Circular RNAs (circRNAs) have been identified as potential biomarkers for many cancer, including colon cancer (CC). However, the function and mechanism of circPPP1R12A in CC have not been fully elucidated. Quantitative realtime PCR was employed to assess the expression of circPPP1R12A, microRNA (miR)-375 and catenin beta-1 (CTNNB1). The proliferation, apoptosis, migration and invasion of cells were determined using colony formation assay, flow cytometry, wound healing assay and transwell assay. The protein levels of cell cyclin-related markers and CTNNB1 were detected by western blot analysis. The interaction between miR-375 and circPPP1R12A or CTNNB1 was verified by dual-luciferase reporter assay. Xenograft models were built to evaluate the effect of circPPP1R12A silencing and CTNNB1 overexpression on CC tumor growth in vivo. Our results showed that circPPP1R12A was a highly expressed circRNA in CC tissues and cells. Silenced circPPP1R12A suppressed the proliferation, promoted the apoptosis, and inhibited the migration and invasion of CC cells. MiR-375 could be sponged by circPPP1R12A, and its inhibitor could reverse the inhibition of circPPP1R12A silencing on CC progression. Furthermore, CTNNB1 was a target of miR-375, and its overexpression also abolished the suppression of miR-375 on CC progression. Moreover, circPPP1R12A indirectly regulated CTNNB1 expression by sponging miR-375. Importantly, circPPP1R12A knockdown reduced the tumor growth of CC in vivo, and this effect also could be reversed by overexpressing CTNNB1. Our study proposed that circPPP1R12A might play an oncogenic role in CC, which could act as a potential therapeutic target for CC.

No abstract

Currently, state-of-the-art semi-supervised learning (SSL) segmentation methods employ pseudo labels to train their models, which is an optimistic training manner that supposes the predicted pseudo labels are correct. However, their models will be optimized incorrectly when the above assumption does not hold. In this paper, we propose a Pessimistic Consistency Regularization (PCR) which considers a pessimistic case that pseudo labels are not always correct. PCR makes it possible for our model to learn the ground truth (GT) in pessimism by adaptively providing a candidate label set containing K proposals for each unlabeled pixel. Specifically, we propose a pessimistic consistency loss which trains our model to learn the possible GT from multiple candidate labels. In addition, we develop a candidate label proposal method to adaptively decide which pseudo labels are provided for each pixel. Our method is easy to implement and could be applied to existing baselines without changing their frameworks. Theoretical analysis and experiments on various benchmarks demonstrate the superiority of our approach to state-of-the-art alternatives.