LAMP3 induces apoptosis and autoantigen release in Sjögren’s syndrome patients

Primary Sjögren’s syndrome (pSS) is a complex autoimmune disease characterized by dysfunction of secretory epithelia with only palliative therapy. Patients present with a constellation of symptoms, and the diversity of symptomatic presentation has made it difficult to understand the underlying disease mechanisms. In this study, aggregation of unbiased transcriptome profiling data sets of minor salivary gland biopsies from controls and Sjögren’s syndrome patients identified increased expression of lysosome-associated membrane protein 3 (LAMP3/CD208/DC-LAMP) in a subset of Sjögren’s syndrome cases. Stratification of patients based on their clinical characteristics suggested an association between increased LAMP3 expression and the presence of serum autoantibodies including anti-Ro/SSA, anti-La/SSB, anti-nuclear antibodies. In vitro studies demonstrated that LAMP3 expression induces epithelial cell dysfunction leading to apoptosis. Interestingly, LAMP3 expression resulted in the accumulation and release of intracellular TRIM21 (one component of SSA), La (SSB), and α-fodrin protein, common autoantigens in Sjögren’s syndrome, via extracellular vesicles in an apoptosis-independent mechanism. This study defines a clear role for LAMP3 in the initiation of apoptosis and an independent pathway for the extracellular release of known autoantigens leading to the formation of autoantibodies associated with this disease. ClinicalTrials.gov Identifier: NCT00001196, NCT00001390, NCT02327884.


LAMP3 expression is increased in minor salivary glands of SS patients and associated with the presence of serum autoantibodies. Several independent transcriptomic studies investigating gene
expression changes in the salivary glands of SS patients have been reported [22][23][24][25] . Aggregated and integrated analysis of the available public data was employed to identify common transcriptomic changes in the disease and reduce experimental bias. Here microarray data from three studies investigating MSG biopsies, which met our inclusion criteria were aggregated and used to compare SS patients with three healthy or disease controls groups consisting of healthy volunteers, non-SS patients with sicca symptoms, and patients with IgG4-related disease. This composite analysis identified 279 common differentially expressed genes (Supplementary Table S1) [22][23][24] . Ingenuity pathway analysis demonstrated enrichment of pathways associated with interferon signaling, pattern recognition of pathogens, RhoGDI signaling, and apoptosis, some of which have been previously reported associated with SS (Supplementary Table S2). Decreased calcium signaling, G protein-coupled receptor mediated nutrient sensing, and amyloid processing were also found in the SS patients compared with controls (Supplementary Table S2).
One of the most significantly upregulated genes, excluding interferon associated genes, was LAMP3. Confirmatory qRT-PCR experiments demonstrated increased LAMP3 mRNA expression in a separate cohort of MSG samples from SS patients compared with healthy volunteers (HV) or MSG samples from non-SS patients with other autoimmune diseases (Fig. 1A).
Analysis of LAMP3 mRNA expression with clinical covariates demonstrated that high LAMP3 expression (defined as > 2SD over the mean level in healthy volunteers) is associated with the presence of serum autoantibodies. Specifically, a significant positive association with increased LAMP3 mRNA expression was found with anti-SSA autoantibody seropositivity in pSS (Fig. 1B) as well as anti-SSB, ANA, total IgG, and focus score (Supplementary Figure S1A-D). No association was found with other immune components such as complement C4 or C3, beta2 microglobulin levels.
Based on the increased LAMP3 mRNA, we immunolocalized LAMP3 positive cells within the salivary glands. Confocal immunofluorescent imaging demonstrated elevated expression of LAMP3 protein in both the epithelial and infiltrating lymphocytic cell compartments of SS MSG biopsies compared with non-SS control MSG biopsies (P < 0.05) (Fig. 1C,D). Since LAMP3 protein expression was not restricted to infiltrating or resident immune cells, the differential epithelial expression of LAMP3 using primary salivary gland epithelial cells (pSGECs) derived from healthy volunteers and SS patients was confirmed. Interestingly, LAMP3 mRNA was significantly increased in pSGECs derived from SS subjects compared with HV (p < 0.01) (Fig. 1E). These findings suggest an important link between increased LAMP3 expression and classical autoantibody seropositivity in SS. LAMP3 expression has previously been reported in dendritic cells but little is known about its role in epithelia. To examine the effect of LAMP3 on epithelial cell function, we established LAMP3-overexpressing HSG (HSG LAMP3-OE) and A253 (A253 LAMP3-OE) cells by transfection with the LAMP3 encoding plasmid, pME18S-LAMP3. Confocal immunofluorescent microscopy was also used to investigate the effect of LAMP3 overexpression on  26 . TRIM21 immunofluorescence demonstrated a diffuse, fine cytoplasmic and nuclear staining with scattered cells demonstrating a punctate nucleolar-like pattern in control HSG cells. In contrast, LAMP3 transfected cells showed nuclear redistribution of TRIM21 (multiple coarse dots spanning throughout the nucleus) and increased overall expression of TRIM21, including significant increases in the size of aggregates, the number of nuclear TRIM21 aggregates, and overall expression (P < 0.001) ( Fig. 2A,B). SSB demonstrated a similar staining pattern as TRIM21 in control cells and similar nuclear redistribution in LAMP3 overexpressing cell; although overall expression was less compared with TRIM21. LAMP3 overexpression increased the number of nuclear aggregates of SSB (P < 0.001), but did not affect the size; and there was a trend towards increased expression (Fig. 2C,D). In addition, LAMP3-OE cells commonly showed changes in nuclear structure associated with late apoptosis and apoptotic nuclear debris were commonly visible. Furthermore, the pattern of nuclear redistribution of TRIM21 (and to a lesser degree SSB) shown by immunofluorescence is indicative of LAMP3-specific initiation of apoptosis and are similar to apoptitic changes previously reported in SS patients 27 . LAMP3 induces apoptosis via caspase pathway. Analysis of the growth characteristics of LAMP3 overexpressing cells showed a significantly reduced cell growth in both HSG and A253 cell lines compared with control cells transfected with an empty vector, when determined by trypan blue exclusion assay (HSG: 60% decrease, P < 0.001; A253: 30% decrease, P < 0.05) (Fig. 3A,B).

LAMP3 overexpression induces expression and redistribution of SS autoantigens.
To determine whether the changes in cell growth rates were the result of alterations in cell cycle or increased apoptosis, the cell cycle in control and HSG and A253 LAMP3-OE cells was analyzed by Propidium Iodide (PI) and flow cytometry. LAMP3 expression in HSG or A253 cells induced a small but statistically significant change in the percentage of cells in different phases of the cell cycle compared with control treated cells (Supplementary Figure S2). However, these changes were likely too small to account for the observed change in growth.
To further understand the effect of LAMP3 increased expression in epithelial cells, markers of apoptosis were studied. HSG and A253 LAMP3-OE cells exhibit a significant time-and dose-dependent increase in the number  Figure S3). To test whether these results were specific to LAMP3 overexpression, HSG and A253 cells were co-transfected with a GFP reporter plasmid and either a LAMP3 or LAMP1 expression vectors and the number of apoptotic cells in the GFP + population were counted. The number of apoptotic cells were significantly increased in the LAMP3 co-transfected GFP + cells compared with GFP + control cells and LAMP1 co-transfected GFP + cells at both early and late stages of apoptosis ( Fig. 3I-L). Furthermore, no differences in the number of apoptotic cells were found between the LAMP1-OE cells and control cells for both cell lines ( Fig. 3I-L) suggesting the increase in apoptosis was specific to LAMP3 expression. Caspases are well-known initiators of apoptosis 28,29 . Treatment of LAMP3-OE cells with the pan-caspase inhibitor, Z-VAD, which also inhibits caspase-3 30 , showed a significant decrease in apoptosis compared with LAMP3-OE controls (LAMP3 + Z-VAD vs. LAMP3: 6.4 ± 0.3% vs. 14.5 ± 1.2%, P < 0.01) (Fig. 3M,N). Taken together, these results imply that LAMP3-associated decrease in cell growth is caused by caspase-dependent induction of apoptosis.
LAMP3 induces apoptosis via caspase-3 independently of endoplasmic reticulum stress. LAMP3-OE cells were also used to investigate the effect of LAMP3 on caspase-3 expression, caspase-3 mRNA and protein levels. Athough no difference in caspase-3 mRNA expression between LAMP3-OE and control cells (Fig. 4A,D) was found, caspase-3 protein expression was decreased in LAMP3-OE cells compared with control cells (Fig. 4B,E), suggesting that LAMP3 overexpression alters caspase-3 cleavage and activation. Endoplasmic reticulum (ER) stress is often linked to apoptosis via caspase activation 31,32 . Therefore, LAMP3-induced ER stress was monitored by following the change in expression of ATF4 and activation of XBP-1, well-known markers of ER stress. No increase in expression of ATF4 or activation of XBP1 was observed in LAMP3-OE cells compared with control cells (Fig. 4C,F), suggesting that LAMP3 induced apoptosis was independent of ER stress.
LAMP3 induces accumulation and release of autoantigens via extracellular vesicles. Extracellular vesicles (EVs), including apoptotic bodies, are thought to be sources of neoantigens within the pathogen- www.nature.com/scientificreports/ esis of autoimmune diseases [33][34][35][36] . Based on the finding that LAMP3 expression in the MSG biopsies from SS patients impacted the survival of the cells by inducing caspase-mediated apoptosis and changed the expression and localization of key autoantigens associated with SS ( Fig. 2). We hypothesized that LAMP3-induced release of TRIM21/SSA and La/SSB might be a source of autoantigens and account for the association between MSG LAMP3 expression and the presence of autoantibodies in pSS patients. Using confocal immunofluorescent microscopy, a membranous expression pattern of LAMP3, particularly in structures appearing to bud from the plasma membrane, was observed with the appearance of EVs. In LAMP3-OE cells a clear colocalization of LAMP3 with TRIM21/SSA and SSB was observed within EVs (open triangles, Fig. 5A). Western blotting showed a difference in TRIM21 expression compared with control cells. However, no difference in the expression of SSB or α-fodrin was observed (Fig. 5B,C). Analysis of the protein content of the EVs isolated from HSG LAMP3-OE showed an increase compared with EVs isolated from control cells (Fig. 5D). Moreover, western blotting of the EVs showed a significant increase in the levels of TRIM21, SSB, and α-fodrin in EVs from LAMP3-OE cells compared with control cells (Fig. 5E-G). However, no difference was detected in the amount of cleaved α-fodrin protein in the EVs from LAMP3-OE cells compared with control cells (Fig. 5G). These results suggest that in addition to the increase in LAMP3-associated apoptosis, LAMP3 over expression induces the release of autoantigens from the cell.

LAMP3-induced autoantigens release is independent of apoptosis. One important question was
whether LAMP3-associated apoptosis was responsible for antigen release. To clarify this, HSG LAMP3-OE cells were treated with Z-VAD, an inhibitor of apoptosis, and the release of autoantigens in EVs was measured. No siginificantly difference was found in the protein content of the EVs isolated from HSG LAMP3-OE cells treated with Z-VAD compared with EVs from LAMP3-OE control cells, in which both treated and untreated HSG-OE cells showed increased EVs compared with control cells (Fig. 6A). Western blotting also showed no difference in autoantigens, TRIM21, SSB and α-fodrin protein levels in the EVs isolated from HSG LAMP3-OE cells treated with Z-VAD compared with HSG LAMP3-OE control cells ( Fig. 6B-E). These results suggested that LAMP3induced autoantigen release is independent of apoptosis.

Discussion
Sjögren's syndrome is a complex disorder of uncertain etiology and pathogenesis. Patients exhibit a heterogenous clinical presentation with multiple molecular subsets of patients which is likely responsible for the lack of universal predictive biomarkers or effective treatments. However, one common feature shared in approximately 70% of the patients is the presence of specific autoantibodies. To gain further insight into the drivers of disease, we employed aggregated transcriptome analysis using previously published transcriptomic studies to analyze the MSG biopsies from SS patients and controls to illuminate common genes and pathways. From this approach, we found increased LAMP3 mRNA expression in a subset of pSS patients. Unexpectedly, we found that patients exhibiting increased LAMP3 expression also were patients with anti-SSA and anti-SSB autoantibodies. Futhermore, this increase could be observed in primary cultures from the patients suggesting SS-related transcriptional reprogramming may not entirely dependent upon immune microenvironment stimulation. Our investigation of the biological impact of LAMP3 upregulation in epithelial cells suggests that it can initiate cell death and apoptosis independent autoantigen release.
Serum autoantibodies against several intracellular proteins (e.g., TRIM21 (Ro52), La/SSB) are found in approximately 70% of pSS patients who meet diagnostic criteria 3,4 . Given the commonality of these serum autoantibodies across a majority of patients, it is likely that these proteins play key roles in the pathogenesis of SS. It has been proposed that phagocytosis of apoptotic bodies and debris leads to the presentation of autoantigens and subsequently to the induction of autoimmunity. However, the mechanism responsible cell death and the resultant release of autoantigens that triggers these aspects of the disease remains largely unknown.
Although anti-Ro/SSA and anti-La/SSB autoantibodies are sometimes found in other autoimmune diseases such as systemic lupus erythematosus, a correlation with a specific differentially expressed gene has not been established. This may be the result of the localized nature of SS tissue destruction to secretory epithelial tissues and limited access to the affected tissue. Our key finding represents a previously unrecognized connection between autoantibody formation, lysosomal protein expression, and apoptosis. Further inquiry into changes in the expression of lysosomal proteins and the induction of autoantibodies in other autoimmune diseases is warranted.
In this study we identified a novel mechanism for apoptosis induction in the course of developing autoimmunity. In turn apoptosis can induce the cell surface expression of autoantigens such as TRIM21, La/SSB, and DNA, and the release of autoantigens. TRIM21 is an attractive protein for the development of autoimmunity given its role in the cellular innate immune response and surveillance to intracellular pathogens such as viruses. TRIM21 recognizes intracellular pathogens, binds to the Fc portion of the human IgG-bound to pathogens marking them for degradation and initiates the innate immune response 37,38 . Additional studies indicate that immunizing with antigens such as TRIM21, La/SSB, and α-fodrin leads to autoantibody production and the autoimmune disease [39][40][41][42] . Additionally, cell surface autoantigen expression induced by apoptosis elevates autoantibody production via antigen presenting cells (APCs) 43 . Moreover, apoptotic bodies can stimulate dendritic cells, and immunizing with apoptotic cells leads to auto-IgG production including autoantibodies against TRIM21 and SSB 44 . LAMP3 expression increased the accumulation and aggregation of TRIM21, and the release of TRIM21, SSB, and α-fodrin (Figs. 2, 5A-G), suggesting that LAMP3 can function as a potent initiator of the release of neoantigens via EVs. Taken together, these results support the idea that LAMP3 plays an important role in the activation of autoimmune responses through accumulation of TRIM21, and the release of TRIM21, La/SSB, and α-fodrin as autoantigen via EVs independent of apoptosis. www.nature.com/scientificreports/ Endoplasmic reticulum stress has been implicated in the pathogenesis of a wide variety of human diseases, including inflammatory disease 45 and autoimmune disease 46 . Increased ER stress markers in MSG biopsies from SS patients have been reported 47 . The unfolded protein response (UPR) is activated by, and actively modulates, ER stress, to maintain homeostasis and energy balance within cells. LAMP3 is activated downstream of the PERK-ATF4 signaling pathway, which is part of the UPR pathway, and is induced by ER stress 48 . Although ATF4 expression was not increased by LAMP3 expression in this study, we postulate that it could be increased by the same environmental stimuli that are likely responsible for the resultant increase in LAMP3 expression and represents a central aspect to the initiation of SS. This study showed a clear link between caspase-dependent apoptosis and LAMP3 expression. Further studies are needed to investigate the effect of LAMP3 on the balance between apoptotic and cell survival pathways.
LAMP3 overexpressing cells showed cell death typical for apoptosis. LAMP3 is classically associated with the lysosome, a main organelle central to autophagy. This cell survival mechanism is also known to kill cells under certain conditions in a process called autophagic cell death, which involves pathways and mediators different from those of apoptosis [49][50][51] . Although our data supports an apoptotic pathway in SS tissue destruction, it is possible that LAMP3 expression may also affect autophagy, leading to apoptotic cell death. Cross talk between apoptosis and autophagy via the lysosome maybe central to the underlying pathogenesis in Sjögren's syndrome.
This study shows that LAMP3 expression contributes to the induction of apoptosis, which has long been associated with SS by an unknown mechanism. Interestingly the LAMP3 associated release of autoantigens is independent of apoptosis. The results presented here suggest a previously unrecognized context for autoantigen presentation and defines a central role for LAMP3 in pSS. Autoantibody formation is not unique to Sjögren's syndrome and this mechanism of autoantigen release maybe at work in other autoimmune diseases. Currently therapeutics are being developed to block the caspase-dependent apoptosis. Additional therapeutic targets may emerge by investigating how LAMP3 is specifically inducing apoptosis and preventing the development of autoantibodies following antigen release.

Materials and methods
Study selection and data aggregation. Previously published microarray studies on Sjögren's syndrome were reviewed using inclusion and exclusion criteria [22][23][24] . Inclusion criteria were for three or more patients, array platforms with more than 40,000 genes, available raw data sets, clearly defined diagnostic criteria based on AECG 2002 criteria 4 , and use of high quality RNA isolated from minor salivary glands. Exclusion criteria included low density arrays, samples on mice and those that used blood samples.
Selected data sets were collected from NCBI gene expression omnibus (GEO) using GeneSpring Multi-Omic Analysis version 14.9 52 . A total of three data sets representing four clinically-defined subsets of patients were identified in GEO: GSE40568, GSE23117, GSE127952.
Each data set was treated as a separate differential gene expression analysis experiment and was analyzed without further normalization. Differentially expressed genes were then aggregated by GeneSpring and used as input for Ingenuity Pathways Analysis (IPA) to identify common genes and pathways.
Clinical studies. MSG biopsies used in the transcriptional studies were obtained from two centers: the Sjögren's Syndrome Clinic at the National Institute of Dental and Craniofacial Research (NIDCR), National Institutes of Health (NIH), in Bethesda, MD, and from Clinics Hospital of the Medical School of Ribeirao Preto (CHRMSRP), University of São Paulo, São Paulo, Brazil. Primary salivary gland epithelial cells (pSGECs) were drived from MSG biopsies obtained from human subjects in the Sjögren's Syndrome Clinic at NIDCR. All studies using human tissues were carried out in accordance with approved NIH guidelines. Subjects under study provided informed consent prior to the initiation of any study procedures. NIH human tissues were obtained from NIH Institutional Review Board (IRB) approved protocols (ClinicalTrials.gov Identifiers: NCT00001390, NCT02327884, or NCT00001196) in the Sjögren's Syndrome Clinic at the NIDCR, NIH in Bethesda, MD. Likewise, CHMSRP human tissues were obtained from subjects who provided informed consent to an IRB-approved  . These cells were confirmed as mycoplasma free by using MycoAlert (Lonza, Allendale, NJ, USA). pME18S-empty and pME18S-LAMP3 plasmids were prepared by cloning the LAMP3 open reading frame into pME18S expression vector containing a Kozak consensus sequence 54 . The full nucleotide sequences   www.nature.com/scientificreports/ Measurement of LAMP3 gene expression in culture of human pSGECs. LAMP3 mRNA expression was measured using quantitative real-time polymerase chain reaction and Taqman primer sets. Briefly, equal amount of total RNAs of each pSGECs were first reverse-transcribed into cDNA using iScript supermix (BioRad, Hercules, CA, USA). cDNAs were amplified for ACTB (Hs01060665_g1) and LAMP3 (Hs01111316_ m1) and data was collected using a StepOnePlus (Applied Biosystems). ACTB was used as an internal control for normalization of input cDNA, and the difference of the cycle threshold (Ct) LAMP3 was calculated using the ΔCt method and used to determine the relative quantitation (RQ) values (2 −ΔΔCt ), which represent the relative level of fold change to control.   Apoptosis assay. Five × 10 5 HSG and A253 cells were transfected with a total amount of 1.5 μg pME18Sempty or pME18S-LAMP3 plasmid, and 5 × 10 5 HSG and A253 cells were co-transfected with 0.75 μg AAV2-GFP and 0.75 μg in pME18S-empty, pME18S-LAMP3, pUC or pCMV-LAMP1 plasmid using Lipofectamine 3000. Twenty-four hours after transfection, the cells were re-plated at 3 × 10 5 cells per well in 6-well plate, then 24-or 48-h after re-plating, cells were used to detect apoptotic cells. To examine whether the apoptosis induced by LAMP3 was via activated caspase, cells were treated with or without 20 μM Z-VAD-FMK (Z-VAD) 24 h after transfection. After incubation for 20 h, cells were trypsinized and used to detect apoptotic cells by flow cytometry using the APC Annexin V Apoptosis Detection Kit with 7-AAD (640930, Biolegend, San Diego, CA, USA) and the BD Accuri (BD Biosciences, San Jose, CA, USA) using the BD CSampler software.
Statistical analysis. Statistical analysis of microarray studies was performed as described above. Twotailed Student's t-or analysis of variance (ANoVA) tests, where appropriate, were employed as indicated using JMP 13.2.0 (SAS Institute, Cary, NC). P-values less than 0.05 were considered statistically significant. Sample size calculations were based on an independent two-sample t-test with unequal variances to show that group sizes of www.nature.com/scientificreports/ at least 16 would be sufficient to have at least 80% power to detect differences where the difference in the means was at least twofold and the standard deviation equal to the difference in the means at a significant 5% level.
Human subjects research declaration. All clinical investigations were conducted in accordance with the Declaration of Helsinki principles. Written informed consent to IRB-approved protocols (as described above) were obtained from all participants prior to inclusion to the studies described herein. All human studies were approved by the appropriate institutional review boards at each site where tissues were procured. www.nature.com/scientificreports/ Publisher's note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http:// creat iveco mmons. org/ licen ses/ by/4. 0/. This is a U.S. Government work and not under copyright protection in the US; foreign copyright protection may apply 2020