In vitro glucocorticoid sensitivity is associated with clinical glucocorticoid therapy outcome in rheumatoid arthritis

Abstract

Introduction: Genetic and disease-related factors give rise to a wide spectrum of glucocorticoid (GC) sensitivity in rheumatoid arthritis (RA). In clinical practice, GC treatment is not adapted to these differences in GC sensitivity. In vitro assessment of GC sensitivity before the start of therapy could allow more individualized GC therapy. The aim of the study was to investigate the association between in vitro and in vivo GC sensitivity in RA.

Methods: Thirty-eight early and 37 established RA patients were prospectively studied. In vitro GC sensitivity was assessed with dexamethasone-induced effects on interleukin-2 (IL-2) and glucocorticoid-induced leucine zipper (GILZ) messenger RNA expression in peripheral blood mononuclear cells (PBMCs). A whole-cell dexamethasone-binding assay was used to measure number and affinity (1/KD) of glucocorticoid receptors (GRs).

Results: GR number was positively correlated with improvement in DAS. IL-2-EC₅₀ and GILZ-EC₅₀ values both had weak near-significant correlations with clinical improvement in DAS in intramuscularly treated patients only. HAQ responders had lower GILZ-EC₅₀ values and higher GR number and KD.

Conclusions: Baseline cellular in vitro glucocorticoid sensitivity is modestly associated with in vivo improvement in DAS and HAQ-DI score after GC bridging therapy in RA. Further studies are needed to evaluate whether in vitro GC sensitivity may support the development of tailor-made GC therapy in RA.

Figure 1

Flow chart of the tREACH and FLARE study. The baseline work-up in tREACH patients included the GILZ/IL-2 assays. In FLARE patients, both a GC-binding assay and GILZ/IL-2 assays could be performed. In all patients, disease activity score (DAS) was measured at baseline and after 2 weeks of GC bridging therapy. ^#Only patients in the high-probability group eventually fulfilling the 1987 ACR criteria for RA were included in the final analysis. *Prednisone is tapered according to the following schedule: weeks 1 through 4, 15 mg/day, weeks 5 and 6, 10 mg/day; week 7 and 8, 5 mg/day; weeks 9 and 10, 2.5 mg/day. Intramuscular GCs could be either methylprednisolone, 120 mg, or triamcinolone acetonide, 80 mg. Baseline workup and start of standardized treatment occurred on the same day. GC, glucocorticoids; HAQ-DI, health assessment questionnaire disability index; HCQ, hydroxychloroquine; MTX, methotrexate; SSZ, sulfasalazine.

Figure 2

Baseline in vitro parameters of GC sensitivity in healthy controls, tREACH, and FLARE patients. The IL-2 assay (A) and GILZ-assay (C) were performed in both tREACH and FLARE patients (bioassays in 32 patients, GC-binding assay in 32 patients, bioassays and GC-binding assay in 27 patients; control groups for the bioassays (n = 20) and binding assay (n = 16) were not the same). As secondary outcome, IL-2 repression (B) and GILZ induction (D) was calculated as follows: $IL-2repression=100\times \frac{\mathsf{\text{(IL}}2-\mathsf{\text{expression}},\mathsf{\text{PHA)}}-\mathsf{\text{(IL2}}-\mathsf{\text{expression,}}\mathsf{\text{333nM)}}}{\left(IL2-expression,PHA\right)}$$GILZinduction=100\times \frac{\mathsf{\text{(GILZ}}-\mathsf{\text{expression, 333nM)}}}{\mathsf{\text{(GILZ}}-\mathsf{\text{expression, PHA)}}}$ The numbers of GRs (E) and the affinity of the receptor (F) were determined in FLARE patients only. EC₅₀, half maximal effective concentration. P values were calculated by using ANOVA and Bonferroni post hoc correction; normalized data were used where appropriate.

Figure 3

Correlation between in vitro and in vivo glucocorticoid sensitivity in intramuscularly treated RA patients. In vivo glucocorticoid sensitivity is presented as percentage improvement in DAS according to the following formula: $100\times \frac{DAS,baseline-DAS,after2weeks}{DAS,baseline}$ Correlations between √IL-2-EC₅₀ values (A), GILZ-EC₅₀ values (B), natural logarithm of the number of GRs per cell (C), and natural logarithm of the K_D of the receptor (D) and percentage improvement DAS. R², square of the Pearson correlation coefficient; proportion explained variability. Triangles (▲) represent the tREACH patients, and solid circles (●) represent FLARE patients.

Figure 4

In vitro glucocorticoid sensitivity and improvement in HAQ-DI score in patients with established rheumatoid arthritis (FLARE study). Boxplots (each box shows the mean and interquartiles) and outliers (●) of IL-2-EC₅₀ (A), GILZ-EC₅₀ (B), number of GRs (C) and K_D of the GRs (D) in HAQ-DI responders and HAQ-DI-nonresponders. Patients are defined as responders if their HAQ-DI sum scores were at least 0.25 lower after GC therapy.