A hydrophobic gelatin fiber sheet promotes secretion of endogenous vascular endothelial growth factor and stimulates angiogenesis

In tissue engineering and regenerative medicine, the formation of vascular beds is an effective method to supply oxygen and nutrients to implanted cells or tissues to improve their survival and promote normal cellular functions. Various types of angiogenic materials have been developed by incorporating growth factors, such as vascular endothelial growth factor, in biocompatible materials. However, these exogenous growth factors suffer from instability and inactivation under physiological conditions. In this study, we designed a novel angiogenic electrospun fiber sheet (C16-FS) composed of Alaska pollock-derived gelatin (ApGltn) modified with hexadecyl (C16) groups to induce localized and sustained angiogenesis without growth factors. C16-FS was thermally crosslinked to enhance its stability. We demonstrated that C16-FS swells in phosphate-buffered saline for over 24 h and resists degradation. Laser doppler perfusion imaging showed that C16-FS induced increased blood perfusion when implanted subcutaneously in rats compared with unmodified ApGltn-fiber sheets (Org-FS) and the sham control. Furthermore, angiogenesis was sustained for up to 7 days following implantation. Immunohistochemical studies revealed elevated nuclear factor-κB and CD31 levels around the C16-FS implantation site compared with the Org-FS implantation site and the control incision site. These results demonstrate that C16-FS is a promising angiogenic material to promote the formation of vascular beds for cell and tissue transplantation without the need for growth factors.


Introduction
Tissue engineering and regenerative medicine is a promising approach to repair or replace damaged tissues and organs.Clinical studies in this eld are mainly based on injecting cell suspensions or transplanting engineered tissues at a target site, where they are expected to serve as a substitute for the damaged tissue.However, the survivability of implanted cells and tissues is low because the supply of oxygen and nutrients via native blood vessels is insufficient. 1,2][5] Biomaterials researchers have developed various angiogenic materials by incorporating growth factors (GFs), such as vascular endothelial growth factor (VEGF), basic broblast growth factor (bFGF), and stromal cell-derived factor (SDF)-1, in hydrogels, [6][7][8] particles, 9,10 and porous scaffolds. 11,12The sustained release of GFs from these materials promotes angiogenesis in vitro and in vivo as well as wound healing, [13][14][15] regeneration in an infarcted heart, 16 and survivability of implanted islets. 17][20][21] On the other hand, Kanno et al. reported that inoculation with Pseudomonas aeruginosa, a Gram-negative bacterium, promotes re-epithelialization and angiogenesis by attracting neutrophils to the wound site. 22However, the bacteria proliferate rapidly, leading to an unstoppable cytokine storm and sepsis.These inammatory response and angiogenesis are caused by lipopolysaccharide (LPS), a constituent of the cell wall of Gram-negative bacteria.4][25][26][27][28] We previously reported that SFA-modied Alaska-pollock-derived gelatin (ApGltn), self-assembles into a hydrogel when hydrated with phosphate-buffered saline (PBS) and stimulated VEGF secretion and angiogenesis in vitro and in vivo. 29However, the SFA-ApGltn hydrogels could not maintain their shapes under physiological conditions as they are vulnerable to enzymatic degradation.Chemical, 30,31 enzymatic, 32 and physical 33,34 methods have been used to make gelatin-based materials resistant to enzymatic degradation in vivo. 35However, in the case of electrospinning, it is difficult to attain homogenous crosslinking and remove residual crosslinker and byproducts with chemical and enzymatic methods.][38] In the present study, we used a hexadecyl group (C16) as an immune activator 25 to synthesize C16-modied ApGltn (C16-ApGltn), then used electrospinning to fabricated C16-ApGltnbased ber sheets (C16-FSs) then thermally crosslinked to increase the crosslinking density and enhance their stability.The physicochemical properties of the C16-FSs with different crosslinking densities were evaluated in terms of the water contact angle, swelling ratio, and degradation rate.Moreover, the in vivo angiogenic properties of the C16-FSs in rats were quantied by laser doppler perfusion imaging (LDPI) and immunohistochemical staining.

Synthesis of C16-ApGltn
0][41] Briey, 100 ml of 0.1 w/v% ApGltn dissolved in 1 : 1 DMSO/H 2 O was dispensed into each well of a 48-well plate.100 ml of a solution of 0.1% TNBS and 0.1% triethylamine in 1 : 1 DMSO/H 2 O was then added to each well.The plate was then shaken for 1 min using a microplate shaker and incubated or 2 h at 37 C.Then, the absorbance at 335 nm in each well was measured with a microplate reader (Spark 10M; Tecan, Männedorf, Switzerland) to calculate the number of amino groups.2-Aminoethanol (which has a known concentration of amino groups) was used instead of ApGltn to generate a calibration curve.
C16-ApGltn was synthesized by reductive amination [42][43][44] as described previously 41 (Fig. S1 †).Briey, 20 w/v% ApGltn was dispersed in a 30% ethanol aqueous solution and dissolved by heating at 50 C. Hexadecanal and 2-picoline borane were dissolved in ethanol and added to the ApGltn solution in 1.5-fold molar excesses relative to the remaining amino groups in the ApGltn.The reaction was carried out for 18 h.The reaction solution was then added dropwise to cold ethanol (10 times the reaction solution volume) to precipitate the C16-ApGltn, which was then isolated by suction ltration using a glass lter.The C16-ApGltn was then re-dispersed in ethanol three times to remove remaining the hexadecanal and 2-picoline borane.The C16-ApGltn precipitate was nally vacuum-dried at 25 C, yielding a white powder.Unmodied ApGltn (Org-ApGltn) was prepared by the same protocol used for C16-ApGltn without the addition of hexadecanal and 2-picoline borane.The modication ratio of hexadecyl groups in C16-ApGltn was quantied using the TNBS method and comparing the absorbance at 335 nm with that of Org-ApGltn.The degree of modication was further conrmed by 1 H nuclear magnetic resonance ( 1 H-NMR) (JNM-AL300; JEOL, Tokyo, Japan).

Fabrication and characterizations of C16-FS
Microber sheets were fabricated by electrospinning the gelatin solution.Briey, Org-ApGltn was dissolved in a 40% aqueous ethanol solution at 30 w/v%, and C16-ApGltn was dissolved in a 50% EtOH aqueous solution at 15 w/v%.Electrospinning was conducted using a NANON-03 instrument (MECC, Japan) at room temperature (25 C) (Fig. 1A).Org-and C16-ApGltn solutions were extruded from an 18 G needle at a feed rate of 1 ml h À1 with voltages of 20 and 25 kV, respectively.The obtained Org-FS and C16-FS were observed with a scanning electron microscope (SEM) (S-4800 ultrahigh-resolution SEM, HITACHI, Japan), and the ber diameters were measured using Image J soware (n ¼ 60).Fourier transform infrared spectroscopy (FT-IR) (ALPHA II; Bruker Corp., Billerica, MA, USA) was conducted on a C16-FS to conrm the presence of hexadecyl groups.Fiber sheets were then thermally crosslinked in a vacuum oven at 150 C for 4 h to improve their degradation resistance.

Water contact angle measurements
To evaluate the hydrophobicity of the ber sheets, water contact angle measurements were conducted using a DropMaster DM-701 (Kyowa Interface Science, Japan).Contact angle measurement was initiated 0.5 s aer a 3 ml PBS droplet was dropped on the ber sheet and continued for up to 15 s.The surface contact angle of the PBS droplet on the ber sheet was then analyzed using FAMAS soware.

Swelling ratio
Aer 4 h of thermal crosslinking, ber sheets were cut into 10 mm-diameter circles, each of which was weighed to determine the dry weight (W d ).Fiber sheets were then immersed in PBS at 37 C for up to 1440 min.The swollen ber sheets were weighed at each time point to determine the swollen weight (W s ).Hence, the swelling ratio of each sheet was calculated using the following equation:

Enzymatic degradation of C16-FS
The enzymatic degradation of Org-FS and C16-FS was evaluated using collagenase.Briey, aer 4 h of thermal crosslinking, Org-FS and C16-FS were cut into 10 mm-diameter circles and immersed in PBS at 37 C for 1 h for pre-swelling.The swollen FSs were then immersed in a solution of 5 U ml À1 collagenase in PBS and incubated at 37 C for up to 240 min.The FSs were weighed at each time point to determine the degradation proles.

Evaluation of angiogenesis in mice
Animal experiments were approved by the Animal Care and Use Committee of National Institute for Material Science (NIMS), Japan, and were performed in accordance with NIMS Regulations Pertaining to Animal Testing.To evaluate the angiogenic properties of C16-FS, the FSs were implanted subcutaneously in mice backs and evaluated by blood ow imaging and histological observation.Thermally crosslinked Org-FS and C16-FS were cut into 10 mm-diameter circles.A total of 24 hairless mice (Hos:HR-1, female, 4 weeks old; Hoshino Laboratory Animals, Japan) were anesthetized by isourane inhalation.An incision was made in the back of each mouse, the FS was implanted, and the incision was sutured closed.Sham surgery was also conducted by making an incision and then sutured without implanting FSs.The FSs were implanted near the mice's tail roots to avoid the blood ow instability caused by the movement associated with breathing.
The blood ows around the implant site was measured using an LDPI system (OZ-2, OMEGAWAVE, Japan) immediately aer and 1, 2, 3, 7, 14, and 24 days aer the implantation.To stabilize and control the blood ow rate in the mice for imaging, the isourane concentration was adjusted until each mouse's breathing rate was 1 breath per s.To analyze the mechanisms of angiogenesis in vivo, immunohistochemical staining was conducted on different mice from those used for the LDPI experiment.3 days aer the subcutaneous implantation of Org-FS and C16-FS, the mice were sacriced using somnopentyl.The FSs and the surrounding tissue were dissected and xed in a 10% formaldehyde neutral buffer solution.Tissue sections were stained with hematoxylin and eosin (H&E); nuclear factor-kB (NF-kB), VEGF, CD31, and myeloperoxidase (MPO) were stained as markers of inammation and angiogenesis.Stained areas of NF-kB, VEGF, and CD31 were quantied using Image J.

Statistical analysis
Data are shown as mean AE standard deviation (SD).Differences between groups were evaluated for statistical signicance with the Student's t-test.Differences were considered statistically signicant when p < 0.05.

Synthesis of C16-ApGltn
C16-ApGltn was synthesized by reacting the residual amino groups in ApGltn with hexadecanal followed by reduction with 2-picoline borane.The yields of Org-and C16-ApGltn were 90.2% and 82.0%, respectively (Table 1).The modication ratio of C16-ApGltn was 24.0 mol%, as measured by the TNBS method.The 1 H-NMR spectra shown in Fig. S2 † show peaks at 1.4 ppm in the C16-ApGltn spectrum, which correlates with a strong peak at the same chemical shi in the hexadecanal spectrum, conrming the substitution of hexadecyl groups into ApGltn.

FS fabrication
The FT-IR spectra of the FSs (Fig. S3 †) show that the absorbance of C-H at 2853 and 2927 cm À1 was increased in the C16-FS spectrum compared with the Org-FS spectrum, which was  1).The SEM images of Org-FS and C16-FS shown in Fig. 1B show that the bers were successfully formed and molded into a non-woven fabric-like structure.The mean ber diameter of C16-FS (0.78 AE 0.20 mm) was larger than that of Org-FS (0.58 AE 0.20 mm), as measured from the SEM images.Histograms of the ber diameters are shown in Fig. 1C and D; these data show that over 70% of the bers in the Org-FS were between 0.3 and 0.8 mm, whereas the same percent in the C16-FS were between 0.5 and 1.0 mm.Aer the thermal crosslinking, ber diameters of Org-and C16-FS were 0.61 AE 0.13 mm and 0.72 AE 0.29 mm, respectively, which were not signicantly different with that before crosslinking.Besides, ber morphologies and histograms of Org-and C16-FS were quite similar to that before crosslinking (Fig. S4 †).ApGltn has signicantly lower sol-gel transition temperature (21.2 C) than porcine and bovine-derived gelatin (31.2 C) due to the lower content of proline and hydroxyproline, which contribute to forming the helical structure of gelatin. 45,46Therefore, ApGltn in FSs was considered to have a random coil and amorphous structure because electrospinning was conducted at room temperature (25 C). 47 The previous study reported that uncrosslinked and thermally-crosslinked gelatin bers demonstrated similar dependence on elastic modulus and tan d, which suggested that the thermally-crosslinked FSs retained amorphous structure. 48

FS hydrophobicity
The surface contact angle on each FS was measured to evaluate the surface hydrophobicity, which may affect the water absorption rate.Fig. 2A shows the behavior of PBS droplets on Org-FS and C16-FS.0.5 s aer dropping the PBS droplets, the surface contact angle of that on Org-FS was already smaller than that on C16-FS.The PBS droplets were then quickly absorbed into the Org-FS compared with the C16-FS.Fig. 2B shows the rates of decrease in the contact angles of PBS droplets on Org-FS and C16-FS over 15 s.The surface contact angle was consistently higher on C16-FS than on Org-FS over 15 s because the modi-cation of hexadecyl groups in C16-ApGltn resulted in higher hydrophobicity.Moreover, the hexadecyl groups in C16-ApGltn were considered to be localized on the surface of the C16-FS because the FSs were fabricated in air, and hydrophobic molecules tend to move toward air-water interfaces. 49We previously conrmed that hydrophobic groups conjugated to gelatin molecules tended to localize on the surfaces of thermally crosslinked microparticles by X-ray photoelectron spectroscopy. 50Owing to this surface hydrophobicity, the C16-FS repelled PBS droplets and delayed their absorption into the FS.

Swelling ratio
The swelling ratios of the FSs were measured to evaluate the effect of the hexadecyl groups in C16-FS on its stability in water (Fig. 3A).Aer immersing the FSs in PBS, they quickly absorbed PBS, as evidenced by signicant weight gain in the 1 h of swelling.Both FSs then continued absorbing water and gaining weight over time, reaching steady states aer 8 h.Although the swelling ratio of C16-FS was signicantly lower than that of Org-  FS for the rst 2 h of swelling, no signicant difference between them was observed aer 4 h until 24 h.The swelling ratio basically depends on the crosslinking density in a polymer network; higher crosslinking densities result in less swelling, while lower crosslinking densities lead to increased swelling.However, the crosslinking density of C16-FS was considered to be lower than that of Org-FS due to the substitution of hexadecyl groups for amino groups, which contributes to amide bonding via dehydration condensation; this should lead to a higher swelling ratio in C16-FS.Therefore, the relatively swelling ratio of C16-FS despite its lower crosslinking density was attributed to physical crosslinking between hexadecyl groups (C16) in the C16-ApGltn molecules.We previously reported that the modi-cation of hydrophobic groups in gelatin molecules contributes to decreased swelling ratios and increased mechanical strength in hydrogels, which is consistent with the results of the present study. 39zymatic degradation of FSs The effect of the hexadecyl groups in C16-FS on degradation rate was evaluated by applying collagenase, an enzyme that cleaves peptide bonds in gelatin and collagen, and measuring the weight over time.Before the degradation test, Org-FS and C16-FSs were pre-swelled in PBS for 1 h.Fig. 3B shows the degradation proles of Org-FS and C16-FS in a collagenase solution.
Results show that the weights of these FSs gradually decreased with time, and the weight of remaining C16-FS was higher than that of Org-FS aer 120 min.These results were attributed to the suppressed swelling ratio of C16-FS for the rst 2 h, which lead the delayed uptake of collagenase molecules into bers.However, the weights of Org-FS and C16-FS were not signicantly different aer immersion in the collagenase solution for 180 min because both FSs broke into small fragments, leading to rapid degradation.

Angiogenesis evaluation in vivo
The in vivo angiogenic properties of C16-FS were evaluated using LDPI and histological observations.Fig. 4A shows that the LDPI results aer subcutaneous FSs implantation in mice, followed by quantication of blood ow at the implantation site.There were no signicant differences between the LDPI images or blood ow in mice with Org-FS implanted, mice with C16-FS implanted, and control mice that received the sham operation (control) immediately aer the implantation (Day 0), or on Day 1.However, on Day 3, the blood ow in the tissues around the C16-FS implantation site was increased, and visible blood vessels were observed, indicating an angiogenic response; however, the blood ow decreased aer Day 3 until Day 24 (Fig. 4A).The quantitative analysis of blood ow over 24 days is shown in Fig. 4B.The blood ow measured on each day was normalized to that of the sham group on the corresponding; thus, the sham group was used as a reference and consistently exhibited 100% blood ow.There was no statistically signicant difference in the mice with Org-FS implanted compared with the sham group over 24 days.However, in the C16-FS group, the blood ow was elevated (122%) on Day 3 and sustained until Day 7, aer which time it gradually decreased to 104% on Day 24.The time-dependent decrease in blood ow around the C16-FS was attributed to the ongoing enzymatic degradation of the C16-ApGltn molecules in the FS, as demonstrated in the in vitro degradation test (Fig. 3B).On Day 24, the blood ow in the C16-FS groups returned to an approximately normal level, and the C16-FS was considered to be almost completely degraded.We also analyzed the difference in blood ow between groups at several time points (Fig. 4C).Although there was no statistically signicant difference between the blood ows in the sham, Org-FS, and C16-FS groups on Day 1, C16-FS induced signicantly higher blood ow on Day 3.Moreover, the C16-FS group sustained signicantly higher blood ow up to Day 7 compared with the sham and Org-FS groups.However, the increased blood ow associated with C16-FS compared with Org-FS was no longer observed on Day 14. Comparing with the results of our previous study with an angiogenic self-assembled hydrogel, 27 the C16-FS prepared with thermal crosslinking suppressed the subcutaneous spreading of material and successfully induced sustained localized angiogenesis for 14 days.
Fig. 5A shows tissue sections with immunohistochemical staining of NF-kB, CD31, MPO, and HE.The area of NF-kB staining revealed signicantly increased NF-kB expression in the tissue around the C16-FS (Fig. 5A and B).NF-kB promotes genes associated with inammatory cytokines and growth factors, such as VEGF, and we have previously shown that SFA-modied gelatin stimulates NF-kB via TLR4-mediated pathways. 27However, there were no statistically signicant differences between the levels of VEGF staining in the Sham, Org-FS, and C16-FS implanted site, although the higher levels of VEGF staining was conrmed inside of C16-FS compared with Org-FS (Fig. S5 †).We previously reported an increased affinity of VEGF to hydrophobically-modied gelatin based on surface plasmon resonance measurements. 51Therefore, VEGF was considered to be entrapped in the C16-FS by hydrophobic interactions between VEGF and the hexadecyl groups in C16-ApGltn, which results were correlated with the increased water contact angle of C16-FS.Furthermore, the level of the stained area of CD31, an endothelial marker, was elevated around the C16-FS implantation site compared with the sham and Org-FS (Fig. 5A and C).In addition, the number of blood vessels (indicated by red arrows in the CD31-stained sections) was higher in the C16-FS sample than in the Org-FS and sham samples, indicating that the C16-FS enhanced angiogenesis (Fig. 5A).The H&E staining showed that there was blood perfusion inside of the newly formed blood vessels around C16-FS (Fig. 5A).We further investigated the accumulation of granulocytes (including monocytes, neutrophils, and macrophages) by MPO staining, as shown in Fig. 5A  and D. The observed increased accumulation of granulocytes around C16-FS was attributed to an increased foreign body reaction by the enhanced protein adsorption on C16-FS or a direct stimulation of inammatory cells by hexadecyl groups.
In summary, C16-FS activated NF-kB secretion by stimulating TLR4-mediated pathways.NF-kB then activated the expression of VEGF, which became entrapped within the C16-FS and gradually released upon the degradation of C16-FS.The secreted VEGF promoted endothelial cell proliferation and, consequently, angiogenesis, which gave rise to increased blood perfusion (Fig. S5 †).Based on these ndings, it was concluded that C16-FS is a promising method to induce angiogenesis for prevascularization in cell or tissue transplantation without the addition of any growth factors.However, various types of inammatory cytokines are secreted in response to VEGF, which is stimulated by NF-kB; the effects of those cytokines on implanted cells need to be further assessed in future studies.Moreover, long-term inammation may lead to brosis, 52 which would be a signicant challenge for clinical application.Therefore, the incorporation of anti-inammatory drugs may alleviate the inammatory responses and modulate the balance of pro-and anti-inammatory responses to lead the angiogenesis without brosis.The modication of cyclodextrin group in FSs enables not only the incorporation of anti-inammatory drugs by forming inclusion complexes, but increased hydrophilicity, which prevents the adhesion of proteins on FSs in vivo followed by the foreign body reaction. 53Furthermore, we are now investigating how changing the thermal crosslinking time may inuence the time-dependent modulation of the angiogenic response elicited by C16-FS.

Conclusion
Here, we developed a C16-FS with thermal crosslinking to achieve sustained and localized angiogenesis in vivo.The C16-FS elicited a high water contact angle and exhibited a lower water absorption rate, indicating increased hydrophobicity of the C16-FS compared with the Org-FS.In addition, the C16-FS was stable in PBS for 24 h due to the thermal crosslinking, but the C16-FS was degraded by collagenase in 4 h (as was the Org-FS).In vivo experiments followed by LDPI observations revealed that the C16-FS induced signicantly higher blood perfusion when implanted subcutaneously in mice compared with the Org-FS.Moreover, C16-FS induced sustained and localized angiogenesis for up to 7 days, followed by a decrease in blood ow to a normal level aer 24 days.Immunohistochemical evaluations revealed that C16-FS promoted NF-kB secretion, leading to VEGF secretion, which was entrapped within the C16-FS and released, resulting in elevated CD31 expression and angiogenesis with blood perfusion.These results demonstrate that the C16-FS is a promising material for making vascular beads to improve the survivability of implanted cells or tissues without the need for drugs or growth factors.

Fig. 2
Fig. 2 Surface contact angle measurements of Org-FS and C16-FS.(A) Representative images of the behavior of PBS droplets over 15 s on Org-FS and C16-FS after 4 h of thermal crosslinking.(B) Quantification of the decrease in the contact angle over time.Error bars represent SDs of each point (n ¼ 3).

Fig. 4 (
Fig. 4 (A) LDPI images of mice backs after the sham implantation and implantation of Org-FS or C16-FS.The hues in images indicate the intensity of blood perfusion: red and blue indicate higher and lower perfusion, respectively.(B) Blood flow over 24 days, as quantified from the LDPI images.The blood flow values were normalized to that in the mice treated with the sham operation (control) for each day.Error bars represent SDs of each point (n ¼ 4-5).(C) Comparison of mean blood flows in different groups on each day (* indicates p < 0.05, ** indicates p < 0.01, and # indicates p > 0.05).

Fig. 5 (
Fig. 5 (A) Immunohistochemical studies of FSs and surrounding tissues 3 days after the sham operation and implantation of Org-FS or C16-FS (* denotes the FS, and the red arrows in the CD31-stained tissue sections indicate blood vessels).(B-D) Active areas of NF-kB, CD31, and MPO, respectively, as quantified using Image J (* indicates p < 0.05, ** indicates p < 0.01, and # indicates p > 0.05).

Table 1
Characterization of Org-FS and C16-FS to the modication of the alkyl chains with hexadecyl groups in C16-ApGltn.Therefore, C16-ApGltn molecules but C16-FS were expected to be more hydrophobic than Org-ApGltn and Org-FS.Org-FS and C16-FS were fabricated by accumulating electrospun bers into non-woven sheets.The mean thicknesses of Org-FS and C16-FS were 223 AE 17 mm and 276 AE 26 mm, respectively (Table attributed