Keywords

1 Chitosan Structure and Solubility

According to the Web of Science, the topic of “chitosan and platelets” started gaining attention by the early 2000s, with around 30 papers published per year on this topic. Since then, the number of publications has progressively soared to over 3,200 papers in 2020 alone, with a particular focus in the areas of hemorrhage control and wound repair. To appreciate the molecular and cellular basis of chitosan–platelet interactions, it is first important to understand that chitosan is a family of polymers with distinct chemical structure and physical forms that collectively influence the way the polymer “presents” to platelets.

Chitosan is obtained by chemical N-deacetylation of chitin, a naturally occurring polysaccharide with linearly arranged β-(O)-1-4-linked N-acetyl-d-glucosamine (GlcNAc) residues [1]. The deacetylation step can be simply achieved by autoclaving chitin particles in a 25% w/v sodium hydroxide solution [2, 3]. During this treatment, exposed acetyl groups are stripped from GlcNAc to generate glucosamine (Glc). It was proposed that GlcNAc groups buried in the nucleus of an insoluble chitin particle are protected from deacetylation, giving rise to a “block” acetylation pattern [4]. “Block” acetylation refers to consecutive clusters of GlcNAc residues (i.e., AADA, AAAA, ADAA) interspersed throughout the poly-Glc chain [5, 6]. Repeated autoclaving of chitosan under alkaline conditions can be used to reach >98% DDA, which for all practical purposes is considered fully deacetylated [2]. Because each autoclave cycle produces chain scission, chitosan Mn is inevitably diminished compared to the chitin starting material. Fully deacetylated chitosan can be reacetylated to different % DDA levels using acetic anhydride; this produces chitosans with a random acetylation pattern [2, 7]. All of these production steps create heterogeneities in each chitosan preparation. It is important to recognize that each batch of chitosan has a number-average molecular weight (Mn), degree of deacetylation (DDA), and pattern of acetylation, along with a certain level of polydispersity for each of these characteristics. Each of these structural features can influence the manner in which chitosan interfaces with blood plasma.

As the major structural component of exoskeletons of crustaceans and insects, chitin has evolved to remain firmly insoluble under aqueous conditions. Therefore, chitin particles and scaffolds are by nature insoluble in blood and present slightly hydrophobic poly-GlcNA surfaces to blood components [8, 9]. By contrast, chitosan powders can be dissolved in slightly acidic solutions when at least half of the Glc subunits are protonated (i.e., R-NH3+) [10]. In other words, chitosan will only remain soluble at a pH equal to or below the chitosan pK0. Because all chitosans (60–100% DDA) have a pK0 6.0–6.5 that is altogether below neutral pH [10, 11], this means that acid-soluble chitosan chains will spontaneously transition into insoluble microparticles upon mixing with neutral pH cell culture medium, blood plasma, or whole blood [12,13,14] (Fig. 1). When soluble chitosan chains shift to insoluble microparticles in the presence of serum, the microparticles adopt a slight negative zeta potential suggesting that chitosan-anionic serum factor complexes form spontaneously [12, 14]. Solid freeze-dried chitosan scaffolds generated from fully protonated chitosan solutions will spontaneously solubilize in whole blood or platelet-rich plasma and then disperse as microparticles in the coagulum [15,16,17]. By contrast, chitosan matrices treated with alkaline solutions and/or organic solvents show depressed surface energy and wettability compared to Glc and GlcNAc monomers [18]. Chitosan matrices cured in alkaline conditions have neutral amine groups and present neutral, insoluble surfaces to blood components (Fig. 1b). Chitosan neutral–solubility is enhanced by decreasing molecular weight and by very low deacetylation levels (i.e., 50% DDA) [7, 19]. Small chitosan oligomer chains (≤5 kDa) are fully soluble at neutral pH for all DDA levels [2, 20].

Fig. 1
figure 1

For chitosan preparations with a molecular weight ≥10 kDa and 60–100% DDA, solution pH controls whether chitosan is (a) cationic and soluble or (b) neutral and mostly insoluble

Full size image

2 Platelets

Platelets are small anuclear disc-shaped cells released by megakaryocytes into the blood circulation [21]. Megakaryocytes are large multinuclear cells derived from a common myeloid progenitor that normally develop and reside in the bone marrow. At full maturity these cells can reach a ploidy up to 64 N [22]. Megakaryocytes are also found in the circulation and are known to populate heart and lung tissues [22,23,24,25]. According to the “pulmonary platelet production model”, a significant portion of platelets could be produced in the lung [22]. It was suggested that shear forces in the lung vasculature shred apart “proplatelets” released by bone marrow megakaryocytes, while other studies using transgenic mice carrying green fluorescent protein-expressing megakaryocytes revealed that circulating megakaryocytes can lodge in pulmonary vessels and eject platelets directly into the pulmonary vasculature [22, 24, 26]. Circulating platelet levels are under the control of a humoral factor, thrombopoietin (TPO), that is released to the circulation mainly by the liver [22]. Platelet counts in normal adult peripheral blood range from around 115 to 400 × 106/mL. Platelet counts vary with sex and ethnicity, with higher average levels reported for Caucasians compared to Afro-Caribbean and African groups and relatively higher levels in women compared to men [27]. Activated platelets can generate small vesicles called microparticles; microparticles are sub-micron lipid vesicles released from a variety of cell types, including platelets, that wind up in the circulation [28]. Several markers are used to identify and characterize platelets, including membrane receptors P-selectin (CD62P), GPIb/IX/V (CD42), and the GPIIb/IIIa integrin complex (CD41/CD61) (Table 1).

Table 1 Platelet characteristics
Full size table

Like red blood cells, platelets have anionic surfaces that repel their spontaneous adhesion to endothelial cells, erythrocytes, and leukocytes whose surfaces are also negatively charged [29]. The platelet glycocalyx is around 20–30 nm thick, an anionic coating of integral membrane glycoproteins, proteoglycans, glycolipids, and factors adsorbed from the blood plasma [30]. Platelets also carry blood group antigens at variable levels on several transmembrane receptors including GPIb and GPIIa and GPIIIa [31]. ABH antigens are fucosylated carbohydrate modifications of blood glycoproteins and glycolipids known to influence the risk of thrombosis in certain disease states [32,33,34]. In a cohort of 313 Japanese healthy volunteers, platelets showed a “high ABH antigen expressor” phenotype in 7% of the donors [31]. These data highlight donor-specific variations in the platelet glycome. In keeping with other cells, platelets show membrane asymmetry with negatively charged phospholipids hidden in the internal membrane.

Platelets can be activated through distinct mechanisms in static whole blood samples and under shear stress generated by blood flow. During whole blood coagulation initiated in vitro by the extrinsic pathway, thrombin activation initiates platelet activation, and activated platelets help propagate clot formation. Platelet activation is often monitored by the release of alpha granule contents. In a kinetic studies of tissue factor (TF)-induced clotting of recalcified citrated whole blood [35, 36], early thrombin activation (as measured by the appearance of fibrinogen peptide A) occurred at 1 min, platelet activation (measured by release of osteonectin, or platelet factor 4, PF4/CXCL4) occurred at 2–3.5 min, and clotting time (measured by appearance of “clumps” on the sidewalls of the tubes) was observed at 4.7 min [35, 36]. It was estimated that platelets had degranulated around 50% of their alpha granule contents at clotting time [36]. In addition to PF4 and osteonectin, platelet alpha granules release a plethora of bioactive factors: platelet-derived growth factor (PDGF), transforming growth factor β1 (TGF-β1), vascular endothelial growth factor (VEGF), as well as procoagulant factors Factor V (FV), von Willebrand factor (vWF), fibrinogen, and thrombospondin [37]. Activated platelets release other factors from delta granules: adenosine nucleotide diphosphate (ADP) and adenosine triphosphate (ATP), serotonin, calcium, and polyphosphate to name a few (Table 1). Upon activation, platelets change their shape from disc-shaped to spiked spheres, adhere to surfaces, and aggregate. Activated platelets undergo cytosolic calcium spikes, release alpha and delta granule contents, express GPIIb/IIIa on the platelet surface, internalize GPIb, and lose plasma membrane asymmetry which exposes phosphatidylserine on the outer membrane [30]. Phosphatidylserine is a docking site for calcium-dependent binding of activated gla-domain coagulation factor assemblies at the platelet membrane [38]. Generation of thromboxane and thrombin at the platelet surface, along with delta granule release of serotonin and ADP, further promotes platelet activation via platelet surface receptors for each of these agonists [39].

Shear-stress induced platelet activation uses vWF to immobilize platelets at the damaged endothelial cell surface. vWF is a large multimeric glycoprotein that is present in the circulation and stored in platelet alpha granules and in Weibel-Palade bodies of endothelial cells [40]. Upon endothelial damage under blood flow-induced shear stress, vWF “strings” are ejected into the blood stream, forming an attachment site for platelets through GPIb/IX/V receptors [41]. Following platelet activation, the platelet integrin receptor GPIIb/IIIa becomes clustered which enables it to engage with subendothelial collagen fibers, fibrin, vWF and thrombospondin through an RGD sequence [30, 42]. Platelet aggregation during coagulation, and anchoring to fibrin through GPIIb/IIIa receptor interactions, is an essential step in clot retraction and hemostasis [42]. The platelet plug forms a “white thrombus” to seal a damaged blood vessel. Given the essential role for platelets in hemostasis, biomaterial features that stimulate platelet activation are clearly important for applications in hemorrhage control. In other applications involving vascular stents, however, attention has focused on formulating chitosan in a manner to inhibit platelet activation [43].

All of the studies analyzed in this chapter used “normal” or “healthy” donor platelets for chitosan interaction studies, however it is important to keep in mind that platelets from unhealthy individuals could potentially show different responses [33]. Evidence has been accumulating to suggest that platelet composition could be altered by certain disease states. Bone marrow megakaryocytes in normal individuals experience a process termed “emperipolesis” [30, 44]. Emperipolesis is a phenomenon distinct from phagocytosis and involves megakaryocyte ingestion of neutrophils, eosinophils, erythrocytes, and lymphocytes without destruction of ingested cell integrity [30]. This remarkable event was originally suggested to serve as a mechanism for neutrophil transit from the bone marrow into the vasculature [30]. Emperipolesis is enhanced by endotoxin [45] and in experimental models of idiopathic myelofibrosis [44]. Video-epifluorescence microscopy showed that megakaryocyte uptake of neutrophils resulted in mingling of megakaryocyte-neutrophil membranes and the appearance of neutrophil membranes on circulating platelets [46]. The potential implications of these observations, notably in contexts where emperipolesis involves the trans-cellular transit of inflammatory neutrophils, remain to be identified.

3 Chitosan Interfacing with Blood Plasma

Chitosan membranes, sponges, and solid scaffolds are readily created by processes that neutralize or chemically cross-link glucosamine residues, either before or after electrospinning, surface coating, freeze-drying, precipitation or gelation [1]. These chitosan matrices and substrates are intended to remain intact after immersion in blood or blood plasma [47, 48]. Extensive modification of chitosan amine groups by chemical cross-linking reduces the density of cationic residues available for interaction with anionic blood factors. Furthermore, when solid chitosan matrices are treated with alkaline solutions or organic solvents, they strongly resist dissolving, even when hydrated in acidic aqueous solutions. In hemorrhage control, use of a solid matrix is important because chitosan solutions or powders were found to slow bleeding from capillaries but were unable to arrest arterial bleeding without additional pressure to seal off the broken blood vessel.

In certain wound-repair applications, formulations that generate chitosan microparticles are preferred [16, 49]. For example, when a liquid 80% DDA chitosan isotonic solution pH 4.5 is dispersed into unmodified whole blood and allowed to clot in a glass tube for 30 min at 37°C, the soluble chitosan chains can be observed to precipitate due to the neutral pH environment to form a hybrid biomaterial clot with chitosan microparticles interspersed in the fibrin network (Fig. 2a, b). In another approach, a freeze-dried (FD)-chitosan formulation is prepared with a fully protonated chitosan solution (pH 2.5) under controlled lyophilization conditions (inset, Fig. 2c). When this FD-chitosan scaffold is immersed in recalcified citrated platelet-poor plasma and kept at 37°C for an hour to permit fibrin clot formation (Fig. 2c), the chitosan scaffold rehydrates and disperses to form micro-hydrogel particles within the resulting fibrin clot network (Fig. 2d) [16]. Other FD-chitosan formulations intended for PRP mixing were optimized to contain ~83% DDA ~40 kDa chitosan, lyoprotectant, and calcium chloride. These FD-chitosan cakes solubilize, disperse in platelet-rich plasma (PRP), and form a micro-hydrogel particle scaffold network within the PRP fibrin clot structure [50].

Fig. 2
figure 2

In certain wound-repair applications, chitosan microparticles are preferred. Two chitosan formulations incorporated with rhodamine isothiocyanate (RITC)-chitosan tracer were used to form hybrid whole blood or recalcified platelet-poor plasma clots and imaged by epifluorescence microscopy. A 1.6% w/v chitosan/100 mM glycerol phosphate solution (RITC-chitosan 80.6% DDA, Mn 46.7 kDa) was mixed at a 1:3 v/v ratio with peripheral whole rabbit blood collected under IRB-approved protocols, and solidified ex vivo in a glass tube to form (a) a cylindrical hybrid clot implant with (b) RITC-chitosan particles interspersed in the blood clot. (c) FD-chitosan (1% w/v chitosan/HCl, pH 2.5, inset) was placed in recalcified citrated human platelet-poor plasma (Precision Biologics, Halifax, Canada) with glass beads to activate coagulation. The FD-chitosan scaffold spontaneously rehydrated and dispersed as micro-hydrogel particles in the fibrin clot (d). Panel (d) is reprinted from Materialia, Vol 9, Hoemann CD et al. Guided bone marrow stimulation for articular cartilage repair through a freeze-dried chitosan microparticle approach, 100609, 2020, with permission from Elsevier

Full size image

As a cationic polysaccharide, chitosan is capable of forming electrostatic complexes with a variety of blood factors, most of which are anionic. Benesch and Tengvall showed that chitosan-coated surfaces develop a 10 nm thick layer of serum protein that includes serum albumin and complement C3 [51]. It was subsequently shown that in platelet-poor plasma (PPP), chitosan microparticles form electrostatic complexes with anionic zymogens C3, C5, and Factor B without activating complement [14]. When liquid chitosan is mixed with blood, the chitosan chains form complexes with red blood cells and induce red blood cell agglutination [52]; this bioactivity is attributed to electrostatic binding between chitosan and sialic acid residues on the RBC surface [53]. Platelet activation also activates the complement cascade, therefore chitosan preparations that activate platelets in whole blood or PRP may also drive complement activation indirectly at the platelet surface [14, 54].

The blood stream is maintained in a fluid state partly through its anionic environment. Endothelial cells are coated with a 0.5 mm thick glycocalyx with dense negatively charged heparan sulfate, chondroitin sulfate, and hyaluronan that is shed into the circulation [55]. The platelet surface carries GPIb receptors with O-linked glycosylations capped with sialic acid, which contribute significantly to the platelet’s negatively charged surface [42]. vWF carbohydrate chains are also capped with sialic acid [40, 41]. Desialylation of vWF by neuraminidase leads to its clearance by the liver through the asialoglycoprotein receptor (ASGPR), also called the Ashwell receptor [41, 56]. The Ashwell receptor was also found to be responsible for clearing desialylated platelets that arise during sepsis by the action of bacterial neuraminidase [57]. Fibrinogen is also sialylated [58]. Sialyl groups in fibrinogen were suggested to serve as low-affinity calcium binding sites that favor fibrin self-assembly and branching of fibrin fibers [58]. Because platelet–vWF and platelet–fibrin interactions are mediated through integrin protein receptor interactions (GPIb:vWF; GPIIb/IIIa:fibrin), sialic acid residues in vWF, fibrin, and GPIb are free to interact with other binding partners including cationic glucosamine residues in chitosan. To summarize, the sialic acid modification of glycosylations in many blood factors, blood cell surface receptors, and platelet receptors represents a “sialome” [59] that creates a physicochemical basis for cationic chitosan complex formation. Other anionic blood proteins have a pKa 6.0 or lower, which enables their binding to chitosan through electrostatic protein–polysaccharide interactions [14].

4 Chitosan–Platelet Interactions

One of the earliest clues that chitosan has procoagulant activity came from experiments by Malette et al., who observed that a porous DeBakey Dacron graft soaked in a 2 mg/mL chitosan solution, but not saline, prevented acute fatal bleed-out when stitched over the excised aorta of heparinized canines [60]. One day later, the chitosan coagulum was found “inside the graft, plugging the interstices” [60]. Klokkevold et al. subsequently observed that lingual hemostasis in rabbits could be accelerated by application of an 80% DDA chitosan solution in acetic acid [61]. In these early reports, the hemostatic mechanism of action was assumed to result from chitosan-induced red blood cell agglutination which creates a “red” thrombus. It was subsequently shown that chitosan-coated microtiter plates, acidic chitosan solutions and chitosan or chitin solid microparticles induce rapid platelet adhesion, aggregation, and elevated surface expression of GPIIb/IIIa in the absence of calcium [62, 63], with one report that chitosan particles could enhance the release of platelet alpha granule contents under calcium-free assay conditions [63] (Table 2). By contrast, neutral-soluble chitosans including chitosan oligomers showed no ability to induce platelet aggregation in vitro and were even slightly anticoagulant [64, 65].

Table 2 Studies on chitosan–platelet interactions
Full size table

Since these original reports, a variety of hemostatic chitosan devices have been developed, the majority of which are solid matrices, films, coatings, gels, filaments, or sponges that produce hemostasis through mechanisms purporting to involve red blood cell agglutination and/or platelet activation [66, 67]. These devices are intended to promote hemostasis and reside temporarily in the wound. It is important to note that platelet–chitosan interaction assays have been carried out either in calcium-free (citrated) plasma or buffers, or in the presence of calcium which permits induction of thrombin and subsequent platelet activation by thrombin and a variety of other endogenous agonists.

Depending on the physical form of a chitosan device, platelets may interact with a 2-D or 3-D chitosan surface, filament, microparticle, or soluble chitosan chain. The Syvek patch is composed of solid chitin fibers, whereas Clo-sur PAD and ChitoSeal are solid patches containing chitosan. Platelets were found to adhere to all three scaffolds, with evidence of induced P-selectin and GPIIb/IIIa expression when calcium-activated PRP was combined with a slurry of GlcNAc fibers [8, 9]. Platelets also adhered and aggregated on the smoother surfaces of chitosan-based Clo-sur PAD after 3 min, but were mainly observed adhering to exposed polyethylene terephthalate (PET) polymer surfaces on ChitoSeal [8]. In a swine splenic capsular stripping hemorrhage model, the Syvek patch produced hemostasis after three compressions compared to 8–10 for gauze, Clo-sur PAD or ChitoSeal [8]. In another study analyzing platelet adhesion under calcium-free conditions by Romani et al., alkaline-cured chitosan films supported platelet adhesion but the platelets remained discoid in shape and scarcely expressed P-selectin, while platelets adhering to glass coverslips became spiked and had strong P-selectin expression [68]. In another study by Sundaram et al., a chitosan-aluminum sulfate-calcium hydrogel was created that induced platelet aggregation, RBC agglutination, and more rapid hemostasis in rat hemorrhage models [69] (Table 2).

For wound-healing applications, some chitosan devices have been formulated to intermingle as micron-sized hydrogel particles within blood or platelet-rich plasma clots. As mentioned above, viscous isotonic chitosan solutions can be mixed with unmodified whole blood and allowed to coagulate to form a hybrid blood clot [53]. Mixture of liquid chitosan-glycerol phosphate solutions into whole blood produces a hybrid clot that, like whole blood, solidifies through the common pathway of the coagulation enzyme cascade [70]. In thromboelastography (TEG) tests, hybrid clots containing 80% DDA chitosan showed thrombin activation starting at 20 min, followed by burst platelet activation around 30 min marked by release of PF4/CXCL4 and PDGF-ββ into the serum [70, 71]. The dispersion of 80% DDA chitosan microparticles in the hybrid clot implant was found to inhibit clot retraction, protect the clot from lysis by plasmin, promote longer in situ clot residency, attract neutrophils and M2 macrophages, and exert favorable effects on osteochondral wound repair [72, 73].

Here we report the effect of chitosan structure on platelet activation in whole blood, using a library of chitosans with 80% or 95% DDA and low (80L, 95L) or high viscosity (80M, 95M) [12]. Each chitosan was prepared as a 1.6% w/v solution in 100 mM glycerol phosphate (pH 6.0) and mixed or not with unmodified human whole blood from 4 healthy donors. Samples were either harvested immediately, submitted to a TEG assay at 37°C for 160 min to assess clotting time and clot tensile strength, or cultured at 37°C for 240 min in glass tubes. As a model system, blood clots held at body temperature have the potential to reflect platelet responses that could occur in a blood clot device in vivo. At different time points (t = 0, 160 or 240 min), using previously described methods [70], samples were vortexed in ice cold quench buffer, centrifuged, and the supernatant analyzed by Western blotting for PF4, or ELISA assay for thrombin-antithrombin (TAT) complex (Dade) (Fig. 3).

Fig. 3
figure 3

Chitosan structure influenced the timing of platelet degranulation, clotting time and clot tensile strength in unmodified whole human blood. Under IRB-approved protocols, whole blood from 4 healthy human donors was mixed at 3:1 ratio with chitosan-glycerol phosphate solutions (80L, 80M, 95L, 95M) to evaluate (a) PF4 levels in plasma or serum using equal amounts of protein per lane and purified PF4 as a control, (b) TAT complex formation in plasma or serum, (c) clotting time (R, min), and (d) clot amplitude at 160 min (A, mm) by TEG assay. Chitosan (DDA and Mn) were 80L (80.2%, 108 kDa), 80M (80.6%, 144 kDa), 95L (94.6%, 102 kDa), 95M (94.6%, 177 kDa) [12]. Graphs show the median, 50% inter-quartile range, min-max, and differences due to condition were evaluated by Least Squares Means Differences with Tukey HSD post-hoc (JMP Pro 14.1.0, SAS). R clotting time, NA not available, WB whole blood, TAT thrombin-antithrombin, PF4 platelet factor 4, M male, F female

Full size image

95M chitosan (high DDA and medium viscosity) had a unique effect on inducing instant release of PF4 at t = 0 in all 4 donors (lane 4, Fig. 3a), prior to thrombin activation (Fig. 3b). 95M also created a rapid increase in blood viscosity that was “read” by the TEG instrument as a more rapid “clotting time” compared to all other chitosans (Fig. 3c). All chitosan-GP/blood mixtures coagulated and showed robust TAT generation at 160 and 240 min (Fig. 3b). However lower molecular weight chitosans (80L, 95L) delayed clotting time (Fig. 3c), and 95L induced a weaker clot tensile strength compared to other clot samples (Fig. 3d). After 240 min at 37°C, PF4 was detected in serum of all chitosan-GP/blood samples, with the highest PF4 levels in 95L and 95M serum (lanes 5–8, Fig. 3a).

In this experiment, unmodified whole blood had a median 18.9 min clotting time, 47.0 mm clot tensile strength (Fig. 3c, d), robust TAT levels after 240 min (48 μg/mL, Fig. 3b), and variable serum PF4 levels at 160 min (lane 9, Fig. 3a). Donor-to-donor differences in serum PF4 levels at 160 min were not explained by selected donor characteristics (Fig. 3a). These results are consistent with previous analyses of healthy donor whole blood: 15.3 ± 4.8 min clotting time, 55.2 ± 5.7 mm maximal amplitude, ~37 μg/mL TAT, and burst appearance of PF4 at 30 min that decayed over time [70]. Altogether, these data suggested that highly deacetylated cationic chitosan chains formed complexes with platelets and red blood cells and may have formed inhibitory complexes with factors that degrade PF4. These results also suggested that platelet responses to chitosan can be influenced by the chitosan charge state at the time of platelet contact and that cationic chitosan is an effective platelet-activating biomaterial. However strong chitosan platelet–interactions by 95L were accompanied by prolonged clotting time and weaker clot tensile strength, which are considered anti-hemostatic effects.

Many experiments analyzing chitosan–platelet interactions are carried out with PRP (Table 2). PRP is generated from citrated whole blood by sequential centrifugation (200 g 15 min to pellet RBC and WBC, then 1,300 g 10 min to pellet platelets). Citrate chelates all plasma calcium and serves as an anticoagulant because calcium is required for propagation of the coagulation factor enzyme cascade leading to thrombin activation and conversion of fibrinogen to fibrin. Therefore, PRP must be recalcified to produce a PRP fibrin clot. Mixture of freeze-dried chitosan containing calcium chloride with PRP resulted in solubilization of chitosan in plasma followed by chitosan coating of platelet surfaces, hybrid chitosan/PRP clot formation, inhibition of PRP fibrin clot retraction, enhanced release of PDGF-ββ, and higher P-selectin expression [15, 17, 50]. Unlike other chitosan–platelet interactions, this FD-chitosan preparation enhanced platelet degranulation while inhibiting platelet aggregation [17]. It was suggested that chitosan inhibited clot retraction in these PRP samples by coating the platelet surfaces and preventing their aggregation [17]. FD-chitosan/PRP showed beneficial biological effects compared to PRP-alone in promoting chondrogenesis in a rabbit cartilage repair model, with greater repair tissue collagen type II content compared to PRP-alone [74]. In other applications, a shape-memory gel containing glycidyl methacrylate functionalized quaternized chitosan, with a permanent positive charge, showed platelet adhesion and aggregation on the surfaces after incubating with citrated PRP for 1 h at 37°C [75]. In the latter study, calcium was not added to the PRP preparation, suggesting the interaction was based on electrostatic interactions between anionic platelet surfaces and the cationic quaternized chitosan.

Platelet adhesion is an undesired event for vascular stents. Coating of a NiTi surface with layer-by-layer (LbL) hyaluronic acid (HA) and chitosan, with the final layer being HA was shown to reduce platelet adhesion from 1 million to 0.6 million platelets/cm2 [43]. The persistent platelet interactions with LbL-HA(Chitosan/HA)4 could be due to an irregular surface coating by HA after only 4 bi-layers that left exposed chitosan, as another study suggested that at least 6 alternating layers are needed to create a uniform anionic polyelectrolyte surface [76]. By contrast, surface coatings created with chitosans derivatized with hydrophobic butyryl functional groups and covalently attached to cellulose were shown to repel platelets and inhibit platelet adhesion and activation [77] (Table 2).

5 Conclusions and Future Perspectives

To summarize, current data suggest that chitosan–platelet interactions occur through two mechanisms. By one mechanism, electrostatic binding between the anionic platelet surface and cationic chitosan poly-glucosamine residues produces platelet adhesion that can trigger a rise in intracellular calcium, degranulation, and appearance of integrin and P-selectin receptors at the platelet surface [8, 62, 69, 75] (Fig. 4). Platelet GPIb is heavily sialylated and could potentially mediate this interaction. Neutral chitosan and chitin have little capacity to induce these responses. Although most solid matrices used as hemostatic devices have no charge (poly-GlcNAc), or are neutralized during manufacturing, recent advances have created quaternary chitosans with a permanent cationic charge state [75], which enables mechanism 1. By a second mechanism that is likely to take place in blood or PRP, abundant anionic blood plasma proteins (i.e., sialyl-fibrinogen, fibronectin, sialyl-vWF, albumin) rapidly bind to chitosan surfaces. Blood proteins are also expected to deposit on chitin surfaces. Deposition of these matricellular factors, also called “biofouling” presents a surface that permits platelet adhesion, activation, GPIIb/IIIa clustering, and the potential for specific receptor–ligand interactions between platelet integrins and RGD motifs in the protein coating (Fig. 4). Evidence for this second mechanism comes from slightly greater platelet aggregation induced by 2.8–6.9 μm chitin or chitosan particles in the presence of plasma proteins compared to platelets washed to remove plasma proteins [63]. Future directions on this topic will benefit from further elucidation of molecular mechanisms involved in chitosan-platelet adhesion and experiments that show the biological relevance of these interactions in whole blood where erythrocytes vastly outnumber platelets. More information is also needed on the performance of platelet-activating chitosan devices in clinical contexts, most notably where platelet physiology is altered by medications, trauma, or disease state.

Fig. 4
figure 4

Two mechanisms are proposed to explain chitosan–platelet interactions. By mechanism 1, platelet anionic surfaces form non-specific electrostatic interactions with chitosan amine groups that are initially protonated at the time of contact with whole blood or PRP. Because anionic blood proteins (including sialylated proteins such as fibrinogen) vastly outnumber platelets, most platelet interactions in PRP or blood likely occur through mechanism 2. In mechanism 2, soluble factors in blood plasma form complexes with chitosan or chitin and display binding sites, some of which may be specific high-affinity interactions via platelet surface receptors (i.e., integrin GPIIb/IIIa, GPIV, GPIc/IIa, GPIc*/IIa, GPIb/IX/V). The schematics show only one side of the platelet surface and are not to scale: platelets are on average 2 μm in diameter [30]. Note that chitosan microparticles that form in blood plasma can have a similar or smaller size compared to platelets, or a much larger size when chitosan aggregates are formed [12]

Full size image