GLDC promotes colorectal cancer metastasis through epithelial–mesenchymal transition mediated by Hippo signaling pathway

Cancer metastasis remains a major cause of death in cancer patients, and epithelial–mesenchymal transition (EMT) plays a decisive role in cancer metastasis. Recently, abnormal expression of Glycine Decarboxylase (GLDC) has been demonstrated in tumor progression, and GLDC is up-regulated in cancers, such as lung, prostate, bladder, and cervical cancers. However, the exact role of GLDC in colorectal cancer (CRC) progression remains to be elucidated. The aim of our study was to explore the role of GLDC in CRC metastasis. The GSE75117 database was used to investigate GLDC expression in tumor center and invasive front tissues and we found that GLDC expression levels were higher in the invasive front tissue. GLDC expression levels were negatively correlated with the prognosis of CRC patients. In vitro studies have showed that GLDC can promote invasion and migration of CRC cells by inhibiting the Hippo signaling pathway and regulating the EMT process. Blocking the Hippo signaling pathway with Verteporfin reduced the effect of GLDC on CRC metastasis. In vivo metastasis assays further confirmed that tail vein injection of GLDC+/+ cells induced more lung metastasis, compared to normal CRC cells. The results of this study suggest that GLDC promotes EMT through the Hippo signaling pathway, providing a new therapeutic target for CRC metastasis.


Introduction
Colorectal cancer (CRC) has the third highest incidence and the second highest mortality rate among cancers worldwide [1]. It was estimated that by 2030, there will be approximately 2.2 million new cases of CRC and 1.1 million deaths from CRC worldwide [2]. Surgery is the preferred treatment for CRC, but approximately 50% of CRC patients develop metastases within 2 years after radical resection, and the survival rate for patients with CRC metastases is only 10% [3].
Glycine Decarboxylase (GLDC) gene is located on the short arm of chromosome 9 (9p24.1). Its main role was first found to break down the glycine into S-aminomethyl dihydrofatty acyl protein. Recently, aberrant expression of GLDC has been implicated in the development of multiple types of tumors [4][5][6]. Zhang WC et al. reported that GLDC was up-regulated in lung, prostate, bladder, and cervical cancers, and lung cancer patients with higher GLDC expression had worse prognosis [7]. In addition, Li X et al. reported that GLDC was able to promote the proliferation of p53-mutated B lymphoma cells [6]. GLDC can regulate biological processes through its metabolites or affecting the expression of other genes [4,8]. Go MK et al. reported that GLDC promotes tumor formation by synthesizing serine through the decarboxylation of glycine [8]. Woo CC et al. found that GLDC promotes the metabolism of pyruvate and lactate, thus promoting the development of liver, prostate, cervical, and lung cancers [9]. In addition, GLDC can be involved in disease formation by regulating gene expression. Wei HY Hao Yu and Xueqing Hu have contributed equally to this work. et al. found that GLDC can up-regulate DNA methyltransferases (DNMTs) and promote lung cancer metastasis [4].
The study of Hippo signaling pathway began with the discovery of the tumor suppressor gene-"Wts" in Drosophila in 1995. In mammals, the core components of the Hippo signaling pathway include MST1/2, Salvador family WW domain protein 1 (SAV1), LATS1/2 (Wts in Drosophila), and MOB kinase activator 1 (MOB1), where SAV1 and MOB1 are adapter proteins that bind to and enhance the phosphorylation of MST1/2 and LATS1/2, respectively. YAP is a major effector molecule downstream of this pathway and can be directly phosphorylated by LATS1/2. P-YAP binds to proteins in the cytoplasm and is subsequently ubiquitinated and degraded, thereby inhibiting the growth-promoting, anti-apoptotic function of YAP. The Hippo signaling pathway negatively regulates YAP activity through this cascade of phosphorylation reactions. In contrast, when the Hippo signaling pathway is inhibited, YAP can be transported into the nucleus to bind to transcription factors such as TEADs to promote expression of target genes. Hippo signaling pathway is mainly responsible for regulating biological processes, such as cell proliferation, differentiation, aging, and organ growth [10,11]. Recent studies have shown that Hippo signaling pathway plays an important role in regulating cancer development and stem cell function. Inactivation of the Hippo signaling pathway can promote metastasis in breast, lung, liver, and other tumors [12][13][14]. Guo PD et al. reported that inhibition of the Hippo signaling pathway was able to promote CRC metastasis [15]. Hippo signaling pathway is regulated by various signals, and the expression of MST1 can be regulated by apoptotic and stress stimuli, such as Staurosporine, hydrogen peroxide, retinoic acid, and some anticancer drugs [16][17][18]. Guo Z et al. reported that DNMT1 can up-regulate MST1 expression in glioma cells [19]. Kuser-Abali G et al. reported that MYC and EZH2 could co-regulate MST1 expression in prostate cancer cells [20]. The upstream regulatory mechanisms of Hippo signaling pathway in CRC have not been elucidated. Those above studies suggest that Hippo signaling pathway is regulated by GLDC and that the Hippo signaling pathway is a potential pathway for GLDC to promote CRC metastasis.
In this paper, we investigated the biological function of GLDC-mediated Hippo signaling pathway in EMT and metastasis of CRC, providing a potential therapeutic target for CRC progression.

Data processing methods
The GEO database, established by the National Center for Biotechnology Information (NCBI), is a gene expression database and online genome resource that collects highthroughput gene expression data uploaded by research institutes around the world. "Colorectal cancer" was entered as search objects. The gene expression microarray datasets GSE75117 was selected and downloaded, which was based on the GPL16699 platform. GSE75117 includes 133 multiples spatially separated samples from 8 patients with metastatic CRC and 8 patients with non-metastatic CRC. The Cancer Genome Atlas (TCGA) system was first launched in 2006 by the National Human Genome Research Institute and the National Cancer Institute to map cancer genes, in order to understand potential pathways of cancer and improve the ability to inhibit cancer progression and accurately diagnose and cure cancer. Genome sequences from 380 CRC patients and 51 normal colorectal tissues were downloaded from TCGA database for correlation analysis.

Plasmid construction and cell transfection
The chemically synthesized human GLDC (Gene ID: 2731) gene was used in this study. The GLDC +/+ sequence was subcloned into the vector (PGMLV-CMV-MCS-3 × Flag-PGK-Puro, Genomeditech (Shanghai) Co). Control and GLDC +/+ lentiviruses were transfected into SW620 and Lovo expression. Before lentiviral transfection, SW620 and Lovo cells were inoculated into 6-well plates and placed in an incubator at 37 °C overnight. After cells were adhered, control lentivirus and GLDC +/+ lentivirus were added into wells. 12 h after transfection, the medium containing lentivirus was removed and fresh medium was added.

Cell wound healing assay
When the cell growth density reached 90% or more and the growth status was good, the bottom of the culture dish was scratched with the tip and the detached cells were gently washed away with PBS 3 times. Verteporfin and 1% fetal bovine serum were added to the drug-treated group, and only 1% fetal bovine serum was added to the control cells and placed in an incubator. Pictures were taken at the same location of the scratch at 0 h, 24 h, and 48 h to observe cell wound healing. Each group of experiments was repeated three times. Scratch area was calculated and statistically analyzed with ImageJ software.

Cell migration assay
Cells in a logarithmic growth phase were prepared into single-cell suspension by adding 0.25% trypsin to digest the cells. The cells were diluted with serum-free medium so that the final concentration of the cell suspension was 2 × 10 5 cells/mL. Matrigel was diluted with culture medium and spread on the bottom of the test chamber. After polymerization, the basement membrane is hydrated. Transwell inserts were placed into 24-well plates, 600-μL complete serum medium or complete serum medium containing Verteporfin was added to the lower chamber, and 200 μL cell suspension was added to the upper chamber. After incubation for 48 h, migrated cells were analyzed after paraformaldehyde fixation and crystal violet staining, and five views were randomly selected, counted, and pictured with an inverted microscope at 300X magnification.

Western blot assay
After 48 h of cell culture in each group, the medium was discarded, and the cells were washed three times with PBS. Total protein, cytoplasmatic protein and nuclear protein were extracted according to the procedures of NucleoProtein Extraction Kit (Solarbio, China). The extracted proteins were quantified with BCA and added into SDS-PAGE gel electrophoresis system. Then the proteins in the gel were transferred to PVDF membranes and blocked in 10% BSA. Next, PVDF membranes were incubated with primary antibodies overnight at 4 °C, washed three times with TBST, and incubated with HRP-labeled goat anti-rabbit IgG for 2 h at room temperature. Visualization was performed by enhanced chemiluminescence, direct photography, and quantification.

Quantitative real-time PCR
Total RNA was extracted with Trizol (Beyotime) according to the instructions. cDNA was synthesized with HiScript III RT SuperMix (Vazyme) using 500 ng of total RNA as a template. ChamQ SYBR qPCR Master Mix (Vazyme) was used for real-time PCR analysis. GAPDH was used as a loading control to detect the relative mRNA expression of GLDC. Real-time PCR results were defined by the threshold cycle (Ct), and relative expression was calculated using the 2 −ΔΔCt method. PCR was performed with ABI 7500 instrument (Applied Biosystems, USA). Primers for GAPDH (forward primer: 5-TCG GAG TCA ACG GAT TTG GT -3, reverse primer: 5-TTC CCG TTC TCA GCC TTG AC -3) and GLDC (forward primer: 5-GGC CCA TCG GAG TGA AGA AA-3, reverse primer: 5-TAT CGC AGT TTC CGT GGC TT-3) were used.

Immunofluorescence staining
Sterilized slides were place the into a 24-well plate, and cells in a logarithmic growth phase were seeded into the slides at the density of 4 × 10 4 cells per well. After cell attached, the culture medium was discarded, and cells were fixed in methanol and were permeabilized with 0.1% Triton-X-100 and blocked with 5% BSA. The cells were first conjugated with YAP mouse antibody and then stained with Alexa Fluor 488-conjugated goat anti-mouse IgG (Beyotime). Nuclear staining was performed with 4′, 6-diamidino-2-phenylindole (DAPI) solution. Cell imaging was performed with a Leica SP-8 laser scanning confocal microscope (Leica, Germany).

In vivo study
All animal experimental studies were approved by the Animal Ethics Committee of Shanghai University of Traditional Chinese Medicine (PZSHUTCM210913009). Sixweek old male BALB/c nude mice were purchased from SHANGHAI SLAC LABORATORY ANIMAL CO. All mice were housed at 24-25 °C and 60-65% humidity in cages with a 12-h light/dark cycle and were free to water and food. After 1 week of habitual feeding, cancer metastasis model was established in nude mice by tail vein injection of CRC cells in control groups and GLDC +/+ groups. Mice in the GLDC +/+ + VP group received intraperitoneal injection of verteporfin (100 mg/kg) every 3 days for 3 weeks. Nude mice were weighed every 4 days. Nude mice were euthanized 36 days after injection, and the number of metastatic nodules was counted and metastatic tumor tissues were collected.

Hematoxylin and eosin (HE) and immunohistochemistry staining analysis
Tissues from nude mice were fixed in 4% paraformaldehyde for 48 h, dehydrated in alcohol, and embedded in paraffin. Tissues were sectioned and stained with routine HE, and tissue slides were air-dried and scanned and photographed with a Jiangfeng digital pathological section scanner. Following the HE staining procedure described above, tissues from nude mice were embedded, sectioned, and de-paraffinized. Next, tissues were processed by adding 3% H 2 O 2 to de-paraffinized hydrated sections and incubated at room temperature with YAP antibody and the corresponding secondary antibody. Slides were visualized using the DAB baseplate toolkit (Vector Laboratories, Burlingame, CA, USA). Those slides were visualized with DAB staining solution, nuclei were stained with hematoxylin solution, and slides were finally loaded into neutral gel. The expression of YAP in tissues was observed under a light microscope.

Statistical analysis
Statistical analysis was performed using SPSS software (version 26.0, USA) and GraphPad Prism 9 (GraphPad Software, USA). Data were expressed as Mean ± SD, and one-way ANOVA was used for comparison between multiple groups, and Student's t test was used to calculate the difference between two groups. All experiments were independently repeated at least three times, and P < 0.05 indicates statistical difference.

GLDC-mediated Hippo signaling pathway is a potential mechanism for CRC metastasis
To investigate the relationship between GLDC-mediated Hippo signaling pathway and CRC metastasis, we analyzed the expression of GLDC and genes regulated by Hippo signaling pathway in CRC tissues using GSE75117 dataset. There were 27 samples in the dataset, including 15 invasive front samples and 12 tumor center samples. Data analysis showed that compared with central CRC tissues, the expression levels of GLDC, CTGF, and CYR61 were higher in the invasive front tissues (Fig. 1A). Three hundred and eighty CRC tissues and 51 normal colorectal tissues were downloaded from the TCGA database to analyze the correlation between GLDC expression and the expression levels of ANKRD1, CTGF, and CYR61, which are regulated by Hippo signaling pathway. The statistical results showed correlation between GLDC expression and the expression levels of CTGF (R = 0.20, P < 0.005; Fig. 1B), ANKRD1 (R = 0.14 P < 0.01; Fig. 1C), and CYR61 (R = 0.22, P < 0.005; Fig. 1D). MUC1 is a signature gene of epithelial cells, and FN1, CDH11, CDH2, and VIM were signature genes of mesenchymal cells, while CTNNB1, MMP2, and CDH1 were important signature genes of EMT. Analysis showed that the expression levels of GLDC, CTGF, and CYR61 were negatively correlated with MUC1 expression and positively correlated with expressions of FN1, CDH11, CDH2, VIM, CTNNB1, and MMP2. All of these suggest that GLDC may be involved in the EMT process in CRC cells through Hippo signaling pathway (Fig. 1E).

GLDC promotes EMT in CRC cells
Bioinformatics analysis has showed that GLDC is closely associated with CRC metastasis, we next explore the role of GLDC in CRC metastasis in vitro. First, we examined GLDC expression in NCM460 and 7 CRC cell lines, and GLDC was highly expressed in SW620 cells and Lovo cells compared with other cells (Fig. 2A). To elucidate the function of GLDC, we overexpressed the GLDC gene in CRC cells. As confirmed by RT-qPCR data, GLDC mRNA expression was significantly up-regulated in the GLDC +/+ group, compared with the empty vector group. Western blot assay was  (Fig. 2B), and the results were consistent with the RT-qPCR results. We observe the effect of GLDC on CRC cell migration by transwell assay and wound healing assay. Transwell analysis showed that GLDC +/+ group could significantly promote the migration of SW620 and Lovo cells compared with empty vector group (Fig. 2C). The results of the wound healing assay were similar to those of the transwell assay. Migration of SW620 cells and Lovo cells was stronger in GLDC +/+ group than that in the empty vector group (Fig. 2D). EMT is a critical step in the invasion and migration of CRC. When EMT occurs, in order to gain motility and invasion, many epithelial phenotypes are discarded, and a series of significant changes will occur as cancer cells detached from the epithelial layer. During this process, expression of E-cadherin, a marker of epithelial cells, is suppressed, initiating the expression of Vimentin, a component of the mesenchymal cytoskeleton. Meanwhile, N-cadherin replaces E-cadherin protein. Activation of Vimentin regulates organelle and membrane-associated proteins and promotes cell motility. Therefore, E-cadherin, N-cadherin, Vimentin, and Snail are widely considered as the markers of EMT. Our data showed a significant decreased level of E-cadherin and increased levels of N-cadherin, Snail, and Vimentin in SW620 or Lovo cells in GLDC +/+ group compared to empty vector group (Fig. 2E). All results indicated that GLDC could promote EMT process, as well as invasion and metastasis of CRC cells.

GLDC inhibits Hippo signaling in CRC
A series of studies have shown that Hippo signaling pathway is involved in tumor invasion and metastasis. To explain the role of GLDC in YAP expression and localization, we examined YAP protein localization in CRC cells after overexpression of GLDC by immunofluorescence. The results showed that GLDC increased the nuclear translocation of YAP in CRC cells (Fig. 3A). To further understand the mechanism of GLDC in the Hippo signaling pathway, this study investigated the regulatory effect of GLDC on the expression of Hippo-related proteins, such as p-MST1, MST1, YAP, p-YAP. As shown in the data, YAP expression levels were higher in the GLDC + / + group compared to the empty vector group, while levels of p-MST1 and p-YAP were lower in total proteins of CRC cells (Fig. 3B). Next, we examined YAP expression in the cytoplasm and nucleus. As shown in Fig. 3C, YAP expression in the cytoplasm was decreased and p-YAP expression was decreased after GLDC expression, while YAP expression in the nucleus was increased. The above study indicates that GLDC can inhibit Hippo signaling pathway and promote translocation of YAP protein to the nucleus of CRC cells.

GLDC inhibits EMT in CRC cells through the Hippo signaling pathway
To verify that GLDC promotes EMT progression through the Hippo signaling pathway, we observed the effect of Verteporfin, a YAP inhibitor that disrupts the YAP-TEAD interaction, on CRC cell migration by transwell and wound heanling assays. Transwell analysis showed that GLDC +/+ significantly promoted the invasion of SW620 and Lovo cells, but the intervention of Verteporfin greatly reduced this effect (Fig. 4A). Similar results were obtained in the wound healing assay, where migration was significantly attenuated after Verteporfin intervention compared with the GLDC +/+ group (Fig. 4B). GLDC +/+ SW620 and Lovo cells were treated with or without Verteporfin for 48 h. EMT-related protein levels were measured by Western blot assay. The results showed that GLDC +/+ promoted the progression of EMT compared with empty vector group, but this effect was greatly attenuated after Verteporfin blocked the YAP-TEAD interaction (Fig. 4C). In conclusion, GLDC exerted an inhibitory effect on the EMT progression by activating the Hippo signaling pathway in CRC cells.

GLDC regulates CRC metastasis in Vivo
To validate the role of GLDC in CRC metastasis in vivo, we established a metastasis model in nude mice by tail intravenous injection and then divided them into three groups, including empty vector group, GLDC +/+ group, and GLDC +/+ + VP group (Fig. 5A). There was no statistical difference in weight among the three groups (Fig. 5B). More lung metastatic sites were observed in the GLDC +/+ group and fewer lung metastases in GLDC +/+ + VP group (Fig. 5C). The metastatic lesions were then excised and examined by HE staining. According to the imaging results, there were more lung metastases and tumores were larger in the GLDC +/+ group compared with empty vector group, but the metastases were significantly improved in GLDC +/+ + VP group (Fig. 5D). The expression level of YAP was detected by immunohistochemistry. YAP reached a higher level in the GLDC +/+ group compared to the empty vector group (Fig. 5E). Taken together, the results of this study suggest that GLDC can promote the lung metastasis of CRC by regulating the Hippo signaling pathway.

Discussion
In recent years, an increasing number of studies have shown that GLDC plays an important role in cancer progression, but its exact mechanism in CRC remains unclear. In our study, the data showed that GLDC inhibited the activation of Glycine is a non-essential amino acid and is one of the important metabolites of many proteins. High levels of glycine are toxic through conversion to metabolites such as aminoacetone and methylglyoxal. Various studies have demonstrated that glycine metabolism is essential for tumorigenesis [21][22][23]. Glycine Cleavage System (GCS) controls glycine decarboxylation and deamination through a multistage reaction to generate CO 2 , NH 3 , NADH, and 5,10-methylene-THF [24,25]. GCS is a multi-enzyme complex, consisting of glycine decarboxylase (GLDC, also called P protein), amino-methyltransferase (T protein), dihydrolipoamide dehydrogenase (L protein), and the hydrogen carrier protein (H protein) [26]. GLDC is an oxidoreductase that is mainly involved in intracellular amino acid metabolic process and affects various metabolic pathways, such as glycolysis and pyrimidine synthesis. It has also been shown that GLDC is highly active in different cancer cells and plays an important role in cancer metastasis. For example, Woo CC et al. found that GLDC was able to promote metabolism of pyruvate and lactate, thereby promoting tumor development including liver, prostate, cervical, and lung cancers.
To observe the role of GLDC in CRC, we first downloaded the GSE75117 dataset from GEO database, which included 15 invasive front samples and 12 tumor center tissues. Data analysis showed that the expression levels of GLDC, CTGF, and CYR61 were higher in invasive front, compared with tumor center of CRC. Then the correlation between the expression of GLDC and ANKRD1, CTGF, and CYR61 regulated by Hippo signaling pathway was analyzed by TCGA database. The results showed that GLDC showed a linear relationship with the expression levels of CTGF, ANKRD1, and CYR61 in 380 CRC cases. The expression of genes regulated by Hippo signaling pathway was negatively correlated with the expression of maker genes of epithelial cells and positively correlated with marker genes of mesenchymal cells and other signature genes for EMT, indicating that GLDC may be involved in the EMT process of CRC through Hippo signaling pathway. To test this conjecture, we observed the effect of GLDC on CRC cell invasion and migration. The data showed that GLDC could promote the invasion and migration of CRC cells and further studies showed that GLDC could significantly promote the EMT process of CRC cells, as shown by the down-regulation of E-cadherin and up-regulation of N-cadherin, Vimentin, and Snail expression. Therefore, the results of the above in vitro experiments showed that GLDC can promote the invasion and migration of CRC cells by promoting the EMT process, which was similar to previous results examining the important role of GLDC in lung, prostate, bladder, and cervical cancer. Overexpressing GLDC in CRC cells line with low GLDC expression, e.g., NVM460, HT-29, and Caco-2 are required to confirm the results.
We then considered the potential mechanism by which GLDC regulates EMT and metastasis in CRC. The Hippo signaling pathway effector protein YAP promotes the expression of target genes by entering the nucleus and binding to transcription factors, such as TEADs, triggering the transcription of their downstream target genes and promoting the development and progression of various cancers. The interaction between GLDC and the Hippo signaling pathway increases the entry of YAP into the nucleus. Therefore, we further investigated the effect of GLDC on the Hippo signaling pathway. The data showed that GLDC inhibited the Hippo signaling pathway by decreasing the phosphorylation level of MST1, and by increasing YAP accumulation in the nucleus. We then tested this mechanism by blocking the nuclear binding of YAP to TEADs and found that Verteporfin significantly reduced the CRC metastasis and EMT by GLDC, suggesting an important role of GLDC in promoting EMT in CRC progression partly through the Hippo signaling pathway. Taken together, our results suggest that GLDC has an important role in the regulating EMT in CRC cells and that the GLDC-mediated Hippo signaling pathway is an important mechanism of CRC metastasis.

Conclusion
In this study, we found that GLDC can inhibit the activation of Hippo signaling pathway and promote the entry of YAP protein into the nucleus, which in turn promotes the EMT process of CRC and promote its metastasis. Therefore, GLDC is a potential target for CRC metastasis, which provides theoretical support for the subsequent clinical prevention and treatment of CRC metastasis.