Biology Reference
In-Depth Information
curvature around the circumference of the cylinder is balanced by increasing negative curva-
ture along its length (
Figure 15.6
). In extremis, the process continues until the cylinder breaks
up completely to leave isolated spherical bubbles.
40
This instability of long cylinders under 'surface tension'
)
creates a real problem for organ-
isms that need to organize epithelia as long tubules. This fact was pointed out at least as long
ago as 1917 by D'Arcy Thompson.
40
There are several possible solutions to the problem. Most
epithelia are embedded in surrounding mesenchyme and it is possible that the mesenchyme
itself would resist the collapse of an underlying epithelium. However, since mesenchymes
are generally flexible enough, through literal flexing and also through remodelling, to permit
morphogenetic movements of epithelia there would have to be a very subtle sensing system
to allow them to tell the difference between morphogeneis and collapse. Alternatively, it is
possible that the epithelia themselves possess the ability to resist collapse. This cannot be
simply by there being a maximum radius of curvature through which an epithelial cell can
bend, because real tubules vary in diameter from microns to centimetres, so there must be
more specific mechanisms. A recent set of experiments in Drosophila melanogaster has identi-
fied a protein, without which tubules 'collapse' into disconnected cysts. Although the
researchers describing the abnormality do not mention it, this phenotype is so reminiscent
of the catastrophic collapse predicted by Thompson that it may lead directly to the systems
that epithelia have evolved to avoid that collapse.
The molecule concerned is a cadherin of the Fat family, a so-called 'atypical' cadherin of
which there are two members (Fat and Fat-like) in D. melanogaster
42,43
and at least three
(Fat1,2,3) in mammals.
44
e
46
Fat proteins are unusually large by the standards of the cadherin
family and have 34 cadherin-type repeats, and also laminin-like domains, in their extracel-
lular portions. By comparison, D. melanogaster's E-cadherin has just six cadherin-type
repeats.
47
Fat-like is expressed in tubular epithelia such as the tracheal trunks
43
while in
mammals Fat1 is expressed in lung and renal epithelia, as well as in some non-epithelial
sites.
44,48
e
50
When expression of Fat-like is prevented by interfering RNAs, the tracheal
system is severely malformed and collapses in places.
43
Significantly, perhaps, Fat1 is
expressed in the slit diaphragms that form the filters of mammalian kidneys; without the
protein, the small spaces of the slit diaphragm collapse, and the filter is completely blocked.
51
Fat1 does, therefore, seem to have a function in keeping epithelia apart.
How might Fat function? The Fat cadherins are located just apical to the zone of adherens
junctions in normal epithelial cells, and they interact heterophilically with another type of
atypical cadherin, Dachsous, on a neighbouring cell. The interaction can polarize cells within
the plane of the epithelium,
52
feeding into the same type of polarity system that is described
in detail in chapter 16, probably rather independently of the signalling systems described in
that chapter.
53
This type of polarity can determine the direction of cell division in an epithe-
lium, and thus control the rates of increase of a tubule's length compared with an increase of
its diameter. Generally, the effect of Fat loss on tubule diameter is considered in the light of
this,
54
but anything that can polarize cell division could also polarize other aspects of the
cytoskeleton. Polarizing the actin cytoskeleton, in particular, could result in anisotropic
mechanical properties along a tubule, making longitudinal stiffness/'surface tension'
)
The phrase is in inverted commas because the tension shown by epithelia is not strictly 'surface tension',
but it is mechanically analogous to it.
