A Composite Dermal Filler Comprising Cross-Linked Hyaluronic Acid and Human Collagen for Tissue Reconstruction

Abstract

In this study, we developed a composite filler comprising cross-linked hyaluronic acid (HA) and human collagen (COL) derived from the human umbilical cord with the aim of improving its biocompatibility and longevity compared with commercially available fillers. After HA/COL composite fillers were made in two different ratios (10:1 and 5:1), the physical properties of the fillers were evaluated. The interior morphologies and in vivo weight change of these hydrogels were also characterized at 1-16 weeks after injection into mice. To evaluate their biocompatibility and durability in vivo, we injected the composite fillers into nude mice subcutaneously. The variations of injected gel weight were measured and compared with the commercial dermal fillers (Restylane and TheraFill). The composites showed improved or similar physical properties (complex viscosity of 19-22 × 10⁵ cP, and injection force of 10-12 N) over the commercial dermal fillers. Sixteen weeks following the injection, the ratio of remaining composite filler weight to initial weight (75.5 ± 16.9%; 10:1) was shown to be greater than that of the commercial fillers (43.2 ± 8.1%, Restylane; 12.3 ± 5.3%, TheraFill). In addition, immunohistochemical analysis with angiogenesis-related markers such as isolectin and vWF revealed newly formed blood vessels and cellular influx into the composite filler, which were not observed in the other fillers. These results clearly suggest that the HA/COL composite filler is a superior candidate for soft tissue reconstruction. The filler we developed may be a suitable candidate as an injectable dermal filler for tissue augmentation in humans.

Introduction

Injectable dermal fillers have recently been the subject of great interest for aesthetic and cosmetic improvement of skin as well as tissue augmentation owing to the associated simple and short surgical procedures and rapid facial rejuvenation after surgery [4]. A dermal filler can also correct void space and increase tissue volume [5], and it can be used for soft tissue defects, such as skin patches, wrinkles, and tissue cavities caused by surgery. As dermal filler surgeries are becoming popular, various commercial dermal fillers have been developed to meet the increasing demand. Dermal fillers are classified by their durability as permanent, semi-permanent, or non-permanent, depending on the characteristics of the base material [18].

An effective dermal filler must, most importantly, satisfy safety needs, including excellent biocompatibility, nonimmunogenicity (or a weak immune response), biodegradability without a toxic secretome, and minimal risk of infection [4,11]. It should also be long-lasting, with easy installation, and inexpensive [7]. Although it is difficult to attain all of these desired characteristics, recently commercialized fillers are coming close to meeting these goals.

Among commercial dermal fillers, hyaluronic acid (HA) and collagen (COL) are the most widely used natural polymers for implantable materials. HA, a naturally occurring biodegradable polymer, provides viscoelasticity to the dermis, fascia, and most fluid media in humans [19]. In particular, there is a high concentration of HA in soft connective tissues, extracellular matrices, hyaline cartilage, synovial joint fluid, disc nuclei, umbilical cord, and skin dermis [17,24]. In addition to occurring naturally in the human body, HA has been reported to have excellent biocompatibility, to be capable of attracting a large amount of water, and to be non-immunogenic. Thus, HA-based fillers of various origins (e.g., animals, humans, and microbes) are currently the most prevalent fillers, with brands including Restylane, Juvederm, Hylaform, and Captique [2,15,28]. HA-based fillers are generally modified by cross-linking agents such as butanediol diglycidyl ether (BDDE), divinyl sulfone, and biscarbodiimide to enhance their properties, including stability and longevity in the skin. However, despite these enhancements, they often last for only 6 months in vivo, and their effects significantly decrease over time [1]. Thus, dermal fillers may need to be repeatedly injected to maintain their properties [25]. For a dermal filler to be effective, a tissue residue time of only a few days or months is not sufficient, as longer durations in tissues may be required to maintain its optimal correction [9]. Furthermore, when HA is used as the sole substance in a dermal filler, the degree of cell affinity (adherence) for HA decreases owing to a lack of extracellular matrix producers [29]. Therefore, it will not significantly affect cell viability or proliferative activity and will not encourage local production of matrix proteins.

Collagen is found in the interstitial tissue of virtually all parenchymal organs, where it contributes to the stability and structure of tissues and organs owing to its characteristic molecular structure. Collagen as a superior resorbable material not only exhibits low inflammatory and good biocompatibility properties, along with natural degradation, but it also promotes cell migration, cell proliferation, and angiogenesis [21]. Thus, collagen-based fillers of various origins (human, animal) have also been commercialized, with brand names including Zyderm and Cosmoderm. Despite its widespread acceptance as a safe and multifunctional material, the outlook for commercial collagen-based dermal fillers, which are generally of bovine origin, is not optimistic, as bovine collagen has the potential to evoke immune responses and allergic reactions in recipients [3]. The risk of bovine spongiform encephalopathy (BSE) is another recent concern [22,29].

To offset the disadvantages of HA and collagen while making use of the advantages of these materials, we fabricated a new type of dermal filler by adding human collagen to cross-linked HA to improve its biocompatibility. Our group has developed a cytocompatible filler that is effective for dermal reconstruction using composite hydrogels made of cross-linked hyaluronic acid and human umbilicalcord-derived collagen, which is able to eliminate the concerns mentioned above. We prepared two different ratios of HA/COL composite hydrogels. These hydrogels were assessed for direct effects in vivo, including biocompatibility and changes in weight, and compared against two commercially available dermal fillers (TheraFill and Restylane), which are based on collagen and HA, respectively. In addition, histocompatibility in vivo after injection was observed by hematoxylin and eosin (H&E) staining, immunofluorescence, fluorescent staining for the expression of isolectin and von Willebrand factor (vWF), and DAPI staining. We evaluated the biocompatibility and degradability of the hydrogels in vivo through a histological analysis.

 

Materials and Methods

Preparation of HA and Umbilical-Cord-Derived COL

The cross-linked HA that was used was HyaFilia (CHA Meditech Co. Ltd., Daejeon, Korea), which is of non-animal origin and cross-linked using 1, 4-butanediol diglycidyl ether. It is granular with a mean particle size of 479 µm. Type I collagen (pH 6.5 ± 1.0, viscosity 4–40 × 105 cP) was isolated from a human umbilical cord donated by a healthy volunteer at CHA Hospital (Seoul, Korea) and used within 24 h. This study was performed with the approval of the Institutional Review Board of CHA Hospital (Approval No. 2009-032). All information pertaining to subjects and human samples were used in compliance with Korean legislation, and participants gave written informed consent. Collagen was extracted by dissociating the human umbilical cord into 1 cm pieces using sterile surgical scissors and washing it three times in distilled water. For virus inactivation, the dissociated tissues were immersed in 70% ethanol for 24 h at 4℃, washed in distilled water, and immersed in 3% H2O2 on a magnetic stirrer for 24 h at 4℃. The tissue then was washed twice in distilled water, homogenized in 0.5 M acetic acid, and transferred to pepsin (Sigma-Aldrich, St. Louis, MO, USA) for 24 h at 4℃. For pepsin inactivation, the tissue suspensions were centrifuged at 15,000 ×g for 30 min at 4℃ after adjusting the pH of the suspension to 7 with 10 N NaOH. Supernatant proteins were precipitated with 2.4 M NaCl for 12 h. The mixture was clarified at 15,000 ×g for 30 min at 4℃, and the pellet was subsequently desalted and concentrated using an ultrafiltration system. We obtained 18 ± 2.4 mg of collagen protein from 1 g of umbilical cord.

Analysis of Collagen Purity by SDS-PAGE

Sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE) was conducted for analysis of the purity of umbilicalcord-isolated collagen using 6% polyacrylamide gels and 5% stacking gels at room temperature. Before the electrophoresis, the sample was heated in the presence of sample buffer (70 mM Tris-HCl (pH 6.8), 11.4% (v/v) glycerol, 3% (w/v) SDS, 0.01% (w/v) bromophenol blue, and 5% (v/v) β-mercaptoethanol) at 100℃ for 5 min. After the sample (20 µg) was loaded in the wells, electrophoresis was performed in a gel electrophoresis system (Power Pac; Bio-Rad Inc., Singapore). The gel was dipped in 0.1% Coomassie brilliant blue R-250. Destaining was achieved by placing the stained gels in destaining solution (MeOH:Acetic acid:Distilled water, 1:1:8).

Fabrication of HA/COL Composite Hydrogels

The HA/COL composite hydrogels were prepared by blending 2% HA and 2% COL at two different ratios (10:1 and 5:1) using a PT 1200E homogenizer (Kinematica, Luzern, Switzerland) for 3 min. The hydrogels were autoclaved at 121℃ for 20 min for sterilization and were then aliquoted into 1 ml sterilized syringes for experimental use.

Fourier Transform Infrared (FTIR) Spectrometry Analysis

FTIR spectroscopy was performed using an IRPrestige-21 spectrophotometer (Shimadzu Corp., Kyoto, Japan) to identify the extent of cross-linking among the cross-linked HA and the incorporation of the HA and COL. The spectra were obtained over a range of 750-4,000 cm-1 .

Measurement of Complex Viscosity

Complex viscosities of HA/COL composites, cross-linked HA, and Restylane were measured using a Physica MCR 301 rheometer (Anton Paar, Graz, Austria) fitted with a cone-plate geometry. All measurements were performed using 0.6 ml of sample with a 35 mm/18 titanium cone sensor at 258℃. Oscillation measurements were performed using a frequency of 0.02 Hz.

Injection Force

The injection force was measured using an IM-010 syringeability test machine (Ganatech Co., Ltd., Daejeon, Korea) with a 1 ml syringe equipped with a 27-gauge needle operating at an injection speed of 12 mm/min. Experiments were performed in triplicate.

Injection and In Vivo Weight Change Tests of Hydrogels

The in vivo biocompatibility of hydrogels was assessed by injecting the hydrogel into 5-week-old male BALB/c-nu Slc mice (Orient Bio Inc., Seongnam, Korea). Before performing the hydrogel injection experiments, the animals were quarantined for a week and allowed free access to food and water to adapt to the laboratory environment but not given any antibiotics. The mice were housed at a controlled temperature of 24℃, a relative humidity of 55%, and a 12 h light cycle. Commercial dermal fillers, Restylane (based on non-animal HA, manufactured by Q-Med., Uppsala, Sweden) and TheraFill (based on collagen, manufactured by Sewon Cellontech, Seoul, Korea), were used as positive controls for tissue augmentation. Each hydrogel sample for the in vivo weight change test was prepared in a syringe (1.0 ml). For this study, 36 mice were selected to serve as host animals of each injected hydrogel. Two hundred microliters of each prepared HA/COL hydrogel was injected to evenly fill into the backs of the mice. For comparison, mice were also injected with Restylane or TheraFill as positive controls, as shown in Fig. 1. A total of 144 injections were performed, and the mice were housed in a pathogen-free environment. In total, 12 mice from each group were sacrificed at 1, 8, and 16 weeks after injection, and the hydrogels were carefully removed using surgical scissors. The degree of hydrogel weight change was determined by measuring the weight and determining its ratio relative to the initial weight.

Fig. 1.Mice models injected subcutaneously with 200 µl of (1) Restylane and (2) TheraFill (controls) and (3) HA/COL (10:1) and (4) HA/COL (5:1) (experimental hydrogels). The images were taken at 1 week after the injection.

Histological Analysis

The extracted gels were immediately fixed with 4% formalin and embedded in paraffin. The embedded specimens were sectioned at 4 µm intervals along the longitudinal axis of the implant and slide-mounted, and the slides were stained with H&E (SigmaAldrich). The slides for immunochemistry were deparaffinized and rehydrated using a graded series of ethanol solutions. The slides were washed with 0.05% Tween 20 in phosphate buffered saline (PBS-T) and blocked with 5% bovine serum albumin (Bovogen, Keilor East, Australia) or 5% horse serum (Invitrogen, Carlsbad, CA, USA) in PBS-T for 1 h at 37℃. The sections were then incubated overnight at 4℃ with primary antibody diluted with PBS-T containing 10% horse serum. Dilution of the following primary antibodies was performed: rabbit anti-vWF (1:50; H-300, Santa Cruz Biotech. Inc., Dallas, TX, USA), fluorescein isothiocyanate (FITC)-conjugated anti-isolectin B4 (1:25; Vector Laboratories, Burlingame, CA, USA). After washing with PBS-T for 5 min three times, the sections were incubated with Cy3 goat anti-rabbit immunoglobulin G (IgG; Life Technologies, Carlsbad, CA, USA) diluted to 1:250 with PBS-T containing 10% goat serum for 3 h at room temperature in the dark. The slides were washed again with PBS-T, counterstained with 4’,6-diamidino-2-phenylindole dihydrochloride (DAPI; Sigma-Aldrich), and then mounted with a fluorescence mounting solution (DAKO, Carpinteria, CA, USA). Immunofluorescence was visualized using an Axio Imager A1 microscope (Carl Zeiss Microimaging GmbH, Göttingen, Germany) equipped with AxioVision Rel. 4.8 software (Carl Zeiss Microimaging GmbH). Before the acquisition of the immunofluorescent images, the delimitation between the gel and the host tissue was determined by differential interference contrast optical microscopy.

 

Results

Characterization of HA/COL Composite Hydrogels

After isolating collagen from human umbilical cord, its characteristics were analyzed. We obtained 18 ± 2.4 mg from 1 g of umbilical cord through the collagen isolation process, and the purity of the collagen was 97% based on SDS-PAGE analysis (Supplemental data 1). We prepared homogeneously dispersed HA/COL composite hydrogels. To identify the cross-linking of HA and the incorporation of the HA and COL, the FTIR spectra of HA/COL (10:1), cross-linked HA, and COL were analyzed (Supplemental data 2A). The collagen spectra showed characteristic amide bending (1,636, 1,553, and 1,237 cm-1). For enhancement of the mechanical properties of HA, BDDE was used as a nontoxic cross-linking agent. The HA spectra had a peak at 1,300 cm-1 that did not appear in the uncross-linked HA because the carboxyl groups and hydroxyl groups of HA can be converted into ether chains with BDDE through the cross-linking process. Therefore, the ether linkages could make the cross-linked HA less susceptible to enzymatic degradation. In addition, the HA/COL spectra clearly showed specific peaks corresponding to HA and COL, suggesting that their properties were not affected by the mixing process. As shown in Supplemental data 2B, we compared the complex viscosity of Restylane, cross-linked HA, HA/COL (10:1), and HA/COL (5:1) composite hydrogels to assess differences in their physical properties and to correlate these physical properties with performance. We found that Restylane had a higher viscosity than crosslinked HA and HA/COL. Additionally, HA combined with the collagen hydrogels had a higher complex viscosity. The swelling ratio of the HA/COL composite hydrogel was influenced by the proportion of HA. Increasing the concentration of collagen within the composite hydrogel decreased the injection force.

In Vivo Tissue Augmentation of HA/COL Composite Hydrogels

As controls, we selected the commercial dermal fillers Restylane and TheraFill, which are composed of crosslinked HA and porcine skin collagen, respectively. The spontaneous formation of a hydrogel in vivo was confirmed by injecting 200 µl of hydrogel subcutaneously into mice with a syringe needle. The injected hydrogel formed a regular, round skin protrusion without immediate resorption or dispersion into the surrounding body tissues (Figs. 1 and 2). The mice were sacrificed 1, 8, and 16 weeks after injection, and the changes to the injected hydrogels were observed. Restylane and HA/COL (10:1 and 5:1) composite hydrogels were clear, transparent, and stable. In contrast, the TheraFill hydrogel was white and opaque, and shrank considerably after 1 week. Therefore, we investigated the weight change by measuring the remaining weight of the HA/COL composite hydrogel in vivo.

Fig. 2.Images of hydrogels removed from mice at 1 week, 8 weeks, and 16 weeks after injection.

In Vivo Weight Change of HA/COL Composite Hydrogels

After injecting the various hydrogels into the mice, changes in the weight and morphology of the gels were observed at 1, 8, and 16 weeks. The TheraFill weight decreased rapidly from 200 mg at the week of injection to 30 mg at 16 weeks without swelling, as observed for the other tested HA-based hydrogels. In contrast, the Restylane and HA/COL hydrogels swelled approximately 1.8-fold by 1 week compared with their initial weight (200 mg), due to the high hydration property of HA. After 1 week, those gels began to gradually decrease in weight, following a similar pattern (Fig. 3A). The 10:1 HA/COL hydrogel showed the lowest weight change rate (24.5 ± 16.9%) among the gels after 16 weeks (Fig. 3B).

Fig. 3.In vivo degradation of controls and experimental hydrogels. (A) Difference in hydrogel weights after 1-16 weeks, and (B) rate of decrease in hydrogel weights at 16 weeks. Error bars represent the standard error of the mean (compared with Restylane, *p < 0.05; compared with TheraFill, #p < 0.5, ##p < 0.05).

Histological Analysis After Tissue Augmentation

Histological analysis of cell infiltration into the hydrogels was performed with H&E staining at 1 and 16 weeks after injection into the dorsal skin of mice. As shown in Figs. 4E and 4F, fibroblast cells did not enter or integrate into the injected Restylane and TheraFill hydrogels. A number of fibroblast cells were located on the surface of the hydrogels. Conversely, in the injected HA/COL composite hydrogels, a number of fibroblasts were observed inside the hydrogel (Figs. 4G and 4H) and there were significantly more fibroblasts than in the Restylane and TheraFill hydrogels. The cells from the surrounding tissues migrated to the injected HA/COL composite hydrogels and regenerated a new tissue structure. Some of these areas became vascularized, exhibiting arterioles and capillaries. We investigated whether cell influx into the HA/COL composite hydrogels to support vascular formation had occurred. In vivo HA/COL composite hydrogels were examined by staining with angiogenic markers, isolectin B4 and vWF antibody, and DAPI was used as a nuclear stain 16 weeks after injection (Fig. 5). Angiogenesis of new blood vessels from fluorescent staining of the isolectin (green) and vWF (red) markers was markedly observed for HA/COL (10:1) and (5:1).

Fig. 4.Hydrogels stained in cross-section with H&E at 1 and 16 weeks after injection. Restylane (A), (E); TheraFill (B), (F); HA/COL (10:1) (C), (G); HA/COL (5:1) (D), (H).

Fig. 5.Fluorescent staining of hydrogels with antibodies to isolectin (green) and vWF (red), and with DAPI (blue), 16 weeks after injection. Restylane (A), (B), (C), (D); TheraFill (E), (F), (G), (H); HA/COL (5:1) (I), (J), (K), (L); HA/COL (10:1) (M), (N), (O), (P).

 

Discussion

Injectable dermal fillers made of various sources (e.g., human, animals, and bacteria) and by various methods (e.g., filler formulation, modifications, and cross-linking) have recently been developed for aesthetic use as well as clinical remedies [12]. Hence, the demand is continuously growing, and consumer expectations of effective fillers are also greater. In fact, recent injectable fillers are becoming more similar to human skin or tissue [29]. However, there are still numerous concerns regarding biocompatibility, biosafety, adverse reactions, allergic reactions, inflammation, durability, physical properties, and cost [1]. Approved dermal fillers have been shown to be relatively safe, but varying degrees of resorption make recipients require repeated percutaneous injections to maintain the expected level of collection. Therefore, new dermal filler hydrogels should be able to offer in vivo stability to ensure the longevity of the injectable implant as well as biosafety.

For this reason, we considered which materials would be suitable for advanced dermal filler hydrogels. Among the material candidates, collagen, being the major protein of the natural extracellular matrix, contains basic residues such as lysine and arginine. It also has specific cell adhesion sites such as arginine-glycine-aspartate (RGD) peptides [20]. The RGD group actively induces cellular adhesion by binding to integrin receptors, and this interaction plays an important role in cell growth and differentiation. Eventually, it could be used to enhance the overall regulation of cell function [11,23]. However, collagen as a dermal filler has drawbacks, including a relatively short duration of use in vivo, the possibility of recipient hypersensitivity, and BSE concerns when using bovine-derived collagen [6]. Thus, the use of autologous human collagen and collagen derived from various human organs or tissues has been investigated for medical applications to overcome these problems [14].

We used human collagen isolated from Warton’s jelly of umbilical cords, which are discarded as medical waste, for dermal filler substance. Kim et al. [16] reported that triplehelical proteins (γ form) were considerably abundant in umbilical-cord-derived COL in comparison with commercial products derived from human placenta and rat tail tendon. This result indicates that the collagen from umbilical cords is well-conserved in its active sites, which promotes cell proliferation and migration, as mentioned above. However, umbilical-cord-derived COL also rapidly decomposes in vivo when it is used as the sole component of a dermal filler. Furthermore, if collagen is implanted alone, its viscosity increases, which acts as a barrier to cell migration.

Cross-linked HA is less susceptible to enzymatic degradation by elimination of carboxyls and is widely used. It has been developed as a next-generation dermal filler owing to its superior properties over collagen, including no requirement of a skin test, longer duration of use, and no chemical or molecular differences between species [10]. However, the presence of a carboxyl group (COO-) in HA can cause poor cell adhesion.

Hence we fabricated a prototype of a material that possesses high biocompatibility by mixing human umbilicalcord-derived collagen into cross-linked HA with two ratios (HA:COL = 10:1 and 5:1), because a higher ratio of COL than HA in the hydrogel can result in short durability in vivo, as mentioned above. It can be expected that this composite material will show excellent longevity in clinical applications and good biocompatibility. It has been reported that the combination of HA and collagen often has a positive effect in tissue engineering applications [27]. For example, Davidenko et al. [8] demonstrated that a collagenHA scaffold is able to enhance mammary stromal tissue development. Incorporation of HA into the collagen matrix stimulated chondrocyte and fibroblast expression [26]. Despite the synergistic effect of HA and collagen composites, there is no report on dermal filler applications. Thus, to offset the disadvantages of HA and collagen as well as to make use of the advantages of these materials, we fabricated a new type of dermal filler by the addition of human collagen to cross-linked HA to improve the biocompatibility of the filler.

To demonstrate its efficiency, we prepared a material by mixing umbilical-cord-derived COL into HA composite hydrogels. Sixteen weeks after injecting this mixture into mice, in vivo longevity and changes in size and appearance were observed and compared with the effects of commercial dermal fillers. Among the injected fillers, HA/COL (10:1) resulted in a lower weight decrease than Restylane or TheraFill after 16 weeks (Fig. 2). The reduced changes in size and weight with time indicate that the corrective effect and aesthetic results can be maintained for a long duration, and the number of injections may therefore be potentially reducible. A 10% or 20% addition of collagen into the HA did not significantly affect its longevity (Fig. 3), while the weight of TheraFill decreased by 88% at 16 weeks.

As with all injectable materials, histocompatibility and biodegradability are important clinical variables. If the gel scaffold loses its stability, the injury site will be subject to compressive stress, leading to the acceleration of cell death and inflammation [11]. As shown in Fig. 4, a large number of fibroblast infiltrations were observed in the HA/COL hydrogels, whereas the cells could not be observed in the commercial gels, because the collagen-based filler (TheraFill) has higher viscosity acting as cell barriers, and the HAbased filler (Restylane) has poor cell adhesion due to carboxyl groups in HA. Although a large amount of HA was subcutaneously injected, it moved easily under the skin owing to its hydrophilicity. In this regard, our developed filler demonstrated its superior cytocompatibility through cell infiltration in vivo over fillers made of a single substance. These results not only indicate its mobility through cointegration with surrounding tissue, but they also suggest the promotion of vascular ingrowth. Thus, the filler would be fixed at the site of injection. Consequently, we could expect to prolong dermal correction and to augment tissue post-injection. Although the two HA/COL fillers (10:1 and 5:1) showed better performances than the others, 10:1 (HA/COL) is preferable owing to its longevity in vivo compared with 5:1.

In this study, we constructed injectable fillers from safe sources using simple and easy fabrication methods and evaluated their effectiveness in vivo. In our demonstrations, we achieved a long-lasting effect as well as cytocompatibility in vivo compared with commercial fillers. The filler we developed may be a suitable candidate as an injectable dermal filler for tissue augmentation in humans. Further studies and tests should follow to ensure its safety and efficacy.