Modulation of E-Cadherin Function through the AmotL2 Isoforms Promotes Ameboid Cell Invasion

Round 1

Reviewer 1 Report

In the manuscript Subramani et al. submitted to Cells the authors investigate the role of a short isoform of the Angiomotin Like 2 protein, discovered previously by the same group, in collective cell behavior. The authors document certain phenotypes associated with defects in cell collectiveness, which could be caused either by the depletion of the long p100 isoform of AmotL2 or by the overexpression of the short one, p60. The p60 isoform is shown to interact to its long counterpart, thus blocking its functioning. The most striking observation were done in the organoid culture, where the short isoform induced cell scattering, when being overexpressed in cells. At the same time, the scattering of invasive cell lines expressing this isoform can be blocked by the depletion of p60.

Many observation in the manuscript are too preliminary to draw conclusions. The methods are not fully described, the manuscript contains many inaccuracies and misses some important controls.

I cannot recommend the publication of this manuscript in its present form.

1_Did authors confirm at least once any of the phenotypes obtained upon shRNA-mediated depletion of AMOTL2 by using another shRNA sequence or by performing a rescue?

2_Authors claim, that they could specifically deplete the short p60 isoform of AMOTL2 by using an RNA sequence targeting the junction between the exons 8 and 9. I don’t understand the specificity since the long isoform also contains exon 8 and 9. The bigger isoform is said to contain an additional triplet CAG between exon 8 and 9. Where does it come from ? Does it translate into a Glutamine? All this is not clear.

3_Why does the Dox concentration used in the experiments varies from unreasonably high (5µg/ml) to as low as 100ng/ml (line 558) ? What does it mean “…the cells were induced for the mutation with 100ng/ml of doxycycline…” (lines 557-558).

4_In many figures it is indicated AmotL2, when the authors imply the longer isoform. This is confusing.

5_Many methods are missing. For example, how nuclear oscillation in 2E-H, cell scattering and invasion in 2B, 3B, laminAC/B ratio in 5D, nuclear solidity in 5F etc… were assessed? In section 4.3, the protocol for IP is not written at all.

6_The figure legends are inaccurate and lack some important information.

-        The lengths of a scale bar are not referred in many cases (1E, 3C,D and the others) There are problems with color scale bars, see the comments below.

-        What is shown on the graphs, mean±what ? What was the N? Figures 2D-G, 5B,D and the others. For the box and whiskers plot N is also not indicated.

-        Statistical tests specific for each dataset are not indicated in the figure legends.

-        On the graph with intensity profiles, numbers cannot be seen (1F, S1E, S1G).

7_Who is A.R. in the section Author contribution (lines 640-643)? The first author Aravindh Subramani ?

8_Figure S6 is referred (line 144), but it does not exist.

9_Figure 1 is one of the most confusing

How do the authors explain the IP data in 1B-D ? Specifically, how does binding of p100 to p60 abolish the interaction of p100 to actin and E-cadherin, if the binding sites do not overlap?

In Fig. 1B, the long isoform was precipitated from the cells, but the authors do not give any detail of the IP experiments in the Methods section. Was it an endogenous precipitation with the Amotl2 p100-specific antibodies? In this case, which antibody was used for the staining of WB, regarding, that we can see both isoforms in the input as well as in IP ?

Line 134 Authors claim , that they use p100 AmotL2-specific antibody for the staining in 1E, whereas in the other cases they probably use the antibody reacting with both isoforms. In the Materials in Method section only one antibody is mentioned (lines 459-460).

What can be concluded from the data in 1E-F is not clear. When looking at the images, as well as the profiles of intensity, one can observe, that the total intensity of Amotl2 p100 is increasing dramatically, what contradicts with the WB in 1B, where the level of p100 remains unaltered upon p60 induction. At the same time the intensity of p100 at the cell-cell junctions is increasing, that is clear from the line scans (it increasing from 30 a.u. to as much as 100 a.u. at the peaks). Therefore the specificity of the antibody to the longest isoform is doubtful. It needs to be demonstrated to exclude the antibody binding to the short isoform.

By the way, what are the arbitrary units on Y-axis? Gray scales from a 8-bit image? Along which line the profile of intensity was measured? Numbers on the Y-axis cannot be seen. The same applies to the results in Figure S1. The numbers are also not seen there.

Overall, this part is weakly performed and poorly presented as one line scans from one image does not suffice to draw any conclusions. Better images could be obtained and the intensities of AmotL2 staining could be assessed quantitatively in several cells-cells junctions from different field of view. Also, both isoform could be tagged independently to reveal the shifting of the longer isoform from the cell junction, what the authors strived to demonstrate.

10_Figure 2

What are the radial fibers, that disassemble upon AmotL2 depletion or overexpression of p60? They look like just regular stress fibers spanning throughout a cell body.

It this confocal microscopy? If so, what do we see on the images, max intensity projections ?

What is AmotL2 staining in green (2A)? Is it E-cadherin as mentioned in the legend?

There is an issue with the comparison of cell areas in 2C. Cell areas can be compared either in a sparse culture or in a fully jammed monolayer. On the image one can observe cell monolayers reaching different densities. How can authors make sure, that in all conditions cells reached the jammed state?

The image in A (AmotL2 depletion) does not support the quantifications in 2G. Most of the cells on the image seem to have 5 or 6 neighbors.

Lines 165-167 “D-G) Quantification of cell geometry, as analyzed by the number of neighboring cells; ***p < 0.001.” Why p is mentioned there, as nothing is shown on the Figures D-G? 

11_Figure 3

What do we see in in 3A as if the cells are stained, but it seems, that this is not the case according to the figure legend?

In Fig 3C nothing is seen, the black is saturated. The way it is depicted in S3C is much better for a visual impression.

3D, lower panel, actin is overexposed, the color of the nuclei does not correspond to the one in the legend.

3E The analysis of the nuclear oscillations of is not detailed enough. What do the symbols in E,F,H mean? Different nuclei? Only 10 nuclei per experiment were assessed?

In the text It is not clear why do authors talk about angiogenesis, while doing experiments mostly on MDCK cells (lines202-203)

Lines 211-212, we cannot conclude that, that the increased oscillations of nuclei are due to disconnection from the actin cytoskeleton

12_Figure 4

In 4A some colors are saturated

CD Images do not fully support the quantifications as 8 or 16 cell on average are scattered from the bulk in the case of AmotL2-p60 depletion or in the control, respectively, while in C we see either “normal” cyst with no scattering at all or a fully scattered one.

Perhaps, this criterion used in 4B and D is not the best to characterized the observed phenotypes.

13_Figure 5

C DAPI is not visible at all. Lamin A is visible only in the 2 leader cells in the control.

Actin is saturated at the left. At the right it is not clear, how the stalk is defined as everything is scattered. Therefore, who is the leader on the graph 5D in the +Dox condition?

D Author do not comment, why to use Lamin AC to B ratio. One can suggest, that the levels of Lamin B should be stable, regarding its more pronounced role in cell physiology. Where is the Lamin B staining in this case? From this part we can likely conclude only, that the level of lamin A is homogenized upon p60 overexpression due to the lack of organized sprouts of cells and therefore, lack of leaders.

Lines 245-246 Collagen fiber length cannot be even considered as an indirect measure of applied force.

Lines 255-256 “The data in Figure 3 indicated that p60AmotL2 uncouples the nucleus from cellular actin filaments” it can not be stated as such.

14_Figure6

6A In all conditions except the control one, the Lamin A staining is lost, that should not be the case al least for the Ctrl (+Dox) condition.

How come the actin fibers go through the nuclei? Are these images max intensity projections of confocal stacks?

3E-F How is it possible, that AmotL2 co-precipitates with Lamin A, regarding the fact, that the actin filaments, connecting the nucleus to the plasma membrane, even if they exist, would be completely disassembled upon all IP conditions ? Again, the IP protocol is not provided in the Methods section.

Lines 272 “We evaluated the nuclear lamina and cytoskeletal interactions using immunofluorescent staining of actin and Lamin A/C”. Interactions cannot be evaluated by an IF staining

Line 282 “Images were processed by ImageJ, using nano J SRRF module” No details of the analysis are provided.

Lines 293-295 “In summary, our data suggest a direct connection between E-cadherin in the outer plasma membrane and the nuclear lamina (Schematic illustrated in Figure 6F)”. This cannot be proven, based on the data provided.

15_Figure 7

It is not clear, how FRET of Nesphrin is connected to the previous data.

7B images are not well aligned, some cropped parts left on the top in the 2 -Dox conditions

Why the scale bar on the bottom has up to 80000 gray values. This is more, that we can expect on a 16-bit image. In addition, all images in the second line are overexposed. What should we see there, fluorescence of an acceptor?

What were the controls used in the FRET experiment to ensure the specificity of the FRET signal? It is not indicated in the Methods.

7E color scale bar at the bottom – colors are missing

16_In the introduction, when discussing force transmission through the cell-cell junction and the actin cytoskeleton, more recent articles on this subject could be cited.

17_ if the authors study the protein regulating cell contacts and thus collective cell behavior, why did not they attempt to apply a simply wound healing assay? It could provide a lot of information for the study.

Author Response

Please see attachment

Author Response File: Author Response.pdf

Reviewer 2 Report

The results of this study clearly demonstrates the coupling of E-cadherin tp the actin cytoskeleton  directly connected to the nuclear membrane. The results of this study also show that. expression of p60AmotL2 inactivates this connection and alters the properties of the nuclear lamina, potentiating the invasion of cells into micropores of the extracellular matrix. The findings of this study are interesting and can be of interest to a broad readership of Cells. However, a few minor revisions need to be accomplished before this study is amenable for publication.

1. The references cited need to be more up-to-date. There have been significant new studies related to this study and should ne cited.

2. Portions of methods section need to be rewritten and at present there are ambiguous statement that need to be edited so that the results of study can be reproduced by other researchers in the ffield. 

The overall English used in this manuscript is adequate.

Author Response

1. The references cited need to be more up-to-date. There have been significant new studies related to this study and should ne cited.

Thank you for your positive feedback on the study and its potential interest to a broad readership. You're right to suggest that more recent research should be included to provide a comprehensive and up-to-date context for our study. There has been considerable research on force transmission through cell-cell junctions and the actin cytoskeleton in the past few years. We updated the references in the introduction to provide a more comprehensive overview of the current understanding of force transmission through cell-cell junctions and the actin cytoskeleton. 

2. Portions of methods section need to be rewritten and at present there are ambiguous statement that need to be edited so that the results of study can be reproduced by other researchers in the ffield. 

We have now carefully reviewed the methods section of the manuscript and hopefully made the necessary revisions to improve clarity and remove any ambiguities.

Round 2

Reviewer 1 Report

The revised version is suitable for publication in Cells.