Materials-based hair follicle engineering: Basic components and recent advances

Abstract

The hair follicle (HF) is a significant skin appendage whose primary function is to produce the hair shaft. HFs are a non-renewable resource; skin damage or follicle closure may lead to permanent hair loss. Advances in biomaterials and biomedical engineering enable the feasibility of manipulating the HF-associated cell function for follicle reconstruction via rational design. The regeneration of bioengineered HF addresses the issue of limited resources and contributes to advancements in research and applications in hair loss treatment, HF development, and drug screening. Based on these requirements, this review summarizes the basic and recent advances in hair follicle regulation, including four components: acquisition of stem cells, signaling pathways, materials, and engineering methods. Recent studies have focused on efficiently combining these components and reproducing functionality, which would boost fabrication in HF rebuilding ex vivo, thereby eliminating the obstacles of transplantation into animals to promote mature development.

Keywords: Engineering methods; Hair follicle; Materials; Signaling pathways; Stem cells.

Graphical abstract

Fig. 1

Schematic illustration of the components involved in the fabrication of hair follicles.

Fig. 2

Three methods for obtaining seed cells. (a) Extraction from tissue. (a–i) Dermal papilla cells (DPCs) are obtained from the base of the HF, while hair follicle stem cells (HFSCs) are extracted from the bulge of the HF. (a-ii) Extraction from neonatal skin. Neonatal skin tissue is directly digested to obtain skin-derived precursors (SKPs). When the epithelial and mesenchymal layers of the tissue are digested separately, epithelial stem cells (EpSCs) and mesenchymal stem cells (MSCs) can be harvested. (b) Stem cell induction. (b–i) Adipose stem cells (ASCs) are induced to differentiate into DPC-like cells by cultured with CAO1/2FP medium and DPC extracellular vesicles (DPC-EVs). (b-ii) Hair follicle cell-inducing potential of induced pluripotent stem cells (iPSCs). IPSCs can differentiate into neural progenitor cells (NPCs), which can further differentiate into DPC-like cells. Alternatively, iPSCs can differentiate into induced mesenchymal cells (iMCs) and then into DP-substituting cells (iDPSCs). Additionally, iPSCs can be directly induced into skin organoids with intact HFs. (c) Reprogramming of somatic L929 cells into DPC-like cells using CHIR99021, TTNPB, and Forskolin.

Fig. 3

Signals involved in HF morphogenesis and circulation. HF morphogenesis occurs through six stages with distinct signals. Placode formation is initiated by Wnt signals in the dermis, while Wnt/β-catenin, ectodysplasin (Eda)/NF-kB, sonic hedgehog (SHH), and noggin promote HF placode formation. Conversely, bone morphogenetic protein 2 (BMP2), bone morphogenetic protein 4 (BMP4), and Notch inhibit placode formation. Wnt/β-catenin facilitates dermal papilla formation, platelet-derived growth factor-A (PDGF-A), and SHH signaling. Subsequently, hair peg formation is promoted by Wnt/β-catenin, SHH, and transforming growth factor α/epidermal growth factor receptor (TGF-α/EGFR) signaling. Boundary formation of the HF involves Wnt/β-catenin, Notch, BMP2, and BMP4. During anagen, follicle formation is stimulated by Wnt/β-catenin, SHH, Notch, fibroblast growth factor 10 (FGF10), and fibroblast growth factor 12 (FGF12). The transition from anagen to catagen phase is induced by fibroblast growth factor 5 (FGF5), BMP, and transforming growth factor β (TGF-β). Finally, fibroblast growth factor 18 (FGF18) maintains the telogen phase and inhibits the transition of HFs into anagen.

Fig. 4

Schematic depictions of the classic Wnt and Hedgehog (HH) signaling pathways. (a) The classic Wnt pathway. In the absence of Wnt, cellular β-catenin is targeted for degradation by a complex consisting of glycogen synthase 3 (GSK3), Axin, casein kinase 1 (CK1), and adenomatous polyposis coli tumor suppressor protein (APC), resulting in the silencing of targeted genes in the nucleus. In the presence of Wnt, the enzymatic complex fails, leading to the release of β-catenin. Subsequently, β-catenin translocates to the nucleus, where it interacts with the lymphoid enhancer factor/T cell factor (LEF/TCF) family, facilitating the normal transcription of the target genes. (b) The classic HH pathway. In the absence of HH (e.g., SHH), the Hh receptor Patch inhibits the smoothened (Smo) activity of protein kinases that includes protein kinase A (PKA), GSK3, and CK1. This inhibition leads to the cleavage of Gli into the truncated form GliR, acting as a deterrent to target gene expression. In the presence of HH, the HH ligand binds to Ptch and derepresses Smo. This action signals Sufu to release the Gli activator (GliA), which subsequently migrates to the nucleus and activates the expression of target genes.

Fig. 5

Schematic depictions of the canonical Notch and transforming growth factor-β/bone morphogenetic protein (TGF-β/BMP) signaling pathways. (a) The Notch pathway. The receptor Notch binds to the ligand (Delta or Jagged) and then undergoes cleavage by the γ-secretase complex, releasing the active fragment of the Notch protein, NICD. NICD translocates to the nucleus and binds to the transcription factors CBF-1/suppressor of hairless/Lag1 (CSL) and mastermind-like (MAML) to regulate downstream gene expression. In the absence of NICD, the CSL co-inhibitor binds to silence target genes. (b) The TGF-β/BMP pathway. TGF-β or BMP binds to type I and II receptors, recruiting and phosphorylating downstream Smads (Smad2/3 in TGF-β, Smad1/5/8 in BMP). p-Smads form a trimeric complex with Smad4, which translocates to the nucleus to regulate the transcription of target genes.

Fig. 6

Acquisition and application of decellularized matrix. (a) Process for the decellularization and solubilization of porcine skin extracellular matrix. Reproduced with permission [175]. Copyright 2022, Elsevier. (b) Localization of EpSCs and MSCs cultured with or without Matrigel supplementation after 2 days of culture. (c) Schematic of different structures formed by EpSCs and MSCs in the presence or absence of Matrigel. (d) In hair follicloids constructed from epithelial cells and mesenchymal cells, the number of genes up- and down-regulated due to Matrigel supplementation. (e) In hair follicloids, the changes in gene expression of ECM and ECM binding related proteins. (b–e) reproduced with permission [82]. Copyright 2022, The American Association for the Advancement of Science.

Fig. 7

Natural polymer applications in skin repair and HF regeneration. (a) SEM images of the native dermal ECM and fibronectin (Fn) scaffolds. (b) Representative images of the untreated group (Control) and Fn nanofiber-treated group on days 2, 8, and 16 in wound repair experiments. The insets below are shown in the enlarged image, showing that the FN treatment group has a better wound healing effect (highlighted with the dashed line). (a–b) reproduced with permission [189]. Copyright 2018, Elsevier. (c) Gross appearance of fibrin solution (left) and hydrogel (right) at concentrations of 20, 40, and 80 mg/mL (d) SEM images of fibrin hydrogels at concentrations of 20, 40, and 80 mg/mL. (e) Real-time PCR analysis of SKPs cultured with fibrin hydrogels for 3 days and the expression of HF induction-associated genes. (f) Representative back images of nude mice after 4 weeks of transplantation. (c–f) reproduced with permission [191]. Copyright 2022, The Authors.

Fig. 8

Synthetic polymer applications in HF regeneration. (a) The confocal image of RFP‐expressing HaCaT surrounding GFP‐expressing DP in the middle slice. Scale bar: 200 μm. (b) The confocal image of 3D tri‐cultured aggregates. The white arrows indicate the position of HDFs that have polarized and migrated around with DP‐HaCaT aggregates in the middle slice. Scale bar: 200 μm. (a–b) reproduced with permission [202]. Copyright 2019, The Authors. (c–d) Digital images of the 3D printed dermis. (c) Top and (d) lateral views of the 3D printed skin equivalent. (e) Scanning electron microscopy (SEM) images of cryo-sectioned GelMA and GelMA/HAMA. (f) 3D projection of the live/dead assay in skin equivalent. (g) Depth coding of the live cell signal in skin equivalent. (c–g) reproduced with permission [204]. Copyright 2022, Wiley‐VCH GmbH. (h) Preparation process of the Chitosan/PVA nanofiber sponge. Scale bar: 5 μm. (i) After 3 days of culture, DP microtissues can form within the internal structure of the nanofiber sponge. DP microtissues were mixed with epidermal cells and transplanted into the back of nude mice. After 4 weeks, HFs can be observed to regenerate. (j) SEM images of the cell morphological change in the 2D and 3D after 1 and 3 days of culture. Scale bar: 25 μm. (h–j) reproduced with permission [207]. Copyright 2020, American Chemical Society.

Fig. 9

Fabrication of self-assembled spheres of HF cells. (a) Formation of microtissues from highly passaged DPC cells using the hanging-drop approach. Reproduced with permission [18]. Copyright 2016, American Chemical Society. (b) Preparation of vHFGs using DPCs, epithelial cells, and HUVECs after 2 days of self-organization using HFG chip. Transplanting vHFGs to the back of nude mice can achieve hair regeneration. (c) Digital image of HFG chip. The inset shows cultured vHFGs in microwells. (b–c) reproduced with permission [19]. Copyright 2021, The Authors. (d) Schematic illustration of LBL-DP preparation. DPCs coated with gelatin (red) and alginate (green), and then LBL-DPCs were crosslinked with calcium ions to prepare LBL-DP. (e) TEM images of DPCs and LBL-DPs. Red arrows indicate the nano-scale ultrathin ECM. (f) Subcutaneous images and HE staining of transplant sites after three weeks post-injection. There is no hair regeneration in the DPCs group; in contrast, green arrows indicate LBL-DP can induce a large number of HF-like structures, and yellow arrows indicate numerous de novo hairs were generated in vascular DP. Scale bars: 100 μm (HE images) and 500 μm (stereoscopic images). (d–f) reproduced with permission [222]. Copyright 2022, The Authors.

Fig. 10

Applications of microfluidics in hair culture and cell spheres preparation. (a) Digital image of multi-organ-chip with built-in micropumps to provide a pulsatile flow of the medium. (b) Schematic diagram of labeled areas for culturing in vitro skin models, ex vitro skin, and hair follicular units in transwells. (a–b) reproduced with permission [230]. Copyright 2013, Royal Society of Chemistry. (c) Digital image of the T-junction microfluidic chip. (d) The diameters of IGMs vary with the oil and aqueous phase flow rates (Flow rate of aqueous phase: flow rate of oil phase). (c–d) reproduced with permission [20]. Copyright 2022, The Authors. (e) The GelMA/HAD microspheres encapsulate MSCs and EPCs using the microfluidic method. (f) Picture of the microfluidic chip. (e–f) reproduced with permission [21]. Copyright 2009, IOP Publishing.

Fig. 11

Applications of 3D bioprinting in HF regeneration. (a) Bioprinting robot performs printing work on the back of nude mice. (b) After 4 weeks, HFs were generated after robotic bioprinting (P) and hand implantation (H). Scale bar: 2 mm. (a–b) reproduced with permission [182]. Copyright 2022, The Authors. (c–d) The mold generated by 3D printing has 255 HF per cm² in grafts. Scale bar: 4 mm. (e–f) Within 4–6 weeks of grafting high follicle-density HSCs onto immune-deficient nude mice, hair grew in the grafts. Scale bar: 2 mm. (c–f) reproduced with permission [22]. Copyright 2018, The Authors. (g) The printed hair microgels (HMG) in both macro and micro views. (h) The three separate tissue grafts produce hair shafts. gHMG: guide- HMG, RVE: upper side, mesenchymal bead; bottom side: epithelial bead, FWD: upper side, epithelial bead; bottom side: mesenchymal bead, RDM: random directions. Three weeks following the transplant, the dorsal skin of the nude mice was examined in the transplanted areas. (g–h) reproduced with permission [23]. Copyright 2023, The Authors.