Biology Reference
In-Depth Information
controlling the leading edge and the pathways diverge right up at the level of the G proteins.
The PI(3,4,5)P
3
/ Rac path proceeds via Gi, while two other G proteins, G12 and G13, act via
Rho to control myosin. Again, the 'front' and 'back' signals antagonize one another as they
do in myxamoebae.
35
One other interesting difference between neutrophils and myxamoe-
bae is that the leading edge of neutrophils is much more sensitive to chemoattractants than
the trailing edge.
36
This creates a marked reluctance of the cells to repolarize, and they tend
to respond to sudden changes in external gradient by performing a U-turn rather than by
reversing.
37
Overall, though, the similarities between neutrophils and myxamoebae are
striking, and while the vertebrate cells seem to be complicated by several parallel pathways
triggered by different signals,
38
the basic ideas remain the same. As with D. discoideum, there
seem to be other pathways that run in parallel with PI-3-kinase to encourage a leading
edge.
39
Not all chemotactic signals proceed via G proteins. The mammalian wound healing
response includes chemotactic migration of fibroblasts up a gradient of platelet derived
growth factor (PDGF) which is released by the wound itself. PDGF signals use receptor
tyrosine kinases rather than G-proteins. Nevertheless, the chemotactic signalling mechan-
ism still relies on PI-3-kinases, a requirement that can be demonstrated by pharmacological
inhibition of PI-3-kinase
40
and by mutants that cannot activate PI-3-kinase in response to
PDGF signalling.
40
e
42
It also involves the translocation of Akt to the leading edge of the
cell, where the (shallow) external gradient of PDGF is at its highest, a fact that has been
demonstrated using Akt tagged with GFP.
43
This translocation fails when PI-3-kinase is
inhibited.
43
Again, the internal gradient of Akt is far steeper than the external gradient
of PDGF.
There has been too little work done on metazoan chemotaxis for any certain conclusions to
be drawn, but the balance of evidence so far suggests that there is essentially one broadmech-
anism for chemotactic crawling migration that is conserved between cell types within an
organism and that has been conserved for the hundreds of millions of years that separate
D. discoideum and mammals.
CHEMOREPULSION
Chemotaxis does not always involve chemoattraction, and some morphogenetic events in
metazoan development depend on the ability of tissues to repel specific migratory cells.
Sometimes repulsion is mediated by fixed molecules (see Chapter 11) but, in other cases,
diffusible chemorepellants are used.
One example of a diffusible chemorepellent system is provided by the Slit family of
secreted proteins. In mammals, these proteins repel many cell types, including neutrophils,
44
dendritic cells,
45
endothelial cells,
46
glia
47
and neurons.
48,49
As well as being widespread
within one organism, Slit proteins are used as repellents in a broad range of other metazoa
including insects (Drosophila melanogaster)
50
and nematodes (Caenorhabditis elegans).
51
Slit
proteins signal via transmembrane receptors of the Robo (Roundabout) family.
52
Binding of
Slit activates a specific intracellular domain of Robo, CC3, that allows it to interact with
the SH3 domain of a GTPase activating protein (GAP) that increases the intrinsic GTPase
activity of cdc42.
53
This alters the cdc42-GTP
cdc42-GDP equilibrium to the right, and
4
