Monocytes are essential for inhibition of synovial T-cell glucocorticoid-mediated apoptosis in rheumatoid arthritis

Rheumatoid arthritis (RA) is a chronic inflammatory disease that is characterized by excessive synovial infiltration and proliferation of mononuclear cells (MCs) partly due to a defective apoptotic process [1]. RA synovial T cells express a phenotype suggesting chronic immune activation but have been found to be anergic [2] and resistant to apoptosis [3, 4]. It has been suggested that factors such as chronic exposure to tumor necrosis factor (TNF) [5], exposure to interleukin-2 receptor (IL-2R) γ chain cytokines, and inhibitory signals received through interaction with stromal cells [3] might contribute to the T cell-specific phenotype of the rheumatoid synovium. This phenotype has been associated with the overexpression of two intracellular molecules, Bcl-2 and Bcl-xl [3, 6, 7], capable of blocking mitochondria-induced apoptosis.

Glucocorticoids are potent anti-inflammatory agents that modulate apoptosis of immune cells. Glucocorticoid activities can be divided in (a) genomic effects mediated through cytosolic glucocorticoid receptors (GRs) that need hours to become evident at the cellular and tissue levels and (b) nongenomic effects mediated through membrane-bound GR or nonspecific physicochemical interaction with the cell membrane which might explain some of the immediate effects observed with glucocorticoid administration in vivo [8]. One of the classic effects of glucocorticoids is induction of apoptosis. In vitro, synthetic glucocorticoids induce apoptosis of human thymocytes and activated T cells of human peripheral blood [9, 10]. The mechanism of T-cell glucocorticoid-induced apoptosis is primarily mediated through the mitochondrial cell death pathway [11] and is thought to be essentially dependent on genomic effects [12]. Two of the main mechanisms for resistance to glucocorticoid apoptosis are defects in the GR signaling and/or defects of the cell apoptotic machinery, such as disregulation of the Bcl-2 rheostat [13]. To date, several synthetic glucocorticoids such as triamcinolone (for local intra-articular administration) and methylprednisolone (for both local and systemic administration) are currently used in clinical practice. Differences in the mechanisms of action of these two compounds have been previously reported [14].

We have previously demonstrated that treatment with intra-articular glucocorticoids reduces the number of synovial tissue (ST) T cells in a wide range of arthritis types and suggested that this finding might be the consequence of reduced inflammatory cell trafficking to the joints [15]. However, apoptosis induction by glucocorticoids might be an additional mechanism. In this study, we used sequential arthroscopic biopsies to characterize the effect of glucocorticoids on synovial cellularity and apoptosis levels in patients with RA. We further investigated ex vivo the link between synovial-derived immune cell interactions and sensitivity to glucocorticoid-induced apoptosis. We demonstrate that monocytes rescue synovial T cells from glucocorticoid-induced apoptosis through a soluble factor(s)-mediated mechanism, a feature that is specific for RA.

Clinical response following intra-articular glucocorticoids is accompanied by a decrease in the number of ST T cells. All patients included in the study were clinical responders as evaluated by physician assessment during arthroscopies. The clinical response was paralleled by a significant decrease in the number of ST T cells (from a mean ± standard error of the mean [SEM] of 15.9 ± 4.1 to a mean ± SEM of 5.4 ± 1.9), as evaluated by CD3 staining without changes in the number of ST macrophages, as evaluated by both CD68 and CD163 staining (data not shown).

The decrease in the ST T-cell population is not mediated through apoptosis induction. Synovial apoptosis evaluated by TUNEL and staining for active caspase-3 did not show changes following intra-articular glucocorticoid injection. ST lymphoid aggregates showed absent to minimal apoptosis levels with both methods both before and after intra-articular glucocorticoid injection (Figure 1). This was confirmed by dual-immunofluorescence demonstrating minimal (<2%) levels of apoptosis (TUNEL) in CD3⁺ cells both before and after treatment (Figure 2).

RA SF-derived T cells are resistant to glucocorticoid-induced apoptosis in the presence of SF-derived monocytes. To further investigate the effect of glucocorticoids on T-cell apoptosis, SF MCs containing both monocytes and lymphocytes but no fibroblast cells were incubated ex vivo with dexamethasone. T cells in cocultures with monocytes of RA-derived (Figure 3b) but not psoriatic arthritis-derived (Figure 3a) SF were resistant to dexamethasone-induced apoptosis. As different synthetic glucocorticoid compounds might have distinct effects, triamcinolone (Figure 3c) and methylprednisolone (Figure 3d) were also tested but both failed to induce T-cell apoptosis in monocyte-T cell cocultures derived from RA SF.

T cell-monocyte interaction is essential to render RA SF T cells resistant to glucocorticoid-induced apoptosis. We hypothesized that the synovial RA environment with close contact between different subsets of inflammatory cells and presence of mediators contributes to the glucocorticoid-induced apoptosis-resistant phenotype of SF-derived T cells. To confirm this, SF-isolated T cells were treated in vitro with different synthetic glucocorticoid compounds. All three tested compounds at equivalent doses resulted in a significant fold increase of the apoptosis levels of isolated T cells to a maximum of 1.7 ± 0.2 for dexamethasone, 1.8 ± 0.2 for triamcinolone, and 3.0 ± 0.8 for methylprednisolone (all values expressed as mean ± SEM) (Figure 4).

Soluble factor(s) rather than cellular interaction are essential for the induction of the T-cell apoptosis-resistant phenotype. To further investigate the mechanism responsible for the resistance of T cells to glucocorticoid-induced apoptosis, we analyzed the importance of cellular contact versus soluble factor(s). Isolated T cells were cultured in the presence of, but without direct contact with, isolated SF-derived monocytes. Incubation with dexamethasone did not result in apoptosis of the T cells (mean ± SEM of 1.0 ± 0.1-fold increase as compared with control), suggesting that soluble factor(s) rather than cellular contact are primarily responsible for induction of the apoptosis-resistant phenotype of the RA synovial T cells (Figure 5).

Intra-articular glucocorticoids are a powerful adjuvant therapy for a variety of inflammatory joint diseases which efficiently reduces local joint inflammation. We demonstrate here that, in RA patients, this effect is mediated through the reduction of the synovial T-cell population as previously suggested in a cohort of patients with arthritis of different pathogenesis [18]. Furthermore, we provide evidence for the first time that RA-derived synovial T cells are resistant to apoptosis induction by glucocorticoids due to a soluble factor(s)-mediated interaction with monocytes.

Our immunohistochemistry results demonstrate that local administration of glucocorticoids decreases the number of lymphocytes without changes in the monocyte/macrophage population, evaluated as both CD68⁺ and CD163⁺ cells. The T cell-specific effect of locally administrated glucocorticoids might reside in the imbalance between the two alternatively spliced transcripts of the GR that have been suggested to have different functional characteristics. Exposure of cells to proinflammatory stimuli such as TNF and IL-1 can lead to induction of β-isoform of human glucocorticoid receptor (hGRβ) and suppression of hGRα, resulting in diminished glucocorticoid responsiveness [19]. Furthermore, within the same tissues, the levels of hGRβ may vary considerably between different types of cells [20]. Thus, the local proinflammatory milieu in an inflamed joint might contribute to the cell type-specific effect of locally administrated glucocorticoids.

Our findings suggest a distinct effect of local as compared with systemic administration of glucocorticoids which has been shown to decrease both lymphocyte and macrophage populations [21]. The difference might reside in the use of distinct synthetic glucocorticoid compounds for local versus systemic administration (that is, triamcinolone versus prednisolone/methylprednisolone). It has been suggested that, at equivalent doses, the effects of triamcinolone and dexamethasone, but not of methylprednisolone, are suppressed by overexpression of the hGRβ that acts as a natural dominant negative inhibitor of the transactivation of glucocorticoid-responsive genes [14]. However, when we tested the three compounds at equivalent doses, we did not observe differences in the in vitro effect of any of the compounds in any cell population studied. An alternative explanation is the apparently specific effect of systemically as compared with locally administrated high-dose glucocorticoids to induce profound monocytopenia in the peripheral blood [22] that would interfere with local synovial accumulation of monocytes/macrophages.

The observed reduction in the number of synovial T cells might be due either to a lower rate of recruitment or to a higher rate of clearance at the site of inflammation. We have previously demonstrated that intra-articular glucocorticoids decrease synovial expression of ICAM-1 (intracellular adhesion molecule-1), an adhesion molecule essential for leukocyte migration, despite minimal changes in the inflammatory phenotype of the endothelial synovial cells [15]. Our current results showing resistance of RA synovial T cells to glucocorticoid-induced apoptosis provide further indirect support for decreased leukocyte recruitment as the major mechanism responsible for the decreased cellularity observed after treatment with intra-articular glucocorticoids.

In RA, synovial-derived T cells have a phenotype suggestive of chronic immune activation but express low levels of cytokines and show signs of anergy [2]. These cells are resistant to apoptosis, partly due to their interaction with other cell populations present in the RA synovial inflamed milieu. It has been previously demonstrated that synovial-derived isolated T cells are rescued from spontaneous apoptosis through an integrin-ligand interaction with stromal cells, an effect that was mimicked by the addition of several members of the IL-2R γ chain cytokines, such as IL-15 [3]. Along the same line, coculture of autologous synovial RA T cells with monocytes induces homeostatic proliferation of T cells which is dependent on the membrane-bound TNF on monocytes [23]. We demonstrate that not only spontaneous but also glucocorticoid-induced apoptosis is dependent on the complex cell-cell interaction in the rheumatoid synovium. The essential factor in this situation appears to be the T cell-monocyte interaction to the extent that T-cell isolation renders the cells sensitive to apoptosis, while coculture of T cells with monocytes in the absence of fibroblasts prevented the effect of all tested glucocorticoid compounds. Furthermore, we propose that the main mechanism by which monocytes are able to rescue T cells is a soluble factor(s)-mediated interaction rather than cell-cell contact. It has been demonstrated, for example, that monocytes isolated from RA SF express IL-15 [24], a cytokine able to upregulate Bcl-2 expression [3] and to render activated T cells resistant to glucocorticoid-mediated apoptosis [25]. The mechanism appears to be RA-specific given that T-cell apoptosis induction was observed in cocultures of cells obtained from psoriatic arthritis in the presence of dexamethasone at similar doses.

We demonstrate that monocytes are essential in rescuing synovial T cells from glucocorticoid-induced apoptosis through a soluble factor(s)-mediated mechanism, a feature that is specific for RA-derived synovial T cells. We propose that this might be overcome by the combination of locally administrated glucocorticoids with monocyte-targeted therapies rather than T-cell apoptosis-inducing therapies.