DNA Damage Regulates Senescence-Associated Extracellular Vesicle Release via the Ceramide Pathway to Prevent Excessive Inflammatory Responses

DNA damage, caused by various oncogenic stresses, can induce cell death or cellular senescence as an important tumor suppressor mechanism. Senescent cells display the features of a senescence-associated secretory phenotype (SASP), secreting inflammatory proteins into surrounding tissues, and contributing to various age-related pathologies. In addition to this inflammatory protein secretion, the release of extracellular vesicles (EVs) is also upregulated in senescent cells. However, the molecular mechanism underlying this phenomenon remains unclear. Here, we show that DNA damage activates the ceramide synthetic pathway, via the downregulation of sphingomyelin synthase 2 (SMS2) and the upregulation of neutral sphingomyelinase 2 (nSMase2), leading to an increase in senescence-associated EV (SA-EV) biogenesis. The EV biogenesis pathway, together with the autophagy-mediated degradation pathway, functions to block apoptosis by removing cytoplasmic DNA fragments derived from chromosomal DNA or bacterial infections. Our data suggest that this SA-EV pathway may play a prominent role in cellular homeostasis, particularly in senescent cells. In summary, DNA damage provokes SA-EV release by activating the ceramide pathway to protect cells from excessive inflammatory responses.


Introduction
DNA damage can have numerous chemical and biological effects on cellular function and may induce apoptotic cell death or cellular senescence in healthy cells [1,2]. Cellular senescence is an important tumor suppression mechanism. It acts as a barrier against various oncogenic stresses, such as telomere shortening, oncogene activation, irradiation, or stimuli that damage DNA [3][4][5]. These stressors activate p53, followed by an increase in p21 WAF1/CIP1 and/or p16 INK4a expression. These are cyclin-dependent kinase inhibitors that promote the activation of retinoblastoma protein (RB),

DNA Damage Induces Small EV Secretion from Normal Human Fibroblasts and Epithelial Cells
To evaluate the effect of DNA damage on small EV secretion from normal human diploid fibroblasts (HDFs), TIG-3 cells were treated with a DNA-damaging agent, doxorubicin (DXR). Treatment with DXR increased the signs of the DNA damage response (DDR) in a dose-dependent manner, as judged by DNA damage foci (the phosphorylation of H2AX and the consensus target sequences of ATM/ATR) ( Figure 1A). Consistent with several previous reports [24,26,32,33], nanoparticle tracking analysis (NTA) revealed that an increase in small EV secretion was concomitant with the level of DNA damage ( Figure 1B). Since DNA damage is known to cause cellular senescence in normal cells, we investigated the molecular mechanism of small EV secretion, induced by DNA damage, using both young and senescent HDFs. We confirmed the induction of cellular senescence by RT-qPCR analysis of p16 INK4a , a well-established marker of cellular senescence, in DNA damage-or oncogene (HRasV12)-induced senescent cells ( Figure 1C) [6][7][8]. Recent studies have shown that there are many factors involved in regulating the biogenesis and release of small EVs, such as the ceramide pathway, ESCRT (endosomal sorting complexes required for transport), and Rab-family small GTPases [34][35][36][37]. Of these, we focused on lipid-related proteins because the sphingolipid pathway might modulate the budding of intraluminal membrane vesicles (ILVs) into multivesicular bodies (MVBs) and regulate small EV biogenesis [34,38]. Here, we revealed that the gene expression of sphingomyelin synthase 2 (SMS2) was downregulated significantly while, conversely, that of neutral sphingomyelinase 2 (nSMase2) was upregulated in both types of senescent cells ( Figure 1C,D). Sphingomyelinase activation and sphingomyelin synthase inhibition result in ceramide production, which is related to the promotion of small EV biogenesis and increased EV release from several cell types [34,[38][39][40][41]. Therefore, we speculated that activating the ceramide synthetic pathway after DNA damage might be important for small EV release from both healthy control and senescent cells. To confirm our findings in another cell line, human retinal pigment epithelial cells (RPE-1 cells) were also treated with DXR to induce cellular senescence. In accordance with DDR induction and p16 INK4a upregulation ( Figure 1E, F), the gene expression of both SMS2 and nSMase2 in the epithelial cells was changed in a similar manner to that seen in senescent HDFs ( Figure 1F,G). Next, small EVs were subject to NTA, demonstrating that the release of small EVs also increased in senescent epithelial cells, compared with their release in healthy control cells ( Figure 1H). In addition, transmission electron microscopy analysis using immuno-gold labelling for CD63, a well-known exosome marker [42], showed that these cells were secreting exosomes among the small EVs ( Figure 1I). These data indicated that DNA damage activates the ceramide synthetic pathway in both HDFs and epithelial cells.  To confirm our findings in another cell line, human retinal pigment epithelial cells (RPE-1 cells) were also treated with DXR to induce cellular senescence. In accordance with DDR induction and p16 INK4a upregulation ( Figure 1E, F), the gene expression of both SMS2 and nSMase2 in the epithelial cells was changed in a similar manner to that seen in senescent HDFs ( Figure 1F,G). Next, small EVs were subject to NTA, demonstrating that the release of small EVs also increased in senescent epithelial cells, compared with their release in healthy control cells ( Figure 1H). In addition, transmission electron microscopy analysis using immuno-gold labelling for CD63, a well-known exosome marker [42], showed that these cells were secreting exosomes among the small EVs ( Figure  1I). These data indicated that DNA damage activates the ceramide synthetic pathway in both HDFs and epithelial cells.

Activation of the Ceramide Synthetic Pathway Promotes Small EV Release from Cells
The expression levels of both SMS2 and nSMase2 changed in senescent cells; therefore we investigated these proteins' roles in small EV release from HDFs. First, we used small interfering RNA (siRNA) to knock-down SMS2 [43], causing a significant induction of small EV secretion from HDFs, as determined by NTA (Figure 2A-C). Conversely, SMS2 overexpression reduced the level of small EV secretion after DXR treatment ( Figure 2D,E). Second, nSMase2 depletion substantially reduced small EV secretion ( Figure 2F-H) [38]. Importantly, inhibiting small EV secretion provoked the aberrant activation of DNA damage signaling in normal HDFs, as previously reported ( Figure 2I) [24]. Furthermore, nSMase2 overexpression resulted in remarkably enhanced small EV release ( Figure 2J,K). Taken together, these results revealed that activating the ceramide synthetic pathway promotes the release of small EV from cells.

Activation of the Ceramide Synthetic Pathway Promotes Small EV Release from Cells.
The expression levels of both SMS2 and nSMase2 changed in senescent cells; therefore we investigated these proteins' roles in small EV release from HDFs. First, we used small interfering RNA (siRNA) to knock-down SMS2 [43], causing a significant induction of small EV secretion from HDFs, as determined by NTA (Figures 2A-C). Conversely, SMS2 overexpression reduced the level of small EV secretion after DXR treatment ( Figure 2D, E). Second, nSMase2 depletion substantially reduced small EV secretion ( Figures 2F-H) [38]. Importantly, inhibiting small EV secretion provoked the aberrant activation of DNA damage signaling in normal HDFs, as previously reported ( Figure 2I) [24]. Furthermore, nSMase2 overexpression resulted in remarkably enhanced small EV release ( Figure 2J,K). Taken together, these results revealed that activating the ceramide synthetic pathway promotes the release of small EV from cells.  The percentage of nuclei that contain more than 3 DNA damaging foci positive were shown in the histograms (I). (J,K) Pre-senescent TIG-3 cells were infected with retrovirus encoding FLAG-tagged nSMase2 or empty vector. After selection with puromycin, cells were subjected to western blotting (J), or to NanoSight analysis of isolated small EV particles (K). For all graphs, error bars indicate mean + standard deviation (s.d.) of triplicate measurements. P values was calculated by unpaired two-tailed Student's t-test (** p < 0.01, *** p < 0.001).

Small EV Release Via the Ceramide Pathway Prevents DNA Damage Accumulation in Mice
In order to examine the effect of the ceramide synthetic pathway on both small EV release and tissue homeostasis in vivo, we used a chemical inhibitor of nSMase, spiroepoxide, which blocks small EV production in human cells [24,41]. We also observed the same effects in mouse embryonic fibroblasts (MEFs) by spiroepoxide treatment ( Figure 3A). It is notable that inhibiting the ceramide pathway clearly induced cell cycle arrest and DNA damage accumulation in MEFs ( Figure 3B,C). Next, we treated mice with spiroepoxide for 14 days. As expected, the inhibitor treatment reduced small EV release from the small intestine and accumulated DNA damage in mice tissues ( Figure 3D,E). Collectively, our data strongly suggested that the ceramide pathway plays a crucial role in vivo in releasing small EVs and maintaining tissue homeostasis to avoid DNA damage accumulation. cells were subjected to western blotting (J), or to NanoSight analysis of isolated small EV particles (K). For all graphs, error bars indicate mean +standard deviation (s.d.) of triplicate measurements. P values was calculated by unpaired two-tailed Student's t-test (**P < 0.01, ***P < 0.001).

Small EV Release Via the Ceramide Pathway Prevents DNA Damage Accumulation in Mice.
In order to examine the effect of the ceramide synthetic pathway on both small EV release and tissue homeostasis in vivo, we used a chemical inhibitor of nSMase, spiroepoxide, which blocks small EV production in human cells [24,41]. We also observed the same effects in mouse embryonic fibroblasts (MEFs) by spiroepoxide treatment ( Figure 3A). It is notable that inhibiting the ceramide pathway clearly induced cell cycle arrest and DNA damage accumulation in MEFs ( Figure 3B,C). Next, we treated mice with spiroepoxide for 14 days. As expected, the inhibitor treatment reduced small EV release from the small intestine and accumulated DNA damage in mice tissues ( Figure 3D, E). Collectively, our data strongly suggested that the ceramide pathway plays a crucial role in vivo in releasing small EVs and maintaining tissue homeostasis to avoid DNA damage accumulation. The histograms indicate the percentage of nuclei that contain more than 3 foci positive for 53BP1 staining. At least 100 cells were scored per group (C). (D, E) ICR (CD1) mice were intraperitoneally injected with spiroepoxide every two days. After 14 days, the mice were euthanized and small intestines were subjected to NanoSight analysis (NTA) of isolated small EV particles (D) or to immunofluorescence analysis of intestine section (E). Section of intestines were subjected to immunofluorescence staining for markers of DNA damage (53BP1 [red] and DAPI [blue]) (E). The representative data from three independent experiments are shown. For all graphs, error bars indicate mean +standard deviation (s.d.) of triplicate measurements. P values was calculated by unpaired two-tailed Student's t-test (*P < 0.05, **P < 0.01, ***P < 0.001). The histograms indicate the percentage of nuclei that contain more than 3 foci positive for 53BP1 staining. At least 100 cells were scored per group (C). (D,E) ICR (CD1) mice were intraperitoneally injected with spiroepoxide every two days. After 14 days, the mice were euthanized and small intestines were subjected to NanoSight analysis (NTA) of isolated small EV particles (D) or to immunofluorescence analysis of intestine section (E). Section of intestines were subjected to immunofluorescence staining for markers of DNA damage (53BP1 [red] and DAPI [blue]) (E). The representative data from three independent experiments are shown. For all graphs, error bars indicate mean +standard deviation (s.d.) of triplicate measurements. p values was calculated by unpaired two-tailed Student's t-test (* p < 0.05, ** p < 0.01, *** p < 0.001).

The Autophagy Pathway Prevents Accumulation of Chromosomal DNA Fragments in the Cytoplasm to Block SASP-Factor Gene Expression in Accompany with Small EV Release
The accumulation of chromosomal DNA fragments in the cytoplasm promotes aberrant activation of the DNA-sensing pathway and causes global SASP-factor gene expression, both of which are associated with chronic inflammation and various age-related pathologies [15,16]. Cytoplasmic DNA fragments are so dangerous that normal cells prevent their accumulation via several degradation mechanisms, such as cytoplasmic DNases and the autophagy pathway [44]. We previously reported that the expression of the cytoplasmic DNases, DNase2 and TREX1, is regulated by E2F transcription factors and thereby downregulated by DNA damage [22]. In addition, the small EV secretion pathway also removes harmful cytoplasmic DNA fragments from cells [24,45]. However, the relationship between small EV release and autophagy in the regulation of cytoplasmic DNA fragments has remained unclear. Previous reports indicated that activating ceramide pathway blocks the autophagy-mediated degradation pathway [46]. Therefore, we speculated that DNA damage might block the autophagy pathway through ceramide pathway activation and, in turn, promote small EV release to prevent the accumulation of chromosomal DNA fragments in the cytoplasm of normal cells. We treated TIG-3 cells with autophagy pathway inhibitors that target lysosomes, chloroquine (CQ) and bafilomycin A1 (BafA1) [47]. Autophagy pathway inhibition caused an accumulation of chromosomal DNA fragments in the cytoplasm and cell cycle arrest in normal HDFs ( Figure 4A-C). Significantly, the gene expression of a number of SASP factors was upregulated by these inhibitors ( Figure 4D). Intriguingly, these treated cells also showed increases in the release of small EVs and double strand DNA (dsDNA) fragments from cells ( Figure 4E,F). Similar to the situation in HDFs, DNA damage enhanced SASP-factor gene expression and small EV release in ATG5 knockout cells compared with their expression and release, respectively, in wild-type cells (WT) (Supplementary Figure S1) [48]. Research has previously shown that autophagy dysregulations can be detected in aged cells, and autophagy activation tended to improve senescent phenotypes [49,50]. Therefore, we treated senescent HDFs with rapamycin, an mTOR signaling inhibitor and autophagy inducer [51][52][53]. Strikingly, rapamycin treatment blocked IFN-β and CXCL10 expression in senescent cells (Supplementary Figure S2). Taken together, these data strongly indicated that downregulation of the autophagy pathway promotes small EV release for the secretion of harmful chromosomal DNA fragments from cells ( Figure 4G). (G) A model of the molecular mechanism. DNA damage activates the ceramide synthetic pathway, both through SMS2 downregulation and nSMase2 upregulation, blocking autophagy and promoting the release of small EVs. For all graphs, error bars indicate mean ± standard deviation (s.d.) of triplicate measurements. P values was calculated by unpaired two-tailed Student's t-test (**P < 0.01, ***P < 0.001).

Small EV Release and Autophagy Cooperatively Prevent Cell Death Caused by Bacterial Infection
Previously, we reported that small EV secretion prevented adenovirus infection by excluding viral DNA from cells [24]. Additionally, we assessed the interplay between the release of small EVs and autophagy in preventing inflammation caused by bacterial infection. After differentiation into mature macrophage-like cells by Phorbol 12-myristate 13-acae-tate (PMA) stimulation for 7 days, human monocytic leukemia cells (THP-1 cells) were then infected with Bacillus Calmette-Guérin (BCG) vaccine, which is an attenuated form of Mycobacterium bovis, with or without inhibiting small

Small EV Release and Autophagy Cooperatively Prevent Cell Death Caused by Bacterial Infection
Previously, we reported that small EV secretion prevented adenovirus infection by excluding viral DNA from cells [24]. Additionally, we assessed the interplay between the release of small EVs and autophagy in preventing inflammation caused by bacterial infection. After differentiation into mature macrophage-like cells by Phorbol 12-myristate 13-acae-tate (PMA) stimulation for 7 days, human monocytic leukemia cells (THP-1 cells) were then infected with Bacillus Calmette-Guérin (BCG) vaccine, which is an attenuated form of Mycobacterium bovis, with or without inhibiting small EV biogenesis or secretion using siRNA oligos against Alix or Rab27a, as previously described ( Figure 5A,B) [24]. Bacterial infection activates inflammatory responses in host cells. Indeed, inhibiting the small EV pathway dramatically increased the expressions of inflammatory genes, such as IFN-β and CXCL10 (Figure 5C), resulting in apoptotic cell death ( Figure 5D). Notably, the cleaved forms of both Caspase 3 and Caspase 1, markers of inflammasome activation, were clearly detected by the inhibition of small EV secretion in THP-1 cells following infection with BCG ( Figure S3). Since bacterial DNA activates inflammasomes via the cGAS-STING pathway [54][55][56][57], STING depletion using previously validated siRNA oligos blocked both inflammatory gene expression and apoptotic cell death, suggesting that small EV release prevents excessive inflammatory responses caused by STING activation. The autophagy pathway also acts as a barrier against bacterial infection [58]. The levels of bacterial genomic DNA in host cells and small EV release increased considerably when autophagy was inhibited by adding 3-MA, an autophagy inhibitor ( Figure S4A,B) [59]. Importantly, apoptotic cell death increased significantly by inhibiting both autophagy and EV biogenesis (Supplementary Figure S4C). Collectively, these results demonstrate that small EV release and autophagy function cooperatively to prevent cellular inflammatory responses against bacterial infection. forms of both Caspase 3 and Caspase 1, markers of inflammasome activation, were clearly detected by the inhibition of small EV secretion in THP-1 cells following infection with BCG ( Figure S3). Since bacterial DNA activates inflammasomes via the cGAS-STING pathway [54][55][56][57], STING depletion using previously validated siRNA oligos blocked both inflammatory gene expression and apoptotic cell death, suggesting that small EV release prevents excessive inflammatory responses caused by STING activation. The autophagy pathway also acts as a barrier against bacterial infection [58]. The levels of bacterial genomic DNA in host cells and small EV release increased considerably when autophagy was inhibited by adding 3-MA, an autophagy inhibitor (Figure S4 A, B) [59]. Importantly, apoptotic cell death increased significantly by inhibiting both autophagy and EV biogenesis (Supplementary Figure S4C). Collectively, these results demonstrate that small EV release and autophagy function cooperatively to prevent cellular inflammatory responses against bacterial infection.

Discussion
Senescent cells are metabolically active in a state of stable cell cycle arrest; they accumulate in the living body during the aging process [9][10][11]. These cells have been reported to play both beneficial and deleterious roles in our health through secreting SASP factors [12][13][14][15][16][17], and they also release many types of EVs, such as exosomes, microvesicles, nucleosomes, and apoptotic bodies, which are characterized by their size and secretory machinery [60][61][62][63]. Previously, we and other groups have reported that small EVs are actively released from senescent cells, functioning as harmful SASP factors by regulating the growth and viability of cancer cells [24,25,31]. According to recent studies, the biological functions of small EVs released from senescent cells change drastically because of changes in the composition of their proteins, lipids, and nucleic acids during cellular senescence [25,26,60,[62][63][64]. In this study, we demonstrated that DNA damage is a key trigger, not only of senescence induction but also of small EV biogenesis via ceramide synthesis in senescent cells. The most common downstream mediator of DNA damage signaling, p53, reportedly induces the expression of several genes involved in endosome regulation and promote small EVs production [65]. In agreement with a previous report, we observed that DNA damage increased nSMase2 expression in HDFs and epithelial cells, suggesting that p53 activation might be involved in this phenomenon [66]. Additionally, we discovered that the expression of SMS2, an antagonistic enzyme of nSMase2 ceramide synthesis, reduced dramatically during cellular senescence ( Figure 1C,D,F,G). The activation of sphingolipid-metabolizing enzymes is important for clearing amyloid beta (Aβ) protein from the brain is associated with preventing Alzheimer's disease [40]. Therefore, it would appear to be important to reveal the mechanism involved in regulating SMS2 expression by DNA damage, and further investigations are required.
Cytoplasmic DNA fragments, derived from endogenous chromosomal or mitochondrial DNA resulting from DNA damage or from exogenous viral or bacterial infection, are detected by specific DNA-sensing machinery, such as the cGAS-STING pathway. Recent studies have shown that the cytoplasmic DNA-sensing pathway is critical for SASP-factor gene expression in senescent cells [19][20][21][22][23]. The innate immune response is important in fighting viral or bacterial infections, although excessive activation of this pathway can be dangerous for cells because the sustained activation of inflammasomes can result in additional DNA damage and/or cell death. Therefore, both DNA degradation enzymes and autophagy work to clear harmful cytoplasmic DNA fragments. However, in senescent cells, these cellular fail-safe mechanisms may not function normally because of persistent DNA damage signaling. Thus, the EV-mediated secretion pathway may also play a role in clearing cytoplasmic DNA fragments (see the model in Figure 6).
Altogether, we have elucidated that DNA damage can provoke senescence-associated EV (SA-EV) secretion by activating the ceramide pathway, thereby helping to eliminate hazardous DNA fragments from cells. It is notable that there might be alternative mechanisms for SA-EV secretion, and these could be potential therapeutic targets for the prevention of SASP, along with senolytic drugs to eliminate senescent cells [67]. Therefore, we will conduct further studies to address the molecular mechanisms underlying SA-EV biogenesis. Figure 6. A model of our research. In healthy conditions, endogenous dsDNA fragments, derived from chromosomal or mitochondrial DNA, or exogenous dsDNA fragments, derived from bacterial or viral infection, are degraded by autophagy or released via small EVs. Therefore, these pathways function as defence mechanisms to prevent aberrant activation of innate immune responses. In senescent cells, irreparable DNA damage inhibits dsDNA degradation by autophagy and upregulates small EV secretion by activating the ceramide pathway. If neither pathway can function sufficiently, inflammasome activation occurs, leading to apoptosis. In summary, autophagy and the small EV pathway cooperatively play key roles to support cellular homeostasis.

Cell Culture
TIG-3 and RPE-1 cells were obtained from Japanese Cancer Research Resources Bank (JCRB, Osaka, Japan) and mouse primary fibroblasts (MEFs) were established from day 13.5 mouse embryos. Wild type (WT) and atg5-/-immortalized MEFs [48] were obtained from RIKEN Cell Bank (RCB2710 and RCB2711, respectively) (RIKEN BRC, Ibaraki, Japan). These cells were cultured in Dulbecco's Modified Eagle's (DME) medium (Nacalai Tesque, Kyoto, Japan) supplemented with 10% fetal bovine serum (FBS). THP-1 cells were obtained from Cell Resource Center for Biomedical Research, Tohoku University. They were cultured in RPMI 1620 medium (Nacalai Tesque) supplemented with 10% FBS. In this research, young control cells were less than 40 population doublings and replicative senescent cells were more than 70 population doublings. TIG-3 cells were infected with recombinant Figure 6. A model of our research. In healthy conditions, endogenous dsDNA fragments, derived from chromosomal or mitochondrial DNA, or exogenous dsDNA fragments, derived from bacterial or viral infection, are degraded by autophagy or released via small EVs. Therefore, these pathways function as defence mechanisms to prevent aberrant activation of innate immune responses. In senescent cells, irreparable DNA damage inhibits dsDNA degradation by autophagy and upregulates small EV secretion by activating the ceramide pathway. If neither pathway can function sufficiently, inflammasome activation occurs, leading to apoptosis. In summary, autophagy and the small EV pathway cooperatively play key roles to support cellular homeostasis.

Cell Proliferation Assay
Cells were plated on 35 mm dishes with 2 mm grids (Thermo Fisher Scientific, Waltham, MA, USA). The number of cells in each grid was counted every day, and the relative number of cells was calculated based on an adjusted cell number at day 1 set at 1.0 as described previously [7].

Apoptosis Assay
Apoptotic cells were judged by FITC-Annexin V staining using an apoptotic/healthy cells detection kit (PromoKine, Heidelberg, Germany) as described previously [24]. After an incubation at room temperature for 15 min, fluorescence signals were measured with a Wallac ARVO 1420 Multilabel counter (PerkinElmer Co., Ltd., Waltham, MA, USA).

Small Extracellular Vesicle Isolation from Cells
Small EVs were obtained from cell supernatants, as previously described with some modifications [24,27,69]. In brief, cells were incubated in DME medium with ultracentrifuged 5% FBS for 48 hours. The supernatants were collected and centrifuged at 300 g for 5 min and then at 2000× g for 10 min. The supernatant was then centrifuged at 10,000× g for 30 min, followed by filtration through a 0.2-µm pore filter (17597K, Sartorius, Gottingen, Lower Saxony, Germany). The collected supernatant was then subjected to preparation procedures by either ultracentrifugation at 100,000× g for 70 min as previously described [24,25] or an affinity-based method using MagCapture (Fujifilm Wako Chemicals, Tokyo, Japan) [70] for small EVs isolation. Nanoparticle tracking analysis were performed using a NanoSight LM10 system (Malvern Panalytical, Westborough, MA, USA). The amount of dsDNA in EVs was determined by quantitative Real-Time PCR using these primers: human LINE1, 5 -CAAACACCGCATATTCTCACTCA-3 (forward), and 5 -CTTCCTGTGTCCATGTGATCTCA-3 (reverse) [23,24].

Electron Microscopy
Small EVs isolated from RPE-1 cells were absorbed to formvar carbon coated nickel grids and immune-labelled with an anti-CD63 antibody (556019, BD Biosciences, NJ, USA), followed by 5 nM of a gold-labelled secondary antibody (British BioCell International Ltd., UK). The samples were fixed in 2% glutaraldehyde in 0.1 M phosphate buffer. The grids were placed in 2% glutaraldehyde in 0.1 M phosphate buffer and dried, then stained with 2% uranyl acetate for 15 min and a Lead stain solution (Sigma-Aldrich). The samples were observed with a transmission electron microscope (JEM-1400Plus, JEOL Ltd., Tokyo, Japan) at 80 kV. Digital images were obtained with a CCD camera (VELETA, Olympus Soft imaging solutions GmbH, Olympus, Tokyo, Japan) [24].

Plasmids
The epitope tagged cDNAs of SMS2 and nSMase2 was cloned into the pMarX-puro retrovirus vector. All cDNAs were sequenced on a Genetic Analyzer 3130 (Applied Biosystems, Waltham, MA, USA) using a BigDye Terminator v3.1 Cycle Sequencing Kit (Applied Biosystems).

Animal Experiments
CD1 (ICR) mice were purchased from Charles River Inc. (Wilmington, MA, USA). An N-SMase inhibitor, spiroepoxide (Santa Cruz), was intraperitoneally injected into 50-day-old male mice at 3.5 mg/kg every two day. Fourteen days later, mice treated with spiroepoxide were euthanized and the small intestine sections were subjected to small EV collection or immunofluorescence analysis as described previously [24]. All animal care was performed according to the protocols approved by the Committee for the Use and Care of Experimental Animals of the Japanese Foundation for Cancer Research (No.17-02-1, 31 5 2017).

Statistical Analysis
Statistical significance was determined using unpaired two-tailed Student's t-test by Excell (Microsoft, Redmond, WA, USA). p-values less than 0.05 were considered significant.

Conclusions
We demonstrated how DNA damage activates the ceramide pathway and leads to increase senescence-associated EV (SA-EV) secretion. The SA-EV pathway is crucial for cellular homeostasis to protect cells from excessive inflammatory responses.
Supplementary Materials: The following are available online at http://www.mdpi.com/1422-0067/21/10/3720/ s1, Figure S1: DNA damage enhanced small EV release from ATG5 knockout mouse embryonic fibroblasts (MEFs) compared with their release in wild-type MEFs. Figure S2: Activation of the autophagy pathway prevents inflammatory gene expression in senescent cells. Figure S3: Inhibiting the small EV pathway provokes inflammasome activation and apoptosis in Bacillus Calmette-Guérin (BCG)-infected cells. Figure S4: Inhibiting both autophagy and small EV biogenesis promotes apoptotic cell death following Bacillus Calmette-Guérin (BCG) infection. Acknowledgments: We thank Eiji Hara (Professor, Osaka University) for helpful discussion during the preparation of this manuscript; Noboru Mizushima (Professor, The University of Tokyo) for providing the WT and atg5-/-MEF cells; Yosuke Ishihara (Tokai-Denshi Inc., Sizuoka, Japan) for technical supports for TEM analysis.

Conflicts of Interest:
The authors declare no competing financial interests.

DDR
DNA damage response EV Extracellular vesicle NTA Nanoparticle tracking analysis SASP Senescence-associated secretory phenotype cGAS The cyclic GMP-AMP synthase STING Stimulator of interferon genes WT Wild-type dsDNA Double strand DNA