A self-fused hydrogel for the treatment of glottic insufficiency through outstanding durability, extracellular matrix-inducing bioactivity and function preservation

Abstract

Injection laryngoplasty with biomaterials is an effective technique to treat glottic insufficiency. However, the inadequate durability, deficient pro-secretion of extracellular matrix (ECM) and poor functional preservation of current biomaterials have yielded an unsatisfactory therapeutic effect. Herein, a self-fusing bioactive hydrogel comprising modified carboxymethyl chitosan and sodium alginate is developed through a dual-crosslinking mechanism (photo-triggered and dynamic covalent bonds). Owing to its characteristic networks, the synergistic effect of the hydrogel for vocal folds (VFs) vibration and phonation is adequately demonstrated. Notably, owing to its inherent bioactivity of polysaccharides, the hydrogel could significantly enhance the secretion of major components (type I/III collagen and elastin) in the lamina propria of the VFs both in vivo and in vitro. In a rabbit model for glottic insufficiency, the optimized hydrogel (C₁A₁) has demonstrated a durability far superior to that of the commercially made hyaluronic acid (HA) Gel. More importantly, owing to the ECM-inducing bioactivity, the physiological functions of the VFs treated with the C₁A₁ hydrogel also outperformed that of the HA Gel, and were similar to those of the normal VFs. Taken together, through a simple-yet-effective strategy, the novel hydrogel has demonstrated outstanding durability, ECM-inducing bioactivity and physiological function preservation, therefore has an appealing clinical value for treating glottic insufficiency.

Graphical abstract

Highlights

• •The self-fused hydrogel has been developed with a dual-crosslinking strategy.
• •The synergistic effect of the hydrogel has been validated by vocal folds vibration and phonation assessment.
• •The self-fused hydrogel has shown outstanding durability and extracellular matrix-inducing bioactivity.
• •Preservation of the vocal folds functions has been realized by the hydrogels through multiple mechanisms.

1. Introduction

As a common dilemma for the otorhinolaryngology department, glottic insufficiency often occurs secondary to various pathological conditions of the vocal folds (VFs) including paralysis, polyps, grooving, scarring, aging larynx, etc. The clinical manifestations include persistent hoarseness and obvious breath sound, and choking and aspiration in severe cases, which can cause great inconvenience to the patient [[1], [2], [3]]. Clinically, considerable efforts have been made to optimize the treatment of glottic insufficiency, and several procedures have been developed. Among these, injection laryngoplasty can improve the voice quality by augmenting the VFs, which has emerged as a high-profile strategy for both patients and surgeons owing to its minimal invasiveness, ease of operation and affordable cost [4,5].

Various materials, including polydimethylsiloxane, hyaluronic acid (HA), autologous fat, etc., have been tested for the injection laryngoplasty for VFs augmentation. However, drawbacks of such materials like intractable complications, scarce bioactivity and insufficient durability, may affect the effect of treatment and hinder their clinical applications [[6], [7], [8]]. As we all known, the maintenance of innate extracellular matrix (ECM) components of the VFs is imperative for its physiological functions [9]. When the VF is paralyzed or the vibration is weakened, the ECM in the lamina propria (LP) may develop disuse atrophy, i.e., degradation of the innate ECM of the VFs. Hence, in addition to catering for the fundamental requirements such as injectability, biocompatibility and volume stability, the desired material for injection laryngoplasty should also be capable to enhance the synthesis of ECM in the VFs, which is a major hurdle clinically [10,11].

With the advance of biomaterial research, tissue-inducing materials have opened a new revenue for regenerative medicine [[12], [13], [14]]. Several promising strategies for ECM-inducing bioactivity have been proposed. One of these was to modulate the immune response of the host towards the materials. Nevertheless, prolonged inflammation during the filling process carries a risk for long-term adverse reactions, particularly granulomatosis [15]. Moreover, exogenous growth factors, as effective stimulator for tissue regeneration, have been immobilized on microspheres to directly regulate the remodeling of the ECM. However, they may increase patients’ financial burden and also have a short shelf-life. Above disadvantages have limited the clinical application of such tissue-inducing materials [16,17]. Therefore, to activate endogenous cells and stimulate ECM formation in a simple yet effective way may provide a better strategy for injection laryngoplasty. For their excellent biocompatibility, abundant source and ease of modification, natural polysaccharides such as carboxymethyl chitosan (CMCS) and sodium alginate (SA) have attracted a great interest in regenerative medicine recently [[18], [19], [20]]. Another noteworthy feature is its potential inherent bioactivity to stimulate tissue repair [21,22]. As reported, an alginate-based supramolecular hydrogel without additional growth factors has achieved accelerated wound healing in vivo by facilitating granulation tissue formation and fibroblast proliferation [23]. Similarly, CMCS, a soluble derivative of chitosan, also showed a positive effect on promoting cell proliferation and activation, cytokine secretion and vascularization [[24], [25], [26]]. Consequently, as demonstrated by many studies, natural polysaccharides have a great potential in inducing the synthesis of the ECM.

In this study, based on a dual-crosslinking strategy, a self-fused hydrogel comprised of CMCS and SA has been designed and prepared for the VF augmentation in a rabbit model for glottic insufficiency (Scheme 1). The results demonstrated that the hydrogel could effectively augment the volume of the VFs and, more importantly, enhance the synthesis of the ECM (e.g., collagen and elastin) both in vitro and in vivo. Furthermore, the acoustic functions of the VFs in the treatment group were better maintained compared with the control group (HA Gel) during the filling. The advantages of this self-fused hydrogel, including biocompatibility, outstanding durability, preservation of acoustic function, and particularly ECM-inducing bioactivity, make it a promising biomaterial for injection laryngoplasty.

Scheme 1.

a) Schematic illustration for the production of the self-fused hydrogel through a dual-crosslinking mechanism; b) The different treatment (hydrogel injection and sham treatment) of glottic insufficiency in a rabbit model; c) Effects of various treatments on LP of VFs during the filling process.

2. Materials and methods

2.1. Preparation of the hydrogels

Both the modified CMCS (CMCS-MA-Cat) and SA (ADA-PBA) were synthesized via two-step modification. Thereafter, based on the proportion of CMCS-MA-Cat and ADA-PBA, a series of hydrogels (C₂A₁, C₁A₁ and C₁A₂) were prepared, whose formation were triggered by pH value and 405-nm light. Here, the C₂A₁, C₁A₁ and C₁A₂ refer to the hydrogels with the volume ratios of CMCS-MA-Cat and ADA-PBA being 2 : 1, 1 : 1 and 1 : 2, respectively. The details are provided in the Supplementary Materials.

2.2. Physical properties of the hydrogels

The morphology and chemical structures of the C₂A₁, C₁A₁ and C₁A₂ hydrogels were characterized by scanning electron microscopy (SEM, EVO MA10, ZEISS, Germany) and X-ray photoelectron spectroscopy (XPS, Axis Ultra DLD, Kratos, UK), respectively. Specifically, the spectra for C 1s, N 1s, O 1s and B 1s in the XPS were further obtained and fitted using CasaXPS software by setting the hydrocarbon peak of C 1s as 284.6 eV. The swelling and degradation properties of the C₂A₁, C₁A₁ and C₁A₂ hydrogels were measured by the swelling ratio and weight remaining ratio, respectively. In addition, the rheological properties of the C₂A₁, C₁A₁ and C₁A₂ hydrogels were measured with a rotational rheometer (Physica MCR302, Anton Paar, Austria). The self-fused properties of the hydrogels, mainly comprised of self-healing and injectable properties, were validated by a variety of subjective and objective evaluation. The details are provided in the Supplementary Materials.

2.3. In-situ vibration and acoustic evaluation of the hydrogels

The acoustic evaluation of the C₂A₁, C₁A₁ and C₁A₂ hydrogels was accomplished by using a rabbit model (n = 3) for the VF injection under a high-speed photographic optical system. The sound produced by the vibration was collected by a voice recording equipment (Newsmy, V03), and the clinical parameters including the fundamental frequency (F₀), frequency perturbation (Jitter), amplitude perturbation (Shimmer) were analyzed. Subsequently, the VFs were carefully separated from the larynx, and the corresponding rheological properties were evaluated with a rotational rheometer (Physica MCR302, Anton Paar, Austria). The details are provided in the Supplementary Materials.

2.4. Cell isolation, identification and culture

Human VF fibroblasts (hVFFs) and human embryonic lung fibroblasts (MRC-5) were employed for the in vitro evaluation of the hydrogels. The acquisition of human VFs tissue in the clinics was approved by the Biomedical Ethics Review Committee (No.2022–800, West China Hospital, Sichuan University). The hVFFs were acquired with a previously described tissue block adherent method [27,28]. The details are provided in the Supplementary Materials.

2.5. Effect of the hydrogels on cell morphology, migration and invasion

Extracts of the C₂A₁, C₁A₁ and C₁A₂ hydrogels were prepared in complete medium (30 mg mL⁻¹) at 37 °C for 24 h. The influence of the hydrogels on the morphology of the hVFFs and MRC-5 was assessed by cytoskeleton staining, with cells cultured in a complete medium as the control. Moreover, the effect of C₂A₁, C₁A₁ and C₁A₂ hydrogels on the cell migration and invasion was evaluated with scratch test and Transwell assay, respectively. The details are provided in the Supplementary Materials.

2.6. ECM-inducing bioactivity of the hydrogels in vitro

After 3 and 7 days’ incubation with the hydrogels, total RNA of the hVFFs and MRC-5 was extracted and reversely transcribed into cDNA. The expression of type I collagen (COL I), type III collagen (COL III) and elastin was quantified by RT-PCR. The primers are listed in Table S1. After incubating for 3 days, type I and III collagen in the hVFFs and MRC-5 were further determined by immunofluorescence staining. High-throughput RNA sequencing (RNA-seq) was used to explore the changes in gene transcription of MRC-5 before and after the C₁A₁ hydrogel treatment. Differentially expressed genes (DEGs) were then subjected to enrichment analysis of Gene Ontology (GO) functions and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathways. The details are provided in the Supplementary Materials.

2.7. Biocompatibility of the hydrogels

Biocompatibility of the C₂A₁, C₁A₁ and C₁A₂ hydrogels was evaluated from the aspects of cytocompatibility, hemocompatibility and histocompatibility. The cytocompatibility of the hydrogels was measured with Cell Counting Kit-8 (CCK-8) assay, Live/Dead test and apoptosis assay. Their hemocompatibility was evaluated with a hemolysis test, whilst the histocompatibility was estimated by subcutaneous implantation in rats. All animal experiments have been approved by the Sichuan University Animal Care and Use Committee (No. 20220701002) and conformed to the Principles of Laboratory Animal Care formulation by The National Society for Medical Research. The details are provided in the Supplementary Materials.

2.8. Subcutaneous degradation behavior of the hydrogels in rats

The stability and degradation of the C₂A₁, C₁A₁ and C₁A₂ hydrogels in vivo was evaluated by subcutaneous injection assay as described above. To eliminate the difference between individual animals, each rat had received the injections (C₂A₁, C₁A₁, C₁A₂, HA Gel) in 4 regions on the back. The HA Gel (PureRegen® Gel Sinus, BioRegen Biomedical (Changzhou) Co., Ltd.) was used as the positive control. The residual hydrogels were assessed by ultrasonography and weighing. The details are provided in the Supplementary Materials.

2.9. Injection laryngoplasty in a rabbit model for paralyzed VFs

Male New Zealand white rabbits (weighing 2.5–3.0 kg each) were anesthetized with pentobarbital sodium (1.1 mL kg⁻¹) and placed in supine position. The C₁A₁ hydrogel was selected to injection laryngoplasty. The HA Gel was used as the positive control, and the blank group had received sham treatment. At 4, 8 and 12 weeks after the operation, the VFs were examined and photographed under a laryngoscope, with 3 rabbits per group at each time point. The details are provided in the Supplementary Materials.

2.10. Histological and functional evaluation of the VFs

After euthanasia, laryngeal tissues were sampled for H&E staining to observe the remaining material on the VFs. Masson staining and Sirius red staining were carried out to determine the types of collagen deposition and morphology around the VFs. COL I (bs-0578r, Bioss) and COL III (NB600-408, NOVUS) immunohistochemistry were carried out to assess the distribution of the collagen. Furthermore, Victoria blue staining was carried out by following the manufacturer's instruction (Solarbio, G1596, China) to observe the arrangement of elastic fibers in the tissue. The vibration of the VFs and the voice generated by the larynx were recorded with a high-speed camera system and voice recording equipment, respectively. The details are provided in the Supplementary Materials.

2.11. Statistical analysis

All data were presented as mean ± standard deviation. Statistical significance between the groups was determined by one-way ANOVA and Student's t-test. P < 0.05 was considered to be statistically significant (*P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001).

3. Results and discussion

3.1. Preparation and characterization of the hydrogels

The modified SA and CMCS were firstly synthesized through two-step reaction as shown in Supplementary Figs. S1a and S1b. Specifically, aldehyde and boronophenyl groups were introduced into the SA. The oxidation degree of the oxidized sodium alginate (ADA), which can reflect the content of aldehyde groups, was determined as 39% by a hydroxylamine hydrochloride method [29]. In Supplementary Fig. S1c, the peaks from 7 to 8 in proton nuclear magnetic resonance (¹H NMR) have illustrated successful grafting of phenylboric acid, and the grafting ratio of boronophenyl groups in the ADA-PBA was determined as 31%. Meanwhile, the CMCS was successfully grafted by the methacrylate and catechol groups, which were validated by ¹H NMR. The two separate peaks (δ = 5.54 and 5.76 ppm) and three separate peaks (δ = 6.72, 6.82 and 6.89 ppm) in Supplementary Fig. S1d were attributed to the protons in the methacrylate and catechol groups, with their grafting ratios being 23% and 17%, respectively.

The hydrogels were prepared through two steps based on a dual-crosslinking strategy (Fig. 1a). Firstly, under pH = 7.5, the dynamic covalent bonds network produced from the borate ester bonds and Schiff base bonds was preliminarily introduced into the hydrogel. The dynamic covalent bonds based on the click chemistry could endow the materials with self-healing, reversibility and unique mechanical properties, which has potential application value in many fields, especially in the case of high-frequency vibration and large deformations in the VFs [[30], [31], [32]]. Subsequently, under the excitation of 405 nm light, the C C bonds in the methacrylate group can polymerize, and the photo-triggered covalent bonds network can enhance the stability of the hydrogel, which is essential to the treatment of glottic insufficiency. According to their components, three hydrogels with the volume ratios of CMCS-MA-Cat and ADA-PBA being 2 : 1, 1 : 1 and 1 : 2 were prepared and named as C₂A₁, C₁A₁ and C₁A₂, respectively.

Fig. 1.

a) The scheme of the CMCS-MA-Cat and ADA-PBA, photographs of the mixed solution and formed hydrogels; b) SEM images and c) pore diameter of the C₂A₁, C₁A₁ and C₁A₂, Scale bar = 20 μm; d) The swelling curves, e) degradation curves and f) remain mass after 28 days' degradation of the hydrogels; N 1s spectra in XPS of g) C₂A₁, h) C₁A₁ and i) C₁A₂ (*P < 0.05, **P < 0.01, ****P < 0.0001).

As shown in Fig. 1b and c, the C₂A₁, C₁A₁ and C₁A₂ hydrogels all possessed microporous structure by scanning electron microscopy (SEM) with the pore diameter of 10.42 ± 1.77 μm, 8.76 ± 1.74 μm and 12.71 ± 2.13 μm, respectively. Thereinto, the C₁A₁ hydrogel has the densest microstructure and can form more crosslinking sites in its network compared with the other two. The swelling and degradation tests were carried out to reflect the stability of the hydrogels. Fig. 1d manifested the C₂A₁, C₁A₁ and C₁A₂ hydrogels have all absorbed water rapidly at the initial stage and reached a swelling equilibrium in about 30 h. The swelling ratio of C₁A₁ hydrogel was 25.97 ± 3.86% at 48 h, which was significantly lower than that of the other two. The degradation behavior of the C₂A₁, C₁A₁ and C₁A₂ hydrogels in the PBS solution for 28 days was monitored. As depicted in Fig. 1e and f, the remaining mass of the C₁A₁ hydrogel was 47.97 ± 6.37%, which was 1.56 and 1.86 times that of C₂A₁ and C₁A₂, respectively. Taken together, the C₁A₁ hydrogel has a more stable structure, which may be attributed to its difference in the crosslinking mechanism.

Alteration of the chemical bonds in the three hydrogels was determined with X-ray photoelectron spectroscopy (XPS), with an emphasis on the status of C, N, O and B. The C 1s, O 1s and B 1s spectra of the C₁A₁ hydrogel were presented in Supplementary Fig. S2. The deconvolution of C 1s spectrum has indicated the presence of C N (287.2 eV) in the hydrogel network, which can be attributed to the formation of Schiff base bonds. Typical peaks attributable to the B–O were detected in the O 1s and B 1s spectra located in 530.7 eV and 190.1 eV, respectively, which has confirmed the formation of borate ester bond. Furthermore, as shown in Fig. 1g–i, the peak-differentiating and imitating of N 1s spectrum for the C₂A₁, C₁A₁ and C₁A₂ hydrogels were carried out, and the peaks at 401.3 eV and 398.5 eV were ascribed to the C N and –NH₂. Interestingly, the proportion of C N in the C₁A₁ hydrogel (17.31%) was significantly higher than those in the C₂A₁ (13.97%) and C₁A₂ (13.95%) due to the formation of more Schiff base bonds in the hydrogel network. Correspondingly, the free amino group in the C₁A₁ hydrogel was reduced by 7.39% and 7.27% as compared with C₂A₁ and C₁A₂, respectively. A plausible explanation is that the C₁A₁ hydrogel has an appropriate ratio of CMCS-MA-Cat and ADA-PBA of 1 : 1, at which more Schiff bases bonds could be formed. This may also explain why C₁A₁ has the smallest pore size, the lowest swelling ratio and the slowest degradation rate among the three hydrogels.

3.2. Rheology and self-fusion of the hydrogels

Dynamic mechanical behaviors of the C₂A₁, C₁A₁ and C₁A₂ hydrogels were determined with a rheometer under different modes (Fig. 2a–c). Under the time sweep mode, the storage modulus (G′) of the three hydrogels without photocuring were higher than loss storage (G″). As shown in Supplementary Fig. S3, all hydrogels have possessed typical viscoelasticity. Before photocuring, the G′ of the C₁A₁ hydrogel was 2291 Pa, which was approximately 1.31 and 1.51 times that of C₂A₁ and C₁A₂, which indicated a more stable network owing to the optimized composition ratio. Excited by 405-nm light, the photo-triggered covalent bonds network has enhanced the dynamic mechanic properties of the hydrogels, with the G′ of the C₂A₁, C₁A₁ and C₁A₂ hydrogels being 4035 ± 205.2 Pa, 5928 ± 407 Pa and 3698 ± 552 Pa, respectively (Fig. 2a and Supplementary Fig. S4). Moreover, as revealed by Fig. 2b, the intact network of the hydrogels has been maintained under various frequencies (ranging from 1 Hz to 100 Hz). The strain amplitude sweep mode was used to evaluate the maximum strain of the hydrogels until breaking (G’’>G’), namely the strain at the flow point (τ_f). Fig. 2c showed the rheological curves of the C₂A₁, C₁A₁ and C₁A₂ hydrogels under an increasing strain, with the τ_f being 55.2%, 56.5% and 48.2%, respectively. As a mechanically dynamic tissue, the VF is characterized for its high-frequency (100 Hz) and small-amplitude (50%) oscillations during phonation [30,33]. Based on the above results, all hydrogels have owned noticeable dynamic mechanical properties and could maintain their integrity under these extreme conditions, showing a potential to be used for injection laryngoplasty.

Fig. 2.

Rheological characterization of the C₂A₁, C₁A₁ and C₁A₂ hydrogels under a) time sweep mode, b) frequency sweep mode, c) strain amplitude sweep mode, and d) alternate step strain sweep mode (low strain: 1%, high strain: 300%); e) Self-healing property of the dual dyed C₁A₁; f) The rheological curves of the hydrogels under shear rate sweep mode; g) Photograph of the dyed C₁A₁ injected through a syringe (25 G); h) The injectability and shape integrity evaluation of the dyed C₁A₁ injected into the S, C and U-shaped templates.

Considering the unique mechanical properties of the VFs, the self-fusion properties of the hydrogels including self-healing and injectability were investigated [34,35]. Firstly, as illustrated in Fig. 2d, the C₂A₁, C₁A₁ and C₁A₂ hydrogels have shown changes between reconstruction (G’ > G″, at 1% strain) and damage (G’ < G″, at 300% strain) under cyclic oscillatory strains, indicating that they are capable of self-healing. Two dyed hydrogels were put into touch for 5 min and could be picked up together with tweezers (Fig. 2e). Secondly, by rheological test under shear rate sweep mode (Fig. 2f), the C₂A₁, C₁A₁ and C₁A₂ hydrogels have possessed moderate viscosity, and the shear thinning test has signified their injectability. As shown in Fig. 2g, the C₁A₁ hydrogel could be injected into a bottle with a 25G syringe without dispersion in wet conditions. Thereafter, the C₁A₁ hydrogel was extruded and filled into the irregular S, C and U-shaped templates (Fig. 2h). Of note, the hydrogel has kept the shape of the templates after removal, illustrating obvious self-fusion. This property can be attributed to the unique dual-crosslinking network in the hydrogel. In addition, the dynamic covalent bonds network composed of Schiff base and borate ester bonds have conferred the hydrogels with self-healing property, making the hydrogel network to spontaneously reform and maintain integrity under significant deformation situations. Meanwhile, the C–C bonds network triggered by 405-nm light has enhanced the formability as well as the mechanical properties and stability of the hydrogels, which is of great significance for the long-term filling in the VFs. Hence, based on the dual-crosslinking strategy, the hydrogels coupled hardness with softness, which is essential to the treatment of glottic insufficiency.

3.3. Synergistic effect of the hydrogels on the physiological functions of the VFs

The mechanical properties of the VFs are essential to the vibration and vocalization [36]. In this study, the viscoelasticity of the VFs injected with the C₂A₁, C₁A₁ and C₁A₂ hydrogels was assessed by rheological test under the time sweep mode and frequency sweep mode (Supplementary Fig. S5). Compared with the normal VFs group, similar rheological curves were derived for the hydrogel groups, suggesting that they have preserved the original dynamic mechanical properties of the VFs.

The symmetrical vibration of VFs is a prerequisite for phonation [37]. In this study, the C₂A₁, C₁A₁ and C₁A₂ hydrogels were injected into the VFs in situ to assess their influence on the vibration capacity of the VFs with a high-speed digital camera system. As shown in Fig. 3a and Supplementary Movie S1, the glottic opening and closing was initially observed in serial vibration images in 4 ms using a normal VFs. Similarly, the VFs still vibrated normally after being injected with the C₁A₁ hydrogel. Fig. 3b showed the corresponding amplitude of the VFs, which suggested similar vibration periods and decreasing amplitude in the VFs after the injection. Moreover, compared with the normal group, the maximum and minimum area of the glottal gap in the C₂A₁, C₁A₁ and C₁A₂ groups were significantly decreased owning to the hydrogels’ augmenting effect (Fig. 3c and d). On the other hand, as an important parameter for the extent of glottic closure, the glottal gap of the hydrogel groups did not significantly differ from that of the normal group (P > 0.05) (Fig. 3e).

Fig. 3.

a) Serial images of the VFs vibration in the normal and C₁A₁ groups captured by a high-speed camera and b) analysis of the corresponding vibration wave amplitude; Analysis of c) maximum area, d) minimum area and e) glottal gap in the C₂A₁, C₁A₁ and C₁A₂ groups; Analysis of f) F₀, g) Jitter and h) Shimmer in the normal, C₂A₁, C₁A₁ and C₁A₂ groups (*P < 0.05, **P < 0.01, ***P < 0.001, ns: no significant difference).

Supplementary data related to this article can be found at https://doi.org/10.1016/j.bioactmat.2022.12.006.

The following is the supplementary data related to this article:

In this study, the voice generated by airflow was analyzed with a recording equipment and professional sound analysis system, Clinically, the F₀, Jitter and Shimmer were used as the key indicators to evaluate the voice. The F₀, defined as the lowest frequency of vibration, is related to the basic characteristics of VFs, e.g., length, mass and tension. Jitter and Shimmer are indexes of acoustic perturbation which can reflect the glottal irregularities during phonation [38,39]. As shown in Fig. 3f, the F₀ of the VFs injected with C₂A₁, C₁A₁ and C₁A₂ hydrogels were 250.37 ± 21.21, 259.48 ± 2.64 and 240.73 ± 7.98 Hz, respectively, all being significantly lower than that of the normal VFs (P < 0.001). The hydrogels have increased the mass of the VFs, causing a decrease of the F₀. As revealed in Fig. 3g and h, the Jitter of the C₂A₁, C₁A₁ and C₁A₂ hydrogels have measured 5.88, 5.47 and 5.28 times that of the normal VFs (0.26 ± 0.12), respectively. Similarly, the shimmer of the C₂A₁, C₁A₁ and C₁A₂ hydrogels was 1.81, 1.79 and 1.78 times that of the normal VFs (6.74 ± 2.25), respectively. In total, after the injection of targeted materials, the F₀, Shimmer and Jitter would be of instant changes because of the increasing of weight of vocal fold and subsequent subglottic pressure; this was in good consistency with previous report, where hyaluronate injection has improved the subglottal pressure whilst improved the Jitter [40]. Taken together, as verified by various functional assessment, the C₂A₁, C₁A₁ and C₁A₂ hydrogels have all shown a synergistic effect on the VFs functions. This is due to the hydrogels have been shown to own unique mechanical properties endowed by the dual-crosslinking network. The hydrogels exhibited viscoelasticity similar to VFs and could effectively assist the transmission of mucosal waves.

3.4. ECM-inducing bioactivity of the hydrogels

As we all know, the time of sound off is usually 2 weeks after the injection laryngoplasty required by most laryngologists in clinic, and this is another bottleneck of all materials at present. Thus, how to shorten the time of sound off is a hurdle that are urgent to settle. The voice parameters of F₀, Shimmer and Jitter are closely related to the physiological functions which are mostly determined by the innate ECM components of the VFs. Accordingly, the desirable material for the injection laryngoplasty should also be able to enhance the synthesis of ECM and possess suffice biocompatibility to the VFs, which in turn may shorten the time of sound off after the injection laryngoplasty. Therefore, we have first verified the in vitro ECM-inducing bioactivity of the hydrogels.

Propagation of the mucosal wave during the phonation is related to the LP of the VFs and consisted of interstitial fluid, collagen (∼43.4%), elastin (∼8.5%) and other components. Fibroblasts, capable of secreting the ECM, e.g., collagen, elastin and so on, play an important role in the VFs. Accordingly, we have used the hVFFs and MRC-5 to evaluate the ECM-inducing ability of the hydrogels. The hVFFs, isolated from human VFs, were identified by immunofluorescent staining of E-cadherin, Vimentin and α-SMA (Supplementary Fig. S6). The effects of the C₂A₁, C₁A₁ and C₁A₂ hydrogels on the morphology of the hVFFs through cytoskeleton staining and corresponding high-content analysis were illustrated Supplementary Figs. S7a–c, which have indicated a significant increase in the cell area and perimeter. In particular, the area of the hVFFs has differed significantly before and after the incubation with the hydrogels (P < 0.05), with the normal, C₂A₁, C₁A₁ and C₁A₂ groups being 1453 ± 127.0 μm², 2427 ± 183.1 μm², 2459 ± 235.7 μm² and 2218 ± 272.8 μm², respectively. Similar changes were also observed in the MRC-5 (Supplementary Figs. S7d–f). In indeed, the increased area of the fibroblasts indicated activation of cellular functions, which has an important impact on wound healing [41].

As shown by previous studies, collagen and elastin secreted by the fibroblasts play a vital role in phonation. Collagen as an important component will provide strength to the VFs tissue. The collagens of the VFs are comprised type I and type III, where COL III is mainly found in the LP. Elastin holds the pliability and elasticity in VFs, and has a great influence on its function [13,33]. As shown in Fig. 4a and b and Supplement Fig. S8, assessment of the expression of type I collagen alpha 1 chain (COL1A1), type III collagen alpha 1 chain (COL3A1) and elastin in the hVFFs and MRC-5 by RT-PCR has demonstrated markedly up-regulated expression in the hydrogels-treated cells. Thereinto, the COL1A1, COL3A1 and elastin expression in the hVFFs treated by the C₁A₁ hydrogel for 7 days were up-regulated by 4.14, 2.83 and 2.01 times that of the normal cells, respectively. In addition, based on the immunofluorescence images in Fig. 4c, the expression of COL1A1 and COL3A1 were remarkably enhanced by the C₂A₁, C₁A₁ and C₁A₂ hydrogels. In keeping with this, as shown in Supplementary Fig. S9, enhanced collagen expression in the MRC-5 were also observed after the incubation with the hydrogels.

Fig. 4.

Expression of a) COL1A1 and b) COL3A1 in the hVFFs and MRC-5 after 3 and 7 days' treatment evaluated by RT-PCR; c) COL1A1 and COL3A1 immunofluorescence staining of the hVFFs after 3 days' treatment, Scale bar = 100 μm; d) GO function and e) KEGG pathway enrichment analysis of the DEGs in MRC-5 (C₁A₁ group vs control group) (*P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001, ns: no significant difference).

The effect of the hydrogels on the migration of the hVFFs was assessed with the Transwell migration and scratch assays. Presented in Supplementary Fig. S10, after incubation with the hydrogels, the cells have crawled faster towards the scratch area with the elapse of time, with the migration rates at 36 h being 46.07 ± 6.34%, 53.91 ± 2.83%, 64.54 ± 2.18% and 62.08 ± 8.16% for the control, C₂A₁, C₁A₁ and C₁A₂ groups, respectively. Similarly, more hVFFs have passed through the Transwell membrane when stimulated by the hydrogels for 24 h compared with the complete medium group (Supplementary Fig. S11). Collectively, hydrogels not only recruit the resident fibroblasts around the tissue, but also promote the synthesis and secretion of ECM-related proteins (COL I, COL III and elastin), enhancing the accumulation of ECM of the VFs.

Furthermore, the transcriptome analysis was employed to validated the effect of hydrogels on biological functions of MRC-5. As shown in Fig. 4d, many GO terms associated with cellular component were specifically enriched, such as collagen-containing ECM and extracellular region/space. As to molecular function, growth factor activity and positive regulation of cell migration were enriched, showing bioactive C₁A₁ hydrogel could improve cell proliferation, which is consistent with the previous results. Moreover, the cytokine-mediated signaling pathway in GO analysis means a series of molecular signals produced by cytokines binding to cell receptors may play a role in biological functions. Through KEGG pathway enrichment analysis in Fig. 4e, some classical signaling pathways were identified including the transforming growth factor-β (TGF-β), mitogen activated protein kinases (MAPK) and phosphoinositide 3-kinase/Akt (PI3K/Akt) signaling pathways, which have been reported to be associated with cytokines and ECM secretion [32,42,43]. Briefly, the RNA-Seq results verified the role of hydrogel in promoting ECM secretion and cell migration, and suggested that this role may be related to cytokines. Furthermore, Huang et al. demonstrated that chitosan could enhance wound healing via direct fibroblast activation and cytokine production [24]. And some studies have verified that polysaccharides can indirectly affect collagen secretion of fibroblasts by influencing the phenotype of immune-related cells such as macrophages [7,25]. Therefore, it may be speculated that the pro-secretion effect of the C₁A₁ hydrogel on ECM-related proteins may be due to autocrine effect of cytokines produced by fibroblasts and/or paracrine effect of cytokines produced by other cells (such as macrophages).

3.5. In-vitro and in-vivo biocompatibility of the hydrogels

Regarded as the most essential characteristics, in vitro and in vivo biocompatibility of the biomaterial have the core of the research, especially for the VFs which require better durability and volume stability [44]. As verified by the CCK-8 assay, the proliferation of the hVFFs and MRC-5 showed no significant difference from the control (Fig. 5a and Supplementary Fig. S12a). Live/Dead fluorescence images indicated that the hVFFs and MRC-5 under vibrant condition had a long-spindle shape and almost no dead cells after incubating with the C₂A₁, C₁A₁ and C₁A₂ hydrogels for 1, 3 and 5 days (Fig. 5b, Supplementary Figs. S12b and S13). As measured by flow cytometry, the apoptosis rates of the hVFFs in the C₂A₁, C₁A₁ and C₁A₂ groups were 0.21 ± 0.06%, 0.21 ± 0.10% and 0.20 ± 0.10%, respectively, which showed no significant differences from the control group (Supplementary Fig. S14). Furthermore, by the hemolysis test, the C₂A₁, C₁A₁ and C₁A₂ hydrogels have all shown sound hemocompatibility with almost no rupture of erythrocytes, with the hemolysis rate of all groups being less than 5% (Fig. 5c and Supplementary Fig. S15). The erythrocytes on the surface of the hydrogels was also observed with SEM. As shown in Supplementary Fig. S16, the erythrocytes all had the typical double concave disk shape and scattered over the hydrogels.

Fig. 5.

a) The effect of the C₂A₁, C₁A₁ and C₁A₂ hydrogels on the proliferation of the hVFFs; b) Live/Dead staining images of the hVFFs treated by the hydrogels for 5 days, Scale bar = 500 μm; c) Hemolysis ratio of the hydrogels (NC: PBS, PC: 0.1% Triton X-100); d) H&E staining of the hydrogels after 7 days of implantation, Scale bar = 200 μm; e) CD86 immunofluorescence staining and f) CD86⁺/DAPI relative area of the hydrogels after 1 and 2 weeks of implantation, Scale bar = 200 μm (*P < 0.05, ns: no significant difference).

During subsequent in vivo experiment, the results of H&E staining after subcutaneous injection of the C₂A₁, C₁A₁ and C₁A₂ hydrogels showed that the hydrogels were closely connected with the around tissue, which indicated good integration without forming obvious capsule layer (Fig. 5d). Moreover, 7 days after the subcutaneous injection, a large number of inflammatory cells have infiltrated into the hydrogels, suggesting the occurrence of typical inflammatory reaction, in which macrophages play a key role. Among these, the M1-subtype macrophages (CD86⁺) have a pro-inflammatory effect [45]. Hence, CD86 immunofluorescence staining was conducted to assess the changes of inflammatory response over time. As shown in Fig. 5e and f, CD86⁺ stained area has obviously decreased from day 7 to day 14, which indicated a reduced inflammation response, albeit that no significant difference was found between the C₂A₁, C₁A₁ and C₁A₂ groups. The liver and renal functions of the rats were assessed by blood biochemical and routine tests. As shown in Supplementary Fig. S17, the indexes of the C₂A₁, C₁A₁ and C₁A₂ groups did not differ significantly from those of the control group (P > 0.05). The influence of the hydrogels on the major organs (heart, liver, spleen, lung and kidney) were explored by H&E staining (Supplementary Fig. S18). All organs had normal structural characteristics without pathological changes after 7 days of hydrogel implantation. By in vitro and in vivo evaluations, the C₂A₁, C₁A₁ and C₁A₂ hydrogels have all shown suffice biocompatibility and have not induced excessive immune response.

3.6. Subcutaneous filling effectiveness of the hydrogels in rats

As shown in Fig. 6a, the filling effectiveness of the C₂A₁, C₁A₁ and C₁A₂ hydrogels was evaluated in the rat model for subcutaneous injection by gross observation, ultrasonography, and histology. To minimize the discrepancy between individual animals, each rat was implanted with four materials in the back, with the HA Gel as the control. A remarkably difference in the gross appearance (Fig. 6b) was noted between the C₁A₁ and HA Gel groups. Two weeks after the filling, the C₁A₁ hydrogel still maintained its spherical shape and original filling volume, whilst the HA Gel was degraded and almost absorbed by the host, leaving only a tiny volume subcutaneously. The remaining volume of the C₁A₁ hydrogel was almost the same after 4 and 6 weeks of filling, indicating that it has stabilized. By contrast, no subcutaneous substance was observed in the HA Gel group. Similar result was derived from histological examination. As shown in Supplementary Fig. S19, the hydrogels infiltrated by cells were easily observed by H&E staining after 2, 4 and 6 weeks, whilst little HA Gel has left 2 weeks after the implantation.

Fig. 6.

a) Scheme of the operation and assessment of subcutaneous degradation of the hydrogels in vivo; b) Gross appearance of the C₁A₁ and HA Gels after 2, 4 and 6 weeks of subcutaneous implantation; c) The remaining mass of the C₂A₁, C₁A₁ and C₁A₂ hydrogels after the subcutaneous implantation; d) Ultrasonography of the hydrogels after the subcutaneous implantation; e) The width of the hydrogels after subcutaneous injection; f) The thickness for various hydrogels after subcutaneous implantation (at 0, 2, 4 and 6 weeks) (*P < 0.05, **P < 0.01, ns: no significant difference).

All of the three hydrogels have shown approving filling results, but with some difference. According to the photographs in Fig. 6b and Supplementary Fig. S20, although the C₂A₁ and C₁A₂ hydrogels have attained the similar filling effect as the C₁A₁ hydrogel by week 2, it was obvious that with the elapse of time, they have shown a faster degradation rate compared with the C₁A₁ hydrogel, which resulted in poorer filling. Fig. 6c depicted that the remaining mass of the C₁A₁ hydrogel was greater than the other two hydrogels, with 55.81 ± 10.03% of the initial weight retained by week 6. The maintained hydrogel was monitored by the means of ultrasound at weeks 2, 4 and 6. In Fig. 6d, there was a obvious boundary between the hydrogels and the surrounding tissue, which may be used to determine the thickness and width of the hydrogels. As shown in Fig. 6e, the implant width of the C₂A₁, C₁A₁ and C₁A₂ hydrogels immediately after the injection was significantly smaller than that of the HA Gel, suggesting that the HA Gel has a tendency to spread out rather than condense. This is because that the dual-crosslinking networks (dynamic covalent bonds network and photo-triggered bonds network) in the hydrogels have conferred them with a self-fusing property, allowing them to maintain the shape after the injection. Furthermore, the implant thickness among the four groups was analyzed during the filling process (Fig. 6f). Immediately after the injection, the thickness of the C₂A₁, C₁A₁ and C₁A₂ implants was 1.83, 1.77 and 1.67 times that of the HA Gel (0.30 ± 0.02 cm), respectively, which suggested an excellent augmentation effect. The difference in the filling effectiveness of the hydrogels was also determined with the thickness of the implant. Six weeks after the injection, the thickness of the C₁A₁ hydrogel was 0.27 ± 0.05 cm, which has retained 51.6% of the initial value and was 1.45 and 2.04 times that of C₂A₁ and C₁A₂ hydrogels, respectively. Above results have indicated that the C₁A₁ hydrogel has attained the best filling effect in the rat model, and this was in keeping with the results of the SEM and the swelling and degradation tests, suggesting that the C₁A₁ hydrogel has the optimum stability. As discussed above, the volume ratio of modified CMCS and SA in the C₁A₁ hydrogel (1 : 1) have resulted in formation of more Schiff base bonds in the hydrogel network. Hence, considering the above results, the C₁A₁ hydrogel was identified as the most suitable among the three, and was selected for treating the glottic insufficiency in the rabbit model.

3.7. Effect of the hydrogels on glottic insufficiency in rabbit

In view of the similar anatomical structure, rabbits were used to construct a glottic insufficiency model by the vagectomy. As shown in Fig. S21 and Supplementary Movie S2, the surgical procedures included location and dissection of the vagus nerve, and endoscopic examination of the VFs vibration and injection. The paralysis of VFs under endoscope has confirmed successful modeling. The HA Gel, which was approved for clinical treatment of glottic insufficiency, was used as the control [46]. As shown in Fig. 7a, obvious augmentation was noted on the right hand VFs (the modeling side) immediately after the injection of the C₁A₁ and HA Gel. During the 12 weeks' consecutive examination, the glottic gap of the C₁A₁ injected VFs was significantly smaller than the blank and HA Gel groups, suggesting a lasting augmentation effect, whilst decomposition, displacement and other adverse events of the material were not noted. Notably, in the blank group, the right VFs showed visible atrophy compared with the opposite side, which indicated degradation of intrinsic tissue. However, this was not observed in the other two groups. By H&E staining (Fig. 7b), the remainder hydrogel could be observed over the 12 weeks’ treatment period, when the HA Gel had completely disappeared. Moreover, many cells have infiltrated into the interior of the C₁A₁ hydrogel, and the material became closely connected to the surrounding tissues without fibrous capsules, which suggested excellent biocompatibility and tissue integration.

Fig. 7.

a) Laryngoscopy of the VFs in the blank, HA Gel, and C₁A₁ groups before and 0, 4, 8 and 12 weeks after the injection (Yellow dotted line: remaining material); b) H&E staining of the VFs 4, 8 and 12 weeks after the HA Gel and C₁A₁ injection, Scale bar = 100 μm; c) Serial images of the VFs vibration in the normal, blank, HA Gel, C₁A₁ groups as captured by a high-speed camera; Analysis of d) maximum area, e) minimum area, f) Jitter and g) Shimmer in each group (*P < 0.05, **P < 0.01, ***P < 0.001, ns: no significant difference).

Supplementary data related to this article can be found at https://doi.org/10.1016/j.bioactmat.2022.12.006.

The following is the supplementary data related to this article:

To assess the effect of various treatments on the vibratory capacity, vibration of the VFs was recorded using a high-speed imaging system (Supplementary Movie S3), and the result (Fig. 7c) has indicated the mucosal wave of the VFs treated by the C₁A₁ hydrogel for 12 weeks was similar to that of the normal VFs. The maximum area in the C₁A₁ group have measured 70.0% and 68.8% that of the blank and HA Gel groups, respectively, but had no significant difference from that of the normal group (P > 0.05) (Fig. 7d). Similar results were obtained for the minimum area and glottal gap (Fig. 7e and Supplementary Fig. S22), suggesting that the vibration function of the VFs has been maintained after the 12-week treatment, which is a prerequisite for normal phonation. Furthermore, by evaluating the acoustic function of the VFs, the Jitter and Shimmer in the blank group were higher than that of the normal VFs, which demonstrated a rougher and hoarse voice, consisting with the clinical symptoms of glottic insufficiency. After injected with the C₁A₁ hydrogel for 12 weeks, as shown in Fig. 7f and g, the Jitter and Shimmer were reduced to the same levels as the normal VFs (0.98 and 9.18, respectively; P > 0.05). Similarly, the F₀ (283.36 Hz) of the C₁A₁ group was 0.62 and 0.81 times that of the blank and HA Gel groups, respectively, suggesting that the sound quality was unaffected. Taking together, compared with the HA Gel, the C₁A₁ hydrogel showed a lasting augmentation effect, as well as a better reservation of the physiological functions (vibration and phonation) of the VFs.

Supplementary data related to this article can be found at https://doi.org/10.1016/j.bioactmat.2022.12.006.

The following is the supplementary data related to this article:

To further assess the influence of the hydrogels on the ECM of the VFs, the collagen deposition around the materials in the LP was evaluated. By Masson staining (Fig. 8a), collagen fibers (blue area) could be clearly seen in the HA Gel and C₁A₁ groups, whilst the LP in the blank group was loosely organized due to the degeneration. As shown in Fig. 8b, the collagen volume fraction (CVF) of the VFs after the 12-week treatment of C₁A₁ has reached 50.79 ± 6.73%, whilst those of the blank and HA Gel groups had measured only 14.96 ± 1.24% and 34.69 ± 2.08%, respectively. In keeping with this, among the three groups, the C₁A₁-treated VFs by the 12th week have presented the thickest LP (723.8 ± 91.8 μm) (Supplementary Fig. S23). Furthermore, as shown in Fig. 8c, the proportion and distribution of various types of collagen in the VFs were determined by Sirius red staining. Under polarized light, denser collagen fibers were observed in the C₁A₁ group compared with the other two groups. As shown in Supplementary Fig. S24, the content of COL III in the C₁A₁ group was 1.73 times that of the HA Gel group, while the COL I was 0.69 times that of the HA Gel group. Meanwhile, the content of COL I in the VFs after C₁A₁ treatment was 0.39 times that of COL III (Fig. 8d). Similar results were noted with immunohistochemical staining of COL I and COL III. As shown in Fig. 8e, significantly more expression of COL I and COL III in the LP was observed in the C₁A₁ group compared with the HA Gel group. By contrast, the collagens were loosely arranged in the LP, which suggested disuse atrophy of the VFs by endoscopy. Furthermore, specific staining by Victoria blue dye has indicated that the elastin in the LP of the C₁A₁ group were denser compared with the other groups (Fig. 8f), which suggested that the viscoelasticity of the VFs has been reserved, which is essential for the normal function of the VFs.

Fig. 8.

a) Masson staining and b) CVF in the blank, HA Gel and C₁A₁ groups after the 12-week injection (LP: lamina propria; ML: muscular layer); c) Sirius red staining and d) COL I/COL III in each group after the 12-week treatment; e) immunohistochemistry of COL I and COL III and f) Victoria blue staining of the VFs in the blank, HA Gel and C₁A₁ groups after the 12-week treatment, Scale bar = 50 μm (**P < 0.01, ***P < 0.001, ns: no significant difference).

In conclusion, the above results have confirmed that the C₁A₁ hydrogel had enhanced the secretion of collagen and elastin in the VFs and maintained the normal composition of the LP, which may also explain its excellent filling effect and physiological function preservation. The preferable durability of the C₁A₁ hydrogel may be attributed to two factors. First, owing to the dual-crosslinking mechanism, the self-fusion property and stability of the C₁A₁ hydrogel has enabled it to more effectively resist the degradation by various enzymes, e.g., collagenase, hyaluronidase and lipase, in the VFs tissue. Second, as verified by in vitro and in vivo experiments, the secretion of ECM proteins enhanced by the C₁A₁ hydrogel containing bioactive components, e.g., CMCS and SA, therefore can avoid the degradation of LP tissue in the VFs.

In addition to outstanding durability, the C₁A₁ hydrogel also sustained the vibration of the VFs and preserved its acoustic functions, which is the ultimate goal for clinical treatment. On the one hand, the effective filling has allowed heterolateral VFs to contact facilely, and it is proved that the hydrogel itself does not hinder the phonation. On the other hand, as mentioned above, the ECM of LP determines the biomechanics of the VFs, and is crucial for phonation. In a simple but effective way, the C₁A₁ could significantly upregulate the expression of COL I, COL III and elastin, ensuring the integrality of the LP during the long-term treatment. Taken together, the novel self-fuse hydrogels, in particular C₁A₁, hold great promise for clinical applications to treat glottic insufficiency through enhancing materials durability, the ECM secretion, and preservation of acoustic functions of the VFs.

4. Conclusion

In this study, a series of hydrogels containing modified CMCS and SA have been produced through a dual-crosslinking mechanism (dynamic covalent bonds and photo-triggered covalent bonds) which has endowed the hydrogels with enhanced self-fusion and stability. By evaluating the viscoelasticity, vibration and acoustics of the VFs after injection, the synergistic effect of the hydrogels on the VFs physiological functions was demonstrated. Notably, owing to the inherent bioactivity of the hydrogels, more ECM-associated proteins including COL I, COL III and elastin have been synthesized, which, as verified by both in vivo and in vitro experiments, were 4.06, 2.76 and 2.94 times that of the control after the treatment with C₁A₁ for 7 days in the hVFFs, respectively. Furthermore, the hydrogels showed reliable filling effectiveness and biocompatibility in a rat model for subcutaneous injection. In the rabbit model for injection laryngoplasty, the C₁A₁ attained superior augmenting effect after a longer time (12 weeks) compared with the commercial HA Gel. More importantly, owing to the enhanced secretion of ECM by the hydrogel, C₁A₁-treated VFs acquired physiological functions (vibration and phonation) indistinguishable from the normal, while this was not observed with the HA Gel. Conceivably, dispensed with exogenous active substances, our simple yet effective strategy to promote ECM synthesis and reserve the acoustic function of the VFs may be of great value for the design of novel materials for the treatment of glottic insufficiency.

CRediT authorship contribution statement

Chen-Yu Zou: Formal analysis, Methodology, Writing – original draft. Juan-Juan Hu: Formal analysis, Methodology, Writing – original draft. Dan Lu: Methodology, Formal analysis. Qian-Jin Li: Methodology, Formal analysis. Yan-Lin Jiang: Methodology, Formal analysis. Rui Wang: Methodology, Formal analysis. Hai-Yang Wang: Methodology, Formal analysis. Xiong-Xin Lei: Methodology, Formal analysis. Jesse Li-Ling: Formal analysis, Writing – original draft. Hui Yang: Conceptualization, Formal analysis, Writing – original draft, Funding acquisition. Hui-Qi Xie: Conceptualization, Formal analysis, Writing – original draft, Funding acquisition.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationship which could have appeared to influence the work reported in this paper.

Appendix A. Supplementary data

The following are the Supplementary data to this article: