Axons emanating from dendrites: phylogenetic repercussions with Cajalian hues

INTRODUCTION The observation that some axons, in hippocampal pyramidal neurons, emanate from dendrites rather than the somatic envelope, and the contributory synaptic input privilege (Thome et al., 2014) was highlighted as a new dimension in understanding input-output transformations (Kaifosh and Losonczy, 2014). Applying state-of-art immunocytochemical, neurophysiological, and computational methods to Thy1-DsRed transgenic, wild-type C57BL/6J mice, and Wistar rats, Thome et al. (2014) found that one-half of CA1, one-third of CA3, and one-fifth of subicular pyramidal cells feature axoncarrying dendrites (ACDs). Two-photon glutamate uncaging onto ACD spines led to excitatory postsynaptic potentials (EPSPs) and output action potentials (APs) at lower activation -thresholds compared to nonACDs (Thome et al., 2014). EPSPs arising at nonACDs pass the soma and attenuate before reaching the axon initial segment (AIS), whereas in ACDs the electrotonic distance between excitatory synapses and the AP trigger zone is shorter. ACDs have a higher propensity to generate active dendritic d-spikes, primarily mediated by voltage-gated Na+ channels, and consistent with a lower functional density of A-type K+ channels (Kaifosh and Losonczy, 2014). A question outstanding is to what extent this dendrito-axonal particularity is evolutionarily regulated across species (Kaifosh and Losonczy, 2014). I weave data from comparative studies to illuminate its phylogenetic basis. MAMMALS TO INSECTS Axons occasionally emanate from dendrites in neocortical pyramidal cells (Peters et al., 1968; Sloper and Powell, 1979), midbrain dopamine neurons (Häusser et al., 1995), GABA interneurons targeting CA1 pyramidal dendrites (Martina et al., 2000), and neuroendocrine cells (Herde et al., 2013). Generalities regarding mechanisms are greater at lower than higher levels; to decipher connections leading to motor patterns, neurons and networks are analyzed in molluscs, crustaceans, lampreys, worms, amphibians, and arthropods (Burrows, 1992). Intracellular recordings from insect motoneurons reinforce and enlarge on synaptic integration principles from the mammalian spinal cord (Simmons and Young, 2010). In the flight motoneurons of the locust, which control the forewing or hindleg muscles, the dendritic arbor and the axon are combined into one process, attached to the soma by a slender stalk. The soma is not in the signal flow pathway. While axonal spikes are 100 mV in amplitude and <1 ms in duration, somatic signals conduct passively (10 mV lasting several ms). Dendrites with voltage-sensitive channels in their membrane can boost the amplitude of input signals. Thus, when an electrode is inserted into the motoneuron dendritic fan, near the axon origin rather than the soma, EPSPs rise and fall more sharply (Burrows, 1992; Gabbiani et al., 2002; Simmons and Young, 2010). RAMÓN AND CAJAL Axons arising from dendrites in vertebrates and invertebrates were documented by Santiago Ramón y Cajal and his younger brother Pedro Ramón. In 1897 Cajal realized that “contrary to the general opinion, the soma does not always participate in the conduction of received nerve impulses; the afferent wave is sometimes propagated directly from dendrites to axon” (Ramón y Cajal, 1937). Accordingly, Cajal substituted his principle of dynamic polarization with axopetal polarization, whereby soma and dendrites conduct the waves of the nervous excitation toward the axon, which in turn carries impulses toward its terminal arborizations. “Currents flowing into the axon do not pass through the soma except when the latter is between the dendritic and the axonal apparatus” (Ramón y Cajal, 1897, 1937). Given the integrative processes of neuronal dynamics, one gathers that, when the axon fires ahead of the soma, it must emerge from a dendrite (Häusser et al., 1995; Yuste and Tank, 1996). Dendrite-derived axons occur in unipolar neurons of invertebrate abdominal ganglia (Ramón y Cajal, 1899, chapter V, “Physiological inductions of neuron morphology, and connections”). The segment that connects the dendritic stem to the initial part of the axon was called accessory process by Retzius (Figure 1). Dendrite-derived axons were documented by Ramón and by Cajal in the corpúsculos del cayado (“crosier” or “shepherd’s crook” cells), and in elongated fusiform cells of the optic lobe of birds, reptiles, amphibians, and fish (Ramón, 1890, 1891,