Oncolytic vaccinia virus GLV-1h68 exhibits profound antitumoral activities in cell lines originating from neuroendocrine neoplasms

The oncolytic vaccinia virus GLV-h168 was kindly provided by Genelux Corporation (San Diego, CA, USA). GLV-1 h68 is a genetically engineered OV originating from the vaccinia Lister strain and also known under the proprietary name GL-ONC1 [11]. It was genetically modified by inserting three transgenes allowing therapeutic monitoring in its genome; RUC-GFP is employed for monitoring via fluorescence microscopy in this study.

The six cell lines derived from NENs are outlined in Table 1. H727, UMC-1, QGP-1, and NEC-DUE1 cells were maintained in RPMI-1640 medium (Gibco, Waltham, MA, USA) supplemented with 10% fetal calf serum (FCS, Biochrom, Berlin, Germany). BON-1 cells were cultured in Dulbecco’s modified Eagle’s Medium (DMEM, Sigma-Aldrich, St. Louis, MO, USA) supplemented with 10% FCS and HROC-57 cells required DMEM/F12 medium (Gibco) with 10% FCS. CV-1 African green monkey kidney cells were purchased from ATCC (CCL-70) and cultured in DMEM supplemented with 10% FCS. All cells were cultured at 37 °C and 5% CO₂ in a humidified atmosphere and seeded in 6- and 24-well plates for the respective assays.

For infection, cells were seeded 24 h before. GLV-1 h68 was diluted in the respective amount of DMEM supplemented with 2% (v/v) FCS to prepare the infection medium. The dilution ratio was calculated to ensure infection with a specific multiplicity of infection (MOI, effector target ratio, i.e. viral particles per cell). Cells were rinsed with phosphate buffered saline (PBS, Sigma-Aldrich) prior to infection, shortly before the respective amount of infection medium was added. Virus infection was allowed to take place for 1 h with swaying every 15 min. Then, infection medium was replaced with normal cell culture medium. Mock treatment was conducted with DMEM supplemented with 2% (v/v) FCS. For sole treatment with everolimus (Sellekchem, Munich, Germany), cell culture medium was replaced with medium containing everolimus at the respective concentration at 24 h post cell seeding. For combinatorial treatment with GLV-1 h68 and everolimus, infection medium was replaced with cell culture medium containing everolimus in the respective concentration.

To assess tumor cell viabilities at 72 and 96 h post infection (hpi), the Sulforhodamine B (SRB) assay was employed. This viability assay measures cell density compared to mock treatment by quantifying the number of adherent (viable) cells [34]. For this purpose, NET/NEC cells were seeded in 24-well plates and infected with OV, mock treated, treated with everolimus or OV and everolimus together. At the respective time point of analysis (at 72/96 hpi), cells were fixed with 10% (v/v) trichloroacetic acid (Carl Roth, Karlsruhe, Germany) after rinsing them with 4 °C cold PBS. Fixation was allowed for at least 30 min at 4 °C. Next, cell cultures were washed with water. Then, fixed cells were stained with SRB dye (0.4% (w/v) in 1% (v/v) acetic acid; Sigma-Aldrich) for at least 10 min and rinsed afterwards with 1% (v/v) acetic acid (VWR, Radnor, PA, USA) to remove unbound SRB dye. After drying for another 24 h, 10 mM TRIS base (pH 10.5; Carl Roth) was added to solve remaining SRB dye. To measure the amount of bound SRB dye, the absorbance of the inoculum at a wavelength of 550 nm was determined in duplicates (using a Tecan Genios Plus Microplate Reader). As the SRB dye binds to cellular proteins, the absorbance correlates with cell density. In the figures, cell density of mock treated cells was adjusted as 100%; percentages refer to mock treatment.

For microscopy, an Olympus IX 50 microscope with a PhL phase contrast filter and a fluorescence filter for GFP detection was used. Pictures were taken with the F-View Soft Imaging System (Olympus) and were colored and overlaid afterwards with the analySIS image-processing software and Apple Preview 10.0 software.

H727 cells were seeded in 96-well plates (E-Plate 96, Roche Applied Science, Mannheim, Germany). The xCELLigence RTCA SP system (Roche Applied Science) was employed to observe impedance of the cell layer in 30 min intervals over 120 h. 24 h after seeding, cells were infected with GLV-1 h68 using MOIs 0.1 and 0.25 or treated with 0.1% (v/v) Triton for lysis control. The measured impedance was used to calculate Cell Index values with the RTCA Software (1.0.0.0805).

Plaque assays were conducted in order to determine the concentration of viral particles in cell cultures as described previously [11]. H727 and BON-1 cells were seeded in 6-well plates and infected with MOIs which led to approx. 50% reduction of tumor cell densities. One hour after virus infection, plates were carefully washed with PBS to remove all extracellular viral particles; then culture medium was added. Every 24 h and at 1 hpi, infected cells and medium were harvested by scraping them into the culture medium. Subsequently, the harvested samples were frozen at − 80 °C. For analysis of the samples, the CV-1 indicator cells were infected with the frozen samples. For this purpose, thawed samples were titrated in duplicates in 10-fold dilutions (10^− 1 to 10^− 6) on the indicator cells. Cells were incubated for 1 h and plates were moved every 15 min to ensure sufficient virus infection. Next, cells were overlaid with 1 ml of 1.5% (w/v) carboxymethylcellulose (CMC, Sigma-Aldrich) in DMEM with 5% (v/v) FCS and 1% (v/v) Pen/Strep per well. As the CMC medium prevents viral spread through the culture medium, each infective viral particle creates a plaque by radial infective spread after 48 h. After 48 h, cell layers were stained with crystal violet staining solution (0.1% (w/v) in 5% (v/v) ethanol, 10% (v/v) formaldehyde, Fluka Chemie AG) for 4 h. Then, the culture plate was washed with water and plaques could be counted. With the plaque count and titrated dilutions, viral titers (plaque forming units (PFU) per ml) could be calculated.

First, effects of a monotherapy of the six NET/NEC cell lines employing the vaccinia virus vector GLV-1 h68 were studied. In this purpose, SRB viability assays were conducted to evaluate cytostatic and cytotoxic effects of the OV on neuroendocrine cancer cells and to identify oncolysis-sensitive and -resistant tumor cell lines. Further, microscopic fluorescence pictures were taken to visualize oncolysis and directly detect and prove virotherapeutic vector-based transgene (GFP) expression. Next, a real-time cell monitoring assay was employed to distinguish between cytostatic and cytotoxic nature of the effect and study the dose dependency of this circumstance. Finally, the production of viral progeny, which forms the basis of the intratumoral infectious spread of an OV, was studied by assessing virus titers sequentially over time.

All NET/NEC cell lines were infected with multiplicities of infection (MOIs) of GLV-1 h68 in logarithmic steps, ranging from 0.0001 to 1. Taking the first results of the SRB viability assays into account, the MOIs were modified by adding MOI 0.5 instead of MOI 0.0001 for all cell lines except BON-1; MOIs 0.025 and 0.05 were added for BON-1 cells, while MOIs 0.0001 and 0.001 were left out. A threshold for clinically relevant anti-tumor activities was set at 60% of tumor cells being residual in SRB viability assays after an infection period of 96 h (Fig. 1, dotted horizontal lines). Three categories to classify cellular response to GLV-1 h68 virotherapy were introduced: (i) highly permissive cell lines, meeting the 60% threshold with MOI 0.1 or less after 96 h; (ii) permissive cell lines requiring MOI 0.5 to meet the threshold at 96 hpi, and (iii) resistant cell lines which required more than MOI 0.5 to meet the threshold at 96 hpi.

It was found that GLV-1 h68 is able to infect and kill all six NET/NEC cell lines, requiring different MOIs for the same effect. For all tumor cell lines, a dose dependency was observed, meaning that a higher MOI resulted in a lower number of residual tumor cells at the end of the observation period, i.e. at 96 hpi. As a result, three highly permissive, three permissive and no resistant cell lines could be identified. BON-1 pancreatic NET (pNET) cells were found to be most sensitive to GLV-1 h68-mediated oncolysis, exhibiting a remaining tumor cell mass of 60% at 96 hpi when using a MOI of only 0.01 (Fig. 1c). For all other NET/NEC cell lines higher MOIs had to be applied in order to meet the 60% threshold at 96 hpi: MOI 0.1 was sufficient for H727 and HROC-57 cells (Fig. 1a and e); accordingly, BON-1, H727, and HROC-57 cells were classified as highly permissive. In contrast, UMC-11, QGP-1, and NEC-DUE1 cells required MOI 0.5 and were classified as permissive. An equal response pattern could be found in all three permissive cell lines. All three showed a significant reduction of remnant tumor cells with MOI 0.1 and met the threshold with MOI 0.5. Finally, a remaining tumor cell count of approx. 15% was reached with MOI 1 in all three cell lines (Fig. 1b, d, and e).

Overall, GLV-1 h68 was able to reduce the tumor cell masses to a minimum of less than 10% in 3 out of 6 NET/NEC cell lines.

In summary, no neuroendocrine cancer cell line turned out to be resistant to GLV-1 h68-mediated oncolysis. The three highly permissive cell lines were found to be BON-1 originating from a pNET, HROC-57 originating from a colon NEC and the lung NET derived cell line H727. Given that the three other cell lines showed very similar responses, no obvious relation between anatomical origin and treatment response could be identified in this experiment.

As GLV-1 h68 encodes a fluorescent GFP transgene for therapeutic monitoring, microscopic pictures were taken to prove viral infection and replication via transgene expression and observation of cell layer densities (Figs. 2 and S1). The same MOIs as in the SRB viability assay (Fig. 1) were applied. As a result, a loss of cell density could be observed in all infected neuroendocrine cancer cell lines, consistent with results from the SRB viability assay, where all tumor cell lines were found to respond to virus infections. Moreover, all analyzed NET/NEC cell lines were found to express the GFP transgene when being infected with GLV-1 h68. Of note, lower cell confluency and intensities of the fluorescence signals were found to correlate to the MOIs being applied (Figs. 2 and S1). This does not apply for HROC-57 cells, as the confluency was also low in uninfected cells (mock). However, with the highly permissive cell lines (H727, BON-1, HROC-57), the highest MOI displayed lower transgene expression, most likely because of a high rate of oncolysis and therefore a lower cell count expressing the fluorescent GFP transgene. This phenomenon is also visible with permissive QGP-1 cells and on the respective pictures taken at 72 hpi, although to a lesser extent (Figure S1). Mock treatment did not display any fluorescence at all.

To precisely investigate the nature of the effect of GLV-1 h68 on neuroendocrine cancer cells, a real-time cell monitoring assay was employed. The lung NET cell line H727 was picked as representative cell line because it showed a stable, average response to GLV-1 h68 in the experiments described above. Two MOIs (0.1 and 0.25), which resulted in remaining tumor cell numbers of around 50% according to SRB viability assay performed at 96 hpi, were chosen for infection. The xCELLigence RTCA assay measures cellular impedance, which was shown to correlate with cell number, cell size/morphology and cell attachment quality [35]. Taking the previous SRB viability assays and applied cell lysis control with Triton X-100 into account, the Cell Index can be seen as a surrogate for cell viability in this context.

Different treatment modalities were initiated at 24 h after cell seeding. As expected, treatment with the cell lysis control Triton X-100 immediately resulted in a complete tumor cell lysis (Fig. 3; green dotted line). In contrast, virus infections showed similar results to mock treatment in the first 24 hpi. In the further course of the experiment, the impedance of infected cells decreased continuously, indicating not only a cytostatic but also a cytotoxic effect of GLV-1 h68. The higher MOI (0.25) results in lower cell viability in the end, but not in a faster mechanism of action, also showing the first impairment of tumor cell growth at 24 hpi and the peak of cell viability at 36 hpi (Fig. 3; line with grey squares). Taken together, GLV-1 h68 was proven to exhibit a pronounced oncolytic effect on the neuroendocrine tumor cell line H727 and also a dependency on the infectious dose being applied. Thus, findings of the SRB viability assay could be confirmed.

As the production of viral progeny is an important step in the underlying mechanism of oncolytic virotherapy, virus titers obtained by neuroendocrine cancer host cells were sequentially determined every 24 h during the whole period of infection. Hence, the lung NET cell line H727 (Fig. 4a) and the pNET cell line BON-1 (Fig. 4b), which was found to be the tumor cell line being most sensitive to GLV-1 h68 treatment, were picked to further investigate tumor cells being established from different anatomical origins. Both NET cell lines were infected with MOIs achieving around 50% reductions of tumor cell counts in the SRB viability assays. Shortly after virus infection, all extracellular viral particles were removed so that only viral particles which had already entered the cells after a 1-h infection period could produce viral progeny.

As a result, high levels of viral replication could be detected in both tumor cell lines (Fig. 4). Titers over 10⁷ plaque forming units (PFU)/ml were easily reached within 72 h. A stagnation of virus titer growth could be observed for H727 after 72 h and a reduction of viral titer between 72 and 96 h was detected with BON-1 cells.

Next, a combinatorial treatment with the mTOR inhibitor everolimus was evaluated by comparing a combinatorial approach (GLV-1 h68 + everolimus) to GLV-1 h68 monotherapy. In this purpose, SRB viability assay and virus quantification were conducted.

SRB viability assays were carried out using the lung NET cell line H727 and the NEC cell line NEC-DUE1, which both are tumor cell lines being generated from different anatomical origins. H727 cells were classified as highly permissive to GLV-1 h68 monotherapy whereas NEC-DUE1 cells were classified as permissive (Fig. 1). Again, MOIs leading to around 50% tumor cell reductions were chosen (0.1 and 0.25 for H727; 0.25 and 0.5 for NEC-DUE1). Everolimus was administered in concentrations of 1 nM for H727 cells and 0.25 nM for NEC-DUE1 cells, respectively.

As a result, the addition of GLV-1 h68 to sole everolimus treatment was found to be able to further reduce the remaining tumor cell count (Fig. 5). This was observed in both cell lines tested and with both MOIs employed in each cell line. With both cell lines, no statistical significance was found for the addition of the respective lower MOI to Everolimus treatment alone (p > 0.05). By adding MOI 0.25 for H727 cells and MOI 0.5 for NEC-DUE1 cells, the combinatorial treatment was able to reduce tumor cells significantly more than Everolimus alone (Fig. 5).

With H727 cells, the addition of MOI 0.25 to Everolimus reduced the remaining tumor cell count by 11% from 65 to 54%. Interestingly, the benefit of the combinatorial therapy appeared to be more pronounced in NEC-DUE1 cells. By adding MOI 0.5 to Everolimus treatment alone, the remnant tumor cells were reduced by 17% from 59 to 42%. However, the extent of this effect was limited, thereby not representing any additive mechanism of action.

To investigate whether everolimus has any impact on virus replication, virus titers were assessed when GLV-1 h68 was employed in a combinatorial setting with everolimus (Fig. 6, dotted lines). In both NET cell lines (H727, BON-1), where virus replication was determined previously, everolimus did not affect the production of viral progeny in any way.

Taking the results from both assays into account, the final benefit of the combinatorial therapy after 96 h is visible but only small. Everolimus did not limit virus replication in a particular way. Given that evidence base, the combinatorial therapy of GLV-1 h68 with everolimus was not found to be inferior to either monotherapy and can be regarded as a possible future combinatorial treatment option for metastatic neuroendocrine cancer.