Human umbilical vein endothelial cells (HUVECs), human leukemic monocyte lymphoma cell line U937, Burkitt’s lymphoma cell line Raji and human bone marrow stromal cell line HS-5 were purchased form American Type Culture Collection (ATCC, Rockville, Maryland, USA). Cells were cultured in high-content glucose DMEM medium (cat. no. 11965092; Thermo Fisher Scientific, Waltham, MA, USA) supplemented with 10% FBS (cat. no. 10091148; Thermo Fisher Scientific), 1% penicillin-streptomycin (cat. no. 15070063; Thermo Fisher Scientific). Cells were incubated in a humidified incubator containing 5% CO₂ at 37°C.

U937 and Raji cells (5000 cells/well) were seed onto 96-well plates and incubated overnight. After that, U937 and Raji cells were treated with 5 ng/mL HHT or/and 10 μM curcumin for 72 h. Later on, 10 μL CCK‐8 reagent (cat. no. C0038; Beyotime) was added into each well, and cells were incubated for another 2 h. Subsequently, the absorbance at a wavelength of 450 nm was measured using the Multiskan™ FC Microplate Photometer (Thermo Fisher Scientific).

Cells were fixed with 4% paraformaldehyde (cat. no. AS1018; ASPEN Biotechnology) for 15 min, and then incubated with 0.1% Triton X-100 (cat. no. ST797; Beyotime) for 5 min. Later on, cells were blocked with 1% BSA (cat. no. 10735078001; Roche) for 30 min, followed by incubation with the Ki67 antibody (cat. no. ab92742; Abcam Cambridge, MA, USA) overnight at 4°C. After that, the cells were incubated with a secondary antibody (cat. no. ab150077; Abcam) at 37°C for 1 h. Nuclei were counterstained with DAPI (100 μl; cat. no. C1005, Beyotime, Shanghai, China). Subsequently, fluorescence signals were imaged with a fluorescence microscope (Carl Zeiss, Jena, Germany).

Transwell migration and invasion assays were performed using the Matrigel-uncoated or Matrigel-coated transwell chambers (8 μm; Corning, USA) respectively. U937 and Raji cells (4 x 10⁴ cells/well) were suspended in 200 μL serum-free culture medium and then seed into the upper chamber, and 700 μL of DMEM medium containing 2% FBS was added into the lower chamber. Twenty-four hours later, cells that migrated or invaded through the transwell membrane were fixed with 70% methanol, and then stained with 1% crystal violet (cat. no. AS1086; ASPEN Biotechnology). Subsequently, the migrated or invaded cells were observed using a fluorescence microscope and counted in five randomly selected fields.

The samples of the supernatant were collected from Raji cells. After that, ELISA kits for VEGF-A (cat. no. ELK1129), VEGF-B (cat. no. ELK2321), VEGF-C (cat. no. ELK1194), VEGF-D (cat. no. ELK1195) were purchased from ELK Biotechnonlgy (Wuhan, China), and the assay was performed according to the manufacturer’s protocols. Absorbance was assessed using a Multiskan™ FC Microplate Photometer (Thermo Fisher Scientific).

Total RNA was extract using the TRIpure Total RNA Extraction Reagent (cat. no. EP013; ELK Biotechnology, Wuhan, China). After that, cDNA was synthesized using EntiLink™ 1st Strand cDNA Synthesis Kit (cat. no. EQ003; ELK Biotechnology). Later on, qPCR was performed on the StepOne™ Real-Time PCR System (Thermo Fisher Scientific) using the EnTurbo™ SYBR Green PCR SuperMix (cat. no. EQ001; ELK Biotechnology). Expression of target genes (2^−ΔΔCt) was normalized against β-actin and U6, as described previously (28). U6: 5’-CTCGCTTCGGCAGCACAT-3’ (F); 5’- AACGCTTCACGAATTTGCGT-3’ (R). miR-150: 5’- TCCCAACCCTTGTACCAGTG-3’ (F); 5’-CTCAACTGGTGTCGTGGAGTC-3’ (R). miR-22: 5’-GCCAGTTGAAGAACTGTCTCAAC-3’ (F); 5’- CTCAACTGGTGTCGTGGAGTC-3’ (R). miR-206: 5’- TGGAATGTAAGGAAGTGTGTGG-3’ (F); 5’-CTCAACTGGTGTCGTGGAGTC-3’ (R). miR-34a: 5’-TGGCAGTGTCTTAGCTGGTTG-3’ (F); 5’- CTCAACTGGTGTCGTGGAGTC-3’ (R). β-actin: 5’- GTCCACCGCAAATGCTTCTA-3’ (F); 5’-TGCTGTCACCTTCACCGTTC-3’ (R). VEGFA: 5’-GAACTTTCTGCTGTCTTGGGTG-3’ (F); 5’- GGCAGTAGCTGCGCTGATAG-3’ (R).

Total proteins were quantified using Bradford protein assay (Bio-Rad, CA, USA), and then equal proteins (30 μg per lane) were separated by 10% SDS-PAGE (cat. no. AS1012; ASPEN). After that, proteins were transferred onto polyvinylidene difluoride (PVDF) membrane (Thermo Fisher Scientific). Later on, the membrane was blocked in 5% skimmed milk (cat. no. AS1033; ASPEN) for 1 h at room temperature, followed by incubating with the primary antibodies against p-VEGFR2 (cat. no. ab5473; 1:1000, Abcam), VEGFR2 (cat. no. ab134191; 1:1000, Abcam), p-Akt (cat. no. ab81283; 1:1000, Abcam), Akt (cat. no. ab179463; 1:1000, Abcam), p-eNOS (cat. no. ab215717; 1:1000, Abcam), eNOS (cat. no. ab252439; 1:1000, Abcam), p-p38 (cat. no. #4631; 1:1000, Cell signaling technology), p38 (cat. no. #8690, 1:1000, Cell signaling technology), p-HSP27 (cat. no. #9709; 1:1000, Cell signaling technology), HSP37 (cat. no. #95357; 1:1000, Cell signaling technology), p-ERK (cat. no. ab201015; 1:1000, Abcam), ERK (cat. no. ab184699; 1:1000, Abcam), p-JNK (cat. no. ab112501; 1:1000, Abcam), JNK (cat. no. ab4821; 1:1000, Abcam), VEGFA (cat. no. ab52917; 1:1000, Abcam), angiogenin-1 (cat. no. ab189207; 1:1000, Abcam), MMP2 (cat. no. ab92536; 1:1000, Abcam), MMP9 (cat. no. ab76003; 1:1000, Abcam) p-STAT3 (cat. no. ab267373; 1:1000, Abcam), STAT3 (cat. no. ab68153; 1:1000, Abcam), SP1 (cat. no. ab124804; 1:1000, Abcam) and anti-β-actin (cat. no. ab8227; 1:1000) at 4°C overnight. Next, the membrane was incubated with horseradish peroxidase-conjugated secondary antibodies (cat. no. ab97051; 1: 5000) for 1 h. Subsequently, the blots were visualized using an electrochemiluminescence (Thermo Fisher Scientific). β-actin was acted as an internal control.

Cell apoptosis were carried out using the fluorescein isothiocyanate (FITC) Annexin V Apoptosis Detection Kit (cat. no. 559763; BD Biosciences). A FACSCalibur flow cytometer (BD Biosciences) was used to analyze the number of apoptotic cells.

Exosomes were purified from Raji cell-derived conditioned media (CM) and mouse serum by ultracentrifugation, following previously described standard procedures (29). The exosomes pellets were washed with PBS followed by ultracentrifugation at 100,000g at 4°C for 60 min, and then suspended in PBS for further processed for protein or RNA extraction. The nanoparticle tracking analysis (NTA; Malvern Panalytical, Malvern, UK) was used to assess the size and quantity of exosomes. Western blot assay was performed to detect exosome surface markers (TSG101 and CD63).

Exosomes isolated from Raji cells were fluorescently labeled with PKH26 membrane dye (cat. no. MINI26; Sigma Aldrich, St. Louis, MO, USA) and incubated for 15 min at 37°C. Later on, labeled exosomes were washed exosome-free PBS by ultracentrifugation at 110,000g, and then resuspended in exosome-free PBS. After that, PKH26-traced exosomes were co-cultured with HUVECs for 24 h. F-actin was stained using phalloidin-FITC. Subsequently, internalization of exosomes by HUVECs were observed by a fluorescence microscope.

100 μL growth factor enriched Matrigel (cat. no. 354248; Corning) was plated in 24-well plate and allowed to polymerize for 30 min at 37°C. After that, the treated HUVECs were resuspended in serum-free medium and seeded on the matrigel-coated well. The tube formation of HUVECs was observed and photographed after 12 h incubation under a microscope. The tube formation ability was determined by measuring the branch points, mesh area and tube numbers with five randomly selected fields.

BALB/c nude mice (6 – 8 weeks old) were obtained from the Shanghai Slac Animal Center (Shanghai, China). All animal experiments were approved by the First Affiliated Hospital of Zhejiang Chinese Medical University and animals were maintained following the recommended procedures of the National Institutes of Health Guide for the Care and Use of laboratory animals. Raji cells (5 x 10⁶ cells) were injected subcutaneously into the left flank of nude mice. When the tumors volume reached to 180 mm³, animals were randomized into two groups: control and Exo^{-HHT + Curcumin} groups. The control group received normal saline only. In addition, mice in the Exo^{-HHT + Curcumin} group were intravenously injected with Exo^{-HHT + Curcumin} every 2 days. Tumor volume was obtained by calipers once a week. After that, animals were sacrificed under anesthesia at day 28 and the tumors were removed.

All statistical analyses were performed using GraphPad Prism software (version 7.0, La Jolla, CA, USA). The normality of the data was tested using Shapiro-Wilk test. One-way analysis of variance (ANOVA) and Tukey’s tests were carried out for multiple group comparisons. Each experiment was carried out in triplicate. Data are presented as the mean ± standard deviation (S.D.). *P < 0.05 was considered to be statistically significant.

To explore the role of curcumin and HHT on the viability of lymphoma cells, CCK-8 assay was performed. As shown in Figure 1A , treatment of both U937 and Raji cells with HHT or curcumin alone significantly decreased the viability compared with the control group (P < 0.01). Remarkably, combination treatment with HHT and curcumin reduced cell viability more in U937 and Raji cells than either HHT or curcumin alone (P < 0.01; Figure 1A ). In addition, the results of ki67 immunofluorescence assay indicated that HHT treatment notably inhibited the proliferation of U937 and Raji cells (P < 0.01; Figure 1B ). As expected, combined treatment with curcumin potentiated the anti-proliferative effects of HHT on U937 and Raji cells (P < 0.01; Figure 1B ). These data suggested that combination of curcumin with HHT could inhibit the proliferation of lymphoma cells.

Next, the effects of HHT/curcumin combination on cell migration and invasion were detected with transwell assay. As shown in Figures 2A–D , combination treatment significantly inhibited the migration and invasion abilities of U937 and Raji cells; however, these changes were obviously reversed by VEGF treatment (All P < 0.01). These results indicated that combination of curcumin with HHT could suppress the migration and invasion abilities of lymphoma cells.

VEGF is released by cancer cells to induce tumor growth, migration and angiogenesis (30). ELISA assay was used to detect the secreted levels of VEGF-A, VEGF-B, VEGF-C and VEGF-D in Raji cells. As shown in Figures 3A–D , combination of HHT with curcumin decreased the secreted levels of VEGFA (P < 0.05), VEGFB (P < 0.05), VEGFC (P < 0.01) and VEGFD (P < 0.01) in Raji cells compared with HHT treatment group; however, these effects were reversed by VEGF treatment. To further explore whether combination treatment inhibited the viability and migration via regulation of VEGFA, we used two siRNAs to downregulate the level of VEGFA in Raji cells ( Supplementary Figure 1A ). Additionally, VEGFA siRNAs enhanced the inhibitory effects of combination treatment on the viability (P < 0.01) and migration (P < 0.01) of Raji cells ( Supplementary Figures 1B, C ). These data suggested that combination treatment could inhibit the viability and migration of lymphoma cells via downregulation of VEGFA.

In addition, it has been shown that Sp1 and STAT3 are transcription factors that regulate VEGF expression and VEGF secretion of cancer cells (31, 32). As shown in Supplementary Figures 2A, B , combination of HHT with curcumin remarkedly decreased the expressions of p-STAT3 and SP1 in Raji cells; however, these effects were reversed by VEGF treatment (All P < 0.01). These results indicated HHT/curcumin combination inhibited VEGF level in Raji cells via inactivation of the transcription factors STAT3 and SP1.

Moreover, evidences have shown that VEGFA expression can be regulated by various microRNAs (miRNAs), including of miR-150, miR-22, miR-206 and miR-34a (33–36). Meanwhile, the levels of miR-150, miR-22, miR-206 and miR-34a have been found to be downregulated in human cancers (37–40). Thus, we investigated whether HHT combined with curcumin reduced VEGFA level in Raji cells by these miRNAs. As shown in Supplementary Figure 3 , combination treatment significantly upregulated the levels of miR-150 (P < 0.01), miR-22 (P < 0.01), miR-206 (P < 0.05) and miR-34a (P < 0.01) in Raji cells. Collectively, HHT/curcumin combination decreased VEGFA level in Raji cells via upregulation of miR-150, miR-22, miR-206 and miR-34a.

It has been reported that PI3K/Akt and MAPK signaling pathways could be mediated by VEGF (41, 42). Thus, we next to examine whether HHT/curcumin combination inhibited Raji cell growth via regulation of PI3K/Akt and MAPK signaling pathways ( Figures 3E–N ). As indicated in Figures 3E–G, I, K, M, N , combination treatment led to decreased phosphorylation levels of VEGFR2 (P < 0.01), Akt (P < 0.01), p-HSP27 (P < 0.01) proteins and increased phosphorylated and total JNK (P < 0.01) levels in Raji cells; however, these levels were reversed in the presence of VEGF. Meanwhile, combination treatment caused no difference in the activation of eNOS, p38 and ERK1/2 in Raji cells ( Figures 3E, H, I, J, L ). These data illustrated that combination of curcumin with HHT could inhibit the growth of lymphoma cells via regulation of VEGF/Akt and JNK signaling pathways.

It has been shown that cancer-secreted exosomes are responsible for tumor growth and angiogenesis (43). In order to explore whether lymphoma cell-secreted exosomes perform the above functions, we extracted exosomes from the cultural supernatants of Raji cells. Nanoparticle tracking analysis (NTA) showed that the lymphomas exosomes were approximately 50-150 nm in diameter ( Figure 4A ). In addition, western blot revealed that vesicles were positive for exosome markers TSG101 and CD63 ( Figure 4B ). RT-qPCR showed that exosomal VEGF was found to be enriched in the conditioned medium of Raji cells (P < 0.01). However, Raji cells treated with HHT and curcumin led to the downregulation of exosomal VEGF; whereas additional of exogenous VEGF reversed the effect of HHT/curcumin combination on the level of VEGF in lymphoma cell-secreted exosomes (All P < 0.01; Figure 4C ). These data indicated that combination of curcumin with HHT could inhibit VEGF levels in exosomes derived from Raji cells.

Next, Raji cell derived PKH26-labeled exosomes were incubated with HUVECs. After 24 h of co-cultivation, PKH26-labeled dye was observed in HUVECs ( Figure 4D ). These results suggested that lymphomas exosomes are absorbed by HUVECs. We further detected the expression of VEGF in HUVECs incubated with exosomes isolated from Raji cells. As shown in Figures 4E, F , the expression of VEGF was markedly upregulated in HUVECs incubated with lymphoma cell-secreted exosomes; however, VEGF expression in HUVECs was markedly downregulated by the exosomes isolated from HHT/curcumin combination-treated Raji cells (Raji/combination exosomes) compared to those isolated from control Raji cells (All P < 0.01). These results suggested that HHT combined with curcumin treatment reduced VEGF secretion from Raji cells.

To further explore whether Raji/combination exosomes could inhibit the proliferation and migration of HUVECs, CCK-8, Ki67 immunofluorescence and transwell migration assays were applied. As shown in Figures 5A–E , Raji cell-derived exosomes significantly promoted the viability, proliferation and migration of HUVECs (All P < 0.01). In contrast, the viability, proliferation and migration of HUVECs were markedly inhibited by the Raji/combination exosomes compared to those purified from control Raji cells; however, these changes were reversed by the exosomes loaded with exogenous VEGF (All P < 0.01; Figures 5A–E ). These results indicated that Raji/combination exosomes could inhibit proliferation and migration of HUVECs by reducing VEGF secretion from Raji cells.

Tube formation assay were then performed to assess whether Raji/combination exosomes could regulate angiogenesis in HUVECs. As indicated in Figure 6A , exosomes from Raji cells notably promoted angiogenesis in HUVECs, as shown by branch points (+52%; P < 0.01), mesh area (+33%; P < 0.01) and tube number (+42%; P < 0.01) increases. On the contrary, Raji/combination exosomes dampened angiogenesis in Raji cells (P < 0.01; Figure 6A ). In addition, western blot assay indicated that the expressions of angiogenin-1, MMP2, MMP9 and p-Akt were obviously decreased in HUVECs by the Raji/combination exosomes compared to those purified from control Raji cells; however, these levels were reversed by the exosomes loaded with exogenous VEGF (All P < 0.01; Figures 6B–F ). These data indicated that Raji/combination exosomes could inhibit the angiogenesis of HUVECs by reducing VEGF secretion from Raji cells.

Additionally, stromal cells have been characterized as important component of the tumor microenvironment as well, which play an important role in cancer development (44). Meanwhile, adhesion to cultured stromal cells could protect lymphoma cells from apoptosis induced by anti-tumor drugs (45). Thus, we established HS-5-Raji co-culture model in direct contact mode to simulate the bone marrow microenvironment in lymphoma (45). As shown in Supplementary Figure 4B , the inhibitory effect of combination treatment on cell viability was reversed by co-culturing of Raji cells with HS-5 (P<0.01). In addition, co-cultures of Raji cells with HS-5 led to a marked decrease of combination treatment-induced apoptosis compared to cells cultured on a plastic surface (17% vs. 25%, P<0.01). Collectively, interactions with the microenvironment could protect Raji cells from combination treatment-induced apoptosis. Again, these results proved that interactions with the microenvironment might be critical for mediating tumor therapy.

Next, we investigated the anti-tumor effect of Raji/combination exosomes in mouse Raji xenograft model in vivo. As shown in Figures 7A–C , Raji/combination exosomes markedly reduced the tumor volume (P < 0.01) and tumor weight (P < 0.01) of Raji transplanted tumor in vivo. Collectively, Raji/combination exosomes could inhibit tumorigenesis in Raji xenograft in vivo.