Abstract

Background

   Tumor resistance is the primary reason for treatment failure in
   patients with cancer, while oncolytic viruses (OVs), as a novel
   therapy, have been rapidly advancing through clinical evaluation and
   are typically assessed in recurrent tumors that are refractory to
   standard chemotherapy. However, whether the adaptive process that
   fosters chemotherapy resistance influences the efficacy of OV therapy
   is unknown.

Methods

   We analyzed chemo-resistant colorectal cancer (CRC) using in vitro, in
   vivo, and patient-derived organoid models to assess sensitivity to OVs.
   Through RNA sequencing analysis and immunohistochemistry were performed
   in clinical samples that indicated TANK-binding kinase 1 (TBK1)
   expression. Using single-cell RNA sequencing, flow cytometry, and in
   vivo neutralization assays to demonstrate that the combination of TBK1
   inhibitor (TBK1i) and OVs reprograms the tumor immune microenvironment,
   particularly by activating natural killer (NK) cells. Through RNA
   sequencing analysis, we identified intercellular cell adhesion
   molecule-1 (ICAM-1) as a potential target responsible for NK cell
   activation. Subsequently, we designed and conducted rescue experiments,
   both in vitro and in vivo, to validate the influence of ICAM-1 on NK
   cell activity.

Results

   We demonstrated that chemo-resistant CRC showed decreased sensitivity
   to the OV in vitro, in vivo, and patient-derived organoids. Further
   investigation revealed aberrant activation of TBK1 in chemo-resistant
   CRC, which mediated the activation of the type I interferon pathway and
   impaired viral replication. TBK1 inhibition enhanced intratumor viral
   replication and direct oncolysis effect in vitro and augmented the
   antitumor immunity elicited by OVs in vivo. Immune cell profiles
   presented that OV/TBK1i combination reshaped the tumor microenvironment
   and especially activated the NK cell response. Immune cell depletion
   studies demonstrated that NK cells were required for the synergistic
   therapeutic activity of the OV/TBK1i combination. Mechanistically, TBK1
   inhibition synergized with VSVΔ51 to increase ICAM1 expression in a
   RIPK1-dependent manner, promoting NK cell-mediated tumor killing.

Conclusion

   This study presents a promising approach for treating chemo-resistant
   CRC by combining OVs and TBK1i.

   Keywords: Chemotherapy, Colorectal Cancer, Combination therapy, Natural
   killer - NK, Oncolytic virus
     __________________________________________________________________

WHAT IS ALREADY KNOWN ON THIS TOPIC

     * Oncolytic viruses (OVs) represent a promising therapeutic strategy
       for cancer treatment via directly lysing tumor cells and driving
       antitumor immunity. However, the therapeutic efficacy of OVs in
       clinical applications has been less than expected and remains
       limited. Investigating combinational strategies and identifying the
       small molecular compounds that exert synergistic effects with OVs
       may optimize therapeutic outcomes.

WHAT THIS STUDY ADDS

     * This study indicates that TANK-binding kinase 1 (TBK1) acts as a
       barrier limiting the efficacy of OVs in chemo-resistant colorectal
       cancer (CRC). Inhibition of TBK1 enhances the viral replication of
       OVs in chemo-resistant CRC and activates intercellular cell
       adhesion molecule 1-mediated natural killer cell immunity, thereby
       presenting TBK1 inhibitors (TBK1i) as effective sensitizers for OV
       therapy in the treatment of chemo-resistant CRC.

HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE OR POLICY

     * The findings of this study provide new insights into the resistance
       to OV therapy and present a promising approach for treating
       chemo-resistant CRC by combining OVs and TBK1i.

Introduction

   Colorectal cancer (CRC) is the second most prevalent cancer and the
   third leading cause of cancer-related mortality worldwide.[51]^1
   Current treatment strategies encompass chemotherapy, radiotherapy,
   surgical intervention, and targeted therapies.[52]^2 Clinically,
   oxaliplatin (OXA) and 5-fluorouracil (5-FU)-based regimens are the
   front-line chemotherapy approaches for patients with CRC. Despite
   advancements in treatment, chemotherapy resistance remains a primary
   cause of therapeutic failure.[53]^3 4 This resistance is typically
   associated with complex etiologies, including genetic mutations,[54]^5
   apoptosis inhibition,[55]^6 tumor microenvironment (TME) changes,[56]^7
   and enhanced repair of damaged DNA.[57]^8 Consequently, there is an
   urgent need for novel therapeutic strategies to combat resistant and
   metastatic CRC and to induce sustained antitumor immunity, thereby
   improving patient outcomes.

   Oncolytic viruses (OVs) represent a promising new approach in cancer
   therapy. OVs include naturally occurring, attenuated, or genetically
   engineered viruses that can selectively infect and lyse tumor cells
   while retaining low toxicity to normal cells.[58]^9 Current approved
   OVs include H101 (China) engineered adenovirus for head and neck
   cancer, Talimogene Laherparepvec (T-VEC) engineered herpes simplex
   virus (HSV-1, USA) for unresectable metastatic melanoma, and DELYTACT
   (Japan) for glioblastoma. However, clinical application faces multiple
   challenges: the heterogeneity of tumor tissues and the complexity of
   cancer cells mean that a single OV may not completely eradicate all
   cancerous cells, necessitating a combination with other
   therapies.[59]^10 Furthermore, immune responses triggered by viral
   invasion can limit their replication capacity.[60]^11 Existing clinical
   studies demonstrate that while T-VEC combined with pembrolizumab
   increases the ORR of advanced melanoma to 62%, a subset of patients
   still experiences limited clinical benefit. This may arise from
   multifactorial resistance mechanisms, including viral neutralization by
   host immunity, stromal barriers impeding viral spread, or tumor
   cell-intrinsic evasion of viral lysis.[61]^12 13 It could undermine the
   strong safety record that OVs have maintained.[62]^14 15

   TANK-binding kinase 1 (TBK1) is a multifunctional signaling protein
   kinase crucial in regulating immune responses, cell proliferation, and
   apoptosis, and tumor progression.[63]^16 TBK1 mediates the induction of
   type I interferons (I-IFNs) (such as IFN-α/β) and antiviral innate
   immunity via Toll-like receptor (TLR) or Retinoic acid-inducible gene I
   (RIG-I)-dependent pathways in the innate immune system.[64]^17 18
   Increasing evidence[65]19,[66]21 suggests that abnormal activation of
   the TBK1 kinase is closely linked to cancer development, including in
   lung, pancreatic, and CRCs. Chien et al[67]^22 discovered that the
   RalB-Sec5 complex directly recruits and activates the TBK1 kinase
   signaling pathway, bridging innate immune and cancer signaling. This
   pathway can inhibit cancer cell apoptosis. However, despite these
   findings, the relationship between TBK1 and reduced oncolytic
   sensitivity of OVs in chemo-resistant CRC remains unclear, and the
   precise roles and molecular mechanisms involved in overcoming
   chemotherapy resistance need further clarification.

   Employing small molecules to selectively boost the growth and
   replication of OVs in tumor sites has shown great potential as an
   effective strategy.[68]^23 24 In this study, the VSVΔ51 was constructed
   based on the wild-type VSV by deleting the matrix protein M51 site,
   which enhances IFN induction capability, significantly reduces
   toxicity, and retains oncolytic activity. We demonstrate for the first
   time that TBK1 inhibitors (TBK1i) can recover the effective replication
   of VSVΔ51 in chemo-resistant CRC and activate intercellular cell
   adhesion molecule 1 (ICAM1)-mediated natural killer (NK) cell immunity.
   TBK1 inhibition enhances VSVΔ51-mediated oncolysis and antitumor
   immunity, offering new insights into OV therapy and providing a novel
   therapeutic avenue for overcoming tumor chemotherapy resistance.

Materials and methods

Ethics approval and consent to participate

   All animal experiments were performed based on the protocol approved by
   the Institutional Animal Care and Use Committee of Zhujiang Hospital of
   Southern Medical University (LAEC-2023–068).

Patients and samples

   10 chemo-responsive and chemo-unresponsive CRC tissues were obtained
   from the Guangdong Provincial People’s Hospital between 2022 and 2024.
   The detailed clinical information were shown in [69]online supplemental
   table S1.

Cell culture and virus

   HCT116, SW620, HCT8, LOVO, MC38, VERO, and HEK-293T were obtained from
   the Cell Bank of the Chinese Academy of Sciences (Shanghai, China).
   HCT116/OXA, SW620/5FU, HCT8/T, LOVO/5FU, and MC38/OXA were provided by
   Professor Zhi Shi (Jinan University, Guangzhou, China). All cells were
   maintained in a medium supplemented with 10% fetal bovine serum (Gibco,
   USA), 100 U/mL penicillin, and 100 µg/mL streptomycin (Fudebio, China)
   at 37°C in a 5% humidified CO[2] atmosphere. For detailed information,
   refer to online supplementary materials.

   The OV VSVΔ51 was provided by Professor Deyin Guo (Guangzhou National
   Laboratory, Guangzhou, P R China) and was replicated and amplified in
   BHK-21 cells. For the production of pseudotyped VSVΔ51ΔG, the backbone
   of the pseudotyped virus comes from the VSVΔ51 virus, in which the G
   gene was deleted (VSVΔG). The recombinant VSVΔ51ΔG was recovered in
   293 T cells transfected with the plasmid carrying the VSV-G protein.

Construct of TBK1/ICAM1 knockout cell lines

   pLV3-U6-TBK1 (human/mouse)-sgRNA-Cas9-Puro and an empty vector;
   pLV3-U6-ICAM1 (mouse)-sgRNA-Cas9-Puro and an empty vector from Miao
   Ling Technology Company (Wuhan, China). HEK-293T cells were transfected
   with the core vector and packaging plasmids psPAX2 and pMD2.G using
   polyethyleneimine. After changing the medium 8 hours post-transfection,
   lentivirus-containing supernatants were collected after 48 hours, and
   HCT116-OXA/MC38-OXA cells were exposed to the lentivirus. Following
   this, cells were selected with 2 µg/mL puromycin for 7–10 days to
   create stable cell lines.

LDH cytotoxicity assay

   Following the manufacturer’s instructions, cell viability was assessed
   using the lactate dehydrogenase (LDH) Release Assay Kit (Beyotime,
   China). Cells were plated at a density of 10,000 cells per well in
   0.1 mL of medium in 96-well plates. After the treatment, gently collect
   the supernatant from the cell culture medium. According to the
   instructions of the LDH assay kit, add an appropriate amount of the
   supernatant to a 96-well plate, then add the LDH substrate and reaction
   buffer, and mix well. Incubate at room temperature for 30 min, then add
   the stop solution to terminate the reaction. Measure the absorbance
   (Optical Density, OD value) of each well at a wavelength of 490 nm
   using a spectrophotometer. Each experimental group was set up with
   three technical replicates and performed with three independent
   biological repeats. The line graph (bar values) represents the mean
   values of each experimental group.

Quantitative reverse transcription PCR

   Total RNA was isolated using TRIzol reagent (Invitrogen). Complementary
   DNA synthesis was performed using a reverse transcription system from
   Takara. For quantitative reverse transcription PCR, SuperReal PreMix
   SYBR Green (AG) and an ABI Q5 Detection System were used. The messenger
   RNA (mRNA) results were normalized to the expression of ACTB. The
   primer sequences were provided in [70]online supplemental table S2.

Plaque assay

   Briefly, after infection, supernatants or sera-containing viruses were
   collected and diluted to infect HCT116-OXA/SW620-5FU cells plated in
   24-well plates at 90% confluence. After 48 hours of infection, the
   supernatants were discarded, and the cells were washed multiple times
   with phosphate-buffered saline (PBS) before adding 0.8%
   methylcellulose. 48 hours later, the cells were stained with 0.2%
   crystal violet for 12 hours, and the plaques (plaque forming unit,
   pfu/mL) were counted.

Immunofluorescence

   HCT116/OXA cells (2×10^5 per dish) were seeded into glass-bottom dishes
   and incubated with MRT (5 µM) and VSVΔ51 (multiplicities of infection
   (MOI)=0.001) for 24 hours. Tumor cells were fixed with 4%
   paraformaldehyde for 15 min at room temperature, rinsed with 1×PBS, and
   blocked with 0.3% Triton X-100 and 5% Bovine Serum Albumin (BSA) for 30
   min at 37°C. Cells were then incubated with p-MLKL or p-RIPK1
   antibodies (1:100 or 1:400 dilution) for 1 hour, followed by
   anti-rabbit Alexa Fluor 555 (1:1,000 dilution). After
   4',6-Diamidino-2-Phenylindole (DAPI) staining, dishes were rinsed and
   analyzed using a Laser Scanning Confocal Microscope, counting foci in
   100 cells per condition.

Western blot analysis

   Cells were lysed using the RIPA Lysis Buffer (P0013, Beyotime,
   Shanghai), and sodium dodecyl sulfate-polyacrylamide gel
   electrophoresis was performed. Membranes were visualized on a
   chemiluminescence imaging system (UVITEC, Alliance Q9, UK) using
   Immobilon Western Chemiluminescent HRP Substrate (Millipore). A certain
   protein was quantified as the ratio of itself and β-ACTIN/Tubulin
   intensity bands by ImageJ software.

Drugs and animal models

   This study was approved by the Institutional Animal Care and Use
   Committee at Zhujiang Hospital of Southern Medical University. The TBK1
   inhibitor (GSK8612) was obtained from TargetMOI, while Necrostatin-1
   (Nec-1) and ICAM1 inhibitor (A205804) were purchased from MCE. GSK8612
   and Nec-1 were dissolved in Dimethyl Sulfoxide (DMSO), with a maximum
   concentration of 0.001% confirmed as non-toxic to mice. In the
   subcutaneous xenograft model, dissociated MC38/MC38-OXA cells (1×10^6)
   in 50 µL PBS were injected subcutaneously into 5-week-old C57BL/6 mice.
   After 5 days, tumors (∼100 mm³) formed, and mice were assigned to four
   groups. From days 5 to 9 and 12 to 16, mice received intraperitoneal
   injections of GSK8612 (10 mg/kg/day) and Nec-1 (1.5 mg/kg/day) for 2
   weeks, with intratumoral injections of VSVΔ51 (2×10^6 PFU/day) on days
   5 and 9. Tumor length and width were measured every 2 days. After
   approximately 2 weeks, mice were euthanized, and tumors were extracted
   and weighed. Tumor volume was calculated as (length×width²)/2. Some
   tumors underwent single-cell sequencing and flow cytometry, while
   others were fixed for H&E staining and immunohistochemistry analysis.

Multicolor flow cytometry analysis

   C57BL/6 mice with MC38/OXA subcutaneous tumors were harvested, minced,
   incubated with a Tumor Dissociation Kit, mouse (Miltenyi, 130–096–730),
   triturated, passed through a 70 mm screen, resuspended in
   Fluorescence-Activated Cell Sorting (FACS) buffer, and stained with
   fluorochrome-conjugated anti-mouse antibodies from BioLegend or
   eBioscience, as well as appropriate isotype control antibodies. A
   Zombie Red Fixable Viability Kit (BioLegend) was used to stain dead
   cells. We followed a “no-wash” sequential staining protocol (BioLegend)
   to stain dead cells and for surface staining. Intracellular IFN-γ
   staining followed the IFN-γ intracellular staining protocol
   (BioLegend). For single-color compensation controls, Ultracomp eBeads
   (eBioscience) were used and stained with each of nine fluorescently
   conjugated antibodies according to the manufacturer’s instructions. For
   the Zombie Red assay, cells from the non-tumor and tumor quadrants,
   respectively, were used as single-color compensation controls. All
   samples were run in a BD FACSymphony S6. Data were analyzed with FlowJo
   software. Technicians acquiring and gating the data were blinded to the
   treatments.

Co-culture experiments

   HCT116/OXA cells carrying the luciferase reporter gene
   (HCT116/OXA-luc), were added to 96-well plates in different proportions
   (2,000, 5,000). On attachment, the cells were treated with either
   Vehicle, the combination of MRT (5 µM)/VSVΔ51 (MOI=0.001) (Comb), and
   Nec-1 (20 µM)/ICAM1 inhibitor (100 µM) before combining for 6 hours.
   Afterward, NK-92 cells were added at 2,000 seeding densities. The
   luciferase fluorescence signals were analyzed at 0, 6, 12, and 24 hours
   using the IVIS Lumina system to monitor the co-culture killing effects
   of HCT116/OXA-luc and NK-92 cells. The results were analyzed by Living
   Image software.

   RNA-seq and relative data analysis, Single-cell RNAseq analysis,
   Construction of human CRC organoids, and Human NK cell sorting were
   provided in [71]online supplemental file 2.

Statistical analysis

   All statistical analyses were performed using GraphPad Prism software.
   Comparisons between different groups were made using Student’s t-test
   or analysis of variance (ANOVA) as appropriate in the in vitro study.
   Values of tumor volume were analyzed by two-way ANOVA. All error bars
   indicate SD. Differences were considered significant if the p value
   was<0.05.

Data availability

   Gene expression data had been deposited in the Gene Expression Omnibus
   (GEO) repository with the accession code GSE293436. All other data that
   support the findings of this study were available from the
   corresponding authors on reasonable request.

Results

Aberrant activation of TBK1 in chemo-resistant CRC reduces the sensitivity to
OVs

   To explore the therapeutic potential of OVs for combating
   chemo-resistant CRC, we first tested the oncolytic efficacy of two OV
   strains, VSVΔ51 and HSV-1, in five chemo-resistant cell lines and their
   parental CRC cell lines. These cell lines included the OXA-resistant
   HCT116 (HCT116/OXA), OXA-resistant MC38 (MC38/OXA), 5-FU-resistant
   SW620 (SW620/5FU), 5-FU-resistant LOVO (LOVO/5FU), and
   paclitaxel-resistant HCT8 (HCT8/T). CRC cells were treated with varying
   concentrations of chemotherapeutic agents or different MOI with VSVΔ51
   and HSV-1. The sensitivity to different drugs or OVs was quantified
   using the half-maximal inhibitory concentration (IC[50]). IC[50]
   measurements showed that the sensitivities to VSVΔ51 and HSV-1 were
   significantly reduced in chemo-resistant cell lines compared with their
   parental CRC cell lines via cell counting kit-8 assay ([72]online
   supplemental figure S1A–D) and LDH release assay ([73]figure 1A,B).
   Correlation analysis of IC[50] values indicated that the IC[50] values
   for OVs were positively correlated with those for chemotherapeutic
   agents, suggesting that chemo-resistant CRC cell lines showed reduced
   sensitivity to OVs compared with their parental CRC cell lines
   ([74]online supplemental figure S1E).

Figure 1. Aberrant activation of TBK1 in chemo-resistant CRC reduced the
sensitivity to OVs. (A–B) The measurement for oncolytic effect of VSVΔ51 and
HSV-1 in CRC cell lines. HCT116/OXA, MC38/OXA, SW620/5FU, LOVO/5FU, HCT8/T
cells and their parental cells were treated with escalating titers of VSVΔ51
or HSV-1 for 24 hours, then the cell death rate was measured by LDH assay and
the sensitivity to VSVΔ51 (A) or HSV-1 (B) was quantified via calculating
IC[50] value. (C) C57 mice with MC38 and MC38/OXA subcutaneous tumors were
treated with VSVΔ51 (2×10^6 PFU/day, intratumoral injection). (D)
Representative images of tumor tissue and tumor weights for each group from
D. (E) Tumor volumes were recorded every 2 days (n=5). (F–H) The measurement
for the oncolytic effect of VSVΔ51 in CRC organoid. The CRC organoids derived
from chemo-responsive patients (n=3) and chemo-unresponsive patients (n=3)
were infected with/without VSVΔ51 (MOI=0.01) or untreated for 48 hours. Scale
bar=100 µm. #1, patient 1; #2, patient 2. (F) Schematic illustration for the
CRC organoid construction model. (G) Representative microscope images of CRC
organoid after VSVΔ51 treatment. The images shown represent samples from
n=3/6 patients in each group (Chemo-Responsive Group: 3 typical responding
cases selected from n=6; Chemo-Unresponsive Group: 3 typical non-responding
cases selected from n=6). (H) Cell viability of CRC organoid was detected via
measuring the intracellular ATP level. (I) KEGG pathway analysis of RNA-seq
data showed the upregulation of IFN-related pathways in HCT116/OXA compared
with HCT116 cells. (J) Volcano plot of RNA-seq data showed the significant
upregulation of TBK1 in HCT116/OXA cells. (K) Immunoblot analysis of TBK1 in
the five pairs of CRC lines. Quantification of p-TBK1 levels represents the
ratio of chemo-unresponsive to chemo-responsive cells, normalized to β-actin
(mean±SD, n=3 independent experiments): HCT116/OXA (3.387±0.533), SW620/5FU
(5.522±0.205), LOVO/5FU (11.390±0.928), HCT8/T (4.042±0.891), MC38/OXA
(9.990±0.224). (L) Immunohistochemical analysis of TBK1 expression in
clinical CRC patient samples (n=10). Scale bar=50 µm. (M) The bilaterally
implanted mice were treated with VSVΔ51 (3×10^7 PFU/day) by intravenous
injection. MC38 and MC38/OXA cells were implanted at the left side (L) and
right side (R) of mice. Tumors were resected 24 hours post-injection. Viral
titers of subcutaneous tumors were detected by plaque assay. (N)
Immunofluorescence staining was used to evaluate the expression of GFP
(reporter gene for VSVΔ51). Representative images of n=5. Scale bar=100 µm.
Statistical significance was determined using one-way ANOVA in A, B, two-way
ANOVA in E, F, or two-tailed Student’s t-test in M, N. Data represent the
mean±SD. n=3 biological replicates in A, B, L. n=5 biological replicates in
E, F, N. *, p<0.05; **, p<0.01; ***, p<0.001; ****, p<0.0001; n.s.,
non-significant. ANOVA, analysis of variance; CRC, colorectal cancer; 5-FU,
5-fluorouracil; HSV-1, herpes simplex virus 1; IC[50], half-maximal
inhibitory concentration; LDH, lactate dehydrogenase; MOI, multiplicities of
infection; OV, oncolytic virus; OXA, oxaliplatin; TBK1, TANK-binding kinase
1; ATP, adenosine triphosphate; JAK-STAT, janus kinase–signal transducer and
activator of transduction pathway; PFU, plaque forming unit; KEGG, kyoto
encyclopedia of genes and genomes.

   [75]Figure 1
   [76]Open in a new tab

   To confirm the above findings, we next tested the oncolytic effects of
   VSVΔ51 in vivo and patient-derived tumor organoids. For the in vivo
   assays, we established subcutaneous tumors using MC38 or MC38/OXA cells
   in immune-competent C57BL/6 mice ([77]figure 1C). We found that VSVΔ51
   treatment via intratumoral injection significantly inhibited tumor
   growth and tumor burden in MC38 subcutaneous tumors, but not in
   MC38/OXA subcutaneous tumors ([78]figure 1D,E). Furthermore, we
   collected three pairs of chemo-responsive and chemo-unresponsive tumor
   organoids derived from patients with CRC to evaluate the oncolytic
   effect of VSVΔ51 ([79]figure 1F). Time-lapse microscopic examination
   revealed that VSVΔ51 treatment triggered tumor sphere lysis in
   chemo-responsive organoids; however, chemo-unresponsive organoids
   showed no significant morphological changes under VSVΔ51 treatment
   ([80]figure 1G and [81]online supplemental figure S1F). The
   intracellular ATP measurement confirmed that VSVΔ51 treatment did not
   significantly reduce the cell viability of chemo-unresponsive organoids
   ([82]figure 1H). These data demonstrated that the sensitivity of CRC
   tumors to OVs was decreased during the acquisition of chemo-resistance.

   To investigate the underlying mechanisms, RNA sequencing (RNA-Seq) was
   conducted in HCT116 and HCT116/OXA cells without stimulation. KEGG
   pathway enrichment analysis revealed that HCT116/OXA cells upregulated
   the genes relating to I-IFN signaling pathway and JAK-STAT cascade
   compared with HCT116 cells ([83]figure 1I). Strikingly, the mRNA level
   of TBK1, a critical kinase in the I-IFN signaling pathway, was
   upregulated in HCT116/OXA cells ([84]figure 1J). We further evaluated
   the mRNA expression of genes related to the I-IFN pathway at virus-free
   conditions and found that both HCT116/OXA and SW620/5FU cells showed
   higher expression of IFNB1, ISG15, and CXCL10 compared with their
   parental CRC cell lines ([85]online supplemental figure S1G). Western
   blot analysis of five paired CRC cell lines also confirmed the
   upregulation of the protein level of TBK1 and the phosphorylation of
   TBK1 and IRF3 in chemo-resistant cells ([86]figure 1K). To clinically
   validate the upregulation of TBK1, immunohistochemistry was performed
   in clinical samples from 10 paired chemo-responsive and
   chemo-unresponsive patients with CRC ([87]online supplemental table
   S1). The results indicated TBK1 overexpression in the
   chemo-unresponsive samples ([88]figure 1L).

   Considering the pivotal role of TBK1 in antiviral immunity, we next
   evaluated the viral replication of OVs in chemo-resistant CRC tumors in
   vivo. A bilateral tumor model was established to analyze the viral
   replication and tropism of VSVΔ51 ([89]figure 1M). Viral titer
   detection and immunofluorescence analyses demonstrated that MC38/OXA
   tumors showed the reduction of virus replication of VSVΔ51 in vivo,
   compared with MC38 tumors ([90]figure 1N). The above data suggested
   that the aberrant activation of TBK1 led to the reduced oncolysis and
   sensitivity of OVs in chemo-resistant CRC.

TBK1 inhibition potentiates oncolytic efficacy of VSVΔ51 in chemo-resistant
CRC

   The aberrant activation of TBK1 contributes to the proliferation and
   survival of tumor cells, particularly in specific mutational or
   tumorous contexts.[91]^25 26 TBK1 is also identified as an
   immune-evasion gene, and targeting TBK1 can enhance the response to
   PD-1 blockade by sensitizing tumor cells to effector-cytokine-induced
   cell death.[92]^27 These findings drove us to test whether the
   combination of TBK1i could overcome the resistance to OV in
   chemo-resistant CRC. We conducted a combination treatment using VSVΔ51
   and MRT67307/GSK8612, a classical inhibitor of TBK1, in the HCT116/OXA
   and SW620/5FU cell lines. Cell viability was measured after treatment
   with increasing doses of MRT67307/GSK8612 in the presence or absence of
   VSVΔ51. Differences in the area under the curve (DAUC) were calculated
   based on the presence or absence of VSVΔ51. The results demonstrated
   that MRT67307/GSK8612 treatment significantly enhanced the sensitivity
   of VSVΔ51 (MRT67307 DAUC=0.73 for HCT116/OXA and 0.51 for SW620/5FU;
   GSK8612 DAUC=0.53 for HCT116/OXA and 0.35 for SW620/5FU) ([93]figure 2A
   and [94]online supplemental figure S2A). We further quantified this
   sensitization effect by measuring changes in the IC50 shift. Treatment
   with MRT67307/GSK8612 sensitized the chemo-resistant CRC to VSVΔ51 to
   varying degrees, with an IC50 shift of MRT67307 being 3,008-fold for
   HCT116/OXA and 3,029-fold for SW620/5FU, GSK8612 being 1,059-fold for
   HCT116/OXA and 266-fold for SW620/5FU ([95]figure 2B and [96]online
   supplemental figure S2B). Fluorescence imaging and flow cytometry
   results revealed that MRT67307/GSK8612 treatment facilitated the
   replication of VSVΔ51 in the HCT116/OXA and SW620/5FU cell lines
   ([97]figure 2C and [98]online supplemental figure S2C). Furthermore, a
   single treatment of VSVΔ51 or MRT67307/VSVΔ51 combination can exert a
   cytotoxic effect in chemo-sensitive HCT116 and SW620 cell lines
   ([99]online supplemental figure S2D). To exclude the off-target effects
   of the TBK1 chemical inhibitor, we performed CRISPR–Cas9 knockout using
   two different single-guide RNAs targeting TBK1 and confirmed the
   successful knockout through immunoblot ([100]online supplemental figure
   S2E). Knockout of TBK1 significantly increased the oncolytic effect of
   VSVΔ51 in HCT116/OXA and MC38/OXA cells ([101]figure 2D and [102]online
   supplemental figure S2F). Viral titer assay confirmed the enhancement
   of VSVΔ51 production following MRT67307 treatment ([103]figure 2E). In
   chemo-unresponsive CRC organoids, the combination treatment of MRT67307
   and VSVΔ51 also induced severe lysis of tumor spheres and reduced cell
   viability ([104]figure 2F).

Figure 2. TBK1 inhibition potentiated the sensitivity of chemo-resistant CRC
to VSVΔ51 in vitro and in vivo. (A) HCT116/OXA and SW620/5FU cells were
treated with increasing doses of the TBK1 inhibitor MRT67307 (MRT) for 24
hours, either in the absence or presence of VSVΔ51 (MOI=0.001), then the cell
viability was assessed. The differences in the area under the curve (DAUC)
were calculated according to the formula
(AUC[single]–AUC[combined])/AUC[combined]; orange areas represent DAUC. (B)
IC[50] shifts were determined treated with escalating titers of VSVΔ51, with
or without MRT (5 µM) for 24 hours. (C) MRT treatment promoted the viral
infection of VSVΔ51. Cells were infected with VSVΔ51 (MOI=0.001) in the
presence or absence of MRT (5 µM). The fluorescence imaging was monitored for
viral reporter GFP. Scale bar=100 µm. The ratio of infected cells (GFP^+) was
measured by flow cytometry. (D) TBK1 knockout was performed in HCT116/OXA
cells by CRISPR–Cas9 system using different sgRNAs. Cells were infected with
VSVΔ51 (MOI=0.001) for 24 hours, then cell viability was measured by CCK-8.
(E) Plaque assays were conducted to evaluate the titer of VSVΔ51. Cells were
infected with VSVΔ51 (MOI=0.001) in the presence or absence of MRT (5 µM).
Then the cellular supernatant was harvested for the plaque assay. (F) MRT
treatment enhanced the oncolytic effect of VSVΔ51 in chemo-unresponsive CRC
organoids. The viability of CRC organoids was monitored at 48 hours after
treatment of VSVΔ51 and MRT via measuring the intracellular ATP. Scale
bar=200 µm. (G) Nude mice with HCT116/OXA subcutaneous tumors were treated
with TBK1 inhibitor GSK8612 (10 mg/kg/day, intraperitoneal injection), VSVΔ51
(2×10^6 PFU/days, intratumoral injection), or a combination treatment. (H)
Representative tumor images for each group from G. Tumor volume and weight
were presented as the mean±SD (n=5). (I) MC38/OXA subcutaneous xenografts in
C57 mice were treated with GSK8612 (10 mg/kg/day, intraperitoneal injection),
VSVΔ51 (2×10^6 PFU/days, intratumoral injection), or a combination treatment.
(J) Representative images, weights, and tumor volume for each group mentioned
in I. Tumor weight and volume were presented as the mean±SD (n=5). (K) Tumor
size was recorded every 2 days and the tumor-free ratio was calculated at the
end of the experiment. (L) Viral gene VSV-G was measured by qPCR in C57 mice
tumors. Statistical significance was determined using one-way ANOVA in A, B,
F, two-way ANOVA in H, J, K, or two-tailed Student’s t-test in D, E, L. Data
represent the mean±SD. n=3 biological replicates in A, B, C, D, E and F. n=5
biological replicates in H, J, K and L. *, p<0.05; **, p<0.01; ***, p<0.001;
****, p<0.0001; n.s., non-significant. ANOVA, analysis of variance; CRC,
colorectal cancer; 5-FU, 5-fluorouracil; IC[50], half-maximal inhibitory
concentration; MOI, multiplicities of infection; mRNA, messenger RNA; OXA,
oxaliplatin; PBS, phosphate-buffered saline; qPCR, quantitative PCR; sgRNA,
single-guide RNA; TBK1, TANK-binding kinase 1; ATP, adenosine triphosphate;
PFU, plaque forming unit.

   [105]Figure 2
   [106]Open in a new tab

   We next evaluated the therapeutic potential of the VSVΔ51/TBK1i
   combination in vivo. GSK8612 is a specific inhibitor of TBK1 and is
   suitable for assessing in vivo. Nude mice with HCT116/OXA subcutaneous
   tumors were randomized into four groups and treated with (1) vehicle,
   (2) intraperitoneal injection of GSK8612, (3) intratumoral injection of
   VSVΔ51, and (4) a combination of GSK8612 and VSVΔ51 ([107]figure 2G).
   The VSVΔ51/GSK8612 combination treatment markedly inhibited tumor
   growth and burden ([108]figure 2H). Next, we developed subcutaneous
   tumors using MC38/OXA cells in immunocompetent C57BL/6 mice
   ([109]figure 2I). The VSVΔ51/GSK8612 combination treatment induced
   tumor regression in the mice and, at the same time, led to the complete
   tumor elimination in one of five mice ([110]figure 2J,K), indicating
   that the VSVΔ51/GSK8612 combination contributed to antitumor immunity.
   Meanwhile, GSK8612 treatment significantly elevated VSV-G expression in
   tumor tissues but not in the heart, liver, spleen, lungs, and kidneys
   ([111]figure 2L and [112]online supplemental figure S2G,H).
   Furthermore, this combination therapy exhibited no significant toxicity
   to major organs such as the heart, liver, spleen, lungs, and kidneys,
   indicating a favorable biosafety profile ([113]online supplemental
   figure S2I). These data suggest that TBK1 inhibition facilitates
   oncolysis by amplifying viral replication and enhancing antitumor
   immunity.

VSVΔ51/TBK1i combination reshapes the immune profiles and enhances natural
killer cell response

   The remarkable therapeutic effect observed in the immune-competent
   model prompted us to investigate the impact of the VSVΔ51/GSK8612
   combination on the immune profile. We conducted single-cell RNA
   sequencing (scRNA-seq) analysis on CD45^+ cells isolated from MC38/OXA
   tumors in C57BL/6 mice treated with either TBK1i, VSVΔ51, or the
   VSVΔ51/TBK1i combination ([114]figure 3A). After standard data
   processing and quality control procedures, we obtained transcriptomic
   profiles for 90,651 cells. Unsupervised clustering identified 19
   distinct clusters ([115]figure 3B and [116]online supplemental figure
   S3A). We further classified these clusters into 10 cell types by
   canonical cell type specific markers, including T cells,
   monocytes/macrophages, neutrophils, basophils, dendritic cells (DCs),
   plasmacytoid DCs, NK cells, B cells, fibroblasts, and erythrocytes
   ([117]figure 3C,D). By quantifying the cell composition of the samples,
   we found that combination treatment increased the frequency of
   macrophages and NK cells; unexpectedly, the frequency of T cells
   decreased in the combination group ([118]figure 3E).

Figure 3. TBK1 inhibitor combined with VSVΔ51 remodels the tumor immune
microenvironment. (A) Single-cell sequencing workflow diagram. (B) The
two-dimensional plots of UMAP dimensionality reduction for cell type. (C) Dot
plot showing marker genes for 10 cell types. (D) The two-dimensional plots of
UMAP dimensionality reduction for 10 clusters. (E) The percentage of 10
clusters between each sample was represented on a proportion chart. (F–H)
UMAP of reclustered macrophages, showing changes in the proportions of
various subtypes between different groups. (I–K) UMAP of reclustered T and NK
cells, showing changes in the proportions of various subtypes between
different groups. (L) Violin plots show the differential expression of
immune-related genes among subtypes of T/NK cells. (M) Volcano plot shows
that the expression of Gzmc in the combination group was significantly higher
than in the other groups. n=2 biological replicates. DC, dendritic cell; NK,
natural killer; pDC, plasmacytoid dendritic cell; scRNA-seq, single-cell RNA
sequencing; TBK1, TANK-binding kinase 1; UMAP, uniform manifold approximation
and projection.

   [119]Figure 3
   [120]Open in a new tab

   We then reanalyzed macrophages and got six functional subclusters
   ([121]figure 3F and [122]online supplemental figure S3B). The
   composition of subclusters in each treatment group showed no
   significant differences; however, the proportion of the C1_C1qa
   subcluster was significantly increased in the combination treatment
   group ([123]figure 3G,H). Studies have shown that C1q^+ macrophages
   express multiple inhibitory molecules, and their high infiltration was
   significantly correlated with poor prognosis.[124]^28 We continued to
   analyze T and NK cells, which were divided into 14 subgroups with
   distinguishable signatures ([125]figure 3I and [126]online supplemental
   figure S3C). Based on the internal statistical proportions of the
   samples, the cell proportions of CD4T_C1_Treg, CD8T_C4_Gzmf,
   CD8T_C5_Ifng, and NK_C1_Ccl5 were elevated in the combination or GSK
   groups, with a much higher proportion of NK_C2_Gzmc in combination
   groups ([127]figure 3J). The increase in NK subclusters was also
   observed in the T/NK group preferences by Ratio of observed to expected