MiR-7e-5p downregulation promotes malignant transformation of follicular lymphoma by modulating an immunosuppressive stroma through the upregulation of FasL in M1 macrophages

In follicular lymphoma (FL), histologic transformation to high-grade FL and diffuse large B-cell lymphoma (DLBCL) is a critical adverse step in disease progression. Activation of the oncogene c-MYC and tumor microenvironment remodeling account for FL progression. A panel of miroRNA (miRNA) was downregulated in transformed FL. Differentially expressed miRNAs were systematically analyzed in tissue samples from patients at different stages of disease. In FL cells, transcriptional regulation of the oncogene c-MYC on its target miR-7e-5p was revealed by Chromatin Immunoprecipitation (ChIP) assay. Exosome, carrying differentially expressed miR-7e-5p was isolated and visualized by transmission electron microscope and uorescence tracing. The effect of miR-7e-5p on recipient macrophage was determined by target gene quantication, ow cytometry, and TUNEL method in a cocultured system with miR-7e-5p-mimics or inhibitor treatment. Expression of miR-7e-5p targets, macrophage proportions, and clinical parameters were included for correlation analysis.


Abstract Background
In follicular lymphoma (FL), histologic transformation to high-grade FL and diffuse large B-cell lymphoma (DLBCL) is a critical adverse step in disease progression. Activation of the oncogene c-MYC and tumor microenvironment remodeling account for FL progression. A panel of miroRNA (miRNA) was downregulated in transformed FL.

Methods
Differentially expressed miRNAs were systematically analyzed in tissue samples from patients at different stages of disease. In FL cells, transcriptional regulation of the oncogene c-MYC on its target miR-7e-5p was revealed by Chromatin Immunoprecipitation (ChIP) assay. Exosome, carrying differentially expressed miR-7e-5p was isolated and visualized by transmission electron microscope and uorescence tracing. The effect of miR-7e-5p on recipient macrophage was determined by target gene quanti cation, ow cytometry, and TUNEL method in a cocultured system with miR-7e-5p-mimics or inhibitor treatment.
Expression of miR-7e-5p targets, macrophage proportions, and clinical parameters were included for correlation analysis.

Results
We determined that downregulation of miR-7e-5p, driven by c-MYC overexpression, was associated with poorer prognosis in FL patients. The decreased expression of miR-7e-5p in lymphoma cells led to a reduced exosomal transfer to surrounding macrophages. As a result, the target gene of miR-7e-5p, Fas ligand (FasL), was upregulated and activated the caspase signaling, which led to the apoptosis of M1 macrophages in tumor stroma. Finally, in transformed FL tissues, overexpression of FasL and activation of caspase proteins was detected in tumor stromal macrophages. Downregulation of miR-7e-5p was associated with poorer clinical outcomes.

Conclusion
Downregulation of exosomal miR-7e-5p induces stromal M1 macrophage apoptosis, which leads to immunosurveillance and transformation of FL.
Background Follicular lymphoma (FL), an indolent B-cell lymphoma, can histologically transform into high-grade FL and a highly aggressive malignancy, diffuse large B-cell lymphoma (DLBCL), under certain conditions. The malignant transformation of FL is considered an adverse step for patients and is associated with a much poorer clinical course and outcomes (1). Transformation has been found to occur in approximately 30% of patients at 10 years, which is associated with a decline in survival to less than two years (1). For this reason, understanding the pathogenesis of FL transformation is helpful for the prediction and treatment of FLs. Accumulation of genetic alterations has been found to be important for FL transformation. For example, the activation of the oncogene c-MYC drives a highly proliferative signature, which is frequently found in high-grade transformed FL and DLBCL (2).
In addition to genetic alterations, the tumor microenvironment is another important aspect that affects FL progression (3). Decreased subpopulations of CD4/CD8 T cells, macrophages and dendritic cells in patients are associated with FL transformation and are predictors of worse survival (3)(4)(5). In a large cohort of 197 FL patients, elevated levels of programmed death (PD)-1-positive cells and the CD8/CD4 ratio were associated with activated immune surveillance and correlated with a better clinical outcome (4). A more speci c subgrouping of PD-1-positive cells, which further separates them into T follicular helper (TFH) cells and exhausted T cells, enables the prediction of acute progressive FL according to the ratio between TFH and exhausted T cells (3). These population-based clinical data provide reliable evidence that the homeostasis of the immune microenvironment strongly in uences the fate of FL.
Recently, exosomes have been found to act as essential mediators in the maintenance and modulation of the tumor microenvironment (6). One of the active substances transported by exosomes is microRNA (miRNA), which is a single-stranded RNA that ne tunes the expression of target genes at the posttranscriptional level (7). Being transported by exosomes, tumor-derived miRNAs are delivered to neighboring immune cells and suppress their cytotoxic activity against tumor cells (7). In addition to tumor parenchyma, miRNAs can also be produced by stromal cells, which in turn affects tumor growth and invasion (8). In a mouse model, miR-298-5p from activated CD8-positive T cells induced the apoptosis of mesenchymal stem cells, which prevented the formation of mesenchymal tumor stroma and tumor cell metastasis (8).
Exosomal miRNAs act as important mediators that maintain the tumor microenvironment and are regarded as potential therapeutic methods for different tumors (7,8). In the context of FL, a panel of miRNAs that includes miR-7e, miR-30a, and miR-199a has been found to be associated with disease progression (9,10). Speci cally, miR-150 negatively regulates forkhead box transcription factor (FOX)-P1, which inhibits the progression of lymphoid malignancies, while miR-31 targets E2F transcription factor 2 and class II PI3 kinases (PIK3C2A) and inhibits tumor cell proliferation in FL (9,10). Although these studies reveal the tumor suppressive role of miRNAs, little is known about their function in modulating the tumor microenvironment. Given the diverse functions of miRNAs observed in other tumor types, certain miRNAs may contribute to the maintenance of microenvironmental homeostasis, which prevents the progression of FL.
In this study, we screened for miRNA candidates in patient tissues that play important roles in FL transformation. We identi ed miR-7e-5p, which was downregulated from indolent FL to transformed FL (tFL) and DLBCL. By using a coculture system and exosome isolation, we detected the transfer of miR-7e-5p from tumor cells to surrounding M1 macrophages. Repressed expression of miR-7e-5p led to the upregulation of FasL and activation of caspase signaling, which accounted for the apoptosis of M1 macrophages. Clinical data demonstrated the correlation of miR-7e-5p downregulation, upregulation of downstream targets and clinicopathologic features in FL transformation.
Mimic for miR-7e-5p (Supplemental Table 2, Guangzhou RiboBio, Guangzhou, China) were synthesized and transfected into both WSU-NHL and A20 cells using Lipofectamine® 3000 (Life Technologies) according to the manufacturer's protocol with a media change 6 hours after transfection. Smallinterfering RNAs (siRNAs) for miR-7e-5p (Supplemental Table 2 Macrophages from mouse were isolated at the age of 6 weeks. Mice were injected with 1 ml liquid para n intraperitoneally three days before isolation according to Ray's method (11). At the time of isolation, mice were euthanized and sterilized with 70% ethanol. It was immediately followed by the intraperitoneally injection of 5 ml cold RPMI using a 27g needle. After gently massage the peritoneum, ascitic uid was collected using a 25g needle. Peritoneum was opened by incision and the remaining uid was collected by Pasteur pipette. Cell suspension was washed once with cold PBS and treated with RBC lysis buffer (C3702, Beyotime, Shanghai, China) for one minute to get rid of erythrocytes. Three hours after seeding, the adherent cells were gently washed once with PBS and cultured in full RPMI medium for further analysis. Enriched macrophages were identi ed and counted by ow cytometry (CytoFLEX, Beckman Coulter, CA, USA).

Real-time PCR and Western blotting
Isolation of miRNA from cultured cells was performed by using the HiPure Universl miRNA Kit (Magen biotechnology, Guangzhou, China). Total RNA was extracted from cells by 1ml TRIZOL and miRNA were isolated according to the manufacturer's instructions. To isolate miRNA from human tissues, 10μm para n embedded tissues was depara nized by using xylene and followed by the isolation of miRNA using the MagMAX™ FFPE DNA/RNA Ultra Kit kit (A31811, Thermo Fisher Scienti c, Woodward St, Austin). Proteins from cultured cells was extracted by using the 10x Cell lysis buffer (Cell Signaling) and supplemented with 0.1mM PMSF and proteinase-inhibitor (5871, Cell Signaling Technology, MA, USA).
Reverse transcription for speci c miRNAs was performed using 2 μg miRNA and respective primers for reverse transcription (listed in supplementary table 2) according to the Stem-loop method (12).
Quantitative real-time PCR were carried out using the ABsolute qPCR SYBR Green ROX Mix (Thermo Fisher Scienti c) and the following cycling condition: 95°C for 10 min, 40 cycles of 95°C for 15 sec, and 60°C for 60 sec. StepOne Plus device (Thermo Fisher Scienti c) was used for the detection. The snRNA U6 was used for normalization.
For Western blotting, 30 μg total proteins measured by BCA assay (Beyotime, Shanghai, China) was separated in sodium dodecyl sulfate (SDS) polyacrylamide gel electrophoresis and transferred to a PVDF membrane (Merck Millipore,MA,USA). It was followed by blocking the membrane with 5% skim milk in TBST (Tris-buffered saline/0.1% Tween 20). Membranes were incubated with primary antibody at 4°C overnight and followed by the incubation of HRP-conjugated secondary antibody in skim milk (1:5000; Boster, Wuhan, China). Electronic chemical Laboratory (ECL) detection kit was used for the signal development (Merck Millipore).

Chromatin Immunoprecipitation (ChIP)
After transfection, 1x107 Cells in cultured T75 cell culture ask were collected and incubated with formaldehyde/PBS (1%) for 12 minutes to allow cross-linking of DNA and protein according to a ChIP protocol previously described (13). Protein in the complex was diluted to 1 mg/ml with RIPA buffer and pre-cleared with 30 µl of Dynabeads Protein G (Nanoeast Biotech, Jiangsu, China). Another tube of 50 µl Dynabeads Protein G beads was blocked with 15 µg sperm DNA and 50 µg BSA. Pre-cleared samples were incubated with blocked Dynabeads and 4 µg primary antibody at 4°C overnight to allow precipitation. The resulting immunocomplex-bound-beads were washed carefully with RIPA buffer and resuspended in TE buffer. Reversal crosslinking was achieved by incubation the complex in 4 M NaCl at 65°C for 5 hours. DNA was extracted and prepared for real-time PCRs analysis. Precipitated promoter fragments were normalized to a standard curve of genomic DNA. Primers binding to an unspeci c sequence of genomic DNA was used as a negative control.
All IHC stains of lymphoma tissues were de ned by two pathologist (H.N and LL.G) according to the intensity and percentage of positive tumor cells. Intensity was analyzed: 0=negative, 1=low nuclear stain, 2=medium nuclear stain, 3=strong nuclear stain. Percentage of positivity was de ned: 0=no positive cells, 1=less than 1%, 2=less than 10%, 3=10-50%, 4=more than 50%. Final scoring was multiplying of qualitative and quantitative parameters.

Data acquisition and statistical analysis
The data are presented as mean + SD. Protein quanti cation after Western blotting was achieved by using the SkanIt™ Software (Thermo Fisher Scienti c). Comparison of miRNA and proteins levels between nonpaired groups were done by the nonparametric Mann-Whitney U test. Comparison of miRNA expression between paired groups of patients relied on the t-test. Signi cance levels were as following: p≤0.05, p≤0.01, and p≤0.001 (IBM SPSS Statistics 19.0, Armonk, NY, USA). Experiments were repeated three times for statistical analysis.

Results
1. miR-7e-5p is downregulated in FL progression to high-grade FL and DLBCL To identify miRNAs that may affect the transformation of indolent FL to high-grade FL and DLBCL, six miRNAs (miR-30a, miR-31, miR-20c, miR-501, miR-199a and miR-7e-5p) previously shown to be dramatically downregulated in FL transformation were selected as our candidates (10). The expression of miRNAs was analyzed in FL (Grade 1 and 2, 3A n=4), tFL (Grade 3B, n=3) and DLBCL (n=4) tumor tissues. All six of these miRNAs were downregulated in DLBCL (fold change=0.36, P 0.01; Figure 1A) compared with low-grade FLs. miR-7e-5p was even downregulated in both tFL(fold change=0.45, P 0.05; Figure 1A) and DLBCL (fold change=0.34, P 0.05), which suggests its role in the early stages of transformation. To validate the relevance of miR-7e-5p in the progression of lymphoma, we increased the number of samples from different stages (FL: n=14, tFL: n=16, and DLBCL: n=16; Figure 1B), and miR-7e-5p was found to be signi cantly downregulated in DLBCL compared with FL (fold change=0.05, P 0.05) and tFL (fold change=0.26, P 0.05). However, the differential expression of miR-7e-5p was not detected in tFL compared with FL. To avoid heterogeneity among patients, we further analyzed the expression of miR-7e-5p in ve paired FL-tFL samples ( Figure 1C). In this cohort, miR-7e-5p was dramatically downregulated in the tFL samples compared with FL samples (fold change=0.19, P 0.001).
To further analyze the function of miR-7e-5p in FL progression in vitro, miRNA expression was compared in different lymphoma cell lines, namely, DLBCL cell lines (OCl-LY3, SUDLH-8, TMD-8, and Farage), FL cell lines (DoHH-2 and WSU-NHL), mantle cell lymphoma (JeKo-1) and Burkitt's lymphoma (Daudi). As expected, the levels of miR-7e-5p were much higher in the FL cell lines than in the DLBCL cell lines (fold change=4.89, P 0.001; Figure 1D). Taken together, these data revealed the potential role of miR-7e-5p in the suppression of the transformation of indolent FL to tFL and DLBCL.

miR-7e-5p is negatively regulated by c-MYC during FL transformation
Genetic ampli cation of the oncogene c-MYC has been found to be a critical step of FL transformation via the induction of tumor growth and chromosomal instability (14). Previously, c-MYC has been reported to regulate the expression of miR-7e-5p at the transcriptional level in cancer cell lines (15,16). We next analyzed the relevance of c-MYC regulating miR-7e-5p in the progression of lymphoma. In WSU-NHL cells, quantitative PCR (q-PCR) revealed a 2.57-fold increase in miR-7e-5p expression after the inhibition of c-MYC by siRNA (P 0.001; Figure 2A and Supporting Figure S1). To test whether c-MYC regulates miR-7e-5p expression through direct interaction during transcription, a potential binding site of c-MYC in the promoter region of miR-7e-5p ( Figure 2B; position -17 to -6) was predicted according to the JASPAR database (17). Q-PCR of chromatin immunoprecipitation eluates revealed ampli cation of the promoter region of pri-miR-7e, which was not detectable when c-MYC expression was inhibited by siRNA treatment. This indicates that miR-7e-5p is transcriptionally downregulated by c-MYC in WSU-NHL cells ( Figure 2B).
To con rm the relevance of c-MYC in the transformation of FL, c-MYC protein levels were detected in ve paired samples from available patient tissues ( Figure 2C). Tumor tissues from t-FL samples had strong c-MYC staining in the majority cells (>40% cell positivity), while FL samples had weak or negative staining (0-10% cell positivity; Figure 2C). Elevated expression of nuclear c-MYC was detected during FL transformation (P 0.001; Figure 2D). Taken together, these data demonstrate that miR-7e-5p is negatively regulated by c-MYC through direct binding to the miR-7e-5p promoter region during FL transformation.

Exosomes mediate the transfer of miR-7e-5p from tumor cells to macrophages
Remodeling of the tumor microenvironment has been shown to be an aspect as equally important as genetic modi cation in FL progression and is characterized by inactivation of cytotoxic T cells and macrophages and elevated levels of TFH cells (3)(4)(5). We hypothesized that miR-7e-5p produced by tumor cells changes the immune context of FL, which may facilitate the survival of tumor cells. Previously, researchers found that lymphoma cells can modulate the phagocytic activity of classically activated macrophages (also known as M1 macrophages) (18). To determine whether miR-7e-5p is responsible for the regulation of M1 macrophages, Cy3-labeled miR-7e-5p-mimics were transfected into the mouse lymphoma cell line A20, and the cells were then coincubated with mouse macrophages in a Transwell system with 0.4 μm pores. Flow cytometry analysis demonstrated that the majority (83%) of peritoneal macrophages were of the M1 phenotype (Supporting Figure S2). After 24 hours of coincubation, obvious red uorescence was detected in the cytoplasm of macrophages, demonstrating the transfer of miR-7e-5p-mimics into macrophages by lymphoma cell-derived particles ( Figure 3A).
Next, we evaluated whether exosomes were responsible for the transport of miR-7e-5p in the tumor microenvironment. Exosomes isolated from the supernatant of A20 cells were identi ed and visualized by both transmission electron microscopy ( Figure 3B) and NanoSight particle tracking analysis ( Figure 3C). After transfection of miR-7e-5p-mimics or inhibitor into A20 cells, exosomes carrying higher or lower levels of miR-7e-5p were collected and con rmed to express exosome markers (CD63 and CD81) by Western immunoblotting (Supporting Figure S3A). Real-time PCR detected increased levels of mature miR-7e-5p in macrophages treated with exosomes carrying miR-7e-5p-mimics (fold change=15, P 0.01), whereas a reduced level of mature miR-7e-5p was detected in macrophages when its expression in lymphoma cells was inhibited by miR-7e-5p inhibitor (fold change=0.05, P 0.001; Figure 3D and Supporting Figure S3B). Taken together, these results demonstrate that miR-7e-5p was delivered from lymphoma cells to recipient macrophages.

Exosomal miR-7e-5pinhibits FasL expression and inactivates the apoptotic pathway in stromal macrophages
To identify the target genes of miR-7e-5p, three miRNA target prediction databases (miRTarBase, TargetScan, and miRDB) were used for the identi cation of miR-7e-5p targets (Agarwal, Bell, Nam, & Bartel, 2015; Chou et al., 2018). Fifty-three potential target genes were found in all three databases, among which 21 had clear biological functions in cellular processes (19). We further analyzed the relevance of the potential target genes in leukemia and lymphoma according to the GCBI database. Seven genes were associated with patient prognosis, among which Fas ligand (FasL), N-myc protooncogene protein (MYCN), and insulin-like growth factor 1 receptor (IGF1R) were associated with cell death and apoptosis ( Figure 4A).
FasL is a strong apoptotic factor that binds to its receptors to activate the caspase cascade and regulate the activation of proteins involved in cell proliferation (Holler et al., 2003). To evaluate the effect of miR-7e-5p on FasL expression, macrophages isolated from mice were coincubated with A20 cells previously transfected with miR-7e-5p-mimics or inhibitor. With the presence of aclarubicin in cultured medium (2.7 μg/ml), Western blotting revealed higher levels of FasL and cleaved poly-ADP-ribosyltransferase (PARP) in the macrophages when miR-7e-5p production was inhibited in A20 cells, whereas decreased FasL and cleaved PARP were detected with miR-7e-5p-mimics treatment ( Figure 4B). We further evaluated the importance of exosomes in target gene regulation. Macrophages were incubated with exosomes isolated from A20 cells with differential expression of miR-7e-5p after miRNA-mimics or inhibitor treatments. Realtime PCR detected a downregulation of FasL in macrophages after incubation with exosomes from miR-7e-5p-mimic-expressing cells (fold change=3.03, P 0.01), whereas an upregulation of FasL was detected when macrophages were incubated with exosomes from miR-7e-5p inhibitor-treated cells (fold change=0.11, P 0.01; Figure 4C). Taken together, these results suggest that miR-7e-5p targets FasL and suppresses the activity of the caspase cascade in M1 macrophages.

Reduced exosomal miR-7e-5p induces apoptosis in stromal macrophages
Downregulation of exosomal miR-7e-5p led to the upregulation of FasL and activation of the apoptotic signaling pathway. To further con rm the effect of miR-7e-5p on cell apoptosis, macrophages were incubated with exosomes from differentially treated A20 cells in the presence of aclarubicin. Western blotting revealed a dramatic decrease in FasL, cleaved PARP and Caspase 3 when exosomal miR-7e-5p was blocked by GW4869 treatment ( Figure 5A). Flow cytometry detected fewer apoptotic macrophages in the group treated with exosomes from miR-7e-5p-mimics overexpressing cells compared with mock cells (59% and 62%), whereas signi cantly more apoptosis was detected when cells were treated with exosomes from miR-7e-5p-silenced cells (68%; Figure 5B). The effect of miR-7e-5p on the apoptosis of macrophages was abolished when the release of exosomes was blocked by GW4869 ( Figure 5B). Similarly, in situ TUNEL assays detected less apoptosis in macrophages incubated with a high level of exosomal miR-7e-5p-mimics, while more apoptosis was detected when cells were incubated with exosomes carrying less miR-7e-5p via inhibitor treatment ( Figure 5C). The effect of miR-7e-5p on macrophages was restored when exosome production was blocked by GW4869 treatment ( Figure 5C). These results demonstrated that reduced exosomal miR-7e-5p levels promoted the apoptosis of recipient macrophages.
. miR-7e-5p and its target genes correlate with FL patient progression Then, we tested whether the downregulation of miR-7e-5p and the upregulation of FasL were associated with the progression of FL. In the tumor areas, a loss of or weaker Fas expression and a clear extracellular FasL were detected in tFL samples ( Figure 6A). A downregulation of Fas and upregulation of FasL were observed during the disease transformation (FL: n=3, tFL: n=3, P 0.05; Figure 6B). In the tumor stroma, a decrease of M1 macrophages, an induction of M2 macrophages and an upregulation of miR-7e-5p downstream targets (FasL, Caspase 3 and Caspase 8) were detected in tFL ( Figure 6C). The proportion of M1 macrophages was signi cantly lower, and the expression of miR-7e-5p target genes was higher in tFL than in FL (FL: n=3, tFL: n=3 Figure 6D). Moreover, overexpression of miR-7e-5p was associated with unfavorable clinicopathological factors, including B symptoms and higher serum LDH levels (n=46, P 0.01; Table 1). These ndings demonstrate that the decreased expression of miR-7e-5p is associated with the induction of FasL and apoptotic signaling in stromal immune cells, which leads to poorer clinical outcomes in FL patients.

Discussion
The transformation of FL in 30% of patients to high-grade FL and DLBCL is widely accepted as an adverse event that dramatically reduces patient survival (1). The decreased expression of different miRNAs, including miR-31, miR-7e, miR-30a and miR-199a, is associated with rapid disease progression, and these miRNAs are regarded as potential tumor suppressors in FL transformation (9,10). However, the underlying mechanisms by which miRNAs inhibit the malignant transformation of FL remain unclear. In this study, we found that miR-7e-5p is downregulated in lymphoma cells during the transformation of FL, which is responsible for the remodeling of the tumor microenvironment via the induction of immune escape and tumor progression.
The oncogene c-MYC is located at chromosome 8q24, which is ampli ed and translocated in 5-15% of DLBCL (20). During the transformation of FL to high-grade FL and DLBCL, c-MYC is activated and consequently induces a subset of target genes involved in promoting cell growth and inhibiting apoptosis (14,21). Previously, let-7 family miRNAs were found to be widely repressed by c-MYC in B cell lymphoma (21). The transcriptional regulation of miR-7e-5p by c-MYC was identi ed by an in vitro a nity puri cation approach in HEK293 cells (16). In this study, we con rmed the direct binding of c-MYC to the miR-7e-5p promoter, which negatively regulates the transcription of miR-7e-5p and its tumor-suppressive activity in the transformation of FL. Therefore, the downregulation of miR-7e-5p during the transformation of FL to DLBCL can be explained by c-MYC ampli cation.
We demonstrated the downregulation of miR-7e-5p in tFL and DLBCL, which contributes to disease progression. The tumor suppressor role of miRNAs can be achieved by the direct inhibition of oncogenes in tumor cells and the regulation of the tumor cell microenvironment through exosome secretion (7,8).
Our results reveal the importance of exosomes in mediating the transfer of miRNA from tumor cells to stromal cells. In the Transwell coculture system, miR-7e-5p was transferred from mouse lymphoma cells to macrophages. In this context, downregulation of miR-7e-5p in lymphoma cells led to a reduced amount of miR-7e-5p in surrounding macrophages. Therefore, miR-7e-5p expressed in normal tissue and indolent FL may function as a tumor suppressor; however, downregulation of miR-7e-5p in the transformation of lymphoma reduces the amount of miR-7e-5p and its function in recipient macrophages.
Based on miRNA databases and enrichment analysis, FasL was selected as a potential target gene of miR-7e-5p. FasL belongs to the tumor necrosis factor (TNF) family, which interacts with and activates Fas receptor (Fas) and induces the caspase cascade during cellular apoptosis (22). By binding to its ligand, the intracellular death domain of Fas interact with each other and serve as a platform for the initiation of apoptosis through the enrichment of Caspase-8 protein (23). After dimerization, Caspase-8 undergoes autocatalytic cleavage and triggers the activation of Caspase-3 and Caspase-7, which affects the function of many genes (e.g., PARP) involved in cell proliferation and transformation (23). As expected, we detected an inhibition of FasL expression at both the transcriptional and protein levels in recipient macrophages with elevated exosomal miR-7e-5p levels in the stroma. Consequently, the cleavage of PARP was reduced in the coincubation experimental setup with miR-7e-5p-mimic overexpression. Interestingly, the reduction in cleaved PARP and cleaved Caspase-3 was more obvious in macrophages treated with isolated exomes containing a high level of miR-7e-5p. The negative regulation of macrophage apoptosis by exosomal miR-7e-5p was demonstrated by FACS analysis and in situ TUNEL assays. Overall, su cient exosomal miR-7e-5p facilitates the survival of macrophages. The decreased production of exosomal miR-7e-5p in lymphoma cells activates apoptotic signaling via its target FasL in stromal macrophages.
Although FasL is upregulated by the suppressive effect of miR-7e-5p during FL transformation, it probably cannot activate the apoptosis of lymphoma cells due to multiple immune escape mechanisms (3). Loss of Fas is present in 17% of FL and 51% of DLBCL, while the death domain of FAS is mutated in 6% of FL and 20% of DLBCL, which leads to inactivation of the Fas-mediated apoptotic signaling pathway (24). However, upregulation of FasL in macrophages according to our results induces immune suppression due to the Fas-FasL interaction. In addition, FasL was previously found to bind to and form a hexamer with cytotoxic T lymphocyte associated protein 4 (CTLA4), which abrogated costimulation by binding to B7 receptor-expressing immune cells (25). Thus, FasL induces a strong apoptotic signal not only by activating the caspase cascade but also by switching off antiapoptotic signaling in immune cells  Table S1-2.

Competing interests
The authors disclose no potential con icts of interest.   Transfer of miR-7e-5p from tumor cells to macrophages via exosomes. A) Scheme showing the transfer of miR-7e-5p-mimics from A20 cells to mouse macrophages. A20 cells were transfected with 50 nM Cy3-miR-7e-5p-mimics (red uorescence) and seeded in the upper chamber six hours after transfection. Red uorescence was detected in macrophages after coculturing with A20 cells for 24 hours. Scale bars: 50 μm. B) Images from transmission electron microscopy showed bilayer round-shaped particles with diameters ranging from 30 nm to 200 nm collected from the supernatant of A20 cells. Scale bars: 1 μm for lower magni cation and 2 μm for higher magni cation. C) NanoSight particle tracking analysis (NTA) of exosomes revealed that most exosomes had a diameter of approximately 100 to 200 nm. The concentration of vesicles was 2.4×108/ml at 150 nm. D) Real-time PCR indicated the expression of miR-7e-5p in macrophages after incubation with exosomes extracted from A20 cells after 24 hours. A20 cells were previously transfected with miR-7e-5p mimics (50 nM) or inhibitor (100 nM) for 24 hours. Cells treated with transfection reagent only served as a control (mock). Macrophages seeded on six-well plates were incubated with 1.2×107 exosomes for 24 hours. Statistical test: t-test.

Figure 4
Exosomal miR-7e-5p inhibits macrophage apoptosis by targeting FASL. A) Target genes of miR-7e-5p were predicted using miRTarBase, miRDB and TargetScan. Intersections of the results were used for functional analysis using the Panther and GCBI online databases. B) Western blotting detecting PARP, cleaved PARP and FASL levels in macrophages is shown. The macrophages in each well were cocultured with miR-7e-5p-mimics or inhibitor-treated A20 cells. Actin served as a control. Macrophages were pretreated with aclarubicin at 2.67 μg/mL. C) Real-time PCR analysis of FasL expression in macrophages. To each well of macrophages, 1.2×107 exosomes from miR-7e-5p mimics (30 nM, 50 nM, 100 nM) or inhibitor (50 nM, 75 nM, 100 nM)-treated A20 cells were added. Statistical test: t-test.

Figure 5
Treatment with exosomes in uences the apoptosis of macrophages. A) Western blotting detecting PARP, cleaved PARP, Caspase-3, cleaved Caspase-3 and FASL in macrophages treated with exosomes from A20 cells transfected with miR-7e-5p mimics (50 nM). GW4869 (10 μM) was added to A20 cells for 24 hours to inhibit exosome secretion. B) Flow cytometry was used to detect apoptotic macrophages by Annexin V-FITC and propidium iodide (PI) double staining. Macrophages were treated as before. C) Representative images of TUNEL assays to assess macrophage apoptosis. Macrophages were pretreated with for FasL, CD68, CD163, Caspase3 and Caspase 8 expression levels in macrophage cells in three FL and three tFL samples. Statistical test: t-test.

Figure 7
Downregulation of miR-7e-5p induces FL transformation via the activation of apoptotic signaling in macrophages. Scheme illustrating the mechanism the regulatory function of miR-7e-5p in recipient M1 macrophages. Downregulation of lymphoma-derived miR-7e-5p activates Fas-FasL interaction and downstream caspase signaling in macrophages.

Supplementary Files
This is a list of supplementary les associated with this preprint. Click to download.