HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

This Article Full Text (PDF) Alert me when this article is cited Alert me if a correction is posted Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via Google Scholar Google Scholar Articles by Richerson, G. B. Search for Related Content PubMed PubMed Citation Articles by Richerson, G. B. Related Collections Exchange of Views

¹ Departments of Neurology and Cellular & Molecular Physiology, Yale University School of Medicine, New Haven, CT, USA

This issue of Experimental Physiology contains a review by Guyenet et al. (2005) on central respiratory chemoreceptors (CRCs) in which he makes two major conclusions: (1) glutamatergic neurones of the retrotrapezoid nucleus (RTN) are CRCs; and (2) medullary serotonergic neurones are not. In an accompanying review (Richerson et al. 2005), we discuss why we believe the existing evidence does not yet support either of these opinions. We will not repeat the arguments presented there, except to point out that the authors have made many of the assumptions identified as having not been validated, including that the primary stimulus is pH, and that CRCs (1) are localized on the surface of the ventrolateral medulla (VLM); (2) must have a response to acidosis that is nearly as large as the system as a whole; (3) project exclusively to the respiratory central pattern generator; and (4) must be stimulated by respiratory acidosis (not inhibited). Here, we will discuss some additional issues that Guyenet et al. raise in their review.

Both of our groups agree, as do probably all investigators in the field (Richerson, 1998; Putnam et al. 2004), that a cell must be intrinsically chemosensitive to be defined as a CRC. It is curious then that Guyenet et al. advance the hypothesis that glutamatergic RTN neurones are chemoreceptors, since they concede that it is possible that they ‘are not intrinsically pH-sensitive but, instead, owe their pH response to the release of a transmitter by nearby non-neuronal cells.’ We agree that it is unclear whether RTN neurones are intrinsically pH sensitive, but given that they may be stimulated by other cells releasing neurotransmitter, these other cells could also be neurones synapsing onto RTN neurones. The fundamental problem is that it is not possible to determine whether the response of a neurone is intrinsic or synaptic with the type of in vivo approach used by Mulkey et al. (2004). Thus, it seems premature to make any conclusions about a role as chemoreceptors until it is first established that these neurones meet the most fundamental criterion for a CRC: intrinsic chemosensitivity.

Perhaps more important than whether RTN neurones are intrinsically chemosensitive is whether their very large response to hypercapnia in vivo is due entirely to intrinsic mechanisms. If only a small component of the in vivo response is intrinsic, then the conclusion that RTN neurones are ‘the long sought-after’ chemoreceptors becomes less compelling. To support their conclusion that RTN neurones are extremely chemosensitive, Guyenet et al. state that there is ‘mounting evidence that RTN neurones are much more robustly activated by pH in vitro than neurones located elsewhere.’ This statement is based on two sources of data. The first source was their own recordings from brain slices (Mulkey et al. 2004; Guyenet et al. 2005). However, pH was changed from 7.5 to 6.9 in those experiments; such a large change that it may not be relevant to respiratory control, and much larger than that used by other investigators in the field – typically 7.4–7.2 (Wang et al. 2002; Putnam et al. 2004). If the response of RTN neurones is linear, it could be predicted based on the data shown by Guyenet et al. that a decrease in pH from 7.4 to 7.2 would increase the firing rate of RTN neurones from about 0.4 Hz to 1.0 Hz at room temperature. This is less sensitive than other CRC candidates that have typically been studied at room temperature (Wang et al. 2002; Putnam et al. 2004). The second source of data for their statement is a review (Putnam et al. 2004) that described preliminary data from three RTN neurones. This work has now been expanded (Putnam et al. 2005), and the response of RTN neurones is smaller than originally estimated, with a chemosensitivity index of 200%. While this degree of chemosensitivity is larger than that of some other candidates for CRCs (Putnam et al. 2004), it is smaller than that of midline raphé serotonergic neurones, which have a mean chemosensitivity index of 300% (Wang et al. 2002; Richerson, 2004). In addition, synaptic inputs were not completely blocked in either brain slice study of RTN neurones (Putnam et al. 2005; Mulkey et al. 2004), so it is not known how much of their response in vitro is intrinsic. Finally, the conclusion that ‘the dynamic range of [the response of RTN neurones to pH in vitro at 35°C]... approaches values found in vivo’ is not valid, because the stimuli used under these two conditions were very different. The response in brain slices was induced by a change in pH of 0.6 units. In contrast, the response in vivo was induced by an increase in end-tidal CO₂ of 10%, which is approximately 80 mm Hg, which only decreases brain pH by 0.1–0.15 units (Richerson et al. 2005). If the response in slices is linear, then an equivalent change in pH (from 7.4 to 7.25–7.3) would only lead to an increase in firing rate from 2 Hz to 4– 5 Hz at 35°C in vitro, rather than the increase from 0 Hz to 14 Hz seen in vivo. Therefore, these neurones actually do not have a uniquely large response to pH in vitro, and some of this response may not be intrinsic. Indeed, if the argument is to be made that a large intrinsic response to pH makes a neurone more likely to be a CRC, then serotonergic neurones are better candidates (Richerson, 2004; Wang et al. 2002). However, there is no evidence for this assumption, although the magnitude of the response may be an important factor (Putnam et al. 2004). Thus, the evidence from the study of Mulkey et al. (2004) does not support the conclusion that glutamatergic RTN neurones have uniquely large intrinsic chemosensitivity in vivo or in vitro.

The authors discuss some other lines of evidence in support of the hypothesis that RTN neurones are uniquely qualified to be the CRCs of their ‘specialized chemoreceptor theory’, including: (1) data from older physiological studies highlighting the importance of the ventral medullary surface in respiratory control; and (2) hypercapnia-induced Fos expression. However, the authors themselves dismiss these two approaches as having inherent limitations, for reasons with which we agree. The authors also suggest that glutamatergic neurones of the RTN are well-equipped to be CRCs because they are weakly modulated by the respiratory rhythm, but CO₂-sensitive neurones in the raphé and locus coeruleus also have weak respiratory modulation. Thus, none of these other lines of evidence are compelling by themselves without strong evidence for a high degree of intrinsic chemosensitivity of RTN neurones. Since this latter evidence does not yet exist, it must be concluded that RTN neurones have not been shown to fulfil the requirements of the specialized chemoreceptor theory.

Is the specialized chemoreceptor theory still a realistic possibility? It has been the dream of many investigators in this field to find neurones in the brainstem with an intrinsic response to changes in CO₂ that is just as large as ventilation at the whole animal level. However, no one has found them, despite hundreds searching over several decades, including many that have looked in and around the RTN. Maybe this expectation is unrealistic. Although Guyenet et al. state that the other possibility is that chemosensitivity is distributed among ‘countless neurones’ throughout the brainstem, there are many other alternatives. First, there may only be one set of CRCs, but they may have a smaller intrinsic response than we expect. This scenario would still meet the requirements for a specialized chemoreceptor theory. Serotonergic neurones could play this role. Their intrinsic response is not quite as large as ventilation at the whole animal level, but it is close (Wang et al. 2002), and it might be larger if in vitro experiments were performed at body temperature. Second, there may be a wide distribution of completely redundant CRCs, but restricted to only specific subsets of neurones. Intrinsic chemosensitivity has not been shown to be as ubiquitous as suggested by Guyenet and colleagues (Wang & Richerson, 2000; Richerson, 1998; Putnam et al. 2004). Third, there may be several sets of CRCs, each with a different role. For example, some may only respond during sleep or others only during extreme pH changes (Feldman et al. 2003). Fourth, there may be one dominant group of CRCs under normal conditions and in the event of pathologic loss of these CRCs, others can immediately compensate, analogous to generation of the heartbeat. Fifth, there may be plasticity in the system, where one set of CRCs dominate under normal conditions, and sometime after loss of the dominant set others become functional, as in the case of the normally quiescent aortic chemoreceptors becoming functional after bilateral loss of the dominant carotid chemoreceptors in piglets (Serra et al. 2002). We favour the fourth or fifth alternative, where a subset of medullary serotonergic neurones serve as the dominant chemoreceptors. However, the existing data are inconclusive, and any of the alternatives described (and others) remain possible. Regardless of which one is true, it is likely that the intrinsic response of individual CRCs is smaller than that of the network as a whole. Otherwise, someone would have likely found hypersensitive neurones after all these years of looking. If CRCs are not hypersensitive, respiratory output is probably large due to amplification within the network (Richerson et al. 2005).

The second major point of Guyenet et al. is that serotonergic neurones are not CRCs. The data in support of this conclusion are limited, and in our accompanying review we discuss some of the deficiencies in those data. There is also a body of data supporting the opposite conclusion that is larger and more compelling than suggested by Guyenet et al. We refer the reader to a recent review for a more complete description of this evidence (Richerson, 2004).

One of the key elements of the authors' argument against a role of serotonergic neurones as CRCs is that ‘5-HT neurones, regardless of location... are largely unresponsive to hypercapnia in vivo.’ However, as discussed in our accompanying review, there are only two published papers (Mulkey et al. 2004; Veasey et al. 1995) reporting a small number of recordings in vivo from a restricted subset of medullary serotonergic neurones, and there are significant methodological problems in one of those studies (Mulkey et al. 2004). In the only study from unanaesthetized animals, Veasey et al. (1995) found that 6 of 27 neurones in the medullary raphé increased their firing rate in response to hypercapnia. Six is indeed only a ‘few CO₂-responsive neurones’, but it represents 22% of the recordings. It is possible that this is a large percentage of those that project to respiratory neurones. While it is correct that Veasey et al. (1995)‘did not verify that the few CO₂ responsive neurones encountered were serotonergic’, it has been shown by Mason (1997) (using juxtacellular labelling) that serotonergic neurones of the medullary raphé can be identified in vivo with a > 90% accuracy based on firing rate and firing pattern alone. Guyenet's group verified that these two criteria are accurate in identifying VLM serotonergic neurones 89% of the time (Mulkey et al. 2004). Veasey et al. (1995) used even more rigid criteria to define neurones as serotonergic: regular firing rate; wide action potentials; and suppression of firing during REM sleep. These neurones also had a low firing rate, albeit slightly higher than in anaesthetized animals. While it is conceivable that some nonserotonergic neurones could have been misclassified, the false positive rate would have certainly been less than the 10% error using only two criteria. Thus, the study by Veasey et al. (1995) has clearly established that some serotonergic neurones are responsive to hypercapnia in vivo. Since these experiments are so heroic, the primary problem is that only a small number of recordings have been made from unanaesthetized, behaving animals, and only from a restricted subset of serotonergic neurones.

Guyenet suggests that there is a consensus that ‘serotonergic neurones ... in the medullary raphé[are CRCs] during sleep.’ However, this question has not been settled. There is a single paper showing that ventilation increases in rats in response to CO₂ microdialysis during sleep, but not during wakefulness (Nattie & Li, 2001). However, using the same approach in goats, Hodges et al. (2004b, 2004a) found that CO₂ microdialysis in the raphé stimulates ventilation during wakefulness, but not during sleep. Nattie et al. (2004) have also shown that lesions of the raphé blunt the hypercapnic ventilatory response of rats in vivo during both sleep and wakefulness. Thus, the interaction between sleep and chemoreception in the raphé has not yet been clearly defined. In addition, in contrast to the statement that ‘the effect of CO₂ on these raphé cells disappeared during sleep’, Veasey et al. (1995) only saw a reduced response during sleep, not complete loss of the response, and this was based on only four neurones. If this finding is replicated, we would suggest that blunting of the response during sleep actually supports, not refutes (as concluded by the authors), the hypothesis that serotonergic neurones are CRCs, because the reduced chemosensitivity of these neurones during sleep could be the underlying cellular mechanism for the hypoventilation and blunting of the hypercapnic ventilatory response that occur normally during sleep in vivo. Finally, the authors conclude that the reduced hypercapnic response of serotonergic neurones during sleep ‘suggests that their activation by hypercapnia could have been due to behavioural changes.’ However, it is not clear what behavioural changes Guyenet et al. are referring to, since Veasey et al. (1995) controlled for changes in sleep state, and ‘trials were rejected if the cat's body position was altered during the trial.’ This also ignores the fact that serotonergic neurones have a high degree of intrinsic chemosensitivity, which is a more parsimonious explanation for why they respond to hypercapnia in vivo.

The conclusion that ‘many fewer medullary neurones are intrinsically responsive to CO₂in vivo than might have been anticipated from prior experimentation in vitro’ presupposes that chemosensitivity can be defined as intrinsic in vivo. Even if it were possible to define neurones as intrinsically chemosensitivity using an in vivo approach, the authors' conclusion is based on a limited number of recordings from a subset of serotonergic neurones whose chemosensitivity has not yet been studied in vitro. It is difficult to understand how results from a single group of neurones could be used to make a conclusion about all other CRC candidates. While it is conceivable that intrinsic chemosensitivity could somehow be turned off in vivo, there are no data to support this idea. Instead, the intrinsic response of neurones identified in vitro is likely to be at play in vivo.

The review by Guyenet et al. brings up a number of important points that remain to be resolved. For example, what is the role of chemosensitivity in serotonergic neurones that are not involved in the control of breathing? We have proposed that they may modulate nonrespiratory brain functions normally influenced by hypercapnia, such as arousal, limbic function and autonomic output (Richerson, 2004). There is no direct evidence for this, but the veracity of the hypothesis that some serotonergic neurones are CRCs does not depend upon whether all serotonergic neurones are chemoreceptors. Another question is why can an increase in ventilation be induced by focal acidosis in so many different nuclei? This effect is much more selective than suggested by Guyenet et al. (Feldman et al. 2003), but it still remains to be determined what cellular mechanisms lead from focal acidosis to an increase in ventilation, and what relationship this has to normal chemoreception. The relative role of each of the CRC candidates also remains to be defined, as does the effect of state (awake versus sleep) on the contribution from each one.

Thus, there are many questions that remain. Glutamatergic neurones of the RTN are viable candidates for CRCs, but the support for this is based on a single paper (Mulkey et al. 2004), and these neurones have not yet met one of the most important criteria for chemoreceptors: intrinsic chemosensitivity. The limited available data are not compelling that RTN neurones are uniquely important. Serotonergic neurones are also good candidates for CRCs, with support from many different investigators using a variety of approaches (Richerson, 2004), but there remain many important issues that need to be resolved before it can be concluded that they are CRCs. Thus, rather than using either type of neurone as a litmus test, or a yardstick to which all other CRC candidates must measure up to, it is prudent to obtain more data on all of the candidates before any of them are included or excluded as CRCs. What is exciting is that we have progressed to the point of identifying specific CRC candidates based on their neurotransmitter content. Regardless of the outcome of future studies on the current CRC candidates, we will ultimately arrive at a more complete understanding of the molecular, cellular and network mechanisms of chemoreception, and of human diseases in which breathing is abnormal.

References

Feldman JL, Mitchell GS & Nattie EE (2003). Breathing: Rhythmicity, plasticity, chemosensitivity. Ann Rev Neurosci 26, 239–266.[CrossRef][Medline]

Guyenet PG, Stornetta RL, Bayliss DA & Mulkey DK (2005). Retrotrapezoid nucleus: a litmus test for the identification of central chemoreceptors. Exp Physiol, Press in press.

Hodges MR, Klum L, Leekley T, Brozoski DT, Bastasic J, Davis S, Wenninger JM, Feroah TR, Pan LG & Forster HV (2004a). Effects on breathing in awake and sleeping goats of focalacidosis in the medullary raphé. J Appl Physiol 96, 1815–1824.[Abstract/Free Full Text]

Hodges MR, Martino P, Davis S, Opansky C, Pan LG & Forster HV (2004b). Effects on breathing of focal acidosis at multiple medullary raphe sites in awake goats. J Appl Physiol 97, 2303–2309.[Abstract/Free Full Text]

Mason P (1997). Physiological identification of pontomedullary serotonergic neurons in the rat. J Neurophysiol 77, 1087–1098.[Abstract/Free Full Text]

Mulkey DK, Stornetta RL, Weston MC, Simmons JR, Parker A, Bayliss DA & Guyenet PG (2004). Respiratory control by ventral surface chemoreceptor neurons in rats. Nat Neurosci 7, 1360–1369.[CrossRef][Medline]

Nattie EE & Li A (2001). CO2 dialysis in the medullary raphe of the rat increases ventilation in sleep. J Appl Physiol 90, 1247–1257.[Abstract/Free Full Text]

Nattie EE, Li A, Richerson GB & Lappi DA (2004). Medullary serotonergic neurons and adjacent neurons that express neurokinin-1 receptors are both involved in chemoreception in vivo. J Physiol 556, 235–253.[Abstract/Free Full Text]

Putnam RW, Filosa JA & Ritucci NA (2004). Cellular mechanisms involved in CO2 and acid signaling in chemosensitive neurons. Am J Physiol Cell Physiol 287, C1493–C1526.[Abstract/Free Full Text]

Putnam RW, Ritucci NA, Erlichman JS & Leiter JC (2005). The effects of hypercapnia on membrane potential (Vm) and intracellular pH (pHi) in neurons and astrocytes from the retrotrapezoid nucleus (RTN). FASEB Abstract. 19, 369.21.

Richerson GB (1998). Cellular mechanisms of sensitivity to pH in the mammalian respiratory system. In Ph and Brain Function , ed. Kaila K & Ransom BR.), pp. 509–533. John Wiley & Sons, Inc, New York , NY .

Richerson GB (2004). Serotonin neurons as CO2 sensors that maintain pH homeostasis. Nature Rev Neuroscience 5, 449–461.

Richerson GB, Wang W, Hodges MR, Dohle CI & Diez-Sampedro A (2005). Homing in on the specific phenotype (s) of central respiratory chemoreceptors. Exp Physiol, 90, 259–269.[Abstract/Free Full Text]

Serra A, Brozoski D, Hodges M, Roethle S, Franciosi R & Forster HV (2002). Effects of carotid and aortic chemoreceptor denervation in newborn piglets. J Appl Physiol 92, 893–900.[Abstract/Free Full Text]

Veasey SC, Fornal CA, Metzler CW & Jacobs BL (1995). Response of serotonergic caudal raphe neurons in relation to specific motor activities in freely moving cats. J Neurosci 15, 5346–5359.[Abstract]

Wang W, Bradley SR & Richerson GB (2002). Quantification of the response of rat medullary raphe neurones to independent changes in pHo and PCO2. J Physiol 540, 951–970.[Abstract/Free Full Text]

Wang W & Richerson GB (2000). Chemosensitivity of nonrespiratory rat CNS neurons in tissue culture. Brain Res 860, 119–129.[CrossRef][Medline]

This Article Full Text (PDF) Alert me when this article is cited Alert me if a correction is posted Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via Google Scholar Google Scholar Articles by Richerson, G. B. Search for Related Content PubMed PubMed Citation Articles by Richerson, G. B. Related Collections Exchange of Views