Neuron

• Dr. Rodolfo Llinas, NYU Medical Center, New York, NY, USA

Figure 1: Drawing of neurons showing soma and dendrites in black, axon in red. A. Cortical pyramidal cell. B. Cerebellar Purkinje cell. C. Spinal motoneuron. D. Inferior olivary cell. (A-D drawings by Cajal) E. Cell sensitive to water movement in the medicinal leech, (Gascoigne & A McVean 1991).

Neurons are the cells that underlie the function of the nervous system including the brain, spinal cord, peripheral sensory systems and enteric (gut) nervous system. The anatomical variation of these neurons is large, but the general morphology allows these cells to be classed as “neurons” (coined in 1891 by Wilhelm von Waldeyer).

Neuronal form

Neurons are generically characterized by a central cell body or soma that comes in different shapes. The body houses the cell nucleus and most of the genomic expression and synthetic machinery that elaborates the proteins, lipid, and sugars that constitute the neuronal cytoplasm. The membranous system bounds and defines different intracellular compartments and includes the outer cell membrane (the plasmalemma) that encopasses the global compartment defining cellularity itself.

Associated with the plasmalemma (but also with other intracellular elements) are transmembrane macromolecules that control the functionality of nerve cells (voltage and ligand activated ionic channels, ionic pumps, non-gated ‘leakage’ channels) and the machinery that keeps the cells alive by continuously taking up and replacing the molecular modules that constitute the cell’s functional matrix.

Figure 2: A. Avian optic lobe neuron showing axon emerging from dendrite (Cajal). B. Spinal motor neuron labeled with antibodies to sodium channels at the axon initial segment (green) and neurofilaments (red) (reproduced with permission). C. Polarization of a developing spinal motor neuron that accompanies the clustering of sodium channels near the axon hillock (red). (reproduced with permission)

In general terms neurons are viewed as having an input and an output pole.

• The receiving or input pole generally consists of extensively branching tree-like extensions of the soma membrane known as dendrites (coined in 1889 by William His from dendros (Greek) meaning tree) which arises in vertebrate neurons directly from the cell body (the body is also a receiving site in most neurons). With invertebrate neurons the dendrites arise most commonly from the axon (see Fig. <ref>F3</ref>A and Fig. <ref>F4</ref>).
• The output pole, called the axon (coined in 1896 by Rudolph Albert von Kolliker; red in Fig. <ref>F2</ref>), arises as a single structure from the soma (and occasionally from a dendrite). The axon conducts propagating electrochemical signals termed action potentials (usually initiated at the axon hillock, green in Fig. <ref>F3</ref>B) away from the soma. There are some dendrites that also serve as output systems (see dendro dendritic synapses by G. Shepherd).

There are, however, exceptions to these general rules of neuronal organization. In some neurons, e.g. peripheral sensory neurons, the input occurs via axons.

Figure 3: Invertebrate neurons. A. A dye-filled antennal lobe projection neuron, one of the second-order neurons of the Drosophila olfactory system (http://wilson.med.harvard.edu). B. Giant neuron in a cricket (?) ganglion (reproduced with permission). C Confocal image of a leech ganglion from an embryo in which mechanosensory P cells were injected with a combination of the fluorescent dyes Alexa 488 and neurobiotin (visualized with Steptavidin-Cy3) and a local bending interneuron was injected with Alexa 488. The electrically coupled cells appear in red. The interneuron appears in yellow due to colocalization of the dyes (reproduced with permission). D. Two neurons in the auditory pathway of crickets named AN2 (red) and ON1 (green). (Dr. Gerald Pollack).

Examples of the great variety of neuronal form are shown in Figs. <ref>F1</ref>-<ref>F6</ref>. Those in images <ref>F4</ref>-<ref>F6</ref> have been made using more recent staining techniques. This set of images represents the common variety of neurons as well as the extremes of morphology to illustrate the enormous variety neurons can generate morphologically.

Neuronal function

Their electrophysiological properties are as rich as their morphology, and endow neurons with a vast set of electrical properties and functional styles. The voltage and ligand dependent ionic conductances that generate and modulate such excitability can implement autorhythmic properties either as single cell oscillators or resonators that ultimately dictate network oscillatory properties. Early electrophysiological results from the study of motoneurons led to the view that neurons do precious little, other than integrate and fire. Intelligence is in the network was the initial credo in those early years. Many other newer parameters have entered electrophysiology over the last three decades. Plasticity (mostly as Long Term Potentiation and Long Term Depression) and intrinsic electrical properties have been particularly significant.

Neurons are characterized by four main functional properties; a) electrical excitability (mostly across the plasmalemma), b) secretion (mostly vesicular and peptide extruding channel dependent), c) molecular synthesis (mostly proteins), and d) growth and plasticity.

This short review will deal mainly with electrical excitability. Secretion, mostly as the mechanism responsible for excitatory or inhibitory synaptic transmission, will be dealt in other entries.

Electrical excitability

When considering neuronal function, electrical excitability is indeed one of the main themes of concern. Four aspects are generally considered

Passive electrical properties

Passive properties refer to the capacitative and resistive aspects inherent in neuronal membranes, along with the resistivity inherent in the cytoplasm and the extracellular melieu. Together, these properties provide an electrical resemblance between neuronal processes (axons and dendrites) and conduction in electrical cables and hence are termed cable properties. Across this capacitive/resistive membrane an electric field and a voltage difference is maintained by the action of selective ion pumps. While the basic assumption of most electrophysiologists is that the membrane potential may be initially considered as having a resting value (the resting potential) that is uniformly distributed along neuronal compartments, this is an oversimplification, as the ionic conductances (and pumps) which are responsible for setting the resting potential need not have a fixed density throughout the neuronal membrane. Even so, isopotentiality is inherent in most initial cable property assumptions. The value of the electrical field (in mV) is related to the driving force (emf) for each of the ionic species that can move across the membrane and the magnitude of the conductance for each ionic species (i.e. the number of permeable channels available for any ion species at any given moment multiplied by the single channel conductance). While a more or less uniform permeability may be found for the leakage channels (a constantly open channel permeable to potassium), voltage gated channels allowing sodium, potassium, calcium, and chloride ions across the plasmalemma also contribute to the membrane potential to the extent of the magnitude of their conductance at a given time. Note however that these latter conductances, because they are voltage-gated, deviate from the simple passive character of the leakage channel.

Because of the complexity afforded by the branching pattern of the dendritic tree, the passive electrical properties of neurons were initially difficult to envision without a proper mathematical model. In the mid sixties of last century Wilfred Rall (see Rall Model) was the first to define the electrotonic cable properties of branching dendritic trees and to show the importance of such branching patterns on synaptic summation at different sites in the dendritic arbor.

Neurons conduct waves of membrane potential passively (electrotonically) a short distance along their processes as the result of currents that flow intracellularly along the longitudinal resistance and simultaneously across the plasmalemmal membrane as resistive or capacitative current. When active properties are engaged these changes can travel the entire length of these processes.

Active electrical properties

By active electrical properties it is meant that the electrical potentials across the plasma membrane may be affected by the activation of voltage, ligand, or second messenger gated transmembrane ionic channels. The generation of action potentials is an example of electrical properties brought about by active, voltage-dependent means. Here the electric field across the membrane will act on the voltage sensors of transmembrane ionic channels (channel sites with dipole moment properties that will trigger conformational changes, often allosteric, that will change channel ionic conductance). In the specific case of action potentials voltage-gated channels, the inflow of sodium or calcium ions depolarizes the plasma membrane. In turn, opening voltage-gated potassium channels and the resulting current flow repolarizes the plasma membrane. Although the conductance of most voltage-gated channels are increased by membrane depolarization, the conductance of some channels is increased when the membrane is hyperpolarized.

Other examples of active electrical properties are those brought about by ligand-gated ionic conductances, where the binding of a neurotransmitter will gate ionic conductances allowing the generation of excitatory or inhibitory synaptic potentials. Yet another form of electrical activity is represented by intrinsic subthreshold oscillations, where the excitability of the cell is gated in such a fashion that the membrane potential is not uniform but rather in a state of continuous fluctuation, generating an oscillatory sinusoidal-like membrane profile—often with phase reset properties indicating chaotic, dynamic kinetics.

Figure 4: Evolution of the neuron. Parker (1919). B. Drawing of Purkinje cell (Purkinje 1837).C. Drawings of dissected motor neurons (Dieters 1865).

Neuronal integration

In an active neuron the superposition of passive and active electrical properties serve to allow the cell the possibility of summing the transmembrane potential either linearly or non-linearly and to reach depolarization levels sufficiently high to trigger action potentials. These can be conducted either along the length of the axon or dendritic tree, in an all-or-none continuous manner, in a saltatory fashion, or in a decremental mode.

Axonal spike

Neurons have but one axon, that is, a single process leaving the soma or a dendrite. However, axons branch either in the form of collaterals along the axonal length or at a terminal arborization known as the telodendrion (distal dendrite). These terminals are usually the site of presynaptic boutons that establish synaptic contacts with other neurons, muscles, or glands.

Because they originate from a single initial segment axons send very similar spike sequences to all their branches. However, spike failure at branch points or changes in conduction velocity, secondary to changes in axonal diameter after branching, can reset conduction patterns and conduction times. A good example is the isochrony of spike conduction time is found in the axons from the inferior olive that terminate in the cerebellum as what are called climbing fibers.

Dendritic spikes

In addition to action potentials in axons, dendrites may also generate regenerative events. For the most part these potentials decrement with distance but may reach the most distal dendritic branches. Such action potentials can be centripetal (towards the soma) or centrifugal (back propagating away from the soma) depending on the dendritic morphology and the distribution and density of voltage gated ionic channels over the dendritic arbor. [I would revise to note that back propagation entirely depends upon the presence of Na channels in the dendritic membrane, even in the apical dendrite of cortical pyramids.] Thus, dendrites with a few main branches such as cortical pyramidal cells will show back propagation because the scarcity of lateral branches allow sufficient longitudinal current to flow along the main dendrite and depolarize the distal segment sufficiently to reach threshold level. By contrast, a dendrite like that of a cerebellar Purkinje cell presents a cable structure where the branch number increases with distance. This presents the advancing longitudinal current with an ever growing capacitative load and an expanding set of parallel resistive paths. Both of these properties work to reduce the effectiveness of back propagation. Interestingly, activity in a distal dendritic branch is more likely to reach the soma since the path from dendritic branchletts to wider dendritic segments presents an easier path for longitudinal current flow—except at branch points were dendritic spikes may fail unless such sites are endowed with an increased density of voltage gated channels.

Figure 5: Variety of neurons in which various markers were used to distinguish individual neurons and parts of neurons. A. Hippocampal neuron transfected (micro-injected) with GABA(A) receptor (red) and co-stained for neurofilaments (green) (reproduced with permission). B. Triple stained neuron made by Wim Annaert. C. Three rabbit retinal ganglion cells injected with Alexa Fluor 568 hydrazide (top cell, red), Alexa Fluor 488 hydrazide (middle cell, green), or both the Alexa Fluor 488 and the Alexa Fluor 568 hydrazides (bottom cell, yellow at colocalization). (Image contributed by Michiel van Wyk, University of Queensland, Australia). D. Neuron transfected with doublecortin (DCX) DsRed and mannose phosphate receptor (MPR) GFP (reproduced with permission).

Intrinsic electrical properties and subthreshold oscillations

Beyond action potentials and synaptic transmission there is the question of the electrical activity generated autonomously by neurons. In most cases the autonomous intrinsic activity results in a modulation of the resting potential and so the overall state of the neuron in the sense not only of synaptic modulation or the modulation produced by peptidergic, hormonal and metabolic activity, but also of other parameters such as pH and free radical (NO and CO) activity modulation.

Beyond modulation, the parameter that is most relevant in defining intrinsic activity, other than resting potential, is the types and distributions of plasma membrane channels and second messenger activated modulation of channels.

Subthreshold oscillatory activity was originally discovered in the inferior olive and then found in many other central nervous system (CNS) neurons (Llinas 1988). In addition to the olive such neurons have been found in: entorhinal cortex layer 2 and 3, thalamus, cortical inhibitory interneurons, neostriatum, olfactory bulb, primary sensory ganglion, hippocampus pyramids in CA1 and CA3, dorsal column, globus pallidus, mesencephalic V, supraoptic nucleus, substancia nigra, locus coeruleus, subthalamus, pedunculo-pontine nucleus, layer V and VI pyramids, amygdala, and cingulus, among others. These oscillations are supported by persistent sodium currents and calcium current (Cav3 and Cav1 subunits), in conjunction with voltage and calcium gated potassium currents. The frequency of these oscillations can vary from high gamma band (80 Hz), Mu (25 Hz), beta (15 Hz), alpha (10 Hz), theta (6 Hz,) delta (3 Hz) and slow sleep oscillation bellow 1Hz.

These oscillations are viewed as supporting time-locked neuronal coherence and resonance. As such they are thought to be of significance in determining large functional brain states such as sleep, wakefulness, and dreaming and abnormal states such as epilepsy and thalamocortical dysrhythmia.

Non-uniformity in channels density

Finally, the electrical signature of neurons are defined by the passive integrative properties of the dendrites and soma and the non-linear electrical properties superimposed by the presence of voltage, ligand, second messenger, and metabotropic conductances supported by specialized ionic plasma membrane bound channels. These modulate excitability by their number, functional phenotype, and distribution over the dendritic, somatic and axonal neuronal segments.

While certain characteristic properties can be assign to given cellular phenotypes, the fact is that every neuron is unique both in its individually detailed shape and its connectivity. Perhaps it is the diversity of such parameters that allow the CNS to be as reliable as it actually is. It was Von Neuman who first realized the genesis of reliability from unreliable elements as one of the central character of “neuron-ness”.

Figure 6: Neurons marked and photographed in place. A. Pyramidal neurons of mouse cortex. (Picture: Tobias Bonhoeffer, MPI of Neurobiology). B. Three Golden Columns -(reproduced with permission). C. A pyramidal shaped neuron in the cortex and a second neuron in the underlying thalamus (?) (reproduced with permission). D. Neurons in Drosophilia, (Andreas Prokop). E-G. Retinal cells (Wallace B. Thoreson). H. Retroviral vector expressing GFP in adult mouse hippocampus. GFP expression in the dentate gyrus 4 weeks after virus injection express mature neuronal markers. Neuronal marker is NerN (red, Cy3). (Nature Vol 415, 28 Feb 2002, pg 1030-1034). I. Neurons within song control nucleus RA. (Timothy J. DeVoogd). J. Cortical neuron showing axon projection. (R. David Andrew).

A little history

Neurons evolved from primitive cells that are capable of sensitivity (irritability) and contractility (exercising force by changing shape) as drawn by Parker in 1919 (Fig. <ref>F1</ref>Aa) (also see Meech & Mackie 2007). The initial differentiation of these cells into true neurons or muscle cells precedes the further differentiation of nerve cells into sensory neurons (in communication relation to the external world) or motor neurons (in direct communication with muscles or glands) (Fig. <ref>F1</ref>Ab). The final stage occurs with the development of interneurons (establishing contacts between sensory and motor neurons) (Fig. <ref>F1</ref>Ac). The latter represent the vast majority of neurons in the brain and ganglia.

The first central nervous system neurons were described by J.E. Purkinje in the cerebellum and named by him gangliosen zellen (1837) (Fig. <ref>F1</ref>B). These cells were found to have a cell body (soma), and fine processes by simple macroscopic inspection of freshly hand-dissected motoneurons (Deiters, 1865) (Fig. <ref>F1</ref>C). However, the true variety and organization of these cells was not realized until the development of the silver impregnation technique by Camilo Golgi (1873) (see Figs. xxx). Golgi recognized that neurons were cells characterized by a cell body and a set of thin filamentous extensions. Golgi believed these extensions fused, making the central nervous system a continuous connectivity network. Concerning their general form, Golgi divided neurons into two main categories.

• Type I were large projection neurons, each with a long axon that projected to places distant from its site of origin on the cell body (see Figs.<ref>F2</ref>A,B, <ref>F5</ref>A, B, J).
• Type II neurons projected locally at short distances from their cell body (see Fig. <ref>F5</ref>I, C, <ref>F6</ref>C).

In contrast, Ramon y Cajal (1888) demonstrated that neurons were individual separate elements that communicated specifically with each other and viewed the nervous system as being organized with nerve conduction going from the dendrites to the soma and axon to the next cell in the network (Fig. <ref>F3</ref>A and others). This allowed him to establish the network direction of conduction and the possibility of understanding CNS function as an orderly unidirectional information processing system. It is now know that the unidirectionality of network conduction is due to the unidirectional nature of synaptic transmission (see Synapse). Wilhem His made an important contribution to the neuron theory in 1886 with his discovery that each nerve fiber stems from a single nerve cell.

After thoughts

Not yet firmly established and sometimes not even recognized, or actively shunned away from, is the cytoplasmic/molecular machinery that lurks under the membrane. This is mostly assumed to be the necessary scaffolding that keeps the electrical membrane properties in place. Very briefly, consider a thought experiment where we instantaneously remove the interior “bulk” content of the neurons in the brain leaving only the membrane bound proteins, enzymes, vesicles and intracellular fluid (a bit like the preparation known as extruded axoplasm in the squid giant axon (Baker et al 1962). How many steps would an individual take, under those conditions, before paralysis or coma would occur? Perhaps less than ten steps. Clearly, if the neuronal membrane were to disappear brain death would be “instantaneous”. So would paralysis if we were to “disappear” the contractile apparatus of muscle cells. In the latter case the issue is that the active sliding of muscle filament provides the effector system for force generation. That is, movement effectors are provided by cytoplasmic properties. If one asks the parallel question “Are neurons effector systems for anything?” the answer may be, yes, “subjectivity”. At this point three opposing thoughts are:

1. Subjectivity does not exist!
2. Subjectivity does not belong to the realm of sciences.
3. Subjectivity is what neurons do intrinsically and it becomes macroscopic when organized by connectivity.

In other words, we have a lot to do yet before we understand the nature of neurons.

References

Baker P.F, Hodgkin A.L. and Shaw T.I. Replacement of the axoplasm of giant nerve fibers with artificial solutions. J. Physiol. 164: 330-354, 1962

Cajal, R (1888) Estructura de los nerviosos de lo saves. Rev. Trimest. Histol.1:305-315. (Structure of neural centers in birds.)

Deiters, OFK (1865) Untersuchungen uber und Ruckermark des Menshen und der Saugerthiere, Braunschweig, Berlin. (Investigations of the Brain and Spinal Cord of Humans and Mammals.)

Gascoigne, L and McVean, A (1991) Water Movement Sensitive Cells in Leech CNS Philosophical Transactions: Biological Sciences, Vol. 332, No. 1264. pp. 261-270, Fig. 2(b)(i).

Golgi, C (1873) Untersuchungen ¸ber den feineren Bau des Centralen und peripherischen Nervensystems. Gazzetta Medica Italian, Lombardia, 6:244-246. (Investingations on the fine structure of the central and peripheral nervous systems.)

His, W (1886) Zur Geschichte der menschlichen R¸ckenmarkes und der Nervenwurzeln (1886). Abhandlungen der Kˆniglichen s‰chsischen Gesellschaft der Wissenschaften (1886), 1887. Mathematisch-physische Classe, Leipzig 13: 477-514..

Llinas, R. (1988) The intrinsic electrophysiological properties of mammalian neurons: A new insight into CNS function. Science 242, 1654-1664,

Meech, R and Mackie, G (2007) Evolution of excitability in lower metazoans in Invertebrate Neurobiology edited by North, G and Greenspan, Cold Spring Harbor Laboratory Press. Chapter 26.

Parker, G. H. (1919) The Elementary Nervous System. Lippincott, Philadelphia.

Purkinje, JE (1837) Bericht ¸ber die Versammlung deutscher Naturforscher und Aerzte in Prague Anat. Physiol. Ver. 3:177-180. (Report on the conference of German scientists and doctors in Prague.)

External links

Neurons in the retina

http://retina.umh.es/Webvision/index.html

History of Bioelectricity

http://www.cerebromente.org.br/n06/historia/bioelectr_i.htm

History of neuroscience

http://faculty.washington.edu/chudler/hist.html

http://neurophilosophy.wordpress.com/2006/08/29/the-discovery-of-the-neuron/

http://www.ibro.info/Pub_Main_Display.asp?Main_ID=11

See also

Brain, Dendritic Processing, Models of Neurons, Neuroanatomy, Synapse, Synaptic Plasticity

Sponsored by: Eugene M. Izhikevich, Editor-in-Chief of Scholarpedia, the peer-reviewed open-access encyclopedia

Reviewed by: Anonymous

Reviewed by: Dr. Nelson Spruston, HHMI Janelia Research Campus, Ashburn, VA, USA

Accepted on: 2008-08-21 19:00:41 GMT