Bifidobacterium infantis-Mediated Herpes Simplex Virus-TK/Ganciclovir Treatment Inhibits Cancer Metastasis in Mouse Model

Previous studies have found that Bifidobacterium infantis-mediated herpes simplex virus-TK/ganciclovir (BF-TK/GCV) reduces the expression of VEGF and CD146, implying tumor metastasis inhibition. However, the mechanism by which BF-TK/GCV inhibits tumor metastasis is not fully studied. Here, we comprehensively identified and quantified protein expression profiling for the first time in gastric cancer (GC) cells MKN−45 upon BF-TK/GCV treatment using quantitative proteomics. A total of 159 and 72 differential expression proteins (DEPs) were significantly changed in the BF-TK/GCV/BF-TK and BF-TK/GCV/BF/GCV comparative analysis. Kyoto encyclopedia of genes and genomes (KEGG) pathway analysis enriched some metastasis-related pathways such as gap junction and cell adhesion molecules pathways. Moreover, the transwell assay proved that BF-TK/GCV inhibited the invasion and migration of tumor cells. Furthermore, immunohistochemistry (IHC) demonstrated that BF-TK/GCV reduced the expression of HIF−1α, mTOR, NF-κB1-p105, VCAM1, MMP13, CXCL12, ATG16, and CEBPB, which were associated with tumor metastasis. In summary, BF-TK/GCV inhibited tumor metastasis, which deepened and expanded the understanding of the antitumor mechanism of BF-TK/GCV.


Introduction
Tumor metastasis is the main cause of poor comprehensive treatment and survival prognosis [1,2]. Despite recent therapeutic advances in cancer treatment, many data show that more than 90% of tumor patients die of distal target organ metastasis [3]. Unlike primary tumors, which can often be cured using local surgery or radiation, metastatic cancers are largely non-curable for circulating tumor cells [1,2,4]. Therefore, the inhibition of tumor metastasis is still a challenge in current antitumor research.
To reduce the high relapse rate after radical surgery and conventional therapies (radiotherapy, chemotherapy, and immunotherapy) [5], novel approaches such as cancer gene therapy have raised hope that the survival rate of tumor patients can be significantly improved [6][7][8][9]. Some gene therapy products have been approved for clinical use, for example, recombinant adenovirus-p53 [10]. Of the gene delivery systems, suicide gene therapy is the most frequently used method for solid tumors, which is used to eliminate tumor cells in three ways: the direct killing effect, bystander effect, and cell signaling pathway [11,12]. The herpes simplex virus thymidine kinase (TK) gene combined with ganciclovir (GCV) is one of the systems that has been studied in the most depth [13][14][15][16][17][18][19].
In this system, viral TK is expressed and simultaneously metabolizes the prodrug GCV to mono-phosphorylated GCV, which will be further converted into GCV triphosphate GCV, an analogue of deoxyguanosine triphosphate. Consequently, triphosphate GCV inhibits DNA synthesis and leads to tumor cell death [18,20].
A safe, effective, and controllable gene delivery system is important for gene therapy. There are three main types of vectors: viral vectors (adenoviral, adeno-associated viral), nonviral vectors (polymers, liposomes), and bacterial vectors (Bifidobacterium, Escherichia coli, and Salmonella) [21][22][23]. Previous studies have demonstrated that some obligate anaerobes (Bifidobacterium, Clostridium) or facultative anaerobes (Salmonella, Escherichia coli, Listeria monocytogenes) can selectively accumulate and proliferate within tumors and suppress tumor growth [24,25]. However, fundamental issues such as safety and targeting remain to be resolved before engineered Salmonella can be used in the clinic [24]. S. typhimurium mutant VNP20009 and its derivative strain TAPET-CD have even been applied in human clinical trials [26,27]. Furthermore, Listeria monocytogenes, an intracellular pathogen, has also been used to develop a tumor vaccine [28]. Nevertheless, some cancer gene therapeutic vectors can still infect normal tissue [29,30]. Bifidobacterium is an obligate anaerobe that selectively localizes and proliferates within hypoxic regions of tumors as a non-pathogenic bacterium. It has been considered an alternative strategy in tumor therapy for its acknowledged recognized safety [19,[31][32][33][34][35][36].
Our previous studies have confirmed that Bifidobacterium infantis (BF)-TK/GCV targeted solid tumors and inhibited the growth of gastric cancer (GC) and several other cancer cell lines through activating extrinsic and intrinsic apoptosis pathways [16,17,19,[37][38][39]. In addition, we detected a significant decrease in CD146 expression after BF-TK/GCV treatment [39]. CD146 is reported to contribute to tumor metastasis and invasion [40,41]. So far, it remains unclear how many molecules are involved in inhibiting tumor cell metastasis after BF-TK/GCV treatment. This study aimed to decipher the mechanism of BF-TK/GCV on inhibiting tumor metastasis.
GC is one of the most common malignant tumors and the fourth leading cause of cancer death worldwide [42]. Despite the gradually declining mortality, GC still burdens many countries in East Asia [43]. Therefore, we comprehensively identified and quantified protein expression profiling in MKN-45 cells by BF-TK/GCV treatment using tandem mass tags (TMTs)-based quantitative proteomics and verified the expression of metastasis-related proteins by immunohistochemistry (IHC) in this study.

Identification of Differentially Expressed Proteins (DEPs)
To better understand the mechanism by which BF-TK/GCV inhibited tumor cell metastasis occurs in GC, MKN−45 cells were treated with PBS, BF-TK, BF/GCV, and BF-TK/GCV, respectively. Then, the quantitative proteomic profiling was detected by 10-plex TMT. The workflow of the TMT-based proteomic analysis is demonstrated in Figure 1A (Figure 1).
Volcano plot filtering was used to identify DEPs with >1.2-fold changes and p values < 0.05. In the BF-TK/GCV/BF-TK group, 102 proteins were upregulated and 57 proteins were downregulated ( Figure 2A). In the BF-TK/GCV/BF/GCV groups, 32 proteins were upregulated and 40 proteins were downregulated (Figure 2A).
Venn diagrams showed that eight and three proteins were observed with similar regulation between the BF-TK/GCV/BF-TK and BF-TK/GCV/BF/GCV groups, which suggested that they were important proteins for the effect of BF-TK/GCV (Figure 2A). Further analysis of the DEPs by unsupervised hierarchical clustering classified these samples into three different cohorts, which also reflected the three distinct characteristics ( Figure 2B).

Functional Enrichment Analysis
To gain insight into the biological classifications between two groups, the DEPs between groups BF-TK/GCV/BF-TK and BF-TK/GCV/BF/GCV were analyzed separately using R software based on the GO database. In the BF-TK/GCV/BF-TK group, 2594, 346, and 402 terms in the BP, CC, and MF were enriched, respectively. The top 10 enriched BP terms are displayed in Figure 3A. BP analysis revealed significant enrichment for regulating dendrite development and lysosomal transport. Meanwhile, 1337, 189, and 214 terms were enriched in the BP, CC, and MF in the BF-TK/GCV/BF/GCV groups. The top 10 enriched Volcano plot filtering was used to identify DEPs with >1.2-fold changes and p va < 0.05. In the BF-TK/GCV/BF-TK group, 102 proteins were upregulated and 57 pro were downregulated (Figure 2A). In the BF-TK/GCV/BF/GCV groups, 32 proteins samples into three different cohorts, which also reflected the three distinct characteristics ( Figure 2B).

Functional Enrichment Analysis
To gain insight into the biological classifications between two groups, the DEPs between groups BF-TK/GCV/BF-TK and BF-TK/GCV/BF/GCV were analyzed separately using R software based on the GO database. In the BF-TK/GCV/BF-TK group, 2594, 346, and 402 terms in the BP, CC, and MF were enriched, respectively. The top 10 enriched BP terms are displayed in Figure 3A. BP analysis revealed significant enrichment for regulating dendrite development and lysosomal transport. Meanwhile, 1337, 189, and 214 terms were enriched in the BP, CC, and MF in the BF-TK/GCV/BF/GCV groups. The top 10 enriched BP terms are shown in Figure 3B. BP category analysis indicated that most genes were related to response to nutrient levels and endoplasmic reticulum stress ( Figure  3). Using R software, KEGG pathway enrichment analysis was performed on the two groups (BF-TK/GCV/BF-TK and BF-TK/GCV/BF/GCV). The top 30 enriched pathways are shown in Figure 3C,D. In the BF-TK/GCV/BF-TK group, KEGG pathway enrichment revealed that the DEPs were primarily implicated in the p53 signaling pathway (TNFRSF10B and Sulforaphane (SFN)) and necroptosis (TNFRSF10B, histone H2A variant H2AX (H2AX), and ATP/ADP translocase (SLC25A4)). The KEGG pathway analysis in the BF-TK/GCV/BF/GCV groups revealed the DEPs mainly participated in apoptosis (BCL2 associated X (BAX), myeloid cell leukemia-1 (MCL1), and α1 tubulin (TUBA1A)). These pathways are associated with cell death consistent with our previous studies [39,41]. In addition, there were pathways such as the Wnt signaling pathway (frizzled class receptor 6 (FZD6), peroxisome proliferator-activated receptor delta (PPARD), and secreted frizzled-related protein 4 (SFRP4)), gap junction (adenylate cyclase 8 (ADCY8) and neural tissue-specific F-actin-binding protein I (Neurabin-1, PPP1R9A)), and cell adhesion molecules (v-set immunoregulatory receptor (VSIR) and leukocyte antigen c (HLA-C)) associated with tumor metastasis. Subsequently, the Venn network showed the relationships between the top 30 enriched pathways and genes in the BF-TK/GCV/BF-TK and BF-TK/GCV/BF/GCV groups ( Figure 3E,F).  Using R software, KEGG pathway enrichment analysis was performed on the two groups (BF-TK/GCV/BF-TK and BF-TK/GCV/BF/GCV). The top 30 enriched pathways are shown in Figure 3C,D. In the BF-TK/GCV/BF-TK group, KEGG pathway enrichment revealed that the DEPs were primarily implicated in the p53 signaling pathway

BF-TK/GCV Inhibits Gastric Cancer Metastasis
The Wnt signaling pathway, gap junction, and cell adhesion molecules signaling pathways were related to tumor metastasis. The MKN−45 cells treated by BF-TK/GCV

BF-TK/GCV Inhibits Gastric Cancer Metastasis
The Wnt signaling pathway, gap junction, and cell adhesion molecules signaling pathways were related to tumor metastasis. The MKN−45 cells treated by BF-TK/GCV demonstrated significantly reduced migration and invasion capabilities compared to the BF-TK or BF/GCV treatment in transwell assays ( Figure 5). Moreover, WB and IHC assays further independently validated the several metastasis-related proteins mentioned above ( Figure 6). The expression of HIF−1α, mTOR, NF-κB1-p105, VCAM1, CEBPB, MMP13, betacatenin, and CXCL12 in the BF-TK/GCV group were significantly decreased compared with the BF-TK or BF/GCV groups ( Figure 6). Obviously, ATG16 was significantly increased in the BF-TK/GCV group ( Figure 6A). However, p-CREB was upregulated in the BFassociated groups compared with the PBS or GCV group ( Figure 6D). 24, x FOR PEER REVIEW demonstrated significantly reduced migration and invasion capabilities compare BF-TK or BF/GCV treatment in transwell assays ( Figure 5). Moreover, WB and IHC further independently validated the several metastasis-related proteins mentione ( Figure 6). The expression of HIF−1α, mTOR, NF-κB1-p105, VCAM1, CEBPB, M beta-catenin, and CXCL12 in the BF-TK/GCV group were significantly de compared with the BF-TK or BF/GCV groups ( Figure 6). Obviously, ATG significantly increased in the BF-TK/GCV group ( Figure 6A). However, p-CR upregulated in the BF-associated groups compared with the PBS or GCV group 6D).  demonstrated significantly reduced migration and invasion capabilities compared to the BF-TK or BF/GCV treatment in transwell assays ( Figure 5). Moreover, WB and IHC assays further independently validated the several metastasis-related proteins mentioned above ( Figure 6). The expression of HIF−1α, mTOR, NF-κB1-p105, VCAM1, CEBPB, MMP13, beta-catenin, and CXCL12 in the BF-TK/GCV group were significantly decreased compared with the BF-TK or BF/GCV groups ( Figure 6). Obviously, ATG16 was significantly increased in the BF-TK/GCV group ( Figure 6A). However, p-CREB was upregulated in the BF-associated groups compared with the PBS or GCV group ( Figure  6D).

Clinical Significance of HIF−1α and VCAM1
TCGA data were used to validate the clinical significance of these protein results revealed that HIF−1α and VCAM1 were significantly upregulated in STA regional lymph nodes metastasis (N) compared with no regional lymph node met (N0) ( Figure 7A). Moreover, the results suggested that STAD patients with a expression of VCAM1 exhibited a poorer overall survival rate (p = 0.026). HIF−1α similar but not significant trend, probably because of the smaller number of patien 0.064) ( Figure 7B,C). Furthermore, the expression of HIF−1α was positively cor with that of VCAM1 (R = 0.28, p = 9.7 × 10 −9 ), suggesting that HIF−1α may play a role to VCAM1 in the progression of GC ( Figure 7D). BF-TK/GCV treatment dec

Clinical Significance of HIF−1α and VCAM1
TCGA data were used to validate the clinical significance of these proteins. The results revealed that HIF−1α and VCAM1 were significantly upregulated in STAD with regional lymph nodes metastasis (N) compared with no regional lymph node metastasis (N0) ( Figure 7A). Moreover, the results suggested that STAD patients with a higher expression of VCAM1 exhibited a poorer overall survival rate (p = 0.026). HIF−1α had a similar but not significant trend, probably because of the smaller number of patients (p = 0.064) ( Figure 7B,C). Furthermore, the expression of HIF−1α was positively correlated with that of VCAM1 (R = 0.28, p = 9.7 × 10 −9 ), suggesting that HIF−1α may play a similar role to VCAM1 in the progression of GC ( Figure 7D). BF-TK/GCV treatment decreased the expression of HIF−1α, VCAM1, and other factors, indicating that BF-TK/GCV treatment could improve the future overall survival rate.  The correlation between HIF−1Α and VCAM1 in TCGA-STAD. Data re mean ± SD. * p < 0.05 vs. N0 group. N0: no regional lymph node metastasis.

Discussion
Although great progress has been made in cancer treatment over the past deca effective prevention of tumor metastasis and invasion is still challenging for c treatment [1,47]. In this study, we found that BF-TK/GCV inhibited tumor meta providing a new antitumor metastasis strategy.
As a model of this study, GC is the fourth leading cause of death and the fifth common cancer [42]. Although systemic chemotherapy, radiotherapy, su immunotherapy, and targeted therapy have been shown to be effective in GC trea [48], the five-year survival rate of GC patients is still poor because of cancer recu due to metastasis [49]. Thus, the ability of BF-TK/GCV to inhibit GC metastasis pro a new method and hope for prolonging the life of GC patients by inhibiting metastasis and invasion.
The delivery vector is crucial to the HSV-TK/GCV system. Some viral-mediated TK/GCV can inhibit the growth or metastasis of insulinoma, glioblastoma, and p cancer cells. Nevertheless, severe immune and inflammatory responses and production costs limit the use of these viral vectors [15,50]. Based on our pr research, here, we further validated that BF-TK/GCV also inhibited tumor metastas invasion. BF is a nonpathogenic, intrinsic, and anaerobic bacterium that sele localizes and proliferates in the central area of solid tumors after administration v The correlation between HIF−1A and VCAM1 in TCGA-STAD. Data represent mean ± SD. * p < 0.05 vs. N0 group. N0: no regional lymph node metastasis.

Discussion
Although great progress has been made in cancer treatment over the past decade, the effective prevention of tumor metastasis and invasion is still challenging for clinical treatment [1,47]. In this study, we found that BF-TK/GCV inhibited tumor metastasis, providing a new antitumor metastasis strategy.
As a model of this study, GC is the fourth leading cause of death and the fifth most common cancer [42]. Although systemic chemotherapy, radiotherapy, surgery, immunotherapy, and targeted therapy have been shown to be effective in GC treatment [48], the five-year survival rate of GC patients is still poor because of cancer recurrence due to metastasis [49]. Thus, the ability of BF-TK/GCV to inhibit GC metastasis provided a new method and hope for prolonging the life of GC patients by inhibiting tumor metastasis and invasion.
The delivery vector is crucial to the HSV-TK/GCV system. Some viral-mediated HSV-TK/GCV can inhibit the growth or metastasis of insulinoma, glioblastoma, and prostate cancer cells. Nevertheless, severe immune and inflammatory responses and high production costs limit the use of these viral vectors [15,50]. Based on our previous research, here, we further validated that BF-TK/GCV also inhibited tumor metastasis and invasion. BF is a nonpathogenic, intrinsic, and anaerobic bacterium that selectively localizes and proliferates in the central area of solid tumors after administration vein or intratumor injection. BF-TK/GCV inhibited the growth of tumors through multiple mechanisms in our previous studies [16,17,19,[37][38][39]51]. In this study, we further confirmed that several metastasis-related proteins decreased expression after BF-TK/GCV treatment, consistent with our previous findings that BF-TK/GCV induced metastasis-related VEGF and CD146 downregulation [19,39].
Consistent with our previous findings, this study also proved that several cancer deathassociated signaling pathways were enriched in the BF-TK/GCV group, including apoptosis, necroptosis, and p53 signaling pathways to exert antitumor effect ( Figure 3C,D) [37,39]. Moreover, vacuolar protein sorting (VPS) genes encode proteins involved in vesicular trafficking. VPS52 has been reported to inhibit the viability and induced apoptosis of GC cells in vitro [52]. Otherwise, VPS41 restrained methuosis and autophagy in cancer [53]. Fortunately, VP41 was downregulated and VPS52 was upregulated by BF-TK/GCV treatment, implying inhibition of tumor viability and growth ( Figure 3A).
In this study, HPX, HRG, and GIG25 were significantly downregulated in the BF-TK/GCV/BF-TK group. Overexpression of HRG (histidine-rich glycoprotein) inhibits cell proliferation and increases apoptosis in hepatocellular carcinoma [54]. Hemopexin (HPX) promotes the invasion and metastasis of pancreatic cancer and colorectal carcinoma cells [55]. GIG25 is a serpin family A member 3 (or SERPINA3), and overexpression of SERPINA3 promotes tumor invasion and migration, and epithelial-mesenchymal transition in triple-negative breast cancer cells [56]. PELO was upregulated by BF-TK/GCV treatment and enriched in the mRNA surveillance pathway. However, it was listed outside of the top 30 pathways (data not shown in Figure 2A). PELO negatively regulates cell migration and metastasis in vivo [56].The results implied that BF-TK/GCV inhibited tumor metastasis through these proteins.
We found that gap junction signaling pathways (ADCY8 and PPP1R9A) were upregulated and cell adhesion signaling pathways (VSIR and HLA-C) were downregulated, which is associated with tumor growth metastasis.
Bax (BCL-2 Associated X) activation leads to the release of apoptotic factor cytochrome C and consequently to cancer cell apoptosis [57]. BF-TK/GCV significantly upregulated BAX ( Figure 3B). Combined with the conclusion of our previous study that BF can induce tumor cell apoptosis, BF-TK/GCV was a promising antitumor agent for clinical treatment [17,19,38,51].
The low expression of ADCY8 is correlated with poor overall survival and progressionfree survival in lung adenocarcinoma [58]. Similarly, VSIR (V-set immunoregulatory receptor) is a negative immune checkpoint regulator that inhibits antitumor immune responses. Anti-VSIR antibody treatment significantly reduces the number of metastatic nodules in the livers of mouse models of PDAC with liver metastases [59]. Our research showed that BF-TK/GCV downregulated VSIR, indicating its antitumor metastasis mechanism.
The tumor microenvironment of metastasis (TMEM) proteins can be described as tumor suppressors or oncogenes [60]. It is reported that the overexpression of circTMEM59 suppresses cell growth, enhances the cell death, and represses the metastatic behaviors of colorectal cancer cells [61]. TMEM59 was upregulated by BF-TK/GCV treatment ( Figure 3A).
The IHC assay confirmed several metastasis-related proteins were downregulated by BF-TK/GCV treatment ( Figure 6C,D). It was reported that the reduction in the gap junction and HIF1α and CXCR4 expression significantly inhibited the metastasis of breast cancer cells [62]. Moreover, neural cell adhesion molecule (NCAM) knockdown inhibits the metastasis of human melanoma cells via the Src/Akt/mTOR/cofilin pathway [63].
HIF-1 is the major regulator of oxygen homeostasis and consists of HIF−1α and HIF-1β subunits. The activation of HIF−1α promotes ovarian cancer cell migration and omental metastasis [64]. In this study, HIF−1α was reduced significantly in the BF-TK/GCV group compared with the BF-TK or BF/GCV groups. mTOR is a serine/threonine protein kinase with two types of protein complex: mTORC1 and mTORC2. Furthermore, it promotes cancer cell metastasis via miR-451 and suppresses glioma cell proliferation and invasion in vitro and in vivo via suppression of the mTOR/HIF−1α/VEGF signaling pathway by targeting CAB39 [65]. In this study, mTOR and HIF−1α were significantly reduced in the BF-TK/GCV group compared to the BF-TK or BF/GCV groups ( Figure 6A,B).
VCAM1 is an important member of the immunoglobulin superfamily and promotes the invasion and metastasis of colorectal cancer by inducing transendothelial migration [67]. The downregulation of VCAM1 expression blocks breast cancer cell metastasis [68]. In this study, VCAM1 was decreased significantly in the BF-TK/GCV group compared with the BF-TK group.
CEBPB is a member of the transcription factor family of CEBP. A previous study found that STAT3 promoted melanoma metastasis by upregulating the expression of CEBPB family members [69]. In this study, CEBPB was significantly reduced in the BF-TK/GCV group compared to the BF-TK or BF/GCV groups.
CXCL12 is a ligand of CXCR4, and the activation of CXCL12/CXCR4 makes M2 polarized macrophages promote liver cancer metastasis by secreting VEGF [70]. The CXCL12/CXCR4 axis plays a pivotal role in tumor development, survival, angiogenesis, metastasis, and tumor microenvironment. For example, the CXCL12/CXCR4 axis activates the NF-κB signaling pathway resulting in the induction of migration, invasion, and EMT processes [71]. BF-TK/GCV downregulated CXCL12 and NF-κB1-p105, indicating its antitumor metastasis mechanism via the downregulation of the CXCL12/CXCR4 axis/NF-κB signaling pathway.
ATG16 plays a key role in autophagy and forms the Atg5-Atg12-Atg16 complex [72]. Autophagy inhibition strongly promotes metastasis to the lung in breast cancer models [73]. Endogenous TMEM59 interacts with ATG16L1 and mediates autophagy in response to Staphylococcus aureus infection [74]. It was suggested that BF-TK/GCV significantly upregulated TMEM59 and ATG16, indicating the potential of inhibiting tumor metastasis.
BF-TK/GCV inhibited GC cell metastasis by downregulating the mTOR/HIF−1α pathway and the CXCL12/CXCR4 axis/NF-κB signaling pathway. More specifically, BF-TK/GCV treatment reduced the expression of HIF−1α, mTOR, NF-κB1-p105, VCAM1, CEBPB, and CXCL12, confirmed by IHC assay or other metastasis-related proteins (HLA-C, HPX, HRG, VSIR, and GIG25) and confirmed by proteomics assay. To better understand the underlying mechanisms, the exact role of these proteins will need to be further explored in the future.

Cell and Animal Treatment
MKN−45 (RRID: CVCL_0434) cells were obtained from the Cell Bank of Type Culture Collection of the Chinese Academy of Sciences (Shanghai Institute of Cell Biology, Shanghai, China) and have been authenticated in the past three years. MKN−45 cells were maintained in a complete growth medium, RPMI 1640 medium with 10% fetal bovine serum, and all experiments were performed with mycoplasma-free cells. The cells were cultured in 100 mm culture dishes in a humidified, mixed environment of 37 • C and 5% CO 2 . The BF-TK/GCV suicide gene therapeutic system was constructed as described previously [17,19]. BF-TK/GCV through intratumoral injections (BF or BF-TK was 1.0 × 10 6 cells/tumor) on day 1 and day 4. GCV (5.0 mg/kg) was intraperitoneal injected daily for 7 days after BF and BF-TK administration. All mice were sacrificed on day 8 of GCV injection, and the subcutaneous tumors were completely excised. All mice were killed by sodium pentobarbital anesthesia for cervical dislocation.
This study was performed according to the guidelines for reporting in vivo experimental studies in animals and the study is reported in accordance with ARRIVE guidelines. The protocol was approved by the Committee on the Ethics of Animal Experiments at the Chongqing Medical University (SYXK2012-0001). All surgery was performed under sodium pentobarbital anesthesia, and all efforts were made to minimize suffering.

Total Protein Extraction
Total protein was extracted from the tumors. Cells were suspended in lysis buffer (1.0% sodium deoxycholate (SDS), 8 M urea) with appropriate protease inhibitors to inhibit protease activity. The mixture was allowed to settle at 4 • C for 30 min, vortexed every 5 min, and treated by ultrasound at 40 kHz and 40 W for 2 min. After centrifugation at 16,000× g at 4 • C for 30 min, the concentration of protein supernatant was determined by the bicinchoninic acid (BCA) method by BCA Protein Assay Kit (Pierce, Thermo, Rockford, IL, USA). Protein quantification was performed according to the kit protocol.

Liquid Chromatography-Tandem Mass Spectrometry (LC-MS/MS) Analysis
Protein digestion was performed according to the standard procedure, and the resulting peptide mixture was labeled using the 10-plex TMT reagent (Thermo Fisher, Art. No. 90111, Waltham, MA, USA) according to the manufacturer's instructions [75]. Then, the pooled samples were fractionated into fractions by ACQUITY Ultra Performance liquid chromatography (Waters, Milford, MA, USA) with ACQUITY UPLC BEH C18 Column (1.7 µm, 2.1 mm × 150 mm, Waters, Milford, MA, USA) to increase proteomic depth. Finally, labeled peptides were analyzed by online nanoflow liquid chromatography-tandem mass spectrometry performed on a 9RKFSG2_NCS-3500R system (Thermo, Waltham, MA, USA) connected to a QExactive Plus quadrupole orbitrap mass spectrometer (Thermo, Waltham, MA, USA) through a nanoelectrospray ion source.

Proteomics Data Analysis
The RAW data were analyzed using Proteome Discoverer (Thermo Scientific, version 2.2, Waltham, MA, USA) against the uniprot-proteome_UP000005640-Homo sapiens (Human) [9606]-74468s-20190823.fasta database. The MS/MS search criteria were as follows: mass tolerance of 10 ppm for MS and 0.02 Da for MS/MS tolerance, trypsin as the enzyme with 2 missed cleavages allowed, carbamidomethylation of cysteine and the TMT of N-terminus and lysine side chains of peptides as fixed modification, and methionine oxidation as dynamic modifications, respectively. The false discovery rate (FDR) of peptide identification was set as FDR ≤ 0.01. A minimum of one unique peptide identification was used to support protein identification. The absolute value of fold change (FC) > 1.20 or < 0.83 and p-value < 0.05 were adopted as criteria for determining the significance of differential expression of a particular protein [76]. The subsequent data used for bioinformatics analysis were the RAW data processed by the difference analysis of BF/GCV vs. PBS, BF-TK vs. PBS, and BF-TK/GCV vs. PBS, so the PBS group data would no longer appear separately in the subsequent analysis.

Transwell Assay
The polycarbonate membrane was coated in the transwell chamber with matrigel (BD Falcon, Bedford, MA, USA) for the invasion assay. MKN−45 cells (approximately 1 × 10 5 in 200 µL serum-free medium) were transferred into the top chamber and added medium with serum in the bottom chamber for 24 h at 37 • C. After the cells on the top side of the membrane were removed, the cells were fixed at the bottom side with 4% paraformaldehyde (PFA) and stained with 0.1% crystal violet. Cells at the undersurface of the chamber were imaged and counted.

Western Blot Analysis
Proteins were taken from the "total protein extraction" section of this study. Protein quantitation was performed by BCA protein assay reagent (Beyotime, Shanghai, China). Equal amounts of protein from the different groups were denatured in SDS sample buffer and separated on 8-10% polyacrylamide-SDS gel based on the protein molecular weight. Proteins were transferred to a polyvinylidene difluoride membrane. respectively. After being washed, tissues were incubated with biotin-labeled secondary antibody for 30 min, followed by Streptavidin-HRP conjugate for 20 min at RT. The presence of the expected protein was visualized by DAB staining and examined under a microscope. Stains with control IgG were used as negative controls.

Statistical Analysis
Statistical analyses were performed using SPSS (Version 25.0, IBM, Chicago, IL, USA). One-way ANOVA was applied for the comparisons among multiple independent variables. The LSD Method was used for multiple pairwise comparisons. An independent sample t-test was used for two sample comparisons. The probability level at which the difference was considered significant was p < 0.05.

Conclusions
BF-TK/GCV inhibited tumor metastasis, which deepened and expanded understanding of the antitumor mechanism of BF-TK/GCV. The result is an important supplement to our previous research on the antitumor molecular mechanism of BF-TK/GCV [17,19,51]. It provides valuable data for the further study of BF-TK/GCV antitumor gene therapy.