Biology Reference
In-Depth Information
Although both receptors can signal through a variety of pathways, it appears that PI-3-kinase
is essential for HGF-induced branching but not for the formation of alveoli, while MAP-
kinase is essential for alveoli but not branching morphogenesis, at least in a simple culture
system in which mammary epithelia are suspended in a three-dimensional matrix.
60
HO
W CONSERVED IS BRANCHING BY SPROUTIN
G?
Having reviewed branching systems ranging from simple three-dimensional cultures to
small fly tracheae to the large branching systems of mammals, it is possible to revisit the
question of whether all are variations on a single mechanistic theme.
There does seem to be strong conservation of the pathways that control branching, both
externally (for example, very frequent use of the FGF family) and internally (even where
a system uses a ramogen other than FGF, it still tends to use one that stimulates MAP-kinase
and PI-3-kinase). This is an argument for homology between the mechanisms used by tissues
such as airways in different phyla and for homology between the mechanisms used by
different branching tissues in the same organism (technically, this is 'serial homology',
a term coined by the nineteenth century anatomist Richard Owen, for homology between
similar parts of the same animal, such as vertebrae). There is, however, a danger in this argu-
ment. It is a valuable principle of biosemiotics
61
that the relationship between signifier and
signified is not universal across systems: what a molecular message means depends on the
state of the cell that is receiving it. To take FGF10 as an example: the molecule controls epithe-
lial branching in developing lung and pancreas, cortical neuroglial stem cell differentiation,
62
tongue muscle development,
63
stratification of the vaginal epithelium
64
and hormone secre-
tion in ovary cycling, amongst other things. To argue that FGF10's involvement in lung and
pancreas branching proves that the two organs must use the same branching mechanism
would, logically, mean accepting that development of neuroglia, tongue muscles and
hormone secretion also uses this mechanism. This is clearly nonsense (the other systems
have nothing to do with branching). To resolve this by dividing the sites of FGF10's action
into a group that involves branching and a group that does not, and discarding the group
that does not, is a logical error, for the act of grouping presupposes the similarity of the
branching organs. One cannot prove the validity of an assumption by an argument that relies
on that assumption being true.
)
Therefore similarity of control systems is not in itself a proof
of similarity of morphogenetic effector mechanisms.
In terms of morphology and behaviour, the systems can be divided into those that advance
by the leading edge motility of a single tip cell (the MDCKmodel, tracheae in
D. melanogaster
)
and those that advance by internal mechanisms such as apical constriction coordinated
across large populations of tip cells with no signs of classic leading edge motility (collecting
ducts, mammalian airways, mammary ducts
65
). One way of approaching the question of
whether these morphologies are fundamentally similar would be to see the same system
showing one or other type of branching depending on some simple, known difference in
)
One can
refute
a hypothesis by an argument that assumes it to be true and ends in a contradiction (Euclid's
proof of the infinite number of primes works this way, for example), but one cannot prove a hypothesis by an
argument that
assumes
it to be true.
