Abstract
Background: Vasculogenic mimicry (VM) plays an important role in invasion and metastasis of malignant tumor. High expression of vascular endothelial cadherin (VE-cahderin) in malignant tumor cells can promote the formation of VM. High expression of SOX4 (sex-determining region Y-related high-mobility group box 4) was found in esophageal squamous cell carcinoma (ESCC). It can promote the development of epithelial stromal transformation. Then, SOX4 can promote the formation of VM in ESCC. Methods: Paraffin-embedded specimens of ESCC (with complete clinicopathological data) and normal esophageal mucosa adjacent to carcinoma (> 5 cm) were collected from January to December 2013. CD34/PAS was used to detect VM. The expression of VE-cadherin and SOX4 was used by immunohistochemistry. The patients were followed up in detail (survival time and survival status). Results: SOX4, VM, and VE-cadherin were highly expressed in ESCC. Moreover, they were positively correlated. Survival analysis shows that the expressions of SOX4, VM, and VE-cadherin are associated with the patient’s prognosis and can be independent prognostic factors for ESCC. Conclusions: Studies suggests that SOX4, which is highly expressed in ESCC, is involved in the formation of VM. The combined detection of SOX4, VE-cadherin and VM expression can be used as biomarkers for invasion and metastasis of ESCC. These three markers can be used as powerful prognostic factors in patients with ESCC.
Keywords: Esophageal squamous cell carcinoma, vasculogenic mimicry, SOX4, vascular endothelial cadherin, epithelial-mesenchymal transition
Introduction
Esophageal cancer is a malignant tumor of the eighth in the world [1]. Esophageal squamous cell carcinoma (ESCC) is one of the most common types, accounting for more than 90% of esophageal cancer [2]. Although there are many treatments at present, the survival time of patients has indeed improved However, the five-year survival rate of patients is still less than 50% [3]. Cancer has always been a big problem for humans to overcome. The detection of cancer mysteries from a genetic perspective has been a hot topic in recent years. However, the development and prognosis of cancer such as esophageal cancer remains unclear.
Vasculogenic mimicry (VM) was first reported in 1999 by scholars to study malignant melanoma and was gradually recognized and agreed upon. VM is described as promoting the direct interconnection of tumor cells to form blood vessel-like channels for transporting blood [4]. VM can circulate tumor cells through the blood circulation, and at the same time, through the formation of an extracellular mechanism (ECM) network structure. The tumor cells can be directly introduced into the microcirculation through the blood circulation and transferred to other organs. In summary, VM plays an important role in promoting tumor cell growth and metastasis [5]. In recent years, various invasive tumors have reported the formation of a channel-like structure that mimics the formation of blood cells and nutrients, such as cervical cancer [6], gastric cancer [7], pancreatic cancer [8], and hepatocellular cancer [9]. VM is also reported to be present in esophageal tumors, including esophageal stromal tumors [10] and ESCC [11].
Vascular endothelial cadherin (VE-cadherin), which was a member of vascular endothelial cells express of the cadherin superfamily, is the major adhesion receptor for endothelial cell adhesion junctions [12]. Studies have shown that VE-cadherin plays an important role in the formation of VM [5,13]. VE-cadherin has been reported to be over-expressed in a variety of tumor cells, and involved in neo-vascularization, growth, and progression of tumors [14]. It has been demonstrated that the abnormal expression of VE-cadherin promoted the progression and metastasis of malignant melanoma and participated in the VM of this tumor [15,16]. Tang NN et al. predicted that HIF-1α can directly or indirectly regulate the expression of VE-cadherin, EphA2, and Laminin 5γ2 (LN5γ2) genes. It can also promote the formation of VM and provide blood supply for ESCC [17]. However, it is still unclear whether the correlation between VM and VE-cadherin in ESCC is related to the abnormal expression of other factors.
SOX4, called sex-determining region Y-related high-mobility group box 4, plays an important role in tumor development. SOX4 is reported to be highly expressed in a variety of tumors, including colorectal cancer [18], ESCC [19], and breast and lung adenocarcinoma cancer [20]. In recent years, it has been reported that SOX4 participates in epithelial-mesenchymal transition (EMT) by regulating the modification of the genetic factor Ezh2 [21]. Sasaki A et al. also suggest that ADAM28 gene may be involved in cancer cell invasion through SOX4-EMT mediated with cancer cell [20]. In conclusion, SOX4 is closely related to EMT and involved in cancer progression.
It has been demonstrated that the formation of VM is related to EMT. EMT can promote angiogenesis and VM construction [23]. The purpose of this experiment is to investigate whether SOX4 in ESCC participates in the formation of VM, thereby improving certain candidate genes for targeted therapy.
Patients and methods
Patients and specimens
170 cases of ESCC paraffin-embedded (preoperative radiotherapy, chemotherapy, or other anti-tumor treatment) and 60 cases of normal esophageal epithelial tissue were archived in the Department of the First Affiliated Hospital of Bengbu Medical College from January 2013 to December 2013. All cases are required to have complete clinical and pathological data. 89 cases were males and 81 females, aged from 29 to 82 years (80 cases ≤ 60 years and 90 cases > 60 years). Gross typing showed the medullary type was 61 cases, ulceration type 51 cases, constricted type 36 cases, and fungating type 22 cases. Location: upper 18 cases, middle 117 cases, and lower 35 cases; tumor size: diameter ≤ 3.5 cm 96 cases, > 3.5 cm 74 cases; without lymph node metastasis 47 cases, and lymph node metastasis 123 cases; TNM stage: I+II 130 patients and III+IV 40 patients; histological grade: 81 with well differentiation, 68 with moderate differentiation, and 21 with poor differentiation; depth of infiltration: over serous membrance 60 cases and above serous membrance 110 cases. 60 cases of normal esophageal tissue were taken from ESCC tissue > 5.0 cm was used as controls. This experiment was approved by the Ethics Committee of Bengbu Medical College.
Immunohistochemistry
ESCC specimens were fixed in 4% neutral formalin solution, dehydrated, and embedded in paraffin. The wax block was taken out, serially sliced in a thickness of 4 μm, dewaxed with xylene, washed with a gradient ethanol solution, and rinsed with running water. The endogenous peroxidase activity was blocked by incubation in 3% H2O2 for 10 min at room temperature and then heated to 95°C for 30 min for antigen retrieval. SOX4 (rabbit polyclonal, diluted 1:200, Abcam), VE-cadherin (rabbit polyclonal, diluted 1:200, Abcam), and CD34 (rabbit monoclonal, diluted 1:250, Abcam) were stained according to the instructions of the kit. Drip Envision reagent and diaminobenzidine (DAB) were used for color development. Hematoxylin was appropriately counterstained, dehydrated, and transparently mounted. A positive control was used as a positive control and phosphate buffer (PBS, pH 7.2) was used instead of the primary antibody as a negative drop. The above experiments were performed according to the ElivisionTM Plus detection kit instructions (Lab Vision, USA).
CD34 and PAS double overdying
After antibody-CD34 to tissue block was added and DAB was developed, the color reaction was terminated by running water for 1 min and oxidized in a 0.5% periodate solution for 10 min. After washing for 2 min in running water, the Schiff solution was stained for 15-30 min, then rinsed with distilled water for 3 times, 1 min/time; then hematoxylin counterstained, hydrochloric acid ethanol differentiated, returned to blue, dehydrated, transparent, neutral gum seal.
Interpretation of immunohistochemistry
SOX4 is positive for brownish yellow mainly in the nucleus. VE-cadherin positive expression is located in the cytoplasm and cell membrane. The determination was made by the secondary scoring method of dyeing intensity and coloring range. Each specimen was randomly counted for 10 high power fields (×400 magnification), first scored by staining intensity: 0 for colorless, 1 for light yellow, 2 for brown, and 3 for tan. Then scored from the coloring range: positive cells < 10%, count 0 points; 11% to 25%, count 1 point; 25% to 50%, count 2 points; 51% to 75% count 3 points, and 76% ~ count 4 points. The product of the scores of the two is used as the criterion for the result judgment, the integral ≤ 3 is negative and > 3 is positive. The results of immunohistochemistry were determined by two pathology doctors.
VM interpretation
The antibody CD34 immunohistochemical step was carried out to the DAB color reaction. After the vascular endothelium was observed under the microscope, the water flow was washed for 1 min to terminate the color reaction. Add 15% periodic acid 1 drop, reduce for 10 min, rinse with water for 10 min. After the Schiff reagent 1 drop, the sample was placed in a dark box and placed in an incubator at 37°C for 20 min. The tissue was red-colored until the reaction was stopped and the trickle water was rinsed for 20 min. Hematoxylin was counterstained for 2 min. 50°C warm water back to blue for 1 min, conventional dehydration, and sealing. Evaluation methods for VM and control organizations in ESCC organizations refer to Zhang et al. [11]. VM is defined as an irregular vessel-like structure surrounded by CD34-negative and PAS-positive tumor cells. Red blood cells, perivascular inflammatory cells, bleeding, and necrosis were seen. The normal endothelial vessels were positive for both CD34 and PAS, and the circular duct structure was regular.
Statistical analysis
All experimental data were analyzed by statistical software SPSS 22. 0. The positive differences were tested by Chi-square test or Fisher’s exact test. Spearman correlation among VM, VE-cadherin and SOX4 positive and negative groups were analyzed. Univariate and multivariate analyses were used to detect the effects of VM, VE-cadherin, and SOX4 of survival. The comparison between groups was performed by the Log-rank test with Kaplan-Meier method. The Cox regression model was used by multivariate analyses. P < 0.05 was considered statistically significant.
Results
Association between VM, VE-cadherin and SOX4 and relationship with clinicopathological features
In this experiment, SOX4 protein was up to 64.1% (109/170) in ESCC, which was higher than that in other normal esophageal mucosa tissues. The difference was statistically significant (P < 0.05). The expression rate of VE-cadherin in ESCC was 56.5% (96/170), but it was not expressed in normal esophageal mucosa (P < 0.05). The positive rate of the VM structure was 51.2% (87/170) in ESCC, but there was no expression of VM in normal esophageal mucosa. The difference was statistically significant (P < 0.05). The expression of SOX4, VE-cadherin and VM in ESCC is shown in Figure 1.
Figure 1.
Immunostaining of SOX4, VE-cadherin and VM in ESCC or normal tissue. A. Positive staining of SOX4 mainly in the nucleus of ESCC cells (×400 magnification); B. Weakly positive staining of SOX4 in the nucleus of normal tissue cells (×400 magnification); C. Positive staining of VE-cadherin in the cytoplasm and membrane of the ESCC cells (×400 magnification); D. Negative staining of VE-cadherin in normal tissue cells (×400 magnification); E and F. Positive staining of VM in ESCC cells (×400 magnification, red arrow is VM structure, black arrow is microvessel).
The positive expression of SOX4 in ESCC was correlated with tumor gross, diameter, location, depth of infiltration, histological grade, and TNM stage (P < 0.05), and was not associated with gender, age, or lymph node metastasis (P > 0.05). VE-cadherin positive in ESCC was related to patients’ age, tumor diameter, depth of infiltration, lymph node metastasis, histological grade, and TNM stage (P < 0.05). However, it was not correlated with gender, tumor location, and gross type (P > 0.05). With the increase of VM positive expression in ESCC, the tumor diameter was larger, the depth of infiltration was deeper, the histological grade was lower, the TNM stage was later, and tumor with lymph node metastasis was more likely (P < 0.05). However, the positive expression of VM was not associated with gender, age, tumor location, and gross type (P > 0.05). The data is shown in Table 1.
Table 1.
Correlation of SOX4, VE-cadherin, and VM expression with the clinicopathological characteristics of ESCC
| Variable | SOX4 | χ2 | P | VE-cadherin | χ2 | P | VM | χ2 | P | |||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
|
|
|
|
||||||||||
| + | - | + | - | + | - | |||||||
| Gender | 0.883 | 0.347 | 0.291 | 0.590 | 2.806 | 0.094 | ||||||
| Male | 60 | 29 | 52 | 37 | 51 | 38 | ||||||
| Female | 49 | 32 | 44 | 37 | 36 | 45 | ||||||
| Age (years) | 0.051 | 0.821 | 4.947 | 0.026 | 3.336 | 0.068 | ||||||
| ≤ 60 | 52 | 28 | 38 | 42 | 35 | 45 | ||||||
| > 60 | 57 | 33 | 58 | 32 | 52 | 38 | ||||||
| Gross | 14.793 | 0.002 | 7.010 | 0.072 | 2.852 | 0.415 | ||||||
| Medullary | 47 | 14 | 39 | 22 | 36 | 25 | ||||||
| Ulceration | 22 | 29 | 21 | 30 | 22 | 29 | ||||||
| Constricted | 25 | 11 | 22 | 14 | 18 | 18 | ||||||
| Fungating | 15 | 7 | 14 | 8 | 11 | 11 | ||||||
| Location | 0.131 | 0.937 | 0.736 | 0.692 | 0.462 | 0.794 | ||||||
| Upper | 11 | 7 | 10 | 8 | 8 | 10 | ||||||
| Middle | 76 | 41 | 64 | 53 | 60 | 57 | ||||||
| Lower | 22 | 13 | 22 | 13 | 19 | 16 | ||||||
| Diameter (cm) | 7.609 | 0.006 | 6.565 | 0.010 | 24.916 | < 0.001 | ||||||
| ≤ 3.5 | 53 | 43 | 46 | 50 | 33 | 63 | ||||||
| > 3.5 | 56 | 18 | 50 | 24 | 54 | 20 | ||||||
| Depth of infiltration | 72.631 | < 0.001 | 87.411 | < 0.001 | 62.923 | < 0.001 | ||||||
| Over serous membrane | 13 | 47 | 5 | 55 | 6 | 54 | ||||||
| Above serous membrane | 96 | 14 | 91 | 19 | 81 | 29 | ||||||
| Lymph node metastasis | 3.025 | 0.082 | 6.656 | 0.010 | 22.893 | < 0.001 | ||||||
| No | 74 | 49 | 62 | 61 | 38 | 9 | ||||||
| Yes | 35 | 12 | 34 | 13 | 49 | 74 | ||||||
| TNM stage | 7.683 | 0.006 | 11.781 | 0.001 | 20.540 | < 0.001 | ||||||
| I and II | 76 | 54 | 64 | 66 | 54 | 76 | ||||||
| III and IV | 33 | 7 | 32 | 8 | 33 | 7 | ||||||
| Histological grade | 18.821 | < 0.001 | 15.900 | < 0.001 | 12.979 | 0.002 | ||||||
| Well | 39 | 42 | 33 | 48 | 30 | 51 | ||||||
| Moderate | 51 | 17 | 47 | 21 | 42 | 26 | ||||||
| Poor | 19 | 2 | 16 | 5 | 15 | 6 | ||||||
Univariate and multivariate analyes
In this experiment, the overall survival (OS) time follow-up was death or survival until December 2018, the OS time of patients was from 5 to 60 months, the median survival time was 28 months, and the OS rate of patients was 40.6% (69/170). Disease-free survival (DFS) time follow-up is patient death, ESCC recurrence, or survival until December 2018, the patient’s DFS time is from 3 to 60 months, the median survival time is 25 months, and the patient DFS rate is 35.3% (60/170).
The data showed that the OS time of ESCC patients with SOX4 positive expression group was shorter than that of SOX4 negative expression group (log-rank = 68.169, P < 0.001; Figure 2A). OS time of VE-cadherin positive patients was significantly shorter than those of VE-cadherin negative patients (log-rank = 128.551, P < 0.001; Figure 3A). Lastly, the OS time with the expression of VM positive was shorter than that of VM negative group (log-rank = 129.826, P < 0.001; Figure 4A). Similarly, in this experiment, the DFS time analysis showed that the DFS time of SOX4-positive group, VE-cadherin-positive group and VM-positive group was smaller than that of the corresponding negative group (log-rank = 84.542, 132.027, 126.127, P < 0.001, Figures 2B, 3B, 4B). More interestingly, this experiment demonstrated that the overall survival time and disease-free survival time of co-expressed SOX4, VE-cadherin, and VM-positive groups were significantly shorter than co-expression of SOX4, VE-cadherin, VM-negative or other groups (log-rank = 135.720,143.257, P < 0.001, Figure 5A, 5B). In addition, OS time and DFS time were positively correlated with tumor tissue size, depth of invasion, lymph node metastasis, and histological grade (P < 0.001, respectively). The results of univariate logistic regression analysis of OS and DFS were shown in Tables 2 and 3. Through multi-factor Cox analysis of OS and DFS, this experiment demonstrated that SOX4, VE-cadherin, VM, tumor diameter, and lymph node metastasis can be used as independent prognostic factors for patients with ESCC (P < 0.05, Tables 4 and 5).
Figure 2.
Kaplan-Meier analysis of the survival rate of ESCC patients in relation to SOX4. A. OS of all patients (log-rank = 68.169, P < 0.001); B. DFS of all patients (log-rank = 84.542, P < 0.001).
Figure 3.
Kaplan-Meier analysis of the survival rate of ESCC patients in relation to VE-cadherin. A. OS of all patients (log-rank = 128.551, P < 0.001); B. DFS of all patients (log-rank = 132.027, P < 0.001).
Figure 4.
Kaplan-Meier analysis of the survival rate of ESCC patients in relation to VM. A. OS of all patients (log-rank = 129.826, P < 0.001); B. DFS of all patients (log-rank = 126.127, P < 0.001).
Figure 5.
Kaplan-Meier analysis of the survival rate of ESCC patients in relation to co-expressed SOX4, VE-cadherin, and VM. A. OS of all patients (log-rank = 135.720, P < 0.001); B. DFS of all patients (log-rank = 143.257, P < 0.001).
Table 2.
Results of univariate logistic regression analyses of OS
| Variable | N | Mean OS time (months) | Log-rank | P |
|---|---|---|---|---|
| Gender | 0.458 | 0.499 | ||
| Male | 89 | 34.202±2.289 | ||
| Female | 81 | 36.728±2.442 | ||
| Age (years) | 7.705 | 0.006 | ||
| ≤ 60 | 80 | 40.413±2.408 | ||
| > 60 | 90 | 30.956±2.223 | ||
| Gross | 5.094 | 0.165 | ||
| Medullary | 61 | 31.426±2.699 | ||
| Ulceration | 51 | 41.294±3.055 | ||
| Constricted | 36 | 33.389±3.485 | ||
| Fungating | 22 | 35.091±4.515 | ||
| Location | 2.127 | 0.345 | ||
| Upper | 18 | 33.722±5.269 | ||
| Middle | 117 | 36.940±2.014 | ||
| Lower | 35 | 30.829±3.476 | ||
| Diameter (cm) | 20.712 | < 0.001 | ||
| ≤ 3.5 | 96 | 41.875±2.104 | ||
| > 3.5 | 74 | 27.014±2.375 | ||
| Depth of infiltration | 86.543 | < 0.001 | ||
| Over serous membrane | 60 | 55.967±1.730 | ||
| Above serous membrane | 110 | 24.191±1.599 | ||
| Lymph node metastasis | 67.604 | < 0.001 | ||
| No | 123 | 42.203±1.764 | ||
| Yes | 47 | 17.489±2.404 | ||
| TNM stage | 72.078 | < 0.001 | ||
| I and II | 130 | 41.338±1.755 | ||
| III and IV | 40 | 16.025±2.381 | ||
| Histological grade | 17.437 | < 0.001 | ||
| Well | 81 | 42.556±2.289 | ||
| Moderate | 68 | 28.691±2.582 | ||
| Poor | 21 | 28.429±3.793 | ||
| SOX4 | 68.169 | < 0.001 | ||
| Negative | 61 | 54.000±1.879 | ||
| Positive | 109 | 24.835±1.683 | ||
| VE-cadherin | 128.551 | < 0.001 | ||
| Negative | 74 | 55.257±1.563 | ||
| Positive | 96 | 20.021±1.288 | ||
| VM | 129.826 | < 0.001 | ||
| Negative | 83 | 52.578±1.732 | ||
| Positive | 87 | 18.931±1.249 |
Table 3.
Results of univariate logistic regression analyses of DFS
| Variable | N | Mean DFS time (months) | Log-rank | P |
|---|---|---|---|---|
| Gender | 0.500 | 0.479 | ||
| Male | 89 | 31.539±2.303 | ||
| Female | 81 | 34.039±2.459 | ||
| Age (years) | 5.273 | 0.022 | ||
| ≤ 60 | 80 | 37.251±2.431 | ||
| > 60 | 90 | 28.736±2.251 | ||
| Gross | 5.467 | 0.141 | ||
| Medullary | 61 | 29.138±2.680 | ||
| Ulceration | 51 | 39.034±3.142 | ||
| Constricted | 36 | 30.503±3.426 | ||
| Fungating | 22 | 31.056±4.630 | ||
| Location | 0.663 | 0.718 | ||
| Upper | 18 | 30.907±5.239 | ||
| Middle | 117 | 33.781±2.028 | ||
| Lower | 35 | 29.727±1.684 | ||
| Diameter (cm) | 17.798 | < 0.001 | ||
| ≤ 3.5 | 96 | 38.892±2.155 | ||
| > 3.5 | 74 | 24.717±2.369 | ||
| Depth of infiltration | 95.754 | < 0.001 | ||
| Over serous membrane | 60 | 55.402±1.819 | ||
| Above serous membrane | 110 | 20.559±1.415 | ||
| Lymph node metastasis | 64.068 | < 0.001 | ||
| No | 123 | 39.254±1.837 | ||
| Yes | 47 | 14.511±1.863 | ||
| TNM stage | 64.973 | < 0.001 | ||
| I and II | 130 | 38.324±1.810 | ||
| III and IV | 40 | 13.275±1.795 | ||
| Histological grade | 23.821 | < 0.001 | ||
| Well | 81 | 40.626±2.411 | ||
| Moderate | 68 | 25.698±2.437 | ||
| Poor | 21 | 22.098±2.591 | ||
| SOX4 | 84.542 | < 0.001 | ||
| Negative | 61 | 53.004±2.013 | ||
| Positive | 109 | 21.287±1.495 | ||
| VE-cadherin | 132.027 | < 0.001 | ||
| Negative | 74 | 53.023±1.703 | ||
| Positive | 96 | 17.208±1.164 | ||
| VM | 126.127 | < 0.001 | ||
| Negative | 83 | 49.934±1.868 | ||
| Positive | 87 | 16.356±1.140 |
Table 4.
Results of multivariate logistic regression analyses of OS
| B | SE | Wald | P | Exp(B) | 95.0% CI for Exp(B) | ||
|---|---|---|---|---|---|---|---|
|
| |||||||
| Lower | Upper | ||||||
| Gender | 0.059 | 0.228 | 0.066 | 0.797 | 1.060 | 0.678 | 1.658 |
| Age (years) | 0.394 | 0.229 | 2.969 | 0.085 | 1.483 | 0.947 | 2.321 |
| Location | -0.151 | 0.206 | 0.536 | 0.464 | 0.860 | 0.574 | 1.288 |
| Gross | -0.035 | 0.098 | 0.124 | 0.725 | 0.966 | 0.797 | 1.171 |
| Diameter (cm) | 0.480 | 0.221 | 4.735 | 0.030 | 1.616 | 1.049 | 2.491 |
| Depth of infiltration | 1.086 | 0.545 | 3.970 | 0.046 | 2.964 | 1.018 | 8.630 |
| Lymph node metastasis | -3.467 | 0.648 | 28.672 | < 0.001 | 0.031 | 0.009 | 0.111 |
| Histological grade | -0.083 | 0.159 | 0.272 | 0.602 | 0.920 | 0.674 | 1.257 |
| TNM stage | -1.555 | 0.634 | 6.013 | 0.014 | 0.211 | 0.061 | 0.732 |
| SOX4 | 1.012 | 0.469 | 4.659 | 0.031 | 2.752 | 1.098 | 6.899 |
| VE-cadherin | 1.933 | 0.488 | 15.700 | < 0.001 | 6.911 | 2.656 | 17.983 |
| VM | 0.777 | 0.346 | 5.042 | 0.025 | 2.174 | 1.104 | 4.283 |
Table 5.
Results of multivariate logistic regression analyses of DFS
| B | SE | Wald | P | Exp(B) | 95.0% CI for Exp(B) | ||
|---|---|---|---|---|---|---|---|
|
| |||||||
| Lower | Upper | ||||||
| Gender | -0.182 | 0.210 | 0.752 | 0.386 | 0.834 | 0.553 | 1.258 |
| Age (years) | 0.170 | 0.219 | 0.605 | 0.437 | 1.186 | 0.772 | 1.820 |
| Location | -0.115 | 0.207 | 0.312 | 0.577 | 0.891 | .594 | 1.336 |
| Gross | 0.001 | 0.093 | 0.000 | 0.988 | 1.001 | 0.834 | 1.202 |
| Diameter (cm) | 0.516 | 0.200 | 6.663 | 0.010 | 1.675 | 1.132 | 2.479 |
| Depth of infiltration | 0.017 | 0.381 | 0.002 | 0.965 | 1.017 | 0.482 | 2.145 |
| Lymph node metastasis | -2.201 | 0.550 | 16.008 | < 0.001 | 0.111 | 0.038 | 0.325 |
| Histological grade | -0.162 | 0.174 | 0.870 | 0.351 | 0.850 | 0.605 | 1.196 |
| TNM stage | -0.341 | 0.559 | 0.372 | 0.542 | 0.711 | 0.237 | 2.128 |
| SOX4 | 0.970 | 0.401 | 5.847 | 0.016 | 2.637 | 1.202 | 5.786 |
| VE-cadherin | 1.810 | 0.423 | 18.291 | < 0.001 | 6.109 | 2.666 | 14.002 |
| VM | 1.462 | 0.307 | 22.718 | < 0.001 | 4.316 | 2.366 | 7.875 |
Correlation of SOX4, VE-cadherin and VM in ESCC
There was a positive correlation between SOX4 and VE-cadherin (r = 0.728, P < 0.001) and the same correlation between SOX4 and VM (r = 0.570, P < 0.001). In this experiment, we also detected a significant positive correlation between VM and VE-cadherin protein expression (r = 0.685, P < 0.001). The experimental results are shown in Table 6.
Table 6.
Correlation among SOX4, VE-cadherin, and VM expression in ESCC
| Variable | SOX4 | r | P | VE-cadherin | r | P | VM | r | P | |||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
|
|
|
|
||||||||||
| + | - | + | - | + | - | |||||||
| VE-cadherin | 0.728 | < 0.001# | 0.685 | < 0.001# | ||||||||
| + | 91 | 5 | 78 | 18 | ||||||||
| - | 18 | 56 | 9 | 65 | ||||||||
| VM | 0.570 | < 0.001# | 0.685 | < 0.001# | ||||||||
| + | 79 | 8 | 78 | 9 | ||||||||
| - | 30 | 53 | 18 | 65 | ||||||||
positive association.
Discussion
SOX4 can be used as a proto-oncogene, which is over-expressed and participates in the development of various tumors [18-20]. It was also found that SOX4 decreased expression in some tumors and inhibited tumor progression [23]. In this experiment, we found that the positive expression rate of SOX4 in patients with ESCC was 64.1%, which was significantly higher than that in normal tissues (P < 0.05). This is consistent with previous research results [24,25]. SOX4 positively was found to be significantly correlated with tumor of infiltration depth, histological grade, and TNM stage (P < 0.05). Correlation between lymph node metastasis and SOX4 protein expression in tumors showed P = 0.082, which was close to 0.05, indicating that lymph node metastasis also had an effect on SOX4 expression. The above experimental results suggest that SOX4 is involved in the progression and metastasis of ESCC. Han et al. proved that SOX4 is highly expressed in ESCC. Meanwhile, they found that knocking out the SOX4 gene can inhibit tumor cell proliferation and enhance doxorubicin-induced tumor cell senescence in vitro [25]. SOX4 was reported to be essential to EMT by regulating the epigenetic modifier Ezh2 [21]. EMT refers to the transition of cells from epithelial phenotype to interstitial phenotype. Once EMT occurs, cells lose cell-to-cell contact, polarity disappears and migration, and invasion type [26]. Zhang et al. reported that over-expression of SOX4 in mammary epithelial cells can enhance the stem cell characteristics of cells and promote their invasion and metastasis. This process is related to the activation of TGF-β signaling pathway and the induction of EMT [27]. Moreover, our experimental results also have demonstrated that SOX4 was associated with patient OS time and DFS time (P < 0.05). The SOX4 expression was higher, OS time and DFS time of the patients were shorter.
After Folkman et al. proposed the concept of vascular-dependent growth of tumors in the 1970s, the influencing factors and mechanisms of tumor angiogenesis became one of the hotspots of cancer research [28]. The VM refers to the ability of some tumors to simulate endothelial cells and form tube-like structures without the involvement of vascular endothelial cells and can communicate with the host vessel. In this study, we confirmed the high expression of VM in ESCC by immunohistochemical method, but no expression of VM in normal esophageal mucosa (P < 0.05). The high-expression VM was related to the depth of infiltration, the histological grade, the lymph node metastasis, and TNM stage in ESCC (P < 0.05). It is suggested that VM was involved in the process of invasion and metastasis of ESCC. The Kaplan-Meier survival analysis showed that the OS and DFS of the positive expression of the VM group in patients with ESCC was significantly lower than that of the negative expression VM group. Moreover, Cox analysis showed that VM might be an independent prognostic factor in patients with ESCC. In summary, VM can be considered as a potential candidate therapeutic target. These are consistent with previous reports [27,29].
Vascular epithelial cells express various members of the cadherin superfamily, particularly VE-cadherin, which is the primary adhesion receptor for endothelial adhesion junctions. In some cancers, Aberrant extra-vascular expression of VE-cadherin has been reported to be involved in the formation of VM [30]. In this study, we found the expression of VE-cadherin in ESCC was significantly higher than that in normal esophageal mucosa (P < 0.05). Similarly, the high expression of VE-cadherin was positively correlated with TNM stage, lymph node metastasis, and histological grade (P < 0.05). Survival analysis showed that the expression of VE-cadherin could be regarded as an independent prognostic factor in patients with ESCC. Furthermore, the OS rate and DFS rate of patients with ESCC were reduced with the decrease of VE-cadherin expression level. These findings demonstrated that the high expression of VE-cadherin promoted the invasion and metastasis of ESCC.
In this experiment, we found that there was a significant positive correlation between VM and VE-cadherin. Liu Z et al. suggested that HIF-1α may influence the formation of VM by regulating the expression of VE-cadherin in esophageal carcinoma [31]. The main function of VE-cadherin in VM formation is to activate EphA2 (epithelial cell kinase) [32-34]. The binding of EphA2 to its ligand results in phosphorylation of EphA2, activates PIK-3 via focal adhesion kinase (FAK) and ERK1/2, and finally, activation of MMP-2 and cleavage of Laminin5r2 form VM [35]. Furthermore, we also confirmed the positive correlation between SOX4 and VM, VE-cadherin in this experiment (P < 0.05).
Current studies have clearly shown that SOX4 can promote epithelial stromal transformation in tumor tissues, including TGF-β signaling pathway [27], Ezh2 [21,36]. It has been pointed out that VM is the process of tumor cell remodeling, which is characterized by the differentiation of epithelial cells (endoderm) into endothelial cells (mesoderm) and the characteristic of obtaining mesenchymal cells. Therefore, it can be speculated that the formation of VM is similar to the process of EMT [37]. We speculate that in ESCC, the high expression of SOX4 promotes the development of EMT, leading to the ability of tumor cells to remodel, the tumor cells with VM structure were separated from the lumen by only one layer of PAS-positive substance. At this time, the tumor cells with reduced adhesion were subjected to both the degradation of basement membrane-like substance by MMP-2 and the impact of blood flow. It is easy to develop lymph node metastasis or even distant metastasis from the primary foci [32].
Conclusions
This study suggests that SOX4, which is highly expressed in ESCC, is involved in the formation of VM in tumor cells. The combined detection of SOX4, VE-cadherin and VM expression can be used as biomarkers for invasion and metastasis of ESCC. These three expressions can be used as powerful prognostic factors in patients with ESCC.
Acknowledgements
This study was supported by grants from the Nature Science Key Program of College and Universityof Anhui Province (NO. KJ2019A0344 and No. KJ2018ZD024), the Nature Science Key Program of Bengbu Medical College (NO. BYKY1817ZD), the Nature Science Technological Development of Bengbu Medical College (NO. BYKF1809), and the National University Students Innovation and Entrepreneurship Projects (No. 201810367004 and No. 201810367016).
Disclosure of conflict of interest
None.
References
- 1.Parkin DM, Bray F, Ferlay J, Pisani P. Global cancers statistics, 2002. CA Cancer J Clin. 2005;55:74–108. doi: 10.3322/canjclin.55.2.74. [DOI] [PubMed] [Google Scholar]
- 2.Pennathur A, Gibson MK, Jobe BA, Luketich JD. Oesophageal carcinoma. Lancet. 2013;381:400–412. doi: 10.1016/S0140-6736(12)60643-6. [DOI] [PubMed] [Google Scholar]
- 3.Ishiguro H, Wakasugi T, Terashita Y, Sakamoto N, Tanaka T, Sagawa H, Okubo T, Takeyama H. Nuclear expression of TCF4/TCF7L2 is correlated with poor prognosis in patients with esophageal squamous cell carcinoma. Cell Mol Biol Lett. 2016;21:5. doi: 10.1186/s11658-016-0006-0. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Maniotis AJ, Folberg R, Hess A, Seftor EA, Gardner LM, Pe’er J, Trent JM, Meltzer PS, Hendrix MJ. Vascular channel formation by human melanoma cells in vivo and in vitro: vasculogenic mimicry. Am J Pathol. 1999;155:739–752. doi: 10.1016/S0002-9440(10)65173-5. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Sun T, Zhao N, Zhao XL, Gu Q, Zhang SW, Che N, Wang XH, Du J, Liu YX, Sun BC. Expression and functional significance of Twist1 in hepatocellular carcinoma: its role in vasculogenic mimicry. Hepatology. 2010;51:545–556. doi: 10.1002/hep.23311. [DOI] [PubMed] [Google Scholar]
- 6.Ngernyuang N, Shao R, Suwannarurk K, Limpaiboon T. Chitinase 3 like 1 (CHI3L1) promotes vasculogenic mimicry formation in cervical cancer. Pathology. 2018;50:293–297. doi: 10.1016/j.pathol.2017.09.015. [DOI] [PubMed] [Google Scholar]
- 7.Liu W, Meng M, Zhang B, Du L, Pan Y, Yang P, Gu Z, Zhou Q, Cao Z. Dehydroeffusol effectively inhibits human gastric cancer cell-mediated vasculogenic mimicry with low toxicity. Toxicol Appl Pharmacol. 2015;287:98–110. doi: 10.1016/j.taap.2015.05.003. [DOI] [PubMed] [Google Scholar]
- 8.Xu XD, Yang L, Zheng LY, Pan YY, Cao ZF, Zhang ZQ, Zhou QS, Yang B, Cao C. Suberoylanilide hydroxamic acid, an inhibitor of histone deacetylase suppresses vasculogenic mimicry and proliferation of highly aggressive pancreatic cancer PaTu8988 cells. BMC Cancer. 2014;14:373. doi: 10.1186/1471-2407-14-373. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Liu WB, Xu GL, Jia WD, Li JS, Ma JL, Chen K, Wang ZH, Ge YS, Ren WH, Yu JH, Wang W, Wang XJ. Prognostic significance and mechanisms of patterned matrix vasculogenic mimicry in hepatocellular carcinoma. Med Oncol. 2011;28:S228–238. doi: 10.1007/s12032-010-9706-x. [DOI] [PubMed] [Google Scholar]
- 10.Zhao H, Gu XM. Study on vasculogenic mimicry in malignant esophageal stromal tumors. World J Gastroenterol. 2008;14:2430–2433. doi: 10.3748/wjg.14.2430. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Zhang J, Zhang G, Hu P, Deng G, Liu Q, Qiao L, Luo H, Zhang J. Vasculogenic mimicry is associated with increased tumor-infiltrating neutrophil and poor outcome in esophageal squamous cell carcinoma. Onco Targets Ther. 2017;10:2923–2930. doi: 10.2147/OTT.S135477. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Delgado-Bellido D, Serrano-Saenz S, Fernández-Cortés M, Oliver FJ. Vasculogenic mimicry signaling revisited: focus on non-vascular VE-cadherin. Mol Cancer. 2017;16:65. doi: 10.1186/s12943-017-0631-x. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Zhao N, Sun BC, Sun T, Ma YM, Zhao XL, Liu ZY, Dong XY, Che N, Mo J, Gu Q. Hypoxia-induced vasculogenic mimicry formation via VE-cadherin regulation by Bcl-2. Med Oncol. 2012;29:3599–3607. doi: 10.1007/s12032-012-0245-5. [DOI] [PubMed] [Google Scholar]
- 14.Liu R, Cao Z, Tu J, Pan Y, Shang B, Zhang G, Bao M, Zhang S, Yang P, Zhou Q. Lycorine hydrochloride inhibits metastatic melanoma cell-dominant vasculogenic mimicry. Pigment Cell Melanoma Res. 2012;25:630–638. doi: 10.1111/j.1755-148X.2012.01036.x. [DOI] [PubMed] [Google Scholar]
- 15.Hendrix MJ, Seftor EA, Meltzer PS, Gardner LM, Hess AR, Kirschmann DA, Schatteman GC, Seftor RE. Expression and functional significance of VE-cadherin in aggressive human melanoma cells: role in vasculogenic mimicry. Proc Natl Acad Sci U S A. 2001;98:8018–8023. doi: 10.1073/pnas.131209798. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.van der Schaft DW, Hillen F, Pauwels P, Kirschmann DA, Castermans K, Egbrink MG, Tran MG, Sciot R, Hauben E, Hogendoorn PC, Delattre O, Maxwell PH, Hendrix MJ, Griffioen AW. Tumor cell plasticity in Ewing sarcoma, an alternative circulatory system stimulated by hypoxia. Cancer Res. 2005;65:11520–11528. doi: 10.1158/0008-5472.CAN-05-2468. [DOI] [PubMed] [Google Scholar]
- 17.Tang NN, Zhu H, Zhang HJ, Zhang WF, Jin HL, Wang L, Wang P, He GJ, Hao B, Shi RH. HIF-1α induces VE-cadherin expression and modulates vasculogenic mimicry in esophageal carcinoma cells. World J Gastroenterol. 2014;20:17894–17904. doi: 10.3748/wjg.v20.i47.17894. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Wang B, Li Y, Tan F, Xiao Z. Increased expression of SOX4 is associated with colorectal cancer progression. Tumour Biol. 2016;37:9131–9137. doi: 10.1007/s13277-015-4756-5. [DOI] [PubMed] [Google Scholar]
- 19.Tao Y, Chai D, Ma L, Zhang T, Feng Z, Cheng Z, Wu S, Qin Y, Lai M. Identification of distinct gene expression profiles between esophageal squamous cell carcinoma and adjacent normal epithelial tissues. Tohoku J Exp Med. 2012;226:301–311. doi: 10.1620/tjem.226.301. [DOI] [PubMed] [Google Scholar]
- 20.Sasaki A, Abe H, Mochizuki S, Shimoda M, Okada Y. SOX4, an epithelial-mesenchymal transition inducer, transactivates ADAM28 gene expression, and co-localizes with ADAM28 at the invasive front of human breast and lung carcinomas. Pathol Int. 2018 doi: 10.1111/pin.12685. [Epub ahead of print] [DOI] [PubMed] [Google Scholar]
- 21.Tiwari N, Tiwari VK, Waldmeier L, Balwierz PJ, Arnold P, Pachkov M, Meyer-Schaller N, Schübeler D, van Nimwegen E, Christofori G. Sox4 is a master regulator of epithelial- mesenchymal transition by controlling Ezh2 expression and epigenetic reprogramming. Cancer Cell. 2013;23:768–83. doi: 10.1016/j.ccr.2013.04.020. [DOI] [PubMed] [Google Scholar]
- 22.Shi Q, Diao Y, Jin F, Ding Z. Anti-metastatic effects of Aidi on human esophageal squamous cell carcinoma by inhibiting epithelial-mesenchymal transition and angiogenesis. Mol Med Rep. 2018;18:131–138. doi: 10.3892/mmr.2018.8976. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Vervoort SJ, van Boxtel R, Coffer PJ. The role of SRY-related HMG box transcription factor 4 (SOX4) in tumorigenesis and metastasis: friend or foe? Oncogene. 2013;32:3397–3409. doi: 10.1038/onc.2012.506. [DOI] [PubMed] [Google Scholar]
- 24.Tao Y, Chai D, Ma L, Zhang T, Feng Z, Cheng Z, Wu S, Qin Y, Lai M. Identification of distinct gene expression profiles between esophageal squamous cell carcinoma and adjacent normal epithelial tissues. Tohoku J Exp Med. 2012;226:301–11. doi: 10.1620/tjem.226.301. [DOI] [PubMed] [Google Scholar]
- 25.Han R, Huang S, Bao Y, Liu X, Peng X, Chen Z, Wang Q, Wang J, Zhang Q, Wang T, Zheng D, Yang W. Upregulation of SOX4 antagonizes cellular senescence in esophageal squamous cell carcinoma. Oncol Lett. 2016;12:1367–1372. doi: 10.3892/ol.2016.4799. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Thiery JP, Acloque H, Huang RY, Nieto MA. Epithelial-mesenchymal transitions in development and disease. Cell. 2009;139:871–890. doi: 10.1016/j.cell.2009.11.007. [DOI] [PubMed] [Google Scholar]
- 27.Zhang J, Liang Q, Lei Y, Yao M, Li L, Gao X, Feng J, Zhang Y, Gao H, Liu DX, Lu J, Huang B. SOX4 induces epithelial-mesenchymal transition and contributes to breast cancer progression. Cancer Res. 2012;72:4597–4608. doi: 10.1158/0008-5472.CAN-12-1045. [DOI] [PubMed] [Google Scholar]
- 28.Folkman J. What is the evidence that tumors are angiogenesis dependent? J Natl Cancer Inst. 1990;82:4–6. doi: 10.1093/jnci/82.1.4. [DOI] [PubMed] [Google Scholar]
- 29.Wang F, Li XK, Xu HY, Shan ZZ, Wang T, Yang ZC, He W, Wang LX, Fan QX. N-cadherin participated in invasion and metastasis of human esophageal squamous cell carcinoma via taking part in the formation of vasculogenic mimicry. Med Oncol. 2015;32:480. doi: 10.1007/s12032-014-0480-z. [DOI] [PubMed] [Google Scholar]
- 30.Delgado-Bellido D, Serrano-Saenz S, Fernández-Cortés M, Oliver FJ. Vasculogenic mimicry signaling revisited: focus on nonvascular VE-cadherin. Mol Cancer. 2017;16:65. doi: 10.1186/s12943-017-0631-x. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.Liu Z, Sun B, Qi L, Li H, Gao J, Leng X. Zinc finger E-box binding homeobox 1 promotes vasculogenic mimicry in colorectal cancer through induction of epithelial-to-mesenchymal transition. Cancer Sci. 2012;103:813–820. doi: 10.1111/j.1349-7006.2011.02199.x. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32.Fan YZ, Sun W. Molecular regulation of vasculogenic mimicry in tumors and potential tumor-target therapy. World J Gastrointest. 2010;2:117–127. doi: 10.4240/wjgs.v2.i4.117. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33.Hess AR, Seftor EA, Gardner LM, Carles-Kinch K, Schneider GB, Seftor RE, Kinch MS, Hendrix MJ. Molecular regulation of tumor cell vasculogenic mimicry by tyrosine phosphorylation: role of epithelial cell kinase (Eck/EphA2) Cancer Res. 2001;61:3250–3255. [PubMed] [Google Scholar]
- 34.Margaryan NV, Strizzi L, Abbott DE, Seftor EA, Rao MS, Hendrix MJ, Hess AR. EphA2 as a promoter of melanoma tumorigenicity. Cancer Biol Ther. 2009;8:279–288. doi: 10.4161/cbt.8.3.7485. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35.Hess AR, Margaryan NV, Seftor EA, Hendrix MJ. Deciphering the signaling events that promote melanoma tumor cell vasculogenic mimicry and their link to embryonic vasculogenesis: role of the Eph receptors. Dev Dyn. 2007;236:3283–3296. doi: 10.1002/dvdy.21190. [DOI] [PubMed] [Google Scholar]
- 36.Koumangoye RB, Andl T, Taubenslag KJ, Zilberman ST, Taylor CJ, Loomans HA, Andl CD. SOX4 interacts with EZH2 and HDAC3 to suppress microRNA-31 in invasive esophageal cancer cells. Mol Cancer. 2015;14:24. doi: 10.1186/s12943-014-0284-y. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 37.Fan YL, Zheng M, Tang YL, Liang XH. A new perspective of vasculogenic mimicry: EMT and cancer stem cells. Oncol Lett. 2013;6:1174–1180. doi: 10.3892/ol.2013.1555. [DOI] [PMC free article] [PubMed] [Google Scholar]