Oncolytic H-1 Parvovirus Hijacks Galectin-1 to Enter Cancer Cells

Clinical studies in glioblastoma and pancreatic carcinoma patients strongly support the further development of H-1 protoparvovirus (H-1PV)-based anticancer therapies. The identification of cellular factors involved in the H-1PV life cycle may provide the knowledge to improve H-1PV anticancer potential. Recently, we showed that sialylated laminins mediate H-1PV attachment at the cell membrane. In this study, we revealed that H-1PV also interacts at the cell surface with galectin-1 and uses this glycoprotein to enter cancer cells. Indeed, knockdown/out of LGALS1, the gene encoding galectin-1, strongly decreases the ability of H-1PV to infect and kill cancer cells. This ability is rescued by the re-introduction of LGALS1 into cancer cells. Pre-treatment with lactose, which is able to bind to galectins and modulate their cellular functions, decreased H-1PV infectivity in a dose dependent manner. In silico analysis reveals that LGALS1 is overexpressed in various tumours including glioblastoma and pancreatic carcinoma. We show by immunohistochemistry analysis of 122 glioblastoma biopsies that galectin-1 protein levels vary between tumours, with levels in recurrent glioblastoma higher than those in primary tumours or normal tissues. We also find a direct correlation between LGALS1 transcript levels and H-1PV oncolytic activity in 53 cancer cell lines from different tumour origins. Strikingly, the addition of purified galectin-1 sensitises poorly susceptible GBM cell lines to H-1PV killing activity by rescuing cell entry. Together, these findings demonstrate that galectin-1 is a crucial determinant of the H-1PV life cycle.


Introduction
Oncolytic viruses selectively infect and destroy cancer cells while sparing normal tissues [1]. They can also stimulate strong anti-tumour immune responses and destroy tumour vasculature [2]. No fewer than 40 oncolytic viruses are currently under evaluation in clinical trials as treatments against a variety of cancers. Among them is H-1 rat protoparvovirus (H-1PV), a member of the Parvoviridae family in the genus Protoparvovirus [3,4]. This genus in addition to H-1PV includes Kilham rat virus, LuIII virus, minute virus of cells. HeLa cells were seeded at a density of 4 × 10 4 cells/well in 24-well plates and then pre-treated with increasing amounts (50,100,150, and 200 mM) of lactose for 30 min and then cells were infected with recH-1PV-EGFP for 4 h and grown for an additional 20 h. Cells were then processed as described in viral transduction assay and fluorescent microscopy sections. Numbers represent the arithmetic mean percentage of EGFP-positive cells relative to the number of EGFP-positive cells observed in untreated cells, which was arbitrarily set as 100%.

Generation of the LGALS1 Knockout Cell Line
CRISPR/Cas9-mediated knockout of LGALS1 in NCH125 was accomplished using galectin-1 Double Nickase Plasmid ([h]sc-400941-NIC), whereas the CRISPR/Cas9 negative control was obtained using the Control CRISPR/Cas9 Plasmid (sc-418922; both from Santa Cruz). NCH125 cells were seeded in a 6-well plate at about 70% confluency. After 24 h, 2 µg of DNA were transfected using Lipofectamine LTX (Thermo Fisher Scientific, Carlsbad, CA, USA) according to the vendor's protocol. Transfected cells were selected in normal growth medium containing 1 µg/mL puromycin (Thermo Fisher Scientific, Shanghai, China) for 72 h. Individual clones were obtained by limiting dilution. Knockout was confirmed by Western blotting.

Plasmid Transfection
To rescue LGALS1 expression, the plasmid encoding LGALS1 gene was used (SC118705; OriGene Technologies, Inc. Rockville, MD, USA). NCH125 Control and LGALS1 KO cells were seeded at a density of 3x10 5 cells/well in a 6-well plate. The next day, cells were transfected with 2.5 µg of DNA using Lipofectamine LTX or mock-transfected for 48 h.

Confocal Microscopy
Cells were seeded at a density of 3.5 × 10 3 cells/spot on spot slides and grown in 50 µL of complete cellular medium. The next day, cells were infected with wild-type H-1PV at an MOI of 500 pfu/cell in a total of 70 µL of 5% fetal bovine serum (FCS)-containing medium. At 2 h post-infection, cells were fixed with 3.7% paraformaldehyde on ice for 15 min and permeabilised with 1% Triton X-100 for 10 min. Immunostaining was carried out with the following antibodies, all used at 1:500 dilution for 1 h: mouse monoclonal anti-H-1PV capsid [29] and rabbit polyclonal anti-galectin-1 (HPA000646; Sigma-Aldrich, Darmstadt, Germany). Anti-mouse Alexa Fluor 594 IgG (A11005; Thermo Fisher Scientific, Carlsbad, CA, USA) or anti-rabbit Alexa Fluor 488 IgG (A11008; Thermo Fisher Scientific, Carlsbad, CA, USA) were used as secondary antibodies. Nuclei were stained with DAPI. Images of randomly assigned cells in the green channel (galectin-1), red channel (H-1PV), or blue channel (DAPI) were acquired with a confocal microscope (Leica TCS SP5 II, Wetzlar, Germany). Picture analysis was carried out using the LAS X Software (Leica, Wetzlar, Germany).

MTT Viability Assay
To determine cell viability after virus infection, the conversion of 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyl-2H-tetrazolium bromide (MTT) (Sigma-Aldrich Chemie GmbH, Stein-Viruses 2022, 14, 1018 5 of 25 heim, Germany) was measured. For this purpose, cells were seeded on a 96-well plate at a density of 2000 cells/well in 50 µL of culture medium supplemented with 10% FCS. The next day, 50 µL of serum-free medium containing wild-type H-1PV were added on top of the cells. In rescue experiments, cells were treated with H-1PV at an MOI of 5 pfu/cell, or 5 µg/mL of recombinant galectin-1 (ab50237; Abcam, Cambridge, UK), or both simultaneously. Every 24 h post-treatment, for a total of 4 time points, 10 µL of 5 mg/mL MTT were added and subsequently incubated for 2 h at 37 • C. Thereafter, the supernatant was aspirated, and the plates were air-dried at 37 • C overnight. To solubilise the formazan product, cells were then incubated with 100 µL of isopropanol for 20 min with moderate shaking and the absorbance was read with an ELISA reader at 570 nm. Viability of treated cells was expressed as a ratio of the measured absorbance (arithmetic mean of three replicates per condition) to the corresponding absorbance of untreated cells (arbitrarily defined as 100%).

Binding-Only and Binding and Entry Assays
First, the culture medium was removed and replaced with 200 µL serum-free medium containing H-1PV at an MOI of 5 pfu/cell (or H-1PV at an MOI of 5 pfu/cell and 5 µg/mL of recombinant galectin-1 simultaneously in rescue experiments). Infection was performed for 1 h at 4 • C to allow only cell surface virus binding or for 4 h at 37 • C to also allow virus cell internalisation. Thereafter, cells were extensively washed with PBS, trypsinised for 5 min, quenched with serum-containing medium and subjected to three snap freeze-thaw cycles to release cell-associated viral particles. Viral DNA was purified from cell lysates using the QiAamp MinElute Virus Spin kit (Qiagen, Hilden, Germany) according to the manufacturer's instructions. Cell-associated H-1PV genomes were quantified by following a parvovirus-specific quantitative PCR (qPCR) protocol, as previously described [12]. A minimum of three independent experiments were performed in triplicate for every condition tested.

Flow Cytometry
Cells were seeded at a density of 5 × 10 5 cells/well in a 6-well plate. The next day, cells were infected with H-1PV at an MOI of 25 pfu/cell for 1 h at 4 • C. Cells were washed with ice-cold PBS and then gently scraped off with a cell lifter on ice. Cells were then fixed with 2% formaldehyde for 15 min at 4 • C and blocked with 2.5% albumin bovine fraction V (BSA; SERVA Electrophoresis, Heidelberg, Germany)/PBS for 20 min at 4 • C. Cells were then incubated with H-1PV anti-capsid antibody (dilution 1:500) for 30 min at 4 • C, and subsequently with Alexa Fluor 488 goat α-mouse (1:500) for 30 min at 4 • C. Three washes with 2.5% BSA/PBS were performed between each staining step. Analysis was carried out using a FACS Calibur (BD Biosciences, San Jose, CA, USA).

Tissue Microarray
Patient material for the tissue microarray was derived from paraffin embedded GBM biopsies obtained from the Department of Pathology, Haukeland University Hospital, Bergen, Norway. The project (number 2017/2505) was approved by the Regional Ethics Committee (Bergen, Norway). Control tissues (brain, liver and tonsil) were derived from autopsy material. The microarray included 61 primary GBM and 49 recurrent GBM biopsies, plus 12 biopsies from normal tissues (four each from brain, tonsil and liver). Immunohistochemical staining was carried out as described previously [30] using galectin-1 antibody (sc-166618; Santa Cruz, Santa Cruz, CA, USA) at a dilution of 1:200 followed by a biotinylated anti-mouse antibody (Vector Laboratories, Burlingame, CA, USA) at a dilution of 1:1100. Galectin-1-positive cells were counted via automated counting, as described previously [31].

Correlation Analysis between Gene Expression of Cancer Cell Lines and H-1PV-Induced Oncolysis
Gene expression data were taken from two databases: 53 cancer cell lines of NCI-60 dataset (https://discover.nci.nih.gov/cellminer, accessed on 23 January 2021, RNAseq data) and the Cancer Cell Line Encyclopedia (CCLE, 52 cancer cell lines) [32]. Simple linear regression models were built for 2 considered genes of interest (LGALS1 and LAMC1) and 2 control genes (LGALS3 and GAPDH), predicting experimentally observed EC50 values. Additionally, a two-variable regression was built predicting EC50 with both genes LGALS3 and GAPDH. Significance of the models was characterised by the p-values (p), coefficients of determination (R 2 ) and Pearson correlations (R).

Measurement of Transcript Levels
The mRNA expression levels of the target of interest LGALS1, as well as of the reference genes ACTB, GAPDH and PGK1, were quantified at the nCounter Core Facility on a SPRINT Profiler system by nCounter technology (NanoString Technologies (Seattle, WA, USA), as described previously [14]. Accession numbers and target sequences of analysed genes are: LGALS1 gene (accession number NM_002305.

xCELLigence
Cell proliferation was monitored in real time through the xCelligence system (ACEA Biosciences Inc. San Diego, CA, USA) according to the manufacturer's instructions. Briefly, 8 × 10 4 cells per well were seeded in a 96-well E-plate (Roche, Mannheim, Germany) in a total volume of 100 µL of complete DMEM medium. Cells were treated during the cellular growth phase with H-1PV at an MOI of 5 pfu/cell or 5 µg/mL of recombinant galectin-1, or both simultaneously. Cell proliferation was monitored every 30 min in real time. Data are expressed as "Cell index" (n = 3) calculated by the RTCA software 1.2.1 (Agilent) as a measure of cell adhesion and, therefore, cell viability.

Statistical Analysis
Results are shown as the arithmetic mean of biological replicates ± standard deviation (SD) from a representative experiment. Statistical significance was determined by a paired two-tailed Student's t-test, unless stated otherwise, using Microsoft Excel 365 and/or GraphPad Prism version 8. Only values below 0.05 were considered significant: p ≤ 0.05 (*), p ≤ 0.01 (**) and p ≤ 0.001 (***).

Knockdown of LGALS1, but Not LGALS3, Hampers H-1PV Infection
To identify cellular modulators of the H-1PV life cycle, we previously carried out a high-throughput siRNA library screening in cervical carcinoma-derived HeLa cells using a siRNA library targeting the human druggable genome (6961 genes, each targeted by a pool of four siRNAs/gene) [14]. This led to the identification of laminins, in particular those containing the laminin γ1 chain, as factors used by H-1PV to attach at the cell surface and to enter cancer cells [14]. In the same screening, LGALS1-the gene encoding Gal-1-  [14]. These findings, along with the fact that galectins interact with laminins [17], prompted us to hypothesise that Gal-1 is involved in H-1PV infection at the level of cell entry. This hypothesis was also supported by the discovery that MVM (a closely related protoparvovirus) requires Gal-3 to efficiently infect mouse cells [23,24].
We confirmed the results of the siRNA library screening by performing a knockdown of LGALS1 using an independent siRNA. Given the role of Gal-3 in host cell entry by MVM, we also investigated the effect of siRNA-mediated silencing of LGALS3. A recombinant H-1PV expressing the EGFP reporter gene (recH-1PV-EGFP) was used for these experiments [25]. This non-replicative parvovirus shares the same capsid of the wild type but harbours the EGFP gene under the control of the natural P38 late promoter, whose activity is regulated by the NS1 viral protein. Therefore, the EGFP signal directly correlates with the ability of the virus to reach the nucleus and initiate its own gene transcription.
Cervical cancer-derived HeLa, glioma-derived NCH125 and pancreatic carcinomaderived BxPC3 cell lines were transfected with siRNAs targeting LGALS1 or LGALS3, or a scrambled siRNA. After 48 h, cells were infected with recH-1PV-EGFP and grown for a further 24 h. Efficient gene silencing was achieved for both genes in cells transfected with their respective siRNAs ( Figure S1). However, only the siRNA targeting LGALS1 significantly decreased H-1PV transduction (by more than 55%). As opposed to the role of Gal-3 in MVM infection, silencing of LGALS3 did not significantly alter H-1PV transduction when compared with scramble controls (Figure 1). These results confirm the results of our siRNA library screening, which indicated that Gal-1, but not Gal-3, plays a key role in H-1PV infection in the cell lines tested.

Pre-Treatment with Lactose Inhibits H-1PV Infection
Gal-1 is a member of the galectins which belongs to a sub-family of lectins, defined by their highly conserved carbohydrate recognition domain (CRD) with the ability to bind to a number of beta-galactosides. Among the beta-galactosides, lactose is known to modulate the function of Gal-1 by directly binding to the CRD [33]. Hence, we hypothesised that treatment with lactose may interfere with H-1PV infectivity by competing with the virus for the interaction with Gal-1. To test this hypothesis, HeLa cells were pre-treated with increasing concentrations of soluble β-lactose before being infected with recH-1PV-EGFP. Pre-treatment with lactose decreased H-1PV transduction in a dose dependent manner ( Figure 2). These results provide further evidence that H-1PV interaction with Gal-1 plays a crucial role in modulating H-1PV infectivity.  Numbers represent the arithmetic mean percentage of EGFP-positive cells relative to the number of EGFP-positive cells observed in cells transfected with control siRNA, which was arbitrarily set as 100%. The independent experiment shown was repeated thrice each with three biologically independent samples (ns-not significant; *** p ≤ 0.001, calculated by using a one-way ANOVA).

Pre-Treatment with Lactose Inhibits H-1PV Infection
Gal-1 is a member of the galectins which belongs to a sub-family of lectins, defined by their highly conserved carbohydrate recognition domain (CRD) with the ability to bind to a number of beta-galactosides. Among the beta-galactosides, lactose is known to modulate the function of Gal-1 by directly binding to the CRD [33]. Hence, we hypothesised that treatment with lactose may interfere with H-1PV infectivity by competing with the virus for the interaction with Gal-1. To test this hypothesis, HeLa cells were pre-treated with increasing concentrations of soluble β-lactose before being infected with recH-1PV-EGFP. Pre-treatment with lactose decreased H-1PV transduction in a dose dependent manner ( Figure 2). These results provide further evidence that H-1PV interaction with Gal-1 plays a crucial role in modulating H-1PV infectivity. which was arbitrarily set as 100%. The independent experiment shown was repeated thrice each with three biologically independent samples (ns-not significant; ** p > 0.05; *** p ≤ 0.001, calculated by using a one-way ANOVA).

Galectin-1 Knockout Impairs H-1PV Infection in NCH125 Cells
To further investigate the biological role of Gal-1 in H-1PV infection, we took advantage of the CRISPR-Cas9 technology and established the NCH125 LGALS1 knockout cell line (LGALS1 KO). We also established the NCH125 control cell line (Control) using a non-targeting guide RNA control sequence. Given that H-1PV requires S-phase factors expressed in proliferating cells for a productive infection [8], we evaluated the proliferation of LGALS1 KO versus Control cells. Both cell lines proliferated at a similar rate, as shown by real-time monitoring of cell growth and viability via xCELLigence ( Figure S2). The independent experiment shown was repeated thrice each with three biologically independent samples (ns-not significant; ** p > 0.05; *** p ≤ 0.001, calculated by using a one-way ANOVA).

Galectin-1 Knockout Impairs H-1PV Infection in NCH125 Cells
To further investigate the biological role of Gal-1 in H-1PV infection, we took advantage of the CRISPR-Cas9 technology and established the NCH125 LGALS1 knockout cell line (LGALS1 KO). We also established the NCH125 control cell line (Control) using a non-targeting guide RNA control sequence. Given that H-1PV requires S-phase factors expressed in proliferating cells for a productive infection [8], we evaluated the proliferation of LGALS1 KO versus Control cells. Both cell lines proliferated at a similar rate, as shown by real-time monitoring of cell growth and viability via xCELLigence ( Figure S2).
We infected LGALS1 KO and Control cell lines with wild-type H-1PV and analysed the number of internalised virus particles by immunofluorescence using a specific anti-capsid antibody [29]. Two hours post-infection, fluorescence was significantly lower in LGALS1 KO cells than in Control cells ( Figure 3A). Then, we evaluated the levels of viral proteins 48 h post-infection with the wild-type H-1PV. In agreement with previous results, Western blotting analysis revealed that NS1 and VP1 protein levels were lower in LGALS1 KO cells than in Control cells ( Figure 3B). Finally, we analysed H-1PV transduction efficiency by infecting both cell lines with recH-1PV-EGFP. Consistent with reduced H-1PV entry, a significant decrease in transduction activity was found in LGALS1 KO cells (47%) in comparison with Control cells ( Figure 3C). Strikingly, transient transfection of LGALS1 KO cells with a plasmid encoding LGALS1 48 h prior to infection, by re-establishing Gal-1 protein levels to values similar to those observed in Control cells ( Figure 3C: Western blotting analysis), rescued the reduction in H-1PV transduction in these cells ( Figure 3C).   LGALS1.
LGALS1 KO cells were transfected with a plasmid encoding LGALS1, treated only with lipofectamine LTX (mock transfection) or left untreated. In addition, 48 h post-transfection, cells were infected with recH-1PV-EGFP for 24 h. Control cells were also included, and the level of virus transduction was set arbitrarily at 100%. The independent experiment shown was repeated twice each with four biologically independent samples (ns: p > 0.05; *** p ≤ 0.001, calculating using a one-way ANOVA). On the right side, Western blotting analysis shows the levels of Gal-1 at the time of infection. Beta-tubulin was used as a loading control.

Galectin-1 Knockout Decreases H-1PV Oncolytic Activity in NCH125 Cells
As LGALS1 knockdown/out decreased the overall amount of internalised H-1PV, we assessed whether this would result in reduced oncolytic activity. For this purpose, we assessed the susceptibility of LGALS1 KO and Control cell lines to H-1PV infection in a time course experiment in which viability of infected cells was assessed every 24 h for a total of 96 h. Whereas the viability of Control cells decreased progressively over time, the viability of LGALS1 KO cells, which were less sensitive to H-1PV infection, remained high throughout the experiment (above 75%; Figure 4A). Remarkably, the susceptibility of LGALS1 KO cells to H-1PV oncotoxicity was re-established by infecting the cells together with human recombinant Gal-1. Indeed, the viability of LGALS1 KO cells infected with H-1PV dropped from 74% to 18% in the presence of Gal-1; a control experiment showed that the protein itself was not toxic to the cells at the concentrations used ( Figure 4B). Together, these results highlight the critical role of Gal-1 in H-1PV infection in NCH125 cells.

Galectin-1 Plays a Role in H-1PV Virus Entry Rather Than Cell Surface Attachment
The results shown above indicate that Gal-1 is involved in the early steps of H-1PV infection. However, whether Gal-1 is required for H-1PV attachment at the cell surface, internalisation, or both events, remains to be determined. To elucidate the role of Gal-1 in H-1PV infection, we first performed virus binding and entry assays. LGALS1 KO and Control cell lines were infected with wild-type H-1PV at 37 • C for different times (0.5, 1, 2 and 4 h) and then cell-associated viral DNA was quantified by PCR. In agreement with previous results, we observed less cell-associated H-1PV DNA in LGALS1 KO cells than in Control cells ( Figure 5A). The addition of recombinant Gal-1 increased the number of cell-associated H-1PV genomes in LGALS1 KO cells to values that were similar to those found in H-1PV-infected Control cells ( Figure 5B). These results confirm the role of Gal-1 in H-1PV binding and/or entry. throughout the experiment (above 75%; Figure 4A). Remarkably, the susceptibility of LGALS1 KO cells to H-1PV oncotoxicity was re-established by infecting the cells together with human recombinant Gal-1. Indeed, the viability of LGALS1 KO cells infected with H-1PV dropped from 74% to 18% in the presence of Gal-1; a control experiment showed that the protein itself was not toxic to the cells at the concentrations used ( Figure 4B). Together, these results highlight the critical role of Gal-1 in H-1PV infection in NCH125 cells.  Concerning the possible involvement of Gal-1 in H-1PV cell surface attachment, LGALS1 KO and Control cell lines were inoculated with H-1PV at 4 • C for 1 h. Under these conditions, only virus attachment at the cell surface virus occurs, while cell entry is prevented [14]. After removing unbound H-1PV particles, those that remained attached to the cell surface were stained with an anti-capsid antibody and Flow cytometry analysis was performed. No significant differences in the fluorescence signal (H-1PV binding) were observed between the LGALS1 KO and Control cells ( Figure 5C). The same findings were obtained by quantitative PCR analysis ( Figure 5D). These results speak for an involvement of Gal-1 in H-1PV entry rather than in H-1PV binding to the cell surface.

Gal-1 Cooperates with Laminins in Mediating H-1PV Infection
Given that laminins containing the γ1 chain play determinant roles in H-1PV attachment at the cell surface as well as in H-1PV cell entry, we asked whether siRNAmediated silencing of LAMC1 in LGALS1 KO cells would further decrease H-1PV infection ( Figure S3). In agreement with the results shown above, reduced H-1PV cell uptake was observed in LGALS1 KO cells. Furthermore, given the important role of laminins in H-1PV cell attachment and entry, knockdown of LAMC1 gene expression also strongly reduced cell-associated H-1PV genomes in NCH125 cells. The reduction observed upon LAMC1 knockdown was not significantly different between LGALS1 KO cells and Control cells, indicating that, in the absence of Gal-1, the depletion of LAMC1 does not further reduce H-1PV cell uptake ( Figure 6A). To confirm these results, we repeated the double knock-down experiment in HeLa cells ( Figure S3). To this end, we silenced LGALS1 gene expression in the previously established HeLa LAMC1 KD cell line (in which the LAMC1 gene was knocked down via CRISPR-Cas9) [14]. In agreement with published results, infection of HeLa LAMC1 KD cells presented a 40% reduction in cell-associated H-1PV genomes in comparison with Control cells. The silencing of LGALS1 also decreased H-1PV cell uptake in HeLa cells, yet no significant difference was observed between HeLa control and HeLa LAMC1 KD cell lines ( Figure 6B). The fact that removal/reduction of both laminin γ1 and Gal-1 does not synergistically inhibit infection suggests that the two factors may act on the same H-1PV entry pathway. At the same time, the finding that, under these conditions, a fraction of viral particles is still able to penetrate cells supports the idea that H-1PV may also use alternative pathways and exploit other cellular factors to infect cells.

Gal-1 Is a Marker of Bad Prognosis in Various Tumour Types including GBM
Growing evidence indicates that overexpression of LGALS1 is associated with metastasis formation, tumour recurrence and poor tumour prognosis [34]. Our analysis of brain tumour expression datasets using the GlioVis web application (http://gliovis.bioinfo.cnio.es/, accessed on 23 January 2021) revealed that LGALS1 overexpression is associated with worse overall survival for brain tumours. Focusing particularly on GBM, we observed that these tissues have significantly higher expression of LGALS1 in comparison to those from healthy individuals ( Figure S4A) and LGALS1 levels increase with the severity of the malignancy from WHO grade II to IV ( Figure S4B). High LGALS1 expression is associated with poor prognosis in glioma ( Figure S5).
Next, we investigated whether Gal-1 protein levels varied between normal tissues, primary and recurrent GBM tissues. To this end, we used an in-house protein tissue microarray including a cohort of 110 GBM patient biopsies (61 primary and 49 recurrent GBM) and 12 biopsies from normal tissues (four each from brain, liver, and tonsil) and performed immunohistochemistry using an anti-galectin-1 antibody. Levels of Gal-1 were higher overall in GBM biopsies than in normal tissues. Among the GBM biopsies, we found a diversified Gal-1 expression profile with recurrent GBM tissues expressing significantly higher levels of Gal-1 than primary GBM tissues (45% of recurrent GBM tissues expressed medium or high levels of Gal-1, compared with 20% of primary GBM tissues; Figure 7). comparison with Control cells. The silencing of LGALS1 also decreased H-1PV cell uptake in HeLa cells, yet no significant difference was observed between HeLa control and HeLa LAMC1 KD cell lines ( Figure 6B). The fact that removal/reduction of both laminin γ1 and Gal-1 does not synergistically inhibit infection suggests that the two factors may act on the same H-1PV entry pathway. At the same time, the finding that, under these conditions, a fraction of viral particles is still able to penetrate cells supports the idea that H-1PV may also use alternative pathways and exploit other cellular factors to infect cells.

LGALS1 Expression Profile of NCI-60 Cells Positively Correlates with H-1PV Oncotoxicity
Given the important role of Gal-1 in H-1PV infection, we looked for a putative correlation between LGALS1 expression levels and H-1PV oncotoxicity. To this end, we screened 53 cancer cell lines from the NCI-60 panel for their susceptibility to H-1PV [14]. The gene expression profiles of these cell lines are fully characterised and publicly available [35]. We assessed virus-mediated oncotoxicity by monitoring cell viability in real time using xCELLigence [14]. We calculated the viral MOI responsible for killing 50% of the cell population at 72 h post-infection (EC50). Using gene expression data from NCI-60 and the Cancer Cell Line Encyclopedia (CCLE), we found that LGALS1 mRNA expression levels anti-correlated with EC50 values, suggesting that cells expressing higher levels of LGALS1 may be more susceptible to virus killing activity (Figure 8).

LGALS1 Expression Positively Correlates with H-1PV Oncolysis in Glioma Cell Lines
Glioma cancer cell lines are generally susceptible to H-1PV oncolysis [36]. However, not all cancer cell lines respond similarly to H-1PV oncolysis: they range from highly to A similar anti-correlation was previously shown for LAMC1 [14]. Therefore, we asked whether we could better predict the susceptibility of a certain cancer cell to H-1PV infectivity and cell killing by analysing the expression levels of the two genes together. As negative controls, LGALS3 (selected for showing no apparent role in H-1PV infection) and GAPDH were chosen. Using both NCI-60 and the CCLE databases, we found that a slight increase in predictability of EC50 may be achieved by combining LAMC1 and LGALS1 in a single linear regression model. For NCI-60, R 2 of the combined model was 0.313 (p = 8.3 × 10 −5 ), while simple models gave R 2 of 0.188 (LAMC1, p = 1.2 × 10 −3 , R = −0.433) and 0.260 (LGALS1, p = 9.8 × 10 −5 , R = -0.510). Control genes show no significant linear relation (LGALS3: R 2 = 0.015, p = 0.38, GAPDH: R 2 = 0.018, p = 0.34) ( Figure S6). Similar results were obtained using the CCLE dataset ( Figure S7). These results are in line with the important role that laminin γ1 and Gal-1 play in H-1PV cell surface recognition and entry.

LGALS1 Expression Positively Correlates with H-1PV Oncolysis in Glioma Cell Lines
Glioma cancer cell lines are generally susceptible to H-1PV oncolysis [36]. However, not all cancer cell lines respond similarly to H-1PV oncolysis: they range from highly to lowly permissive, or are even resistant. We recently described four glioma cell lines that are semi-permissive to H-1PV infection, namely U251, LN308, T98G and A172-MG, which all express low levels of LAMC1 mRNA [14]. However, it is possible that other cell components may account for the poor susceptibility of these cell lines to virus infection. Using the NanoString technology, we found that LGALS1 mRNA levels were lower in the four aforementioned cell lines, as well as in the control normal human astrocytes, than in two H-1PV-sensitive glioma cell lines (NCH125 and NCH37) ( Figure 9A). Monitoring of cell viability in real time by xCELLigence confirmed our previous results showing that NCH125 and NCH37 cell lines are efficiently killed by H-1PV at an MOI of 5 (pfu/cell). By contrast, U251, LN308, T98G, A172-MG cell lines as well as control normal human astrocytes were not. Remarkably, susceptibility of the four semi-permissive glioma cell lines to H-1PV oncotoxicity was substantially enhanced by the addition of human recombinant Gal-1 ( Figure 9B). Consistent with our previous results, the addition of exogenous Gal-1 promoted H-1PV entry in the four glioma cell lines leading to an increase in the number of cell-associated viral genomes by 1.5-to 2.8-fold ( Figure 9D), while not interfering with viral binding to the cell surface ( Figure 9C). Together, these results confirm that Gal-1 plays a critical role in H-1PV infection at the level of virus entry and provide evidence that Gal-1 levels can determine the outcome of H-1PV infection. These results further support the importance of Gal-1 to H-1PV oncolytic activity and pave the way for its use in predicting the success of H-1PV infection.

Discussion
Following more than five decades of preclinical research, H-1PV monotherapy has been evaluated in patients with recurrent GBM and pancreatic carcinoma in early-phase clinical trials. H-1PV treatment was demonstrated to be safe, well tolerated and associated with surrogate evidence of anticancer efficacy, including immunoconversion of the tumour microenvironment and improved progression-free survival and overall survival in comparison with historical controls [9]. These promising results have motivated research aiming at further improving H-1PV efficacy [11,37].
The virus life cycle is a multistep process that is heavily dependent on the presence and abundance of viral (co-)receptors, processing enzymes and proteins required for a productive infection. The levels and activities of these components may vary in different cancer cells, determining their susceptibility to a particular virus. We anticipate that a better understanding of the H-1PV life cycle may provide the cues to further develop H-1PV-based therapies. For instance, this knowledge may help to identify new drugs that enhance H-1PV replication in cancer cells and/or oncolytic activity by modulating H-1PV-related cellular pathways. Furthermore, the cellular factors involved in the H-1PV life cycle may also serve as biomarkers to predict whether a certain tumour is susceptible or resistant to H-1PV infection.
Recently, we found that H-1PV enters cancer cells via clathrin-mediated endocytosis, a process that involves dynamin and requires a low pH in the endocytic compartments [15]. We also reported that laminins, in particular those containing the laminin γ1 chain, act as attachment factors at the cell surface for a successful H-1PV infection [14]. In particular, we found that sialic acid moieties in the laminins provide a docking place for the virus to anchor to at the cell surface [14]. Laminin γ1 was originally identified as a modulator of the H-1PV life cycle in a siRNA screening using a druggable-genome library performed in HeLa cells. The same screening revealed Gal-1 as another top activator of H-1PV infection. Indeed, silencing of LGALS1 impaired H-1PV virus transduction by approximately 70% in HeLa cells. These results prompted us to explore whether Gal-1 is involved in the early steps of H-1PV infection.
In the present study, we show that Gal-1 plays a central role in H-1PV infection at the level of H-1PV cell entry but not cell attachment, indicating a role that is distinct from that of laminins in the virus cell cycle. By contrast, knockdown of LGALS3 did not impair H-1PV infection (Figure 1), further supporting the specificity of the interaction between H-1PV and Gal-1.
Until the present study, Gal-3 was the only galectin that had been implicated in a protoparvovirus infection. Indeed, knockdown of LGALS3 rendered LA9 mouse fibroblasts less susceptible to MVM infection. This phenotype was not due to reduced binding to the cell surface; instead, Gal-3 was responsible for promoting efficient virus uptake [24]. Our results indicate that Gal-1 is essential for a productive H-1PV infection at the level of cell entry, with no evidence of its requirement in viral binding to the plasma membrane in NCH125 cell lines. Therefore, these findings suggest that the mechanisms through which Gal-1 mediates H-1PV entry are similar to those of Gal-3 in MVM infection [23]. The fact that our results do not support an involvement of Gal-3 or clathrin-independent endocytosis in the H-1PV entry process [15] suggests that H-1PV and MVM engage different galectins for their entry processes which may contribute to their different tropism. At this point, we cannot also exclude that, in addition to Gal-1, other members of the galectin family may participate in the H-1PV entry process.
Based on our results, we envision a model in which H-1PV interacts with different classes of molecules, rather than with a single cell surface receptor, in order to enter cancer cells. Laminins containing γ1 chains would accumulate virus in the vicinity of the cell surface via sialic acid, while Gal-1 would promote the efficient internalisation of virus particles into a clathrin-coated pit. After engagement of these factors, H-1PV would penetrate the cells preferentially via clathrin-mediated endocytosis [15].
One possibility arises that H-1PV hijacks extracellular Gal-1 to enter the cells together with the protein. Our results support this idea by showing that the addition of exogenous purified Gal-1 boosts H-1PV infection at the level of virus entry, thereby sensitising semipermissive cancer cells to H-1PV-mediated oncolysis. On the other hand, we cannot exclude that Gal-1 may play additional roles at post-entry levels, including viral trafficking and DNA uncoating.
Galectins are known to be synthesised in the cytoplasm and accumulate there until they are secreted via a poorly characterised pathway [38]. The exact mechanism through which galectin(s) translocate across the cell membrane remains also poorly understood [39]. However, previous research has shown that inhibition of the lipid raft-dependent pathway does not impede Gal-1 internalisation; instead, a total block of Gal-1 internalisation was observed only when both clathrin-mediated endocytosis and lipid rafts were disrupted, demonstrating that Gal-1 enters cells through various mechanisms, including clathrinmediated endocytosis [40,41]. Therefore, it may be possible that H-1PV uses Gal-1 to enter cancer cells through clathrin-mediated endocytosis.
Alternatively, Gal-1 could mediate the binding of H-1PV to other cellular factors involved in its entry, e.g. a transmembrane receptor or a co-receptor. A number of studies have shown that the multivalent binding activity of Gal-1 and other galectins is able to cross-link carbohydrates and glycoconjugates [42,43]. For instance, Gal-1 cross-linking has the ability to massively redistribute a diverse population of glycoproteins on the cell surface of T cells and segregate them into membrane microdomains [44]. Gal-1 has also been associated with the assembly and remodelling of the extracellular matrix [45], and has been shown to bind to various components present there, especially those containing polylactosamine chains, such as laminins [45,46]. In this respect, LAMC1 knockdown in NCH125 LGALS1 KO cells, or LGALS1 knockdown in HeLa LAMC1 KD cells did not further decrease H-1PV cell uptake ( Figure 6), suggesting that laminins and Gal-1 may cooperate in the early steps of H-1PV infection presumably by performing overlapping roles. However, as there is still residual internalisation of H-1PV ( Figure 3C), it is likely that H-1PV may use alternative pathways to enter cells and that other unidentified cell factors are involved in this process. On the other hand, other laminins with or without the laminin-γ1 chain may contribute to residual H-1PV entry, independently from Gal-1. Future studies are needed to further characterize the pathways involved in H-1PV cell binding-entry and to determine whether other cellular factors besides laminins and Gal-1 are involved in these events.
Gal-1 and galectins in general have been described as playing important roles in different aspects of various viral infections, leading to their promotion or inhibition. For instance, Gal-1 stabilised the binding of HIV-1 to CD4 + T cells by cross-linking the viral gp120 and the host CD4 receptor, thereby helping HIV-1 to infect these cells [21]. Furthermore, soluble Gal-1 enhanced the uptake of HIV-1 by monocyte-derived macrophages, whereas Gal-3 had no effect on infection [47]. Enterovirus 71 is another example where Gal-1 has a supporting role. Gal-1 facilitates infection by interacting with the carbohydrate residues in VP1 and VP3 domains, leading to a more efficient release and dissemination of virus to other cells [48]. During Nipah virus (NiV) infection, Gal-1 enhances virus cell attachment to primary human endothelial cells [49]. However, later in the NiV replication cycle, Gal-1 seems to exert an inhibitory effect. Indeed, Gal-1 specifically binds to the viral glycoproteins NiV-F and NiV-G, which are responsible for cell-cell fusion and syncytia formation, thus blocking virus infection [50,51]. The inhibitory effect of Gal-1 is also observed in Influenza A infection, both in vitro and in vivo. Gal-1 binds directly to the envelope glycoproteins, stopping influenza from inducing hemagglutination and thereby impairing infectivity. Accordingly, influenza infection led to poorer survival rates in LGALS1 KO mice than in wild-type mice [22].
Apart from their role in virus infections, galectins are also linked to apoptosis, angiogenesis, cell migration and tumour-immune escape [52]. In particular, high levels of Gal-1 are associated with cancer progression, poor prognosis and recurrence (reviewed in [34]). Several cancer types have been implicated, including gastric cancer [53], ovarian cancer [54], pancreatic cancer [55] and GBM [56]. In agreement with previous studies, our in silico analysis revealed that GBM presents significantly higher levels of LGALS1 than normal tissues, and that LGALS1 expression increases from grade II to IV gliomas ( Figure S4). In terms of survival, high LGALS1 expression is associated with a poor prognosis in glioma ( Figure S5). To complement the bioinformatic analysis, we assessed a cohort of 122 patient biopsies by immunohistochemistry. We showed that Gal-1 protein levels vary across GBM with higher levels found in biopsies from patients with recurrent versus primary GBM, while in normal tissues the levels were relatively low ( Figure 7). These findings corroborate previous studies showing that elevated levels of Gal-1 are associated with GBM progression [57][58][59]. They also further support the use of H-1PV to treat GBM, especially those cases with high Gal-1 protein content, given the key role that this protein has in virus entry and oncolysis.
We also found a correlation between LGALS1 expression levels and the ability of H-1PV to induce oncolysis in 59 cancer cell lines (Figures 8 and 9). These results suggest that tumours with elevated LGALS1 expression levels are likely to be more susceptible to H-1PV oncolytic activity. Building on these findings, we show that, while virus attachment is unaffected, virus entry is enhanced in the U251, LN308, U87 and A172-MG semi-permissive cell lines when H-1PV is administrated together with recombinant Gal-1 (Figure 9). Under these conditions (2 h from infection), H-1PV is still likely at the very early stages of infection, strongly supporting that Gal-1 plays a role in the H-1PV cell entry process. In agreement with enhanced infection, we found that susceptibility of these semi-permissive glioma cells to H-1PV oncolytic activity increases upon addition of exogenous Gal-1, suggesting that a certain level of Gal-1 is required for an efficient and productive H-1PV infection. Together, these findings support the idea that Gal-1 may represent a limiting factor for H-1PV oncolysis, and therefore, that tumours with high Gal-1 expression are more likely to respond to H-1PV treatment.
A similar correlation was previously shown for LAMC1 where cell lines highly expressing this gene were found to be more susceptible to H-1PV oncotoxicity [14]. Anticorrelation analysis of LAMC1 obtained in the previous work (Pearson correlation R = −0.52, R 2 = 0.27) stated a slightly higher anti-correlation than the one obtained in this re-analysis (LAMC1 R = −0.433, R 2 = 0.188). Using the current datasets, LGALS1 anti-correlation (LGALS1, R = -0.510, R 2 = 0.260) is moderately higher than that of LAMC1. Update of the datasets with acquisition of novel mRNA sequencing data may account for these differences. Interestingly, the anti-correlation is strongest when the LGALS1 and LAMC1 expression levels are analysed together, suggesting that their combined expression analysis may better predict the success of H-1PV infection against a certain tumour. This is in line with the concept that both genes play an important role in H-1PV infectivity. Our results also present new scenarios for treatment in which exogenous administration of recombinant Gal-1 could constitute a promising adjunct in H-1PV-based therapies. However, given the role of Gal-1 in carcinogenesis [60], its possible use with H-1PV must be carefully evaluated.
Supplementary Materials: The following are available online at https://www.mdpi.com/article/10 .3390/v14051018/s1, Figure S1: LGALS1 and LGALS3 knockdown in HeLa, NCH125 and BxPC3 cell lines. Figure S2: Cell proliferation of NCH125 Control versus NCH125 LGALS1 KO. Figure S3: Depletion of both LAMC1 and LGALS1 in NCH125 and HeLa cells. Figure S4: LGALS1 overexpression in high-grade GBM. Figure S5: Higher levels of LGALS1 are associated with poor prognosis in glioma. Figure S6: NCI60 database retrieved LGALS1 and LAMC1 expression levels showed anti-correlation with EC50 values and a slight increase in the ability to predict susceptibility to H-1PV induced oncotoxicity. Figure S7: CCLE database that retrieved LGALS1 and LAMC1 expression levels showed anti-correlation with EC50 values and showed a slight increase in the ability to predict susceptibility to H-1PV induced oncotoxicity. Funding: This study was supported initially by a seeding grant from Institut National du Cancer (INCA) and at later stages by a grant from ORYX GmbH to A.M. Our deepest gratitude also goes to André Welter, the Luxembourg Cancer Foundation and Télévie for supporting oncolytic virus immuno-therapy.

Institutional Review Board Statement:
The tissue microarray prepared from GBM biopsies was approved by the Regional Ethics Committee of Bergen, (Norway) with the number 2017/2505.

Informed Consent Statement:
Informed consent was obtained from all subjects involved in the study.
Data Availability Statement: All relevant data supporting the findings of this study are available within the paper and its Supplementary material. All other data are available from the corresponding author on request. Figure 8, Figure S6 and Figure S7 were generated employing the data sets publicly available at the CellMiner™ (https://discover.nci.nih.gov/cellminer, accessed on 23 January 2021) and at The Cancer Cell Line Encyclopedia portals (https://portals.broadinstitute.org/ccle, accessed on 23 January 2021). Figures S4 and S5 were generated using the Glio-Vis data portal (http://gliovis.bioinfo.cnio.es/, accessed on 23 January 2021).