Down-regulation of KLF9 RNA and protein levels in breast cancer patients.
To explore the potential role of KLF9 in the occurrence and development of breast cancer, we downloaded RNA sequence transcriptome data and clinical information data of 1096 breast cancer cases and 112 corresponding normal tissues from TCGA database. Next, the differential gene expression data and clinical information data of breast cancer tissues and normal tissues are analyzed. The results showed that the expression level of KLF9 in breast cancer tissues was lower than that in normal tissues (Fig. 1A). To explore whether the expression of KLF9 is related to the clinical stage of patients, we analyzed the relevant data. The results showed that the expression of KLF9 in breast cancer tissue samples was related to the clinical stages of breast cancer patients, and the expression of KLF9 in late clinical patients was lower than that in early clinical patients (Fig. 1B). We also analyzed the expression of KLF9 in patients of different age groups, and found that the expression of KLF9 in patients aged > 65 was lower than that in patients aged ≤ 65 (Fig. 1C).
TNM staging system is the most common tumor staging system in the world at present. TNM staging system was put forward by Frenchman Pierre Denoix from 1943 to 1952, and then gradually improved by American Joint Committee on Cancer (AJCC) and union for international cancer control (UICC). In 1968, the first edition of TNM Classification of Malignant Tumors was officially published. The TNM staging system of each tumor is different, so the meaning of letters and numbers in TNM staging is different in different tumors. At present, TNM staging system has become the standard method for clinicians and medical scientists to stage malignant tumors(47). In TNM staging system, "T" (tumor) refers to the situation of the primary tumor, which is represented by T1 ~ T4 in turn with the increase of tumor volume and adjacent tissue involvement. "N" (node) refers to the involvement of Regional Lymph Node. When lymph nodes are not involved, they are denoted by N0, and with the increase of the degree and scope of lymph node involvement, they are denoted by N1 ~ N3 in turn. "M" (Metastasis) refers to distant metastasis (usually hematogenous metastasis). Those without distant metastasis are denoted by M0, and those with distant metastasis are denoted by M1. Therefore, a specific tumor stage can be expressed by the combination of three indicators of TNM. The TNM stage of breast cancer is shown in Table 1(48, 49). The results showed that the expression of KLF9 in breast cancer tissue samples was significantly correlated with some different T stages, different N stages and different M stages of breast cancer patients, and the low expression of KLF9 in breast cancer tissue was associated with tumor volume increase and distant metastasis (Fig. 1D-F).
Table 1
TNM staging of breast cancer
T(Tumor) | N(Lymph Node) | M(Metastasis) |
T0: No primary cancer was detected. | N0: There are no enlarged lymph nodes in the ipsilateral armpit. | M0: No distant metastasis. |
Tis: Preinvasive carcinoma. | N1: There are enlarged lymph nodes in the ipsilateral armpit, which can be pushed. | M1: Distant metastasis. |
T1: The tumor size is 0−2cm. | N2: The enlarged lymph nodes in the ipsilateral armpit fuse or adhere to the surrounding tissues. | |
T2: The tumor size is 2−5cm. | N3: There were ipsilateral parasternal lymph node metastasis and ipsilateral supraclavicular lymph node metastasis. | |
T3: Tumors larger than 5cm. | | |
T4: The tumor has penetrated the skin or attached to the chest wall. | | |
At present, the diagnosis of breast cancer mainly depends on the combination of clinical physical examination, imaging auxiliary examination and pathological examination, and pathological examination is still the main "gold standard" of diagnosis. The expression of Estrogen Receptor (ER), Progesterone Receptor (PR), human epidermal growth factor-2 (HER-2) and Ki67 in cells were detected by whole gene expression profile, and breast cancer was divided into five subtypes. They are Basal-like, HER2 overexpression (HER2+), Luminal-A, Luminal-B and Normal like(50, 51). Luminal-A breast cancer is hormone receptor positive (estrogen receptor or progesterone receptor positive), HER2 negative, and the protein Ki-67 level is low, which helps to control the growth rate of cancer cells. Luminal-A cancer is a low-grade cancer, which tends to grow slowly and has the best prognosis. Luminal-B breast cancer is hormone receptor positive (estrogen receptor or progesterone receptor positive), HER2 positive or HER2 negative, and the protein Ki-67 level. Luminal-B cancer usually grows slightly faster than luminal-A cancer, and its prognosis is slightly worse. Basal breast cancer is hormone receptor negative (estrogen receptor and progesterone receptor negative) and HER2 negative, so it is also called Triple-negative breast cancer. This type of cancer is more common in women with BRCA1 gene mutation. HER2 overexpression breast cancer is hormone receptor negative (estrogen receptor and progesterone receptor negative) and HER2 positive. HER2-enriched cancers tend to grow faster and have a worse prognosis than intracavitary cancers. For this kind of breast cancer, targeted therapy for HER2 protein is usually used in combination with corresponding drugs, such as Herceptin (chemical name: trastuzumab), Perjeta (chemical name: pertuzumab), Tykerb (chemical name: lapatinib), Nerlynx (chemical name: neratinib) and Kadcyla (chemical name: T-DM1 or ADO). Normal breast cancer, similar to Luminal-A, is hormone receptor positive (estrogen receptor or progesterone receptor positive), HER2 negative, and the level of protein Ki-67 is low, which helps to control the growth of cancer cells. However, although the prognosis of normal breast cancer is good, its prognosis is slightly lower than that of lumen cancer(52–54). See Table 2 for the classification of breast cancer subtypes.
Table 2
Subtype | ER | PR | HER2 | Ki−67 | Prognosis |
Luminal-A | + | + | - | < 14% | Better |
Luminal-B | + | + | - | ≥ 14% | Good |
| + | + | + | Any | Good |
HER2+ | - | - | + | Any | Poor |
Triple-negative/Basal-like | - | - | - | Any | Poor |
Normal like | + | + | - | < 14% | Good |
Next, we explore whether the expression of KLF9 is related to the subtypes of breast cancer patients. The results showed that the expression of KLF9 was significantly different among different subtypes, and there was a certain relationship between the low expression of KLF9 in breast cancer samples and the poor prognosis of breast cancer patients (Fig. 1G). Kaplan-Meier survival curve of breast cancer patients was generated based on TCGA data set. The results showed that there was no significant correlation between KLF9 expression in breast cancer samples and the overall survival rate of patients (Fig. 1H). The correlation analysis of gene expression between KLF9 and CDH1 (encoding E-cadherin) showed that there was a significant negative correlation between the expression levels of KLF9 and CDH1 (Fig. 1I).
Twenty tissue samples of 4 breast cancer adjacent tissues and 16 breast cancer tissues were specifically labeled with anti-KLF9 antibody, and the images of the samples were obtained at 100× magnification. It is obvious from the image that the expression of KLF9 in breast cancer tissue is lower than that in breast cancer adjacent tissue (Fig. 1J). We then randomly select the images of cell areas from each IHC slice, and use Image-J software to quantify the average optical density (AOD) of the images. The AOD data is plotted and visualized in the GraphPad Prism 9.4.0 software. The results showed that the AOD value of breast cancer tissue was significantly lower than that of breast cancer adjacent tissue, which indicated that the expression of KLF9 protein in breast cancer tissue was significantly lower than that in breast cancer adjacent tissue (Fig. 1K).
Through the analysis of transcriptome data and immunohistochemistry in TCGA database, we found that KLF9 was at a low expression level in breast cancer patients. Therefore, we have reason to believe that KLF9 has a certain correlation with the occurrence of breast cancer, and we will continue to explore it through follow-up experiments.
KLF9 inhibits the migration and invasion of human breast cancer cells
Recent studies have shown that KLF9 acts as a cancer-inhibitory effector in breast cancer cells(32, 35). The diffusion of tumor cells is often related to Epithelial interstitial transformation (EMT) process. EMT is a multifunctional cell process, which is accompanied by the characteristics of cell polarity loss, adhesion decline, cytoskeleton change, cell migration and cell mobility enhancement. Among them, the gradual decrease of E-cadherin expression on the cell surface is an important feature of EMT process. Besides, KLF9 could transcriptionally down-regulate MMP9 expression and inhibited the metastasis of breast cancer cells(35). However, the relationship between KLF9 and E-cadherin, an important marker of EMT, remains largely unknown.
In order to explore the association between KLF9 and E-cadherin in the occurrence and development of breast cancer, we determined the protein level in breast cancer cell lines containing MCF-7, T47D, ZR-75-30 cells. Western blot analysis revealed almost no endogenous KLF9 was detected in ZR-75-30 cells, which displayed mesenchymal-like morphology and lesser cell–cell adhesion. Relatively higher-level expression of KLF9 was identified in MCF-7 cells than in T47D cells, consistent with the expression of E-cadherin in breast cancer cells (Fig. 2A and 2B). The protein expression of KLF9 in ZR-75-30 breast cancer cell line of Luminal-B subtype is lower than that of T47D and MCF7 breast cancer cell lines of Luminal-A subtype, which is consistent with the transcription level analysis of KLF9 in different breast cancer subtypes in TCGA database. These studies suggested that KLF9 expression is higher in non-invasive cancer cells and may inhibit breast cancer by increasing cell-cell adhesion.
To verify the role of KLF9 in cell motility in breast cancer cells, scratch wound-healing and transwell assays were carried out to demonstrate the migration and invasion capacity. Compared with control cells, overexpressing of KLF9 markedly inhibited cell motility (Fig. 2C and 2D) and reduced the cell numbers that migrated through the Matrigel-coated membrane (Fig. 2G and 2H) in ZR-75-30 cells. In contrast, knockdown of endogenous KLF9 in ZR-75-30 cells significantly enhanced cell migration ability (Fig. 2E and 2F) and increased the capacity of cells to traverse the Matrigel-coated membrane (Fig. 2I and 2J), compared with control cells. These results indicated that KLF9 can inhibit the invasion and metastasis in breast cancer cells.
KLF9 increases E-cadherin expression via transcriptional activation in breast cancer cells
As E-cadherin is an important inhibitor in the process of EMT(55, 56), which has the same function as KLF9, we speculate whether KLF9 inhibits breast cancer metastasis by regulating the expression of E-cadherin. As we predicted, Reverse transcription-PCR(RT-PCR) showed that overexpression of KLF9 increased the mRNA level of E-cadherin compared with the control group in ZR-75-30 cells (Fig. 3A and 3B). In contrast, the mRNA level of E-cadherin decreased after siKLF9 (pRNA T-U6.1 vector) was transfected in ZR-75-30 cells, suggesting that KLF9 could promote the transcription of E-cadherin (Fig. 3A and 3B). Western blot results showed that increased level of E-cadherin protein by overexpression of KLF9, and decreased level of E-cadherin protein following siKLF9 (pRNAT-U6.1 vector), consistent with the change of mRNA level (Fig. 3C and 3D). These data clearly indicated that increased E-cadherin protein expression was mediated by KLF9, occurred mainly at the mRNA level regulation.
Luciferase reporter gene assay also revealed that increasing doses of KLF9 induced a dose-dependent growth of E-cadherin-promoter-driven luciferase activity (Fig. 3E), while knockdown of KLF9 resulted in a decrease in E-cadherin-promoter-driven luciferase activity in ZR-75-30 cells (Fig. 3F). These data showed that KLF9 could activate the E-cadherin promoter.
Taken together, these results indicated that KLF9 could up-regulate the level of E-cadherin protein by transcriptionally activating the activity of the E-cadherin promoter.
KLF9 binds to the CACCC motif of the E-cadherin promoter through its DNA binding domain
To explore the specific sites and regions of KLF9 acting on E-cadherin promoter, we constructed four truncated E-cadherin promoter and fused each promoter with luciferase reporter gene to yield a reporter construct (Fig. 4A).
The reporter construct comprising E-cad-a promoter (-999 ~ + 47) increased more than two-fold of the reporter activity when ZR-75-30 cells overexpressed KLF9 compared with control. Interestingly, the reporter activity disappeared when deleted nucleotide from − 206 to + 47 (E-cad-b), suggesting that KLF9 binding site might be existed in nucleotide − 206 to + 47. The result showed that after the deletion of nucleotide − 999 to -206 (E-cad-c), the activation multiple of KLF9 overexpression was almost the same as that of the E-cad-a, which proved that nucleotide − 999 to -206 (E-cad-b) probably contained no regulatory element. These data indicate that nucleotide − 206 to + 47 (E-cad-c) on the E-cadherin promoter was contain response elements required for KLF9 to activate its transcriptional activity. Furthermore, due to KLF9 recognize sequences with a preference for the 5′-CACCC-3′ core motif in the promoters and enhancers(31), nucleotide − 12 to + 8 "CACCC" located within nucleotide − 206 to + 47 was mutated into "CATTT" (E-cad-d), the activation of KLF9 was almost completely lost, indicating that KLF9 was exactly bound to the "CACCC" sequence of E-cadherin promoter (Fig. 4B).
Since KLF9 DNA-binding domain (zinc finger domain, ZNF) could recognize the CACCC element, we speculated that DNA-binding domain of KLF9 might bind to the E-cadherin promoter to regulate the transcription activity of E-cadherin. Therefore, to verify this hypothesis, we constructed a deletion of KLF9 DNA binding domain (pcDNA3.1-3×Flag-KLF9△DBD) (Fig. 4C). CHIP experiment further showed that KLF9 can specifically bind to the E-cad-c promoter, which included "CACCC" sequence, consistent with the results of previous reporter gene assay (Fig. 4D, top). KLF9-∆DBD cannot achieve specific binding to E-cad-c promoter. We also amplified the GAPDH fragment as a control (Fig. 4D, bottom). Taken together, these results suggest that KLF9 may promote E-cadherin transcription through its DNA-binding domain (DBD) interaction with CACCC motif located within nucleotides − 206 to + 47 of E-cadherin promoters. Consistently, compared to full length KLF9, the activation of E-cad-c promoter by KLF9△DBD was disappeared (Fig. 4E and 4F ), suggesting that the DNA-binding domain (DBD) of KLF9 might be responsible for the E-cadherin promotion. This was subsequently confirmed by ChIP assay. ChIP assay also showed that the KLF9△DBD experimental group, compared with the wild-type KLF9, did not detect specific DNA bands on E-cad-c promoter, indicating that KLF9 DNA-binding domain specifically bind to E-cad-c promoter (Fig. 4D).
Collectively, these results suggested that KLF9 binds to the CACCC motif of E-cadherin promoter through its DNA binding domain and the DNA-binding domain of KLF9 is essential for its tumor-suppressive role in breast cancer cells.
Furthermore, MCF-7 cells were tested for their motility, migration ability and invasion ability by wound healing and transwell experiment respectively. When KLF9 is overexpressed, the migration ability of cells is obviously reduced (Fig. 4G and 4H), and the ability of cells to pass through the matrix gel coating film is also reduced (Fig. 4I and 4J). However, when Flag-KLF9-△DBD was overexpressed, compared with the control group, the migration ability of cells did not change significantly (Fig. 4G and 4H), and the ability of cells to pass through the matrix gel coating film remained unchanged (Fig. 4I and 4J). To sum up, after deleting the DNA binding region of KLF9, the inhibitory effect of overexpression of KLF9 on the motility, migration and invasion of MCF7 cells disappeared, which proved that the DNA binding region of KLF9 was the necessary region to inhibit the motility, migration and invasion of breast cancer cells.
KLF9 affects the transcriptional regulation of E-cadherin by competing with SNAI1
The transcription factor SNAI1 has been recognized as a direct repressor of E-cadherin expression, which recruits HDAC and the corepressor mSin3A to form a multimolecular complex to repress E-cadherin(6, 26, 29). Because of the target sites located in E-cad-c, we speculated that KLF9 may be involved in coordinating the interaction between SNAI1 and E-cadherin.
Firstly, we utilized luciferase reporter detection to further evaluate our hypothesis. Overexpression of SNAI1 alone inhibited E-cadherin-Luc activity, whereas overexpression of SNAI1 and KLF9 relieved this inhibition (Fig. 5A). Moreover, western blot results also showed that overexpression of SNAI1 alone repressed the expression of E-cadherin protein, whereas overexpression of SNAI1 and KLF9 relieved this repression (Fig. 5B and 5C). Quantitative RT-PCR showed that E-cadherin mRNA levels were significantly decreased after transfection of SNAI1 alone, and KLF9 partially rescued the decrease of E-cadherin mRNA, consistent with the change of protein level (Fig. 5D and 5E).
To determine whether KLF9 affects the regulation of E-cadherin promoter by SNAI1, ChIP assays were carried out using ZR-75-30 cells. Overexpression of KLF9 effectively prevented the recruitment of SNAI1 to the E-cad-c promoter (Fig. 5F). Taken together, these findings revealed that KLF9 may enhance the activity of E-cadherin by alleviating its transcriptional suppression exerted by SNAI1.
To further confirm, using scratch wound-healing and transwell assays to demonstrate the cell motility, migration and invasion capacity respectively in ZR-75-30 cells. When the ZR-75-30 cells co-overexpressed KLF9 and SNAI1, the migration ability of cells are significantly reduced (Fig. 5G and 5H), and the ability of cells to pass through the matrix gel coating film are reduced (Fig. 5I and 5J), compared to the cells expressed SNAI1 alone. Collectively, these results indicate that SNAI1 is involved in KLF9-mediated suppression of breast cancer invasion and metastasis.