Engineering a halloysite nanotube-enhanced hydrogel 3D skin model for modulated inflammation and accelerated wound healing

Abstract

The medicine field continues to encounter obstacles in understanding the etiology of skin inflammation and the process of skin wound repair. Developing sustainable and effective three-dimensional (3D) skin models for investigating inflammatory skin biology remains a challenge. By incorporating halloysite nanotubes (HNT) into a composite collagen/alginate/hyaluronic acid hydrogel, we created a novel 3D skin model and introduced keratinocytes and fibroblast cells into it. We demonstrate that the modified 3D skin model is capable of enhancing the differentiation and adhesion behaviors of keratinocytes and fibroblast cells in vitro and promoting wound healing in vivo. These characteristics highlight the potential of this approach for the study of skin inflammation, wound healing, regeneration, aging, and beyond.

Keywords: Halloysite nanotubes; Hydrogel; Inflammation; Skin; Wound healing.

Graphical abstract

Fig. 1

Toxicity assessment and optimization of HNT content in hydrogel scaffold. (a) Schematic of hydrogel-based scaffold. Natural hydrogel formulation comprises collagen type I, hyaluronic acid sodium salt, alginic acid sodium salt stock solution, 10 × DMEM, 7.5 μM CaCl₂ and cell culture medium. The modified hydrogel formulation incorporates HNT into the natural hydrogel matrix. (b) Relative survival fraction of HFFs and HaCaTs assessed by CCK-8 analysis within hydrogels that contain different concentrations of HNT. Each dot represents one independent experiment. Data are represented as mean ± SD. ns, not significant, ∗P < 0.05, ∗∗∗P < 0.001, ∗∗∗∗P < 0.0001 (unpaired student's t-test). (c) Representative images of HFFs stained for Ki-67 within hydrogels that contain different concentrations of HNT. Scale bar = 50 μm. (d) Percentage of HFFs that are positive for Ki-67 within hydrogels that contain different concentrations of HNT. Data are represented as mean ± SD. Each dot represents one independent experiment. ns, not significant, ∗∗P < 0.01 (unpaired student's t-test). (e) Representative images of HaCaTs stained for Ki-67 and K5 within hydrogels that contain different concentrations of HNT. Scale bar = 50 μm. (f) Percentage of HaCaTs that are positive for K5 within hydrogels that contain different concentrations of HNT. Each dot represents one independent experiment. Data are represented as mean ± SD. ns, not significant, ∗∗P < 0.01 (unpaired student's t-test). (g) Percentage of HaCaTs that are positive for Ki-67 within hydrogels that contain different concentrations of HNT. Each dot represents one independent experiment. Data are represented as mean ± SD. ns, not significant, ∗∗P < 0.01 (unpaired student's t-test).

Fig. 2

Characterization of the modified hydrogel scaffold incorporating human fibroblast cells. (a) Dynamic rheological analysis of hydrogel without HNT and HNT (1 μg/mL)-based hydrogel. (b) Distribution of effect Young's modulus on both the natural hydrogel and the modified hydrogels that contain HNT of different concentrations. (c) Compressive stress-strain curves of HNT-based hydrogels on day 0 and day 4. (d) Swelling assay of hydrogels. The height of the natural hydrogel and the modified hydrogel was compared on days 0, 2, and 5. Each dot represents one independent experiment. Data are represented as mean ± SD. ∗∗∗P < 0.001 (two-way ANOVA analysis). (e) SEM images of hydrogels in different magnifications and energy spectrum images of the distribution of Si, Ca, P, C and N elements in the natural hydrogel and the modified hydrogel.

Fig. 3

Viability and functionality assessment of HFFs in HNT-modified hydrogel (a) Representative images and percentage of HFFs positive for vimentin by IF staining on days 0, 3, and 7 in natural hydrogel and 1 μg/mL HNT-based hydrogel. Data are represented as mean ± SD. ∗P < 0.05 (two-way ANOVA analysis). Scale bar = 20 μm. (b) Relative expression levels of talin2, plasminogen, fibronectin, and paxillin assessed by RT-PCR tests in cultured HFFs on day 3. Each dot represents one independent experiment. Data are represented as mean ± SD. ns, not significant, ∗P < 0.05; ∗∗P < 0.01 (unpaired Student's t-test). (c) Representative images and percentage of HFFs positive for vimentin by IF staining in natural hydrogel and 1 μg/mL HNT-based hydrogel. The cells were stimulated with cytokines (TNF-α, IL-17A, and IL-22) in different concentrations (0, 25, 50, 100 ng/mL). Scale bar = 100 μm. (d) Representative 3D images from the top view of HFFs positive for vimentin in I collagen-based hydrogel, natural hydrogel, and 1 μg/mL HNT-based hydrogel on days 1, 3 and 7. The cells were stimulated without or with cytokines (TNF-α, IL-17A, and IL-22, 50 ng/mL for each cytokine). Scale bar = 100 μm. (e) Relative expression levels of vimentin in HFFs in I collagen-based hydrogel, natural hydrogel, and 1 μg/mL HNT-based hydrogel on days 1, 3, and 7 without cytokine stimulation. Data are represented as mean ± SD. ∗P < 0.05 (two-way ANOVA analysis). (f) Relative expression levels of vimentin in HFFs in I collagen-based hydrogel, natural hydrogel, and 1 μg/mL HNT-based hydrogel on days 1, 3, and 7 with cytokine (TNF-α, IL-17A, and IL-22, 50 ng/mL for each) stimulation. Data are represented as mean ± SD. ∗∗P < 0.01 (two-way ANOVA analysis).

Fig. 4

The modified 3D hydrogel skin model promotes epidermal thickness and differentiation of keratinocytes (a) Schematic of the 3D skin model building. Human keratinocyte cells line (HaCaTs) was seeded on hydrogels on day 1 and day 3. The seeded cells were prepared in a density of approximately 1 × 10⁶ cells/mL. (b) Representative images of the 3D skin model stained by HE from the side-view of hydrogel. The thickness of epidermal layer was presented as a time-dependent graph. Data are represented as mean ± SD. ∗∗∗P < 0.001 (two-way ANOVA analysis). (c) Representative images of the upper-layer of the 3D skin model stained for fibronectin from the side-view of hydrogel. Scale bar = 20 μm. (d) Relative expression levels of fibronectin in the epidermis on the natural hydrogel matrix or HNT-based hydrogel. Data are represented as mean ± SD. ∗∗P < 0.01 (two-way ANOVA analysis). (e) Representative images of the upper-layer of the 3D skin model stained for K5 from the top-view of hydrogel. (f) Representative images of the upper-layer of the 3D skin model stained for K5 and K10 from the side-view of hydrogel. Scale bar = 20 μm. (g) Relative expression levels of K5 and K10 in the epidermis on the natural hydrogel matrix or HNT-based hydrogel. Data are represented as mean ± SD. ∗P < 0.05 (unpaired Student's t-test). (h) Relative expression levels of talin2, plasminogen, fibronectin, and paxillin assessed by RT-PCR tests in HaCaTs. Each dot represents one independent experiment. Data are represented as mean ± SD. ns, not significant, ∗P < 0.05 (unpaired Student's t-test).

Fig. 5

Inflammatory bioactivity of modified hydrogel and its potential for cell proliferation, migration, and adhesion. (a) Schematic of RNA-seq of the HFFs in hydrogel group. Four groups were set as group A (without HNT), group B (with HNT of 1 μg/mL), group C (without HNT, stimulated with TNF-α, IL-17A, and IL-22, 50 ng/mL) and group D (with HNT of 1 μg/mL, stimulated with with TNF-α, IL-17A, and IL-22, 50 ng/mL). (b) Heatmap of relative expression levels of genes associated with inflammation, proliferation and migration in group B (with HNT) compared to group A (without HNT). (c) KEGG pathway analysis of enriched genes showing the relevant functional terms in group D compared to group B. (d) Specific inflammatory clusters of genes related to KEGG analysis showing the relevant functional terms in group D compared to group B. (e) KEGG pathway analysis of enriched genes showing the relevant functional terms in group A compared to group C as well as group D compared to group B. (f) Venn diagram of differentially expressed genes overlapped in the four groups. (g) Representative 3D images of K17 expression in upper-layer from the top and side-view of the modified 3D skin model. (h) Relative expression levels of K17 in 0 μg/mL and 1 μg/mL HNT-based hydrogel with cytokine (TNF-α, IL-17A, and IL-22, 50 ng/mL) stimulation. Each dot represents one independent experiment. Data are represented as mean ± SD. ∗P < 0.05 (unpaired student's t-test).

Fig. 6

HNT-Based hydrogel accelerates the healing process in full-thickness wound model in mice (a) Schematic of hydrogels in a skin wound healing murine model. (b) Processes of full skin defect repair in C57BL/6 mice as a wound healing process (n = 3). The diameter of defect skin area is 8 mm. (c) Representative images of wound area in groups that were administered with blank, natural hydrogel, HNT-based hydrogel on days 0, 2, 5, 7, and 14. (d) Quantification analysis of residual wound area. (e) Relative ratio of wound area of groups that were administered with blank, natural hydrogel, HNT-based hydrogel on days 2, 5, 7, and 14. n = 3. Data are represented as mean ± SD. ∗P < 0.05, ∗∗P < 0.01 (unpaired student's t-test). (f) Representative images of HE staining on day 14 in groups that were administered with blank, natural hydrogel, HNT-based hydrogel. Scale bar = 400 μm.

Fig. 7

HNT-based hydrogel enhances epithelial proliferation and adhesion in full-thickness wound model in mice (a) Representative images of whole-mount IF staining for K5 in upper-layer of skin and fibronectin expression between the epidermis and dermis. Scale bar = 50 μm. (b) Relative expression of K5 and fibronectin in groups treated with blank, natural hydrogel, HNT-based hydrogel, compared to the average expression level in control group. Each dot represents one independent experiment. Data are represented as mean ± SD. ns, not significant, ∗∗P < 0.01 (unpaired student's t-test).