Antiinflammatory glucocorticoid receptor ligand with reduced side effects exhibits an altered protein-protein interaction profile

Abstract

Glucocorticoids are commonly used antiinflammatory agents whose use is limited by side effects. We have developed a series of glucocorticoid receptor (GR) ligands that retain the strong antiinflammatory activity of conventional glucocorticoids with reduced side effects. We present a compound, LGD5552, that binds the receptor efficiently and strongly represses inflammatory gene expression. LGD5552 bound to GR activates gene expression somewhat differently than glucocorticoids. It activates some genes with an efficacy similar to that of the glucocorticoids. However, other glucocorticoid-activated genes are not regulated by LGD5552. These differences may be because of the more efficient binding of corepressor in the presence of LGD5552, compared with glucocorticoid agonists. This class of nonsteroidal, GR-dependent antiinflammatory drugs may offer a safer alternative to steroidal glucocorticoids in the treatment of inflammatory disease.

Fig. 1.

The benzylidene LGD5552 exhibits selective binding to human GR (hGR). The structure of LGD5552 is shown above a table with the binding affinity (K_i in nM) of the compound for human GR (hGR), human progesterone (hPR-A), human androgen (hAR), and human mineralocorticoid (hMR) receptors. This compound, (5Z)-5-[(2-fluoro-3-methylphenyl)methylene]-2,5-dihydro-10-methoxy-2,2,4-trimethyl-1H-[1]benzopyrano[3,4-f]quinolin-9-ol, has a molecular weight of 443.5 g/mole (C₂₈H₂₆FNO₃).

Fig. 2.

Transcriptional activation and antagonist activity of LGD5552. (A) LGD5552 is not an agonist in the MMTV:luciferase reporter assay. Dexamethasone (filled circles) induces transcription from this promoter >1,000-fold. LGD5552 and RU486 exhibit <5% efficacy compared with dexamethasone. (B) LGD5552 is an antagonist of dexamethasone-mediated activation at MMTV. Dexamethasone is added at a half-maximal concentration of 0.3 nM Dex, and LGD5552 (open squares) or RU486 (filled squares) are titrated in dose–response. (C) LGD5552 and dexamethasone are highly efficacious at repression of the inflammatory marker E-selectin. The E-selectin promoter was induced with inflammatory stimuli (TNF and IL-1β), and compounds were added in dose–response. Strong repression is correlated with antiinflammatory activity. Consistent with their binding affinity, LGD5552 (ED₅₀ = 2 nM, efficacy 100%) is less potent than dexamethasone (ED₅₀ = 0.1 nM, efficacy 100%). However, LGD5552 is equally potent and efficacious as prednisolone in the E-selectin assay (Pred: ED₅₀ = 4 nM, efficacy 98%) (data not shown).

Fig. 3.

Differential gene regulation of LGD5552 and prednisolone. H4IIE liver cells with endogenous GR were assayed for response to either prednisolone or LGD5552. RT-PCRs were used to quantify RNA from a variety of different genes. (A) (Left) Relative RNA levels for the PEPCK gene. (Right) Relative RNA levels for the PDK4. (B) (Left and Center) COX2 and APOCIII RNA levels were measured in response to vehicle, prednisolone, or LGD5552 in HEK293 cells transfected with hGR. (Right) POMC RNA was measured in ATT20 cells in response to vehicle, dexamethasone, prednisolone, or LGD5552 treatment in dose–response. Vehicle control is shown at the far left.

Fig. 4.

Differential coactivator/corepressor interactions induced by LGD5552. M2H assay to measure ligand-dependent interaction between GR and various peptides. (A) M2H assay with GR and peptides that interact to a greater degree in the presence of the GR agonists, dexamethasone or prednisolone, than in the presence of the GR antagonist, RU486. (B) M2H assay with GR and peptides that interact to a greater degree in the presence of the GR antagonist, RU486, than in the presence of the GR agonists, dexamethasone or prednisolone.

Fig. 5.

LGD5552 has full efficacy inhibiting CIA. The impact of treatment by vehicle, prednisolone, or LGD5552 on arthritis disease score in a mouse model of arthritis is shown. Animals injected with adjuvant and collagen develop severe disease (vehicle), whereas untreated animals (normal) do not. Glucocorticoids and LGD5552 are capable of blocking existing disease in these animals. Both compounds are active at <3 mg/kg. ∗, P < 0.05 versus vehicle. Statistical analysis used a one-way ANOVA followed by Fischer's LSD.

Fig. 6.

LGD5552 has less impact on bone formation and percentage of body fat than a steroid. Three-month-old male Swiss–Webster mice were treated orally with prednisolone or LGD5552 for 4 weeks. Treatment groups consisted of vehicle, prednisolone (3, 10, and 30 mg/kg), and LGD5552 (3, 10, and 30 mg/kg). Prednisolone (gray bars) raises the percentage of body fat at both 10 and 30 mg/kg, whereas LGD5552 (white bars) does not. Bone formation rate is significantly suppressed by prednisolone at all doses tested, whereas LGD5552 (white bars) suppresses at 30 mg/kg. LGD5552 is not significantly different from vehicle at 3 and 10 mg/kg.