Human hepatocellular carcinoma cell line HUH-7 was purchased from Vacsera (Giza, Egypt), sorafenib p- Toluenesulfonate salt was purchased from LC Laboratories (Woburn, MA, United States), Avastin^® (bevacizumab, Genentech, United States) was purchased, in its formulated commercial preparation, from a community Pharmacy (Cairo, Egypt) and fucoidan extracted from Laminaria Japonica was purchased as a crude extract from Buchem BV (Holland). Dulbecco’s modified Eagle’s medium (DMEM) medium and fetal bovine serum (FBS) were purchased from Gibco^®, Thermo Fisher Scientific, United States. Antibiotic-antimycotic mixture (100 U/ml of penicillin, 0.1 U/ml streptomycin and 0.25 μg/ml of Amphotericin B) was purchased from Lonza^®, Walkersville, MD, United States. All other chemicals were purchased from Sigma Aldrich, United States unless otherwise specified.

HUH-7 cells were maintained in culture media consisting of DMEM supplemented with 10% fetal bovine serum (FBS), 100 U/ml of penicillin, 0.1 U/ml streptomycin, 0.25 μg/ml of Amphotericin B at 37°C in a humidified incubator containing 5% CO₂. All cell culture procedures were performed in a class II laminar flow hood and incubations were done inside the incubator at 37°C, unless otherwise specified.

Cytotoxicity was evaluated using the MTT (3-(4,5-Dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide) cell viability assay. 20,000 cells/well were seeded in a 96-well plate and left to attach overnight (o/n). Next day, the cells were treated with increasing concentrations of Avastin (0.94–30 µM), sorafenib (0.3–20 µM) and fucoidan (0.15—5 mg/ml) for 72 h. Next, media was discarded and replaced with 100 μl/well of fresh DMEM containing 12 mM (5 mg/ml) MTT stock in complete media and further incubated for 2 h inside the incubator at 37°C. Finally, the media was discarded and replaced with 100 μl of DMSO to dissolve the formed formazan crystals and then the absorbance was measured at 570 nm. Inhibitory concentration 50 (IC₅₀) was calculated using GraphPad Prism Software using the non-linear regression analysis.

For combination therapies, cells were seeded as previously described then next day treated with increasing concentrations of Avastin (0.94–30 µM), sorafenib (0.3–20 µM) either alone or in combination with IC₃ of fucoidan (55 μg/ml) for 72 h. The pharmacological interaction between fucoidan and sorafenib or Avastin when used in combination was evaluated by applying the isobologram equation shown below to calculate the combination indices (Chou, 2006). C o m b i n a t i o n   i n d e x   C I = D 1   / D n 1 + D 2 / D n 2 Where (D_n)1 and (D_n)2 are the concentrations of each drug alone to exert n% effect, while (D)1 and (D)2 are the drug concentrations when used in combination to exert the same effect. The pharmacological interaction of the combination is estimated to be synergistic if CI < 0.8; additive if CI ranges from 0.8 to 1.264 or antagonistic if CI > 1.2.

Briefly, 10⁶ HUH-7 cells were seeded in six 6-well plates and allowed to attach overnight (o/n). Once cells reached confluency, the cell monolayer was scratched using a 200 μL pipette tip held vertically followed by washing twice with PBS to remove floating cells. The cells in each plate were then treated with complete DMEM medium alone (untreated control) or added to it either 5 µM sorafenib, 25.22 µM Avastin, 55 μg/ml fucoidan or the combinations of sorafenib and fucoidan (S + F) or Avastin and fucoidan (A + F). The wound area was imaged on day 0, 1, 2, 3 and 4 post treatment using an inverted microscope (Labomed Inc., LA, CA, United States) connected to a digital camera. Wound width was calculated using a wound healing size plugin for ImageJ^® (NIH, United States) as described by (Suarez-Arnedo et al., 2020).

HUH-7 cells were seeded in T-25 flasks and allowed to attached o/n. Next day, cells were treated with 5 µM sorafenib, 25.22 µM Avastin, 55 μg/ml fucoidan or the combinations of sorafenib and fucoidan (S + F) or Avastin and fucoidan (A + F) for 72 h. Next, cells were detached using trypsin, rinsed with cold 1x PBS, centrifuged twice at 280 x g for 5 min, resuspended in cold PBS and kept on ice until analysis. To a 100 µl aliquot of each cell suspension, 1 µl of PI stock (prepared at 100 μg/ml) and 5 µl Annexin V-FITC were added for 15 min at room temperature in the dark. Finally, 400 µl of 1x Annexin binding buffer was added to each sample and analyzed by CytoFlex flow cytometer (Beckman Coulter, CA, United States) according to the manufacturers’ instructions. A minimum of 10,000 events were recorded for each sample. Data analysis was done using FlowJo software (Treestar Inc., San Carlos, CA, United States).

Cells were grown in six 6-well plates and each plate was subsequently treated with either 5 µM sorafenib, 25.22 µM Avastin, 55 μg/ml fucoidan or the combinations of sorafenib and fucoidan (S + F) or Avastin and fucoidan (A + F) for 48 h. Next, cells were harvested, washed with cold 1x PBS, centrifuged twice at 280 x g for 5 min and resuspended in cold PBS and kept on ice until analysis. Cell pellets were fixed by being re-suspended in 2 ml of 60% ice-cold ethanol at 4°C for 1 h. Fixed cells were subsequently washed at least twice with PBS (pH 7.4). The cell pellet was then re-suspended in 1 ml of nuclei acid staining mixture (10 μg/ml propidium iodide (PI) and 50 μg/ml RNAase A in PBS) for 20 min in the dark at 37°C. Cells were analyzed for DNA contents using flow cytometry analysis using CytoFlex flow cytometer (Beckman Coulter, CA, United States) according to the manufacturers’ instructions. For each sample, 10,000 events were acquired. Cell cycle distribution was calculated using CytExpert software (Beckman Coulter, CA, United States).

HUH-7 cells were seeded in T-75 flasks overnight at a seeding density of 5 × 10⁶ cells/flask. Next day, the media was discarded and replaced with DMEM medium alone (untreated control) or added to it either 5 µM sorafenib, 25.22 µM Avastin, 55 μg/ml fucoidan or the combinations of sorafenib and fucoidan (S + F) or Avastin and fucoidan (A + F) and the cells were incubated for 72 h. Next, total RNA was isolated using RNeasy^® Kits (Qiagen, Hilden, Germany) according to the manufacturer’s protocol. RNA samples were then assessed to detect purity by measuring the absorbance of the RNA samples using NanoDrop Spectrophotometer (BMG LABTECH, Ortenberg, Germany) at 260 nm (ng/μl) and calculating the A260/280 ratio. cDNA was synthesized using the Revertaid cDNA synthesis kit (K1621; Thermo Fisher Scientific, MA, United States), according to the manufacturer’s instructions. Primers used were purchased from Thermo Fisher (MA, United States). Gene expression levels were calculated as follows: 2^−ΔΔCT ± standard deviation. Full list of primers’ sequences is shown in Table 1.

HUH-7 cells were seeded in T-75 flasks overnight at a seeding density of 5 × 10⁶ cells/flask. Next day, the media was discarded and replaced with DMEM medium alone (untreated control) or added to it either 5 µM sorafenib, 25.22 µM Avastin, 55 μg/ml fucoidan or the combinations of sorafenib and fucoidan (S + F) or Avastin and fucoidan (A + F) and the cells were incubated for 72 h. Next, cells were collected via trypsinization and then washed twice with cold PBS and re-pelleted. The cell pellet was resuspended in 100 μl cell lysis buffer (1 ml of RIPA lysis buffer (Thermo Scientific, United States, Catalogue number: 89,900) + 10 µl of HALT™ protease inhibitor (Thermo Scientific, United States, Catalogue number: 78,410) + 10 µl of HALT™ phosphatase inhibitor (Thermo Scientific, Catalogue number: 78,420) and incubated for 30 min on a rotor at 4°C. Next, the samples were centrifuged at 280 x g for 10 min at 4°C and the supernatants were transferred to fresh tubes and placed on ice. Total protein concentration was measured using BCA protein assay as per the manufacturer’s protocol (Thermo Scientific, United States). The samples were stored at −80°C until processing.

Total protein was quantified in whole cell lysates as described above; all samples were diluted with dH₂O to contain 1 mg/ml total protein. 80 µl of each sample (whole cell lysate) was used in each ELISA and the manufacturer’s protocol was followed. All ELISA kits used were from Invitrogen, Thermo Scientific, United States as follows: human VEGF ELISA kit (catalogue number KHG0111); human caspase 3 (active) ELISA kit (catalogue number KHO1091), human caspase 8 ELISA kit (catalogue number BMS2024) and human caspase 9 ELISA kit (catalogue number BMS2025).

All animal procedures met the Animals (Scientific Procedures) Act 1986/ASPA Amendment Regulations 2012 and were done after the approval of the Ethical Review committee at the British University in Egypt (Ethics approval number: EX-2218).

46 Sprague Dawley male rats, 5–6 weeks old (100–150 g) were injected intraperitonially (I.P) with 50 mg/kg DEN dissolved in saline once a week for 16 weeks (Schiffer et al., 2005). At the beginning of the 17^th week, rats were randomly divided into 6 groups (n = 7–8): Group 1: Untreated control, group 2: Received 200 µg per rat Avastin I.P once per week. Group 3: Received sorafenib at 10 mg/kg via oral gavage once per day for five consecutive days. Group 4: Received fucoidan at 20 mg/kg I.P once per day for five consecutive days while Group 5 received the combination of Avastin and fucoidan and Group 6 received the combination of sorafenib and fucoidan. Treatments were continued for 4 weeks then mice were sacrificed by cervical dislocations. Blood and tissue samples were collected and stored appropriately until further analysis.

Alanine aminotransferase (ALT) and aspartate aminotransferase (AST) were assessed in the serum samples taken from rats using colorimetric commercial kits from Spectrum diagnostics, Cairo, Egypt as per the manufacturer’s protocol. Serum alpha fetoprotein (AFP) was measured using rat AFP ELISA kit (Elabscience, United States, catalogue number E-EL-R01047) as per the manufacturer’s protocol.

For histopathological examination, liver tissues were preserved in 10% formalin solution for at least 48 h. Samples were then rinsed with tap water and treated with serial dilutions of methanol, ethanol and absolute ethanol for tissue dehydration followed by xylene. Tissue samples were embedded in paraffin wax and kept at 56°C for 24 h in a hot air oven. Next, the tissue blocks were sliced into 4 μm thick sections and placed on glass slides, deparaffinized, subsequently stained with hematoxylin & eosin (H&E) and examined with light microscopy (Banchroft et al., 1996).

Liver tissue homogenates extracted from rats of the different treatment groups were prepared and evaluated for the levels of VEGF, VEGFR and AFP using ELISA kits as per the manufacturer’s protocol. ELISA kits used were rat VEGFR1/FLT1 ELISA kit (Elabscience, United States, catalogue number E-EL-R011108); rat VEGF ELISA kit (Creative Diagnostics, United States, catalogue number DEIA1173) and rat AFP ELISA kit (Elabscience, United States, catalogue number E-EL-R01047).

Avidin-Biotin immunoperoxidase complex technique (ABC) was used (Hsu et al., 1981). Briefly, paraffin embedded tissues were deparaffinized and rehydrated through graded alcohol series. Antigen retrieval was done by incubating tissue sections in an antigen retrieval citrate buffer in a microwave oven for 5 min at 700 w, then sections were left to cool. To quench endogenous peroxidase activity, 2 drops of peroxidase blocking serum was added for 10 min, then slides were rinsed with PBS (pH 7.4). Next, two drops of protein blocking serum were added for 10 min followed by incubation with either of the following primary antibodies: caspase-3 rabbit polyclonal antibody (catalogue number A2156, ABclonal, Woburn, MA, United States), CD34 rabbit monoclonal antibody (catalogue number A19015, ABclonal, Woburn, MA, United States) or Ki67 rabbit monoclonal antibody (Clone QR015, Quartett, Berlin, Germany) for 30 min at room temperature. Next, pre-diluted biotinylated secondary antibody was added to each section for 45 min followed by rinsing with PBS. Horseradish peroxidase conjugated streptavidin was added for 20 min followed by rinsing with PBS. Substrate/chromogen (DAB) mixture (which was prepared immediately before use) was added and slides were incubated for 5–10 min followed by rinsing with dH₂O. Finally counterstaining was done with Harris’s hematoxylin and sections were dehydrated with graded alcohol series, cleared in xylene, and finally mounted by DPX. Sections were imaged using Leica Application software for tissue sections analysis (Leica Microsystems GmbH, Germany).

Experimental results were plotted as mean values ±standard error of mean (SEM) unless otherwise specified. Statistical analysis for multiple experimental groups was done using one-way analysis of variance (ANOVA) followed by post hoc Tukey-Kramer test (unless otherwise specified) with p values considered statistically significant if less than or equal 0.05. All statistical analyses and data plotting were done using GraphPad Prism software, version 5.00 (GraphPad Software, Inc. La Jolla, CA, United States). All experiments were repeated at least 3 times with 3-6 replicates per treatment. Representative data is shown.

HUH-7 cells were treated with various concentrations of Avastin, sorafenib and fucoidan as single therapies for 72 h. Inhibitory concentrations 50 (IC₅₀) were calculated for each drug individually (data not shown). Next, fucoidan was combined with the anti-angiogenic drugs and results showed a reduction in the IC₅₀ of sorafenib from 4.9 to 0.4439 µM and Avastin from 25.22 to 11.55 µM when combined with fucoidan with a calculated combination indices of 0.2 and 0.6, respectively. These findings suggest a synergistic interaction between the anti-angiogenic drugs and fucoidan (Chou, 2006). (As shown in Figure 2)

The ability of sorafenib, Avastin and fucoidan and their respective combinations to alter HUH-7 cells migration was analyzed via the scratch wound healing assay.

Throughout the span of the experiment (4 days); sorafenib, A + F and S + F treated cells consistently showed an unhealed wound indicating the inhibition of cancer cell proliferation and migration unlike the untreated control and fucoidan treated cells where the wound rapidly disappeared throughout the span of the experiment (Figure 3). While Avastin treated cells had a small wound remaining on day 1 and 2 with an almost completely healed wound on day 4. The group S + F followed by sorafenib then A + F showed the best results with almost a slightly smaller wound area observed on day 4 compared to day 0. Next, wound width was quantified using ImageJ software and a plugin designed by (Suarez-Arnedo et al., 2020). Sorafenib, A + F or S + F treated cells consistently showed a significantly smaller % wound closure (sorafenib: 59%–63%; A + F: 69%–70% and S + F: 51%–58%) throughout the incubation period versus the untreated control in which the wound almost completely healed (91%–100%) (p < 0.05, two-way-ANOVA, Tukey’s post hoc).

To examine the role of apoptosis in the cytotoxic effect of the drugs, the percentage of apoptotic cells was detected via Annexin V/PI staining that was measured by flow cytometry. In the untreated control and Avastin, 98.1% and 97.66% of the cell population was alive with no significant difference between both groups. Sorafenib monotherapy caused a significant increase in the percentage of cells in late apoptosis (0.52%) and necrosis (6.6%) and a significant decrease in the percentage of live cells (92.8%) compared to untreated control (p < 0.0001). While fucoidan, A + F and S + F showed similar patterns with a significant decrease in live cells to 91.6%, 94.66%, and 92.83%, respectively (p < 0.0001) with an increase of necrotic cells to 8.3%, 4.4%, and 6.9%, respectively (p < 0.0001) compared to untreated control. The combination of A + F was significantly better than the Avastin monotherapy as it showed higher percentage of cells in early apoptosis and necrosis. While the S + F combination showed very similar findings to sorafenib monotherapy (Figure 4).

We used this assay to examine the effect of the different drugs on the progression of the cell cycle. Results (Figure 5) revealed no significant differences between the untreated control and sorafenib treated cells with around 74% of the cells in the G0/G1 phase, 21% in S-phase and 3.5% in the G2/M phase. In the Avastin monotherapy group, significantly higher portion of cells were arrested in the G2/M phase compared to the untreated control (7% versus 3.5%, p < 0.05) while 70.4% and 21.7% were detected in the in G0/G1 and S-phases, respectively. In the fucoidan, A + F and S + F treated cells the pattern was different, with more cells arrested in the S-phase and G2/M phases compared to the untreated control. This pattern appears to be induced by fucoidan as both A + F and S + F had a significantly higher percentage of cells in the S-phase (26.6% and 23.1%, respectively) and G2/M phase (12.9% and 8.2%, respectively) versus the monotherapies (p < 0.05).

The effect of the drugs was investigated on the three caspases that are the key players in both the intrinsic and extrinsic pathways of apoptosis (Van Opdenbosch and Lamkanfi, 2019). ELISA results (Figure 6) revealed that all drugs and their combinations, except Avastin monotherapy, significantly increased caspases 3 and 8 with S + F group showing 40- and 16-times higher protein levels, respectively (p < 0.05, versus untreated control). While for caspase 9, a similar pattern was observed but to a lesser extent.

Next, we evaluated the effect of the drug alone and in combinations on the key pro-angiogenic pathways: VEGF (using ELISA) and its downstream signaling molecules, PI3K/AKT/mTOR and KRAS/BRAF/MAPK pathways via mRNA expression levels by RT-qPCR. Fucoidan, sorafenib, A + F and S + F significantly reduced the expression of the pro-angiogenic PI3K/AKT/mTOR and KRAS/BRAF/MAPK pathways by up to 3 folds (p < 0.05 versus untreated control). The most affected were PI3K and KRAS with a significantly decreased mRNA levels of 1.5 and 2.4 folds with A + F or 2.6 and 3.2 folds with S + F, respectively (p < 0.05 versus untreated control) (see Figure 7).

As previously described in the methodology, rats were given weekly injections of DEN for 16 weeks then treated for a month with the different drugs and respective combinations. At the end of the experiment, rats were sacrificed, and blood and tissue samples were collected (summary of experimental timeline is shown in Figure 8).

Assessment of liver function was done by measuring aspartate aminotransferase (AST), alanine aminotransferase (ALT) and alkaline phosphatase (ALP) in the rats’ serum samples using commercially available colorimetric kits. While Alpha fetoprotein (AFP), a key marker of tumour burden and prognosis (Morse et al., 2019), was assessed using ELISA. Results (Figure 9) showed that the untreated controls had the highest level of ALT, AST, ALP and AFP indicating liver damage and high tumour burden. While all treated groups showed significantly lower levels of these biomarkers compared to the untreated control with no significant difference amongst the different treatments.

Next, the histopathological changes induced by DEN was evaluated using hematoxylin and eosin staining (H&E) done on livers, kidneys and lungs tissue sections. All DEN treated rats developed at least cirrhosis with 25 out of 35 rats progressing to HCC, therefore, the success rate of this model was 71.4%. Both acinar (grade I) (Figures 10A, B) and trabecular (grade II) (Figures 10C, D) HCC nodules were observed within the liver tissues with acinar HCC nodules being the most common. Metastatic tumour masses were also observed in the lungs and kidneys of DEN-treated rats regardless of the treatment given (Figure 11).

Upon evaluation of the tumour sections of the untreated control, we observed large nodules of HCC with markedly pleomorphic cells accompanied by prominent nucleoli, scattered apoptosis, mild inflammatory infiltrate, mildly congested blood vessels and marked areas of hemorrhage and inflammatory infiltrate. While smaller tumour nodules as well as marked areas of apoptosis and necrosis were observed in the tumour nodules of rats treated with the different drugs with more prominent apoptotic and necrotic areas observed in the combination treated groups (Figure 12A).

Liver tissues extracted from DEN HCC model were analyzed using ELISA for the levels of AFP, VEGF, and VEGFR.

AFP, a key marker of liver tumour burden (Morse et al., 2019), was assessed in the different treatment groups. Results (Figure 12B) revealed that Avastin and sorafenib monotherapies significantly decreased the levels of AFP while fucoidan was similar to the untreated control. Although the combination therapies decreased the levels of AFP, the results were not statistically significant compared to the untreated control.

The key angiogenic protein VEGF and its receptor VEGFR were also assessed (Figure 12B). In both instances Avastin, sorafenib and A + F significantly reduced their levels compared to the untreated control. While S + F also markedly reduced their levels, but the results were insignificant. Surprisingly, for VEGF, fucoidan alone slightly but significantly increased its levels compared to the control unlike VEGFR as fucoidan had very similar levels to the control.

Liver tissues were stained for the key apoptotic executioner caspase 3, one of the main proliferation markers (Ki67) (Morse et al., 2019) and CD34, the highly sensitive marker for endothelial cells and is generally used as a marker of angiogenesis in tumours (Vieira et al., 2005).

Caspase 3 reactivity was predominantly cytoplasmic with some nuclear staining. The interpretation of the results considered both the staining intensity (Figure 13A) and the percentage of positive cells (Figure 13B). The reactivity was classified as: negative (0), weak (+), moderate (++) or marked (+++). In the untreated control, liver showed moderate cytoplasmic reactivity (++) for caspase-3 in cirrhotic and tumor nodules. The most significant findings were observed in rats treated with Avastin and S + F, liver tissues showed moderate cytoplasmic reactivity (++) for caspase-3 in liver tissue and marked reactivity (+++) in tumor nodules. While Avastin, fucoidan and A + F were very similar to the untreated control with moderate cytoplasmic reactivity (++) for caspase-3 in tumor nodules and sorafenib alone showing weak reactivity (+) in tumor nodules. The results indicate that the combination of S + F was significantly better than sorafenib alone, but an opposite pattern was observed for Avastin (Figure 13B).

Ki67 positivity was evaluated according to percentage of positive cells into four degree: (-) < 24%, (+) 25%–50% (Isolated), (++) 51%–74% (Focal) or (+++) > 75% (Diffuse) (Mocanu et al., 2012). In the untreated control, liver showed focal reactivity (++) for Ki67 in cirrhotic and tumor nodules. This pattern was reduced in all treatment groups to various degrees. In Avastin treated groups, liver tissues showed isolated reactivity (+) for Ki67 in cirrhotic nodules and negative reactivity (-) in tumor nodules. For sorafenib, liver showed isolated reactivity (+) for Ki67 in cirrhotic and tumor nodules. In fucoidan, negative reactivity (-) for Ki67 in tumor nodules was detected. In the combination groups, A + F liver showed isolated reactivity (+) for Ki67 in tumor nodules, while S + F liver showed negative reactivity (-) for Ki67 in cirrhotic nodules, and isolated reactivity (+) in tumor nodules (as shown in Figure 14A).

Finally, for CD34 staining, the number of positive cells was evaluated as follows: (0 = 0%, 1 = 1–25%, 2 = 26–50%, 3 = 51–75%, and 4 = 76–100% of cells) as described by (Kamat et al., 2006). Untreated control liver showed moderate reactivity (2) for CD34 in cirrhotic and tumor nodules while in the treated groups weak or negative reactivity was detected. For Avastin and sorafenib weak reactivity (1) for CD34 was observed in cirrhotic and tumor nodules. Fucoidan showed negative reactivity (0) for CD34 in tumor nodules. For the combinations, A + F liver showed negative reactivity (0) for CD34 in cirrhotic nodules, and weak reactivity (+) in tumor nodules while S + F showed an opposite pattern with weak reactivity (+) for CD34 in cirrhotic nodules, and negative reactivity (0) in tumor nodules (as shown in Figure 14B).