The oncolytic vesicular stomatitis virus VSV-GP shows profound oncolytic activity in NUT-carcinoma cell lines
Rieka C. Buchenau, Anne M. Schiller, Julia Beil, Susanne Berchtold, Andrea Schenk, Irina Smirnow, Mary E. Carter, Martin Schaller, Birgit Fehrenbacher, Anneli Vollert, Benjamin Schoeps, Rainer Kleemann, Ulrich M. Lauer, Linus D. Kloker
TL;DR
This study shows that the oncolytic virus VSV-GP effectively kills NUT carcinoma cells and identifies factors influencing its effectiveness.
Contribution
The first evaluation of VSV-GP's virotherapeutic efficacy in NUT carcinoma cell lines.
Findings
VSV-GP caused significant oncolysis and apoptosis in five out of seven NUT carcinoma cell lines.
Resistance mechanisms were linked to functional IFN-β signaling despite suppressed IFN-β synthesis.
VSV-GP's entry receptor was equally expressed in both permissive and resistant cell lines.
Abstract
NUT carcinoma (NC) is a rare, yet aggressive disease (median survival of 6.5 months) of the young defined by a translocation of the nuclear protein in testis gene 1 (NUTM1). Neither surgery nor radiochemotherapy or targeted therapies provide effective disease control, thereby creating a need for innovative therapeutic strategies. Recently, the recombinant oncolytic virus (OV) VSV-GP entered phase I clinical testing (NCT05155332). We (1) visualized VSV-GPs’ replication cycle in transmission electron microscopy; (2) analyzed VSV-GP’s oncolytic efficacy and replication kinetics using viability assays, xCELLigence analyses, RT-qPCR, TCID50 assays, and FACS analysis; as well as (3) explored potential resistance mechanisms in seven human NC cell lines. Upon infection with VSV-GP, we noticed profound oncolysis, apoptosis, and high replication kinetics in five out of seven NC cell lines.…
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
Figure 6
Figure 7
Figure 8Peer 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
TopicsVirus-based gene therapy research · Virology and Viral Diseases · Poxvirus research and outbreaks
Introduction
NUT carcinoma (NC) is a rare, yet aggressive tumor disease that is defined by a reciprocal chromosomal translocation of the NUTM1 (nuclear protein in testis gene 1, chromosome 15q14) gene. In over 70% of cases, NUTM1 is fused to BRD4, a representative of the genes encoding for the bromodomain and extraterminal motif (BET) protein family localized on chromosome 19p13.1.1 Other fusion partners include BRD3 and NSD3. NC is often reported to arise from epithelial structures along the midline of the body.1 Among other locations, NC was observed in the lungs,2 thymus,3 salivary,4^,^5^,^6 and adrenal glands.7 Due to rapid lympho- and hematogenic metastasis, the diagnosis is often made in a stage of advanced disease. With a median survival of only 6.5 months, prognosis is grim for affected patients, and, regrettably, many of them are adolescents or young adults (median age 23.6 years).8^,^9
Oncolytic virotherapy employs viruses as therapeutic agents to infect and kill tumor cells as well as enhance the body’s own anti-tumor immune reaction. Viruses are capable of self-amplification10 and exhibit a tumor cell specificity caused by an aberrant receptor expression11^,^12^,^13 or metabolism in tumor cells14 and, above all, by a deterioration of antiviral defense mechanisms such as the type I and II interferon (IFN) signaling pathways.15^,^16^,^17 The herpes simplex virus (HSV) T-VEC (talimogene laherparepvec)—the first OV approved by both the US Food and Drug Administration and the European Medicines Agency for the treatment of malignant melanoma—already proved to be highly effective in NC in vitro and in vivo.18^,^19^,^20
The recombinant vesicular stomatitis virus VSV-GP is an enveloped, negative-stranded RNA virus from the Vesiculovirus genus pseudotyped with the envelope glycoprotein GP from the lymphocytic choriomeningitis virus (LCMV; variant LCMV-WE-HPI) by replacing the VSV-G locus with LCMV-GP.21 It is currently being investigated with respect to its oncolytic potential in a clinical phase I trial in various solid advanced and metastatic cancers (NCT05155332). VSV-GP is a promising candidate due to a number of favorable characteristics: fast replication kinetics, effective cytotoxicity, and a distinct tumor specificity due to its inherent sensitivity toward the anti-viral cytokine IFN-β.22^,^23 Unlike other OVs, where an anti-viral immune response often diminishes oncolytic efficacy, pre-existing immunity against VSV-GP is very rare in humans.24 Additionally, the pseudotyping with the GP derived from LCMV does not only abrogate neurotoxicity21 but also helps to circumvent anti-VSV humoral immunity, which arises from the formation of neutralizing antibodies21^,^25 or complement inactivation.26 Thus, VSV-GP is the first OV to retain its oncolytic efficacy throughout sequential administration.21
In this study, we examined VSV-GP’s oncolytic efficacy in seven established human NC cell lines by employing a VSV-GP recombinant carrying a green fluorescent protein (GFP) transgene (VSV-GP-GFP). The efficacy of VSV-GP-GFP-mediated oncolysis was investigated in all seven NC cell lines using sulforhodamine B cell (SRB) viability assay and continuous xCELLigence real-time measurement of cell proliferation. We then assessed the viral replication cycle in NC cells by using a TCID_50_ (50% tissue culture infectious dose) assay and performing sequential RT-qPCR analysis. Transmission electron microscopy (TEM) was implemented to typify morphological alterations in NC cells in response to VSV-GP-GFP infection via immunofluorescence detection of stress and apoptosis markers. Additionally, we identified markers of cell death and cellular resistance to viral infection with fluorescence-activated cell sorting (FACS) analysis. In search of potential antiviral resistance mechanisms, we examined expression levels of α-dystroglycan (a major viral entry receptor for VSV-GP) and analyzed the cellular ability to produce and respond to the antiviral cytokine IFN-β in the context of an infection with VSV-GP-GFP. Finally, in a mechanistic approach, we studied the impact of VSV-GP-GFP infection on expression of the tumor-driving NUT fusion protein.
Results
Most NC cell lines are permissive to VSV-GP-GFP-mediated cytotoxicity
For initial screening of permissivity, all seven NC cell lines were infected with VSV-GP-GFP at an MOI of 0.01 prior to SRB viability analysis performed at 72 h post-infection (hpi) (Figure 1A). In analogy to our previous research, we classified the permissivity of NC cell lines to VSV-GP-GFP depending on the remnant tumor cell mass after infection with an MOI of 0.01 analyzed at 72 hpi: if initial tumor cell mass was reduced by > 50%, NC cell lines were defined as permissive to VSV-GP-GFP-mediated oncolysis. NC cell lines with a remnant tumor cell mass of 50%–75% were deemed semi-permissive; those mustering a remnant cell mass >75% were defined as resistant.27^,^28 Based on these results, five out of six NC cell lines exhibiting the BRD4::NUTM1 fusion (143100, 14169 JCM, 10-15, 690100) were found to be permissive to VSV-GP-GFP-mediated oncolysis, while only NC cell line HCC2429 was considered resistant (Figure 1A). The NC cell line SNU-3178S, which harbors a BRD3::NUTM1 fusion gene, was also found to be resistant to VSV-GP-GFP-mediated oncolysis. All NC cell lines expressed the NUT fusion protein (NUT::BRD4 [200–240 kDa] in cell lines 10-15, 14169, HCC2429, 143100, JCM, and 690100; NUT::BRD3 [160 kDa] in cell line SNU-3178S) (Figures S1A and S1B), while the human colon cancer control cell line HT-29 did not. Even though number and size of the bands differed between cell lines, neither the fusion partner, strength, nor pattern of expression correlated with cell line permissivity to VSV-GP-GFP. Fluorescence microscopy was conducted to detect the expression of the VSV-GP-GFP-encoded transgene GFP, which was linked to oncolytic efficacy in the corresponding NC cell lines (Figure 1B). In permissive cell lines, treatment with VSV-GP-GFP led to strong expression of the GFP marker gene. It was most pronounced in the highly permissive NC cell line 143100, where it was associated with cell viability: high expression of GFP was found at 48 hpi but appeared to drop significantly at 72 hpi, in association with a reduction of cell confluency (Figure 1B, top row). In NC cell line 690100, signs of an effective infection, such as GFP expression and tumor cell rounding—a process that has been described for VSV infection previously29—could be seen, as the cytopathic effects of the virus were first observed at a later time point (72 hpi; Figure 1B, middle row). In contrast, GFP expression could not be detected at any time in the resistant NC cell line SNU-3178S (Figure 1B, bottom row).Figure 1. Permissivity of NC cell lines to VSV-GP-GFP(A) Viability of seven NC cell lines after infection with VSV-GP-GFP (MOI 0.01) at 72 hpi. Dotted lines represent cutoff permissivity values for each group (remnant tumor cell masses at 72 hpi >75%, >50%, and ≤50%, respectively). Values show the mean ± SD of at least two independent SRB viability assays. (B) GFP expression and cell density in NC cell lines 143100, 690100, and SNU-3178S after infection with VSV-GP-GFP (MOI 0.0005) as documented by fluorescent light and transmitted light microscopy at 48 and 72 hpi. (C–H) NC cell lines 143100, 690100, and SNU-3178S were infected with VSV-GP-GFP (MOI 0.00005–0.01 for SRB; 0.0001–0.1 for xCELLigence) and analyzed at 48 and 72 hpi for SRB (left, C–E), as well as continuously in 30-min intervals over 96 h for xCELLigence (right, F–H). Values show the mean ± SD of at least two independent experiments for SRB, as well as four biological replicates for xCELLigence. ∗p < 0.05 (C,D).
For NC cell lines 143100, 690100, and SNU-3178S, results of detailed MOI testing at 48 and 72 hpi are shown in Figures 1C–1E. VSV-GP-GFP-mediated cytotoxicity was found to be time- and dose-dependent in all permissive NC cell lines (Figures 1C, 1D, and S2A–S2C, left panels). In contrast, the same MOIs (0.00005–0.01) had no relevant effect in the resistant NC cell lines SNU-3178S and HCC2429 (Figures 1E and S2D, left panel). However, in the NC cell line SNU-3178S MOIs as high as 10 viral particles per tumor cell (MOI 10) did show a relevant cytopathic effect, thereby indicating that resistance may be a relative phenomenon (Figure S3D). In contrast, MOIs as low as 0.00005 or 0.0001 potentiated the risk of an inoculum not containing any viral particles at all, which—especially in highly permissive NC cell lines—led to either their complete eradication or no oncolytic effect on the cells. Error bars were thus large under these conditions. Nonetheless, in cell line 143100, MOI 0.0001 led to a significant cytotoxic effect at 48 hpi (p = 0.029) and MOI 0.00005 at 72 hpi (p = 0.020) (Figure 1C).
To demonstrate the fast replication of VSV-GP-GFP, cell density was continuously monitored (30 min intervals over 96 h) via xCELLigence real-time cell analysis (Figures 1F–1H). A decrease of cell index curves was seen as early as 12 hpi (36 h) for the highly permissive NC cell line 143100 (Figure 1F). In cell lines 143100 and 690100, higher MOIs led to earlier decreases in the viability curves. However, dynamics in NC cell line 690100 were found to be distinct from those in NC cell line 143100: while NC cell line 143100 responded rapidly to infection with VSV-GP-GFP even at low MOIs, a drop of the cell index in NC cell line 690100 occurred only at a later time point (52–72 hpi) and at higher MOIs (0.01–0.1). These drops of the cell index curves further demonstrated not only antiproliferative but also cytotoxic effects of VSV-GP-GFP (Figures 1F and 1G). MOIs lower than 0.01 did not have any significant effect on cell index curves in NC cell line 690100 (Figure 1G), which confirms the results of the SRB viability data (p < 0.05 for MOCK vs. MOI 0.01 48 hpi and MOI 0.001 72 hpi) (Figure 1D). Infection with VSV-GP-GFP did not affect cell index curves in the resistant NC cell line SNU-3178S (Figure 1H).
To evaluate whether resistance phenomena were caused by impaired virus replication, viral RNA copy numbers and viral titers were measured every 24 h with RT-qPCR (Figure 2A) and TCID_50_ assays (Figures 2B–2D) after infection with VSV-GP-GFP (MOI 0.01 for RNA copies; MOI 0.0001 and MOI 0.01 for viral titers). In NC cell lines 143100 and 690100, RT-qPCR showed a steep increase in RNA copy numbers over time after infection, with values peaking around 10^10^ RNA copies/mL at 72 hpi. Notably, RNA copy numbers also rose in NC cell line SNU-3178S (though significantly less—p = 0.0369 and p = 0.0366 for cell lines 143100 and 690100 vs. SNU-3178S respectively—and attaining peak levels of only 10^8^ at 72 hpi) (Figure 2A) even though previous experiments had failed to demonstrate virus-mediated cytotoxicity or transgene expression in this NC cell line (Figures 1B–1E and 1H).Figure 2RNA copy levels and viral titers correlate with NC cell line permissivityNC cell lines 143100, 690100, and SNU-3178S were infected with VSV-GP-GFP (MOI 0.01 for qPCR; MOI 0.0001 and 0.01 for TCID_50_). RNA copy numbers and viral titers were measured at 1, 24, 48, and 72 hpi with RT-qPCR (A) and in a TCID_50_ assay (B–D), respectively. Dotted lines in (B–D) represent a threshold titer of 10^7^ TCID_50_/mL. Values show the mean ± SD. ns: not significant, ∗p < 0.05 (A).
Similarly, in NC cell line 143100, viral titers (as determined per TCID_50_ assay) rose to 10^10^ TCID_50_/mL at 48 hpi at an MOI as low as 0.0001. A plateau could be seen for the last 24 h of observation (Figure 2B). The increase of viral titers was less pronounced in NC cell line 690100; titers in proximity of those displayed in NC cell line 143100 were only reached with a higher MOI of 0.01 at 72 hpi (Figure 2C). In NC cell line SNU-3178S, kinetics were found to depend on the MOI employed: while a minimal increase of viral titers within the first 24 hpi was followed by a return to initial titers at MOI 0.0001, titers increased and peaked at about 10^6^ TCID_50_/mL at MOI 0.01. Nonetheless, while VSV-GP-GFP exceeded titers of 10^7^ TCID_50_/mL within 48 hpi at MOI 0.0001 in all five permissive NC cell lines (Figures 2B, 2C, and S2A–S2C, right panels), titers in the resistant NC cell lines were never found to exceed 10^7^ TCID_50_/mL, even when employing the higher MOI of 0.01 (Figures 2D and S2D, right panel).
Visualization of VSV-GP-GFP replication in NC cells
To characterize the replication cycle of VSV-GP-GFP in NC cells, TEM images of the permissive NC cell line JCM were taken at different points of time in the viral replication cycle (Figure 3A). TEM images depict the bullet-shaped, 150–200 nm long, enveloped RNA virus docked to the tumor cell surface at 48 hpi (Figure 3A, upper right). During virus replication (at 48 hpi), secretory vesicles containing viral progeny were observed in close proximity to the cell membrane (Figure 3A, middle right, red arrows). In addition, TEM showed the release of viral particles from an oncolysed NC tumor cell infecting a neighboring NC tumor cell at 72 hpi (Figure 3A, bottom, red arrows), thereby visualizing viral spread throughout the tumor. The infected NC cell demonstrated plasma membrane disassembly (orange arrows, bottom left panel) as well as nuclear chromatin decondensation (green arrows, bottom right panel).Figure 3. Viral replication cycle, entry receptor expression, and stress granule formation in NC cell lines(A) VSV-GP-GFP infects NC cell line JCM. Top left: uninfected tumor cell (scale bars: 1,000 nm). Top right: VSV-GP-GFP particle close to the tumor cell surface at 48 hpi (arrow; scale bars: 100 nm). Middle row: secretory vesicles containing viral progeny (arrows; scale bars, left: 1000 nm; scale bars, right: 100 nm). Right side shows magnification of the marked section. Bottom row: viral particles (red arrows), released from an oncolysed NC cell exhibiting membrane disassembly (orange arrows) at 72 hpi, infect a neighboring NC cell, causing chromatin decondensation (green arrows) (scale bars, left: 1,000 nm; scale bars, right: 500 nm). (B) Expression of the VSV-GP-GFP entry receptor α-dystroglycan (α-DG, purple) and the nuclear stain DAPI (blue) in uninfected NC cell lines 143100 and SNU-3178S in immunofluorescence (C) NC cell line JCM shows stress granule formation after infection with VSV-GP-GFP. Top row: a cytoplasmic stress granule (TEM; scale bars: 1,000 nm). Right side shows magnification of the marked middle row: MOCK control. Cells are positive for the nuclear stain DAPI (white) but negative for GFP, TIA-1 (blue), and phospho-eIF2⍺ (red) (scale bars, left: 50 μm; scale bars, right: 10 μm). Bottom row: after infection with VSV-GP-GFP (72 hpi), cell line JCM expresses GFP (green) as well as TIA-1 (blue) and phospho-eIF2⍺ (red) (scale bars, left: 50 μm; scale bars, right: 10 μm).
Immunofluorescence (IF) staining for α-dystroglycan was conducted to assess whether a potentially reduced expression of the viral entry receptor of VSV-GP could be responsible for the resistance observed in individual NC cell lines. However, equal expression patterns of α-dystroglycan were observed in both the highly permissive NC cell line 143100 and the resistant NC cell line SNU-3178S (Figure 3B).
VSV-GP-GFP infection of cell line JCM frequently resulted in the formation of cytoplasmic stress granule-like structures (Figure 3C, top row). To confirm our observations in TEM, we employed IF for two canonical markers of stress granules, TIA-1 (T cell intracellular antigen 1) and p-eIF2ɑ (phosphorylated eukaryotic initiation factor 2). After infection with VSV-GP-GFP, JCM cells expressed the viral transgene product GFP (green), TIA-1 (blue), and p-eIF2ɑ (red) (Figure 3C, bottom row)—all of which were absent in the MOCK control (Figure 3C, middle row). TIA-1 and p-eIF2ɑ were frequently found in apoptotic cells displaying nuclear fragmentation (as depicted by a nuclear DAPI stain in white) or cell detritus. Both markers rarely colocalized with the stain for VSV-GP-expressed GFP.
Permissive NC cells express VSV-N and undergo apoptosis upon viral infection
To further investigate resistance mechanisms, the question arose whether NC cell lines displayed differences regarding the translation of viral genes depending on their localization within the viral genome. We thus analyzed the expression of the marker protein GFP (gene located in proximity of the 5′ genome end) and the viral protein VSV-N (gene located at the 3′ genome end of the viral genome) after infection with VSV-GP-GFP at 48 hpi via FACS. Cells were differentiated by their viability with the cell death staining dye Zombie; only viable cells were investigated for expression of GFP and VSV-N (Figure 4A). As a result, FACS analysis showed the permissive NC cell lines 143100 and 690100 expressed the marker protein GFP and the viral protein VSV-N at 48 (143100) and 72 (690100) hpi. Both NC cell lines demonstrated similarly high expression levels of GFP and VSV-N (30%–40%), with the majority of cells being double-positive for both proteins. On the contrary, GFP and VSV-N were only remotely detectable in the resistant NC cell line SNU-3178S at 72 hpi. More Zombie-positive cells were counted after viral infection than after MOCK treatment in the permissive NC cell lines 143100 and 690100, whereas in the NC cell line SNU-3178S, a higher overall rate of Zombie-positive cells was found irrespective of MOCK or viral pretreatment (Figure 4A).Figure 4. Viral infection leads to VSV-N expression and apoptotic cell death(A) FACS analysis of GFP, VSV-N, and the live/dead stain Zombie Aqua in NC cell lines 143100, 690100, and SNU-3178S after infection with VSV-GP-GFP (MOI 0.01, 0.1, and 1, respectively, 48 hpi). One representative result of two independent experiments is shown. Values show the mean ± SD (B, C). FACS expression of apoptosis marker Annexin V and GFP over time (48 and 72 hpi) and with increasing MOIs in cell lines 690100 and HCC2429. ns: not significant, ∗p < 0.05 (B, C). For Annexin V, only PI negative events are shown in % of total cell population (gates shown in Figure S5).
As infection with VSV-GP-GFP led to a larger population of dead cells in permissive NC cell lines, we conducted FACS analysis for apoptosis and necrosis in the permissive NC cell line 690100 and the resistant NC cell line HCC2429 (Figure 4). Viral load was quantified via GFP expression, which initially increased significantly over time and with higher MOIs in NC cell line 690100 (Figure 4B, top panel; p = 0.0434 for MOCK vs. MOI 0.01 at 72 hpi). Comparatively, GFP expression was reduced 72 hpi with the highest MOI of 0.1 (p = 0.0652 for MOCK vs. MOI 0.1 at 72 hpi). This is most likely related to the overall reduction in cell viability as previously observed in fluorescence and transmitted light microscopy (Figure 1B). A similar difference was observed for the early apoptosis marker Annexin V (Figure 4B, lower panel; p = 0.0229 for MOCK vs. MOI 0.1 at 72 hpi). In the resistant NC cell line HCC2429, GFP expression was only significantly detectable at 72 hpi with MOI 1 (Figure 4C, top panel; p = 0.0186 for MOCK vs. MOI 1 at 72 hpi), which correlated with a small, albeit insignificant increase of Annexin-V-positive, apoptotic cells in this condition (Figure 4C, bottom panel; p = 0.0825 for MOCK vs. MOI 1 at 72 hpi). The scatterplots (Figure S5) showed a clear separation of Annexin-V-positive and propidium iodide (PI)-negative events at 48 hpi. FACS data suggest that apoptosis is the main mode of cell death. This finding was further supported by IF staining for the apoptosis marker caspase-3, which was found to be positive in cell line JCM after infection with VSV-GP-GFP (Figure S4).
IFN-β effectively abrogates VSV-GP-GFP-mediated cytotoxicity in NC cells
To investigate whether active intracellular interferon (IFN) signaling contributed to cellular resistance to VSV-GP-GFP, we analyzed the influence of pretreatment with IFN-β and the JAK1/2-STAT inhibitor ruxolitinib on VSV-GP-GFPs’ oncolytic efficacy.
ELISA quantification of intrinsic IFN-β production at 72 hpi with VSV-GP-GFP showed no detectable IFN-β levels in NC cell line 143100—even with high MOIs (Figure 5A). Nonetheless, NC cell line 143100 was found to be capable of IFN-β production as demonstrated by a positive control with an oncolytic measles vaccine virus (MeV), where IFN-β values at 72 hpi reached levels as high as 400 pg/mL (with MOI 0.1) and 700 pg/mL (with MOI 1) (Figure 5A, right side).Figure 5IFN-β pretreatment effectively inhibits VSV-GP-GFP-mediated cytotoxicity in NC cells(A) IFN-β production in NC cell line 143100 after infection with VSV-GP-GFP (left) and measles virus (MeV) at 72 hpi (right). Values show the mean ± SD of technical triplicates (MOIs 0.0001–1) measured by ELISA. (B) Cell density of NC cell line 690100 after pretreatment with IFN-β or ruxolitinib 24 h before infection with VSV-GP-GFP. Cell viability was continuously monitored via xCELLigence in 30 min intervals over 120 h. Values show the mean ± SD of 4–8 biological replicates. (C) Cell viability after IFN-β pretreatment and infection with VSV-GP-GFP (MOI 0.01 for cell line 143100 and MOI 10 for cell lines 690100 and SNU-3178S). Viability was analyzed at 72 hpi using an SRB assay. Values show the mean ± SD of three independent experiments. (D) Cell viability after pretreatment with JAK inhibitor ruxolitinib. NC cell lines were pretreated with IFN-β or ruxolitinib, infected with VSV-GP-GFP (MOI 0.00005 for cell line 143100 and 0.01 for cell lines 690100 and SNU-3178S), and analyzed at 72 hpi using an SRB assay. Values show the mean ± SD of three independent experiments. ns: not significant, ∗∗∗∗p < 0.0001 (D).
Continuous monitoring of NC cell viability after infection with VSV-GP-GFP via xCELLigence analysis revealed the effects of IFN-β and ruxolitinib treatment on efficacy of virus-mediated oncolysis (Figure 5B). In untreated cells, infection with VSV-GP-GFP led to a decrease in cell viability beginning between 60 and 72 h and a complete loss of cell viability after 108 h (Figure 5B, gray curve). Treatment with IFN-β (Figure 5B, light blue curve) and ruxolitinib (Figure 5B, pink curve) alone without viral infection induced a flattening of the growth curve (with especially IFN-β exhibiting antiproliferative but no cytotoxic effects). Interestingly, treatment with IFN-β fully abrogated VSV-GP-GFP-mediated cytotoxicity (Figure 5B, dark blue curve). However, combination of VSV-GP-GFP infection with ruxolitinib treatment did not alter VSV-GP-GFP-mediated cytotoxicity (Figure 5B, purple curve).
Similarly, an SRB assay, which compared the viability of NC cell lines 143100, 690100, and SNU-3178S pretreated with IFN-β before infection with VSV-GP-GFP and analyzed at 72 hpi, confirmed the complete reversal of VSV-GP-GFP-mediated cytotoxicity by IFN-β even when high MOIs were used (0.01 for the highly susceptible NC cell line 143100—an MOI that led to a cell viability reduction below 10% in dose-dependent efficacy testing (Figure 1C)—and 10 for NC cell lines 690100 and SNU-3178S) (Figure 5C). In contrast, SRB analysis revealed that pretreatment with JAK1/2-STAT inhibitor ruxolitinib did not have any effect on tumor cell density after infection with VSV-GP-GFP (Figure 5D; p = 0.0778 and p = 0.3859 for monotherapy with VSV-GP-GFP versus infection with VSV-GP-GFP after pretreatment with ruxolitinib for cell lines 143100 and 690100, respectively). As such, ruxolitinib was not able to reverse resistance to VSV-GP-GFP in the resistant NC cell line SNU-3178S; even though cell viability was significantly reduced when comparing monotherapy with VSV-GP-GFP and infection with VSV-GP-GFP after pretreatment with ruxolitinib (p < 0.0001), this reduction can be attributed to the cytotoxic effect of ruxolitinib itself (p = 0.7578 for monotherapy with ruxolitinib versus infection with VSV-GP-GFP after pretreatment with ruxolitinib) (Figure 5D).
NC cells fail to activate IFN-β signaling in response to infection with VSV-GP-GFP
To analyze the capability of NC cell lines 143100, 690100, and SNU-3178S to activate the IFN pathway, we performed a western blot analysis of IFN-β downstream proteins STAT1 and P-STAT1 (Figure 6).Figure 6NC cell lines fail to activate IFN-β signaling after infection with VSV-GP-GFPNC cell lines were pretreated for 24 h with IFN-β or ruxolitinib or left untreated before infection with VSV-GP-GFP (MOI 0.1). STAT1 and P-STAT expression were assessed at 24 hpi using immunoblotting with Vinculin as a loading control. For P-STAT1-positive controls (right) were treated for 1 h with IFN-β.
Positive quantification of STAT1 and P-STAT1 after stimulation with IFN-β indicated a functioning IFN signaling system in all three NC cell lines, when P-STAT1 was detected at 1 h and STAT1 at 48 h after a short 1-h pretreatment with IFN-β (Figures 6 and S6). Additional infection with VSV-GP-GFP after pretreatment with IFN-β did not alter STAT1 expression when compared to mono-treatment with IFN-β.
Without IFN-β pretreatment, neither STAT1 nor P-STAT1 could be detected at significant levels at any point of time after infection, indicating that NC cells were incapable of activating IFN downstream signaling in response to infection with VSV-GP-GFP (Figures 6 and S7).
Immunofluorescence staining of G3BP1 (Ras-Gap binding protein 1), a stress granule initiator and mediator of the cellular IFN response to viral infections, showed only a moderate increase of G3BP1 expression in cells infected with VSV-GP-GFP in comparison to a MOCK control (Figure S4), which supports the hypothesis that the antiviral IFN response is impaired in NC cell lines.
VSV-GP-GFP influences NUT fusion protein expression in NC cell lines
Finally, the NUT protein was quantified with IF and western blotting to explore the influence of infection with VSV-GP-GFP on the expression of the tumor-driving NUT fusion protein (Figure 7). IF imaging of NUT in the permissive NC cell line 143100 (BRD4::NUTM1 fusion) showed a speckled nuclear stain, the typical unique staining pattern for NC cells (Figure 7A, top row). After viral infection, the NUT fusion protein was still detectable, but the nuclear staining pattern was found to be blurred instead (Figure 7A, middle row). NUT fusion proteins and wtNUT could not be detected in the human colon carcinoma cell line HT-29, which served as a negative control (Figure 7A, bottom row; Figures S1A and S6D). Western blot analysis of the NUT fusion protein shows bands corresponding to different sizes of the NUT fusion protein in the cell lines (Figures 7B, S1A, and S6D). NC cell lines 143100, 690100, and HCC2429 exhibit distinct breakpoints of the BRD4::NUTM1 fusion gene, which resulted in differently sized fusion proteins (237 kDa, 215 kDa, and 200 kDa, respectively). The detection of multiple bands in the NC cell lines 143100, 690100,s and HCC2429 can be attributed to the expression of different BRD4-splice isoforms within the fusion protein (Figures 7 and S1A). NC cell line SNU-3178S, which harbors the BRD3::NUTM1 fusion gene, expressed a markedly smaller fusion protein due to the minor size of BRD3 (157 kDa). The fusion proteins in NC cell lines HCC2429 and SNU-3178S generated stronger bands. In the resistant NC cell lines HCC2429 and SNU-3178S, fusion protein expression was not found to be affected by viral infection. In contrast, the fusion protein expression was weakened in the permissive NC cell lines 143100 and 690100 after infection with VSV-GP-GFP when compared to MOCK (Figures 7B and S6D). The NUT-fusion protein partner BRD4 was only detected in NC cell line SNU-3178S, as BRD3 serves as a fusion protein partner in this cell line instead of BRD4, and the epitope site for antibody-binding is blocked when BRD4 is fused to NUT (Figure S8). Interestingly, BRD4 expression in SNU-3178S was affected negatively by infection with VSV-GP-GFP.Figure 7. Effect of VSV-GP-GFP on NUT fusion protein expression(A) NC cell line 143100 was infected with VSV-GP-GFP (MOI 0.0005) and stained with an NUT antibody and DAPI at 48 hpi. Human colon carcinoma cell line HT-29 served as a negative control. (B) Western blot of NUT fusion proteins in NC cell lines before and after infection with VSV-GP-GFP. HT-29 served as a negative control. Samples were collected at 48 hpi with MOIs, resulting in a cell mass reduction of ∼50%. (C) Viability of NC cell lines 143100, 690100, and SNU-3178S after monotherapy with iBET BI894999 versus combination treatment of VSV-GP-GFP after a pretreatment with BI89499 for 24 h. Analysis was conducted 96 and 72 hpi with an SRB assay. Values show the mean ± SD of 8–12 biological replicates. ns: not significant, ∗p < 0.05, ∗∗∗p < 0.001, ∗∗∗∗p < 0.0001 (C).
To determine whether decreased NUT-fusion-protein expression after infection with VSV-GP-GFP was due to a direct interaction of VSV-GP-GFP with the fusion protein, we measured cell viability after treatment with the BET-protein inhibitor (iBET) BI894999 and compared it to cell viability after combination treatment with both BI894999 and VSV-GP-GFP (Figure 7C). Analysis of cell viability in an SRB assay showed infection with VSV-GP-GFP always exhibited a significant additional cytotoxic effect in combination with iBET versus iBET alone in the 1 nM concentration (p = 0.0002, p = 0.0379, and p < 0.0001 for cell lines 143100, 690100, and SNU-3178S, respectively). In the 2.5 nM concentration, additional effects of VSV-GP-GFP infection were significant in cell lines 143100 and SNU-3178S (p = 0.0002 and p < 0.0001, respectively) but not in cell line 690100 (p = 0.2345).
Discussion
As a rare, yet aggressive disease, NC therapy faces two challenges: a lack of awareness for the entity on the one hand and a shortness of effective therapeutic strategies due to the often advanced disease stages at presentation and rapid progression thereafter on the other hand. Oncolytic virotherapy with VSV-GP could help overcome some of these challenges due to the tumor-specific, self-replicating, and self-enhancing characteristics of this treatment approach.
This is the first occasion that VSV-GP-GFP has been tested as an OV candidate in an in vitro model of NC. The overall efficacy of VSV-GP-GFP was found to be high in most NC cell lines, even when employing very low MOIs. The strong cytotoxic efficacy and quick replication kinetics observed in the majority of NC cell lines are characteristic for VSV-GP.30 Cytotoxic efficacy of VSV-GP-GFP in NC was comparable to observations made by Muik et al. in a study testing VSV-GP in the NCI-60 panel, where a reduction of tumor cell survival to 50% and lower was achieved at MOIs ranging from 0.0001–10.21 Even so, VSV-GP-GFP seems to possess especially high lytic qualities in NC cell lines as opposed to other cancer entities such as prostate and ovarian cancer. Studies employing VSV-GP in prostate30 and ovarian cancer22 observed peak viral titers ranging from 10^5^ to 10^8^ TCID_50_/mL at either 24 or 48 hpi when applying an MOI of 0.1. In NC cell lines, we used the much lower MOI of 0.0001 and still achieved titers as high as 10^9^ TCID_50_/mL (Figures 2B, 2C, and S1A–S1C). NC cell survival rates at 72 h after treatment with an MOI of 0.01 were comparable to those observed in the entities mentioned above when employing at least 10-fold higher MOIs. In addition, resistance to VSV-GP-GFP is a relative phenomenon in NC cell lines, as it can be overcome with higher MOIs (Figure S2). When comparing the oncolytic efficacy of VSV-GP-GFP with the EMA- and FDA-approved OV T-VEC (a *Herpes-simplex-*virus) in NC cell lines, both viruses attain similar cell killing rates, although VSV-GP reaches its full oncolytic potential at an earlier point of time due to faster replication kinetics. VSV-GP-GFP titers are higher than T-VEC titers in NC.18^,^20
Interestingly, analysis of RNA copy levels and viral titers revealed that VSV-GP-GFP is able to sustain replication in the resistant cell lines SNU-3178S and HCC2429 but requires higher MOIs, remains at a lower level, and fails to affect cell viability. These observations suggest resistance to VSV-GP-GFP-mediated oncolysis is likely allocated to the junction between viral gene transcription and translation. Fluorescence microscopy and FACS analysis of resistant SNU-3178S cells (Figures 1B and 4A) demonstrated that neither the transgene product GFP nor the viral nucleocapsid protein VSV-N was detectable in significant amounts, which implies that although viral RNA is replicated, it is not translated into protein. Similar observations have been made in benign HPDE cells (human pancreatic duct epithelial cells) where the infection of the resistant cell line with VSV resulted in the synthesis of viral mRNA levels similar to those seen in permissive cells, whereas viral protein synthesis was significantly decreased.31 The localization of genes of interest within the VSV-GP-GFP genome was not found to influence the corresponding expression rate: VSV-N, allocated at the 3′ end of the genome, is expressed just as scantly as GFP, positioned in proximity to the 5′ end.21 This indicates translational repression of viral mRNA is generalized rather than selective in NC cell line SNU-3178S.
Moreover, western blot analysis of fusion protein and BRD4 expression revealed NUT fusion protein expression was decreased after infection with VSV-GP-GFP in cell line 143100 (Figures 7B and S6D), and BRD4 expression was decreased in cell line SNU-3178S (Figure S8). Previous research has shown that infection with VSV-GP causes host cell translational shut-down, while viral replication continues unhindered, or is suppressed at a later point of time.32^,^33 Stress granules, as observed in TEM and IF in our study (Figures 3C and S4), could play a role as effectors of host cell translational repression. Stress granules are membrane-less compartments formed as a result of liquid-liquid phase separation crucial to the cellular stress response.34 As a dynamic, transient compartment originating from the accumulation of stalled pre-initiation complexes, stress granules serve a multitude of functions associated with RNA control.35^,^36 Infection with VSV frequently leads to the formation of stress granules,32^,^37 as a number of kinases responsible for the phosphorylation of the translation initiation factor eIF2ɑ—a core prerequisite for stress granule formation38—sense and respond to the presence of viral RNA. These stress-response kinases include PKR,39^,^40 PERK,41 and GCN2.42 We here show that VSV-GP-GFP-infected NC cells exhibit stress granule formation and stain positively in IF for canonical stress granule markers such as p-eIF2ɑ and TIA-1. These stress granule markers often do not co-localize with virus-mediated expression of the GFP marker protein and are expressed in cells displaying nuclear fragmentation as a sign of apoptosis. Our results most likely represent the distinct kinetics of viral infection, in which GFP expression occurs first, followed by the reaction of the host cell with formation of stress granules. Stress granule formation may either occur as a direct answer to viral infection or due to specific effects of infection with VSV-GP-GFP on host cell function. For example, the viral matrix protein M functions as a suppressor of host gene expression through interruption of transcription, nuclear-cytoplasmic transport, translation,43^,^44 and mitosis.45 It generates an accumulation of stalled host pre-initiation complexes, a powerful inductor of stress granule formation. Therefore, reduction of protein expression after viral infection as observed in cell lines 143100 and SNU-3178S (Figures 7B and S8) are most likely a result of global host cell translational repression. However, in cell line SNU-3178S, expression of BRD4 is reduced (Figure S8), while expression of the NUT fusion protein remains stable after infection (Figure 7B). This is likely due to distinct translational control mechanisms: endogenous BRD4 is subject to cellular stress responses and antiviral signaling pathways, while the NUT fusion protein is constitutively expressed and dissociated from normal transcriptional and translational activity regulations.46^,^47^,^48 While this shows that infection with VSV-GP-GFP has an effect on host cell functions in the resistant cell line SNU-3178S, viral protein translation is also suppressed, and infection fails to overwhelm the host cell. Previous studies suggest a kinetic model of translational repression in which cell lines resistant to infection with VSV proceeded to shutdown of viral protein translation after initial host cell translational repression.33^,^44 Suppression of viral translation also leads to reduced expression of viral proteins necessary for secondary transcription, one of which is VSV-N.49^,^50 This explains why RNA copy numbers and viral titers in resistant cell lines remain below those of permissive cell lines.
Additionally, it should be stressed that the reduced expression of the NUT fusion protein after infection with VSV-GP-GFP in cell line 143100 is not due to a direct interaction of the virus with the fusion protein but much rather a bystander effect of large-scale host cell translational shutdown after virus infection. As shown in the viability assessment of NC cells after treatment with iBET BI894999 before and after infection with VSV-GP-GFP, iBET treatment did not mitigate the oncolytic effects of VSV-GP-GFP (Figure 7C). iBETs disrupt the interaction of BET proteins such as BRD4 with chromatin, hampering the tumor-driving effect of the NUT fusion protein.51 Previous studies showed iBET pretreatment did not negatively affect the oncolytic qualities of OV T-VEC.18^,^20 If the reduction of cell viability seen after infection with VSV-GP-GFP in most NC cell lines depended on a direct interaction of the virus with the NUT fusion protein, pretreatment with the iBET would decrease or abrogate VSV-mediated cytotoxicity. Instead, infection with VSV-GP-GFP always led to an additional reduction of cell viability (Figure 7C). Nonetheless, the reduced expression of the NUT fusion protein observed in cell line 143100 could lead to an additional apoptotic effect due to the loss of the oncogenic cell-survival function of the fusion protein. Apoptosis and cell differentiation have previously been observed after fusion protein knockdown.52
Also, the NUTM1 fusion partner does not seem to play a role in resistance to VSV-GP-GFP: resistant SNU-3178S cells harbor the BRD3::NUTM1 fusion, while resistant HCC2429 cells harbor a BRD4::NUTM1 fusion. Western blot analysis of the NUT fusion protein detects multiple bands in the permissive NC cell lines 143100, 690100, and HCC2429. These can be attributed to both the expression of different BRD4-splice isoforms within the fusion protein (larger protein, upper band, 237 and 215 kDa) as well as an expression of wtNUT (130 kDa, mostly in cell lines 143100 and 690100). Both observations were also found in a previous study, although ectopic expression of wtNUT seems to be a rare phenomenon even in NC cell lines.47
Our exploration of resistance mechanisms ruled out a number of targets, such as a differential expression of the viral entry receptor α-dystroglycan on permissive and resistant cell lines. Previous examinations of the effect of α-dystroglycan expression on VSV-GP efficacy in ovarian cancer cells rendered similar results: expression rates of α-dystroglycan did not necessarily correlate with oncolytic efficacy in the corresponding tumor cell line.22 However, VSV-GP-GFP was pseudotyped with an LCMV strain that has a relatively low affinity to α-dystroglycan.53 As such, other receptors such as heparan sulfates may play an equal role in viral entry for this pseudotype.54
Similarly, an antiviral IFN-β response to VSV-GP infection was found to be absent. As VSV-GP-GFP is highly sensitive to IFN-β, a cellular IFN response to infection results in strongly decreased oncolytic efficacy.17^,^22 Previous investigations showed that NC cells (143100 and 690100) demonstrated an innate IFN-β response to measles vaccine virus, and all cell lines have functional intracellular IFN signaling pathways. This suggests that VSV-GP-GFP proficiently circumvents the antiviral IFN response. Possibly, this is mediated through viral components such as the matrix protein M: the transcriptional and translational shutdown mediated by protein M has been shown to include genes responsible for the expression of IFN.43^,^55 IF staining for G3BP1, which is capable of IFN stimulation through RIG-I activation during the cellular stress response,56 observed only a slight increase of expression when comparing MOCK control and VSV-GP-GFP infection. Since STAT activation does not play a key role in cellular resistance to VSV-GP-GFP in NC, ruxolitinib is not eligible as a potential therapeutic agent to overcome resistance. However, due to preserved intracellular IFN signaling pathways, IFN-β serves as an antiviral agent to embank uncontrolled infection during tumor therapy.
While the lack of an IFN response plays out positively for the oncolytic efficacy in NC cells, the influence of an innate immune response to infection with VSV-GP-GFP remains to be investigated in vivo. IFN response of immune cells in the tumor microenvironment is desirable for an increased immune stimulation and lymphocyte migration to the tumor site.57^,^58 The past 2 years have seen the development of two genetically engineered, immunocompetent NC mouse models (GEMMs). Both models were able to reproduce key characteristics of human NC in the corresponding GEMM, such as morphological heterogeneity, high oncogenic potential of the BRD4::NUTM1 fusion gene across various tissues, and rapid metastasis.59^,^60 In the context of virotherapy, immunocompetent GEMMs could offer an opportunity to explore virus-host interactions. However, murine-to-human translation of observations made in GEMMs might be curbed by altered viral tropism.
While VSV-GP-GFP’s efficacy in NC remains to be determined in vivo, a first case report employing the OV T-VEC in an NC patient yielded promising results in disease stabilization and demonstrated strong viral replication.19 Since NC is characterized by a low mutational burden, OV therapy might not only help diminish tumor mass but also increase anti-tumor immune activation. Among a multitude of current clinical trials investigating different recombinant versions of VSV in malignancies (e.g., NCT05846516 and NCT03647163), some include multimodal treatment approaches such as a combination of VSV-GP treatment with the immune checkpoint inhibitor ezabenlimab (NCT05155332).61 In possible future clinical use, patients might benefit from a pre-therapeutic ex vivo resistance screening in the form of a virogram62 in order to select, in the context of a companion diagnostic, the most efficient viral construct for each individual patient. As our data demonstrate, VSV-GP-GFP exhibits high efficacy under in vitro conditions and hold significant potential for future application.
Materials and methods
Cell lines
Seven human NC cell lines were employed in this study. Cell lines 143100 and HCC2429 were kindly provided by Dr. Xin Zhang from University Hospital Essen, Germany; NC cell lines 14169, 10-15, and JCM by Dr. Christopher French, Boston, MA, USA; NC cell line 690100 by Prof. Jens Siveke, University Hospital Essen, Germany; and NC cell line SNU-3178S was purchased from the Korean Cell Line Bank, Seoul National University College of Medicine. SNU-3178S carries the BRD3::NUTM1 gene, while the other six cell lines harbor the BRD4::NUTM1 fusion gene. The human colon carcinoma cell line HT-29 (NCI-DCTC, National Cancer Institute – Division of Cancer Treatment of Diagnosis) served as a non-NC control. BHK 21, used as an indicator cell line for VSV-GP titration, was purchased from the European Collection of Animal Cell Culture (ECACC). Dulbecco’s modified Eagle medium (DMEM) supplemented with 10% fetal calf serum (FCS) served as culture medium for all cell lines except for the cell line BHK 21, which was cultivated in Glasgow Minimum Essential Medium (GMEM) supplemented with 10% FCS and 5% tryptose phosphate broth. The cell lines were authenticated by short tandem repeat (STR) profiling at the German Collection of Microorganisms and Cell Cultures (DSMZ) in Braunschweig, Germany, and were negatively tested for mycoplasma contamination prior to utilization (MycoTOOL PCR Mycoplasma Detection Kit, Roche, Mannheim, Germany).
Virus infection
The oncolytic vesicular stomatitis virus VSV-GP-GFP was kindly provided by ViraTherapeutics (Boehringer Ingelheim, Rum, Austria).63 The virus was stored at −80°C at a stock titer of 1.78 × 10^10^ TCID_50_/mL and diluted in culture medium to retain the desired MOI required for each experiment. Cells were seeded 24 h prior to infection, washed once with PBS (Carl Roth, Karlsruhe, Germany), and treated for 1 h with virus-containing infection medium, subsequently washed with phosphate buffered saline (PBS) and incubated in virus-free culture medium. Virus-free culture medium was used as uninfected control (MOCK).
Pretreatment with IFN-β and ruxolitinib
For pretreatment, cells were incubated 24 h after seeding with culture medium supplemented with 2 ng/mL IFN-β (PeproTech, Thermo Fisher Scientific) or 10 μM ruxolitinib (Selleck Chemicals, Planegg, Germany) for 24 h. Then, virus infection was conducted in fresh culture medium as described above.
Pre- and continuous treatment with BET inhibitor BI894999
Cell lines 143100 and 690100 were pretreated with iBET BI894999 (Boehringer Ingelheim) for 24 h and cell line SNU-3178S 48 h after seeding. Cells were incubated with 1 or 2.5 nM of BI894999 in culture medium for 24 h before infection with VSV-GP-GFP. Virus infection and MOCK infection were conducted as described above for 1 h in culture medium without BI894999. After infection, treatment with BI894999 was continued in the respective concentrations until fixation in an SRB viability assay 72 hpi.
Immunofluorescence staining and confocal microscopy
For α-dystroglycan and NUT staining, cells were fixed with 4% PFA (PFA; Otto Fischar GmbH & Co. KG, Saarbrücken, Germany) and blocked with 3% bovine serum albumin (BSA) for 30 min before primary antibodies were added in 3% BSA and left to incubate overnight. Anti-NUT (C52B1) and anti-α-dystroglycan-antibody (both rabbit monoclonal antibodies) were used. The Alexa-546-conjugated anti-rabbit immunoglobulin G (IgG) antibody was employed as a secondary antibody and incubated for 1 h. 4′,6-Diamidino-2-phenylindole (DAPI) (Carl Roth) was administered for nuclear staining (Table S1). Analysis was conducted at 20× magnification with a Leica DMi8 microscope. Data acquisition was performed using Leica Application Suite X 3.7.2.22383.
For IF staining of stress and apoptosis markers, cell line JCM was either left untreated (MOCK) or infected with VSV-GP-GFP (MOI 0.00005). Samples were trypsinized at 72 h post-infection (hpi) and fixed in 4% paraformaldehyde (PFA; Otto Fischar GmbH & Co. KG, Saarbrücken, Germany) and kept at room temperature before final storage at 4°C. For cryo-sectioning, cells were frozen in Tissue-Tek (Sakura Finetek, Umkirch, Germany), cut in 5 μm sections, and left to dry for 15 min at room temperature. After fixation in periodate-lysine-paraformaldehyde (PLP) for 2 min, samples were washed first with PBS (Carl Roth, Karlsruhe, Germany) for 5 min, then with PBS containing 0.005% BSA (AURION, Wageningen, the Netherlands) and 0.05% Tween 20 (Carl Roth) for 20 min. Excess protein was blocked with donkey serum (Sigma-Aldrich, Saint Louis, MO, USA) for 30 min (1:20 dilution in PBS). Samples were then stained for TIA1, (phospho-)eIF2ɑ, caspase-3, LC3C, and G3BP1 (Figure S4); they were incubated with their primary antibodies overnight, followed by a 1-h incubation period with the corresponding secondary antibody. Cell nuclei were stained with DAPI (1:2000; Sigma-Aldrich) for 30 min. Samples were embedded in Mowiol (Merck, Darmstadt, Germany) mounting medium and analyzed in a confocal laser scanning microscope (LSM800, Carl Zeiss, Oberkochen, Germany).
Transmission electron microscopy
NC cell line JCM was either left untreated (MOCK) or was infected with VSV-GP-GFP (MOI 0.00005). Samples were trypsinized 24, 48, and 72 hpi and fixed in Karnovsky-A/-B fixative and kept at room temperature before final storage at 4°C. Karnovsky’s fixative was applied once more after cells were embedded in 3.9% agarose (BioWhittaker Molecular Applications, Rockland, ME, USA) at 37°C and coagulated at room temperature. Cells were then washed with a 0.1 M cacodylate buffer (Morphisto, Offenbach am Main, Germany) and incubated with 1.0% osmium tetroxide (Science Services, Munich, Germany) containing 2.5% K-ferrocyanide (Morphisto) for 2 h. After rinsing with distilled water, samples were contrasted with 2% aqueous uranyl acetate (Merck). After stepwise dehydration in alcohol (50%–95%), samples were suspended in propylene oxide and embedded in a polymerized glycidyl ether (48 h at 60°C; Serva, Heidelberg, Germany). Ultra-thin sections (30 nm), cut in an ultra-microtome (Ultracut, Reichert, Vienna, Austria) and mounted on copper grids (Electron Microscopy Sciences, Hatfield, PA, USA), were analyzed using a Zeiss LIBRA 120 transmission electron microscope (Carl Zeiss) operating at 120 kV. This procedure has been described in detail before.18^,^64
Sulforhodamine B cell viability assay
SRB assays were performed as described previously.65 Cells were washed once with PBS and fixed with trichloroacetic acid (TCA, 10%) for 30 min at 4°C. Cells were washed with water, exposed to UV light for 20 min to inactivate viral residues, and then dried at 40°C for at least 4 h. SRB dye (0.4 in 1% acetic acid, Sigma-Aldrich) was applied and incubated for 10 min. Unbound dye was removed by washing with 1% acetic acid glacial (VWR Chemicals, Radnor, PA, USA). Plates were then dried for 4 h at 40°C, and 10 mM Tris buffer (pH 10.5, Trizma base; Sigma-Aldrich) was added. The absorbance—proportional to cell density—was measured in a Tecan Genios Plus multimode microplate reader at a 564 nm. Percentages in this study are normalized to MOCK-treated cells.
Real-time cell proliferation assay
NC cells were seeded in RTCA E-96-well-plates and 24 h later either directly infected with VSV-GP-GFP (MOI 0.0001–0.1), left untreated (MOCK), or pretreated with IFN-β or ruxolitinib as described above and infected 48 h after seeding. Using the xCELLigence RTCA SP system (Roche Applied Science), dynamic real-time cell proliferation was measured continuously in 30 min intervals over 96 or 120 h. Cell index values indicate the mean ± SD of 4–8 replicates and were calculated using the RTCA software (1.0.0.0805; Roche Applied Science).
VSV-N RT-qPCR
Viral replication was analyzed by RT-qPCR in cell lines 143100, 690100, and SNU-3178S, which were infected 24 h after seeding with VSV-GP-GFP or UV-inactivated VSV-GP-GFP as a control (MOI 0.01). Supernatant was collected 1, 24, 48, and 72 hpi and stored at −80°C. RNA was isolated from the supernatant using the QIAamp Viral RNA Mini Kit (Qiagen, Hilden, Germany). An aliquot of 2 μL isolated RNA or DNA standards (2 × 10^1^–2x10^7^ copies/μL) were added to a 96-well-plate containing 8 μL mastermix, formulated with 0.2 μL forward primer (10 pmol/μL; VSV N_fwd 5′AGT ACC GGA TTG ACG ACT AAT-3′) and 0.2 μL reverse primer (10 pmol/μL; VSV N_rev 5′-TCA AAC CAT CCG AGC CAT TC-3′), 0.2 μL probe (10 pmol/μL; N_probe 5'-[FAM]A CCG CCA CAA GGC AGA GAT GTG GT[BHQ1]-3′) (Eurofins Genomics, Ebersberg, Germany), 0.25 μL reverse transcriptase, 5 μL 2 × reaction mix (iTaq Universal Probes One-Step Kit; Bio-Rad Laboratories Inc., Hercules, CA, USA), and 2.25 μL H_2_O for PCR (ROTH). Two technical and two biological replicates of each sample were analyzed. RT-qPCR was performed at the LightCycler 96 (Roche Applied Science).
Virus quantification with TCID50 assay
VSV-GP-GFP-infected cells were collected 1, 24, 48, and 72 hpi and frozen at −80°C. A dilution series ranging from pure sample to a dilution of 10^−7^ was created in medium for each cell line and time point. An aliquot of 50 μL of each dilution was added to BHK 21 cells seeded 24 h previously. After an incubation period of 72 h, GFP expression was observed via fluorescence light microscopy and deemed positive if a fluorescence signal was discernible, regardless of intensity. The highest dilution at which a cytopathic effect (CPE) (fluorescence signal) could be observed, the dilution factor, and CPE-rate for each dilution were factored into the calculation of the TCID_50_ as implied in the improved Spearman-Kärber method.66
Flow cytometry
Cells were analyzed by FACS at 48 and 72 hpi. They were washed with PBS, detached with Accutase solution (Sigma-Aldrich), and rinsed with FACS buffer (PBS + 1% FBS). Cells were stained with Zombie Aqua (1:300; BioLegend, CA, USA) and then fixed using fix-/perm-solution (True-Nuclear Transcription Factor Buffer Set, BioLegend). After an incubation period of 30 min, cells were permeabilized and subsequently stained with the primary antibody anti-VSV-N and the secondary, BV421-conjugated antibody anti-mIgG2a (Table S1). Annexin V and propidium iodide (PI) staining was performed using the Pacific Blue Annexin V Apoptosis Detection Kit with PI (BioLegend) according to the manufacturer’s instructions. Analysis was conducted using an Attune NxT Acoustic Focusing Cytometer and its respective software (v.3.2.1, Thermo Fisher Scientific). Gates shown in S5 were applied for analysis of GFP and VSV-N expression. All events were counted for Annexin/PI cell death analysis.
Western blot
NUT fusion protein and BRD4 expression after viral infection
NC cells were infected at MOI 0.0005 (143100) or 0.01 (690100, HCC2429, SNU-3178S) or left untreated (MOCK) and collected at 48 hpi. Primary anti-NUT C52B1 rabbit, anti-BRD4 rabbit monoclonal antibody (mAb) (BL-151-6F11), and anti-vinculin mouse mAb were employed together with a secondary (H + L)-HRP-conjugated antibody (Table S1).
NUT fusion protein expression
Untreated lysates from all seven NC cell lines and the control cell line HT29 were collected 72 h after seeding. Primary anti-NUT C52B1 rabbit and anti-vinculin mouse mAb were employed together with a secondary (H + L)-HRP-conjugated antibody (Table S1).
Intracellular IFN-β signaling
For analysis of NUT and P-STAT1/STAT1 expression, NC cells were treated 24 h after seeding with IFN-β or ruxolitinib as described above and were infected 24 h later with VSV-GP-GFP or left untreated (MOCK).
For blotting 24 hpi, cells were detached, suspended in lysis buffer, and subjected to three freeze-thaw lysis cycles. After centrifugation, the supernatant was used for western blot analysis. A suspension of loading buffer and supernatant at a 1:7 ratio was incubated at 95°C for 5 min. A fraction of the suspension containing 100 μg protein (calculated by Bradford protein determination) was loaded onto a 6% or 8% SDS-PAGE, and gel electrophoresis was performed. Proteins were blotted on a P 0.45 PVDF membrane (Amersham Hybond, Cytiva, MA, USA) and blocked with 5% non-fat milk in TRIS-buffered saline including 0.02% Tween 20 (TBS-T), cut into sections if necessary, and incubated with the corresponding primary antibodies anti-P-STAT1 rabbit mAb, anti-STAT1 mouse mAb, and monoclonal anti-vinculin antibody (Table S1) overnight. Membranes were washed and incubated with corresponding secondary antibodies (H+L)-HRP-conjugated goat anti-mouse IgG and (H+L)-HRP-conjugated goat anti-rabbit IgG (Table S1) for 1 h and washed with TBS-T. ECL Western Blotting Detection reagents (Amersham Cytiva, MA, USA) were added and membranes analyzed in the ChemiDoc MP Imaging System using Image Lab software v.5.1 (Bio-Rad Laboratories Inc.).
IFN-β ELISA
NC cell line 143100 was infected with VSV-GP-GFP or measles vaccine virus (MeV) at MOI 0.0001–1 or remained uninfected (MOCK) 24 h after seeding. Supernatants were collected 72 hpi, and IFN-β ELISA was performed according to the manufacturer’s protocol (VeriKine-HS Human Interferon-Beta ELISA Kit; PBL Assay Science, NJ, USA). Values show the mean ± SD of technical triplicates.
Statistical analysis
All data are provided as mean ± standard deviation (SD). Statistical analysis and illustration were performed with GraphPad Prism v.9.4.1 software (GraphPad, LLC, San Diego, CA, USA). p values <0.05 were considered statistically significant and are indicated with asterisks (∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001, ∗∗∗∗p < 0.0001). For SRB viability assays with VSV-GP-GFP mono-treatment and pretreatment with IFN-beta or ruxolitinib, as well as for qPCR and FACS analysis of Annexin V and GFP expression, Welch’s one-way ANOVA for inhomogeneous variances and Dunnett T3 post-hoc tests were performed if ANOVA was statistically significant (p < 0.05). For SRB viability assays with iBET BI 894999 pre-treatment, Mann-Whitney U tests were performed.
Data and code availability
The data generated or analyzed during this study are available from the corresponding author upon request.
Acknowledgments
We thank ViraTherapeutics GmbH (Rum, Austria) and Boehringer Ingelheim Pharma GmbH & Co. KG (Ingelheim, Germany) for providing the oncolytic virus VSV-GP-GFP. L.D.K. the received funding from intramural fortüne program of the Faculty of Medicine, 10.13039/501100002345University of Tübingen (proposal number 3007-0-0), and from the Dres. Bayer Stiftung (Württembergischer Krebspreis - Kategorie Nachwuchs).
Author contributions
Conceptualization, U.M.L. and L.D.K. Formal analysis, R.C.B., A.M.S., J.B., S.B., and L.D.K. Funding acquisition, M.S., R.K., U.M.L., and L.D.K. Investigation, R.C.B., A.M.S., J.B., S.B., B.F., A.V., and L.D.K. Methodology, R.C.B., A.M.S., A.S., I.S., B.F., and A.V. Project administration, U.M.L. and L.D.K. Resources, M.S., B.S., R.K., and U.M.L. Software, R.C.B., A.M.S., and L.D.K. Supervision, J.B., S.B., B.F., U.M.L., and L.D.K. Validation, J.B., S.B., B.S., U.M.L., and L.D.K. Visualization, R.C.B., A.M.S., and A.V. Writing—original draft, R.C.B., A.M.S., and L.D.K. Writing—review & editing, J.B., S.B., M.E.C., B.F., B.S., U.M.L., and L.D.K.
Declaration of interests
B.S. is an employee of ViraTherapeutics GmbH, and R.K. is an employee of Boehringer Ingelheim Pharma GmbH & Co. KG.
The reference list from the paper itself. Each links out to its DOI / PubMed record.
- 1French C.A.NUT Carcinoma: Clinicopathologic features, pathogenesis, and treatment Pathol. Int.68201858359510.1111/pin.1272730362654 · doi ↗ · pubmed ↗
- 2Sholl L.M.Nishino M.Pokharel S.Mino-Kenudson M.French C.A.Janne P.A.Lathan C.Primary Pulmonary NUT Midline Carcinoma: Clinical, Radiographic, and Pathologic Characterizations J. Thorac. Oncol.10201595195910.1097/jto.000000000000054526001144 PMC 4443847 · doi ↗ · pubmed ↗
- 3Kubonishi I.Takehara N.Iwata J.Sonobe H.Ohtsuki Y.Abe T.Miyoshi I.Novel t(15;19)(q 15;p 13) chromosome abnormality in a thymic carcinoma Cancer Res.511991332733282040007 · pubmed ↗
- 4Chau N.G.Hurwitz S.Mitchell C.M.Aserlind A.Grunfeld N.Kaplan L.Hsi P.Bauer D.E.Lathan C.S.Rodriguez-Galindo C.Intensive treatment and survival outcomes in NUT midline carcinoma of the head and neck Cancer 12220163632364010.1002/cncr.3024227509377 PMC 5361614 · doi ↗ · pubmed ↗
- 5Ziai J.French C.A.Zambrano E.NUT gene rearrangement in a poorly-differentiated carcinoma of the submandibular gland Head Neck Pathol.4201016316810.1007/s 12105-010-0174-620352379 PMC 2878624 · doi ↗ · pubmed ↗
- 6den Bakker M.A.Beverloo B.H.van den Heuvel-Eibrink M.M.Meeuwis C.A.Tan L.M.Johnson L.A.French C.A.van Leenders G.J.L.H.NUT midline carcinoma of the parotid gland with mesenchymal differentiation Am. J. Surg. Pathol.3320091253125810.1097/PAS.0b 013e 3181 abe 12019561446 · doi ↗ · pubmed ↗
- 7French C.A.Pathogenesis of NUT midline carcinoma Annu. Rev. Pathol.7201224726510.1146/annurev-pathol-011811-13243822017582 · doi ↗ · pubmed ↗
- 8Chau N.G.Ma C.Danga K.Al-Sayegh H.Nardi V.Barrette R.Lathan C.S.Du Bois S.G.Haddad R.I.Shapiro G.I.An Anatomical Site and Genetic-Based Prognostic Model for Patients With Nuclear Protein in Testis (NUT) Midline Carcinoma: Analysis of 124 Patients JNCI Cancer Spectr.42020 pkz 09410.1093/jncics/pkz 094PMC 716580332328562 · doi ↗ · pubmed ↗