A novel method to quantify fibrin-fibrin and fibrin-α2AP cross - links in thrombi formed from human trauma patient plasma .

Background The widespread use of the anti-fibrinolytic agent, tranexamic acid (TXA), interferes with the quantification of fibrinolysis by dynamic laboratory assays such as clot lysis, making it difficult to measure fibrinolysis in many trauma patients. At the final stage of coagulation, Factor XIIIa (FXIIIa) catalyses the formation of fibrin-fibrin and fibrin-α 2 - antiplasmin (α 2 AP) cross-links which increases clot mechanical strength and resistance to fibrinolysis. Objectives Here, we develop a method to quantify fibrin-fibrin and fibrin-α 2 AP cross-links that avoids the challenges posed by TXA in determining fibrinolytic resistance in conventional assays

A novel method to quantify fibrin-fibrin and fibrin-α2AP cross-links in thrombi formed from human trauma patient plasma.

Introduction
Traumatic injury accounts for 4.9 million deaths globally every year (1), and is the leading cause of death in persons under the age of 44 (2).Uncontrolled bleeding accounts for 1.2 million deaths each year, 25 % of all injury-related deaths (3), and is exacerbated by the development of trauma induced coagulopathy (TIC).TIC is associated with an increased need for massive transfusion and 3-4 fold increased risk of death(4).
Major studies have shown that there are two fundamental changes that underpin major haemorrhage during TIC; hypofibrinogenaemia (fibrinogen depletion) and hyperfibrinolysis (excessive clot degradation) (5).Fibrinogen is the key pro-coagulant factor for stable clot formation and is the first coagulation protein to reach critically low levels during traumatic haemorrhage (6)(7)(8).
Fibrinogen is cleaved by thrombin to insoluble fibrin, a viscoelastic polymer that is crucial in determining the physical and mechanical characteristics of the clot (9).At the final stage of coagulation the transglutaminase, activated Factor FXIII (FXIIIa), catalyses the formation of crosslinks between neighbouring fibrin molecules to enhance clot stability against mechanical stress (10,11).FXIIIa also exerts an anti-fibrinolytic effect by cross-linking the principal fibrinolysis inhibitor, alpha 2-antiplasmin ( 2 AP), to fibrin to stabilise the clot against premature degradation by plasmin (12,13) and thus plays a key role in regulating fibrinolysis.Early during TIC, fibrinolysis is amplified, demonstrated by the increased levels of plasmin- 2 AP (PAP) complexes, D-dimer (degradation product of cross-linked fibrin) and tissue plasminogen activator (tPA) (14).Depletion of fibrinolytic inhibitors, including  2 AP and plasminogen activator inhibitor 1 (PAI-1), are likely contributors to hyperfibrinolysis in TIC.
The anti-fibrinolytic agent, tranexamic acid (TXA), is administered to trauma patients to stop bleeding.
The CRASH-2, CRASH-3 and WOMAN trials found that early anti-fibrinolytic treatment with TXA effectively reduced mortality in trauma and obstetric haemorrhage (15)(16)(17).TXA is widely used to treat patients with major haemorrhage due to its effectiveness and low cost.TXA inhibits fibrinolysis by binding to the lysine binding sites of plasminogen.This prevents plasminogen binding to fibrin and its activation to plasmin (18).Consequently, TXA interferes with commonly used dynamic fibrinolysis tests such as the clot lysis assay, and alternative approaches are required to clearly understand the effects of factor replacement therapy on clot strength and susceptibility to fibrinolysis in trauma patients.Here we describe a novel method to assess clot susceptibility to fibrinolysis indirectly that J o u r n a l P r e -p r o o f avoids the effect of TXA; by quantification of fibrin-fibrin and fibrin-α 2 AP cross-links in trauma patients recruited to the Fibrinogen Early In Severe Trauma studY (FEISTY; NCT02745041) (19,20).FEISTY trial participants were randomised to receive one of two fibrinogen replacement therapies; either cryoprecipitate (cryo) or fibrinogen concentrate (Fg-C).
Our previous data have shown that FXIII levels were significantly increased post-cryo transfusion and significantly decreased post Fg-C transfusion.Analysis of individual fibrin fibres revealed that clots formed from patients receiving cryo were composed of fibres that were more resistant to mechanical disruption (21).Coagulation FXIIIa plays a critical role in the formation of fibrin-fibrin and fibrin-α 2 AP cross-links (10)(11)(12), therefore, we hypothesised that the abundance of cross-links may differ between patients who received cryo and Fg-C.Here, we utilise liquid chromatography-mass spectrometry (LC-MS) to quantify cross-linked peptides in clots formed from FEISTY trauma patient plasma (Figure 1).(22).Parallel reaction monitoring (PRM) (23) was used to target crosslinks between fibrinogen alpha chains (FGA-FGA), fibrinogen gamma chains (FGG-FGG) and fibrinogen alpha and α 2 AP chains (FGA-SERPINF2).We report using LC-MS to quantify the relative abundance of cross-links as a novel method to indirectly quantify fibrinolytic resistance in patients who receive TXA and adjunctive transfusion therapy.
J o u r n a l P r e -p r o o f

Eligibility Criteria & Randomisation
Full details of the FEISTY study are available (20,21).Patients were randomly assigned to receive either Fg-C (Riastap; CSL Behring, Pennsylvania, USA) or cryo (whole blood or apheresis cryo; Australian Red Cross Lifeblood) replacement during active trauma haemorrhage.All blood samples were collected in 0.13 M trisodium citrate vacutainers and platelet poor plasma was obtained by centrifugation of whole blood samples at 2500 g for 30 min at 4°C.Plasma samples were stored at -80°C until analysis.
Baseline characteristics of the patient cohorts were previously published and included as Table S1 (21).

Mass Spectrometry Quantification of Fibrin Cross-links
A workflow is shown in Figure 1.

Clot Formation
Plasma clots were formed at 37°C for 30 min and clotting was initiated by addition of 1 nM thrombin (Sigma Aldrich, Missouri, USA), 2 mM MgCl 2 and 18 mM CaCl 2 .Clots were centrifuged at 17,000 g for 1 min and the supernatant discarded.

Clot Purification
Clots were treated with 6 M guanidine hydrochloride (Gnd-HCl) and 100 mM ammonium bicarbonate (ABC) pH 9 to remove any constituents that are not covalently linked to the fibrin clot.The tubes were placed on a vortex shaker for 5 min at room temperature.Clots were centrifuged at 17,000 g for 1 min to remove supernatant.This purification step was repeated 3 times to ensure the clots were translucent, each time the vortex time was extended to 15 min, 30 min and finally 16 h, which was performed at 4°C.

Clot Digestion
Our protocol was adapted from Schmitt et al (24).Clots were treated with hydroxylamine solution (1M NH 2 OH-HCl, 4.5M Gnd-HCl, 0.2M K 2 CO 3 pH 9) at 45°C on a vortex shaker for 16 h.Following incubation, the samples were centrifuged at 17,000 g for 1 min and the supernatant removed.Clots were stored at -80°C until ready for analysis.Prior to mass spectrometry analysis clots were subjected to proteolytic digestion.Clots were washed with 100 mM triethylammonium bicarbonate buffer pH 8.5 (TEAB) prior to centrifugation at 17,000 g for 2 min at 4°C and removal of the supernatant.Proteolytic J o u r n a l P r e -p r o o f digestion was performed using 25 µg PTMScan Lys-C Protease (Cell Signaling Technology, Massachusetts, USA) and 5 µg TPCK trypsin (Sigma Aldrich, Missouri USA) in 100 mM TEAB at 300 RPM for 16 h at 37°C.A further 5 µg TPCK trypsin was added, and the digestion continued for another 2.5 h.

Peptide Clean Up
Following enzymatic digestion, 1% trifluoracetic acid (TFA) was added to acidify the digests and the samples were centrifuged at 17,000 g for 30 min at 4°C.The supernatant was desalted on C18 cartridges using the AssayMAP Peptide Cleanup protocol on the Agilent Bravo liquid handling platform (All from Agilent, California, USA).Eluates were dried by vacuum centrifugation and stored at -80°C.

Liquid chromatography-mass spectrometry (LC-MS)
Cross-link enriched peptide samples were analysed on a Dionex 3000 Ultimate high-performance liquid chromatography (HPLC) system coupled to an Orbitrap Fusion Lumos mass spectrometer (Thermo Fisher Scientific, Massachusetts, USA).The peptides were loaded onto a trap column (PepMapC18; 300 µm x 5 mm, 5 µm particle size; Thermo Fisher Scientific, Massachusetts, USA) for 1 min at a flowrate of 20 μL/min.The loaded peptides were separated on a 75 µm x 500 mm C18 EASY-Spray chromatographic column (Thermo Fisher Scientific, Massachusetts, USA) with a 2-35 % acetonitrile gradient in 0.1% formic acid and 5% DMSO at a flow rate of 250 nL/min for 60 min.The Orbitrap Fusion Lumos was operated in parallel reaction monitoring (PRM) mode within a 10 min retention window (Table 1).The MS1 spectra were acquired at a scan range of 350-1500 m/z at 120,000 resolution, an automated gain control (AGC) target of 4 x 10 5 and a maximum injection time J o u r n a l P r e -p r o o f of 50 ms.MS2 spectra were acquired over a scan range of 400-1500 m/z at 7500 resolution, an isolation window of 1.2 m/z and 25% higher-energy collision dissociation (HCD).Extracted ion chromatograms (XICs) of cross-linked peptides were generated from the raw data using Skyline v22.2 (MacCoss Lab, University of Washington) (25).XICs were assessed for retention time alignment of transitions, mass error ppm (maximum ppm 10), and the similarity of the measured fragment spectrum to the library spectrum (dotp value in Skyline).The sum of the fragment peak intensities (Max Height) and total fragment area were extracted from Skyline.The total fragment area parameter was used for quality control and the fragment peak intensity was used to quantify the cross-linked peptides (cross-link abundance).

Thrombin Generation
TG was triggered with 1 pM tissue factor (TF) in the presence of 4 µM phospholipids, CaCl 2 and a fluorogenic substrate for thrombin (Diagnostica Stago, Asnieres, France).Thrombin generation was measured using the calibrated automated thrombogram (CAT) (26) and thrombinoscope v5 software.
The peak height and endogenous thrombin potential (ETP) parameters were extracted from the thrombogram and exported for statistical analysis.

Lateral atomic force microscopy
This method is described in detail by Duval et al (27).Briefly, plasma was diluted to obtain a fibrinogen concentration of 0.5 mg/ml and clotting initiated with 0.5 U/ml thrombin and 10.6 mM CaCl 2 .Clots were allowed to form for 1.5 h in a humid chamber prior to a washing step with tris buffered saline.

Data Analysis
Results are expressed as mean and standard deviation (SD).Chord diagrams were prepared in Flourish (London, UK).Statistical analysis was performed using GraphPad Prism Software (v9.J o u r n a l P r e -p r o o f

Results
Initial experiments utilised untargeted LC-MS analysis of fibrin clots formed from FEISTY trauma patient plasma to identify proteins of interest (fibrinogen).Analysis of the untargeted LC-MS spectral count revealed that the FGA, fibrinogen beta (FGB) and FGG chains were the most abundant proteins in the fibrin clot (Figure 2A).FGA, FGB and FGG had a spectral count of 100, 63 and 70, respectively.
We then developed a targeted LC-MS method to quantify the abundance of FGA-FGA, FGG-FGG and FGA-SERPINF2 cross-links in fibrin clots formed from FEISTY trauma patient plasma (Table 1, Figure 1).
Using the spectral count from the initial untargeted experiments, we utilised parallel reaction monitoring (PRM) to identify cross-links in fibrin clots formed from FEISTY trauma patient plasma (Table 1).The MS identified the cross-links based on their known mass and charge within a known time frame (the retention time, Table 1).
Analysis of the PRM LC-MS spectra identified 44 different cross-linked peptides in fibrin clots formed from FEISTY trauma patient plasma; 26 FGA-FGA, 5 FGG-FGG and 13 FGA-SERPINF2 cross-links (Figure 3).Analysis of the cross-linked peptides in FEISTY trauma patient plasma was performed using Skyline(25) by extracting the max height (cross-link abundance) and total area parameters.We used fragment peak intensity to quantify fragment abundance, reasoning that it was less vulnerable to signal-to-noise differences between samples, however, the correlation with fragment total area in any case was extremely high (r 2 = 0.98, p < 0.0001; Figure 4A).
The FEISTY trauma patients were analysed based on the fibrinogen replacement therapy they received (Cryo or Fg-C) and were split into 4 groups; pre-cryo transfusion, post-cryo transfusion, pre-Fg-C transfusion and post-Fg-C transfusion.Patients who received cryo had a 2.2-fold increase in FGA-FGA cross-links, 1.8-fold increase in FGG-FGG cross-links and 1.4-fold increase in FGA-SERPINF2 cross-links (p < 0.0001; Figure 4B-D).In contrast, patients who received Fg-C displayed a small decrease in FGA-FGA, FGG-FGG and FGA-SERPINF2 cross-links (not significant; Figure 4B-D).
Analysis of each of the 44 individual cross-links in the FEISTY trauma patients revealed that patients who received cryo showed a trend towards an increase in cross-link abundance, whereas those who J o u r n a l P r e -p r o o f received Fg-C were likely to trend a decrease (Table 2).Examples of a significant change in cross-link abundance pre-and post-transfusion are the FGA-FGA cross-link (MADEAGSEADHEGTHSTKRGHAK-ALTDMPQMR) and FGA-SERPINF2 (TVTKTVIGPDGHK-NQEQVSPLTLLK) (Figure 4E and 4F).Analysis of all 26 FGA-FGA cross-links revealed that 62% of these cross-links were significantly increased postcryo in FEISTY trauma patients, whereas no FGA-FGA cross-links were significantly increased in those who received Fg-C (Table 2).Out of the 5 FGG-FGG cross-links, 4 cross-links were significantly increased post-cryo (p < 0.05, p < 0.01), with no statistically significant changes observed post Fg-C transfusion (Table 2).Finally, 92 % of the FGA-SERPINF2 cross-links were significantly increased postcryo transfusion (p < 0.05, p < 0.01), whereas only 15 % of these cross-links were significantly increased post-Fg-C (Table 2).
The abundance of cross-links pre-and post-fibrinogen replacement therapy in all FEISTY trauma patients was correlated with the toughness of individual fibrin fibres (Figure 5A-B).The toughness was measured by atomic force microscopy to determine the mechanical strength of individual fibrin fibres upon lateral stretching.There was no correlation between the two parameters pre-transfusion (Figure 5A), but a positive correlation was observed post-fibrinogen transfusion (Figure 5B).This was observed as an increase in individual fibrin fibre toughness that was associated with an increase in the abundance of cross-links (r 2 = 0.5, p < 0.05).The peak amount of thrombin generated in FEISTY patient plasma did not show any association with the abundance of cross-links pre-fibrinogen transfusion (Figure 5C).Post-fibrinogen transfusion, there was a significant correlation between the abundance of cross-links and the peak thrombin generated; an increase in peak thrombin was associated with an increase in the abundance of cross-links (r 2 = 0.5, p < 0.05; Figure 5D).Finally, FXIII antigen levels were correlated with the abundance of cross-links pre-and post-fibrinogen transfusion (Figure 5E-F).There was no association between FXIII antigen and the abundance of cross-links pre-transfusion (Figure 5E), however, a weak but significant correlation was observed between FXIII antigen and the abundance of cross-links post-fibrinogen replacement therapy (r 2 = 0.3, p < 0.05).

Discussion
The value of the PRM LC-MS method developed for this study is that it can be used to quantify the abundance of cross-links in patients with acute bleeding disorders who have received TXA and cannot be analysed using conventional fibrinolysis assays.We anticipate that this method may also have wider applicability, in other haemorrhagic disorders and for analysis of pathological thrombi.Despite vast understanding of the molecular cascades that regulate coagulation, less is known about the end product, the insoluble fibrin matrix.Further development of this method could expand our understanding of venous and arterial thrombus structure at a molecular level and lead to identification of novel diagnostic targets of abnormal coagulation and fibrinolysis.
In the manuscript we use PRM LC-MS to evaluate differences in two fibrinogen replacement therapies; cryo and Fg-C; in FEISTY trauma patients.Cryo is a pooled blood component derived from whole blood donations that has a variable but high fibrinogen concentration (8-16 g/L) (28).Cryo is rich in a number of other coagulation factors that are not present in Fg-C (29).These include anti-fibrinolytic factors such as PAI-1, α 2 AP and FXIII.Fg-C has a standard concentration of 20 g/l and has a favourable safety profile evident from its long term use in inherited dysfibrinogenaemia, afibrinogenaemia and hypofibrinogenaemia (30).Our previous study showed that cryo supplementation reduced fibrinolytic activity, demonstrated by the attenuation of plasmin activation, an increase in PAI-1 and maintenance of thrombin activatable fibrinolysis inhibitor (TAFI) concentration (21).Cryo clots displayed a more homogeneous fibrin network with an increased number of fibres than those formed post Fg-C (21).
Analysis of FXIII and individual fibrin fibres indicated that these clots were composed of fibres that were more resistant to mechanical disruption (21).These PRM LC-MS cross-linking data agree with our previous findings, in that patients who received cryo exhibit a significant increase in abundance of cross-links post-transfusion compared to those who received Fg-C.
Here we focus on cross-links formed between FGA-FGA, FGG-FGG and FGA-SERPINF2.However, it is important to note that other fibrinolytic inhibitors and proteins can also be cross-linked to the fibrin clot by FXIIIa, including, but not limited to, plasminogen activator inhibitor 2 (PAI-2)(31), fibronectin (32), von Willebrand Factor (33), TAFI (34) and complement C3 (35).These cross-links were not searched for in our trauma patient samples but may warrant future study.Our findings highlight the importance of FXIIIa, the transglutaminase responsible for forming cross-links between neighbouring fibrin molecules and the fibrinolytic inhibitor, α 2 AP (12).FXIII is present in cryo at a concentration of 57.7 µg/ml, whereas only trace amounts (0.4 µg/ml) were detected in Fg-C (21).The Fg-C used in the FEISTY study is the CSL Behring product, Riastap.Previous in vitro studies have J o u r n a l P r e -p r o o f 13 evaluated different Fg-C preparations and have shown that Fg-C's that contain FXIII perform similarly to cryo in measures of clot strength and stability (36,37).Our previous data suggest there are differences in fibrin polymerisation and degradation in patients who received cryo or Fg-C, most likely due to the differences in FXIII content (21).Here we show that these differences in FXIII levels correlate with significant differences in the degree of cross-linking between the cryo and Fg-C cohorts.
Combined, these studies suggest that using a Fg-C that contains FXIII may have some benefits over one without to increase clot stability.Interestingly, the extent of fibrin cross-linking did not correlate with the fibrinogen concentration, however this may be due to the small number of patients analysed.
Future studies might investigate the relationship with fibrinogen further and any cross-talk with inflammation.
A limitation of our study is the small number of patient samples; 13 in the Fg-C arm and 9 in the cryo arm.Despite the small number of patient samples, we have demonstrated statistically significant differences in fibrin clot polymerisation between the two patient cohorts.The FEISTY trial was a pilot study to inform viability of a larger RCT and these on-going studies form part of the FEISTY II trial (NCT05449834), that will recruit 850 trauma patients across Australia and New Zealand.There are of course limitations to the PRM LC-MS method, which would impact its use in routine clinical practice, namely the requirement for highly specialised pieces of equipment and expert personnel in addition to a lack in real-time data analysis capacity and platform certification.
The results of CRYOSTAT-2, a RCT evaluating whether early fibrinogen replacement with cryo improves survival in major trauma haemorrhage, were recently published (38).There was no difference in mortality at 28 days if cryo was given early (<60 min), but interestingly there was an improvement in patients who received cryo later (between 60-90 min) (38).The findings highlight the importance of moving away from empiric transfusion and focus on the development of a personalised approach to diagnose coagulopathy and identify the best treatment.To do so, we need to improve our understanding of trauma coagulopathy at a molecular and cellular level which requires additional studies of a similar nature to this one.Access to our targeted PRM LC-MS method and similar technologies will allow scientists and clinicians to improve our understanding of trauma coagulopathy to develop precision transfusion therapies improving patient care and outcomes.
J o u r n a l P r e -p r o o f

Table 1-Parallel reaction monitoring method
Parallel reaction monitoring mass spectrometry (PRM-MS) was utilised for analysis of FGA-FGA, FGG-FGG and FGA-SERPINF2 cross-linked peptides in clots formed from FEISTY trauma patient samples.
The table lists the protein name, peptide cross-link sequence, mass to charge ratio (m/z), charge (z) and the elution window in which the peptides were detected (retention time).This data was used as the method on the MS for detection of the cross-linked peptides.

Table 2-The percentage change in individual cross-links post-fibrinogen replacement therapy with cryo or Fg-C
The percentage increase or decrease in the cross-link abundance for each individual cross-link posttransfusion of fibrinogen replacement with either cryo or Fg-C is shown.* p < 0.05, ** p < 0.01, ns = not significant.

Fibers
were then incubated with 20-nm-yellow-green carboxylate FluoSpheres (Thermo Fisher Scientific, Massachusetts, USA) for 10 min, then washed again.A MFP3D atomic force microscope (Asylum Research, Oxford Instruments, California, USA) combined with an Axiovert 200 optical fluorescence microscope (Zeiss, Jena, Germany) were used to measure the mechanical response of individual fibrin fibres upon lateral stretching.Individual fibres were pulled by CSC38 cantilevers (MikroMasch, Tallinn, Estonia) until rupture, while the deformation was visualised with the fluorescence microscope.For each fibre a stress (calculated from the lateral defection of the cantilever) vs. strain (calculated from the position of the cantilever) curve was plotted and analysed to obtain a range of parameters to quantify fibre mechanical properties.In this work, we report on J o u r n a l P r e -p r o o f 9 toughness.Toughness is calculated as the area under the stress-strain curve and corresponds to the amount of energy required to rupture the fibre.

Figure 2 -Figure 3 -
Figure 2-Protein abundance and identification of cross-links in the fibrin clot.

J o u r
n a l P r e -p r o o f J o u r n a l P r e -p r o o f J o u r n a l P r e -p r o o f J o u r n a l P r e -p r o o f J o u r n a l P r e -p r o o f