Introduction

P2X receptors are a family of ligand-gated ion channels responsive to ATP. Seven subtypes have been identified which form homomultimeric or heteromultimeric pores. The P2X3 receptor subunit was cloned in 1995 and shown to be predominantly localised in a subpopulation of small nociceptive sensory neurons in the dorsal root ganglia (DRG), trigeminal ganglia (TG) and nodose ganglia (NG). Subsequent immunohistochemical studies showed that these neurons were labelled with the isolectin B4 (IB4; Chen et al. 1995; Cook et al. 1997; Vulchanova et al. 1997, 1998; Bradbury et al. 1998; Llewellyn-Smith and Burnstock 1998; Xiang et al. 1998; Barden and Bennett 2000).

In DRG, intensive P2X3 immunoreactivity is found predominantly in a subset of small- and medium-diameter neurons, which is absent from most large neurons (Vulchanova et al. 1997, 1998; Bradbury et al. 1998; Xiang et al. 1998; Novakovic et al. 1999). The P2X3 subunit is present in approximately equal numbers of neurons projecting to the skin and viscera but in very few of those innervating skeletal muscle (Bradbury et al. 1998). After sciatic nerve axotomy, P2X3 receptor expression decreases by about 50% in lumbar region L4/5 DRG, and this downregulation is reversed by glial cell line-derived neurotrophic factor (Bradbury et al. 1998). In contrast, after chronic constriction injury to the sciatic nerve, the number of P2X3-positive small- and medium-diameter DRG neurons appeared to have increased (Novakovic et al. 1999). The reason why these different forms of nerve damage produce opposite effects of P2X3 subunit expression is at present unclear. At the peripheral terminals of DRG neurons, P2X3 immunoreactivity was observed in nerve fibres in rat and mouse glabrous skin, where the fibres appear to extend superficially and terminate in the epidermis (Vulchanova et al. 1998). P2X3 immunoreactivity was also seen on sensory nerve terminals in the tongue (Cook et al. 1997; Bo et al. 1999; Rong et al. 2000), tooth pulp (Cook et al. 1997; Alavi et al. 2001) and on the suburothelial nerve plexus of mouse urinary bladder, with positive staining on terminals of small nerve fibres that embed in the urothelium (Cockayne et al. 2000). Following sciatic nerve ligation, P2X3 immunoreactivity accumulates proximal to the ligation site, indicating that P2X3 is synthesised in the cell body and transported along the processes into both central and peripheral terminals (Vulchanova et al. 1997; Novakovic et al. 1999).

In TG, P2X3 receptor immunoreactivity occurs in both small and large nerve cell bodies and their processes (Cook et al. 1997; Llewellyn-Smith and Burnstock 1998; Xiang et al. 1998).

Electrophysiological studies have also been carried out to characterise P2X3 receptors in recombinant receptors expressed in Xenopus oocytes and in sensory ganglia (Dunn et al. 2000, 2001; Liu et al. 2001; Zhong et al. 2001) and recordings made in sensory nerves following activation of terminals of these nerves in bladder (Vlaskovska et al. 2001; Rong et al. 2002). The P2X3 receptor is observed mainly in capsaicin receptor vanilloid receptor subtype 1 (VR1)-positive sensory neurons (Guo et al. 1999). In capsaicin-sensitive DRG neurons, ATP and its analogue α,β-methylene ATP evoke an inward current with fast activation and inactivation (Ueno et al. 1999), which is lacking in P2X3-deficient mice (Cockayne et al. 2000; Souslova et al. 2000), indicating that functional P2X3 receptors are expressed in nociceptors. These results all suggest that P2X3 receptor subunits may play important roles in sensory transmission.

There is growing evidence to suggest that purinergic signalling is involved in early embryonic development. Purinoceptors were shown to be one of the first functionally active membrane receptors in chick embryo cells during gastrulation, in which, by means of purinoceptors, ATP induced rapid accumulation of inositol phosphate and Ca2+ mobilisation in a similar way and to the same extent as acetylcholine (Laasberg 1990). Recent reports also implicate ATP as a key regulator of the development of various organs and systems in frog and chick as well as in mammalian embryos (reviewed by Burnstock 1996, 2001). P2X3 receptor expression at the early stages of mouse embryogenesis has been reported, but it is interesting to note that the P2X3 expression patterns were not identical in rat and mouse embryos. Mouse embryos express P2X3 in many tissues (such as DRG and TG) at embryonic day 9.5 (E9.5) embryos where no corresponding expression was found in the rat (Boldogkö et al. 2002; Cheung and Burnstock 2002). Neurons in DRG began to express P2X3 receptors in E12.5 rat embryos. For the TG in E11–11.5 rat embryos, P2X3 was found in the neurons and fibres (Cheung and Burnstock 2002). However, we do not know the changes in numbers of sensory neurons expressing P2X3 receptors during mouse embryonic and postnatal development. The relationship between P2X3 receptors and other phenotypic markers has also been little studied in developing mouse. In this study, the location, quantity of P2X3 receptors and relationship between P2X3 receptors and other phenotypic markers including IB4, calcitonin gene-related peptide (CGRP), neurofilament 200 (NF200) and calbindin was studied in the sensory ganglia from E14 to E18, postnatal and adult mice.

Materials and methods

Tissue preparation

The expression of P2X3 receptor protein in embryonic sensory ganglia was studied in albino outbred mouse embryos of E14, E16 and E18 (five animals/stage) using immunohistochemical techniques. Postnatal mouse sensory ganglia on postnatal day 1 (P1), P7, P14 and adult (five animals/stage) were chosen for examination. The day of identification of the presence of a vaginal plug was designated as day zero (E0). Pregnant mice were killed by asphyxiation with a rising concentration of CO2 (between 0% and 100%), and death was confirmed by cervical dislocation according to Home Office (UK) regulations covering Schedule One procedures. Embryos were collected and fixed in 4% paraformaldehyde in 0.1 M phosphate buffer (PB; pH 7.2) at 4°C, overnight. Embryos were then washed in 0.1 M phosphate-buffered saline (PBS; pH 7.2) and dehydrated using 10% sucrose, 20% sucrose and, finally, 30% sucrose. Thereafter, lumbar DRG, TG and NG were dissected and immersed in OCT-embedding medium and frozen in precooled isopropanol (−70°C) for cryosectioning. Frozen sections (10 μm) were cut and mounted on gelatin-coated slides and dried at room temperature. Postnatal mouse lumbar DRG, TG and NG were dissected after cervical dislocation and the ganglia were fixed and processed as described above. Frozen sections (10 μm) were cut and mounted.

Antisera

Rabbit anti-P2X3 receptor (1:400) antibody was provided by Roche Bioscience (Palo Alto, CA, USA). The immunogens used for production of polyclonal P2X3 antibody were synthetic peptides corresponding to the carboxyl terminal of the cloned rat P2X3 receptors, covalently linked to keyhole limpet hemocyanin. The peptide sequences of the P2X3 receptors are of amino acid sequence 383–397 (VEKQSTDSGAYSIGH). Other antibodies used are mouse anti-NF200 (clone N52; Sigma; 1:400), mouse anti-CGRP (Affiniti; 1:2,000) and mouse anti-calbindin (Swant; 1:20,000). We also used biotin-conjugated IB4 from Griffonia simplicifolia (Sigma; 10 μg/ml).

Immunohistochemistry

Immunohistochemistry for P2X receptors was performed by using rabbit polyclonal antibodies against a unique peptide sequence of P2X3 receptor subtypes. For immunostaining of cryosections, the standard avidin-biotin complex (ABC) technique was used. Sections were postfixed with 4% paraformaldehyde for 2 min at room temperature. Endogenous peroxidase was blocked by 0.5% H2O2 and 50% methanol (methanol:PBS, 1:1) for 20 min. The P2X3 primary antibody was used at a concentration of 1:400, prepared in 10% normal horse serum (NHS) containing 0.2% Triton X-100, overnight. Subsequently, the sections were incubated with biotinylated donkey anti-rabbit IgG (Jackson Immunoresearch, West Grove, PA, USA) at a dilution of 1:500 in PBS containing 1% NHS for 1 h. The sections were then incubated in ExtrAvidin peroxidase diluted 1:1,000 in PBS for 1 h at room temperature. For colour reactivity, a solution containing 0.05% 3,3′-diaminobenzidine (DAB), 0.04% nickel ammonium sulphate, 0.2% β-d-glucose, 0.004% ammonium nitrate and 1.2 U/ml glucose oxidase in 0.1 M PB (pH 7.2) was applied. Sections were washed three times with 0.1 M PBS after each of the above steps (except for serum preincubation). Slides were mounted with Eukitt (BDH Laboratory, UK) and examined with light microscopy. The control experiments were carried out with the primary antibodies preadsorbed with the peptides for immunising the rabbits or the primary antibody replaced with NHS.

Following immunohistochemistry, a group of sections was counterstained with 1% neutral red. Positively and negatively immunostained neurons were counted to calculate the proportion of positive neurons. To ensure counting was unbiased, the procedure was conducted ‘blind’ ensuring that the operator was not aware of the developmental stage or immunostaining performed on each slide.

Immunofluorescence double labelling

In colocalisation studies investigating the coexpression of P2X3 receptors and NF200, CGRP or calbindin, sections were incubated with the P2X3 antibody overnight and detected with Cy3-conjugated donkey anti-rabbit IgG, incubated with either monoclonal NF200, CGRP or calbindin antibody (4°C, overnight) and detected with fluorescein isothiocyanate (FITC)-conjugated mouse antibody (raised in goat; Sigma; 1:200, 1 h). For colocalisation with IB4, sections were immunostained for the P2X3 receptor, as above, and then incubated with biotin-conjugated IB4 (16 h) and detected with streptavidin-FITC (1:200;1 h). The sections were washed and mounted in Citifluor.

Sections on which counts of IB4-, NF200-, CGRP- and calbindin-positive cells had been performed were marked and then counterstained with toluidine blue (2.5% in 0.1 M PB for 2 min followed by dehydration through increasing grades of alcohol, cleared in xylene and coverslipped with DPX mounting medium). This enabled the total cell profile numbers to be determined by counting all profiles in the marked sections under bright-field illumination. The percentages of total cell numbers that were IB4-, NF200-, CGRP- and calbindin-positive were then calculated.

Photomicroscopy

Images of DAB immunochemical staining and immunofluorescence labelling were taken with the Leica DC 200 digital camera (Leica, Switzerland) attached to a Zeiss Axioplan microscope (Zeiss, Germany). Images were imported into a graphics package (Adobe Photoshop 5.0, USA). The two-channel readings for green and red fluorescence were merged by using Adobe Photoshop 5.0.

Data analysis

All analyses were performed at ×20 objective magnification. P2X3 receptor, IB4-, NF200-, CGRP- and calbindin-positive cells in ganglia were determined by counting all immunoreactive-positive cell profiles in every sixth section throughout the ganglia. Total cell profiles were counted and the percentages of total cell profiles that were P2X3 receptor-, IB4-, NF200-, CGRP- and calbindin-positive were then calculated.

To calculate percentages of P2X3 receptor colocalisation with other markers, four randomly selected ganglia sections were chosen for each pair of markers, for each animal. For each section, counts were made of the number of profiles that were positive for the P2X3 receptor, the number of profiles positive for the other markers (CGRP, calbindin, NF200 or IB4) and the number of profiles expressing both antigens and percentages were calculated. All values are given as the mean ± SEM (%) and the results were statistically analysed by a one-way analysis of variance followed by a Tukey’s post hoc test where more than five age groups were analysed. For analyses of less than five age groups, a Bonferroni’s post hoc test was used. Although all grouping of data were compared by this test, statistical significance between each group and the one immediately preceding it has been highlighted in the tables; * P<0.05, ** P<0.01, *** P<0.001. For the data in Table 1, a Tukey’s post hoc test was performed to compare embryonic days E14, E16 and E18 with postnatal ages P7, P14 and adult; *** P<0.001.

Results

P2X3 immunoreactivity in the sensory ganglia during mouse embryonic and postnatal development

Immunohistochemistry was performed to investigate the pattern of P2X3 receptor protein expression in sensory ganglia during mouse embryonic and postnatal development. Using the standard ABC method with nickel and DAB as chromogens, black staining indicated positive immunoreactivity. In this study, mouse embryos of E14, E16, E18 and postnatal P1, P7, P14 and adult mice were chosen to investigate P2X3 immunoreactivity and the results are summarised in Table 1 and Fig. 1.

Table 1 Expression of P2X3 receptor immunoreactivity in dorsal root ganglion (DGR), trigeminal ganglion (TG), nodose ganglion (NG) in mouse embryonic and postnatal development. n Number of P2X3 receptor-positive cell profiles and the total number of cell profiles counted, respectively (five animals/stage). Note statistical significance indicated by asterisks relates to embryonic days E14, E16 and E18, as compared with postnatal ages P7, P14 and adult
Full size table
Fig. 1A–C
figure 1

Percentage of P2X3-immunoreactive nerve cell bodies in sensory ganglia of mouse in embryonic and postnatal development. A Dorsal root ganglia. B Trigeminal ganglia. C Nodose ganglia. Note statistical significance indicated by asterisks relates to postnatal ages P7, P14 and adult as compared with embryonic days E14, E16 and E18. *** P<0.001

Full size image

The intensive expression of P2X3 in DRG, TG and NG persists from E14 through later stages of development and into the adult. In embryos (E14–E18), P2X3 is expressed in a homogeneous pattern in almost all DRG, TG and NG neurons. By contrast, in later developmental stages (postnatal P1–P14), P2X3 receptor expression is significantly reduced in DRG, TG and NG. From P14 to adult mouse, P2X3 receptor expression is largely restricted to a subpopulation of small-diameter neurons, with fewer large-diameter cells strongly immunopositive. The staining was evenly distributed throughout the cytoplasm of these cells and positively stained cells were apparently randomly distributed throughout the ganglia (Fig. 2).

Fig. 2A–O
figure 2

Localisation of P2X3 receptor immunoreactivity in sensory ganglia of mouse in embryonic and postnatal development. A–E P2X3 receptor immunoreactivity in dorsal root ganglion (DRG). A E14. B E16. C P1. D P7. E Adult. F–J P2X3 receptor immunoreactivity in trigeminal ganglion (TG). F E14. G E16. H P1. I P7. J Adult. K–O P2X3 receptor immunoreactivity in nodose ganglion (NG). K E14. L E16. M P1. N P7. O Adult. Scale bar 100 μm

Full size image

Colocalisation of P2X3 receptors with cytochemical markers of sensory neurons

In order to further identify the sensory neurons that express P2X3 receptors, a number of established markers were used. In colocalisation experiments, immunolabelling for the P2X3 receptor was compared with that for IB4, CGRP, NF200 and calbindin. Briefly, as stated above, the P2X3 receptor immunolabelling was evenly distributed throughout the cytoplasm of the cell body. IB4 produced a labelled plasmalemma and diffuse speckling in the cytoplasm (Fig. 3). CGRP immunolabelling produced diffuse cytoplasmic speckling and a more diffuse perinuclear ring (Fig. 4). NF200 labelling was evenly distributed throughout the cytoplasm of the cell body, but found mainly in large neurons (Fig. 5). Calbindin labelling was seen throughout the whole cell (Fig. 6).

Fig. 3A–R
figure 3

Colocalisation (yellow/orange) of P2X3 receptor immunoreactivity (-IR; red) with isolectin B4 (IB 4 ; green) immunostaining in the developmental mouse sensory ganglia. A–C Double staining for P2X3-IR and IB4 in DRG of P7 mouse. D–F Double staining for P2X3-IR and IB4 in DRG of adult mouse. G–I Single section double stained for P2X3-IR and IB4 in TG of P7 mouse. J–L Double staining for P2X3-IR and IB4 in TG of adult mouse. M–O Double staining for P2X3-IR and IB4 in NG of P7 mouse. P–R Double staining for P2X3-IR and IB4 in NG of adult mouse. Scale bar 100 μm

Full size image
Fig. 4A–L
figure 4

Colocalisation (yellow/orange) of P2X3 receptor immunoreactivity (red) with calcitonin gene-related peptide (CGRP) immunoreactivity (green) in the developmental mouse sensory ganglia. A–D Double staining for P2X3-IR and CGRP-IR in DRG of mouse embryonic and postnatal development (A E16; B P1; C P7; D adult). E–H Double staining for P2X3-IR and CGRP-IR in TG of mouse embryonic and postnatal development (E E16; F P1; G P7; H adult). I–L Double staining for P2X3-IR and CGRP-IR in NG of mouse embryonic and postnatal development (I E16; J P1; K P7; L adult). Scale bar 100 μm

Full size image
Fig. 5A–L
figure 5

Colocalisation (yellow/orange) of P2X3 receptor immunoreactivity (red) with neurofilament 200 (NF200) immunoreactivity (green) in the developmental mouse sensory ganglia. A–D Double staining for P2X3-IR and NF200-IR in DRG of mouse embryonic and postnatal development (A E16; B P1; C P7; D adult). E–H Double staining for P2X3-IR and NF200-IR in TG of mouse embryonic and postnatal development (E E16; F P1; G P7; H adult). I–L Double staining for P2X3-IR and NF200-IR in NG of mouse embryonic and postnatal development (I E16; J P1; K P7; L adult). Scale bar 100 μm

Full size image
Fig. 6A–L
figure 6

Colocalisation (yellow/orange) of P2X3 receptor immunoreactivity (red) with calbindin immunoreactivity (green) in the developmental mouse sensory ganglia. A–D Double staining for P2X3-IR and calbindin-IR in DRG of mouse embryonic and postnatal development (A E16; B P1; C P7; D adult). E–H Double staining for P2X3-IR and calbindin-IR in TG of mouse embryonic and postnatal development (E E16; F P1; G P7; H adult). I–L Double staining for P2X3-IR and calbindin-IR in NG of mouse embryonic and postnatal development (I E16; J P1; K P7; L adult). Scale bar 100 μm

Full size image

P2X3/IB4

The population of small [predominantly anti-neurofilament H (RT-97)-negative] DRG neurons can be subdivided into two populations, one which binds IB4 and another which contains the neuropeptide CGRP (Silverman and Kruger 1990; Alvarez et al. 1991). These are important markers because several lines of evidence suggest that they identify subpopulations of putative nociceptors in DRG.

Isolectin B4 was used as a marker for fluoride-resistant acid phosphatase (FRAP)-containing neurons. IB4 labelling was observed in 56.7%, 38.3% and 49.2% of all examined adult mouse DRG, TG and NG neurons, respectively. IB4 labelling in DRG, TG and NG of embryos was not detected until postnatal stages and was restricted to the small cell population. In P7 mouse, the number of sensory neurons binding IB4 was increased significantly and reached adult levels at P14 (Table 2). All IB4-positive neurons showed colocalisation with the P2X3 receptor in DRG, TG and NG of P7 mice. However, the number of neurons expressing the P2X3 receptor was greater than the number of IB4-positive neurons; only 43.2%, 29.6% and 21.5% of P2X3 receptor-positive neurons were also positive for IB4 in DRG, TG and NG of P7 mice, respectively. However, from P14 there was a significant increase of 99.2%, 70.3% and 60.3% of P2X3 neurons containing IB4 in DRG, TG and NG, respectively (Table 3; Fig. 3).

Table 2 Changes of isolectin B4 (IB4), medium molecular weight neurofilament (NF200), calcitonin gene-related peptide (CGRP) and calbindin immunoreactivity in mouse embryonic and postnatal development. − Not detected, n number of IB4-, NF200-, CGRP- or calbindin-positive cell profiles and the total number of cell profiles counted, respectively (five animals/stage). Note statistical significance indicated by asterisks relates to comparison of a developmental stage with the preceding age group
Full size table
Table 3 Colocalisation of P2X3 receptor immunoreactivity with IB4 in the developing mouse sensory neurons. n Number of double-labelled cell profiles and the total number of cell profiles counted, respectively, for each combination of receptor (five animals/stage). Note statistical significance indicated by asterisks relates to comparison of a developmental stage with the preceding age group
Full size table

P2X3/CGRP

Calcitonin gene-related peptide immunoreactivity was low in DRG, TG and NG of mouse embryos, but the percentage of DRG and TG neurons expressing the peptide significantly increased from E18 (8.5%, 7.9%) to P1 (21.1%, 18.3%) and continued increasing until 2 weeks after birth, P14. The proportion of CGRP-immunoreactive neurons stabilised after the second postnatal week, but adult values were not statistically different from P14. In NG, CGRP immunoreactivity gradually increased after birth and reached adult levels at P14 (Table 2). Double immunofluorescence revealed that almost all neurons that were positive for CGRP (90–100%) also showed colocalisation with P2X3 receptors in DRG, TG and NG of mouse embryos (Fig. 4). However, in adult mouse, only 18–34% of CGRP neuron profiles were also positive for the P2X3 receptor in DRG, TG and NG. Conversely, 17.6%, 31.9% and 4.3% of P2X3 receptor neurons coexpressed CGRP in DRG, TG and NG, respectively (Table 4; Fig. 4).

Table 4 Colocalisation of P2X3 receptor immunoreactivity with CGRP in the developmental mouse sensory neurons. n Number of double-labelled cell profiles and the total number of cell profiles counted, respectively, for each combination of receptor (five animals/stage). Note statistical significance indicated by asterisks relates to comparison of a developmental stage with the preceding age group
Full size table

P2X3/NF200

Neurofilament 200 is an anti-neurofilament antibody that stains the classically defined large neurons. In DRG and NG of embryos, approximately 3–15% neurons were NF200-immunoreactive positive cells. After birth, NF200 immunoreactivity was significantly increased and reached adult levels at P14 and this percentage did not significantly change in postnatal and adult mice (Table 2). Unlike DRG and NG, almost half of the neurons were NF200 immunoreactive in TG of embryos. Double immunofluorescence histochemistry showed that all NF200-immunoreactive neurons were also positive for P2X3 in DRG, TG and NG in E16. However, in adult mouse, only 0.5%, 15.2%, and 49.8% of NF200-immunoreactive neuron profiles were also positive for P2X3 in DRG, TG and NG, respectively (Table 5; Fig. 5).

Table 5 Colocalisation of P2X3 receptor immunoreactivity with NF200 in the developmental mouse sensory neurons. n Number of double-labelled cell profiles and the total number of cell profiles counted, respectively, for each combination of receptor (five animals/stage). Note statistical significance indicated by asterisks relates to comparison of a developmental stage with the preceding age group
Full size table

P2X3/calbindin

Calbindin-positive immunostaining was found in 5–16% of neurons in DRG, TG and NG of mouse embryos (E16). It significantly increased at postnatal P14 and remained the same in adults (Table 2). These neurons were large- or medium-sized in the DRG, TG and NG of developing mouse. Almost all calbindin-positive neurons were colocalised with P2X3 receptors in DRG, TG and NG of mouse embryos. However, in adult mouse only 0.6% and 11.1% of calbindin-positive neurons were also positive for the P2X3 receptor in DRG and TG, respectively. However, in NG, the number of neurons coexpressing P2X3 and calbindin (31.1%) was greater than that in DRG and TG (Table 6; Fig. 6).

Table 6 Colocalisation of P2X3 receptor immunoreactivity with calbindin in the developmental mouse sensory neurons. n Number of double-labelled cell profiles and the total number of cell profiles counted, respectively, for each combination of receptor (five animals/stage). Note statistical significance indicated by asterisks relates to comparison of a developmental stage with the preceding age group
Full size table

Discussion

P2X receptors are ligand-gated ion channels activated by extracellular ATP that mediate rapid cation permeability and fast excitatory neurotransmission in both the central and peripheral nervous systems (reviewed by Ralevic and Burnstock 1998). One of the P2X receptor subunits, P2X3, was cloned from rat dorsal root ganglia (Chen et al. 1995; Lewis et al. 1995) and is known to be largely restricted to a subset of sensory neurons (TG, NG and DRG). In this study, we present the detailed expression changes of the P2X3 receptor at different stages of rat embryonic and postnatal development. In addition, the coexpression of the P2X3 receptor and other phenotypic markers was also examined during embryonic and postnatal development. The results obtained have shown that the P2X3 receptor is the dominant receptor subtype among the P2X receptor family in the embryonic sensory ganglion and that P2X3 receptor expression is downregulated in the postnatal sensory ganglion of mouse.

In this study, we observed that the P2X3 receptor is intensively and homogeneously expressed in the DRG, TG and NG from E14 to P1. However, after birth there was a gradual decline to about 50% of sensory neurons that were positive for P2X3 at P14 with little further change in adults. From E14 to P7 there were mostly only small neurons, while from P7 onwards both large and small neurons were present. At P7 to adult, some large neurons stained for P2X3 in TG and NG, but in DRG, for the most part, only small neurons were P2X3 positive, which corresponds to the finding that approximately 40% of the DRG neurons express P2X3 mRNA (Chen et al. 1995). The P2X3 receptor has been demonstrated on small- or medium-sized neurons in the sensory ganglia in rat and monkey (Cook et al. 1997; Vulchanova et al. 1997; Xiang et al. 1998). The expression of P2X3 receptor subunits in sensory ganglia may be involved in signal transduction and modulation as well as in regulating cellular maturation during development.

In the present work, we identified the particular class of sensory neuron that expresses P2X3 receptors with the use of double immunofluorescence immunohistochemistry. In the sensory ganglia, about 60% of the total neurons are the classically defined small population. These cells predominantly have unmyelinated C-fibre axons. Electrophysiological studies have shown that most of these cells (approximately 90%) in mouse, rat, monkey and human are nociceptive in function (Snider and McMahon 1998). These small cells can be further subdivided into two major classes according to their neurochemical phenotype (Snider and McMahon 1998): peptidergic and non-peptidergic neurons. The former, about one-half, express two major peptidergic neuromodulators, substance P and CGRP, and also the p75 neurotrophin receptor and TrkA, the NGF-specific tyrosine kinase receptor. The central terminals of these neurons project to lamina I/IIo of the superficial dorsal horn of the spinal cord, a region where many small afferents terminate. Few cells here were found to coexpress CGRP and the P2X3 receptor. CGRP is the most prevalent neuropeptide identified in subpopulations of primary afferent neurons, constituting 40–50% of DRG and TG neurons (Lee et al. 1985; Ju et al. 1987; Kai-Kai 1989). The other half of the C-fibre population does not express NGF receptors (McMahon et al. 1994; Averill et al. 1995) but can be identified by a number of markers: they bind IB4, express the enzyme thiamine monophosphatase (TMP) and stain with the antibody LA4 (marker of a population of primary afferents in the dorsal horn). They also express the heat-responsive vanilloid receptor, VR1 (Caterina et al. 1997), which implicates them in thermal nociception. The central terminals of these neurons project to inner lamina 2 of the superficial dorsal horn of the spinal cord, another region where many fine afferents terminate. In vivo studies (Gerke and Plenderleith 2001) demonstrate that IB4-binding sites are selectively expressed by nociceptive afferents. Physiological studies demonstrate that some IB4-binding DRG neurons are deep tissue, high-threshold mechanoreceptors and polymodal nociceptors. In our experiments, we found that IB4-positive neurons in DRG, TG and NG did not appear until birth, but the numbers increased to about 50% by P14 with little further change in adults. We also found that about 10% of neurons in DRG, TG and NG were positive for CGRP in embryos, but this increased postnatally to about 35–40% in adults, except for NG where CGRP immunoreactivity only reached about 15% in adults.

To characterise the cytochemical profile of P2X3-positive neurons, we examined the colocalisation of the P2X3 receptor with IB4 and CGRP. At P7, a few P2X3-positive neurons in DRG,TG and NG stained for IB4, but by P14 and in adults a high proportion of IB4-positive neurons also labelled P2X3. In contrast, in the embryo nearly 100% of CGRP-positive neurons, albeit relatively few, stained for P2X3, but by postnatal P14 and in adult ganglia only about 20% in DRG and NG and 35% in TG showed colocalisation of CGRP and P2X3 receptors. Since IB4 has been shown to label almost all FRAP-positive neurons (Wang et al. 1994), and because we found little colocalisation of the P2X3 receptor with CGRP in adults, our results indicate that P2X3 is predominantly expressed by FRAP-containing DRG and TG neurons. This conclusion is supported by the extensive colocalisation of P2X3 and LA4-IR; LA4 also labels almost all FRAP-positive cells (Dodd and Jessell 1985). It has been proposed that FRAP-containing and peptide-containing sensory neurons represent two parallel systems for processing of nociceptive information (Hunt and Rossi 1985). A nociceptive role for FRAP-positive neurons is also suggested by the localisation of the enzyme activity in nerve fibres in cornea (Silverman and Kruger 1988) and in tooth pulp primary afferents (Fried et al. 1989), as these peripheral targets are thought to be innervated primarily by nociceptors. Unlike DRG and TG, in NG, the highest intensity of P2X3 neurons was most often observed within the small IB4-negative population. The types of overlap in the binding of IB4 and coexpression of P2X3 receptor, described in the present study for NG neurons, is both similar and different from the patterns described for DRG and TG neurons. A continued focus using these and other population markers for each of these functionally different neuronal populations (i.e. NG versus DRG, TG) has the potential for identifying a unique neuronal fingerprint for the processing of visceral sensory information.

The large cells within the sensory neurons are rich in neurofilament. We identified these cells with the antibody NF200 and found that NF200 was expressed in about 50% of neurons in TG in the embryo, which persisted in postnatal and adult mice. However, a relatively low percentage of NF200-positive neurons were present in embryonic DRG and NG, but the percentage increased substantially by P14 and in adults, when many of the neurons became large, especially in NG. In the embryo, all NF200-positive neurons stained for P2X3 in DRG, TG and NG, but by adulthood this was reduced to 0.5%, 15% and 50% in these ganglia, respectively. The P2X3 receptor was expressed not only in small- and medium-sized neurons, but also in large neurons of TG and NG. The large neurofilament-rich neurons are known to possess myelinated axons and to be predominantly responsive to non-nociceptive innocuous mechanical stimuli.

Calbindin-D28 k is widely represented in the nervous system and is used as a cellular marker for neuroanatomical studies along with other calcium-binding proteins such as parvalbumin and calretinin (Andressen et al. 1993). Calbindin-D28 k immunoreactivity has been observed in the sensory and autonomic ganglia such as the DRG, TG, jugular ganglion and superior cervical ganglion (Carr et al. 1989; Ichikawa and Helke 1995; Ichikawa et al. 1996; Wakisaka et al. 1996; Grkovic and Anderson 1997). In the present study, we observed calbindin-positive immunostaining in 5–12% of neurons in DRG, TG and NG of mouse embryos. Expression is increased gradually postnatally and reached adult levels at P14. Almost all calbindin-positive neurons were colocalised with P2X3 receptors in DRG, TG and NG of mouse embryos. In contrast, in adult DRG and TG, few showed colocalisation, 0.6% and 11%, respectively (because calbindin is mostly present in large neurons), while in adult NG, where there are many large neurons, 31% showed colocalisation. These results are consistent with those previously reported (Kashiba et al. 1990; Honda 1995; Ichikawa and Helke 1995; Ichikawa et al. 1996). These findings suggest that the P2X3 receptor is related to cutaneous sensory neurons in DRG and TG and splanchnic sensory neurons in NG.