Neuronal release of ATP was first demonstrated from the peripheral terminals of primary afferent fibres (Holton, 1959). Since then, co-storage and release of ATP from the terminals of both central and peripheral neurons has been widely described, and its role as a neurotransmitter is well established (see Burnstock, 1972, 1976, 2003). When released from nerve terminals, ATP can produce both rapid effects through the activation of a family of ligand gated ion channels (the P2X receptors), and slower and longer-lasting actions, which are usually mediated via the G protein–coupled P2Y receptors (Ralevic and Burnstock, 1998). To date, seven P2X receptor sub-units have been identified, which can assemble to form either homomeric or heteromeric receptors (North, 2002). In the periphery, the activation of P2X receptors is important for neuromuscular transmission in the vas deferens (Mulryan et al., 2000) and for the activation of primary afferent fibres in the urinary bladder (Cockayne et al., 2000; Vlaskovska et al., 2001) and carotid body (Prasad et al., 2001; Rong et al., 2003). P2X receptors are also found on most autonomic ganglion neurons (see Dunn et al., 2001). Since ATP is released from preganglionic nerve terminals (Vizi et al., 1997), these receptors may play a role in ganglionic neurotransmission.
In rat and mouse sympathetic ganglia, immunohistochemical and molecular biology suggest the presence of a variety of P2X subunits. However, most neurons respond to ATP, but not to α,β-meATP, which in combination with other pharmacological data indicates the presence of homomeric P2X2 receptors (see Dunn et al., 2001).
Nevertheless, a few neurons in these ganglia do respond to α,β-meATP (Khakh et al., 1995; Schadlich et al., 2001; Calvert and Evans, 2004). This agonist is considered to be selective for P2X receptors containing P2X1 or P2X3 subunits (North, 2002), although some activation of receptors containing P2X5 or P2X6 subunits has been reported (Wildman et al., 2002; Jones et al, 2004). Studies using P2X1 knockout mice have led to the suggestions that at least some α,β-meATP-evoked responses in mouse superior cervical ganglion neurones are mediated by receptors containing the P2X1 subunit (Calvert and Evans, 2004). However, autonomic ganglion neurons share the same embryological origins as dorsal root ganglion neurones, which express predominantly P2X3 homomeric or P2X2/3 heteromeric receptors. It is therefore possible that some α,β-meATP responses in sympathetic neurons might be mediated by P2X3-containing receptors.
There are many examples of changes in receptor expression during development due to the expression of different receptor subunits, for example, the γ to ϵ switch in nicotinic acetylcholine receptors at the neuromuscular junction (Mishina et al., 1986) and changes in GABA and NMDA receptors in the cerebellum (Gutierrez et al., 1997; Cathala et al., 2000). Changes in expression of P2 receptors also occur during embryonic and postnatal development (see Burnstock, 2001). For example, P2X receptors are expressed transiently in developing rat and chick skeletal muscle (Wells et al., 1995; Meyer et al., 1999; Ryten et al., 2001), and transient expression of the P2X3 subunit has been observed in the brain and spinal cord during rat (Cheung and Burnstock, 2002) and mouse (Boldogkoi et al., 2002) embryogenesis. Dynamic expression of P2Y receptor subtypes has also been described in the embryonic rat brain (Cheung et al., 2003).
In this study, we have used whole cell patch clamp recording and immunohistochemistry to investigate the expression of the P2X3 receptor subunit and the formation of heteromeric P2X2/3 receptors in rat superior cervical ganglion neurons during late embryonic and early post-natal development.
Responses of P1 and P17 Superior Cervical Ganglion Neurons
We initially compared the responsiveness of superior cervical ganglion neurons from early (P1) and late (P17) post-natal rats to ATP and α,β-meATP. All neurons tested responded to the nicotinic agonist DMPP. All superior cervical ganglion neurons from P17 animals responded to 100 μM ATP with a sustained inward current, with current densities ranging from 3 to 19 pA/pF (Fig. 1A,B). In keeping with previous studies on adult neurons (Khakh et al., 1995; Calvert and Evans, 2004), 100 μM α,β-meATP failed to evoke any response in most neurons from P17 animals. However, in one culture, two neurons did give significant responses to α,β-meATP (see Fig. 2Bi). In superior cervical ganglion neurons from P1 animals, the responsiveness to ATP was significantly greater than that in P17 animals, and the majority of neurons also responded to α,β-meATP with an inward current (Fig. 1A,B)
In P17 ganglia, most neurons failed to show any significant staining for P2X3. However a small sub-population of neurons did exhibit distinct immunoreactivity (Fig. 2Ai). In ganglia from P1 animals, many neurons were immunopositive for P2X3, and the strongest staining was considerably more intense than that seen in P17 ganglia (Fig. 2Aii). This pattern of staining was consistent with the responses to α,β-meATP of neurons from P17 and P1 superior cervical ganglion (Fig. 2B). In contrast, neurons from both P1 (Fig. 2Ci) and P17 (Fig. 2Cii) superior cervical ganglia exhibited quite uniform immunoreactivity for the P2X2 subunit.
Temporal Change in Agonist Responses
Having established that in contrast to adult neurons, the majority of neurons in P1 superior cervical ganglion respond to α,β-meATP, we investigated the time dependence of the change in agonist responsiveness, by looking at neurons from E18, P1, P7, and P17 animals. Neurons from E18 embryos all responded to 100 μM α,β-meATP with small inward currents, while more robust currents were produced by 100 μM ATP and 10 μM DMPP (Fig. 3). Response to all three agonists was maximal between birth and seven days. Following this, the current density to the nicotinic agonist DMPP was maintained, but there was a marked reduction in the response to ATP and few neurons responded to α,β-meATP (Fig. 3).
To further investigate the nature of the receptors responsible for the response to α,β-meATP in P1 sympathetic ganglia, we carried out pharmacological characterization of this receptor. The response to α,β-meATP was concentration dependent. Fitting the data with the Hill equation yielded an EC50 value of 29.9 ± 5 μM, with a Hill coefficient of 1.1. (Fig. 4A).
Trinitrophenyl-ATP is a subtype selective antagonist with nanomolar affinity for P2X receptors containing the P2X1 and P2X3 subunits (Virginio et al., 1998). TNP-ATP produced a reversible concentration-dependent antagonism of the response to α,β-meATP. Fitting the Hill equation to this data gave an IC50 of 13 ± 8 nM, with a Hill coefficient close to unity (Fig. 4B).
A characteristic of P2X receptors involving the P2X2 sub-unit is that they exhibit positive allosteric modulation by Zn2+ and H+ ions (see North, 2002). We, therefore, investigated the effects of these two ions on response of P1 superior cervical ganglion neurons to α,β-meATP. Lowering the pH from 7.4 to 6.8 produced a dramatic increase in the response to 20 μM α,β-meATP (Fig. 4C). However co-application of 10 μM Zn2+ produced no significant change in the response to α,β-meATP. In contrast, this concentration of Zn2+ more than doubled the response to 20 μM ATP in these neurons (Fig. 4C). Adult nodose ganglion neurons respond to α,β-meATP with a sustained inward current due to the presence of heteromeric P2X2/3 receptors (Lewis et al., 1995). We therefore investigated the effect of Zn2+ on responses to α,β-meATP on nodose ganglion neurons taken from newborn rats. On these neurons, Zn2+ produced a small but significant increase in the amplitude of the α,β-meATP response (Fig. 4C).
The main finding of this study is that there is a marked change in the expression of α,β-meATP-sensitive P2X2/3 receptors in sympathetic neurons of the rat superior cervical ganglion. The levels of this receptor peak soon after birth, then decline to very low levels by the time animals are about 17 days old. The expression of P2X receptors may be altered by dissociation and cell culture (Smith et al., 2001), possibly as a result of ATP release due to metabolic stress, ischemia, or trauma (see Volonte et al., 2003). Although our results show agreement between functional experiments and immunohistochemistry, we cannot rule out the possibility of changes resulting from the use of cell culture.
P2X receptors responding to α,β-meATP are believed to require the presence of either the P2X1 or P2X3 (North, 2002), although there is some evidence that P2X5 or P2X6 receptors may also respond to this agonist (Wildman et al., 2002; Jones et al., 2004). Much of our data suggests that the α,β-meATP-sensitive receptor expressed in sympathetic neurons from newborn rats is the heteromeric P2X2/3 receptor. Firstly, we observed considerable levels of P2X3 immunoreactivity in ganglia from P1 animals, which was virtually absent in ganglia from animals more than 17 days old. The sustained nature of the responses would argue against the involvement of homomeric P2X1 or P2X3 receptors, which give rapidly desensitizing responses (North, 2002). The EC50 value we obtained for α,β-meATP (30 μM) is similar to the value of 39 μM reported for nodose ganglion neurons (Dunn et al., 2000), but slightly greater than the value of 9 μM reported for heteromeric P2X2/3 receptors expressed in Xenopus oocytes (Liu et al., 2001). The sensitivity of this receptor to the antagonist TNP-ATP, with an IC50 of 13 nM, is quite similar to the values of 7 and 11 nM reported for recombinant P2X2/3 receptors (Virginio et al., 1998; Liu et al., 2001) and 21 nM for receptors in the rat nodose ganglion (Dunn et al., 2000). The potency of this antagonist is, however, considerably less than the low nanomolar values reported for the homomeric P2X3 receptor (see North, 2002). The potentiation of α,β-meATP responses by low pH is also in keeping with the properties of the heteromeric P2X2/3 receptor (Liu et al., 2001), and contrasts with the negative allosteric action of protons at the homomeric P2X3 receptor (North, 2002). Responses at the P2X2/3 receptor are also potentiated by Zn2+, although this effect is less pronounced than at the homomeric P2X2 receptor (Liu et al., 2001). In our experiments, Zn2+ enhanced responses to ATP in P1 superior cervical ganglion neurons, and also increased responses to α,β-meATP in nodose ganglion neurons from neonatal rats. However, we failed to observe potentiation of α,β-meATP responses in P1 superior cervical ganglion neurons. The reason for this is at present unclear, but might indicate the involvement of other P2X subunits or spliced variants. Studies using P2X1 knockout mice have indicated that a small percentage of superior cervical ganglion neurons respond to α,β-meATP through activation of P2X1 receptors (Calvert and Evans, 2004). Although we cannot exclude involvement of P2X1 subunits in a heteromeric receptor, the kinetic and pharmacological properties of the response we observed do not match those of the homomeric P2X1 receptor.
ATP is co-released with acetylcholine from pre-ganglionic nerve terminals (Vizi et al., 1997), and may thus play a role in synaptic transmission. This notion is supported by observation of synaptic responses, which are resistant to nicotinic receptor antagonists in some ganglia (Seabrook et al., 1990; Callister et al., 1997). P2X receptors are also present on the terminals of postganglionic sympathetic neurons, where they can modulate the release of noradrenaline (Sperlagh, et al., 2000; Queiroz et al., 2003). In the central nervous system, many P2X3-containing receptors are localized to presynaptic terminals. Thus, the loss of the P2X3 subunit in P17 SCG neurons may reflect the targeting of these subunits to the nerve terminal.
P2X2/3 receptors exhibit a higher affinity for ATP than the homomeric P2X2 receptors present on adult SCG neurons. This is likely to account for the high responsiveness to ATP, which we observed in embryonic and P1 ganglion neurons. Interestingly, this change in P2X receptor expression occurs at a time when synaptogenesis is taking place in the superior cervical ganglion (Smolen and Raisman, 1980; Mills and Smith, 1983), which might indicate a role for purinergic receptors in this process.
In conclusion, we have shown that sympathetic neurons of the rat superior cervical ganglion exhibit larger responses to ATP and α,β-meATP at birth and during the early post-natal period. This appears to be due, at least in part, to the expression of the P2X3 subunit, giving rise to the presence of heteromeric P2X2/3 receptors. Sensitivity to purinergic agonists then declines. It is tempting to speculate that the role of these receptors may be in some way related to synapse formation, which occurs during the early post-natal period.
Superior cervical ganglion neurons were cultured from E18, P1, P7, and P17 rats. Post-natal rats and pregnant females were killed by inhalation of a rising concentration of CO2 and death was confirmed by cardiac haemorrhage. Embryos were removed from pregnant females and placed in Leibovitz L-15 medium (Life Technologies, Paisley, UK). Neonatal animals were killed by cervical dislocation followed by decapitation. Superior cervical ganglia were rapidly dissected out, and placed in L-15 medium. The ganglia were then desheathed, cut, and incubated in 4 ml Ca2+- and Mg2+-free Hanks' balanced salt solution (HBSS; Life Technologies, Bethesda, MD) with 10 mM Hepes buffer (pH 7.4) containing 1.5 mg ml-1 collagenase (Class II, Worthington Biochemical Corporation, UK) and 6 mg ml-1 bovine serum albumin (Sigma Chemical Co., Poole, UK) at 37°C for 45 min. The ganglia were then incubated in 4 ml HBSS containing 1 mg ml-1 trypsin (Sigma) at 37°C for 15 min. The solution was replaced with 1 ml growth medium comprising L-15 medium supplemented with 10% bovine serum, 50 ng ml-1 nerve growth factor, 2 mg ml-1 NaHCO3, 5.5 mg ml-1 glucose, 200 i.u. ml-1 penicillin, and 200 μg ml-1 streptomycin. The ganglia were dissociated into single neurons by gentle trituration. The cell suspension was diluted to 8 ml, and centrifuged at 160g for 5 min. The pellet was resuspended in 0.8 ml growth medium and plated onto 35-mm Petri dishes coated with 10 μg ml-1 laminin (Sigma). Cells were maintained at 37°C in a humidified atmosphere containing 5% CO2, and used on the following day.
Whole-cell voltage-clamp recording was carried out at room temperature using an Axopatch 200B amplifier (Axon Instruments, Foster City, CA). Membrane potential was held at −60 mV. External solution contained (mM): NaCl 154, KCl 4.7, MgCl2 1.2, CaCl2 2.5, Hepes 10, and glucose 5.6; the pH was adjusted to 7.4 using NaOH. Recording electrodes (resistance 2–4 MΩ) were filled with internal solution that contained (mM): KCl 120, Hepes 10 and tripotassium citrate 10, EGTA 0.1; the pH was adjusted to 7.2 using KOH. In some experiments, a similar solution was used in which K+ was replaced by Cs+. No difference in response was observed between the two internal solutions. Data were acquired using pCLAMP software (Axon Instruments). Signals were filtered at 2 kHz (−3 dB frequency, Bessel filter, 80 dB per decade). Cells were confirmed as neurones by the presence of a fast rapidly inactivating inward current upon depolarization to 0 mV.
Drugs were applied rapidly through a manifold comprising 6 capillaries made of fused silica coated with polyimide, with a 250-μm internal diameter (SGE, Milton Keynes, UK), connected to a single outlet made of the same tubing, which was placed about 200 μm from the cell. Solutions were delivered by gravity flow from independent reservoirs. One barrel was used to apply drug-free solution to enable rapid termination of drug applications. Solution exchange measured by changes in open tip current was complete in 200 msec; however, complete exchange of solution around an intact cell was considerably slower (1 sec). Nevertheless, solution exchange using this system is fast enough to observe rapidly desensitising responses in dorsal root ganglion neurons (see Dunn et al., 2000). Agonists were separately applied for 1–5 sec at 2-min intervals, a time sufficient for responses to be reproducible. Antagonists were present for 2 min before and during the reapplication of agonists.
Rats were killed as described above and the superior cervical ganglia were dissected out. For immunohistochemistry on sections, the ganglia were fixed in 4% formaldehyde (in 0.1 M phosphate buffer) containing 0.03% picric acid (pH 7.4) for 120 min, then they were rapidly frozen by immersion in isopentane at −70°C for 2 min, cut into 10-μm sections using a cryostat, thaw-mounted on gelatin-coated poly-L-lysine-coated slides, and air-dried at room temperature. For immunohistochemstry on cultured neurons, ganglia were dissociated as above, plated in chamber slides and maintained in culture for 24 hr. They were fixed in 4% formaldehyde (in 0.1 M phosphate buffer) containing 0.03% picric acid (pH 7.4) for 120 min, then washed with distilled water three times.
Antibodies against rat P2X2 and P2X3 subunits (see Oglesby et al., 1999) were used in this study with an indirect three-layer immunofluorescent method. Primary antibody to P2X subunits were raised in rabbits, detected with biotinylated donkey anti-rabbit IgG secondary antibody (Jackson Immunoresearch, West Grove, PA) and visualised with Streptavidin-Texas Red (red fluorophore, Sigma). Briefly, the sections or cells were incubated overnight with the primary antibodies diluted to 3 μg/ml with 10% normal horse serum (NHS) in PBS containing 0.05% Merthiolate and 0.2% Triton X-100. Subsequently, the slides were incubated with biotinylated donkey anti-rabbit IgG (Jackson Immunoresearch) diluted 1:500 in 1% NHS in PBS containing 0.05% Merthiolate for 1 hr, followed by incubation in Streptavidin-Texas Red diluted 1:200 in PBS containing 0.05% Merthiolate for 1 hr. All incubations were held at room temperature and separated by three 5-min washes in PBS. Slides were mounted with citiflour and examined with fluorescence microscopy. Control experiments were performed both by using an excess of the appropriate homologue peptide antigen to absorb the primary antibodies and by omission of the primary antibody to confirm the specificity of the immunoreaction.
All responses were normalized with respect to cell capacitance, to give a current density in pA/pF, unless otherwise stated. All data are expressed as the means ± S.E.M. Statistical analysis (Student's t-test) was performed using Origin 4.1 (Microcal, Northampton, MA). Concentration-response data were fitted with the Hill equation:
where A is the maximum effect, K is the EC50, and nH is the Hill coefficient. The combined data from the given number of cells were fitted, and the results are presented as values ± S.E., determined by the fitting routine. Traces were acquired using Fetchex (pCLAMP software) and plotted using Origin 4.1.
ATP, αβ-meATP, and 1,1-Dimethyl-4-phenylpiperazinium (DMPP) were obtained from Sigma. TNP-ATP was from Molecular Probes Europe (Leiden, The Netherlands). Solutions of ATP and other drugs were prepared using deionized water and stored frozen. All drugs were then diluted in extracellular bathing solution to the final concentration.
We are grateful to Dr. Chrystalla Orphaides for editorial assistance.