To explore the molecular mechanism by which mitochondrial matrix import factor 23 (MIX23) affects triple-negative breast cancer (TNBC).

The RNA-Seq expression profile and corresponding clinical data of breast cancer patients were downloaded from The Cancer Genome Atlas (TCGA) database, including 1 097 primary tumor tissue samples and 114 adjacent normal tissue samples. Immunohistochemistry staining images of MIX23 in breast cancer tissues were obtained from the Human Protein Atlas (HPA) database for quantitative analysis of protein expression levels. Expression profile data of the METABRIC breast cancer cohort were downloaded as an independent validation set. Based on the mean expression level of MIX23 in breast cancer patients, the TCGA and METABRIC datasets were each stratified into high- and low-expression groups (TCGA: 480 cases in high-expression group, 534 in low-expression group; METABRIC: 969 in high-expression group, 1,011 in low-expression group). Kaplan-Meier method was used to plot survival curves, and log-rank test was used to evaluate the difference in survival rate between the two groups. RT-qPCR and Western blot were used to detect the mRNA and protein expression of MIX23 in TNBC cell line. Lentiviral vector was used to construct MDA-MB-231 cells with stable MIX23 knockdown (sh-MIX23 group), with the cells transfected with lentiviral vectors carrying non-targeted scramble shRNA as control group. Cell clone formation assay, CCK-8 assay and cell scratch assay were adopted to detect cell proliferation and migration abilities. Differential gene analysis was performed based on TCGA database to identify downstream target genes.

In the TCGA cohort, the mRNA expression level of MIX23 was significantly higher in breast cancer tissues than in adjacent normal breast tissues (7.756±0.605 vs 7.145±0.341, t=16.613, P<0.001). HPA cohort analysis also showed that MIX23 expression was significantly higher in tumor tissues than in normal tissues (43.978±4.158 vs 55.211±7.339, t=-2.980, P=0.018). The expression levels of MIX23 in TNBC, HER-2 positive, Luminal A and Luminal B subtypes were 8.154±0.743, 7.858±0.628, 7.613±0.535 and 7.744±0.477, respectively, with significant difference (F=28.260, P<0.0001). Kaplan-Meier survival analysis revealed that in the TCGA cohort, high MIX23 expression was significantly associated with shorter disease-free interval (HR=1.667, 95%CI: 1.087–2.555, P=0.018). After MIX23 knockdown, compared with control group, the colony formation ability of cells in the sh-MIX23 group was significantly decreased (184.5±9.576 vs 110.2±6.976, t=12.578, P<0.001), the cell proliferation was markedly inhibited at 24, 48 and 72 h (0 h: 0.185±0.009 vs 0.182±0.007, t=0.515, P=0.626; 24 h: 0.399±0.057 vs 0.246±0.004, t=4.333, P=0.006; 48 h: 0.704±0.058 vs 0.403±0.021, t=9.642, P<0.001; 72 h: 0.999±0.0416 vs 0.452±0.031, t=21.110, P<0.0001). The relative migration area of cells in the sh-MIX23 group was significantly smaller than that in the control group at 24 h after scratching (62.861±12.252 vs 35.971±8.024, t=5.193, P<0.01). Compared with the control group, MIX23 could up-regulate the expression of osteoglycin (1.195±0.254 vs 1.958±0.223, t=-3.945, P=0.018) and inhibit the expression of Akt, mTOR and EGFR (1.000±0.023 vs 0.054±0.006, t=68.352, P<0.0001; 1.000±0.052 vs 0.058±0.007, t=31.092, P<0.0001; 1.000±0.094 vs 0.036±0.004, t=17.785, P<0.0001).

MIX23 is highly expressed in breast cancer and correlated with poor prognosis. Knockdown of MIX23 affects the growth and metastasis of breast cancer by up-regulating osteoglycin and inhibiting the downstream EGFR/Akt/mTOR signaling pathway.