Introduction
The Reproducibility Project: Cancer Biology (RP:CB) is a collaboration between the Center for Open Science and Science Exchange that seeks to address concerns about reproducibility in scientific research by conducting replications of selected experiments from a number of high-profile papers in the field of cancer biology (Errington et al., 2014). For each of these papers, a Registered Report detailing the proposed experimental designs and protocols for the replications was peer reviewed and published prior to data collection. The present paper is a Replication Study that reports the results of the replication experiments detailed in the Registered Report (Phelps et al., 2016) for a paper by Tay et al., 2011 and uses a number of approaches to compare the outcomes of the original experiments and the replications.
In 2011, Tay et al. reported PTEN was modulated through competing endogenous RNAs (ceRNAs), which are protein-coding RNA transcripts that compete for microRNAs through common micoRNA response elements (MREs). Testing four candidate PTEN ceRNAs Tay and colleagues reported that for three of these candidate ceRNAS (SERINC1, VAPA, and CNOT6L) silencing the ceRNAs impacted the activity of a luciferase construct engineered with the 3’UTR of PTEN (Tay et al., 2011). Using the same reporter construct, overexpression of the 3’UTRs of these ceRNAs were reported to increase luciferase activity suggesting inhibition of PTEN 3’UTR was relieved (Tay et al., 2011). These effects were reported to be dependent on microRNAs since inhibition of PTEN protein expression when ceRNAs were depleted was abrogated when DICER, a key part of the microRNA machinery, was disrupted (Tay et al., 2011). Two ceRNAs, VAPA and CNOT6L, when depleted also resulted in increased cell proliferation and phosphorylation of AKT when depleted, which was attenuated when DICER was disrupted (Tay et al., 2011).
The Registered Report for the paper by Tay et al., 2011 described the experiments to be replicated (Figures 3C-D, 3G-H, 5A-B, and Supplemental Figures S3A-B), and summarized the current evidence for these findings (Phelps et al., 2016). Since that publication additional studies have reported finding other ceRNAs of PTEN. TNRC6B was identified as a ceRNA of PTEN with depletion of TNRC6B reported to decrease PTEN mRNA and protein expression in the prostate cancer cell lines DU145, 22RV1, and BM1604 and increase cell proliferation in DU145 and PC3 cells (Zarringhalam et al., 2017). Zarringhalam et al., 2017 also used CNOT6L in their study and reported depletion of CNOT6L produced similar results as TNRC6B depletion. DNMT3B and TET3 were recently reported to be ceRNAs of PTEN with miR-4465 identified as a microRNA regulating these three transcripts via their 3’UTRs (Roquid et al., 2019). Multiple studies have reported a growing list of potential ceRNAs, which includes mRNAs, pseudogenes, lncRNAs, and circRNAs (Gebert and MacRae, 2019; Li et al., 2018; Wang et al., 2016; Yang et al., 2016). For example, the lncRNA BGL3 has been identified as a ceRNA for PTEN to regulate Bcr-Abl-mediated cellular transformation in chronic myeloid leukemia (Guo et al., 2015) and c-Myc has been reported as a potential ceRNA for PML/RARα in acute promyelocytic leukemia (Ding et al., 2016). Prediction of putative ceRNAs are being reported using a variety of computational methods and data sources that construct ceRNA interaction networks (Chen et al., 2018; Chiu et al., 2017; Feng et al., 2019; Liu et al., 2019; Park et al., 2018b; Song et al., 2016; Sun et al., 2016; Swain and Mallick, 2018; Wang et al., 2019; Yue et al., 2019). Additionally, recently it has been reported that 3’UTR shortening of transcripts of predicted ceRNAs could be a potential mechanism of repressing tumor-suppressor genes, including PTEN, in trans by disrupting ceRNA cross-talk (Park et al., 2018a).
At the same time, whether physiological changes of ceRNAs can modulate microRNA activities remains controversial (Cai and Wan, 2018; Thomson and Dinger, 2016). Experiments designed to test the feasibility of the ceRNA hypothesis, have reported that microRNA-binding sites are generally much higher than the number of microRNA molecules (Denzler et al., 2014). This would suggest that under physiological conditions ceRNA perturbation would likely lead to a change too small to be detected and to produce biological consequences (Broderick and Zamore, 2014; Denzler et al., 2014). Mullokandov and colleagues reported that only the most abundant microRNAs mediate target suppression as over 60% of detected microRNAs have no discernable activity (Mullokandov et al., 2012). Bosson and colleagues suggested that the microRNA-target ratios determined the respective susceptibility to ceRNA-mediated regulation (Bosson et al., 2014). This model has been further examined and Denzler and colleagues reported that while microRNA levels did not affect site competition, they defined microRNA-mediated repression (Denzler et al., 2016). The experimental strategies currently used for studying the ceRNA hypothesis are also limited, especially when attempting to represent the in vivo levels of endogenous RNAs (Cai and Wan, 2018; Jens and Rajewsky, 2015; Thomson and Dinger, 2016).
The outcome measures reported in this Replication Study will be aggregated with those from the other Replication Studies to create a dataset that will be examined to provide evidence about reproducibility of cancer biology research, and to identify factors that influence reproducibility more generally.