Culture experiments were performed with the hiPSC line CMC-003, donated from the National Stem Cell Bank of Korea (Korea National Institute of Health), and originally provided by The Catholic University of Korea. Pluripotent stem cells were cultured on six-well plates coated with 0.5% Matrigel (Corning, NY, USA) and maintained in mTeSR-1 Basal Medium (StemCell Technologies Inc., Vancouver, BC, Canada) with mTeSR supplement (StemCell Technologies Inc.). The medium was replenished every day, and the CMC-hiPSC-003 cells were passaged as tiny clusters at around 40% to 60% confluency (generally every 3 to 4 days) using Gentle Cell Dissociation Reagent (StemCell Technology Inc.).

For differentiation, colonies of hiPSCs were detached from the culture dish using 10% dispase (StemCell Technologies Inc.) to generate embryo bodies. After dissociation, cells were collected as tiny clusters in E8 medium containing 10 μM Y27632. The appropriate number of live cells needed for an experiment was transferred into an E8 medium containing 20 μM Y27632 in Primaria cell culture dishes (Corning) and incubated at 37°C under 5% CO₂ for 48 hours. We designated this time point as day-2. After 24 hours of incubation (day-1), 8 mL of fresh E8 medium was added to each dish to promote cell proliferation and aggregate growth. On day 0, to start differentiation, all cell aggregates were individually collected and transferred to new dishes with 3 mL of E6-based differentiation medium containing 2% Matrigel, 10 μM SB431542 (Tocris, Bristol, UK), 4 ng/mL basic fibroblast growth factor (bFGF; R&D Systems, Minneapolis, MN, USA) and 2.5 ng/mL BMP4 (PeproTech, East Windsor, NJ, USA) to initiate non-neural ectoderm formation. On day 3 of differentiation, to induce cranial neural crest (CNC) cell formation, 200 ng/mL LDN-193189 (LDN) (a BMP inhibitor; Sigma Aldrich, St. Louis, MO, USA) and 50 μg/mL bFGF were added in a volume of 1.2 mL, thus making the final volume 3 mL per dish. On day 6 of differentiation, 1.8 mL of fresh E6 medium was added, bringing the final volume to 4.8 mL. Half of the medium was changed (removal of 2.4 mL of spent medium and addition of 2.4 mL of fresh E6 medium) on days 8 and 10. On day 12, to induce self-assembly of the epidermis, all aggregates were transferred into low-attachment plates (Thermo Fisher Scientific, Waltham, MA, USA) in 4.8 mL of organoid maturation medium (OMM) containing 0.5% Matrigel. OMM is composed of advanced Dulbecco’s modified Eagle medium: nutrient mixture F-12 (DMEM/F-12; Gibco, Waltham, MA, USA) and neurobasal medium (Gibco) at a 1:1 ratio, 0.5× B-27 minus vitamin A (Gibco) and 0.5×N-2 (Gibco) supplements, 0.1 mM 2-mercaptoethanol (Gibco) and 100 μg/mL Normocin (InvivoGen, San Diego, CA, USA). On differentiation day 12, half of the spent medium was replenished (removal of 2.4 mL of spent medium and addition of 2.4 mL of fresh medium) with OMM containing 0.5% Matrigel until day 46. On day 47, the whole medium was replenished with fresh OMM without Matrigel. Starting from day 50 (or longer), half of the medium was replenished (removal of 2.4 mL of spent medium and addition of 2.4 mL) with OMM every 3 days.

Total RNA from organoids was extracted using the TRIzol reagent (Invitrogen, Waltham, MA, USA) according to the manufacturer’s instructions. The extracted RNA of all samples was stored at -20°C until quantitative reverse-transcription polymerase chain reaction (qRT-PCR) was performed. The iScript Reverse Transcription Supermix for qRT-PCR (iScript RT Supermix; Bio-Rad, Hercules, CA, USA) was used to synthesize cDNA from total RNA using a Master cycler (Eppendorf, Hamburg, Germany). The concentration of RNA was measured using a Nano Drop spectrophotometer (Thermo Fisher Scientific). Primers were designed using Primer-BLAST (NCBI, https://www.ncbi.nlm.nih.gov/tools/primer-blast). The primer sequences for quantitative PCR are as follows: Lhx2, forward: 5’-GTGGACAAGTCGACAGACGC-3’, reverse: 5’-AGGGTGGGGCTAGTCAAGTC-3’; Pcad, forward: 5’-TCTCGCGTCTCTCCTCCTTC-3’, reverse: 5’-CGCCTCCAAGGTCACTTCAG-3’; Edar, forward: 5’-GCTACCAGATATGCAGGCGT-3’, reverse: 5’-ATGTAGTAGCCAGGGAGGCA-3’; Pdgfrα, forward: 5’-CTGCCTGACATTGACCCTGT-3’, reverse: 5’-CTGCTGGAACCCGTCTCAAT-3’; Ngfr(p75), forward: 5’-AGATCCCTGGCCGTTGGATT-3’, reverse: 5’-AGGTCTTGTTCTGGAGGTGC-3’; Sox2, forward: 5’-TACAGCATGTCCTACTCGCAG-3’, reverse: 5’-GAGGAAGAGGTAACCACAGGG-3’; KRT5, forward: 5’-TCTTGCCGGAGGTAGCAGTG-3’, reverse: 5’-CTGCTTGCACTGAAGCCAGAG-3’; KRT15, forward: 5’-ATGGAGGCTCAGAACCAGGA-3’, reverse: 5’-TCCTGAAGAGGCTTCCCTGA-3’. qRT-PCR was carried out using the StepOne Real-Time PCR System (Thermo Fisher Scientific) with the iTaq Universal SYBR Green Master Mix (Bio-Rad). The conditions were 95°C for 10 minutes followed by 50 cycles at 95°C for 15 s and 60°C for 3 s. Triplicate reactions were carried out for each sample. Gene expression levels were compared statistically using the one-way analysis of variance (ANOVA) (the Kruskal-Wallis test) to quantify differences between experimental groups. p-values <0.05 were considered significant.

In brief, samples were fixed in 4% (v/v) paraformaldehyde overnight at 4°C and then washed in phosphate-buffered saline (PBS). After the tissues were embedded in paraffin, they were cut into 4-μm-thick sections. For immunohistochemical staining, the fixed samples were deparaffinized and antigen retrieval was performed. All samples were blocked in 5% bovine serum albumin (BSA) for 30 minutes and then incubated with primary antibodies overnight at a 4°C refrigerator. The next day, the slides were then incubated with fluorescently labeled secondary antibodies in 1% BSA for 1 hour at room temperature after 3 PBS washes for 10 minutes and counterstained with 4′,6-diamidino-2-phenylindole (DAPI). Images were obtained using confocal microscopy (K1 Fluo; Nanoscope Systems Inc., Daejeon, Korea). We sectioned 5 to 6 organoids on day 110 for immunohistochemistry (IHC) analysis. All images are representative of samples obtained from 4 separate experiments.

The antibodies used in this study were CD49f (mouse, ab20142; Abcam Inc., Cambridge, MA, USA,), cytokeratin 5 (KRT5) (mouse, ab259429; Abcam Inc.), cytokeratin 15 (KRT15) (rabbit, ab52816; Abcam Inc.), cytokeratin 17 (KRT17) (mouse, ab212553; Abcam Inc.), loricrin (rabbit, ab85679; Abcam Inc.), S100β (mouse, ab11178; Abcam Inc.), SOX2 (rabbit, ab97959; Abcam Inc.), tubulin β-3 (Tuj1) (rabbit, B240864; BioLegend, San Diego, CA, USA) and premelanosome protein (PMEL) (rabbit, ab137078; Abcam Inc.), and Ki-67 (mouse, ab15580; Abcam Inc.).

The organoid specimens used for histology and qRT-PCR analysis were selected at random and not screened for quality or specific morphological characteristics. Each experiment was performed 4 times (n=4). All statistics were performed using Excel statistical tools or Prism 5 (GraphPad Software, San Diego, CA, USA). Data are reported as the mean±standard deviation. The statistical significance of differences between groups was analyzed using one-way ANOVA (the Kruskal-Wallis test). A p-value less than 0.05 was considered significant: *p<0.05, **p<0.01, ***p<0.001.

To induce 3D HF skin organoids, we sought to guide hiPSCs toward the surface ectoderm lineage. hiPSCs of the CMC-003 cell line were detached into small pieces of cell clumps and plated on ultra-low attachment plates to create embryoid bodies. We then transferred the embryoid bodies to new plates containing a differentiation medium with key factors, namely bone morphogenetic protein 4 (BMP4) and a transforming growth factor-beta inhibitor (SB431542), to promote epidermal induction (Fig. 1A). This differentiation strategy produced uniform epithelial cysts with a diameter ranging from 500 to 1,000 μm [16]. Next, co-treatment of bFGF and a BMP inhibitor (LDN) was delivered, facilitating the induction of CNC-like cells that give rise to mesenchymal cells. As described elsewhere [16], by day 16, the organoids had an exterior layer of CNC-like cells surrounding the epithelial cells. Remarkably, from day 18, the organoids experienced bipolarization, with an emerging asymmetric structure with a head (the pole composed of epidermal cyst) and tail (the opposite pole characterized by an opaque cell mass) [16]. By day 30, lipid-rich regions emerged in epithelial cysts (Fig. 1B). With extended culture, the tail structure shrank gradually and disappeared in some organoids, eventually forming a matured structure of skin organoids that comprised different cell types. Theoretically, the development of human HFs initiates around weeks 8-10 of gestation; therefore, the culture period was prolonged, allowing the skin construct to form hair bud structures. By day 60, hair-germ-like buds protruded radially on the surface of attached organoids. On day 110 of culture, these hair placode germs significantly extended.

To confirm whether our HF skin organoids morphologically resembled in vivo tissue, we assessed the mRNA expression of different markers through qRT-PCR analysis. The results showed that organoids with protruding hair placodes expressed epidermal germ cell markers, including LIM homeobox protein 2 (LHX2), P-cadherin (PCAD), and ectodysplasin A receptor (EDAR) (Fig. 2B). The expression level of LHX2 and EDAR, which was similar in the hiPSC and embryoid body day 5 (EB5) stages, significantly increased at day 100. Meanwhile, the PCAD expression level at EB100, which was relatively similar to that of EB5, was significantly reduced in comparison with the initial stage. In addition, a significantly increased level of platelet-derived growth factor receptor alpha (PDGFRα) and the neuroglia-associated marker NGFR (P75) was detected in embryoid bodies by day 100, proving the formation of dermis (Fig. 2C). The high transcription of the marker genes of interest within the constructed skin organoids demonstrated the results of stratification during this period, which constituted the fundamental layers of the skin.

To further demonstrate whether the 3D organoids resembled and mimicked skin maturation and HF formation in native tissues, we confirmed their architecture based on the expression of key proteins. Specifically, KRT5 (a basal cell marker) and KRT15 (a peridermal marker) were highly expressed on EB100 (Fig. 3A), which was also revealed in IHC analysis results (Fig. 3B). Additionally, the formation of the basal layer and granular layer of epidermis was demonstrated, respectively characterized by CD49F and loricrin as markers (Fig. 3C). Regarding HFs, a growth in the bulge stem cells and outer root sheath in our skin organoids was proven by the elevated level of KRT15 by day 100 (Fig. 3A). In additions, qRT-PCR analysis suggested a remarkably enhanced level of Sox2 transcription at the mature stage, indicating the presence of dermal condensate and dermal papilla cells (Fig. 3D). These cells were shown to accompany the periderm-like layers (KRT5⁺), immunolabeled by the co-expression of Sox2 (Fig. 3E). Furthermore, we observed that this transcription factor interwove with KRT17⁺ cells (Fig. 3F), demonstrating that HF precursors were distributed in proximity to the epithelium and HFs. Together, these results demonstrated that the skin organoids were a facsimile of native skin tissue in terms of higher-order skin morphogenesis and HF formation during the maturation process.

To assess the formation of HFs within the generated 3D skin organoids, we examined the construction of prominent cells and the neuron network. Specifically, we identified the proliferation of HF stem cells on EB100, as evident from the accelerated level of transcription of nuclear factor of activated T cells 1 (NFATC1) and leucine-rich repeat-containing G-protein coupled receptor 5 (LGR5), compared to hiPSCs and EB5 (Fig. 4A). The development of the Merkel cell population, indicated by a significantly increased level of KRT20 transcripts, were also detected in mature skin organoids (Fig. 4B). Since organoid HFs are pigmented [17], we assessed the growth of melanocytes through the expression of pattern of selected melanocytes (PMEL). qRT-PCR analysis of organoids with protruding placodes showed a significant increase in the PMEL level by day 100 (Fig. 4C). The result was confirmed by IHC, as PMEL and Ki-67 were expressed highly with a similar pattern (Fig. 4D), implying the proliferation of pigmented cells. The results, hence, suggested that HF skin organoids can recapitulate the skin in terms of pigmentation.

Next, the contact of constituted HFs to neuronal cells was considered. We observed that early neurons were characterized by Tuj1-wrapped HFs (KRT17⁺) and contacted the surrounding epithelial cells (KRT17⁺) (Fig. 4E). Here, a neuron cell body was clearly exhibited, accompanied by the scattered distribution of a neuron network over the surrounding region that enhanced the connection to the HF. Furthermore, staining results revealed that satellite glial and Schwann cells (S100β⁺) were distributed surrounding the outer root sheath of the HF (KRT15⁺) and elongated to their ends, thereby extending their connection (Fig. 4F). Concurrently, the distribution of these sensory cells overlapped the region of epithelial cells (KRT15⁺) (Fig. 4F). In summary, the data suggested that our skin organoids, with the formation of precursor and pigment cells accompanied by a sensory neuron network, were sufficient to physiologically recapitulate native human skin tissue.