Hypothesis:

By compilation of available evidence, Yook et al state that through hyperactive Wnt signalling, β-catenin-T-Cell factor [TCF] triggers epithelial mesenchymal transition [EMT] in human breast cancer cells. They then hypothesise that canonical Wnt signalling can result in tumour cell dedifferentiation and tissue-invasive activity - through an Axin2-dependent pathway. They aim to demonstrate this can occur by the stabilisation of the zinc transcription factor, Snail1 that they state plays an imperative regulation role of EMT programmes.

Comment: The context of the experiment including present-day scientific evidence and present-day understanding was well established on the onset. However, the opening paragraph is a somewhat convoluted blend of the author’s experimental aims supported by an array of scientific facts. It’s hard to discern which of these were deduced prior-to and which post- their experimentation.
A description of the experiments and hypotheses being tested: including the controls used, number of experimental replicates and the statistical analysis used. A critique of how effectively the data is presented. Finally, does the experimental data really show what the authors say it does?

Figure 1 is subdivided into four panels: 1a-1d – each visually recording results of experiments investigating particular questions that collectively challenged the hypothesis to test its validity.

Microscopy results of mock-transfected fluorescently-labelled MCF-7 cells and Wnt expressing tumour MCF-7 cells cultured on the vascularised epithelial-stromal tissue of the chorioallantoic membrane [CAM] of 11-day-old chick embryos. The authors hypothesised the mock-transferred MCF-7 cells would remain confined to the upper CAM surface as they were unable to express a tissue-invasive phenotype as expected with the Wnt expressing tumour cells [able to invade the stromal matrix]. These two variants acted as the control and tested cell lines respectively. Both the invasion and depth of invasive front were recorded using a mean of three repetitions taking into account the standard error of the mean [representing the spread of the mean of the sample i.e. an indication of how accurate the mean is]. To test the true statistical significance however, the P value was also shown – providing clear quantitative data. The advantages of confocal microscopy were used maximally – it was easy to distinguish between cells, distinguish the CAM surface and follow the points of invasion through. Even more so, the proposed down regulation of E-cadherin by the cancerous cells can be seen extremely well by the loss of the red-occupied area. One may argue, however, that the images showing β-catenin shifting into the nuclear zone aren’t very clear as the nuclear border hasn’t been defined in these single-colour images – somewhat questioning their conclusion drawn from this.

Next the authors aim to show that Wnt-1-transduced MCF-7 cell activity is linked to increased Snail1 protein levels as well as the repression of E-cadherin transcription. With actin included as a loading control, a section of a western blot is presented clearly showing the induction of nuclear Snail1 proteins in MCF7-Wnt cells [compared to the mock]. Only the best results have been presented and no repetitions are shown. Representing the down-regulation of E-cadherin, a bicolour vertical bar chart with confidence intervals is plotted – again recorded using a mean of three repetitions taking into account the standard error of the mean. This shows that the authors recognised the underlying uncertainty - failure to do this would likely lead to misunderstanding and bias, thus reducing data reliability. However, it is unclear why E-box mutants were included in this graph as there was no subsequent discussion about these results.

To determine whether the Wnt-initiated EMT programme can be mediated by the canonical pathway, MCF-7 cells were transfected with a dominant-negative TCF-4 construct. This was an extremely effective method as the gene product would adversely affect the wild-type gene product. As expected E-cadherin repression was relieved and invasion completely suppressed. Diagrammatically this is clear using another confocal-microscopy image that allows direct comparison to the MCF-7-mock result to deduce that in fact Wnt-TCFDN conditions mimicked the MCF-7-mock.

Figure 1c boasts all the same qualities of image 1a - but contains nice use of asterisks to differentiate between the P-values versus mock/siRNA experiments. The diagram effectively supports the hypothesisthat β-cateninS33Y-induced CAM invasion is blocked by Snail1 siRNA or shRNA. Although the confocal-microscopy images support this case for siRNA, there is no data provided to support the shRNA claim made. The controls contained no RNA and again the mean of three repetitions was taken.