1 Introduction

The severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) infection affects multiple organs such as the lungs, gut, kidney, brain, and skin.[1-4] Previous reports have shown that coronavirus disease 2019 (COVID-19) patients developed acral areas of erythema with vesicles or pustules (pseudochilblain), acral areas with erythematous rash, widespread urticaria, chickenpox-like vesicles, and maculopapular eruptions.[5, 6] In addition, it has been reported that depending on the clinical severity of the disease COVID-19 patients experienced hair loss on their scalp.[7] However, there is no evidence yet about any abnormality observed in the hair follicles of the COVID-19 patients.

Epidermal stem cells (EpSCs) of the basal layer of skin comprise of two types of cells including interfollicular epidermal stem cells and hair follicle stem cells (HFSCs), which maintain homeostasis and wound healing in the skin. EpSCs have undifferentiated proliferative progenitor cells expressing keratins, including keratin 5 (K5) and keratin 14 (K14).[8] These progenitors not only replenish the basal layer through self-renewal, but also progressively migrate upward through the epidermis, differentiating to form mature keratinocytes expressing keratin 1 (K1), keratin 10 (K10), and involucrin, and finally the outer layers of the terminally differentiated, dead stratum corneum cells.[9] EpSCs can be linked to the adjacent basement membrane (BM) zone through integrin α3β1-rich focal adhesions and hemidesmosome complex including type IV collagens, nidogens, laminins, etc. to maintain their proliferative capacity and their ability to migrate in response to injury.[10] In addition, cell-cell interaction associated with proteins including E-cadherin (ECAD) and desmosome can coordinate cytoskeleton dynamics and directed movements. Loss of these intercellular and cell-matrix junctions could lead to the impairment of the microenvironment and the mechanical strength of EpSCs. During hair follicle formation, the epidermis first forms placodes with dermal condensates. Then, cells of the placode form hair germs, and the matrix cells encapsulate the dermal papilla (DP). The matrix produces a three-layered inner root sheath (IRS) and the outer root sheath (ORS), whose cells contact the BM zone directly.[10] Wnt/β-catenin signaling is crucial for the maintenance of interfollicular epidermal stem cells along with self-renewal and formation of hair follicles.[11] Also, the Wnt signal is involved in the central nervous system, directing progenitor neuronal cells into neuronal characteristics.[12] Interestingly, it has been reported that hair follicles are the important target tissues for classical neurohormones, neurotrophins, and neuropeptides, signifying the communication between hair follicles and the nervous system.[13]

Organoids displaying key structural and functional features of natural tissues and organs can be grown, mimic the physiological characteristics of tissue growth and wound repair. Therefore, organoids are widely used for the investigation of mechanisms of tissue regeneration, disease model, and drug screening.[14] Several kinds of organ models have been established, such as intestine,[15] brain,[16] kidney,[17] lung,[18] liver,[19] pancreas,[20] heart,[21] retina,[22] and skin.[23] During the COVID-19 pandemic, multiple human pluripotent stem cell (hiPSC)-derived organoids have been used to model SARS-CoV-2 infection in many organs.[24-28] This powerful and revolutionary technology has promoted the emergence of new biomedical models, as well as the research and treatment of various endogenous and exogenous injuries or infectious diseases in human beings. However, so far, there has been no report about the use of 3D skin models for studying the pathological mechanisms in SARS-CoV-2 infected skin tissues. In the present work, we have established a hiPSC-derived skin organoid with hair follicles and nervous system, which is sensitive to external damage. The skin organoid was used to investigate the cell types affected by SARS-CoV-2 infection and the pathological characteristics associated with a skin infection. Based on the results, we present evidence of the functional consequences of skin infected by SARS-CoV-2 at cellular and molecular levels.

2 Results 2.1 Comprehensive Proteomics Landscape of Clinical COVID-19 Skin Samples

Previous studies have shown that the skin can be infected by SARS-CoV-2.[29] Our results revealed that the lymphocyte infiltration was observed in COVID-19 skin—particularly adjacent to the epidermis and accessory glands in the dermis (Figure 1a and Figure S1a, Supporting Information). These infiltrating lymphocytes includes CD3+/CD8+ T cells and CD68+ macrophages but not CD4+ T cells, CD19+/CD20+ B cells or myeloperoxidase- positive neutrophils (Figure S1b, Supporting Information). In addition, capillary endothelial cells swell in dermis and organized microthrombus could be see occasionally (Figure S1a, Supporting Information). Furthermore, the staining of spike and nucleoprotein (NP) proteins, as well as the electron microscopy results confirmed the presence of SARS-CoV-2 particles in the COVID-19 skin tissues (Figure 1a). An integrated quantitative proteomics approach was used to detect proteomic changes in the SARS-CoV-2 infected skin (Figure 1b). Principal component analysis (PCA) revealed that the proteins identified in the skin tissues from the patients with COVID-19 and the controls formed independent clusters (Figure S1c, Supporting Information). Functional analysis showed that the proteins upregulated during SARS-CoV-2 skin infection were mainly enriched in the biological processes involved in the response to microorganisms (bacteria, fungi, and viruses), apoptosis, immune response, and estrogen associated signaling; while the downregulated proteins were related to regulation of blood circulation, tissue remodeling and the development of epidermis, nervous system, and blood vessels, as compared to the control (Figure 1c). In addition, the most downregulated proteins in the cellular components category were associated with extracellular matrix (ECM), focal adhesion, cell-cell junction, myelin sheath, melanosome, and cytoskeleton (Figure S1d, Supporting Information). These results indicate that ECM microenvironment, epithelial development, nervous system, circulation system, as well as cell junctions in skin tissue might be affected in patients with COVID-19.

Quantitative proteome profiling of protein signatures in COVID-19 skin tissues. a) H&E, spike, and NP proteins staining, and transmission electron microscope analysis of SARS-CoV-2 infected skin tissues (Scale bar: 200 and 100 µm, 500 and 200 nm). The white solid arrows indicate the SARS-CoV-2 proteins in COVID-19 skin tissues. b) Schematic representation of the experimental workflow of the skin organoids culture, SARS-CoV-2 infection, quantitative proteomics, bioinformatics analysis, and biological validation. c) Biological process analyses of the upregulated and downregulated-expressed proteins between COVID-19 (n = 5) and control (n = 6) skin tissues. Upregulated-expressed and downregulated-expressed proteins: t-test, Benjamini–Hochberg (BH) adjusted p-value < 0.01 and log2COVID-19/Control > 1, and BH adjusted p-value < 0.01 and log2COVID-19/Control < −1.

2.2 Formation of Skin Organoids with Various hiPSC-Derived Skin Cells

Next, to study the different cell types involved in the COVID-19 skin, a skin organoid model was constructed using hiPSCs and subjected to SARS-CoV-2 infection (Figure 2a). To obtain uniform epithelial cysts, a stepwise differentiation of hiPSCs into ectoderm was achieved through the treatment with bone morphogenetic protein 4 (BMP4) and a TGFβ inhibitor (SB431542) on Matrigel, following the previous study[23] (Figure S2a, Supporting Information). From day 16 to 35, the organoids displayed TFAP2A+ ECAD+ epithelial cells and two subtypes of cranial neural crest (CNC)-like cells, expressing mesenchyme-associated marker, either platelet-derived growth factor receptor alpha (PDGFRα) or the neuroglia-associated markers (SOX10 and P75) (Figure 2a,b). Remarkably, by day 35, the organoids displayed asymmetric structures, which were divided into head and tail (Figure 2a). With continuous culture, the tail structure decreased gradually and would disappear in some organoids, and finally the skin organoids gained matured structure with different cell types. By day 55, the KRT15+ KRT5+ epithelial-like stem cells stably expanded and the tail portion was reduced. Further, the organoids presented hair follicle structures, expressing the epidermal germ cell markers (LIM Homeobox Protein 2 [LHX2], P-cadherin (PCAD), and ectodysplasin a receptor (EDAR)) and showed the presence of dermal condensates (SOX2, PDGFRα, and P75) by day 75 (Figure 2a,b), indicating that the hair follicle structures were formed during this period. Our previous study showed that TGFβ induced protein (TGFBI) can enhance epidermal stem cell growth (Figure S2b, Supporting Information). After 15 d, the 3D/organoid cultures treated with TGFBI for a month, exhibited the typical stratified epithelium structure, with the expression of epidermal basal cell markers (P63, KRT14, and KRT5), mature epithelium markers (KRT10, KRT1, and involucrin), adhesive receptor (ITGB1) and desmosomes (DSG1 and DSG2) (Figure 2d and Figure S2c, Supporting Information). In addition, the organoids are also comprised of the components of the dermal-epidermal junction zone, called basement membrane, which is the key niche of EpSCs, such as laminin, type IV collagen, and nidogen (Figure S2d, Supporting Information). Dermal components (Aggrecan and COL2A1) were also identified in these organoids (Figure S2e, Supporting Information). In addition, mature hair follicle structures were also formed in the skin organoids (Video S1, Supporting Information).

Formation of skin organoids in vitro. a) A schematic overview of the hiPSC-skin organoid culture. The bright field representing organoids culture on days 16, 35, 55, 75, and 90 (Scale bar: 400 µm). b) Immunofluorescence of the epithelial cells markers (TFAP2A, ECAD, and KRT17) at day 15–35, mesenchyme-associated marker (PDGFRα), neuroglia-associated markers (SOX10 and P75), proliferation marker (PCNA), epithelial stem cells markers (KRT15 and KRT5) at day 55, epidermal germ cell markers (LHX2, PCAD, and EDAR), dermal condensates (PDGFRα and P75) at day 75. c) Bright field and immunofluorescence represent hair follicles of skin organoids at day 90 (Scale bar: 200 and 50 µm). d) Immunofluorescence of the epithelial cell marker (ECAD), epidermal basal cells markers (KRT15, KRT5, KRT14, and P63), mature epithelium markers (KRT1 and KRT10), proliferation marker (PCNA), desmosome (DSG1), and adhesive receptor (ITGB1) at day 120 (Scale bar: 200 and 50 µm).

To further prove that the skin organoids resemble and mimic the physiological function of human skin tissue, we established the proteomic profile of mature skin organoids at day 140 after the culture was established (Table S1, Supporting Information). Results showed that the proteins identified in these organs were mainly enriched in epidermal cells, nerve cells, stromal cells, fibroblasts, muscle cells, and vascular-related cells, after a comparison with the reported database (Figure 3a). Biological process analysis showed that the proteins identified from skin organoids were mainly enriched in the processes of skin development including epidermal development, nervous system development, and stem cell division (Figure 3b). Several signaling pathways including canonical and noncanonical Wnt, epidermal growth factor receptor, fibroblast growth factor (FGF), and TGFβ and TGFg associated with skin development were found to be enriched in the skin organoids. Especially, proteins associated with epidermal development that participated in the processes of establishment of the skin barrier (e.g., KRT1 and KRT16), epithelial cell differentiation (e.g., KRT14 and KRT4), and melanosome were identified in the skin organoids (Figure 3c and Figure S3a, Supporting Information). In addition, skeletal tissue associated processes including ECM and cytoskeleton organization and adhesion, as well as the lipid and energy metabolism associated processes were enriched in the skin organoids (Figure S3b, Supporting Information). 111 ECM proteins including core ECM proteins like 15 collagens, 31 glycoproteins, and 10 proteoglycans, as well as ECM associated proteins including 16 ECM regulators, 18 ECM-affiliated proteins and 21 secreted factors, were identified (Figure S3c and Table S2, Supporting Information). These results indicate that we successfully constructed skin organoids with multiple structures and cellular functions.

Quantitative proteome profiling of protein signatures in skin organoids. a) Proteins identified in the skin organoids (n = 3) were associated with different cell types according to the cell type signature gene sets in the Molecular Signatures Database (cell type enrichment p-value < 0.001). b) Functional analysis of proteins identified in the control skin organoids. c) Heatmap analysis of the proteins identified in the control skin organoids that were associated with the establishment of the skin barrier and epithelial cell differentiation. Red and blue boxes indicate proteins with high and low intensities, respectively. Exp1, Exp2, and Exp 3 represent different biological repeat of the proteomics experiment (n = 3).

2.3 SARS-CoV-2 Can Infect the Hair Follicle

COVID-19 patients are prone to hair loss.[7] Further, we found that proteins associated with hair follicle development (KRT33B, KRT31, KRT71, KRT14, KRT17, CTNNB1, etc.) and epidermal development were significantly downregulated in the skin tissues of COVID-19 patients (Figure 4a). We previously found that BM structure is an important microenvironment supporting the fate and functional polarity of EpSCs.[30] We observed that the proteins (type IV collagen, laminins, HSPG2, NID1, etc.) of BM structure and hemidesmosome complex (ITGA6) were severely damaged in the COVID-19 skin tissues (Figure 4a). Immunofluorescence staining also showed that the expression of KRT14, KRT10, proliferating cell nuclear antigen (PCNA), P63, COL4A1, and ITGA6 downregulated in the COVID-19 skin tissues or organoids, compared with the control group, which was consistent with the proteomics results (Figure 4b and Figure S4a, Supporting Information). These results suggest that the hair follicle development in COVID-19 patients could be affected by SARS-CoV-2. To further examine if hair follicle might be one of the targets of SARS-CoV-2, we induced SARS-CoV-2 infection in these skin organoids.

SARS-CoV-2 infects hair follicles. a) An interactome network of downregulated proteins from COVID-19 skin tissues, when compared to the control group. b) Immunofluorescence of PCNA, KRT14, KRT10, and hemidesmosome complex (ITGA6) in the COVID-19 and control skin tissues (Scale bar: 50 µm). c) A schematic of hair follicle construction. Immunofluorescence of hair follicles’ markers with different construction: d) Mtx (Ki67 and PMEL), e) DP (SOX2 and PDGFα), DS (αSMA and ITGA8), ORS (KRT15), Cp (KRT17), IRS (Ki67, KRT71, GATA3, and KRT74), Ch (AE13 and KRT82), Co (AE13), and Me (KRT5 and KRT75) (Scale bar: 50 µm). Mtx: matrix; DP: dermal papilla; DS: dermal sheath; ORS: outer root sheath; Cp: companion layer; IRS: inner root sheath; Ch: cuticle; Co: cortex; Me: medulla. f) In the normal organoids at day 140. Immunofluorescence of g) SARS-CoV-2 proteins (NP and spike), KRT71 and KRT71, h) spike and PDGFα in the SARS-CoV-2 infected organoids (Scale bar: 10 µm). i) Immunofluorescence of NP and KRT17 in the COVID-19 skin tissues (Scale bar: 10 and 100 µm). j) Downregulated proteins between COVID-19 (n = 3) and control skin tissues (n = 3) with functions of the nervous system, epidermal development, and cell cycle in the skin organoids were associated with different cell types according to the cell type signature gene sets in Molecular Signatures Database (cell type enrichment p-value < 0.001). Downregulated-expressed proteins: t-test, BH adjusted p-value < 0.01, and log2COVID-19/Control < −1.

After the mature hair follicle structure was formed with the germ elongated into a peg-like structure, in the skin organoids, Ki67+ epithelial cells were found to express premelanosome protein (PMEL) around the DP structure along with markers SOX2 and PDGFRα (Figure 4c–e). Further, we found that the hair follicles of skin organoids shared an outermost layer of alpha-smooth muscle actin (αSMA) + integrin alpha 8 (ITGA8) + dermal sheath with characteristic arrector pili muscle. In addition, ORS markers (KRT15), ceruloplasmin markers (KRT17), IRS makers (Ki67, KRT71, GATA3, and KRT74), cuticle and cortex markers (AE13 and KRT82), and medulla markers (KRT5 and KRT75) of hair follicles were also found to be expressed in the skin organoids (Figure 4f). This indicates that all the hair follicle lineages formed from the outermost dermal sheath to the center of the medulla were well established in the organoids. The proteins associated with hair follicle development, hemidesmosome complex, desmosome, as well as BM were successfully identified in these skin organoids (Figure S4b, Supporting Information). Next, we exposed the mature organoids (day 140) to SARS-CoV-2. Testing the target structure of SARS-CoV-2 in these organoids revealed that SARS-CoV-2 could mostly target the KRT17+ hair follicle cells (Figure 4g and Figure S4c,d and Video S2, Supporting Information), however, less target the KRT71+ and PDGFRα+ hair follicle cells (Figure 4h). Pathological staining of the skin tissues infected by SARS-CoV-2 further proved that the virus could chiefly infect the KRT17+ hair follicle cells (Figure 4i and Figure S4e, Supporting Information). Besides, the downregulated proteins in the COVID-19 skin tissues were mainly enriched in EpSCs, cell cycle, and epidermal development (Figures 4j and 5a). The epithelization morphology of epidermal cells around KRT17 + hair follicles disappeared and the ability of cell proliferation decreased in the SARS-CoV-2 infected organoids, compared to the normal controls (Figure 5b,c). According to the pathological features, it is found that the cell cycle of infected organoids could be caused by DNA damage after SARS-CoV-2 infection (Figure 5a). We found that H2A.X proteins in the nucleus of the epidermal cells around the hair follicles in the infected organoids were phosphorylated, indicating that SARS-CoV-2 infection could lead to double-strand breaks (DSB) of the epidermal cells around the hair follicles (Figure 5d). Further, results also showed that a redox sensitive transcription factor (NRF2) were activated in the epidermal cells with DNA damage, suggesting that DSB could be caused by oxidative stress after SARS-CoV-2 infection (Figure 5d). These results indicate hair follicles could be the target domain of SARS-CoV-2 in the COVID-19 skin, which could lead to the polarity and function of epidermal cells around hair follicles, possibly by affecting DNA damage and oxidative stress of the epidermal cells.

Pathological features of SARS-CoV-2 infected skin organoids. a) Functional analyses of the upregulated and downregulated-expressed proteins identified in the SARS-CoV-2-infected skin organoids compared with the controls. Upregulated-expressed and downregulated-expressed proteins: t-test, BH adjusted p-value <0.01 and log2Infected/Control >1, and BH adjusted p-value < 0.01 and log2Infected/Control < −1. b) Immunofluorescence of β-catenin, epithelial cell marker (ECAD) and KRT17, c) proliferation marker (PCNA) and Cyclin D1, and d) H2A.Xser and NRF2 (Scale bar: 100 and 50 µm).

2.4 SARS-CoV-2 Can Infect the Nervous System of Skin

We have found that the function of the nervous system in the COVID-19 skin tissues was severely affected (Figure 1c). Therefore, we investigated whether SARS-CoV-2 could directly infect the nerve cells in the skin. The cultured skin organoids showed the presence of the nervous system, expressing pan-neuronal marker tubulin β class III (TUJ1), especially around the hair follicles (Figure 6a). Neurofilament heavy chain (NEFH), peripherin (PRPH), and insulin gene enhancer protein (ISL1) were found to be expressed in the neuronal network along with S100 Calcium Binding Protein B (S100β)+ Schwann-like cells and parvalbumin (PVALB)+ neuron somas (Figure 6a–c and Video S3, Supporting Information). The proteins associated with the development of glial cells, axons, and neurons were successfully identified in the skin organoids. (Figure 6d). After SARS-CoV-2 infection, we found that SARS-CoV-2 could target TUJ1+, NEFH+, PRPH+, and PVALB+ neurons (Figure 6e,f and Video S4, Supporting Information). S100β+ Schwann-like cells were also found to be infected by SARS-CoV-2 (Figure 6f). Further, the COVID-19 skin tissues, also showed the TUJ1+, NEFH+, PRPH+, PVALB+, and S100β+ nerve cells were infected by SARS-CoV-2 (Figure 6g and Figure S4f, Supporting Information). Besides, the downregulated proteins in the COVID-19 skin tissues were mainly enriched in neuroendocrine cells, microglia, astrocytes, satellite cells, neuron and nervous development (Figure 4j). These results indicate that the nervous system could also be the target of SARS-CoV-2 in the COVID-19 skin.

SARS-CoV-2 infects the nervous system of the skin. Immunofluorescence of a nervous marker with different cell types: a) KRT17, TUJ1 (pan-neuronal marker), NEFH (neurofilament heavy chain), and PRPH (peripherin), b) KRT20, TUJ1, and ISL1, and c) PVALB, S100β, and NEFH in the normal skin organoids (Scale bar: 400 and 50 µm). d) Heatmap analysis of the proteins identified in the skin organoids that were associated with the development of glial cells, axons, and neurons. Red and blue boxes indicate proteins with high and low intensities, respectively. Immunofluorescence of e) spike and ISL1, f) spike, NEFH, PRPH, PVALB, and S100β in the SARS-CoV-2 infected skin organoids (Scale bar: 50µm). g) Immunofluorescence of the spike, TUJ1, NEFH, PRPH, PVALB, and S100β in the COVID-19 skin tissues (Scale bar: 10 µm). Exp1, Exp2, and Exp 3 represent different biological repeat of the proteomics experiment (n = 3).

3 Discussion

In tissues with a regenerative capacity, the body needs to mobilize a variety of cell types for tissue repair during the process of wound healing, such as activating quiescent reserve stem cells and stromal cells secreting matrix proteins for tissue remodeling. Skin is the first line of defense against external damage in the human body. It is the first organ exposed to chemical damage (e.g., radiation and ultraviolet rays), physical damage (e.g., acute or chronic wounds), and microbial infection (e.g., bacteria or viruses). The integumentary system consists of the epidermis, dermis, dermal appendages, such as hair follicle, sweat glands, sebaceous glands, and nerve cell endings. The interaction of different cell types from epidermis and dermis promotes skin self-renewal and regeneration during skin homeostasis and injury. For example, the basement membrane from the dermis can maintain the cell fate and function of epidermal stem cells (EpSCs)[30] during wound healing. The re-epithelialization process requires cytokines, inflammatory factors, and ECM secreted by various types of cells in the dermis to promote the proliferation and migration of EpSCs.[31, 32] In the healing phase of ECM remodeling, the enzyme system secreted by epidermal cells promotes the degradation and assembly of dermal matrix.[30] Therefore, to investigate the mechanism of skin injury, we need to establish a relatively ideal model composed of a variety of cells in in vitro culture system, to find a potential therapeutic method for skin repair.

Here, we constructed a skin organoid model induced by hiPSCs and proved that it can respond well to the microbial damage. Through this model, we first assessed that SARS-CoV-2 can attack hair follicles directly, which could be related to the hair loss in COVID-19 patients reported previously.[7] HFSCs maintain homeostasis partly from the stem cells in the basal layer (EpSCs) in the skin.[33] We found that the proteins associated with the development of EpSCs were acutely downregulated both in the skin tissues from COVID-19 patients and SARS-CoV-2 infected organoids. In addition, we also found the components of BM, an important niche of EpSCs, downregulated, which is consistent with the results from lungs[3] and liver.[2] This might also be one of the reasons for skin blistering in the COVID-19 patients. Results also showed that the proliferation ability of the epidermal cells around hair follicles infected by SARS-CoV-2 decreased, probably through SARS-CoV-2 infection leading to the DSB and oxidative stress, which is consistent with the our previously work.[2, 3] It has also been reported that some virus can attack the hair follicles, for example, the varicella zoster virus (VZV), a highly human-specific virus. Studies showed that hair follicles is the initial site of VZV replication in skin.[34] There were clinical cases reported the relation of the localized hair loss with VZV infection,[35] which is similar with our findings that SAR-CoV-2 infection may cause hair loss. Since the use of 3D skin organoid models for studying the pathological mechanisms in SARS-CoV-2 infection has not been reported so far, it will not be possible to compare our findings with other skin organoids infected by the virus. Further, whether COVID-19 can directly attack HFSCs and lead to hair follicle necrosis requires further investigation.

We also found that SARS-CoV-2 can attack the neurons in the skin, which is consistent with the previous reports that SARS-CoV-2 was able to directly attack the nerve cells in the brain organoids.[24] It has been previously reported that mediators secreted by the nervous system play an important role in hair follicle's growth.[36] Our results also suggest a possible correlation between the somatic nervous system and hair follicle development. Several studies have found that the herpes simplex virus type 1 can infect skin neurons, which could result in release of cytokines and neuropeptides and affect the neural function and immune system.[37] This suggests that SARS-CoV-2 infection of the skin nervous system may lead to the disorder of the central system and immune system. In addition, we also found a considerable depletion of proteins related to vascular development and regulation of blood circulation, which may be caused by abnormal blood coagulation of COVID-19.[2, 3] Vascular development and hair follicle development are also inseparable. Therefore, there are many direct and indirect factors leading to hair loss in COVID-19 patients, which need to be explored further. In conclusion, our research suggests that COVID-19 can directly affect the hair follicles and nerves in the skin. The infection model that we established provides a guideline for subsequent research on the mechanism of SARS-CoV-2 infection and drug screening in COVID-19 patients with affected skin.

Moreover, the skin organoid constructed in this study includes both the surface (epidermis and dermis) and deep (cartilage and subcutaneous fat) structures of skin, which could make it an ideal model for other skin disease research. For example, the patients with scleroderma have high skin fibrosis.[38] In addition to dermal vascular malformation and epidermal appendage atrophy, they also show the phenotype of articular cartilage damage and thin adipose layer. For the disease with such complex clinical phenotypes, it could be a good application to treat it using the skin organoids with complete tissue structure and cell types.

Acknowledgements

J.M., D.G., J.L., and X.L. contributed equally to this work. The authors thank Dr. Mansheng Li from National Center for Protein Sciences (Beijing) for helpful discussions on proteomics data analysis. This work was supported by the National Key Research and Development Program of China (2020YFE0202200), the Chinese Academy of Medical Sciences Innovation Fund for Medical Sciences (CIFMS) (2020-I2M-CoV19-001), and the Hubei Natural Science Foundation for Distinguished Young Scholar (2021CFA050).

Conflict of Interest

The authors declare no conflict of interest.

Author Contributions

L.L. and J.M. conceived the overall study and designed experiments. L.L., M.W., Y.Z., and S.Z. had full access to all the data in the study and took responsibility for the accuracy of the data analysis. L.L., D.G., Q.Z., and L.Lv contributed to the culture of skin organoids. Y.Z., Q.L., and L.M. prepared the skin tissues from COVID-19 patients and control individuals. M.W., J.L., Z.H., H.H., and Y.L. performed viral infection experiments. J.M., L.L., X.L., and Y.Z. performed proteomics experiments and bioinformatics analysis. L.L., D.G., L.Lv, Y.W., and Z.W. performed most of pathological staining experiments. L.L., J.M., J.L., Y.Z., and M.W. wrote and edited the paper. S.Z., Y.Z., Z.W., and Z.H. provided funding support. All authors made important comments to the paper.