14-3-3η is a novel mediator associated with the pathogenesis of rheumatoid arthritis and joint damage

Rheumatoid arthritis (RA) is a chronic autoimmune disease characterised by inflammation of the synovium [1]. When left untreated, RA typically leads to severe joint destruction [2]. Early diagnosis, combined with an accurate prognostic assessment at presentation, is a central tenet in the effective management of RA patients [3]. There is consensus amongst rheumatologists that the development of a risk stratification strategy to group patients into low-, medium- and high-risk categories for radiographic progression at presentation is a clinical imperative [4]. Currently, this remains a major limitation in patient management because risk factors such as rheumatoid factor (RF), anticitrullinated protein antibodies (ACPAs; often detected by performing an anti–cyclic citrullinated peptide antibody test) and C-reactive protein (CRP) together account for only 32% of the total variance in predicting joint destruction, leaving 68% of the variance unaccounted for [4].

Serum 14-3-3η, which was first described in 2007 [5] as being elevated in arthritis, has been reported to be significantly higher in RA patients than in healthy individuals and in various disease controls including, but not limited to, osteoarthritis (OA), ankylosing spondylitis (AS) and gout [6, 7]. On the basis of receiver operating characteristic curve analyses defining optimal sensitivity and specificity, serum 14-3-3η positivity is denoted by a concentration ≥0.19 ng/ml [6, 7]. We have previously shown that (1) serum 14-3-3η expression is not strongly correlated with standard clinical and serological measures, is detected in both early and established RA, with sensitivities of 60% to 82% and 78%, respectively, and (2) that 14-3-3η positivity adds incrementally to both RF and ACPA positivity for diagnostic sensitivity [8]. In a previous study of an early RA cohort, we reported that 60% of patients were positive for 14-3-3η, 32% for RF, 44% for ACPA and 72% for at least one of those three markers [8].

The 14-3-3 family of conserved regulatory proteins consists of seven isoforms: α/β, γ, δ/ζ, ε, η, θ/τ and σ. Under normal circumstances, these proteins exist as intracellular adapters that can either homo- or heterodimerise to form a cuplike structure known as the amphipathic groove, which allows them to interact with more than 200 intracellular proteins to modulate their activities. Interactions include an array of biological processes, such as protein trafficking and cellular signalling. Externalisation of 14-3-3η, similarly to its process in RA [5], is believed to be mediated in part through an exosomal process in a like that of other intracellular proteins, such as the release of heat shock proteins (HSPs) from human peripheral blood mononuclear cells (PBMCs) or B lymphocytes under heat stress [9, 10]. Importantly, these two families of proteins—14-3-3 and HSPs—have been described as key components of exosomal microvesicles [11], and it is now widely accepted that members of the HSP superfamily, such as HSP60, behave as proinflammatory factors priming the immune system to respond to a noxious agent [12].

In 2007, we reported a strong correlation between the expression of 14-3-3η and matrix metalloproteinases (MMPs) and demonstrated that extracellular 14-3-3η possesses MMP-1-inducing activity in vitro on the basis of 14-3-3η levels at the upper range of serum levels observed in a small subset of RA patients [5‐7]. MMPs are serine proteases that play a critical role in maintaining tissue homeostasis, but, in the context of RA, the imbalance between expression of these proteolytic enzymes and their cognate inhibitors leads to the breakdown of cartilage [13].

MMP expression has been reported to be regulated through the transcription factor activator protein 1 (AP-1), which resides downstream from intracellular signalling factors such as mitogen-activated protein kinase (MAPK) [14‐17]. The MAPK family, which includes extracellular regulated kinase (ERK), c-Jun N-terminal kinase/stress-activated protein kinase (JNK/SAPK) and p38MAPK, has been investigated extensively in RA [18]. Recently, de Launay et al. reported a specific increase in ERK and JNK activation, but not in p38MAPK activation, in early RA patients with progressive joint destruction, underscoring their possible relevance to the aetiology of RA [19].

In our present study, given the strong correlation of 14-3-3η with MMP-1 and MMP-3 in synovial fluid and serum [5], we sought to advance the understanding of the role of 14-3-3η in RA by assessing whether 14-3-3η (1) possesses extracellular ligand properties leading to activation of MAPK, (2) induces mediators of inflammation and joint destruction and (3) contributes prognostic information in addition to clinical and serological measures used in the management of RA.

14-3-3η is one of seven members of the 14-3-3 family that are preferentially expressed at higher concentrations in certain tissues, underscoring the importance of specific isoforms in the regulation of tissue-specific functions [25‐27]. Data that we reported in 2007 indicated that 14-3-3η is expressed at significantly higher levels than the other isoforms in the synovial fluid of patients with arthritis and that these levels were three to five times higher than corresponding levels in the serum of matched donors, citing the joint as the likely source of serum 14-3-3η [5]. Since that first report, we have demonstrated that 14-3-3η is an RA-specific marker that complements both RF and ACPA, increasing their diagnostic value [6‐8]. We also reported a positive association between 14-3-3η and MMPs and suggested that 14-3-3η may have a role in the pathogenesis of RA [28].

To expand upon our understanding of the biological relevance of extracellular 14-3-3η, we performed in vitro cell signalling studies to determine if 14-3-3η signals through the MAPK family, as well as through which family members. This family was selected because the transcription factor AP-1, which resides downstream in the MAPK signalling nexus, is a key regulator of MMP expression. The data produced in this study indicate that stimulation of cells with 14-3-3η, similarly to TNFα, results in the phosphorylation of both ERK and JNK/SAPK in a time-dependent manner. However, unlike TNFα, 14-3-3η had no impact on p38MAPK phosphorylation. The differential effects of 14-3-3η in relation to TNFα on the activation of MAPK family members indicates that these two ligands, which are present in the serum of patients with RA, are likely to signal by divergent mechanisms. Phosphorylation of these two MAPK members, ERK and JNK/SAPK, together with the finding that 14-3-3η is present in the serum of 60% to 82% of early RA patients [8], is consistent with the observations reported by de Launay et al.: that ERK and JNK, but not p38MAPK, are preferentially phosphorylated in early RA [19]. On the basis of these findings, we propose that the relationship between 14-3-3η expression and the phosphorylation status of the MAPK family members, especially in early RA, be investigated in human synovial tissue.

It is well-established that phosphorylation of both ERK and JNK/SAPK at the TXY motif within the activation loop of these kinases is reflective of their activation. In turn, these MAPKs directly and/or indirectly phosphorylate the transcription factor c-Jun. Phosphorylation of c-Jun results in its association with c-fos, the formation of the AP-1 transcription complex and the transactivation or switching-on of genes containing the AP-1 binding site [29]. In this head-to-head analysis, we demonstrate that 14-3-3η at a concentration similar to that of TNFα, is capable of inducing genes that have been described in the literature as being TNFα-responsive. Our 14-3-3η dose escalation studies demonstrate that these TNFα-responsive genes are highly sensitive to 14-3-3η stimulation and can be induced with as little as 1.0 ng/ml, a 14-3-3η concentration that approximates median concentrations measured in the serum of RA patients [6‐8]. On the basis of these in vitro findings, it is evident that 14-3-3η may play a role in perpetuating inflammation through the induction of factors such as IL-6 and by exacerbating joint destruction via MMPs and RANKL. Examining 14-3-3η’s expression in relation to clinical outcomes in RA patients will be of utmost importance in understanding how 14-3-3η serum expression used as a diagnostic test can assist clinicians with patient management, to further substantiate its relevance to the pathophysiology of RA and to gain a better understanding of how 14-3-3η biomarker expression can be aligned with the advancement of a 14-3-3η-targeted therapy to identify the subset of patients who are more likely to respond to such an approach.

High levels of other 14-3-3 isoforms have been associated with more severe disease or worse outcomes in many patient groups, including, but not limited to, carcinoma [30‐33], Creutzfeldt-Jakob disease [34], Alzheimer’s disease [35], multiple sclerosis [36], infarction [37] and HIV [38]. Analysis of serum 14-3-3η expression in relation to joint damage and progression revealed significantly higher levels of 14-3-3η in patients who already had radiographic evidence of damage at study baseline, as well as in those who developed progression by the end of the follow-up period. Amongst other laboratory variables tested, none were significantly associated with radiographic status at baseline in the early RA cohort and only RF was significantly associated with radiographic progression. Nevertheless, 14-3-3η contributed to the variance of a regression model for radiographic progression (R²) that included RF. The association of 14-3-3η with radiographic status and progression was less prominent in the BeSt cohort than in the early RA cohort, but DAS28 control was already very well-established by year 3 in the majority of patients in the BeSt cohort (mean DAS28 = 1.9). Of the three laboratory variables examined—CRP, RF, and ACPA—only CRP was associated with radiographic progression. 14-3-3η contributed to the variance of a regression model for radiographic progression that included CRP.

In this study, we observed that serum 14-3-3η did not correlate with disease activity (DAS28) or with titres of CRP. We noted a modest correlation with ACPA and RF in the early RA cohort, but not in the established RA cohort. These findings indicate that 14-3-3η may act independently of these measures, which is one of the key requirements proposed by the Outcome Measures in Rheumatology (OMERACT) soluble biomarker working group for a biomarker reflecting joint damage [39]. The OMERACT guidelines state that, in order to be useful at informing structural damage endpoints, a marker should be independent of other commonly available variables associated with radiographic progression and should complement them in predictive models. Regression analyses revealed that 14-3-3η added incrementally to the predictive power of the model. On the basis of these findings, a prospective study designed to examine 14-3-3η’s capacity as an independent predictor of radiographic damage status and progression is underway.

As with proof-of-concept studies, one of the main limitations of the present study is the small sample size for both RA cohorts. In addition, relatively few patients demonstrated substantial joint damage progression. The results of this study are being used in the design of larger studies to formally investigate the merits of 14-3-3η in marking joint damage and predicting radiographic progression in combination with other clinical variables. In addition, we have previously shown that 14-3-3η is modifiable by both anti-TNF and standard DMARD therapies [40, 41]. Confounding by treatment is a significant limitation of studies in which the predictive role of biomarkers is assessed when only baseline biosamples are available. It is to be anticipated that levels of 14-3-3η are likely to change in both cohorts when treatment changes are driven by DAS28 targets, which will limit the predictive capacity of biosample data that are available only at baseline. This was particularly noteworthy in a previous report of BeSt study patients in which baseline levels of acute-phase reactants, RF and ACPA were not shown to be predictive of radiographic progression in the two patient groups undergoing intensive therapy. However, RF and ACPA were predictive in patients being treated with monotherapy or combination therapy with standard DMARDs [24]. Consequently, future studies should assess biosamples at multiple time points.

PPT is currently with GlaxoSmithKline, Stevenage, UK.