BCKDK Promotes EOC Proliferation and Migration By Activating The MEK/ERK Signaling Pathway

Background: Ovarian cancer is the most fatal gynecologic cancer, and epithelial ovarian cancer (EOC) is the most common type. The branched-chain α-keto acid dehydrogenase kinase (BCKDK) plays an important role in many serious human diseases, including cancers. Its function in promoting cell proliferation and migration has been reported in various cancers. However, the biological role of BCKDK and its molecular mechanisms underlying EOC initiation and progression are unclear. Methods: First, the expression level of BCKDK in EOC cell lines or tissues was determined using tissue microarray (TMA)-based immunohistochemistry or western blotting. Then, growth curve analysis, anchorage-independent cell transformation assays, wound healing assays, cell migration assays, and tumor xenografts were used to test whether BCKDK could promote cell transformation or metastasis. Finally, the signaling pathways involved in this process were investigated by western blotting or immunoprecipitation. Results: We found that the expression of BCKDK was upregulated in EOC tissues and that high expression of BCKDK was correlated with an advanced pathological grade in patients. The ectopic overexpression of BCKDK promoted the proliferation and migration of EOC cells, and the knockdown of BCKDK with shRNAs inhibited the proliferation and migration of EOC ex vivo and in vivo. Moreover, BCKDK promoted EOC proliferation and migration by targeting MEK. Conclusions: Our results demonstrate that BCKDK promotes EOC proliferation and migration by activating the MEK/ERK signaling pathway. Targeting the BCKDK-MEK axis may provide a new therapeutic strategy for treating patients with EOC. clean swab. Then cells on the bottom side of the membrane were xed with 4% paraformaldehyde for 30 min, then stained with 0.1% of crystal violet (Sangon Biotech) for 15 min. Finally, the stained cells were counted by microscopy. Results represent the average number of cells in three elds per membrane in triplicate inserts.


Background
Ovarian cancer (OC) is the second most common malignancy after breast cancer in women over the age of 40 1

.
However, it is the most fatal gynecologic cancer 2 . Because it remains a challenge to diagnose early. There are three main types of ovarian cancer: epithelial, germ cell, and sex-cord stromal. Epithelial ovarian cancer (EOC) is the most common type, comprising about 95% of all ovarian cancers 2,3 . About 75% of patients are diagnosed at an advanced stage because of the asymptomatic nature of EOC.
Endometriosis has been linked to some EOC. However, no evidence shows that removal of endometriosis lesions will decrease a woman's chances of developing OC 4 . The biomarker, such as serum cancer antigen 125 (CA125) or human epididymis protein 4, combined with transvaginal sonography was tested for some OC. A randomized controlled trial of over 200000 women assessing annual multimodal screening with CA125, did not identify signi cant mortality reduction when the risk for ovarian cancer algorithm was used, versus annual transvaginal ultrasound screening, versus no screening 5 . Therefore, over the last 30 years, mortality rates from OC, including EOC, have narrowly dropped 1,6 . On the other hand, PARP1 inhibitors (Niraparib, Olaparib and Rucaparib) maintenance therapy was recently shown to substantially improve progression-free survival in ovarian cancer patients [7][8][9][10] . However, not all patients respond to PARP inhibitor therapy, either due to intrinsic or acquired resistance to PARP inhibitors [10][11][12] . Therefore, the development of alternative synthetic lethality targets is urgently needed in EOC.
Branched-chain α-keto acid dehydrogenase kinase (BCKDK) located in the mitochondrial matrix, belonging to pyruvate dehydrogenase kinases (PDKs) family 13 , which promoted the proliferation and metastasis of various tumors and was considered to be a strong therapeutic target for preventing tumors development [14][15][16][17][18][19] . Dysfunction of BCKDK was closely related to various human diseases, especially maple syrup urine disease. Bravo-Alonso and Oyarzabal found the excessive function of BCKDK resulted in maple syrup urine disease [20][21][22] . As diabetics had increased susceptibility to ovarian cancer, always divided into late stages when the rst diagnosis and had a poor prognosis 23 . Over-expression of BCKDK resulted in branched-chain amino acid (BCAA) increase, elevated plasma levels of BCAA were associated with a greater than 2-fold increased risk of future pancreatic cancer diagnosis 24 .
Leu promoted adipose tissue protein synthesis through mTOR pathway 25 , and then adipocytes promoted ovarian cancer metastasis and provided energy for rapid tumor growth 26 . BCKDK promoted colorectal cancer and hepatocellular carcinoma metastasis and proliferation via the ERK signaling pathway [27][28][29] . Furthermore, BCKDK was highly expressed in DOX-induced ovarian cancer drug-resistant cell lines, and its expression level was twice as high as that of sensitive one 30 . Inhibition of BCKDK increased the sensitivity of ovarian cancer cells to paclitaxel 31 .
We can't help but wonder if BCKDK could promote ovarian cancer proliferation and metastasis? Which pathways worked in this process?
In this study, We showed that the BCKDK had a higher expression in EOC patients versus normal tissues. The high expression of BCKDK was correlated with the advanced pathological grade for patients. Overexpression of BCKDK increased the clone formation and migration ability of SKOV3 and OVCAR3 cells ex vivo. Knockdown of BCKDK inhibited EOC tumor progression ex vivo and in vivo. And we identi ed BCKDK as an upstream kinase of MEK, which up-regulated MEK/ERK signaling by interacting with MEK. The above results suggest that BCKDK may promote EOC proliferation and migration through enhancing the MEK/ERK signaling pathway. and anti-sense oligonucleotides were synthesized, annealed and cloned into the pLKO.1-TRC cloning vector at the EcoRI and AgeI sites as described by the manufacturer 32 .

Western blot
Cells (0.8-2×10 6 ) were cultured in 10 cm diameter dishes to 70-80% con uence, and then starved no serum for 24h. Then the cells were treated with 40 ng/mL epidermal growth factor (EGF) (R&D catalog: 236-EG-200) for 15min. EGF is a well-known tumor promotion agent used to study malignant cell transformation in animal and cell models of cancer 33 . After this, cells were washed twice in PBS before being lysed in RIPA buffer (Coolaber, China). Then, samples were sonicated in 15 seconds intervals three times, and insoluble debris was removed by centrifugation at 13000 rpm for 15 min. Protein content was determined by BCA method (Coolaber, China). 30-120 µg of protein was separated by 10% SDS-PAGE and visualized by chemiluminescence (BIO-RAD, USA) in triplicate.
Growth curve analysis 2×10 5 cells were plated in each dish and counted at different times in triplicate, using a hemacytometer to generate a growth curve.
Tumor xenografts and the tissue microarray (TMA) assay Female athymic Balb/c nude mice (4-6-week-old) were purchased from Chongqing Tengxin Beer Experimental Animal Sales CO, LTD ( Chongqing, China). Mice were kept in speci c pathogen-free conditions according to the National Guidelines for Animal Usage in Research (set by the Chinese government) at the Chongqing Population and Family Planning Science and Technology Research Institute. Mice were divided and randomized into three groups. Each of the different cell lines (3×10 6 in 200 µl PBS) was injected subcutaneously into the right ank. The tumor volumes were measured every three day and were calculated with the formula: V = 0.52 (length ×width × height). The tumor tissues were prepared with para n sections after xation with formalin, and then stained with hematoxylin and eosin (H&E).
The TMA underwent EOC (catalog: FOV 1006) were purchased from Xi'an Elena Bio Co., Ltd. Samples were obtained with informed consent. The TMA was stained with anti-BCKDK. The immuno-scores were calculated following the Remmele score method (Regitnig et al., 2002), and the scores greater than 2 were used as positive group, the others were used as negative group.

Wound healing assay
The wound healing assays were applied to determine the migration ability of cells. 2×10 5 cells were cultured in a six-well plate until 80-90% con uence and then carefully scratched with a 10 µL pipette tip. After washing three times with 1×PBS to remove detached cells, images in 10 different wound elds were captured at respective time points (0 h and 48 h) to evaluate the migration of cells.
Cell migration assay (trans-well assay) Chambers (catalog: 3422, 8µm pore, Corning, NY, USA) were used to investigate migration ability of cells. 1× 10 5 cells suspended in 150 µL serum-free medium were seeded onto the upper chamber of 24-well plates, and 700 µL of medium with 10% FBS was added to the lower chamber. 48 h later, the medium was removed from the upper chamber. The non-migrating cells on the upper side of the chamber were removed thoroughly with a clean cotton swab. Then cells on the bottom side of the membrane were xed with 4% paraformaldehyde for 30 min, then stained with 0.1% of crystal violet (Sangon Biotech) for 15 min. Finally, the stained cells were counted by microscopy. Results represent the average number of cells in three elds per membrane in triplicate inserts. Immunoprecipitation HO8910-PM were seeded in 10cm dishes for 24h. Then, cells were harvested in IP buffer (150mM NaCl, 50mM tris-HCl pH 7.4, 1% NP40, 1mM DTT and 1mM EDTA). 2mg proteins were subjected to immunoprecipitation following the manufacturer's instructions. (Http://www.scbt.com/protocols.html?protocol=immunop recipitation). The mouse source antibody was used for IP and the rabbit source antibody was used for western blotting.

Statistical analysis
All quantitative data in the present study were performed at least in triplicate. The results are expressed as the mean ± standard deviation. A two-tailed ANOVA or Student's t-test was used to evaluate the data. Correlation data were determined by using Pearson correlation coe cients. And P < 0.05 was considered signi cant (*p < 0.05, ** p < 0.01, ***p < 0.001).
Results BCKDK over-expression is associated with advanced pathological grade in EOC patients.
First, BCKDK expression level was analyzed in 5 EOC cell lines and 1 normal ovarian epithelial cell line (Fig. 1a). The results showed that BCKDK was highly expressed in HO8910-PM and SW626 cells, moderately expressed in HO8910 cells, and poorly expressed in SKOV3 and OVCAR3 cells. The BCKDK level of normal IOSE80 cells was the lowest. Then, the expression level of BCKDK was analyzed in EOC tissue and corresponding tumor adjacent tissue samples. The results indicated that expression level of BCKDK was higher in cancer tissue than adjacent tissue (Fig. 1b, 1c), and is correlated with advanced pathological grade for patients (Fig. 1e). Detailed information on patients is shown in Fig. 1d.

BCKDK promotes EOC cell proliferation
To test whether BCKDK can promote cell transformation, BCKDK was overexpressed in SKOV3 and OVCAR3 cells which poorly expressed BCKDK. SKOV3 and OVCAR3 stable cell lines that overexpressed the pCMV-c-Flag or pCMV-BCKDK-Flag were generated, and the growth curves of SKOV3-Mock and SKOV3-BCKDK cells, or OVCAR3-Mock and OVCAR3-BCKDK cells were compared. The results showed that SKOV3-BCKDK cells grew faster than SKOV3-Mock cells (Fig. 2a, inner session indicating BCKDK overexpression). Next, the anchorage-independent growth of SKOV3-Mock or SKOV3-BCKDK was compared, and the result indicated that the number of colonies in SKOV3-BCKDK cell cultures was much more than that in SKOV3-Mock cell cultures ( Fig. 2c left panel). The corresponding statistical signi cance is shown in the right panel of Fig. 2c. Similar results were observed in the cultures of OVCAR3-Mock or OVCAR3-BCKDK stable cells (Fig. 2b, 2d). These results indicate that BCKDK promotes cell proliferation.

BCKDK promotes EOC cell migration
Since BCKDK is tightly linked to tumor migration in colorectal caner 28 , we wondered whether BCKDK also regulated EOC metastasis. To test this speculation, wound healing cell migration and transwell assays were performed to investigate the effects of BCKDK on the migration of SKOV3 and OVCAR3 cells. The outcomes are shown in Fig. 3a demonstrating that BCKDK could accelerate the migration and invasion of EOC cells. These results suggest that BCKDK in uences EOC metastasis. Wound healing assays indicated that over-expression of BCKDK accelerate the migration of SKOV3 and OVCAR3 cells (Fig. 3a, 3b). In addition, transwell assays showed that over-expression of BCKDK enhanced the invasive ability of SKOV3 and OVCAR3 cells (Fig. 3c, 3d).

Knockdown of BCKDK attenuates EOC tumor properties
To verify this idea further, BCKDK was knocked down in HO8910-PM EOC cells to generate the stable shBCKDK cell lines and the stable shMock cell lines (HO8910-PM-shBCKDK, HO8910-PM-shMock). As shown in Fig. 4a inner session of left panel by the result of Western blot, BCKDK was knocked down by shRNA sequence for lines 2 and 4.
Growth curves of HO8910-PM-shMock, -shBCKDK2, or -shBCKDK4 cells were tested, and the results demonstrated that HO8910-PM-shBCKDK cells grew dramatically slower than HO8910-PM-shMock cells (Fig. 4a). Next, the anchorage-independent growth of the HO8910-PM-shMock or HO8910-PM-shBCKDK cell lines was evaluated, and the results indicated that the number of colonies in HO8910-PM-shBCKDK cell cultures was much less than in HO8910-PM-shMock cell cultures (Fig. 4b). And wound healing assays and transwell assays of the HO8910-PM-shMock or HO8910-PM-shBCKDK cell lines were also performed, and the results indicated that knockdown of BCKDK attenuated the migration and invasion of HO8910-PM cells (Fig. 4c, 4d). Therefore, these results indicated that knockdown of BCKDK in ovarian tumor cells inhibited tumorigenesis and metastasis ex vivo. Then, tumor xenograft assays were preformed in female athymic Balb/c nude mice. We injected HO8910-PM-shMock, -shBCKDK2, or -shBCKDK4 cells (3×10 6 ) subcutaneously into the right ank, with tumor size assessed over time.
Tumors in HO8910-PM-shMock-inoculated mice grew to a much larger size compared to those in HO8910-PM-shBCKDK-inoculated mice (Fig. 5a, 5b). The tumor growth curve was shown in Fig. 5c. The nal weight of tumor was shown in Fig. 5d. And the tumor tissues dissected from the xenografts in the study were stained with hematoxylin & eosin (H&E) to con rm these tissues belong to tumor tissues (Fig. 5e). These data demonstrated that blocking BCKDK expression in EOC cells signi cantly reduces their tumorigenic properties ex vivo and in vivo, and further con rmed that BCKDK promotes cell proliferation and migration.
BCKDK promotes tumor properties through up-regulating the MEK-ERK signaling pathway We con rmed that BCKDK promoted cell proliferation and migration ex vivo and in vivo. Next, we wanted to know which signaling pathway was involved in this process. BCKDK promotes carcinogenesis has been reported in coloreactal cancer and hepatocellular carcinoma 27,29 . Therefore, p-MEK1/2 (ser221) and p-ERK1/2 (T202/Y204) levels were tested in OVCAR3 and SKOV3 BCKDK stable cell lines, and the results indicated that the level of p-MEK and p-ERK were up-regulated when BCKDK was overexpressed (Fig. 6a, 6b). These results suggest that BCKDK promotes EOC through up-regulating MEK-ERK activity. Next, the level of p-MEK and p-ERK were tested in the HO8910-PM-shBCKDK cell lines. Both p-MEK and p-ERK were decreased signi cantly when BCKDK was knocked down in HO8910-PM cell lines (Fig. 6c). Next, the expression level of p-MEK1/2 (Ser221), p-ERK1/2(T202/Y204), and BCKDK were detected in dissected tumor tissues by Western Blotting. The results con rmed that the level of p-MEK1/2 (Ser221), p-ERK1/2(T202/Y204), and BCKDK were higher in the tumor tissue of HO8910-PM-shMock mice than in the tumor tissue of HO8910-PM-shBCKDK mice (Fig. 6d). These data con rmed that BCKDK promoted tumor properties through up-regulating the MEK-ERK signaling pathway.

BCKDK interacts with MEK
We found BCKDK directly interacted with MEK in colerectal cancer cells in before research 27 . Therefore, We wounder whether BCKDK also interacts with MEK in EOC cells? Then, BCKDK was immunoprecipitated from HO8910-PM cells, and was detected by Western blotting with a MEK antibody. The results indicated that BCKDK could co-immunoprecipitates with MEK in HO8910-PM cells (Fig. 6e).
Taken together, our study indicates that BCKDK promotes EOC proliferation and migration through enhancing the MEK/ERK signaling pathway.

Discussion
OC is one of the most lethal cancers among women. As OC usually develops without well-de ned clinical symptoms, it is diagnosed mostly in advanced stages with poor ve-year survival rates of 15-45% 34 . On the other hand, FIGO stage I OC is associated with a 90% or higher survival rate. To improve the diagnostic procedures and identify OC patients in early stages of the disease, various serum markers have been developed and used. CA125 was rst reported as a tumor marker to detect ovarian tumors in the 1980s 35 . However, the sensitivity and speci city of these markers are not high 36, 37 . Therefore, the development of speci c serum markers for screening without internal examinations is urgently needed, particularly for patients with ovarian cancer, who typically do not present any symptoms until an advanced stage. In addition, over the last 30 years, mortality rates from OC, including EOC, have slightly dropped 6, 38 . The development of alternative synthetic lethal targets and diagnostic biomarkers is urgently needed.
The amino acid pro le has been identi ed as an effective diagnostic tool in different cancers previously [39][40][41][42] , and some amino acids are associated with OC [43][44][45][46] , for example, leu 47 . Branched-chain amino acid (BCAA) catabolism is closely related to tumors. The loss of BCAA catabolism promotes tumor development and growth 48, 49 . The suppression of BCAA catabolic enzyme expression led to BCAA accumulation in tumors but not in regenerating liver tissues 48 . Current research focuses on BCAT1 or BCAT2, which works in the rst step of BCAA catabolism, while there are relatively few studies on BCKDK, which is a key negative regulatory enzyme in BCAA catabolism [50][51][52][53] .
Despite this, studies have shown that the overexpression of BCKDK promotes the growth and metastasis of various tumors [27][28][29] . In this study, we determined that BCKDK promoted the proliferation and metastasis of EOC, and BCKDK was expressed at higher levels in EOC tissues than in adjacent normal tissues (Fig. 1b) and is correlated with advanced pathological grade for patients (Fig. 1e). This suggests that BCKDK could be another potential biomarker for the treatment of EOC. Inhibitors targeting BCKDK will be examined in future research.
Current research of BCAA catabolism worked in tumors focused on BCAT. As the BCAT reaction is reversible and near equilibrium, its direction should respond to changes in concentrations of BCAA and BCKAs, and availability of the donors and acceptors of nitrogen, to some extent, the conclusion was opposite in different researches. For example, some studies con rmed that the high expression of BCAT promoted the transfer of the BCAA amino group to α-ketoglutarate (α-KG) to form glutamate and the corresponding branched-chain keto acids (BCKAs). The BCAA catabolism was increasing, then the BCA-CoA entering into tricarboxylic acid cycle that provided energy for tumor cells proliferation and growth 53,54 . The other studies veri ed that the catabolism of BCAA in tumor cells was decreasing, and the high expression of BCAT promoted the conversion of BCKAs to BCAA and α-KG, and then providing essential nutrients and energy for cancer growth [48][49]55 . Our research supports the second. The overexpression of BCKDK inhibits the conversion of BCKAs to BCA-CoA, which leads to the accumulation of BCKAs.
Furthermore, the accumulation of BCKAs inhibits BCAA catabolism. Therefore, BCKAs are converted into BCAAs again through amination with the BCAT enzyme. In addition, there are studies showing that BCKDK and PPM1K make up a ChREBP-regulated node that integrates BCAA and lipid metabolism and promotes BCAAs as a material for fat synthesis for fat cells, which provide energy for tumor growth 56 . It has been proven that leu was increased in OC 47 . Other studies also found that the overexpression of BCAT promoted OC proliferation [51][52][53] . Therefore, our research gave a further understand of BCAA catabolism worked in the ovarian cancer. While, how BCKDK coordinated with BCAT to balance the BCAA metabolism? And whether they could directly regulate each other was still unclear.
Furthermore, to con rm the function of BCKDK in EOC, BCKDK was overexpressed in SKOV3 and OVCAR3 cells which poorly expressed BCKDK. BCKDK gain signi cantly promoted the proliferation and migration ex vivo, whereas knocked down the expression of BCKDK in HO8910-PM EOC cells reduced the proliferation and migration ex vivo and inhibited the tumor growth in vivo. Hence, these data supported the tumor-promoting function of BCKDK in EOC. These results were also consistent with previous ndings demonstrating that BCKDK is a key regulator of cell proliferation and metastasis in colorectal cancer and hepatocellular carcinoma [27][28][29] . Moreover, BCKDK promotes EOC proliferation and migration by activating the MEK/ERK signaling pathway. In agreement with our previous study, our previous study demonstrated that BCKDK promoted colorectal cancer proliferation by targetting the MEK1 27 . Another previous study also veri ed that BCKDK promoted hepatocellular cancer proliferation by MEK/ERK signaling pathway 29 . To our knowledge, this study was the rst to report the ectopic expression of BCKDK in EOC, and uncovered the mechanism that BCKDK regulates EOC proliferation and migration by MEK/ERK signaling pathway.
Other studies also showed that BCKDK was closely related to lipid metabolism, which was upregulated by APN 55 . In addition, BCKDK promoted tumor growth and metastasis by interacting with SRC or mTOR in colon cancer or hepatic carcinoma (Fig. 7) 28, 29 . Therefore, in addition to the MEK-ERK pathway, whether the APN, SRC, or mTOR signaling pathways are also involved in this process still needs further examination. In addition, the drug resistance of ovarian cancer is a thorny issue, and and the mitochondria are closely related to apoptosis and autophagyinduced drug resistance 58-60 . As BCKDK is located in the mitochondria, and related to drug resistance in ovarian cancer [30][31] . What is the relevant mechanism? Many questions need to be addressed in the future. Due to the limited samples in this study, future studies need to expand the number of clinical samples and collect more clinical information.

Conclusions
These ndings indicated that the expression of BCKDK was upregulated in EOC tissues and that high expression of

Consent for publication
All authors involved in the authorship are consent for publication in the current form.

Availability of data and materials
Not applicable.

Competing interests
The authors declare that they have no competing interests.