Patients and follow-up
HCC and the corresponding adjacent non-tumorous liver tissue (ANLT) specimens were collected from 102 cases of HCC patients who underwent operation at Guangzhou Red Cross Hospital, Jinan University from February 2012 to December 2019. All tissue samples were obtained and frozen in liquid nitrogen immediately after operation and then transferred to an ultra-low temperature freezer (Meling Biology & Medical, Hefei, China) and stored at -80°C until detection. The diagnoses of HCC were confirmed by histopathological observation. The clinicopathologic data of the patients including age, gender, etiologies, liver cirrhosis status, serum α-fetoprotein (AFP) level, maximal tumor size, tumor cell differentiation (Edmondson- Steiner grade), tumor nodule number, capsular formation, vein invasion (including portal vein invasion, venous invasion or microscopic vessel invasion), and tumor node metastasis (TNM) stage were also collected. Follow-up data were obtained after hepatic resection for all 102 patients. The follow-up period was defined as the interval between the date of operation and that of patient’s death or the last follow-up. Deaths from other causes were treated as censored cases. Recurrence and metastasis were diagnosed by clinical examination, serial AFP level mensuration, and ultrasonography or computed tomography (CT) scan. Prior written informed consent of all patients involved in the study was obtained before operation and the protocol of the present study was approved by the Ethics Committee of Guangzhou Red Cross Hospital.
Cell culture and reagents
HCCLM3, Hep3B, and HepG2 liver cancer cell lines were purchased from GeneChem Technology (Shanghai, China) and all cultured at 37°C in a humidified incubator (SANYO, Osaka, Japan) with 5% CO2 by using high glucose DMEM (Dulbecco’s modified Eagle’s medium) containing 10% fetal bovine serum (FBS, Invitrogen, Frederick, MD, USA). The rabbit polyclonal antibody against human EGFL8, Notch1, NICD, Hes1 and Hey1 were purchased from Abcam (Cambridge, UK). The mouse monoclonal antibody against human GAPDH was purchased from Santa Cruz Biotechnology (Santa Cruz, CA, USA). The γ-secretase inhibitor N-[N-(3,5-Difluorophenacetyl- L-alanyl)]-(S)-phenylglycine t-butyl ester (DAPT, CAS: 208255-80-5) was also purchased from Abcam and the dimethylsulfoxide (DMSO) was purchased from Sigma (St. Louis, MO, USA).
RNA preparation and reverse transcription-quantitative polymerase chain reaction (RT-qPCR)
The total RNA of HCC tissues, adjacent non-tumorous liver tissues (ANLTs), liver cancer cell lines was extracted by using TRIzol reagent (Invitrogen, Carlsbad, CA, USA) as previously described [14-16]. The protocol of the first-strand cDNA generation and was also performed as described previously [16]. The quantitative polymerase chain reaction (qPCR) was consisted of 40 cycles after an initial denaturation step (95 °C for 3 min) and every cycle including 95 °C for 5 sec and then 60°C for 30 sec. As an internal control, β‑actin gene in the same samples was also detected. The primers for EGFL8 and β-actin qPCR amplification were designed by Primer Premier 5.0 software (Premier Biosoft International, Palo Alto, CA, USA). EGFL8: forward, 5’-CCCGCTCCACTACAACGAGT-3’; reverse, 5’-AACGCGGTACATGGTCCTGT- 3’. Beta-actin: forward, 5’-GCATGGGTCAGAAGGATTCCT-3’; reverse, 5’-TCG- TCCCAGTTGGTGACGAT-3’. All the qPCR analyses were performed in triplicate and the results were calculated by 2-ΔΔCt method. The relative expression of EGFL8 mRNA in tissue samples and liver cancer cell lines were normalized to β-actin and the EGFL8 expression was defined as down-regulation when the relative expression of EGFL8 in HCC tissue < the relative expression of EGFL8 in the corresponding ANLT tissue. Egfl8 expression in HCC specimens were also divided into high EGFL8 expression group (EGFL8 expression level ≥ median of EGFL8 expression levels in all HCC specimens) and low EGFL8 expression group (EGFL8 expression level < median of EGFL8 expression levels in all HCC specimens).
Western blot
Total protein of HCC tissues, ANLTs and liver cancer cell lines was extracted and separated by SDS-PAGE and subsequently transferred onto PVDF (Polyvinylidene Fluoride) membrane (Merck Millipore, Etobicoke, Ontario, Canada), which were incubated successively with primary antibody against EGFL8, Notch1, NICD, Hes1, Hey1 and the corresponding secondary antibody. GAPDH protein in the same samples was also detected as a loading control. The expression levels of EGFL8, Notch1, NICD, Hes1 and Hey1 proteins were normalized to GAPDH. Down-regulation of EGFL8 protein was defined as positive when the relative expression of EGFL8 protein in HCC tissue < the relative expression of EGFL8 protein in the corresponding ANLT tissue.
Lentivirus-mediated EGFL8 transfection
Full-length human EGFL8 cDNA was amplificated by polymerase chain reaction as described previously [17]. Then EGFL8 expression vector was generated by subcloning EGFL8 cDNA into the GV143 plasmid (purchased from GeneChem). The package and infection of lentivirus containing EGFL8 expression vector or empty vector were also described previously [17]. The HCCLM3 cells infected with lentivirus containing EGFL8 expression vector were named as HCCLM3EGFL8 and those HCCLM3 cells infected with lentivirus containing empty vector were named as HCCLM3Vector. And Western blot analysis was subsequently used to determine the expression levels of EGFL8 protein in HCCLM3EGFL8 and HCCLM3Vector cells.
Lentivirus-mediated short hairpin RNA (shRNA)
The GV248 shRNA expressing vector was purchased from GeneChem Technology (Shanghai, China). The sequences of shRNA targeting EGFL8 (selected from three putative candidate sequences, data not shown) were: sense: 5’-CCGGCAACCAGTGCCAGCATACTCACTCGAGTGAGTATGCT- GGCACTGGTTGTTTTTG-3’; anti- sense: 5’-AATTCAAAAAGAGTTGGTA- CTGGATCACATTCTCGAGAATGTGATCCAGTACCAACTC-3’. And the negative control shRNA sequences (not targeting any known gene) were: sense: 5’-CCGGTTCTCCGAACGTGTCACGTTTCAAGAGAACGTGACAC- GTTCGGAGAATTTTTG-3’; antisense: 5’-AATTCAAAAATTCTCCGAACGT- GTCACGTAAGTTCTCTACGTGACACGTTCGGAGAA-3’. The package and infection of lentivirus containing shRNA sequence targeting EGFL8 or negative control sequences were described previously [18]. The Hep3B cells infected with lentivirus containing shRNA sequence targeting EGFL8 were named as Hep3BshEGFL8 and those Hep3B cells infected with lentivirus containing negative control sequence were named as Hep3BshCtrl. The expression levels of EGFL8 protein in Hep3BshEGFL8 and Hep3BshCtrl cells were also detected by Western blot analysis.
Wound healing assay
Liver cancer cells were seeded on 96-well plates (3×104 cells/well). The cells were incubated for 24~48 h and when the confluency reached about 80%, a linear wound was made by a cell scraper (1.2 mm width) across the cell monolayer. Then the cells were washed twice and incubated with the high glucose DMEM without FBS. And photographs were taken at 0, 24 and 72 h by a microscope (MicroPublisher 3.3RTV, OLYMPUS, Tokyo, Japan). The 24h / 72h wound closure rate = (0h width of wound - 24h / 72h width of wound) / 0h width of wound × 100%. The scattered spherical cells in the scratch wound of the original image, which were scratched out or washed out by PDS and adhered again in the scratch wound after PDS washing, did not influence the assessment of the results. Experiments were carried out in triplicate.
Transwell invasion assay
Transwell invasion assay were carried out according to the same methods described previously [19]. These experiments were performed in triplicate.
MTT assay and cell apoptosis analysis
The protocol of MTT assay and cell flow cytometric for apoptosis analysis were also described previously [17, 18, 20].
HCC metastatic mouse model
A total of 20 nude mice, six to eight weeks old, were purchased from Tongji University Experimental Animal Center (Shanghai, China) and reared in cages under SPF (specific pathogen free) conditions at 21‑25˚C with 40‑70% humidity, 12/12 light cycles and free access to food and water. These mice were randomized into the HCCLM3EGFL8 group and the HCCLM3Vector group (10 mice/group). HCCLM3EGFL8 or HCCLM3Vector cells were subcutaneously inoculated into the mice in HCCLM3EGFL8 or HCCLM3Vector group on their right upper flank regions (1×107 cells/mouse). Since the subcutaneous tumors emerged from these mice, the volume (V) of tumors was measured every other day until the ninth measurement (in 16 days after the emergence of the tumor) and calculated by using the formula: V = a × b2/2, in which a means the largest tumor diameter and b means the smallest one. During the whole time of experiment, the health and behavior of these mice were observed every other day. If the tumor burden of a mouse was evaluated to be high or a mouse was found to be difficult to feed, this mouse would be euthanized early. However, no early euthanasia was actually executed. After the ninth measurement of the subcutaneous tumors, which was in 26~31 days after HCCLM3EGFL8 or HCCLM3Vector cells inoculation, all 20 mice were euthanized with pentobarbital sodium by intraperitoneal injection at a dose of 150 mg/kg and then the euthanasia was confirmed by cervical dislocation. Subsequently, the subcutaneous tumors were dissected out and cut into 4μm thick slices for hematoxylin and eosin (H&E) staining and immunohistochemistry. Lung of each mouse was made into serially sections and observed under a microscope after H&E stain as described previously [12]. Once metastatic HCC cells were found on any slide of lung sections of a mouse, which was considered as lung metastasis positive. All animal experiments were designed to minimize pain or discomfort to the animals and complied with the ARRIVE guidelines and the AVMA guidelines for the euthanasia of animals (2013 Edition). The animals were acclimatized to laboratory conditions (23 °C, 12 h/12 hlight/dark, 50% humidity, ad libitum access to food and water) for 2 wk prior to experimentation. All the experiment protocols involving animals were reviewed and approved by the Ethics Committee of Guangzhou Red Cross Hospital and this study was carried out in compliance with the ARRIVE guidelines.
Immunohistochemistry assay
As described previously, the expression of EGFL8 and Notch1 in mouse xenograft tumors was determined by immunohistochemistry (IHC) assay [12, 17, 18]. Briefly, the sections made from xenograft tumors were deparaffinized, rehydrated and incubated with 3% H2O2. Then, these sections were put in citrate buffer (0.01M, pH 6.0) and heated by a microwave oven at high power for two times, each for 7 minutes, for antigen retrieval. After rinsing with PBS, the sections were incubated successively with primary antibodies and HRP-conjugated second antibodies. Finally, the sections were visualized by using 3,3-diaminobenzidine tetrahydrochloride (DAB) and counterstained with hematoxylin.
DAPT treatment
To verify the role of Notch signaling pathway in the biological effects of EGFL8 on liver cancer cells, EGFL8 depleted Hep3B cells were treated with DAPT (50 μM, dissolved by DMSO), an inhibitor of Notch signaling pathway, for 48 h [21].
Statistics analyses
The expression levels of EGFL8 in HCC tissues were compared with those in ANLTs by Mann-Whitney test. Data come from three or five independent experiments are presented as mean ± SEM (standard error of mean). The differences between two groups were examined by unpaired t test. The comparisons of multi-group were performed by one-way analysis of variance (ANOVA) with Tukey test as post hoc test. The growth curves of liver cancer cells and xenograft tumors were compared by ANOVA with Tukey test. Survival curves were constructed using the Kaplan-Meier method and evaluated using the Log-Rank test. These analyses were all completed by using Graphpad Prism 7.0 software (Graphpad Software, La Jolla, CA, U.S.A.). All analyses were two-sided and P < 0.05 was considered as significant.