Leptin Induces Hypertension Acting on Transient Receptor Potential Melastatin 7 Channel in the Carotid Body.

RATIONALE: Obesity leads to resistant hypertension and mechanisms are poorly understood, but high plasma levels of leptin have been implicated. Leptin increases blood pressure acting both centrally in the dorsomedial hypothalamus and peripherally. Sites of the peripheral hypertensive effect of leptin have not been identified. We previously reported that leptin enhanced activity of the carotid sinus nerve, which transmits chemosensory input from the carotid bodies (CBs) to the medullary centers, and this effect was abolished by nonselective blockers of Trp (transient receptor potential) channels. We searched our mouse CB transcriptome database and found that the Trpm7 (transient receptor potential melastatin 7) channel was the most abundant Trp channel. OBJECTIVE: To examine if leptin induces hypertension acting on the CB Trpm7. METHODS AND RESULTS: C57BL/6J (n=79), leptin receptor (LepRb) deficient db/db mice (n=22), and LepRb-EGFP (n=4) mice were used. CB Trpm7 and LepRb gene expression was determined and immunohistochemistry was performed; CB glomus cells were isolated and Trpm7-like current was recorded. Blood pressure was recorded continuously in (1) leptin-treated C57BL/6J mice with intact and denervated CB; (2) leptin-treated C57BL/6J mice, which also received a nonselective Trpm7 blocker FTY720 administered systemically or topically to the CB area; (3) leptin-treated C57BL/6J mice transfected with Trpm7 small hairpin RNA to the CB, and (4) Leprb deficient obese db/db mice before and after Leprb expression in CB. Leptin receptor and Trpm7 colocalized in the CB glomus cells. Leptin induced a nonselective cation current in these cells, which was inhibited by Trpm7 blockers. Leptin induced hypertension in C57BL/6J mice, which was abolished by CB denervation, Trpm 7 blockers, and Trpm7 small hairpin RNA applied to CBs. Leprb overexpression in CB of Leprb-deficient db/db mice demethylated the Trpm7 promoter, increased Trpm7 gene expression, and induced hypertension. CONCLUSIONS: We conclude that leptin induces hypertension acting on Trmp7 in CB, which opens horizons for new therapy.

Leptin acts in the carotid bodies to increase minute ventilation during wakefulness and sleep and augment the hypoxic ventilatory response.

KEY POINTS: Leptin is a potent respiratory stimulant. A long functional isoform of leptin receptor, LepRb , was detected in the carotid body (CB), a key peripheral hypoxia sensor. However, the effect of leptin on minute ventilation (VE ) and the hypoxic ventilatory response (HVR) has not been sufficiently studied. We report that LepRb is present in approximately 74% of the CB glomus cells. Leptin increased carotid sinus nerve activity at baseline and in response to hypoxia in vivo. Subcutaneous infusion of leptin increased VE and HVR in C57BL/6J mice and this effect was abolished by CB denervation. Expression of LepRb in the carotid bodies of LepRb deficient obese db/db mice increased VE during wakefulness and sleep and augmented the HVR. We conclude that leptin acts on LepRb in the CBs to stimulate breathing and HVR, which may protect against sleep disordered breathing in obesity. ABSTRACT: Leptin is a potent respiratory stimulant. The carotid bodies (CB) express the long functional isoform of leptin receptor, LepRb , but the role of leptin in CB has not been fully elucidated. The objectives of the current study were (1) to examine the effect of subcutaneous leptin infusion on minute ventilation (VE ) and the hypoxic ventilatory response to 10% O2 (HVR) in C57BL/6J mice before and after CB denervation; (2) to express LepRb in CB of LepRb -deficient obese db/db mice and examine its effects on breathing during sleep and wakefulness and on HVR. We found that leptin enhanced carotid sinus nerve activity at baseline and in response to 10% O2 in vivo. In C57BL/6J mice, leptin increased VE from 1.1 to 1.5 mL/min/g during normoxia (P < 0.01) and from 3.6 to 4.7 mL/min/g during hypoxia (P < 0.001), augmenting HVR from 0.23 to 0.31 mL/min/g/Δ FIO2 (P < 0.001). The effects of leptin on VE and HVR were abolished by CB denervation. In db/db mice, LepRb expression in CB increased VE from 1.1 to 1.3 mL/min/g during normoxia (P < 0.05) and from 2.8 to 3.2 mL/min/g during hypoxia (P < 0.02), increasing HVR. Compared to control db/db mice, LepRb transfected mice showed significantly higher VE throughout non-rapid eye movement (20.1 vs. -27.7 mL/min respectively, P < 0.05) and rapid eye movement sleep (16.5 vs 23.4 mL/min, P < 0.05). We conclude that leptin acts in CB to augment VE and HVR, which may protect against sleep disordered breathing in obesity.

A Short-Term Fasting in Neonates Induces Breathing Instability and Epigenetic Modification in the Carotid Body.

The respiratory control system is not fully developed in newborn, and data suggest that adequate nutrition is important for the development of the respiratory control system. Infants need to be fed every 2-4 h to maintain appropriate energy levels, but a skip of feeding can occur due to social economical reasons or mild sickness of infants. Here, we asked questions if a short-term fasting (1) alters carotid body (CB) chemoreceptor activity and integrated function of the respiratory control system; (2) causes epigenetic modification within the respiratory control system. Mouse pups (

The human carotid body transcriptome with focus on oxygen sensing and inflammation--a comparative analysis.

The carotid body (CB) is the key oxygen sensing organ. While the expression of CB specific genes is relatively well studied in animals, corresponding data for the human CB are missing. In this study we used five surgically removed human CBs to characterize the CB transcriptome with microarray and PCR analyses, and compared the results with mice data. In silico approaches demonstrated a unique gene expression profile of the human and mouse CB transcriptomes and an unexpected upregulation of both human and mouse CB genes involved in the inflammatory response compared to brain and adrenal gland data. Human CBs express most of the genes previously proposed to be involved in oxygen sensing and signalling based on animal studies, including NOX2, AMPK, CSE and oxygen sensitive K+ channels. In the TASK subfamily of K+ channels, TASK-1 is expressed in human CBs, while TASK-3 and TASK-5 are absent, although we demonstrated both TASK-1 and TASK-3 in one of the mouse reference strains. Maxi-K was expressed exclusively as the spliced variant ZERO in the human CB. In summary, the human CB transcriptome shares important features with the mouse CB, but also differs significantly in the expression of a number of CB chemosensory genes. This study provides key information for future functional investigations on the human carotid body.

ATP release from the carotid bodies of DBA/2J and A/J inbred mouse strains.

The purposes of this study were to: (1) establish an effective method to measure the release of ATP from the mouse carotid body (CB) and (2) determine the release of ATP from the CB of the DBA/2 J (high hypoxic responder) and A/J (low hypoxic responder) mouse in response to hypoxia and hypercapnia. An incubation chamber was constructed utilizing a Costar® Spin-X Centrifuge Tube Filter. The filter was coated with low melting point agarose to hold 4 CBs or 4 superior cervical ganglia (SCG). Hypoxia did not increase ATP release from the CB of either strain. ATP increased in response to a normoxic/hypercapnic challenge in the DBA/2 J's CB but not in the A/J's CB. ATP release from the SCG was affected by neither hypoxia nor hypercapnia in both strains. Thus, we have concluded: (1) we successfully established a chamber system to measure ATP released from the mouse CB; (2) ATP may not be an excitatory neurotransmitter in the CB of these mice under hypoxia; (3) ATP may be a neurotransmitter in the CB of the DBA/2 J mouse strain during hypercapnia.

Oxygen sensing in the carotid body and its relation to heart failure.

This brief review first touches on the origins of the earth's oxygen. It then identifies and locates the principal oxygen sensor in vertebrates, the carotid body (CB). The CB is unique in that in human subjects, it is the only sensor of lower than normal levels in the partial pressure of oxygen (hypoxia, HH). Another oxygen sensor, the aortic bodies, are mostly vestigial in higher vertebrates. At least they play a much smaller role than the CB. In such an important role, the many reflexes in response to CB stimulation by HH are presented. After briefly reviewing what CB stimulation does, the next topic is to describe how the CB chemotransduces HH into neural signals to the brain. Several mechanisms are known, but critical steps in the mechanisms of chemosensation and chemotransduction are still under investigation. Finally, a brief glance at the operation of the CB in chronic heart failure patients is presented. Specifically, the role of nitric oxide, NO, is discussed.

Role of acetylcholine in neurotransmission of the carotid body.

Acetylcholine (ACh) has been considered an important excitatory neurotransmitter in the carotid body (CB). Its physiological and pharmacological effects, metabolism, release, and receptors have been well documented in several species. Various nicotinic and muscarinic ACh receptors are present in both afferent nerve endings and glomus cells. Therefore, ACh can depolarize or hyperpolarize the cell membrane depending on the available receptor type in the vicinity. Binding of ACh to its receptor can create a wide variety of cellular responses including opening cation channels (nicotinic ACh receptor activation), releasing Ca(2+) from intracellular storage sites (via muscarinic ACh receptors), and modulating activities of K(+) and Ca(2+) channels. Interactions between ACh and other neurotransmitters (dopamine, adenosine, nitric oxide) have been known, and they may induce complicated responses. Cholinergic biology in the CB differs among species and even within the same species due to different genetic composition. Development and environment influence cholinergic biology. We discuss these issues in light of current knowledge of neuroscience.

The impact of PCO2 and H+ on the release of acetylcholine from the cat carotid body.

The carotid body (CB) is a sensor of oxygen, carbon dioxide, hydrogen ion, and glucose in the arterial blood. Many studies of the CB's responses to low oxygen (hypoxia) have been reported. Recently attention has been increasingly focused on its responses to elevated CO2 (hypercapnia). An increase in ventilation or carotid body neural output (CBNO) can result from stimulating the CB with blood or perfusion fluids having an elevated CO2 or H+. The increase in ventilation seen with a hypoxic stimulus is accompanied with an increase in CBNO and an increased release of both acetylcholine (ACh) and ATP from the CB. The present in vitro study using both CBs harvested from six cats was undertaken to determine if hypercapnia also provoked an increased release of ACh from the incubated CBs. The anesthetizing, handling, and euthanizing of the animals were according to the guidelines of the Johns Hopkins Animal Care and Use Committee which are totally consonant with those of the NIH. CBs, once harvested and prepared for the experimental protocol, were subjected to the following steps each lasting 10 min: (1) control; (2) stress; (3) recovery. The stresses were respiratory acidosis (RAC; acidic hypercapnia), compensated respiratory acidosis (CRAC; isohydric hypercapnia), and metabolic acidosis (MtAC). The first and last forms of acidosis generated small but significant increases in the release of ACh from the CBs; the second generated a very small and insignificant increase in ACh release. Since it is generally accepted that ACh is a key excitatory neurotransmitter in the CB along with ATP, these data are consistent with other studies measuring the increase in ventilation in response to a small increase in CO2 and those studies recording CBNO in response to hypercapnia. In five of the six animals the responses to RAC and MtAC were compared to the responses to hypoxia. The latter were statistically indistinguishable from the former two.

Lipopolysaccharide exposure during the early postnatal period adversely affects the structure and function of the developing rat carotid body.

The carotid body (CB) substantially influences breathing in premature infants by affecting the frequency of apnea and periodic breathing. In adult animals, inflammation alters the structure and chemosensitivity of the CB, yet it is not known if this pertains to neonates. We hypothesized that early postnatal inflammation leads to morphological and functional changes in the developing rat CB, which persists for 1 wk after the initial provoking insult. To test our hypothesis, we exposed rat pups at postnatal day 2 (P2) to lipopolysaccharide (LPS; 100 µg/kg) or saline (SAL) intraperitoneally. At P9-10 (1 wk after treatment), LPS-exposed animals had significantly more spontaneous intermittent hypoxic (IH) events, attenuated ventilatory responses to changes in oxygen tension (measured by whole body plethysmography), and attenuated hypoxic chemosensitivity of the carotid sinus nerve (measured in vitro), compared with SAL-exposed controls. These functional changes were associated with the following: 1) increased inflammatory cytokine mRNA levels; 2) decreased volume of supportive type II cells; and 3) elevated dopamine levels (a major inhibitory neuromodulator) within the CB. These findings suggest that early postnatal inflammation in newborn rats adversely affects the structure and function of the CB and is associated with increased frequency of intermittent desaturations, similar to the phenomenon observed in premature infants. Furthermore, this is the first newborn model of spontaneous intermittent desaturations that may be used to understand the mechanisms contributing to IH events in newborns.

Genetic background affects cardiovascular responses to obstructive and simulated apnea.

We have recently demonstrated that genetic background significantly impacts the blood pressure and heart rate response to hypoxia (Campen MJ, Tagaito Y, Li J, Balbir A, Tankersley CG, Smith P, Schwartz A, and O'Donnell CP. Physiol Genomics 20: 15-20, 2005). Because hypoxia is considered a mediator of the acute and chronic cardiovascular complications of obstructive sleep apnea, we investigated whether genetic factors also influence the cardiovascular response to experimentally induced obstructive apnea (OA) and simulated apnea (SA). In three strains of inbred mice (C57BL/6J, DBA/2J, and FVB/J) anesthetized with urethane (1.2 g/kg), apnea was induced at end-expiration for 5- and 10-s periods in spontaneously breathing (OA) and mechanically ventilated (SA; pancuronium, 0.2 mg/kg bolus + 0.003 mg.kg(-1).min(-1)) animals before and after administration of an autonomic ganglionic blocker (hexamethonium, 20 mg/kg). In contrast to our previous findings with hypoxia, OA produced a marked hypertensive response in all three strains. However, strain impacted on the degree of bradycardia during OA, which was large in C57BL/6J and FVB/J mice and effectively absent in DBA/2J mice. In C57BL/6J but not FVB/J mice, the bradycardia was abolished with SA under mechanical ventilation. Cardiovascular responses to SA in all strains were eliminated by autonomic blockade. These data show that 1) DBA/2J mice, in contrast to the previous demonstration of marked bradycardia during hypoxia, unexpectedly do not produce bradycardia during apnea; 2) C57BL/6J mice exhibit a bradycardia that is dependent on input from thoracic afferents; and 3) FVB/J mice exhibit a bradycardia despite the loss of thoracic afferent input, consistent with a potent pressure response eliciting a baroreceptor-mediated bradycardia. Thus genetic background can affect both the pattern and magnitude of the cardiovascular response to apnea.

The effect of a nitric oxide donor, sodium nitroprusside, on the release of acetylcholine from the in vitro cat carotid body.

The purpose of the present study was to determine the impact of a nitric oxide (NO) donor, sodium nitroprusside (SNP), on the release of acetylcholine (ACh), an essential excitatory neurotransmitter, from the in vitro cat carotid body (CB). Bilateral CBs were harvested from five deeply anesthetized cats according to the regulations contained in the policies of the Johns Hopkins Animal Care and Use Committee. After recovering from the surgical procedures for extraction and cleaning, the CBs were taken through a 15-step protocol in which they were exposed to a hyperoxic gas mixture (40% O2/5% CO2; 20 min), then a hypoxic gas mixture (6% O2/5% CO2; 20 min), and a final 10 min hyperoxic mixture. This sequence was applied twice, followed by the same sequence in the presence, first, of 5 microM SNP, and secondly in the presence of 10 microM SNP. After washing and a recovery period the CBs were again exposed to the gases as in the first two non-SNP trials. The SNP exposures significantly reduced the overall release of ACh by about 20% (P=0.039). Further, SNP significantly reduced the hypoxia-induced increase in ACh release (without SNP: 82.4+/-19.1 fmol/20 microL versus with SNP: 49.7+/-15.0 fmol/20 microL; mean+/-S.E.M.; P=0.032). Trials #1 and #2 which preceded the application of SNP and Trial #3 which followed SNP were statistically indistinguishable. The CBs had recovered their original status. The data support the hypothesis that the frequently reported NO-induced reduction in CB neural output during hypoxia is at least in part due to the reduction in ACh release. The results are consistent with a previous report in which l-arginine, an NO precursor, had the same reducing effect. Possible mechanisms are discussed.

L-arginine's effect on the hypoxia-induced release of acetylcholine from the in vitro cat carotid body.

NO is known to reduce the hypoxia-induced increase in carotid body neural activity (CBNA). Acetylcholine (ACh), a known excitatory transmitter in the cat carotid body (CB), is released during hypoxia. This study addressed the impact of an NO precursor on ACh release during hypoxia. Both CBs from nine cats were prepared for incubation, then inserted into a medium and bubbled with three consecutive gas mixtures, hyperoxic, hypoxic, and a final hyperoxic mixture. This series of exposures was performed in the absence of L-arginine, followed by the three exposures in a 1mM L-arginine medium, and followed, thirdly, in a 10mM L-arginine medium. L-Arginine significantly attenuated the hypoxia-induced release of ACh. Two post-arginine procedures suggested strongly that the reduction in the ACh release was not due to a gradual exhaustion of carotid body ACh stores over the course of the experiment. The data are consistent with those reports showing that NO donors and precursors reduce the hypoxia-induced increase in CBNA, and further support a role for ACh in the hypoxia-induced increase in CBNA.

Structural and functional differences of the carotid body between DBA/2J and A/J strains of mice.

In a previous study, DBA/2J and A/J inbred mice showed extremely different hypoxic ventilatory responses, suggesting variations in their carotid bodies. We have assessed the morphological and functional differences of the carotid bodies in these mice. Histological examination revealed a clearly delineated carotid body only in the DBA/2J mice. Many typical glomus cells and glomeruli appeared in the DBA/2J but not in the A/J mice. The size of the carotid body in the DBA/2J and A/J mice was 6.3 +/- 0.5 x 10(6) and 1.5 +/- 0.3 x 10(6) micro m(3), respectively. The area immunostained for tyrosine hydroxylase, an estimation of the glomus cell quantity, was four times larger in the DBA/2J mice than in the A/J mice. The individual data points in the DBA/2J mice segregated from those in the A/J mice. ACh increased intracellular Ca(2+) in most clusters (81%) of cultured carotid body cells from the DBA/2J mice, but only in 18% of clusters in the A/J mice. These data suggest that genetic determinants account for the strain differences in the structure and function of the carotid body.

Characterization of nicotinic acetylcholine receptors in cultured arterial chemoreceptor cells of the cat.

Neurotransmitters appear to be involved in chemotransmission of the carotid body, a major arterial chemoreceptor. Substantial data indicate that acetylcholine (ACh) is an excitatory neurotransmitter in the carotid body, regulating the excitability of afferent nerve endings and glomus cells (putative chemoreceptor cells). In this study we characterized properties of nicotinic ACh receptors (nAChRs) in cultured cat glomus cells using immunocytochemistry and whole cell patch clamp techniques. Cultured glomus cells expressed immunoreactivity for alpha3, alpha4, and beta2 subunits of nAChRs. An application of ACh elicited inward current. Nicotinic AChRs of glomus cells showed high affinity for ACh. The current-voltage relationship showed strong inward rectification at positive membrane potential. alpha-Conotoxin MII (20 nM), dihydro-beta-erythroidine (DHbetaE; 1 nM), and hexamethonium (300 microM) significantly inhibited ACh-induced current. These results indicate that cultured cat glomus cells possess functional nAChRs, and that their characteristics are consistent with those of alpha3, alpha4 and beta2 containing nAChRs.

Patch clamp study of mouse glomus cells using a whole carotid body.

Some electrophysiological characteristics of mouse glomus cells (DBA/2J strain) were investigated using an undissociated carotid body. The carotid body with major carotid arteries was placed in a recording chamber, and glomus cells were visualized with a water immersion lens combined with an infrared differential interference video camera. Patch clamp experiments revealed that voltage-gated outward current, but not inward current, was easily observed in glomus cells. Pharmacological experiments and the kinetics of the current suggest that outward current is via delayed rectifier, A type, and large conductance calcium-activated K channels. Furthermore, K current was reversibly attenuated by mild hypoxia. The results suggest electrophysiological similarities of glomus cells among the cat, the rat, and the DBA/2J mouse. The method appears useful for physiological experiments.

The impact of adenosine on the release of acetylcholine, dopamine, and norepinephrine from the cat carotid body.

Exogenously administered adenosine provokes an increase in respiration in both animal models and in man. Administered near the carotid body adenosine increases neural output from the carotid body in rats and cats. Hypoxia has the same effect. Hypoxia also provokes a release of acetylcholine (ACh), dopamine (DA), and norepinephrine (NE) from the carotid body. The present study aimed to determine the effect of exogenous adenosine on the release of ACh, DA, and NE from the carotid bodies of cats. After a recovery period (from surgery) carotid bodies were first incubated for 10 (DA, NE) or 15 (ACh) min in Eppendorf tubes containing 85 microL of a physiological salt solution equilibrated with 40% O2/5% CO2 at 37 degrees C (hyperoxia). At the end of the incubation period the medium was drawn off, and measured for ACh, DA, and NE using HPLC-ECD methods. Next 85 microL of the medium and the tubes were equilibrated with a hypoxic gas mixture (4% O2/5% CO2) and the carotid bodies were incubated for 10 (DA, NE) or 15 (ACh) min, at the end of which the medium was drawn off and measured for ACh, DA, and NE. In the ACh studies there followed a post-hypoxic hyperoxic exposure (40% O2/5% CO2). ACh tubes were then made 100 microM with respect to adenosine, and the hyperoxic, hypoxic, and post-hypoxic hyperoxic challenges were repeated. One of the two DA, NE tubes had the 100 microM adenosine from the start. Adenosine significantly increased the release of ACh, but significantly decreased the hypoxia-induced release of DA. Potential mechanisms for these changes are reviewed.

Carotid body denervation prevents fasting hyperglycemia during chronic intermittent hypoxia.

Obstructive sleep apnea causes chronic intermittent hypoxia (IH) and is associated with impaired glucose metabolism, but mechanisms are unknown. Carotid bodies orchestrate physiological responses to hypoxemia by activating the sympathetic nervous system. Therefore, we hypothesized that carotid body denervation would abolish glucose intolerance and insulin resistance induced by chronic IH. Male C57BL/6J mice underwent carotid sinus nerve dissection (CSND) or sham surgery and then were exposed to IH or intermittent air (IA) for 4 or 6 wk. Hypoxia was administered by decreasing a fraction of inspired oxygen from 20.9% to 6.5% once per minute, during the 12-h light phase (9 a.m.-9 p.m.). As expected, denervated mice exhibited blunted hypoxic ventilatory responses. In sham-operated mice, IH increased fasting blood glucose, baseline hepatic glucose output (HGO), and expression of a rate-liming hepatic enzyme of gluconeogenesis phosphoenolpyruvate carboxykinase (PEPCK), whereas the whole body glucose flux during hyperinsulinemic euglycemic clamp was not changed. IH did not affect glucose tolerance after adjustment for fasting hyperglycemia in the intraperitoneal glucose tolerance test. CSND prevented IH-induced fasting hyperglycemia and increases in baseline HGO and liver PEPCK expression. CSND trended to augment the insulin-stimulated glucose flux and enhanced liver Akt phosphorylation at both hypoxic and normoxic conditions. IH increased serum epinephrine levels and liver sympathetic innervation, and both increases were abolished by CSND. We conclude that chronic IH induces fasting hyperglycemia increasing baseline HGO via the CSN sympathetic output from carotid body chemoreceptors, but does not significantly impair whole body insulin sensitivity.

Carotid chemoreceptor development in mice.

Mice are the most suitable species for understanding genetic aspects of postnatal developments of the carotid body due to the availability of many inbred strains and knockout mice. Our study has shown that the carotid body grows differentially in different mouse strains, indicating the involvement of genes. However, the small size hampers investigating functional development of the carotid body. Hypoxic and/or hyperoxic ventilatory responses have been investigated in newborn mice, but these responses are indirect assessment of the carotid body function. Therefore, we need to develop techniques of measuring carotid chemoreceptor neural activity from young mice. Many studies have taken advantage of the knockout mice to understand chemoreceptor function of the carotid body, but they are not always suitable for addressing postnatal development of the carotid body due to lethality during perinatal periods. Various inbred strains with well-designed experiments will provide useful information regarding genetic mechanisms of the postnatal carotid chemoreceptor development. Also, targeted gene deletion is a critical approach.

Inflammation in the carotid body during development and its contribution to apnea of prematurity.

Breathing is a complex function that is dynamic, responsive, automatic and often unstable during early development. The carotid body senses dynamic changes in arterial oxygen and carbon dioxide tension and reflexly alters ventilation and plays an essential role in terminating apnea. The carotid body contributes 10-40% to baseline ventilation in newborns and has the greatest influence on breathing in premature infants who characteristically have unstable breathing leading to apnea of prematurity. In this review, we will discuss how both excessive and minimal contributions from the carotid body destabilizes breathing in premature infants and how exposures to hypoxia or infection can lead to changes in the sensitivity of the carotid body. We propose that inflammation/infection during a critical period of carotid body development causes acute and chronic changes in the carotid body contributing to a protracted course of intractable and severe apnea known to occur in a subset of premature infants.