Physical Address
304 North Cardinal St.
Dorchester Center, MA 02124
Neuroanatomical Substrates of Spinal Nociception
SUMMARY
The spinal dorsal horn receives input from a wide variety of primary afferent axons, including nociceptors , which respond to tissue-damaging stimuli from the skin, muscles, joints, and viscera. The patterns of termination of primary afferents within the spinal cord are related to axonal diameter and receptive field modality. Most nociceptive primary afferents have slowly conducting fine myelinated or unmyelinated axons , and they terminate mainly in the superficial part of the dorsal horn. Primary afferents release a variety of chemical mediators, but all use glutamate as their principal neurotransmitter, and on entering the dorsal horn they form excitatory synapses with neurons located within it. These neurons include projection cells with axons that convey information to various parts of the brain and interneurons with axons that remain in the spinal cord and contribute to local neuronal circuits. Interneurons make up the great majority of the neuronal population in the dorsal horn and can be divided into two main functional classes: inhibitory interneurons, which use γ-aminobutyric acid (GABA) and/or glycine as a transmitter, and excitatory interneurons, which are glutamatergic. The organization of interneurons in the dorsal horn is very complex, and we still know little about the neuronal circuits in which they take part. Intrathecal administration of GABA A or glycine receptor antagonists can cause allodynia, in which brushing of the skin becomes an aversive stimulus. This suggests that one function of inhibitory interneurons is to suppress activity evoked by tactile afferents so that it is not perceived as painful.
The dorsal horn also receives input from descending axons that originate in various parts of the brain. An important group in terms of pain mechanisms consists of axons that release serotonin or norepinephrine. These axons are thought to play a role in controlling transmission of nociceptive information through the dorsal horn and contributing to stimulation-produced analgesia.
Introduction
The dorsal horn of the spinal cord is the major receiving zone for primary afferent axons that transmit information from sensory receptors in the skin, viscera, joints, and muscles of the trunk and limbs to the central nervous system. Nociceptive primary afferent axons (i.e., those that respond to tissue-damaging stimuli) terminate almost exclusively in the dorsal horn, which is therefore the site of the first synapse in ascending pathways conveying the sensory information that underlies conscious perception of pain. In addition, it contains neuronal circuits involved in generating local reflexes.
In the Gate Control Theory of pain, proposed that inhibitory interneurons in the superficial part of the dorsal horn play a crucial role in controlling incoming sensory information before it is transmitted to the brain. This theory aroused a great deal of interest in organization of the dorsal horn. However, despite intensive study since then, our knowledge of the neuronal circuitry of the region remains limited. The dorsal horn contains four neuronal components: (1) central terminals of primary afferent axons, which arborize in different areas, depending on their diameter and the type of sensory stimulus that they respond to; (2) interneurons, with axons that remain in the spinal cord, either terminating locally or extending into other spinal segments; (3) projection neurons, with axons that pass rostrally in white matter to reach various parts of the brain; and (4) descending axons that pass caudally from several brain regions and play an important role in modulating the transmission of nociceptive information. In this chapter we review the anatomical organization of the mammalian dorsal horn, with particular emphasis on primary afferents and interneurons. Certain features of projection neurons are covered here, but they are described in more detail in Chapter 12 . Descending modulatory systems are dealt with in Chapter 8 , but here we discuss possible targets of the monoamine neurotransmitters released by axons projecting from the brain stem. Because many of the anatomical studies of the dorsal horn have been carried out on cats or rodents, our account is based on these species.
The Laminae of Rexed
divided the dorsal horn of the cat spinal cord into six parallel laminae based on differences in the size and packing density of neurons (cytoarchitectonics). This scheme has since been extended to other species, including human, monkey, and rat ( Fig. 5-1 ), and serves as a useful basis for describing its anatomical organization. Lamina II is often subdivided into two parts: inner (IIi) and outer (IIo). Laminae I and II, which are referred to as the superficial dorsal horn, constitute the main target for nociceptive primary afferents (see later). We concentrate our account on this region, partly because of its obvious importance in pain mechanisms and partly because more is known about its neuronal organization. However, the deeper laminae (III–VI) also have an important role in pain: some nociceptive primary afferents terminate in this region, and many neurons in these laminae (including some projection cells) are activated by noxious stimulation. In addition, low-threshold afferents that terminate in laminae IIi–V are at least partially responsible for the tactile allodynia (pain felt in response to touch) that occurs in certain pathological pain states ( ).
Lamina I, also known as the marginal layer, forms a thin sheet covering the dorsal aspect of the dorsal horn and contains both projection neurons and interneurons. Although this lamina contains the highest density of projection neurons in the dorsal horn, they are thought to make up only ≈5% of its neuronal population, with the remainder being interneurons ( ). Most of the cells have dendrites that remain within the lamina. Lamina I neurons vary considerably in size and shape, with projection cells being larger than interneurons ( ). A few particularly large projection neurons, known as marginal cells of Waldeyer, can be recognized. Lamina II is also known as the substantia gelatinosa because the lack of myelinated fibers gives it a translucent appearance. Virtually all the neurons in this lamina are interneurons, and they are densely packed in its outer part. Lamina III also contains a high density of neurons. Most are interneurons and are generally somewhat larger than those of lamina II, but scattered large projection neurons are also present. Although Rexed’s scheme was based on cytoarchitectonic criteria, the border between laminae II and III can be identified more easily by the absence of myelinated axons in lamina IIi and their presence in lamina III. This can be seen with myelin stains or dark-field microscopy of unstained sections. It should be noted that the correlation between the substantia gelatinosa and Rexed’s lamina II, originally determined in cats, may differ in rodents ( ) because lamina IIi receives abundant input from some large myelinated low-threshold mechanoreceptive afferent fibers in rodents but not in cats ( , , ). Laminae IV–VI are more heterogeneous, with neurons of various size, some of which are projection cells. The borders between these laminae are difficult to determine with certainty.
Primary Afferent Fibers
Primary sensory neurons provide constant feedback on the external environment, as well as the ongoing state of the body. The somata of those that innervate the limbs and trunk are located in sensory ganglia associated with spinal nerves (dorsal root ganglia). Their axons bifurcate within the ganglion and give rise to a peripheral branch that innervates various tissues and a central branch that travels through a dorsal root to enter the spinal cord, where it forms synapses with second-order neurons. The peripheral targets of these fibers provide a convenient means for classification. Fibers innervating skin are described as cutaneous sensory neurons. Likewise, those innervating abdominal or pelvic viscera are termed visceral afferents. Within these populations, fibers can respond to various sensory modalities, including mechanical, thermal, and chemical stimuli. Modality-specific groups are further divided according to the intensity of their adequate peripheral stimuli. Those that respond to gentle mechanical force or innocuous thermal stimuli are low-threshold mechanoreceptors or innocuous cooling or warming afferents. Fibers responding only to stimulus intensities considered tissue threatening or potentially tissue damaging are termed nociceptors.
As a group, primary sensory neurons exhibit a rich diversity in morphological and functional properties, including somatic membrane properties, laminar location of central projections, neurochemical content, and response properties of the central networks that they activate ( , ). The most common means of classifying primary sensory neurons is based on the conduction velocity of their peripheral axons, which is directly related to axon diameter and whether the axon is myelinated. From the distribution of these peripheral conduction velocities, primary sensory neurons are routinely divided into different groups: Aα/β, Aδ, and C.
The Aα/β group consists of large myelinated axons with the fastest peripheral conduction velocity, the Aδ group contains smaller fibers that are thinly myelinated and conduct at an intermediate velocity, and the C group consists of the smallest, unmyelinated, and most slowly conducting fibers. Within each group there is a wide range of functional types of primary afferents, as defined by sensory modality. Most sensory neurons with fibers conducting in the Aα/β range respond to innocuous mechanical stimuli, do not encode noxious stimulus intensities, and are classified as low-threshold mechanoreceptors. Some of these fibers, however, respond to relatively innocuous mechanical stimuli but also encode stimulus intensities in the noxious range and in some cases respond to noxious heating of the skin. This trend reverses with decreasing conduction velocity, with a majority of Aδ fibers and most C fibers being classified as nociceptors. The relative number of functional types in specific conduction velocity groups varies between species and the areas of the body that the fibers innervate. However, it is important to point out that both nociceptors and non-nociceptors exist in all three conduction velocity groups ( Fig. 5-2 ).
Vesicular Glutamate Transporters
All primary afferents are thought to use glutamate as their principal neurotransmitter since the excitatory post-synaptic currents produced by these afferents can be blocked with antagonists of the α-amino-3-hydroxy-5-methyl-4-isoxazoleproprionate (AMPA)-type glutamate receptor ( ), glutamate is enriched in their central terminals ( ), and these are associated with post-synaptic AMPA receptors ( ). Until recently it was difficult to identify glutamatergic axons, but this changed with the discovery of the vesicular glutamate transporters (VGLUTs). Two of these transporters, VGLUT1 and VGLUT2, are widely distributed in the spinal gray matter ( , ), whereas the third (VGLUT3) is present at much lower levels ( ). Within the dorsal horn, VGLUT1 is largely restricted to the deeper part (laminae IIi–VI) and is apparently present in the central terminals of all low-threshold mechanoreceptive myelinated primary afferents, as well as in some descending glutamatergic axons ( ). VGLUT2-immunoreactive axons are found throughout the gray matter, and most belong to local interneurons (see later). The central terminals of many myelinated nociceptors in lamina I also show strong VGLUT2 immunoreactivity. However, although most C fibers appear to express VGLUT2 ( ), it is present at very low or undetectable levels in their central terminals ( ). For example, dorsal rhizotomy does not lead to detectable loss of VGLUT2 in the dorsal horn, but there is a substantial reduction in VGLUT1 ( , ). recently identified a population of low-threshold mechanoreceptive unmyelinated primary afferents that express VGLUT3 and project mainly to the innermost part of lamina II and the dorsal part of lamina III.
Physiological Properties
Early studies of primary sensory neurons revealed heterogeneity in the shapes of somatic action potentials ( , ), and subsequent studies have shown strong correlation between these shapes and receptor function (e.g., , ). Myelinated fibers that respond to innocuous mechanical stimulation of the skin have narrow somatic action potentials without breaks in the rising or falling phase. Unmyelinated and myelinated nociceptive fibers have broad somatic action potentials that most often have a distinct inflection on the falling phase. This correlation is very consistent for myelinated fibers. However, all unmyelinated fibers have broad inflected somatic action potentials regardless of their peripheral response properties.
Spinal Projections of Primary Sensory Neurons
The central projections of primary sensory neurons have been visualized with various labeling methods, including the Golgi technique, detection of degenerating axon terminals, bulk-labeling techniques in which tracer substances are administered to a nerve or peripheral tissue and transported by many primary afferents, and intracellular staining of individual identified fibers. In his landmark studies, used the Golgi technique and suggested that fine primary afferent fibers project to the superficial part of the dorsal horn. As newer bulk-labeling techniques were introduced, the organization of central projections of different fiber types was further refined (e.g., , ). However, the uncertainties inherent in bulk-labeling techniques left many questions unanswered. The use of intracellular labeling has clarified the specific central targets of different afferent fiber types (e.g., , ). We will therefore focus on the findings of studies using these intracellular staining techniques.
Cutaneous Sensory Neurons
Myelinated Low - Threshold Mechanoreceptors
Low-threshold mechanoreceptive fibers enter the spinal cord and bifurcate into main ascending and descending branches that travel in the dorsal columns and migrate medially as they move away from their point of entry. Collateral fibers arise from these main branches, turn ventrally, and pass through the dorsal horn before terminating in dense arborizations that lie within a region extending from lamina IIi to lamina V. The morphological characteristics of these projections vary with the specific fiber type and the target tissue innervated, their relative position within the dorsal horn, and the distance from the point of entry into the spinal cord ( , , , ). In all cases the most superficial and dense central projections lie near the point of entry. Away from the entry zone they become more diffuse and occupy more ventral and medial positions ( ). Another consistent feature is that those innervating hair follicles terminate more superficially than do those innervating slowly adapting receptors. Among fibers innervating hair follicles, those conducting in the Aδ range (D-hair afferents) occupy the most superficial position of all low-threshold fibers and project extensively into lamina IIi ( , , ).
Myelinated Nociceptive Afferent Fibers
Myelinated nociceptors, which are thought to signal fast pricking or sharp pain, were first identified in the pioneering studies of and have since been studied extensively by other groups (e.g., , , ). They span a very large range of conduction velocities, from the slowest in the Aδ spectrum to well into the Aβ range ( , ). They respond to different stimulus modalities (e.g., mechanical and thermal) and have threshold stimulus intensities ranging from innocuous to noxious (e.g., , ). Recent studies also show that they exhibit different central morphologies and have a range of neurochemical phenotypes ( , , , , , ).
Most early studies examining individual myelinated nociceptors used extracellular recording techniques, either from peripheral nerves using microelectrodes ( ) or from nerve strands draped over metal electrodes ( ). With these techniques investigators are limited to using peripheral response properties, such as mechanical threshold, to identify myelinated nociceptors. Although it has long been known that some putative myelinated nociceptors could be activated by non-noxious moderate pressure ( ), threshold criteria were commonly used to avoid ambiguity, and thus neurons that code for both non-noxious and noxious mechanical stimuli have largely been overlooked.
The recent development of procedures that allow intracellular recordings from the cell soma combined with labeling of the central projections of the recorded fiber has increased the number of criteria that can be used to identify nociceptive sensory neurons ( , ). For example, myelinated nociceptors have broad inflected somatic action potentials that can easily be distinguished from those of low-threshold mechanoreceptors ( , ). Although relatively little information is available on the neurochemical properties of myelinated nociceptors, it is known that some contain neuropeptides, such as substance P and calcitonin gene–related peptide (CGRP), and express the high-affinity neurotrophin receptor TrkA, thus being responsive to nerve growth factor (NGF) ( ). Others contain TrkC and are sensitive to neurotrophic factor 3 (NT3) ( ). In addition, these fibers often possess acid-sensing ion channel 3 (ASIC3) and the transient receptor potential vanilloid type 2 (TRPV2) receptor ( , ).
The central projections of myelinated nociceptors were first described by , who focused on fibers conducting in the Aδ range in cats. They found that on entry into the spinal cord, the main branches were located laterally in the dorsal column, often in or near Lissauer’s tract. Terminal arbors from these afferents were centered primarily on laminae I and IIo, with some passing ventrally to terminate in lamina V. However, some of the faster-conducting fibers had collateral branches that penetrated deeply into the dorsal horn and then recurved dorsally and projected into the ventral part of lamina IV. The findings of this landmark study led to the widespread belief that myelinated nociceptors project only to laminae I and V.
More recently, Woodbury and Koerber used an ex vivo preparation consisting of isolated spinal cord and attached innervated skin to examine the projections of these and other afferent types in neonatal and adult mice ( ). Two distinct morphological types of myelinated nociceptor were observed. The first closely resembled the thinly myelinated nociceptors described in the cat by . On entry into the spinal cord, their axons bifurcated to give rise to ascending and descending branches that extended over several segments. Some of these afferents gave rise to axons that ascended in the dorsal columns. However, most maintained a lateral position in the vicinity of Lissauer’s tract. The second type had main branches that ascended and descended in the dorsal columns and gave rise to numerous collaterals that penetrated ventrally through the depth of the dorsal horn before recurving dorsally, as seen with many low-threshold myelinated mechanoreceptive afferents. However, in marked contrast to the latter, the arbors of this group extended through the full depth of the dorsal horn, including laminae IIo and I ( Fig. 5-3 ). Once in lamina I, they typically turned to run along the rostrocaudal axis while continuing to arborize. In general, the arbors of these fibers were somewhat more diffuse than those of low-threshold mechanoreceptors with similar peripheral conduction velocities. Interestingly, those with central projections focused in laminae I and IIo had higher mechanical thresholds than did those with projections spanning the dorsal horn. Some fibers of each type also responded to noxious heating of the skin ( , ).
You're Reading a Preview
Become a Clinical Tree membership for Full access and enjoy Unlimited articles
If you are a member. Log in here