Conduits based on the combination of hyaluronic acid and silk fibroin: characterization, in vitro studies and in vivo biocompatibility

We address the production of structures intended as conduits made from natural biopolymers, capable of promoting the regeneration of axonal tracts. We combine hyaluronic acid (HA) and silk fibroin (SF) with the aim of improving mechanical and biological properties of HA. The results show that SF can be efficiently incorporated into the production process, obtaining conduits with tubular structure with a matrix of HA-SF blend. HA-SF has better mechanical properties than sole HA, which is a very soft hydrogel, facilitating manipulation. Culture of rat Schwann cells shows that cell adhesion and proliferation are higher than in pure HA, maybe due to the binding motifs contributed by the SF protein. This increased proliferation accelerates the formation of a tight cell layer, which covers the inner channel surface of the HA-SF tubes. Biocompatibility of the scaffolds was studied in immunocompetent mice. Both HA and HA-SF scaffolds were accepted by the host with no residual immune response at 8 weeks. New collagen extracellular matrix and new blood vessels were visible and they were present earlier when SF was present. The results show that incorporation of SF enhances the mechanical properties of the materials and results in promising biocompatible conduits for tubulization strategies.


Introduction
Injuries and diseases affecting the nervous system have very high incidence and prevalence, affecting millions of people around the world. New concepts and solutions are in demand for both the central nervous system (CNS) and the peripheral nervous system (PNS). In the human CNS axons do not regenerate spontaneously after damage or degeneration, causing life-long functional deficit as in the case of spinal cord injury, traumatic brain damage or cerebral infarction [1]. Failure of axonal regeneration is probably due, among other factors, to the formation of an impenetrable glial scar, composed mainly of astrocytes, that inhibits axonal growth and myelination [2]. This causes regenerative neurons to be blocked and unable to reach their synaptic target. In the PNS there is successful endogenous axon regeneration in adult humans for distances of few centimetres [1]. Here damage is followed by a degeneration of the distal segment of the nerve, known as Wallerian degeneration, while the proximal segment of the nerve suffers only minimal damage [3,4]. After the distal segment degeneration, Schwann cells and phagocytic cells such as macrophages remove myelin and axonal debris while producing cytokines that improve the axonal growth, creating a growth cone able to direct axon regrowth until reconnection with the distal target. Finally, Schwann cells myelinate the re-growing axon forming new myelin sheaths [4,5]. Nonetheless, in cases of larger damages the process of regeneration is not successful.
In order to improve over the limitations of current therapies tissue engineering approaches invoke the combination of cell supply, bioactive molecules and scaffolds based on biomaterials [6,7]. conduits typically consist in hollow tubes used to connect the two ends of a damaged nerve, guiding the axonal regeneration between both sides of the injury [8]. Their main goal is to protect the spontaneous regeneration process.
Conduits must guide tissue regeneration from one end to the other of the lesion, avoiding the growth of fibrous tissue and retaining the neurotrophic factors secreted by the damaged nerve remnants [9]. Tubular scaffolds should have a porosity large enough to allow the exchange of oxygen and nutrients, and small enough to prevent the infiltration of inflammatory cells into the conduit and to minimize the outflow of cells and growth factors out of the structure; furthermore, they should possess adequate mechanical properties to allow its manipulation during surgery and be flexible and soft enough not to compress the regenerated axons [6,9].
Here we present conduits made from hyaluronic acid (HA), a natural biomaterial that is widely present in the extracellular matrix of most tissues [10]. HA is a non-branched glycosaminoglycan formed by repetitions of D-glucuronic acid and D-Nacetylglucosamine [10]. HA has an excellent biocompatibility and unique biological properties, as it regulates the immune and inflammatory response, the cell differentiation, the vascularization, the scar tissue formation and the cell adhesion to extracellular matrix (ECM) proteins [10,11]. In addition, HA modulates the behaviour of glial and immune cells [10]. A very important characteristic of HA is its high hydrophilicity, due to hydrogen bonding interactions between its chains and water [12,13]. Scaffolds made out of HA have been studied for the regeneration of the CNS, showing their biocompatibility and ability to induce neural regeneration [14][15][16][17][18].
Tubular scaffolds made out of HA are able to induce seeded Schwann cells to form highly organized structures in vitro [19,20]. The HA tubes acted as a template in organizing the growth of the Schwann cells into cylindrical cell sheets. To improve the properties of such biohybrids (structures composed by NGCs with cells) and their biological response here we incorporate Bombyx mori silk fibroin (SF) in the production process of the scaffolds. SF is the term conventionally used for referring to the heavy chain protein which is the main constituent of the silk fibers of the cocoons [21] . SF is a protein with a molecular weight in excess of 300 kDa [22] that presents the specific motif of sequence -GAGAGS-. Due to its biological function, SF has robust mechanical properties and, in addition, a notable biocompatibility [23] and remarkable cell adhesion. This latter property might be favoured by the presence of arginine residues in the sequence of SF. The arginine residues might be recognized by cell membrane adhesion proteins such as integrins and would contribute to the good biological response to the material [22][23][24][25]. Besides its excellent in vitro and in vivo biocompatibility SF also shows a slow proteolytic biodegradation [23,[26][27][28].
The present work aimed at combining SF and HA in a way such that the celltemplating ability of the HA tubes was preserved, while their mechanical properties and biological response were improved. Our interest thus was in combining both materials but with a majority fraction of HA. It may be noted that both HA and SF are materials currently approved by the United States Food and Drug Administration (FDA) for different uses in biomedical applications in humans, particularly in guidance of peripheral nerve regeneration. [45][46][47][48].

SF preparation
Fibroin was extracted from Bombyx mori cocoons. Initially, the cocoons were degummed in an aqueous Na 2 CO 3 solution (0.2 w/v) and heated to 121 °C in an autoclave (Sturdy SA-252F) for 50 minutes. The degumming process solubilizes the sericin coating that covers the fibroin fibers, so that it can be removed. Fibroin was washed with distilled water for 20 minutes and the washing step was repeated three times. The fibers were allowed to dry overnight and then immersed in a 9.3 M LiBr solution and introduced in a furnace at 60 °C for 4 hours, so that a fibroin solution was obtained. After being dissolved, the silk solution was dialyzed against deionized water to remove the LiBr salt. Dialysis membranes (BioDesign Dialysis Tubing) with a MWCO of 3500 Da were used. Water was changed every 8 hours and the whole process proceeded for 48 hours. After dialysis the fibroin solution was centrifuged at 5000 rpm for 20 minutes at 21 °C to remove any debris. Finally, the fibroin was lyophilized and stored in a freezer at -20 °C.  The solution   178  179  180  181  182  183  184  185  186  187  188  189  190  191  192  193  194  195  196  197  198  199  200  201  202  203  204  205  206  207  208  209  210  211  212  213  214  215  216  217  218  219  220  221  222  223  224  225  226  227  228  229  230  231  232  233  234  235  236   was then homogenized by agitation at NCTP for 20 minutes and mixed with the divinyl   sulfone cross-linker (DVS; V3700, Sigma-Aldrich) in a 9:10 DVS:HA monomeric units molar ratio. This mixture (hereinafter referred to as HA) was kept under stirring for 10 seconds to achieve the correct mixing and to begin the crosslinking process.

HA and HA-SF films preparation
The elaboration of biomaterials based on the combination of HA and SF is analogous to that of HA, with the exception that, after the elaboration of the HA solution, it was mixed with SF in a concentration of 3 % w/v (hereinafter referred to as HA-SF). The

HA and HA-SF conduits preparation
To obtain the consuits, the biomaterials obtained as described in 2.1.2 were injected by pipetting into a PTFE mold formed by elongated channels with a square profile of 2 mm per side. A polycaprolactone fiber (PCL; Polysciences) with a diameter between 950 and 1050 μm was introduced into the center of each channel to form the lumen of the conduit. At both ends of the PCL fiber, a PTFE sheath was inserted with a thickness slightly greater than 2 mm. By pressing the ends of the fiber it was fixed in the center of the channel at the same approximate distance from all the walls. After the injection of the material into the mold, it was left for 10 minutes in NCTP to achieve the complete crosslinking of the material. Subsequently, it was subjected to freezing and lyophilization with the same parameters used for films. Finally, the PCL fibers were removed from the inside, forming the lumen of the conduit.
To study the effect of the dissolution of SF in a basic medium, unprocessed SF (hereinafter referred to as untreated SF) was compared with a 3 % w/v SF solution in deionized water (hereinafter referred to as water-treated SF) and with a 3 % w/v SF solution in NaOH 0.2 M (hereinafter referred to as NaOH-treated SF). The solutions were homogenized by stirring in NCTP for 20 minutes and then they were spilled into a PTFE plate to give them a film shape. They were then frozen and lyophilized with the same parameters as the other films.

Morphology characterization
For the characterization of the surface morphology and of the geometry of axial and cross-sectional cuts of conduits, a field emission scanning electron microscope (FESEM; ULTRA 55, ZEISS Oxford Instruments) was used. The preparation of the samples consisted primarily in a desiccation under vacuum conditions during the 24 hours prior to the test to avoid interferences due to evaporated water. Subsequently samples were placed on a carbon tape and a carbon bridge was created between the sample and the carbon tape. Finally, samples were coated with a thin layer of platinum.
Conduits were cut longitudinally and transversally to observe the interior of the lumen and the morphology of the porosity of the walls. The voltage used was 2 kV. The study was performed with three different samples (n = 3) made of HA and HA-SF.

Elemental analysis
The FESEM was also used to carry out an elemental analysis through an energy dispersive X-ray spectroscopy (EDS) to verify the presence of SF, by studying the nitrogen (N) content. Samples were prepared as above. The elemental analysis provided the mass fraction (ω) of N, oxygen (O) and sulfur (S) in each selected area of the material. The voltage used was 2 kV. The study was performed with three different samples (n = 3) of films made of HA and HA-SF.

Stability against swelling and drying cycles
This study was carried out to evaluate the possible loss of material in cycles of swelling and drying. For this purpose, both the amount of non-crosslinked mass and the water uptake were studied. Disc-shaped films of 8 mm dry diameter ( ) of HA and D 0 HA-SF were used. Samples were first vacuum desiccated for 6 hours at 50 °C. They were then weighed ( ) using a balance (AX205, Mettler-Toledo Inc., sensitivity of m 0 0.01 mg). After that, discs were swollen in deionized water for 24 hours to attain the maximum degree of swelling and the diameter was measured ( ). The process was D 1 then carried out a second time: samples were first air-dried for 6 hours and then dried in a desiccator for 6 hours with continuous vacuum and a temperature of 50 °C; they were then weighed ( ). Finally, discs were swollen in deionized water for 24 hours and the m diameter was measured ( ). To study the loss of mass, the parameter was D 2 m/m 0 calculated, while the water uptake was characterized through the parameter for D i /D 0 the first (i=1) and the second (i=2) swelling. The swelling ratio (Q) was also calculated as for the first (i=1) and the second (i=2) swelling. The procedure was ( / 0 ) 3 repeated for 4 different discs (n = 4) of each material.

Dimensional change in swelling
To study the swelling characteristics , conduits dimensions were measured before and after the absorption of water. For this purpose, conduits made of HA and HA-SF were prepared with two different lengths: samples of approximately 4 mm in length to study the change in length and samples of approximately 1 mm in length to study changes in the lumen diameter. Samples were examined and photographed using a binocular magnifying glass (MZ APO, Leica Microsystems). Once images of dry samples were acquired, they were swollen in deionized water for 24 hours and images in swelling state were acquired. The measurement of the dimensions was performed using the ImageJ/FIJI image processing software [49]. The dimensions (length, lumen diameter) in dry and swollen conditions were calculated for 4 different samples (n=4).
The variation of the dimensions of the conduits was calculated by the ratios L/L0 (length in swelling state (L) divided by length in dry state (L0)) and D/D0 (lumen diameter in swelling state (D) divided by lumen diameter in dry state (D0)).

Density and porosity
Density and porosity of samples were determined gravimetrically, using a balance (AX205, Mettler-Toledo Inc., sensitivity of 0.01 mg) equipped with a density measurement kit. First, the weight of the NGC in air ( ) was measured. Subsequently ,   355  356  357  358  359  360  361  362  363  364  365  366  367  368  369  370  371  372  373  374  375  376  377  378  379  380  381  382  383  384  385  386  387  388  389  390  391  392  393  394  395  396  397  398  399  400  401  402  403  404  405  406  407  408  409  410  411  412  413 the sample was immersed in n-octane (n-octane: 412236, Sigma-Aldrich) and placed under vacuum for 30 minutes to replace the air inside the pores by n-octane. After removing the n-octane from the lumen of the conduit with an absorbent paper, the weight of the conduits was measured with n-octane in the pores ( ). Finally, the NGC with n-octane in the pores was completely submerged in n-octane and the apparent weight of the conduits immersed in n-octane ( ) was measured. Following Archimedes, the density of the sample without considering the air inside the pores was obtained applying equation (1), were is the density of n-octane. The porosity was ρ o then calculated applying equation (2). This procedure was performed for 3 samples (n = 3) of HA and HA-SF.

Thermogravimetric analysis (TGA)
A thermogravimetric analyzer (TGA/SDTA 851 Mettler-Toledo operated using the STARexx software) was used to study the thermal degradation and composition of the materials. To study the effect of the basic manufacture conditions on SF, untreated, water-treated and NaOH-treated samples of SF were measured. Samples were first maintained at a temperature of 30 °C for 2 minutes. Then a heating ramp was applied from 30 °C to 120 °C with a heating rate of 10 °C/min. After that, a preheating stage consisting of maintaining a temperature of 120 °C for 30 minutes was applied to eliminate all the water content of the sample. Finally, a heating ramp was applied from 120 °C to 720 °C with a heating rate of 10 °C/min. The procedure was performed under a positive nitrogen flow ( ) of 20 ml/min and was repeated for three different samples N 2 (n=3) of each of the materials mentioned above. As a result, thermograms in which the mass loss of the sample is represented as a function of temperature were obtained.
These curves were normalized dividing through the weight existing at the end point of the preheating plateau. Assuming that thermal degradation proceeds in composite HA-SF samples as it would for both pure separate components, an estimate of mass loss in composites is given by equation (3), were and are the experimental curves of pure HA ∆m HA ∆m SF and SF, respectively, and is the mass fraction of SF.
This equation was employed to deduce a value of from the measured curves of ω SF . For this purpose, given by equation (3) was least-squares ∆m HA , ∆m SF , ∆m exp ∆m comp fitted to taking as fitting parameter. ∆m exp ω SF

Characterization of mechanical properties
The mechanical characterization of the materials was performed by rotational rheometry (Discovery Hybrid Rheometer DHR, TA Instruments) in order to study the effect of incorporating SF on the shear modulus (G) of the material. The samples (circular films 20 mm in diameter) were studied in a state of swelling in deionized water for 24 hours.
From the images obtained by confocal microscopy, and thanks to staining with FITC-phalloidin, the total area occupied by the cells cytoskeleton was calculated.

Effect of preparation procedures on material composition
The crosslinking reaction of HA with DVS takes place in basic pH [10]. Since an aqueous solution of SF was mixed with the basic HA-DVS solution during the material preparation process, it was necessary to assess the effect of exposure of SF to the basic solution. A thermogravimetric analysis (TGA) of untreated SF, water-treated SF and NaOH-treated SF (Fig. 1A) shows slight modifications of the temperature-dependent mass loss. Prolonged exposure of SF to a basic medium causes hydrolysis of peptide bonds [50]. Therefore, when SF is added to the solution of 5 % HA in 0.2 M NaOH, it is subjected to a basic medium throughout the process and it suffers the hydrolysis of a number of peptide bonds. Chemical alteration of silk by immersion in a NaOH solution accumulates to the reduction in the molecular weight induced by the degumming process. It was found that this reduction is larger when Na 2 CO 3 is added to the degumming medium due to the resulting pH of the solution [51]. TGA data reflects an alteration as a difference in both curves in the range between 500 °C and 700 °C. The different N content of SF and HA is reflected in the EDS data in Table 1.
Theoretically, the mass fractions of N in HA and in HA-SF samples are 2.60 % ( ) and 6.75 %, respectively. As shown in Table 1, the experimental mass fractions of N were lower than the theoretical ones. It should be noted that the elemental analysis is performed on very small areas of material, so a non-homogeneous distribution of SF could produce this difference. In addition, elemental analysis by EDS is sensitive solely to surface composition. Table 1 shows also the ratios between and the mass fractions of O and S. All these parameters were substantially higher in the HA-SF samples than in the HA samples, proving the presence of SF in the matrices. 709  710  711  712  713  714  715  716  717  718  719  720  721  722  723  724  725  726  727  728  729  730  731  732  733  734  735  736  737  738  739  740  741  742  743  744  745  746  747  748  749  750  751  752  753  754  755  756  757  758  759  760  761  762  763  764  765 766 767

Stability against swelling and drying cycles
The loss of mass of HA and HA-SF films after prolonged immersion in water was analysed to determine the stability against solution of the SF incorporated to the HA matrix, and whether the presence of SF could alter the sol fraction of HA in the crosslinked gel. As seen on Table 1, there were no significant differences in the mass loss between HA and HA-SF films after being swollen in water. A similar sol fraction (around 30 %) obtained in both systems.
Dimensional change upon swelling was expressed by computing the ratio of measured diameters of discs after swelling (D) and dry (D 0 ). Though small, the effect of SF is to constrain swelling and dimensional change, as expected (  768  769  770  771  772  773  774  775  776  777  778  779  780  781  782  783  784  785  786  787  788  789  790  791  792  793  794  795  796  797  798  799  800  801  802  803  804  805  806  807  808  809  810  811  812  813  814  815  816  817  818  819  820  821  822  823  824  825  826 The mechanical behaviour of HA and HA-SF discs was assessed in shear experiments, in a rotational rheometer. Fig. 2A shows considerably. The fact that this difference is sensibly higher at lower deformations may suggest that it be related to the diffusion of water molecules in both gels, which is altered in the presence of the protein molecules. As already commented, the stiffness of the HA-SF material, as characterized by G', is greater than that of HA; however, its viscoelasticity, as characterized by the ratio G'/G'', doesn't change very much (see Table   2).

HA
HA-SF

Adhesion and proliferation of Schwann cells on HA-SF films
Rat Schwann cells (rSCs) were seeded and cultured for 1 and 5 days on HA and HA-SF films in order to assess differences in adherence and proliferation. For the visualization of the cells by confocal microscopy, a staining process was performed with FITC-phalloidin 488 to visualize their cytoskeletal actin fibers (green color) and with DAPI to visualize the cell nucleus (blue color). For the day 1, Fig. 3A and 3C show representative images of both studied cases in order to appreciate the differences in the total extent of the cytoskeleton, which is indicative of the cell adhesion characteristics on the material. On HA substrates (Fig. 3A) early cell adhesion was scarce, and the cells had a marked globular morphology, in conformity with known facts [52]. This is a consequence of the highly hydrophilic nature of HA: the enormous amounts of water molecules which surround the HA chains hinder ECM protein deposition on its surface and thus difficult cell adhesion. This means that cell adhesion on the surface of HA materials must rely primarily on physical phenomena such as the surface roughness, where differential accumulation of HA molecules offers more stable anchoring sites to ECM proteins. With the material containing SF (Fig. 3C) early cell adhesion was clearly greater, with the cells exhibiting a much more developed cytoskeleton. Several studies have shown that SF has good cell adhesion properties [25,53]. Although the RGD (Arg-Gly-Asp) cell binding motif as such is not present in the aminoacid sequence of Bombyx mori SF, an amino acid sequence has been observed in a non-repetitive C-terminal region with a large number of basic amino acids, especially arginine, which play a key role in the cell adhesion properties of SF [25].
Arginine residues can be recognized by transmembrane cell adhesion proteins such as integrins, which mediate in the interaction between cell and substrate [54,55].
In order to quantify somehow the extent of cell adhesion, the total area occupied by the cytoskeleton of the cells was measured (Fig. 3E), confirming the conclusions obtained in a qualitative way. It can be seen that early cell adhesion in HA films is very low, while addition of SF clearly enhances early cell adhesion.

Morphology and physicochemical properties of conduits
Longitudinal (axial) and transverse cuts of conduits show the lumina of the conduits and the porous walls, Fig. 4. Both in HA and HA-SF conduits the lumen surface is always tighter and less porous than the walls, which is important to ensure that the cells seeded within the lumen are retained inside the conduit and cannot protrude across its walls. This is due to the fact that the surface of the PCL fiber used as a template of the lumen is hydrophobic; thus, in its vicinity non-swollen HA chains are predominant [19,20]. Water crystallization and subsequent lyophilization then produce a thin layer of denser HA close to the lumen. The pores observed in the conduit's wall matrix are produced by the phase-separation of water during the crystallization process.
This structure ensures good diffusion of molecules across the conduit (water, oxygen, nutrients, factors), while efficiently retains the cells seeded inside the tube. The addition of SF to the HA matrix of the conduits thus didn't modify the main features of the original HA conduits. Fig. 4D' shows a detail of the internal surface of the HA-SF conduit, where elongated structures arranged on the surface (indicated with arrows) could correspond to SF fibers.
The swelling behaviour of HA and HA-SF conduits was characterized by the change in dimensions of conduit length and lumen diameter, Table 3. There is no significant difference in dimensional change between HA and HA-SF samples. Though very small, the effect of SF seems to be to introduce some anisotropy in the conduit's swelling as compared with the HA tubes: the HA-SF tubes swell somewhat less than the HA tubes transversally to their axis, while they swell somewhat more axially. This may have to do with anisotropies of the SF fibril distribution induced during the phase separation of crystallized water, a process which is determined by the geometry of the heat transfer between gel and mold during cooling.  Table 3 shows the density and porosity of conduits made of HA and HA-SF.
Regarding the density values, there's a slight decrease for HA-SF conduits, but it is not enough to show a statistical difference. This slight decrease may be due to the incorporation of SF, since SF has a lower density than HA [56,57]. Regarding the porosity values, the slight differences are not significant. This could be expected, since the pores are produced by the crystallization of the large amounts of water present in the samples, which are insignificantly influenced by the presence of the SF molecules.

HA
HA-SF  Table 3. Physicochemical properties of HA and HA-SF conduits. Density was calculated by equation (1) and porosity by equation (2). Swelling ratios L/L0 (length) and D/D0 (diameter) are explained in the text. Values are expressed as mean ± SD.

HA-SF-Schwann cells biohybrids
Schwann cells seeded within the lumen of HA tubular scaffolds create a continuous cylinder which adapts to the inner channel surface, which acts as a template for the formation of such a cell sheath [19,20]. In HA tubes this process needs 9 days, and it is the result of the interplay of cell proliferation and cell···cell contacts being preferred over cell···material contacts, due to the highly hydrophilic nature of the HA matrix. To study this process in HA-SF tubes rSCs were seeded and cultured for 1, 5 and 10 days. Staining with FITC-phalloidin and DAPI permit to visualize the cell cytoskeleton (green in Fig. 5) and nuclei (blue), respectively. A comparison of Fig. 5A and 5D shows that at day 1 there are already differences between both materials, cell density and spread being higher in the HA-SF tubes. At day 5 ( Fig. 5B and 5E) considerable proliferation of SCs has occurred in both cases. However, while the One week after the scaffold implantation, a moderate immune response can be observed for both HA (Fig. 8 A-D) and HA-SF ( Fig. 9 A-D) scaffolds. In both cases a fibrous capsule surrounding the scaffold is clearly visible, which is a typical reaction after the implantation of a biomaterial. There is also a dense population of granulocytes surrounding the scaffolds. Notably, some collagen structures and some cross sections of blood vessels and capillaries are present, although small and isolated.
After 4 weeks, a reduction of the immune response can be observed, corresponding to a mild foreign body reaction for both HA (Fig. 8 E-H) and HA-SF ( Fig. 9 E-H) scaffolds. The population of granulocytes has been reduced and it has infiltrated inside the scaffold. Now, a greater presence of macrophages is observable, which indicates that the bioresorption process of the scaffold has begun. Remarkably, there is an increase of the number and extension of collagen structures, indicating that new extracellular matrix is being deposited. In addition, more cross sections of blood vessels and capillaries and more red blood cells are observable around and inside the scaffold. It seems that this process of collagen deposition and vascularization inside the scaffold is accelerated when SF is present, since more collagen structures and more blood vessels can be observed inside the scaffold.   1124  1125  1126  1127  1128  1129  1130  1131  1132  1133  1134  1135  1136  1137  1138  1139  1140  1141  1142  1143  1144  1145  1146  1147  1148  1149  1150  1151  1152  1153  1154  1155  1156  1157  1158  1159  1160  1161  1162  1163  1164  1165  1166  1167  1168  1169  1170  1171  1172  1173  1174  1175  1176  1177  1178  1179  1180 These results suggest that both HA and HA-SF scaffolds are biointegrated by the host after 8 weeks of implantation and that, furthermore, the processes of extracellular matrix deposition and angiogenesis are accelerated when SF is present.

Conflicts of interest
There are no conflicts to declare.     The cytoskeleton is shown in green and the nuclei are shown in blue. Scale bars = 100 μm. The presence of SF seems to accelerate the formation of the cell sheath, which is completely developed in the HA-SF conduits at day 5 (arrows in E showing a fold of this structure), while it took 10 days to fully develop in HA tubes (arrows in C).   7. HA and HA-SF scaffolds were implanted subcutaneously by means of two independent skin incision of approximately 1 cm on the back of each mouse. The skin sections containing the scaffolds were photographed at 1 (A and B for HA and HA-SF scaffolds, respectively), 4 (C and D for HA and HA-SF scaffolds, respectively) and 8 (E and F for HA and HA-SF scaffolds, respectively) weeks after their implantation. Healthy tissue can be observed surrounding the scaffolds since week 1. A high capillarity (white arrows) that is near to or in direct contact with the scaffolds is also visible since the first week. Scale bars = 2 mm. Fig. 8. Cross sections of representative HA scaffolds stained with H&E and Masson's trichrome. One week after the implantation of the scaffold, an acute moderate foreign body reaction can be observed, with a large infiltration of granulocyte populations (A and B for H&E staining, C and D for Masson's trichrome staining). A fibrous capsule surrounding the scaffold is clearly visible which is a typical reaction after the implantation of a biomaterial. After 4 weeks, a reduction of the immune response can be observed, corresponding to a mild foreign body reaction (E and F for H&E staining, G and H for Masson's trichrome staining). The population of granulocytes has been reduced and there is a greater presence of macrophages. After 8 weeks, the scaffold has been completely resorbed and assimilated into the native tissue (I and J for H&E staining, K and L for Masson's trichrome staining). The fibrous capsule has disappeared and there are no granulocytes present. Now the cells present are mainly macrophages and active fibroblasts. Also, an increase of collagen structures inside the scaffold can be observed thanks to the fact that active fibroblasts have managed to invade the biomaterial. Many cross-sections of blood vessels are also visible after 8 weeks. Scale bar = 2 mm (A, C, E, G, I and K) and 500 μm (B, D, F, H, J and L). An infiltration of granulocyte populations is also observed. After 4 weeks, a clear reduction of the immune response can be observed, since there is a clear presence of blood vessels and collagen networks inside the scaffold and the population of granulocytes has been reduced (E and F for H&E staining, G and H for Masson's trichrome staining). After 8 weeks, the fibrous capsule has disappeared and there are no granulocytes present, so the scaffold has been completely resorbed and assimilated into the native tissue (I and J for H&E staining, K and L for Masson's trichrome staining). Now the cells present are mainly macrophages and active fibroblasts. Many collagen structures and cross-sections of blood vessels can be observed. Scale bar = 2 mm (A, C, E, G, I and K) and 500 μm (B, D, F, H, J and L).