Biology Reference
In-Depth Information
drug is removed and does so even when the epithelium is not in physical contact with the
mesenchyme.
76
It is not clear whether the actin is needed for events such as cell wedging or
for another purpose entirely; perhaps myosin-actin contraction, acting via integrin adhesions,
helps to pull collagen fibrils tight and drive the cleft hard back into the epithelium.
Clefting, then, seems to be initiated by epithelial cells that synthesize, secrete and organize
fibronectin proteins that then organize collagens I and III into fibrils capable of making a cleft.
The fibronectin also causes epithelial cell-cell adhesions to weaken. What is not at all clear is
how the epithelial cells organize themselves so that some initiate clefting while others,
between them, do not. The mesenchyme-derived signals mentioned above are required for
the stimulation of clefting, but it is highly unlikely that the position of clefts is patterned
by local variations in mesenchymal secretion because the epithelium branches normally
even in mesenchyme-free culture systems. It is possible that a lateral-inhibition system oper-
ates in the epithelium itself, in which cells already expressing fibronectin inhibit their neigh-
bours from deciding to initiate clefts themselves. Such a lateral inhibition systemmay rely on
chemical messengers or it may rely on the mechanical properties of the epithelium as it is
bent. If, for example, fibronectin expression were increased by concave curvature as motility
was decreased in the MDCK cell system described at the beginning of this chapter, the
required positive feedback would exist. There is some evidence, although no mechanism
yet, for the latter possibility, because altering the amount of collagen available to make fibrils,
using collagenases or inhibitors of endogenous collagenases, alters the number of clefts.
74,75
This suggests that collagen-rich clefts are themselves part of the feedback system that
controls spacing, and not just a downstream consequence of that system.
PATTERNING THE BRANCHING TREE
Although the positions and directions of the largest (first-formed) tubules of branched
systems tend to be laid down in stereotyped positions under the control of specific patterning
molecules, the detailed arrangements of smaller branches are normally more flexible and the
tree organizes itself according to the tissues that it has to serve. This self-organization seems
to be achieved by a combination of autocrine and paracrine signalling.
Many mesenchyme-derived ramogens act as chemoattractants to branching epithelia. If,
for example, a bead soaked in the renal ramogen GDNF is placed adjacent to a kidney rudi-
ment growing in culture, the nearby parts of the developing ureteric bud tree grow towards
it.
77
Indeed, such beads can even induce the formation of new branches from parts of the
excretory system that would not normally branch, such as the nephric duct. Cells derived
from the ureteric bud also show chemotactic responses for GDNF in two-dimensional
culture.
78
It is not clear how this chemotaxis works in the context of a branch tip, in terms
of what moves where, but if branching tips do show 'conventional' locomotory behaviour
(filopodia, and so on), it may use an adaptation of the chemotactic mechanisms described
in Chapter 9. Whatever its mechanism, the directional response would have the effect of
bringing epithelial branches towards the strongest sources of attractants and would
encourage further branching there.
Computer modelling of the ureteric bud, with GDNF-stimulated proliferation and GDNF-
directed chemotaxis, simulates typical branching events well.
79
In particular, if motility is
