The cell state splitter: Embryogenesis Explained: A review by David Miller

When I was an undergraduate at Glasgow University back in the 1970s, it was Adam Curtis and his team who first introduced me to cell and developmental biology. Those early days provided me with the foundations of a hypothetical framework for the establishment of positional information in the embryo, first postulated by Louis Wolpert with his ‘French Flag Theory’ of pattern formation and cell determination and fate [Wolpert 1969]. The more recent consolidation of the evidence for a morphogenic gradient guiding early embryo development in Drosophila [Ephrussi and St Johnston 2004] and Wolpert’s own work with retinoic acid on chick limb development [Tickle et al. 1982] seemed to provide the best links between cell signalling, positioning, and fate available to us at the time I was studying them. Michael Abercrombie’s work on fibroblast movement migration and contact inhibition also made a lasting impression [Abercrombie et al. 1971].

To my mind, these phenomena could all be neatly tied into cell fate and determination in both embryological and neoplastic contexts (somewhat interchangeable terms in relation to the fascinating differentiation of teratomas). So, Natalie and Richard Gordon’s recent book on amphibian embryogenesis Embryogenesis Explained [Gordon and Gordon 2016a] took me back to those early days and my first encounters with cell structure, function, movement, and differentiation. The book is somewhat ambiguously titled as before reading it, I had assumed that the mechanisms underlying embryology as a developmental process, while sufficiently well understood by the experts at least, now only needed explaining to the interested lay reader in a more accessible format. What I was not expecting was a whole new and potentially paradigm shifting concept in our understanding of what drives cell determination and fate in the developing embryo, in the form of the cell state splitter and differentiation trees. I am not alone in being (up till now at any rate) ignorant of these fascinating ideas. At a recent major fertility meeting I attended, not a single ‘expert’ I spoke with was aware of the Gordons’ concepts, although in all fairness to the Gordons, most of the attendees could hardly be described as classical embryologists or even developmental biologists.

The Gordons’ book offers the reader a rich and fascinating insight not only to the science and history of embryology and developmental cell biology but also to the minds of its writers, who the reader will discover, have been championing this radically ‘new’ way of looking at and explaining embryo morphogenesis for over twenty years. To a much greater extent than their book suggests from its cover (a salamander pondering its own differentiation tree), Embryogenesis Explained is a veritable potpourri of perspectives and concepts in genetics and epigenetics, signalling, cell and developmental biology, and even going so far as beginning and ending with broader speculations on our biological purpose as observers of the (observable) universe and the anthropic principle. Regarding the latter, there did come a point towards the end of the book where I thought the Gordons were about to set off for the more unchartered waters of the quantum nature of consciousness and the mysteries of microtubules therein, but I was happily spared that diversion!

The book considers the big bang and the origins of life on our planet (and possibly other planets) and includes whole chapters detailing the intricate mechanisms controlling meiosis, mitosis, differentiation, genetics, and epigenetics. The book is written in a relatively accessible format for the lay and more expert reader alike although some sections covering mainly theoretical considerations and philosophical musings may be more challenging for some. I must confess that until I finished the book, I had no idea that the mechanisms of cell determination and differentiation in the early embryo when the body plan is laid down for the first time were still so controversial and poorly understood. Much of my own limited understanding stemmed from those early, pioneering experiments dissecting out and exchanging the organizer regions of amphibian embryos from different species to see the effects on development [Spemann and Mangold 2001]. As the authors acknowledge, these experiments are as pertinent today as they were when first reported in 1923. Moreover, I was under the impression that the central role of morphogenic gradients in giving cells positional information was by now the accepted model for fate determination in the developing embryo although even Wolpert has lamented recently that the science underpinning pattern formation is far from complete [Wolpert and Vicente 2015]. So, the Gordons’ entirely reasonable criticisms of the prevailing gradient hypothesis and their detailed descriptions of the ‘cell state splitter’ organelle and its mechanical signalling mechanism as described in some detail in this book are completely new to me and I daresay, will be to most of you reading this review. To some extent at least, my ignorance arises from the reductionist ‘silo’ approach to basic investigation that the Gordons eschew as far as possible but which has grown to dominate our contemporary Western research culture, particularly when it comes to funding decisions. It did not surprise me to learn that among their first reported theoretical and practical considerations for the cell state splitter was an article published in a Russian development journal [Martin and Gordon 1997; Bjorklund and Gordon 1994].

Reading selections from the comprehensive background literature provided for each chapter, the mechano-chemical system the Gordons predicted and then began to test and which is described in some detail in the book’s most salient chapters (6 and 7) dates from the late 80s and early 90s. Their ‘cell state splitter’ is actually a cytoskeletal organelle, related to but independent of the spindle, sitting at the apical end of ectoderm cells in the developing axolotl embryo and signalling mechanical deformation in the form of cellular expansion or contraction to the cell nucleus. The nucleus ‘acknowledges’ receipt of this signal by switching its gene expression program in one of two possible ways, depending on whether the signal was from an expansion or a contraction event. It is this binary choice that gives the ‘splitter’ its name. The expansion or contraction of the cell is transferred to adjacent cells via adherans junctions leading to waves of expansion or contraction that have been recorded sweeping across the embryo (albeit rather sedately) and propagating the mechano-chemical sensory signal on to neighbouring cells [Brodland et al. 1994]. The book devotes a whole chapter to cell signalling and includes a good solid framework for how mechanical sensing and biochemical/molecular signalling can be merged and reconciled. The Gordons even provide a simple coding system based on the number and order of expansion or contraction waves that cells encounter and propagate, determining their differentiation pathways, migration, and final fates in the developing embryo. None of these events requires or has any need for a diffusible morphogen and together, they offer a compelling alternative explanation to the still extant consensus that synergistic or antagonistic gradients of diffusible morphogens govern differentiation and cell fate in the developing embryo [Wolpert 1989]. It is hard to see how such a synthesis could have been developed without the polymathic breadth and depth of knowledge that comes through clearly from the Gordons’ writing. I tip my hat at Richard Gordon’s seemingly encyclopedic knowledge that made it possible for him to construct this model.

I am, however, left wondering whether the Gordons might be doing themselves no favours by insisting that the splitter and diffusible morphogen models are mutually exclusive. My take on the more recent literature is that they could also be working in synergy. A recent, highly relevant and appropriate review on the mechanics of developmental processes [Miller and Davidson 2013] not only cites some of Gordon’s earlier work, but also cites research that while not mentioning the cell splitter directly, to my mind describes quite similar mechanosensors. This review does not include anything on diffusible morphogens and pattern formation, which might suggest a wider failure to connect them with mechanosensory signalling or consider them in that context. Assuming the Gordons view of morphogens is broadly correct, the problem as ever is proving which of them is the primary determinant. Furthermore, as revealed by another, more recent review on the subject, the whole business of diffusible morphogens as mediators of embryonic development and fate is still very much alive with Bone Morphogenetic Protein (BMP) signalling the focus of considerable contemporary research in both vertebrate and invertebrate lineages, although the same anti-morphogen arguments highlighted by the Gordons may, of course, still apply [Bier and De Robertis 2015]. So to my way of thinking, the main argument is not so much with the existence of the cell state splitter (or something similar), but proving that mechanosensory signalling and differentiation trees are all that is needed to explain embryogenesis. There lies the paradigm shift that could relegate diffusible morphogens to second fiddle. I can certainly sympathize with the argument that it is not they who are at fault, but a community refractory to new ways of thinking. But I also cannot help thinking that their fight with diffusible morphogens is a distraction that does not serve them well either. It is surely not outside the Gordons’ or others’ expertise to develop computer models that can ‘predict’ cell fate based on their splitter and differentiation tree mechanism and virtualise their model in silico. That way, people could design experiments based on the computer models to test their hypotheses and predictions in situ.

So, in many ways, like the Gordons’ earlier work on the subject [Brodland et al. 1994; Gordon and Brodland 1987], Embryogenesis Explained is really a plea to the scientific community and to classical embryologists alike to free themselves from the straightjacket posed by gradients, recognize the existence and role of the cell state splitter, and start working to independently verify his and his co-workers findings. As anyone (including the Gordons) trying to find money to pay for such paradigm shifting research will confirm, the challenge is far from trivial. Writing this review from the perspective of the original literature (or rather what time I have had to read of it), my conclusion is that Embryology Explained is an excellent primer for people wanting to learn something about developmental biology and its underpinning disciplines. But reading between the lines, there is little that I can see dividing the Gordons and the ‘mainstream’ community of developmental biologists. I am sure they would not disagree with the Gordons’ emphasis on mechanosensory information in the guidance of embryo development. Even the cell state splitter is something that has already been described if not yet recognised as the Gordons’ mechanosensory organelle. The main difference between the Gordons and the ‘rest of the world’ is the former’s assertion that the splitter and not diffusible morphogens determines cell fate. I am not sure the wider community would even argue about that if more evidence was available. In that regard, it is surely not beyond the capabilities of a code-literate champion of the splitter hypothesis to develop simplified computer models of virtualised embryos and use the differentiation fate rules described by the Gordons to mimic furrowing, pattern formation and cell fate determination in silico. Such modeling would certainly help design and justify experiments aimed at confirming the hypothesis in vivo. If the UK law is ever changed to permit development in vitro beyond formation of the primitive streak, the sophisticated embryo recording equipment now installed in many of our IVF labs could be gainfully employed to confirm the sweep of developmental furrows in the embryos of our own species.

An excerpted and distilled summary of the concepts and ideas so broadly described in the book is available to anyone wishing to learn a bit more about the Gordons’ work and the cell state splitter model [Gordon and Gordon 2016b].

Declaration of interests

The author has no conflict of interest to report.