting GC concentration may limit inflammation and prevent fibrosis, potentially improving clinical outcomes.

Supplementary Material

supinfo2

Supplemental Figure Legends. Legends corresponding to Supplemental Figures.

supinfo1

Supplemental Materials and Methods. Comprehensive description of materials and methods employed in the current investigation.

supinfo3

Supplemental Figures 1–5. Dose response curves for dexamethasone, methylprednisolone, and triamcinolone.