TFEB, SIRT1, CARM1, Beclin-1 expression and PITX2 methylation in breast cancer chemoresistance: a retrospective study

Abstract

Background: Breast cancer chemoresistance is attributed to a wide variety of mechanisms, including autophagy. Transcription factor EB (TFEB) has been recently identified and characterized as one major regulator of autophagy and lysosomal genesis.

Objective: This study aims to evaluate the prognostic impact of TFEB and its pathway in breast cancer chemoresistance.

Methods: This retrospective study analyzes the expression of TFEB, CARM1, SIRT1, and Beclin-1 and the methylation of PITX2 in breast carcinoma. A group of breast cancer patients treated with chemotherapy, who relapsed within 12 months from treatment initiation, were compared to a sub-cohort of chemo-treated patients who did not recur within 12 months of follow-up. The expression of TFEB, CARM1, SIRT1, and Belcin-1 was analyzed using immunohistochemistry or RT-PCR on formalin-fixed paraffin-embedded samples. PITX2 methylation was tested with the diagnostic CE-marked kit Therascreen PITX2 RGQ PCR. In the final model, 136 cases of chemo-treated breast cancer were included.

Results: A higher TFEB and Beclin-1 expression correlate with shorter survival in patients with chemo-treated invasive breast cancer (respectively HR 3.46, CI.95 1.27-9.47, p < 0.05 and 7.11, CI.95 2.54-19.9). TFEB, CARM1, and SIRT1 are positively correlated with Beclin-1. The protein expression of SIRT1 is significantly associated with TFEB and CARM1 so that a very low SIRT1 expression (lower than the first quartile of the H-score distribution) correlates with a low expression of TFEB and CARM1 and with longer survival. SIRT1 seems to have a lower H-score in the basal-like and HER2-enriched tumors than the luminal subtypes. Beclin-1 and TFEB seem to have a higher H-score in the basal-like and HER2-enriched tumors than the luminal subtypes. PITX2 methylation analysis was feasible only in 65% of the selected samples, but no significant differences between cases and controls were found, and there was also no correlation with the expression of the TFEB pathway.

Conclusions: TFEB, SIRT1, and Beclin-1 seem to have a potential prognostic significance in patients with chemo-treated breast cancer, likely because of their role in the regulation of autophagy. In addition, no correlation between TFEB and PITX2 methylation was found, likely because they perform two different roles within the autophagy process.

Keywords: Beclin-1; Breast cancer; CARM1; Chemoresistance; PITX2; SIRT1; TFEB.

Fig. 1

Immunohistochemical staining. A Image at 20x (and in the box at 40x) of primary breast cancer tissue TFEB immunohistochemical staining in a patient with breast cancer recurrence within 12 months of follow-up. B Image at 20x (and in the box at 40x) of TFEB immunohistochemical staining in a patient’s primary breast cancer tissue without recurrence within 12 months of follow-up. C 20x image (and in the 40x square) of CARM1 immunohistochemical staining in primary breast cancer tissue of a woman with recurrent breast cancer within 12 months of follow-up. D 20x image (and in the 40x frame) of CARM1 immunohistochemical staining in a primary breast cancer specimen of a patient which did not recure within 12 months of follow-up. E Image at 20x (and in the box at 40x) of SIRT1 immunohistochemical staining in a primary breast cancer specimen of a woman with recurrent breast cancer within 12 months of follow-up. F Image at 20x (and in the box at 40x) of SIRT1 immunohistochemical staining in a primary breast cancer specimen of a woman without tumor recurrence within 12 months of follow-up. G Image at 20x (and in the box at 40x) of Beclin-1 immunohistochemical staining in a primary breast cancer specimen of a woman with recurrent breast cancer within 12 months of follow-up. H Image at 20x (and in the box at 40x) of Beclin-1 immunohistochemical staining in a primary breast cancer specimen of a woman without tumor recurrence within 12 months of follow-up

Fig. 2

Kaplan-Meier survival curves and p-values refers to Log-rank tests. A Curve based on TFEB expression (high expression consists in an H-score greater than the distribution median [> 185] and low expression in an H-score lower or equal to the distribution median) (p < 0.05). B Analysis based on CARM1 expression (i.e., high with an H-score higher than the median distribution [> 85] or low, i.e., with an H-score lower than or equal to the distribution median) (p = 0.156). C Analysis of survival based on SIRT1 expression (i.e., high with an H-score higher than the median distribution [190] or low, i.e., with an H-score lower than or equal to the distribution median) (p = 0.322). D Analysis of survival based on the expression of Beclin-1 (i.e., high with an H-score higher than the median distribution [> 90] or low, i.e., with an H-score lower than or equal to the distribution median) (p < 0.05). E Analysis of survival based on the expression of SIRT1 with a different cut-off (i.e., high with an H-score higher than the first quartile (*) of the distribution [> 100] or low, i.e., with an H-score lower than or equal to the first quartile of the distribution) (p < 0.05). F Subgroup analysis of survival (luminal A, luminal B, and luminal Her sub-types) based on the expression of SIRT1 (i.e., high with an H-score higher than the first quartile (*) of the distribution [132] or low, i.e., with an H-score lower than or equal to the first quartile of the distribution) (p = 0.055). Panel G: Subgroup analysis of survival (Her enriched and Basal-like sub-types) based on the expression of SIRT1 (i.e., high with an H-score higher than the first quartile (*) of the distribution [132] or low, i.e., with an H-score lower than or equal to the first quartile of the distribution) (p = 0.294)