Cocktail Cell‐Reprogrammed Hydrogel Microspheres Achieving Scarless Hair Follicle Regeneration

Abstract The scar repair inevitably causes damage of skin function and loss of skin appendages such as hair follicles (HF). It is of great challenge in wound repair that how to intervene in scar formation while simultaneously remodeling HF niche and inducing in situ HF regeneration. Here, chemical reprogramming techniques are used to identify a clinically chemical cocktail (Tideglusib and Tamibarotene) that can drive fibroblasts toward dermal papilla cell (DPC) fate. Considering the advantage of biomaterials in tissue repair and their regulation in cell behavior that may contributes to cellular reprogramming, the artificial HF seeding (AHFS) hydrogel microspheres, inspired by the natural processes of “seeding and harvest”, are constructed via using a combination of liposome nanoparticle drug delivery system, photoresponsive hydrogel shell, positively charged polyamide modification, microfluidic and photocrosslinking techniques. The identified chemical cocktail is as the core nucleus of AHFS. In vitro and in vivo studies show that AHFS can regulate fibroblast fate, induce fibroblast‐to‐DPC reprogramming by activating the PI3K/AKT pathway, finally promoting wound healing and in situ HF regeneration while inhibiting scar formation in a two‐pronged translational approach. In conclusion, AHFS provides a new and effective strategy for functional repair of skin wounds.


Experimental Section
Chemical cocktail screening "Fibroblasts" and " Dermal papilla cells" as a search term in the GEO database (https://www.ncbi.nlm.nih.gov/geo/), the expression data of DPC and HDF was downloaded (GSE31324), which included 4 samples of "dermal papilla freshly-microdissected" (GSM776959-6962) and 4 samples of "fibroblasts cultured" (GSM776967-6970), and GPL571 was selected as the detection platform ([HG-U133A_2] Affymetrix Human Genome U133A 2.0 Array).GEO2R was used to identify the differentially expressed genes (DEG) of HDF and DPC.Firstly, background correction and normalization were performed on samples, and then DEG analysis was performed using the limma package of R language to determine the enriched genes.Log (Fold Change) >1 or < -1 and adjusted P<0.05 were considered as significantly expressed DEG. GO enrichment analysis and KEGG enrichment analysis of DEG were performed using STRING database (STRING: functional protein association networks (string-db.org)) [1].
Connectivity Map [clue.io], a biological database of associations of genes, diseases, and drugs based on gene expression profiles, was as the platform for high-throughput mechanism-driven phenotype compound screening [2] .The principle of this database is to sequence thousands of drugs after treating different cells and record their gene expression profiles.Then, we can compare the list of differential genes with the reference data set of the database by sequencing the up-regulated and down-regulated DEG from tissue or cells.According to the enrichment of DEG in the reference gene expression profile, a correlation score (-100~100) was obtained.
Positive numbers indicate that the up-regulated and down-regulated differentially expressed genes have similar expression profiles with reference genes, while negative numbers are the opposite.Ultimately, small molecule drugs that might play a role are sought based on the ratio of expression differences.Briefly, genome-wide chemical-induced gene expression is dependent on cMap for drug discovery or repurposing [3] .By uploading the DEG of HDF and DPC into the cMap database, a large number of drugs that may contribute to the conversion of HDF to DPC were identified.Then, according to the reported mechanisms in cellular reprogramming, FDA approval status, and mRNA level of DPC-specific markers, optimized small molecular compounds with high reprogramming efficiency and clinical transformation prospect are gradually selected.

Preparation of the AHFS seed microspheres
Liposome carrying drugs were composed of lecithin, cholesterol, and fat-soluble drugs (Ti and T) that drive fibroblast fate toward DPC.The steps to synthesize liposomes are as follows.The mixture of 6.87 mg lecithin, 2.9 mg cholesterol, and moderate fat-soluble drugs was dissolved in 1 ml of chloroform and then evaporated for 20 min at 37℃ to form a uniform film.1.5 ml deionized water was added and ultrasound with a probe (30% power) for 7min.Finally, the filters (0.45μm or 0.22μm) were used to remove free fat-soluble drugs.GelMA was prepared.Briefly, gelatin at a concentration of 10% (w/v) was completely melted in polyphosphate-buffered saline (PBS) at (60 °C, added into methacrylic acid (MA) solution, and mixed and reacted at 50°C for 4 h.An additional PBS diluent was used 5 times to stop the reaction, and then the compound was analyzed at 40°C for 1 week to filter impurities (14 kDa cut-off analysis tube).The GelMA aqueous solution was frozen and dried, resulting in a white, milky, porous structure foam.AHFS microspheres were prepared using an improved micro-liquid fluidity-focusing device.In simple terms, the aqueous phase (10 wt% GelMA and 2.5% liposome evenly mixed in PBS with 0.5% light stabilizer) and the oil phase (5 wt% Span80 in paraffin oil) were introduced into the micro-liquid device, and the water flow of the aqueous and oil phases was controlled using a syringe connected to a syringe pump.The resulting single-dispersive emulsion drops were optically cross-linked under UV light.The microspheres were placed into each microtubule.Then, 1 mL isopropyl alcohol was added, washed by oscillating and centrifuged at 4000 rpm to collect the microspheres.Subsequently, the microspheres were treated with 10% Poly allylamine hydrochloride (PAH) solution for 2h at 37°C under shaking at (")rpm.Then, the microspheres were washed to remove the excess dopamine.

Characterization of AHFS seed microspheres
The morphology and diameter of microspheres were determined by the bright-field microscope (LSM800, ZEISS, Germany).SEM (FEI Sirion 200, USA) was used to scan the surface morphology and microstructure of microspheres.Energy dispersive spectroscopy (EDS) and elemental mapping were utilized for the measurement of microsphere composition and element distribution.Microspheres were incubated with PBS containing hyaluronidase to assess its degradability, and the changes in microsphere morphology were monitored by microscopic observation at the indicated time.In the swelling test, 3mg microsphere is added to a 1.5mL tube and the weight of the microsphere and tube is measured before adding 1ml deionized water.Adjust the pH of the suspension to 7.4 and place the tube in a shaking incubator at 37°C and 80 rpm.At the specified time point, centrifuge the tube (3000 rpm, 3 min) and then remove the supernatant.Use filter paper to remove excess water before weighing.
Repeat the whole process until you get a constant weight.For drug loading and release, all the chemicals used in this study were purchased from MCE, and the concentrations of the chemicals were measured via an ultraviolet spectrophotometer.
Both Ti and T concentrations in vitro experiment were 5 μmol/L, while that were 50 μmol/L in vivo studies.Then, the microspheres were centrifuged and the unbound chemicals in the supernatant were collected to calculate the loading efficiency and loading capacity.The drug loading was calculated by ng/mg microsphere.The loaded microsphere was then incubated at 37℃ in 1ml PBS (pH = 7.4) containing 0.1% w/v BSA and stirred at 80 rpm.Chemicals released from the supernatant were measured at an indicated time point of 120 hours.

Cell Biocompatibility
Fibroblasts, epidermal keratinocytes, and vascular endothelial cells were all wound-resident cells that participate in wound healing.Considering the application of microspheres in the skin, the cell biocompatibility of microspheres with HDF, human epidermal keratinocyte (HEK), and human umbilical vein endothelial cells (HUVEC)

Quantitative real-time PCR
According to the manufacturer's advice, using TRIzol reagent (Invitrogen, USA) to extract total RNA from cells or tissues.Reversing transcription of total RNA (500 ng) with a PrimeScript RT reagent kit (TaKaRa, Japan) to generate cDNA.In quantitative real-time polymeric chain enzyme reaction (qRT-PCR), targeted genes were analyzed quantitatively by SYBR Green Supermix (Bio-Rad, USA).The details of primer information were listed in Table S2.

ALP staining
For detection of ALP activity to evaluate the characteristics of AHFS-treated cells, the HDF and microsphere-treated cells (1x 10 6 ) were fixed with 4% paraformaldehyde for 10 min and then incubated with ALP staining solution (BCIP/NBT alkaline phosphatase color development kit, C3206, Beyotime) for 30 min at room temperature away from light.The color reaction can be terminated by removing the ALP staining solution and washing 1-2 times with distilled water.

Cell cycle analysis
Cells were fixed in 70% precooled ethanol at 4 °C overnight.After washing with PBS, cells were re-suspended with 100µl RNase A solution in 37℃ water baths for 30min, followed by 400µl PI staining solution at 4 °C for 30 min.Finally, the treated cell suspension was detected and the red fluorescence at 488nm was recorded.

Western blotting
Using RIPA buffer (Sigma-Aldrich) to obtain total protein from cells or tissue.

Figure S1 .
Figure S1.Transcriptomic differences based on Transcriptomic s between primary DPC and HDF.a) Normalized signal intensity of HDF and DPC samples, showing the good uniformity.b) Principal Components Analysis (PCA) of HDF and DPC samples.c) Heatmap analysis between HDF and primary DPC.d) Volcano plot showing the differentially expressed genes (DEG) of primary DPC compared to HDF.Not sig, not significant.Up-regulated DEG enriched in DPC function and feature: e) cell activity.f) Osteogenic differentiation and stem cell property.g) Hair follicle and skin development.h) Interleukin and chemokine production.

Figure S2 .
Figure S2.The information of predicted drug candidates that induce fibroblast-DPC transition.a) The cMap score of drug candidates.b) Clinical status and biochemical characteristics of drug candidates.

Figure S3 .
Figure S3.The property of AHFS microspheres.a) The micromorphology of drug-loaded liposomes; b) Macroscopic and local images of microspheres, AHFS, and a mixture of both; c) the zeta potential of the microspheres without polyamide modification and AHFS; d) Images of microspheres adhesion on the skin surface of mice before and after rinsing with running water; e) The number of microspheres remaining on the skin surface after rinsing.As the control, MS in this experiment was non-adhesive.Data are expressed as mean ± S.D.; n = 3. ***, p < 0.001.

Figure S4 .
Figure S4.The biocompatibility of AHFS microspheres.a) Images of HEK live/dead cell staining on day 1, day2 and day 3, among control, MS microsphere and AHFS microsphere group.Scale bar = 50 μm.b) Quantitative analysis for the percentage of alive HEK cells in control, MS microsphere, and AHFS microsphere group.Results are expressed as mean ± S.D.; ns, not significant.c) Images of HUVEC live/dead cell staining on day 1, day2 and day 3, among control, MS microsphere and AHFS microsphere group.Scale bar = 50 μm.d) Quantitative analysis for the percentage of alive HUVEC cells in control, MS microsphere, and AHFS microsphere group.Results are expressed as mean ± S.D.; ns, not significant.

Figure S5 .
Figure S5.The effect of MS microspheres on HDF.a) Phase contrast images showing the morphological characteristics of HDF and MS microsphere-treated DPC in 2D culture.Scale bar = 200 μm.b) Multicellular sphere formation assay of HDF and MS microsphere-treated DPC in 3D culture.Scale bar = 100 μm.c) qRT-PCR analysis of transcriptional expression of DPC-associated maker α-SMA, CXCR4, BMP2, BMP4, ALPL, LEF1, β-CATENIN and FOXO1 in HDF and MS microsphere-treated DPC after 5 days of induction.All data in qRT-PCR analysis are expressed as mean ± S.D.; n = 3. ns, not significant.

Figure S6 .Figure S7 .
Figure S6.PI3K inhibitor blocked the AHFS microspheres-induced fibroblast-DPC transition.a) KEGG analysis of upregulated DEG between primary DPC and HDF, showing PI3K/AKT signaling pathway was significantly enriched.b) Phase contrast images showing the morphological characteristics of HDF with or without AHFS microsphere+Alpelisib treatment in 2D culture.Scale bar = 200 μm.c) Multicellular sphere formation assay of HDF with or without AHFS microsphere+Alpelisib treatment in 3D culture.Scale bar = 100 μm

Figure S8 .
Figure S8.Hair follicle regeneration and in vivo mechanism verification.a) The efficiency of hair regeneration in healed wound (POD 28); b) Western blot analysis of wound tissue in phosphorylation AKT, AKT and DPC marker β-CATENIN in control group, MS group, AHFS group and TiT group; c) Qualification of β-CATENIN, and p-AKT/AKT immunoblots was performed using Image J software.GAPDH was used as internal loading control.***, p < 0.001, ****, p < 0.0001.