Fig 2: Signaling pathways for HCC cell proliferation and apoptosis affected by NK-exo. A: Hep3B cells were co-cultured with increased doses (100 and 200 µg) of NK-exo for 10 h. The expressions of specific apoptosis markers (caspase-3, -7, -8, -9, cytochrome c, and PARP) and cytotoxic proteins (perforin, granzyme B, FasL, and TRAIL) were analyzed using western blotting. Caspase activation occurs upon cleavage. B and C: The expression levels were quantified with ImageJ software, and normalized to that of ß-actin. D: Hep3B cells were co-cultured with increased doses (100 and 200 µg) of NK-exo for 10 h. The phosphorylation of serine/threonine kinases (p-AKT and p-ERK1/2) was analyzed using western blotting. E: The histograms were quantified with ImageJ software, and the protein expressions were normalized to that of total proteins. F and H: Hep3B cells were co-cultured with NK-exo (200 µg), AKT inhibitor (124005, 20 µM), and ERK1/2 inhibitor (U0126, 20 µM) for 10 h. The inhibition of p-AKT at Ser473 (F) and p-ERK1/2 at Thr202/Tyr204 (H) were analyzed using western blotting. G and I: The histograms were quantified with ImageJ software, and the protein expressions were normalized to that of total proteins. All data are expressed as the mean ± S.D. of at least three experiments. *p < .05, **p < .01, and ***p < .001 vs. untreated Hep3B cells. CD63, as an exosomal marker and ß-actin, as a loading control were used.

Fig 3: In vivo anti-tumor effects of NK-exo in an orthotopic HCC xenograft mouse model. A: Schematic depiction of the experimental protocol for the evaluation of NK-exo targeting ability in orthotopic xenograft mouse models of HCC. 50 μL of Hep3B cells (2 × 106 cells) were inoculated directly into the livers of mice. After 2 weeks, the HCC-bearing mice were administered PBS (control) or different doses of NK-exo via intravenous injection, once every 2 days for 12 days. The HCC-bearing mice were sacrificed and then the anti-tumor effects in each group were assessed. B: Representative images of a dissected liver from HCC-bearing mice treated with PBS (control) or NK-exo (50, 100, 200, and 500 μg) are shown. The yellow-dotted circles indicate tumors. C–E: Changes in body weight (C), tumor weight (D), and tumor inhibition rate (E) of HCC-bearing mice treated with PBS (control) or NK-exo (50, 100, 200, and 500 μg) are shown. Body weight was measured until 12 days. Tumor weight was determined after being harvested from the end of experiment. The tumor inhibition rate was calculated as: [1 – (TWNK-exo treated/TWPBS control)] × 100 (%). All data are presented as mean ± S.D. of at least three independent experiments. *p < .05, **p < .01, and N.S. (no significance) vs. PBS control group. F: Representative histologic images of H&E staining (top panel) and IHC staining (bottom panel) against apoptosis marker cleaved PARP in tumor tissues of HCC-bearing mice treated with PBS (control) or NK-exo (50, 100, 200, and 500 μg). Yellow-dotted lines indicate the boundary of normal tissues (N) and tumor tissues (T). Scale bars: 200 μm (top panel) and 50 μm (bottom panel). G: The apoptosis signaling pathway in tumor tissues of HCC-bearing mice treated with NK-exo (200 and 500 μg) was evaluated. The expressions of specific apoptotic makers (caspase-3, -7, -8, -9, and PARP) and cytotoxic proteins (perforin, granzyme B, FasL, and TRAIL) were analyzed using western blotting. Caspase activation occurs upon cleavage. CD63, as an exosome marker and β-actin, as a loading control were used. H and I: The histograms were quantified with ImageJ software, and normalized to that of β-actin. All data are presented as mean ± S.D. of at least three independent measurements. *p < .05 and **p < .01 vs. untreated NK-exo group. TWPBS control, tumor weight of PBS control group; TWNK-exo treated, tumor weight of NK-exo treated group; H&E, hematoxylin and eosin; IHC, immunohistochemistry.

Fig 4: Tet1 and Tet2 regulate FASL through demethylation of DKK-1.a Immunocytofluorescent staining showed that CD146-positive hPDLSCs express DKK-1. Scale bar, 50 µm. b Western blot analysis showed decreased levels of DKK-1 in Tet1 and Tet2 siRNA-treated hPDLSCs compared to the control hPDLSCs. c qPCR analysis showed significantly decreased expression of DKK-1 in Tet1 and Tet2 siRNA-treated hPDLSCs compared to the control hPDLSCs. d The CpG island of DKK-1. e Tet1 and Tet2 binding on the promoter of DKK-1 in hPDLSCs, as assessed by ChIP-qPCR. IgG was used as a control. f Enrichment of 5-hmC in the DKK-1 promoter decreased in Tet1 and Tet2 siRNA-treated hPDLSCs. g OxBS-sequencing analysis showed that Tet1 and Tet2 siRNA-treated hPDLSCs displayed elevated methylation in the promoter of DKK-1 locus compared to control hPDLSCs. Empty dots indicate unmethylated GpGs; black dots indicate methylated CpGs. h DKK-1 treatment increased the level of DKK-1 and decreased the expression levels of FASL and active β-catenin in Tet1 and Tet2 siRNA-treated hPDLSCs, assessed by Western blot. i In vitro coculture showed that DKK-1 treatment decreases the capacity of Tet1 and Tet2 siRNA-treated hPDLSCs to induce apoptosis (AnnexinV+7AAD+) of T cells. ***P < 0.001, **p < 0.01; p values calculated using two-tailed Student’s t test (mean ± SD)

Fig 5: ASA treatment rescues the impaired immunomodulatory properties of iGMSCs. a In vitro co-culture system showed a significantly increased capacity of ASA-treated iGMSCs to induce Annexin V+/7AAD+ T cell apoptosis when compared with the iGMSC group. b Western blot analysis showed that ASA treatment elevated the expression of FasL in iGMSCs. c–e ASA-treated iGMSCs showed restored immunomodulation capacity compared with untreated iGMSCs, as assessed by c amelioration of losing body weight, d decreased disease activity index (DAI), and e alleviation of colitis histologic activity index (HAI). f FACS analysis showed that ASA-treated iGMSC infusion rescued the iGMSCs’ ability to reduce Th1 and Th17 cells and elevate Treg cells in colitis mice. g ELISA analysis showed that ASA-treated iGMSC infusion rescued the iGMSCs’ ability to downregulate the levels of TNF-a and IL-17 in colitis mice. n = 5 in each group. Scale bar = 200 µm. *P < 0.05. **P < 0.01. ***P < 0.005. Error bars: mean ± SD