Impaired GATE16-mediated exocytosis in exocrine tissues causes Sjögren’s syndrome-like exocrinopathy

Abstract

Sjögren’s syndrome (SjS) is a chronic autoimmune disease characterized by immune cell infiltration of the exocrine glands, mainly the salivary and lacrimal glands. Despite recent advances in the clinical and mechanistic characterization of the disease, its etiology remains largely unknown. Here, we report that mice with a deficiency for either Atg7 or Atg3, which are enzymes involved in the ubiquitin modification pathway, in the salivary glands exhibit a SjS-like phenotype, characterized by immune cell infiltration with autoantibody detection, acinar cell death, and dry mouth. Prior to the onset of the SjS-like phenotype in these null mice, we detected an accumulation of secretory vesicles in the acinar cells of the salivary glands and found that GATE16, an uncharacterized autophagy-related molecule activated by ATG7 (E1-like enzyme) and ATG3 (E2-like enzyme), was highly expressed in these cells. Notably, GATE16 was activated by isoproterenol, an exocytosis inducer, and localized on the secretory vesicles in the acinar cells of the salivary glands. Failure to activate GATE16 was correlated with exocytosis defects in the acinar cells of the salivary glands in Atg7 and Atg3 cKO mice. Taken together, our results show that GATE16 activation regulated by the autophagic machinery is crucial for exocytosis and that defects in this pathway cause SjS.

Supplementary Information

The online version contains supplementary material available at 10.1007/s00018-022-04334-x.

Introduction

The acinar cells in lobular salivary glands (SGs) produce and secrete salivary proteins via exocytosis, enabling the secretion of prepackaged secretory granules containing key functional salivary components, including amylase, mucins, and immunoglobulins, which play important roles in lubrication, digestion, immunity, and in the overall maintenance of homeostasis within the oral cavity. Previous studies indicate that the proteomic profiles of saliva are altered in individuals with metabolic syndromes, diabetes, and obesity, suggesting that exocytosis in exocrine tissues may be dysregulated in these metabolic conditions [1–5]. Soluble N-ethylmaleimide-sensitive factor attachment protein receptor (SNARE) proteins have been identified as regulatory molecules that are crucial for exocytosis [6].

Sjögren’s syndrome (SjS) is a chronic autoimmune disease characterized by immune cell infiltration of the exocrine glands, mainly the salivary and lacrimal glands. The differential diagnosis and etiology of SjS is extensive, since there are several diseases with similar symptoms and various factors associated with SjS. Current diagnostic criteria for SjS use a combination of histological examinations of the minor SGs for immune cell foci, serology for autoantibodies, and dry eye/mouth evaluation. Since patients are diagnosed only at later stages of the disease, the earlier events prior to clinical manifestations are largely unknown. Interestingly, the expression of vesicle-associated membrane protein 8 (VAMP8; a.k.a. Endobrevin), a v-SNARE that plays an essential role in exocytosis at the secretory vesicle membrane [7], is altered from the apical to the basal side in acinar cells of the lacrimal glands in individuals with dry eyes [8] and takes place in the entire cytoplasm in patients with SjS [9]. In addition, the intracellular expression patterns of syntaxins, t-SNARE proteins that play an essential role in exocytosis at the targeting membrane [10], is altered in patients with SjS [9]. Corresponding to the disruption of SNARE complexes, the distribution of secretory mucins is not limited to the cytoplasm but also extends to the extracellular matrix in SjS [9, 11]. Thus, the accumulating evidence in the literature suggests a potential link between exocytosis and SjS.

An increasing number of studies indicates that molecules initially identified in autophagy, a protein/organelle degradation system, have additional functions in a variety of non-autophagy pathways (e.g., formation of various types of membranes and vesicles) [12]. Interestingly, autophagy-related molecules participate in the secretion of the contents of granules in yeast, and of secondary lysosomes in mammalian cells, through the regulation of the structural formation of these membranes [13]. Autophagy and the autophagic machinery are regulated by two ubiquitin (Ub)-like conjugation pathways, the ATG12-5 and ATG8 pathways, which are essential for the extension of membranes originating from the endoplasmic reticulum. ATG7, an E1-like enzyme, activates both ATG12-5 and ATG8 pathways, whereas ATG3, an E2-like enzyme, is specifically involved in the ATG8 pathway [14]. To date, there are six homologs of yeast Atg8 in mammals (i.e., microtubule-associated protein 1 light chain 3A [MAP1LC3, a.k.a. LC3A], MAP1LC3 beta [MAP1LC3B], MAP1LC3 beta 2 [MAP1LC3B2], microtubule associated protein 1 light chain 3 gamma [MAP1LC3C], γ-aminobutyric acid receptor-associated protein [GABARAP], GABA type A receptor associated protein like 1 [GABARAPL1], and Golgi-associated ATPase enhancer of 16 kDa [GATE16, a.k.a. GABARAPL2] [15–18].

The ATG8 homologs in the cytosol (LC3A/B/C-I, GABARAP-I, GABARAPL1-I, and GATE16-I) are lipidated with phosphatidylethanolamine (PE) (LC3A/B/C-II, GABARAP-II, GABARAPL1-II, and GATE16-II) by E1- and E2-like ubiquitin ligases (ATG7, ATG4, and ATG3). ATG8 homologs have been extensively investigated in several studies that revealed that each homolog uniquely functions in both the canonical and noncanonical autophagic machinery [19–23]. For example, while LC3-II is involved in the elongation of the isolation membrane during autophagy, GABARAP and GATE16 are required for the maturation of the autophagosome and subsequent fusion with the lysosomes [24, 25]. GABARAPL1 has been suggested to bind with LC3-II on the autophagosomal membrane [26]; GABARAP is expressed in the brain and is essential for steady-state constitutive autophagy [27]. The regulation and function of GATE16 are less characterized in mammals, and its contribution to autophagy seems to be limited [15, 28]. Thus, the noncanonical machinery is still under investigation in various membrane trafficking systems under physiological and pathological conditions.

A detailed understanding of the mechanism(s) by which the autophagic machinery regulates exocytosis, and how defects in this cascade result in SjS-like exocrinopathy, will provide new insights on the key functions of exocytosis and the pathological mechanisms of SjS-like exocrinopathy, not only in the SGs but also in other secretory organs, including the mammary glands, the pancreas, and the prostate.

Results

The autophagic machinery plays a crucial role in exocytosis

To investigate whether and how the autophagic machinery functions in exocytosis, we analyzed mice deficient for Atg7 in epithelial tissues (Atg7^F/F; K14-Cre mice, hereafter Atg7 conditional knockout [cKO] mice) (Fig. 1A). While there were no developmental defects in these mice (Fig. 1B, C; SI Appendix, Fig. S1), we found that salivary protein secretion stimulated by pilocarpine was significantly decreased in the Atg7 cKO mice compared to wild-type control mice (Fig. 1D). The secretory vesicles in the acinar cells of the SMGs were accumulated in Atg7 cKO mice after 2 months of age (Fig. 1E; SI Appendix, Figs. S1H, I). We used transmission electron microscopy to quantify the size and number of secretory vesicles in the SMGs of Atg7 cKO and control mice and found increased number and size of secretory vesicles, leading to increased cell size (Fig. 1F, G). To further confirm the accumulation of secretory vesicles in the SMGs of Atg7 cKO mice, we conducted both quantitative RT-PCR and immunoblotting for molecules related to exocytosis. As expected, MUC10 (mucin specifically expressed in the acinar cells of the SMGs), AMY1 (amylase secreted from the acinar cells of the SGs), and VAMP2 and VAMP8, which locate on secretory vesicle membranes, accumulated in Atg7 cKO mice, while the expression of these genes was not changed (Fig. 1H, I). The adrenergic receptor beta 1 (ADRB1) and protein kinase A (PKA) pathway, which can induce secretory vesicle formation, was not altered in the SMGs, indicating that exocytosis, but not induction of secretory vesicle formation, was compromised in the SMGs of Atg7 cKO mice. The accumulation of MUC10 was further confirmed with immunohistochemical analysis (Fig. 1J).

Fig. 1.

Exocytosis is regulated by the autophagic machinery. A Immunoblotting of ATG7 in the submandibular glands (SMGs) of newborn wild-type (WT) control and Atg7 conditional knockout (cKO) mice. GAPDH was used as a loading control. B Body weight of WT control (blue bar) and Atg7 cKO (yellow bar) mice in males and females at 2 months of age. n = 6 per group. C Weight of the SMGs per body weight in WT (blue bar) and Atg7 cKO (yellow bar) mice at 2 months of age. n = 6 per group. D Amount of secreted protein (mg per g of SMGs) induced by pilocarpine (10 mg/kg body weight, intraperitoneal injection) in WT (blue bar) and Atg7 cKO (yellow bar) mice at 2 months of age. n = 6 per group. **p < 0.01. E Hematoxylin/eosin staining of the SMGs of WT and Atg7 cKO mice at 1 month of age. The boxed areas are enlarged in the lower panels. The SMGs from Atg7 cKO mice appear to contain more granules compared to WT mice. Scale bars: 50 μm in the upper panels, and 10 μm in the lower panels. F Transmission electron microscopy images from the SMGs of WT and Atg7 cKO male mice at 2 months of age. G The graphs show the quantification of the size and number (per cell) of secretory vesicles (SV), and the size of acinar cells in WT (blue bars) and Atg7 cKO (yellow bars) mice at 2 months of age. Scale bars: 10 μm. *p < 0.05, **p < 0.01, ***p < 0.001. H Quantitative RT-PCR analyses of Adrb1, Amy1, Prol1/Muc10, Vamp2, and Vamp8 in the SMGs of WT (blue bars) and Atg7 cKO (yellow bars) mice at 1 month of age. n = 6 per group. I Immunoblotting analyses of the indicated molecules using samples from the SMGs of WT and Atg7 cKO mice at 1 month of age. GAPDH was used as a loading control. J Immunohistochemical analyses for MUC10 in the SMGs of WT and Atg7 cKO mice at 1 month of age. The boxed areas are enlarged in the lower panels. Scale bars: 50 μm in the upper panels, and 25 μm in the lower panels

GATE16 plays a crucial role in exocytosis

Autophagy and the autophagic machinery are regulated by two ubiquitin-like conjugation pathways, the ATG12-5 and ATG8 pathways [29], but it remains unclear how mammalian Atg8 homologs are differentially regulated and how they function. For instance, the role and regulation mechanism(s) of GATE16 (a.k.a. GABARAPL2 [15, 16]) have remained mostly unknown to date, and its contribution to autophagy seems to be limited and differentially activated [15, 28]. To investigate the expression levels and pattern of GATE16 in various mouse tissues, we performed immunoblotting analyses on tissues from C57BL/6J mice and found that GATE16 was highly expressed in tissues that contained secretory components: the acinar cells of the SMGs and pancreas, and the goblet cells of the intestine (Fig. 2A). While Gate16 gene expression was not altered in the SMGs of Atg7 cKO mice compared to littermate controls (Fig. 2B), GATE16 was accumulated in Atg7 cKO mice under both feeding and starvation conditions (Fig. 2C). Although autophagy activity, as monitored by LC3 lipidation (LC3-II), was induced under starvation conditions in the liver, autophagy was not induced in the SMGs under starvation conditions (Fig. 2C). Our previous study showed that autophagy is only detectable in the duct cells in the SMGs [30]. Notably, we found that GATE16-II (the activated, lipidated form) was induced by isoproterenol, an exocytosis inducer, in the SMGs of wild-type control mice, while this inducer failed to activate GATE16 in the SMGs of Atg7 cKO mice (Fig. 2D). To test the functional significance of Gate16 in exocytosis, we performed shRNA knockdown for Gate16 in SMG tissue explants obtained from C57BL/6J mice and found that MUC10 accumulated after Gate16 shRNA knockdown while Muc10 expression was not affected by Gate16 shRNA knockdown (Fig. 2E, F). This indicates that GATE16 plays a crucial role in exocytosis. Indeed, immunohistochemical analysis for GATE16 showed that GATE16 was explicitly expressed in the acinar cells of SGs, specifically located at the apical edge of the plasma membrane and in secretory vesicles attached to the plasma membrane in wild-type control mice. By contrast, GATE16 was diffused in the cytoplasm of acinar cells in the SMGs of Atg7 cKO mice (Fig. 2G). To further confirm the distribution of GATE16, we generated Gate16-GFP transgenic mice (Fig. 2H, I) and detected a similar expression pattern of GATE16 in the SMGs (Fig. 2J), with GATE16-II co-localizing with and nearby AQP5 at the apical membrane (Fig. 2K). Recent studies indicate that SNARE molecules, which are crucial for exocytosis, play a role in the fusion of the autophagosome with a lysosome [31]. This suggests that there are some shared molecular mechanisms between autophagy and exocytosis. To test the interactions between GATE16 and SNARE proteins during exocytosis, we conducted co-immunoprecipitation using SMG tissues from control and Atg7 cKO mice and found that GATE16 formed a complex with VAMP2 (v-SNARE), but not with VAMP8 and SNAP23 (t-SNARE), in the SMGs of control mice, but not in those of Atg7 cKO mice (SI Appendix, Fig. S2A). The interaction between GATE16 and VAMP2 was further confirmed with co-immunoprecipitation using overexpressed tagged-proteins (SI Appendix, Fig. S2B). These results indicate that VAMP2 can form a complex with GATE16-II on secretory vesicle membranes in the SMGs of WT mice, while VAMP2 fails to form such complex in the SMGs of Atg7 cKO mice.

Fig. 2.

GATE16 plays a crucial role in exocytosis. A Immunoblotting of GATE16 using the indicated organs from C57BL/6J male mice at 2 months of age. Pro, pro-GATE16 form; I, inactivated form; II, activated form. B Quantitative RT-PCR analyses for Gate16 in the submandibular glands (SMG) of wild-type (WT) and Atg7 conditional knockout (cKO) mice at 1 month of age. n = 3 per group. C Immunoblotting for the indicated molecules using samples from the liver and SMGs of WT and Atg7 cKO mice at 1 month of age under fed (Fed) and starved (Stv) conditions. D Time-course immunoblotting for GATE16 after treatment with isoproterenol (IPR) in the SMGs of WT and Atg7 cKO mice at 1 month of age. E Quantitative RT-PCR analyses for the indicated genes, under Fed and Stv conditions, after treatment with shRNA for Gate16 (shGate16). n = 6 per group. F Immunoblotting for the indicated molecules, under Fed and Stv conditions, after treatment with shGate16. G Immunohistochemical analyses for GATE16 in the SMGs of WT and Atg7 cKO mice at 2 months of age. d duct. Scale bars: 10 μm. H RT-PCR for Gfp in the SMGs of WT and Gate16-GFP transgenic (Tg) mice at 2 months of age. I Immunoblotting for GFP in the SMGs of WT and Tg mice at 2 months of age. J Immunohistochemical analyses for GATE16 in the SMGs of transgenic (Tg) mice at 2 months of age. D duct. Scale bar: 10 μm. K Immunohistochemical analysis for GATE16 and AQP5 in the SMGs of WT and Atg7 cKO mice at 2 months of age. Scale bar: 25 μm

Loss of Atg7 causes SjS-like inflammatory lesions

To investigate the consequence of exocytosis defects, we analyzed the SG phenotype in Atg7 cKO mice at 2, 3, 4, 5, 6, and 12 months of age. Notably, we detected elevated apoptosis in the SMGs of Atg7 cKO mice in both males and females at 2 months of age (Fig. 3A, B). In addition, we detected elevated inflammatory cytokine levels at 2 and 6 months of age (Fig. 3C; SI Appendix, Fig. S3). In agreement with the inflammation observations, the submandibular lymph nodes (SLNs) were swollen at 2 and 6 months of age (Fig. 3D, E). In addition, we detected activated T cells that stained with methylene blue in the SLNs of Atg7 cKO mice (Fig. 3F). Importantly, we detected lymphocyte infiltration foci formation after 5 months of age (Fig. 3G; SI Appendix, Fig. S3C). CD4-positive helper T cells, CD8-positive cytotoxic T cells, Tregs, B cells, MHC class II-positive antigen-presenting cells, and macrophages were detectable at inflammatory lesions in the SMGs of Atg7 cKO mice (Fig. 3H; SI Appendix, Figs. S3D). Moreover, serum anti-SSA and anti-SSB autoantibodies were detected, as seen in patients with SjS (Fig. 3I; SI Appendix, Fig. S3E). Thus, mice deficient for Atg7 meet all the current criteria for SjS in humans. Moreover, to confirm that the impairment in tissue homeostasis appeared prior to autoimmune activation, we performed flow cytometry using the SMGs and spleens collected from Atg7 cKO and control mice. As expected, representative immune cells were significantly increased in the SMGs of Atg7 mice, compared to control mice, at 8 weeks of age, although there was no change in the profile of representative immune cells in the spleen at that age (Fig. 3J; SI Appendix, Fig. S4). These results strongly suggest that impaired SMG tissue homeostasis can trigger the onset of SjS-like inflammation.

Fig. 3.

Exocytosis deficiency causes a Sjögren’s syndrome-like phenotype. A TUNEL assays in the submandibular glands (SMGs) of male and female wild-type (WT) control and Atg7 cKO mice at 2 months of age. Nuclei were stained with DAPI (blue). Scale bars: 25 μm. B Percentage of TUNEL-positive cells per total number of cells in the SMGs of WT control (blue bars) and Atg7 cKO (yellow bars) mice. ***p < 0.001. C Amount of inflammation markers in the SMGs of female Atg7 cKO mice at 2 months of age. n = 3 per group. *p < 0.05, **p < 0.01, ***p < 0.001. D Gross picture of the submandibular lymph nodes dissected from WT control and Atg7 cKO mice at 2 and 6 months (2 m and 6 m) of age. Scale bars: 1 mm. E Quantification of the weight of lymph nodes dissected from WT control (WT; blue bar) and Atg7 cKO mice at 2, 3, 4, 5, and 6 months (2 m, 3 m, 4 m, 5 m, and 6 m) of age. ***p < 0.001. F Hematoxylin/eosin (H&E) and methylene blue staining of submandibular lymph nodes collected from WT and Atg7 cKO mice at 6 months of age. Scale bars: 200 μm in top panels and 25 μm in middle and bottom panels. G H&E staining of the SMGs of male and female wild-type (WT) control and Atg7 cKO mice at 6 months of age. Arrows indicate the infiltration of lymphocytes in the acinus. Scale bars: 100 μm. H Immunohistochemical analysis for the indicated immune cell types in the SMGs of female WT and Atg7 cKO mice at 6 months of age. The boxed areas are enlarged. Scale bars: 50 μm. I Quantification of IgA anti-Ro60/SSA (top) and anti-La/SSB (bottom) antibodies in female WT control and Atg7 cKO mice at 2, 3, 4, 5, 6, and 12 months (2 m, 3 m, 4 m, 5 m, 6 m, and 12 m) of age. Blue bars, WT; yellow bars, Atg7 cKO. n = 6 per group. ***p < 0.001. J Graphs with dot plots for quantification of each immune marker in the SMGs. n = 3 per group. **p < 0.01, ***p < 0.001

Loss of Atg3 causes SjS-like inflammation

Since ATG7 can activate both the ATG12-5 and ATG8 pathways, we further analyzed mice deficient for Atg3 in epithelial tissues (Atg3^F/F;K14-Cre mice, hereafter Atg3 cKO mice) to clarify the specific contribution of the ATG8 pathway to exocytosis (Fig. 4A). We found that Atg3 cKO mice displayed exocytosis defects (Fig. 4A–C) similar to those seen in Atg7 cKO mice. In addition, we observed similar consequences of SjS-like phenotypes such as increased cell death (Fig. 4D, E), elevated inflammation-related cytokines in the SMGs (Fig. 4F; and SI Appendix, Figs. S5A), swollen SLNs (Fig. 4G–I), immune cell infiltration in the SMGs and LGs (Fig. 4J; SI Appendix, Figs. S5B, C and S6A), and elevated autoantibodies in the serum (Fig. 4K; SI Appendix, Fig. S6B) of Atg3 cKO mice compared to control mice. As seen in Atg7 cKO mice, immune cells were significantly increased in the SMGs, but not in the spleen, of Atg3 cKO mice at 8 weeks of age (SI Appendix, Fig. S7). Taken together, our results indicate that ATG7/3-mediated GATE16 activation is responsible for exocytosis, and a failure in this pathway leads to SjS-like exocrinopathy.

Fig. 4.

Atg3 cKO mice display a Sjögren’s syndrome-like phenotype. A Immunoblotting for the indicated molecules in female wild-type (WT) and Atg3 cKO mice at 1 month of age. B Hematoxylin/eosin (H&E) staining of female submandibular glands (SMGs) of WT and Atg3 cKO mice at 2 months of age. Scale bars: 25 μm. C Immunohistochemical analyses for MUC10 in the SMGs of male WT and Atg3 cKO mice at 2 and 6 months of age. Scale bars: 50 μm. D TUNEL assays in the SMGs of WT and Atg3 cKO mice at 2 months of age. Nuclei were stained with DAPI (blue). Scale bars: 25 μm. E Percentage of TUNEL-positive cells per total number of cells in the SMGs of WT (blue bars) and Atg3 cKO (green bars) mice at 1 month of age. n = 6 per group. ***p < 0.001. F Amount of inflammation markers in the SMGs of female Atg3 cKO mice at 2 months of age. n = 3 per group. *p < 0.05, **p < 0.01, ***p < 0.001. G Gross picture of the submandibular lymph nodes dissected from WT and Atg3 cKO mice at 2 and 6 months (2 m and 6 m) of age. Scale bars: 1 mm. H Quantification of the weight of lymph nodes dissected from WT (blue bars) and Atg3 cKO (green bars) mice at 2, 3, 4, 5, and 6 months (2 m, 3 m, 4 m, 5 m, and 6 m) of age. ***p < 0.001. I H&E and methylene blue staining of submandibular lymph nodes collected from WT and Atg3 cKO mice at 6 months of age. Scale bars: 200 μm in top panels and 25 μm in middle and bottom panels. J H&E staining of the SMGs of WT and Atg3 cKO mice at 6 months of age. Arrows indicate the infiltration of lymphocytes in the acinus. Scale bars: 100 μm. K Quantification of IgA anti-Ro60/SSA and anti-La/SSB antibodies in female WT and Atg3 cKO mice at 2, 3, 4, 5, 6, and 12 months (2 m, 3 m, 4 m, 5 m, 6 m, and 12 m) of age. Blue bars, WT; green bars, Atg3 cKO. n = 6 per group. ***p < 0.001. L Graphs with dot plots for quantification of each immune marker in the SMGs. n = 3 per group. *p < 0.05, **p < 0.01, ***p < 0.001

Discussion

The lubricating properties of saliva depend on both the salivary flow and the quality of secretory products like mucins. A defect in any step of exocytosis results in altered secretion of salivary proteins, which play crucial roles in the digestion of food, lubrication of the mouth, and prevention of dental caries, periodontitis, candidiasis, and halitosis [32, 33]. Previous studies show that secretory vesicles are accumulated in the acinar cells [34], which contain enlarged and accumulated secretory vesicles, prior to the onset of overt SjS symptoms in humans [35, 36] and mice [34], suggesting that exocytosis defects are associated with SjS. Moreover, patients with SjS [9, 34, 36, 37] and mouse models for SjS (e.g., non-obese diabetic (NOD) mice) [38, 39] show reduced salivary protein secretion and exocytosis defects. For example, mucins are deposited in the extracellular matrix in the SGs of SjS patients [9], and the expression pattern and levels of RAB proteins, which are associated with secretory vesicle trafficking, are differently altered in the SGs of SjS patients, suggesting that there are several pathogenic subtypes of SjS. In this study, we demonstrated that the autophagic machinery plays a role in the trafficking and secretion of secretory vesicles through activation of GATE16 in the SGs. Dysfunctions in the autophagic machinery–regulated exocytosis lead to acinar cell death, which triggers autoimmunity and inflammation. Thus, our results uncover a new mechanism of SjS-like exocrinopathy without defects in immune cells.

Mouse models are essential tools to investigate the pathogenic mechanism of SjS, and substantial advancements have derived from their analysis [40]. However, no single mouse model to date has replicated each and every aspect of human SjS [40]. Interestingly, in this study we found that Atg7 and Atg3 cKO mice met all of the current diagnosis criteria for the human disease, and the exocytosis defects were detectable prior to other SjS-like phenotypes. Although the existing mouse models for SjS (e.g., spontaneous models, conventional mutant models) have mutations in genes related to the immune system, the Atg7 and Atg3 cKO mice exhibited immune cell infiltration in the SMGs and the lacrimal glands (LGs) without genetic mutations in immune cells. Thus, it is important to understand how exocytosis defects trigger the onset of SjS.

An increasing number of studies show that the autophagic machineries play crucial roles in various cellular systems, including secretion and exocytosis [12]. For example, the autophagic machinery regulates the secretion of the contents of granules in yeast, and of hormones and secretory lysosomes in mammalian cells [13, 41]. The autophagic machinery also plays a role in the secretion of lysozymes in response to bacterial infection in the Paneth cells of the small intestine, which are specialized epithelial cells that function in part by secreting antimicrobial peptides, through the Atg12-Atg5-Atg16 pathway [42]. In the acinar cells of SMGs, exocytosis is modestly affected in Atg5 cKO mice (Atg5^F/F;Aqp5-Cre mice) under physiological conditions [43], while Atg5 cKO mice exhibit homeostasis defects under ductal ligation [44] and irradiation conditions [45]. In this study, we investigated the contribution of the Atg8 pathway to exocytosis, and found that GATE16 plays crucial roles in the ATG7/3-mediated regulatory mechanism in exocytosis in the acinar cells of the SGs; on the secretory vesicles, GATE16 activated by the ATG7/3-conjugation system forms a complex with VAMP2, a v-SNARE protein, and then the GATE16-VAMP2 complex binds to the plasma membrane at the apical side to secrete its content. In Atg7 and Atg3 cKO mice, a failure in GATE16 activation led to the accumulation of secretory vesicles in the cytosol of acinar cells. Since GATE16 is highly expressed in the acinar cells, but not in the duct cells, of the SMGs, activation of the ATG7/3-mediated pathways may depend on cell type and function. For instance, MRL/Mp-Fas^lpr mice, a model for secondary SjS, show decreased autophagic activity in the SMGs compared to those of control mice, but exocytosis was not examined [46]. Phosphopeptid P140, an inhibitor of autophagic activity, decreases tertiary lymphoid structure formation in NOD.H-24h4 mice, a SjS model, through the suppression of CD4-positive T cells and subsequent B cell activation [47]. Interestingly, autophagic activity is upregulated in lymphocytes infiltrated into the inflammatory lesions of the SGs in SjS patients [48, 49], suggesting that the autophagic machinery plays a crucial role in SjS progression. Moreover, autophagic flux and activity are upregulated in primary SG epithelial cells (SGECs) isolated from SjS patients [50]. In addition, mouse SG stem cells, either knocked down for Atg13 or isolated from Atg5 null mice, show a failure in self-renewal and activation, which are important for tissue homeostasis and regeneration, in a three-dimensional (3D) organoid system [51]. Mice with a deficiency for Atg7 in pancreatic acinar cells (Ipf1-Cre;Atg7 cKO mice) exhibit acinar cell dysfunctions, acinar-to-ductal metaplasia, and inflammation due to ER stress, oxidative stress, decreased protein synthetic capacity, and mitochondrial dysfunctions [52]. Taken together, these results suggest that a basal level of autophagic activity is required for acinar cell homeostasis and function. Interestingly, GATE16 expression is upregulated in the corneal tissue of the benzalkonium-induced dry eye mouse model and in human corneal epithelial cells (HCECs) cultured under hyperosmolarity conditions [53]. Normalization of GATE16 expression restores inflammatory cytokine production in hyperosmolarity HCECs [53]. Moreover, autoantibodies against GATE16 are detected in the serum of patients with rheumatoid arthritis at early disease stage [54]. Thus, GATE16 may be involved in various autoimmune diseases and inflammation.

To date, expression of and genetic mutations in genes related to this pathway have not been analyzed in SjS patients; therefore, it is important to evaluate the role of the autophagic machinery in SjS patients in a future study. Since defective exocytosis is detectable at an early stage, or a stage at which there are no manifestations of SjS in Atg7 and Atg3 cKO mice, such an exocytosis mechanism will constitute an applicable target for therapy and early diagnosis of SjS in humans.

Materials and methods

Mice

Mice with the R26R promotor [55] and K14-Cre [56] mice were obtained from The Jackson Laboratory. Atg7^F/F mice were generated by flanking exon 14 [20]. Atg3^F/F mice, gifts from Dr. You-Wen He (Duke University, North Carolina, USA), were generated by flanking exons 8, 9, and 10 [57]. Atg7^F/F and Atg3^F/F mice were crossed with K14-Cre mice to generate Atg7 cKO and Atg3 cKO mice. Genotyping was performed by PCR, as previously described [30]. To measure the salivary flow rate in the mice, saliva secretion was stimulated through administration of pilocarpine hydrochloride (a muscarinic type of cholinergic receptor agonist, 10 mg/kg body weight) and isoproterenol (a β-adrenergic receptor agonist, 10 mg/kg body weight) by intraperitoneal injection (IP). All animal experiments were reviewed and approved by the Animal Welfare Committee and the Institutional Animal Care and Use Committee of UTHealth (AWC-19–0079).

GATE16-GFP transgenic mice

A 1.1-kb NheI-SalI DNA fragment containing the human GATE-16 cDNA fused to EGFP at the N-terminus (GFP-GATE-16) was excised from pGFP-GATE-16 [14], blunted with KOD DNA polymerase, and inserted into the blunt-ended XhoI site of pCAGGS, downstream of the cytomegalovirus immediate-early (CMVie) enhancer and chicken-actin (CAG) promoter [58], to generate pCAG-GFP-GATE-16. The 3.4-kb SalI-PstI fragment was isolated from pCAG-GFP-GATE-16 and microinjected into the pronuclei of fertilized one-cell eggs from B6D2F1/Crj mice (C57BL/6NCrj x DBA/ 2NCrj). Then, 487 microinjected eggs were transferred into the oviducts of pseudo pregnant ICR mice. After extraction of DNA from tail biopsies, 30 mice were screened by PCR analysis for the incorporation of the transgene, using the primers CAGGS-F (5'-GGCTTCTGGCGTGTGACC-3') and CAGGS-Rv (5'-AGCCACCACCTTCTGA T AG-3'). The results showed that three founder (F0) mice were positive for the transgene. These F0 mice were backcrossed with C57BL/6J Crj six times to establish lines and were maintained as heterozygotes for the GFP-GATE-16 transgene. One of the transgenic lines was used for all the experiments described here. Mice were housed in standard animal cages with free access to standard chow in pathogen-free facilities. All experiments were performed according to the guidelines of the Committee on Animal Experiments of Juntendo University, Tokyo, Japan.

Plasmid construction

The cDNA of the coding region of the Vamp2 (NM_009497), Vamp8 (NM_016794), Snap23 (NM_00117792), and Gate16 (NM_026693) genes was amplified from the SMGs of male C57BL/6J mice, using the Expand™ High Fidelity PCR System (Roche), and then subcloned into a pMD20 vector (Takara Bio, Shiga, Japan) and named pMD20-Vamp2, pMD20-Vamp8, pMD20-Snap23, and pMD20-Gate16, respectively. For Vamp2, Vamp8, or Snap23, the fragments amplified with specific forward and common reverse primers with Phusion® High-Fidelity DNA Polymerase (New England Biolabs) were digested by EcoRI and EcoRV (New England Biolabs) and then cloned into the FLAG-HA-pcDNA3.1 vector (addgene #52,535); these constructs were named pcDNA-FLAG-HA-Vamp2, pcDNA-FLAG-HA-Vamp8, and pcDNA-FLAG-HA-Snap23, respectively. The V5-epitope tag fragment was cloned into a pCAG-GFP vector at the SmaI site (addgene #11,150), which was named pCAG-V5-GFP. The coding region of Gate16, amplified with Phusion^® High-Fidelity DNA Polymerase, was cloned into a pCAG-V5-GFP vector at the SmaI site, which was named pCAG-V5-Gate16-GFP. The V5-Gate16 fragment amplified with specific primers and with Phusion^® High-Fidelity DNA Polymerase was digested by NheI and EcoRV (New England Biolabs) and then cloned into a FLAG-HA-pcDNA3.1 vector; this construct was named pcDNA-V5-Gate16.

Quantitative RT-PCR

Total RNAs isolated from the SMGs (n = 6 per group) were dissected with the QIAshredder and RNeasy mini extraction kit (QIAGEN), as previously described [59]. The following PCR primers were used for further specific analysis: Muc10, 5´-TCTCCTACCCAGGAGCAACATT-3´ and 5´-TAGTTCCTGGAGGGCACATCT-3´; Adrb1, 5´-CTGCTCATCGTGGTGGGTAA-3´ and 5´-GAAAGGCACCACCACGCAATC-3´; Amy1, 5´-ACGTTTAATCTGGCCTTTTGTTTG-3´ and 5´-CATTGGGTGGAGAGACCTGC-3´; Gate16, 5´-AAGGAGGACCACTCTCTGGAA-3´ and 5´-TTTTCCACGATCACCGGAACT-3´; Vamp2; 5´-ATGGGGATCCTCGTCCTTCT-3´ and 5´-AGAAAACCTGGTGTGTGGGG-3´; Vamp8; 5´-CTGGACCACCTCCGAAACAA-3´ and 5´-AGATGCAGCTACAGTCCAGC-3´; and Gapdh, 5´-AACTTTGGCATTTGGAAGG-3´ and 5´-ACACATTGGGGGTAGGAACA-3´. Information about the RT-PCR primers used in this study is included in Table S2.

Immunoblotting

The SMGs were dissected from Atg7 cKO, Atg3 cKO, and control mice at 8 weeks of age. Immunoblots were performed as previously described [60]. Information about the antibodies used in this study is included in Table S1.

Immunoprecipitation for endogenous proteins

The SMGs dissected from mice at 8 weeks of age were homogenized with IP lysis buffer (Thermo Fisher Scientific). The homogenates were centrifuged at 15,000 rpm for 10 min at 4 ℃, and the supernatants were used for subsequent immunoprecipitation. The Pierce Crosslink Immunoprecipitation kit (Pierce) was used for protein immunoprecipitation following the manufacturer’s protocol. Each antibody against VAMP2 (Cell Signaling Technology), VAMP8 (abcam), SNAP23 (abcam), and rabbit control IgG (Santa Cruz Biotechnology) was incubated and crosslinked with magnetic beads. Tissue lysates were incubated with each magnetic bead complex overnight at 4 ℃. Immunoprecipitated complexes were washed three times with IP lysis buffer, and one time with ultra-pure water at 4 ℃. Proteins were then eluted with 100 μL elution buffer and neutralized with 10 μL neutralization buffer. The immunoprecipitated proteins were analyzed by immunoblotting.

Immunoprecipitation for the overexpressed tagged proteins

COS-7 cells were obtained from the American Type Culture Collection (CRL-1651; ATCC) and maintained under Dulbecco’s Modified Eagle Medium (DMEM), supplemented with 10% fetal bovine serum (FBS) and penicillin/streptomycin, at 37 °C in a humidified atmosphere with 5% CO₂. COS-7 cells were plated onto a 60-mm dish at a density of 30,000 cells per dish. After 24 h, the cells were transfected with 2 μg of pcDNA-V5-Gate16 and 2 μg of pcDNA-FLAG-HA-Vamp2, pcDNA-FLAG-HA-Vamp8, or pcDNA-FLAG-HA-Snap23 using Lipofectamine 3000 transfection reagent (Thermo Fisher Scientific), according to the manufacturer’s protocol. After 48 h, the cells were lysed with lysis buffer (150 mM NaCl, 10 mM HEPES pH 7.4, 0.5% Triton-X) containing a protease inhibitor cocktail (Roche). The cell lysate was centrifuged at 18,000 × g for 20 min at 4 °C. The supernatant of each sample was collected, and the protein level was determined using the BCA protein kit (Pierce). For the immunoprecipitation experiments, each antibody against V5-tag (Thermo Fisher Scientific), FLAG-tag (Sigma Aldrich), or HA-tag (Thermo Fisher Scientific) was incubated with 25 μL of magnetic beads (Pierce) for 1 h at room temperature. The cell lysates (150 μg) were then incubated with each antibody-bead complex overnight at 4℃. The immunoprecipitated complexes were washed three times in IP lysis buffer (Thermo Fisher Scientific), recovered in Laemmli buffer, boiled for 5 min, and then analyzed by immunoblotting.

Enzyme-linked immunosorbent assay (ELISA)

Mouse anti-Ro60/SSA and anti-La/SSB ELISAs (Signosis), and multi-analyte ELISA arrays for mouse inflammatory cytokines (MEM-004A; Qiagen), were used according to the manufacturer’s protocol. Serum samples from mice were diluted at 1:50. The amount in tissues was normalized with SMG tissue volume.

Fluorescence-activated cell sorting (FACS) analysis

The cells were isolated from the spleens and the SMGs of Atg7 cKO, Atg3 cKO, and control mice at 8 weeks of age, and cell clumps were removed using a 70-μm nylon mesh strainer. The cells were then suspended with Cell Staining Buffer (Thermo Fisher Scientific) at a concentration of 5 × 10⁶ cells/mL and centrifuged at 500g for 5 min at 4 °C. The collected cells were incubated with Fc blocker (Biolegend) at 4 °C for 5 min and incubated for 30 min in the dark on ice with fluorescence-labeled antibodies listed in Table S1. The samples in 500 μL Cell Staining Buffer were analyzed with FACS Aria II (BD Biosciences). All data were analyzed using FlowJo (BD Biosciences).

Tissue explant culture

Tissue explants of the SGs were dissected out from 3-week-old C57BL6/J mice and incubated with 2 mg/mL collagenase in DMEM/F-12 medium (Thermo Fisher Scientific) supplemented with 5% FBS, 1% Insulin-Transferrin-Selenium-G supplement (MilliporeSigma), 2 nM triiodothyronine (MilliporeSigma), 1 μM hydrocortisone (MilliporeSigma), 50 ng/mL EGF (BioVision Inc.), and penicillin/streptomycin in an incubator at 37 °C for two days. According to the manufacturer's protocol, the tissue explants were treated with shRNA for Gate16 and control (Santa Cruz Biotechnology).

Histology

H&E, β-galactosidase, BrdU staining, immunohistochemistry, and TUNEL assays were performed as previously described [61]. Periodic acid-Schiff staining was performed with 0.5% Periodic acid and Schiff reagent (3,952,016, Sigma-Aldrich), followed by 1% alcian blue staining (pH 2.5). 0.04% Methylene blue staining was performed for detection of T cells. Information about the antibodies used for immunohistochemistry is included in Table S1. According to the manufacturer's instructions, the Click-iT^® Plus TUNEL Assay with Alexa 594 (Molecular probes, C10618) was used for detection of apoptotic cells. Fluorescence images were obtained using a confocal microscope (Ti-C2, Nikon), and color images were obtained using a light microscope (BX43, Olympus). A total of six animals per group was used in each experiment.

Transmission electron microscopy

Mice were fixed by cardiac perfusion with 2% paraformaldehyde and 2.5% glutaraldehyde in 0.1 M cacodylate buffer (pH 7.4) at 8 weeks of age. The SMGs were dissected and post-fixed with 2% OsO₄ in 0.1 M cacodylate buffer, and then dehydrated with a graded series of ethanol and embedded in Embed 812 (Electron microscopy sciences). 1-µm-thick semi-thin sections and 70 nm ultra-thin sections were cut with a Leica UC7 ultramicrotome (Leica Microsystems, Vienna, Austria) and stained with either toluidine blue or positive staining with uranyl acetate and lead citrate, respectively. Contrasted ultrathin sections were analyzed under a JEM 1010 transmission electron microscope (JEOL, USA, Inc., Peabody, MA) at an accelerating voltage of 80 kV. Digital images were obtained using the AMT Imaging System (Advanced Microscopy Techniques Corp, Danvers, MA) (n = 6 per group).

Statistical analysis

All results were obtained from at least three independent experiments. All experimental data were analyzed using Prism software (GraphPad Software). Two-tailed paired student’s t-tests were applied to compare two groups; one-way analysis of variance (ANOVA) with Dunnett's multiple comparison test was used for comparison of multiple groups. A p value ≤ 0.05 was considered statistically significant. For all graphs, data are represented as mean ± standard deviation (SD).

Supplementary Information

Below is the link to the electronic supplementary material.

Authors' contributions

AS and JI designed research; AS, CI, KO, HY, JS, and JI performed research; AS, IT, MK, NT, and JI contributed to generate new mouse models; AS, CI, KO, HY, and JI analyzed data; and AS and JI wrote the paper.

Funding

This study was supported by grants from the National Institute of Dental and Craniofacial Research (DE026767 to JI) and UTHealth School of Dentistry faculty funding to JI.

Availability of data and material

The data that support the findings in this study are available from the corresponding author upon reasonable request.

Declarations

Conflict of interest

The authors declare no competing interests.

Ethics approval and consent to participate

Not applicable.

Consent for publication

Not applicable.