M/G‐Ratio‐Tuned Calcium Alginate Hydrogel Microcarrier for Controlled Delivery of Adeno‐Associated Virus for Gene Therapy

ABSTRACT

This paper proposes the controlled release of adeno‐associated virus (AAV) by polymer mesh structures of alginate hydrogel microbeads (gel‐beads). The polymer mesh structure of alginate hydrogel can be controlled by adjusting the number of cross‐linking points depending on the molecular structure (M/G ratio) of the alginate hydrogel. Scanning electron microscopic analysis on the mesh structures of alginate hydrogels with M/G ratios of 0.96 (MG‐0.96) and 0.42 (MG‐0.42) indicates that the mesh size distribution of MG‐0.42 hydrogel was smaller than that of MG‐0.96 hydrogel. Using these characteristics, controlled release of AAVs (AAV‐1 or AAV‐5) from M/G‐ratio‐tuned gel‐beads (MG‐0.96 or MG‐0.42), Janus‐shaped gel‐beads consisting of hemispheres of the MG‐0.96 and MG‐0.42, and MG‐0.64 gel‐beads consisting of a mixture of MG‐0.96 and MG‐0.42 were examined with ELISA for AAV capsids and the GFP fluorescent signals from the transfected cells. The precise control of AAV release by the mesh structure of the alginate hydrogels could be an effective method for creating AAV micro‐carriers for gene therapy.

1. Introduction

Gene therapy is a method of treating diseases by introducing genes to recover normal function, to decrease mutant gene expression, or to edit genes (Anguela and High 2019). This therapy has been expected to treat intractable or genetic diseases such as hearing loss and retinal diseases (Han et al. 2023; Isgrig et al. 2023). For efficient gene transfer into cells, vectors incorporating genes are essential (Giacca and Zacchigna 2012). In particular, gene therapy using adeno‐associated virus (AAV) vector has attracted attention because of its non‐pathogenic and stable gene expression (Gonçalves 2005; Mendell et al. 2021). Previous studies have confirmed the restoration of hearing after AAV administration by injection into the inner ear of mice treated with gene knockout (Tao et al. 2022) and children with autosomal recessive hearing loss (Lv et al. 2024). Furthermore, gene treatment using AAV has also been effective for long‐term expression of factors in hemophilia, and improvement of retinal function and suppression of degeneration in mice with retinal degeneration (George et al. 2021; Karali et al. 2020).

For using AAV in vivo, it is necessary to suppress immune responses and protect AAV from clearance in the body (Alvarez‐Rivera et al. 2020; Wen et al. 2000). For these purposes, encapsulating AAV in a carrier to control the release of AAV from the carriers is effective (Kim et al. 2012). Hydrogels, such as 2‐hydroxyethyl methacrylate and chitosan used in contact lenses and artificial skin, respectively, have been popular materials as the carriers of AAV (Alvarez‐Rivera et al. 2020; Cheng et al. 2023). Among such hydrogels, alginate hydrogels have been studied as carriers because of their biocompatibility and suitability for medical use (Szekalska et al. 2016; Remes et al. 2021; Ruvinov and Cohen 2016). In a previous study, encapsulation of AAVs in alginate hydrogel could fix AAV locally in vivo (Madrigal et al. 2019). On the other hand, the size of AAV is smaller than the polymer mesh structure of the alginate hydrogel, causing AAV to leak out by diffusion. To solve this problem, poloxamer was added to control the release of AAV by changing the polymer mesh structure according to temperature (Diaz‐Rodriguez et al. 2015). However, the addition of other materials complicates the system in several ways, including preparation and application for clinical use.

Here, we propose a method to control AAV release by focusing on the molecular composition of alginate hydrogels to adjust the polymer mesh structure determined by cross‐linking density (Figure 1). Alginate is generally extracted from natural marine sources (Chandía 2001; Fenoradosoa et al. 2010). Extracted alginate polymers are composed of two monomers, mannuronic acid (M) and guluronic acid (G) (Agulhon et al. 2012) (Figure 1a). The ratio of these components is called the M/G ratio. In the alginate polymer, the divalent cations can be highly coordinated only between G‐sections to form a three‐dimensional gel network (Turco et al. 2011; Liu et al. 2016) (Figure 1b). Therefore, the increase in G‐sections in alginate (low M/G ratio) leads to the increase in the cross‐linking points of the alginate gel network (C. Hu et al. 2021), resulting in decreasing the polymer mesh size in the gel network to control AAV release ratio (Figure 1c). In this study, two types of alginate hydrogel microbeads (gel‐beads) (~200 µm in diameter) with different M/G ratios available to us (MG‐0.96: M/G ratio 0.96, MG‐0.42: M/G ratio 0.42) were prepared to control AAV release by using this property (Figure 1d). While these types of gel‐beads are effective in controlling AAV release, obtaining alginate with a specific M/G ratio remains challenging due to its nature as a natural material. To address this limitation and expand the strategies for controlling AAV release, we propose two additional types of alginate hydrogel microbead carriers: Janus‐shaped hydrogel microbeads composed of MG‐0.96 and MG‐0.42 hemispheres, and MG‐0.64 gel‐beads obtained by mixing MG‐0.96 and MG‐0.42. These designs complement the properties of MG‐0.96 and MG‐0.42, providing enhanced flexibility and precision in regulating AAV release.

Figure 1.

Concept of controlled release of adeno‐associated virus (AAV) by alginate hydrogel microbeads (gel‐beads) using M/G ratio for gene therapy. (a) Molecular structure of alginate and (b) cross‐linking points. (c) Cross‐linking points and mesh structure of the alginate hydrogel. (d) Alginate hydrogel microbeads with different mesh structures encapsulating AAV.

2. Results

2.1. Observation of the Mesh Structure in Alginate Hydrogels

To control the AAV release from alginate hydrogel microbeads, we focused on the relationship between the M/G ratio of alginate and the size of the polymer mesh structure. To investigate the relationship experimentally, the mesh structures of alginate hydrogels with the M/G ratio of 0.96 (MG‐0.96) and 0.42 (MG‐0.42) were analyzed using a scanning electron microscope (SEM). The two different alginate hydrogels used in this study (MG‐0.96 and MG‐0.42, 2% [w/v], cross‐linked with 100 mM CaCl₂ aq.) were cut to obtain the cross‐section of the alginate hydrogels for SEM observation. In this experiment, freeze‐dried MG‐0.96 and MG‐0.42 were prepared because samples need to be dehydrated to be observed by SEM (Fischer et al. 2024; Santana et al. 2015). In other words, the hydrogels used for SEM observation were dehydrated and differed from the hydrogels actually used as AAV release carriers. It is known that the mesh of the hydrogel in a dehydrated state shrinks or does not maintain its natural state (Waje et al. 2005; Tomlins et al. 2004). Therefore, this experiment was designed to compare the relative mesh sizes of MG‐0.96 and MG‐0.42.

The obtained SEM images show the mesh structures formed in the MG‐0.96 and MG‐0.42 alginate hydrogels (Figure 2a). By using image analysis (ImageJ), the size distributions of the pores in the mesh structures were quantified (Figure 2b). The averaged pore sizes of the mesh structures were 1.43 × 10⁻⁹ mm² (diameter assuming a regular circle R _assumption = 42.7 nm) in the MG‐0.96 hydrogel and 5.01 × 10⁻¹⁰ mm² (R _assumption = 25.3 nm) in the MG‐0.42 hydrogel. In addition, the number of pores larger than 4.0 × 10⁻⁹ mm² (R _assumption = 22.7 nm) is higher in MG‐0.96 hydrogel than that in MG‐0.42 hydrogel. On the other hand, the number of pores smaller than 4.0 × 10⁻¹⁰ mm² is higher in MG‐0.42 hydrogel than that in MG‐0.96 hydrogel. Furthermore, the average mesh size range of the MG‐0.96 and MG‐0.42 hydrogels was determined (Figure 2c). The range of mesh size in MG‐0.96 was 9.7 × 10⁻¹⁰–1.9 × 10⁻⁹ mm² (R _assumption = 35.1–49.2 nm), and in the case of MG‐0.42, it was 3.7 × 10⁻¹⁰–9.6 × 10⁻¹⁰ mm² (R _assumption = 21.7–35.0 nm). Therefore, the average mesh size range of MG‐0.42 was smaller than that of MG‐0.96, confirming a significant difference. These results indicate that the MG‐0.42 hydrogel has smaller mesh structures than the MG‐0.96 hydrogel.

Figure 2.

Mesh structures of 2% alginate hydrogel. (a) SEM images of cross‐section of MG‐0.96 hydrogel and MG‐0.42 hydrogel (scale bars: 500 nm). (b) Mesh size distributions of MG‐0.96 hydrogel and MG‐0.42 hydrogel (red: MG‐0.96, green: MG‐0.42). (c) The range of average mesh sizes in MG‐0.96 hydrogel and MG‐0.42 hydrogel (red: MG‐0.96, green: MG‐0.42). *p < 0.05. The horizontal axis of (b) and (c) represents the area of the mesh (mm²) and its diameter, assuming the mesh is a regular circle (nm).

2.2. Fabrication of Alginate Hydrogel Microbeads (Gel‐Beads)

Using the MG‐0.96 and MG‐0.42 hydrogels above, we produced two types gel‐beads by a microfluidic device using centrifugal force (Maeda et al. 2012; Takatsuka et al. 2023) (Figure 3a): The first one is single‐aspect gel‐bead consisting of one type of alginate hydrogel (MG‐0.96, MG‐0.42, or MG‐0.64), and the second one is Janus‐shaped gel‐bead consisting of hemispheres of two types of alginate hydrogels (MG‐0.96 and MG‐0.42). The fabrication process of the gel‐beads is as follows: Pre‐gel solutions were prepared with 2% (w/v) Na‐Alg. aq. with 1% (v/v) 200‐nm‐diameter fluorescent polystyrene particles (red: MG‐0.96, green: MG‐0.42). Subsequently, the pre‐gel solution (10 µL) was introduced into a pulled glass capillary (0.9 mm inner diameter). The glass capillary was set in a 3D‐printed jig and inserted into a 1.5‐mL microtube filled with 100 mM CaCl₂ aq. (100–120 µL) (Figure 3b). The devices were centrifuged at ~2500 rpm for 120 s using a tabletop centrifuge.

Figure 3.

Fabrication of gel‐beads by centrifuge‐based microfluidic device. (a) A schematic image of droplets detached by centrifugal force and gelated. (b) Images of (i) a glass capillary for one‐aspect gel‐beads, (ii) a theta‐shaped glass capillary for Janus‐shaped gel‐beads, (iii) and a device with the capillary, Jig, and CaCl₂ aq. All scale bars are 10 mm except the left images of (i), (ii), 50 µm. (iv) Image of the device being centrifuged by a table‐top centrifuge. (c) Diameter of the tip of the glass capillary and the diameter of the fabricated gel‐beads at a fixed centrifugal force of ~2500 rpm (red: MG‐0.96 gel‐beads, green: MG‐0.42 gel‐beads, blue: Janus‐shaped gel‐beads, and yellow: MG‐0.64 gel‐beads). (d) Phase contrast images (scale bar: 200 μm), fluorescence images of gel‐beads encapsulating fluorescent particles instead of AAV (scale bar: 200 μm), and distribution of gel‐bead diameters for gel‐beads fabricated to be 200 μm. MG‐0.96 gel‐beads contain red particles, MG‐0.42 gel‐beads contain green particles, and MG‐0.64 gel‐beads contain red particles in MG‐0.96 and green particles in MG‐0.42.

During the centrifuge (Figure 3a, inset), the pre‐gel solution of the Na‐Alg. aq. was pushed out from the glass capillary to form a droplet at the tip of the glass capillary. When the centrifugal force, F _g, applied to the pre‐gel droplet became larger than the reaction force, F _s, acting between the droplet and the tip, the droplet detached and dove into CaCl₂ aq. to form gel‐beads. Therefore, when the droplet is detached:

$$F_{\mathrm{g}}=F_{\mathrm{s}}$$ (1)

is satisfied (Tate's law) (Zhao et al. 2020; Haeberle et al. 2008). In this system, the centrifugal force F _g is expressed by the diameter of the droplet d _p, the density of the pre‐gel solution ρ, and the centrifugal force G. The reaction force F _s is expressed by the surface tension σ and the tip diameter of the glass capillary d ₀. Therefore, Tate's law is expressed as:

$$d_{\mathrm{p}}=\sqrt[3]{\frac{6\sigma d_0}{\rho G}}$$ (2)

To examine that the diameter of the fabricated gel‐beads followed Tate's law, we conducted varied fabrication conditions by changing the tip diameter of the glass capillary (50–120 µm) for MG‐0.96 gel‐beads, MG‐0.42 gel‐beads, Janus‐shaped gel‐beads (MG‐0.96 and MG‐0.42 for each hemisphere), and MG‐0.64 gel‐beads (mixture of MG‐0.96 and MG‐0.42) (Figure 3c). Theoretical bead diameters derived from Tate's law were also plotted as a dashed line, showing that the diameters of the fabricated gel‐beads reasonably followed Tate's law. The diameter of MG‐0.42 gel‐beads tended to be larger than that of MG‐0.96 gel‐beads and Janus‐shaped gel‐beads. We considered that the MG‐0.42 gel‐beads were larger due to the fact that alginate hydrogels with smaller M/G ratios tend to swell (Ramos et al. 2018).

In this study, it was necessary to fabricate gel‐beads of the same volume to compare the diffusion of AAV. Therefore, we tuned the diameter of the fabricated gel‐beads to be 200 µm (RPM: ~2500 rpm, the tip diameter of the glass capillary: 50–80 µm). In addition, in this experiment, fluorescent particles (MG‐0.96: red, MG‐0.42: green) were encapsulated in gel‐beads to clearly visualize them. Observation by phase contrast and fluorescent microscopy shows (Figure 3d) that the averaged diameter of MG‐0.96 gel‐beads, MG‐0.42 gel‐beads, Janus‐shaped gel‐beads, and MG‐0.64 gel‐beads were 201.8 ± 6.3, 198.3 ± 4.0, 201.4 ± 4.9, and 200.52 µm (mean ± SD), respectively. Thus, we obtained 200‐µm‐diameter gel‐beads with high monodispersity (CV ≤ 3.5%).

2.3. Controlled Release of AAV From Gel‐Beads

We then evaluated the release of AAV from the MG‐0.96 gel‐beads, MG‐0.42 gel‐beads, and Janus‐shaped gel‐beads (200 µm in diameter) (Figure 4a). In this experiment, we used AAV‐5 providing cytomegalovirus (CMV)‐driven green fluorescent protein (GFP). The gel‐beads were fabricated using 10 µL of pre‐gel solution (2% (w/v) Na‐Alg. aq. and 1.3 × 10⁹ GC/mL AAV‐5) by the centrifugal microfluidic method described above. After the fabrication of gel‐beads, the gel‐beads were placed in 24 wells with 1 mL of Dulbecco's modified Eagle's medium (DMEM) at 37°C. Then, the supernatant was collected at 12, 24, 48, and 96 h. The concentration of AAV‐5 in the supernatant was measured using enzyme‐linked immunosorbent assay (ELISA). Note that 400 µL of the supernatant from the DMEM containing the gel‐beads was collected, followed by the addition of 400 µL of fresh DMEM.

Figure 4.

AAV release from gel‐beads over time. (a) Schematic image of the AAV release experiment. (b) Cumulative AAV release ratio and (c) AAV release speed from each gel‐bead over 4 days (red: MG‐0.96 gel‐beads, green: MG‐0.42 gel‐beads, and blue: Janus‐shaped gel‐beads).

The initial release ratio of AAV‐5 from the MG‐0.96 gel‐beads was approximately 70% at 12 h. After 96 h, 92% of the encapsulated AAV‐5 was released from the gel‐beads (Figure 4b). On the other hand, in the case of the MG‐0.42 gel‐beads, AAV‐5 release ratios were initially 12% at 12 h, and finally 34% of the total AAV‐5 release at 96 h. These results show that AAV release could be controlled by the M/G ratios of the gel‐beads. Generally, the AAV released from the carrier could be affected by the charge of the alginate hydrogel and the pH (McConnell et al. 2014; Zhang et al. 2016). However, in this study, all conditions except the M/G ratio of the alginate were unified during the fabrication of the beads and in the experiments. Therefore, we consider that the AAV release was suppressed due to the fact that the MG‐0.42 hydrogel had a smaller mesh structure and could physically capture more AAVs than the MG‐0.96 hydrogel, not due to the charge or the pH. Additionally, in the case of the Janus‐shaped gel‐beads, AAV‐5 release ratios were approximately 26% after 12 h and 47% after 96 h. The AAV‐5 release from Janus‐shaped gel‐beads was between those in MG‐0.96 and MG‐0.42 gel‐beads. These results indicate that the release of AAV‐5 can be controlled by the combination of MG‐0.96 and MG‐0.42 hydrogels.

Based on the results above, we hypothesized that AAV release from gel‐beads depends on the diffusion of AAV in the hydrogel. Therefore, we calculated the diffusion coefficient of AAV from the gel‐beads based on Fick's law. The initial stage of substance release is assumed to follow Fick's second law under transient conditions. For spherical gel‐beads, Fick's law is expressed as:

$$\frac{\partial c}{\partial t}=D\left(\frac{{\partial }^{2}c}{{\partial r}^2}+\frac{2}{r}\frac{\partial c}{\partial r}\right)$$ (3)

where c is the concentration of the substance (AAV), t is time, and r is the radius of the gel‐bead (Crank 1979). Assuming that D is constant and the value of t is small (the initial stage of the AAV release), Equation (3) can be integrated over time. As a result, the relative substance release is finally represented as follows (Favre et al. 2001):

$$\frac{M_{\mathrm{t}}}{M_{\mathrm{\infty }}}=\frac{6}{r}{\left(\frac{\mathit{Dt}}{\pi }\right)}^{0.5}-3\frac{\mathit{Dt}}{r^2}$$ (4)

where M _t is the amount of substance released after time t, and M _∞ is the amount of substance released after infinite time.

By using Equation (4), the diffusion coefficient D was calculated from the AAV‐5 release ratio after 12 h. The diffusion coefficient, D _MG‐0.96, for MG‐0.96 gel‐beads and D _MG‐0.42, for MG‐0.42 gel‐beads, were 1.7 × 10⁻¹⁴ and 3.0 × 10⁻¹⁶ m²/s, respectively. The theoretically calculated AAV‐release ratios over time for the MG‐0.96 and MG‐0.42 gel‐beads using Equation (4) and the diffusion coefficients were plotted as dashed lines (Figure 4b). The plots were reasonably close to the experimental values we obtained. Moreover, the AAV‐5 release ratio for the Janus‐shaped gel‐beads was also estimated by using the theoretical values of D _MG‐0.96 and D _MG‐0.42. The experimental AAV‐5 release ratio from the Janus‐shaped gel‐beads was slightly lower than the estimated AAV release ratio over time. The gap between the two types of values could be due to the possibility that the AAV release from the MG‐0.42 hydrogel hemisphere in the Janus‐shaped gel‐beads was dominant. In addition, it was confirmed that by being Janus‐shaped, the swelling of MG‐0.42 was suppressed (Figure 3a). Therefore, it is possible that the AAV release was suppressed due to the shrinkage of the pores in the MG‐0.42 hydrogel in Janus‐shaped gel‐beads caused by the suppression of swelling. Even so, the estimated values roughly matched the actual experimental values. These results show that the release of AAV‐5 from the gel‐beads could be explained by the diffusion principle in the hydrogel.

Furthermore, the AAV‐5 release speed was calculated from the cumulative AAV‐5 release over elapsed times (Figure 4c). The AAV‐5 release speed for the MG‐0.96 gel‐beads decreased significantly from 6.0 × 10⁸ to 1.0 × 10⁷ capsids/h for 96 h. In contrast, in the case of the MG‐0.42 gel‐beads, AAV‐5 release speed decreased relatively slowly from 1.0 × 10⁸ to 1.9 × 10⁷ capsids/h for 96 h. These results show that the release of AAV‐5 from the MG‐0.42 gel‐beads was relatively constant compared to that from the MG‐0.96 gel‐beads. In addition, the AAV‐5 release speed from the Janus‐shaped gel‐beads was approximately 2.2 × 10⁸ capsids/h at 12 h and 1.3 × 10⁷ capsids/h at 96 h. The AAV‐5 release speed for the Janus‐shaped gel‐beads was intermediate between the MG‐0.96 and MG‐0.42 gel‐beads, indicating that the release of AAV‐5 could be controlled by combining the different alginate hydrogels into a single bead.

2.4. Gene Transfer to HeLa Cells by Released AAV From Gel‐Beads

Next, we confirmed the gene transfer ratio into cultured cells on a well by AAVs released from the gel‐beads (Figure 5a). Gel‐beads encapsulating AAV‐5 were fabricated using 3.85 µL of pre‐gel solution (2% [w/v] Na‐Alg. aq. and 1.3 × 10⁹ GC AAV‐5 providing CMV‐driven GFP). The fabricated gel‐beads were dispersed into 96‐well plates seeded with HeLa cells (5.0 × 10³ cells/well). The HeLa cells were then cultured with the gel‐beads for 72 h. After that, the GFP‐positive ratio of the cells was analyzed from the microscopic fluorescent images.

Figure 5.

Gene transfer into cells by releasing AAV. (a) Schematic image of gene transfer experiment. (b) Fluorescent images of gene‐transferred cells after 72 h (scale bar: 200 μm). (c) GFP‐positive ratio of HeLa cells by AAV for each condition. *p < 0.05 and **p < 0.01.

The obtained fluorescence images show that the cells expressing GFP signals by the infection of AAV‐5 released from the gel‐beads were observed (Figure 5b). When AAV‐5 was added directly to the cells (positive) and in the absence of AAV (negative), the averaged GFP‐positive ratios were approximately 61% and 1%, respectively (Figure 5c). On the other hand, in the case of the MG‐0.96 gel‐beads, MG‐0.42 gel‐beads, and Janus‐shaped gel‐beads, the averaged GFP‐positive ratios were approximately 27%, 2%, and 14%, respectively. These results clearly show that the GFP‐positive ratio became lower when the AAV‐5 release ratio was lower. This indicates that AAV‐triggered gene transfer to cells can be controlled by the molecular composition (M/G ratio) and combination of the different molecular compositions (Janus‐shape) for alginate hydrogel microbeads.

2.5. Application of Release Control System to Different Serotype AAV

Then, to confirm the applicability of our approach to different serotypes of AAV, we tested AAV‐1 serotype instead of AAV‐5 (used in Sections 2.4 and 2.5) to examine the release from gel‐beads and gene transfer to cells in the same manner. Note that in this experiment, MG‐0.64 gel‐beads prepared by mixing pre‐gel solutions of MG‐0.96 and MG‐0.42 at a 1:1 ratio (expected M/G ratio: ~0.64) were used instead of the Janus‐shaped gel‐beads (MG‐0.96 and MG‐0.42 for each hemisphere).

ELISA analyses on the released AAV‐1 from the different gel‐beads show that in the case of the MG‐0.96 and MG‐0.42 gel‐beads, AAV‐1 release ratios were approximately 79% and 30% at 96 h, respectively (Figure 6a). In addition, AAV‐1 release ratios from the MG‐0.64 gel‐beads were approximately 23% at 12 h and finally reached 60% at 96 h. As well as the results for AAV‐5, the release ratio of AAV‐1 over time became lower as the M/G ratio of the gel‐beads was smaller. Furthermore, the gene transfection ratios to HeLa cells (expressing GFP) by the released AAV‐1 were investigated. Note that the positive (AAV‐1 only) and negative (no AAV‐1 and no gel‐bead) controls showed the average GFP‐positive ratios of 54% and 0%, respectively. Regarding the MG‐0.96, MG‐0.42, and MG‐0.64 gel‐beads, the averaged GFP‐positive ratios were 37%, 3% and 33%, respectively (Figure 6b). Similar to the release ratio, the transfection ratios also became lower as the M/G ratio of the gel‐beads was smaller. As in Section 2.3, this experiment was conducted under uniform experimental conditions except for the M/G ratio of the alginate hydrogel. The results above show that the proposed controlled release system by the mesh structures in the gel‐beads could be applied not only to AAV5 but also to AAV1, indicating the versatility of the alginate‐based gel‐beads for different serotypes of AAV.

Figure 6.

Evaluation of gel‐beads consisting of alginate (MG‐0.64) mixed with MG‐0.96 and MG‐0.42 in equal amounts. (a) Cumulative AAV release ratio from each gel‐bead over 4 days (red: MG‐0.96 gel‐beads, green: MG‐0.42 gel‐beads, orange: MG‐0.64 gel‐beads, and black: mixture of MG‐0.96 gel‐beads and MG‐0.42 gel‐beads in half quantities in a well). (b) GFP‐positive ratio of HeLa cells by AAV for each condition (red: MG‐0.96 gel‐beads, green: MG‐0.42 gel‐beads, blue: Janus‐shaped gel‐beads, and orange: MG‐0.64 gel‐beads). *p < 0.05 and **p < 0.01. (c) Density of cross‐linking points by G‐section in gel‐beads and AAV release ratio from gel‐beads after 4 days. (d) Comparison of diffusion coefficients (D) of AAV‐1 and AAV‐5 in each gel bead.

The above experimental results regarding the controlled release clearly show that the release ratios of both AAV‐1 and AAV‐5 depended on the density of the G‐section in the alginate hydrogel. In addition, since it is known that the permeability of low molecular weight compounds and proteins differs depending on the cross‐linking points (mesh structure) of the G‐section, the difference in AAV release could also be attributed to the cross‐linking points formed by the G‐section (Ramos et al. 2018; Khanna et al. 2010). Furthermore, the summarized plots of the AAV release ratios at 96 h relative to the density of cross‐linking points by the G‐section for both AAV‐1 and AAV‐5 (Figure 6c) show that the release of both AAVs was suppressed as the density of cross‐linking points increased. Note that the density of cross‐linking points for Janus‐shaped gel‐beads was estimated as the average of those for MG‐0.96 and MG‐0.42. From these results, as a method of tuning the density of cross‐linking points by the G‐section, both the spatial compartmentalization (the arrangement of polymers with different ratios of M and G) and the composite mixtures (the blend of polymers with different ratios of M and G) were found to be effective approaches. Furthermore, the release speeds of AAV‐1 and AAV‐5 were compared (Figure 6d). For MG‐0.96, the diffusion coefficient of AAV‐1 was 3.4 × 10⁻¹⁵ m²/s, that of AAV‐5 was 1.7 × 10⁻¹⁴ m²/s, and for MG‐0.42, 1.6 × 10⁻¹⁶ m²/s (AAV‐1) and 3.0 × 10⁻¹⁶ m²/s (AAV‐5), respectively. This comparison indicates that the release speed of AAV‐5 from the gel‐beads was faster than that of AAV‐1. This difference in the release speeds would be caused by the differences in the charge and hydrophobicity of the AAV surface depending on the serotype (Heldt et al. 2023), which could affect the intermolecular interactions between the alginate polymer chains and the surface of AAVs.

2.6. Comparison of Janus‐Shaped Gel‐Beads and MG‐0.64 Gel‐Beads

Finally, we compared AAV release behavior from Janus‐shaped gel‐beads (consisting of MG‐0.96 and MG‐0.42 hemispheres) and MG‐0.64 gel‐beads (a mixture of MG‐0.96 and MG‐0.42 alginates) (Figure 7a,b). In Figure 7, the cumulative AAV release from Janus‐shaped gel‐beads and MG‐0.64 gel‐beads at 12, 24, 48, and 96 h is plotted, with the release of AAV‐5 and AAV‐1 from MG‐0.64 gel‐beads normalized to 1.0. The results showed that the release of AAV from Janus‐shaped gel‐beads and MG‐0.64 gel‐beads closely matched. Furthermore, the release of AAV from Janus‐shaped gel‐beads tended to be slightly lower than that from MG‐0.64 gel‐beads. This tendency was consistent with the results of the AAV release experiments in Sections 2.3 and 2.5. In addition, the amounts of AAV‐5 and AAV‐1 remaining in Janus‐shaped gel‐beads and MG‐0.64 gel‐beads at 96 h, respectively, were measured to be approximately 60% in both cases (Figure S1), consistent with the amount of AAV released at 96 h. These results suggest that when the number of G‐sections in the gel‐beads is the same, the AAV release speed from those gel‐beads is also the same in the adjustment of G‐sections by the structure of gel‐beads using hydrogels with different M/G ratios (Janus‐shaped) and by blending alginates with different M/G ratios (MG‐0.64). Thus, both Janus‐shaped gel‐beads and MG‐0.64 gel‐beads proposed in this study are viable as extensive methods for the release control of AAV using specific M/G ratios.

Figure 7.

Comparison of cumulative AAV release from Janus‐shaped gel‐beads and MG‐0.64 gel‐beads. (a) In the case of AAV5, and (b) in the case of AAV1, comparison of Janus‐shaped and MG‐0.64 gel‐beads (cumulative AAV release at 96 h from MG‐0.64 for AAV5 and AAV1, respectively, calculated as 1).

3. Discussion

In this study, we demonstrated the controlled release of AAV by gel‐beads using alginate hydrogels with M/G ratios of 0.96 and 0.42. Previous studies (Madrigal et al. 2019) attempted to control AAV release by the degradation ratio of the alginate hydrogels prepared from oxidized and unoxidized alginate. However, the control of the AAV diffusion from the alginate hydrogel was not demonstrated. Another study (Diaz‐Rodriguez et al. 2015) succeeded in controlling the release of AAV by tuning the hydrogel mesh structure with a composite material prepared from alginate and poloxamer. Compared to those previous research, we have achieved the controlled release of AAV with pure alginate hydrogel without any additional materials: The release of AAV can be controlled by tuning the mesh structure of the alginate hydrogels determined by the molecular composition (M/G ratio) of alginate polymer. This approach is simple and suitable for practical applications because pure alginate itself was already approved for medical use (Sun and Tan 2013). Furthermore, we confirmed the controlled release of the two different serotypes of AAV, AAV‐1 and AAV‐5, from our alginate gel‐beads. These results show that the alginate hydrogel carrier proposed in this study for controlling AAV release could be applied to various serotypes of AAV with different directivity.

The alginate used in this study had the largest M/G ratio of 0.96 and the smallest M/G ratio of 0.42 available to us. The M/G ratio of alginate varies depending on various factors, such as the type of marine resource, location, and timing, with reported M/G ratios ranging from 0.34 to 1.73 (Chandía 2001; Fenoradosoa et al. 2010). Therefore, it would be possible to improve the sustainability of AAV release by using alginate with a smaller M/G ratio and to enhance the immediacy of AAV release over a short period by using alginate with a larger M/G ratio.

The demonstration of the AAV release and gene transfer to cells in this study was performed in vitro experiment. For further development to practical use of this technology, the applicability of our controlled release system to in vivo environment would be examined. Kurashina et al. (2024) demonstrated that alginate microspheres encapsulating AAV can be fixed in the body by using collagen to enable transfection with AAV in vivo. Their study confirmed that the local transfection of AAV to biological tissues by using alginate microspheres and collagen. However, the speed of AAV release from the AAV carriers has yet to be achieved. By incorporating our proposed control of AAV release by the M/G ratio of alginate into their results, it is expected to provide a suitable release speed of AAV to each biological tissue in vivo. The establishment of such a system of controlling AAV release in vivo is expected to improve efficacy in gene therapy.

One of the practical targets of the proposed system is gene therapy for hearing loss (Isgrig et al. 2023). In particular, sensorineural hearing loss caused by a defect in hair cells in the inner ear requires gene transfer to the inner ear (Chien et al. 2015). However, the low efficiency of gene transfer to the inner ear has been a problem (X. Hu et al. 2019). The problem can be solved by encapsulating an additional enhancer drug that increases gene transfer efficiency (Wang et al. 2012; Shibata et al. 2012). In this case, our Janus‐shaped gel‐beads could be effective for independent controlled release of both the AAV and the enhancer drug. In conclusion, alginate hydrogel microbeads, by adjusting the mesh structure composed of alginate polymers, are expected to provide a promising platform for the precise controlled release of AAV and gene vectors in a clinical setting.

4. Conclusion

We proposed gel‐beads for controlling AAV release by using the M/G ratio of alginate hydrogel. In the MG‐0.96 and MG‐0.42 hydrogels, the number of cross‐linking points in the molecule was higher in the MG‐0.42 hydrogel, resulting in a smaller mesh structure. Using this difference in the mesh structure, we demonstrated that the release of AAV from those alginate hydrogels is controllable, based on the AAV‐5 released from the gel‐beads and gene transfer to the cells. In addition, we conducted similar experiments not only with AAV‐5 but also with AAV‐1, and confirmed that this system can control the release of AAVs regardless of AAV serotype. With the growing demand for AAV‐based gene therapy for restoration of hearing loss and retinal disease, controlled release of AAV by alginate hydrogel is expected to improve target‐directedness and efficiency of gene delivery.

5. Materials and Methods

5.1. Materials

For the preparation of alginate hydrogel, endotoxin‐free sodium alginate (Na‐Alg., M/G ratio: 0.96 and 0.42) was provided from Mochida Pharmaceutical Co. Ltd. Both provided sodium alginates were approximately 200 mPa s in viscosity (2%) and 150 kDa on average molecular weight. AAV‐1 (pAAV.CMV.PI.EGFP.WPRE.bGH, catalog #105530‐AAV1) and AAV‐5 (pAAV.CMV.PI.EGFP.WPRE.bGH, catalog #105530‐AAV5) were purchased from Addgene. For dissolving and diluting Na‐Alg., we used water (milliQ, Merck Millipore). For gelation of Na‐Alg. aq., calcium chloride solution (CaCl₂ aq.) was prepared by dissolving calcium chloride (CaCl₂, 038‐24985) purchased from Wako Pure Chemical Industries with water (milliQ). For cell culture, DMEM (D5796) and antibiotics (penicillin–streptomycin, P4458) were purchased from Sigma‐Aldrich, and fetal bovine serum (FBS; S‐FBS‐NL‐015) was purchased from SERANA EUROPE GMBH. For washing and detaching cells, phosphate‐buffered saline (PBS; 163‐25265) was purchased from Wako Pure Chemical Industries, and trypsin (Trypsin‐EDTA, 25200072) was purchased from Thermo Fisher Scientific.

5.2. Preparation of AAV

This study was approved by the Recombinant Gene Research Safety Committee of Keio University (approval number: 2022‐21). AAV‐1 and AAV‐5 were stored in a deep freezer at −80°C, thawed just before spreading to cultured cells or mixing with the pre‐gel solution.

5.3. Cell Culture

The human cervical cancer cell line, HeLa (RCB4676, Riken Cell Bank), was used to evaluate the AAV infection in vitro. The cells were maintained in DMEM supplemented with 10% FBS and 1% antibiotics at 37°C and water‐saturated 5% CO₂ atmosphere. Before confluence, the cells were passaged at a 10:1 ratio with trypsin treatment.

5.4. Observation of the Cross‐Section of Alginate Hydrogel

To observe the cross‐section of alginate hydrogel, 2% Na‐Alg. aq. was prepared. A glass capillary (Inner diameter: 0.9 mm, outer diameter: 1.5 mm, G‐1.5, Narishige) was processed to a tip diameter of 120 µm using a puller (PC‐10) and a microforge (MF‐900). The 2% Na‐Alg. aq. was injected into the pulled glass capillary. To assemble a centrifuge‐based microfluidic device (Maeda et al. 2012), 1.5 mL microtube (616 261, Greiner Bio‐One) was filled with 300 µL of 100 mM CaCl₂ aq. and the pulled glass capillary with the Na‐Alg. aq. was fixed to the microtube using a jig. The tip of the glass capillary was immersed in the 100 mM CaCl₂ aq. The microfluidic device was centrifuged (~2500 rpm) in a tabletop swing‐type centrifuge (ATT‐101). The extruded Na‐Alg. aq. gelated to form a calcium alginate hydrogel fiber. After replacing the water in the prepared alginate hydrogel fiber with t‐butyl alcohol (025‐03396, Wako), the fibers were freeze‐dried (Vacuum Freeze Dryer, VFD‐30, Vacuum Device Inc.) and broken to obtain the cross‐sectional plane of the fiber. The fiber was then coated with osmium (HPC‐20, Vacuum Device Inc.) and observed by SEM (FE‐SEM‐H2, Hitachi High‐Tech Co.). To measure the size of the mesh structure of the alginate hydrogel, the obtained SEM images were binarized and the area and number of pores in the images were measured by ImageJ.

5.5. Fabrication of Alginate Hydrogel Microbeads Encapsulating Fluorescent Particles

To fabricate alginate hydrogel microbeads encapsulating fluorescent polystyrene particles, 2% Na‐Alg. aq. mixed with 1% (v/v) 200‐nm‐diameter fluorescent polystyrene particles (red: MG‐0.96, green: MG‐0.42). was prepared as a pre‐gel solution. A glass capillary was processed to a tip diameter of 50 ‐ 120 µm using a puller and a microforge. Then, 10 µL of the prepared pre‐gel solution with fluorescent particles was injected into the glass capillary. In the case of Janus‐shaped gel‐beads, 5 µL of each pre‐gel solution (MG‐0.96, MG‐0.42) was injected into each compartment of a processed theta‐shaped glass capillary (TST150‐6, World Precision Instruments) (10 µL injected). To assemble a centrifuge‐based microfluidic device, a 1.5 mL microtube was filled with 120–140 µL of 100 mM CaCl₂ aq. and the glass capillary containing the pre‐gel solution was fixed to the microtube using a jig. The distance between the tip of the glass capillary and the liquid surface of 100 mM CaCl₂ aq. was 2 mm. The microfluidic device was centrifuged (~2500 rpm) in a small tabletop swing‐type centrifuge. The ejecting droplets of the pre‐gel solution gelated to form alginate hydrogel microbeads encapsulating fluorescent particles. The fabricated gel‐beads were placed on a glass plate (24 × 50 mm, 863‐14‐01‐11, Matsunami Glass Ind. Ltd.). The gel‐beads were observed with a microscope (IX73PI‐22FL/PH, Olympus Co.), and the diameter of the gel‐beads was measured by Imaging Software (cellSens, CS‐ST‐SET, Olympus).

5.6. Fabrication of Alginate Hydrogel Microbeads Encapsulating AAV

To fabricate alginate hydrogel microbeads encapsulating AAV, 2% Na‐Alg. aq. mixed with AAV (AAV‐5: 1.3 × 10¹² GC/mL, AAV‐1: 1.7 × 10¹² GC/mL) was prepared as a pre‐gel solution. A glass capillary was processed to a tip diameter of 50–70 µm using a puller and a microforge. After that, 10 µL of the prepared pre‐gel solution with AAV was injected into the glass capillary. In the case of Janus‐shaped gel‐beads, 5 µL of each pre‐gel solution (MG‐0.96, MG‐0.42) was injected into each compartment of the processed theta‐shaped glass capillary (10 µL injected). To assemble a centrifuge‐based microfluidic device, a 1.5 mL microtube was filled with 120–140 µL of 100 mM CaCl₂ aq. and the glass capillary was fixed to the microtube using a jig. The distance between the glass capillary and the liquid surface of 100 mM CaCl₂ aq. was 2 mm. The microfluidic device was centrifuged (~2500 rpm) in a small tabletop swing‐type centrifuge. The ejecting droplets of the pre‐gel solution gelated to form alginate hydrogel microbeads encapsulating AAV.

After the gel‐beads were created centrifugally, 1 mL of fresh DMEM was added to the microtubes to suspend the fabricated microbeads. After that, the microtubes were centrifuged for 10 s, and 1 mL of the supernatant solution was removed. These steps were repeated three times to wash the gel‐beads to remove CaCl₂ aq. from the gel‐bead suspension.

5.7. Evaluation of Release Ratio of AAV from the Alginate Hydrogel Microbeads Using ELISA

To evaluate the amount of AAV released from the gel beads, the concentration of AAV in the supernatant of the gel‐bead suspension was measured by ELISA. The fabricated gel‐beads were incubated with 1 mL of fresh DMEM in a 24‐well plate (353090, Greiner Bio‐One). After 12, 24, 48, and 96 h, the supernatant of the gel‐bead suspension (400 µL) was collected. After collecting the supernatant, 400 µL of fresh DMEM was added to each well to keep the volume of the gel‐bead suspension constant. The collected supernatant was stored in 1.5 mL microtubes and incubated at 37°C and 5% CO₂ until all samples were collected. After all samples were obtained (96 h), the collected supernatants were analyzed using an ELISA kit (PRAAV5, PROGEN Biotechnik GmbH). The absorbance at a wavelength of 450 nm was measured using a plate reader (51 119 000, Thermo Fisher Scientific). From the measured absorbance, the AAV capsid concentration in the supernatant was obtained. The AAV release ratios from gel‐beads were calculated by (the amount of released AAV from all gel‐beads)/(the initial amount of AAV encapsulated in all gel‐beads). At this time, the concentration of AAV in the supernatant solution at each time obtained by ELISA was corrected using the following equation to calculate the cumulative AAV release from the gel‐beads. Then,

$$R_i=\frac{{\sum }_{j=1}^i(C_{j-1}\times V_{\text{supernatant}})+C_j\times V_{\text{Total}}}{A_{\text{Total}}}$$ (5)

where R is AAV release ratio and C is the AAV concentration of the supernatant at each time, V _Total is the total volume of the solution containing the gel‐beads in 24 wells, V _Supernatant is the volume of solution collected, and A _Total is the amount of AAV encapsulated in the gel‐beads immediately after fabrication. In addition, the subscripts i = 1–4 represent 12, 24, 48, and 96 h, respectively. The AAV remaining in the gel‐beads was calculated by resolving the gel‐beads in 100 mM sodium chloride solution (7647‐14‐5, MANAC Incorporated) and measuring the supernatant.

5.8. Evaluation of AAV Infection With GFP‐Positive Ratio

For the AAV infection experiment, the HeLa cells were trypsinized and seeded in a 96‐well plate (655090, Greiner Bio‐One) at 5.0 × 10³ cells/well with 5 h of incubation. During incubation of the cells, the gel‐beads were fabricated (3.85 µL Na‐Alg. pre‐gel solution with AAV (AAV‐5: 1.3 × 10¹² GC/mL, AAV‐1: 1.7 × 10¹² GC/mL)). The obtained gel‐bead suspensions were washed with fresh DMEM to remove CaCl₂ aq. Then, 1 mL of DMEM with 10% FBS and 1% PS was added to the gel‐beads suspension to resuspend the gel‐beads. After that, the microtubes were centrifuged for 10 s, and 1 mL of the supernatant solution was removed. By this process, the gel beads were properly suspended in the cell culture medium. The gel‐beads suspensions were dispersed over the prepared HeLa cells (5.0 × 10³ cells/well) in the 96‐well plate with culture medium (200 µL). The cells were incubated with the gel‐beads at 37°C and 5% CO₂ for 72 h. After 72 h, the gel beads were removed with an aspirator, and the cells were washed three times with PBS for 30 s. For the evaluation of the AAV infection ratio with fluorescent signals of GFP (infected cell marker), cell nuclei were stained for counting the number of the cultured cells as follows: After washing cells with PBS three times, the cells were fixed by treating cells with 4% paraformaldehyde (163‐20145, Wako) for 15 min. Subsequently, cell nuclei were stained with DAPI (D1306, Thermo Fisher Scientific) for 30 min. Finally, the cells were washed three times by immersing them in PBS for 30 s.

Three fluorescence microscopic images of DAPI and GFP were taken per single well using the microscope. To measure the GFP‐positive ratio, the obtained fluorescent images were analyzed by ImageJ. The fluorescent images of DAPI were binarized, and the number of cells in the image was counted (Threshold, Watershed, and Analyze Particles). These images were stacked on the binarized GFP image as a mask (Convert to Mask), and the cells expressing GFP were counted to calculate the GFP‐positive ratio (Image Calculator and Analyze Particles).

5.9. Statistical Analyses on Morphometric Data

Sample analyses were performed using the Student's t‐test (Ruxton 2006). A value of *p < 0.05 or **p < 0.01 was considered significant. In the line graphs, error bars indicate standard deviation. For the diameter distribution of the gel‐beads, the coefficient of variation (CV) is defined as the ratio of the standard deviation to the mean. For the box‐and‐whisker diagram of gene transfer ratios, values more than 1.3 times the quartile range away from the first and third quartiles were treated as outliers.

Supporting information

Supporting Information Hioki.

Hioki, A. , Takatsuka S., Kurashina Y., and Onoe H.. 2025. “M/G‐Ratio‐Tuned Calcium Alginate Hydrogel Microcarrier for Controlled Delivery of Adeno‐Associated Virus for Gene Therapy.” Biotechnology and Bioengineering 122: 3433–3445. 10.1002/bit.70066.

Data Availability Statement

The data that support the findings of this study are available from the corresponding author upon reasonable request.