Injectable Thermosensitive Chitosan-Collagen Hydrogel as A Delivery System for Marine Polysaccharide Fucoidan

Fucoidans, sulfated polysaccharides from brown algae, possess multiple bioactivities in regard to osteogenesis, angiogenesis, and inflammation, all representing key molecular processes for successful bone regeneration. To utilize fucoidans in regenerative medicine, a delivery system is needed which temporarily immobilizes the polysaccharide at the injured site. Hydrogels have become increasingly interesting biomaterials for the support of bone regeneration. Their structural resemblance with the extracellular matrix, their flexible shape, and capacity to deliver bioactive compounds or stem cells into the affected tissue make them promising materials for the support of healing processes. Especially injectable hydrogels stand out due to their minimal invasive application. In the current study, we developed an injectable thermosensitive hydrogel for the delivery of fucoidan based on chitosan, collagen, and β-glycerophosphate (β-GP). Physicochemical parameters such as gelation time, gelation temperature, swelling capacity, pH, and internal microstructure were studied. Further, human bone-derived mesenchymal stem cells (MSC) and human outgrowth endothelial cells (OEC) were cultured on top (2D) or inside the hydrogels (3D) to assess the biocompatibility. We found that the sol-gel transition occurred after approximately 1 min at 37 °C. Fucoidan integration into the hydrogel had no or only a minor impact on the mentioned physicochemical parameters compared to hydrogels which did not contain fucoidan. Release assays showed that 60% and 80% of the fucoidan was released from the hydrogel after two and six days, respectively. The hydrogel was biocompatible with MSC and OEC with a limitation for OEC encapsulation. This study demonstrates the potential of thermosensitive chitosan-collagen hydrogels as a delivery system for fucoidan and MSC for the use in regenerative medicine.


Introduction
Bone tissue has good self-healing capacities [1]. However, large and irregular fractures or diseases such as diabetes can affect the regeneration of the tissue. It is estimated that 5-10% of bone fractures fail to heal, causing long-term disability in the patient [2,3]. The local application of bioactive compounds or growth factors can stimulate molecular processes to support tissue regeneration [4,5]. Fucoidans, sulfate-rich polysaccharides from the cell wall of brown algae, have shown to be promising candidates for the use in regenerative medicine [6,7]. These polysaccharides exhibit multifaceted bioactivities in processes which are essential for successful bone regeneration, including angiogenesis, osteogenesis, and inflammation [8][9][10][11]. For the local application of fucoidan, an appropriate

Preparation and Physicochemical Characterization of the Fucoidan Delivery System
In the presented study, an injectable delivery system for fucoidan and stem cells based on chitosan, collagen, and β-GP was developed. To avoid fucoidan aggregates, it was necessary to premix the fucoidan solution with the β-GP solution immediately before adding it to the chitosan-collagen mixture. Preliminary experiments revealed that a β-GP content of 7% resulted in adequate gelation times. As demonstrated in Figure 1a, the clear sol turned into an opaque hydrogel at 37 • C. The tube-inverting method was applied to roughly estimate the time until gelation. After 3-4 min, flow of the sols could not be observed any longer, indicating the gelation of the hydrogels. The incorporation of fucoidan had no visible influence on the flow characteristics ( Figure 1b). To study the gelation time and duration of the gelation process in more detail, time sweep tests were performed at 37 • C. Therefore, the sol was pipetted onto the preheated platform and the storage (G ) and loss modulus (G ) were measured for 16 min. The gel point is reached when the storage modulus, describing the elastic portion of the material, becomes larger than the loss modulus, describing the viscous portion of the material. As shown in Figure 1c, the sol-gel transition point is marked with a dashed line. Over time, storage and loss modulus increased, indicating the hydrogel formation, until reaching a plateau. The end of the gelation process is marked with another dashed line. It was found that the gel point of hydrogels without fucoidan occurred after approximately 75 s at 37 • C. The gel point of hydrogels containing 100 µg/mL fucoidan occurred after approximately 59 s at 37 • C ( Figure 1d). The gelation process of both hydrogels, with and without fucoidan, lasted approximately for 6 min until reaching a stable state (Figure 1e).  Further rheological measurements were performed to determine the exact gelation temperature of the hydrogels without and with fucoidan. Temperature sweep tests as shown in Figure 2a revealed that hydrogels without fucoidan started to gelate at 36.8 • C. Hydrogels containing 100 µg/mL fucoidan already started to gelate at 35.3 • C. These results indicate that the incorporation of fucoidan into the hydrogel tentatively decreased the temperature at which the gel point is reached (Figure 2b).  The gelation temperature was plotted for hydrogels without and with 100 µ g/mL fucoidan. Experiments were repeated three times (n = 3).
Next, the pH, swelling properties, and the turbidity of the hydrogels containing different concentrations of fucoidan were examined. As shown in Figure 3a, the pH of the sols was approximately 6.8. After gelation, the pH (~6.6) was slightly more acidic. The incorporation of fucoidan had no impact on the pH before and after gelation. The swelling capacities of the hydrogels were quantified by determining the equilibrium swelling ratio (ESR). Therefore, hydrogels were lyophilized and rehydrated in PBS until reaching equilibrium. The ESR is defined as ESR = (ws − wd/wd) with ws and wd being the weight of the swollen and the dried gels, respectively. High fucoidan concentrations affected the swelling capacities of the hydrogel as demonstrated in Figure 3b. Hydrogels containing 500 Next, the pH, swelling properties, and the turbidity of the hydrogels containing different concentrations of fucoidan were examined. As shown in Figure 3a, the pH of the sols was approximately 6.8. After gelation, the pH (~6.6) was slightly more acidic. The incorporation of fucoidan had no impact on the pH before and after gelation. The swelling capacities of the hydrogels were quantified by determining the equilibrium swelling ratio (ESR). Therefore, hydrogels were lyophilized and rehydrated in PBS until reaching equilibrium. The ESR is defined as ESR = (w s − w d /w d ) with w s and w d being the weight of the swollen and the dried gels, respectively. High fucoidan concentrations affected the swelling capacities of the hydrogel as demonstrated in Figure 3b. Hydrogels containing 500 µg/mL fucoidan had three times lower ESR than hydrogels without fucoidan. Incorporation of high-fucoidan concentrations also affected the appearance of the hydrogel as depicted in Figure 3d. Hydrogels containing 500 µg/mL fucoidan were visibly more turbid. Absorption measurements revealed that hydrogels with 500 µg/mL fucoidan were four times more turbid than hydrogels with 0, 10, or 100 µg/mL fucoidan (Figure 3c).
The internal microstructure of the hydrogels was studied by scanning electron microscopy. Therefore, samples were dried using critical point drying beforehand. The incorporation of fucoidan caused no visible changes in the morphology of the hydrogels. Figure 4 shows the morphology of a hydrogel without and with 100 µg/mL fucoidan. All hydrogels consisted of fluffy spheres that were entangled by thin long filaments. The distribution of particles and filaments seemed to occur randomly. The particles tended to merge together into larger aggregates causing the formation of irregular pores. The fluffy spheres and thin filaments likely represent the chitosan and collagen portion, respectively. µg/mL fucoidan had three times lower ESR than hydrogels without fucoidan. Incorporation of high-fucoidan concentrations also affected the appearance of the hydrogel as depicted in Figure 3d. Hydrogels containing 500 µ g/mL fucoidan were visibly more turbid. Absorption measurements revealed that hydrogels with 500 µ g/mL fucoidan were four times more turbid than hydrogels with 0, 10, or 100 µ g/mL fucoidan (Figure 3c). Figure 3. pH, swelling, and turbidity of the injectable chitosan-collagen hydrogel containing different fucoidan concentrations. (a) The pH was measured before and after gelation. (b) The swelling capacity of the hydrogels was estimated by determination of the equilibrium swelling ratio (ESR). Hydrogels were lyophilized and rehydrated until reaching equilibrium. ESR = (ws − wd)/wd); ws, wd = weight of swollen and dried hydrogel, respectively. (c) The turbidity of the sols was quantified by measuring the absorbance at 600 nm. (d) Sols containing 500 µ g/mL fucoidan are visibly more turbid. Scale bar = 1 cm. Experiments were repeated three times (n = 3). Significances compared to the control (hydrogel with 0 µ g/mL Fucoidan) were calculated with the unpaired t-test (* p < 0.05, *** p < 0.001, **** p < 0.0001).
The internal microstructure of the hydrogels was studied by scanning electron microscopy. Therefore, samples were dried using critical point drying beforehand. The incorporation of fucoidan caused no visible changes in the morphology of the hydrogels. Figure 4 shows the morphology of a hydrogel without and with 100 µg/mL fucoidan. All hydrogels consisted of fluffy spheres that were entangled by thin long filaments. The distribution of particles and filaments seemed to occur randomly. The particles tended to merge together into larger aggregates causing the formation of irregular pores. The fluffy spheres and thin filaments likely represent the chitosan and collagen portion, respectively.  Hydrogels were lyophilized and rehydrated until reaching equilibrium. ESR = (w s − w d )/w d ); w s , w d = weight of swollen and dried hydrogel, respectively. (c) The turbidity of the sols was quantified by measuring the absorbance at 600 nm. (d) Sols containing 500 µg/mL fucoidan are visibly more turbid. Scale bar = 1 cm. Experiments were repeated three times (n = 3). Significances compared to the control (hydrogel with 0 µg/mL Fucoidan) were calculated with the unpaired t-test (* p < 0.05, *** p < 0.001, **** p < 0.0001). Figure 3. pH, swelling, and turbidity of the injectable chitosan-collagen hydrogel contain ent fucoidan concentrations. (a) The pH was measured before and after gelation. (b) Th capacity of the hydrogels was estimated by determination of the equilibrium swelling ra Hydrogels were lyophilized and rehydrated until reaching equilibrium. ESR = (ws − wd)/w = weight of swollen and dried hydrogel, respectively. (c) The turbidity of the sols was qua measuring the absorbance at 600 nm. (d) Sols containing 500 µ g/mL fucoidan are visibly bid. Scale bar = 1 cm. Experiments were repeated three times (n = 3). Significances compa control (hydrogel with 0 µ g/mL Fucoidan) were calculated with the unpaired t-test (* p < < 0.001, **** p < 0.0001).
The internal microstructure of the hydrogels was studied by scanning elec croscopy. Therefore, samples were dried using critical point drying beforehand corporation of fucoidan caused no visible changes in the morphology of the hy Figure 4 shows the morphology of a hydrogel without and with 100 µg/mL fuco hydrogels consisted of fluffy spheres that were entangled by thin long filaments. tribution of particles and filaments seemed to occur randomly. The particles t merge together into larger aggregates causing the formation of irregular pores. T spheres and thin filaments likely represent the chitosan and collagen portion, resp  In difference to many proteins, no simple and established procedure exists to quantify fucoidans, especially inside a hydrogel. Consequently, it is also challenging to determine the fucoidan release from a hydrogel over time. To prove the incorporation of fucoidan into the hydrogel, a protocol published by Yamazaki and colleagues [34] was adapted. Yamazaki and colleagues demonstrated that the nucleic acid stain SYBR Gold can be used to quantify the concentration of fucoidan solutions. Similar to DNA, the fluorescent dye binds to the fucoidan molecules. Subsequently, the amount of bound dye can be quantified by measuring the fluorescence. The detected fluorescence has a linear relation to the amount of fucoidan for a specific range of concentrations and can therefore be interpolated using a standard curve with known fucoidan concentrations. We were able to adapt the method to detect fucoidan inside the hydrogel. Even though, the determination of absolute fucoidan concentrations inside the hydrogel using a standard curve was not possible, Figure 5a clearly shows that hydrogels with more incorporated fucoidan resulted in an increased Mar. Drugs 2022, 20, 402 6 of 15 fluorescent signal. The same method was applied to study the amount of fucoidan inside the hydrogel over a time period of six days. The results are shown in Figure 5b. On day two, approximately 40% of the initially detected fucoidan were still incorporated inside the gel. On day six, only 20% of the initially measured fucoidan could be detected. These results indicate that fucoidan was released from the hydrogel over time.
sulted in an increased fluorescent signal. The same method amount of fucoidan inside the hydrogel over a time period o shown in Figure 5b. On day two, approximately 40% of the i were still incorporated inside the gel. On day six, only 20% of coidan could be detected. These results indicate that fucoidan drogel over time. Figure 5. Fucoidan detection inside the hydrogel and fucoidan release tocol from Yamazaki and colleagues [34] was adapted to detect fucoid SYBR Gold nucleic acid stain allowed detecting different fucoidan amo Hydrogels with and without fucoidan were incubated in PBS for 6 days the hydrogel was detected on day 0, 2, and 6 using SYBR Gold. Expe times (n = 3). Significances compared to the control (hydrogel with 0 µ lated with the unpaired t-test (* p < 0.05, ** p < 0.01).

Biocompatibility of the Fucoidan Delivery System with Primary Human Endothelial Cells
To assess the biological compatibility of the chitosan-coll different concentrations of fucoidan, human bone-derived MS cells were seeded on top of the hydrogels (2D culture) or incorp (3D culture) and observed using life/dead staining. The interac drogel was additionally studied using scanning electron microc Life/dead staining of 2D-cultured MSC on day two and hered to the hydrogels as shown in Figure 6a. Incorporated fuc on the cell adhesion and viability. On day six, MSC spreading an Figure 5. Fucoidan detection inside the hydrogel and fucoidan release from the hydrogel. The protocol from Yamazaki and colleagues [34] was adapted to detect fucoidan inside a hydrogel. (a) The SYBR Gold nucleic acid stain allowed detecting different fucoidan amounts inside the hydrogel. (b) Hydrogels with and without fucoidan were incubated in PBS for 6 days. The fucoidan content inside the hydrogel was detected on day 0, 2, and 6 using SYBR Gold. Experiments were repeated three times (n = 3). Significances compared to the control (hydrogel with 0 µg/mL Fucoidan) were calculated with the unpaired t-test (* p < 0.05, ** p < 0.01).

Biocompatibility of the Fucoidan Delivery System with Primary Bone-Derived MSC and Human Endothelial Cells
To assess the biological compatibility of the chitosan-collagen hydrogel containing different concentrations of fucoidan, human bone-derived MSC and human endothelial cells were seeded on top of the hydrogels (2D culture) or incorporated into the hydrogels (3D culture) and observed using life/dead staining. The interaction of MSC with the hydrogel was additionally studied using scanning electron microcopy.
Life/dead staining of 2D-cultured MSC on day two and six revealed that cells adhered to the hydrogels as shown in Figure 6a. Incorporated fucoidan had no visible effect on the cell adhesion and viability. On day six, MSC spreading and adhesion was increased compared to day two. In a next step, the viability of MSC encapsulated into the hydrogel was studied on day two and six. As shown in Figure 6b, MSC reacted more sensitive to an encapsulation into the hydrogel than to the 2D culture. The red stained cells on day two and six indicate that dead cells remained trapped inside the hydrogel. Vital MSC adhered and elongated well inside the hydrogels. Similarly to the 2D culture, cell adhesion and spreading of the vital cells were increased from day two until day six. Again, fucoidan incorporation had no visible effect on the biocompatibility of the hydrogels. Scanning electron microscopy was applied to study the interaction of MSC with the hydrogel on a microscopic scale. As visualized in Figure 6c, MSC were completely embedded into the hydrogel material. Cellular interaction dominantly occurred with the thin filaments which most likely represent the collagen portion of the hydrogel.
2D-cultured OEC were vital and adhered to the hydrogels as shown in Figure 7a. The amount of adherent cells on day two however was larger than on day six, indicating OEC detachment over time. Similar to the culture of MSC, integration of fucoidan into the hydrogels did not change the biocompatibility. OEC were as vital on hydrogels without fucoidan as on hydrogels which contained fucoidan. As visualized in Figure 7b, 3Dcultured OEC were mostly characterized by a round and small shape which did not change from day two until day six. Different to the MSC, OEC were not able to develop an adherent and elongated phenotype inside the hydrogels. Integration of fucoidan into the hydrogels had no effect on the encapsulated OEC.
Drugs 2022, 20, x FOR PEER REVIEW 7 of microscopic scale. As visualized in Figure 6c, MSC were completely embedded into th hydrogel material. Cellular interaction dominantly occurred with the thin filaments whic most likely represent the collagen portion of the hydrogel. 2D-cultured OEC were vital and adhered to the hydrogels as shown in Figure 7a. Th amount of adherent cells on day two however was larger than on day six, indicating OE detachment over time. Similar to the culture of MSC, integration of fucoidan into the h drogels did not change the biocompatibility. OEC were as vital on hydrogels without fu coidan as on hydrogels which contained fucoidan. As visualized in Figure 7b, 3D-culture OEC were mostly characterized by a round and small shape which did not change fro day two until day six. Different to the MSC, OEC were not able to develop an adhere

Discussion
The multifaceted biological activities of fucoidan on, for example, angiogenesis [35,36] and osteogenesis [11,37] are reported in many studies. To benefit from fucoidan's bioactivities, a delivery system is needed which concentrates and immobilizes the molecule inside the tissue of interest. In this context, we developed a hydrogel-based delivery system for fucoidan and studied its suitability for use in regenerative medicine. The composite hydrogel consisted of chitosan and collagen type I; β-GP was added into the chitosan--collagen mixture to achieve thermosensitivity [23]. The effect of fucoidan incorporation on different physicochemical characteristics such as gelation time, gelation temperature, pH, appearance, swelling, and internal microscopic structure were studied. To evaluate the biocompatibility of the hydrogels, human bone-derived MSC and OEC isolated from peripheral blood were cultured on top and inside the hydrogels for six days. Fluorescence microscopy and SEM were applied to study the viability of the cells and their interaction with the hydrogel material.
Preliminary experiments have demonstrated that increased β-GP concentrations reduced the time until gelation. Accordingly, many studies have shown that β-GP has an impact on the gelation time [38][39][40]. It is however also reported that very high β-GP contents can be toxic to cells. Yan and colleagues demonstrated an impaired viability of rat BMSC on hydrogels containing 80% β-GP [41]. Ahmadi et al. found that hydrogels containing >10% β-GP resulted in increased cytotoxicity for goat MSC [38]. Our results showed that an amount of 7% β-GP resulted in an adequate gelation time and was not cytotoxic to our model cells.
Rheological measurements revealed that hydrogels containing 100 µ g/mL fucoidan gelated after 1 min at 37 °C. This was comparable to hydrogels without fucoidan which needed around 75 s to reach the gel point. This gelation time seems appropriate for medical applications; it is long enough to allow the application via injection and to allow the

Discussion
The multifaceted biological activities of fucoidan on, for example, angiogenesis [35,36] and osteogenesis [11,37] are reported in many studies. To benefit from fucoidan's bioactivities, a delivery system is needed which concentrates and immobilizes the molecule inside the tissue of interest. In this context, we developed a hydrogel-based delivery system for fucoidan and studied its suitability for use in regenerative medicine. The composite hydrogel consisted of chitosan and collagen type I; β-GP was added into the chitosancollagen mixture to achieve thermosensitivity [23]. The effect of fucoidan incorporation on different physicochemical characteristics such as gelation time, gelation temperature, pH, appearance, swelling, and internal microscopic structure were studied. To evaluate the biocompatibility of the hydrogels, human bone-derived MSC and OEC isolated from peripheral blood were cultured on top and inside the hydrogels for six days. Fluorescence microscopy and SEM were applied to study the viability of the cells and their interaction with the hydrogel material.
Preliminary experiments have demonstrated that increased β-GP concentrations reduced the time until gelation. Accordingly, many studies have shown that β-GP has an impact on the gelation time [38][39][40]. It is however also reported that very high β-GP contents can be toxic to cells. Yan and colleagues demonstrated an impaired viability of rat BMSC on hydrogels containing 80% β-GP [41]. Ahmadi et al. found that hydrogels containing >10% β-GP resulted in increased cytotoxicity for goat MSC [38]. Our results showed that an amount of 7% β-GP resulted in an adequate gelation time and was not cytotoxic to our model cells.
Rheological measurements revealed that hydrogels containing 100 µg/mL fucoidan gelated after 1 min at 37 • C. This was comparable to hydrogels without fucoidan which needed around 75 s to reach the gel point. This gelation time seems appropriate for medical applications; it is long enough to allow the application via injection and to allow the gel adapting to the defect shape. On the other hand, it is short enough to prevent long waiting times and to avoid the immediate diffusion of integrated fucoidan.
Temperature sweep tests demonstrated that the gelation temperature for hydrogels containing 100 µg/mL fucoidan was 35.3 • C. The gel point of hydrogels without fucoidan occurred at 36.8 • C. Hence, hydrogels with fucoidan tentatively started to gelate at slightly lower temperatures than gels without fucoidan. The gelation of the hydrogels at 37 • C or slightly lower temperatures is advantageous for a medical application. The cooled sol can be applied at room temperature, but it will quickly start to gelate once exposed to the physiological body temperature. Several studies have shown that the gelation temperature is determined by a combination of different parameters, such as β-GP content, chitosan concentration, and molecular weight, as well as degree of deacetylation [39,42,43]. Increased β-GP content and increased chitosan concentrations result in a decreased gelation temperature [39,44]. Further, higher molecular weight and degree of deacetylation of chitosan decrease the gelation temperature as well [42].
Even though, aqueous fucoidan solutions exhibit low viscosity [43], fucoidan alone is not able to form hydrogels. Due to its sulfate residues, fucoidan usually has an overall negative charge. Together with oppositely charged polysaccharides such as chitosan, hydrogels or films can be formed based on electrostatic interactions [43,45]. The results from the rheological measurements, swelling experiments, and the visual appearance (turbidity) indicate that the integration of fucoidan into the chitosan hydrogel system increased the degree of interconnectivity inside the gel. Swelling experiments revealed that hydrogels containing high amounts (500 µg/mL) of fucoidan absorbed less water than hydrogels with no or less fucoidan. We assume that the additional molecular interactions resulted in faster gelation at lower temperatures and that the increased matrix interconnectivity hindered water uptake into the gel, therefore reducing the swelling capacities of hydrogels containing 500 µg/mL fucoidan.
To our knowledge, no study exists which developed and characterized thermosensitive chitosan-collagen hydrogels with integrated fucoidan. To date, most studies investigated the potential of fucoidan as a component of nanoparticles for the delivery of specific drugs. The putative applications are manifold: anti-cancer, anti-bacterial, to treat pulmonary diseases or diabetes [46][47][48][49]. However, some studies developed hydrogel systems containing fucoidan. Carvalho and colleagues prepared ionically-crosslinked hydrogels with collagen, chitosan and fucoidan and found that gels containing all three polysaccharides possessed better mechanical properties than hydrogels with only two polysaccharides [45]. Sezer and colleagues produced chitosan-fucoidan hydrogels for the treatment of dermal burns [50]. In contrast to our results, Sezer et al. found that hydrogels with fucoidan absorbed more water, indicating greater swelling capacities. This was associated with the hydrophilic nature of fucoidan. Another explanation for the different result in the experiments of Sezer and colleagues is the absence of collagen and β-GP which are additional crosslinking molecules. Other studies describe the preparation of chitosan-fucoidan hydrogels for the encapsulation of angiogenic growth factors [51] or the development of alginate-fucoidan composite hydrogels for enhancing chondrogenesis of stem cells [52]. Additionally, some approaches are described which utilize covalent cross-linking to create hydrogels containing fucoidan. Lu and colleagues prepared genipin crosslinked hyaluronic acid-fucoidan-gelatin hydrogels for the delivery of platelet-rich plasma [53]. A study from Hsu and colleagues presented the production of methacrylated hyaluronan and fucoidan to create photo-crosslinkable hydrogels [6].
In almost all biomaterial approaches which include fucoidan, the polysaccharide rather represents a structural element with beneficial properties, such as enhancing the bioactivity of the encapsulated drug or altering the mechanical properties of the hydrogel [45,51]. To our knowledge, no study exists which integrates fucoidan into a delivery system and examines its release comparable to a drug. Our experiments indicated that approximately 60% of the fucoidan was released from the hydrogel within two days, and 80% of the fucoidan was released after six days. These results suggest that fucoidan was not permanently immobilized inside the hydrogel. Its interactions with the other components were reversible, allowing the release of fucoidan into the supernatant over a specific period of time.
To use a biomaterial in the medical context, biocompatibility is a crucial requirement. It is defined as "the ability of a material to perform with an appropriate host response in a specific application" [54]. Interaction of human cells with the material is especially important for applications dealing with tissue regeneration. Here, the biomaterial is continuously in direct contact with the injured tissue, providing a platform for adjacent cells to migrate into the defect and initiate its regeneration [13]. Life/dead stainings and SEM revealed that human MSC were vital when cultured inside (3D) or on top (2D) of the injectable chitosan-collagen hydrogels with integrated fucoidan. Further, we showed by fluorescence microscopy that MSC adhered and had an elongated phenotype, indicating a good biocompatibility of the material with MSC. Life/dead stainings showed that 2Dcultured OEC were vital and adhered to the material. However, encapsulated OEC had a round and small phenotype indicating difficulties of OEC to spread well inside the material. These results demonstrate that MSC can be encapsulated along with fucoidan into the hydrogel to support healing processes. However, to encapsulate OEC alone, structural adjustments of the hydrogel are needed. In addition, a co-encpasulation of OEC with another cell type such as MSC might help to improve the biocompatibility.

Conclusions
To conclude, we were able to develop an injectable hydrogel-based delivery system for fucoidan. The hydrogel was injectable at room temperature and gelated within 1 min at 37 • C. The incorporation of fucoidan into the hydrogel had only minor impacts on the determined physicochemical properties. Only the incorporation of high fucoidan concentrations (500 µg/mL) increased the turbidity and decreased the swelling capacity of the hydrogel. Further, the biomaterial was compatible with human MSC and OEC with a limitation for OEC encapsulation. Due to its short gelation time, physiological gelation temperature, and biocompatibility, the hydrogel represents a promising delivery system for fucoidan for the use in regenerative medicine.

Ethical Approval
The use of human material was approved by the local ethics committee of the University Medical Center Schleswig-Holstein. Isolation of primary cells from human tissue was performed with the consent of the donors.

Hydrogel Preparation
All steps were performed on ice and under constant stirring. Two percent (in 0.1 M acetic acid) 4 • C cooled chitosan (Chitosan 95/100, Heppe Medical Chitosan GmbH, Halle, Germany) solution was pipetted into a glass vial, followed by addition of rat tail collagen type I (Corning, Bedford, MA, USA) to reach a final concentration of 1.5 mg/mL. Cold 50% β-glycerophosphate (Sigma-Aldrich, St. Louis, MO, USA, end concentration 7%) was mixed with aqueous 5 mg/mL fucoidan solution (≥95% from Fucus vesiculosus, Sigma-Aldrich, end concentrations 0-500 µg/mL) and immediately added dropwise into the chitosan-collagen mixture. The mixture was incubated on ice under constant stirring for 15 min to ensure proper mixing of all components. The sol was always used on the same day for experiments.

Gelation Time by Tube-Inverting Method
The gelation time was roughly estimated by using the tube-inverting method. The sol was incubated in a water bath at 37 • C and inverted every minute. The sol was defined as gelated when no flow could be observed any longer.

Rheological Measurements
The rheological studies were performed with the CVO 120 HRNF (Bohlin Instruments, Pforzheim, Germany) plate-to-plate rheometer (plate diameter: 40 mm). One mL of sample covered with a solvent trap to prevent evaporation was used for each measurement. The storage (G ) and loss moduli (G") were determined to characterize the viscoelastic properties of the hydrogels with and without fucoidan. At the beginning, an amplitude sweep test was carried out at 37 • C and at a constant frequency of 1 Hz to determine the viscoelastic region of the hydrogels. The start of the gelation process and its duration were assessed by time sweep tests at 37 • C, 1 Hz, and 5% strain. To determine the exact gelation temperature of the hydrogels, temperature sweep tests were performed. Therefore, G and G" were measured at 10-60 • C (2 • C/min) at a constant frequency of 1 Hz and 5% strain.

pH
The pH of the sols was measured before and after gelation using a SevenEasy pH meter S20 with an InLab Micro Pro-ISM pH electrode (Mettler-Toledo GmbH, Gießen, Germany).

Swelling
To quantify the swelling capacity of the hydrogels, the equilibrium swelling ratio (ESR) was determined. Therefore, hydrogels were prepared and quick-frozen in liquid nitrogen, followed by lyophilization (ScanVac CoolSafe Freeze Dryer, LaboGene, Allerød, Denmark). Dried hydrogels were weighted and incubated in PBS at 37 • C for three days. After three days, the weight of the swollen hydrogels was determined. The ESR was calculated using the following equation: ESR = ((w s − w d )/w d ), with w s and w d = weight of swollen and dried hydrogel, respectively.

Turbidity
The turbidity of the sol was quantified by measuring the absorbance at 600 nm using a plate reader (infinite M200 Pro, Tecan, Männedorf, Switzerland).

SEM
Hydrogels were prepared in 8-well slides and incubated in PBS for 24 h. On the next day, hydrogels were fixed for 30 min at room temperature using 3% glutaraldehyde (Sigma-Aldrich) in 4% paraformaldehyde (Morphisto, Frankfurt am Main, Germany). After 3 times washing for 5 min with PBS, aqueous buffer was replaced gradually by ethanol (Merck, Darmstadt, Germany). Each ethanol dilution (50%, 60%, 70%, 80%, 90%, 95%) was incubated on the hydrogels for 15 min until reaching 100% ethanol. Subsequently, hydrogels were carefully removed from the wells and cut in half for imaging the gel morphology at the cross-section or left intact for imaging the surface of the gels. Hydrogels were dried using a critical-point dryer (Leica EM CPD300, Nussloch, Germany) and glued to sample holders using conductive carbon tape. Subsequently, samples were sputtered with gold for 90 s and imaged with a Supra 55VP (Zeiss, Jena, Germany).

Fucoidan Detection Inside the Hydrogel
Fucoidan incorporation into the hydrogel was proven using the SYBR Gold Nucleic Acid Gel Stain (Invitrogen, Eugene, Oregon, USA) using an adapted protocol from Yamazaki and colleagues [34]. One-hundred µL sol was pipetted into a 96-well plate and gelated at 37 • C for 30 min. SYBR Gold stain solution was diluted 1:625 in 80 mM Tris-HCl (pH 7.5) and 50 µL was applied onto the hydrogels. After 1 h incubation at room temperature, the staining solution was removed and hydrogels were washed with PBS. The fluorescence was measured with a plate reader at 470 nm and 600 nm excitation and emission wavelength, respectively.
For the fucoidan release assay, 100 µL sol was pipetted into a 96-well plate and gelated at 37 • C. Sols were incubated at 37 • C in PBS. On days 0, 2, and 6, fucoidan was detected using SYBR Gold Nucleic Acid Gel Stain as described above.

Human Outgrowth Endothelial Cells
Human outgrowth endothelial cells (OEC) were isolated from peripheral blood as described before [55]. OEC were cultured in fibronectin-coated flasks using endothelial growth medium-2 (EGM-2), containing endothelial basal medium (EBM-2), EGM-2-associated supplements (Promocell, Heidelberg, Germany), 5% FBS, and 1% Penicillin/Streptomycin. OEC were sub-cultured every two or three days after reaching confluency. The medium was exchanged every second or third day. 5.6. Cell Culture Inside and on Hydrogels 5.6.1. 2D Cell Culture Sols were pipetted into the wells of an adequate dish and gelated at 37 • C for 30 min. 80,000 cells/cm 2 were seeded on top of the hydrogel after gelation. The medium was gently exchanged each or every second day.

3D Cell Culture
For cell encapsulation into the hydrogel, the cell suspension was centrifuged at 400× g for 5 min and the pellet was resuspended in cold sol in a concentration of 10 6 cells/mL by pipetting up and down. Sols including cells were pipetted into wells and gelated for 30 min at 37 • C. Subsequently, warm culture medium was added on top of the hydrogels. The medium was gently exchanged each or every second day.

Cell Life/Dead Staining
For life/dead stainings, cells were cultured 2D or 3D in µ-Slides Angiogenesis (Ibidi, Gräfelfind, Germany). On day two or six, the culture medium was gently removed and replaced by fresh culture medium containing 10 µg/mL CalceinAM (Sigma-Aldrich) and 2 µg/mL propidium iodide (Sigma-Aldrich) for staining vital and dead cells, respectively. The staining solution was incubated for 30 min at 37 • C. Subsequently, the staining solution was removed and replaced by fresh medium. Cells were imaged with the Evos FL Auto 2 fluorescence microscope (Thermo Fisher Scientific, Waltham, MA, USA).

Statistical Analysis
Unpaired t-test was used to calculate the statistical significances in Prism 7.03. (Graph-Pad, San Diego, CA, USA) Values were considered as significant when p < 0.05.