Nanostructured fibrin agarose hydrogel as a novel haemostatic agent

Abstract Blood loss remains a major concern during surgery and can increase the morbidity of the intervention. The use of topical haemostatic agents to overcome this issue therefore becomes necessary. Fibrin sealants are promising haemostatic agents due to their capacity to promote coagulation, but their effectiveness and applicability need to be improved. We have compared the haemostatic efficacy of a novel nanostructured fibrin‐agarose hydrogel patch, with (c‐NFAH) or without cells (a‐NFAH), against two commercially available haemostatic agents in a rat model of hepatic resection. Hepatic resections were performed by making short or long incisions (mild or severe model, respectively), and haemostatic agents were applied to evaluate time to haemostasis, presence of haematoma, post‐operative adhesions to adjacent tissues, and inflammation factors. We found a significantly higher haemostatic success rate (time to haemostasis) with a‐NFAH than with other commercial haemostatic agents. Furthermore, other relevant outcomes investigated were also improved in the a‐NFAH group, including no presence of haematoma, lower adhesions, and lower grades of haemorrhage, inflammation, and necrosis in histological analysis. Overall, these findings identify a‐NFAH as a promising haemostatic agent in liver resection and likely in a range of surgical procedures.

Several haemostatic agents approved by Food and Drug Administration have been developed to maintain haemostasis in multiple surgical procedures. Topical haemostatic agents can be classified according to their composition in matrix-based, fibrin and/or thrombin-based and combined products (matrix + fibrin and/or thrombin-based haemostatic agent). Matrix-based products form a barrier that interrupt blood flow facilitating clot formation (e.g., Gelfoam, Hemopatch) whereas fibrin and/or thrombin-based products provide factors and compounds that promote generation of fibrin clots (e.g., Tissucol, Evicel). Combined products (e.g., Tachosil, Evarrest) seem to achieve better results (Brustia et al., 2016).
Fibrin sealant derived from human plasma has become a popular haemostatic agent due to its capacity to imitate and promote the final step of the coagulation cascade (Mankad & Codispoti, 2001;MOSESSON, SIEBENLIST, & MEH, 2006). Furthermore, fibrin sealant is the only Food and Drug Administration clinical material approved as a haemostatic sealant and adhesive agent for therapeutic use.
Despite recent advances in the field, however, overall haemostatic suitability in terms of effectiveness (time to haemostasis), applicability, or absence of rebleeding or undesired adhesion remains far from perfect and could be further improved (Spotnitz, 2014).
We have previously developed a nanostructured fibrin and type VII agarose hydrogel (NFAH) as a scaffolding biomaterial compatible with multiple tissue models (Alaminos et al., 2006;Carriel et al., 2012Carriel et al., , 2013Carriel, Garzón, Campos, Cornelissen, & Alaminos, 2014;Garzón et al., 2014;Sanchez-Quevedo et al., 2007). This biomaterial can be seeded with different cell types and has been characterized as a substitute for cornea, oral mucosa, and nerve (Carriel et al., 2017;Ionescu et al., 2011;Rodríguez et al., 2012). This stromal-like substitute is composed mainly of human plasma and shows high flexible and elastic properties as well as mechanical strength (Mosesson et al., 2006;Scionti et al., 2014). Moreover, fibroblasts and other cell types can be seeded inside the scaffold with good viability and function (Carriel et al., 2013).
37°C with the secondary antibody, an Alexa Fluor 594 anti-mouse IgG antibody (A11005; Thermo Fisher Scientific). Nuclei were stained and mounted with ProLong™ Gold Antifade Mountant with DAPI (P36931; Thermo Fisher Scientific). Fluorescent microscopy was performed on a Nikon TiS microscope (Nikon Instruments, Amsterdam, The Netherlands).

| Karyotyping
For cytogenetic analysis, the karyotype with Bands-G of the fibroblast

| Generation of the NFAH pad
The protocol for NFAH production is an adaptation of methods described previously (Alaminos et al., 2006;Carriel et al., 2012;2013;2014;Garzón et al., 2014;Scionti et al., 2014). To obtain the volume of fibrin-agarose hydrogel (FAH) needed to coat a six-well plate with 24 mm transwell inserts (3450; Corning Incorporated, Kennebunk, ME, USA), a 30-ml mixture was prepared as follows: 25 ml of human plasma (Tebu-bio, Le-Perray-en-Yvelines, France), 0.5 ml of tranexamic acid (Amchafibrin 500 mg, Rottapharm, Milan, Italy), and 2 ml of culture medium, with or without fibroblasts, was placed into a 50 ml conical tube. Subsequently, a solution containing 1.8 ml of 10% calcium chloride (B.Braun, Melsungen, Germany), 1.2-ml PBS, and 1.5 ml of liquid 2.2% type VII agarose (Sigma-Aldrich) was added. After mixing, 5 ml of the resulting solution was poured into each transwell, and the plate was left at 37°C for 2 hr. Once gelation was complete, the FAH was covered with culture medium and maintained at 37°C for 24 hr (plates remained in the incubator for a further 7 days when FAH contained cells), prior to nanostructuration (Figure 1a).
Nanostructuration is a compression and dehydration process. This technique preserves the fibrin structure while removing most of its water content. As a consequence, rheological properties of FAH are changed increasing elasticity and resistance (Ionescu et al., 2011). To this purpose, FAH was placed between two extra-thick western blotting filter paper (88620; Thermo Fisher Scientific). Two 10-μm nylon net filters (NY-1009000; Merck Millipore, Burlington, MA) were positioned between the sample and the blotting paper to prevent adherence. Rapidly, a flat glass surface of 0.25 kg is positioned on the top for 1 min and 40 s for compression. The final acellular or cellularized NFAH has a high density with approximately 80% dehydration and a thickness of 50-60 μm.

| Animal protocol and hepatic resection
Male Wistar rats (200-250 g) were distributed in two experimental groups, according to the extension of tissue resection. The first short incision or mild experimental model (n = 20) includes two animal groups distributed in PCC (n = 10) and c-NFAH (n = 10). The second long incision or severe experimental model (n = 45) includes four animal groups distributed in PCC (n = 15), FTC (n = 10), c-NFAH (n = 10), and a-NFAH (n = 10). Anaesthesia was induced by subcutaneous injection of 80-mg/ kg ketamine and 10-mg/kg xylazine and was maintained by isoflurane inhalation. After abdominal shaving, animals underwent a longitudinal laparotomy 1 cm below the xiphoid process in the craniocaudal direction, leaving the hepatic median lobe exposed for the procedure. Hepatic resection was performed in animals with an incision of 0.5 (short incision or mild model) or 1.5 cm (long incision or severe model) in length.
All haemostatic agents were applied as round-shaped pads of 12-mm diameter for the short incision and 24 mm for the long incision. Time to haemostasis was defined as the time needed for the arrest of free blood leakage after application of the haemostatic agent. The timer was started at the moment of application of the haemostatic agent over the hepatic surface and was stopped when a complete cessation of blood extravasation through the hepatic resection surface was achieved.
Immediately postrecovery, rats were housed in individually ventilated cages with free access to food and water. If animals required analgesia, ketoprofen was administrated every 12 hr.
Rats were euthanized by cardiac puncture 24 hr after surgery to evaluate post-operative haemorrhage, presence of haematoma, migration of the haemostatic agent, and intra-abdominal adhesion of the haemostatic agent to adjacent intact tissues. Blood was also drawn for the measurement of cytokines by Enzyme-Linked ImmunoSorbent Assay (ELISA). The grade of adhesion was determined by a score of 0-2: 0, no adhesion; 1, thin adhesions separable by gravity; 2, thick adhesions not separable by gravity.
Sections of hepatic tissue attached to the haemostatic agent were fixed in 4% paraformaldehyde for histopathological analysis.

| Microscopy analysis
Paraformaldehyde-fixed paraffin-embedded blocks of liver tissue were cut at a thickness of 4 μm, and sections were stained with the Trichrome Stain (Masson) Kit (Catalog N°HT15-1KT, Sigma-Aldrich) and haematoxylin-eosin (Catalog N°GHS316 and HT110116, Sigma-Aldrich), by standard methods. Sectioned liver samples were analysed by investigators blinded to the treatment groups.
Histological variables were studied by microscopic analysis and categorized as shown in Table 1.

| Statistical analysis
Based on our preliminary data, the sample size was calculated to detect a significant effect d = 0.55, with a power of 80% at a level of significance of 5%. This requires a minimum of 10 animals per group (https://www.anzmtg.org/stats/PowerCalculator/PowerANOVA).
Data are presented as mean ± SEM. Significance was determined using the Mann-Whitney U test or the Kruskal-Wallis analysis of variance test with Dunn's post hoc multiple comparison test. Differences were considered significant at p ≤ 0.05. All statistical analyses were performed using GraphPad Prism 7.0 (GraphPad Software Inc., San Diego, CA). We next tested the haemostatic effect of the c-NFAH pad in a severe model of hepatic resection (long, 1.5-cm incision). We also included three additional treatment groups in this analysis: a-NFAH, to evaluate whether fibroblasts had a beneficial effect on the haemostatic process; FTC, to compare the effect of c-NFAH with another fibrin-based commercial haemostatic product; and PCC, to compare the fibrin-based products with a collagen-based matrix.
Regarding applicability, PCC and FTC were applied by manual pressure as indicated by the manufacturers. By contrast, NFAHs were placed onto the wound without applying any additional pressure ( Figure 1b). This application procedure was sufficient to reach haemostasis with NFAH treatment. After 24 hr, no presence of rebleeding was found in any treated group, and only one FTC pad was displaced in one animal.

| NFAH has a favourable profile outcome for tissue response and inflammatory features
Visual inspection of the wound 24 hr after the procedure in severe model revealed no incidence of macroscopic haematoma in either c-NFAH or a-NFAH groups. By contrast, perilesional haematoma was present in rats treated with PCC and FTC (73.3% and 40%, respectively; Figure 2a). Similar results were obtained in the mild model ( Figure S2B). Examples of haematoma with PCC and no haematoma with a-NFAH are shown in Figure 2b and 2c, respectively.
We also evaluated the presence of undesired adhesions to surrounding healthy tissues at the time of sacrifice. Results showed significant differences among groups, with rats treated with c-NFAH or a-NFAH presenting similar results. Remarkably, some of these animals showed no adhesion at all, with most presenting only thin adhesions. As expected, given that PCC an FTC have strong adhesive properties (they have sealant indications), animals treated with PCC and FTC consistently showed thicker adhesion to adjacent organs ( Figure 2d). Examples of high-grade adhesion induced by PCC and no adhesion by a-NFAH are shown in Figure 2e and 2f, respectively. We next measured the levels of inflammatory factors in blood (at sacrifice) by ELISA, to test the inflammatory response at 24 hr following surgery (minimum of N = 7 animals per group). Regarding CRP levels, there were no significant differences between the a-NFAH, PCC, or FTC groups. However, CRP levels in the c-NFAH group were significantly higher than in the a-NFAH (p < 0.01), PCC (p < 0.05), or FTC (p < 0.001) groups (Figure 3a). By contrast, the PCC group had significantly higher IL-1β levels than the a-NAFH or FTC groups (p < 0.05 and p < 0.001, respectively), whereas no differences were observed between c-NAFH, a-NAFH, and FTC groups (Figure 3b). Furthermore, there were no significant differences in TNF-α levels between the a-NAFH group and the other three groups. However, the c-NAFH and FTC groups presented significantly higher levels than the PCC group (p < 0.01 and p < 0.05, respectively; Figure 3c).
Histopathological variables, including haemorrhage, inflammation, and necrosis, were scored by blinded investigators according to the criteria described in Section 2. Representative images of the different grades of these variables are shown in Figure 4a. Both NFAH treatments showed lower haemorrhage, inflammatory response, and necrosis than PCC or FTC treatment. Further, the PCC group displayed haemostasis when compared with PCC and FTC; however, no significant differences were found between c-NFAH and a-NFAH.
Interestingly, in our hepatic surgical model, better haemostasis was reached using PCC than FTC, in line with a previous report (Lewis et al., 2014).
In addition to time to haemostasis, we investigated other relevant outcomes such as incidence of haematoma formation, rebleeding and/or adhesions, levels of some inflammatory factors, and histology.
In contrast to the PCC and FTC treatments, both NFAH treatments showed a complete absence of haematoma formation and a low incidence of thick adhesions. These results might be due to a combination of the fast adherence of the NFAH pads, their lower time to haemostasis and their easier applicability to the wound surface. Contrary to a previous study, PCC had a higher incidence of haematoma formation than FTC in our model (Lewis et al., 2014). Overall, a-NFAH was superior to the other groups regarding inflammatory factors, with levels equal to or lower than those found with the commercial products. However, c-NFAH had significant higher CRP levels than a-NFAH, likely due to the presence of cells. Similarly, NFAH treatment groups had lower grades of haemorrhage, inflammation, and necrosis.
An additional advantage of NFAHs is the smaller adherent effect on surrounding tissue in contrast with hemostats PCC and FTC. This feature can be a great asset to certain types of surgeries.
Our novel NFAH matrix is composed of a nanostructured fibrin-type VII agarose hydrogel that is completely absorbable and biodegradable and displays optimal biomechanical and rheological properties (flexibility, elasticity, and mechanical strength), with no tissue reactivity or antigenicity (Garzón et al., 2009;Fernández-Valadés Gámez et al., 2017). Agarose is a polysaccharide extracted from certain algae that, besides being an inert product from an immunological perspective, offers natural adhesiveness. This feature, as well as the high fibrin content (such that it does not rely on fibrinogen conversion as FTC, for example), may contribute to the haemostatic effectiveness of NFAH and the rapid sealing of damaged tissue.
We need to mention that, besides the fibrin, other agents used in the formulation may contribute to the haemostatic effect of our NFAH. Tranexamic acid is a procoagulant drug that has an antifibrinolytic action and has been successfully used to reduce blood loss in several types of surgery (Meng, Pan, Xiong, & Liu, 2018;Queiroz et al., 2018;Suh, Kyung, Han, Cheong, & Lee, 2018;Zilinsky et al., 2019). Indeed, it has been shown that the use of tranexamic acid and fibrinogen reduces blood loss and improves coagulation measurements in a porcine model of liver injury (Zentai et al., 2016). However, the final content of tranexamic acid in the NFAH (1.6% or 8.3 mg) is quite low compared with the concentrations used for topical application (100-500 mg) or intravenously (1 g). Similarly, calcium chloride accelerates fibrin polymerization and clot formation (Brass, Forman, Edwards, & Lindan, 1978), and although it is used to stabilize the matrix, we cannot exclude that the haemostatic effect of the NFAH is enhanced by the presence of calcium. As mentioned, no optimal haemostatic agent exists to date, and uncontrolled surgical bleeding remains an unmet clinical need. Any improvement in terms of reduction of time to haemostasis and/or FIGURE 3 Inflammation factor levels in rats subjected to long hepatic incision after haemostatic treatment. (a) CRP, (b) IL-1β, and (c) TNF-α blood levels measured by ELISA after 24 hr in rats with long hepatic incision treated with c-NFAH, a-NFAH, PCC, or FTC. Data represent mean ± SEM. Data were analysed using Kruskal-Wallis analysis of variance test followed by Dunn's post hoc test (N = 7-15); *p < 0.05; **p < 0.01; ***p < 0.001. c-NFAH: cellularizednanostructurated fibrin-agarose hydrogel; a-NFAH: acellularnanostructurated fibrin-agarose hydrogel; FTC: fibrinogen/thrombincoated collagen pad (Tachosil®); PCC: protein-reactive polyethylene glycol-coated collagen pad (Hemopatch®); CRP: C-reactive protein; IL-1β: interleukin 1 beta; TNF-α: tumour necrosis factor alpha rebleeding, along with ease of use and applicability, could pose a huge impact not only in hospital costs but also in patients' quality of life. We show here that NFAH has excellent haemostatic and sealing properties, and we believe that these results will pave the way for this new product to be part of the arsenal of haemostatic agents in multiple surgical procedures in the future.

ACKNOWLEDGEMENTS
The authors are grateful to Gloria Carmona and Rosario Sanchez-Pernaute for scientific advice. We also thank Cynthia Morata for her writing assistance. This work was supported by preclinical research funds from the Regional Government of Andalusia through the Andalusian Initiative for Advanced Therapies.   Table 1 for categorization). c-NFAH: cellularized-nanostructurated fibrin-agarose hydrogel; a-NFAH: acellular-nanostructurated fibrin-agarose hydrogel; FTC: fibrinogen/ thrombin-coated collagen pad (Tachosil®); PCC: protein-reactive polyethylene glycolcoated collagen pad (Hemopatch®) [Colour figure can be viewed at wileyonlinelibrary. com]