Proc. Nati. Acad. Sci. USA Vol. 82, pp. 1448-1450, March 1985 Cell Biology Oxygen chemoreception by carotid body cells in culture (glomus cells/dopamine/hypoxia/sensory cells) MARK C. FISHMAN*t#§¶, WILLIAM L. GREENEt§, AND DoRos PLATIKAt¶ *Howard Hughes Medical Institute, tThe Developmental Biology Laboratory (Medical Services), and *Neurology Services of the Massachusetts General Hospital, Boston, MA 02114; and Departments of §Medicine and ¶Neurology, Harvard Medical School, Boston, MA 02115 Communicated by Jerome Gross, October 29, 1984 ABSTRACT Chemoreceptors for oxygen reside within the enrichment procedure. Hypoxic conditions were identical carotid body, but it is not known which cells actually sense except that the Po2 was diminished to 36 mm Hg. Since there hypoxia and by what mechanisms they transduce this informa- are only a few thousand cells in the carotid body, each ex- tion into afferent signals in the carotid sinus nerve. We have periment necessitated the pooling of dissociated cells of 10- developed systems for the growth of glomus cells of the carotid 12 rats and subsequent distribution to two plates, one to be body in dissociated cell culture. Here we demonstrate that, as used for control and one for hypoxic conditions. Cells in in vivo, these cells contain the putative neurotransmitters do- both plates for each experiment were grown in culture for pamine, serotonin, and norepinephrine. Oxygen tension regu- the same period of time. Experiments were performed after lates the rate of dopamine secretion from the glomus cells. culturing of the cells for 5-7 days. Similar to chemically stimulated catecholamine secretion from Prior to testing the cells were washed twice in Hanks' bal- other adrenergic cells this hypoxia-stimulated release requires anced salt solution supplemented with 100 ,uM pargyline/5 extracellular calcium. These results are compatible with the mM ascorbate/4.5 g of glucose per liter, and then exposed in suggestion that the glomus cells of the carotid body are chemo- 500 ,ll of the same solution for 1-2 hr to either Po2 151 mm receptor cells and that they signal hypoxia by regulated secre- Hg or Po2 36 mm Hg. The Po2 of the cells beneath 2 mm of tion of dopamine. culture fluid can be estimated to be diminished below that of the surrounding air by <5 mm Hg (8). Prior to analysis, the Since the work of Heymans (1) it has been known that the cells were sonicated and freeze-thawed in HPLC running hypoxic ventilatory drive originates primarily within the ca- buffer. This buffer consisted of 0.75 M dibasic sodium phos- rotid body. Subsequent work has outlined the role of this phate/0.1 M citric acid/1.5 mM sodium heptane sulfonate sensory organ in control ofbreathing at sea level and at high- (Kodak)/14% (vol/vol) methanol. Particulate matter was re- er altitude (1, 2) and in the sensation of dyspnea (3). Howev- moved from both the sonicated cellular extract and the su- er, definition of the molecular basis of oxygen transduction pernatant by centrifugation in an Airfuge (Beckman) for 5 has been hampered by the complexity of the carotid body. It min. [Concentration of the catecholamines by alumina ex- contains several cell types, including endothelial, sustentac- traction (9) proved infeasible because of the very small ular, and glomus cells as well as the terminals ofafferent and amounts of tissue.] The supernatant from the centrifugation sympathetic nerves (4, 5), but it is not known which are actu- was separated by the use of HPLC (Beckman 420) with a C- ally capable of sensing the level of oxygen. The glomus cells 18 column (Altex Ultrasphere ODS), and detection was by have been attractive candidates to be chemoreceptor cells an electrochemical detector (BAS LC-3A). Elution rate was because they contain dopamine, a transmitter secreted from 1.5 ml/min, and the sample injection size was 200 /.. The the carotid body during hypoxia, and because they form electrochemical detector was used at 0.70 V. The identity of morphological synapses with afferent nerve terminals (5, 6). peaks was confirmed by exact superimposition of standard We have designed systems for the growth of glomus cells in peaks with presumptive norepinephrine, dopamine, and ser- culture in order to investigate the neurophysiology of these otonin peaks from the cellular extract, which represented cells and their potential role in chemoreception (7). Here, we cell content, or the culture fluid, which reflected the secret- report evidence that carotid body glomus cells in culture ed catecholamines. The sensitivity of the HPLC-EC system synthesize and secrete putative neurotransmitters and mani- for measuring catecholamines was "'100 pg ofdopamine, 150 fest chemosensitivity in their control ofdopamine secretion. pg of norepinephrine, and 200 pg of serotonin. RESULTS MATERIALS AND METHODS Glomus cells in vivo contain several putative neurotransmit- Dissociated newborn rat carotid body cells were cultured as ters, including norepinephrine , dopamine, and serotonin (5, described (7) in 24-well plates and were grown in a humidi- 10). As shown in Fig. 1, measurement of the extracted con- fied atmosphere of Po2 151 mm Hg, Pco2 36 mm Hg in sup- tents of cultured glomus cells by HPLC combined with elec- plemented F-12 medium (pH 7.2). Nerve growth factor was trochemical detection (HPLC-EC) documented their synthe- not added to the medium. Enrichment for glomus cells could sis of norepinephrine, dopamine, and serotonin. Dopamine be achieved by the use of tyrosine-free medium (7), but suc- and serotonin secretory rates were great enough to be de- cessful elimination of other cells required several weeks of tected in the culture fluid after 1 hr of incubation, as shown growth under these conditions. Since the very few cells in in Fig. 1. Secreted norepinephrine was not detected, but the carotid body was the limiting factor in the number of very low levels might have been obscured by interference experiments that were possible and since there was a slow from other oxidizable compounds. Interestingly, carotid but noticeable decrement in catecholamine content in cul- body cells cultured for <4 days synthesized dopamine but tures >7-10 days old, we elected not to use this long-term did not secrete measurable quantities into the medium. In each of the six experiments using the cultured carotid body cells, hypoxia caused an increase in dopamine secre- The publication costs of this article were defrayed in part by page charge ± payment. This article must therefore be hereby marked "advertisement" tion. The mean dopamine secretion increased from 0.092 in accordance with 18 U.S.C. §1734 solely to indicate this fact. 0.029 ng per carotid body per hr to 0.211 ± 0.091 ng per 1448 Downloaded by guest on September 25, 2021 Cell Biology: Fishman et aL Proc. NatL Acad Sci USA 82 (1985) 1449 ic> I CU a)a -0- .6 CZ mv 0.05 IL FIG. 1. (Left) Cellular content of neurotransmitters synthesized o - by cultured carotid body cells and revealed by HPLC-EC include norepinephrine (NE), dopamine (DA), and serotonin (5HT). Cali- CL bration bar length represents 2 min, and height represents 125 pg of NE, 200 pg of DA, and 333 pg of 5HT. (Right) HPLC-EC of culture CY) fluid, revealing secreted DA and 5HT. ........... carotid body per hr (mean ± SD; n = 6 independent experi- P < ments; 0.01, Student's t test). As shown in Fig. 2, this FIG. 3. Extracellular calcium is necesbary for stimulated dopa- enhanced secretion was accomplished without significant mine secretion. Baseline secretory rate of dopamine (Left) is un- diminution in cellular content of dopamine. This suggests changed by hypoxia in the absence ofextracellular calcium (Middle) first of all that hypoxia did not injure the cell and thereby but is increased significantly by hypoxia when calcium is present cause a nonspecific leak of dopamine. In addition, the in- (Right) (mean ± SEM). crease in secretion without diminution in content suggests that hypoxia also stimulated enhanced synthesis of dopa- mine, an effect of hypoxia that has been noted using ex- To assess further the specificity of the oxygen sensitivity, planted rabbit carotid bodies (11). we investigated the effect of hypoxia on epinephrine secre- Decrease in extracellular calcium diminishes hypoxia- tion by cultured rat adrenal medullary cells. Glomus cells stimulated release of dopamine from explanted carotid bod- resemble adrenal medullary cells morphologically, and both ies (6). Since calcium influx is integral to many secretory cell types derive embryologically from the neural crest (13, processes that involve vesicular release (12), it was of inter- 14). In vivo, the predominant secreted catecholamine from est to evaluate the effect of calcium on glomus cell function. the adrenal is epinephrine. Adrenal cells cultured for 4 days Omission of calcium from the medium, with magnesium con- secreted epinephrine, as shown in Fig. 4. Hypoxia, however, centration left unaltered, did not significantly affect baseline did not cause an increase in their rate of secretion. dopamine secretion from the cultured carotid body cells (not shown). However, as shown in Fig. 3, the oxygen-sensitive DISCUSSION component was abolished. Under the same conditions but with the addition of calcium to the medium, hypoxia induced The hypoxic stimulus to ventilatory drive originates predom- a significant increase in the mean dopamine secretory rate inantly within the carotid body. As arterial Po2 decreases, from 0.040 ± 0.020 ng per carotid body per hr to 0.077 ± the discharge rate of the carotid sinus sensory nerve in- 0.023 ng per carotid body per hr (mean + SD; n = 4; P < creases. It has been suggested, however, that it may not be 0.05). Extracellular calcium might, of course, be involved the nerve itself that is chemosensory, but rather other cells not only in but at of the carotid body, because a cut carotid sinus nerve must secretion also other steps in the sensory regenerate to the carotid body to manifest chemosensitivity transduction mechanism.