Toll-like Receptor-4 Activation Boosts the Immunosuppressive Properties of Tumor Cells-derived Exosomes

The biology of tumor-derived exosomes (TEX) is only partially understood and much remains to be studied in order to define the effect that the tumor microenvironment or the activation of tumor cells exerts on their composition and functions. Increased expression and activity of toll-like receptor 4 (TLR4) in chronic infectious and inflammatory conditions is related with cancer progression: its activation induces an inflammatory signaling that increases the tumorigenic potential of cancer cells promoting their immune evasion. We investigated the immune modulatory properties of TEX released upon cell TLR4 activation, and we found that, although differences were observed depending on the type of the tumor, the treatment influences TEX composition and boosts their immunosuppressive ability. Our results suggest that the activation of TLR4 supports tumor progression by stimulating the release of more effective immunosuppressive exosomes, which allow tumor cells to escape immune surveillance and probably even play a role in the metastatic process.


Isolation of monocytes and T cell population.
For the in vitro experiments, samples of whole blood from healthy donors were collected in EDTA-tubes by the Department of Transfusion Medicine (University Hospital of Udine). All healthy donors gave their informed consent to this study according to the Declaration of Helsinki and to the Italian legislation (Authorization of the Privacy Guarantor No. 9, 12 th of December 2013). Only anonymized leftovers samples from routine clinical practice were used and their use was not subjected to ethics review, according to the International Standard ISO 15189 "Medical Laboratories. Particular requirements for quality and Competence, 2nd Ed. 2007" and Food and Drug Administration OBM Control No 0910-0582 "Guidance of informed consent for in vitro diagnostic device studies using leftover human specimen that are not individually identified".
Treg cell induction. Isolated CD4+ T lymphocytes were incubated with 5 × 10 8 exosomes and stimulated with anti-CD3 and anti-CD28 in the concentrations described above. The recombinant IL2 at the concentration of 250U/ml was also added as a stimulus. After 3 days, the percentage of FoxP3+ regulatory T cells in the total population of stimulated CD4+ cells, was determined by CD25 surface staining and FoxP3 intracellular staining using Foxp3/Transcription Factor Staining Buffer Set (eBiosciences) 20 . In a separate set of experiments, SW480-, SW620-and U87-MG-derived exosomes were pre-incubated for 15 min at 37 °C with 0.5 µg/ml of anti-TGFβ1 neutralizing antibody (R&D system) before adding to cells.
NKG2D expression levels. Resting PBMCs or purified NK cells obtained from healthy donors were co-incubated with or without 5 × 10 8 exosomes for 24 hours, then cells were stained with anti NKG2D APC and anti-CD8 FITC or anti-CD16 FITC/anti-CD56 PE (all from eBiosciences) and tested by flow cytometry. In a separate set of experiments, SW480-and SW620-derived exosomes were pre-incubated for 15 min at 37 °C with 5 µg/ml of anti-MICA/B neutralizing antibody (R&D system) before adding to cells.
Real-time reaction was performed on the Applied Biosystems QuantStudio 3 System. MiRNA relative concentrations were normalized using relative standard curve method obtained by serial dilutions of cel-miR39 (1 nM-100 fM) 19 .
Statistical analysis. The data are expressed as mean ± standard deviation. Statistical analysis has been performed using the GraphPad Software (version 7). Data were tested for normal distribution using the Kolmogorov-Smirnov test. Numerical data were processed for significance by one-way analysis of variance followed by the Bonferroni or Dunnett post test. A value of P-value of 0.005 was considered significant.

Effect of TLR4 activation by LPS on tumor cell.
To evaluate whether TLR4 activation may affect morphology, proliferation and viability of tumor cells, cells were cultured for 24 hours in the presence of 1 µg/ml LPS, a TLR4 specific ligand. Results are reported in Fig. 1A through 1D. We show that the treatment with LPS upregulate the expression of TLR4, as verified by immunoblot analysis of cell homogenates (Fig. 1A). We observed that the treatment did not induce morphological changes (Fig. 1B), while the cell proliferation was decreased by about 25% in all the analyzed cell lines (Fig. 1C). Moreover, we observed that cell viability was not affected by treatment with LPS (Fig. 1D).

Validation of exosomes released by tumor cells upon TLR4 activation by LPS. Tumor cells were
pre-treated with LPS, then an enriched fraction of exosomes was isolated from the cells supernatants by polymer precipitation method.
As shown in Fig. 2A and Supplementary Fig. 1A, exosomes and cell lysates expressed the specific exosomal markers CD63, TSG101, CD81 and CD9, while no signal was detected with the Exoquick-derived supernatants, which were used as negative controls. To exclude any possible contaminations in our exosomes preparations, we also evaluated the expression of proteins associated with subcellular compartments. Our data showed lack of calnexin and GRP94 (endoplasmic reticulum proteins), indicating a successful enrichment. The expression of exosomal markers was further confirmed by flow cytometric analysis of exosomes coupled to Exo-Flow beads. As shown in Fig. 2B, beads bound to exosomes expressed high levels of CD81, CD9 and CD63 and reacted with the ExoFITC reagent. It is interesting to note that the expression profile of CD9 highlighted the presence of three distinct exosomes subpopulations and its expression did not change after the treatment with LPS (Fig. 2C).
The concentration of exosomes released by tumor cells was determined measuring the activity of AChE with the Exocet kit (Fig. 2D) and by Nanoparticle tracking analysis (NTA) (Fig. 2E and Supplementary Fig. 1B). These www.nature.com/scientificreports www.nature.com/scientificreports/ quantification methods provided different absolute values, probably due to the different principle on which the respective analyzes are based. Specifically, NTA analysis reported a concentration of about 100 times higher than that measured with the Exocet kit. However, both quantification methods showed that cellular activation with LPS by stimulation of TLR4 did not influence the number of exosomes released, although a slight non-significant increase, was observed for exosomes produced by U87-MG. It is interesting to report that untreated MDA-MB-231 cells released the highest number of exosomes. The number of exosomes used as treatment for subsequent experiments was quantified by Exocet.
As reported in the dimensional profile of the representative NTA analysis (Fig. 2F), tumor exosomes are detected as nanoparticles with dimensions comparable to those described for physiological exosomes (range from 50 to 120 nm in diameter). Specifically, the average size of the collected vesicles was 113 ± 21.1 nm for SW480-derived exosomes, 81.5 ± 19.7 for SW620-derived exosomes, 86.4 ± 15.1 for MDA-MB-231-derived exosomes and 129 ± 16.1 for U87-MG. Finally, the size of exosomes was not influenced by LPS treatment of producing cells (Fig. 2G).

Exosomes released upon LPS treatment inhibit T cell proliferation in a dose-dependent manner.
To determinate whether activation of TLR4 expressed on tumor cells influences the immune-modulatory properties of released exosomes, CFSE-labelled PBMCs, isolated from healthy donors, were incubated for 24 hours with scalar doses (from 5 × 10 6 to 5 × 10 9 ) of exosomes and then stimulated with anti-CD3/-CD28. We found that only exosomes released upon LPS treatment, up to 5 × 10 8 , significantly inhibited CD3+ T cell proliferation in a dose-dependent manner. In all other cases, only the highest dose (5 × 10 9 ) of exosomes released by the untreated cells was able to inhibit the proliferation of CD3+ T cells. With lower doses, the inhibitory effect of the exosomes released by the unstimulated cells disappeared, while a stimulation effect on T cell proliferation appeared (Fig. 3A,B).  www.nature.com/scientificreports www.nature.com/scientificreports/ In order to evaluate the effect of the exosomes on the proliferation of different T cell subtypes, the PBMCs were incubated with 5 × 10 8 exosomes and labeled for the expression of CD4 and CD8. As the dose of exosomes to use in the experiments, the smallest dose able to induce a different effect on the proliferation of T lymphocytes between control and LPS-treated exosomes was chosen, based on previous results.
We found that only exosomes released after LPS treatment were able to significantly inhibit the proliferation of both, CD4+ (

Internalization of exosomes by immune cells. To investigate cellular interactions responsible for the
inhibitor effect of exosomes on T cell proliferation, we stained the vesicles with DiD dye and monitored their uptake by T cells and monocytes isolated from PBMCs. As reported in Fig. 4A, CD14+ monocytes quickly and efficiently internalized DiD-labelled exosomes. After 6 hours, about 90% of monocytes were positive for DiD (data not shown) and their mean fluorescence intensity (MFI) increased in a time-depend manner (Fig. 4B). Surprisingly, CD3+ T cells slightly internalized tumor exosomes, even after 48 hours of co-incubation (Fig. 4C). No differences were observed between exosomes released by the different cell lines (data not shown). Exosomes released by cells after LPS stimulation were internalized with the same kinetics observed for control exosomes. Our results suggested that the internalization of tumor-derived exosomes was not necessary for deliver the signals that result in inhibition of T cells proliferation. They also suggest that surface contact might be sufficient for inducing the changes observed. SW480-and U87-MG-derived exosomes induce the release of IL-1α (Fig. 5A), IL-8 (Fig. 5E) and MCP-1 (Fig. 5G) by monocytes, while exosomes released from all tumor cell lines after LPS treatment trigger the www.nature.com/scientificreports www.nature.com/scientificreports/ production of IL1β (Fig. 5B), IL-10 (Fig. 5C), IL-6 ( Fig. 5D), MIP-1α (Fig. 5H), MIP-1β (Fig. 5I) and TNF-α (Fig. 5L) cytokines. The release of slCAM was increased in monocytes treated with LPS-stimulated SW480-, SW620-and U87-MG-derived exosomes (Fig. 5K). On the contrary, MDA-MB-231 and U87-MG-derived exosomes and SW480-and SW620-derived exosomes released after LPS treatment inhibited the production of IP-10 (Fig. 5F).
Finally, to ensure that the inflammatory proteins measured were not contaminants derived from the isolation process, their presence was explored directly in purified exosomes. The results indicate that all analytes were below the level of detection or expressed at low levels respect the concentration measured on monocytes supernatants (data not shown).

Expression of TGFβ on exosomes surface and its effect on regulatory T cells expansion.
To investigate the molecular mechanisms underlying the inhibitory effects of exosomes released by tumor cells upon TLR4 activation by LPS, we examined whether TGFβ was present on exosome surface, as this cytokine is strongly implicated in mechanisms of immune evasion and may be responsible for the antiproliferative effect www.nature.com/scientificreports www.nature.com/scientificreports/ observed in our study. We evaluated the expression of the three different isoforms of TGFβ by multiplex assay, finding that only TGFβ1 and TGFβ2 were measurable in our exosomal samples, while TGFβ3 was undetectable (data not shown). Specifically, TGFβ1 was expressed at a very low level on the surface of exosomes released by MDA-MB-231 as compared to the other cells lines (Fig. 6A). Interestingly, we found that the expression of TGFβ1 was increased in exosomes released by SW480, SW620 and U87-MG cells upon TLR4 activation by LPS, while no differences was observed for MDA-MB-231-derived exosomes. The expression of TGFβ2 on exosome surface was not influenced by cell TLR4 activation by LPS (Fig. 6B).
The expression of TGFβ1 on the exosomes surface was also confirmed by FACS analysis of exosomes coupled to Exo-Flow beads (Fig. 6C).
In addition, we hypothesized that TGFβ1 expressed on the surface of exosomes could also induce the conversion of CD4+ CD25− T cells into Foxp3-expressing regulatory T cells. Stimulation by exosomes secreted by untreated SW480 and U87-MG increased the percentage of CD4+/CD25+/Foxp3+ cells in the PBMC population (Fig. 6D,E) and this stimulatory effect becomes statistically more significant with exosomes released by cells upon TLR4 activation. Interestingly, only exosomes produced by SW620 cells upon TLR4 activation become capable of inducing conversion to regulatory T cells. Moreover, the pre-incubation of exosomes released by SW480, SW620 and U87-MG LPS-activated cells with neutralizing anti-TGFβ1 antibody resulted in a decrease of the percentage of CD4+/CD25+/Foxp3+ cells, suggesting that in vitro induction of regulatory T cells by exosomes is mediated by TGFβ1 (Fig. 6F).

Expression of NKG2D ligands on the surface of exosomes and modulation of NKG2D expression on NK and CD8+ T cells. There is mounting evidence showing that tumor cells escape immune
surveillance by the release of soluble NKG2D ligands, which trigger a general downregulation of the NKG2D receptor on NK cells and CD8+ T cells 21 . The expression of MICA/B and ULBP-1, a NKG2D ligands, on the surface of exosomes coupled to Exo-Flow beads was evaluated by FACS analysis. As shown in the representative plot of the FACS analysis, tumor exosomes expressed, even if at low levels, MICA/B (Fig. 7A,B), but not ULPB-1 (Fig. 7C). Interestingly, the expression of MICA/B was increased following activation of TLR4 with LPS only in the exosomes released by SW480 and SW620. Moreover, they were able to reduce the expression of NKG2D in CD8+ T lymphocytes (Fig. 7D). The pre-incubation with neutralizing anti-MICA antibody of exosomes released by LPS-activated SW480 and SW620 cells resulted in an increased expression of NKG2D in CD8+ cells, although this effect was statistically significant only for SW480-derived exosomes (Fig. 7E). In contrast, the expression of NKG2D in purified NK cells was slightly reduced by exosomes released from all tumor cell lines, but no further effect was observed after activation of TLR4 with LPS. (Fig. 7F).

Expression of miR-21, miR-34a and miR-155 in tumor-derived exosomes. Several specific miR-
NAs have recently emerged as important regulators of immune-cell function in the context of different solid tumors 22 . Among these, miR-21, miR-155, and miR-34a are mediators of immune suppression, as they promote the expression of immunosuppressive factors. To analyze the expression of these miRNAs, total RNA from tumor-derived exosomes released by cells upon TLR4 activation was extracted. The expression of miRNA-21 ( Fig. 8A) was significantly higher in exosomes produced by SW480 pre-treated with LPS compared to those of untreated cells, while no difference was observed for exosomes produced by other cell lines. On the other hand, we found miR-34a only in exosomes produced by U87-MG and its expression was significantly decreased after treatment with LPS (Fig. 8B). Finally, exosomes produced by all the cell lines analyzed were did not show expression of the miRNA-155 (data not shown).

Discussion
It has been well established that the immune system plays an important role in monitoring tumor development and progression. However, a number of mechanisms have been discovered that allow tumors to escape from the host immune defenses by a process known as "immune evasion". Among various molecular and/or genetic changes that tumors induce in the immune cells, the production of exosomes is emerging as novel mechanism of immune suppression 23 . Specifically, it has been described that tumor-derived-exosomes (TEX) impair dendritic cell differentiation and maturation, decrease the cell cytotoxicity of NK cells, push the differentiation of myeloid precursor cells toward MDSCs, reduce T cells response and induce regulatory T and B cells 3,8 .
Until now, many studies have been directed to investigate the mechanisms by which tumor-derived exosomes are able to support tumor growth but little is known about the effect that the microenvironment or the activation of cancer cells exert on the composition and functionality of TEXs released in the environment. It has been proposed that the content and number of exosomes produced may change if the cells are exposed to different stressors or stimuli, and that exosomes produced by the same cells can contain different constituents. A recent study has shown that exosomes mediate hypoxia-dependent intercellular signaling of glioblastoma multiforme, a highly malignant brain tumor, since the profile of the proteome and mRNA species found in the vesicles closely reflect the oxygenation status of glioma cells in patients' tumors 24 .
To further investigate those functions, we evaluated the effect of toll-like receptor 4 (TLR4) activation on the immune modulatory properties of exosomes released by tumor cells. TLR4, belongs to the family of pattern recognition receptors (PRRs), highly conserved receptors that recognize pathogen-associated molecular patterns (PAMPs) and damage-associated molecular pattern (DAMPs), thus representing the first line of defense against infections 25 . Accumulating evidence demonstrated that TLRs may promote carcinogenesis through pro-inflammatory, anti-apoptotic, proliferative and profibrogenic signals in either the tumor microenvironment or tumor cells themselves 26 . These effects can be either induced directly in TLR-expressing target cells, or mediated by TLR-induced cytokines 26 . Specifically, tumor-promoting functions of TLR4 are best established in the colon 27 , liver 28 , breast 14 , pancreas 29 and skin 30 cancers. Likewise, it has been reported that TLR4 is expressed on (2019) 9:8457 | https://doi.org/10.1038/s41598-019-44949-y www.nature.com/scientificreports www.nature.com/scientificreports/  www.nature.com/scientificreports www.nature.com/scientificreports/ Our data suggest that the activation of TLR4 on tumor cells did not influence the number or size of released TEXs, instead, it increased their immunosuppressive potential. As already reported in the literature 37,38 , we found that exosomes secreted by untreated tumor cells are able to inhibit T cells proliferation, however the effect occurs only with high-dose treatments, while low doses stimulate the proliferation of T cells. Instead, exosomes produced by tumor cells after TLR4 activation were always immunosuppressive, suggesting that the treatment influences TEX cargo. The molecular profile show that TEX may express a dual capability of mediating either immune suppression or immune stimulation, presumably depending upon the microenvironment 3 . The immunostimulatory properties of TEXs are probably due to the presence of tumor-associated antigens (TAA) and costimulatory molecules on TEX surface 9 , which stimulate the immune response. It has been reported that exosomes produced www.nature.com/scientificreports www.nature.com/scientificreports/ by lymphoma cells after a heat-shock contained high levels of Hsp70 and were reported to stimulate a direct Th1-polarized immune response in a MHC-independent manner in autologous and allogeneic murine models 39 .
Upon secretion, exosomes are able to fuse with the plasma membranes of recipient cells to deliver their content into the cytoplasm. Alternatively, proteins present on their surface can bind cell surface receptors on target cells and influence intracellular signaling 40 .
We reported that tumor-derived exosomes were efficiently internalized by CD14+ monocytes and were able to modulate their production of inflammatory proteins. Specifically, exosomes produced by tumor cells after TLR4 activation strongly induce the release of pro-inflammatory cytokines and chemokines in monocytes. It is widely recognized that abnormalities in cytokine expression are implicated in tumorigenesis as they promote growth, reduce apoptosis and facilitate invasion and metastasis of cancer cells 41,42 . Moreover, it has been reported that tumor-derived EVs induce in primary monocytes an activated phenotype, which is also observed in tumor-associated macrophages 43 .
We observed that while monocytes rapidly take up exosomes, T cells did not internalize TEXs, even after 48 hours of co-incubation. According with other studies in which the Authors show that T lymphocytes, unlike other mononuclear cells, do not internalize exosomes 44 , our data suggest that TEXs deliver signals to receptors present on the T cell surface. It has been previously proposed that exosomes produced by cancer cells may induce regulatory T cells growth through the TGF-β1 expressed on their surface 45,46 . Our data demonstrate that TLR4 activation increases the expression of TGFβ1 on TEX surface and promotes the expansion of regulatory T cells, which in turn may be responsible for the reported antiproliferative effects of TEXs on T cells. We also found that exosomes secreted by the breast cancer cell line MDA-MB-231, expressed TGF-β1 at a very low level, as compared to exosomes released by the other cells lines, even after TLR4 cell activation, and fail to induce regulatory T cells.
The production of NKG2D-ligand-bearing exosomes is a newly described mechanism for cancer cell immune evasion 21,47,48 . We found that exosomes released by all cell lines analyzed express on their surface MICA/B, a NKG2D ligand, although only in colorectal cancer cells-derived exosomes TLR4 activation increased its expression. It is therefore possible to hypothesize that colorectal cancer cells are more sensitive to the activation of TLR4. In fact, a correlation between the chronic activation of TLR4 and the progress of CRC through the release of immunosuppressive factors which promote tumor escape has been widely shown 24,37,38 .
Increasing evidences suggest that exosomal microRNAs transported by cancer cells can be affect other cells in the local microenvironment leading to reprogramming of the target cell transcriptome and influencing cancer progression in a paracrine manner 49 . MiR-21 and miR-155 are well-characterized oncomiRs that promote both, cancer development and metastasis by targeting numerous mRNAs. In contrast, miR-34a has been shown to suppress cancer growth and metastasis by inducing apoptosis, cell cycle arrest and senescence. We found that the expression of miR-21 was increased in SW480-derived exosomes after TLR4 activation. It has been proposed that in the early stage of inflammation, pri-miR-21 exerts pro-inflammatory effects and polarizes macrophages towards the pro inflammatory M1 phenotype 50 . Moreover, miR-21 present in cancer cell-secreted exosomes can be transferred to surrounding immune cells and can bind to endosomal Toll-like receptors. By binding to TLRs, it induces cytokines secretion by the immune cells, leading to a prometastatic inflammatory response that ultimately may lead to cancer growth and metastasis 51 .
In conclusion, the present study demonstrates that the activation of TLR4 contributes to cancer progression by promoting the release of more effective immunosuppressive exosomes, which allow tumor cells to escape immune surveillance. As a matter of fact, exosomes released by tumor cells after LPS treatment can inhibit the proliferation of T lymphocytes more effectively. Further investigations will be needed to identify the common mechanisms underlying the observed effects.