A novel interaction between extracellular vimentin and fibrinogen in fibrin formation

Abstract

Introduction

Thrombosis is frequently manifested in critically ill patients with systemic inflammation, including sepsis and COVID-19. The coagulopathy in systemic inflammation is often associated with increased levels of fibrinogen and D-dimer. Because elevated levels of vimentin have been detected in sepsis, we sought to investigate the relationship between vimentin and the increased fibrin formation potential observed in these patients.

Materials and methods

This hypothesis was examined by using recombinant human vimentin, anti-vimentin antibodies, plasma derived from healthy and critically ill patients, confocal microscopy, co-immunoprecipitation assays, and size exclusion chromatography.

Results

The level of vimentin in plasma derived from critically ill subjects with systemic inflammation was on average two-fold higher than that of healthy volunteers. We determined that vimentin directly interacts with fibrinogen and enhances fibrin formation. Anti-vimentin antibody effectively blocked fibrin formation ex vivo and caused changes in the fibrin structure in plasma. Additionally, confocal imaging demonstrated plasma vimentin enmeshed in the fibrin fibrils. Size exclusion chromatography column and co-immunoprecipitation assays demonstrated a direct interaction between extracellular vimentin and fibrinogen in plasma from critically ill patients but not in healthy plasma.

Conclusions

The results describe that extracellular vimentin engages fibrinogen in fibrin formation. In addition, the data suggest that elevated levels of an apparent aberrant extracellular vimentin potentiate fibrin clot formation in critically ill patients with systemic inflammation; consistent with the notion that plasma vimentin contributes to the pathogenesis of thrombosis.

1. Introduction

Thrombosis is frequently seen in critically ill patients with systemic inflammation including sepsis, trauma, burns, and infection with the SARS-CoV-2 virus that causes COVID-19 [1], [2], [3]. Similarities in the prothrombotic and antifibrinolytic state in these patients include an enhanced fibrin formation and widespread fibrin deposition in small to mid-size blood vessels that eventually cause organ ischemia and dysfunction [2], [4], [5]. The high levels of fibrinogen and D-dimer, a fibrin degradation product, reveal active fibrin formation and fibrinolysis [6], [7]. Moreover, autopsy findings from patients who succumbed to sepsis-induced coagulopathy [8] or COVID-19 [9], [10], [11] often reveal disseminated fibrin-rich microthrombi suggesting the association between widespread fibrin deposition and poor outcomes.

We and others have reported that vimentin, a cytoplasmic intermediate filament protein [12], is expressed on the cell surface of different cell types [13], [14], [15], [16] and is also detected in blood [16], [17]. An increasing number of studies have described cell surface vimentin as a receptor for viruses, bacteria and plasma proteins [13], [16], [18], [19], [20], [21], [22]. On the other hand, extracellular vimentin has been reported to be found in blood from healthy subjects and at elevated levels in plasma from patients with clinical conditions such as coronary artery disease [23], rheumatoid arthritis (citrullinated vimentin) [24], cancer [25], [26], and sepsis [27]. However, the function of extracellular plasma vimentin remains elusive.

In an effort to understand the mechanisms of systemic inflammation-associated coagulopathy, we performed a study that, to our knowledge, is the first one to characterize the potential relationship between vimentin and fibrinogen and the impact of this association to coagulation. The objective of our study was to explore the novel role of plasma vimentin in fibrin formation, particularly, in clinical conditions associated with systemic inflammation such as sepsis and COVID-19.

2. Materials and methods

2.1. Plasma from patients and healthy donors

The plasma samples from critically ill patients used in this study were obtained in a previous study focused on fibrin clot structure in critical illnesses [28]. To obtain healthy human blood, informed consent was provided based on the recommendations of the Declaration of Helsinki. Approval was attained from the Baylor College of Medicine institutional review board (IRB) for these studies.

2.2. Reagents

Purified recombinant human vimentin was purchased from SinoBiological (Wayne, PA). Human fibrinogen from Calbiochem (Gibbstown, NJ), and thrombin from Sigma (St. Louis, MO). Sheep anti-Vimentin antibody was obtained from Affinity Biologicals (Ancaster, Canada), and the rabbit anti-Vimentin antibody was purchased from Proteintech (Rosemont, IL). This rabbit anti-Vimentin antibody was validated by testing its reactivity to purified recombinant vimentin and plasma vimentin by ELISA and western blot. Anti-vimentin antibody V9 and isotype immunoglobulins (IgGs) were purchased from Sigma. Recombinant human vimentin rod-domain (rhRod) was expressed and purified as described [29].

2.3. Gel filtration chromatography

The plasma samples were subjected to size exclusion chromatography using a Superose 6™ 10/300 GL column (Cytiva) equilibrated with 25 mM Tris-HCl, 150 mM NaCl, pH 7.4 (TBS) and a constant flow rate of 0.5 ml/min [30]. The collected fractions were analyzed by immunoblotting to verify the presence of vimentin [16].

2.4. Bio-layer interferometry

We used bio-layer interferometry for kinetic binding measurements as previously described [29]. Briefly, we immobilized purified human fibrinogen 50 μg/ml onto amine reactive 2nd generation sensors (AR2G; Pall ForteBio LLC) and the sensors were used to assess binding kinetics to recombinant human vimentin at concentrations ranging from 0 to 2500 nM. We analyzed the data using the Octet Data Acquisition Software version 11.1.

2.5. ELISA to measure plasma vimentin

Sheep anti-vimentin antibody was diluted to 1.0 μg/ml with carbonate buffer (pH -9.6), added into microtiter wells (50 μl/well) and incubated overnight at 4 °C. After washed and blocked with 3 % BSA (150 μl/well) for 1 h at 37 °C, 50 μl of diluted plasma (1:20) or increasing concentrations of purified vimentin (for standard curve) in Tris-buffer saline-0.05 % Tween-20 (TBS-T) were incubated in the coated wells for 1 h at 37 °C. After washed the wells three times with TBS-T, they were incubated with rabbit anti-vimentin antibody (1:1000 in TBS-T- 50 μl/well) at 37 °C for 1 h. The wells were washed four times with TBS-T and incubated with a secondary goat anti-rabbit IgG antibody (1:10,000 in TBS-T) horseradish peroxidase (HRP) conjugate at 37 °C for 1 h.

2.6. Fibrin porosity

As described [28], [31], the fibrin clot was visualized with confocal microscopy and porosity was quantified using Image J software.

2.7. Fibrin polymerization assays

Fibrin formation and degradation were performed as we described [31]. Briefly, 20 % (16 % for fibrinolytic studies) plasma in Tris buffered saline (TBS) (50 mmol/l Tris, 0.15 mol/l NaCl, pH 7.4) was mixed with anti-vimentin antibody or isotype IgG in the presence or absence of tissue plasminogen activator (tPA, 250 ng/ml) (Cathflo Activase). Polymerization and/or fibrinolysis was initiated with the addition of 1.0 U of human thrombin. When using plasma, the enzymatic reaction was evaluated by tracking turbidity using a spectrophotometer set to 350 nm. In another set of experiments of fibrin formation and degradation, we adjusted the volume of plasma to load fibrinogen at a concentration (0.5 mg/ml) that was equivalent between healthy donors and patients. In addition, we also performed experiments of fibrin polymerization in which we adjusted the concentration of fibrinogen (3.4 mg/ml) to be equivalent between healthy donor and patient by adding purified human fibrinogen to healthy plasma.

The fibrin polymerization of a purified system (human fibrinogen, calcium, and thrombin) was performed using the recombinant human vimentin or buffer with 1.0 U of human thrombin and 20 mmol/l CaCl₂ (final concentrations) [31]. The enzymatic reaction using the purified system was evaluated by tracking turbidity using a spectrophotometer set to 405 nm.

2.8. Imaging of the fibrin-clot structure

Briefly, plasma was supplemented with 2 % (w/w) of human fibrinogen conjugated to Alexa Fluor 488 (Thermo Scientific). Clot formation was initiated with the addition of 1 U thrombin (EMD Millipore) in the presence of 2.4 mM calcium as described [31].

For fibrin-vimentin interaction, plasma was supplemented with 2 % (w/w) of human fibrinogen conjugated to Alexa Fluor 647 (Thermo Scientific) and with mouse anti-human Vimentin (V9, Sigma-Aldrich) or mouse IgG (isotype control) prior to initiating clot formation with the addition of 1 U thrombin (EMD Millipore) in the presence of 2.4 mM calcium as described [31]. Fibrin clots were washed three times with phosphate buffered saline (PBS) and incubated immediately with a secondary anti-mouse IgG-CF488A (1:1000, Sigma) conjugate for 1 h. Then, fibrin-clot were washed three times with PBS and fixed with 4 % paraformaldehyde for 10 min and washed three times for visualization by confocal microscopy.

2.9. Statistical analysis

GraphPad Prism 8 software (San Diego, CA) was used to perform statistical analyses. Comparisons between groups were conducted by t-test and ANOVA. P values were 2-sided, and statistical significance was determined by a P value <0.05.

3. Results

3.1. Vimentin binds to fibrinogen and enhances the fibrin formation potential

We and others have reported the presence of circulating, extracellular vimentin [16], [17] whose levels become elevated in sepsis, a condition associated with marked inflammation [27]. Thus, we first determined the levels of vimentin in plasma from healthy donors (the median for fibrinogen level in healthy subjects used in this study: 2.86 (2.1–3.6) mg/ml) and critically ill patients (median: 3.54 (0.9–8.2) mg/ml), for further clinical characteristics see reference [28]. By using ELISA, we confirmed that critically ill [27], systemically inflamed patients had significantly elevated circulating vimentin levels as shown in Fig. S1 (patients 543.5 ± 253 ng/ml (n = 28) vs. healthy subjects 325.7 ± 103 ng/ml (n = 10), mean ± SD, ***p < 0.0006). High levels of plasma vimentin in parallel with hypercoagulability associated with aberrant fibrin structures [28] seen in critically ill patients opened the possibility of an interaction between fibrinogen and vimentin. Biolayer interferometry demonstrated binding of recombinant vimentin to purified fibrinogen in a concentration-dependent and saturable manner with a binding constant of K_D = 580.0 ± 0.9 nM (Fig. 1A and B). Additionally, the blot overlay (Fig. S2) suggests that vimentin may preferentially bind the reduced/denatured form of fibrinogen alpha chain. This interaction of vimentin with fibrinogen may influence fibrin polymerization; thus, we tested this hypothesis using a purified system [31], with recombinant vimentin at concentrations comparable to those found in plasma from healthy subjects (Fig. S1). Recombinant vimentin significantly increased fibrin formation potential (maximum absorbance) in the purified system in a dose dependent manner (Fig. 2A). Note that the initial rate of change (slope) increased at the highest concentrations of vimentin (4.0 and 5.0 nM), although the protein did not have an apparent effect in the lag time (Supplemental Table 1). The initial rate of change was determined as described [31]. To examine whether the capacity of vimentin to potentiate fibrin formation depends on the levels of fibrinogen, we performed fibrin polymerization in a purified system by mixing increasing concentrations of purified human fibrinogen (0.3, 0.5, and 0.7 mg/ml) with buffer or a fixed concentration of recombinant vimentin (5.0 nM). Fig. 2B shows that, in comparison to samples without vimentin, vimentin positively affected fibrin formation (greater maximum absorbance) with increasing concentration of fibrinogen, suggesting an interaction between vimentin and fibrinogen. Importantly, it is also possible that the changes observed in the turbidity assays (Fig. 2A and B) represent changes in the resultant fibrin clot structure, which was examined below. We further tested our hypothesis of functional interaction between vimentin and fibrinogen by measuring fibrin formation in the presence of anti-vimentin antibodies. In comparison to the corresponding isotype IgG (Fig. S3A), fibrin formation was blocked with a rabbit-polyclonal anti-vimentin antibody and a mouse-monoclonal (V9) antibody by 60 % and 15 %, respectively, (Fig. S3B). The sheep anti-human vimentin antibody did not inhibit the effect of the recombinant vimentin on fibrin formation (Fig. S3B). These different outcomes may be due to the different epitopes recognized by each antibody in the vimentin structure (Fig. S4A), and suggest a putative binding site for fibrinogen within the rod domain of vimentin. We tested this postulate by characterizing the binding of our recombinant human vimentin rod domain (rROD) [29] to fibrinogen. Fig. S4B demonstrates that rROD bound to fibrinogen with a binding constant of K_D = 260 ± 0.6 nM. These results point to a specific interaction between vimentin and fibrinogen and that the vimentin rod domain may interface with fibrinogen alpha chain.

Fig. 1.

Recombinant vimentin binds to fibrinogen. (A) BLI was performed with immobilized fibrinogen. The phases of measurement are shown as association phase and dissociation phase. (B) Steady state binding analysis curve of Req versus concentration of recombinant vimentin to fibrinogen. Vimentin interacted with a binding constant of K_D = 580 nM ± 0.9 nM, mean ± SEM of three determinations.

Fig. 2.

Vimentin enhances fibrin formation potential. (A) Fibrin was formed from purified human fibrinogen (0.3 mg/ml), calcium, thrombin and recombinant vimentin protein was added at concentrations as indicated. Turbidity was measured at 405 nm. Each curve is the average of four separate experiments. Higher levels of recombinant vimentin (4.0 and 5.0 nM) resulted in significantly more fibrin compared to the absence of vimentin (**p < 0.001). (B) Fibrin formation was performed as described for A. Increasing concentrations of human fibrinogen (indicated) were mixed with buffer or recombinant vimentin (5.0 nM). Each curve represents the average of three separate experiments. The positive effect of vimentin on fibrin polymerization increments the maximum absorbance as the fibrinogen concentration increases (*p < 0.01; ***p < 0.0004).

3.2. Extracellular plasma vimentin contributes to fibrin formation and is incorporated into the fibrin clot

Fibrinogen levels are elevated in systemic inflammation [7], [32]. As vimentin is more abundant in systemic inflammation and enhances fibrin formation in vitro, we hypothesized that the effect of extracellular vimentin on fibrin polymerization is greater in plasma from critically ill patients with systemic inflammation than in healthy plasma. To test this hypothesis, we used anti-vimentin antibody to block the effect of extracellular vimentin in fibrin polymerization ex vivo. The rabbit anti-vimentin antibody, which reduced fibrin formation in plasma in a dose dependent manner (Fig. S5), inhibited fibrin formation in plasma from patients by 50.1 ± 8.1 %, mean ± SD (n = 6) (Fig. 3A and C). Note that the antibody had a greater inhibitory effect in plasma from patients as compared to plasma from healthy donors (inhibited 20 ± 6.2 %, mean ± SD, n = 3) (Fig. 3B and C), indicating that vimentin may participate in clot formation even at low levels. To reduce the variability in fibrin formation among patients and healthy donors, we tested for differences in the effect of the antibody between healthy and patients using an equivalent fibrinogen concentration (3.4 mg/ml). Fig. S6B clearly demonstrates that the rabbit anti-vimentin antibody reduced fibrin formation (maximal absorbance) by 42 % in patient plasma and 27 % in healthy plasma containing the same fibrinogen concentration and volume of plasma. Note that the antibody had an inhibitory effect comparable to those shown in Fig. 3C using plasma with different levels of fibrinogen. This outcome indicates that apparently the antibody preferably interacts with plasma vimentin independent of the fibrinogen level.

Fig. 3.

Anti-vimentin antibody reduced fibrin formation in plasma. The tracings shown representative experiments of plasma from (A) patient or (B) healthy subject. Fibrin was formed from 20 % plasma. Plasma was mixed with rabbit (Rb) anti-vimentin antibody or isotype IgG (20 μg/ml). Turbidity was measured at 350 nm. (C) Anti-vimentin antibody blocked fibrin formation in plasma from critically ill patients by 50.8 ± 8.1 %, mean ± SD (n = 6 patients), while it blocked in healthy plasma by 20 ± 3.5 %, mean ± SD (n = 3 donors).

Next, we investigated the effect of blocking plasma vimentin on the resultant fibrin network structure. Figs. 4A and S7 show representative clot structures generated from plasma of a healthy subject and critically ill patients. The antibody against vimentin provoked a change in the resultant fibrin network structure. It increased the area unoccupied by fibrin formed from plasma, suggesting an increase in porosity area (Fig. 4B). In comparison to isotype IgG, the anti-vimentin antibody at 2.5 μg/ml significantly increased the porosity area of the fibrin network in plasma. This change in the clot architecture also suggested that vimentin could be incorporated into the resultant network. To test our speculation that vimentin is enmeshed in the clot, we visualized fibrin-associated vimentin from plasma of critically ill patients by confocal immunofluorescence microscopy. Punctate staining for vimentin was found distributed along the length and branching points of the fibrin fibrils and at sites of fibrin fibrils branching (Fig. 5 , lower panels), illustrating the incorporation of vimentin into fibrin. Since the fibrin clot architecture influences fibrinolysis [33], we further examined the effect of vimentin on fibrinolytic susceptibility to plasmin. Fig. 6A depicts representative fibrin polymerization/proteolysis curves of a patient plasma in the presence of tissue plasminogen activator. In comparison to isotype IgG, analyses of the area under the curve (AUC) revealed that the anti-vimentin antibody significantly attenuated fibrin formation, consistent with fibrin polymerization studies described above (Fig. 3) and thereby reducing the fibrinolytic burden in plasma from critically ill patients (Fig. 6B) and healthy subjects (Fig. 6C). To reduce the variability in fibrin formation and degradation (AUC) among patients and healthy donors, we tested for differences in fibrinolytic susceptibility between healthy and patients using an equivalent fibrinogen concentration (0.5 mg/ml). Consistent with the fibrinolysis assays starting with the same volume of plasma, Fig. S8 clearly shows the inhibitory effect of the anti-vimentin antibody in the absence of any variation in plasma fibrinogen concentration. These results reveal a novel finding that extracellular vimentin in plasma plays a role in fibrin clot and structure formation, possibly impeding efficient fibrinolysis.

Fig. 4.

Anti-vimentin antibody induced changes in the fibrin clot structure. (A) Representative confocal microscopy images of fibrin clots at magnification of 120× formed in plasma from healthy subject or critically ill patient. The anti-vim antibody clearly changed the architecture of resultant clot network in the plasma from the healthy and patient at 2.5 μg/ml (right panel). (B) Anti-vimentin antibody changed the resultant fibrin network structure in plasma from healthy donors (blue symbols) and critically ill patients (black symbols), increasing the area of unoccupied by fibrin formed from plasma (increased porosity area). (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

Fig. 5.

Localization of vimentin in plasma fibrin clot. Representative of three experiments. The confocal microscopy images show fibrin clots at magnification of 177× formed from patient plasma. Plasma was supplemented with human fibrinogen conjugated to Alexa Fluor 647. The lower-left panel shows the optical section through the plasma clot stained for vimentin with the monoclonal mouse anti-vimentin antibody (V9), and detected with a FITC-conjugated secondary antibody. The upper-left panel shows fibrin stained with an isotype control to anti-vimentin and with the same fluorescent secondary as the bottom right panel. Note the staining pattern for vimentin along the fibrin fibrils and the lower-right panel shows the punctate foci for vimentin (white arrows). Scale bar, 5 μm.

Fig. 6.

Anti-vimentin antibody affected fibrin formation and fibrinolysis in plasma. (A) Fibrin formation and fibrinolysis were initiated with thrombin, calcium, and in the presence of recombinant tissue-type plasminogen activator, Rb anti-vim antibody or isotype IgG (20 μg/ml). Turbidity was measured at 350 nm. Representative curves for fibrin formation and fibrinolysis; each curve is the mean of duplicate reactions using plasma of a patient. B and C: Measurements of both fibrin formation and fibrinolysis are presented as the area under the curve (AUC). The Rb anti-vim antibody significantly reduced fibrin formation and thus diminishing the clot burden to lyse in plasma from patients (***p = 0.0003), and healthy donors (*p = 0.03).

3.3. Circulating, extracellular vimentin is in complex with plasma fibrinogen

Compared to fibrin formation using purified proteins (Fig. 2A), plasma from critically ill patients robustly formed fibrin with almost twice the concentration of vimentin (recombinant 5 nM = ~285 ng/ml in Fig. 2A vs. endogenous 9.5 nM = ~540 ng/ml in Fig. S1), suggesting the possibility of a strong, heteromeric interaction between endogenous vimentin and fibrinogen in circulation. To investigate this possibility, we fractioned patient or healthy plasma by size-exclusion chromatography and analyzed the eluates by Western blot using an anti-vimentin antibody as we previously reported [16]. In contrast to healthy plasma, vimentin in plasma from critically ill patients co-eluted with fibrinogen from the column (Fig. 7A, inset, and Fig. S9A-C). A Coomasie-blue stained SDS-PAGE under reduced conditions demonstrated the presence of fibrinogen for both patient and healthy plasma in their corresponding elution peaks (Fig. S9D). Co-immunoprecipitation of fibrinogen with vimentin from eluates of peak 2 from critically ill patients but not healthy donors confirmed the interaction between circulating vimentin and fibrinogen (Fig. 7B) and suggested that this interaction may be part of the inflammatory spectrum.

Fig. 7.

Gel filtration and Western blot analyses of extracellular vimentin in plasma. (A) Plasma from a critically ill patient or a healthy subject was subjected to gel filtration to separate vimentin (expected MW = 57 kDa) from fibrinogen (expected MW = 340 kDa). Using a Superose 6 gel filtration column with a TBS mobile phase, the proteins were detected by their absorbance at 280 nm. The graph represents the elution profile for a patient plasma showing the elution time for fibrinogen (elution peak). mAu, milliabsorbance units. Inset, in sharp contrast to healthy plasma, the Western blot verified the presence of extracellular vimentin in the fibrinogen elution peak for a critically ill patient plasma. Figure represents one of four independents experiments. (B) Binding of extracellular vimentin to fibrinogen in solution (elution peak) was further demonstrated by co-immunoprecipitation. Vimentin was immunoprecipitated (IP) from gel-filtered fractions corresponding to the elution peak and immunoblotted for fibrinogen.

4. Discussion

Coagulopathy is a clinical manifestation seen in critically ill patients with systemic inflammation, including sepsis (bacterial, fungal or viral), trauma, and burns [3], [8], [34]. Several factors have been proposed to be associated with the hypercoagulable state in systemic inflammation. Among the markers of coagulopathy in these patients are fibrinogen and D-dimer, a fibrin proteolytic product [5], [6]. Elevated levels of these markers are most likely indicative of an active fibrin formation and fibrinolysis. This study describes a novel role for extracellular vimentin in interacting and actively participating in fibrin formation, particularly in a greater magnitude in plasma from critically ill patients with systemic inflammation.

This is the first study to characterize the binding of vimentin to fibrinogen by several means. First, we used a purified recombinant human vimentin to demonstrate the interaction by binding kinetic and fibrin polymerization studies. Our data suggest that the vimentin rod domain and fibrinogen alpha chain comprise this interaction. Second, the interaction between endogenous vimentin and fibrinogen in plasma was revealed with co-immunoprecipitation assays, indicating this heterocomplex exits in solution. Lastly, the capacity of an anti-vimentin antibody to block fibrin polymerization in plasma validated the relationship between vimentin and fibrinogen in circulation. The vimentin-fibrinogen interaction seems to be stronger in the context of systemic inflammation because vimentin, which has a molecular mass of ~57 kDa co-eluted with fibrinogen (molecular mass of ~340 kDa) from the size exclusion chromatography. It is possible that the extracellular vimentin protein forms a complex with fibrinogen in plasma as implied by the co-immunoprecipitation assay of the co-eluted proteins. If the anti-vimentin antibody does not displace the fibrinogen-bound vimentin one can suggest that the antibody bound to vimentin in complex with fibrinogen perturbs the first step in fibrin fiber formation. The addition of the antibody in plasma apparently affected both lag time (protofibrils formation) and initial rate (lateral aggregation) of fibrin polymerization (Fig. 6). Future studies will be needed to address the underpinning mechanisms by which anti-vimentin antibody changes fibrin formation kinetics and structural features of the resultant fibrin network structure. A number of studies have reported that the functions described for vimentin in health and diseases are linked to posttranslational modifications (PTMs) in the protein [35], [36], [37]. One can argue that PTMs in vimentin cause structural modifications that change its binding properties for fibrinogen, consequently altering the kinetics of fibrin polymerization. In fact, the difference on the binding constant for fibrinogen between the commercially available insect cell-derived recombinant vimentin and our bacterial-derived rROD protein could be explained by the fact that the former is glycosylated. The mechanism for generating a vimentin-fibrinogen complex remains unknown.

This study reports elevated levels of extracellular vimentin in plasma from patients with systemic inflammation, similar to a previous study that reported elevated vimentin levels in sepsis [27]. Although others and we have detected vimentin in plasma from patients with other clinical conditions or healthy donors [16], [38], [39], normal ranges for plasma vimentin is undefined. Notwithstanding, the comparative analysis between patients and healthy donors in this report demonstrate that plasma levels of vimentin may vary in health and become elevated with systemic inflammation. Additionally, the source for the increased extracellular vimentin in the plasma remains elusive. The increment of plasma vimentin in sepsis may represent its involvement in the immune response to the infection or injury, possibly, being secreted by activated macrophages and/or endothelial cells [22], [40]. Further investigations are necessary to identify the cell types and the mechanisms by which pathogens or conditions associated with systemic inflammation induce the cells to secrete vimentin or a different form of vimentin.

Another fascinating and clinically relevant result obtained from this study is the role of extracellular vimentin in mediating both fibrin formation and fibrin clot structure, especially in plasma from critically ill patients. This novel function of vimentin in plasma was demonstrated by the inhibitory effect conferred by the anti-vimentin antibody in attenuating fibrin formation and reducing the fibrinolytic burden in plasma. Furthermore, the effect of the antibody in the architecture of the resultant fibrin clot structure (i.e. increase porosity area) seems to be similar in plasma from both healthy volunteers and critically ill patients. Consequently, the fibrinogen-bound vimentin alters the conversion of fibrinogen to fibrin after thrombin cleavage, causing the formation of a denser or aberrant clot structure as we and other recently reported for patients with severe COVID-19 [28], [41]. Thus, we reason that extracellular vimentin may be a potential procoagulant agent in critical illness with systemic inflammation, including severe cases of COVID-19. Further work is necessary to investigate the mechanism(s) by which vimentin and fibrinogen interact.

The anti-vimentin antibody was also effective at reducing (to a lower degree) fibrin formation in healthy plasma, suggesting a normal role for vimentin in coagulation. This outcome appears to be in line with our previous study in which we reported a modest prolonged tail bleeding time in the knock out mice as compared to wild type mice [16]. Despite the different phenotypes (>25) reported in vimentin null mice (review [17]), bleeding is not one of them. Vimentin is expressed in several cell types of mesenchymal origin, explaining its participation in many cell functions (review [20]), and therefore, the vimentin null mice may not be suitable to determine the source of the plasma vimentin that interacts with fibrinogen. Nonetheless, all the outcomes from our studies strongly provide evidence that the binding of extracellular vimentin to fibrinogen modulates fibrin formation.

The interaction of vimentin with fibrinogen could have other clinical implications. 1) Elevated levels of an aberrant extracellular vimentin in thromboinflammatory diseases could serve as a marker for an altered fibrin formation and thrombosis. 2) It is known that various types of cancer overexpress vimentin [42], [43], and elevated levels of extracellular vimentin detected in circulation could serve as a marker for cancer-associated thrombosis [44], [45].

In summary, the results of this study identify extracellular vimentin as an active mediator of fibrin formation. Additionally, the data suggest that elevated level of an apparent aberrant vimentin may increase the risk for thrombosis in patients with persistent systemic inflammation. One can propose that strategies to inhibit vimentin may represent a novel therapeutic approach to attenuate the prothrombotic state in critically ill patients with systemic inflammation.

CRediT authorship contribution statement

M. Martinez, L. Brubaker, performed experiments, analyzed data and contributed in writing the manuscript. N. Seshadri and A. Cebula performed experiments. M. Loor, L. Brubaker and T.K. Rosengart provided blood samples from patients as previously described [28]. R. Rumbaut, A. Yee and M. A. Cruz designed experiments, analyzed data, wrote, and edited the manuscript.

Declaration of competing interest

M.A. Cruz is the founder and CSO of A2 Therapeutics, Inc.

Appendix A. Supplementary data

Supplementary material