The Atypical Fibrin Fibre Network in Rheumatoid Arthritis and its Relation to Autoimmunity, Inflammation and Thrombosis

Objective The risk of cardiovascular events in patients with RA is disproportionately heightened as a result of systemic inflammation. The relative effect of autoimmune-associated citrullination on the structure and thrombotic potential of fibrin(ogen) remains unknown. We therefore compared indices of vascular function, inflammation, coagulation and fibrin clot composition in RA patients with healthy controls and evaluated inter-parameter relationships. Methods Blood samples were collected from 30 RA patients and 25 age- and gender-matched healthy volunteers. Levels of SAA, CRP, ICAM-1 and VCAM-1 was measured using a sandwich immunoassay. Whole blood coagulation was assessed using Thromboelastography. Fibrin clot networks and fiber structure was investigated using Scanning Electron Microscopy. The detection and quantification of citrullination in formed fibrin clots were performed using a fluorescently labeled Citrulline monoclonal antibody with Confocal Microscopy. Results Concentrations of SAA, CRP and ICAM-1 were significantly elevated in RA patients compared to controls. TEG parameters relating to coagulation initiation (R and K), rate of fibrin cross-linking (α-Angle), and time to reach maximum thrombus generation (TMRTG) were attenuated in RA patients. Parameters relating to clot strength (MA, MRTG, TGG) did not statistically differ between RA and controls. Logistic regression modelling revealed stronger association between acute phase reactants (CRP, SAA) with TEG parameters than endothelial function markers. Microscopic analysis revealed denser networks of thicker fibrin fibers in RA patients compared to controls [median (interquartile range) 214 (170-285) vs 120 (100-144) nm respectively, p<0.0001, Odds ratio=22.7). Detection of multiple citrullinated regions within fibrin clot structures in RA patients, which was less prevalent in control samples (p<0.05, OR=2.2). Conclusion Patients with active RA display a coagulation profile that is dissimilar to general findings associated with other inflammatory conditions. The alteration of protein structures by autoimmune linked citrullination could play a role in determining the structure of fibrin and the potential of conferring a heightened thrombotic risk in RA patients.


Introduction
Rheumatoid Arthritis (RA) is a chronic, systemic autoimmune disease characterized by both peripheral joint and extra-articular site inflammation, with an increased predisposition to a higher incidence of cardiovascular disease (CVD) (1,2). CVD (including stroke and myocardial infarction) is almost 50% more common in RA patients than the general population and is the most frequent cause of early mortality (3). Traditional risk factors for CVD (age, hypertension, obesity, etc.), do not fully account for the elevated occurrence of CVD events, and thus RA (genetics and disease characteristics) has been identified as a strong independent risk factor (4).
Key intermediaries of this manifestation are the structural components of formed thrombi. Soluble fibrinogen is cleaved by thrombin in order to form dense matrices of thin fibrous protein known as fibrin (24). Polymerized fibrin networks are essential for wound healing and other occlusive physiological processes (24). However, exposure to inflammatory biomarker stimuli [such as CRP (25), SAA (26), and proinflammatory cytokines (27,28)] can result in the alteration of mechanical and viscoelastic properties of fibrin clots into a prothrombotic phenotype. This phenomenon has previously been observed in RA plasma clots (29,30). Various immunopathogenic processes related to RA development can exert upstream amplification of the coagulation cascade as well as impairing fibrin clot dissolution (9,19,31).
Fibrin(ogen) is also a potent pro-inflammatory signaling entity itself, mainly through ligand-receptor interactions with immune cells that further propagates pro-inflammatory effects (32)(33)(34)(35)(36). The deimination of particular arginine residues in fibrin(ogen), known as citrullination, is a distinctive RA posttranslational modification that alters normal protein structure and function that confers antigenicity to modified proteins (37)(38)(39)(40)(41)(42)(43). The functional relationship between citrullination and the presence of a prothrombotic fibrin clot phenotype is still poorly understood. Some studies have shown that citrullination of fibrinogen prevents thrombin digestion and subsequent fibrinogenesis (44)(45)(46). However, the experimental conditions upon which these findings are based do not reflect physiological coagulation and is inconsistent with a predominantly hypercoagulable state seen in RA (47).
Inflammation-induced fibrin formation is equally present in RA synovial spaces as it is in circulation (Refer to Figure 1). Synovial coagulation is a key step in pannus formation, where fibrin provides the structural scaffold for immune cells that are responsible for synovial membrane disintegration and eventual joint damage (31,48). Endothelial tissue dysfunction is a key process that facilitates this ubiquitous distribution of aberrant fibrin deposition in both synovia and vasculature. This pathophysiological state is characterized by the expression of cell adhesion molecules [intercellular cell adhesion molecule-1 (ICAM-1) and vascular cell adhesion molecule-1 (VCAM-1)], pro-inflammatory cytokines and pro-thrombotic markers (49,50). This allows for the recruitment, translocation and propagation of inflammatory and thrombotic mediators across the synovial barrier (51)(52)(53).
There is significant overlap in inflammatory pathways responsible for joint damage in RA and hypercoagulation, coupled with the fact that disease severity has been correlated to more adverse cardiovascular complications (21,54,55). It is therefore prudent that these processes and their relevant markers be examined systemically in RA, and not isolated to either vascular or synovial compartment. The aim of this study was to examine the extent to which the coagulation profiles and fibrin network architecture of RA patients are influenced by acute phase inflammation, endothelial dysfunction an autoimmune-related protein modification.    Table 1.

Blood sampling
Whole blood (WB) samples were collected in vacutainer tubes using 3.8% sodium citrate as anticoagulant. Blood drawing on all participants was performed by a qualified nurse, or phlebotomist by sterile puncture of the antecubital vein. Blood tubes were incubated at room temperature for a minimum duration of 30 minutes prior to the commencement of any whole blood analysis. In order to obtain platelet poor plasma (PPP), sodium citrated blood tubes were centrifuged at 3000xg for 15 minutes, aliquoted into Eppendorf tubes and stored at -80 o C until further analysis.  Analyte concentrations were calculated from the calibration curve generated by absorbance measurements of manufacturer supplied calibrator standards.

Confocal laser scanning microscopy
In order to determine the extent of protein deamination in fibrin networks, PPP aliquots of RA samples (n=10) and control samples (n=10) were thawed and fibrin clots prepared (refer to SEM method) on glass microscope slides in a dark room.
Samples were fixed with 4% PFA, washed 3x with PBS, and blocked with 5% Goat serum solution (Abcam, ab7481) for 30 minutes. Clots were then stained with a 1:50 dilution Citrulline Monoclonal Antibody (2D3.1) (Thermo Fisher Scientific, Cat #MA5-27573) and incubated for 1 hour. Following another 3x PBS wash to remove unbound antibodies, samples were then stained with 1:200 dilution Goat Anti-Mouse IgG Secondary antibody conjugated to AlexaFluor 488 (Thermo Fisher Scientific, Cat #A327273) and incubated for 1 hour. Slides were washed 3x with PBS to remove unbound antibody, allowed to dry, and mounted with a glass coverslip. Samples were viewed with a Zeiss LSM 780 Confocal laser scanning microscope (CLSM) with a Plan-Apochromat 63x/1.4 oil DIC M27 objective. AlexaFluor488 was excited with the 488nm laser and emission was detected at 508-570nm. Three representative micrographs per sample were analysed for fluorescent particle distribution using ImageJ (Version 1.52p). Images were calibrated to scale, and a global threshold (27 pixel cut-off) applied to all analysed micrographs.

Statistical analysis
Statistical analysis was performed using R version 4.0. Specifically, univariate logistic regression was performed to determine odds ratios (OR) for experimental variables using the logistic model in the rstanarm package (with default priors). ORs and 95% confidence intervals were extracted in the corresponding unit system (i.e. not z-scaled) for all variables except Fibrin fiber diameter and citrulline particle number shown in Table 4 which are z-scaled to aid interpretation. Tables 2 and 3 show ORs after adjustment for age and gender with unadjusted analysis identifying the same significance and effect sizes. In addition, results for classical statistical tests are reported as follows. The distribution of sample datasets for each variable experimental was determined using the Shapiro-Wilk test. Accordingly, p-values for each variable comparing RA to healthy controls were calculated using either a Mann-Whitney U test for nonparametric data or a Student t-test for parametric data.
Statistical significance was set at p<0.05. One can see close alignment between all these and the OR results.

Subjects
Demographic information of all study participants is listed in Table 1. The RA sample group closely resembles the general population distribution for age (median: 54 years) and sex (80% female) of the disease (58). The control group of healthy volunteers was closely matched to the RA group with regards to age (median: 50 years) and sex (73% female). The RA sample group was heterogeneous with respect to clinical presentation, with most patients on an anti-rheumatic drug therapy regime. The majority of RA patients also presented with positive titers for anti-cyclic citrullinated peptide (CCP) (77%) and rheumatoid factor (97%) autoantibodies.  (59). For full demographic information and specified list of disease modifying anti-rheumatic drugs (DMARDs) please refer to full study database in supplementary material.

Confirmation of altered inflammatory and vascular function profile in RA
Circulating concentrations of endothelial function and acute phase markers are shown in Table 2. Previous studies have shown these markers (CRP, SAA, sVCAM-1 and sICAM-1) to be associated with a prothrombotic state and increased CVD risk in RA (7,13,(60)(61)(62), and was therefore measured in this study to determine the extent to which systemic inflammation influences viscoelastic and structural clot

Functional coagulation assessment indicates a prothrombotic tendency in RA
Whole blood coagulation parameters as measured by TEG ® are listed in Table 3.
Limited viscoelastic assessment of coagulation in RA has been performed to date (60,63,64), with TEG ® not commonly used in rheumatology practice (65)(66)(67) Measures of overall clot strength (MA) and growth (TTG) were attenuated in RA but did not statistically differ from those of controls.      Figure   5A (Healthy control sample), which demonstrates thinner protein strands and a more permeable fibrin network. Evaluation of plasma clots using image analysis software (ImageJ) based techniques for representative RA (n=10) and Healthy control (n=10) subjects. Fibrin fiber diameters were determined from SEM micrographs, while fluorescent particle analysis was calculated from CLSM micrographs. Values are expressed as median [interquartile range]. Scaled odds ratios for the logistic regression model are listed with their 95% confidence intervals, while level of statistical significance was set at p<0.05 for calculated Mann-Whitney test values. *Denotes statistically significant differences in variables between RA and controls.

Discussion
There is a need to bridge translational gaps between RA immunopathogenesis and systemic vascular and hemostatic irregularities. The development of RA autoimmunogenicity precedes the onset of joint disease (69,70). The link between RA autoimmune patterns and its possible role in exacerbating thrombosis is still poorly understood. Crosstalk between immune and hemostatic systems with the endothelium represents a critical interface in which both arthritic and cardiovascular pathologies are initiated and propagated. We therefore analyzed a panel of biomarkers that are representative of this dynamic milieu, is associated with RA disease severity and CVD.
Levels of both acute phase reactants (CRP and SAA) were significantly elevated in RA patients ( Table 2) and showed a strong association (Figure 4). This was expected as acute phase reactant concentrations rise dramatically under acute inflammatory states, with both CRP and SAA shown to reliably predict disease severity and CVD risk in RA (13,14,22,71,72). CRP can bind to surface receptors on monocytes, endothelial cells and platelets thereby altering their function to propagate hypercoagulable conditions (73,74). The ability of CRP and SAA to induce TF expression has been demonstrated in various cell types (13,14,22,71,72,(75)(76)(77)(78)(79)(80)(81) and by in vivo studies (82,83). Additionally, acute phase reactants can also suppress fibrinolysis by promoting expression of plasminogen activator inbibitor-1 (PAI-1) (82) and inhibiting TF pathway inhibitor (TFPI) expression (81,83). CRP is also able to activate endothelial cells and cause the expression of cell adhesion molecules (ICAM-1 and VCAM-1) (84,85). It should however be noted that cellular effects of CRP have been disputed as being primarily caused by bacterial contaminants in CRP preparations rather than the protein itself (86). SAA is the precursor to amyloid A (AA) protein, which form insoluble fibrillar depositions in major organs as a result of long-term inflammation (87). RA has been prominently implicated as a pre-existing condition for the development of potentially fatal AA amyloidosis (88,89). Known for being primarily hepatically synthesized, synovial tissue cells (90)(91)(92)(93) and chondrocytes (94) can be articular sources of SAA production. Within the synovium SAA promotes pro-arthritic processes through the expression of cell adhesion molecules (95,96), cytokines (96), and matrix degrading enzymes (92,95) by local tissue. SAA showed slightly weaker association with sICAM-1 compared with CRP in our studies (Figure 4).
Levels of soluble cell adhesion molecules (sICAM-1 and sVCAM-1) are the most accessible form in which to determine endothelial activity. They are also strong predictive biomarkers for CVD in RA (61,62,97). Increased levels of CAMs indicate endothelial dysfunction that facilitates pro-inflammatory and prothrombotic conditions (7). ICAM-1 is a prominent receptor for fibrinogen (98), with interaction fortifying endothelial adhesion and migration of leukocytes (99)(100)(101), endothelial-platelet adhesion (102)(103)(104), and causes vasoconstriction (105,106). ICAM-1 signaling pathways can also promote endothelial tissue factor expression (107). The role of VCAM-1 in directly promoting coagulation is not as well understood, but does contribute to atherosclerotic plaque formation as a result of adhering to PBMCs (108, 109) sICAM-1 and sVCAM-1 concentrations were elevated in RA ( Table 2) but were not as strongly associated with coagulation function as CRP and SAA (Figure 4).
This belies the fact that CAMs present on cell surfaces play more facilitative roles in thrombotic diseases rather than instigating them. Our findings show that coagulation initiation was amplified in RA patients with shortened velocity parameters of clot formation (R, K, α , TMRTG) ( Table 3) and these indices were moderately associated with levels of SAA and CRP (Figure 3).
Parameters relating to clot strength (MA, TTG) were attenuated in the RA sample group, but did not statistically differ from healthy controls ( Table 3). Thus, although the blood clots form rapidly it leads to a weak clot. The mechanics of clot formation as measured by TEG ® have been related to the risk of adverse ischaemic events (111). Excessive hepatic production of fibrinogen is highly prevalent in RA (10).
Increased plasma fibrinogen concentration is a strong contributing factor to hypercoagulation (112). Fibrin(ogen) is susceptible to structural and functional modifications by certain inflammatory molecules, including CRP (25), SAA (26), and certain bacterial virulence factors (113)(114)(115). Fibrin(ogen) is also prone to posttranslational modification that relates to the generation of auto-immunogenicity in RA -the relevance of this process was investigated and is discussed below.
Evaluating fibrin gel matrices visually can reveal much about thrombotic potential under inflammatory conditions. Denser fibrin fiber networks are accompanied with increased resistance to fibrinolysis and is associated with the risk for thrombotic events (reviewed by Undas and Ariëns, 2011) (116). These structural properties can be viewed and functionally assessed with a high degree of resolution using SEM (57) Our analysis revealed denser fibrin networks in RA prepared ex vivo PPP clots compared to controls (Figures 2 and 5). This is consistent with a prothrombotic phenotype observed in previous studies that have inspected the fibrin network in RA (29,30). Furthermore, we determined a relative measure of fibrin fiber diameter using an image analysis software-based technique (Figure 2). The average diameter of fibrin fibers was larger in RA clots compared to controls (Figure 7). This can be clearly seen in the presented micrographs, were there is an apparent amalgamation of single fibrin fibers in RA preparations (Figures 2 and 5).  (120). It is apparent that our investigation into the structural properties of fibrin networks in RA and its relation to hemostatic function has a rather deceptive appearance. Our group has previously shown that fibrin networks in multiple inflammatory conditions (113,(121)(122)(123) and in plasma exposed to inflammatory stimuli (26,114,124)  PADs are usually localized to intracellular environments and requires calcium for catalysis (127). PADs become active under inflammatory and apoptotic conditions where enzymes are liberated to extracellular spaces and exposed to sufficient calcium concentrations for catalytic function (126). Two isoforms of PAD (PAD2 and PAD4) are primarily responsible for generating citrullinated antigens that incite the generation of anti-citrullinated protein antibodies (ACPAs), a hallmark serological feature of RA (128). Fibrinogen and fibrin are prominent substrates for PAD2 and PAD4 and autoantibodies targeting citrullinated fibrin(ogen) have been identified (40,43,68,(129)(130)(131). The pathogenicity of citrullinated fibrin(ogen) immune complexes have been demonstrated both in vitro (33) and in vivo (132,133). Citrullinated fibrin deposits are also common manifestations within synovial cavities, where it contributes to self-perpetuating inflammatory processes (48,134). Our findings provide novel evidence for the citrullination of fibrin within vasculature which is more prominent in RA plasma compared to control plasma (Figures 6 and 7). Previously the presence of citrullinated fibrinogen could only be detected in RA synovial fluid (41). Later research by  confirmed the presence of citrullinated fibrinogen containing immune complexes in RA plasma (131). The insolubility of fibrin may increase the likelihood of it being citrullinated in circulation. As we could not stain specifically for citrullinated residues in fibrin only, detected fluorescence could also have originated from citrullinated histone derived from neutrophil extracellular traps (135). Nevertheless, binding of ACPAs to fibrin could render it less degradable, by decreasing available binding surface to plasmin (136). There remains conjecture as to the effect of citrullination on hemostatic outcome. Citrullination of proteins results in structural unfolding (42) and loss of function (137), which increases its antigenic shelf-life. It has been demonstrated that citrullinated fibrinogen is resistant to thrombin digestion, as preferential epitopes for PADs overlap with thrombin binding sites (45,46,138). Despite this, fibrinogenesis in RA is by no means impaired, as evidenced by this study and others. It is plausible that high levels of fibrinogen (10) and thrombin activity (17,139) in RA has much stronger influence on the fate of fibrinogen than PAD enzymes. There is also evidence that upstream coagulation factors and fibrinolytic components are susceptible to citrullination (140,141). It is therefore difficult to predict a hemostatic endpoint based on overall citrullination and the effect of citrullination on thrombosis cannot be postulated on singular reactions. The implications that citrullination could have on fibrin, being the end-product of coagulation and a major determinant of thrombotic risk, remains intriguing and should be further investigated.
Future investigations would be to determine the effect of citrullination on the thrombotic potential of formed thrombi. Our research has previously implicated the role of amyloid resembling structures in fibrin clots that confer a prothrombotic phenotype. There is evidence to suggest that the processes of citrullination and amyloidogenesis may overlap. Citrullination has been shown to affect the aggregation and oligomerization of β -amyloid proteins (142) and that citrullination of human myelin oligodendrocyte glycoprotein (MOG) can lead to amyloid-like behavior shift that has pathogenic implications for multiple sclerosis (143).
This study did present some limitations and challenges. The relatively small sample population complicates the determination of correlative associations between inflammatory and hemostatic indices. The determination of true overall fibrin diameter was also not possible with current techniques -statistical analysis revealed a measurement accuracy of 82%. The detection of citrullination in fibrin networks using fluorescent techniques were unspecific to fibrin(ogen) in this study. However, the current study included this analysis only as a preliminary probe into determining if citrullination of clots was a discerning factor between RA and non-RA individuals.

Concluding remarks
Inflammatory and thrombotic processes are highly pertinent to the development of

Competing interests
The authors declare that they have no competing interests.

Consent for publication
All authors approved submission of the paper.

Funding
We thank the Medical Research Council of South Africa (MRC) (Self-Initiated Research Program: A0X331) for supporting this collaboration.

Raw data sharing
All raw data is available at: https://1drv.ms/u/s!AgoCOmY3bkKHiqIaVlNkRzHB8cUo6Q?e=FusgPo     Confocal micrographs of representative control and RA samples. Each row represents identical areas captured of fibrin clot preparations using transmitted light to illustrate fibrin network topography (left-hand column) and a green channel (middle column) to identify fluorescent particles corresponding to citrulline residues. Fluorescent images are also provided in binary (right-hand column) to better illustrate differences seen between RA and control samples with respect to citrulline-coupled fluorescence. Particle analysis confirmed the observable presence of enhanced fluorescent signal in RA samples (n=10) versus healthy controls (n=10).