Anesthetic Propofol Promotes Tumor Metastasis in Lungs via GABAAR‐Dependent TRIM21 Modulation of Src Expression

Abstract Generation of circulating tumor cells (CTCs), a key step in tumor metastasis, occurs during surgical tumor resection, often performed under general anesthesia. Propofol is the commonly used anesthetic, but its effects on CTCs and tumor metastasis remain largely unknown. Propofol effects are investigated in an experimental metastasis model by injecting tumor cells and, subsequently, low‐ or standard‐dose propofol to nude mice through tail vein. Propofol‐ or vehicle‐treated tumor cells are also injected to the mice. An in vitro tumor cell–vascular endothelial cell adhesion assay, immunofluorescence, and other methods are employed to assess how propofol affects tumor cell adhesion and extension. Propofol induces more lung tumor metastasis in mice than control. Mechanistically, propofol enhances tumor cell adhesion and extension through GABAAR to downregulate TRIM21 expression, leading to upregulation of Src, a protein associated with cell adhesion. These results demonstrate that propofol may promote tumor metastasis through GABAAR–TRIM21–Src mechanism.


Introduction
Perioperative factors including anesthesia and surgery have been reported to associate with poor prognosis in tumor patients, period (e.g., 24 to 72 h) or multiple administrations of propofol treatment. Therefore, it is important to establish more clinically relevant models to determine the effects of propofol on CTCs and tumor metastasis.
Recently, Li et al. systematically compared the difference of anesthetics on tumor metastasis and reported that inhalation anesthetic sevoflurane led to significantly more lung metastasis of breast cancer in mice after surgical resection of tumor as compared to propofol by changing the tumor microenvironment in lungs. [9] However, the study did not demonstrate whether anesthetic itself (vs nonanesthesia condition) could promote or inhibit tumor progression. Therefore, it remains unknown whether standard-dose propofol can promote tumor metastasis compared to low-dose propofol or nonanesthesia condition.
To assess this, here, we employed an experimental metastasis mouse model, often used to study tumor metastasis, [10] rather than a spontaneous metastasis model, to determine the effects of standard-dose propofol on tumor metastasis as compared to low-dose propofol or nonanesthesia condition. This is because although spontaneous metastasis models are valuable for in vivo tumor research, it is not possible to perform tumor resection with low-dose propofol or nonanesthesia conditions in spontaneous metastasis models.
We injected tumor cells and, subsequently, low-or standarddose propofol into nude mice via the tail vein to conceptually mimic the condition by which propofol mixes with CTCs. We also injected propofol-or vehicle-treated tumor cells in nude mice via the tail vein to further determine the effects of propofol versus nonanesthesia condition on tumor metastasis. These two methods could conceptually represent the clinical condition of general anesthesia with standard-dose propofol versus regional anesthesia with or without low-dose propofol for sedation.
Surgical tumor resection can generate CTCs in the blood. [11] Tumor metastasis begins with local migration and invasion of cancer cells from their primary sites into blood (intravasation), and is followed by adhesion of CTCs to vascular endothelial cells (VECs); finally, CTCs leave the blood stream (extravasation). [7b,12] During extravasation, CTCs adhere to VECs, [10b,13] and extension of these tumor cells with membrane protrusions makes it easier for the cells to adhere to and further penetrate through VECs, exiting the blood vessel [10b,14] to infiltrate and colonize different organs. [15] Thus, adhesion and extension of CTCs to VECs are critical steps during tumor metastasis. In the present study, we specifically focused on investigating the effects of propofol on extravasation (e.g., adhesion and extension), but not intravasation (e.g., migration and invasion).
Propofol is a type-A -aminobutyric receptor (GABA A R) agonist. [16] GABA A Rs exist in peripheral tissues [17] and promote migration and invasion of some tumor cells. [18] Therefore, we sought to determine whether propofol could promote tumor metastasis via enhancing adhesion and extension of tumor cells in vivo in nude mice and in vitro through action on GABA A R.
Other factors are important regulators of tumor metastasis. Src is a nonreceptor cytoplasmic kinase that mediates formation of focal adhesions [19] and arrangement of actin filaments for cell extension, [20] which can regulate tumor cell metastasis. [21] Upregulation of Src expression critically contributes to Src activation, [22] which can be detected in many tumors including colon, lung, breast, and endometrial tumors. [23] Src can regulate migration and invasion of tumor cells. [24] Activation of Src promotes breast tumor cell extravasation into the brain parenchyma. [25] Inhibition of Src activity effectively reduces adhesion of colorectal cancer cell line SW620 and HT29 to fibroblasts. [26] Notably, PP2, a specific inhibitor of Src activation, can reduce tumor metastasis. [27] Further, Src can be quickly metabolized via ubiquitination. [28] We, therefore, assessed whether propofol could act on GABA A R to regulate Src expression by affecting Src ubiquitination, consequently influencing tumor metastasis.
Additionally, members of the tripartite motif (TRIM) protein family contribute to tumor cell metastasis. TRIM21 suppresses progression of tumor metastasis by ubiquitylation and degradation of Snail [29] and I B kinase . [30] In addition, TRIM21 can regulate adhesion of monocytes to endothelial cells. [31] TRIM21 also interacts with cytoskeletal proteins. TRIM21 colocalizes with Factin, which can inhibit TRIM21 function. [32] Therefore, it is possible that TRIM21 may regulate the effects of propofol on tumor metastasis, adhesion between tumor cells and endothelial cells, and cytoskeletal protein (e.g., F-actin). However, whether propofol can act on GABA A R to regulate TRIM21 expression, leading to changes in Src expression and consequently alterations in adhesion and extension, remains unknown.
Therefore, the objective of the present study was to investigate the effects of propofol on tumor cell metastasis and the underlying mechanisms of these effects. We hypothesized that standard-dose propofol promotes tumor metastasis as compared to low-dose propofol or nonanesthesia condition through GABA A R-TRIM21-Src mechanism. We found that propofol activated GABA A R to decrease expression of TRIM21 and increase expression of Src, which enhanced tumor cell adhesion and extension, leading to promote tumor metastasis in lungs of mice.

Propofol Promotes Tumor Cell Metastasis Mediated by GABA A R
We injected HCT116 cells and, subsequently, a low or standard dose of propofol into nude mice via the tail vein (Figure 1a). Ex vivo bioluminescence imaging of mouse lungs four weeks after injections showed that treatment with standard-dose propofol (240 mg kg −1 h −1 over 1 h) led to significant increases in tumor metastasis, represented by photon intensity (7.59 ± 0.43 vs 9.14 ± 1.11, P = 0.019, Student's t-test) (Figure 1b,c) and number of nodules (16.55 ± 15.49 vs 39.6 ± 18.57, P = 0.031, Student's ttest) (Figure 1d,e) as compared to treatment with low-dose propofol (20 mg kg −1 ). These data suggest that propofol may promote lung metastasis of tumor cells in vivo.
Next, we compared the effects of propofol anesthesia versus the nonanesthesia condition on tumor metastasis by injecting propofol-treated versus vehicle (DMSO)-treated HCT116 cells into nude mice via the tail vein ( Figure 1f). We found that the propofol-treated tumor cells had increased photon intensity (7.31 ± 0.47 vs 7.98 ± 0.16, P = 0.019, Student's t-test) (Figure 1g,h) and number of nodules (15.44 ± 7.21 vs 42.33 ± 14.62, P = 0.001, Student's t-test) (Figure 1i,j) as compared to DMSO-treated tumor cells (vehicle control condition) four weeks after injection. These a) Study diagram: Injection of HCT116 cells overexpressing luciferase was followed by injection of low-dose or standard-dose propofol into the tail vein of BALB/c nude mice. b) Tumor metastasis in lungs of mice injected with low-versus standard-doses of propofol, by ex vivo bioluminescent assay. c) Quantification of bioluminescent photon intensity (N = 10 in low-dose group, N = 7 in standard-dose group; mean ± SD, Student's t-test, P = 0.019). d) H&E staining of metastatic lung nodules (scale bar = 200 μm). e) Quantification of metastatic nodules (N = 9 in low-dose group, N = data further suggest that propofol may promote tumor metastasis to lungs in the mice. In vitro, treatment with propofol caused more HCT116 cells to adhere to a monolayer of human umbilical vein endothelial cells (HUVECs) than control (Figure 1k,l). Propofol also induced greater cell membrane protrusion extensions with more apparent F-actin structures (Figure 1m). Propofol increased focal adhesion sites detected by immunofluorescence staining of focal adhesion proteins Talin and Vinculin (Figure 1n; Figure S1a Propofol is a GABA A R agonist. We found that GABA A R antagonist bicuculine induced depolarization of membrane potential in HCT116 cells, indicating the cells possess functional GABA A R ( Figure S2a,b, Supporting Information). Further, like propofol, the GABA A R agonist muscimol-treated HCT116 cells also increased lung metastasis of tumor compared to control HCT116 cells (Figure 1o Figure  S5g-i, Supporting Information). However, NMDA receptor antagonist MK-801 and acetylcholine receptor antagonist benzethonium did not regulate adhesion of HCT116 cells ( Figure S6a,b, Supporting Information). These data suggest that propofol promotes metastasis of tumor cells by enhancing their adhesion and extension, specifically via acting on GABA A R.   Figure S11f,g, Supporting Information). These data suggest that propofol inhibits Src ubiquitination by acting on GABA A R, leading to accumulation of Src, which then promotes tumor cell adhesion and extension.

GABA A R Upregulates Src via TRIM 21
Treatment of HCT116 cells with propofol (Figure 3a; Figure  S12a Overexpression of TRIM21 also attenuated the propofolinduced upregulation of Src and promotion of adhesion of HCT116 cells on HUVECs (Figure 3l,m). Overexpression of TRIM21 attenuated the muscimol-induced enhancement of HCT116 cell adhesion and extension (Figure 3n-q; Figure S14, Supporting Information). These data suggest GABA A R activation increases Src by reducing TRIM21, leading to enhancement of adhesion and extension in HCT116 cells.

Inhibition of Src Attenuates Propofol-or Muscimol-Promoted Tumor Metastasis in the Lungs
Finally, we used a pharmacological but not a genetic approach to determine the more clinically relevant effects of inhibition of Src on propofol's effects. Ex vivo bioluminescent assays showed that treatment with PP2, an inhibitor of Src, significantly attenuated the propofol-induced increases in the number of tumor nodules in mouse lungs (Figure 4a,b). Histological analysis further confirmed that PP2 significantly attenuated the propofolpromoted metastasis of HCT116 cells in mice lungs (Figure 4c,d). Ex vivo bioluminescent and histological analysis also confirmed that PP2 significantly attenuated the muscimol-promoted metastasis of HCT116 cells in mice (Figure 4e-h). These data suggest that PP2 can inhibit the effects of propofol on tumor cell metastasis in mouse lungs.
Collectively, these results suggest a potential pathway wherein propofol activated GABA A R, which caused downregulation of TRIM21 and consequent upregulation of Src, leading to enhancement of adhesion and extension of the tumor cells with VECs, potentially promoting tumor metastasis ( Figure 5).

Discussion
The data in this proof-of-concept study demonstrate that intravenous administration of standard-dose propofol was associated with greater tumor metastasis in the lungs of nude mice as compared to low-dose propofol or nonanesthesia condition. Mechanistically, propofol enhanced adhesion and extension of tumor cells to VECs, a critical step in tumor metastasis, [33] by acting on GABA A R. Activation of GABA A R decreased expression of the cytosolic ubiquitin ligase TRIM21, leading to increases in the expression of Src, a protein associated with cell adhesion with VECs.
We showed that treatment with standard-dose propofol promoted tumor metastasis to the lungs as compared to low-dose   (Figure 1a-e). We compared the effects of low versus standard dose of propofol on tumor metastasis by injecting different amounts of propofol via tail vein. However, this setting could not allow us to determine the effects of propofol anesthesia versus nonanesthesia condition because we could not use the pump to administer the same volume (600 μL) of intralipid (the vehicle of propofol) via tail vein without general anesthesia over 1 h. The injection of HCT116 cells without intralipid to mice would cause confounding influence because no vehicle of propofol was administered to the mice. We, therefore, used low-dose propofol as the control condition in the experiment, which also represented the clinical condition for patients receiving standard-dose propofol for general anesthesia or lowdose propofol for sedation during regional anesthesia.
To compare the effects of propofol anesthesia versus nonanesthesia condition on tumor metastasis, we established another system by pretreating the tumor cells with propofol dissolved in DMSO versus DMSO (vehicle) alone for 3 h. We then injected these pretreated tumor cells into mice via the tail vein. We found that propofol-treated tumor cells also resulted in greater metastasis in lungs of the nude mice as compared to the vehicletreated tumor cells (Figure 1f-j). These results further suggest that propofol would promote tumor metastasis in mice as compared to nonanesthesia condition.
Mechanistically, propofol enhanced adhesion and extension of tumor cells to VECs, a critical step in tumor metastasis, [33] by acting on GABA A R. Activation of GABA A R decreased expression of the cytosolic ubiquitin ligase TRIM21, increasing expression of Src, a protein associated with cell adhesion with VECs.
Given that our objective was to determine the effects of standard-dose propofol on tumor metastasis as compared to low-dose propofol or nonanesthesia condition, the experimental metastasis model was the most suitable choice. While spontaneous metastasis models are valuable for in vivo research, in such mouse models, it is not possible to perform tumor resection with low-dose propofol or nonanesthesia conditions. The experimental metastasis mouse model circumvents this limitation by injecting tumor cells and propofol to the mice via tail vein.
Propofol is a commonly used anesthetic in patients undergoing surgery, and surgical tumor resection may be associated with generation of CTCs, [11] a critical step for tumor metastasis. However, previous studies focused on assessing the effects of propofol on tumor cell proliferation, migration, and invasion [8, 34] but not on CTCs adhesion. In the present study, we specifically assessed the effects of propofol on CTCs by mixing the tumor cells with propofol in the blood of mice through injecting the tumor cells immediately before injecting propofol into the mouse tail vein (Figure 1). We also mixed tumor cells with propofol in cell culture medium before injecting the propofol-treated cells to the mice via tail vein (Figure 1).
A recent study illustrated that propofol led to less lung metastasis than inhalational anesthetic sevoflurane in both syngeneic murine 4T1 and xenograft human MDA-MB-231 breast cancer models, demonstrating the effects of different anesthetics on tumor metastasis. [9] Specifically, sevoflurane promoted the lung metastasis by modulating inflammatory cytokines and activating STAT3 pathway in the lungs. [9] However, the study did not compare the difference of standard-dose anesthetic (e.g., propofol) versus a low dose of the same anesthetic or nonanesthesia (control condition) on tumor metastasis. Moreover, this study determined the effects of anesthetics on the microenvironment of tumor, but not tumor cells themselves, as the underlying mechanisms.
In contrast, the present study, by employing both pharmacological and genetic regulation of the expression of GABA A R and TRIM21, revealed that propofol acted on GABA A R-dependent TRIM21 modulation of Src expression of the tumor cells to lead to more lung metastasis as compared to nonanesthesia control condition. These findings will promote future in vivo approaches (e.g., employment of two-photon microscopy to visualize the acute dynamic process of cell adhesion) to further reveal the role of GABA A R, TRIM21, and Src, in the effects of propofol on tumor metastasis in vivo.
Another study reported that propofol (2-10 μg mL −1 for 6 h) promotes migration and invasion of oral squamous carcinoma cells; [34a] Meng et al. also reported that propofol (2-10 μg mL −1 for 1-12 h) increases proliferation and migration of human breast tumor cells; [8c] and Garib et al. showed that propofol (6 μg mL −1 ) enhances migration of breast carcinoma cells. [34b] Consistently, our results showed that propofol-treated cells adhered to VECs more easily, had more focal adhesion sites, and demonstrated greater extension, promoting tumor progression.
Activation of Src contributes to epithelial-mesenchymal transition (EMT), which occurs during intravasation. [35] Src can phosphorylate focal adhesion proteins [36] to further regulate tumor metastasis. [37] In addition, Src can promote mesenchymalepithelial transition (MET)-related processes. Inhibition of Src activity suppresses adhesion capacity of lung cancer A549 cells [38] and inhibits adhesion of colorectal cancer cell line SW620 and HT29 to fibroblasts. [26] Thus, Src can regulate both EMT-and   MET-related processes in tumor cells. Accordingly, we found that decreasing Src inhibited the propofol-promoted tumor cell adhesion and extension.
Interestingly, PP2, a Src kinase inhibitor, decreased both Src and phosphorylated Src amounts in the present study. The underlying mechanism of such observation is not known at present. We postulated that reduction in phosphorylated Src could lead to increased degradation of Src as some sites of protein phosphorylation could block the action of degradation enzyme of the protein. [39] Consistently, a previous study also demonstrated that PP2 decreased both Src and phosphorylated Src protein levels. [40] Moreover, we did not demonstrate the direct interaction between E3 ligase-TRIM21 and Src in the Co-IP assay, suggesting that the degradation of Src could be indirectly regulated by TRIM21.
Adhesion of tumor cells occurs within a short period of time, [14b] suggesting that the regulatory process should be quick. Ubiquitination is a rapid process regulating protein expression. [41] In line with this, we found that propofol or activation of GABA A R increased Src expression by inhibiting Src ubiquitination to promote adhesion and extension of tumor cells to VECs. Moreover, colocalization of GABA A R and focal adhesion protein confirmed a direct path of close spatial regulation. Finally, Src inhibitor PP2 blocked GABA A R or propofol-mediated lung metastasis (Figure 4). These results reveal that propofol might enhance adhesion and extension of tumor cells via GABA A R-regulated changes in Src expression. We also showed that TRIM21 could be regulated by propofol and GABA A R agonist. TRIM21 regulated Src expression and, consequently, tumor cell adhesion and extension. Taken together, these data suggest a potential propofol-GABA A R-TRIM21-Src-tumor cell adhesion/extension mechanism. Future studies are warranted to further test whether the presence of such a cascade is responsible for propofol-promoted tumor metastasis.
On the other hand, several studies demonstrated that propofol inhibits tumor cell progression. [42] Zhang et al. reported that continuous exposure to propofol (8 μg mL −1 , every week) inhibits growth of xenografts of primary colorectal tumors. [34c] Wang et al. showed that invasion, migration, and growth of malignant pheochromocytoma cells can be inhibited by propofol (1-10 μg mL −1 ) treatment for 24-48 h. They also found that propofol (35 mg kg −1 ) could inhibit growth of malignant pheochromocytoma tumors in a xenograft model. [34d] The reason for these contrasting findings is not yet known. However, these studies differed from present study in several ways: 1) they did not assess the effects of propofol on tumor metastasis and only used a subcutaneously injected xenograft model; 2) multiple administrations of propofol were used in some studies; 3) treatment times were generally longer.
Several studies demonstrated that propofol can attenuate the proliferation, invasion, and migration of tumor cells. [43] However, most of these studies were performed in vitro with long duration of propofol treatment, including 24, [43c,e] , 48, [43a,d] and 72 h. [43b] In the present study, we only treated the tumor cells with propofol for 3 h, mimicking common surgery time in patients, for both in vitro (adhesion and extension of tumor cells) and in vivo (tumor metastasis) experiments. Moreover, injection of tumor cells followed by injection of standard-dose propofol also increased tumor metastasis as compared to injection of tumor cells followed by injection of low-dose propofol in mice.
The present study has several strengths. First, we established a system to conceptually mimic the clinical situation in which propofol encounters CTCs. Second, we showed that intravenous administration of propofol promoted tumor metastasis to lungs in mice by using experimental metastasis model. Finally, we revealed a GABA A R-TRIM21-Src-tumor cell adhesion/extension may be the underlying mechanism of this propofol-promoted tumor metastasis.
The present study also has several limitations. First, we did not inject the GABA A R knockout or TRIM21 overexpression HCT116 cells into the mice to further determine the role of GABA A R or TRIM21 in the propofol-promoted tumor metastasis in vivo. This is because TRIM21 overexpression or GABA A R knockout itself could have long lasting effects in regulating tumor metastasis even after the clearance of propofol from blood, thus generating potential confounding influence. However, the in vitro data demonstrated the role of GABA A R or TRIM21, by using genetic modifications, in the propofol-promoted tumor cell adhesion to VECs. Future studies using genetically modified tumor cells and two-photon microscopy to visualize the acute dynamic process of cell adhesion will allow us to further determine whether propofol can promote adhesion of CTCs to VECs via interactions of GABA A R, TRIM21, and Src in vivo, leading to penetration of the vascular endothelium to start the process of extravasation. Second, we did not compare the effects of propofol with other anesthetics on tumor metastasis, as demonstrated in another study. [9] However, the objective of the present study was to illustrate that standard-dose anesthetic could promote tumor metastasis in mice as compared to low-dose anesthetic or nonanesthesia control condition. Third, the administration of low-dose propofol (20 mg kg −1 ) was once, but the administration of standard-dose propofol (240 mg kg −1 ) was over 1 h. This was because the mouse would move around under the low-dose propofol and unable to receive 1 h infusion of propofol. Thus, we did a single injection of low-dose propofol to the mouse.
In conclusion, this study demonstrated that the commonly used anesthetic propofol promoted tumor metastasis in mouse lungs. Mechanistically, propofol could promote tumor metastasis by enhancing tumor cell adhesion and extension via interactions of GABA A R, TRIM21, and Src. Finally, inhibition of Src might mitigate the propofol-promoted tumor metastasis. These findings will likely foster more research on anesthesia and tumor metastasis in preclinical and clinical settings, ultimately leading to better patient prognosis after surgical resection of tumors.

Experimental Section
Mice: 6-week-old female BALB/c nude mice (Charles River, China) were employed. The animal protocol was approved by the Standing Committee on Animals at Shanghai Tenth People's Hospital and mice were maintained in accordance with federal guidelines in the Laboratory Animal Center at Tongji University (protocol number: SHDSYY-2018-17310101). All mice were raised in sterilized cages under pathogen-free conditions (22-26°C, 12/12 light/dark cycle) and offered food and water ad libitum. Animal suffering during euthanasia was ameliorated by using CO 2 overdose followed by cervical dislocation. Efforts were made to minimize the number of animals used in the studies. The manuscript was written according to ARRIVE guidelines. www.advancedsciencenews.com www.advancedscience.com Generation of HCT116 Cells Overexpressing Luciferase: Generation of tumor cells with stable overexpression of luciferase enables ex vivo bioluminescence imaging of tumors. Thus, the HCT116 cells were generated with stable overexpression of luciferase to visualize tumors in mouse lung. HCT116 cells were infected with pSLenti-EF1-Luc2-F2A-Puro virus (Oobio, Shanghai, China). Stable cell lines were selected by adding 1 μg mL −1 puromycin (S7417, Selleck, Houston, TX, USA) to culture medium. The cells were digested with 0.25% trypsin (Gibco, Grand Island, NY, USA). A total of 1.5 million cells were centrifugally collected and resuspended in 100 μL culture medium for the injection to mice.
Administration of Low-Dose or Standard-Dose Propofol in Mice: The effects of propofol on tumor metastasis were determined by injecting a low dose (20 mg kg −1 once, about 0.5 mg propofol in 50 μL intralipid per mouse) versus standard dose (240 mg kg −1 h −1 for 1 h, about 6 mg propofol in 600 μL intralipid per mouse over 1 h) of propofol (AstraZeneca, UK) via tail vein in nude mice (6-week-old female BALB/c nude mice). The doses of 20 or 240 mg kg −1 of propofol were chosen according to the methods described in previous studies with modifications. [44] Specifically, for mice in the low-dose propofol group, first, 1.5 million HCT116 cells and then (with 1 min apart) 20 mg kg −1 propofol were injected. The same procedure was repeated for mice in the standard-dose propofol group, but, 3 min after the administration of 20 mg kg −1 propofol, an angel catheter (size 26G, ZiBo Eastmed Healthcare Products Co., Ltd., Zibo, China) was placed in the mouse tail vein. Then, a syringe pump (WH-SP-08, Wenhao Mircofluidic Technology Co., Ltd., Suzhou, China) was used to inject propofol (240 mg kg −1 ) slowly (over 1 h). Four weeks after the injection of HCT116 cells and propofol, lung metastasis of tumor cells was detected by using ex vivo bioluminescence imaging. Tumor cells were injected before the injection of propofol to enable cells to be sufficiently exposed to propofol before propofol, which has half-life of a 2 to 4 min, [45] redistributed from blood into fat tissues. Note that low-dose propofol could not be administered over 1 h because the mouse would move around under the low-dose propofol. Thus, a single injection of low-dose propofol was performed in the mouse.
Bioluminescence Imaging: Bioluminescence imaging of mouse lungs was performed 4 weeks after intravenous injection of the tumor cells. A total of 150 mg kg −1 d-luciferin solution (Yeasen, Shanghai, China) was injected intraperitoneally in the mice under brief isoflurane anesthesia. 15 min after luciferin injection, the mice (N = 6 per group) were euthanized to harvest whole lungs for ex vivo bioluminescent imaging to evaluate lung metastasis with the BLT AniView100 multimodal animal imaging system (Biolight Biotechnology, Guangzhou, China).
Hematoxylin and Eosin (H&E) Staining: Harvested lungs were rinsed in PBS and fixed in 4% paraformaldehyde (Servicebio, Wuhan, China) overnight. Fixed lungs were embedded in paraffin, and sampled sections were taken across each lobe of lungs. Consecutive tissue (3 μm apart) sections were stained using H&E. To evaluate tumor metastasis, the total number of tumor nodules in all lung lobes per mouse (N = 6 to 9 per group) was counted on H&E-stained sections using a microscope with sample identities disguised. Specifically, one section with the biggest diameter was obtained from each lobe of the lungs and the number of nodules in the section was counted for the quantification of tumor metastasis.
Cell Culture: Human colorectal tumor cell lines HCT116 were cultured in McCoy's 5A medium (Gibco) with 10% fetal bovine serum (FBS) (Lonsera, S711-001S, Uruguay). HUVEC line, human nonsmall-cell lung carcinoma cell (NSCLC) line A549, and mouse colorectal tumor cell line CT26 were cultured in RPMI-1640 medium (Gibco) with 10% FBS. Endometrial carcinoma cell line Ishikawa was cultured in DMEM/F12 medium (Gibco) with 10% FBS. The 293FT cell line was cultured in DMEM medium (Gibco) with 10% FBS. Cells were cultured at 37°C, 5% CO 2 atmosphere. Breast tumor cell line MDA-MB-231 was cultured in Leibovitz's L15 medium (Gibco) at 37°C. All cells tested negative for mycoplasma. HCT116, CT26, and MDA-MB-231 cell lines were purchased from the cell bank of the Chinese Academy of Sciences. A549 and Ishikawa cell lines were gifts from Dr. Jiuhong Kang at Tongji University. HUVEC was a gift from Dr. Jialin Charles Zheng's lab at Tongji University. The 293FT cell line was a gift from Dr. Xiaoqing Zhang's lab at Tongji University.
Tumor Cell-HUVEC Adhesion Assay: 1.5 × 10 5 HUVECs per well were seeded in 24-well plates overnight for cell monolayer formation. A total of 5 × 10 4 tumor cells stained with Calcein AM (Yeasen) were seeded on HUVECs per well. Cells were incubated to assess the interaction of tumor cells with HUVECs. Because different tumor cell lines have different adhesion capabilities, the incubation time differed for each cell line: CT26 cells, 2 h; HCT116 cells, 6 h; MDA-MB-231 cells, 3 h; A549 cells, 2 h; and Ishikawa cells, 5 h. After incubation, cells were washed to remove nonadherent cells. Images of cells were captured using a fluorescence inverted microscope with three different visual fields per dish (six dishes per group) with sample identities disguised.
Membrane Potentiometric Analysis of HCT116 Cells: Voltage-sensitive dye bis-(1,3-dibutylbarbituric acid) trimethine oxonol [DiBAC4(3)] [46] (Sigma-Aldrich) (final concentration of 20 × 10 −6 m) and -GABA (Selleck) (final concentration of 250 × 10 −6 m) were added to HCT116 culture medium without FBS, and cells were incubated for 10 min. Green fluorescence of cells were captured using a fluorescence inverted microscope (Zeiss observer Z1, Germany). Cells were then incubated with bicuculine (Selleck) (200 × 10 −6 m) or DMSO for 5 min and green fluorescence of cells was assessed. The experiment was repeated six times, and three fields were viewed per sample. The intensity of the fluorescence in each sample was determined by using ImageJ (version 8.0, NIH, Bethesda, MD).
KO of GABA A R Subunit 3 (GABRB3) and (GABRD) in HCT116 Cells: A vector with clustered regularly interspaced short palindromic repeats