Gene expression profiling in the synovium identifies a predictive signature of absence of response to adalimumab therapy in rheumatoid arthritis

Abstract

Introduction: To identify markers and mechanisms of resistance to adalimumab therapy, we studied global gene expression profiles in synovial tissue specimens obtained from severe rheumatoid arthritis (RA) patients before and after initiation of treatment.

Methods: Paired synovial biopsies were obtained from the affected knee of 25 DMARD (disease-modifying antirheumatic drug)-resistant RA patients at baseline (T0) and 12 weeks (T12) after initiation of adalimumab therapy. DAS28-CRP (disease activity score using 28 joint counts-C-reactive protein) scores were computed at the same time points, and patients were categorized as good, moderate, or poor responders according to European League Against Rheumatism criteria. Global gene expression profiles were performed in a subset of patients by means of GeneChip Human Genome U133 Plus 2.0 Arrays, and confirmatory immunohistochemistry experiments were performed on the entire cohort.

Results: Gene expression studies performed at baseline identified 439 genes associated with poor response to therapy. The majority (n = 411) of these genes were upregulated in poor responders and clustered into two specific pathways: cell division and regulation of immune responses (in particular, cytokines, chemokines, and their receptors). Immunohistochemistry experiments confirmed that high baseline synovial expression of interleukin-7 receptor alpha chain (IL-7R), chemokine (C-X-C motif) ligand 11 (CXCL11), IL-18, IL-18 receptor accessory (IL-18rap), and MKI67 is associated with poor response to adalimumab therapy. In vitro experiments indicated that genes overexpressed in poor responders could be induced in fibroblast-like synoviocytes (FLS) cultures by the addition of tumor necrosis factor-alpha (TNF-alpha) alone, IL-1beta alone, the combination of TNF-alpha and IL-17, and the combination of TNF-alpha and IL-1beta.

Conclusions: Gene expression studies of the RA synovium may be useful in the identification of early markers of response to TNF blockade. Genes significantly overexpressed at baseline in poor responders are induced by several cytokines in FLSs, thereby suggesting a role for these cytokines in the resistance to TNF blockade in RA.

Figure 1

Evolution of disease activity score (DAS) (three variables) in 25 individual rheumatoid arthritis patients before (T0) and 12 weeks after (T12) initiation of adalimumab therapy. Patients are categorized into (good or moderate) responders or non-responders according to European League Against Rheumatism criteria.

Figure 2

Changes in immunohistochemistry parameters in the synovial biopsies of severe rheumatoid arthritis patients. Biopsies were collected prior to (T0) (n = 25) and 12 weeks after (T12) (n = 25) initiation of adalimumab therapy. (a) Characteristic images of the stained markers (sublining C68, CD3, CD20, and CD15) (original magnification × 400). (b) Ratio of surface staining to staining of the nuclei (S/N). Slides stained for CD68, CD3, CD15, and CD20 were analyzed using ImageJ with six digitalized pictures (magnification × 400) obtained for each sample. Open boxes refer to all patients, and closed boxes refer to responders. Results are the mean and standard error of the mean of S/N ratio. *P < 0.05; **P < 0.005 versus good and moderate responders using Wilcoxon matched-pairs signed rank tests.

Figure 3

Genes differentially expressed before (T0) and 12 weeks after (T12) start of adalimumab in synovial biopsy specimens of rheumatoid arthritis patients who responded to therapy. Paired Student t tests indicated that 632 (out of 54,675) genes displayed significant differences in expression between T0 and T12 in six synovial tissue samples obtained from RA patients who responded to adalimumab therapy. Pathway analyses indicated that a significant percentage of these genes clustered into two distinct pathways: genes involved in the regulation of immune responses (a) and genes involved in cell division (b). Fold-change values are the mean level of decreased expression at T12 as compared with T0. (c) Real-time reverse transcriptase-polymerase chain reaction studies of the expression of selected genes in rheumatoid arthritis synovial biopsy tissue before (T0) (n = 10) and 12 weeks after (T12) (n = 8) initiation of adalimumab therapy. Samples were loaded in triplicate, and results are the mean and standard error of the mean of gene expression, relative to the mean gene expression in a standard sample normalized to 1. *P < 0.05. CCL5, chemokine ligand 5; CTLA4, cytotoxic T-lymphocyte-associated antigen 4; LTB, lymphotoxin beta.

Figure 4

Genes differentially expressed at baseline between poor versus moderate and good responders to adalimumab therapy. Five hundred twenty-four genes were found to be differentially expressed among good, moderate, and poor responders at baseline by analysis of variance (P < 0.05). Post hoc (Student-Newman-Keuls) tests were used to discriminate genes that were specifically upregulated (n = 411) or downregulated (n = 28) in poor responders as compared with the two other groups. Pathway analyses indicated that these genes were significantly enriched in genes involved in the regulation of immune responses (a) and genes involved in cell division (b).

Figure 5

Baseline immunostaining for selected synovial markers of response to adalimumab therapy. Synovial samples of rheumatoid arthritis patients who responded or who did not respond to adalimumab therapy were stained at baseline with polyclonal antibodies directed at MKI67, interleukin-7 receptor α chain (IL-7R), interleukin-18 receptor accessory (IL-18rap), IL-18, and chemokine (C-X-C motif) ligand 11 (CXCL11). (a) Characteristic images of the stained markers are shown in responders (n = 20) versus non-responders (n = 5) (original magnification × 400). (b) Ratio of surface staining to staining of the nuclei (S/N). Slides were analyzed using ImageJ with six digitalized pictures (magnification × 400) obtained for each sample. Results are the mean and standard error of the mean of S/N ratio. *P < 0.05, **P < 0.005, ***P < 0.0005 using Wilcoxon matched-pairs signed rank tests.

Figure 6

Genes overexpressed at baseline in poor responders are significantly induced by the combination of tumor necrosis factor-alpha (TNF-α) and interleukin-1β (IL-1β) in fibroblast-like synovial cells (FLSs). FLSs were cultured overnight in the presence of TNF-α (10 ng/mL), IL-1β (10 ng/mL), IL-6 (10 ng/mL), IL-7 (100 ng/mL), IL-17 (50 ng/mL), or combinations of several of these cytokines. RNA was extracted and real-time reverse transcriptase-polymerase chain reaction evaluation of IL-7R, IL-6, INDO, CDC2, GTSE1, and MKI67 was evaluated in at least four different experiments. Results are expressed as the mean fold change in gene expression and standard error of the mean, relative to the mean gene expression of the baseline condition normalized to 1. *P < 0.05, **P < 0.005, ***P < 0.0005 using Wilcoxon signed rank tests.