Autoantibodies: Double Agents in Human Disease

PADS AND PROTEIN CITRULLINATION

Rheumatoid arthritis (RA) is a systemic inflammatory disease hallmarked by a heightened autoimmune response to citrullinated protein targets in the joints. The production of autoantibodies targeting citrullinated proteins has also been implicated in many other diseases that involve autoimmunity, such as multiple sclerosis, psoriasis, sporadic Creutzfeldt-Jakob disease, and Alzheimer’s disease (AD) (1, 2). Citrullination of proteins occurs posttranslationally and results from the deimination of peptidyl-arginine and its conversion to citrulline. This reaction is catalyzed by peptidylarginine deiminases (PADs), and although the exact circumstances and underlying purpose of PAD activation and protein citrullination within cells are not fully understood, PADs have emerged as key participants in the pathogenesis of RA (3, 4). Two PAD isoforms, PAD2 and PAD4, are expressed by cells in the bone and cartilage as well as by neutrophils and monocytes and are present at high levels in RA synovial tissue in regions that also contain citrullinated proteins (3, 5).

PADs require high concentrations of calcium ions (Ca²⁺) for catalytic activity in vitro. Citrullination assays have shown that 5 to 10 mM Ca²⁺ is necessary to achieve maximal PAD4 activation—concentrations that far exceed those possible in vivo within the extracellular synovial fluid (~0.49–0.98 mM) and within most human cells (~100 μM). Yet, citrullinated proteins and activated PAD4 are commonplace in the joints of RA patients. This conundrum has prompted a search for in vivo factors or conditions that may reduce the Ca²⁺ sensitivity of PAD4 and thus facilitate its activity within the synovial fluid and contribution to the pathogenesis of RA. In this issue of Science Translational Medicine, Darrah et al. (6) have tracked down an unexpected culprit that allows PAD4 to bypass its own requirement for very high Ca²⁺ levels. This allows the enzyme to retain high levels of activity outside of cells and to citrullinate proteins—a process that produces disease-exacerbating autoantibodies.

PAD4 AUTOANTIBODIES IN RA

In addition to its role in driving protein citrullination, PAD4 is itself a frequent antigen target in RA. Autoantibodies targeting PAD4 are detectable before disease onset and are often associated with the more erosive version of RA. The authors screened sera from RA patients with radiolabeled PAD3 and detected PAD3 autoantibodies in 18% of sera (6). Unexpectedly, they found that anti-PAD3 was exclusively present in RA patients that were also positive for anti-PAD4 and was not detected in healthy controls or patients with psoriatic arthritis. Competition experiments confirmed that the antibody to PAD3 was cross-reactive to PAD4 but not to PAD2. Last, PAD3/PAD4 cross-reactive autoantibodies were found primarily in RA patients that had the most erosive joint disease as compared with anti-PAD–negative patients or patients with antibodies to PAD4 only.

To investigate whether PAD3/PAD4 cross-reactive autoantibodies play a pathological role in RA, the authors tested their influence on the sensitivity of PAD4 to Ca²⁺ at a range of concentrations. In the absence of PAD3/PAD4 autoantibody, PAD4 activity was maximal at 5 mM Ca²⁺ and fell off sharply below 2 mM. Immunoglobulin G (IgG) isolated from PAD3/PAD4 cross-reactive patients enhanced PAD4 activity, increasing the rate of protein citrullination by an average of 500-fold as compared with that observed with PAD4-only sera at physiologically relevant Ca²⁺ concentrations. This suggests that cross-reactive autoantibodies can effectively stabilize the conformation of PAD4 protein that is ordinarily stabilized by Ca²⁺ binding (Fig. 1). PAD4 can bind up to five calcium ions. Two bind to the C-terminus and are vital for the formation of the active site. Three others are involved in protein-protein interactions and the structural stabilization of PAD4. A pivotal finding of the paper by Darrah et al. is that PAD3/PAD4 cross-reactive autoantibodies act to stabilize this latter region and thus reduce PAD4’s requirement for high Ca²⁺ concentrations in the synovial fluid (6).

• Download high-res image
• Open in new tab
• Download Powerpoint

To explore the effects of Ca²⁺ and the binding of cross-reactive autoantibodies on the structure of PAD4, the authors capitalized on the ability of the serine protease granzyme B to cleave PAD4. When PAD4 was exposed to PAD3/PAD4 autoantibodies before exposure to granzyme B, PAD4 proteolysis was inhibited, strongly supporting the conclusion that these autoantibodies reduce granzyme B cleavage of PAD4 by inducing structural changes in the protein (6). Thus, the binding of PAD3/PAD4 cross-reactive autoantibodies locks PAD4 into a conformation similar to that effected by Ca²⁺ and makes it favorable for catalyzing citrullination, without the requirement for supraphysiological Ca²⁺ concentrations.

INTERPLAY OF ANTIGENS AND AUTOANTIBODIES

The fundamental mechanism for PAD activation and target protein citrullination in RA described here by Darrah et al. (6) may be a common feature of many diseases that also have an autoimmune component. Cell death in regions of evolving pathology may serve as a source of citrullinated or otherwise modified antigens that trigger the production of autoantibodies. For example, PAD4 and citrullinated proteins are selectively coexpressed in pyramidal neurons that populate brain regions known to be most vulnerable to Alzheimer’s disease pathology—the same neurons that exhibit intraneuronal amyloid deposition, profound dendrite and synapse loss, and eventual cell death (1, 7).

Similar to the cells in the cartilage and bone in RA, neuron cell death and lysis in the AD brain leads to the release of their content of citrullinated proteins and their fragments into the brain interstitium and cerebrospinal fluid (CSF) (Fig. 1) (1, 7). These citrullinated products in the CSF are eventually returned to the blood and exposed to immune surveillance and are thought to drive the production of autoantibodies directed against the neuronal debris (Fig. 1) (1, 7). Breakdown of the blood-brain barrier, an inevitable outcome of many neurodegenerative diseases, allows autoantibodies targeting neuronal surface antigens to access the brain tissue and bind to the surfaces of neurons (7). Relentless autoantibody binding forces neurons to clear their surfaces via chronic endocytosis. This stresses these cells by driving intraneuronal amyloid deposition and depleting key cell surface proteins, resulting in neuronal functional deficits throughout the course of the disease (1, 7).

CLINICAL DOUBLE AGENTS

Previous studies have detected soluble PADs and citrullinated autoantigens in RA synovial fluid, suggesting that extracellular PADs remain active in this fluid (5). Given the relatively low levels of Ca²⁺ in the synovial fluid, the expectation is that PADs should not be active under these conditions. However, the work by Darrah et al. showed that PAD3/PAD4 cross-reactive autoantibodies played a key role in unleashing high levels of PAD4 activity within RA synovial fluid (6). The resulting increased generation of citrullinated proteins within synovial tissue and fluid provides a source of highly specific protein targets that trigger autoimmune responses associated with RA.

In addition, this work invites the possibility that PAD4 may be similarly regulated within cells where Ca²⁺ levels are even lower, although the identity of this endogenous modulation factor (or factors) (perhaps autocitrullination and other interactions at the N-terminal region of PAD4) and its mechanism of action remain to be discovered (8, 9). If such a factor were discovered, it would be a natural choice as a target for new therapies for RA. This study has identified PAD3/PAD4 as a previously unknown autoantibody biomarker that may be used to identify RA patients who are likely to develop the most erosive and progressive form of RA (Fig. 1). If so, then this will also identify those RA patients who would most benefit from a therapy aimed at PAD4 inhibition.

The use of disease-specific autoantibodies as biomarkers of specific pathology is not unique to RA. For example, recent studies using human protein microarrays have demonstrated that autoantibodies are far more abundant in the blood than previously thought: several thousand are detectable in any one individual, regardless of age, gender, or health status (10). What is the purpose of such a vast number of circulating autoantibodies? We have proposed that these autoantibodies function as natural debris-clearance agents for the body and that their number and titer reflect the level of the debris generated by the body due to normal wear and tear or ongoing pathology. Supporting this possibility, disease-induced alterations in autoantibody number and titer provide an avenue for accurate disease diagnostics, as has been shown recently in Alzheimer’s (10) and Parkinson’s diseases and as suggested by the work by Darrah et al. on RA (6).

It is becoming increasingly clear that autoantibodies play a “double agent” role in RA, neurodegenerative disease, and potentially other human diseases (Fig. 1). First, they function to enhance the clearance of disease-specific cell and tissue debris. Second, they may trigger or exacerbate disease by chronically binding to cell surfaces, instigating cell stress, cell death, and inflammation. This dual role of autoantibodies provides exciting opportunities for future research: on the one hand, by enabling disease detection and diagnosis, and on the other, by hinting at new possibilities for therapeutic intervention.