- Email: [email protected]
Electrophysiological and immunocytological demonstration of cell-type specific responses to hypoxia in the adult cat carotid body
Brain Research 789 Ž1998. 229–238
Research report
Electrophysiological and immunocytological demonstration of cell-type specific responses to hypoxia in the adult cat carotid body Chung-Long Chou, James S.K. Sham, Brian Schofield, Machiko Shirahata
)
Department of EnÕironmental Health Sciences, School of Hygiene and Public Health, The Johns Hopkins UniÕersity, Baltimore, MD 21205, USA Accepted 25 November 1997
Abstract We have recently shown two types of cat carotid body cells based on the oxygen sensitivity of voltage-gated potassium channels. In the present study, we attempted to determine the correlation between cell types Žglomus cells, sheath cells, and subtypes of glomus cells. and oxygen sensitivity of potassium channels. Further, changes in membrane potentials in response to hypoxia were also examined. Carotid body cells harvested from adult cats were cultured, and a whole cell patch clamp method was applied to determine the oxygen sensitivity of outward current. The tested cells were identified by Lucifer Yellow in the patch pipette. Glomus cells and sheath cells were immunocytochemically identified using tyrosine hydroxylase ŽTH. and glial fibrillary acidic protein ŽGFAP. as markers. The cells whose outward current was inhibited by hypoxia showed TH-immunoreactivity but not GFAP-immunoreactivity. The cells whose outward current was not sensitive to hypoxia were GFAP-positive or TH-negative. One TH-positive cell had oxygen-insensitive outward current. The resting membrane potentials of the cells having oxygen-sensitive outward current were significantly higher Žy55 " 3 mV. than those of the cells having oxygen-insensitive outward current Žy35 " 2 mV.. The former type of cells was depolarized during hypoxia, but not the latter type of cells. These results suggest that most glomus cells of the adult cat carotid body possess oxygen-sensitive potassium channels and are depolarized in response to hypoxia. On the other hand, sheath cells and possibly a small fraction of glomus cells possess oxygen-insensitive potassium channels and their membrane potential is not affected by hypoxia. q 1998 Elsevier Science B.V. Keywords: Chemoreceptor; Glial fibrillary acidic protein; Oxygen; Patch clamp; Potassium channel; Tyrosine hydroxylase
1. Introduction The carotid body is a major arterial chemoreceptor, whose sensitivity to oxygen tension plays an essential role in systemic responses to hypoxia w9,15x. A current hypothesis of carotid body chemoreception implies that inhibition of oxygen-sensitive potassium channels by hypoxia depolarizes glomus cells. Depolarization activates voltage-gated calcium channels and triggers neurotransmitter release w17,24,34x. The presence of voltage-gated oxygen-sensitive potassium channels in carotid body cells has been shown in the rabbit, rat and cat Žw8,10,20,25,26,33,38x; for reviews, see Refs. w24,34x.. In agreement with the hypothesis
) Corresponding author. Division of Physiology, Department of Environmental Health Sciences, School of Hygiene and Public Health, The Johns Hopkins University, 615 N. Wolfe Street, Baltimore, M D 2 1 2 0 5 , U S A . F a x : q 1 -4 1 0 -9 5 5 -0 2 9 9 ; E -m a il: [email protected]
0006-8993r98r$19.00 q 1998 Elsevier Science B.V. All rights reserved. PII S 0 0 0 6 - 8 9 9 3 Ž 9 7 . 0 1 4 7 2 - 8
described above depolarization of carotid body cells during hypoxia was observed in the fetal rabbit w10x. In the adult rabbit glomus cells generated action potentials and the frequencies of action potentials increased during hypoxia w25x. However, there were some inconsistent observations. Biscoe and Duchen w3x reported that carotid body cells of the adult rabbit were hyperpolarized and that outward potassium current was increased, instead of decreased, during hypoxia. Other studies using microelectrodes did not show any consistent changes in membrane potentials during hypoxia w14,31x. The inconsistency may be due to differences in species, age, or experimental procedures. It is also possible that subtypes of glomus cells respond differently. Earlier morphological studies distinguished several types of glomus cells in rats, cats, and human w19,28x. Recent electrophysiological and fluorometric studies have indicated heterogeneous responses of glomus cells to hypoxia and cyanide w4,11,35x. Our previous study has shown that one type of cultured cat carotid body cell has oxygen-sensitive potassium chan-
230
C.-L. Chou et al.r Brain Research 789 (1998) 229–238
nels and the other has oxygen-insensitive potassium channels w8x. The electrophysiological properties of these two types of potassium channels were very similar, and we could not detect any clear morphological differences between the cells which differentially expressed these channels. Questions are whether these cells consist of subtypes of glomus cells, or whether glomus cells and sheath cells differentially express different potassium channels. This study was designed to clarify these questions. We further examined the correlation between the oxygen sensitivity of potassium channels and changes in the membrane potential during hypoxia. A preliminary study was reported previously w7x.
2. Materials and methods 2.1. Cell culture Carotid body cells were cultured as previously reported with a slight modification w36x. In short, carotid bodies were harvested from adult cats that were deeply anesthetized with pentobarbital Ž30–40 mgrkg, i.p., then 50 mgrkg, i.v.. and decapitated. Both carotid bodies were cleaned, and dissociated with collagenase Ž0.1–0.2%, type XI, Sigma. and gentle trituration. The cells were then centrifuged Ž100–200 = g, 5 min., the pellet was resuspended in a culture medium, and the cell suspension was
Fig. 1. Identification of a patch clamped cell using Lucifer Yellow and immunocytochemistry. ŽA. Outward current recorded during a single voltage clamp pulse from y60 mV to q40 mV. The cell was superfused with normocapnic normoxic Krebs solution Ž‘Control’ and ‘Recovery’, PO 2 s 150 mmHg. or normocapnic hypoxic Krebs Ž‘Hypoxia’, PO 2 s 25 mmHg.. ŽB. Fluorescence photomicrograph showing Lucifer Yellow in the studied cell Žarrow.. Bar s 20 m m. ŽC. Phase contrast micrography. An arrow indicates the tested cell. ŽD. Fluorescence photomicrograph after immunofluorescent staining for tyrosine hydroxylase ŽTH.. The tested cell Žarrow. was immunoreactive. ŽB. – ŽD. show the same microscopic field.
C.-L. Chou et al.r Brain Research 789 (1998) 229–238 Table 1 Summary of immunoreactivity of carotid body cells to tyrosine hydroxylase ŽTH. or glial fibrillary acidic protein ŽGFAP. Outward currents
Oxygen-sensitive Oxygen-insensitive
TH
GFAP
q
y
q
y
8 1
0 5
0 5
4 0
After testing oxygen sensitivity of voltage-gated outward current, each cell was immunostained for TH or GFAP.
divided into 10–15 wells. Each well consisted of a glass coverslip Ž25 mm in diameter. on which a 5 mm diameter plastic cloning cylinder ŽResearch Product International. was placed with a silicon glue. The coverslip was previously coated with poly-D-lysine ŽSigma. and diluted basement membrane complex ŽMatrigele, Collaborative Biomedical Products; 1r10 dilution in Dulbecco’s modified Eagle’s medium.. These substrates were washed with a culture medium before plating cells. The well was placed in a regular 35 mm culture dish, and the cells were cultured in a 5% CO 2rair incubator at 378C. The culture medium consisted of a 1:1 mixture of Dulbecco’s modified Eagle’s medium and Ham’s F12 medium supplemented with fatty acid-free bovine serum albumin Ž500 m grml., bovine transferrin Ž20 m grml.,
231
bovine insulin Ž10 m grml., sodium selenite Ž5 ngrml., Ž20 m grml. and 7s-nerve growth factor Ž100 ngrml.. These substances were obtained from Sigma or Life Technologies. The cells were cultured up to 2 weeks, and all experiments were performed between the third and fourteenth day in culture.
L-glutamine
2.2. Patch clamp methods A standard patch clamp technique with a whole-cell configuration was applied w27x. Patch micropipettes were fabricated from a soda–lime ‘microhematocrit’ capillary tube and fire polished. The resistances of micropipettes were between 2 and 5 M V when filled with internal solution composed of Žin mM. K glutamate 90, KCl 33, NaCl 10, EGTA 10, MgATP 5, CaCl 2 1, HEPES 10, pH s 7.2. For the first set of experiments Žsee Section 2.4. Lucifer Yellow Žpotassium salt, 0.1%; Sigma. was included in the solution to label a tested cell. All experiments were performed at room temperature Ž22–248C.. A glass coverslip with cultured cells was assembled to a specially designed recording chamber w8x. The chamber was mounted on an inverted phase contrast microscope ŽOlympus IMT2., and the cells were continuously perfused with modified Krebs solution which con-
Fig. 2. A cell which had oxygen-sensitive outward current was not immunostained for glial fibrillary acidic protein ŽGFAP.. ŽA. Outward current evoked by a single voltage clamp pulse from y60 mV to q40 mV. The ‘Control’, ‘Hypoxia’, and ‘Recovery’ conditions were the same as in Fig. 1. ŽB. Fluorescent appearance of a cell labeled with Lucifer Yellow. Bar s 20 m m. ŽC. Phase contrast micrograph of the tested cell Žarrow.. ŽD. Position of the same cell Žarrow. after immunostaining for GFAP. Many other cells but not the tested cell showed fluorescent labeling. ŽB. – ŽD. show the same microscopic field.
232
C.-L. Chou et al.r Brain Research 789 (1998) 229–238
tained Žin mM.: NaCl 120, KCl 5.4, CaCl 2 2.5, MgCl 2 1.0, NaHCO 3 20, and glucose 10, pH 7.4 Žequilibrated with 5% CO 2 .. Perfusion solution was delivered and withdrawn by a peristaltic pump ŽMinipuls 2, Gilson Medical Electronics.. The volume in the recording chamber was adjusted to approximately 0.65 ml. Oxygen tension in the recording chamber was controlled by switching normoxic and hypoxic Krebs solutions. These solutions were equilibrated with 5% CO 2rair or 5% CO 2rargon in two separate reservoirs. Either gas was delivered Ž20 mlrmin. to the top of the solution in the recording chamber to establish a gas barrier. Oxygen tension in the chamber was
continuously monitored by an oxygen sensor ŽMI-730 Oxygen Electrode, OM-4 oxygen meter, Microelectrodes.. Whole cell currents were induced by step changes of the membrane potential from a holding potential of y60 mV to q40 mV for 40 ms. Voltage clamp was accomplished using a List EPC-7 patch clamp amplifier ŽAdams and List Assoc... The membrane potential was measured without applying any currents to the pipette. A PC-compatible computer installed with pCLAMP software ŽAxon Instruments. was used for pulse generation and data acquisition. Junctional potential between an electrode and the bath solution and capacitative transients were electronically
Fig. 3. A cell which expressed oxygen-insensitive outward current contained glial fibrillary acidic protein ŽGFAP.. ŽA. Outward current evoked by a single voltage clamp pulse from y60 mV to q40 mV. The ‘Control’, ‘Hypoxia’, and ‘Recovery’ conditions were the same as in Fig. 1. ŽB. A single cell labeled with Lucifer Yellow. Bar s 20 m m. ŽC. Phase contrast micrograph showing the position of the tested cell. ŽD. Fluorescence micrograph after staining for GFAP. ŽB. – ŽD. show the same microscopic field and arrows indicate the tested cell.
C.-L. Chou et al.r Brain Research 789 (1998) 229–238
compensated. A leakage current subtraction program in pCLAMP software was applied assuming that a leak current is linear. 2.3. Immunocytochemistry An indirect immunofluorescent technique was used to demonstrate TH and GFAP in the cytoplasm. After voltage clamp experiment which determined the hypoxic sensitivity of outward current of a tested cell, the cells were fixed and permeabilized by perfusion of absolute methanol for 10 min. The patch pipette was gently removed and the cells were washed by perfusing with 0.1 M phosphate buffered saline ŽPBS.. Ten minutes later the perfusate was changed to a blocking solution Ž1.5% normal horse serum in PBS. to reduce nonspecific binding. Cells were incubated in this solution for 1 h. Subsequently, cells were incubated with monoclonal anti-TH antibody ŽBoehringer Mannheim, mouse IgG, dilution: 1:500. or anti-GFAP ŽBoehringer Mannheim, mouse IgG; 1:50. for 45 min. Cells were washed by perfusing with PBS for 15 min and incubated with Texas Red conjugated anti-mouse IgG ŽVector. for 1 h. Finally, the cells were washed with perfusion of PBS for 10 min and an epifluorescent image was viewed using a rhodamine filter.
233
2.4. Experimental protocol 2.4.1. Experimental set 1 Single cells or cells at the edge of the cluster were selected for this set of experiments. Although obtaining steady patch clamp recording was difficult in these cells, identification of these cells by immunofluorescent technique was easier than cells at the middle of the cluster. After obtaining a whole cell configuration the resting membrane potential was measured followed by a voltage clamp experiment which determined the presence of oxygen-sensitive potassium channels. Then, phase contrast micrographs and epifluorescent micrographs with a fluorescein filter were taken for later identification of the tested cell. Subsequently, immunocytochemistry was performed as described above and epifluorescent micrographs with a rhodamine filter were taken. By comparing the three types of images Žphase contrast, Lucifer Yellow staining and immunostaining., immunoreactivity of the tested cell to TH or GFAP was determined. 2.4.2. Experimental set 2 Cells were initially superfused with Krebs solution equilibrated with 5% CO 2 in air Žcontrol.. After obtaining whole cell currents of a cell, the amplifier was switched to
Fig. 4. Oxygen-insensitive outward current and photomicrographs of a cell which was not immunostained for tyrosine hydroxylase ŽTH.. ŽA. Outward current evoked by a single voltage clamp pulse from y60 mV to q40 mV. The ‘Control’, ‘Hypoxia’, and ‘Recovery’ conditions were the same as in Fig. 1. ŽB. Lucifer Yellow labeling of the tested cell Žarrow. in the periphery of a cluster. Bar s 20 m m. ŽC. Phase contrast micrograph showing the position of the tested cell Žarrow.. ŽD. Fluorescence microscopy showing immunostaining for TH. The tested cell Žarrow. did not show TH-immunoreactivity. ŽB. – ŽD. show the same microscopic field.
234
C.-L. Chou et al.r Brain Research 789 (1998) 229–238
the current clamp mode to record the membrane potential. Subsequently, the external solution was changed to Krebs solution equilibrated with 5% CO 2 in argon. When oxygen tension of the recording chamber reached 25 mmHg, voltage clamp was again applied and whole cell currents were recorded. Next, the amplifier was switched back to the current clamp mode and cells were re-exposed to control solution. When oxygen tension in the chamber reached 152 mmHg Žrecovery., the voltage clamp experiment was repeated.
sensitive outward current. Five cells whose outward current was not sensitive to oxygen had a strong GFAP signal. Six other cells with oxygen-insensitive outward current were examined for TH-immunoreactivity, and five did not show TH-immunoreactivity ŽTable 1.. We observed dye coupling of Lucifer Yellow in eight of twelve cells which expressed oxygen-sensitive outward current. In Fig. 5, the cell indicated by an arrow was used for a patch clamp experiment. This cell and two adjacent cells showed Lucifer Yellow fluorescence, and they were
2.5. Statistical analysis Data are reported as mean " S.E.M. when applicable. To determine significant difference between means, Student’s t-test or analysis of variance combined with the Duncan’s multiple range test was used. Values were considered significantly different if p - 0.05.
3. Results 3.1. Experimental set 1 Fig. 1 represents electrophysiological and immunohistological results from a cell which had oxygen-sensitive outward current. The voltage clamp experiment showed that the outward current was reversibly inhibited by hypoxia ŽFig. 1A.. Lucifer Yellow contained in the patch pipette spread out in the cytoplasm and processes and, even after the patch pipette was removed, the patched cell was readily identified under a fluorescent microscope with a fluorescein filter ŽFig. 1B.. This concentration of Lucifer Yellow did not appear to affect the oxygen sensitivity of outward current. After application of anti-TH and Texas Red conjugated anti-mouse IgG, a few cells showed a bright fluorescence of Texas Red ŽFig. 1D., while some cells in the field did not fluoresce. By superimposing images of Lucifer Yellow, phase contrast, and Texas Red fluorescence, the tested cell was identified as TH positive. Eight cells with oxygen-sensitive outward current were tested for TH-immunoreactivity. All cells showed a strong TH-positive signal ŽTable 1.. In separate experiments, cells which expressed oxygen-sensitive outward current were immunostained for GFAP ŽFig. 2.. Lucifer Yellow fluorescence clearly identified the single cell at an edge of a cluster from which oxygen-sensitive outward current was recorded ŽFig. 2A.. Many cells in the cluster were immunoreactive to GFAP, but the patched cell was GFAPnegative ŽFig. 2D.. Similar results were obtained in four cells ŽTable 1.. In other experiments cells with oxygen-insensitive outward current were stained for either TH or GFAP. Fig. 3 shows an example of a GFAP-positive cell. On the other hand, Fig. 4 illustrates a TH-negative cell with oxygen-in-
Fig. 5. The dye coupling phenomenon in cultured carotid body cells. ŽA. Fluorescence photomicrograph showing Lucifer Yellow loading of the tested cell Žarrow. and two adjacent cells. Bar s 20 m m. ŽB. Appearance of carotid body cells by phase contrast microscopy. ŽC. Fluorescence photomicrograph showing TH-immunoreactive cells. The position of dye-coupled cells were indicated by a solid black line in ŽB. and ŽC..
C.-L. Chou et al.r Brain Research 789 (1998) 229–238
235
Fig. 6. Effects of hypoxia on the membrane potential of a cell with oxygen-sensitive outward current. The upper panel: outward currents activated by a voltage step from y60 mV to q40 mV. The middle panel: the changes in the oxygen tension in the recording chamber. I, II, and III denote the timings of voltage clamp experiments. The lower panel: membrane potential Ž Em . was measured using current clamp mode. Hypoxia depolarized the cell.
TH-positive. No cells whose outward current was oxygeninsensitive showed dye coupling of Lucifer Yellow. 3.2. Experimental set 2 Resting membrane potentials were measured in all cells in experimental set 2 and in some cells in experimental set 1. The resting membrane potentials of cells with oxygensensitive outward current were significantly higher than those of cells with oxygen-insensitive potassium current. The average of the resting membrane potentials of the former cells was y55 " 1 mV Ž n s 9., while the latter was y35 " 2 mV Ž n s 11.. These values were significantly different from each other Žunpaired t-test, p - 0.05.. The difference between these cells was also manifested in the changes in membrane potentials during hypoxia. Cells with oxygen-sensitive outward current were depolarized by hypoxia ŽFig. 6.. The decrease of PO 2 was accompanied by a gradual but significant depolarization Žfrom y55 " 3 mV to y27 " 5 mV; n s 4. and at this time the outward current was reduced. Upon re-exposure to the control solution the membrane potential recovered to control levels Žy48 " 7 mV., and the outward current also recovered. On the other hand, the membrane potentials of cells with oxygen-insensitive outward current were not affected by hypoxia. The average membrane potentials of
four cells were y39 " 5 mV, y34 " 8 mV, and y40 " 5 mV during control, hypoxia, and recovery periods, respectively. These values were not significantly different. Summarized data are presented in Fig. 7.
4. Discussion The present study has demonstrated that the cat carotid body cells whose outward current was inhibited by hypoxia were TH-positive or GFAP-negative. The cells whose outward currents were insensitive to hypoxia were GFAPpositive or TH-negative except one case. The resting membrane potentials of cells having oxygen-sensitive outward current were much higher than those of cells having oxygen-insensitive outward current. The former type of cell was depolarized during hypoxia, but not the latter type of cells. The results indicate that most glomus cells of the adult cat carotid body express oxygen-sensitive potassium channels, have high resting membrane potential, and are depolarized by hypoxia. On the other hand, sheath cells, and possibly a small fraction of glomus cells, are not sensitive to hypoxia in their outward current and membrane potential. Cellular components of the carotid body include glomus cells, sheath cells, fibroblasts, endothelial cells of capillar-
236
C.-L. Chou et al.r Brain Research 789 (1998) 229–238
Fig. 7. Changes in membrane potentials in response to hypoxia. The upper panel shows the response of the cells with oxygen-sensitive outward current. The lower panel is the response of the cells with oxygen-insensitive outward current. ), Significantly different from control and recovery Ž p- 0.05..
ies, and some ganglion cells, but glomus and sheath cells are predominant. Morphological differences between these two types of cells in fixed tissues have been noted even under light microscopy, but some characteristics overlap. This makes distinguishing cat glomus cells more troublesome than those of the rat or the rabbit, because the proportions of glomus cells and sheath cells were estimated as two to one in the cat, four to one in the rat, and seven to three in the rabbit wfor a review, see Ref. w28xx. In our experience identifying live cat glomus cells in culture by simple morphology is extremely difficult. We have been investigating specific methods by which the cell of interest is identified and distinguished. In this study we used Lucifer Yellow to mark the tested cell and applied immunocytochemistry to differentiate cell types. TH, a key enzyme for catecholamine synthesis, is expressed only in glomus cells, not in sheath cells, and more than 90% of glomus cells were TH-positive in the cat w40x. GFAP is a structural protein in astrocytes, and has been localized in sheath cells in the cat w1,21,37x and human w18x. Therefore, TH and GFAP are specific cell makers for glomus and sheath cells, respectively. By comparing images of Lucifer Yellow, phase contrast, and immunofluorescence, identifying the tested cell is relatively straightforward in single
cells or in cells at the edge of a cluster as seen in Figs. 1–4. We recorded whole currents in this study and did not use any specific blockers to identify responsible ionic species for the currents. However, in a recent report, we have shown that the net outward current of cultured carotid body cells was mostly through voltage-gated potassium channels w8x. Since the culture and patch clamp methods of this study were the same as the previous study, and since characteristics of the current Žthe activation time, the size, the lack of apparent inactivation, etc.. were not different from those of the previous study, we have reasonably assumed that the outward current was evoked by activation of voltage-gated potassium channels. Nevertheless, a possibility that the inward calcium current was hidden by the large outward potassium current cannot be totally ruled out. In the present experiments, oxygen-sensitive potassium channels were found only in the cells which contained TH. No cells which had GFAP immunoreactivity expressed oxygen-sensitive potassium channels. Furthermore, all cells that expressed oxygen-sensitive potassium channels did not contain GFAP signals. The results indicate that oxygen-sensitive potassium channels are localized only in glomus cells and not in sheath cells. On the other hand, all cells that stained for GFAP had voltage-gated and oxygen-insensitive potassium channels. Four of five cells which did not show TH-immunoreactivity expressed voltage-gated oxygen-insensitive potassium channels. Therefore, it is reasonable to conclude that sheath cells express voltage-gated potassium channels which are not oxygensensitive. The results agree with previous studies indicating that glomus cells possess oxygen-sensitive potassium channels and sheath cells express small voltage-gated oxygen-insensitive potassium channels in the adult rabbit w13,39x. Compared with the adult rabbit carotid body, voltage-gated oxygen-insensitive potassium channels in sheath cells of the cat carotid body had a lower threshold and a larger current. Interestingly, one cell which did not have oxygen-sensitive potassium channels showed positive immunoreactivity to TH ŽTable 1.. One possibility is that this TH-positive cell was a ganglion cell in the carotid body. However, this possibility is remote, because the size of the cell was not large enough to be a ganglion cell. Rather, this may indicate the presence of electrophysiological subtypes of glomus cells in the cat. Morphologically as many as four subtypes of glomus cells have been reported in the cat according to the size of the dense-cored vesicles and their histochemical reaction w30x. Subtypes of glomus cells were also reported in the rat and human wfor reviews, see Refs. w19,28xx. In addition to the morphological subtypes, heterogeneous physiological responses of glomus cells have been also reported. Donnelly w11x found two distinct subtypes of glomus cells in the adult rat carotid body: one with a small-voltage dependent inward current and a large out-
C.-L. Chou et al.r Brain Research 789 (1998) 229–238
ward current, the other with little voltage-dependent current. Bright et al. w4x observed that changes in intracellular calcium in response to hypoxia varied among glomus cells in the rat. Further, Perez-Garcia et al. w35x noted that cells in some cultures did not show ‘the typical inhibition of the potassium currents’, suggesting subtypes of glomus cells in the adult rabbit. Cells with oxygen-sensitive potassium channels were depolarized during hypoxia, whereas cells with oxygen-insensitive voltage-gated potassium channels were not depolarized by hypoxia. Thus, these findings are consistent with the current hypothesis of hypoxic chemotransduction which implies that hypoxia depolarizes glomus cells possibly due to inhibition of oxygen-sensitive potassium channels. However, unlike adult rabbit glomus cells, adult cat glomus cells did not show any action potentials. This fact raises a question whether the size and the speed of depolarization of cat glomus cells are adequate to activate voltage-gated calcium channels for subsequent release of neurotransmitterŽs.. In smooth muscle cells sustained depolarization to y30 mV caused a steady-state elevation of intracellular calcium via activation of L-type of calcium channels w16x. It is not known whether this is true in glomus cells. Another important question is whether the depolarization of glomus cells is due to the inhibition of voltage-gated potassium channels. Simultaneous measurement of carotid sinus nerve activity and potassium current of glomus cells in the adult rat revealed the dissociation of the inhibition of potassium current and increase in nerve activity w6,12x. Further, Buckler w5x has recently found a new type of potassium channels which are active at rest and inhibited by hypoxia. These data suggest that the role of oxygen-sensitive potassium channels in depolarizing glomus cells may be modulatory. The presence of gap junctions in the rat and the cat has been suggested by McDonald w28x. He estimated two to three gap junctions per 100 glomus cells. Electrical coupling between glomus cells was also indicated w2,29x, and in these studies about 30% of cells showed electrical coupling. Recently, Kondo and Iwasa w23x reevaluated the occurrence of gap junctions in the rat carotid body. They reported that presumed gap junctions between glomus cells occurred frequently, and they were present even between glomus cells and sheath cells. In agreement with these studies the phenomenon of dye coupling appeared to be a common phenomenon in the cat in our study. Eight out of twelve cells which have oxygen-sensitive potassium channels showed dye coupling with other cells. Eleven cells with oxygen-insensitive potassium channels did not show any dye-coupling. Thus, intracellular connections may be most frequently found between glomus cells in the cat. For understanding the mechanisms of chemoreception in vivo, one important question is whether cultured glomus cells have the same characteristics as those in situ. Although cultured glomus cells have been widely used, very few studies have addressed this question w21,35x. Our
237
culture system employs Matrigele and nerve growth factor ŽNGF.. Matrigele contains several growth factors, but the possibility that the growth factors in Matrigele would change the expression of channels is remote. We have been using diluted Matrigele, which has been aspirated and washed before plating cells. NGF is well known to promote neuronal differentiation in neural crest-derived cells w32x. However, neurofilament ŽNF., a structural protein specific for neurons, was not detected in cultured cat carotid body cells w21x. Jackson and Nurse w22x have also reported that NGF failed to induce NF 68 kDa and NF 160 kDa in cultured postnatal rat glomus cells. The absence of NF in the cultured glomus cells suggest that NGF does not promote differentiation of glomus cells into neurons. Further, we did not observe any time-dependent changes in outward current during 2 weeks culture period, suggesting that the expression of voltage-gated potassium channels are stable for 2 weeks in our culture conditions. In summary, we have shown that cultured glomus cells of the adult cat carotid body possess oxygen-sensitive potassium channels and are depolarized in response to hypoxia. On the other hand, sheath cells and possibly a small fraction of glomus cells possess oxygen-insensitive potassium channels and their membrane potential is not affected by hypoxia. Although the data are consistent with the notion that glomus cells are the chemosensory cells, the role of oxygen-sensitive potassium channels in carotid body chemoreception in situ requires further investigation.
Acknowledgements The authors are grateful to Drs. Thomas Croxton and Bradly Undem for their helpful suggestions. This study was supported by HL47044 and HL50712.
References w1x A. Abramovici, D.J. Pallot, J.M. Polak, Immunohistochemical approach to the study of the cat carotid body, Acta Anat. 140 Ž1991. 70–74. w2x V. Abudara, C. Eyzaguirre, Electrical coupling between cultured glomus cells of the rat carotid body: observations with current and voltage clamping, Brain Res. 664 Ž1994. 257–265. w3x T.J. Biscoe, M.R. Duchen, Responses of type I cells dissociated from the rabbit carotid body to hypoxia, J. Physiol. 428 Ž1990. 39–59. w4x G.R. Bright, F.H. Agani, U. Haque, J.L. Overholt, N.R. Prabhakar, Heterogeneity in cytosolic calcium responses to hypoxia in carotid body cells, Brain Res. 706 Ž1996. 297–302. w5x K.J. Buckler, A novel oxygen-sensitive potassium current in rat carotid body type I cells, J. Physiol. 498 Ž1997. 649–662. w6x P.M. Cheng, D.F. Donnelly, Relationship between changes of glomus cell current and neural response of rat carotid body, J. Neurophysiol. 74 Ž1995. 2077–2086. w7x C.-L. Chou, B. Schofield, Y. Ishizawa, M. Shirahata, Differential
238
w8x w9x
w10x
w11x w12x w13x
w14x
w15x
w16x
w17x
w18x
w19x w20x
w21x
w22x
w23x
C.-L. Chou et al.r Brain Research 789 (1998) 229–238 expression of voltage-gated K channels in glomus cells and sheath cells of cat carotid body, FASEB J. 10 Ž1996. A406. C.-L. Chou, M. Shirahata, Two types of voltage-gated K channels in carotid body cells of adult cats, Brain Res. 742 Ž1996. 34–42. M. de B. Daly, Interactions between respiration and circulation, in: A.P. Fishman, N.S. Cherniack, J.G. Widdicombe ŽEds.., Handbook of Physiology, Section 3: The Respiratory System, American Physiological Society, Bethesda, MD, 1986, pp. 529–594. M.A. Delpiano, J. Hescheler, Evidence for a PO 2-sensitive Kq channel in the type I cell of the rabbit carotid body, FEBS Lett. 249 Ž1989. 195–198. D.F. Donnelly, Response to cyanide of two types of glomoid cells in mature rat carotid body, Brain Res. 630 Ž1993. 157–168. D.F. Donnelly, Modulation of glomus cell membrane currents of intact rat carotid body, J. Physiol. ŽLondon. 489 Ž1995. 677–688. M.R. Duchen, K.W. Caddy, G.C. Kirby, D.L. Patterson, J. Ponte, T.J. Biscoe, Biophysical studies of the cellular elements of the rabbit carotid body, Neuroscience 26 Ž1988. 291–311. C. Eyzaguirre, L. Monti-Bloch, M. Baron, Y. Hayashida, J.W. Woodbury, Changes in glomus cell membrane properties in response to stimulants and depressants of carotid nerve discharge, Brain Res. 477 Ž1989. 265–279. R.S. Fitzgerald, S. Lahiri, Reflex responses to chemoreceptor stimulation, in: A.P. Fishman, N.S. Cherniack, J.G. Widdicombe ŽEds.., Handbook of Physiology, Section 3: The Respiratory System, American Physiological Society, Bethesda, MD, 1986, pp. 313–362. B.K. Fleischmann, R.K. Murray, M.I. Kotlikoff, Voltage window for sustained elevation of cytosolic calcium in smooth muscle cells, Proc. Natl. Acad. Sci. U.S.A. 91 Ž1994. 11914–11918. C. Gonzalez, L. Almaraz, A. Obeso, R. Rigual, Carotid body chemoreceptors: from natural stimuli to sensory discharges, Physiol. Rev. 74 Ž1994. 829–898. J.O. Habeck, W. Kummer, Neuronal and neuroendocrine markers in the human carotid body in health and disease, Adv. Exp. Med. Biol. 337 Ž1993. 31–35. D. Heath, P. Smith, Diseases of the human carotid body, SpringerVerlag, London, 1992. J. Hescheler, M.A. Delpiano, H. Acker, Ionic currents on type I cells of the rabbit carotid body measured by voltage clamp experiments and the effect of hypoxia, Brain Res. 486 Ž1989. 79–88. Y. Ishizawa, B. Schofield, C.-L. Chou, M. Shirahata, Changes in protein expression of cultured cat carotid body cells. Book of Abstracts ŽISOTT 95—Satellite Symposium: Adaptation to Hypoxia. Ž1995. IV.3. A. Jackson, C. Nurse, Plasticity in cultured carotid body chemoreceptors: environmental modulation of GAP-43 and neurofilament, J. Neurobiol. 26 Ž1995. 485–496. H. Kondo, H. Iwasa, Re-examination of the carotid body ultrastructure with special attention to intracellular membrane appositions, in: P. Zapata, C. Eyzaguirre, R.W. Torrance ŽEds.., Frontiers in Arterial Chemoreception, Plenum, New York, 1996, pp. 45–50.
w24x J. Lopez-Barneo, Oxygen-sensing by ion channels and the regulation of cellular functions, Trends Neurosci. 19 Ž1996. 435–440. w25x J. Lopez-Barneo, J.R. Lopez-Lopez, J. Urena, C. Gonzalez, Chemotransduction in the carotid body: Kq current modulated by PO 2 in type I chemoreceptor cells, Science 241 Ž1988. 580–582. w26x J. Lopez-Lopez, C. Gonzalez, J. Urena, J. Lopez-Barneo, Low pO 2 selectively inhibits K channel activity in chemoreceptor cells of the mammalian carotid body, J. Gen. Physiol. 93 Ž1989. 1001–1015. w27x A. Marty, E. Neher, Tight-seal whole-cell recording, in: B. Sakmann, E. Neher ŽEds.., Single-Channel Recording, Plenum, New York, 1983, pp. 107–122. w28x D.M. McDonald, Peripheral chemoreceptors structure–function relationships of the carotid body, in: T.F. Hornbein ŽEd.., Regulation of Breathing, Marcel Dekker, New York, 1981, pp. 105–319. w29x L. Monti-Bloch, V. Abudara, C. Eyzaguirre, Electrical communication between glomus cells of the rat carotid body, Brain Res. 622 Ž1993. 119–131. w30x E. Morita, S.R. Chiocchio, J.H. Tramezzani, Four types of main cells in the carotid body of the cat, J. Ultrastruct. Res. 28 Ž1969. 399–410. w31x L. Pang, C. Eyzaguirre, Different effects of hypoxia on the membrane potential and input resistance of isolated and clustered carotid body glomus cells, Brain Res. 575 Ž1992. 167–173. w32x P.H. Patterson, Control of cell fate in a vertebrate neurogenic lineage, Cell 62 Ž1990. 1035–1038. w33x C. Peers, Hypoxic suppression of Kq currents in type I carotid body cells: selective effect on the Ca2q-activated Kq current, Neurosci. Lett. 119 Ž1990. 253–256. w34x C. Peers, K.J. Buckler, Transduction of chemostimuli by the type I carotid body cell, J. Membr. Biol. 144 Ž1995. 1–9. w35x M.T. Perez-Garcia, A. Obeso, J.R. Lopez-Lopez, B. Herreros, C. Gonzalez, Characterization of cultured chemoreceptor cells dissociated from adult rabbit carotid body, Am. J. Physiol. 263 Ž1992. C1152–C1159. w36x M. Shirahata, B. Schofield, B.Y. Chin, T.R. Guilarte, Culture of arterial chemoreceptor cells from adult cats in defined medium, Brain Res. 658 Ž1994. 60–66. w37x M. Shirahata, B. Schofield, R.S. Fitzgerald, Immunohistological differentiation of cultured cat carotid body cells, Soc. Neurosci. Abstr. 20 Ž1994. 957. w38x A. Stea, C.A. Nurse, Whole-cell and perforated-patch recordings from O 2 -sensitive rat carotid body cells grown in short- and long-term culture, Pflugers Arch. 418 Ž1991. 93–101. w39x J. Urena, J. Lopez-Lopez, C. Gonzalez, J. Lopez-Barneo, Ionic currents in dispersed chemoreceptor cells of the mammalian carotid body, J. Gen. Physiol. 93 Ž1989. 979–999. w40x Z.Z. Wang, L.J. Stensaas, B. Dinger, S.J. Fidone, Co-existence of tyrosine hydroxylase and dopamine b-hydroxylase immunoreactivity in glomus cells of the cat carotid body, J. Auton. Nerv. Syst. 32 Ž1991. 259–264.
Research report
Electrophysiological and immunocytological demonstration of cell-type specific responses to hypoxia in the adult cat carotid body Chung-Long Chou, James S.K. Sham, Brian Schofield, Machiko Shirahata
)
Department of EnÕironmental Health Sciences, School of Hygiene and Public Health, The Johns Hopkins UniÕersity, Baltimore, MD 21205, USA Accepted 25 November 1997
Abstract We have recently shown two types of cat carotid body cells based on the oxygen sensitivity of voltage-gated potassium channels. In the present study, we attempted to determine the correlation between cell types Žglomus cells, sheath cells, and subtypes of glomus cells. and oxygen sensitivity of potassium channels. Further, changes in membrane potentials in response to hypoxia were also examined. Carotid body cells harvested from adult cats were cultured, and a whole cell patch clamp method was applied to determine the oxygen sensitivity of outward current. The tested cells were identified by Lucifer Yellow in the patch pipette. Glomus cells and sheath cells were immunocytochemically identified using tyrosine hydroxylase ŽTH. and glial fibrillary acidic protein ŽGFAP. as markers. The cells whose outward current was inhibited by hypoxia showed TH-immunoreactivity but not GFAP-immunoreactivity. The cells whose outward current was not sensitive to hypoxia were GFAP-positive or TH-negative. One TH-positive cell had oxygen-insensitive outward current. The resting membrane potentials of the cells having oxygen-sensitive outward current were significantly higher Žy55 " 3 mV. than those of the cells having oxygen-insensitive outward current Žy35 " 2 mV.. The former type of cells was depolarized during hypoxia, but not the latter type of cells. These results suggest that most glomus cells of the adult cat carotid body possess oxygen-sensitive potassium channels and are depolarized in response to hypoxia. On the other hand, sheath cells and possibly a small fraction of glomus cells possess oxygen-insensitive potassium channels and their membrane potential is not affected by hypoxia. q 1998 Elsevier Science B.V. Keywords: Chemoreceptor; Glial fibrillary acidic protein; Oxygen; Patch clamp; Potassium channel; Tyrosine hydroxylase
1. Introduction The carotid body is a major arterial chemoreceptor, whose sensitivity to oxygen tension plays an essential role in systemic responses to hypoxia w9,15x. A current hypothesis of carotid body chemoreception implies that inhibition of oxygen-sensitive potassium channels by hypoxia depolarizes glomus cells. Depolarization activates voltage-gated calcium channels and triggers neurotransmitter release w17,24,34x. The presence of voltage-gated oxygen-sensitive potassium channels in carotid body cells has been shown in the rabbit, rat and cat Žw8,10,20,25,26,33,38x; for reviews, see Refs. w24,34x.. In agreement with the hypothesis
) Corresponding author. Division of Physiology, Department of Environmental Health Sciences, School of Hygiene and Public Health, The Johns Hopkins University, 615 N. Wolfe Street, Baltimore, M D 2 1 2 0 5 , U S A . F a x : q 1 -4 1 0 -9 5 5 -0 2 9 9 ; E -m a il: [email protected]
0006-8993r98r$19.00 q 1998 Elsevier Science B.V. All rights reserved. PII S 0 0 0 6 - 8 9 9 3 Ž 9 7 . 0 1 4 7 2 - 8
described above depolarization of carotid body cells during hypoxia was observed in the fetal rabbit w10x. In the adult rabbit glomus cells generated action potentials and the frequencies of action potentials increased during hypoxia w25x. However, there were some inconsistent observations. Biscoe and Duchen w3x reported that carotid body cells of the adult rabbit were hyperpolarized and that outward potassium current was increased, instead of decreased, during hypoxia. Other studies using microelectrodes did not show any consistent changes in membrane potentials during hypoxia w14,31x. The inconsistency may be due to differences in species, age, or experimental procedures. It is also possible that subtypes of glomus cells respond differently. Earlier morphological studies distinguished several types of glomus cells in rats, cats, and human w19,28x. Recent electrophysiological and fluorometric studies have indicated heterogeneous responses of glomus cells to hypoxia and cyanide w4,11,35x. Our previous study has shown that one type of cultured cat carotid body cell has oxygen-sensitive potassium chan-
230
C.-L. Chou et al.r Brain Research 789 (1998) 229–238
nels and the other has oxygen-insensitive potassium channels w8x. The electrophysiological properties of these two types of potassium channels were very similar, and we could not detect any clear morphological differences between the cells which differentially expressed these channels. Questions are whether these cells consist of subtypes of glomus cells, or whether glomus cells and sheath cells differentially express different potassium channels. This study was designed to clarify these questions. We further examined the correlation between the oxygen sensitivity of potassium channels and changes in the membrane potential during hypoxia. A preliminary study was reported previously w7x.
2. Materials and methods 2.1. Cell culture Carotid body cells were cultured as previously reported with a slight modification w36x. In short, carotid bodies were harvested from adult cats that were deeply anesthetized with pentobarbital Ž30–40 mgrkg, i.p., then 50 mgrkg, i.v.. and decapitated. Both carotid bodies were cleaned, and dissociated with collagenase Ž0.1–0.2%, type XI, Sigma. and gentle trituration. The cells were then centrifuged Ž100–200 = g, 5 min., the pellet was resuspended in a culture medium, and the cell suspension was
Fig. 1. Identification of a patch clamped cell using Lucifer Yellow and immunocytochemistry. ŽA. Outward current recorded during a single voltage clamp pulse from y60 mV to q40 mV. The cell was superfused with normocapnic normoxic Krebs solution Ž‘Control’ and ‘Recovery’, PO 2 s 150 mmHg. or normocapnic hypoxic Krebs Ž‘Hypoxia’, PO 2 s 25 mmHg.. ŽB. Fluorescence photomicrograph showing Lucifer Yellow in the studied cell Žarrow.. Bar s 20 m m. ŽC. Phase contrast micrography. An arrow indicates the tested cell. ŽD. Fluorescence photomicrograph after immunofluorescent staining for tyrosine hydroxylase ŽTH.. The tested cell Žarrow. was immunoreactive. ŽB. – ŽD. show the same microscopic field.
C.-L. Chou et al.r Brain Research 789 (1998) 229–238 Table 1 Summary of immunoreactivity of carotid body cells to tyrosine hydroxylase ŽTH. or glial fibrillary acidic protein ŽGFAP. Outward currents
Oxygen-sensitive Oxygen-insensitive
TH
GFAP
q
y
q
y
8 1
0 5
0 5
4 0
After testing oxygen sensitivity of voltage-gated outward current, each cell was immunostained for TH or GFAP.
divided into 10–15 wells. Each well consisted of a glass coverslip Ž25 mm in diameter. on which a 5 mm diameter plastic cloning cylinder ŽResearch Product International. was placed with a silicon glue. The coverslip was previously coated with poly-D-lysine ŽSigma. and diluted basement membrane complex ŽMatrigele, Collaborative Biomedical Products; 1r10 dilution in Dulbecco’s modified Eagle’s medium.. These substrates were washed with a culture medium before plating cells. The well was placed in a regular 35 mm culture dish, and the cells were cultured in a 5% CO 2rair incubator at 378C. The culture medium consisted of a 1:1 mixture of Dulbecco’s modified Eagle’s medium and Ham’s F12 medium supplemented with fatty acid-free bovine serum albumin Ž500 m grml., bovine transferrin Ž20 m grml.,
231
bovine insulin Ž10 m grml., sodium selenite Ž5 ngrml., Ž20 m grml. and 7s-nerve growth factor Ž100 ngrml.. These substances were obtained from Sigma or Life Technologies. The cells were cultured up to 2 weeks, and all experiments were performed between the third and fourteenth day in culture.
L-glutamine
2.2. Patch clamp methods A standard patch clamp technique with a whole-cell configuration was applied w27x. Patch micropipettes were fabricated from a soda–lime ‘microhematocrit’ capillary tube and fire polished. The resistances of micropipettes were between 2 and 5 M V when filled with internal solution composed of Žin mM. K glutamate 90, KCl 33, NaCl 10, EGTA 10, MgATP 5, CaCl 2 1, HEPES 10, pH s 7.2. For the first set of experiments Žsee Section 2.4. Lucifer Yellow Žpotassium salt, 0.1%; Sigma. was included in the solution to label a tested cell. All experiments were performed at room temperature Ž22–248C.. A glass coverslip with cultured cells was assembled to a specially designed recording chamber w8x. The chamber was mounted on an inverted phase contrast microscope ŽOlympus IMT2., and the cells were continuously perfused with modified Krebs solution which con-
Fig. 2. A cell which had oxygen-sensitive outward current was not immunostained for glial fibrillary acidic protein ŽGFAP.. ŽA. Outward current evoked by a single voltage clamp pulse from y60 mV to q40 mV. The ‘Control’, ‘Hypoxia’, and ‘Recovery’ conditions were the same as in Fig. 1. ŽB. Fluorescent appearance of a cell labeled with Lucifer Yellow. Bar s 20 m m. ŽC. Phase contrast micrograph of the tested cell Žarrow.. ŽD. Position of the same cell Žarrow. after immunostaining for GFAP. Many other cells but not the tested cell showed fluorescent labeling. ŽB. – ŽD. show the same microscopic field.
232
C.-L. Chou et al.r Brain Research 789 (1998) 229–238
tained Žin mM.: NaCl 120, KCl 5.4, CaCl 2 2.5, MgCl 2 1.0, NaHCO 3 20, and glucose 10, pH 7.4 Žequilibrated with 5% CO 2 .. Perfusion solution was delivered and withdrawn by a peristaltic pump ŽMinipuls 2, Gilson Medical Electronics.. The volume in the recording chamber was adjusted to approximately 0.65 ml. Oxygen tension in the recording chamber was controlled by switching normoxic and hypoxic Krebs solutions. These solutions were equilibrated with 5% CO 2rair or 5% CO 2rargon in two separate reservoirs. Either gas was delivered Ž20 mlrmin. to the top of the solution in the recording chamber to establish a gas barrier. Oxygen tension in the chamber was
continuously monitored by an oxygen sensor ŽMI-730 Oxygen Electrode, OM-4 oxygen meter, Microelectrodes.. Whole cell currents were induced by step changes of the membrane potential from a holding potential of y60 mV to q40 mV for 40 ms. Voltage clamp was accomplished using a List EPC-7 patch clamp amplifier ŽAdams and List Assoc... The membrane potential was measured without applying any currents to the pipette. A PC-compatible computer installed with pCLAMP software ŽAxon Instruments. was used for pulse generation and data acquisition. Junctional potential between an electrode and the bath solution and capacitative transients were electronically
Fig. 3. A cell which expressed oxygen-insensitive outward current contained glial fibrillary acidic protein ŽGFAP.. ŽA. Outward current evoked by a single voltage clamp pulse from y60 mV to q40 mV. The ‘Control’, ‘Hypoxia’, and ‘Recovery’ conditions were the same as in Fig. 1. ŽB. A single cell labeled with Lucifer Yellow. Bar s 20 m m. ŽC. Phase contrast micrograph showing the position of the tested cell. ŽD. Fluorescence micrograph after staining for GFAP. ŽB. – ŽD. show the same microscopic field and arrows indicate the tested cell.
C.-L. Chou et al.r Brain Research 789 (1998) 229–238
compensated. A leakage current subtraction program in pCLAMP software was applied assuming that a leak current is linear. 2.3. Immunocytochemistry An indirect immunofluorescent technique was used to demonstrate TH and GFAP in the cytoplasm. After voltage clamp experiment which determined the hypoxic sensitivity of outward current of a tested cell, the cells were fixed and permeabilized by perfusion of absolute methanol for 10 min. The patch pipette was gently removed and the cells were washed by perfusing with 0.1 M phosphate buffered saline ŽPBS.. Ten minutes later the perfusate was changed to a blocking solution Ž1.5% normal horse serum in PBS. to reduce nonspecific binding. Cells were incubated in this solution for 1 h. Subsequently, cells were incubated with monoclonal anti-TH antibody ŽBoehringer Mannheim, mouse IgG, dilution: 1:500. or anti-GFAP ŽBoehringer Mannheim, mouse IgG; 1:50. for 45 min. Cells were washed by perfusing with PBS for 15 min and incubated with Texas Red conjugated anti-mouse IgG ŽVector. for 1 h. Finally, the cells were washed with perfusion of PBS for 10 min and an epifluorescent image was viewed using a rhodamine filter.
233
2.4. Experimental protocol 2.4.1. Experimental set 1 Single cells or cells at the edge of the cluster were selected for this set of experiments. Although obtaining steady patch clamp recording was difficult in these cells, identification of these cells by immunofluorescent technique was easier than cells at the middle of the cluster. After obtaining a whole cell configuration the resting membrane potential was measured followed by a voltage clamp experiment which determined the presence of oxygen-sensitive potassium channels. Then, phase contrast micrographs and epifluorescent micrographs with a fluorescein filter were taken for later identification of the tested cell. Subsequently, immunocytochemistry was performed as described above and epifluorescent micrographs with a rhodamine filter were taken. By comparing the three types of images Žphase contrast, Lucifer Yellow staining and immunostaining., immunoreactivity of the tested cell to TH or GFAP was determined. 2.4.2. Experimental set 2 Cells were initially superfused with Krebs solution equilibrated with 5% CO 2 in air Žcontrol.. After obtaining whole cell currents of a cell, the amplifier was switched to
Fig. 4. Oxygen-insensitive outward current and photomicrographs of a cell which was not immunostained for tyrosine hydroxylase ŽTH.. ŽA. Outward current evoked by a single voltage clamp pulse from y60 mV to q40 mV. The ‘Control’, ‘Hypoxia’, and ‘Recovery’ conditions were the same as in Fig. 1. ŽB. Lucifer Yellow labeling of the tested cell Žarrow. in the periphery of a cluster. Bar s 20 m m. ŽC. Phase contrast micrograph showing the position of the tested cell Žarrow.. ŽD. Fluorescence microscopy showing immunostaining for TH. The tested cell Žarrow. did not show TH-immunoreactivity. ŽB. – ŽD. show the same microscopic field.
234
C.-L. Chou et al.r Brain Research 789 (1998) 229–238
the current clamp mode to record the membrane potential. Subsequently, the external solution was changed to Krebs solution equilibrated with 5% CO 2 in argon. When oxygen tension of the recording chamber reached 25 mmHg, voltage clamp was again applied and whole cell currents were recorded. Next, the amplifier was switched back to the current clamp mode and cells were re-exposed to control solution. When oxygen tension in the chamber reached 152 mmHg Žrecovery., the voltage clamp experiment was repeated.
sensitive outward current. Five cells whose outward current was not sensitive to oxygen had a strong GFAP signal. Six other cells with oxygen-insensitive outward current were examined for TH-immunoreactivity, and five did not show TH-immunoreactivity ŽTable 1.. We observed dye coupling of Lucifer Yellow in eight of twelve cells which expressed oxygen-sensitive outward current. In Fig. 5, the cell indicated by an arrow was used for a patch clamp experiment. This cell and two adjacent cells showed Lucifer Yellow fluorescence, and they were
2.5. Statistical analysis Data are reported as mean " S.E.M. when applicable. To determine significant difference between means, Student’s t-test or analysis of variance combined with the Duncan’s multiple range test was used. Values were considered significantly different if p - 0.05.
3. Results 3.1. Experimental set 1 Fig. 1 represents electrophysiological and immunohistological results from a cell which had oxygen-sensitive outward current. The voltage clamp experiment showed that the outward current was reversibly inhibited by hypoxia ŽFig. 1A.. Lucifer Yellow contained in the patch pipette spread out in the cytoplasm and processes and, even after the patch pipette was removed, the patched cell was readily identified under a fluorescent microscope with a fluorescein filter ŽFig. 1B.. This concentration of Lucifer Yellow did not appear to affect the oxygen sensitivity of outward current. After application of anti-TH and Texas Red conjugated anti-mouse IgG, a few cells showed a bright fluorescence of Texas Red ŽFig. 1D., while some cells in the field did not fluoresce. By superimposing images of Lucifer Yellow, phase contrast, and Texas Red fluorescence, the tested cell was identified as TH positive. Eight cells with oxygen-sensitive outward current were tested for TH-immunoreactivity. All cells showed a strong TH-positive signal ŽTable 1.. In separate experiments, cells which expressed oxygen-sensitive outward current were immunostained for GFAP ŽFig. 2.. Lucifer Yellow fluorescence clearly identified the single cell at an edge of a cluster from which oxygen-sensitive outward current was recorded ŽFig. 2A.. Many cells in the cluster were immunoreactive to GFAP, but the patched cell was GFAPnegative ŽFig. 2D.. Similar results were obtained in four cells ŽTable 1.. In other experiments cells with oxygen-insensitive outward current were stained for either TH or GFAP. Fig. 3 shows an example of a GFAP-positive cell. On the other hand, Fig. 4 illustrates a TH-negative cell with oxygen-in-
Fig. 5. The dye coupling phenomenon in cultured carotid body cells. ŽA. Fluorescence photomicrograph showing Lucifer Yellow loading of the tested cell Žarrow. and two adjacent cells. Bar s 20 m m. ŽB. Appearance of carotid body cells by phase contrast microscopy. ŽC. Fluorescence photomicrograph showing TH-immunoreactive cells. The position of dye-coupled cells were indicated by a solid black line in ŽB. and ŽC..
C.-L. Chou et al.r Brain Research 789 (1998) 229–238
235
Fig. 6. Effects of hypoxia on the membrane potential of a cell with oxygen-sensitive outward current. The upper panel: outward currents activated by a voltage step from y60 mV to q40 mV. The middle panel: the changes in the oxygen tension in the recording chamber. I, II, and III denote the timings of voltage clamp experiments. The lower panel: membrane potential Ž Em . was measured using current clamp mode. Hypoxia depolarized the cell.
TH-positive. No cells whose outward current was oxygeninsensitive showed dye coupling of Lucifer Yellow. 3.2. Experimental set 2 Resting membrane potentials were measured in all cells in experimental set 2 and in some cells in experimental set 1. The resting membrane potentials of cells with oxygensensitive outward current were significantly higher than those of cells with oxygen-insensitive potassium current. The average of the resting membrane potentials of the former cells was y55 " 1 mV Ž n s 9., while the latter was y35 " 2 mV Ž n s 11.. These values were significantly different from each other Žunpaired t-test, p - 0.05.. The difference between these cells was also manifested in the changes in membrane potentials during hypoxia. Cells with oxygen-sensitive outward current were depolarized by hypoxia ŽFig. 6.. The decrease of PO 2 was accompanied by a gradual but significant depolarization Žfrom y55 " 3 mV to y27 " 5 mV; n s 4. and at this time the outward current was reduced. Upon re-exposure to the control solution the membrane potential recovered to control levels Žy48 " 7 mV., and the outward current also recovered. On the other hand, the membrane potentials of cells with oxygen-insensitive outward current were not affected by hypoxia. The average membrane potentials of
four cells were y39 " 5 mV, y34 " 8 mV, and y40 " 5 mV during control, hypoxia, and recovery periods, respectively. These values were not significantly different. Summarized data are presented in Fig. 7.
4. Discussion The present study has demonstrated that the cat carotid body cells whose outward current was inhibited by hypoxia were TH-positive or GFAP-negative. The cells whose outward currents were insensitive to hypoxia were GFAPpositive or TH-negative except one case. The resting membrane potentials of cells having oxygen-sensitive outward current were much higher than those of cells having oxygen-insensitive outward current. The former type of cell was depolarized during hypoxia, but not the latter type of cells. The results indicate that most glomus cells of the adult cat carotid body express oxygen-sensitive potassium channels, have high resting membrane potential, and are depolarized by hypoxia. On the other hand, sheath cells, and possibly a small fraction of glomus cells, are not sensitive to hypoxia in their outward current and membrane potential. Cellular components of the carotid body include glomus cells, sheath cells, fibroblasts, endothelial cells of capillar-
236
C.-L. Chou et al.r Brain Research 789 (1998) 229–238
Fig. 7. Changes in membrane potentials in response to hypoxia. The upper panel shows the response of the cells with oxygen-sensitive outward current. The lower panel is the response of the cells with oxygen-insensitive outward current. ), Significantly different from control and recovery Ž p- 0.05..
ies, and some ganglion cells, but glomus and sheath cells are predominant. Morphological differences between these two types of cells in fixed tissues have been noted even under light microscopy, but some characteristics overlap. This makes distinguishing cat glomus cells more troublesome than those of the rat or the rabbit, because the proportions of glomus cells and sheath cells were estimated as two to one in the cat, four to one in the rat, and seven to three in the rabbit wfor a review, see Ref. w28xx. In our experience identifying live cat glomus cells in culture by simple morphology is extremely difficult. We have been investigating specific methods by which the cell of interest is identified and distinguished. In this study we used Lucifer Yellow to mark the tested cell and applied immunocytochemistry to differentiate cell types. TH, a key enzyme for catecholamine synthesis, is expressed only in glomus cells, not in sheath cells, and more than 90% of glomus cells were TH-positive in the cat w40x. GFAP is a structural protein in astrocytes, and has been localized in sheath cells in the cat w1,21,37x and human w18x. Therefore, TH and GFAP are specific cell makers for glomus and sheath cells, respectively. By comparing images of Lucifer Yellow, phase contrast, and immunofluorescence, identifying the tested cell is relatively straightforward in single
cells or in cells at the edge of a cluster as seen in Figs. 1–4. We recorded whole currents in this study and did not use any specific blockers to identify responsible ionic species for the currents. However, in a recent report, we have shown that the net outward current of cultured carotid body cells was mostly through voltage-gated potassium channels w8x. Since the culture and patch clamp methods of this study were the same as the previous study, and since characteristics of the current Žthe activation time, the size, the lack of apparent inactivation, etc.. were not different from those of the previous study, we have reasonably assumed that the outward current was evoked by activation of voltage-gated potassium channels. Nevertheless, a possibility that the inward calcium current was hidden by the large outward potassium current cannot be totally ruled out. In the present experiments, oxygen-sensitive potassium channels were found only in the cells which contained TH. No cells which had GFAP immunoreactivity expressed oxygen-sensitive potassium channels. Furthermore, all cells that expressed oxygen-sensitive potassium channels did not contain GFAP signals. The results indicate that oxygen-sensitive potassium channels are localized only in glomus cells and not in sheath cells. On the other hand, all cells that stained for GFAP had voltage-gated and oxygen-insensitive potassium channels. Four of five cells which did not show TH-immunoreactivity expressed voltage-gated oxygen-insensitive potassium channels. Therefore, it is reasonable to conclude that sheath cells express voltage-gated potassium channels which are not oxygensensitive. The results agree with previous studies indicating that glomus cells possess oxygen-sensitive potassium channels and sheath cells express small voltage-gated oxygen-insensitive potassium channels in the adult rabbit w13,39x. Compared with the adult rabbit carotid body, voltage-gated oxygen-insensitive potassium channels in sheath cells of the cat carotid body had a lower threshold and a larger current. Interestingly, one cell which did not have oxygen-sensitive potassium channels showed positive immunoreactivity to TH ŽTable 1.. One possibility is that this TH-positive cell was a ganglion cell in the carotid body. However, this possibility is remote, because the size of the cell was not large enough to be a ganglion cell. Rather, this may indicate the presence of electrophysiological subtypes of glomus cells in the cat. Morphologically as many as four subtypes of glomus cells have been reported in the cat according to the size of the dense-cored vesicles and their histochemical reaction w30x. Subtypes of glomus cells were also reported in the rat and human wfor reviews, see Refs. w19,28xx. In addition to the morphological subtypes, heterogeneous physiological responses of glomus cells have been also reported. Donnelly w11x found two distinct subtypes of glomus cells in the adult rat carotid body: one with a small-voltage dependent inward current and a large out-
C.-L. Chou et al.r Brain Research 789 (1998) 229–238
ward current, the other with little voltage-dependent current. Bright et al. w4x observed that changes in intracellular calcium in response to hypoxia varied among glomus cells in the rat. Further, Perez-Garcia et al. w35x noted that cells in some cultures did not show ‘the typical inhibition of the potassium currents’, suggesting subtypes of glomus cells in the adult rabbit. Cells with oxygen-sensitive potassium channels were depolarized during hypoxia, whereas cells with oxygen-insensitive voltage-gated potassium channels were not depolarized by hypoxia. Thus, these findings are consistent with the current hypothesis of hypoxic chemotransduction which implies that hypoxia depolarizes glomus cells possibly due to inhibition of oxygen-sensitive potassium channels. However, unlike adult rabbit glomus cells, adult cat glomus cells did not show any action potentials. This fact raises a question whether the size and the speed of depolarization of cat glomus cells are adequate to activate voltage-gated calcium channels for subsequent release of neurotransmitterŽs.. In smooth muscle cells sustained depolarization to y30 mV caused a steady-state elevation of intracellular calcium via activation of L-type of calcium channels w16x. It is not known whether this is true in glomus cells. Another important question is whether the depolarization of glomus cells is due to the inhibition of voltage-gated potassium channels. Simultaneous measurement of carotid sinus nerve activity and potassium current of glomus cells in the adult rat revealed the dissociation of the inhibition of potassium current and increase in nerve activity w6,12x. Further, Buckler w5x has recently found a new type of potassium channels which are active at rest and inhibited by hypoxia. These data suggest that the role of oxygen-sensitive potassium channels in depolarizing glomus cells may be modulatory. The presence of gap junctions in the rat and the cat has been suggested by McDonald w28x. He estimated two to three gap junctions per 100 glomus cells. Electrical coupling between glomus cells was also indicated w2,29x, and in these studies about 30% of cells showed electrical coupling. Recently, Kondo and Iwasa w23x reevaluated the occurrence of gap junctions in the rat carotid body. They reported that presumed gap junctions between glomus cells occurred frequently, and they were present even between glomus cells and sheath cells. In agreement with these studies the phenomenon of dye coupling appeared to be a common phenomenon in the cat in our study. Eight out of twelve cells which have oxygen-sensitive potassium channels showed dye coupling with other cells. Eleven cells with oxygen-insensitive potassium channels did not show any dye-coupling. Thus, intracellular connections may be most frequently found between glomus cells in the cat. For understanding the mechanisms of chemoreception in vivo, one important question is whether cultured glomus cells have the same characteristics as those in situ. Although cultured glomus cells have been widely used, very few studies have addressed this question w21,35x. Our
237
culture system employs Matrigele and nerve growth factor ŽNGF.. Matrigele contains several growth factors, but the possibility that the growth factors in Matrigele would change the expression of channels is remote. We have been using diluted Matrigele, which has been aspirated and washed before plating cells. NGF is well known to promote neuronal differentiation in neural crest-derived cells w32x. However, neurofilament ŽNF., a structural protein specific for neurons, was not detected in cultured cat carotid body cells w21x. Jackson and Nurse w22x have also reported that NGF failed to induce NF 68 kDa and NF 160 kDa in cultured postnatal rat glomus cells. The absence of NF in the cultured glomus cells suggest that NGF does not promote differentiation of glomus cells into neurons. Further, we did not observe any time-dependent changes in outward current during 2 weeks culture period, suggesting that the expression of voltage-gated potassium channels are stable for 2 weeks in our culture conditions. In summary, we have shown that cultured glomus cells of the adult cat carotid body possess oxygen-sensitive potassium channels and are depolarized in response to hypoxia. On the other hand, sheath cells and possibly a small fraction of glomus cells possess oxygen-insensitive potassium channels and their membrane potential is not affected by hypoxia. Although the data are consistent with the notion that glomus cells are the chemosensory cells, the role of oxygen-sensitive potassium channels in carotid body chemoreception in situ requires further investigation.
Acknowledgements The authors are grateful to Drs. Thomas Croxton and Bradly Undem for their helpful suggestions. This study was supported by HL47044 and HL50712.
References w1x A. Abramovici, D.J. Pallot, J.M. Polak, Immunohistochemical approach to the study of the cat carotid body, Acta Anat. 140 Ž1991. 70–74. w2x V. Abudara, C. Eyzaguirre, Electrical coupling between cultured glomus cells of the rat carotid body: observations with current and voltage clamping, Brain Res. 664 Ž1994. 257–265. w3x T.J. Biscoe, M.R. Duchen, Responses of type I cells dissociated from the rabbit carotid body to hypoxia, J. Physiol. 428 Ž1990. 39–59. w4x G.R. Bright, F.H. Agani, U. Haque, J.L. Overholt, N.R. Prabhakar, Heterogeneity in cytosolic calcium responses to hypoxia in carotid body cells, Brain Res. 706 Ž1996. 297–302. w5x K.J. Buckler, A novel oxygen-sensitive potassium current in rat carotid body type I cells, J. Physiol. 498 Ž1997. 649–662. w6x P.M. Cheng, D.F. Donnelly, Relationship between changes of glomus cell current and neural response of rat carotid body, J. Neurophysiol. 74 Ž1995. 2077–2086. w7x C.-L. Chou, B. Schofield, Y. Ishizawa, M. Shirahata, Differential
238
w8x w9x
w10x
w11x w12x w13x
w14x
w15x
w16x
w17x
w18x
w19x w20x
w21x
w22x
w23x
C.-L. Chou et al.r Brain Research 789 (1998) 229–238 expression of voltage-gated K channels in glomus cells and sheath cells of cat carotid body, FASEB J. 10 Ž1996. A406. C.-L. Chou, M. Shirahata, Two types of voltage-gated K channels in carotid body cells of adult cats, Brain Res. 742 Ž1996. 34–42. M. de B. Daly, Interactions between respiration and circulation, in: A.P. Fishman, N.S. Cherniack, J.G. Widdicombe ŽEds.., Handbook of Physiology, Section 3: The Respiratory System, American Physiological Society, Bethesda, MD, 1986, pp. 529–594. M.A. Delpiano, J. Hescheler, Evidence for a PO 2-sensitive Kq channel in the type I cell of the rabbit carotid body, FEBS Lett. 249 Ž1989. 195–198. D.F. Donnelly, Response to cyanide of two types of glomoid cells in mature rat carotid body, Brain Res. 630 Ž1993. 157–168. D.F. Donnelly, Modulation of glomus cell membrane currents of intact rat carotid body, J. Physiol. ŽLondon. 489 Ž1995. 677–688. M.R. Duchen, K.W. Caddy, G.C. Kirby, D.L. Patterson, J. Ponte, T.J. Biscoe, Biophysical studies of the cellular elements of the rabbit carotid body, Neuroscience 26 Ž1988. 291–311. C. Eyzaguirre, L. Monti-Bloch, M. Baron, Y. Hayashida, J.W. Woodbury, Changes in glomus cell membrane properties in response to stimulants and depressants of carotid nerve discharge, Brain Res. 477 Ž1989. 265–279. R.S. Fitzgerald, S. Lahiri, Reflex responses to chemoreceptor stimulation, in: A.P. Fishman, N.S. Cherniack, J.G. Widdicombe ŽEds.., Handbook of Physiology, Section 3: The Respiratory System, American Physiological Society, Bethesda, MD, 1986, pp. 313–362. B.K. Fleischmann, R.K. Murray, M.I. Kotlikoff, Voltage window for sustained elevation of cytosolic calcium in smooth muscle cells, Proc. Natl. Acad. Sci. U.S.A. 91 Ž1994. 11914–11918. C. Gonzalez, L. Almaraz, A. Obeso, R. Rigual, Carotid body chemoreceptors: from natural stimuli to sensory discharges, Physiol. Rev. 74 Ž1994. 829–898. J.O. Habeck, W. Kummer, Neuronal and neuroendocrine markers in the human carotid body in health and disease, Adv. Exp. Med. Biol. 337 Ž1993. 31–35. D. Heath, P. Smith, Diseases of the human carotid body, SpringerVerlag, London, 1992. J. Hescheler, M.A. Delpiano, H. Acker, Ionic currents on type I cells of the rabbit carotid body measured by voltage clamp experiments and the effect of hypoxia, Brain Res. 486 Ž1989. 79–88. Y. Ishizawa, B. Schofield, C.-L. Chou, M. Shirahata, Changes in protein expression of cultured cat carotid body cells. Book of Abstracts ŽISOTT 95—Satellite Symposium: Adaptation to Hypoxia. Ž1995. IV.3. A. Jackson, C. Nurse, Plasticity in cultured carotid body chemoreceptors: environmental modulation of GAP-43 and neurofilament, J. Neurobiol. 26 Ž1995. 485–496. H. Kondo, H. Iwasa, Re-examination of the carotid body ultrastructure with special attention to intracellular membrane appositions, in: P. Zapata, C. Eyzaguirre, R.W. Torrance ŽEds.., Frontiers in Arterial Chemoreception, Plenum, New York, 1996, pp. 45–50.
w24x J. Lopez-Barneo, Oxygen-sensing by ion channels and the regulation of cellular functions, Trends Neurosci. 19 Ž1996. 435–440. w25x J. Lopez-Barneo, J.R. Lopez-Lopez, J. Urena, C. Gonzalez, Chemotransduction in the carotid body: Kq current modulated by PO 2 in type I chemoreceptor cells, Science 241 Ž1988. 580–582. w26x J. Lopez-Lopez, C. Gonzalez, J. Urena, J. Lopez-Barneo, Low pO 2 selectively inhibits K channel activity in chemoreceptor cells of the mammalian carotid body, J. Gen. Physiol. 93 Ž1989. 1001–1015. w27x A. Marty, E. Neher, Tight-seal whole-cell recording, in: B. Sakmann, E. Neher ŽEds.., Single-Channel Recording, Plenum, New York, 1983, pp. 107–122. w28x D.M. McDonald, Peripheral chemoreceptors structure–function relationships of the carotid body, in: T.F. Hornbein ŽEd.., Regulation of Breathing, Marcel Dekker, New York, 1981, pp. 105–319. w29x L. Monti-Bloch, V. Abudara, C. Eyzaguirre, Electrical communication between glomus cells of the rat carotid body, Brain Res. 622 Ž1993. 119–131. w30x E. Morita, S.R. Chiocchio, J.H. Tramezzani, Four types of main cells in the carotid body of the cat, J. Ultrastruct. Res. 28 Ž1969. 399–410. w31x L. Pang, C. Eyzaguirre, Different effects of hypoxia on the membrane potential and input resistance of isolated and clustered carotid body glomus cells, Brain Res. 575 Ž1992. 167–173. w32x P.H. Patterson, Control of cell fate in a vertebrate neurogenic lineage, Cell 62 Ž1990. 1035–1038. w33x C. Peers, Hypoxic suppression of Kq currents in type I carotid body cells: selective effect on the Ca2q-activated Kq current, Neurosci. Lett. 119 Ž1990. 253–256. w34x C. Peers, K.J. Buckler, Transduction of chemostimuli by the type I carotid body cell, J. Membr. Biol. 144 Ž1995. 1–9. w35x M.T. Perez-Garcia, A. Obeso, J.R. Lopez-Lopez, B. Herreros, C. Gonzalez, Characterization of cultured chemoreceptor cells dissociated from adult rabbit carotid body, Am. J. Physiol. 263 Ž1992. C1152–C1159. w36x M. Shirahata, B. Schofield, B.Y. Chin, T.R. Guilarte, Culture of arterial chemoreceptor cells from adult cats in defined medium, Brain Res. 658 Ž1994. 60–66. w37x M. Shirahata, B. Schofield, R.S. Fitzgerald, Immunohistological differentiation of cultured cat carotid body cells, Soc. Neurosci. Abstr. 20 Ž1994. 957. w38x A. Stea, C.A. Nurse, Whole-cell and perforated-patch recordings from O 2 -sensitive rat carotid body cells grown in short- and long-term culture, Pflugers Arch. 418 Ž1991. 93–101. w39x J. Urena, J. Lopez-Lopez, C. Gonzalez, J. Lopez-Barneo, Ionic currents in dispersed chemoreceptor cells of the mammalian carotid body, J. Gen. Physiol. 93 Ž1989. 979–999. w40x Z.Z. Wang, L.J. Stensaas, B. Dinger, S.J. Fidone, Co-existence of tyrosine hydroxylase and dopamine b-hydroxylase immunoreactivity in glomus cells of the cat carotid body, J. Auton. Nerv. Syst. 32 Ž1991. 259–264.