P2X receptors in peripheral neurons

Abstract

P2X receptors are a family of ligand-gated ion channels, activated by extracellular ATP. The seven subunits cloned (P2X_1–7) can assemble to form homomeric and heteromeric receptors. Peripheral neurons of neural crest origin (e.g. those in dorsal root, trigeminal, sympathetic and enteric ganglia) and placodal origin (e.g. those in nodose and petrosal ganglia) express mRNAs for multiple P2X subunits. In this review, we summarize the molecular biological, electrophysiological and immunohistochemical evidence for P2X receptor subunits in sensory, sympathetic, parasympathetic, pelvic and myenteric neurons and adrenomedullary chromaffin cells. We consider the pharmacological properties of these native P2X receptors and their physiological roles. The responses of peripheral neurons to ATP show considerable heterogeneity between cells in the same ganglia, between ganglia and between species. Nevertheless, these responses can all be accounted for by the presence of P2X₂ and P2X₃ subunits, giving rise to varying proportions of homomeric and heteromeric receptors. While dorsal root ganglion neurons express predominantly P2X₃ and rat sympathetic neurons express mainly P2X₂ receptors, nodose and guinea-pig sympathetic neurons express mixed populations of P2X₂ and heteromeric P2X_2/3 receptors. P2X receptors are important for synaptic transmission in enteric ganglia, although their roles in sympathetic and parasympathetic ganglia are less clear. Their presence on sensory neurons is essential for some processes including detection of filling of the urinary bladder. The regulation of P2X receptor expression in development and in pathological conditions, along with the interactions between purinergic and other signalling systems, may reveal further physiological roles for P2X receptors in autonomic and sensory ganglia.

Introduction

Neurons of the peripheral nervous system have many roles including conducting sensory information from peripheral organs, relaying commands from the central nervous system to the periphery, and controlling visceral organs. Among peripheral neurons, sensory neurons of the dorsal root ganglia (DRG) share with neurons of the sympathetic, parasympathetic and enteric ganglia, along with adrenomedullary chromaffin cells a common embryological origin in the neural crest. In contrast, cranial sensory neurons are derived from the placodes. While these sensory and autonomic neurons exhibit some common properties, they also show very diverse phenotypes commensurate with their diverse physiological roles. In this article, we examine the considerable and ever increasing information on the P2X receptors present on neurons of sensory, sympathetic, parasympathetic and enteric ganglia, as well as adrenomedullary chromaffin cells.

As long ago as 1929, Drury and Szent-Gyorgyi had demonstrated that adenine compounds had potent extracellular actions (Drury and Szent-Györgyi, 1929), and in the 1950s, Holton demonstrated the release of ATP from nerve terminals (Holton, 1959). However, it was not until 1970 that Burnstock and his colleagues suggested that ATP or a related nucleotide might be a neurotransmitter released by non-adrenergic non-cholinergic neurons in the gut (Burnstock et al., 1970, Burnstock, 1972). Implicit in that proposal was the existence of specific purinergic receptors present in the post-junctional cell membrane.

Subsequent pharmacological studies, comparing agonist profiles in different tissues led to the division of these purinergic receptors first into P1 and P2 (Burnstock, 1978) and then subdivision of P2 into P2X and P2Y (Burnstock and Kennedy, 1985) subtypes. Other apparently distinct nucleotide receptors including P2_Z, P2_U and P2_T were subsequently identified before molecular cloning techniques brought about a rationalization of nucleotide receptor classification (see Abbracchio and Burnstock, 1994, Ralevic and Burnstock, 1998, North and Surprenant, 2000). However, the heteromeric assembly of subunits still complicates the characterization of native receptors. There are now two major P2 receptor families: P2Y receptors that are members of the G protein-coupled receptor superfamily and P2X receptors that are ligand-gated ion channels.

To date, seven mammalian P2X receptor subunits (P2X₁–P2X₇) have been identified by molecular cloning (see Table 1; Buell et al., 1996a). In addition, an apparently distinct member of this family had been identified in the chicken and provisionally named cP2X₈ (Bo et al., 2000). The P2X subunits all share the same general structure (Fig. 1), having intracellular N- and C-termini, two membrane-spanning domains and a large extracellular loop containing 10 conserved cysteine residues. While the intracellular N-termini are of similar length, there is considerable variability in the length of the C-terminus from 30 residues in P2X₆ to 240 amino acids in P2X₇ (for reviews, see Buell et al., 1996a, MacKenzie et al., 1999).

Like all other ligand-gated ion channels, functional P2X receptors are formed by the assembly of a number of subunits. These may be identical, giving rise to homo-multimeric receptors, or non-identical forming hetero-oligomers. The number of subunits present in a P2X receptor is thought to be either three (Nicke et al., 1998, Stoop et al., 1999) or four (Kim et al., 1997, Ding and Sachs, 2000). However, the rules governing assembly are still poorly understood. Co-immunoprecipitation studies using human embryonic kidney cells (HEK) indicate that while the P2X₇ subunit is unlikely to assemble with other P2X subunits, P2X₆ is unlikely to form homo-multimers. Others such as P2X₂ and P2X₅ will readily heteropolymerize with most members of the P2X family (Torres et al., 1999; see Table 2). Thus, from these seven subunits, even if one assumes that all the subunits in the receptor must be compatible, there could be 33 distinct trimeric or 55 tetrameric receptors. However, even if all these possibilities exist, whether they can be distinguished by pharmacological, biophysical or other techniques is another matter. It is also possible that different heteropolymerization may occur in other cell types where different auxiliary proteins or processing are present. An additional level of complexity is generated by the existence of spliced variants of many of the P2X receptor subunits, which may produce receptors with modified properties (Brändle et al., 1997, Townsend-Nicholson et al., 1999, Chen et al., 2000), thus potentially producing even more distinct receptors.

The seven homomeric P2X receptors, when studied in heterologous expression systems, exhibit different but overlapping properties in terms of agonist and antagonist sensitivities and rates of desensitization (see Table 3). Thus, P2X₁ and P2X₃ receptors are activated by α,β-methylene ATP (αβ-meATP), while P2X₂ is not. P2X₁ and P2X₃ receptors desensitize rapidly (τ≤1 s) while P2X₂ receptors desensitize slowly (τ>10 s). Although the detailed mechanisms involved in the desensitization of P2X receptors have yet to be elucidated, it is clear that the rate of desensitization can be modified under some conditions, for example by phosphorylation (Boue-Grabot et al., 2000) or changes in the cytoskeleton (Parker, 1998). Consequently, caution should be exercised in its use for receptor characterization.

An interesting property of some P2X receptors, which was first described for P2X₇, is the ability to enable molecules up to 900 Da to enter the cell (Surprenant et al., 1996). While this has been attributed to the dilation of the P2X₇ ion channel (Virginio et al., 1999), an alternative explanation is that activation of the P2X₇ receptor leads to the opening of the same, large-diameter pore, which is opened by maitotoxin (Schilling et al., 1999). A similar increase in membrane permeability has now been reported to occur as a result of activating some neuronal P2X receptors (Virginio et al., 1999).

The expression of some pairs of P2X receptor subunits can give rise to the formation of receptors with novel pharmacological properties. Thus, coexpression of P2X₂ and P2X₃ subunits produces a receptor that responds to αβ-meATP (property of P2X₃) with a slowly desensitizing response (property of P2X₂) (Lewis et al., 1995). Other functional heteromeric receptors so far identified are P2X_4/6 (Lê et al., 1998), P2X_1/5 (Torres et al., 1998) and P2X_2/6 (King et al., 2000).

As long ago as 1954, Feldberg and Sherwood demonstrated profound effects of intracerebroventricular injection of ATP, including ataxia and a tendency to sleep (Feldberg and Sherwood, 1954). Numerous subsequent studies have demonstrated biochemical and electrophysiological effects of purines on the central nervous system (CNS) including excitation of neurons in the chemoreceptor trigger zone (Borison et al., 1975) and the cerebral cortex (Phillis and Edstrom, 1976; for reviews, see Burnstock, 1977, Burnstock, 1997). Synaptic transmission in the brain, mediated by ATP, was first observed in the medial habenula (Edwards et al., 1992) and has now been detected in a number of other brain areas (Nieber et al., 1997, Pankratov et al., 1999).

The earliest report of an effect of ATP on autonomic ganglia was in 1948, when Feldberg and Hebb demonstrated that intra-arterial injection of ATP excited neurons in the cat superior cervical ganglion (SCG) (Feldberg and Hebb, 1948). Later work from de Groat's laboratory showed that in the cat vesical parasympathetic ganglia and SCG, purines inhibited synaptic transmission through P₁ receptors, but high doses of ATP depolarized and excited the post-ganglionic neurons (Theobald and de Groat, 1977, Theobald and de Groat, 1989).

The earliest intracellular recordings of an action of ATP on neurons were obtained in frog sympathetic ganglia. However, here, ATP produced a depolarization through a reduction in K⁺ conductance (Siggins et al., 1977, Akasu et al., 1983), which was probably mediated through P2Y receptors.

In 1983, Jahr and Jessell, using intracellular recordings from cultured neurons, demonstrated that ATP could excite DRG neurons and some neurons from the dorsal horn of the spinal cord (Jahr and Jessell, 1983). These responses were associated with an increase in membrane conductance, which we now know to be due to the activation of P2X receptors.

During early embryological development, the neural ectoderm folds to form the neural tube. Cells in the overlying ectoderm (the neural crest) then migrate within the ectoderm and into the mesoderm. The cells that follow this latter pathway differentiate and mature to become glial cells and neurons. Some become primary afferent neurons of the DRG, while others become the post-ganglionic neurons of the sympathetic and parasympathetic ganglia. A third group of cells go on to form the enteric nervous system. One group of potential sympathetic neurons become surrounded by developing adrenal cortical cells and develop into adrenomedullary chromaffin cells. The sensory neurons of cranial nerves, including those of nodose, petrosal and trigeminal ganglia, however, are derived partly or entirely from the neural placodes (Lindsay, 1996).

In view of their common embryological origins, it is perhaps not surprising that peripheral neurons and chromaffin cells possess some common properties, e.g. N-type calcium channels (Gross and Macdonald, 1987, Bossu et al., 1991, Zhu and Ikeda, 1993) and GABA_A receptors (Adams and Brown, 1975, Bormann and Clapham, 1985, White, 1990). However, in keeping with their diverse roles, there are also many differences between them, e.g. only a distinct subpopulation of primary sensory afferents express the heat sensing vanilloid receptor (VR₁; Guo et al., 1999).

In the last few years, there has been a dramatic increase in our understanding of P2X receptors at the molecular level, and the discovery of new pharmacological tools with which to study them. In addition, demonstration of the coexistence of multiple P2X receptor subtypes in single neurons (Thomas et al., 1998, Grubb and Evans, 1999, Zhong et al., 2000a) and interspecies differences (Khakh et al., 1995, Zhong et al., 2000a, Zhong et al., 2000b) suggest a reappraisal of our knowledge of P2X receptors in peripheral neurons could be valuable.