Rab27-Dependent Exosome Production Inhibits Chronic Inflammation and Enables Acute Responses to Inflammatory Stimuli

The mammalian inflammatory response must maintain an intricate balance between proinflammatory and anti-inflammatory signals. Proper control of inflammation is needed to clear infections as well as maintain commensal microbe populations and overall homeostasis of tissues. When this balance is disrupted, a chronic, low-grade inflammation develops that over time can contribute to a variety of diseases associated with aging, including obesity and type 2 diabetes (1). Chronic inflammation and associated diseases are expanding at an alarming rate putting enormous stress on the US medical system and overall economy (2). There is a pressing need to improve our understanding of the underlying mechanisms that contribute to chronic inflammation to develop effective therapies moving forward. Noncoding RNAs, including microRNAs (miRNAs), are one such mechanism for regulating chronic inflammation as demonstrated by an age-dependent, chronic inflammatory disease seen in mice lacking miR-146a that is driven by another miRNA, miR-155 (2, 3). However, several aspects of how these miRNAs are regulated and function is still being explored.

Over the past decades, researchers have begun to investigate a novel form of intercellular communication between immune cells mediated by small lipid vesicles called exosomes (4–7). The transfer of cellular contents including proteins, RNAs, and other molecules by exosomes has been reported (8). Exosome-mediated transfer of miRNAs has become of particular interest to the field (7, 9, 10) because miRNAs can be transferred between different cell types, including immune cells, to regulate cellular responses (6, 7, 11). For instance, miRNAs can be shuttled between dendritic cells (DCs) and mediate target knockdown in recipient cells (12), as well as be transferred between T cells and DCs at the immunological synapse (13). Further, transfer of miRNAs via exosomes conveyed resistance to hepatitis B and contributed to IFN-α antiviral responses in mice (14). There are also examples of specific miRNAs being transferred between hematopoietic cells and altering responses. For example, the exogenous delivery of exosomal miR-155 leads to an enhanced response to inflammatory challenges, whereas administration of exosomal miR-146a reduces this response (11). Such effects have been observed both systemically and in specific tissues, including those that make up the CNS (15). In addition, miR-155 can also be transferred from acute myeloid leukemia cells to healthy blood cells, resulting in the suppression of c-MYB and compromised hematopoiesis in the context of cancer (16). These examples clearly demonstrate that miRNAs contained within exosomes are involved in the communication between immune and other hematopoietic cells, during both physiological and pathological situations.

Despite the expanding amount of research investigating the functional transfer of miRNAs between immune cells, the key question whether endogenously produced exosomes and their miRNA cargos are important for cellular communication in vivo is just beginning to be addressed as new tools emerge. Rab27a and Rab27b double-knockout (Rab27DKO) mice provide one such reagent with which to study the roles of exosomes in vivo because these mice exhibit significantly reduced exosome release, among other phenotypes, because of the role of these GTPases in the docking and retention of the multivesicular body to the plasma membrane (17–20). Rab27DKO regulatory T cells (Tregs) have previously been used to study the function of an exosomal miRNA, Let7d, in suppression of Th1 cells by Tregs using an adoptive transfer model (9). However, it remains unclear whether endogenously produced exosomes play important roles in other immune cell functions, including innate immunity.

In this study, we found that Rab27DKO mice have a chronic, low-grade inflammatory condition characterized by increased baseline inflammatory cytokines and myeloproliferation. Some of these phenotypes were cell extrinsic and could be rescued, at least in part, by wild-type (WT) bone marrow (BM) cells or injections of WT exosomes isolated from BM cells. This indicates that exosome uptake is important for maintenance of hematopoietic homeostasis and prevention of aberrant chronic inflammation. Chronically inflamed Rab27DKO mice were also hyporesponsive to challenge with LPS, and thus resembled a state of endotoxin tolerance (21). Further, we found that the response of Rab27DKO mice to LPS can be rescued by delivery of WT, but not miR-155^−/−, exosomes, suggesting that miR-155 contained within exosomes is essential for the proper responses to inflammatory cues. SHIP1 and IRAK-M, miR-155 targets that negatively regulate inflammatory responses (22, 23) and are involved in establishing tolerance to LPS (24), are elevated in Rab27DKO mice compared with WT mice. SHIP1 and IRAK-M levels were reduced after administration of WT, but not miR-155^−/−, exosomes in mice challenged with LPS, implicating miR-155 targeting of SHIP1 and IRAK-M in the exosomal rescue of Rab27DKO mouse response to LPS. Together, these results provide evidence that exosomal communication is important for proper maintenance of homeostasis in the immune system and subsequent responses to challenge with LPS.

CD45.2 WT (Jackson Labs), CD45.1 WT (Jackson Labs), and Rab27DKO (Rab27a ash/ash Rab27b^−/−) mice (T. Tolmachova and M.C. Seabra, Imperial College London) are on a C57BL6 genetic background and housed in the animal facility at the University of Utah. Experiments were approved by the Institutional Animal Care and Use Committee at the University of Utah. Mice were age- and sex-matched, and were in the age range of 6 to 16 wk old for all experiments. Both males and females displayed observed phenotypes.

Differential centrifugation was performed to isolate exosomes from BM conditioned medium. Initial spins consisted of 1000 × g for 10 min, 2000 × g for 10 min, and 10,000 × g for 30 min. The supernatant was retained each time. The supernatant was then spun at 100,000 × g for 70 min, and the pellet was resuspended in 25 ml 1× PBS, to dilute remaining soluble factors, followed by another centrifugation at 100,000 × g for 70 min. The final pellet contained the exosomes, which were resuspended in PBS. This protocol is based on previous exosome isolation methods (25). We used a Thermo Scientific Sorvall Lynx 6000 with a T26-8 × 50 rotor. For the in vitro and in vivo experiments, exosomes were isolated from BM cultured in GM-CSF (GM-BM cells) from WT, Rab27DKO, or miR-155^−/− mice. BM cells were incubated for 3 d with 20 ng/ml GM-CSF and then given an additional 5 ml of medium with 20 ng/ml GM-CSF for a total of 7 d of culture. For the in vitro experiments, exosomes isolated from 3 million GM-BM cells were transferred to the same amount of recipient cells. For the in vivo experiments, each mouse was i.p. injected with exosome pellets resuspended in 100 μl of 1× PBS. These exosomes were derived from three 10-cm plates of 3 million GM-BM cells per plate and resuspended in 100 μl of PBS, which yields ∼10⁹ exosomes as previously quantified (11). Protein concentrations in the exosome preparations were also quantified, and similar protein levels were injected except for Rab27DKO exosome pellets; exosomes were previously characterized to be CD63⁺ and of normal size and morphology (11).

BM-derived DCs (BMDCs) were derived from mouse BM by culturing RBC-depleted BM in complete RPMI 1640 (10% FBS, 100 U/ml penicillin, and 100 U/ml streptomycin, glutamate, sodium pyruvate, HEPES, and nonessential amino acids) with 20 ng/ml GM-CSF for 3–4 d at 37°C with 5% CO₂. The cells were then spiked with 5 ml of complete RPMI 1640 with 20 ng/ml GM-CSF for an additional 3–4 d for a total of 7 d in culture. LPS stimulation was performed at a concentration of 500 ng/ml.

miRNA mimics were purchased from Qiagen. Seed mutant and miR-mimic sequences are as follows: miR-155 seed mutant (5′-UUUGCUAAAAUUGUGAUAGGGGU-3′) and miR-155 mimic (5′-UUAAUGCUAAUUGUGAUAGGGGU-3′). Rab27DKO BMDCs were transfected with 30 μl of the Tran-IT siQUEST reagent (Mirus) in 2 ml of OPTI-MEM media with 60 ng of each mimic. After 24 h, the Rab27DKO BMDCs were treated with or without LPS at a concentration of 500 ng/ml, and media were isolated at 2 and 6 h for ELISAs.

Fluorophore-conjugated Abs against the indicated surface markers (eBioscience) were used to stain RBC-depleted splenocytes and BM cells. Stained cells were analyzed with a BD LSR Fortessa flow cytometer, and further data analysis was carried out with FlowJo software.

The ELISA used to quantify mouse IL-6 and TNF-α concentrations were obtained from eBioscience and were performed using the manufacturer’s suggested protocol.

Cellular extract was size fractionated via SDS-PAGE, and immunoblotting was performed in accordance with standard protocols. Specific Abs were used to detect SHIP1 (sc-1964; Santa Cruz) and GAPDH (ab9485; Abcam).

RNA isolation was performed using the miRNeasy kit (Qiagen), according to manufacturer’s instructions. cDNA from total RNA was made with qScript using 90 ng of RNA from each sample (Quanta). qPCR was performed with Promega GoTaq qPCR master mix. Primer sequences are as follows: SHIP1 forward, 5′-GAGCGGGATGAATCCAGTGG-3′; SHIP1 reverse, 5′-GGACCTCGGTTGGCAATGGTA-3′; IRAK-M forward, 5′-CCTGAACATAATGAAAAAGGAACAC-3′; IRAK-M reverse, 5′-ATGCTTGGTTTCGAATGTCC-3′; L32 forward, 5′-AGCTCCCAAAAATAGACGCAC-3′; and L32 reverse, 5′-TTCATAGCAGTAGGCACAAAGG-3′. L32 levels were used to normalize mRNA expression levels.

CD45.1 WT mice were lethally irradiated (1000 rad) using an x-ray source (Rad Source RS200 biological system). After irradiation, mice were injected with 3 million RBC-depleted BM cells via retro-orbital injection and aged for 2 mo before analysis.

Escherichia coli LPS (Sigma) was administered through i.p. injections at a sublethal concentration of 50 μg in 1× PBS per mouse. In exosome injection experiments, exosomes were i.p. injected 24 h before LPS injection and again at 48 h after LPS injection. Mice were harvested at 72 h after LPS injection for analysis of immune cell populations.

Data were analyzed using either Welsh corrected t tests or one-way ANOVA with Tukey multiple comparisons, and adjusted p values are reported. These statistics were done using GraphPad Prism software. The p values were either listed or represented by asterisks as follows: *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001.

Upon phenotyping Rab27DKO mice, we observed that granulocyte-monocyte (GM) myeloid populations, marked by surface expression of CD11b⁺ and GR1⁺, were expanded in the Rab27DKO mice in both the spleen and BM compartments with a corresponding decrease in the B220⁺ B cell population compared with WT controls (Fig. 1A–C, 1E–G). There were signs of extramedullary hematopoiesis indicated by the expansion of Ter119⁺ erythroid precursor cells in the spleen, and a reduction of these cells in the BM (Fig. 1A, 1D, 1E, 1H). Increased spleen weights were also observed in the Rab27DKO versus WT control mice (Fig. 1I). These data suggest that the Rab27DKO mice have a mild myeloproliferative disorder. In contrast, no apparent differences in splenic and thymic T cell populations, hematopoietic stem and progenitor cells (lineage-negative Kit⁺ Sca⁺), and overall hematopoietic cell numbers in the BM and spleen of Rab27DKO mice were observed when compared with WT mice (Supplemental Fig. 1). Beyond hematopoietic population differences, the proinflammatory cytokines TNF-α and IL-6 were elevated above baseline in the serum of Rab27DKO mice (Fig. 1J, 1K). These results indicate that mice deficient in Rab27a/b develop a chronic, low-grade inflammatory condition.

To determine whether these phenotypes are cell intrinsic or extrinsic, we used BM radiation chimeras. CD45.1 WT mice were lethally irradiated and reconstituted with either a 1:1 mixture of CD45.1 WT/CD45.2 WT, Rab27DKO (CD45.2)/WT (CD45.1), or only Rab27DKO BM cells. After 2 mo, we found that CD45.1 and CD45.2 populations were fairly equivalent in both groups (Fig. 2A, 2B), indicating that Rab27DKO hematopoietic stem cells are able to engraft with similar efficiency as WT hematopoietic stem cells. However, Rab27DKO/WT reconstituted mice no longer had an overall accumulation of CD11b⁺ GR1⁺ GM myeloid cells in the spleen as seen in Rab27DKO reconstituted mice (Fig. 2C, 2D). To investigate whether Rab27DKO or WT cells composed this population, we gated on CD45.1 and CD45.2 and found that within the CD11b⁺ GR1⁺ population there were not significantly more WT or Rab27DKO cells making up this population, suggesting that the presence of WT cells was able to rescue the accumulation of these cells (Fig. 2E). In addition, there was some rescue of B220⁺ B cell populations in the spleens of mice reconstituted with the WT/Rab27DKO BM mixture, which was also composed of WT and Rab27DKO-derived cells at equal proportions (Fig. 2F–H). The Ter119⁺ erythroid precursor population was restored back to normal low levels in the spleen in the WT/Rab27DKO BM chimeras (Fig. 2I). This was also accompanied by a decrease in spleen weights (Fig. 2J). Similar trends were seen in the BM with the exception of the Ter119⁺ population, which was not rescued with the presence of WT BM (Supplemental Fig. 2). These results provide evidence that certain Rab27DKO phenotypes are cell extrinsic and potentially regulated by exosomes, whereas others may be intrinsic and regulated by exosome-independent mechanisms.

Due to our observations that some of the Rab27DKO mouse chronic, low-grade inflammatory phenotypes are a result of cell-extrinsic mechanisms, we investigated whether exosome uptake was an important contributing factor to homeostasis within the immune system. To test this, we i.p. injected Rab27DKO mice with exosomes produced by WT GM-BM cells or a control “exosome” pellet from Rab27DKO GM-BM cells, which have been previously shown to produce significantly fewer exosomes (11, 18). We have previously characterized these WT exosomes and found they are CD63⁺ and display a typical morphology and size as observed by electron microscopy (11). Injections were performed two times per week over the course of 4 wk.

WT exosome injections fully rescued the splenomegaly of the Rab27DKO mice, whereas the splenic GR1⁺ CD11b⁺ GM myeloid population was partially reduced toward WT levels (Fig. 3A–C). Ter119⁺ erythroid precursor cell expansion in the Rab27DKO mouse spleens was also reduced after injections of WT exosomes, but the B220⁺ B cell population was not rescued by WT exosome treatment (Fig. 3D–F). The GR1⁺ CD11b⁺ GM population in the BM also recovered after WT exosomes were injected (Fig. 3G, 3H); however, the Rab27DKO Ter119⁺ and B220⁺ BM cellular populations were not altered after delivery of WT exosomes (Fig. 3I–K). Levels of IL-6 and TNF-α in the Rab27DKO serum were also reduced to normal levels after injection of WT exosomes (Fig. 3L, Supplemental Fig. 3). This experiment was repeated with PBS and miR-155^−/− exosome injections as additional controls, and we observed that the injections with PBS displayed similar results as the injections of Rab27DKO exosomal pellets, whereas the injection of miR-155^−/− exosomes showed results similar to WT exosome injections (Supplemental Fig. 3). These results suggest that Rab27DKO mouse steady-state phenotypes that are regulated by exosomes are independent of exosomal miR-155. Of note, miR-155^−/− exosomes have been previously shown to be made at similar levels as WT exosomes (11), which is corroborated by similar protein concentrations in our WT and miR-155^−/− exosomal preparations (Supplemental Fig. 3). Rab27DKO exosomal pellets have significantly decreased protein concentration, as well as significantly decreased levels of miR-155 and miR-146a in their exosomal pellet (Supplemental Fig. 3). A summary of which cellular phenotypes are rescued by both WT BM and WT exosomes can be found in Supplemental Table I. These results suggest that some Rab27DKO mouse phenotypes are dependent on GM-BM exosome uptake, such as Ter119⁺ cell accumulation in the spleen and splenomegaly, increased GM myeloid cells, and elevated cytokine levels, whereas others appear to be independent of GM-BM exosome uptake.

Due to the chronic inflammation observed in the Rab27DKO mice, we next wanted to investigate how these mice respond to a LPS challenge. To do this, we i.p. injected WT or Rab27DKO mice with a nonlethal dose of LPS and analyzed the inflammatory response by isolating serum at 2 and 6 h after LPS injection and performing IL-6 and TNF-α ELISAs. Despite starting out with increased TNF-α and IL-6 basal levels, Rab27DKO mice could not elevate these levels in response to LPS to the same extent as WT mice (Fig. 4A–F). Normally, during a LPS challenge, mice will undergo a switch in hematopoietic development in the BM compartment, termed emergency granulopoiesis, where the GR1⁺ CD11b⁺ GM myeloid population expands and the B220⁺ B and Ter119⁺ erythroid precursor cell populations contract by 72 h after LPS stimulation. Rab27DKO mice were unable to shift their myeloid, B, and erythroid precursor populations in response to LPS, which was observed in their WT counterparts (Fig. 4G–K). These results indicate that chronically inflamed Rab27DKO mice are refractory to stimulation with LPS.

To determine whether the failure of Rab27DKO mice to respond to LPS could be rescued by exosome uptake, we i.p. injected WT exosomes into Rab27DKO mice, or PBS as a control, 24 h before LPS injection (Fig. 5A). Serum levels of TNF-α and IL-6 were significantly increased when Rab27DKO mice were pretreated with WT exosomes compared with PBS (Fig. 5B–E). To analyze changes in emergency granulopoiesis, we injected mice a second time with exosomes 48 h after LPS administration and then harvested them 24 h later (Fig. 5A). Rab27DKO mice given WT exosomes were able to expand their CD11b⁺ GR1⁺ myeloid populations and reduce their B220⁺ B cell and Ter119⁺ erythroid precursor populations in response to LPS (Fig. 5F–J), suggesting that the uptake of WT exosomes is able to restore the ability of Rab27DKO mice to perform emergency granulopoiesis in response to LPS. It is important to note that the two-injection regimen did not rescue the baseline levels of myeloid, B, and erythroid precursor cells, but did rescue the responsiveness of these populations to LPS (Fig. 5F–J). These results demonstrate that uptake of GM-BM–derived exosomes can complement the refractory response to LPS in Rab27DKO mice, and suggest that exosomes are involved in proper responsiveness to endotoxin in vivo.

Based on our findings that Rab27DKO mice have a refractory response to endotoxin that involves exosome uptake, we next wanted to investigate which factor in the exosome could be responsible for this outcome. Based on our previous findings that miR-155–containing exosomes were able to bolster the response to endotoxin (11), and recent observations that exosomal miR-155 levels increase in response to LPS in vivo (26), we performed the following experiment. Rab27DKO, WT, or miR-155^−/− GM-BM–derived exosome pellets were administered to recipient Rab27DKO BMDCs, followed by LPS administration 24 h later. Two hours after LPS treatment, TNF-α levels were increased in the Rab27DKO BMDCs given WT exosomes, but not the Rab27DKO BMDCs that received their own exosome pellet or miR-155^−/− exosomes (Fig. 6A, 6B), indicating that WT exosomes can rescue this defect through a miR-155–dependent mechanism. We next investigated whether SHIP1, a target of miR-155 that negatively regulates inflammation and responses to LPS, was affected by the administration of WT exosomes. We found that WT, but not miR-155^−/−, exosomes were able to reduce SHIP1 and IRAK-M levels in Rab27DKO BMDCs, corresponding to increased TNF-α levels (Fig. 6C, 6D), consistent with previous reports of SHIP1 negatively regulating TNF-α production (27). Further, another negative regulator of inflammation and target of miR-155 (28), IRAK-M, displays a similar pattern as SHIP1 (Fig. 6E), suggesting that miR-155 targets multiple negative regulators of inflammation to promote the LPS response in Rab27DKO BMDCs.

To determine whether miR-155 is sufficient to rescue Rab27DKO BMDC response to LPS, we used miR-155 mimics. Rab27DKO BMDCs were transfected with either a mature miR-155 mimic or a miR-155 mimic where the seed site had been mutated (seed) using a lipid-based reagent. Twenty-four hours after mimic transfection, the cells were treated with or without LPS. Rab27DKO BMDCs that received the miR-155 mimic had increased IL-6 and TNF-α levels compared with those that received the seed mutant mimic (Fig. 6F–I). These results demonstrate that miR-155 is sufficient to rescue Rab27DKO BMDC LPS response, which further supports the idea that the delivery of miR-155 via WT exosomes is responsible for rescuing this response.

Based on our in vitro observations, we investigated the requirement of miR-155 in exosomes for the rescue of Rab27DKO mouse response to LPS. To study this, we injected WT, miR-155^−/−, or Rab27DKO GM-BM–derived exosome pellets into Rab27DKO mice 24 h before LPS injection and again 48 h after LPS injection with the same timeline as Fig. 5A. Rab27DKO mice treated with WT exosomes had enhanced TNF-α production by 2 h after LPS challenge, whereas TNF-α levels in LPS-treated Rab27DKO mice given miR-155^−/− or Rab27DKO exosome pellets were not rescued (Fig. 7A, 7B). We also found that resting Rab27DKO mice had increased levels of SHIP1 and IRAK-M (Fig. 7C–F), suggesting that the Rab27DKO mice are trying to compensate for their chronic inflammatory status by upregulating negative regulators of inflammation. The increase in SHIP1 and IRAK-M at resting conditions could explain why the Rab27DKO mice are hyporesponsive to LPS, and why WT exosome administration aids in the rescue of Rab27DKO response to LPS through an miR-155–dependent mechanism. This hypothesis is supported by the observed reduction in SHIP1 and IRAK-M levels when LPS-treated Rab27DKO mice received WT, but not miR-155^−/−, exosomes (Fig. 7G, 7H). In addition, the GM myeloid population was increased in response to LPS administration after pretreatment with WT, but not miR-155^−/−, exosomes (Fig. 7I, 7J). These results provide evidence that miR-155 is required for exosomes to rescue LPS responsiveness by Rab27DKO mice, and that reductions in miR-155 targets SHIP1 and IRAK-M are involved in mediating this rescue (Fig. 8).

Although it is clear from the literature that exosomes are important for intercellular communication between immune cells (6, 29), the roles of endogenously produced exosomes in vivo are just beginning to be investigated. Previous studies have shown that endogenous exosome production and content are affected by disease states in humans (30, 31) and can be used as biomarkers for diseases such as obesity and type 2 diabetes (32). However, it is unclear from these studies whether exosomes are playing a role in disease or whether they are mere byproducts. In addition, manipulated tumor exosomes have been shown to activate CD8⁺ T cells in vivo (33), whereas endogenous exosomes can mediate Treg suppression of Th1 cells via Let7d in an adoptive transfer system (9). Although these studies provide evidence for the importance of endogenously produced exosomes during immunity, we designed our approach utilizing the Rab27DKO mouse model where we were able to study the relevance of endogenous exosome production with minimal manipulation because these mice have defective exosome production (9, 17).

Although useful for exosome studies, it is important to note that the Rab27DKO mouse model has defects beyond exosome release. Rab27a- and/or Rab27b-deficient mice have been shown to have defective granule release by platelets, cytotoxic T cells, and neutrophils, as well as improper neutrophil chemotaxis in certain contexts (18–20, 34). There may also be a role for the Rab27 proteins in recycling of membrane proteins and phagosome function (35, 36). Based on these additional functions of the Rab27 proteins in immune and cellular functions, it is critical to distinguish these from their role in exosome secretion. To address this, we attempted to rescue observed phenotypes with either WT hematopoietic cells or the administration of WT exosomes. If the defect was rescued by both approaches, then it suggests that a lack of exosome uptake contributes to the phenotype.

Our data demonstrate that exosome uptake is essential for hematopoietic homeostasis and the prevention of chronic inflammation in Rab27DKO mice. Corresponding to increased basal inflammation, the Rab27DKO mice were hyporesponsive to LPS, consistent with previous findings that Rab27a knockout mice (ashen) are protected from LPS sepsis (37). We hypothesized that the Rab27DKO mice are less responsive to LPS because they have increased negative regulators of inflammation, such as SHIP1 (24, 27), as a way to combat their chronic inflammatory condition. This would explain why WT exosomes could rescue Rab27DKO response to LPS, whereas miR-155^−/− exosomes could not because SHIP1 is a known miR-155 target (23), and TNF-α and IL-6 have been previously shown to be inhibited by SHIP1 (27). Further, previous studies have demonstrated that miR-155 upregulation increases TNF-α production through repression of its targets SHIP1 and IRAK-M (26, 28), which agrees with our findings that exosomally delivered miR-155 is able to restore Rab27DKO cellular response to LPS via a mechanism that involves these targets. Our model for how exosome uptake is affecting the LPS response is summarized in Fig. 8. Together our findings implicate endogenously produced, miRNA-containing exosomes in the regulation of the innate immune response in vivo.

Although our findings indicate that exosomal communication is needed to maintain hematopoietic homeostasis during the steady-state, it remains to be clarified which component of the exosome contributes to these important physiological effects. Previous studies have highlighted the importance of certain contents of exosomes for immune cell responses (6, 29). Exosomes can directly present Ag via MHC class II molecules on their surface to T cells (38, 39). Activation of NK cells can be mediated by NKG2D ligands and IL-15Rα on DC exosomes (40). Further, mRNAs and miRNAs can be transferred between cells via exosomes and modulate immune responses (7, 9, 12). These examples and more summarized by Théry et al. (29) and Robbins et al. (6) show that there are many potential molecules contained on or within exosomes that could be contributing to baseline phenotypes observed in the Rab27DKO mice.

Our experiments provide strong evidence that the transfer of exosomes containing miR-155, which is increased within exosomes in response to LPS (41), is important for proper responsiveness to LPS. However, although our experiments indicate that miR-155 itself is involved in this, it is important to note that other exosomal factors could be altered by the deletion of miR-155 in exosome donor cells. This possibility will be explored in future studies. Despite these possibilities, our data strongly support the hypothesis that exosomes are critical for limiting chronic inflammation and the range of phenotypes involved in this condition, as well as for enabling proper responses to inflammatory cues through an miR-155–dependent mechanism.

To further understand the role of endogenously produced exosomes in immune cell communication, novel reagents are required moving forward. The development of additional ways to specifically decrease exosome production in vivo would be helpful to see whether the phenotypes observed in a different model of defective exosome production mimic the phenotypes seen in Rab27DKO mice. A reagent where the production of miRNA-containing exosomes can be specifically blocked in vivo is needed to specifically delve into the role of miRNAs within exosomes. As we begin to identify factors involved in miRNA loading, such as Ybx1, hnRNPA2B1, and SYNCRIP (42–44), novel mouse strains can be created to address these questions. Furthermore, the production of exosomes containing only the miRNA of interest would make a strong tool both for further understanding of that miRNA within exosomes and for therapeutic applications down the road.

Our study implicates exosomes as key regulators of chronic inflammation and presents a valuable model that can be used to better understand the role of exosomes in this condition. Chronic inflammation and associated diseases pose an enormous economic and medical burden, making it essential to understand the underlying mechanisms of these disorders. Due to our observations that exosome delivery could rescue some aspects of chronic inflammation and responsiveness to LPS, we propose that exosomes could potentially be used in a therapeutic manner to promote these outcomes clinically. In addition, the differential response of WT versus miR-155^−/− exosome treatment suggests that miRNAs can impact the function of the exosomes and their ability to alter response to LPS. Therefore, it stands to reason that the miRNA contents of exosomes could be manipulated to alter the therapeutic outcome desired during exosome treatment.

The authors have no financial conflicts of interest.