Multiple Origins of Neurons From Secretory Cells.

Abstract

The elusive diversity of neurons puzzled neuroscientists since discovering the first nerve cells in the 1830s. Quantitative information about neuronal diversity began to flow from the middle of the twentieth century. At that time, microelectrode and histochemical tools were applied to vertebrate and invertebrate preparations. Simpler nervous systems of some gastropod mollusks, annelids, and nematodes revealed identified neurons with defined transmitter specificity and functions. Early systematic studies pointed out that most of the neurons composing their nervous systems might be unique (Bullock and Horridge, 1965). That revelation provided tractable experimental preparations to decipher cellular bases of behaviors (Kandel, 1976, 2001; Kuffler and Nicholls, 1976). Today, with advances in single-cell (epi)genomics and transcriptomics, the astonishing diversity of neuronal cell types exceeds any imagination (Moroz, 2018). The most straightforward question is, how different are the neurons? But more fundamental questions are: Why are neurons different? Why are there so many neurotransmitters? Why are neurotransmitters different? These questions have been addressed by many (Kandel, 1979; Van Vallen, 1982; Bloom, 1984), aiming for functional aspects. In 1968–1974 these questions were asked from an evolutionary standpoint, and Dmitry Sakharov had proposed the hypothesis of neuronal polygeny (=multiple origins of neurons) (Sakharov, 1970a,b, 1972, 1974a,b). Using minimal comparative data available 50 years ago, Sakharov suggested that neurons evolved from genetically different secretory cells. The evolutionary view of neuronal evolution can be summarized as follows. Each of these populations of secretory cells could use chemically distinct transmitter(s) and different (distant) receptors for communications in early neural systems, where synapses are not required. Ancestral diversity of secretory cell types (=secretory phenotypes) has been preserved over 500+ million years of biological evolution, forming lineages of homologous neurons across phyla. Thus, neurons are different because they have different genealogies. Subsequent functional “demands” and specifications could further tune these different ancestral neurosecretory phenotypes. In other words, the traditional one-root genealogy of neurons was transformed into multiple genealogies or a net of phyletic cell/neuronal lineages, as schematically presented in Figure 1. Open in a separate window Figure 1 Multiple origins of neurons and secretory cells. Schematic illustration of the ancestral cell lineages (different color trajectories) that led to the exant neural systems in four basal metazoan clades with Placozoa as nerveless animals. Neural systems might consist of genetically diverged cell types with different ancestries, gene regulatory networks, and signal molecules. This diagram integrates both the hypothesis of independent origins of neurons (as in ctenophores, Moroz et al., 2014) and the sister-cell model (Arendt et al., 2016), which suggests that novel neuronal types arise in pairs, through sub-specialization of ancestral cell types. Thus, sister neuronal subtypes can share gene-regulatory networks, perhaps, evolutionary conserved developmental pathways, and are expected to have more similar expression profiles than each of them compared to other neuronal types. The key prediction of this model is that gene expression profiles from sister-cell types will form a hierarchical tree structure in phylogenetic reconstructions. A complementary model predicts that neurons and novel neuronal subtypes arise through “co-options” or “fusions” of regulatory modules and pathways “recruited” from genetically unrelated cell types. As a result, their expression profiles would be substantially different, leading to net-type rather than tree-type cellular genealogies in phylogenetic reconstructions. We expect that both scenarios can coexist in any given nervous system. But the tremendous diversity of neural systems across phyla suggests variable contributions of each historical scenario. Combining tools of (i) statistical geometry, artificial intelligence and (ii) modern phylogenomics with (iii) massive parallel single-neuron transcriptome profiling would allow us to unbiasedly reconstruct the genealogy of neurons by testing the treeness statistics as it was recently used for cancer and placental cells. The top illustrations are photos of Mnemiopsis, Trichoplax, Podocoryne, Priapulus, and Aplysia. Some cell lineages (different colors) might become eliminated in the course of evolution (loss) or be expanded or evolved in parallel from different secretory cell types.