ATP triggers intracellular Ca²⁺ release in type II cells of the rat carotid body

Abstract

Using a Ca²⁺-imaging technique, we studied the action of ATP on the intracellular Ca²⁺ concentration ([Ca²⁺]_i) of fura-2-loaded mixtures of type I and type II cells dissociated from rat carotid bodies. ATP (100 μm) triggered a transient rise in [Ca²⁺]_i in the spindle-shaped type II (sustentacular) cells, but not the ovoid type I (glomus) cells. When challenged with ionomycin (1 μm), no amperometry signal could be detected from the ATP-responsive type II cells, suggesting that these cells lacked catecholamine-containing granules. In contrast, KCl depolarization triggered robust quantal catecholamine release from type I cells that were not responsive to ATP. In type II cells voltage clamped at −70 mV, the ATP-induced [Ca²⁺]_i rise was not accompanied by any current change, suggesting that P2X receptors are not involved. The ATP-induced Ca²⁺ signal could be observed in the presence of Ni²⁺ (a blocker of voltage-gated Ca²⁺ channels) or in the absence of extracellular Ca²⁺, indicating that Ca²⁺ release from intracellular stores was the dominant mechanism. The order of purinoreceptor agonist potency in triggering the [Ca²⁺]_i rise was UTP > ATP > 2-methylthioATP ≫α,β-methyleneATP, implicating the involvement of P2Y₂ receptors. In carotid body sections, immunofluorescence revealed localization of P2Y₂ receptors on spindle-shaped type II cells that partially enveloped ovoid type I cells. Since ATP is released from type I cells during hypoxia, we suggest that the ATP-induced Ca²⁺ signal in type II cells can mediate paracrine interactions within the carotid bodies.

Oxygen is essential for the survival of mammalian cells. An immediate response to a hypoxic stress is the activation of arterial oxygen receptors in the carotid bodies, which in turn stimulates the brainstem respiratory centre (via the carotid sinus nerve) and leads to a reflex increase in respiratory rate. The carotid body comprises small groups of ovoid, type I (glomus) cells, which are partially enveloped by the spindle-shaped type II (sustentacular) cells (see review by Grönblad, 1983; Gonzalez et al. 1994). It is generally accepted that type I cells are the initial sites of oxygen sensing, and the release of neurotransmitters from type I cells in turn stimulates the carotid sinus nerve (see review by López-Barneo, 1996; Prabhakar, 2000). Type I cells contain a variety of neurotransmitters and hormones, including catecholamines (mostly dopamine), acetylcholine (ACh), ATP, opioids and substance P (Gonzalez et al. 1994). Hypoxia has been shown to stimulate dopamine release from type I cells in many animal species, including rabbit and rat (Donnelly, 1993; Ureña et al. 1994; Pardal et al. 2000). Recent studies in rat carotid bodies have also suggested that the co-release of ATP and ACh from type I cells mediates hypoxic signalling (Zhang et al. 2000). In contrast, the functional role of type II cells is not well understood. Type II cells do not contain any dense-core granules and exhibit no uptake of exogenous monoamines (Grönblad, 1983). Patch-clamp studies on type II cells dissociated from rabbit show that these cells are electrically non-excitable and do not express any voltage-gated Ca²⁺ or Na⁺ current (Duchen et al. 1988; Ureña et al. 1989). Immunohistochemical and ultrastructural studies indicate that type II cells resemble the glial cells of the peripheral nervous system (Kameda, 1996). Traditionally, glial cells are viewed as passive supportive structures. However, recent studies have shown that a variety of neurotransmitters can elicit Ca²⁺ signals in glial cells (reviewed by Verkhratsky et al. 1998). Among these neurotransmitters, ATP has been shown to act via purinoreceptors to elicit Ca²⁺ signals in different types of glial cells, including Schwann cells, astrocytes, oligodendrocytes and microglia (Verkhratsky et al. 1998). Since type I cells release ATP during hypoxia (Zhang et al. 2000) and they are partially enveloped by type II cells (Grönblad, 1983; Kameda, 1996), it is possible that type II cells are exposed to local high concentrations of ATP during hypoxia. In this study, we investigated the action of ATP on single type II cells dissociated from rat carotid bodies. Our results indicate that ATP, acting via P2Y₂ receptors, triggers Ca²⁺ release from the intracellular stores of type II cells. In intact carotid bodies, type II cells are in close contact with type I cells as well as many nerve endings (Grönblad, 1983; Kameda, 1996). Thus, the ATP-induced Ca²⁺ signal in type II cells may play an important role in the hypoxic signalling by modifying the response of type I cells and/or the nerve endings.

METHODS

Cell preparation

The carotid bifurcation was removed from male Sprague-Dawley rats (age 6–7 weeks) killed with an overdose of halothane in accordance with the standards of the Canadian Council on Animal Care. The tissue was placed in ice-cold 100 % oxygen-equilibrated Tyrode solution (mm): 140 NaCl, 5 KCl, 2 CaCl₂, 1.1 MgCl₂, 5 glucose and 10 Hepes, pH 7.4 and the carotid bodies were dissected from the surrounding tissue. The details of the cell dissociation procedure are similar to that described by López-López et al. (1997). Briefly, the carotid bodies were incubated with Ca²⁺- and Mg²⁺-free Tyrode solution containing collagenase (2 mg ml⁻¹, Type IV; Sigma) and deoxyribonuclease (DNase type V, 0.2 mg ml⁻¹; Sigma) for 30 min at 37 °C. Trypsin (type XIII; 0.4 mg ml⁻¹) was then added and the cells were incubated for another 10 min at 37 °C. Following trituration, dissociated cells were plated onto glass coverslips and kept in a medium containing F-12/Dulbecco's modified Eagle's medium (1:1), supplemented with 10 % fetal calf serum, 50 U ml⁻¹ penicillin G and 50 μg ml⁻¹ streptomycin (all from Gibco). Cells were cultured at 37 °C in an incubator circulated with air and 5 % carbon dioxide for 12–36 h before recording.

Solutions

The standard bath solution contained (mm): 150 NaCl, 10 Hepes, 8 glucose, 2.5 KCl, 2 CaCl₂ and 1 MgCl₂ (pH 7.4). In experiments where extracellular Ca²⁺ was removed, Ca²⁺ was omitted from the standard bath solution and 2 mm MgCl₂ and 1 mm Na-EGTA were included. For whole-cell patch-clamp recording of current during ATP challenges, the standard pipette solution contained (mm): 120 K-Asp, 20 KCl, 20 K-Hepes, 2 MgCl₂, 2 Na₂ATP, 0.1 Na₄GTP and 0.1 indo-1 (pH 7.4). For measurement of the voltage-gated K⁺ current, indo-1 was omitted from the standard pipette solution and 10 mm EGTA was included. For isolation of voltage-gated Ca²⁺ current, the pipette solution contained (mm): 120 Cs-Asp, 20 TEA-Cl, 20 Cs-Hepes, 10 EGTA, 2 MgCl₂, 2 Na₂ATP and 0.1 Na₄GTP (pH 7.4). The cells were continuously perfused with control or drug solution. The time for a complete change of bath solution was approximately 80 s.

Electrophysiology

In experiments involving the detection of quantal catecholamine release, carbon fibre (tip diameter 7 μm) amperometry (Wightman et al. 1991) was employed as described in our previous studies (Xu & Tse, 1999; Xu et al. 2002). Briefly, +700 mV was applied to the carbon fibre electrode, using a VA-10 Voltammeter (NPI Electronic, Tamm, Germany) and the cell was stimulated by bath application of a high [K⁺] (50 mm) extracellular solution (the standard bath solution with equal molar replacement of NaCl by KCl). Amperometric currents were first recorded on VCR tapes using a NeuroData PCM recorder (Neuro Data, New York, NY, USA) and digitized later at 10 kHz (filtered at 1 kHz). In experiments involving current recording, single cells were voltage-clamped with the whole-cell gigaseal method (Hamill et al. 1981) using an EPC-7 patch-clamp amplifier. Holding potentials and voltage pulses were controlled using an IBM-compatible PC and the data acquisition program Pulse Control (Herrington & Bookman, 1994). The pipettes were made from haematocrit glass (VWR Scientific Canada, London, ON, Canada) and the resistance was 2–4 MΩ after filling, and 5–10 MΩ during whole-cell recording. All experiments were performed at room temperature (20–23 °C). A −10 mV junction potential was corrected throughout. Values are given in the text as mean ± s.e.m.

[Ca²⁺]_i measurement

In most experiments, [Ca²⁺]_i was monitored with digital imaging using a Tillvision imaging system equipped with a Polychrome II high-speed monochromator (Till Photonics, OR, USA). Cells were loaded with fura-2 AM (2.5 μm) in standard bath solution at 37 °C for 10 min and then washed with standard bath solution for 15 min before recording. Fura-2 was excited sequentially by 340 and 380 nm light delivered from a xenon lamp via a × 40, 1.3 NA UV fluor oil objective (Olympus). Fluorescent images were collected at 510 nm every 10 s by a Peltier-cooled CCD camera. The ratio of fluorescence, R (340 nm/380 nm) from individual cell was analysed with the aid of Tillvision Software 3.02 (Till Photonics) on an IBM-compatible computer, according to the equation (Grynkiewicz et al. 1985):

where all three calibration constants, R_min, R_max and K* were calibrated in situ by dialysing various solutions of known Ca²⁺ concentration into individual cells via the whole-cell pipette, as described previously (Tse & Tse, 1998). For all fura-2 measurements shown here, the values for R_min, R_max and K* were 0.13, 3.4 and 2.72 μm, respectively.

In experiments involving simultaneous measurement of [Ca²⁺]_i and amperometry, cells were incubated with indo-1 AM (2.5 μm), similar to that described above for fura-2 AM. In experiments involving simultaneous measurement of [Ca²⁺]_i and current, the K⁺ salt of indo-1 (100 μm) was dialysed into the cell via the whole-cell patch pipette. Details of the instrumentation and procedures of the [Ca²⁺]_i measurement are as described previously (Tse & Tse, 1998). Briefly, indo-1 was excited by 365 nm (band-pass filtered) light delivered from a HBO 100 W mercury lamp via a × 40, 1.3 NA UV fluor oil objective (Nikon). Emitted photons were collected at 405 and 500 nm by photomultiplier tubes (Hamamatsu H3460–04), translated into logic signals, and then counted simultaneously by a CYCTM-10 counter card (Cyber Research, Branford, CT, USA) in an IBM-compatible computer. The ratio of fluorescence, R (405 nm/500 nm) was used to calculate [Ca²⁺]_i, as described above for fura-2. For all indo-1 measurements shown here, the values for R_min, R_max and K* were 0.29, 2.9 and 2.34 μm, respectively

Immunocytochemistry

S-100 immunostaining

Cultured cells on glass coverslips were fixed with phosphate-buffered saline (PBS) containing 4 % paraformaldehyde for 10 min at room temperature. The coverslips were then treated sequentially with the following solutions at room temperature: (1) goat serum (1:200 dilution; Vector Laboratory, Burlingame, CA, USA) for 30 min; (2) rabbit antibody to S-100 (1:500 dilution; DakoCytomation, Mississauga, ON, Canada) for 3 h; (3) biotinylated goat anti-rabbit antibody (1:200 dilution; Chemicon International, Temecula, CA, USA) for 1 h; (4) biotin–avidin–peroxidase complex (1:100 dilution; Vector) for 30 min; (5) diaminobenzidine and 0.06 % H₂O₂ (Vector) for 5 min. Between each treatment, the coverslips were rinsed with PBS for 5 min. For controls, some coverslips were incubated without the rabbit antibody to S-100. Cells were viewed under a × 100 oil-immersion objective (Nikon) using bright-field microscopy.

P2Y₂ receptor immunostaining

Rat carotid bodies were fixed with PBS containing 4 % paraformaldehyde for 10 h at room temperature and then incubated in PBS containing 25 % sucrose for 15 h. The tissue was immersed in optimal temperature cutting compound (TissueTek, Sakura Finetek, Torrance CA, USA) and frozen at −20 °C. Sections (∼20 μm thickness) were cut in a cryostat and collected on glass slides. After drying at 4 °C for 48 h, the sections were rinsed with PBS for 5 min and then incubated at room temperature for 2 h with a blocking, permeabilizing solution containing 7.5 % horse serum and 0.3 % Triton X-100 in PBS. The sections were then incubated overnight at 4 °C with the primary antibody, a goat polyclonal IgG raised against an intracellular site of P2Y₂ receptors (1:200 dilution; Santa Cruz Biotechnology, Santa Cruz, CA, USA). After three washes (15 min each) with 0.3 % Triton X-100 (in PBS), the sections were incubated at room temperature for ∼1.5 h with a secondary antibody, a rabbit anti-goat IgG conjugated to fluorescein (1:40; Calbiochem, La Jolla, CA, USA). The sections were rinsed twice with 0.3 % Triton X-100 (15 min each) and then washed with PBS for 15 min before viewing with a × 63 water-immersion objective (Zeiss) on a confocal microscope (Zeiss LSM 510) with a fluorescein filter set. For controls, sections were incubated without the primary antibody and no staining could be observed.

Chemicals

Fura-2 AM, 2-methylthioATP (2MeSATP), α,β-methyleneATP (α,β-meATP) and ionomycin were obtained from Calbiochem. Indo-1 and indo-1 AM were obtained from Texas Fluorescence Laboratories (Austin, TX, USA). UTP, ATP, GTP, Triton X-100, PBS and reactive blue 2 (Rb-2) were obtained from Sigma-Aldrich (Oakville, ON, Canada).

RESULTS

ATP triggers a transient [Ca²⁺]_i rise in type II but not type I cells of rat carotid bodies

In cell cultures obtained from enzymatic dissociation of rat carotid bodies, two morphologically distinct cell types could be detected. The majority (∼80 %) of the cells were ovoid in shape and the rest of the cells were spindle shaped (Fig. 1A). Application of ATP (100 μm) to the cell mixture elicited a transient [Ca²⁺]_i elevation in the spindle-shaped cells but not in the ovoid cells (Fig. 1B). On average, ATP (100 μm) elevated [Ca²⁺]_i in spindle-shaped cells, from a resting level of 62 ± 5 nm to 524 ± 51 nm (n = 47). In contrast, ATP failed to elicit any change in [Ca²⁺]_i in the ovoid cells (n = 59). Based on the morphology of the carotid body cells described in the literature (Duchen et al. 1988; Ureña et al. 1989; Kameda, 1996; Chou et al. 1998), the ovoid cells that did not respond to ATP were probably type I (glomus) cells and the ATP-responsive cells were probably type II (sustentacular) cells. Type II cells have been reported to express S-100 protein, a common marker for glial cells (Kondo et al. 1982; Abramovici et al. 1991; Kameda, 1996). To examine whether the spindle-shaped cells are type II cells, we immunostained the cells with an antibody to S-100 protein. Figure 1C (left-hand panel) shows a spindle-shaped cell that was immunostained with the antibody to S100 protein. Note that the intensity of the immunostaining was much higher that that of the control cells, which were stained without the antibody (right-hand panel of Fig. 1C). Thus, spindle-shaped cells express immunoreactivity to S-100 protein. The ovoid cells, on the other hand, typically exhibited a higher background signal than the spindle-shaped cells, possibly due to their catecholamine content. However, no increase over the background signal was observed in the ovoid cells when the antibody to S-100 protein was included in the staining procedure, suggesting that there was no significant expression of S-100 protein in the ovoid, type I cells. This result is consistent with previous studies (Kondo et al. 1982; Kameda, 1996) showing that S-100 protein is expressed selectively in type II cells. Another characteristic of the type II cells is their lack of catecholamine-containing granules. Therefore, we exploited this property to examine further whether the ATP-responsive cells were indeed type II cells. In this experiment, we monitored quantal catecholamine release from individual cells using carbon fibre amperometry. Figure 2A shows a simultaneous measurement of [Ca²⁺]_i and amperometry signal from a spindle-shaped cell. During the two ATP (100 μm) challenges, [Ca²⁺]_i rose to ∼1 μm, but the [Ca²⁺]_i elevations were not accompanied by any amperometry signal. Application of KCl (50 mm) to the same cell elicited neither [Ca²⁺]_i elevation nor amperometry signal. Subsequent application of ionomycin (1 μm) raised [Ca²⁺]_i to > 3 μm, saturating the indo-1 signal. However, even at such a high [Ca²⁺]_i, no amperometry signal was triggered. Similar results were obtained in four other spindle-shaped cells. In contrast, in the ovoid cells, KCl depolarization consistently triggered a robust [Ca²⁺]_i elevation (1–2 μm) that was accompanied by amperometry signals, indicating catecholamine release (Fig. 2B; n = 4). The lack of catecholamine release from the ATP-responsive, spindle-shaped cells suggests that they were indeed type II cells.

Figure 1. ATP triggered a transient [Ca²⁺]_i rise in the S-100 protein-expressing, spindle-shaped type II cells, but not in the ovoid type I cells.

A, bright-field picture showing an ovoid type I (glomus) cell and a spindle-shaped type II (sustentacular) cell. The longer axis of the ovoid type I cell is 9 μm. The two morphologically distinct cell types were obtained from the same cell culture. B, [Ca²⁺]_i measurement of the two cells shown in A when challenged with ATP (100 μm). The cells were loaded with fura-2 AM and [Ca²⁺]_i was monitored with digital imaging. C, bright-field picture showing the localization of S-100 protein immunoreactivity in a spindle-shaped cell. The cell on the left was immunostained with antibody to S-100 protein. With the omission of the antibody to S-100 protein, the cell shown on the right did not exhibit any immunostaining. The shorter axis of the spindle-shaped cell is ∼5.5 μm.

Figure 2. The ATP-responsive type II cells lacked catecholamine-containing granules.

A, simultaneous recording of [Ca²⁺]_i and carbon fibre amperometry from a spindle-shaped type II cell. The ATP-induced [Ca²⁺]_i elevations were not accompanied by any amperometry signal. Application of ionomycin (1 μm) elevated [Ca²⁺]_i to > 3 μm, but no amperometry signal was detected. Note that application of KCl (50 mm) failed to elicit any [Ca²⁺]_i rise. B, simultaneous recording of [Ca²⁺]_i and amperometry from an ovoid type I cell. ATP challenge failed to elicit any [Ca²⁺]_i rise, but KCl depolarization raised [Ca²⁺]_i to ∼1.5 μm, a response that was accompanied by a large number of amperometric spikes, reflecting quantal release of catecholamines. In both A and B, [Ca²⁺]_i was measured with indo-1 fluorometry.

The absence of voltage-gated Ca²⁺ channels in type II cells

Previous studies in rabbit carotid bodies have shown that type II cells lack voltage-gated Na⁺ and Ca²⁺ currents (Duchen et al. 1988; Ureña et al. 1989). The lack of a Ca²⁺ response to KCl depolarization in the ATP-responsive spindle-shaped cells in Fig. 2A suggests that these cells have no significant expression of voltage-gated Ca²⁺ channels. To examine whether the spindle-shaped cells possess any voltage-gated Na⁺ or Ca²⁺ current, single spindle-shaped cells were voltage clamped at −70 mV and depolarized for 50 ms to different potentials (Fig. 3A). In this experiment, outward K⁺ currents were blocked by replacing K⁺ in the pipette solution with Cs⁺ and TEA. Note that at all potentials examined (−60 to +70 mV), no inward current was elicited in the spindle-shaped cells (n = 5), suggesting that both voltage-gated Na⁺ and Ca²⁺ currents were absent in these cells. In contrast, a family of slow inward currents, resembling voltage-gated Ca²⁺ currents, was consistently triggered in the ovoid cells (Fig. 3B; n = 4). Figure 3C shows that K⁺ currents were present in the spindle-shaped cells (n = 3). When the spindle-shaped cells were recorded with standard K⁺-containing pipette solution, depolarization from −70 mV to potentials more positive than −20 mV elicited a family of outward currents that exhibited little time dependence during the 50 ms voltage steps. Outward currents elicited from the ovoid cells under the same conditions are shown in Fig. 3D (n = 4). Note that the outward currents in the ovoid cell have slow activation kinetics and the amplitudes of the outward current were several-fold larger than those elicited from the spindle-shaped cells. These results indicate that the spindle-shaped cells and the ovoid cells of the rat carotid bodies possess a different repertoire of voltage-gated channels. Most importantly, voltage-gated Ca²⁺ channels are absent in the spindle-shaped cells.

Figure 3. Absence of voltage-gated Ca²⁺ current in type II cells.

A, depolarizations failed to elicit any inward current in a type II cell. The cell was held at −70 mV and voltage stepped to different potentials (−70 to +70 mV in 10 mV increments) for 50 ms every 200 ms. The traces shown here are superimposed records of current elicited from depolarization to −40, −20, 0, 20 and 40 mV. The whole-cell pipette solution contained Cs⁺ and TEA to block outward K⁺ currents. B, depolarizations elicited a family of slow inward currents in a type I cell. The voltage protocol and pipette solution were identical to those used in A. C, type II cells possessed an outward K⁺ current. The cell was recorded with a standard pipette solution that included 10 mm EGTA. The voltage protocol was identical to that used in A. The traces shown here are superimposed records of current elicited during depolarization to −10, 10, 30, 50 and 70 mV. Note that the current exhibited little time dependence during the voltage step. D, family of outward currents elicited in a type I cell. Note that the currents had slow activation kinetics and the current amplitudes were severalfold larger than those elicited from type II cells. Same protocol as in C.

The ATP response is mediated via P2Y receptors

ATP may exert its effect via binding to P1 (adenosine) or P2 (purine) receptors. Figure 4A shows that the ATP response involved P2 receptors. In this experiment, the cell was first challenged with ATP to elicit Ca²⁺ response. The cell was then exposed to Rb-2 (100 μm), a non-selective inhibitor of P2 receptors (Ralevic & Burnstock, 1998). In the continued presence of Rb-2, a second ATP challenge failed to elicit any Ca²⁺ signal. This was in contrast with control experiments (e.g. Figs 2A and 4B), where a second ATP challenge consistently elicited a Ca²⁺ signal. For the cell shown in Fig. 4A, following the removal of Rb-2, a subsequent ATP challenge was still able to elicit a Ca²⁺ response. Similar experiments were conducted in 10 cells. Rb-2 completely inhibited the ATP response in six cells and dramatically reduced the ATP-induced Ca²⁺ signals in the remaining four cells.

Figure 4. The ATP response did not involve P2X receptors.

A, the ATP response observed in type II cells was inhibited by Rb-2 (100 μm), a P2 receptor antagonist, implicating the involvement of P2 receptors. Following the removal of Rb-2, the ATP response was partially restored. B, the ATP-induced Ca²⁺ signal in type II cells did not involve Ca²⁺ entry via voltage-gated Ca²⁺ channels. ATP could elicit a robust Ca²⁺ signal in the continued presence of 3 mm Ni²⁺, a blocker of voltage-gated Ca²⁺ channels. C, the ATP-induced increase in [Ca²⁺]_i was not accompanied by any activation of inward current. The two traces are a simultaneous recording of current and [Ca²⁺]_i from a type II cell that was whole-cell voltage clamped at −70 mV. [Ca²⁺]_i was monitored with indo-1 fluorometry.

P2 purinoreceptors can be divided into two main classes: P2X and P2Y (reviewed by Ralevic & Burnstock, 1998). P2X receptors belong to a family of direct ligand-gated cation channels, and P2Y receptors are coupled to G proteins and phospholipase C activation. In cells with voltage-gated Ca²⁺ channels, activation of P2X receptors can lead to depolarization and Ca²⁺ entry via voltage-gated Ca²⁺ channels. In the absence of voltage-gated Ca²⁺ channels, Ca²⁺ entry can still occur via the opening of Ca²⁺-permeable ligand-gated Ca²⁺ channels. The lack of voltage-gated Ca²⁺ channels in the spindle-shaped type II cells (Fig. 3A) suggests that the ATP response could not involve activation of voltage-gated Ca²⁺ channels. Consistent with this, Fig. 4B shows that in the continued presence of Ni²⁺ (3 mm), a blocker of voltage-gated Ca²⁺ channels, repetitive challenges of ATP (100 μm) could still elicit robust Ca²⁺ responses in type II cells. For nine type II cells examined in the presence of Ni²⁺, the peak [Ca²⁺]_i rises elicited by the two consecutive ATP (100 μm) challenges (∼400–500 s apart) were 536 ± 98 and 414 ± 69 nm. Note that in the presence of Ni²⁺, the peak Ca²⁺ response to the first ATP challenge was similar to that of the control cells (524 ± 51 nm; n = 47).

To determine whether ATP elicited Ca²⁺-permeable ligand-gated current in type II cells, we simultaneously monitored the membrane current and [Ca²⁺]_i of type II cells during an ATP challenge. An example of this experiment is shown in Fig. 4C. The cell was held at −70 mV. After establishment of the whole-cell configuration, there was a small and gradual increase in the leak current. Application of ATP elicited a robust rise in [Ca²⁺]_i, but this was not accompanied by any increase in inward current (n = 3). This result indicates that P2X receptors are not involved in the ATP response.

We then examined whether the presence of extracellular Ca²⁺ was essential for the ATP response. In this experiment, the standard bath solution was replaced by a Ca²⁺-free solution (containing 1 mm EGTA). Figure 5 shows that in the absence of extracellular Ca²⁺, the first ATP challenge could still elicit a robust Ca²⁺ response. In 15 type II cells examined in Ca²⁺-free extracellular solution, the peak [Ca²⁺]_i during the first ATP challenge was 438 ± 57 nm, slightly smaller than that of the control cells (524 ± 51 nm; n = 47). The robust Ca²⁺ response in the absence of extracellular Ca²⁺ indicates that Ca²⁺ release from the intracellular stores is the primary mechanism underlying the ATP response. Note, however, that for the cell shown in Fig. 5A, a second ATP challenge (∼300 s later) in the continued absence of extracellular Ca²⁺ failed to elicit any [Ca²⁺]_i rise. In 15 cells examined, six cells failed to generate any [Ca²⁺]_i rise during the second ATP challenge. In the remaining nine cells (e.g. Fig. 5B), the second ATP challenge elicited a much smaller Ca²⁺ response (150 ± 37 nm). The much-diminished Ca²⁺ response to the second ATP challenge was observed only in the absence of extracellular Ca²⁺, but not in the presence of Ni²⁺ (Fig. 5A) or in the control condition (Fig. 2A). Thus, it cannot be due to a decrease in sensitivity of the receptors to ATP. Instead, this result suggests that extracellular Ca²⁺ influx is required to replenish the intracellular stores after one ATP challenge.

Figure 5. The ATP response involved intracellular Ca²⁺ release.

Cells were bathed in Ca²⁺-free bath solution that contained 1 mm of the Ca²⁺ chelator, EGTA. Two ATP challenges were applied ∼400 s apart. A, example of a type II cell in which only the first ATP challenge elicited a rise in [Ca²⁺]_i. B, example of a type II cell in which the second ATP challenge elicited a small Ca²⁺ response.

The release of Ca²⁺ from intracellular stores during an ATP challenge (Fig. 5) suggests that P2Y receptors are probably involved. To further identify the purinoreceptors involved, we examined the potency of different purinoreceptor agonists in triggering the Ca²⁺ response. In this series of experiments, type II cells were exposed to three different concentrations (1, 10 or 100 μm) of purinoreceptor agonists. Figure 6A shows the response of a type II cell when challenged sequentially with different concentrations of ATP. In this example, application of 10 or 100 μm ATP triggered a robust Ca²⁺ response, but ATP at 1 μm failed to elicit any increase in [Ca²⁺]_i. In 30 cells examined, none responded to 1 μm ATP; 11 cells exhibited a [Ca²⁺]_i rise when challenged with 10 μm ATP, and every one of the 30 cells responded to 100 μm ATP. Similar experiments were repeated with UTP (Fig. 6B), 2MeSATP (Fig. 6C) and α,β-meATP (Fig. 6D). In each of these experiments, ATP (100 μm) was applied at the end of the experiment to confirm that each cell could respond to ATP. As shown in Fig. 6B, UTP was a potent agonist in triggering a Ca²⁺ response in type II cells. In 19 cells examined, four cells exhibited a small Ca²⁺ response at 1 μm UTP. When challenged with 10 μm UTP, 16 out of 19 cells responded with a robust Ca²⁺ signal and all 19 cells responded to 100 μm UTP. In contrast, for the 15 cells challenged with 2MeSATP, none exhibited any Ca²⁺ response to either 1 or 10 μm 2MeSATP, and only eight cells responded to 100 μm 2MeSATP (e.g. Fig. 6C). Figure 6D shows that α,β-meATP was a poor agonist in triggering the Ca²⁺ response. At 10 μm (n = 8) or 100 μm (n = 16), α,β-meATP failed to elicit any Ca²⁺ response, but a subsequent challenge with ATP (100 μm) to the same cell triggered a robust Ca²⁺ response. Thus, the order of purinoreceptor agonist potency for eliciting Ca²⁺ signal in type II cells was UTP > ATP > 2MeSATP ≫ α,β-meATP. This result suggests that the ATP-induced Ca²⁺ signals in type II cells involve P2Y₂ receptors.

Figure 6. Actions of different purinoreceptor agonists.

Ca²⁺ response to different concentrations of ATP (A), UTP (B), 2MeSATP (C) and α,β-meATP (D). In B–D, ATP (100 μm) was applied at the end of each experiment to confirm that the same cell was responsive to ATP.

Immunolocalization of P2Y receptors on rat carotid body sections

To determine the localization of the ATP-responsive cells in the intact carotid body, cryostat sections of rat carotid body were immunostained with an antibody against P2Y₂ receptors. The confocal fluorescence image shown in Fig. 7 shows that P2Y₂ receptors were localized on the spindle-shaped type II cells but not on the ovoid type I cells. This is consistent with the lack of ATP response from the type I cells (Figs 1 and 2). Note that the spindle-shaped type II cells partially envelop clusters of ovoid type I cells.

Figure 7. Localization of P2Y₂ receptors on spindle-shaped type II cells from the carotid body.

Confocal image of a carotid body section immunostained with a fluorescently labelled antibody against P2Y₂. Note that the fluorescence is localized in spindle-shaped cells that partially wrap around clusters of ovoid type I cells. In control experiments where the primary antibody against P2Y₂ was omitted, the ovoid type I cells had slightly higher background fluorescence than the type II cells, probably as a result of their stored catecholamines. In this figure, the background level of type I cells was set to almost black. A white line has been drawn around the margin of individual type I cells that could be identified from a comparison between the fluorescence image and the bright-field image. The horizontal length of the field is 50 μm.

DISCUSSION

Although the identity of the neurotransmitter(s) responsible for neurotransmission in carotid body remains controversial, it is generally accepted that the release of transmitter(s) from type I cells is the major mechanism underlying the stimulation of the carotid sinus nerve. Among the various neurotransmitters present in type I cells, dopamine and ACh have been implicated as the major neurotransmitters (reviewed by López-Barneo, 1996; Fitzgerald, 2000). Recently, ATP has also been suggested to have an important role in hypoxic signalling (Zhang et al. 2000). Perfusion of ATP into the cat carotid body was shown to stimulate chemosensory discharge (Spergel & Lahiri, 1993). In co-cultures of rat type I cells and petrosal neurons, hypoxic chemotransmission was partially inhibited by a blocker of P2X receptors, suggesting that ATP released from type I cells during hypoxia stimulates chemosensory discharge via P2X receptors on the nerve terminals (Zhang et al. 2000). The present findings demonstrate that ATP also stimulates type II cells in the carotid body. The identification of the ATP-responsive cells as type II cells was based on cell morphology (spindle shaped), immunoreactivity for S-100 protein, and the lack of catecholamine secretion from these cells (Figs 1 and 2). In contrast to the action of ATP on nerve terminals, the ATP response in type II cells was not mediated via P2X receptors, as the ATP-induced [Ca²⁺]_i rise was not accompanied by activation of any inward current (Fig. 4C). Among the different purinoreceptor agonists, α,β-meATP was inactive (Fig. 6D) and 2MeSATP was weak (Fig. 6C), in triggering Ca²⁺ signal in type II cells. On the other hand, UTP was slightly more potent than ATP (Fig. 6B). This agonist potency profile is consistent with that for the activation of the P2Y₂ receptor subtypes (Ralevic & Burnstock, 1998). The major signal transduction pathway for P2Y receptors is activation of phosopholipase C, which in turn leads to the generation of inositol trisphosphate (IP₃) and Ca²⁺ release from IP₃-sensitive stores (Ralevic & Burnstock, 1998). Consistent with this, ATP could elicit at least one robust transient [Ca²⁺]_i rise in type II cells bathed in Ca²⁺-free extracellular solution (Fig. 5). However, in the continued absence of extracellular Ca²⁺, the ATP-induced Ca²⁺ response diminished with a subsequent ATP challenge. This observation suggests that the first ATP challenge (100 μm) empties most of the intracellular Ca²⁺ store in type II cells, and replenishment of intracellular Ca²⁺ stores requires the presence of extracellular Ca²⁺.

Type II cells are traditionally viewed as glial-like passive supporting structure that are wrapped around clusters of type I cells (Grönblad, 1983; Kameda, 1996). Our current findings suggest that type II cells play an active role in hypoxic signalling in the carotid body. During hypoxia, [Ca²⁺]_i in type I cells is elevated and multiple neurotransmitters, including ATP, are released. ATP in turn activates the P2Y₂ receptors on the neighbouring type II cells (Fig. 7), leading to an increase in [Ca²⁺]_i. The function of the ATP-induced Ca²⁺ signal in type II cells is obscure. However, based on the findings in glial cells, several speculations can be made here. Recent studies have shown that glial Ca²⁺ signalling may be a potent modulator of synaptic transmission (Araque et al. 1998; Kang et al. 1998). Since type II cells in the carotid body are also in close contact with many nerve endings, it is possible that the Ca²⁺ signal in type II cells may modulate chemosensory transmission by acting on the nerve endings. Multiple mechanisms have been postulated to underlie the interaction between glial cells and neurons, including the formation of gap junctions and the release of neuroactive substances from glial cells (reviewed by Verkhratsky et al. 1998). Interestingly, a recent study suggests that ATP release from glial cells may be responsible for the propagation of Ca²⁺ signals from glia to neurons (Cotrina et al. 2000). If Ca²⁺-dependent release of ATP also occurs in the type II cells of the carotid body, the ATP released may act on P2X receptors on the nerve endings to stimulate chemosensory discharge. On the other hand, Ca²⁺ signals have been shown to affect K⁺ channels in the glial cells, and thus alter extracellular [K⁺] and the ionic buffering properties of the glial syncytium (reviewed by Verkhratsky et al. 1998). Since type II cells in carotid bodies possess K⁺ channels (Fig. 3C) and these cells envelop type I cells as well as many nerve endings, a change in their K⁺ buffering property may modulate the electrical properties of both structures.

In summary, although type II cells are electrically non-excitable, they can generate Ca²⁺ signals in response to neurotransmitters, such as ATP. The Ca²⁺ signalling in type II cells may in turn interact with nerve endings or type I cells to co-ordinate chemosensory transmission in the carotid body.