How many times has a centralised nervous system evolved?

Neuroscience interests me because of the idea that an object (the human brain) is attempting to understand itself. For that same reason, I’m fascinated by the evolution and development of the central nervous system. Scientist’s brains are attempting to discover exactly how they came into existence. Our central nervous system obviously appeared a long time ago. Cows, crows and clownfish all have a brain for the same reason we do: it already existed in the most recent common ancestor of these animals. Obviously your brain is a bit different than that of a clownfish, but both are just uniquely modified versions of the same thing. You’ve probably noticed that these animals are all chordates, so what about brains we find in even more distant relatives? Is the fly’s brain homologous with our own? What about a worm? Are your brain and a worm’s brain just derived versions of a brain that existed in the last common ancestor between yourself and worms? Or have the lineages that led to worms and humans invented their brains independently?

Central nervous systems are observed in many species throughout the bilateria, in both protostomes (e.g. arthropods, annelids, molluscs) and deuterostomes (e.g. chordates). The origin of the central nervous system is one of the key debates taking place in evolutionary biology. How many times has it evolved? Have different taxa evolved a central nervous system independently of each other? Or do all central nervous systems share a common ancestry, originating in an ancestor of bilaterians? Attempts to answer these questions involve debates over the phylogenetics of major taxa, the developmental biology of key taxa, and the analysis of deep homology.

The last bilaterian common ancestor

Fossil-calibrated molecular clocks date the origin of the bilaterians (animals with bilateral symmetry) to approximately 700-600 million years ago. The oldest tripartite brain discovered in the fossil record is from a 520 million year old arthropod. The morphology of the last bilaterian common ancestor (LBCA) is currently unknown. One hypothesis is that the LBCA was a relatively simple animal with only a diffuse nerve plexus, suggesting that the central nervous system (CNS) evolved several times independently in various later taxa. So a fly’s brain and your own brain would be independent inventions of evolution.

An alternative hypothesis is that the LBCA was relatively complex and already possessed a tripartite brain, suggesting that the CNS found in both deuterostomes and protostomes is homologous, and that many bilaterian taxa secondarily lost the CNS. With no fossil evidence of the LBCA, evidence must be obtained from studying other fossils and extant taxa. It is therefore essential that phylogenies are reconstructed as accurately as possible if they are to be used to predict the features that were present in the LBCA. However, reconstructed phylogenies are constantly changing as research takes place, further complicating the study of CNS evolution.

Historically, the chordate CNS was thought to be derived and that the simple nervous systems observed in various hemichordates and ascidians resembled those of chordate ancestors, suggesting the CNS had evolved at least twice in bilaterians. It is now known that many of these simpler taxa are not basal to the deuterostomes and that their nervous systems are actually secondarily simplified. Inspired by Geoffroy Saint-Hilaire’s “unité de plan” in 1822, Dohrn proposed that the chordate dorsal cord is homologous to the protostome ventral cord, suggesting that both groups share a CNS that existed in the LBCA. Although the dorso-ventral inversion hypothesis was abandoned for decades, many researchers have recently found evidence supporting the idea.

Evidence that the protostome and deuterostome CNS may be homologous

Some genetic processes involved in dorsoventral patterning are similar in protostomes and deuterostomes but inverted. Left-right patterning is also inverted between various taxa, including left-sided Pitx and Nodal expression in chordates and right-sided expression in the larvae of sea urchins and starfish. If the LBCA and the lineage that led to deuterostomes possessed a CNS, then centralised neurons are likely to be observed in the basal deuterostomes. Echinoderms are bizarrely divergent in almost every sense (body plan, genome, genetic signalling etc) but recent studies of hemichordates may support the DV-inversion hypothesis. Previous research of juvenile Saccoglossus kowalevski (hemichordate “acorn worms”) found only scattered epithelial expression of nervous system patterning genes, suggesting a diffuse nervous system. But recent research found that adults of Saccoglossus kowalevski possess a fully-formed CNS interpreted as a transition between protostome-like and deuterostome-like nerve cords. This hemichordate possesses both ventral and dorsal cords. A reminder: we have a dorsal cord, a protostome such as a fly has a ventral cord. The apparent separation of CNS and PNS in this species, taken with the evidence of BMP patterning in acorn worms, supports the view that the neural plate, collar cord, circumesophageal tract and ventral cord all correspond to the whole chordate CNS. This suggests that chordate ancestors underwent DV-axis inversion, and could be interpreted as evidence that the CNS was present in the LBCA. Understanding the development of various taxa is essential, but the knowledge can only be applied to explaining the origin of the CNS if our understanding of deep phylogenetic relationships is accurate. Many recent studies are rewriting what we know about deep relationships among taxa.

Complications caused by incomplete and changing phylogenies

One way to predict the morphology of the LBCA is to consider the basal bilaterians, but the acoel flatworms previously regarded as basal bilaterians (Acoelomorpha) have recently been reclassified as deuterostomes. Their location in phylogenies may affect the predicted morphology of the LBCA. Both acoels and the Xenoturbellida were first presumed to be Platyhelminthes (protostomes) until the acoels were considered as basal among bilaterians and Xenoturbella was classified as a deuterostome. In 2009, Hejnol et al grouped Xenoturbella and the acoelomorphs together as sister taxa, basal to the bilateria. Philippe et al re-examined Hejnol’s data and kept both taxa together (Xenacoelomorpha) but as deuterostomes, placed as a sister group of the Ambulacraria (echinoderms and hemichordates). This reclassification has been rejected by some researchers, claiming that Philippe et al didn’t include species that contradicted their hypothesis. One of the researchers had previously found that Meara stichopi did not have deuterostome microRNA, despite being closely related to acoels. It is also known that microRNAs can be lost in evolution, so perhaps all bilaterians possessed them but the lineage that led to protostomes lost them. According to a recent outgroup analysis of cephalic neural characters across extant metazoans, the reclassification shouldn’t affect the debate of CNS origins as Xenoturbella and the acoels have a diffuse nervous system, not a CNS with simple ganglia or a complex brain. The reclassification was thought to be compatible with the view that the LBCA possessed a diffuse nerve plexus and the vertebrate CNS evolved in an ancestor of the Chordata.

Bery et al (2010) challenged this compatibility by describing a central nervous system in a juvenile acoel, Symsagittifera roscoffensis, consisting of nerve cords and a compact brain that develops around a central neuropile. If considered basal to the bilateria, the acoel nervous systems could partly represent that of the LBCA, thus suggesting the ancestor of bilaterians possessed a CNS. Although this is compatible with the DV-axis inversion hypothesis, it isn’t the most parsimonious explanation, proposing that the CNS was lost in many independent taxa. The discovery of a CNS in a member of the Xenacoelomorpha affects one school of thought more than the other. Regardless of whether the acoels are considered basal to bilaterians or basal deuterostomes, both views are compatible with the hypothesis that the LBCA already possessed a CNS with a brain. But for the hypothesis that the CNS evolved independently in several taxa, the placement of acoels in bilaterian phylogeny affects the possible number of times the CNS has evolved. Northcutt’s analysis included the Xenacoelomorpha as deuterostomes with a diffuse nervous system and concluded that the CNS has evolved 4 times: in arthropods, annelids, molluscs, and chordates. The acoel CNS as described by Bery et al could either mean that the CNS has evolved 4 times including once in the ancestor of all deuterostomes (and was secondarily lost in some deuterostome taxa), or at least 5-6 times if the examples of a CNS in Xenacoelomorpha and Amulacraria are apomorphies.

Further complications due to the Mollusca and “penis worms”

These hypotheses may be conservative as they assume that the CNS evolved once in the ancestors of the Mollusca. There is ongoing debate regarding Molluscan relationships, leading to several competing hypotheses of Molluscan phylogeny. A recent phylogenomic study supported the “Aculifera hypothesis”, placing the Polyplacophora together with the Aplacophora, and classifying Gastropoda and Bivalvia as sister-taxa within the Conchifera. This hypothesis suggests that the CNS evolved independently in various molluscan taxa. Moroz claims that the CNS may have evolved at least 5 times within the Mollusca and at least 9 times within the bilateria. Molluscan taxonomy is highly controversial and opposing phylogenies are frequently suggested. Even small changes within molluscan phylogenies can impact our estimate for the number of times the CNS has evolved, and some of the proposed hypotheses make drastic changes. Goloboff et al went as far as suggesting that the Scaphopoda and Bivalvia must be entirely removed from Mollusca in order to make it monophyletic. A lot of phylogenomic work clearly still needs to take place for Mollusca before the origins of molluscan central nervous systems can be inferred.

Even the details of widely-accepted taxonomical groups are occasionally challenged. Last year an interesting paper was published that may have consequences for some of the phylogenies already discussed. While other studies often debate the exact relationships within various closely-related taxa, Martín-Durán et al have suggested that the entire protostome-deuterostome split in the bilateria could be more complicated than first thought. The authors found that Priapus caudatus, a priapulid “penis worm”, develops like a deuterostome despite being classified as a primitive protostome. This finding isn’t unique, as many protostomes have been found to show deuterostomic development (anus develops first) or even amphistomic development (mouth and anus develop simultaneously), but Priapus caudatus is so primitive that its development has long been assumed to partly resemble that of the earliest protostomes. Researchers may have to rename the “protostomes” to reflect these findings. Deuterostomic development in the earliest protostomes increases the probability that the LBCA may have developed anus-first. If correct, various taxa within the Protostomia evolved protostomic development independently, and that extra data might further evolve our understanding of phylogenetic relationships among protostome taxa and therefore the number of CNS origins.

Deep homology – just how homologous are morphologies?

Recent evidence supports the hypothesis of a complex common ancestor of animals though the hypothesis that the LBCA possessed a complex CNS is complicated by our understanding of deep homology. Whether the LBCA was simple or complex, its ancestors had already evolved genes for the developmental genetic networks that determine dorso-ventral and anterior-posterior patterning. This has been taken by many researchers as evidence of homology between the CNS in protostomes and deuterostomes. However, diagnosing morphological homology is a thorny issue as some genetic networks determine character identity and others determine character state. We now have more data from genome studies than ever before, and can go further than simply matching homologous genes and morphologies.

Researchers can now construct gene networks for the development of morphological characters. Otherwise, researchers may find convergent morphologies caused by homologous genes and mistakenly conclude that the morphologies themselves are homologous. This problem has been described as the “inverse-paradox”. Similar morphologies can evolve in several taxa even when the genetic toolkits are variable. Even among Chordata, there is surprising variation in the development and gene networks for what appear to be homologous characters. Researchers should be wary of the inverse-paradox when making predictions regarding the nervous system of the LBCA. Gene networks for the development of morphological features need to be carefully studied. Although protostomes and deuterostomes possess homologous genes that are involved in the development of the CNS, their anterior-posterior (or dorsoventral) patterning may have preceded the evolution of a CNS. This is why I’m so interested in co-option during evolution (teaching old genes new tricks).

Further genome sequencing and comparisons of genetic networks across a wide range of taxa are required if we are to answer the question of how many times the CNS has evolved. The related efforts of profiling genetic networks and reconstructing phylogenies will have to feed into one another in future research. We may never know the exact number of times a central nervous system has evolved, but perhaps one day we will know whether or not the last bilaterian common ancestor already possessed a central nervous system. There appears to be evidence that the LBCA had a CNS, but it’s best to remain cautious.