Mode of action of abatacept in rheumatoid arthritis patients having failed tumour necrosis factor blockade: a histological, gene expression and dynamic magnetic resonance imaging pilot study

Abstract

Objectives: Abatacept is the only agent currently approved to treat rheumatoid arthritis (RA) that targets the co-stimulatory signal required for full T-cell activation. No studies have been conducted on its effect on the synovium, the primary site of pathology. The aim of this study was to determine the synovial effect of abatacept in patients with RA and an inadequate response to tumour necrosis factor alpha (TNFalpha) blocking therapy.

Methods: This first mechanistic study incorporated both dynamic contrast-enhanced (DCE) magnetic resonance imaging (MRI) and arthroscopy-acquired synovial biopsies before and 16 weeks after therapy, providing tissue for immunohistochemistry and quantitative real-time PCR analyses.

Results: Sixteen patients (13 women) were studied; all had previously failed TNFalpha-blocking therapy. Fifteen patients completed the study. Synovial biopsies showed a small reduction in cellular content, which was significant only for B cells. The quantitative PCR showed a reduction in expression for most inflammatory genes (Wald statistic of p<0.01 indicating a significant treatment effect), with particular reduction in IFNgamma of -52% (95% CI -73 to -15, p<0.05); this correlated well with MRI improvements. In addition, favourable changes in the osteoprotegerin and receptor activator of nuclear factor kappa B levels were noted. DCE-MRI showed a reduction of 15-40% in MRI parameters.

Conclusion: These results indicate that abatacept reduces the inflammatory status of the synovium without disrupting cellular homeostasis. The reductions in gene expression influence bone positively and suggest a basis for the recently demonstrated radiological improvements that have been seen with abatacept treatment in patients with RA.

Conflict of interest statement

Competing interests: PE has provided expert advice and undertaken clinical trials for Bristol-Myers Squibb.

Figures

Figure 1

(A) Median change in synovial expression of cellular infiltrate after 4 months of treatment with abatacept. Multiple synovial biopsies were obtained from representative inflamed sites and stained with antibodies to a range of synovial markers. Synovial lining layer (LL) and sub-lining layer (SL) expression of each marker was scored semi-quantitatively on a five-point scale (0, minimal infiltration; 4, maximal infilatration) and a median percentage change from baseline calculated. *p

Figure 2

Geometric mean percentage change in synovial gene expression in patients after 4 months of treatment with abatacept in (A) all patients combined (n  =  14†), (B) disease activity score 28 (DAS28) responders (n  =  8†) versus DAS28 non-responders (n  =  6†). RNA was isolated from the synovium using chloroform extraction and reverse transcribed to cDNA. mRNA levels of tumour necrosis factor alpha (TNFα), IL-1β, IL-6, matrix metalloprotease (MMP) 1, MMP-3, IFNγ, receptor activator of nuclear factor kappa B (RANK), RANK ligand (RANKL) and osteoprotegerin (OPG) were quantified using TaqMan quantitative PCR analysis. Data are expressed as the geometric mean percentage change in gene expression relative to baseline. Error bars represent 95% CI. *95% CI do not include zero. †Two patients (one responder and one non-responder) had IFNγ levels below the detectable limit at baseline and one patient (non-responder) had IL-6 levels below the detectable limit at baseline; these patients were excluded from analysis for these cytokines. Wald statistics indicate a significant treatment effect of abatacept on IFNγ, IL-1β, IL-6, MMP-1 and MMP-3 simultaneously (p<0.01).

Figure 3

Median percentage change in key magnetic resonance imaging (MRI) measures. MRI of the knee was performed, and measures of the initial rate of enhancement (IRE) and maximal enhancement (ME) made in the global region as well as in regions of interest (ROI) at the cartilage–pannus junction (CPJ) and the suprapatellar pouch (SSP). Values for the ME and IRE are calculated from the sum of the enhancing pixels in each region of interest, expressed as arbitrary units. Error bars represent 25th and 75th percentiles.

Figure 4

Five characteristic dynamic gadolinium-enhanced magnetic resonance imaging sagittal scans acquired from the knee with superimposed colour data showing values of initial rate of enhancement (IRE) across synovial space following gadolinium enhancement—pixels shown in yellow represent high IRE values whereas red show relatively lower values. Image (A) shows global region of interest outlined in white, which defines the extent of the synovitis at a level of 10 pixels below the tibial plateau to the top of the suprapatellar pouch, avoiding enhancement from muscle and arteries in each of the five sagittal images. Images (B) and (C) are an example of pre and posttreatment, respectively.

Figure 5

The relationship between the IFNγ gene expression and magnetic resonance imaging (MRI) outcomes. Pearson correlation analyses were carried out to evaluate the relationship between gene expression data and MRI outcomes. A good correlation between IFNγ and the initial rate of enhancement (IRE)-global, as well as maximal enhancement (ME)-global was observed.