Cytoplasmic RNA quality control failure engages mTORC1-mediated autoinflammatory disease

Inborn errors of nucleic acid metabolism often cause aberrant activation of nucleic acid sensing pathways, leading to autoimmune or autoinflammatory diseases. The SKIV2L RNA exosome is cytoplasmic RNA degradation machinery that was thought to be essential for preventing the self-RNA–mediated interferon (IFN) response. Here, we demonstrate the physiological function of SKIV2L in mammals. We found that Skiv2l deficiency in mice disrupted epidermal and T cell homeostasis in a cell-intrinsic manner independently of IFN. Skiv2l-deficient mice developed skin inflammation and hair abnormality, which were also observed in a SKIV2L-deficient patient. Epidermis-specific deletion of Skiv2l caused hyperproliferation of keratinocytes and disrupted epidermal stratification, leading to impaired skin barrier with no appreciable IFN activation. Moreover, Skiv2l-deficient T cells were chronically hyperactivated and these T cells attacked lesional skin as well as hair follicles. Mechanistically, SKIV2L loss activated the mTORC1 pathway in both keratinocytes and T cells. Both systemic and topical rapamycin treatment of Skiv2l-deficient mice ameliorated epidermal hyperplasia and skin inflammation. Together, we demonstrate that mTORC1, a classical nutrient sensor, also senses cytoplasmic RNA quality control failure and drives autoinflammatory disease. We also propose SKIV2L-associated trichohepatoenteric syndrome (THES) as a new mTORopathy for which sirolimus may be a promising therapy.


Introduction
Recognition of microbial nucleic acids by the innate immune system and activation of the type I interferon (IFN) response are fundamental mechanisms of host defense against infection. Inappropriate activation of IFN due to inborn errors of nucleic acid metabolizing enzymes can also drive a variety of autoimmune and autoinflammatory diseases (1). One classical example is TREX1/DNaseIII deficiency, which causes accumulation of self-DNA in the cytosol and activation of the DNA-sensing cGAS/STING/IFN pathway. TREX1 mutants are associated with clinical diseases such as Aicardi-Goutières syndrome and systemic lupus erythematosus (2). Similarly, the accumulation of self-RNA activates the RIG-I-mediated IFN response. The SKIV2L RNA exosome has been shown to degrade self-RNAs in cells undergoing ER stress that produce IRE1-cleaved immunostimulatory self-RNA fragments (3). However, physiological functions of the SKIV2L RNA exosome at homeostasis and in vivo remain unknown.
The RNA exosome is an evolutionarily conserved intracellular 3′ to 5′ RNA degradation complex in eukaryotes, and it is involved in RNA processing, maturation, surveillance, and turnover (4). Mammalian RNA exosomes contain shared catalytic subunits as well as distinct cofactor complexes that bring in different RNA substrates (5). The cofactor for the cytoplasmic RNA exosome is the super-killer (SKI) complex, which contains a Ski2-like RNA helicase (SKIV2L), tetratricopeptide repeat domain 37 (TTC37), and WD repeat domain 61 (WDR61; ref. 6). SKIV2L is an RNA helicase that unwinds RNA substrates and threads them through the RNA exosome for degradation, while TTC37 and WDR61 contribute to the structure and activity of the SKI complex. The SKI complex directly associates with the 80S ribosome and extracts mRNA from stalled ribosomes for degradation as part of ribosome-associated mRNA surveillance (7,8). Mutations in either the SKIV2L or TTC37 gene are associated with a rare inherited autosomal recessive disorder, trichohepatoenteric syndrome (THES) (OMIM 222470 THES1, 614602 THES2; refs. 9, 10). THES is clinically characterized by intrauterine growth retardation, intractable diarrhea, primary immunodeficiency, liver diseases, and skin and hair abnormalities (11). However, the molecular mechanism of THES remains unknown and no targeted treatment is currently available.
Here, we report that Skiv2l deficiency causes skin-specific autoinflammation without triggering innate IFN signaling. Loss of SKIV2L disrupts both epidermal and T cell homeostasis in a cell-intrinsic manner through the mTORC1 pathway, which can be targeted to ameliorate disease pathology associated with Skiv2l deficiency.
Inborn errors of nucleic acid metabolism often cause aberrant activation of nucleic acid sensing pathways, leading to autoimmune or autoinflammatory diseases. The SKIV2L RNA exosome is cytoplasmic RNA degradation machinery that was thought to be essential for preventing the self-RNA-mediated interferon (IFN) response. Here, we demonstrate the physiological function of SKIV2L in mammals. We found that Skiv2l deficiency in mice disrupted epidermal and T cell homeostasis in a cell-intrinsic manner independently of IFN. Skiv2l-deficient mice developed skin inflammation and hair abnormality, which were also observed in a SKIV2L-deficient patient. Epidermis-specific deletion of Skiv2l caused hyperproliferation of keratinocytes and disrupted epidermal stratification, leading to impaired skin barrier with no appreciable IFN activation. Moreover, Skiv2l-deficient T cells were chronically hyperactivated and these T cells attacked lesional skin as well as hair follicles. Mechanistically, SKIV2L loss activated the mTORC1 pathway in both keratinocytes and T cells. Both systemic and topical rapamycin treatment of Skiv2l-deficient mice ameliorated epidermal hyperplasia and skin inflammation. Together, we demonstrate that mTORC1, a classical nutrient sensor, also senses cytoplasmic RNA quality control failure and drives autoinflammatory disease. We also propose SKIV2L-associated trichohepatoenteric syndrome (THES) as a new mTORopathy for which sirolimus may be a promising therapy.
Cytoplasmic RNA quality control failure engages mTORC1-mediated autoinflammatory disease pathologies observed in iSkiv2l -/mice ( Figure 2B). No histological abnormality was observed in the vehicle-treated contralateral skin ( Figure 2B). In addition, we observed a clear boundary of skin epidermal thickening only in the 4-OHT-treated area but not in the adjacent nontreated area ( Figure 2C). We also observed lymphadenopathy of inguinal lymph nodes (LNs) from 4-OHT topicaltreated but not vehicle-treated skin ( Figure 2D). These data suggest that inducible Skiv2l gene deletion locally in the adult mouse skin is sufficient to cause tissue pathology.
Second, we crossed Skiv2l fl/fl with K14-Cre mice and generated keratinocyte-specific Skiv2l knockout mice (constitutive Cre expression driven by K14 promoter). Keratinocyte-specific depletion of SKIV2L protein was confirmed by Western blot using Skiv2l fl/fl K14-Cre P0 pup epidermis ( Figure 3A and Supplemental Figure 3A). Skiv2l fl/fl K14-Cre mice and littermates were born at expected Mendelian ratio ( Figure 3B). However, only Skiv2l fl/fl K14-Cre newborn pups showed abnormal 'glassy' skin and died within 24 hours after birth with 100% penetrance ( Figure 3C). Although P0 pups have no body hair, we found that Skiv2l fl/fl K14-Cre pups lacked whiskers or developed distorted thin and fine whiskers ( Figure 3D), suggesting defective hair morphogenesis. Histological analysis revealed thickening of Skiv2l fl/fl K14-Cre epidermis, particularly in the spinous layer, while no apparent immune infiltration was observed in the dermis ( Figure 3E). Skiv2l fl/fl K14-Cre epidermis contained increased proliferating basal keratinocytes compared with controls, as evidenced by positive Ki67 staining and cell-cycle analysis by flow cytometry (Figure 3, F and G). Fluorescence immunohistochemistry analysis further revealed substantial expansion of keratinocytes expressing basal K14 and K5 as well as suprabasal K10 in Skiv2l fl/fl K14-Cre compared with Skiv2l fl/fl epidermis ( Figure 3H). Toluidine blue exclusion assay revealed impaired barrier integrity of Skiv2l fl/fl K14-Cre mouse skin ( Figure  3I). Together, these results suggest that the SKIV2L RNA exosome acts cell-intrinsically in basal keratinocytes to maintain a highly controlled proliferative program during epithelial stratification, and loss of SKIV2L causes uncontrolled expansion of the basal layer and disruption of skin barrier function.
Skiv2l knockout does not activate IFN signaling in vivo. We next assessed whether Skiv2l knockout activates IFN signaling in vivo. In disease-affected Skiv2l fl/fl K14-Cre P0 pup epidermis, we did not detect any increase in the expression of IFN-stimulated genes (ISGs) compared with littermate controls (Supplemental Figure 3B). We also isolated primary keratinocytes from uninduced Skiv2l fl/fl UBC-Cre/ERT2 and Skiv2l fl/fl neonates, then induced Skiv2l deletion ex vivo by treating cell culture with 4-OHT (Supplemental Figure 4, A and B). We did not observe significant changes in the expression of IFN genes, ISGs, or inflammatory genes, indicating the lack of cell-intrinsic IFN signaling when Skiv2l function is lost in keratinocytes (Supplemental Figure 4C). We further generated myeloid-specific Skiv2l knockout mice Skiv2l fl/fl LysM-Cre (Supplemental Figure 5, A-D). Skiv2l fl/fl LysM-Cre mice were born at the expected Mendelian ratio (data not shown) and exhibited normal skin and hair with no signs of skin inflammation or any immunopathology (Supplemental Figure 5, E and F). Cytokines and chemokines in the serum of Skiv2l fl/fl LysM-Cre mice were also indistinguishable from those of Skiv2l fl/fl controls (Supplemental Figure 5G). Together, these data suggest that Skiv2l knockout does

Results
Loss of SKIV2L causes epidermal hyperplasia and impairs skin barrier integrity. Germline whole-body deletion of Skiv2l resulted in early embryonic lethality before E13.5 (Supplemental Figure 1; supplemental material available online with this article; https:// doi.org/10.1172/JCI146176DS1). We therefore generated a loxPflanked Skiv2l conditional allele and bred Skiv2l fl/fl mice with UBC-Cre/ERT2 mice to generate Skiv2l fl/fl UBC-Cre/ERT2 mice, which allow postnatal whole-body deletion of Skiv2l by tamoxifen administration at 4 weeks of age ( Figure 1A, we call this iSkiv2l -/throughout). iSkiv2l -/mice have reduced body weight compared with Skiv2l fl/fl littermate controls (Supplemental Figure  2A). iSkiv2l -/mice developed phenomenal skin lesions and hair loss 4 weeks after tamoxifen injection (at 8 weeks of age, Figure  1B and Supplemental Figure 2B). Skin redness, scaly plaques, and hair abnormalities were first observed in the skin around the eyes, nose, and neck as early as 2 weeks after induction of Skiv2l deletion (Supplemental Figure 2C). By 8 weeks after tamoxifen administration, iSkiv2l -/mice lost most of their body hair, then regrew thin hypopigmented hairs ( Figure 1B). Histopathological examination of iSkiv2l -/dorsal skin revealed striking thickening of the epidermis (epidermal hyperplasia) and massive immune infiltrates in the dermis ( Figure 1C). We also observed dystrophic anagen hair follicles (HFs) in iSkiv2l -/skin (Supplemental Figure 2D). Surprisingly, we did not observe significant inflammation in duodenum, colon, liver, or kidney of iSkiv2l -/mice, nor any changes in cytokines and chemokines in the serum compared with Skiv2l fl/fl littermate controls (Supplemental Figure 2, E and F), suggesting that the inflammation is restricted to the skin even though the Skiv2l gene is deleted in the whole body.
The prominent epidermal hyperplasia in the iSkiv2l -/mouse skin prompted us to examine keratinocyte proliferation and differentiation. Normally, keratinocytes in the epidermis continuously differentiate from a proliferative state at the basal layer to a nondividing state at the cornified layer at the top of the skin. This process is tightly regulated to ensure epidermal barrier integrity ( Figure  1D). Fluorescence immunohistochemistry analysis of proliferation marker Ki67 showed more proliferating keratinocytes at the basal layer in iSkiv2l -/epidermis compared with that in Skiv2l fl/fl controls ( Figure 1E). Cell cycle analysis of keratinocytes by flow cytometry also revealed increased mitotic cells in iSkiv2l -/epidermis compared with Skiv2l fl/fl control ( Figure 1F). Basal keratinocyte markers keratin 14 (K14) and keratin 5 (K5) as well as suprabasal postmitotic marker keratin 10 (K10) all showed massive expansion in iSkiv2l -/mice ( Figure 1G), suggesting dysregulation of epidermal stratification. Functionally, iSkiv2l -/mouse skin displayed impaired epidermal permeability as evidenced by the penetration of toluidine blue dye through the lesioned skin ( Figure 1H), as well as increased transepidermal water loss ( Figure 1I).
SKIV2L acts cell intrinsically in keratinocytes. We next analyzed the role of SKIV2L in keratinocytes in 2 additional knockout mouse models. First, we performed local deletion of Skiv2l by topically applying 4-hydroxyltamoxifen (4-OHT, the active metabolite of tamoxifen) on the right dorsal flank of the Skiv2l fl/fl UBC-Cre/ERT2 mice. We also applied vehicle (DMSO) on the left side of the same mouse as a control (Figure 2A). Topical treatment with 4-OHT induced epidermal thickening and immune infiltration similar to tors of the mTORC1 pathway. Both p-S6 and p-4E-BP1 were drastically increased in Skiv2l fl/fl K14-Cre and iSkiv2l -/epidermis, quantified as signaling intensity per cell, compared with littermate controls (Figure 4, F and G and Supplemental Figure 7, A and B). These data suggest that Skiv2l deficiency activates mTORC1 in keratinocytes.
The SKI complex interacts with the ribosome to extract mRNA from stalled ribosomes for degradation (7,8). We hypothesized that loss of SKIV2L may impair ribosome recycling on certain mRNAs, which could activate mTORC1 and global protein synthesis. To measure global protein synthesis in vivo, we adapted a pulse-labeling assay using O-propargyl-puromycin (OP-Puro) that can be injected in mice and labels nascent polypeptides in cells. Then, we labeled OP-Puro fluorescently by click chemistry and quantitated protein synthesis in single cells by flow cytometry. iSkiv2l -/keratinocytes showed a significant increase in OP-Puro incorporation, suggesting increased global protein synthesis (Supplemental Figure 7C). We also found that Skiv2l-deficient keratinocytes were substantially larger than control cells ( Figure 4H and Supplemental Figure 7D), consistent with the notion that mTORC1 controls mammalian cell size by regulating protein translation (12). These data further support mTORC1 activation in Skiv2l-deficient cells in vivo.
Skiv2l deficiency disrupts T cell homeostasis. Besides epidermal hyperplasia, we also observed immune cell infiltrates, particularly T cells, in iSkiv2l -/mouse dermis and hair follicles (Supplemental Figure 8, A and B). iSkiv2l -/mice exhibited lymphadenopathy and an increase in cellularity of skin-draining lymph nodes (Supplemental Figure 8C). The lymph node in iSkiv2l -/mice exhibits hyperplasia with a larger number of follicles in the paracortex, suggesting an expansion of T cell clusters (Supplemental Figure 8D). We next examined whether T cell immune homeostasis is perturbed in iSkiv2l -/mice. In the spleen, the number of CD3 + T cells and the CD4/CD8 ratio in iSkiv2l -/mice are normal (Supplemental Figure 8, E and F). However, iSkiv2l -/mice showed significantly increased effector memory (CD62L lo CD44 hi ) and central memory (CD62L hi CD44 hi ) CD4 + and CD8 + T cells with a corresponding reduction in naive (CD62L hi CD44 lo ) T cells in the spleen ( Figure 5A and Supplemental Figure 9, A and B), indicating loss of quiescence in Skiv2l-deficient T cells. We further evaluated in vivo proliferation of T cells using a bromodeoxyuridine (BrdU) incorporation assay. iSkiv2l -/mice had more BrdU-incorporated T cells than controls under steady-state conditions ( Figure 5B). After anti-CD3/CD28 stimulation, expression levels of T cell activation markers CD25 and CD69 were much higher in iSkiv2l -/compared with control T cells ( Figure 5C and Supplemental Figure 9C). iSkiv2l -/-T cells also proliferated more rapidly and produced more not trigger cell-intrinsic IFN signaling nor myeloid-driven immunopathology in vivo.
Aberrant activation of the mTOR pathway in Skiv2l-deficient keratinocytes. To explore the underlying mechanism that drives epidermal hyperplasia in Skiv2l-deficient mice, we performed whole transcriptome RNA sequencing analysis comparing Skiv2l fl/fl K14-Cre versus Skiv2l ctl (includes Skiv2l fl/fl , Skiv2l fl/+ , and Skiv2l fl/+ K14-Cre) P0 mouse epidermis. Differentially expressed genes (DEGs) enriched in Skiv2l fl/fl K14-Cre samples include keratinization-related genes (Krt), keratinocyte damage-associated molecular patterns (DAMPs) genes, and genes associated with inflammatory skin diseases (e.g., psoriasis, atopic dermatitis) ( Figure 4A and Supplemental Figure 6, A-C). qRT-PCR analysis of an additional cohort of mice confirmed dysregulated gene expression in Skiv2l fl/fl K14-Cre epidermis compared with controls (Supplemental Figure 6, D-F). Further pathway analysis revealed enrichment of keratinization and cornification pathways, consistent with the epidermal pathology (Figure 4, B and C).
Interestingly, several metabolic pathways (e.g., lipid and glucose metabolism) were also significantly enriched in Skiv2l fl/fl K14-Cre epidermis compared with controls ( Figure 4C). This prompted us to examine the mTOR pathway, which is the master regulator of cellular metabolism. Gene set enrichment analysis (GSEA) using hallmark gene sets revealed enrichment of mTORC1 signaling pathway in Skiv2l fl/fl K14-Cre samples ( Figure 4D). The RNA-seq data set showed broad elevation in the expression of the mTORC1 pathway genes in Skiv2l fl/fl K14-Cre dermis compared with controls, which we further confirmed by qRT-PCR analysis using an additional cohort of mice ( Figure 4E and Supplemental Figure 6G). We next examined mTORC1 activity by directly measuring phosphorylation of S6 ribosome protein and 4E-BP1, 2 downstream effec- became more proliferative compared with controls (Supplemental Figure 10, B-D). These results suggest an intrinsic role for SKIV2L in maintaining T cell homeostasis.
We also observed increased p-S6 staining in both CD8 + and CD4 + T cells of iSkiv2l -/mice compared with controls ( Figure 5E). Western blot analysis of sorted splenic T cells further revealed IFN-γ and granzyme B than controls ( Figure 5D and Supplemental Figure 9, D-F). To further confirm a cell-intrinsic role of Skiv2l deficiency in T cells, we isolated naive CD4 + T cells that have not yet encountered antigen and stimulated them ex vivo with anti-CD3/CD28 antibodies (Supplemental Figure 10A). Skiv2l-deficient CD4 + T cells expressed higher levels of CD25 and CD69 and  increased phosphorylation of mTORC1 substrates S6K and 4E-BP-1, while phosphorylation of mTORC2 downstream targets FoxO1 and FoxO3a was reduced (Supplemental Figure 11, A-C). iSkiv2l -/-CD8 + and CD4 + T cells are also larger in size than control T cells ( Figure 5F). Together, these data suggest that Skiv2l deficiency activates mTORC1 and induces T cell hyperactivation. mTORC1 inhibitor rapamycin ameliorates iSkiv2l -/disease pathology. We next assessed whether targeting mTORC1 can reduce disease pathology associated with Skiv2l deficiency ( Figure  6A). Remarkably, systemic treatment of iSkiv2l -/mice with rapamycin (by intraperitoneal injection) led to significantly reduced skin lesion and hair loss ( Figure 6, B and C). Histopathology analysis revealed that rapamycin dramatically ameliorated epidermal hyperplasia in iSkiv2l -/mice ( Figure 6D). We also observed reduced p-S6 staining and fewer Ki67 + cells within the skin of rapamycin-treated compared with vehicle-treated iSkiv2l -/mice Histopathologic analysis of a lesional skin biopsy identified interface dermatitis with lymphocyte infiltration in the dermis ( Figure 7C). RNA-seq analysis of PBMCs from the patient with THES2 did not identify elevated ISG expression when compared with healthy controls ( Figure 7D and Supplemental Figure 13A). Further immunostaining revealed substantially increased p-S6 in epidermis from the patient with THES2, consistent with activated mTORC1 pathway ( Figure 7E). We also noted an interesting difference between THES2 and an unrelated case of atopic dermatitis (AD). p-S6 staining in THES2 is stronger toward the basal layer of the expanded epidermis whereas p-S6 staining in AD is lacking at the basal layer. Further, we observed a moderate increase in MX1 expression in skin tissue from the patient with THES2 (Supplemental Figure 13B). The findings from the patient with THES2 match those from mice and together they suggest that loss of SKIV2L activates mTORC1 signaling, which induces skin autoinflammation ( Figure 7F).

Discussion
Inborn errors in RNA metabolism have been associated with human autoimmune diseases. For instance, Aicardi-Goutières syndrome mutations in RNase H2 (cleaves RNA-DNA hybrids) or in adenosine deaminase acting on RNA 1 (ADAR1) invoke innate immune sensing of self-nucleic acids (DNA or RNA), leading to type I IFN signaling and systemic inflammation (13). Biochemical activities of the cytoplasmic RNA exosome are well-studied in the yeast, but its physiological functions in mammals are not known. Here, using a panel of germline, inducible, wholebody, and tissue-specific gene knockout mouse models, we demonstrate that SKIV2L RNA exosome loss-of-function causes skin-specific autoinflammation by disrupting both epidermal and T cell homeostasis without provoking innate immune response to self-RNA. Tissue-specific deletion of SKIV2L also did not trigger type I IFN response in keratinocytes or myeloid cells. Although MX1 expression is moderately increased in the epidermis of the patient with THES2, it is possible that overactive T cells that infiltrate skin tissue induce ISGs expression in keratinocytes in a paracrine manner. Moreover, we did not observe systemic or multi-organ inflammation when the Skiv2l gene was deleted globally in adult mice.
Our in vivo data indicates that the cytoplasmic SKIV2L RNA exosome, under physiological conditions, is unlikely to act as a gatekeeper of self-RNA-mediated innate immune activation. This notion is in line with a recent study investigating substrates of mammalian RNA decay pathways, which reveals that SKIV2L participates widely in surveillance of mRNA with aberrant translation (14). A more comprehensive model could be that the mammalian SKIV2L RNA exosome engages in different functions depending on context or cellular cues. Under physiological conditions, SKIV2L extracts and removes mRNA from stalled ribosomes to avoid activation of mTORC1. Under stress conditions such as viral infection, SKIV2L may expand its role to remove immunogenic host or viral RNA to avoid activation of innate immune sensors.
Our study demonstrates that cytoplasmic RNA quality control failure could activate mTORC1 and cause tissue pathology. The SKI complex directly associates with 80S ribosome and extracts mRNA ( Figure 6, E and F). Furthermore, rapamycin treatment attenuated hyperproliferation and IFN-γ production of iSkiv2l -/-T cells ( Figure 6, G and H). Topical rapamycin treatment also reduced epidermal hyperplasia, keratinocyte proliferation and overall skin pathology (Supplemental Figure 12, A-G). These data suggest that mTORC1 is an attractive therapeutic target for SKIV2L deficiency. mTORC1 activation in a SKIV2L-deficient patient with THES2. We identified a pediatric patient with THES2 who carries 2 pathogenic mutations in each of the SKIV2L alleles: c.1452del (p.Val-485Cysfs*45) in exon 14 and c.3541-2A>G in intron 27. Both variants are present at extremely low frequencies in the Genome Aggregation Database (gnomAD) and at the time of diagnosis neither variant is present at the Human Gene Mutation Database (HGMD) or ClinVar. SKIV2L protein was undetectable in skinderived fibroblasts from the patient, and TTC37 protein level was also reduced ( Figure 7A). At 8 weeks the patient was admitted to the neonatal intensive care unit (NICU) for failure to thrive (2.1 kg), watery diarrhea, and diagnosis of Group B Streptococcus urinary tract infection, and was found to be viremic with cytomegalovirus. The patient presented with woolly hair appearance indicative of trichorrhexis nodosa ( Figure 7B). An erythematous raised nonpruritic rash was noted throughout the body since birth (Figure 7B). Liver studies demonstrated persistent transaminitis. At 5 months, a trial of glucocorticoid therapy was started empirically and resulted in mild skin improvement. Liver biopsies at 7 months were consistent with active hepatitis with mild portal fibrosis and early bridging fibrosis (data not shown). At 17 months, off steroid therapy, the patient was noted to have an increase in blood urea nitrogen (BUN) and creatinine, developed hypertension, and was found to have become positive for p-ANCA and serine protease 3. A kidney biopsy revealed pauci-immune necrotizing and crescentic glomerulonephritis (CKD stage V; Table 1 and Supplemental Figure 13, C and D). There is no history of lupus in either parent of the patient. through the mTORC1 pathway and causes epidermal hyperproliferation (15), similar to the Skiv2l-deficient skin phenotypes observed in our study. How stalled ribosomes activate the mTORC1 pathway requires further investigation. Nonetheless, these findings and ours collectively demonstrate how seemingly broad cellular defects from stalled ribosome for degradation (7,8). This mechanism has been proposed to facilitate subsequent dissociation and recycling of stalled ribosomes by the Pelota/Hbs1/ABCE1 complex, though its physiological importance has not been experimentally elucidated. Deletion of Pelota leads to global upregulation of protein translation multicentric Castleman disease (iMCD) (28)(29)(30). Aberrant mTOR signaling also contributes to pathogenic lymphoproliferation in those diseases and administration of sirolimus has been shown to induce cellular or clinical benefits in patients (27)(28)(29)(30)(31)(32).
Patients with THES develop other clinical symptoms besides skin inflammation, including intractable diarrhea, primary immunodeficiency, and liver disease (11). Some patients with THES2 do not develop characteristic diarrhea, suggesting phenotypic heterogeneity of the disease (33). The whole-body inducible iSkiv2l -/mice do not fully replicate all THES phenotypes; this is likely due to incomplete or delayed gene deletion in certain tissues or differences in mouse versus human. Nonetheless, numerous studies have demonstrated a critical role for mTORC1 in intestinal homeostasis (34)(35)(36). We demonstrated remarkable therapeutic efficiency of mTOR inhibitor rapamycin in ameliorating skin disease of Skiv2l-deficient mice. Therefore, our findings provide a scientifically rationalized and immediately actionable treatment for the THES disease.  Figure 1A. PCR genotyping of knockout-first and floxed alleles was carried out using the following primers: knockout first allele, forward 5′-CGAATTCTCAGTTGAG-CACATAGC-3′, reverse 5′-CCACAACGGGTTCTTCTGTTAG-3′; floxed allele, forward 5′-TACACATGCCTCCATTAGCCTG-3′, reverse 5′-GGCTCATGTCCTGTACACTAATG-3′. PCR genotyping of LysM-Cre, UBC-Cre/ERT2, and K14-Cre mice was carried out following protocols of the Jackson Laboratory. All mice were housed in pathogen-free barrier facilities at UT Southwestern Medical Center.
We present a new paradigm of tissue-specific autoinflammation, where metabolically dysregulated T cells (without innate immune activation or priming) are attracted by skin-intrinsic pathology to confer tissue specificity and fuel inflammation. This mechanism is distinct from other skin-specific autoimmune diseases, such as psoriasis and cutaneous lupus erythematosus, which are generally believed to be initiated by aberrant innate immune activation as the primary trigger of pathogenesis (16,17). Notably, the metabolic perturbation in T cells by Skiv2l deficiency seems insufficient to breach global immune tolerance, and we did not observe systemic T cell-mediated autoimmune pathologies that are common in patients with FOXP3 or AIRE mutations (18,19). While hyper-responsiveness of iSkiv2l -/naive T cells to ex vivo TCR stimulation highly suggest a cell-intrinsic role for SKIV2 in regulating T cell response, it's worth noting that soluble or secreted factors that naive T cells have engaged with in vivo could also contribute to enhanced T cell response. The patient with THES2 in this study did not develop hemophagocytic syndrome, which is characterized by hyperactivation of lymphocytes and histiocytes, and which is noted in about 60% of patients with THES (20). However, the patient developed ANCA-associated vasculitis, including pauci-immune necrotizing and crescentic glomerulonephritis, which, to our knowledge, is first reported in patients with THES in the present study. Besides autoreactive B cells, expanded effector T cells that are in a persistent activation state are also involved in the pathogenesis (21). The T cell hyperactivation phenotype in our mouse model merits further clinical investigation in patients with THES. The increased mTORC1 activity and decreased mTORC2 activity in Skiv2l-deficient T cells are similar to that in systemic lupus erythematosus (SLE) (22)(23)(24)(25). Treatment with sirolimus, an mTORC1 inhibitor, corrects the proinflammatory T cell phenotype and improves clinical outcomes in patients with SLE (23,26). Hyperactivation of the mTORC1 pathway has also been associated with other autoimmune hematological diseases, including autoimmune lymphoproliferative syndrome (ALPS) (27) and idiopathic Flow cytometry analysis of cell cycle and protein synthesis in vivo. To analyze cell cycle, keratinocytes were isolated from adult mice tail skin and P0 neonate skin using dispase II digestion (Sigma, catalog To generate postnatal whole-body Skiv2l knockout mice, tamoxifen (Sigma-Aldrich, catalog T5648) was dissolved in corn oil (20 mg/mL) and injected intraperitoneally into 4-week-old Skiv2l fl/fl UBC-Cre/ ERT2 and Skiv2l fl/fl littermate control mice (75 mg/kg body weight) once daily for 5 consecutive days. To delete Skiv2l locally in a small area of skin, Skiv2l fl/fl UBC-Cre/ERT2and Skiv2l fl/fl littermate control mice were depilated on both lower flanks. Five microliters of (Z)-4-OHT, Sigma-Aldrich, catalog H7904, dissolved in DMSO to a concentration of 5 μM) or vehicle DMSO was topically applied on right or left lower flank within a circle area (diameter 1 cm) respectively, once daily for 5 consecutive days.
Human patient study. The patient included in this study is a female (< 10 years old) born to nonconsanguineous non-Hispanic White parents. Genetic and histopathological analyses of the affected patient were performed with informed consent. Skin fibroblasts were obtained from the patient and expanded in the laboratory. Formalin-fixed paraffin-embedded lesional skin tissue was sectioned for fluorescence immunohistochemistry analysis. Deidentified nondiseased skin samples were obtained from the Tissue Management Shared Resource (TMSR) at UT Southwestern Medical Center. Skin biopsy slides from an unrelated case of AD and an unrelated case of seborrheic keratosis (SK) were kindly provided by Richard C. Wang (UTSW) and used for comparison to the THES2 patient's skin biopsy.