CHAPTER 1
Electrophysiology of Magnocellular Neurons In Vivo

Gareth LengNancy Sabatier

Centre for Integrative Physiology, University of Edinburgh, Edinburgh, UK

1.1 Introduction

Neuroendocrinology is the study of things that matter: stress and appetite, metabolism, body rhythms, growth, and all aspects of reproduction from the reproductive cycle, through sexual behavior, pregnancy and parturition, to lactation and maternal behavior-things that matter for our health and happiness. However, neuroendocrine systems are also influential model systems for neuroscience generally, because of their unique value as "windows on the brain." The products of neuroendocrine systems can be measured relatively easily, and their effects are, with wit and persistence, determinable. In consequence, the neuronal activity that gives rise to those products is interpretable to a degree that can only be envied by colleagues in most other areas of neuroscience. If we ask of any neuron in the brain, what does it really do, the answers are often frustratingly incomplete: even if we know how it responds to stimuli, what it makes and where it projects, we may still not know what it does that matters to the behaving organism. By contrast, for the magnocellular vasopressin and oxytocin neurons of the hypothalamus, we can know much of what they do even before we know how they do it. All of these neurons project to the posterior pituitary gland, and what they secrete from there is measurable in the blood, and has measurable consequences for important physiological functions.

Neuroendocrinology began as the study of the secretion of peptides from neurons into the blood, and evolved to be also the study of secretion of peptides within the brain. In both aspects, electrophysiological recordings from single neurons in vivo have been fundamental in defining the physiological significance of mechanisms that have been established by cellular and molecular studies in vitro (Figure 1.1).

Figure 1.1 The milk-ejection reflex. The reflex was uncovered by electrophysiological studies in vivo. In response to suckling, oxytocin cells discharge intermittently in brief synchronized bursts that evoke secretion of pulses of oxytocin, which induce abrupt episodes of milk ejection. Dendritically released oxytocin facilitates the bursting.

1.2 Opening the window on the brain

Single-cell recording has strengths and weaknesses-and in vivo studies have particular problems as well as certain unique advantages. Because in vivo electrophysiology requires an extensive and prolonged investment in skills and expertise, it is wise to understand the nature of the investment that is involved, as well as the potential returns and the likely limitations.

The window on the brain afforded by the magnocellular system was opened by Wakerley and Lincoln (1973), when they used a technique to enable the magnocellular neurosecretory neurons to be rigorously interrogated electrophysiologically. There had been earlier attempts to study these neurons electrophysiologically, but despite careful stereotactical control, these recordings came from a mixture of neuroendocrine and non-neuroendocrine cells. This heterogeneity subverted interpretation: the breakthrough came from the ability to identify individual neurons as neuroendocrine neurons immediately and unequivocally. That breakthrough came with the introduction of antidromic identification (Figure 1.2).

Figure 1.2 Antidromic identification. A stimulating electrode that is placed on the axon of any neuron may be used to trigger a spike that is propagated both orthodromically (green stars), toward the axonal endings, and antidromically (blue stars), toward the cell body. A recording electrode at the cell body will record the antidromic spike at a fixed latency following the stimulus-a latency that reflects the conduction velocity and the axonal length. In general, a stimulus pulse might evoke a spike that is generated by monosynaptic excitation, which would also arise at a nearly constant latency. Antidromic spikes can be distinguished from such orthodromic spikes by two additional tests: frequency following and the collision test. Antidromic spikes will be generated (i.e., will follow) each of a short train of stimulus pulses presented at a high frequency (50-100 Hz); these spikes will maintain a near-constant latency (there is a slight prolongation of latency with each successive pulse). A longer train of stimuli will result in fractionation of the antidromic spike-as the soma becomes refractory to antidromic stimulation, antidromic invasion is progressively delayed and may intermittently fail, while the smaller initial segment spike, which is normally hidden within the soma spike, will be preserved and become visible as a notch on the rising phase of the antidromic spike. However, antidromic spikes will not invade either the initial segment or the soma when they are extinguished by collision with a spontaneous, orthodromically propagated spike. This collision (red X) occurs when an antidromic stimulus pulse immediately follows a spontaneously generated spike-the descending spontaneous orthodromic spike meets the ascending antidromic spike along the axon, and both are extinguished by this collision.

Because magnocellular neurons project to the pituitary, a stimulating electrode placed on the neural stalk can be used to trigger action potentials (spikes) in their axons, and these spikes can be detected at the soma, after antidromic (i.e., backward) propagation, as spikes that follow each stimulus at a constant latency. The rat supraoptic nucleus contains only two types of neuron-magnocellular oxytocin neurons and magnocellular vasopressin neurons, so every neuron recorded from this region that can be antidromically stimulated can be identified as one of these two types. Once Wakerley and Lincoln began to apply this approach, they saw that these two cell types could be distinguished by other tests. These tests were refined by many others, leading to an extensive battery of tests that now allows us to talk of identified oxytocin neurons and identified vasopressin neurons.

From the ability to identify oxytocin and vasopressin neurons in vivo came the cornerstones of our understanding of the physiology of these systems. Wakerley and Lincoln (1973) defined the milk-ejection reflex, as it is known in all good text books (Figure 1.1). From their studies came the realization that, in response to suckling, oxytocin is not released continuously, but intermittently-in pulses that result from the brief intense synchronized bursts of action potentials generated by the oxytocin neurons (Figure 1.3). From Francois Moos and her colleagues (see Richard et al., 1997), we learned that this reflex depends upon central release of oxytocin, and could be facilitated by injecting tiny amounts of oxytocin introcerebroventricularly (i.c.v.)-the first (and still the best) example of an indispensable physiological role of central peptide release. From Summerlee and Lincoln (1981), we learned that the milk-ejection reflex in conscious rats is the same as in anaesthetized rats-and that during parturition, oxytocin is again released in intermittent pulses as a result of similar synchronized bursting (Summerlee, 1981). From Jon Wakerley, Richard Dyball, and Dominique Poulain came the recognition that in response to osmotic challenge, vasopressin neurons fire not continuously, but phasically (Figure 1.4). This phasic firing is asynchronous among the vasopressin neurons, so that it leads to continuously maintained secretion-its significance is not in producing a phasic pattern of secretion, but in optimizing the efficiency of stimulus-secretion coupling at the nerve terminals. From Dyball and others, we learnt that oxytocin neurons are just as responsive as vasopressin neurons to osmotic pressure increases, leading to recognition of oxytocin's role (in many species) in regulating natriuresis. These pioneers laid the cornerstones of our current understanding, inspiring a wealth of hypotheses that have been pursued ever since by subsequent workers. Some of the questions that they posed were particular to these systems: What makes magnocellular neurons osmosensitive? Why is oxytocin released in response to osmotic stimulation? Many others were questions of general significance: Why do cells fire in bursts, and what are the mechanisms that underlie bursting? What mechanisms allow bursts to be synchronized? Why is it important that hormones are released in pulses? How can a peptide change the pattern in which cells fire? How is peptide release in the brain regulated? The mechanistic understanding that flowed from pursuit of the answers to these and related questions has changed our understanding of the brain in profound ways. Most neurons in the brain release one or more peptide messengers in addition to classical neurotransmitters, and these have a myriad of autocrine, paracrine, neuromodulatory and neurohormonal actions.

Figure 1.3 Milk-ejection bursts. Recording from an oxytocin cell in a urethane-anesthetized, lactating rat. Typically, oxytocin cells fire slowly and continuously, but, once the pups are applied, then brief intense bursts start to appear. (A)...