A549, H1975, BEAS-2B, 293T cells and HUVECs were purchased from Chinese Academy of Sciences Cell Bank of Type Culture Collection (Shanghai, China), authenticated for morphology, genotypes, drug response and absence of mycoplasma. A549, H1975, BEAS-2B were cultured with RPMI-1640 medium (Gibco, NY, United States). 293T cells were cultured with DMEM medium (Gibco, NY, United States). Both kinds of medium were supplemented with 10% fetal bovine serum (Gibco), 100 U/ml penicillin and 100 μg/ml streptomycin (Gibco). HUVECs were cultured with Endothelial Cell Growth Medium (Lonza, Basel, Switzerland). A549-Luc cells containing luciferase reporter gene were constructed by Land Biological Technology (Guangzhou, China), screened with 200 μg/ml Geneticin (G-418) Sulfate (Life Technologies, United States). All the cells were placed in 37°C, 5% CO₂ incubator. Curcumol was obtained from Chengdu Must Bio-technology (purity 99%, Chengdu, China). Axibinib was purchased from MedChemExpress (purity 99.94%, NJ, United States).

The 3- (4,5-dimethylthiazol-2-yl) -2,5-diphenyltetrazolium bromide (MTT) assay was conducted to detect the cell viability. A549, H1975, BEAS-2B cells (4000 cells/well) were seeded into 96-well plates, then treated with indicated doses of Curcumol for 24 h, 48 h, and 72 h. Next, cells were cultured with MTT solution (5 mg/ml) at 37°C for at least 4 h, followed by adding DMSO (dimethyl sulfoxide). Microplate reader (Synergy H1, United States) was used to measure OD (optical density) at 490 nm. The experiments were performed independently five times (n = 5) in triplicate.

EdU (5-Ethynyl-2′-deoxyuridine) assay kit (RIBOBIO, Guangzhou, China) was applied to evaluate cell proliferation. A549 and H1975 cells were treated with Axitinib (4 μM) or Curcumol (400 μM) for 24 h, then incubated with EdU for 2 h, fixed with 4% paraformaldehyde for 30 min, incubated in 0.2% TritonX-100 for 10 min, added with 1× Apollo buffer, stained with 1×Hoechest (5 mg/ml). Fluorescence microscope (Eclipse Ti2-E, Japan) was used to capture images.

Cells were seeded in 6-well plates. When cells confluence reached 90%, scratches were made by 200 μL micropipette tip. The wells were gently washed with PBS twice. Then, cells were treated with Axitinb or Curcumol. 0 h and 24 h images were recorded, and relative wound healing rate was calculated by comparing difference of wound width before and after scratching.

The invasion assay was carried out by using 24-well Matrigel chambers (BD Biosciences, United States). 2 × 10⁵ A549 cells in 200 μL RPMI-1640 without FBS were seeded in the upper chamber, which was added with Matrigel in advance. 500 μL RPMI-1640 supplemented with 20% FBS was added to the lower chamber. After 24 h administration of Curcumol or Axitinib, cells on the upper surface of the filter were taken away by cotton swab, while cells on the lower surface were immediately fixed with 4% paraformaldehyde and stained with crystal violet.

A549 cells were cultured in 6-well plates overnight, then supplemented with Axitinib (4 μM) or Curcumol (200, 400 μM). After 24h, CM (conditioned medium) was harvested by taking supernatant after centrifugation. CM was named as following: 1) Con-CM, 2) Axitinib 4 μM-CM, 3) Curcumol 200 μM-CM and 4) Curcumol 400 μM-CM.

HUVECs were seeded into Matrigel-coated 24-well plates at a density of 1 × 10⁵ cells per well and cultured with different conditioned medium as described above for 24 h. Branch points and capillary length were calculated from images by ImageJ software.

Cell total protein was extracted, then resolved on 10% SDS-PAGE, and transferred onto PVDF (polyvinylidene fluoride) membranes. Non-specific binding sites were blocked by 5% non-fat milk for 1 h at room temperature. Subsequently, the membranes cut and probed with corresponding primary antibodies at 4°C overnight, followed by incubation with the secondary antibodies for 2 hours at room temperature. Primary antibodies information is as follows: anti-VEGFR2 (phospho Y951) (ab38473, Abcam), anti-VEGFR2 (ab39638, Abcam), anti-VEGFA (ab46154, Abcam), anti-SP1 (9389S, Cell Signaling Technology, CST for short), anti-β-Tubulin (2128S, CST).

Real-time quantitative Polymerase Chain Reaction (qRT-PCR) was performed to detect the expression of VEGFA, SP1 mRNA and miR-125b-5p. Total RNA (1 μg) was reversed to cDNA by using First Strand cDNA synthesis kit (Vazyme, Nanjing, China). A total of 20 μL mixed liquid, containing cDNA, primers and enzyme mixture was added to the PCR 96-well plates for PCR reaction in PCR instrument (ABI ViiA7, United States). The PCR reaction procedure was as follows: 95°C for 30 s, then 95°C for 10 s plus 60°C for 30 s for 40 cycles. GAPDH or U6 was used as endogenous control. Primers sequences were designed as follows:

VEGFA forward: 5′-TAG​AGT​ACA​TCT​TCA​AGC​CGT​C-3′,

A549 cells were seeded on slides (SPL life sciences, South Korea) then transfected with SP1 overexpression vectors (pcDNA3.1-SP1) and pcDNA3.1 negative control vectors (GeneCopoeia, United States), followed by treating with Curcumol for 24 h. MiR-125b-5p mimics, inhibitors and NC were generated by GenePharma (Shanghai, China). Cells were fixed with 4% paraformaldehyde for 10 min and permeabilized for 15 min with 0.3% TritonX-100, then incubated for an hour in blocking buffer (1% Bovine Serum Albumin) at room temperature. Cells were incubated with primary antibodies against VEGFA (ab39250, Abcam) at 4°C overnight. Next day, cells were incubated with Alexa Fluor 488 anti-rabbit antibodies for 1 h. Anti-fluorescence quenching sealing agent containing DAPI (Beyotime, China) were added onto slides, then analyzed by confocal microscopy (ZEISS LSM710, Germany).

The wild-type and mutation-type VEGFA 3′UTR luciferase vectors were synthesized (GeneCopoeia, MD, United States). SP1 promoter vector with Renilla luciferase reporter as internal reference was also generated by GeneCopoeia. Co-transfected these vectors into cells with either miR-125b-5p mimics or inhibitors and corresponding negative control using Liposomal Transfection Reagent (Yeasen, Shanghai, China). Detection of luciferase activities was undergone by using Luc-Pair™ Dual-Luciferase HS Assay Kit (GeneCopoeia MD, United States). Optical density was normalized with Renilla within each sample.

Subcutaneous xenograft tumor tissue of nude mice was fixed with 4% formaldehyde for about 24 h before paraffin-embedding, slicing. After antigen retrieval with microwave method, the primary antibody of PCNA (13110S, diluted into 1:4000, CST), VEGFA (ab39250, diluted into 1:100, Abcam), SP1 (9389S, diluted into 1:2000, CST), CD31 (3528S, diluted into 1:1000, CST) were all incubated at 4°C overnight, followed by secondary antibody for 1 h. 3,3ʹ- diaminobenzidine (Maixin Biotech, Fuzhou, China) was used as chromogenic agent. Pictures were randomly observed and taken under ×400 magnification from at least five random fields by using upright microscope (BX53+DP72, Olympus Corporation, Japan).

Animal studies were proceeded according to regulations approved by Guangdong Provincial Hospital of Chinese Medicine Institutional Animal Care and Use Committee (Ethics Approval Number 2020076). BALB/c-nu mice of 6-week-old were provided by Beijing Vital River Laboratory Animal Technology Co., Ltd. 2 × 10⁶ A549-Luc cells (Land Biological Technology, Guangzhou, China) were injected subcutaneously into the right flank, near the axillary fossa region. One week later, mice were randomly assigned into four groups (six mice per group) including Control (saline), Axitinib 20 mg/kg, Curcumol 40 mg/kg, Curcumol 80 mg/kg groups. Saline, Axitinib or Curcumol was given via gavage once a day for 25 days. The doses selected referred to other reports (Ma Y. H. et al., 2019; Zuo et al., 2020). Tumor volume was calculated by the formula: volume = (width² × length)/2. Body weight of mice were measured every four days. As for bioluminescence imaging analysis, each goup mice were injected with 10 mg/kg luciferin (Promega, United States). Images were acquired and analyzed by the IVIS-200 Imaging System (Xenogen, Berkeley, United States). After 25 days, all tumor bearing mice were sacrificed, and tumor tissues were isolated and applied for examination of proteins and microRNA expression level.

Statistical analysis was performed by using one-way univariate analysis of variance (ANOVA) with Bonferroni post-hoc test, and values were shown as mean ± standard deviation. Difference with p < 0.05 was considered and interpreted to be statistically significant.

At first, we evaluated the cytotoxicity of Curcumol against normal human bronchial epithelioid cells BEAS-2B cells to explore its safe concentration. Next, we evaluated inhibitory effect of Curcumol on the proliferation of Non-Small Cell Lung Cancer Cells, A549 and H1975. The results showed that Curcumol at the concentration of 400 μM at 24 h or 48 h was acceptable concentration and duration against BEAS-2B cells because it had limited cytotoxicity compared to that of vehicle control. A549 cells viability were reduced by 19.06%, 21.38%, 28.53%, when treated with Curcumol (400 μM) for 24 h, 48 h, 72 h respectively, compared to corresponding non-treated controls. The rate of cell proliferation was found to be significantly decreased (by 31.00%, 50.86% and 47.17%) in H1975 cells treated with Curcumol (400 μM), for 24 h, 48 h, and 72 h respectively, when compared to non-treated controls H1975 cells (Figure 1A). Meanwhile, EdU (Figure 1B) assay also showed that Curcumol could inhibit A549 and H1975 cell viability (EdU positive cells in Curcumol 400 μM group: A549 cells 23.58 ± 6.48%; H1975 cells 27.63 ± 5.34%, after normalization to control groups, shown as mean ± SD), with Axitinib (a commonly clinically used targeting anticancer drug) used as the positive control drug. Furthermore, we found that it also could attenuate the migration (wound healing rate in A549 cells: control group 43.16 ± 7.02%; Curcumol 400 μM group 17.28 ± 3.20%, H1975 cells: control group 70.57 ± 2.57%; Curcumol 400 μM group 34.70 ± 4.95%) (Figures 1C,D) and invasion (invasion cell number in A549 cells: control group 116 ± 7.07; Curcumol 400 μM group 61.40 ± 8.79) (Figures 1E,F) capability of NSCLC cells.

Curcumol is a monomer component of Chinese medicine Curcuma, and Curcuma is considered to have the effect of promoting blood circulation and removing blood stasis (Dosoky and Setzer, 2018). Therefore, we speculated that Curcumol may have a regulatory impact on angiogenesis. We treated A549 cells with Axitinib, Curcumol, or DMSO, then collected the conditioned medium (CM) for HUVEC culture (marked as Axitinib-CM, Curcumol-CM, Con-CM). We cultured HUVEC in these CMs and examined HUVEC tube formation (Figure 2A) and migration capacities (Figures 2B,C). Curcumol 400 μM-CM significantly inhibited HUVEC tube formation (branch points: Con-CM 43.33 ± 5.51; Curcumol 400 μM-CM 16.33 ± 3.51, fold change of capillary length: Con-CM 1.69 ± 0.14; Curcumol 400 μM-CM 0.35 ± 0.10, shown as mean ± SD) and migration, compared with Con-CM, which showed opposite effect. Then we checked the protein level of phosphorylated VEGFR2 and total VEGFR2, we found that phosphorylated/total VEGFR2 protein ratio was decreased in Axitinib-CM group (0.68 ± 0.05) and Curcumol 400 μM-CM group (0.84 ± 0.11), with Con-CM group (1.21 ± 0.19) as comparing object (Figures 2D,E). Moreover, the level of VEGFA expressed by HUVEC cultured with Curcumol 400 μM-CM (0.78 ± 0.12) was significantly down regulated, while up regulated in Con-CM group (1.27 ± 0.05) (Figures 2F,G). These data indicated that Curcumol affected A549 cells-induced VEGF pathway of endothelial cells in tumor microenvironment, modulating angiogenesis of HUVEC.

VEGFA around endothelial cells in the tumor microenvironment is partly derived from autocrine and mostly from paracrine of tumor cells. Therefore, we speculated that Curcumol may directly regulate the expression level of VEGFA in tumor cells. The transcription factors in descending order of score, that may regulate VEGFA transcription were predicted by software JASPAR (https://jaspar.genereg.net/). The list showed that SP1 (specificity protein 1) could be a good candidate (Figure 3A). qRT-PCR results showed that the mRNA levels of VEGFA and SP1 in lung cancer cells were significantly higher than that in normal lung epithelial cells (relative VEGFA mRNA level: BEAS-2B 1.00 ± 0.06; A549 1.62 ± 0.14; H1975 3.36 ± 0.27, relative SP1 mRNA level: BEAS-2B 1.00 ± 0.05; A549 2.09 ± 0.34; H1975 3.76 ± 0.46) (Figure 3B). Western blotting analysis results confirmed our assumptions, that Curcumol (400 μM) could decreased both VEGFA and SP1 protein level in A549 and H1975 cells (relative VEGFA protein level: A549 0.75 ± 0.08; H1975 0.71 ± 0.06, relative SP1 protein level: A549 0.69 ± 0.12; H1975 0.77 ± 0.07) (Figures 3C,D).

It is well known that microRNAs can degrade mRNA and reduce post transcriptional translation of target genes by pairing with the 3′ untranslated region of mRNA. We predicted the microRNAs with more than seven base binding sites at the 3′UTR of VEGFA mRNA by software miRmap (https://mirmap.ezlab.org/app/), and arranged them in descending order of score (Figure 3E). Next, the differential expression of those microRNAs in human normal tissues and tumor tissues was retrieved in the database ENCORI (https://starbase.sysu.edu.cn/). Interestingly, we found that miR-125b-5p was significantly lower in human lung adenocarcinoma tissues (Figure 3F), which was consistent with the results of cell experiments (relative miR-125b-5p expression level: BEAS-2B 0.97 ± 0.04; A549 0.55 ± 0.15; H1975 0.65 ± 0.08) (Figure 3G). Next, we tested whether Curcumol could affect the expression of miR-125b-5p in NSCLC cells and found that Curcumol (400 μM) could promote its expression (relative fold change of miR-125b-5p expression: A549 1.27 ± 0.08; H1975 1.62 ± 0.21) (Figure 3H).

To evaluate the underlying mechanisms associated with the anticancer and antiangiogenic effect of Curcumol, we applied transient transfection technique. SP1 overexpression vectors (pcDNA3.1-SP1) or control vectors (pcDNA3.1) were transfected into A549 and H1975 cells for up to 24 h, followed by treating with Curcumol for an additional 24 h. Overexpression of SP1 reversed the effect of Curcumol-inhibited cell viability of A549 and H1975 cells, determined by MTT assays (fold change of cell viability: in A549 cells pcDNA3.1 group 0.98 ± 0.04; pcDNA3.1 + Curcumol group 0.79 ± 0.05; pcDNA3.1-SP1 + Curcumol group 0.89 ± 0.03, in H1975 cells pcDNA3.1 group 0.92 ± 0.04; pcDNA3.1 + Curcumol group 0.68 ± 0.06; pcDNA3.1-SP1 + Curcumol group 0.80 ± 0.03) (Figure 4A). According to the previous prediction, we conjectured that SP1 might play a role in promoting the growth and angiogenesis of NSCLC cells by increasing the expression of VEGFA. Then, we explored whether SP1 could promote the expression of VEGFA protein through Western blot (Figures 4B,C) and immunofluorescence experiments (Figures 4D,E), and obtained positive results. Interestingly, qRT-PCR results suggested that SP1 could also attenuate the promoting effect of Curcumol on miR-125b-5p in NSCLC cells (fold change of miR-125b‐5p expression level: in A549 cells pcDNA3.1 group 0.97 ± 0.08; pcDNA3.1-SP1 group 0.32 ± 0.03; pcDNA3.1 + Curcumol group 1.20 ± 0.03; pcDNA3.1-SP1 + Curcumol group 0.64 ± 0.11, in H1975 cells pcDNA3.1 group 0.92 ± 0.09; pcDNA3.1-SP1 group 0.42 ± 0.06; pcDNA3.1 + Curcumol group 1.56 ± 0.13; pcDNA3.1-SP1 + Curcumol group 1.11 ± 0.12) (Figure 4F).

Next, we carefully explore the specific role and possible mechanism of miR-125b-5p in Curcumol-inhibited lung cancer cells growth. The results of MTT indicated that exogenous increase of miR-125b-5p level could promote the anti-cancer effect of Curcumol (fold change of cell viability: in A549 cells NC group 0.98 ± 0.09; miR-125b-5p mimic group 0.69 ± 0.05; NC + Curcumol group 0.70 ± 0.03; mimic + Curcumol group 0.56 ± 0.04, in H1975 cells NC group 1.00 ± 0.09; miR-125b-5p mimic group 0.56 ± 0.02; NC + Curcumol group 0.71 ± 0.05; mimic + Curcumol group 0.47 ± 0.04), while decreasing the level of miR-125b-5p could partially reverse the anti-cancer effect of Curcumol (fold change of cell viability: in A549 cells NC group 1.02 ± 0.04; miR-125b-5p inhibitor group 1.28 ± 0.08; NC + Curcumol group 0.72 ± 0.03; inhibitor + Curcumol group 1.04 ± 0.01, in H1975 cells NC group 0.99 ± 0.08; miR-125b-5p inhibitor group 1.02 ± 0.06; NC + Curcumol group 0.70 ± 0.07; inhibitor + Curcumol group 0.93 ± 0.06) (Figures 5A,B).

By utilizing multiple bioinformatics prediction databases, we acknowledged that miR-125b-5p has a relatively conservative binding site in the region of VEGFA mRNA 3′UTR. Subsequently, we created a pair of cDNA fragment of VEGFA containing the wild-type or mutated binding sites of miR-125b-5p. Then, dual-luciferase reporter assay was conducted to detect the luciferase activities, the results indicated that miR-125b-5p mimics significantly reduced the luciferase activities in 293T cells transfected with the wild-type cDNA fragment of VEGFA (1.03 ± 0.04), compared with negative control (NC) group (0.64 ± 0.09) (Figure 5C). Furthermore, miR-125b-5p inhibitor could rescue the inhibitory effect of Curcumol on VEGFA protein expression, verified by Western blot analysis (Figures 5D,E) as well as cell immunofluorescence assay (Figure 5F).

We wondered whether miR-125b-5p also has a regulatory effect on SP1, and if so, what is the possible mechanism. Surprisingly, miR-125b-5p was able to regulate SP1 protein expression (fold change of SP1 protein expression level: in A549 cells NC group 1.02 ± 0.07; NC + Curcumol group 0.66 ± 0.06; miR-125b-5p inhibitor group 1.40 ± 0.14; inhibitor + Curcumol group 1.08 ± 0.19, in H1975 cells NC group 1.01 ± 0.07; NC + Curcumol group 0.75 ± 0.07; miR-125b-5p inhibitor group 1.28 ± 0.06; inhibitor + Curcumol group 1.04 ± 0.11) (Figures 5G,H). Therefore, we continued to design dual-luciferase reporter assay to explore the effect of miR-125b-5p and Curcumol on SP1 promoter activity. At first, A549 and H1975 cells were all transfected with SP1 promoter plasmids, then co-transfected with miR-125b-5p inhibitor or NC, treated with Curcumol (400 μM) for additional 24 h. It turned out that Curcumol could decreased SP1 promoter activity, while miR-125b-5p inhibitor showed opposite effect (fold change of SP1 promoter activity: in A549 cells NC group 1.06 ± 0.09; NC + Curcumol group 0.53 ± 0.07; miR-125b-5p inhibitor group 1.26 ± 0.11; inhibitor + Curcumol group 1.04 ± 0.09, in H1975 cells NC group 1.01 ± 0.06; NC + Curcumol group 0.65 ± 0.05; miR-125b-5p inhibitor group 1.76 ± 0.07; inhibitor + Curcumol group 1.15 ± 0.10) (Figure 5I).

To further validate in vitro results, we constructed the xenograft tumor model with nude mice to test the antitumor effect of Curcumol. Luciferase-expressing A549 cells (A549-Luc) were injected subcutaneously to the right flank of each BALB/c nude mouse. We found that Curcumol could significantly suppress tumor volume (control group: 2544.20 ± 550.13 mm³; Curcumol 80 mg/kg group: 1146.35 ± 244.92 mm³), tumor weight (control group: 1.78 ± 0.43 g; Curcumol 80 mg/kg group: 0.61 ± 0.23 g) as well as tumor luciferase activity (control group: 4.99 ± 1.77; Curcumol 80 mg/kg group: 1.97 ± 0.54), compared with the control group (Figures 6A–C). It is worth noting that during the administration of the drug, nude mice were in good spirits in each group. There was no obvious abnormality in diet, defecation and urination, and no obvious adverse reactions were observed. The body weight was steadily increased, and there was no significant difference among the groups (Figure 6D).

Moreover, consistent with the in vitro results, decreased expression of SP1 and VEGFA protein was observed in both nude mice tumor tissues fixed with paraformaldehyde (Figures 6E,F) and fresh tumor tissues (Figures 6H,I). PCNA, which may well indicate cancer cell proliferation activity, also decreased within tumors from Curcumol 80 mg/kg group (control group: 59.41 ± 8.56%; Curcumol 80 mg/kg group: 32.03 ± 8.97%). Microvessel density (MVD) was calculated through CD31 antibody staining in mice tumor tissues (control group: 35.33 ± 6.51; Curcumol 80 mg/kg group: 16.00 ± 4.58) (Figure 6G). Last but not the least, miR-125b-5p expression increased apparently in Curcumol higher dose group (fold change of miR-125b-5p expression in control group: 1.00 ± 0.30; Curcumol 80 mg/kg group: 2.14 ± 0.46) (Figure 6J). Taken together, above results indicated that Curcumol may repress NSCLC cells growth and angiogenesis by regulating SP1/miR-125b-5p/VEGFA interaction network (Figure 6K).