A Novel Zebrafish Liver‐Specific Metastasis Model Reveals c‐Met as a Driver of Liver Tropism
Merve Basol, Peyda Korhan, Helin Ozaktas, Nese Atabey, Gulcin Cakan‐Akdogan
TL;DR
Researchers created a new zebrafish model to study liver metastasis and found that high c-Met expression in cancer cells increases their spread to the liver.
Contribution
The novel zLiverMet model enables visualization of liver metastasis in zebrafish and identifies c-Met as a key driver of liver tropism.
Findings
Cancer cell lines with high c-Met expression showed strong liver tropism in the zLiverMet model.
Pharmacological inhibition of c-Met reduced liver colonization by cancer cells.
The zLiverMet model successfully mimics intrahepatic metastasis and offers a scalable platform for drug testing.
Abstract
Intrahepatic metastasis negatively impacts the prognosis of several cancers, including hepatocellular and colorectal carcinoma. Zebrafish larval xenografts serve as a robust vertebrate platform that allows direct visualisation of tumour behaviour within a living organism. However, organ‐specific metastasis models in zebrafish remain limited, and liver metastasis has not yet been demonstrated. This study aimed to establish a zebrafish larval intrahepatic metastasis model and to determine the role of c‐Met activation in mediating liver tropism of cancer cells. An intravenous injection–based zebrafish model (zLiverMet) was developed using a liver‐specific fluorescent reporter line to visualise tumour colonisation in vivo. Liver cancer cell lines with distinct c‐Met expression and activation levels were injected into 2‐day post‐fertilisation larvae. The effects of c‐Met overexpression and…
Genes, proteins, chemicals, diseases, species, mutations and cell lines named across the full text — each resolved to its canonical identifier and authoritative record.
Click any figure to enlarge with its caption.
FIGURE 1
FIGURE 2
FIGURE 3
FIGURE 4
FIGURE 5| Cell line | Origin | Morphology | Differentiation status | c‐Met level | zLiverMet rate |
|---|---|---|---|---|---|
| SNU‐398 | HCC | Mesenchymal | Poor | +/− | +/− |
| HuH‐7 | HCC | Epithelial‐like | Well | + | + |
| SNU‐449 | HCC | Mesenchymal | Poor | ++ | ++ |
| SK‐HEP‐1 | Liver AC | Mesenchymal | Poor | +++ | +++ |
| Mahlavu | HCC | Epithelial | Poor | +/− | +++ |
- —European Commission10.13039/501100000780
- —Türkiye Bilimsel ve Teknolojik Araştırma Kurumu10.13039/501100004410
Peer Reviews
No public reviews on file for this paper yet. If you reviewed it on a platform where reviews are public (OpenReview, ICLR, NeurIPS, ICML), you can paste yours below so the community can read it here.
Videos
No videos yet. Explain this paper in a talk, walkthrough, or lecture? Add one.
Taxonomy
TopicsZebrafish Biomedical Research Applications · Liver physiology and pathology · Developmental Biology and Gene Regulation
Introduction
1
More than 90% of the cancer‐related mortalities are attributable to tumour invasion and metastasis [1, 2]. The liver is one of the most common metastatic targets for various cancer types, including hepatocellular carcinoma (HCC) and colorectal cancer (CRC) [2, 3]. Intrahepatic metastases significantly worsen the survival of patients [4, 5, 6, 7, 8]. Cancer cells that complete the metastasis cascade can induce secondary tumours. This multistep process begins with the local invasion of malignant cells into the surrounding tissue and continues with the intravasation of the cancer cells into the vasculature. Cells that withstand the hemodynamic shear stress and evade immune surveillance are termed circulating tumour cells (CTCs) [9]. Intrahepatic metastases require arrest of CTCs within liver sinusoids, adhesion to and extravasation through sinusoidal endothelium. Moreover, interaction with liver‐resident cells is needed to create a supportive niche for survival and colonisation [5, 6].
To recapitulate these processes, cell culture methods have been used to develop liver‐on‐a‐chip models incorporating endothelial and hepatic cell types on microfluidic platforms to mimic vascular and hepatic microenvironments [10, 11, 12, 13]. However, modelling liver metastasis in vitro remains challenging due to difficulties in generating physiologically relevant shear stress, reproducing the complex multicellular tumour microenvironment and modelling extravasation and metastatic colonisation. In vivo models therefore provide a more physiologically relevant approach for studying liver metastasis. Rodent xenograft models have long been employed to assess the metastatic capacity of cancer cells [14, 15]. Intrahepatic metastasis mouse models established through orthotopic implantation or splenic injection of HCC cells have been described [16, 17]. However, the complexity of surgical procedures, incompatibility with whole organism live‐imaging and ethical restrictions on sample size are among the disadvantages of rodent models.
Zebrafish larval xenograft models offer a robust alternative to study cancer metastasis in a small, essentially transparent organism amenable to imaging, drug treatment and screening [18]. Transplantation of fluorescently labelled cancer cells into the yolk sac of 2 days post‐fertilisation (dpf) embryos induces vascular invasion and dissemination of metastatic cells within 4 days, providing a considerable time advantage over rodent models [19, 20]. Zebrafish larval xenografts have been developed for essentially all cancer types, including patient‐derived models used for preclinical drug resistance testing [21, 22]. However, organ‐specific metastasis has been explored in only a few zebrafish studies. Reinhardt et al. demonstrated that MDA‐MB‐231 breast cancer cells injected to the yolk sac of the zebrafish larvae metastasized to the head region preferentially [23]. Paul et al. showed that intravenously injected MDA‐MB‐231 sublines with established brain‐ or bone marrow‐tropic phenotypes retained their organ‐specific preferences within the zebrafish larval body [24]. Whether zebrafish larval organs can serve as organ‐specific metastatic targets analogous to human tissues remains an open question. To date, intrahepatic metastasis and liver colonisation have not been demonstrated in zebrafish larvae.
c‐Met is a receptor tyrosine kinase with multiple functions throughout embryonic development, organogenesis and wound healing, and is located on the cell surface of epithelial cells, including hepatocytes [25, 26]. The ligand of c‐Met is hepatocyte growth factor (HGF), which is secreted by the mesenchymal cells [25]. Upon HGF binding, c‐Met undergoes homodimerization and activation through trans‐phosphorylation of two critical tyrosine residues (Y1234 and Y1235) within its catalytic domain, triggering downstream signalling pathways that regulate proliferation, motility, migration and invasion [27]. Aberrant activation of c‐Met promotes aggressive tumour phenotype and therapy resistance via increased cell scattering, dissociation, invasion and growth [28]. Previously we reported that aberrations in HGF/c‐Met signalling are crucial for the aggressive behaviour of HCC [29, 30, 31]. More recently, we showed that overexpression and activation of c‐Met direct HCC cells preferentially toward hepatocyte‐rich microenvironments in a lab‐on‐a‐chip metastasis model [32]. Conversely, inhibition of c‐Met phosphorylation using a small‐molecule inhibitor reduced extravasation and homing capacity, underscoring c‐Met's role in liver‐specific colonisation by HCC cells [32]. However, the role of c‐Met in regulating liver tropism has not been studied in an in vivo model.
Here, an intravenous injection–based zebrafish larval model (zLiverMet) that enables real‐time visualisation of intrahepatic metastasis in vivo was established. To our knowledge, this is the first zebrafish model that recapitulates intrahepatic metastasis by CTCs. Using the zLiverMet model, activation of the receptor tyrosine kinase c‐Met was found to promote liver tropism of cancer cells. Functional modulation of c‐Met through ectopic overexpression and pharmacological inhibition confirmed its role as a metastatic driver. Finally, application of the zLiverMet model to colorectal cancer (CRC) cells demonstrated its versatility for investigating intrahepatic metastasis across tumour types and for preclinical testing of targeted therapies.
Results
2
A Zebrafish Larval Xenograft Model to Study Liver Metastasis
2.1
Zebrafish liver reporter transgenics that express mCherry in hepatocytes were generated, using the fabp10a promoter (Figure 1A). Liver vascularization was visualised in the double transgenic larvae obtained by crossing to Tg(fli1:EGFP) transgenics (Figure 1B) [33]. The gut was also detected as a highly vascularized organ that emerged as a neighbouring organ to the liver (Figure 1C).
Liver and gut are fully vascularized in 6 dpf larvae. (A) Liver (red) and cardiac muscles (green) express mCherry and GFP, respectively, in Tg(fabp10:mCherry, cmlc2:GFP) zebrafish larvae. Overlay with the transmission image is displayed, and the gut‐intestine area is encircled with a dashed white line. Scale bar: 200 μm. (B) Hepatocytes (red) and vasculature (green) are visualised at 6 dpf Tg(fabp10:mCherry, cmlc2:GFP), Tg(fli1a: GFP) double transgenic line. The stem of the pectoral fin is marked with a dashed yellow line. GFP channel and Cherry channel images are displayed in the middle and right, respectively. Scale bar: 80 μm. (C) A close‐up image of the gut vasculature is displayed, marked with a dashed white line. Scale bar: 80 μm.
The liver reporter transgenic larvae were used to determine liver localization of cancer cells in xenografts. First, the effectiveness of the yolk sac microinjection and the intravenous injection (IV) approaches were compared for generation of a liver metastasis xenograft model (Figure 2A) [34, 35]. When hepatocellular carcinoma cell lines HuH‐7 and Hep3B were injected into the yolk sac, 6.30% ± 3.37% and 20.2% ± 3.79% of the cells metastasized to the liver, respectively. On the other hand, when cells were transplanted to the vasculature, liver colonisation rate increased to 18.30% ± 6.52% and 79.0% ± 2.5% for HuH‐7 and Hep3B, respectively (Figure 2B). Representative images of xenografts generated via yolk or IV injection and localization of metastatic cells were imaged (Figure 2C–F). Non‐metastatic tumours in the yolk‐injected group populated the yolk area without any invasion of the body (Figure 2C). Xenografts with at least 5 cells colonising the liver were considered metastatic. Metastasis to the gut was also observed in some larvae; however, this was not quantified (Figure 2D). In the IV injected larvae, colonisation to the liver or both liver and gut were observed (Figure 2E,F). The cancer cells colonised in the liver were also detected in histopathological sections of the larvae (Figure 2G–I).
*zLiverMet model allows observation of liver metastasis and colonisation. (A) Graphic representation of methodology. Xenografts were generated by injection of DiO stained (green) cancer cells either into yolk sac or vasculature of 2‐days post‐fertilisation (dpf) embryos. The metastasis rates are quantified at 6 dpf, 4 days post injection (created with BioRender.com). (B) Percentages of larvae with liver metastasis were quantified and displayed on bar graphs. For each cell line, at least 2 independent experiments were conducted, and average rates ± STD were plotted. Each independent experiment had at least 3 replicates. *p ≤ 0.05, **p ≤ 0.01, ***p ≤ 0.001, ***p ≤ 0.0001. (C–F) Representative images GFP channel (top), mCherry_GFP overlay (bottom) of xenografted larvae at 6 dpf. Cells (green), liver (red), heart (green, indicated with white star). Liver and gut are encircled with white and yellow dashed lines, respectively. (C, D) Yolk injected larvae. (C) Non‐metastatic tumours appeared as a mass at the injection site. (D) Metastatic cells invaded liver and/or gut. (E, F) IV injected larvae. The cancer cells delivered via IV injection are colonised into (E) liver or (F) liver and gut at 6 dpf. Scale bars: 200 μm. (G–I) H & E‐stained paraffin sections show liver (left) and close‐up images of the hepatocytes (right) of (G) Uninjected, (H) HuH‐7 injected and (I) Hep3B injected zebrafish larvae. Scale bars: 40 μm. The zebrafish liver is marked with a white dashed line and the tumour‐host liver border is marked with a black dashed line; some cancer cells are indicated with a black star.
The IV xenograft method, referred to as zLiverMet from here on, was chosen over the yolk xenograft method to test liver‐specific metastasis in the zebrafish larvae. Next, it was tested whether the zLiverMet model could be used for the detection of varying degrees of liver‐specific metastasis by transplanting liver cancer cells with different metastatic capacities.
Cellular c‐Met Levels of HCC Cell Lines Correlated With Liver Tropism in the zLiverMet Model
2.2
The HCC cell lines SNU‐398, HuH‐7, SNU‐449 and Mahlavu, as well as the liver adenocarcinoma cell line SK‐HEP‐1, were selected to test the strength of the zLiverMet model in the determination of liver tropism of cells. Xenografted larvae were imaged with a fluorescent stereomicroscope to quantify liver metastasis rates (Figure 3A–E). Among the tested cells, SNU‐398 and HuH‐7 cells had the lowest liver colonisation rates of 9.28% ± 2.03% and 18.25% ± 6.52%, respectively. SNU‐449 and SK‐HEP‐1 showed high liver metastasis rates of 36.03% ± 1.5% and 40.93% ± 13.54%, respectively (Figure 3F). Finally, 80.03% ± 0.57% of the xenografts generated with Mahlavu cells showed liver metastasis, showing the highest liver metastasis rate among all cells tested (Figure 3F).
*Liver metastasis rate correlates with c‐Met levels in the zLiverMet model. (A–E) Representative images of zLiverMet xenografts at 6 dpf generated with (A) SNU‐398, (B) HuH‐7, (C) SNU‐449, (D) SK‐HEP‐1 and (E) Mahlavu cells. GFP channel (top), Cherry channel (middle) and overlay (bottom) images are displayed. The liver is encircled with a white dashed line, and the heart is indicated with a white star. Scale bars: 200 μm. (F) Average liver metastasis rates ± STD obtained after at least 2 independent experiments (n > 20 per replicate, 3 replicates each experiment) were plotted. *p ≤ 0.05, **p ≤ 0.01, ***p ≤ 0.001, ***p ≤ 0.0001. (G) Western blots showing p‐c‐Met, c‐Met, p‐MAPK, MAPK and Calnexin proteins in SNU‐398 (1st lane), HuH‐7 (2nd lane), SNU‐449 (3rd lane), SK‐HEP‐1 (4th lane) and Mahlavu (5th lane) cells. (H–K) Flow cytometry analysis with c‐Met antibody showed c‐Met positive cell percentages. (H) 74.1% of HuH‐7, (I) 94.8% of SNU‐449, (J) 90.7% of SK‐HEP‐1 and (K) 98% of Mahlavu cells were c‐Met positive.
It was previously shown that c‐Met is an important regulator of metastasis, and cells with high c‐Met activity have higher hepatic metastasis capacity, in vitro [32]. When we determined expressions of c‐Met, phospho‐c‐Met Y1234/1235 (p‐c‐Met), as well as downstream MAPK, phospho‐MAPK Thr42/44 (p‐MAPK) signalling by western blotting, we observed that neither c‐Met nor p‐c‐Met was detected in SNU‐398 cells. While c‐Met was mildly expressed in HuH‐7 cells, higher c‐Met was detected in SNU‐449, SK‐HEP‐1 and Mahlavu cells (Figure 3G). Aliquots taken from cells used for zLiverMet injections were analysed with flow cytometry to quantify c‐Met‐positive cells. 74.1%, 94.8%, 90.7% and 98% of the HuH‐7, SNU‐449, SK‐HEP‐1 and Mahlavu cell lines were found to be c‐Met positive, respectively (Figure 3H–K, Figures S1 and S2 and Table S1). Furthermore, c‐Met expression was still detected in xenografted HCC cells even 4 days after injection (Figure S3).
As a result, liver metastatic capacities of well‐described liver cancer cell lines in the zLiverMet model paralleled the aggressive phenotype and increased c‐Met signalling (Table 1) [36, 37]. Next, the causality between cellular c‐Met activity and liver metastasis was tested via genetic and chemical perturbation of the system.
Modulation of c‐Met Expression in HCC Cells Altered Liver Tropism in the zLiverMet Model
2.3
SNU‐398 clones stably overexpressing c‐Met (SNU‐398_c‐Met), along with their corresponding control group (SNU‐398_Mock), were established to assess the impact of c‐Met overexpression on liver metastasis and to determine whether c‐Met signalling enhanced metastatic potential. Overexpression and phosphorylation of c‐Met in SNU‐398_c‐Met cells were confirmed with western blot (Figure 4A). To analyse the activation status of the c‐Met signalling pathway, downstream molecules, MAPK activation and expression levels were investigated. MAPK activation and expression levels were found to be parallel to c‐Met activation and expression levels (Figure 4A). Furthermore, only 0.097% c‐Met‐positive cells were detected in SNU‐398_Mock, while 43.7% c‐Met‐positive cells were detected in SNU‐398_c‐Met cells (Figure 4B, Figure S4 and Table S1). Next, zLiverMet xenografts were generated, and as expected, the liver metastasis rates increased upon c‐Met overexpression (Figure 4C,D). While SNU‐398_Mock cells had a 7.29% ± 3.72% liver metastasis rate, SNU‐398_c‐Met had colonised the liver in 47.18% ± 1.60% of the xenografted larvae (Figure 4E). As a result, the overexpression of c‐Met in SNU‐398 cells increased the potential for these cells to form liver metastases by 7‐fold.
*c‐Met overexpression increases the liver metastasis capacity. (A) The western blot image of p‐Met, p‐c‐Met, p‐MAPK, MAPK and Calnexin proteins in SNU‐398_Mock and SNU‐398_c‐Met OE cells. (B) Flow cytometry analysis of the SNU‐398_Mock group showed 0.097%, and SNU‐398_c‐Met OE showed 43.7% c‐Met positive cells. Representative images of (C) SNU‐398_Mock and (D) SNU‐398_c‐Met OE xenografts are shown. GFP channel (top) and overlay (bottom) images are displayed. The liver is encircled with a dashed line, and the heart is indicated with a star. Scale bars: 200 μm. (E) Average liver metastasis ratios were determined with 3 independent experiments, and average rates ± STD were plotted. *p ≤ 0.05, **p ≤ 0.01, ***p ≤ 0.001, ***p ≤ 0.0001.
Inhibition of c‐Met Activity by a Small Molecular Inhibitor in HCC Cells Reversed c‐Met‐Induced Liver Tropism in the zLiverMet Model
2.4
Next, the potential of the zLiverMet xenograft model in testing the effects of a small drug c‐Met inhibitor was evaluated (Figure 5). c‐Met inhibition in SNU‐449 cells was achieved with SU11274, which inhibits the Y1234/1235 phosphorylation [32]. The xenografts were generated via injection of SNU‐449 cells pretreated with either 1 μM SU11274 or an equal volume of DMSO. Since administration of small drug molecules to zebrafish larvae is possible via simple immersion in drug‐containing water, continued inhibition of c‐Met in xenografted cells was ensured via administration of SU11274 to the fish (Figure 5A). To this end, the previously determined non‐toxic dose (2.5 μM) of SU11274 or an equal volume of DMSO was added to fish water between 3 and 5 dpf, and the experiment was completed at 6 dpf (Figure 5A). Liver metastasis was imaged and quantified at 6 dpf (Figure 5B,C). DMSO treatment did not induce any change in liver metastasis rates when compared to untreated samples. Consistent with prior results, the SNU‐449 control group exhibited a 39.92% ± 4.48% liver metastasis rate, which was decreased to 22% ± 3.92% upon SU11274 treatment (Figure 5C).
*c‐Met inhibition decreases liver metastasis rates. (A) Graphical representation of the experimental set‐up (created with BioRender.com). Cells were pretreated with DMSO or 1 μM SU111274, microinjected at 2 dpf and the xenografts were treated with DMSO or 2.5 μM SU111274 between 3 and 5 dpf. (B) Representative images of SNU‐449 xenografts of DMSO‐treated (left) and SU11274‐treated (right) groups. GFP channel (top) and overlay (bottom) images are displayed. The liver is encircled with a dashed line, and the heart is indicated with a star. Scale bars: 200 μm. (D) Average liver metastasis ratios were determined with 3 independent experiments, with 3 replicates each (n > 13/replicate), and average rates ± STD were plotted. *p ≤ 0.05, **p ≤ 0.01, ***p ≤ 0.001, ***p ≤ 0.0001.
The zLiverMet Provided a Suitable In Vivo Model for Testing of Metastatic Capacity of CRC Cells to the Liver
2.5
We next examined the relevance of the zLiverMet xenograft model to examine the liver colonisation capacity of circulating CRC cells. In this regard, liver colonisation capacity of SW480 and SW620 colorectal cancer cell lines was tested using the standard yolk sac injection and IV injection as described in the materials and methods section. SW480 and SW620 CRC cell lines metastasized to the liver in 4.56% ± 1.56% and 4.83% ± 0.02% of the larvae when the cells were injected into the yolk sac. However, zLiverMet xenografts generated via IV injection of the same cells resulted in 22.23% ± 4.42% and 15.92% ± 3.75% liver colonisation of SW480 and SW620 CRC cell lines, respectively (Figure 5D). As a result, similar to HCC cell lines, the potency of the zLiverMet model for the determination of liver‐specific metastasis capacity of cancer cells was demonstrated with at least a 3‐fold increase when compared to the classical yolk‐sac injection method.
Discussion
3
Liver metastasis remains a major clinical challenge due to its frequent occurrence and associated poor prognosis in multiple cancers, particularly HCC and CRC [4, 5]. It profoundly worsens patient outcomes, and relevant test models are essential for understanding the mechanisms of metastasis and the development of inhibitory therapies [8]. While in vitro assays offer valuable mechanistic insights into metastatic processes, they are inherently limited by the absence of native tissue architecture, insufficient representation of extracellular matrix elements and the inability to reproduce in vivo shear stress conditions, which collectively represent essential regulators of metastatic progression [38, 39]. In vivo models that naturally contain complexities of tissues, tumour microenvironment components, such as homing factors, are powerful tools to tackle metastatic spread and drug resistance. However, current in vivo models to study this process are limited. This study complements existing in vitro and rodent models by presenting a novel liver‐specific metastasis model, zLiverMet, enabling real‐time imaging and quantification of liver metastatic colonisation in zebrafish larvae.
The small size of the zebrafish larvae, the short period and small number of cells needed to test metastasis, and amenability for both genetic and chemical manipulation make larval xenograft a prime model for mechanistic and drug screening studies [19, 20]. The optical transparency of larvae allows direct visualisation of host–tumour interactions, creating an ‘organism‐on‐a‐chip’ environment. Despite these advantages, organ‐specific metastasis studies have been scarce, and intrahepatic metastasis had not been demonstrated previously in the zebrafish larvae. The zebrafish larval liver is a physiologically relevant system since human and zebrafish liver functions, cell types and structures are similar [40, 41]. Expansion of the larval liver begins at 3 dpf, and the liver is fully functional at 5 dpf [40, 41]. The emerging liver and hepatocytes are detected as of 3 dpf in liver reporter transgenic lines [42]. Here, using a liver reporter transgenic zebrafish, we were able to track metastatic cell localization in real time.
While the short time of the zebrafish larval xenograft assay is an advantage, it also limits the applicability of this model for testing organ‐specific metastasis. When cells are injected into the yolk sac, they have to interact with the local microenvironment, pass the ECM barrier, intravasate, endure the shear stress before extravasating into the liver parenchyma. With the yolk sac injection method, most metastatic cells are detected in the yolk sac or in the larval vasculature at 4 days post injection [18, 20]. Given these limitations, here an intravenous (IV) injection method was developed, and our results demonstrated that it provides a superior model for organ‐specific metastasis modelling in the zebrafish larvae. In the zLiverMet assay, the cells are delivered into the larval vasculature, thereby bypassing the initial invasion steps required for intravasation. Tumour cells introduced into the circulation can interact with vascular and parenchymal compartments immediately, mimicking natural dissemination. Human CTCs were shown to transit through zebrafish larval vasculature, and the arrest of CTCs at the central vein plexus, where blood flow is lower, was demonstrated in 3 dpf larvae [43, 44]. The extravasation of CTCs via zebrafish endothelial cells was demonstrated [43]. Our findings suggest that IV injection of tumour cells to the zebrafish larvae is critical for modelling organ‐specific metastasis and that intrahepatic colonisation requires dynamic interactions within the vasculature and liver parenchyma. This is consistent with clinical observations that tumour cells metastatic to the liver must survive hemodynamic shear stress, arrest in liver sinusoids, extravasate and adapt to the unique hepatic microenvironment [5, 45]. Further studies are needed to define the hemodynamic forces in the larval sinusoids.
Homing factors also influence organ‐specific colonisation. The zLiverMet model accurately reflected the intrinsic liver tropism of HCC cell lines with different aggressiveness [36, 37]. Differential expression of the receptor tyrosine kinase c‐Met among HCC cell lines strongly correlated with liver metastatic potential in the zLiverMet model. High c‐Met–expressing lines such as Mahlavu and SK‐HEP‐1 exhibited markedly greater liver colonisation than low c‐Met–expressing lines like HuH‐7 and SNU‐398. Forced c‐Met overexpression in SNU‐398 cells significantly enhanced hepatic metastasis, confirming c‐Met as a driver of liver tropism in HCC.
Pharmacological inhibition further validated this link. Treatment with the c‐Met inhibitor SU11274 effectively reduced hepatic metastasis rates in the high c‐Met‐expressing SNU‐449 line, validating this model for preclinical evaluation of anti‐metastatic therapeutics such as small molecular inhibitors targeting c‐Met signalling. Previous studies have shown that daily intravenous administration of SU11274 in mouse models significantly reduced the invasive and metastatic characteristics of HCC cells in xenografts [46]. Results reported here show c‐Met activity in the HCC cell is a determinant of in vivo hepatic metastasis, and they align well with existing rodent models and clinical data linking c‐Met overexpression to aggressive HCC progression and metastasis, underscoring its potential as a therapeutic target [36, 37]. The zebrafish larval xenografts are amenable to screening with pharmacological agents, providing a cost‐effective solution in terms of time and amount of drug needed. The conservation of major signalling pathways may lead to toxicity in the larvae when the drug is systemically delivered. In this context, c‐Met inhibitor dose and application duration were optimised to ensure viability of the larvae.
The zLiverMet model provides a versatile assay for testing hepatic metastasis of cancer cells. CRC cell lines tested in the zLiverMet model exhibited liver colonisation, albeit at lower frequencies than the high c‐Met HCC cells. This reflects the distinct but overlapping metastatic mechanisms where CRC cells exploit venous drainage patterns and tumour‐host interactions for hepatic metastasis. The successful modelling of CRC liver metastasis demonstrates the broader applicability of the zebrafish system for dissecting organ‐specific metastasis mechanisms across cancer types.
Despite some differences in zebrafish liver architecture and immune system maturity relative to humans, the zLiverMet model effectively recapitulates key steps of intrahepatic metastasis and tumour‐microenvironment interactions. Unique advantages of the larval stage include optical transparency, rapid metastasis onset, amenability to small molecule treatment as well as host gene expression modifications. These features make zLiverMet a powerful complementary tool to traditional models that may not capture early metastatic events and microenvironmental cues as dynamically. The model could be further developed to investigate interactions of metastatic cells with liver sinusoidal endothelium and determine networks that regulate arrest and extravasation of metastatic cells. Future applications will benefit from integrating this model with single‐cell and molecular profiling to unravel the complex interplay of metastatic tumour cells, hepatic stromal components and immune elements. Moreover, expanding its use for testing additional molecular drivers and therapeutic interventions could accelerate the development of effective anti‐metastatic strategies specifically targeting intrahepatic colonisation.
In conclusion, this study presents a robust zebrafish liver‐specific metastasis model that confirms the critical role of c‐Met in driving HCC liver tropism and demonstrates the liver colonisation capacity of CRC cells. As an innovative ‘organism‐on‐a‐chip’ model, zLiverMet bridges the gap between in vitro assays and mammalian models, facilitating both mechanistic and preclinical studies aimed at combating liver metastasis.
Materials and Methods
4
Zebrafish Handling
4.1
Zebrafish were maintained under standard conditions at Izmir Biomedicine and Genome Center (IBG) Zebrafish Facility, with a 14/10 h dark/light photoperiod. Embryos were obtained from wild‐type AB (+/+) or TL (+/+), Tg(fabp10a:mCherry, cmlc2:GFP) and Tg(fli1:EGFP) strains [33, 47]. Embryos were reared in E3 medium (5 mm NaCl, 0.17 mm KCl, 0.33 mm CaCl2, 0.33 mm MgSO_4_, pH 7.2) at 28°C, until 2 dpf at a density of 50 embryos/30 mL. Xenografted larvae were incubated at 34°C with a 14/10 h light/dark photoperiod; 20–25 larvae were kept in each well of six well plates with 10 mL E3. Zebrafish experiments were performed according to EU Directive 2010/63 and national regulations, with permission from IBG Local Animal Experiments Ethics Committee with the approval number 2024–025.
Transgenic Line Generation
4.2
fabp10a:mCherry, cmlc2:GFP tol2 plasmid was cloned with Gateway LR Clonase II Enzyme mix (#11791020, Invitrogen, Carlsbad, CA, USA), using the Tol2Kit plasmids pME‐mCherry, p3E‐polyA, pDest‐Tol2CG2 and p5E‐fabp10a plasmid kindly gifted by Donghun Shin [48, 49]. pCS2FA‐transposase plasmid was linearized with NotI, and mRNA was produced with mMESSAGE mMACHINE SP6 Transcription Kit (#AM1340; Invitrogen, Carlsbad, CA, USA). An injection mix containing 250 ng/μl plasmid and 250 ng/μl mRNA was injected into single‐cell embryos. Mosaic fish (F_0_) were screened for bleeding heart (GFP) at 30 h post‐injection, and mCherry signals were detected in their livers at 72 h post‐injection. A stable transgenic line was established.
Cell Culture
4.3
SNU‐398_Mock and SNU‐398‐cMet cells were generated as described previously [32]. Hep3B, SNU‐398, SNU‐449, HuH‐7, SK‐Hep‐1, Mahlavu, SW480 and SW620 cell lines were routinely cultured at 37°C under 5% CO_2_ in Dulbecco's modified Eagle's medium (DMEM) (#31885023, Thermo Fisher Scientific, Waltham, MA, USA) and Roswell Park Memorial Institute (RPMI) 1640 medium (#21875034, Thermo Fisher Scientific, Waltham, MA, USA). Media were supplemented with 10% (v/v) foetal bovine serum (FBS) (#A5256801, Thermo Fisher Scientific, Waltham, MA, USA), 1% (10 U/mL) penicillin/(10 μg/mL) streptomycin (#SLP‐508‐100, Serox GmbH, Mannheim, Germany). 1% MEM non‐essential amino acids (#SRM‐813‐500, Serox GmbH, Mannheim, Germany).
Flow Cytometry
4.4
Previously published protocol was slightly modified [32]. SNU‐398_Mock, SNU‐398_c‐Met OE, SNU‐449, SK‐HEP‐1 and Mahlavu cells were trypsinized and washed with PBS. Unfixed cells were resuspended in a staining buffer containing 5% BSA, 1% sodium azide and PBS at 1 × 10^6^ cells/mL density. Cells were stained with mouse c‐Met antibody (dilution factor 1:50) (#566014, Clone 3D6, BD Pharmingen Alexa Fluor 647) for 60 min at RT. Unstained cells were used for compensation. After incubation, cells were washed twice with PBS and then resuspended in 100 μL PBS for acquisition. Data was analysed using FlowJo VX (BD Bioscience). Compensation matrices were manually checked and adjusted as needed.
Western Blot
4.5
Total proteins were isolated and prepared by using a modified RIPA buffer as described previously [29]. Antibodies against p‐c‐Met (y1234/1235) (#799139, Invitrogen, Waltham, MA, USA), c‐Met (#370100, Invitrogen, Waltham, MA, USA), p‐MAPK (Thr42/44) (#9186, Cell Signalling Technology, Danvers, MA, USA), MAPK (#sc‐514 302; Santa Cruz Biotechnology, Dallas, TX, USA) and Calnexin (#PA5‐34754; Invitrogen, Waltham, MA, USA) were used as described before [29]. Equal loading and transfer were confirmed by repeat probing for calnexin (housekeeping gene). Alexa Fluor 680 Goat (#A21109, Thermo Fisher Scientific, Waltham, MA, USA) and Alexa Fluor Goat 488 (#A‐11017, Thermo Fisher Scientific, Waltham, MA, USA) were used for detection.
Xenograft Assays
4.6
1 × 10^6^ cells in a culture plate were stained with DiO (#V2885, Invitrogen, Carlsbad, CA, USA), added to the culture medium at 1: 1000 dilution, O/N. Before injection, the cells were trypsinized and washed 2 times with FBS and resuspended at 30.000–50.000 cells/μl in complete medium. For inhibitor treatment, SNU‐449 cells were incubated with 1 μM SU11274 (#448101, Calbiochem, San Diego, USA) or 0.05% DMSO (#276855, Sigma Aldrich, Burlington, MA, USA) for 18–20 h in complete medium containing 5% FBS and DiO. After trypsinization, the treated cells were washed twice with complete medium containing 5% FBS and resuspended in the same medium. 2 dpf embryos were anaesthetised with 0.02% tricaine (#A5040‐100G, Merck, Darmstadt, Germany) and aligned in a microinjection mould. 100–150 cells per embryo were injected into the anterior cardinal vein of the 2 dpf embryos. Each injected zebrafish was controlled under the fluorescence stereo microscope to ensure comparable cell transplantation and randomly grouped in 3 biological replicates. Inhibitor treatment of the xenografts was conducted with 2.5 μM SU11274; control groups were treated with an equal volume of DMSO (0.14%), the embryo water was refreshed every 24 h.
Liver Metastasis Quantification and Statistical Analyses
4.7
Larvae that had at least five cells at the liver, when observed under the Olympus SZX16 fluorescence stereomicroscope, were considered to have liver metastasis. The liver metastasis ratio was quantified with the following formula:
Each independent experiment had 3 replicates of blindly randomised xenografted larvae. Liver metastasis potential was calculated as the average of the 3 replicates. In graphs, the average of independent experiments was plotted, and the standard deviation was used for error bars. For statistical analyses, independent experiments with their replicates were tested for normality. Normality distributions of the groups were assessed using the Shapiro–Wilk and Kolmogorov–Smirnov tests. The unpaired parametric t‐test (Welch correlation) or one‐way ANOVA was used for groups with normal distribution. After the statistical analysis, graphs were represented with the technical mean of the experiments and standard deviation. Statistical analyses were performed using the GraphPad PRISM 10.0 (GraphPad Software, Boston, MA, USA). Normality test was analysed with the mean of the replicates.
Zebrafish Histopathology
4.8
6 dpf larvae were euthanized with tricaine and subsequently fixed with 10% neutral buffered formalin (NBF) (#HT501128; Sigma Aldrich, Burlington, MA, USA). The samples were transferred to a series of ethanol (#1.00983.2511, Merck, Darmstadt, Germany) and xylene (#1082972500; Merck, Darmstadt, Germany) and embedded in paraffin (#TK.200661.02502, Tekkim, Istanbul, Türkiye) at 65°C based on previously published protocols [50]. 1.5 μM slices were obtained with a microtome (#RM2245, Leica Microsystems GmbH, Wetzlar Germany). Haematoxylin and eosin (H&E)‐stained slides were imaged Olympus BX61 microscope.
Author Contributions
M.B. and H.O. performed zebrafish experiments. P.K. performed cell culture, flow cytometry and western blot experiments. M.B. and P.K. analysed data, wrote results and prepared figures. G.C.‐A. and N.A. conceived the project, acquired funding and wrote the manuscript. All authors reviewed and approved the submitted version of the manuscript.
Funding
This study was funded by the European Commission HORIZON‐HLTH‐2022‐STAYHLTH‐02 grant #101095679 (Halt‐RONIN) and TUBITAK 1001 Grant #119S698.
Ethics Statement
This research involved animal subjects and it complies with ARRIVE guidelines. Animal experiments were performed with permission from IBG Local Animal Experiments Ethics Committee with the approval number 2024–025.
Conflicts of Interest
The authors declare no conflicts of interest.
Supporting information
Figure S1: Cell flow cytometry negative control antibody graphs are shown in (A) HuH‐7 0.099%, (B) SNU449 0.085% (C) SK‐HEP‐1 0.026% and (D) Mahlavu 0.038% positive cells were detected. Figure S2: The gating strategies of the flow cytometry analysis for c‐Met positive cell percentage calculation. Figure S3: Injected cells have c‐Met expression at 6 dpf. Figure S4: SNU‐398_Mock and SNU‐398_cMet flow cytometry graphs with negative control antibody, and gating strategies. Table S1: Flow cytometry all event counts.
The reference list from the paper itself. Each links out to its DOI / PubMed record.
- 1A. W. Lambert , D. R. Pattabiraman , and R. A. Weinberg , “Emerging Biological Principles of Metastasis,” Cell 168 (2017): 670–691, 10.1016/j.cell.2016.11.037.28187288 PMC 5308465 · doi ↗ · pubmed ↗
- 2Y. Li , F. Liu , Q. Cai , et al., “Invasion and Metastasis in Cancer: Molecular Insights and Therapeutic Targets,” Signal Transduction and Targeted Therapy 10 (2025): 57, 10.1038/s 41392-025-02148-4.39979279 PMC 11842613 · doi ↗ · pubmed ↗
- 3F. Bray , M. Laversanne , H. Sung , et al., “Global Cancer Statistics 2022: GLOBOCAN Estimates of Incidence and Mortality Worldwide for 36 Cancers in 185 Countries,” CA: A Cancer Journal for Clinicians 74 (2024): 229–263, 10.3322/caac.21834.38572751 · doi ↗ · pubmed ↗
- 4H. Zhou , Z. Liu , Y. Wang , et al., “Colorectal Liver Metastasis: Molecular Mechanism and Interventional Therapy,” Signal Transduction and Targeted Therapy 7 (2022): 70, 10.1038/s 41392-022-00922-2.35246503 PMC 8897452 · doi ↗ · pubmed ↗
- 5M. Mitsunobu , A. Toyosaka , T. Oriyama , E. Okamoto , and N. Nakao , “Intrahepatic Metastases in Hepatocellular Carcinoma: The Role of the Portal Vein as an Efferent Vessel,” Clinical & Experimental Metastasis 14 (1996): 520–529, 10.1007/BF 00115112.8970582 · doi ↗ · pubmed ↗
- 6T. Nakashima and M. Kojiro , “Pathologic Characteristics of Hepatocellular Carcinoma,” Seminars in Liver Disease 6 (1986): 259–266, 10.1055/s-2008-1040608.3022387 · doi ↗ · pubmed ↗
- 7S. Yamamoto , Y. Midorikawa , G. Nagae , et al., “Spatial and Temporal Expansion of Intrahepatic Metastasis by Molecularly‐Defined Clonality in Multiple Liver Cancers,” Cancer Science 111 (2020): 601–609, 10.1111/cas.14282.31845427 PMC 7004543 · doi ↗ · pubmed ↗
- 8S. L. Yang , Y. Y. Luo , M. Chen , et al., “A Systematic Review and Meta‐Analysis Comparing the Prognosis of Multicentric Occurrence and vs. Intrahepatic Metastasis in Patients With Recurrent Hepatocellular Carcinoma After Hepatectomy,” HPB: The Official Journal of the International Hepato Pancreato Biliary Association 19 (2017): 835–842, 10.1016/j.hpb.2017.06.002.28734693 · doi ↗ · pubmed ↗