Model

Colorectal cancer is a leading cause of cancer-related mortality worldwide. This project focuses on developing a novel targeted therapy by inhibiting the TEAD4 gene, a key regulator in the Hippo signaling pathway that plays a significant role in colorectal cancer progression. We have designed two plasmids that efficiently suppress TEAD4 expression. Our results indicate that elevated levels of TEAD4-targeting siRNAs in the colorectal cancer cell line SW480 lead to inhibition of tumor cell proliferation and migration, as well as increased reactive oxygen species (ROS) expression.

We have transfected different amount of shRNA plasmids targeted TEAD4 into SW480 cell. Based on the proliferation and migration speed, and ROS expression, we proposed a statistical model. The construction of model involves several steps, including data collection, data preprocessing and evaluation.

In this experiment, SW480 cells were transfected with sh-TEAD4-1 and sh-TEAD4-2 plasmids at varying dosages (0 µg, 0.5 µg, 1 µg, 2 µg), and CCK-8 experiment was used to exam the proliferation speed, the transwell assays was used to detect the migration speed, and intensity of fluorescence was used to quantitatively assess the ROS levels within the cells.

Firstly, we evaluated the global levels of cell proliferation, cell migration and the ROS expression of each group. As shown in the boxplot of Figure 1 to Figure 3, both the plasmid groups showed a significantly higher or lower level of each cellular phenotype compared to the control group. These results confirmed that the plasmid interference was successful.

Next, the performance of the sh-TEAD4-1 plasmid and the sh-TEAD4-2 plasmid was compared. In our data, we observed that the sh-TEAD4-2 plasmid outperformed the sh-TEAD4-1. In the Figure 1, the sh-TEAD4-2 plasmid group showed a lower level of cell proliferation compared to the sh-TEAD4-1 plasmid group. In the Figure 2, the sh-TEAD4-2 plasmid group showed a decreased cell migration compared to the sh-TEAD4-1 plasmid group. In the Figure 3, the sh-TEAD4-2 plasmid group showed an increased level of ROS expression compared to the sh-TEAD4-1 plasmid group. In summary, these data confirmed that sh-TEAD4-2 plasmid performed better during the interference.

To evaluate the predicting effects of the shRNA efficiency, we performed ordinal logistic regression analysis. The analysis used plasmid amount as dependent variable. The results revealed the cell proliferation level (Transwell), cell migration level (CCK8) and ROS expression (ROS) were significantly associated with the dependent variable (Table 1), which means the cell proliferation, cell migration and ROS expression were predictable. Therefore, the statistical model using the coefficient estimate from the Cox regression was defined as:

Y = - 0.0014 x Transwell - 11.5056 x CCK8 + 0.0673 x ROS

Table 1. ordinal logistic regression analysis of plasmid amount

Furthermore, we used dosage as dependent variable to construct the statistical model of multiple logistic regression analysis. The results confirmed the significant association of the cellular phenotypes with the dependent variable. Besides, we observed the dosage 1 and 2 µg showed a higher coefficient estimate, which suggested the dosage 1 and 2 µg have a better performance. The cell migration (CCK8) also showed a higher coefficient estimate compared to other cellular phenotypes, suggesting the cell migration have a better performance of prediction.

Table 2. multivariate Cox regression analysis of plasmid amount

We evaluated the global levels of cell proliferation, cell migration and the ROS expression of different groups (NLS, siRNA, and sh-TEAD4-1) in the boxplot of Figure 4 to Figure 6. In our data, we observed that the NLS group outperformed other groups. In the Figure 4, the NLS group showed a lower level of cell proliferation compared to other groups. In the Figure 5, the NLS group showed a decreased cell migration compared to other groups. In the Figure 6, the NLS group showed an increased level of ROS expression compared to other groups. In summary, these data confirmed that NLS groups have the best effect of inference.

To evaluate the predicting effects of the three types of inference, we performed ordinal logistic regression analysis. The analysis used the dosage as the dependent variable. The results revealed the cell proliferation level (Transwell), cell migration level (CCK8) and ROS expression (ROS) were significantly associated with the dependent variable (Table 3), which means the cell proliferation, cell migration and ROS expression were predictable.

Table 3 ordinal logistic regression analysis of plasmid amount

These findings suggest that this RNA interference vector could serve as a promising therapeutic candidate for colorectal cancer treatment. Our statistical model revealed that these cellular phenotypes was significantly associated with the colorectal cancer treatment in accuracy and robustness. NLS methods presented better therapy method in colorectal cancer compared with siRNA and shRNA targeted to TEAD4.