Tissue-infiltrating macrophages mediate an exosome-based metabolic reprogramming upon DNA damage

DNA damage and metabolic disorders are intimately linked with premature disease onset but the underlying mechanisms remain poorly understood. Here, we show that persistent DNA damage accumulation in tissue-infiltrating macrophages carrying an ERCC1-XPF DNA repair defect (Er1F/−) triggers Golgi dispersal, dilation of endoplasmic reticulum, autophagy and exosome biogenesis leading to the secretion of extracellular vesicles (EVs) in vivo and ex vivo. Macrophage-derived EVs accumulate in Er1F/− animal sera and are secreted in macrophage media after DNA damage. The Er1F/− EV cargo is taken up by recipient cells leading to an increase in insulin-independent glucose transporter levels, enhanced cellular glucose uptake, higher cellular oxygen consumption rate and greater tolerance to glucose challenge in mice. We find that high glucose in EV-targeted cells triggers pro-inflammatory stimuli via mTOR activation. This, in turn, establishes chronic inflammation and tissue pathology in mice with important ramifications for DNA repair-deficient, progeroid syndromes and aging.

T o counteract DNA damage, mammalian cells have evolved partially overlapping DNA repair systems to remove DNA lesions and restore their DNA back to its native form 1,2 . For helix-distorting damage, cells employ the nucleotide excision repair (NER) pathway [3][4][5] , a highly conserved mechanism that recognizes and removes helical distortions throughout the genome or selectively from the actively transcribed strand of genes 6,7 . In humans, defects in NER are causally linked to mutagenesis and cancer initiation as in the cancer-prone syndrome xeroderma pigmentosum (XP, complementation groups XP-A to XP-G) 8 or to developmental and neuronal abnormalities as seen in a heterogeneous group of progeroid syndromes, including the Cockayne syndrome (CS; affected genes: Csa, Csa), Trichothiodystrophy (TTD; affected genes: Xpb, Xpd) or the XPF-ERCC1 syndrome (XFE; affected genes: Ercc1, Xpf) [9][10][11] . The links between persistent DNA damage and the premature onset of age-related metabolic and endocrine perturbations in NER patients and accompanying mouse models 10,[12][13][14][15][16][17][18] are well established. We and others have recently shown that chronic inflammation 19,20 , genotoxic and oxidative stress [21][22][23] contribute significantly in NER progeria and age-related degenerative diseases, but the mechanisms remain unresolved 15,24,25 . XPF-ERCC1 is a highly conserved, heterodimeric, structure-specific endonuclease complex required for lesion excision in NER 26,27 that is thought to play an analogous role in the repair of DNA interstrand crosslinks (ICLs) that covalently link both DNA strands preventing transcription and replication 28 .
Using mice with an engineered ERCC1-XPF defect in tissueinfiltrating macrophages, we provide evidence for a fundamental mechanism by which irreparable DNA damage triggers an exosome-based, metabolic reprogramming that leads to chronic inflammation and tissue pathology in NER progeroid syndromes and likely also during aging.

Results
Cytoplasmic stress responses in Er1 F/− macrophages. To dissect the functional links between irreparable DNA damage and innate immune responses in vivo, we intercrossed animals homozygous for the floxed Ercc1 allele (Ercc1 F/F ) 29 with mice carrying the Lysozyme 2 (Lys2)-Cre transgene in an Ercc1 heterozygous background; Lys2 is a bacteriolytic enzyme that is primarily expressed in the monocyte-macrophage system 30 . Crossing the Lys2-Cre with Rosa YFP transgenic animals confirmed the specificity of Lys2-driven YFP expression to thioglycolate-elicited peritoneal macrophages (TEMs; Fig. 1a) but not to hepatocytes, the primary pancreatic cells (PPCs; Fig. 1b, c) or the pancreas and the white adipose tissue (WAT) that are infiltrated with MAC1possitve macrophages expressing YFP (Fig. 1d). Western blotting confirmed the excision of the floxed Ercc1 allele in Lys2-Ercc1 F/− (referred from now on as Er1 F/− ) peritoneal macrophages (Fig. 1e), neutrophils and monocytes but not in neurons (Supplementary Fig. 1A). Confocal microscopy revealed the absence of ERCC1 expression in Er1 F/− bone marrow-derived macrophages (BMDMs) and TEMs (Fig. 1f, Supplementary Fig. 1B) as well as in monocytes and neutrophils ( Supplementary Fig. 1B). Together, these findings indicate the normative ERCC1 expression levels in Er1 F/− tissues or cells other than the targeted cell populations i.e. monocytes, macrophages and neutrophils. Phosphorylated histone H2A.X (γ-H2A.X)-containing foci accumulate at sites of DNA breaks 31 . In line, the number of γ-H2A.X positive nuclei was significantly higher in the DNA repair-defective Er1 F/− BMDMs (Fig. 1g) and TEMs (Fig. 1h) compared to Lys-Ercc1 F/+ control cells (referred from now on as Er1 F/+ or wild-type; wt.). We also find marked differences in the number of positively stained nuclei for FANCI involved in the repair of DNA ICLs 32 , RAD51 involved in the repair of DSBs by homologous recombination (HR) 33 and phosphorylated ATM, a central mediator of the DNA damage response (DDR; Fig. 1i for BMDMs and Fig. 1j for TEMs). Staining with caspase 3 revealed few, if any, apoptotic cells in Er1 F/− TEMs (Fig. 1k). We also find no decrease in the total number of CD45(+) CD11b(+) macrophages during hematopoiesis (bone marrow) or at peripheral tissues i.e. blood, spleen and pancreas between Er1 F/− and Er1 F/− animals (Supplementary Fig. 1E-H). Staining with a lipophilic, biotin-linked Sudan Black B analogue (GL13; commercially available Sen-TraGor® 34 ) in BMDMs revealed a uniform, >four-fold accumulation of lipofuscin (Fig. 1l). Likewise, senescence-associated (SA)-β-gal assay and western blotting for Lamin B1 revealed an increase in β-galactosidase expression in Er1 F/− BMDMs (Supplementary Fig. 1C) and a mild but consistent senescenceassociated loss of Lamin B1 expression in Er1 F/− BMDMs, respectively ( Supplementary Fig. 1D). Confocal microscopy studies with the HSP chaperone GRP78 known to be required for ER integrity 35 , LC3β that is stably associated with autophagosomal membranes 36 , P62 a reporter of autophagic activity 37 and Gm130, a protein localized to the Golgi 38 revealed a dilated endoplasmic reticulum (ER; Fig. 2a), signs of autophagy with accumulation of autophagosomes ( Fig. 2a; as indicated) and a dispersed and fragmented Golgi in Er1 F/− macrophages (Fig. 2b). Western blotting further confirmed the increased GRP78, P62, and LC3 levels in Er1 F/− macrophages (Fig. 2c). We also performed western blotting (Supplementary Fig. 2A) and/or confocal studies ( Supplementary Fig. 2B) in Er1 F/+ BMDMs treated with nocodazole (known to trigger Golgi dispersal) 39 , chloroquine (known to inhibit the degradation of autophagosomes in lysosomes) 40 and tunicamycin (known to trigger ER stress) 41 further confirming the validity of GRP78, P62, LC3, and GM130 biomarkers used to detect the cytoplasmic stress responses seen in Er1 F/− macrophages ( Supplementary Fig. 2B). To test whether persistent DNA damage triggers similar cytoplasmic stress responses, we exposed wt. macrophages to mitomycin C (MMC), a potent inducer of DNA ICLs. As with Er1 F/− cells, treatment of wt. macrophages with MMC led to dilated ER lumen and autophagy ( Fig. 2d; as shown) and to a dispersed Golgi throughout the cytoplasm (Fig. 2e). Exposure of wt. macrophages to lipopolysaccharide (LPS), a potent activator of monocytes and macrophages partially mimicked the MMC-driven autophagy and Golgi dispersal ( Fig. 2f) but failed to instigate the dilation of ER. Inactivation of Ataxia-telangiectasia mutated (Atm) by exposing MMC-treated macrophages to KU-55933, an ATM inhibitor 42 or Ataxia telangiectasia and Rad3 related (ATR)/Cyclin dependent kinase inhibitor NU6027 (known to inhibit ATR kinase without interfering with irradiation-induced autophosphorylation of DNA-dependent protein kinase or ATM) reversed the Golgi dispersal after DNA damage (Fig. 2g). Likewise, inhibition of ATM and ATR dampened the increase in GRP78 and LC3 protein levels in MMC-treated macrophages (Fig. 2h) indicating that DNA damage is causal to ER dilation and autophagy; notably, the cytoplasmic alterations are reversible and require a functional DDR. Unlike in Er1 F/+ macrophages (Fig. 2i), transmissionelectron microscopy in Er1 F/− macrophages revealed the gradual accumulation of intracellular vesicles ( Fig. 2j; left panel) often organized into larger vacuolar structures ( Fig. 2j; right panel) and the appearance of cytoplasm-filled projections ( Fig. 2k; left panel) that formed a convoluted network of pseudopodia-like structures ( Fig. 2k; right panel) containing vesicles (Fig. 2l). These findings closely resemble the enhanced contractile activity observed in activated macrophages 43 .
The Er1 F/− defect causes systemic metabolic alterations. Er1 F/− mice maintained a lower body weight (Fig. 3a) and a normal breeding efficiency over a period of 22 weeks. To investigate whether tissue-infiltrating Er1 F/− macrophages trigger the onset of systemic metabolic complications in Er1 F/− mice, we performed a glucose tolerance test (GTT) in 2-months-old Er1 F/− and Er1 F/+ mice fed on a normal diet (ND) for a period of 2, 4 or 6 months. Animals were fasted for 16 h prior to administration of a single glucose dose (i.e. 200 mg/dL) and blood glucose levels were measured at consecutive time intervals for a period of 3 h. After 2 months on ND, we find no difference in GTT between Er1 F/− and Er1 F/+ animals ( Fig. 3b; as shown). However, we find a progressive difference in GTT between Er1 F/− and Er1 F/+ animals fed on ND for 4 or 6 months ( Fig. 3b; as indicated) that was further pronounced when animals were fed for 4 months on a Er1 F/+ Er1  high-fat diet ( Fig. 3c; as shown); unlike the Er1 F/− and Er1 F/+ animals fed on ND, Er1 F/− and Er1 F/+ animals fed on high-fat diet manifest no significant differences in body weight (Supplementary Fig. 2C), thereby, uncoupling Er1 F/− leanness from the noticeable differences in GTT and the mild, yet noticeable, hypoglycemia seen in 4-and 6-months-old Er1 F/− and Er1 F/+ animals (Fig. 3d). Further work revealed no difference in insulin serum levels ( Fig. 3e) or insulin sensitivity (Fig. 3f) in the 8months-old Er1 F/− animals and comparable levels of 2-Deoxy-Dglucose (2-DG) uptake in the liver and muscle protein extracts of overnight starved Er1 F/− and Er1 F/+ in response to insulin ( Supplementary Fig. 2D). In line with enhanced glucose tolerance seen in Er1 F/− animals (Fig. 3b), we find that the 2-DG uptake is significantly higher in the tissues of non-insulin treated Er1 F/− mice when compared to Er1 F/+ corresponding controls (Supplementary Fig. 2D). Together these findings indicate that the enhanced glucose tolerance seen in Er1 F/− mice is mediated through alternative, insulin-independent mechanisms. Further work revealed an accumulation of glycogen (Fig. 3g) and the significant increase in the mRNA levels of glycogenin-1 (Gyg-1) and glycogen synthase (Gys) genes involved in the initiation of glycogen synthesis as well as the decrease in the mRNA levels of liver glycogen phosphorylase (Pygl) and glycogen synthase kinase 3 (Gsk3) involved in glycogen breakdown and the regulation of glycogen synthesis in the 6-months-old Er1 F/− livers (Fig. 3h). We find a lower epididymal fat and comparable body size (nasoanal length) in age-matched Er1 F/− and Er1 F/+ animals with lower deposition of triglycerides in Er1 F/− livers compared to agematched Er1 F/+ control animals (Supplementary Fig. 3A; Fig. 3i) that remained substantially lower when Er1 F/− animals were maintained on a long-term high-fat diet ( Fig. 3i; as indicated). The latter is in agreement with previous observations supporting the notion that accumulation of glycogen in the liver reduces food intake and attenuates obesity in mice 44 . The lower body weight in Er1 F/− mice (Fig. 3a) and the recent finding that LyzM-Cre is expressed in the hypothalamus 45 known to regulate appetite 46 prompted us to test for changes in the daily food intake of Er1 F/− mice. We find no differences in the daily food consumption of Er1 F+ and Er1 F/− mice over a period of 14 days ( Supplementary  Fig. 3B). Moreover, western blotting and immunofluorescence studies revealed comparable ERCC1 protein levels and no detectable accumulation of DNA damage-associated γ-H2A.X foci in the 7-months-old Er1 F+ and Er1 F/− hypothalamic regions ( Supplementary Fig. 3C, D; as indicated). In support, the great majority of YFP signal colocalized with MAC1 in Rosa YFP transgenic animals expressing the Lys2-Cre transgene with detection of YFP signal in only a few e.g. 1-2 cells expressing the neuronal marker NeuN (Supplementary Fig. 3D).
Er1 F/− macrophages instigate a chronic inflammatory response. Histological examination in the 8-months-old Er1 F/− animals revealed the infiltration of monocytes and/or lymphocytes in perirenal fat, the kidney ( Fig. 4a and Supplementary  Fig. 3E), the lung and the liver ( Fig. 4b and Supplementary  Fig. 3F) and the presence of lipofuscin-accumulating macrophages in the white adipose tissue (Fig. 4c). Confocal studies revealed the infiltration of CD45+ hematopoietic cells and the accumulation of MAC1+ macrophages in the 8-months-old Er1 F/− livers (Fig. 4d), pancreata (Fig. 4e) and the white adipose tissue (Fig. 4f). These data and the previously documented higher frequencies of CD45 + CD11b + cells in all tissues tested (Supplementary Fig. 1E-H) along with the more differentiated/activated status of Er1 F/− monocytes/macrophages (as evidenced by their higher volume; Supplementary Fig. 3G) further support the presence of ongoing systemic inflammation. Flow cytometry (FACS) in the 8-months-old Er1 F/− and Er1 F/+ animals and in Er1 F/+ animals fed on a high fat diet (Er1 F/+ HFD) (Supplementary Fig. 4A) revealed a comparable increase in the infiltration of F4/80 (+) cells in the epididymal fat of Er1 F/− and Er1 F/+ HFD mice (HFD; Supplementary Fig. 4A); the latter animal group also developed insulin resistance as evidenced by the ITT test ( Supplementary Fig. 4C). Unlike in Er1 F/+ HFD animals, FACS analysis with markers CD11c for M1 47 and CD206 for M2 48 revealed a potent increase in the M2 population of macrophages in Er1 F/− animals ( Supplementary Fig. 4D). This finding indicates that the ERCC1 defect triggers the accumulation of macrophages that are polarized in a direction distinct to that seen in animal models associated with HFD and insulin resistance. In parallel, we find the increased expression of PECAM-1 involved in leukocyte transmigration and angiogenesis 49 , ICAM-1 known to play a role in inflammatory and immune responses 50 and VCAM-1 involved in the adhesion of monocytes to vascular endothelium 51 (Fig. 4d-f). RNA-Seq profiling of Er1 F/− and Er1 F/+ macrophages revealed 1756 differentially expressed genes [meta-FDR ≤ 0.005, fold change ≥ ±1.5, 1090 upregulated genes; 665 downregulated genes; Fig. 4g; Supplementary data 1]. Within this gene set, the gene ontology (GO)-classified biological processes associated with innate immune or GTPase-mediated signaling, membranebound projection and pathways associated with endosomal/ vascular and ER-phagosome interactions and senescence in Er1 F/ − macrophages have a significantly disproportionate number of responsive genes relative to those mapped in the murine genome (false detection rate ≤0.05; Fig. 4h-j). Quantitative PCR for a set of pro-inflammatory cytokine and chemokine genes (Fig. 4k) and our previous findings (Figs. 2a-i and 1l, Supplementary Fig. 1C, D), confirmed the validity of gene expression changes. n > 500 cells counted per genotype). The numbers indicate the average percentage of GFP (+) cells ± SEM), b Lys-Cre-driven Rosa-YFP expression in hepatocytes and primary pancreatic cells (PPCs) shown by confocal microscopy (n > 100 cells counted per genotype; the numbers indicate the average percentage of GFP (+) cells ± SEM) and c western blotting. d Immunofluorescence staining of Lys-Cre-driven Rosa-YFP expression in the Er1 F/− pancreas and the white adipose tissue (WAT) that are infiltrated with MAC1-positive macrophages (indicated by the arrowheads). e Western blotting of ERCC1 protein in whole-cell (w) cytoplasmic (c) and nuclear (n) extracts. Tubulin (TUB), and Fibrillarin (FIB) were used as loading controls (as indicated). The graph represents the fold change (F.C.) of ERCC1 protein levels in Er1 F/− samples compared to corresponding Er1 F/+ controls (n = 3). f Cell type-specific ablation of ERCC1 (indicated by the arrowhead) in bone marrow-derived (BMDMs) and TEMs expressing the macrophage-specific antigen MAC1. The numbers indicate the average percentage of ERCC1 (+) nuclei ± SEM in Er1 F/+ and Er1 F/− BMDMs and TEMs (n > 150 cells were counted per genotype). g Immunofluorescence detection of γ-H2AX in Er1 F/− and Er1 F/+ BMDMs and h TEMs. i Immunofluorescence detection of FANCI, pATM and RAD51 in Er1 F/− and Er1 F/+ BMDMs (in each case n > 200 cells were counted per genotype). j Immunofluorescence detection of pATM in Er1 F/− and Er1 F+ TEMs (n > 150 cells were counted per genotype). k Immunofluorescence detection of Caspase 3 (CASP3) (n > 300 cells were counted per genotype) and l GL13 (indicated by the arrowhead), commercially available SenTraGor®, in Er1 F/− and Er1 F+ BMDMs. Fluorescence intensity was calculated in n > 50 cells per genotype. Gray line is set at 5 μm scale, unless otherwise indicated. Error bars indicate S.E.M. among replicates (n ≥ 3). Asterisk indicates the significance set at p-value: *≤0.05, **≤0.01 (two-tailed Student's t-test).
DNA damage triggers the release of extracellular vesicles. The increased tolerance of Er1 F/− animals to glucose challenge and the activation of innate immune signaling in peripheral tissues prompted us to investigate whether Er1 F/− macrophages acquire a secretory phenotype that exerts systemic, metabolic and proinflammatory stimuli in Er1 F/− animals. To do so, we first employed a high-throughput mass spectrometry approach in isolated Er1 F/− and Er1 F/+ macrophage media (Fig. 5a). This strategy led to the identification of 329 proteins with 211 proteins (64%) being shared by two independent measurements under stringent selection criteria ( Fig. 5b and Supplementary data 2 and Methods). At the confidence interval used (p-value: 5 × 10 −11 ), we find that proteins associated with the presence of (membranebound) extracellular vesicles (EVs) and membrane-bound   To validate the in vivo relevance of these findings, we isolated intact EVs from Er1 F/− and Er1 F/+ sera. Western blot analysis confirmed the enrichment of vesicle-associated protein marker CD9, that is involved in the biogenesis, targeting and function of EVs 61 and ALIX known to be associated with the endosomal sorting complex required for transport 62 along with two members of the Ras superfamily of GTPases i.e. RAB10 and RAC1 (Fig. 5f, Supplementary Fig. 5A). Electron microscopy revealed the presence of EVs in the media of Er1 F/− macrophages with a size ranging between 30 and 80 nm; the latter corresponds to that known for exosomes (40-100 nm diameter) 62 , although the presence of larger (>200 nm) EVs cannot be excluded 63 (Fig. 5g). To test whether DNA damage is the primary instigator of EV secretion in Er1 F/− macrophages, we exposed Er1 F/+ macrophages to the genotoxin MMC. In line with Er1 F/sera or the EV fraction of Er1 F/− macrophage media ( Fig. 5f-h, Supplementary Fig. 5A), we find a substantial enrichment for EV-associated protein markers CD9 and Alix and an accumulation of Ras GTPases RAB10, RAC2 and RAC1 in the EV fraction isolated from the media of MMC-treated cells compared to those of untreated cells (Fig. 5i). Further work revealed a similar accumulation of CD9 and Alix in the EV fraction of Er1 F/− monocytes and neutrophils and a mild to negligible accumulation in the EV fraction of Ercc1 −/− (designated from now on as Er1 −/− ) primary mouse embryonic fibroblasts (MEFs) or the adipocytes, respectively ( Supplementary Fig. 5B). Inactivation of DDR by inhibiting ATM or ATR in MMC-treated macrophages dampened substantially the increase in CD9, ALIX and RAB10, RAC2 and RAC1 protein levels ( Fig. 5i and Supplementary Fig. 5C). The DNA damage-associated EV secretion is further supported by the parallel increase of γ-H2A.X, CD9 and Alix protein levels in BMDM whole-cell extracts or in the EV fraction of Er1 +/+ , Er1 +/− , and Er1 −/− BMDM culture media ( Supplementary Fig. 5D). As detection of low abundance proteins (<100 ng/ml) is typically challenging with current mass spectrometry protocols, we also employed an ELISA-based immunoassay to quantify Interleukin (IL)-1, IL6, IL-8, Interferon-γ, monocyte chemotactic protein 1, and stromal-derived-factor 1 in animal sera and macrophage media. We find substantially higher IL6 and IL8 levels in Er1 F/− sera and macrophage media compared to Er1 F/+ control samples (Fig. 5j); all other cytokines were either non-detectable or did not vary significantly among the sample groups tested. Primary pancreatic cells (PPCs) or the hepatocytes are relevant cell types for coupling growth stimuli with fine-tuning mechanisms involved in nutrientsensing and glucose homeostasis. Confocal studies confirmed the accumulation of RAC1, RAB10, RAC2, and RHOA in the cytoplasm of Er1 F/− PPCs and hepatocytes (Fig. 5k, l and Supplementary Fig. 5E). In agreement with our previous findings on macrophage media (Fig. 5h), we also find that TEMs accumulate the exosome marker CD9 along with RAC1, RAB10 and RAC2 in their cell membrane (Fig. 5m); unlike other Ras GTPases tested, we find that RAC2 also accumulates in the nuclei of Er1 F/− hepatocytes and TEMs (Fig. 5l, m; as shown). To test the functional relevance of RAB10 and RAC1 in the biogenesis/secretion of EVs in Er1 F/− macrophages, we transfected Er1 F/+ macrophages with GFP-tagged RAB10 and RAC1. In line with the pronounced accumulation of CD9 and Alix in the EV fraction Er1 F/− macrophages, we detect the marked accumulation of EV-associated protein markers CD9 and Alix in GFP-tagged RAB10/RAC1-transfected macrophages (Supplementary Fig. 6A). In parallel we treated Er1 F/− macrophages with the RAC1 inhibitor NSC 23766 for 16 h 64 . We find that the inhibition of RAC1 activity in Er1 F/− macrophages leads to a decrease in EV secretion, as evidenced by the decrease in CD9 protein levels ( Supplementary Fig. 6B). Next, we performed live confocal imaging in Er1 F/− and Er1 F/+ macrophages transiently expressing CD9-GFP. Unlike the sedentary appearance of Er1 Er1 F/− EVs promote the glucose uptake in recipient cells. To test that Er1 F/− macrophage-derived EVs (Er1 F/− EVs) are delivered successfully to recipient cells, EVs were labeled with a Fig. 2 Abrogation of ERCC1 triggers cytoplasmic stress responses in Er1 F/− macrophages. a Immunofluorescence detection of GRP78 (n > 250 cells counted per genotype) marking the dilation of ER, LC3 (n > 40 cells counted per genotype) for autophagy, P62 for autophagic activity (n > 15 optical fields per genotype,~150 cells per field) and b Gm130 for Golgi dispersal (~500 cells per genotype; arrowhead) in Er1 F/− and Er1 F/+ BMDMs. For GRP78 and LC3, the colored numbers indicate the average percentage of (+) stained cells ± SEM for the indicated, color-matched protein. For p62, the colored numbers indicate the average mean fluorescence intensity ± SEM of p62 signal. For Gm130, the green-colored numbers indicate the average percentage of (+) stained cells ± SEM showing Golgi dispersal. c Western blot levels of GRP78, P62, LC3, and Tubulin in Er1 F/− and Er1 F+ BMDMs. The graph shows the fold change of indicated protein levels in Er1 F/− BMDMs compared to Er1 F+ corresponding controls (n = 3 per group). d Immunofluorescence detection of GRP78 (for ER stress) (n > 500 cells counted per genotype), LC3 (for autophagy) (n > 750 cells counted per genotype) and e Gm130 (for Golgi dispersal) in MMC-treated BMDMs (n > 500 cells counted per genotype). Numbers indicate the average percentage of (+) stained cells ± SEM for the indicated, color-matched protein. For Gm130, the green-colored numbers indicate the average percentage of (+) stained cells ± SEM showing Golgi dispersal. f Immunofluorescence detection of LC3 (for autophagy) (~250 cells counted per genotype) and Gm130 (for Golgi dispersal) (~950 cells counted per genotype) in LPS-stimulated BMDMs. g Immunofluorescence detection of Gm130 (for Golgi dispersal) in MMC-treated and control BMDMs exposed to ATM (ATMi) (~200 cells counted) or ATR (ATRi) (500 cells counted) inhibitor (as indicated). The green-colored numbers indicate the average percentage of (+) stained cells ± SEM showing Golgi dispersal. h Western blot levels of GRP78 and LC3 in MMC-treated and control (ctrl) macrophages exposed to ATM (ATMi) or ATR (ATRi) inhibitor (as indicated; Tubl.: tubulin, unt: untreated). The graph represents the fold change in indicated protein levels in MMC-treated macrophages exposed to ATM (ATMi) or ATR (ATRi) inhibitor compared to corresponding controls (n = 4 per group) (i-j). Representative transmission electron micrographs of Er1 F+ (i) and  lipophilic green fluorescent dye PKH67 and were injected in the media of PPCs. Using this approach, we detect the presence of green-fluorescent, Er1 F/− EVs in the cytoplasm of targeted PPCs (Fig. 6a). To further confirm that macrophage-derived Er1 F/− EVs also deliver their cargo to recipient PPCs, we transfected Er1 F/− macrophages with GFP-tagged RAB10 or RAC1 ( Supplementary Fig. 7A). Following the exposure of PPCs to Er1 F/− EVs carrying the GFP-tagged RAB10 and RAC1, we were able to detect both proteins in the cytoplasm of EV-recipient PPCs (Fig. 6b,    to cell membrane [65][66][67][68] , we reasoned that the release of Er1 F/− EVs in animal sera and the macrophage media promotes the glucose uptake in EV-recipient cells. Indeed, treatment of PPCs or hepatocytes with Er1 F/− EVs triggers the noticeable uptake of 2-NBDG, a fluorescent tracer used for monitoring glucose uptake into living cells (Fig. 6d). We find similar findings when PPCs are exposed to EVs derived from Er1 F/− sera (ser. Er1 F/− EVs; Fig. 6e) or from Er1 F/+ macrophages exposed to the genotoxic agent MMC (MMC EVs; Fig. 6f).
Because the plasma membrane is impermeable to large polar molecules, such as glucose, the cellular uptake of glucose is accomplished through specific transmembrane transporters. Glucose transporters GLUT1 and GLUT3 are expressed in most tissues and, unlike other family members, they do not rely on insulin for facilitated diffusion of glucose across cell membranes 69,70 . Although GLUT1 was undetectable 3 h after the exposure of wt. PPCs to Er1  7G) and is detectable with GLUT1 for at least 24 h ( Fig. 6i and Supplementary Fig. 7H; as indicated). Using a previously established method 72 , we also confirm a mild but reproducible increase in oxygen consumption rate (OCR) in PPCs treated with MMC-Er1 F/− -or serum-derived Er1 F/− EVs (32%, 74%, and 26%, respectively) compared to cells exposed only to EVs derived from untreated cells or Er1 F/+ sera (Fig. 6j). Western blotting ( Fig. 6k and Supplementary Fig. 8A) and immunofluorescence studies further confirmed the accumulation of GLUT 1 and 3 in the 6-months-old Er1 F/− pancreata (Fig. 6l) and livers ( Supplementary Fig. 8B). Staining for PECAM-1 and GLUT3 in Er1 F/− tissues revealed that, for at least GLUT3, its expression is not restricted to endothelial cells but extends to other cell types in the Er1 F/− livers, the white adipose tissue and the pancreata (Supplementary Fig. 8C).
Glucose uptake activates innate immune signaling. High-glycemic index diets have been associated with the activation of inflammatory processes 73 . These data and the marked inflammation seen in Er1 F/− tissues (Fig. 4a-f, Supplementary  Figs. 1E-H, 3G and 4A-D) led us to investigate whether the EVmediated increase in glucose uptake triggers similar proinflammatory responses. To test this, we exposed PPCs to i. MMC-treated or Er1 F/− macrophage media (the media are rich in IL6 and IL8) (Fig. 5j) that were devoid of EVs or to ii. Er1 F/− or MMC EVs alone (that were devoid of the macrophage media) or iii. to MMC-treated or Er1 F/− macrophage media (the media are rich in IL6 and IL8) containing the secreted EVs. We find that the pro-inflammatory iNOS factor accumulates in PPCs exposed to MMC-treated or Er1 F/− macrophage media containing the EVs (Fig. 7a, b). Interestingly, neither the Er1 F/− EVs nor the MMC EVs alone or the media that were devoid of EVs could trigger a similar effect in iNOS accumulation ( Fig. 7a and Supplementary  Fig. 8D). Likewise, NF-κB that translocates to the nucleus to induce the transcription of pro-inflammatory genes 74,75 accumulates in the nuclei of PPCs exposed to macrophage media containing the MMC (Fig. 7c) or Er1 F/− (Supplementary Fig. 8E) EVs; for NF-κB, exposure of PPCs to the media or the EVs alone also triggered a noticeable, albeit to a lesser extent, nuclear translocation ( Supplementary Fig. 8E). These data suggest that EV-mediated glucose uptake activates iNOS or NF-κB signaling in cells previously primed with pro-inflammatory signals (in this case, IL6 and IL8). To test whether high glucose levels could potentiate the response, PPCs already exposed to MMCtreated or Er1 F/_ macrophage media containing the EVs were cultured under low (5 mmol) or high (15 mmol) glucose concentration. We find that iNOS accumulation and NF-κB nuclear translocation is more profound in PPCs maintained at the highest (15 mmol) glucose concentration (Fig. 7e, Supplementary  Fig. 9A-C). High glucose levels are known to activate the PI3K/ AKT/mTOR signaling pathway that couples cellular activation to environmental cues 76 . In line, we find a substantial increase in the protein levels of phosphorylated Eukaryotic translation initiation factor 4E (eIF4E)-binding protein 1 (P4E-BP1), a translation repressor protein and a well-known target of rapamycin (mTOR) signaling pathway, in Er1 F/− pancreata (Fig. 7f). Inhibition of mTOR by exposing MMC-treated macrophages to rapamycin abrogated the increase in P4E-BP1 protein levels and iNOS accumulation (Fig. 7g). Likewise, rapamycin revokes the nuclear translocation of NF-κB in PPCs treated with Er1 F/− EVs but does not affect glucose uptake (Fig. 7h).
Exogenous delivery of Er1 F/− exosomes triggers inflammation. Having established that macrophage-derived Er1 F/− EVs promote the glucose uptake in recipient cells and the activation of innate immune responses, we sought to test the specificity and in vivo relevance of our findings. To do so, B6 C57BL/6 animals were injected intravenously every 24 h for a period of 10 days with EVs derived from either Er1 −/− or wt. MEFs and Er1 F/+ or Er1 F/− macrophages. Unlike in animals treated with Er1 −/− or wt. MEFderived EVs and with Er1 F/+ macrophage-derived EVs, intravenous injection of animals with Er1 F/− macrophage-derived EVs, leads to the pronounced accumulation of GLUT1 and the Fig. 4 Systemic inflammation and gene expression changes in Er1 F/− mice. a Infiltration of foamy cells (macrophages) in a region of lipogranuloma in Er1 F/− perirenal fat (as indicated) and inflammatory infiltration of lymphocytes and monocytes in Er1 F/− kidneys. b Inflammatory infiltration of lymphocytes and monocytes in Er1 F/− lungs and livers. See also Supplementary Fig. 3E, F for magnified inlays. The graph indicates the percentage of inflammatory (inflam.) infiltrates per tissue area in Er1 F/− (n = 9) and Er1 F/+ mice (n = 7). c Detection of GL13 (+) macrophages in perirenal fat of Er1 F/− (n = 5) and Er1 F/+ (n = 7) mice. d Immunofluorescence detection of PECAM-1, ICAM-1 and VCAM-1 along with CD45 and MAC1 (shown by the arrows) in the liver (n = 4; 4-6 optical fields per animal). e pancreas (n = 3; 3 optical fields per animal) and f the white adipose tissue (WAT) (n = 3; two optical fields per animal) of Er1 F/− and Er1 F+ mice indicating the expression of cell adhesion molecules and the presence of monocytic/lymphocytic infiltrates in Er1 F/− tissues. Colored numbers indicate the average mean fluorescence intensity ± SEM for the color-matched protein (as indicated). g Heat-map representation of significant gene expression changes (n = 1756 genes) in Er1 F/− BMDMs compared to corresponding control cells. h Over-represented GO biological processes and i pathways (Reactome) of Er1 F/− BMDMs compared to corresponding control cells; p: −log of p-value which is calculated by Fisher's exact test right-tailed, R: ratio of number of genes in the indicated pathway divided by the total number of genes that make up that pathway. j Heat-map representation of gene expression changes associated with significantly over-represented biological processes in Er1 F/− BMDMs compared to corresponding control cells (as indicated). k Interleukin and chemokine mRNA levels in Er1 F/− compared to Er1 recruitment of macrophages in the liver of these animals (Fig. 8a,  b). Western blotting further confirmed the accumulation of GLUT1 in the liver and muscle protein extracts of animals injected with Er1 F/− macrophage-derived EVs but not with Er1 −/− and wt. MEF-derived EVs or with Er1 F/+ macrophagederived EVs (Fig. 8c, d). Importantly, mice injected with Er1 F/− macrophage-derived EVs also manifest an enhanced glucose tolerance in GTT compared to corresponding control animals.
No difference in GTT is observed between animals treated with Er1 −/− or wt. MEF-derived EVs (Fig. 8e; as indicated). In agreement, wt. PPCs treated with Er1 F/+ or Er1 F/− EVs derived from either macrophages or monocytes and with EVs derived from Er1 −/− or wt. adipocytes or MEFs leads to the accumulation of GLUT1 and iNOS (Supplementary Fig. 10A) and to the 2-NBDG uptake (Supplementary Fig. 10B) only in recipient cells exposed to Er1 F/− macrophage-derived EVs.

Discussion
How distinct cell types adapt their metabolic demands to counteract deleterious threats remains an intriguing question arguing for tissue-specific responses against irreparable DNA lesions. Our findings provide evidence that persistent DNA damage in circulating Er1 F/− macrophages triggers the release of EVs that gradually surmount a systemic, glucose-based metabolic reprogramming leading to chronic inflammation in mice. The lack of apoptosis in Er1 F/− macrophages indicates that the dilation of ER, Golgi dispersal and autophagy in these cells are not a byproduct of cell death. Instead, these processes occur quickly after DNA damage and are reversed when DDR is inhibited. Er1 F/− macrophages form cytoplasmic projections filled in with EVs which are released in Er1 F/− animal sera and the macrophage media and are rapidly secreted upon exposure of macrophages to DNA damage. The Er1 F/− EV cargo is enriched with Ras GTPases involved in glucose transporter trafficking and vesicle-mediated transport 77 that accumulate in Er1 F/− macrophages, PPCs, the hepatocytes and in wt. PPCs treated with Er1 F/− or MMC EVs. Inhibition of ATM or ATR abrogates the release of MMC EVs highlighting the causal contribution of DNA damage signaling in this response. It remains to be seen whether the accumulation of lipofuscin and β-galactosidase along with the increase in the mRNA levels of p16INK4A, p21CIP1 and protein levels of IL-6 and IL-8 in Er1 F/− macrophages reflect a recently proposed, reversible response of macrophages to physiological stimuli 78 or the premature onset of replicative or stress-induced senescence 79 .
Er1 F/− mice present with systemic inflammation and are hyper-tolerant to glucose challenge, independently, of insulin signaling. Τhe response is progressive; it develops over the course of several months and is exacerbated when animals are fed on a long-term high-fat diet. We find that Er1 F/− EVs tagged with the lipophilic green fluorescent dye PKH67 are successfully delivered to recipient PPCs and that macrophage-derived Er1 F/− EVs carrying GFP-tagged RAB10 or RAC1 are detected in the cytoplasm of EV-recipient PPCs. Importantly, exposure of PPCs to Er1 F/− -, MMC-or serum-derived Er1 F/− EVs leads to the noticeable uptake of glucose tracer 2-NBDG, the increase in GLUT1 protein and mRNA levels, the accumulation of RAB10 in the cytoplasm of PPCs and to higher OCR. Conversely, GLUT1 and GLUT3 known to facilitate cellular glucose uptake in an insulin-independent manner 80 , accumulate in the Er1 F/− livers and pancreata.
Hyperglycemia is acknowledged as a pro-inflammatory condition and lower glucose values are considered to be antiinflammatory [81][82][83] indicating that the EV-mediated glucose uptake is causal to the inflammatory signals seen in Er1 F/− tissues. However, exposure to high glucose concentrations does not cause inflammation unless cells are primed with an inflammatory stimulus 84,85 . Indeed, unlike in PPCs treated with only the IL6-and IL8-rich macrophage media or the EVs alone, exposure of PPCs to these media also containing the Er1 F/− or MMC-EVs led to iNOS accumulation and to the nuclear translocation of NF-κB in EV-recipient cells; importantly, the response is more profound when cells are maintained at the highest glucose concentration. We find that the increase in glucose uptake activates mTOR signaling known to coordinate cell metabolism with environmental inputs, including nutrients 86 . In line, rapamycin, a potent mTOR inhibitor 87 abrogates 4E-BP1 phosphorylation, iNOS accumulation and the NF-κB nuclear translocation but has no effect on glucose uptake itself in PPCs, indicating that mTOR acts downstream of the glucose signaling pathway to activate proinflammatory responses. Recently, EVs have been used as natural nanocarriers for the systemic, in vivo delivery of biologically active cargo to recipient cells 88 . Importantly, we find that chronic exposure of animals to Er1 F/− macrophage-derived EVs alone leads to GLUT1 accumulation, the recruitment of macrophages in parenchymal tissues and to enhanced glucose tolerance in wt. animals.
Previous data on the impact of defective NER in mammalian physiology 9,10,14,15,24,25 and the emerging role of exosomes in cell homeostasis [89][90][91][92] , supports the notion that the DNA damagedriven metabolic adaptation represents a physiologic response that is both beneficial and detrimental for organismal survival. In the short run, activated macrophages could temporarily alert EVrecipient cells to rebuild their glucose reservoirs in order to defend themselves against foreign pathogens or threats. In the long run, however, the slow but steady buildup of irreparable DNA lesions e.g. in macrophages is expected to intensify EV secretion leading to the gradual onset of an exosome-based, metabolic reprogramming and pro-inflammatory signaling in mice. The latter would perpetuate a vicious cycle of persistent DDR signaling, increased glucose uptake and activated innate immune responses leading to the premature onset of chronic inflammation and tissue malfunction in NER progeroid syndromes (Fig. 8f). As DNA damage accumulates over time, a low glycemic diet may, therefore, be promising to delay age-related diseases 24,93 .

Methods
Animal studies. Ercc1 F/F mice containing a floxed allele of the Ercc1 gene and Rosa26-YFP st/st mice were crossed with (Lys2)-Cre transgenic mice to obtain inactivation of the Ercc1 gene or expression of YFP in tissue-infiltrating macrophages, respectively. For insulin tolerance test (ITT), animals were fasted for 6 h Fig. 5 DNA damage promotes the generation and secretion of extracellular vesicles (EVs) in Er1 F/− macrophages. a Schematic representation of the high-throughput MS analysis in Er1 F/− compared to Er1 F/+ BMDMs media. b Venn's diagram of proteins identified in Er1 F/− media from two independent biological replicates. c List of significantly over-represented GO terms associated with Cellular Component. d Number of observed (obs.) and expected (exp.) known protein interactions within the core 211 shared proteins set. e Schematic representation of the major protein complex identified in BMDM media. f Western blot analysis of CD9, ALIX, RAB10, and RAC1 proteins levels in the EV fraction of Er1 F/− and Er1 F/+ sera (n = 6; see also Supplementary Fig. 5A; left panel). g Transmission electron microscopy of EVs marking the presence of exosomes with a size 30-80 nm in Er1 F/− TEM media. h Western blot analysis of CD9, ALIX, RAB10, RAC2, and RAC1 proteins levels in Er1 F/− compared to Er1 F/+ EV fraction of BMDM media (n = 5). A graph showing the fold change and statistical significance of the indicated protein levels is shown in Supplementary Fig. 5A; right panel. i Western blot analysis of CD9, ALIX, RAB10, RAC2, and RAC1 proteins levels in the EV fraction of media derived from the MMC-treated and control BMDMs exposed to ATM (ATMi) or ATR (ATRi) inhibitors (as indicated; n = 3). A graph showing the fold change and statistical significance of the indicated protein levels is shown in Supplementary Fig. 5C. j IL8 and IL6 protein levels in Er1 F/− and Er1 F/+ sera and BMDM media (as indicated). k Immunofluorescence detection of RAC1 (~500 cells per genotype), RAB10 (~150 cells per genotype) and RAC2 (~150 cells per genotype) in Er1 F/− and Er1 F/+ PPCs (n > 400 cells per genotype) (see also Supplementary Fig. 5e for RHOA), l hepatocytes (n > 100 cells per genotype) and m thioglycolate-elicited macrophages (TEMs) (n > 500 cells per genotype). and were injected intraperitoneally with 0.75 Units/kg of body weight insulin (Humulin, Ely Lili). For glucose tolerance tests (GTT), mice were fasted for 16 h. and subsequently were injected intraperitoneally with 1 mg/gr of body weight 35% dextrose solution. Blood glucose levels were measured using CONTOUR® meter, at the indicated time points. To determine steady state glucose levels, 4-and 6months-old animals were fasted for 2 h and glucose was determined. Serum insulin and triglyceride levels were measured with specialized kits (ALPCO and LabAssay triglyceride, Wako Chemicals, respectively). Mice were maintained in grouped cages in a temperature-controlled virus-free facility on a 12-h light/dark cycle and fed either a high-fat diet (60% energy from fat, 20.3% carbohydrate, and 18.41% protein, 58Y1-58126, TestDiet) or a normal diet (Lactamin, Stockholm, Sweden). Mice had access to water ad libitum. Body weight was measured weekly. For food intake experiments, mice were kept individually in separate cages. Defined food quantity was added daily in each cage and food consumption was measured every 24 h. This work received ethical approval by and independent Animal Ethical Committee at the IMBB-FORTH. All relevant ethical guidelines for the work with animals were adhered to during this study. For the in vivo 2-DG uptake assay 6months old male mice were used (three in each group).  MEFs and Er1 F/+ and Er1 F/− BMDMs) as described in the section EV isolation. On day 10 mice were starved O/N and GTT was performed as described above. Mice were sacrificed and liver and muscle tissues were isolated and analyzed with Western blotting and Immunofluorescence. Electron microscopy. For transmission electron microscopy, primary macrophages were washed in PBS followed by fixation in 2% paraformaldehyde, 2% glutaraldehyde in 0.1 M sodium cacodylate buffer (pH 7.42) with 0.1% magnesium chloride and 0.05% calcium chloride. After washes with sodium cacodylate buffer, cells were fixed in 1% osmium tetroxide in sodium cacodylate buffer and samples were dehydrated in ethanol gradient. Samples were then treated with propylene oxide and embedded in Epon/Araldite resin mix. Ultrathin-sections (50-100 nm) were taken on a Leica LKB2088 ultramicrotome and were examined under JEM 100 C/JEOL/Japan Transmission Electron Microscope. Microphotographs were obtained with an ES500W Erlangshen camera and processed with the Digital Micrograph software (Gatan, Germany). For electron microscopy (EM) analysis of whole-mount exosome preparations, fixed EVs were deposited on EM grids and were further fixed with glutaraldehyde. Samples were first contrasted in a solution of uranyl oxalate and then contrasted and embedded in a mixture of 4% uranyl acetate and 2% methyl cellulose.
EV isolation and labelling. Exosomes were purified using the differential ultracentrifugation protocol 94 . Briefly culture medium was centrifuged sequentially at 300 g, (10 min), 2000 g (10 min), and 10000 g (30 min) to remove dead cells and cell debris. Extracellular vesicles were purified with the final step of ultracentrifugation at 100000 g for 2 h. For functional experiments, macrophage-derived EVs were purified five times the number of recipient cells. For PKH67 staining, EVs were incubated with PKH67 (500 mL 0.2 mM) for 5 min at room temperature. Labelled EVs were diluted in 500 mL 1% BSA, and then pelleted at 100,000 g, washed with 1 mL PBS to remove excess dye, re-suspended in 1 mL PBS and then pelleted at 100,000 g before final re-suspension. The same protocol was followed in the absence of EVs (dye alone) to ensure lack of fluorescence. For functional experiments, macrophage-derived EVs were purified five times the number of recipient cells and then added in the medium of PPCs or hepatocytes for 3 or 24 h.
Histology, Immunofluorescence and Immunoblot analysis. Er1 F/+ and Er1 F/+ livers were OCT-embedded, cryosectioned, fixed in 10% formalin, stained with oil red O or Periodic acid-Schiff (PAS), counterstained with Harris's hematoxylin and visualized with DAB chromogen (Sigma). For histological analysis of Er1 F/+ and Er1 F/− tissues, samples were fixed in 4% formaldehyde, paraffin embedded, sectioned and stained with Harris's Hematoxylin and Sudan Black B (SBB)-Analogue (GL13). For immunofluorescence experiments of mouse tissues, minced tissues (liver, pancreas and hypothalamus), BMDMs, TEMs, PPCs and primary hepatocytes were fixed in 4% formaldehyde, permeabilized with 0.5% Triton-X and blocked with 3% normal calf serum and 1% BSA. After overnight incubation with primary antibodies, secondary fluorescent antibodies were added and DAPI was used for nuclear counterstaining. Samples were imaged with SP8 confocal microscope (Leica). For SDS-page analysis, EVs were derived from 20-25 × 106 macrophages, whole-cell extracts and nuclear or cytoplasmic extracts were used. Rapid time-lapse imaging was performed using SP8 confocal microscope (Leica) with a stage-top incubation system creating a 37°C environment. Z-stacks of 3-4 sections with two-channel detection were acquired every. For live cell experiments, in vitro differentiated Er1 F/+ and Er1 F/− BMDMs were transfected with 5 μg of total plasmid DNA (pEGFP-CD9, see below). Transfected cells were seeded on a Mattek culture dish and imaged 24 h post transfection. Time-lapse stacks were analyzed with Fiji. At least 20 cells from each genotype were recorded in total in three independent replicates.
Primary cell cultures, transfection and cell assays. Er1 F/+ and Er1 F/− BMDMs were differentiated from bone marrow precursors. Briefly, bone marrow cells were isolated from mouse femurs and tibias and cultured for 7 days in DMEM containing 10% FBS, 30% L929 conditioned media, 50 μg/ml streptomycin, 50 U/ml penicillin (Sigma) and 2 mM L glutamine (Gibco). For TEM isolation, mice were injected intraperitoneally with 2 ml 4% thioglycolate medium (Brewer). Three days later, mice were sacrificed and peritoneal cells were collected isolated by peritoneal lavage with 10 ml dMEM per mouse. Primary hepatocytes and PPCs were obtained from 20-days old mice. Briefly, liver and pancreas were excised, minced and incubated in 2 mg/ml collagenase type IV at 37°C for 15 min. After centrifugation, cells were resuspended in DMEM containing 10% FBS, 50 μg/ml streptomycin, 50 U/ml penicillin (Sigma) and 2 mM L glutamine (Gibco). Before any experiment, primary hepatocytes were cultured overnight and PPCs for 4 days with medium replacement daily. For isolation of monocytes, marrow cells were cultured for five days in DMEM containing 10% FBS, 30% L929 conditioned media, 50 μg/ml streptomycin, 50 U/ml penicillin (Sigma) and 2 mM L glutamine (Gibco). Monocytes were harvested by collecting the non-adherent cells using an EDTA-free wash buffer. For neutrophil isolation, marrow cells were isolated from the cell suspension by density gradient centrifugation on Percoll. The Percoll density gradient was prepared in a 15 ml tube by layering 2 ml of 67 and 52% Percoll solution on top of 2 ml 75% Percoll solution. Cells were resuspended in 2 ml of PBS and loaded on top of the Percoll density gradient. Neutrophils were separated by centrifugation at 400 g for 30 min at 4°C in a swinging bucket rotor. The cell band formed between the 75 and 67% layer was harvested, cells were diluted with PBS, washed, and diluted in standard medium. Monocytes and neutrophils were cultured for 15 h in low oxygen incubator prior to EV isolation. Primary MEFs and adipocytes were derived from Ercc1 −/− animals. Primary MEFs were induced 2 days after confluency for adipocyte differentiation with standard medium supplemented with an adipogenic cocktail (0,5 mM IBMX, 1 μΜ dexamethasone, 10μgr/ml insulin) 15 . BL/ 6 or Er1 F+ BMDMs or Er1 F/− were treated with 2.5 μg/μl mitomycin C (MMC; AppliChem) or tunicamycin (1μgr/ml), chloroquine (10μΜ), nocodazole (5 μM) and NSC23766 (50μΜ Rac1 inhibitor), for~15 h in standard medium. For ATM or ATR kinase inhibitor assays (ATMi; ATRi), cells were incubated with 0.66 μM ATMi or ATRi (Millipore) for 1 h, prior to MMC addition. For plasmid transfections, approximately 1 × 10 6 differentiated BMDMs were transfected with 5 μg of total plasmid DNA using Amaxa mouse macrophage nucleofector kit according to the manufacturer's instructions (Lonza, VPA1009BMDMs were incubated for 24 h and their media were used for EV isolation. To generate the RAB10-GFP and RAC1-GFP and CD9-GFP, the cDNA encoding the whole open reading frame Fig. 6 Er1 F/− EVs stimulate the glucose uptake in EV-recipient cells. a Immunofluorescence detection of the pKH67-labelled EVs in PPCs. Numbers indicate the average percentage of positive cells ± SEM (n > 800 cells counted in three independent experiments). b Immunofluorescence detection of RAB10-GFP and c RAC1-GFP in PPCs treated with EVs from Er1 F/− BMDMs transfected with RAB10-GFP or RAC1-GFP. Numbers indicate the average percentage of positively stained cells ± SEM for the indicated, color-matched protein. (n > 700 cells counted in three independent experiments). d Immunofluorescence detection of fluorescent tracer 2-NBDG for the monitoring of glucose uptake in PPCs (~2500 cells per genotype) and hepatocytes (~250 cells per genotype) exposed to Er1 F/− and Er1 F/+ -derived EVs (as indicated). e Immunofluorescence detection of fluorescent tracer 2-NBDG for the monitoring of glucose uptake in PPCs exposed to Er1 F/− and Er1 F/+ EVs derived from animal sera (ser.; n > 700 cells per genotype). f Immunofluorescence detection of fluorescent tracer 2-NBDG for the monitoring of glucose uptake in PPCs exposed to MMC EVs (as indicated, n > 2000 cells per treatment). g Western blotting of GLUT1 and GLUT3 protein levels in PPCs exposed to Er1 F/− and Er1 F/+ -derived EVs (n = 4). A graph showing the fold change and statistical significance of GLUT1 and GLUT3 protein levels is shown in Supplementary Fig. 7E. h Glut1 mRNA levels (in fold change; fc) in PPCs exposed to EVs derived from Er1 F/− (Er1 F/− EVs) or MMC-treated (MMC-EVs) macrophages compared to untreated wt. macrophages (red dotted line) for 24 h (see also Supplementary Fig. 7B-D). i Immunofluorescence detection of GLUT1 and RAB10 in PPCs exposed to Er1 F/− and Er1 F+ -derived EVs (see also Supplementary Fig. 7G, H). Numbers indicate the average percenatge of GLUT1, RAB10 double positive cells in each experimental condition (n > 200 cells counted in three independent experiments). j Oxygen consumption rate (OCR) of PPCs maintained in the presence of 15 mmol glucose that were exposed to Er1 F/+ , Er1 F/− or MMC-treated, macrophage-derived EVs (upper panel) and to Er1 F/+ or Er1 F/− serum-derived EVs (lower panel); numbers indicate the % increase in OCR compared to corresponding controls. k Western blotting of GLUT1 and GLUT3 protein levels in Er1 F/− and Er1 F/+ liver and pancreata (n = 3). A graph showing the fold change and statistical significance of GLUT1 and GLUT3 protein levels is shown in Supplementary Fig. 8A. l Immunofluorescence detection of GLUT1 and GLUT3 in the pancreas of Er1 F/− and Er1 Glucose uptake activates mTOR and pro-inflammatory responses in EV-recipient cells. a Immunofluorescence detection of iNOS accumulation (indicated by the arrowhead) in PPCs exposed to culture media (CM) and the EVs derived from MMC-treated and untreated control macrophages (n = 3, >600 cells per treatment) or b Er1 F/− and Er1 F/+ macrophages (n = 3, > 1200 cells/treatment) (see also Supplementary Fig. 8D). c Immunofluorescence detection of NF-kβ nuclear translocation (indicated by the arrowhead) in PPCs exposed to CM and the EVs derived from MMC-treated and untreated control macrophages or (n = 3, >700 cells per treatment). d Er1 F/− and Er1 F/+ macrophages (see also Supplementary Fig. 8E). (n = 3, >500 cells per treatment). e Immunofluorescence detection of NF-kβ (shown by the arrowhead) in PPCs exposed to CM supplemented with EVs from MMC-treated and control macrophages upon low (5 mmol) or high (15 mmol) glucose concentration (n = 4, >750 cells per treatment; see also Supplementary Fig. 9A-C).
f Western blotting of phosphorylated pS6K, phosphorylated p4EBP1, phosphorylated pAKT1, phosphorylated pAKT2 and phosphorylated pAKT protein levels in Er1 F/− and Er1 F/+ pancreata (n = 4). The graph represents the fold change (F.C) of indicated protein levels in Er1 F/− pancreata to wt. controls. g Western blotting of pS6K, p4EBP1 and iNOS protein levels in MMC-treated macrophages exposed to rapamycin (n = 3, Rap/cin; as indicated). The graph represents the fold change of protein levels in MMC-treated macrophages exposed to rapamycin compared to MMC-treated macrophage control (ctrl.) cells. h Immunofluorescence detection of NF-kβ and 2-NBDG in PPCs (indicated by the arrowheads) exposed to CM supplemented with EVs derived from Oxygen consumption rate measurement. For the oxygen consumption rate (OCR) analysis, PPCs (~1.5 × 10 6 ) were treated with EVs for 24 h as described above and were harvested with trypsin/EDTA, rinsed with phosphate-buffered saline (PBS) (with 10 % FBS), and centrifuged at 1200 × g for 3 min. Cells were re-suspended in 1.5 ml of PBS. The OCR (nmoles O 2 /min) was measured for 20 min at 37°C using a Clark-type electrode (Hansatech). The average OCR was determined from 1 min measurements deriving from four different time points (3, 6, 9, 12 min) 72 . OCR data were normalized to total protein content using the Bradford method.
Immunoblot analysis and antibodies. For western blot analysis, cells were pelleted and tissues from Er1 F/− and Er1 F/+ animals were homogenized in Sucrose buffer (0.32 M Sucrose, 15 mM HEPES-KOH, 60 mM KCl, 2 mM EDTA, 0.5 mM EGTA, 0.5%BSA, 0.1%NP-40, pH-7.9 and protease inhibitors). Cell pellets were washed three times with 1x PBS. Cell pellets were then re-suspended in NP-40 lysis buffer (10 mM Tris-HCl pH 7.9, 10 mM NaCl, 3 mM MgCl2, 0.5% NP-40 and protease inhibitors) and incubated for 10 min at 4°C. The supernatant after centrifugation was kept as the cytoplasmic fraction and the pellets was re-suspended in high-salt extraction buffer (10 mM HEPES-KOH pH 7.9, 380 mM KCl, 3 mM MgCl2, 0.2 mM EDTA, 20% glycerol and protease inhibitors) and incubated for 60 min, 4°C. The supernatant after centrifugation was kept as the nuclear fraction 95 . For whole-cell extract preparations, cells pellets were resuspended in 150 mM NaCl, 50 mM Tris pH = 7.5, 5% Glycerol, 1% NP-40, 1 mM MgCl) and incubated on ice for 30 min. For western blot analysis of EVs, EV pellets were resuspended in 2x Laemli buffer. Antibodies against Rac2 (C-11, WB: 1:500, IF:  Fig. 8 Exogenous delivery of Er1 F/− exosomes promote GLUT1 accumulation, glucose uptake and inflammation in vivo. a Immunofluorescence detection of GLUT1 in the liver of 4-weeks-old C57BL/6 animals intravenously injected with EVs derived from either Er1 −/− or wt. MEFs and b Er1 F/+ or Er1 F/− macrophages (as indicated). Numbers show average mean fluorescence intensity ± SEM, n = 4 animals per genotype, >3 optical fields/animal. c Western blotting of GLUT1 in the muscle and liver of C57BL/6 animals intravenously injected with EVs derived from either Er1 −/− or wt. MEFs and d Er1 F/+ or Er1 F/− macrophages (as indicated). The graph represents the fold change (F. C) of GLUT1 protein levels in the liver or muscle tissues of animals treated with Er1 −/− MEF or Er1 F/− macrophage-derived EVs as compared to corresponding tissues of animals treated with Er1 +/+ MEFs or Er1 F/+ macrophage-derived EVs. e Glucose tolerance test (GTT) graphs of C57BL/6 mice injected intraperitoneally every 24 h for a period of 10 days with EVs derived from either Er1 −/− or wt. MEFs (n = 6) and Er1 F/+ or Er1 F/− macrophages (n = 8) (as indicated). f The accumulation of irreparable DNA lesions in tissue-infiltrating macrophages activates the secretion of EVs in vivo and ex vivo. Er1 F/− macrophage-derived EVs are targeted to recipient tissues and cells triggering the expression and translocation of insulin-independent glucose transporters GLUT1 and 3 in cell membrane. This leads to an increase in cellular glucose uptake in cells and greater tolerance to glucose challenge in higher cellular oxygen consumption rate and greater tolerance to glucose challenge in Er1 F/− mice. In turn, high glucose levels activate pro-inflammatory responses in an mTOR-dependent manner leading to chronic inflammation in Er1 F/− animals. Error bars indicate S.E.M. among replicates (n is indicated in each panel). Asterisk indicates the significance set at p-value: *≤0.05, **≤0.01 (two-tailed Student's t-test). Gray line is set at 10 μm scale. a 75μm internal diameter (ThermoFisher Scientific, USA) packed in-house using a packing bomb (Loader kit SP035, Proxeon). Tryptic peptides were separated and eluted in a linear water-acetonitrile gradient and injected into the MS.
RNA-Seq and Quantitative PCR studies. Total RNA was isolated from Er1 F/+ and Er1 F/− BMDMs using a Total RNA isolation kit (Qiagen) as described by the manufacturer. For RNA-Seq studies, libraries were prepared using the Illumina® TruSeq® mRNA stranded sample preparation Kit. Library preparation started with 1 µg total RNA. After poly-A selection (using poly-T oligo-attached magnetic beads), mRNA was purified and fragmented using divalent cations under elevated temperature. The RNA fragments underwent reverse transcription using random primers. This is followed by second strand cDNA synthesis with DNA polymerase I and RNase H. After end repair and A-tailing, indexing adapters were ligated. The products were then purified and amplified to create the final cDNA libraries. After library validation and quantification (Agilent 2100 Bioanalyzer), equimolar amounts of all 12 libraries were pooled. The pool was quantified by using the Peqlab KAPA Library Quantification Kit and the Applied Biosystems 7900HT Sequence Detection System. The pool was sequenced by using an Illumina HiSeq 4000 sequencer with a paired-end (2 × 75 cycles) protocol. Quantitative PCR (Q-PCR) was performed with a DNA Engine Opticon device according to the instructions of the manufacturer (MJ Research). The generation of specific PCR products was confirmed by melting curve analysis and gel electrophoresis. Each primer pair was tested with a logarithmic dilution of a cDNA mix to generate a linear standard curve (crossing point (CP) plotted versus log of template concentration), which was used to calculate the primer pair efficiency (E = 10 (−1/ slope) ). Hypoxanthine guanine phosphoribosyltransferase1 (Hprt-1) mRNA was used as an external standard. For data analysis, the second derivative maximum method was applied: (E 1gene of interest ΔCP (cDNA of wt. mice -cDNA of Ercc1F/-) gene of interest )/(E hprt-1 ΔCP (cDNA wt. mice-cDNA) hprt-1 ). Hprt F: CCCAACATCAACAG GACTCC, Hprt R: CGAAGTGTTGGATACAGGCC, Cxcl5 F:TGCCCCTTCCT CAGTCATAG, Cxcl5 R:GGATCCAGACAGACCTCCTTC, Cxcl24 F:AATTC CAGAAAACCGAGTGG, Cxcl24 R:TGGGCCCCTTTAGAAGGCTGG, Cxcl1 F: CCACACTCAAGAATGGTCGC, Cxcl1 R: GTTGTCAGAAGCCAGCGTTC, Ccl9 F:CCGGGCATCATCTTTATCAG, Ccl9 R:GTCCGTGGTTGTGAGTTTTCC, Cxcl12 F:ACGTCAAGCATCTGAAAATCC, Cxcl12 R:AATTTCGGGTCAATG CACAC, Csf1 F: CCTCATGAGCAGGAGTATTGC, Csf1 R:AAAGGCAAT CTGGCATGAAG, Ccl7 F:TCCCTGGGAAGCTGTTATCTTC, Ccl7 R: TGGAGTTGGGGTTTTCATGTC, Cxcl13 F:TAGATCGGATTCAAGTTACGC, Cxcl13 R:GTAACCATTTGCCACGAGG, Ccl3 F:AGATTCCACGCCAATTCA TC, Ccl3 R:TCAAGCCCCTGCTCTACAC, Ccl5 F:CTCGTGCCCACGTCAAG GAG, Ccl5 R:CCCACTTCTTCTCTGGGTTG, Cxcl9 F:CGGAGATCAAACCT GCCTAG, Cxcl9 R:CTTGAACGACGACGACTTTG.
Data analysis. For Q-PCR data, a two-way t-test was used to extract the statistically significant gene expression data (unless noted otherwise) by means of the IBM SPSS Statistics 19 (IBM, NY, USA), Spotfire (Tibco, CA, USA), Partek (Partek INCoR1porated, MO, USA) and R-statistical package (www.r-project.org/). For mass spectrometry (MS), the MS/MS raw data were loaded in Proteome Discoverer 1.3.0.339 (Thermo Scientific) and run using Mascot 2.3.02 (Matrix Science, London, UK) search algorithm against the Mus musculus theoretical proteome (Last modified 6 July 2015) containing 46,470 entries 96 . A list of common contaminants was included in the database 97 . For protein identification, the following search parameters were used: precursor error tolerance 10 ppm, fragment ion tolerance 0.8 Da, trypsin full specificity, maximum number of missed cleavages 3 and cysteine alkylation as a fixed modification. The resulting.dat and.msf files were subsequently loaded and merged in Scaffold (version 3.04.05, Proteome Software) for further processing and validation of the assigned MS/MS spectra employing PeptideProphet and ProteinProphet algorithms for the identification of proteins [98][99][100] . Thresholds for protein and peptide identification were set to 99 and 95% accordingly, for proteins with minimum 1 different peptides identified, resulting in a protein false discovery rate (FDR) of <0.1%. For single peptide identifications, we applied the same criteria in addition to manual validation. Protein lists were constructed from the respective peptide lists. For label-free relative quantitation of proteins, we applied a label-free relative quantitation method between the different samples (control versus bait) in order to determine unspecific binders during the affinity purification. All.dat and.msf files created by Proteome Discoverer were merged in Scaffold where label-free relative quantification was performed using the total ion current (TIC) from each identified MS/MS spectra. The TIC is the sum of the areas under all the peaks contained in a MS/MS spectrum and total TIC value results by summing the intensity of the peaks contained in the peak list associated to a MS/MS sample. This approach has advantages in comparison to other label-free methods, such as increased dynamic range and quantification for low spectral counts 101 . Protein lists containing the calculated by Scaffold total TIC quantitative value for each protein ware exported from to Microsoft Excel for further processing. The fold change of protein levels was calculated by dividing the mean total TIC quantitative value in bait samples with the mean value of the control samples for each of the proteins. Proteins having ≥80% protein coverage, ≥1 peptide in each sample (Er1 F/− BMDMs) and a fold change ≥1.55 were selected as being significantly enriched in Er1 F/− BMDMs compared to Er1 F/+ BMDM controls. Significant overrepresentation of pathways, protein-protein interactions and protein complexes were derived by STRING 102 (http://string-db.org/). For RNA-Seq data analysis, the data were downloaded as FASTQ files and their quality was checked with FASTQC, a quality control tool for high throughput sequence data: http://www.bioinformatics. babraham.ac.uk/projects/fastqc). The data were aligned to mm10 genome assembly available from UCSC via Bowtie2. The differentially expressed genes were identified with metaseqR. Count normalization was performed based on edaseq algorithm. In parallel, cufflinks pipeline (version 2). Significant overrepresentation of pathways and gene networks was determined by Gene Ontology and DAVID (http://david. abcc.ncifcrf.gov/summary.jsp; through BBID, BIOCARTA and KEGG annotations). All experiments were repeated ≥3 times. The data exhibited normal distribution (where applicable). For animal studies, each biological replicate consists of 3-5 mouse tissues or cell cultures per genotype per time point or treatment (unless stated otherwise). None of the samples or animals was excluded from the experiment. The animals or the experiments were not randomized. The investigators were not blinded to allocation during animal experiments and outcome assessment.
Reporting summary. Further information on research design is available in the Nature Research Reporting Summary linked to this article.

Data availability
The mass spectrometry proteomics data have been deposited to the ProteomeXchange Consortium (http://proteomecentral.proteomexchange.org) via the PRIDE partner repository with the dataset identifier PXD015727 and 10.6019/PXD015727. The RNA-Seq. data are deposited in ArrayExpress (https://www.ebi.ac.uk/arrayexpress/), (E-MTAB-8439). All other data and reagents are available from the authors upon reasonable request. The source data underlying