Integrating Deep Learning and Real-Time Imaging to Visualize In Situ Self-Assembly of Self-Healing Interpenetrating Polymer Networks Formed by Protein and Polysaccharide Fibers

Abstract

Fibrillar protein hydrogels are promising sustainable biomaterials for biomedical applications, but their practical use is often limited by insufficient mechanical strength and stability. To address these challenges, we transformed native proteins into amyloid fibrils (AFs) and incorporated a fibrillar polysaccharide, phytagel (PHY), to engineer interpenetrating polymer network (IPN) hydrogels. Notably, we report for the first time the formation of an amyloid-based hydrogel from apoferritin (APO), with PHY reinforcing the network's mechanical integrity. In situ self-assembly of APO within the PHY matrix yields fully natural, biopolymer-based IPNs. Rheological analyses confirm synergistic interactions between AF and PHY fibers, with the composite hydrogels exhibiting significantly enhanced viscoelastic moduli compared with individual components. The AF-PHY hydrogels also demonstrate excellent self-healing behavior, rapidly restoring their storage modulus after high-strain deformation. A major advancement of this study is the application of deep learning (DL)-based image analysis, using convolutional neural networks, to automate the identification, segmentation, and quantification of fibrillar components in high-resolution scanning electron microscopy images. This AI-driven method enables precise differentiation between AF and PHY fibers and reveals the three-dimensional microarchitecture of the IPN, overcoming key limitations of traditional image analysis. Complementary real-time confocal laser scanning microscopy, with selective fluorescent labeling of protein and polysaccharide components, further validates the IPN structure of the hybrid hydrogels. Our results demonstrate that DL significantly enhances structural characterization and provides insights into gelation processes. This approach sets a new guide for the analysis of complex soft materials and underlines the potential of AF-PHY hydrogels as mechanically robust, self-healing, and fully sustainable biomaterials for biomedical engineering applications.

Keywords: IPN networks; deep learning; fibrillar polysaccharide; hydrogels; protein fibers; real-time CLSM imaging.

1

(A–C) FESEM images and (D–F) corresponding analysis images and (i–iii) histograms of APO, BLG, and LYS hydrogels, respectively. Macroscopic images of self-standing hydrogels are also included (A–C inset).

1. Schematic of the Strategy for Producing Versatile Materials through AF In Situ Polymerization within PHY to Form IPNs

2

(A–C) Hydrogel disc exterior HRSEM images and (D–F) hydrogel disc interior HRSEM images of APO-, BLG-, and LYS-PHY IPN hydrogels, respectively. The red arrows highlight the brighter protein fibers. Macroscopic images of hydrogel discs are also included (insets). (G–L) Corresponding analysis images.

3

(A–C) Hydrogel disc exterior SEM images and (D–F) hydrogel disc interior SEM images of APO-, BLG-, and LYS-PHY IPN hydrogels, respectively. (G,–L) CNN analysis images. Blue color: PHY and magenta color: AF; (i–vi) the corresponding histograms are also shown.

4

(A) Nitrogen gas absorption–desorption isotherms of the AF–PHY hydrogels. (B) Pore size distribution derived from the nitrogen desorption curves. (C) Elemental analysis of pure protein, pure PHY, and AF–PHY hybrid hydrogels (black: carbon, blue: nitrogen, yellow: sulfur, and white: hydrogen). (D) FT-IR spectra of AF–PHY hybrid hydrogels.

5

Rheological properties of AF–PHY hydrogels. (A) Normalized strain amplitude sweeps at a constant frequency (f = 1 Hz). (B) Normalized excitation frequency sweep at a constant strain amplitude (γ₀ = 0.1%). (C) Damping factor (tan δ = G″/G′) as a function of the strain amplitude for the data shown in (A,D) elastic stress (G′γ₀) as a function of the strain amplitude for the data shown in (A).

6

(A) G′ (black dots) and Young’s modulus (red bars) of the AF–PHY hybrids as obtained under compression at a compression velocity of 10 μm·s^–1. (B) Flow stress/yield stress ratio and cohesive energy (G′γ _y ²/2) of AF–PHY hydrogels (black dots: flow stress/yield stress coefficient from viscosity measurements).

7

Stepped strain amplitude tests in dynamic oscillatory tests when alternated between γ₀ = 0.1 and γ0 = 100% over 50 s intervals at f = 1.0 Hz and 25 °C. (A,B) Time dependence of the viscoelastic moduli in a five-intervals test for PHY and APO-PHY hydrogels, respectively. (C) Fraction of recovered storage modulus of the hydrogels investigated in this work. (D) SEM image of APO-PHY hydrogel after self-healing experiments.

8

3D CLSM images of (A) ATTO488-APO, (B) ATTO488-BLG, (C) ATTO488-LYS, (D) pure PHY, (E) ATTO488-APO-PHY, (F) ATTO488-BLG-PHY, (G) ATTO488-LYS-PHY, and (H) ATTO488-PHY hydrogels. (i–iv) Hydrogel discs under white light and (v–viii) under UV light irradiation.

9

CLSM (A) images of the three-component mixture (PHY/ATTO647/ATTO488). The green and red images were acquired in ATTO647 and ATTO488 channels, respectively. The merged image clearly indicates that PHY pure hydrogel is only functionalized with ATTO647. (B) CLSM images of the four-component mixture (APO/PHY/ATTO647/ATTO488). The green and red images were acquired in ATTO647 and ATTO488 channels, respectively. The merged image clearly indicates that ATTO647-PHY and ATTO488-APO fibers are separately and specifically functionalized with one of the two fluorophores. Scale bars, 75 μm. (C) 3D CLSM image (left) of the three-component (PHY/ATTO647/ATTO488) and (right) of the four-component mixture (APO/PHY/ATTO647ATTO488). The 3D CLSM images are constructed from z-stacked x–y slice images. (i) Three-component and (iii) four-component hydrogel discs under white light and (ii,iv) under UV light irradiation.