3.2 Limb patterning in tetrapods

Geoffroy de Saint Hilaire would have be happy to read the mix Fleury made of his observations of structural homologies between tetrapods and insects and genetic homologies, but he probably have laughed with the persistence of his own idea of analogy between ribs and insects appendages armed with current knowledge of embryology (no ectoderm participating in the building of ribs, but then Fleury seems to mess all the time with the embryonic tissues). And he probably have being delighted with the actual knowledge on deep homology1, more then reading cteappv.

Fleury is concerned with limbs positioning, FGF role in limb development, limb positioning and specification, and autopod patterning. He struggles to express the specification of limbs in terms where the role of morphogenic genes is minimized and FGFs role maximized2. Interspersed to his presentation there are a few quite particular points of view, which have more to do with his misunderstanding of empirical approaches in general than with the actual subject.

The main idea is that limbs positioning is not genetically specified but “come together with the limbs in the morphogenetic process“. Which morphogenetic process wouldn’t be genetically specified?
Geometrical positioning seems to be the solution Fleury will try to defend, “from a physics point of view“, with a lousy description of anatomical features, with the idea that genetic determination of the pelvis’ shape should be “strictly instructed“, or that the hindlimb “not truly [grows] ‘along the trunk”“, or the “one would expect to see mutants with a pelvis without limbs at any given level of the thorax, and anywhere else, limbs without pelvis or shoulders, which is certainly not the case.“.

The main conclusion is that “Therefore in vertebrates the concept of antero-posterior gradient is not related to the correct positioning of limbs.” assorted with the idea that the would be a “physical field of morphogenesis“.

What he consider to be a concept is rather an observation: gradients of morphogens (direct or indirect products of genetic expression) are descriptions of embryos constitution, and their relations with morphogenic fields experimentally established. One can’t simply dismiss them if they don’t fit a preconceived phenomenological model unsupported by any realistic approach.
Fleury have already be laughed for his not genetically determined morphogenic process approach, so he is careful to give some credit to genes’ role in the process.

His champions are the FGFs, as controllers of the proximo-distal axis of limbs’ growth, not related to the topology of the patterns which they grow. FGFs are considered mechanosensitive and thus he expect to fit them with a purely “physical field of morphogenesis“, even if he need to imagine how their expression is modulated by physical variables such as pressure. Without further consideration of how pressure variations are specified, to avoid to acknowledge genetic determinism.

Where data exists showing that a particular gene’s expression is linked with limb positioning, specification or patterning, the observed effects are simply minimized:

It has been shown that it is possible to shift slightly the position of outgrowth of the limb by repressing or activating Tbx3 [49], but shifting a limb is not the same thing as inducing a limb17.

[emphasis mine].

Where the complexity is overwhelming for our physicist he stops considering and tries to come with a simple alternative, implying that physics are the correct rationale:

This shows that the relationship between growth factors and the final pattern is somewhat puzzling, there seems to be strong “redundancies”, and “regulatory loops”, although the rationale may be otherwise: as long as quantitative levels of growth factors are not related to physical parameters of growth, it is hard to tell why a limb grows or not.

I hope other physicists don’t flinch when confronted with the complexity of gene regulatory networks or of genetic material organization.

It is quite funny to observe how Fleury struggles to interpret experimental data which he don’t understands but it becomes consternating when he express doubts of their significance.

Let’s deal with a single example of it: limbs type specification by Tbx4, Tbx5 and Pitx1.

One odd comparison is the one between chicks and newts:

Unfortunately Tbx5 and Tbx4 were found in limbs of both types altogether in the newt model [58], which sheds doubt on the relationship between Tbx and limb identity.

OK, genetic determinism of limbs is different between chicks and newts. That doesn’t shed any doubt about the genetic determinism of limbs type in the chick, the reported results of Minguillon et al. [57] and Takeuchi et al. [59] are rather clear. One can expect differences in phylogenetically distant species and avoid comparisons where the data are not equivalent: there is no data concerning the expression of Pitx1, which participates at the specification of hindlimbs in the chicken, at the newt model.

Takeuchi et al. did a great job using in ovo electroporation (a new technique at the time) to study roles of Tbx4 and Tbx5 in chick limbs. Fleury’s comment on these results is the following [p11, col2,§1]:

The success rate of these experiments is very low (∼10%), and puzzling for physicists. Especially one would expect electroporation experiments of Tbx5 in a right leg bud to give either 50% of left wings and 50% of right wings, if limb identity is only defined by Tbx5, or 100% of right wings, if the limb identity is defined by Tbx5 and the context. Still 90% of right legs are actually obtained18.

Neutrino’s detection success rate is several orders of magnitude smaller than Takeuchi et al. experiments success rate. Hopefully there aren’t any physicists puzzled by that. Takeuchi et al. would have be puzzling if the technique was know to have a higher success rate and in the particular case it didn’t worked..

But Fleury don’t like these results and he badly needs to diminish their value. The only way to that is non sense. He use it.

It is about an in vivo competition assay, performed in a wild type embryo, expressing it’s genome normally, including endogenous Tbxs; the transgene have to compete with them and overcome their activity to change the hindlimb identity from wing to leg. Using a low efficiency technique. 100% efficiency may be an ignorant theorist expectation, 10% is a very nice reality based result.

I can’t even imagine what Fleury means by: “Especially one would expect electroporation experiments of Tbx5 in a right leg bud to give either 50% of left wings and 50% of right wings” and why one would expect “50% of left wings” at all at the right side of the chick. I can’t even imagine that this haven’t puzzled the referees of the paper.

The same attitude versus experimental results is found a second time in the same subsection [p13, col2, §2]:

Still, a primitive limb is such a simple thing: a round paddle with rays. A number of facts explain this darkness.

Yeah, right, ask an anatomist how simple a limb is. Fortunately there is a savior coming, with his deep knowledge of embryology to save us poor biologists from the darkness.

First of all Fleury don’t get that, from fecondation to fecondation there is cyclic phenomenon and one can choose any convenient point as the start point of the life process. Usually the formation of the zygote is used as the starting point and one considers at least two different processes (male and female) along the life’s cycle, at least for sexual reproduction. And that life is a continuous and continuously evolving process, from the very first abiogenesis to today.

It is obvious that for one without enough knowledge of a particular scientific domain, it is not clear how the experimentally obtained data can be interpreted and what natural situations experimental settings simulate. But it is exceptionally rare for a scientist (any discipline) to be so short sighted that he can’t understand that a “bead soaked in retinoic acid” simulates either a genetic disorder leading to increased production of retinoic acid, or a retinoic acid intoxication. Fleury is a quite exceptional fellow.

He do a great job reporting from the cited papers in a way that seem to support his views, as long as the reader don’t check the source.

That can be just plainly false, and this is the case of [72]: Minimal ProtoHox cluster inferred from bilaterian and cnidarian Hox complements, D. Chourrout et al., Nature 2006, doi:10.1038/nature04863. One wonders what Fleury understood from this reading.

Or simply ignoring context, e.g. the remark from Serrano & O’Farrell [32] paper: “One of the most mysterious features of morphogenesis is that structures form themselves as they grow.”, the first phrase of the last section of the paper titled “Mysteries exposed”, from a paper dealing with “Limb morphogenesis: connections between patterning and growth”. Shear mystery here.

The same with his citation from Rallis et al. [49]: “However, neither gene deletion nor gene misexpression experiments have provided direct evidence for a role of Hox genes in limb positioning.”, which leaves the erader with the impression that nothing is known about the relations between Hox genes expression patterns and limbs positioning, a phrase from the introduction of a paper where the authors show the implication of Tbx3 in limb’s positioning.

Or by concealing informations, e.g. those in Cohn’s papers [48 a & b], about “Hox9 genes and vertebrate limb specification” or about the “Developmental basis of limblessness and axial patterning in snakes”.

One can’t be sure if Fleury’ didn’t understood what he read (admitting that he really read the papers he reference) or that is is concealing (consciously or not) information that might contradict his point of view. But one can be sure that the reader don’t get a clarification of tetrapods’ embryogenesis, by any mean, but rather a carefully instilled darkness

Things get worse, much worse, when Fleury don’t just wrongly reports other people’s work, but provide his own interpretations. He ignores where the lateral plate comes from (and this is a particularly important point for his model), explains the geometry of limb buds formation and dorsoventral polarity in a completely nonsensical way and when it comes to describe a part of embryogenesis, gastrulation, which is central for his clarification what he delivers is quite terrible:

The initial formless embryo (the “blastula”) is composed of one layer (the epiblast). A second layer, the hypoblast nucleates underneath the epiblast, therefore the hypoblast emanates from the epiblast. During morphogenesis, a third layer appears by invagination, during gastrulation, of the first two layers inside themselves. In the literature, as the body forms, new names are ascribed to the layers: ectoderm (skin on top), endoderm (layer below), and mesoderm (in between).

invagination, […] of the first two layers inside themselves” WTF?3 No notion of primitive and definitive endoderm difference? The hypoblast invaginating? The endoderm being the “layer below” and not intern?

The finale is interesting:

Another problem lies in experimental skill. It seems from oral interview of scientists in limb developmental biology, that, in order to produce extra limbs or extra digits, beads soaked in chemicals have to be put at specific times, and specific places, in an area called “limb field”. Especially, aberrant digit distributions can be obtained by grafting pieces of limb buds onto other limb buds, but at very early stages where digits are supposed not to be present at all, therefore even the timing of digit appearance is not truly understood.

Well, experimental skills are quite important for experimental approaches and Fleury seems to have some experience of that, as he was unable to get data crucial for his point of view in one of his experimental works. Every experimentalist know that, even physicists (at least those I know and those I read).

I very much like how Orzel put it:

Hands are the main thing setting us above dogs, after all, hands and brains. Without those, who knows where we would be.

Hands are for experimental skills. Brains help a lot understanding that you can’t expect to trigger a limb before the mesodermal lateral plate is established of when it looses it’s competence to produce limbs. Brains are also of great help to understand that tissues able to induce particular anatomical features may have to be grafted before those anatomical features appear (except when you use as animal model some species able of limbs regeneration, where additional limbs may be obtained in the adult).


1. Deep homology and the origins of evolutionary novelty, Neil Shubin, Cliff Tabin & Sean Carroll, Nature, Vol 457, 12 February 2009, doi:10.1038/nature07891

2. FGF and the regulation of expression by pressure is important enough for Fleury to invent it without data to support his assertion.

3. A nice reading on the origin of the three -derm layers, reality based: Epiblast and primitive-streak origins of the endoderm in the gastrulating chick embryo, Aaron Lawson and Gary C. Schoenwolf, Development 130, 3491-3501 (2003), doi: 10.1242/dev.00579

3 Responses

  1. […] enfin mis en public : 3.2 Limb patterning in tetrapods. J’ai réussi, je crois, de ne pas dire de gros mots. Suis pas sûr, me suis pas relu, […]

  2. Tonio,
    a bug at the last doi, a duplication of 10.1242/

  3. Thanks, I made the correction.
    You lucky french speaking people check this one

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

%d bloggers like this: