Evolutionary origins of the nervous system

THE human brain is a true marvel of nature. This jelly-like 1.5kg mass inside our skulls, containing hundreds of billions of cells which between them form something like a quadrillion connections, is responsible for our every action, emotion and thought.

How did this remarkable and extraordinarily complex structure evolve? This question poses a huge challenge to researchers; brain evolution surely involved thousands of discrete, incremental steps, which occurred in the mists of deep time across hundreds of millions of years, and which we are unlikely to ever fully understand.

Nevertheless, a number of studies published in recent years have begun to shed some light on the evolutionary origins of the nervous system, and provide clues to some of the earliest stages in the evolution of the human brain. These clues come from the most unexpected of places – from sea sponges, which lack nervous systems altogether, and from the extant descendents of a primitive worm which lived some 600 million years ago.

Despite their differences, vertebrates, worms and insects are all believed to be descended from a common ancestor – a worm-like organism, named Urbilateria, which lived some 600 million years ago. Urbilateria displayed bilateral symmetry – its body was symmetrical along its longitudinal axis – and this body plan was inherited by the diverse array of organisms descended from it. But, according to new research, published in 2007 in the journal Cell, it wasn’t just bilateral symmetry that the descendants of Urbilateria inherited: at the earliest stages of their evolution, vertebrates – including humans – may have inherited the organization of their nervous systems from it as well.

The simplest nervous systems lack a brain, and instead consist of diffuse networks of nerves. The nervous systems of vertebrates and annelid worms, however, are organized in another way, with nerve fibres arranged in centralized cords, and large groups of nerve cells (called ganglia, singular ganglion). The major differences between them are, of course, the level of complexity, and the positioning – the nerve cord of invertebrates is located ventrally (toward the belly), whereas the vertebrate spinal cord is located dorsally (toward the back). In 1875, Anton Dohrn proposed the annelid theory, according to which, the vertebrate central nervous system arose after a proto-vertebrate inverted itself along the dorsal-ventral axis. This conflicts with the view that the vertebrate and invertebrate nervous systems evolved separately, on the dorsal and ventral sides of the body, respectively.

During vertebrate development, different types of neurons are generated in a specific pattern across the dorso-ventral (D-V, or back-to-stomach) axis, as well as along the anterior-posterior (A-P, or longitudinal) axis, of the neural tube (the early developing nervous system). This pattern is brought about by the differential expression of genes encoding diffusible chemical signals. The concentration of a signal in a given location specifies a spatial domain within which a particular type of neuron will be generated. Each domain contains undifferentiated cells that express a unique combination of these chemical signals. The combination of signals in a domain constitutes a “molecular fingerprint” that determines the differentiation pathway of cells within that domain. Each signal is a transcription factor – it binds to DNA and switches a specific gene or set of genes on or off (and each is itself regulated by a diffusible signalling factor belonging to the BMP protein family). This drives the differentiation of the cell along a particular pathway. In this way, sensory neurons are generated in the dorsal half of the neural tube, and motor neurons are generated more ventrally.


Detlev Arendt and his colleagues, of the European Molecular Biology Laboratory in Germany, determined the patterns of these developmental signals in the marine ragworm, Platynereis dumerilii (above), and compared them with the patterns found in development of the zebrafish, which is a vertebrate. The expression patterns of genes encoding the annelid versions of three proteins known to be involved in dorso-ventral patterning of the vertebrate nervous system (Nk2.2, Pax6 and Msx) were determined in worm larvae, by in situ hybridization and antibody staining. Confocal microscopy. It was found that the developing nervous system of the ragworm, just like that of the zebrafish and other vertebrates, is subdivided into discrete domains, each of which is characterized by the same molecular fingerprint as its counterpart in the zebrafish.

Thus, in the ragworm, as in vertebrates, one combination of signals generates motor neurons near the ventral midline, another generates sensory neurons near the dorsal midline, and a third generates interneurons in between. Furthermore, the chemical signals in the ragworm were found to be sensitive to the same regulatory factors that are involved in patterning the D-V axis of the neural tube of the zebra fish and other vertebrates. They locally increased the concentration of a member of the BMP family; this inhibited expression of the chemical signals that would normally be found in the targeted domain, changing the fingerprint of the cells in that domain and the pattern of cell types across the D-V axis.

There are, of course, no Urbilateria fossils in which these gene expression patterns can be investigated. But Platynereis is considered to be a “living fossil”, and, as such, is thought to resemble the common ancestor more closely than most other extant organisms. The study by Arendt’s group therefore supports Dohrn’s annelid theory by providing evidence that vertebrates, worms and insects all inherited their central nervous system from their common ancestor, Urbilateria. However, Chris Lowe, an assistant professor in the University of Chicago’s Department of Organismal Biology and Anatomy, has evidence that suggests otherwise. Lowe works with another descendant of Urbilateria, a worm whose nervous system consists of a diffuse network of cells (unlike vertebrates and Platynereis, in which the nervous system is centralized). He points out that this animal uses the same signals as vertebrates for patterning the anterior-posterior axes, even though its nervous system is organized in a different way. So, in Urbilateria, central nervous system development and axis patterning could well have been separate mechanisms that evolved independently.

“Such a complex arrangement could not have been invented twice throughout evolution. It must be the same system,” says co-author Gáspár Jékely. For him, and other members of Arendt’s lab, the question is: how did the inversion from ventral to dorsal take place? According to Dohrn, the ventral-to-dorsal relocation of the central nervous system was simply the result of an inversion of the entire body, so that the belly became the back, and vice versa, after which gill slits nearest the front of the body formed a mouth. Arendt suggests another possibility – that Urbilateria went from being a burrowing worm that spent much of its life partially buried in the seabed to a free-swimming one. With this adaptive change in lifestyle, the pioneer of the vertebrate lineage would have been surrounded by water in all directions, and the body would have been fixed in a new belly-up orientation.

More insight into the evolutionary origins of the nervous system come from a study of sea sponges, which was published in the open access journal PLoS One at around the same time. Sea sponges are sedentary organisms that attach themselves to the sea bed and filter nutrients from the water that they force through their porous bodies with flagella. They are the most primitive of all multicellular animals, with just four different types of cells making up partially differentiated tissues in a simply organized body. Because of the lifestyle they lead, sea sponges do not need, and therefore lack, nerve cells, muscle cells and internal organs of any kind. However, the PLoS One study shows that at least one species of sea sponge, called Amphimedon queenslandica, synthesizes many of the proteins that are essential for the cell-to-cell communication that takes place within nervous systems. These surprising findings provide clues about how the first neurons may have evolved in the most ancient of animals.

Neurons are specialized to communicate with one another, and this communication takes place at a structure called the synapse, a miniscule gap of about 40 nanometres found at the junction between adjacent cells. For the signalling to be effective, it is crucial that all the proteins involved are organized correctly. On both sides of the synapse (at the presynaptic and postsynaptic membranes) the signalling components are organized by a scaffold of proteins called the pre- and postsynaptic densities. These structures are a specialization of the cytoskeleton found just beneath the pre- and post-synaptic nerve cell membranes. The density is an extremely complex and highly dynamic network – in humans it contains perhaps several hundred different types of protein – which organizes the molecular machinery needed for a neuron to detect and respond to the chemical signals sent to it by adjacent cells, and mediates the movements of the machinery within the membrane in response to neuronal activity.

In the PLoS One study, Kenneth Kosic and his colleagues from the University of California at Santa Barbara analyzed the Amphimedon genome, and found that it contains 36 families of genes known to encode proteins of the post-synaptic density. So, even though it has no neurons, this sea sponge synthesizes an almost complete set of post-synaptic density proteins. A comparison of the DNA sequences from the 36 sea sponge genes with the homologous sequences from humans, Drosophila melanogaster (fruit flies) and Nematostella vectensis (a cnidarian with a simple nervous system consisting of a loose network of nerves) revealed striking similarities between the genes in all four species. One gene, called dlg, encodes a crucial component of the post-synaptic density scaffold. The protein product of that gene contains a number of regions that form the protein-protein bonds that hold the scaffold together. The segment of the dlg gene encoding these binding regions was found to be highly conserved – the DNA sequences in the sea sponge gene were identical to the human sequences. This suggests that in the sea sponge these proteins interact in exactly the same way as they do in the human postsynaptic density.

Amphimedon has nearly all the components required to make a post-synaptic density; only a few of the human postsynaptic density genes are missing from the sea sponge’s genome – those encoding ion channel receptors for the neurotransmitter glutamate. (These genes are, however, present in cnidarians, which express them in its simple nervous system.) In sponges, the genes are expressed predominantly in the flask cells of the free-swimming larvae, where they may be involved in sensing chemical cues found in the organism’s environment. Flask cells with post-synaptic densities may predate the first neurons. If so, the first synapses may have evolved from postsynaptic densities in a process called exaptation, whereby a pre-existing structure is modified slightly to perform a new function. It is, however, also possible that flask cells evolved from simple neurons that lost some of their synaptic components.

More recently, Seth Grant and his colleagues at the University of Oxford carried out a comparative study of the composition of the postsynaptic density in a wide variety of organisms across animal kingdom. Using a technique called proteomics, they analysed more than 600 postsynaptic proteins in 19 different species, including yeast, fruit flies, mice and humans. Traditionally, it has usually been assumed that brains grew larger as they evolved, and there has been a long history of linking brain size with intelligence. However, the analysis performed by Grant’s suggests that this is not necessarily the case – it shows that brain complexity, rather than size, may be the most important factor which determines the behavioural repertoire of an organism.


Denes, A. et al (2007). Molecular Architecture of Annelid Nerve Cord Supports Common Origin of Nervous System Centralization in Bilateria. Cell 129: 277-288. DOI: 10.1016/j.cell.2007.02.040.

Sakarya O., et al (2007). A post-synaptic scaffold at the origin of the animal kingdom. PLoS ONE 2: e506. doi:10.1371/journal.pone.0000506. [Full text]

Emes, R.D. et al (2008). Evolutionary expansion and anatomical specialization of synapse proteome complexity. Nat. Neurosci. DOI: 10.1038/nn.2135.

8 thoughts on “Evolutionary origins of the nervous system

  1. Hi
    The brain from bottom to top seems to advance in complexity (leave alone that every part of brain is important by its own). Does this have anything valuable in terms of Evolution of Brain?

  2. Lowe works with another descendant of Urbilateria, a worm whose nervous system consists of a diffuse network of cells (unlike vertebrates and Platynereis, in which the nervous system is centralized). He points out that this animal uses the same signals as vertebrates for patterning the anterior-posterior axes, even though its nervous system is organized in a different way. So, in Urbilateria, central nervous system development and axis patterning could well have been separate mechanisms that evolved independently.

    This view is challenged by a recent paper, blogged by Zen at NeuroDojo: Centralization of the Deuterostome Nervous System Predates Chordates by Marc Nomaksteinsky, Eric Röttinger, Héloïse D. Dufour, Zoubida Chettouh, Chris J. Lowe, Mark Q. Martindale, and Jean-François Brunet (doi:10.1016/j.cub.2009.05.063):

    […] We show here that juvenile and adult enteropneust worms in fact have a bona fide CNS, i.e., dense agglomerations of neurons associated with a neuropil, forming two cords, ventral and dorsal. The latter is internalized in the collar as a chordate-like neural tube. Contrary to previous assumptions, the greater part of the adult enteropneust skin is nonneural, although elements of the peripheral nervous system (PNS) are found there. We use molecular markers to show that several neuronal types are anatomically segregated in the CNS and PNS. These neuroanatomical features, whatever their homologies with the chordate CNS, imply that nervous system centralization predates the evolutionary separation of chordate and hemichordate lineages. [emphasis mine]

    See also my comment at the NeuroDojo post.

  3. You REALLY need to study some invertebrate zoology before writing a blog like this. I cringe. For starters, Playtnereis, a polychaete annelid, is a long, long way from the Urbilateria. If you want to see a modern version of “Urbi,” try taking a look at any of the Acoelomorph worm.
    This book may help you get some of the concepts right:
    Ruppert, E. E, R. S. Fox and R. D. Barnes. 2003. Invertebrate Zoology, A Functional Evolutionary Approach. 7th Ed. Brooks/Cole-Thomson Learning. Belmont, CA. xvii +963 pp.+ I1-I26 pp.

  4. Hi,
    Nice post, but I have a technical comment though:
    as a PhD student working on Platynereis dumerilii, I can assure you that we are still trying to fuse genes encoding fluorescent proteins and to insert the constructs into larvae… The results published in the Denes et al., although very important, are “only” in situ hybridization (revealing RNA with an antisense probe).
    And yeah, Chris Lowe seems now to have a different opinion about the centralisation of the nervous system among hemichordates… See the Nomaksteinsky et al. published in Current Biology last week.

  5. Good post but…I’m quite puzzled about the possibility
    to understand mental functions by knowing the evolutionary origin of the nervous system. The gap between mind and brain seems to me to be open and very wide!!

Comments are closed.