Neuropilin-2 Is Associated With Increased Hepatoblastoma Cell Viability and Motility

The neuropilins NRP1 and NRP2 are multifunctional glycoproteins that have been implicated in several cancer-related processes including cell survival, migration, and invasion in various tumor types. Here, we examine the role of neuropilins in hepatoblastoma (HB), the most common pediatric liver malignancy. Using a combination of immunohistochemistry, RNA analysis and western blotting, we observed high level expression of NRP1 and NRP2 in 19 of 20 HB specimens and in a majority of human HB cell lines (HUH6 and five cell lines established from patient-derived xenografts) studied but not in normal hepatocytes. Silencing of NRP2 expression in HUH6 and HB-282 HB cells resulted in decreased cell viability, impaired cytoskeleton remodeling, and reduced cell motility, suggesting that NRP2 contributes to the malignant phenotype. We propose that neuropilins warrant further investigation as biomarkers of HB and potential therapeutic targets.


INTRODUCTION
Hepatoblastoma (HB) is the most common primary liver malignancy in the pediatric population with an incidence of 1.9 cases per million (1,2). Preterm birth, low birthweight, and certain genetic conditions such as Beckwith-Wiedemann syndrome and Familial Adenomatous Polyposis are associated with increased risk of HB. In most HB cases, however, the etiology of the disease remains unidentified (3,4). HB histology resembles embryonal or fetal liver, and low differentiation stage associates with poor prognosis (5). The molecular pathways involved in the pathogenesis of HB are not fully understood, although aberrant activation of WNT/β-catenin signaling is present in the majority of these tumors (6,7).
Here, we characterize the expression patterns of NRP1 and NRP2 in human HB specimens and cell lines. Additionally, we examine the functional consequences of NRP2 gene silencing in HB cells.

Patient Samples
Archival formalin-fixed paraffin-embedded (FFPE) HB patient samples (n = 20) and normal liver control (NL, n = 4) samples were obtained from the Helsinki Biobank at Helsinki University Hospital. The HB samples were originally collected at the time of surgical treatment, whereas the NL samples were from liver transplantation donors. This study was approved by Helsinki University Hospital institutional ethical committee (HUS/3319/2018) and conducted in accordance with Finnish bylaws. Informed consent was obtained when samples were deposited to the Helsinki Biobank.

NRP2 Silencing
NRP2 expression was inhibited in HUH6 and HB-282 cells via small interfering RNA (siRNA) transfection. Briefly, adherent HUH6 cells were exposed to 100 nM and HB-282 cells to 25 nM of NRP2 ON-TARGETplus SMARTpool siRNA or ON-TARGETplus non-targeting (NT) control siRNA (both from Horizon Discovery, Cambridge, UK). Lipofectamine RNAiMAX reagent (Invitrogen, Carlsbad, CA, USA) was used to deliver siRNAs into the HUH6 cells and Dharmafect 4 (Horizon Discovery) was utilized for HB-282 cells. Knockdown efficacy was evaluated at mRNA and protein level 72 h after initiation of transfection. Transfection efficacy was assessed with siGLO Green transfection indicator (Horizon Discovery). A detailed transfection protocol for HUH6 cells is described elsewhere (36), and HB-282 cells were transfected following manufacturer's instructions for Dharmafect.

RNA and Protein Extraction
A NucleoSpin RNA/Protein extraction kit was utilized for total RNA and protein extractions (Macherey-Nagel, Düren, Germany) following the manufacturer's instructions.

Statistical Analysis
For qPCR, protein analysis, viability assays, and migration studies, three independent experiments were conducted. Statistical significance was assessed with Student's t-test utilizing JMP Software (JMP Pro; version 15.1.0, SAS Institute Inc.). A p-value < 0.05 was considered as statistically significant. RNA-sequencing data was analyzed with edgeR-package and significance level was set at adjusted p-value < 0.05.

NRPs Are Highly Expressed in HB Tissue and Cell Lines
We assessed NRP1 and NRP2 expression in FFPE samples collected from 20 HB patients treated at Helsinki University Hospital between January 1, 1990 and December 31, 2016. Demographic information of the patients is shown in Table 1.
Healthy liver samples from organ donors were used as controls. Consistent with prior reports (28,38), NRP1 and NRP2 expression was limited to LSECs in healthy liver; hepatocytes did not display specific immunoreactivity (Figures 1A-D) (Figures 1G-L). Next, we evaluated NRP1 and NRP2 expression in six human HB cell lines and primary hepatocyte cultures. NRP1 mRNA was upregulated in four HB cell lines (HB-282, HB-295, HB-279, HB-243) compared to primary hepatocytes (Figure 2A). Similarly, these four cell lines demonstrated a 70-to 700-fold increase in NRP1 protein levels compared to primary hepatocytes ( Figure 2B). Upregulation of NRP2 mRNA was noted in all HB cell lines while expression in primary hepatocytes was barely detectable (Figure 2C). At the protein level, HUH6, HB-282, HB-295, HB-284, and HB-243 demonstrated striking amounts of NRP2 (1,400-to 4,300-fold increase compared to primary hepatocytes; Figure 2D). NRP2 protein expression was undetectable in primary hepatocytes and HB-279 cells ( Figure 2D).

NRP2 Silencing in HUH6 and HB-282 Cell Lines
Since NRP2 expression was prominent in five out of six HB cell models, it was selected as the target for functional studies. We performed NRP2 knockdown in HUH6 and HB-282  Supplementary Figures 2A-F). Following transient siRNA transfections, NRP2 expression was reduced 70-75% at the mRNA and protein level in HUH6 cells (Figures 3A,B) and 60-70% at mRNA and protein level in HB-282 cells (Figures 3C,D).

Knockdown of NRP2 Attenuates HB Cell Viability
To evaluate the impact of NRP2 silencing on cell growth and survival, we performed a clonogenic assay. A statistically significant reduction in the number of colonies was observed in both HUH6 (Figures 4A-C) and HB-282 (Figures 4E-G) cell lines, the decrease in colony numbers being 40 and 60%, respectively. The effect of NRP2 knockdown on ATP availability was assessed as a secondary measure for cell viability. Approximately a 30% decrease in ATP concentration was noted in HUH6 ( Figure 4D) and HB-282 (Figure 4H) cells.

NRP2 Silencing Decreases Stress Fiber Formation and Attenuates HB Cell Motility
A previous study by Wittmann and colleagues demonstrated that NRP2 expression is associated with increased motility of liver carcinoma cells (28). We observed a reduced amount of actin protrusions and depolymerization of stress fibers in NRP2 siRNA treated HUH6 (Figures 5E-H) and HB-282 (Figures 5M-P) cells compared to NT siRNA treated control cells (Figures 5A-D,I-L). Next, we investigated the motility of HUH6 and HB-282 cells exploiting the transwell migration assay. A statistically significant 55% (HUH6) and 85% (HB-282) decrease in the number of migrated cells was noted in HB cells with downregulated NRP2 expression cells (Figures 6B,C,E,F) compared to control cells (Figures 6A,C,D,F).

DISCUSSION
NRPs have been associated with increased malignant potential and poor prognosis in various human cancers (19,22,(39)(40)(41). In addition to their potential as biomarkers, efforts are underway to develop NRP1 or NRP2 as therapeutic targets (42)(43)(44)(45). To the best of our knowledge, the present study is the first to demonstrate that NRP1 and NRP2 are highly expressed in HB. Furthermore, our observations link NRP2 expression with increased cell survival, actin stress fiber polymerization, and migration of HB cells (Figure 7). Currently, chemotherapy with cisplatin and doxorubicin is one of the mainstays of HB therapy, but multiorgan toxicity and chemoresistance limit the usability of these agents (46)(47)(48)(49). Therefore, there is a need to identify tumor specific proteins to enable the development of targeted treatments for HB. In adults with hepatocellular carcinoma, NRP2 expression correlates with shorter disease-free survival and overall survival (19). Moreover, high NRP2 expression was noted in de-differentiated tumors and mesenchymal hepatocellular carcinoma cell lines (19,28). We observed high NRP2 expression in majority of HB tumors and in vitro models. Taken together, it appears that NRP2 expression is a common feature in poorly differentiated hepatic malignancies. Due to the limited number of patient samples, NRP2 expression was not correlated with clinical variables in this study. Therefore, in future investigations with larger sample sets, the association of NRP2 with HB patient survival should be explored. NRP2 expression has been associated both with increased proliferation rate as well as with the capability of tumor cells to escape from programmed cell death in cancer cells (13,50,51). We observed decreased cell viability in NRP2 knockdown cells, but further studies are needed to clarify whether this is a consequence of lower proliferation rate or increased cell death. Interestingly, a recent study linked upregulation of NRP2 with cisplatin-resistance in neuroblastoma cells (52). Poor response to current chemotherapeutical treatments, including cisplatin, have been observed in 20-25% of HB tumors pointing out the need to find approaches to overcome the chemoresistance (49). It would be of interest to study whether cisplatin treatment combined with NRP2 targeting could cause an even more drastic decrease in cell viability in HB cells than cisplatin administration or NRP2 knockdown alone.
Actin stress fibers are a hallmark of mesenchymal phenotype, and remodeling of actin cytoskeleton is a prerequisite for cancer cell motility (53). Polymerized stress fibers form protrusive structures named filipodia, lamellipodia, and invadopodia directing the migration and invasion of cancer cells to new locations (54). This process is a preliminary step in metastasis formation (55). High NRP2 expression has been associated with increased motility and invasiveness in cancer cell models (28,56,57). In line with these findings, NRP2 silencing led to depolymerization of actin stress fibers, and consequently decreased the migration of HB cells. The importance of NRP2 in pre-metastatic processes suggests that NRP2 targeted interventions hold potential in the management of aggressive HB.
A multitude of signaling mediators are known to interact with NRPs. In breast cancer cells, neutralizing NRP2 antibody blocked cytoplasmic C-X-C Motif Chemokine Receptor 4 (CXCR4) expression which was followed by decreased tumor cell migration (22). Interestingly, upregulation of CXCR4 was recently associated with the infiltration of protumorigenic macrophages, neutrophils, and dendritic cells in HB tissue samples (58). Considering the NRP2-CXCR4 axis, NRP2 targeting might attenuate the growth of HB at multiple levels. In addition to the direct effect on HB cells observed in our study, anti-NRP2 approaches might disrupt the tumor microenvironment.
Abnormal WNT/β-catenin signaling is observed in the majority of HBs, and several studies have shown that inhibition of WNT/β-catenin suppresses HB cell growth in vitro (7,(59)(60)(61). Captivatingly, upregulation of secreted WNT antagonists decreased NRP2 expression in osteosarcoma cell models suggesting that NRP2 transcription is regulated by WNT pathway (44). NRP2 expression has also been directly connected to improved β-catenin stability with consequently increased motility and invasiveness of gastrointestinal cancer cells (57). Future investigations should explore whether upregulation of NRP2 in HBs is linked to aberrant WNT/β-catenin signaling activity.
Major limitations of this study were the rather low number of patient samples and challenges linked to the siRNA mediated silencing method. It is known that siRNAs may have off-target effects (62). Therefore, it should be noted that the in vitro results presented in this study may not be explicitly due to the NRP2 silencing. In the future studies, another silencing methods, such as CRISPR-Cas9 mediated gene editing, should be considered to overcome this issue.  In conclusion, NRPs are expressed in the majority of HBs and further studies are warranted to evaluate their potential as prognostic biomarkers. Moreover, RNA interference mediated inhibition of NRP2 suppresses HB cell viability and motility suggesting that NRP2 targeted interventions have potential in the management of aggressive HB.

ETHICS STATEMENT
The studies involving human participants were reviewed and approved by Helsinki University Hospital institutional ethics committee. Written informed consent to participate in this study was provided by the participants' legal guardian/next of kin.