So it looks as if a single epigenetic process, coupled with
threshold responses not only initiates differentiation and morphogenesis
of the nervous system, but also integrates the differentiation
and morphogenesis of all the primary organ systems within the
embryo. This concept of an epigenetic cascade is a useful one,
- a single morphogenetic event setting off a string of others.
If the epigenetic cascade could be modified to include or exclude
certain events then the structure formed might be modified.
Epigenetic cascades (OVERHEAD)
Despite the utility of the concept we know very little about
epigenetic cascades, and have only worked out the detail of a
few of them. Besides the development of the embryonic axis we
know something about the eye and ear, the teeth of fetal mice
and the mesonephric kidney of the newt.
Induction of the sense organs
1. the amphibian eye
The classical story of the development of the eye is often
quoted as an example of embryonic induction. In the future head
the neural ectoderm is specified to form the three major divisions
of the brain by the notochord. The forebrain has two fates (OVERHEAD),
part of it develops into the cerebral hemispheres and the remainder
grows out as optic lobes. These grow towards the overlying ectoderm,
becoming optic vesicles and finally forming the optic cup. in
the optic cup several different cell types differentiate, neural
retina, pigmented retina, iris. Once to optic vesicle has contacted
the overlying ectoderm the vesicle induces the ectoderm to bud
off a lens vesicle which then rounds up and differentiates into
the lens. The lens acts on the ectoderm overlying it to induce
the transparent cornea instead of opaque skin.
So here is an elegant cascade, with one structure (the optic
vesicle) inducing another (the lens) which in turn induces another
(the cornea).
It is also a textbook interpretation of the truth. There are,
in fact, many species of anuran and urodele amphibians in which
the lens can form in the absence of the optic cup, prior interactions
of the ectoderm with endoderm and/or mesoderm being sufficient.
There are genera within which the eye is induced by the optic
cup in some species but self differentiates in others. There
is at least one species (Rana esculentia) where the lens
self differentiates but the optic cup can still induce lens differentiation
from mesoderm. We shall return to the evolutionary significance
of this variation between related species later.
2. the amphibian ear
The situation here is slightly different from that of the
eye. Instead of a cascade of different inducers producing different
tissues we see here a number of inducors co-operating to produce
a single structure, the cartilaginous ear capsule or otic vesicle.
In Ambystoma punctata cranial ectoderm in the region of
the hindbrain is induced to form an otic vesicle in two major
steps. During early to mid neurulation cranial mesoderm induces
cranial ectoderm. As this wanes in mid neurulation a second inductive
wave is produced by the hindbrain. (OVERHEAD) Once the
otic vesicle has formed it acts inductively on adjacent mesoderm
to form the cartilaginous otic capsule.
Timing is obviously important in this sort of multi-stage
process, and this is regulated by the formation of cellular contacts,
including gap junctions, which appear only during the phase of
induction. What regulates the timing of these cellular contacts
is another question. Timing is also regulated by the acquisition
of ectodermal competence to respond to induction. Ectoderm from
the early neural tube will not respond to inductive influences
produced by the hindbrain of late nerulae and ectoderm from beyond
neurulation doesn't respond either.
Structures other than sense organs
1. Mammalian tooth induction.
Teeth (OVERHEAD) are composite structures made up of
the outer white enamel which covers the teeth above the gums
and the inner dentine, a different mineralised tissue forming
the root and interior of the teeth. Dentine and enamel are extracellular
products of two different types of cell, the ameloblasts (enamel)
and odontoblasts (dentine).
Ameloblasts arise from epithelium, the oral epithelium. Odontoblasts
arise in the neural crest. The neural crest loses many special
populations of cells which migrate off to different places within
the body: the pre-odontoblasts are one such group which migrate
into the developing oral cavity and hence along the oral epithelium.
Odontoblasts are formed from these cells after a series of
epigenetic interactions with the overlying epidermis. The critical
factor is that the mesenchymal cells stop migrating and settle
down against the buccal epithelium at the locations of the future
teeth. We might like to ask ourselves why they stop precisely
in those positions, but we shall come back to the whole question
of spatial analysis later.
Dental mesenchyme (OVERHEAD) induces adjacent oral
epithelium to proliferate (interaction 1). Dental epithelium
then induces adjacent mesenchyme to proliferate to form a dental
papilla (2). The dental papilla then induces the epithelium to
form an enamel organ. If these two are artificially or experimentally
separated from each other at this stage no further development
occurs. If left in association the enamel organ induces the dental
papilla to differentiate as preodontoblasts, which develop into
odontoblasts (4). The odontoblasts and the predentine deposited
upon them induces the preameloblasts to become ameloblasts and
to deposit enamel (5,6). Again nothing further occurs if the
pre-ameloblasts and pre-odontoblasts are separated.
So in both the ear and the tooth we have cascades of interactions
following one another and building up a complex whole. In these
examples all cascade steps are positive and promote differentiation
or development. This is not necessarily always so.
2. The newt kidney
In the kidney of the newt (OVERHEAD) another complex
series of interactions is known which starts in the early gastrula
and continues late into larval life. These associations are,
as one would expect between an epithelium and mesenchyme. At
first there is an interaction between endoderm and gut mesoderm,
then nephric ducts and kidney mesenchyme. Notochord also acts
as a positive inducer. Somitic and lateral plate mesoderm however,
initially stimulate, then inhibit, and the neural tube has a
small inhibitory role throughout. The largest inhibitor is, however,
the neural crest.
So how does this relate to the sort of thing we saw in the
tooth? It's more complex: it isn't a sequence of one step interactions:
it isn't simple series of ping pong interactions, it is a complex
interplay of inhibitory as well as stimulatory events. What we
have described as induction, and perhaps seen as a simple switching
on, is in fact a complex sequence of interacting processes which
may switch the induction off as well as on. On balance, over
time they will result in the induction of a kidney: but there
is enormous scope for tinkering. If we changed things a bit we
might end up with a completely different organ!
On a larger scale - Oral and buccal development in the
newt.
Lets now try to see how whole regions of embryos, rather than
individual organ systems are formed. In the mouth region of the
newt four distinct cell populations, (lateral mesoderm, the prechordal
plate, the neural crest and the pharyngeal ectoderm (OVERHEAD)
are all acting together to form mesenchymes and epithelia which
interact together to make teeth, cartilages, bones, loose connective
tissue and the various epithelia around the buccal cavity. Each
element has its own epigenetic cascade, as described before,
but the cascades are integrated: if not we might find a random
mixture of structures such as that seen in a teratoma, where
recognisable organs are formed in no particular order or
arrangement.
In the newt (and other animals) the vomer, palatine, dentary
and splenial bones (in the face) will not form until the trabecular
and Meckel's cartilages (the cartilages of the first arch) have
formed. Teeth will not form unless jaws are present. These are
just a few examples of a widespread phenomenon: this interaction
between cascades is important in both phylogeny and evolution.
We can see immediately the importance of building the right things
in the right order in an embryo, but it is just as important
to see that this sort of constraint, the dependence of one system
on another is what limits evolutionary possibilities. We can't
have three eyes, or a nose on top of our head because these changes
would influence other systems to an extent that would make the
whole process of development untenable.
The tetrapod limb (OVERHEAD).
Once we can begin to understand the way in which
regions like the buccal region develop we ought to be able
to look at rather bigger bits, like the limbs.
We don't need to go into great detail, but the limb is a good
example of morphogenesis, depending as it does on interactions
between an epidermis and underlying mesoderm. Also limbs are
quite variable structures: so as well as the way limbs develop
lets look at some of the ways in which they might vary.
In all vertebrates the limb is organised first as a simple
limb field in the flank. That is to say limb development is limited
to a couple of regions, one anterior and one posterior (OVERHEAD).
This wasn't always the case: some of the first fishes to have
paired fins (which we can regard as homologous to limbs) had
up to seven pairs (OVERHEAD).
In the putative limb regions a ridge of epidermis, the apical
ectodermal ridge, develops as a line of rather tall cells (OVERHEAD).
Beneath the ridge mesenchyme accumulates, setting up a classical
ectoderm/mesoderm situation. Within the ridge dorsal - ventral
and anterior - posterior axis quickly develop (experimentally
verified by rotating cut off tips and seeing how they develop
(OVERHEAD) and the limb characteristically becomes a footplate
at the end of a narrower leg. The mesenchyme in the limb is complex
in origin. In chicks it is relatively easy to show (by experiments
using grafts from quail, which have a characteristic nucleolus,
so that cells are easily identified) that the cells which will
form the muscles originate in the somites, whilst skeletal and
fibroblastic cells come from more lateral mesoderm. Innervation
is rather late and segmental, but the number of segments involved
in a limb, and which ones they are, is rather variable.
There is a general hypothesis, due to Saunders and Zwilling, that the apical ectodermal ridge is maintained by a mesodermal maintenance factor (OVERHEAD). The general characteristics of the limb are controlled by the mesoderm. Early in development a zone of polarising activity, or ZPA, develops on the posterior side of the limb bud. This has many interesting features, such as involvement with many theories of development, position effect, progress zones, retinoic acid etc. At present lets just regard it as an organiser region specifying a-p pattern.
All the cells in a developing embryo are part of a pattern,
and each one has to be in the right place at the right time.
The early development of the egg was a good example of this.
Whether regulatory or determinate the mechanism ensures a pattern
(the blastula) with the right cells in the right place. In the
determinate egg this is quite literal, cell 86 or whatever has
to be there else the embryo won't have a kidney or a gonad.
In the regulatory system the precise instructions have yet to
be issued, but the right sort of cell, perhaps in terms
of size, shape, stickiness, motility etc. is in the right area
of the developing embryo. The influences which govern this early
phase of pattern formation comes more or less passively from
the distribution to daughter cells and clones of cytoplasmic
determinants from the oocyte, from localised signalling via cell
to cell contact and interactions from cell behaviour.
The blastomeres and their derivatives in fact behave as if they have knowledge of their positions in time and space. Horstadius (1973) showed that the cleavage patterns of sea urchin embryos follow a strict internal clock that seems to run on absolute time. Position sense is even more complex: the right cells end up in the right place. Wolpert (1969) suggested that cells have positional information. If a cell has positional information it must:
1. register its position in an embryo according to a set of signalling systems then remember this information
2. Allow subsequent cell behaviour to be governed by this information.
In everyday language a cell is given a password at a certain
point or time, and the password determines the fate of the cell.
An alternative view is that there is a clock in the cell,
and that at a given elapsed time from zero the cell records the
values of external signals and these form the password.
In either case the cells must be responding to a pattern set
up within the embryo. Or more accurately a series of patterns
existing one after the other, and each generated by the one before.
This is difficult stuff, so let's relax bit with the idea
of patterns and how they work.
Pattern (OVERHEAD) is the spatial relationships within
a developing form. An arm and a leg contain much the same tissues,
x % bone, y% cartilage, z% muscle, but a given structure is recognisable
as an arm or a leg. The pattern is different, although similar.
So how are similar cells arranged into dissimilar patterns, and
why is one pattern preferred for the arm and another for the
leg. And again why and how is the pattern in an aardvark different
from that in a sloth?
We can suggest five main types of patterns (OVERHEAD)
1. Unit generated patterns.(OVERHEAD) The arhytypical
unit generated pattern is that seen in an inorganic crystal.
Here ions and atoms are held together by valency forces. The
shape generated is very precise and crystals can be at least
as large as cells. The secret is in the very precise arrangements
of components and the valency forces which hold them together.
But it doesn't need to be that uniform or that precise: globular
protein molecules with two asymmetrical sticky patches on the
surface shaken together would form a helix. If the units were
cylinders, like insulin, the unit would be a four lobed fibre.
Fibres can be monotype, made up of one sort of molecule, like
collagen, or polytypic, like muscle fibres with actin and myosin.
It is easy to see that a suitable unit might make up into a sheet.
2. Instruction generated patterns (OVERHEAD) It is
also possible to make a structure from a series of units plus
a plan. Think of a bricklayer with a pile of bricks and a plan
building a house. The shape of the house is not governed by the
shape of the bricks, but by the instructions in the plan. In
this case the structure is generated diachronically i.e. slowly,
but similar patterns can be generated very quickly. When the
Sergeant-major shouts 'fall in' (instructions) a structure made
up of individual soldiers (bricks) is generated very rapidly.
3. Template generated systems (OVERHEAD) In biological systems it seems most likely that a modification of the house building set-up is likely to occur. Instead of a scaled plan the architect could peg out the pattern of the house directly on the ground. This is just the same as a dressmakers pattern. The template is a structure made of dissimilar materials on or around which the structure is formed. Of course there is a philosophical catch in this: the template is itself a pattern - how do we generate it?
Again templates can be synchronic or diachronic. A diachronic
example might be the formation of a hen's egg. Here the various
layers of the egg are laid down sequentially by different regions
of the oviduct. More usually they are synchronous. More important
however is that the hen's egg is a non copy of the template.
Nowhere in the hen is there the equivalent of a dressmakers pattern
for an egg: nowhere is there an egg mould. The dorsal lip of
the blastopore induces a neural tube: a neural tube is not a
dressmakers pattern of a blastopore lip. So templates can produce
non-copies as well as copies.
4. Condition generated patterns. (OVERHEAD) These are not related to the substance from which they are made, nor do they involve a set of instructions, but arise from the interactions of a series of spatially distributed conditions. Lets start with a cylinder of tissue and impose the following conditions
1. the cells are capable of differentiating into precartilage cells
2. when the density of these cells reaches a certain level they die
3. the threshold for death varies linearly.
Mesoderm differentiates into cartilage under the influence
of oxygen concentration. Therefore a rod of precartilage cells
will appear in the centre of the cylinder. At the end with the
lowest death threshold cells will begin to die. If we impose
another condition, that the cylinder is flattened a little at
the end the rod of cartilage will in fact eventually split into
two. What have we made? a limb with femur, tibia and fibula.
5. Positional Information. Wolpert applied this concept
to the limb in order to demonstrate pattern generation (OVERHEAD).
If a line of cells differentiates according to a gradient so
that one third is red, one third white, one third blue the length
of the line is unimportant. This might be applied to the various
regions of the limb: however the number of thresholds for colour
or region change is large.
In a structure like a limb different pattern generating methodologies
will be acting at the same time. Collagen fibres in the extracellular
matrix will be assembling because of the shape of tropocollagen
molecules, protein will be forming on RNA templates and RNA will
be forming on DNA templates in every cell. Mesoderm and ectoderm
will be interacting to form non-copy template patterns.
Where is the pattern?
The pattern forming properties might be intrinsic, or extrinsic.
Cells might have properties which predispose them to make patterns
(like our proto-limb) or they might be responding to some sort
of stimulus from elsewhere. If the stimulus is internal do all
cells know about the pattern, or only some of them? We can simplify
matters a little by sticking with our limb. Limbs have a limited
number of tissue types. We can rule out the circulatory system
because that develops late, after the patterns have been established.
For the same reason the nervous system can be ruled out: perfectly
shaped limbs can be formed with no innervation. The skin and
its precursor the ectoderm is not responsible. The ectoderm,
of course, interacts with the mesoderm to produce the outgrowth
of the limb, but transplanting ectoderm from forelimb to hindlimb
still leaves a hindlimb, and vice versa. So it looks as if the
pattern lies in the skeleton or the muscles. Both skeleton and
muscles form patterns and the patterns are related to each other:
specific bones carry specific muscles. Is this the same pattern?
In the region of the digits all the relevant cells are skeletogenic
- no muscles, so the skeletogenic cells have the pattern: we
could argue that the skeleton cells only have the pattern and
impose this upon the incoming muscle cells. Or is it the other
way around?
We don't really know, but at least we are narrowing the field. We do know that both types of cell form condensations. Are condensations the result of cell behaviour or receipt of a signal? In other words what sorts of behaviour do cells show to other cells?
This page is maintained by Steve Paxton