The Role of Platelets and von Willebrand Factor in the Procoagulant Phenotype of Inflammatory Bowel Disease

Abstract

Aims

Although the risk of thrombosis is well documented for inflammatory bowel disease [IBD] patients, the underlying pathological mechanism seems to be different from other thrombotic conditions. Determining the factors responsible for the increased risk of thrombosis in IBD would help to improve the management of this frequent complication.

Methods

We studied the interplay between platelets, coagulation, and von Willebrand factor [VWF] in 193 IBD patients and in experimental models [acute and chronic] of colitis in wild-type and VWF-deficient mice.

Results

We found a platelet-dependent increase in thrombin generation in IBD patients and in our mouse model of colitis. Agglutinated platelets were present in the blood of patients and mice. Interestingly, we observed not only a significant increase in total VWF antigen, but we were also able to detect the presence of active VWF [VWF in its platelet-binding conformation; 3.2 ± 2.7 μg/mL] in the plasma of 30% of all IBD patients. In healthy controls, active VWF levels were <0.3 μg/mL. This led us to further explore experimental colitis in VWF-deficient mice and we observed that these mice were protected against the procoagulant state triggered by the colitis. Unexpectedly, these mice also showed a significant worsening of colitis severity in both acute and chronic models.

Conclusion

Platelets and VWF [including its active form] appear to be central players in the procoagulant phenotype in IBD. We observed that the role of VWF in haemostasis differs from its role in colonic tissue healing, potentially opening new therapeutic avenues for a life-threatening complication in IBD patients.

1. Introduction

Inflammatory bowel disease [IBD], which includes Crohn’s disease [CD] and ulcerative colitis [UC], is an independent risk factor for first and recurrent thrombotic events.¹ Meta-analyses in large populations have revealed that the risk of venous thromboembolism [VTE] in IBD patients is increased more than 2-fold as compared to the general population.^2,3 Major arterial thrombotic events such as myocardial infarction, splanchnic ischaemia and stroke are also increased in IBD patients.⁴

The increased risk of thrombosis in IBD patients is poorly documented. However, several changes in haemostasis processes have been identified: [i] increased platelet count, [ii] increased markers of in vivo coagulation activation and increased levels of haemostatic proteins such as von Willebrand factor [VWF], and [iii] increased plasma fibrinogen and a denser fibrin network.⁵

So, despite the presence of intestinal bleedings, the clinical picture in IBD patients remains in favour of a clear imbalance towards a procoagulant phenotype. However, the pathogenesis of thrombosis in IBD patients seems to be different from other conditions where thrombosis may occur. Furthermore, the traditional clotting times, such as prothrombin time or partial thromboplastin time, can be unchanged or even prolonged in IBD and thus are not suitable to evaluate hypercoagulability in such a context.⁵

Until now, anticoagulant treatment in IBD patients is given during hospitalization periods and consists of low-molecular-weight heparin or fondaparinux.⁶ However, the risks of thrombosis and/or of bleeding remain increased during a flare for non-hospitalized patients.⁷ Thus, a better understanding of the relationship between colic and systemic inflammatory and haemostatic pathways in IBD would help to limit the risk and lead to better management of patients.

Here, we studied the interplay between platelets and coagulation in IBD, by analysing thrombin generation in plasma of IBD patients and by using a dextran sulphate sodium [DSS] mouse model of colitis. By identifying both platelets and VWF as potential contributors to the procoagulant profile in IBD, we then explored their role in VWF-deficient mice. VWF-deficient mice were protected against the procoagulant state, while in parallel a worsening of the colitis was observed. These results identify VWF as a key player that is involved in both the procoagulant phenotype and tissue healing in IBD.

2. Methods

2.1. Human samples

This study used 55 healthy volunteers and 193 IBD patients included in the I-BANK cohort [which aims to identify prognostic and predictive circulating biomarkers in IBD; NCT03809728]. Blood was obtained by venipuncture via a 21-gauge needle and collected in 0.129 mol/L sodium citrate [9:1, v/v]. Patients included in the I-BANK study were adult patients with a diagnosis of CD or UC. Disease activity was assessed with the Harvey–Bradshaw Index [HBI] for patients with CD, and the partial Mayo score [PMS] for patients with UC. HBI < 5 and PMS ≤ 2 were defined as clinical remission for CD and UC patients respectively. The distribution of patients among the different experiments is summarized in Supplementary Figure 1. Complete blood counts were determined with an automatic cell counter [Micros 60 ABX model]. Platelet-rich plasma [PRP] was prepared within 2 h following blood sampling with a first centrifugation at 190 g for 10 min at 20°C. The supernatant was removed, and the platelet count was adjusted to 200 × 10⁹/L with platelet-poor plasma [PPP] prepared with 10 min of centrifugation at 1750 g at 20°C. Platelet-free plasma [PFP] was prepared by centrifuging PPP at 13 000 g for 30 min at 4°C. The remaining PFP was stored at −80°C until use.

2.2. Wild-type and VWF-deficient mice

Eight- to 10-week-old C57BL/6N male [n = 36] mice were purchased from Charles River. Male VWF-deficient [VWF KO] mice [n = 26] or their control littermates [n = 28], also on a C57BL/6 background, were used for this study.⁸ Housing and experiments were conducted in accordance with French regulations and the experimental guidelines of the European Community [Directive 2010/63/EU]. Animals were housed under standard conditions and given free access to standard rodent chow and water. The protocols were approved by the Lorraine Animal Welfare and Ethical Review Body [CELMEA], France [#9411-2017032718404787 v4], and by ethical committee CEEA26 [#29959-2021021816061812 v1 and #42514-2023040517378767 v3].

2.3. Acute and chronic models of colitis

Mice were treated with DSS to induce acute colitis by administration of 3% DSS, [molecular weight 36 000–50 000 Da; MP Biomedicals] dissolved in drinking water for 5 days. DSS solution was replaced thereafter by normal drinking water for another 3 days as previously described.⁹ Mice were killed on the 8th day via isoflurane anaesthesia followed by cervical dislocation.

Chronic DSS colitis was induced by three cycles of administration of 2% DSS for 5 days, followed by 5 days of normal drinking water.

2.4. Assessment of Disease Activity Index and cumulative score in mice

Mice were weighed and evaluated for clinical symptoms. The Disease Activity Index [DAI] was determined on a scale from 0 to 4 and calculated as the mean of three individual subscores [body weight loss, stool consistency, and blood in the stool] as previously described.¹⁰ For the chronic DSS model, a cumulative score was performed adding colon length and histological score to DAI. Each parameter was scaled from 1 to 10 before addition.

2.5. Blood sampling, PRP, and PFP preparation in mice

Mice were anaesthetized with isoflurane inhalation [1.5% in 1 L/min oxygen]. Whole blood was collected via heart right ventricle puncture following opening of the thoracic cage diaphragm and mixed with 10% citrate [109 mM] for calibrated automated thrombography [CAT] experiments. PRP was prepared by two successive centrifugations of 190 g for 4 min and 1750 g for 1 min at 20°C. The platelet count was adjusted to 200 × 10⁹/L with PPP prepared with 10 min of centrifugation at 1750 g at 20°C and then supernatants were centrifuged at 13 000 g for 30 min to remove all platelets and cell fragments. PFP was frozen and stored at −80°C until use.

2.6. Thrombin generation assay in human and mice

CAT was performed at 37°C in a microtitre plate fluorometer [Fluoroskan Ascent, ThermoLabsystems] using a dedicated software program [Thrombinoscope BV] as reported previously for human¹¹ and mice.¹² Thrombin generation was triggered by recalcification in the presence of recombinant human tissue factor [0.5 pM; Dade Behring] and, when specified, in the presence of 13.9 nM of activated protein C [APC]. Round-bottom 96-well Greiner blue plates were used for PRP and PFP.

2.7. Measurement of total VWF antigen in human and mice and active VWF in human

Total VWF antigen [VWF:Ag] was measured with the Asserachrom VWF kit [Diagnostica Stago]. Active VWF was detected by an enzyme-linked immunosorbent assay [ELISA], using the llama nanobody AU/VWFa-11 [which detects a gain-of-function conformation of the A1 domain] as the capture antibody and polyclonal VWF antibody to detect bound VWF. The slope of VWF captured by AU/VWFa-11 for test plasma was divided by the slope obtained for pooled normal plasma, and this ratio was designated the ‘VWF activation factor’. Total active VWF was determined by multiplying VWF antigen by the VWF activation factor.¹³

2.8. Measurement of D-dimer in human

D-dimer were measured in PFP by the ELISA [Asserachrom, Stago].

2.9. Platelet analysis in whole blood via flow cytometry in mice

Blood was collected as previously described but with addition of PPACK [80 μM; Sigma Aldrich] and diluted in Tyrode’s buffer [dilution 1:20; 137 mM NaCl, 2 mM KCl, 0.3 mM NaH₂PO₄, 5.5 mM glucose, 5 mM HEPES, 12 mM NaHCO₃, 1 mM MgCl₂, pH 7.4]. To 20 µL of diluted blood, 5 µL of GPIbβ-FITC antibody [Emfret, #M040] was added to gate only platelets, followed by either 5 µL of CD62P-B421V antibody [BD Horizon 564289], 5 µL of Annexin V-PE CF594 antibody [BD Horizon 563544], 5 µL of JON/A-PE [Emfret, #M023-2], or 5 µL CD41/61-PE [Emfret, Leo. F2, #M025-2]. Tyrode buffer was added to a final volume of 100 µL, and the mixture was incubated for 30 min in the dark at 37°C. The reaction was stopped with 300 µL of Tyrode buffer and the samples were analysed with the Celesta cytometer [Becton Dickinson]. Flow cytometry data were analysed with the Kaluza software.

2.10. Platelet agglutination analysis in human and mice

Blood smears were made using 5 µL of fresh blood on polysin adhesion microscope slides. May Grunwald-Giemsa [MGG] staining was performed by the haematology department at Nancy University hospital. For each smear, platelet agglutinates were counted in ten independent fields under a Nikon eclipse Ti2 microscope using the 60× objective. Agglutinates were divided into two groups: two to five platelets, and more than five.

2.11. Histological assessment and scoring in mice

Colon samples were thoroughly washed, prepared using the Swiss-rolling technique,¹⁴ fixed in 10% neutral phosphate-buffered formalin solution for 24 h, and embedded in paraffin. Colitis was histologically assessed on 5-μm-thick sections stained with haematoxylin–eosin–saffron. The histological colitis score was calculated by expert pathologists unaware of the animal experimental groups, as previously described.^15,16 Briefly, disease scoring was based on four histological parameters: acute inflammatory cell infiltrate [polymorphonuclear cells in the lamina propria, cryptitis, and crypt abscesses], chronic inflammatory cell infiltrate [mononuclear cells in the lamina propria], crypt architectural irregularities, and ulceration. Thus, the histological inflammatory activity was graded on a five-point scale corresponding to grade 0 [absence of inflammatory infiltrate], grade 1 [chronic inflammatory infiltrate alone], grade 2 [mild or moderate acute inflammatory infiltrate], grade 3 [moderate or severe acute inflammatory infiltrate] and grade 4 [severe, ulceration]. Immune infiltrate, in particular neutrophils, was counted at ×400 magnification on 5 consecutive fields with a Zeiss Axioskop microscope using a Celltrac cell counter by an expert pathologist. Additionally, a cumulative score was established via the addition of the DAI, weight variations, colon length, and histological score.

2.12. Lipocalin assay in mice

The colon was lysed using a cell lysis kit [Bio-Rad]. Lipocalin-2 was determined in lysed tissue [dilution 1:100] using an ELISA kit [Duoset Lcn-2 ELISA Kit, R&D system].

2.13. FITC dextran

Mice were fasted for 4 h, and FITC-Dextran 4 kDa [SigmaAldrich] dissolved in PBS [80 mg/mL] was administered orally [60 mg/100 g body weight]. Blood was collected 4 h later on heparinized tubes [Microtainer BD Medical].¹⁷ Plasma was prepared as described above and immediately placed in opaque tubes. Fluorescence was measured using a plate reader [FLUOstar OPTIMA, BMG, excitation wavelength 485 nm, emission wavelength 510 nm] and compared to a standard curve for FITC-dextran prepared in plasma.

2.14. Statistical analysis

Two-group comparisons were performed using the Mann–Whitney U-test for non-parametric data or t-test for parametric data. ANOVA with Tukey’s post-hoc test for parametric data or Kruskal–Wallis with Dunn’s correction for non-parametric data were used to compare more than two groups. Statistical significance was considered at p < 0.05. Statistical analyses were performed with GraphPad Prism 6 software [GraphPad Software].

3. Results

3.1. Increased platelet-dependent thrombin generation and platelet agglutinates in IBD patients

In this study, we included 55 healthy volunteers and 193 patients with IBD. Characteristics of patients, who are included in the I-BANK cohort, are shown in Table 1. In this cohort 131 CD and 62 UC patients were recruited; 58% were males and the mean age was 43 years. Concerning in vivo coagulation and fibrinolysis activation, healthy controls and IBD patients from the I-BANK cohort display similar D-dimer plasma concentrations [Figure 1A] and none of the patients displayed an elevated D-dimer concentration. We found an increase in thrombin generation in IBD patients compared to healthy controls [Figure 1B]. Thrombin generation with PRP was higher and faster in patients with IBD, as visualized by an increased thrombin peak and decreased time to peak [Figure 1B]. This increase was observed regardless of the type of IBD [CD or UC] or disease activity. Increased thrombin generation was present in PRP but not in PFP, suggesting a platelet-dependent mechanism. In addition, analysis of blood smears revealed the presence of platelet agglutinates in IBD patients compared to healthy controls [Figure 1C and D]. It appears therefore that IBD patients present a procoagulant phenotype that relies on a platelet-dependent mechanism. To our knowledge, this represents the first report of an increased thrombin generation in PRP in IBD patients.

3.2. Elevated VWF antigen and presence of active VWF in patients with IBD

We first studied APC resistance as this phenomenon can be triggered by activated platelets, but IBD patients did not have higher APC resistance compared to healthy controls [Supplementary Figure 2]. Since platelet agglutination may depend on haemostatic proteins such as VWF,^18,19 we first measured plasma concentration of VWF:Ag. We found increased VWF:Ag in IBD patients compared to healthy controls [Figure 2A]. Interestingly, a positive correlation between VWF:Ag concentration and disease activity score of CD patients was found [Figure 2B].

Beyond the mere increase in VWF:Ag, we detected increased levels of the active, platelet-binding form of VWF [up to 15 μg/mL] in 30% of patients [Figure 2C]. This active form is present to a limited extent in healthy controls [<0.3 μg/mL], and more often below the limit of detection [90 ng/mL]. What was particularly interesting is that patients displaying active VWF have an increased peak thrombin generation and faster time to peak compared to patients without active VWF [Figure 2D]. Together, by showing that thrombin generation in the presence of platelets correlates with the presence of active VWF, our data identify VWF as a new player potentially contributing to the procoagulant phenotype in IBD.

3.3. Mice treated acutely with DSS exhibit a procoagulant phenotype similar to IBD patients

To better understand the mechanisms at play in this increased thrombin generation in PRP, we next used an experimental mouse model of colitis induced by DSS. As in IBD patients, thrombin generation in PRP in DSS-treated mice was increased, indicating the presence of a procoagulant phenotype [Figure 3A]. Similar to IBD patients, DSS-treated mice also displayed increased total VWF in plasma [Figure 3B]. Since platelets appear to play a central role in the procoagulant phenotype present in IBD, we performed flow cytometry analysis to measure their basal activation in whole blood from DSS-treated mice [Figure 3C]. Platelet population was gated using glycoprotein [GP] Ibβ and activation status was determined with P-Selectin, Annexin V, CD41 or JON/A. No variations were visible between the groups [Figure 3C]. However, the ability of platelets to agglutinate was increased in DSS-treated mice, with large agglutinates of more than five platelets being increased almost 4-fold in DSS-treated mice compared to control mice [Figure 3D]. These results suggest that despite being in a non-activated state, platelets can nevertheless contribute to an increased procoagulant potential in IBD.

3.4. Decreased thrombin generation and platelet agglutination in VWF KO mice treated acutely with DSS

To determine the potential role of VWF in these processes, we used VWF KO mice treated with DSS. Vehicle-treated and DSS-treated VWF KO mice showed lower platelet agglutinates compared to WT mice treated with DSS: small agglutinates of two to five platelets were similar to WT DSS-treated mice, whereas agglutinates of more than five platelets were significantly decreased in VWF KO DSS-treated mice [Figure 4A and B]. To investigate the co-influence of VWF and platelets on the procoagulant phenotype triggered by colitis, we assessed thrombin generation in PRP. At baseline, VWF KO mice showed significantly less thrombin generated (decreased endogenous thrombin potential [ETP] with a strong tendency towards a peak decrease) [Figure 4C]. Following DSS treatment, thrombin generation remained lower in the absence of VWF compared to DSS-treated WT mice. However, the most striking observation was that, in contrast to WT mice, DSS treatment did not induce any significant increase in thrombin generation in VWF KO mice compared to the basal situation [p > 0.99 for the ETP and p > 0.74 for the peak]. These results indicated that VWF is crucial in the platelet-dependent increased thrombin generation due to colitis.

3.5. Lack of VWF aggravates colitis in the acute DSS model

Since IBD is characterized not only by an increased risk of thrombosis but also by intestinal bleedings and because VWF is one of the central molecules in haemostasis regulation and is also important for tissue healing, we explored the colitis phenotype of VWF KO mice treated with DSS. Weight loss and DAI were not different between WT and VWF KO mice following DSS treatment [Figure 5A]. We measured lipocalin-2 in mice colonic tissue as a marker for intestinal inflammation and found a strong tendency of increased inflammation in VWF KO mice [Figure 5B]. In addition, histological scores revealed increased colitis severity in VWF KO mice compared to WT mice [Figure 5C and D; Supplementary Figure 4]. Histological observation confirmed that DSS-treated VWF KO mice display a worsened colitis, highlighting the importance of VWF in limiting tissue damage. The typical lesions found showed a moderate to severe acute inflammatory infiltrate of the lamina propria associated with variable images of cryptitis and crypt abscesses. Finally, neutrophil infiltration in colonic tissue was increased with DSS in both WT and VWF KO mice, and despite being not significantly different, more DSS-treated VWF KO mice had an elevated number of cells.

3.6. VWF deficiency is associated with worsened colitis in a mouse model of chronic colitis

Since IBD is a chronic inflammatory disease with active and remission phases, VWF KO mice were submitted to chronic DSS treatment. Following three cycles of 2% DSS, VWF KO DAI was worsened compared to WT mice [Figure 6A]. This was confirmed by histology given that most VWF KO mice were severely affected [Figure 6C; Supplementary Figure 4]. In addition, using a cumulative activity score that included mouse weight variations and colon length, VWF KO mice displayed a worsening of their phenotype compared to WT mice [Figure 6D]. Histological lesions were similar to those in the acute model, but were more in the severe rather than the moderate grade, and with a larger extent. The most severe cases showed ulcerations with complete loss of glands and formation of granulated tissue. Lipocalin-2 was increased similarly in the colon tissue of both WT and VWF KO mice following DSS treatment [Figure 6E]. A strong trend towards increased intestinal epithelium permeability was measured in VWF KO mice treated with DSS and neutrophilic infiltration was more important in VWF KO compared to WT mice after DSS treatment [Figure 6F and G].

4. Discussion

Compared to the general population, the risk of thrombosis in IBD patients is more than 2-fold increased.^2,3 However, the mechanisms underlying such an increased risk are not well identified, especially since the classical molecular players of haemostasis are differently regulated due to the chronic inflammation present in the context of IBD. What is well documented is an increased platelet count in IBD patients, corresponding to reactive thrombocytosis.²⁰ Using a DSS mouse model, we have been able to show that this feature is also present in mice [Supplementary Tables 1 and 2].

Since platelets, through the expression of procoagulant phospholipids, support thrombin generation and subsequent formation of a fibrin clot, the interplay between platelet activation and fibrin clot formation is of central importance in gaining a better understanding of the risk of thrombosis. It is for this reason that we studied thrombin generation in PRP from IBD patients or colitis mouse models. This allowed us to highlight the central importance of platelets in the increased procoagulant phenotype occurring in IBD. Of note, in contrast to others, we did not find basal activation of platelets in the DSS-treated mouse model [Supplementary Figure 5].²¹ This difference can be related to the DSS model itself that is very sensitive to mouse housekeeping. Indeed, despite identical DSS protocols, we noted that our mice where less prone to very severe colitis.

When performing thrombin generation assays in PRP, thrombin is the main activator of platelets. Interestingly in CD, the platelet thrombin receptors, protease activated receptors 1 and 4 [PAR-1 and -4], were found to increase platelet responsiveness in whole blood platelet aggregometry, which is consistent with our results.²²

One important observation concerned the presence of platelet agglutination in both patients and the mouse model of colitis, indicating that other plasmatic factors were playing an important role in the interplay between platelet action and coagulation. Previous studies identified VWF as one of these potential factors.^18,19 VWF is a multimeric plasma glycoprotein that plays a crucial role in bridging platelets and anchoring them to the site of vascular injury to limit blood loss.²³ A VWF increase has been clearly identified to be associated with disease activity in IBD patients.²⁴ VWF also has direct cellular functions related to inflammation, via the recruitment of leukocytes, or by participating in vascular permeability.²⁵ In addition, VWF may induce a pro-inflammatory M1 phenotype to macrophages.²⁶ However, how these functions may affect the course of IBD has not been explored. We confirmed this plasmatic increase of VWF in mice treated with DSS and in IBD patients. One important novelty of our study is that we identified for the first time the presence of active VWF in 30% of all patients, whereas it remained essentially undetectable in healthy volunteers. Active VWF represents circulating VWF that is in its platelet binding conformation, a form that is normally adopted solely when VWF is bound to the subendothelium following vascular injury.¹³ The presence of active VWF has been identified in a number of pathological situations associated with increased thrombotic risk,²⁷ and its identification in a subset of IBD patients is a clear indication that VWF plays a central role in hypercoagulability associated with IBD. In addition, we observed that IBD patients with active VWF displayed a higher peak in thrombin generation and shorter time to peak in PRP, indicating that the presence of active VWF enhances thrombin formation in both a quantitative and a temporal manner. All these results may be useful toward identification of patients at risk of thrombosis who might benefit from preventive anticoagulant treatment. Larger cohort studies are needed to validate a correlation between active VWF and the occurrence of thrombosis in IBD.

To determine the involvement of VWF in colitis procoagulant phenotype, we submitted VWF KO mice to an acute DSS treatment. DSS platelet agglutination was lowered in VWF KO mice, confirming the central implication of VWF in the formation of large platelet agglutinates. In thrombin generation assays in PRP, the thrombin peak was lower in VWF KO mice, consistent with the decrease in coagulation factor VIII secondary to VWF deficiency.⁸ However, it was very striking that DSS treatment did not lead to increased thrombin generation [p > 0.99], in contrast to what is observed in WT mice, indicating that VWF contributes to the platelet-dependent pro-coagulant phenotype in IBD.

Interestingly, despite this protection against an increased pro-coagulant phenotype, VWF KO mice developed a more severe colitis than WT mice and intestinal lesions were worsened. We observed a similar phenotype when using chronic DSS treatment with worsened DAI, and increased histological severity in VWF KO mice compared to WT mice. Additionally, we noted an increased tendency of altered intestinal epithelium permeability in VWF KO mice associated with increased neutrophil infiltration. These results may seem contradictory, but they emphasize the role that VWF plays in haemostasis and its role in inflammation. Additional cellular roles of VWF on vascular smooth muscle cell migration or wound healing have been highlighted in different studies, emphasizing the complexity of this multifaceted protein.^28,29 Indeed, VWF is able to bind several growth factors such as vascular endothelial growth factor-A and platelet-derived growth factor-BB through its heparin-binding domain, leading to increased angiogenesis and accelerating wound healing.²⁸ In addition, we have previously shown that VWF is able to promote vascular smooth muscle cell proliferation though LRP4 and integrin α_vβ₃, providing new insight into mechanisms important for wound healing.²⁹ Of note, we did not observe any difference in tissular immune cell composition in VWF KO mice compared to WT mice, but basal permeability of VWF KO may be impaired [Supplementary Figure 7].

Thus, inhibition of VWF, despite its beneficial effect on protecting against thrombosis, may lead to aggravated colic tissue damage via limitation of inflammation inhibition and wound healing. Such an observation is unfortunately very reminiscent of the pathophysiology of IBD patients who may develop intestinal bleedings, despite being at increased risk of thrombosis. Therein lies the difficulty in developing new treatments for IBD, which need to accommodate this co-occurrence of opposite phenotypes.

In addition, on a more general note, we did not observe an increase in D-dimer in our IBD population indicating that in vivo activation of coagulation [and fibrinolysis] is limited. This result suggests that thrombus formation may be the result of a second hit activation in the presence of increased coagulation reactivity due to platelets linked to active VWF as active VWF is able to bind GPIbα and participate in their activation.

To summarize, in the present study we have identified platelets and VWF and its active form as central players in the pro-coagulant phenotype linked to increased thrombosis in IBD. Using a translational approach and prospectively enrolling 193 well-characterized IBD patients, we observed for the first time an increased thrombin generation in PRP compared to healthy subjects as well as the presence of platelet agglutinates and active VWF in nearly 30% of all patients. To go further in the characterization of the role of platelets and VWF we first used a mouse model of colitis in which we found an increased thrombin generation in PRP and increased platelet agglutinates as in patients without basal activation of platelets. Finally, in VWF KO mice we found protection against the pro-coagulant phenotype. However, VWF KO mice showed impaired colitis.

To conclude, we have shown a possible beneficial effect of VWF inhibition in IBD patients at high risk of thrombosis. The controversial effects of VWF on colonic tissue inflammation need to be further investigated in a study of pathology coupled with cellular and molecular inflammatory markers. Our findings may help in designing new strategies aimed at preventing this life-threatening condition in IBD patients.

Funding

This work was supported by the Fédération Française de Cardiologie [FFC] and Pfizer grant programme. JL was funded thanks to a grant from the Fondation pour la Recherche Médicale [FRM], CS was funded with a grant from Région Grand Est and the French Minister of Higher Education, Research and Innovation, and MUA was funded with a grant from Région Grant Est, CHRU Nancy.

Conflict of Interest

BC has received lecture and/or consulting fees from Abbvie, Amgen, Celltrion, Ferring, Galapagos, Janssen, and Takeda. LPB has received lecture fees from Galapagos, AbbVie, Janssen, Genentech, Ferring, Tillots, Celltrion, Takeda, Pfizer, Sandoz, Biogen, MSD, Amgen, Vifor, Arena, Lilly, Gilead, and Viatris, Medac consulting fees from AbbVie, Alimentiv, Alma Bio Therapeutics, Amgen, Applied Molecular Transport, Arena, Biogen, BMS, Celltrion, CONNECT Biopharm, Cytoki Pharma, Enthera, Ferring, Fresenius Kabi, Galapagos, Genentech, Gilead, Gossamer Bio, GSK, HAC-Pharma, IAG Image Analysis, Index Pharmaceuticals, Inotrem, Janssen, Lilly, Medac, Mopac, Morphic, MSD, Norgine, Novartis, OM Pharma, ONO Pharma, OSE Immunotherapeutics, Pandion Therapeutics, Pfizer, Prometheus, Protagonist, Roche, Sandoz, Takeda, Theravance, Thermo Fisher, Tigenix, Tillots, Viatris, Vifor, Ysopia, and Abivax, and grants from Takeda, Fresenius Kabi, and Celltrion.

Acknowledgements

We thank the cytometry core facility, UMS2008 IBSLor. We thank Dr Alexandre Kauskot for technical support with platelet flow cytometry experiments and Dr Anne-Charlotte Heba for technical support with lipocalin-2 ELISA. We thank the LIMM [Laboratoire d’Immunologie, Microbiologie et Métabolisme] and platform of the UMR1184 [IMVA-HB] for technical assistance with colon immune cell flow cytometry. We thank Dr Stéphanie Roullet for her help with statistical analyses. We thank Dr Frances Yen Potin for English revision of the manuscript.

Author Contributions

CS, JL, MU, PC, HL, DA, NT, BC, FP, JP, PJL, VR, and CVD performed experiments and analysed the data. PJL, VR, PL, CVD, and LPB designed and supervised the study. CS and JL wrote the first version of the manuscript. All authors revised and approved the final version of the manuscript.

Data Availability

The data that support the findings of this study are openly available upon request..

References

Author notes