New Paper Published – Is exposure to ECM enough to cause epithelial transdifferentiation?

New paper out in Frontiers in Medicine (free), from Tiago Ramos.

Tiago did his PhD at the University of Liverpool with primary supervisor of Saj Ahmad, a clinical researcher who has moved to Moorefields in London. His work focused on the question; why do apparently terminally differentiated conjunctival cells change when they grow over the cornea?

It turns out (unsurprisingly to regulars on this blog), that the extracellular matrix deposited by corneal cells provides signals that drive changes to the cells. That bit is no surprise really, what is surprising is that there is plasticity at all; that the corneal cell-derived matrix is sufficient to cause partial de-differentiation and then push the conjunctival cells partially down a corneal differentiation lineage. This same plasticity was also observed in the opposite direction corneal to conj.

Below is an overview of the approaches used to come to these conclusions:

Tiago did an awful lot of characterisation! RT-qPCR plates were piled high as he compared keratin 3, 12, 7, 13, MUC5AC, deltaNp63 and ABCB5 at multiple time points for each line in each condition, then followed it up with blots for the same set of “markers” at the protein level

Magic Matrix

While it was exciting to find out this could happen, the cell-derived matrixes were a black box in terms of what constituents were driving the changes. Although I was confident the answer would be within the laminins, we took a more prosaic approach, using mass spectrometry to generate an unbiased analysis of the different matrixes (following a similar approach to our previous study on conj ECM). These analyses revealed many proteins to be shared between the matrixes, and identified a number that were present in one but not the other.

These data provoked another round of analyses, focusing on confirming the different proteins (more graphs!).

In the list of differentially expressed proteins were four of my favourites; type XVII collagen and three laminin chains, laminin beta1, beta2 and alpha5. Tiago followed these up at the transcript, protein and distribution levels (another massive amount of data). These experiments confirmed that there were clear differences between the two matrixes (e.g. pairs of images below compare the conj matrix on the left of the pair vs the corneal of col XVII then the three laminins)

It had taken quite a while to get to this point. However, now the rest of the supervisor team were prepared to accept that the answer could lie in the laminins. Huzzah, new believers! Although it was fun to say; “told you”, I was actually very pleased that the hypothesis-independent approach led to the same endpoint. Does this now mean laminins are the answer rather than a answer? Probably not, but certainly great for laminin-o-philes.

Now that the team were convinced that laminins could hold the answer… Tiago did a series of experiments comparing the cellular responses to recombinant purified laminins. There is another massive data set comparing the same panel of key marker genes on different laminins and then with different concentrations. This is all in supplemental figures but can summarised; the expression level of differentiation markers is dependent on which laminin cells are interacting with.

With thousands of RT-qPCR performed, Tiago was ready to try to something different! So, he next took the best performing laminin in defining lineage specification and compared how the cells responded to that matrix in terms of key signalling pathways activities. Again, a bucketload of data and another 20 graph figure ensues.

In case you haven’t worked out, there is a lot in this paper.

Interpreting this amount of data into a coherent narrative is difficult. We spent many many many hours discussing what each of these graphs mean and when you put them together, it gets even more complex. Tiago wrote a magnum opus of a PhD thesis and his viva was long! The very first draft of this paper had 16 figures.

The newly published paper should provide readers with a resource they can refer to. Likely Tiago already did the experiment you wanted, and the data are in there for you.

Congrats Tiago and Saj, it’s great to see this work finally out! It’s been a long time coming but the end result should be valuable to the field.

Read the paper here.