Which of the following statements concerning the peripheral chemoreceptors is not true?
|
J Physiol. 2015 Sep 15; 593(Pt 18): 4225–4243. We asked if the type of carotid body (CB) chemoreceptor stimulus influenced the ventilatory gain of the central chemoreceptors to CO2. The effect of CB normoxic hypocapnia, normocapnia and hypercapnia (carotid body Key points
IntroductionPeripheral–central chemoreceptor interactive effects on control of breathing have been observed using animal models with isolated perfusion of the carotid body and/or central chemoreceptors in such varied conditions as eupnoea, apnoea, hypercapnia, and hypoxia (Day & Wilson, 2007,2009; Smith et al. 2007,2010; Blain et al. 2009,2010; Dempsey et al. 2012; Fiamma et al. 2013). However, the exact nature of these chemoreceptor interactions are controversial with studies in a wide variety of experimental preparations and theoretical models claiming additive, hyperadditive, or hypoadditive effects on the control of breathing (Duffin, 1990; Duffin & Mateika, 2013; Teppema & Smith, 2013; Wilson & Day, 2013). Based on these divergent findings some investigators (Wilson & Day, 2013; Guyenet, 2014) have suggested a ‘hybrid’ model as a basis for peripheral–central interaction, whereby variations in both the experimental models and in the prevailing physiological conditions and/or chemoreceptor stimuli may markedly alter the nature of the chemoreceptor interactions (also see Discussion). Accordingly, in the present study we have tested the nature of the peripheral–central interaction under novel experimental conditions consisting of hypercapnic stimulation and hypocapnic inhibition at the level of the isolated CB chemoreceptor. This represents an important advance in addressing the interaction problem for several reasons. First, comparing normoxic hypercapnia/hypocapnia to results from our prior use of hypoxia/hyperoxia at the carotid body (Blain et al. 2010) provides a test of the equivalence of the observed hyperadditive interactive effect in the presence of both major peripheral chemoreceptor stimuli, i.e. CO2 and O2. Second, perturbations in We found that, in the awake dog, specific carotid body stimulation/inhibition by means of hyper- or hypocapnia resulted in hyperadditive interaction when the central chemoreceptors were stimulated by means of increased MethodsEthical approvalExperimental protocols and the roles of each investigator involved were approved by the Animal Care and Use Committee of The University of Wisconsin-Madison School of Medicine and Public Health and were performed with strict adherence to all American Association for the Accreditation of Laboratory Animal Care International (AAALAC) and National Institutes of Health guidelines embodied in the Guide for the Care and Use of Laboratory Animals, 8th edition, published by the National Academies Press, 2011. AnimalsStudies were performed during wakefulness on 13 unanaesthetized, spayed (n = 12) or anoestrus (n = 1), adult, female, mixed-breed dogs (20–25 kg). The dogs were trained to lie quietly in an air conditioned (19–22°C) sound attenuated chamber. The dogs were housed in an AAALAC accredited animal care facility. Seven of the 13 dogs were used in the main perfusion/interaction portion of the study. Four additional dogs were used to assess the anatomical isolation of our perfusion preparation. Two additional dogs were used to further assess the ventilatory effects of unilateral CB denervation. Surgical procedures and chronic instrumentationOur preparation required two surgical procedures performed under general anaesthesia; identical anaesthetic protocols were used for both procedures. All dogs were pre-medicated with acepromazine (0.04–0.08 mg · kg−1s.c. or i.m.) and buprenorphine (0.01–0.03 mg · kg−1i.v. or i.m.); induced with propofol (3–5 mg · kg−1i.v.); intubated immediately with a 7.5 mm ID cuffed endotracheal tube and maintained with 1–2% isoflurane in O2 with mechanical ventilation. Strict sterile surgical techniques and appropriate postoperative analgesics and antibiotics were used. No agents producing neuromuscular blockade (‘paralysing’ agents) were used. In the first procedure, using a ventral abdominal midline approach, 12 of 13 dogs were subjected to ovariectomy and hysterectomy, a chronic indwelling catheter was placed into the abdominal aorta via a branch of the femoral artery, and bipolar electromyogram (EMG) recording electrodes were installed into the costal diaphragm. The catheter and EMG wires were tunnelled subcutaneously to the dorsal aspect of the thorax where they were exteriorized a few centimetres caudal to the scapulae. After a recovery interval of at least 3 weeks a second procedure was performed in which the left carotid body (CB) was denervated (CBD). The right carotid sinus was equipped with a vascular occluder and catheter to permit extracorporeal perfusion of the reversibly isolated carotid sinus–carotid body. A chronic indwelling catheter was also placed in the cranial vena cava via a branch of the jugular vein. Catheters were tunnelled subcutaneously to the lateral aspect of the dog’s neck where they were exteriorized. Dogs recovered for at least 4 days before study. The instrumentation was protected with a heavy nylon jacket and a padded ‘Elizabethan’ collar modified to allow normal eating and drinking. Post-operative analgesia was achieved with buprenorphine (0.005 to 0.02 mg · kg−1i.m. or i.v. q6–12 h; initial dose administered 30 min prior to the end of surgery) for the first 24 h in the first procedure and 12 h in the less invasive second procedure. We also administered one s.c. dose of carprofen (4 mg · kg−1; Pfizer, NY, USA) just after induction of anaesthesia. We continued carprofen p.o. (2.2 mg · kg−1q 12 h or 4.4 mg · kg−1q 24 h) for 3–7 days. A wide-spectrum oral antibiotic (cephalexin, Karalex Pharma, NJ, USA) 20–30 mg · kg−1p.o., BID or cefpodoxime proxetil (Pfizer, NY, USA) 10 mg · kg−1, p.o., SID or enrofloxacin (Bayer, KS, USA) 5–10 mg · kg−1, p.o., SID) was begun 24 h prior to surgery and was continued for 5–7 days post-operatively for the first procedure and 7–21 days for the second procedure. In addition, cefazolin (Sagent, IL, USA) (20–30 mg · kg−1, slow i.v. injection) was administered immediately pre-operatively and every 2 h intra-operatively. It is important to note that we used spayed female dogs (and one ovaries and uterus-intact dog which was in anoestrus) to avoid problems with ventilatory effects of fluctuating ovarian hormone levels. We do not think this introduces a significant bias; we have discussed this in detail in a previous publication (Blain et al. 2009). Carotid sinus perfusionDogs lay unrestrained on a bed within the sound attenuated chamber. The extracorporeal circuit was primed with ∼700 ml of saline, 120 ml of allogenic, packed red cell blood (Animal Blood Resources International, Dixon, CA, USA), and 2000 U of heparin and supplemented with 500–1000 U · h−1. Experimental set-up and measurementsVentilation was measured using a tight-fitting muzzle mask connected to a pneumotachograph (model 3700, Hans Rudolph, Kansas City, MO, USA) that was calibrated before each study with four known flows. The dogs were acclimated to the mask during several weeks of training prior to study. Costal diaphragm EMG (EMGdi) signals were amplified, band-pass filtered (100–1000 Hz), rectified and moving-time-averaged (BMA-931; MA-821RSP, CWE Ardmore, PA, USA). End tidal Arterial and perfusion circuit blood samples (∼0.5–1.0 ml) were analysed for pH, Ventilation and blood pressure signals were digitized (128 Hz sampling frequency) and stored on the hard disk of a PC for subsequent analysis. Key signals were also recorded continuously on a polygraph (AstroMed K2G, West Warwick, RI, USA). All ventilatory data were analysed on a breath-by-breath basis by means of custom analysis software developed in our laboratory. Experimental protocolInteractionEupnoeic ventilation and blood gases and the ventilatory response to hypercapnia via increased For CB perfusion, each test protocol consisted of a 5–7 min control period (eupnoea), during which perfusion of the CB was endogenous, i.e. systemic arterial blood. Two 1 ml blood samples were collected at ∼3–7 min for determination of control value for blood gases and pH. Then, CB perfusion was abruptly switched to the extracorporeal circuit. The dogs were perfused from the extracorporeal circuit with blood gases and pH concentrations matching a given dog’s eupnoeic values (CB normal), or with hypocapnic ( Assessing contamination of vertebral artery blood by perfusateFour dogs were instrumented as described above but underwent bilateral CB denervation. CBD was confirmed by absence of a ventilatory response to intravenous NaCN (20–30 μg · kg−1). We used these dogs to assess the significance of potential CB perfusate contamination of the blood flow to the brain via the vertebral artery. To this end, air-breathing ventilation and blood gases were assessed during carotid sinus region perfusion with hypocapnic and hypercapnic perfusates. In two of these bilaterally carotid body denervated dogs we also tested their ventilatory response to inhaled CO2 (as described above) to verify that these animals were responsive to systemic hypercapnia. KillingUpon completion of the protocol, dogs (n = 12) were humanely killed with an overdose of 200 mg propofol into the implanted i.v. catheter followed by 5 ml Beuthanasia-D Special (Schering Plough, NJ, USA) via the same catheter after full unconsciousness was achieved. One dog was enrolled in a special protocol at the University of Wisconsin that allows removal of instrumentation and subsequent adoption after a suitable period of recovery. StatisticsSlopes of the CO2 responses were determined by linear regression. Significance of the differences in group mean ventilatory responses to central hypercapnia between the three CB perfusion conditions was determined by means of one-way ANOVA. Significant ANOVAs were followed by Tukey’s HSD post hoc analyses with correction for unequal n using the harmonic mean of n(Portney & Watkins, 2009). The significance of the ventilatory effects of unilateral CBD was determined by means of one-way ANOVA. Means are presented as mean ± SD. Differences were considered significant if P ≤ 0.05. ResultsUnilateral carotid body denervationTable1 shows the effects of unilateral CBD on key ventilatory components for the seven dogs used in the main interaction portion of this study. On average, there was a modest, variable and non-significant hypoventilation ( Table 1Eupnoeic values during air breathing when the dogs were intact vs. unilateral CBD(CB not perfused) Effects of CB hypercapnia and hypocapnia on eupnoeaTime courseIn four dogs (of the 7 used in the main perfusion/interaction portion of the study) we obtained technically acceptable transitions from non-perfused to CB perfusion with hypercapnic blood during room air breathing. All four dogs increased In five dogs (of the 7 used in the main perfusion/interaction portion of the study) we obtained technically acceptable transitions over 3–7 min periods from non-perfused to CB perfusion with hypocapnic blood during room air breathing. All five dogs decreased Figure 1 shows examples of transitions in response to the hypercapnic CB perfusion in two representative dogs, one with a brisk ventilatory response to changes in Steady-stateTable3 shows the effect on steady-state, air-breathing eupnoea during specific carotid body normocapnia, hypercapnia and hypocapnia during CB perfusion. We observed that rather marked changes in CB Table 3Air breathing eupnoeic values during CB perfusion Effects of CB hypercapnia and hypocapnia on the central ventilatory response to systemic hypercapniaIn the time course response to CB hyper- and hypocapnia shown in Fig. 1, the ventilatory response to progressive increases in Figure 2 illustrates the steady-state responses of five ventilatory components in a representative dog to increased central Steady-state ventilatory responses of five key ventilatory components to increased FICO2 Data obtained during CB normocapnic ( Central CO2 response slopes for all variables during CB normocapnia, hypercapnia and hypocapnia are summarized as individual slopes and group means for all animals in Fig. 3. The central CO2 response slopes for Individual and mean ventilatory response slopes to increased systemic PaCO2 grouped by CB perfusate condition Each type of symbol represents the responses of one dog; the thick horizontal bars represent the group means; vertical dashed arrows represent ±95% confidence interval. *Mean value significantly different (*P ≤ 0.05; **P ≤ 0.01) from hypocapnic mean; †significantly different (†P ≤ 0.05; ††P ≤ 0.01) from normocapnic mean. Some symbols have been moved to the left where extensive overlap between points occurred. n = 7 for CB hypocapnia ( Figures 8 show individual steady-state response slopes to central hypercapnia for five key components of ventilation comparing only the hypercapnic and hypocapnic CB perfusion conditions. These figures make the point that hyperaddition (i.e. hypercapnic CB perfusion ventilatory response slopes steeper than hypocapnic CB perfusion ventilatory response slopes) is a consistent finding in all seven animals for most variables. The slopes of the ventilatory responses to central Ventilatory responses of DiRR to increased Fico2 during CB hypercapnic and CB hypocapnic perfusion for each of the seven dogs Regression equations and r2 values are adjacent to the regression line they pertain to (See text for details). Symbols as in Figure 4. Ventilatory responses of Regression equations and r2 values are adjacent to the regression line they pertain to. (See text for details.) Ventilatory responses of VT to increased Fico2 during CB hypercapnic and CB hypocapnic perfusion for each of the seven dogs Regression equations and r2 values are adjacent to the regression line they pertain to (See text for details). Symbols as in Figure 4. Ventilatory responses of fb to increased Fico2 during CB hypercapnic and CB hypocapnic perfusion for each of the seven dogs Regression equations and r2 values are adjacent to the regression line they pertain to (See text for details). Symbols as in Figure 4. Ventilatory responses of VT/TI to increased Fico2 during CB hypercapnic and CB hypocapnic perfusion for each of the seven dogs Regression equations and r2 values are adjacent to the regression line they pertain to (See text for details). Symbols as in Figure 4. DiscussionLimitations of our preparationUnilateral carotid body denervationOur animal preparation is awake and intact except for the required denervation of the left carotid body. So the question remains whether our preparation actually responds similarly to a completely intact, awake preparation with both carotid bodies present. Eldridge et al.(1981) showed increased phrenic nerve responses to electrical stimulation of two vs. one carotid sinus nerves in anaesthetized cats and Brooks & Tenney (1966) reported reduced ventilatory responses to hypercapnic-hypoxic combinations in two of three goats following unilateral CBD; there was no change in the third goat. We compared 26 awake dogs before and after unilateral CBD (see detailed findings in Results). During eupnoeic air-breathing, group mean VariabilityOn the one hand our use of awake animals ensures that cardiorespiratory response gains to chemoreceptor stimuli are in the physiological range and not subject to ‘neural saturation’ as might be the case in surgically reduced and/or anaesthetized preparations (Eldridge et al. 1981). On the other hand, a limitation of studying the awake animal is that ventilatory responses to hyperpnoeic stimuli are variable – both within and among animals – likely to be attributable at least in part to variable behavioural responses to chemoreceptor and ventilatory stimulation. For example, in 6 of our 7 unilaterally carotid body denervated animals, each with multiple CO2 steady state response tests, we found average coefficients of variation of 24 ± 17%, and 28 ± 11% for the slopes of the ventilatory response to CO2 for Effects of changes in systemic
In our isolated carotid body perfused model, as compared to some reduced preparations with isolated perfusions of both carotid body and brain stem (Berkenbosch et al. 1984; Day & Wilson, 2007), our increase in Duration of carotid body chemoreceptor stimulationOur current and prior studies demonstrated very similar hyperadditive effects of CB stimulation/inhibition on central CO2 ventilatory responses for alterations in either CB Hyper-, hypo- and/or additive interaction of peripheral with central chemoreceptorsThe long-standing controversy over the nature of central–peripheral chemoreceptor interaction (whether there is hyper-, hypo- and/or additive interaction of peripheral with central chemoreceptors) was recently reviewed (Forster & Smith, 2010; Smith et al. 2010; Guyenet, 2014) and also debated extensively in publications in which three sets of authors defended their positions and suggested new concepts and approaches (Duffin & Mateika, 2013; Teppema & Smith, 2013; Wilson & Day, 2013). Clearly, there are several studies using a wide variety of preparations in animals and also in humans to support additive, hyperadditive, or hypoadditive interactive effects on ventilatory control between peripheral and central chemoreceptors. We now summarize our perception of each of these positions and compare these findings and concepts to our recent and current findings which utilized two different means of stimulation and inhibition of the isolated, perfused carotid sinus region (and, therefore, carotid body chemoreceptor) in the awake canine preparation. We concentrate our analysis on published findings in unanaesthetized preparations and refer the reader to comprehensive recent reviews of this topic (Forster & Smith, 2010; Smith et al. 2010; Guyenet, 2014). HypoadditiveHypoadditive interactive effects between medullary and peripheral chemoreceptors on ventilatory responses to CO2 have been reported consistently primarily in reduced preparations and perhaps most notably in the rat (Day & Wilson, 2007,2009; Tin et al. 2012). The rat preparation has the advantage of including specific, isolated perfusions of the brain stem as well as the carotid chemoreceptor. These preparations have included anaesthesia (Tin et al. 2012) and decerebration as well as ‘reconfigured’ chemosensitive control pathways achieved via removal of vagal, hypothalamic and/or cortical inputs (Day & Wilson, 2007,2009; Guyenet, 2014). In these preparations the ventilatory responses to inhaled CO2 are commonly 10–20% of those in the intact, awake rat of the same strain (Day & Wilson, 2007,2009; Mouradian et al. 2012). We also reported hypoadditive effects of an alkaline CSF perfusate which increased the transient carotid chemoreceptor response to NaCN in the awake goat – but interpretation of this augmented transient response was confounded by a concomitant hypoventilation, CO2 retention and systemic respiratory acidosis (Smith et al. 1984). AdditiveAdditive effects between central and peripheral chemoreceptors have been reported primarily (see exception below) in intact humans using methods which rely on separation of the peripheral and central chemoreceptor responses via temporal dissociation of the ‘early’ (peripheral) and ‘late’ (central) responses to inhaled CO2(Clement et al. 1992,1995; St Croix et al. 1996; Cui et al. 2012). We suggest that any determination of the nature of peripheral–central interaction cannot be adequately quantified in these intact preparations because (a) given that an enhanced or reduced sensory input originating in the carotid chemoreceptors is transduced immediately to central CO2-sensitive neurons (Takakura et al. 2006) (i.e. the concept of chemoreceptor interdependence), then precise, complete, temporal separation of the relative responses of the peripheral and central chemoreceptors is not achievable;(b) clear separation of early- and late-phase CO2 responses is not always measurable perhaps because a significant portion of the so-called ‘late’ response to CO2 might be attributed to peripheral (rather than central) chemoreceptors (Pedersen et al. 1999; Dahan et al. 2007); and (c) the likelihood that a short-term potentiation of respiratory motor output (Eldridge & Gill-Kumar, 1980) will follow the initial peripheral chemoreceptor responses to CO2 would seriously confound attempts at assigning response gains exclusively to the different sets of receptors. These limitations indicate to us that quantification of the interdependence between peripheral and central chemoreceptors requires their anatomical isolation. HyperadditiveHyperadditive effects of peripheral chemoreceptor inhibition/stimulation on the central ventilatory CO2 responses have been demonstrated in two types of awake, cortically and vagally intact preparations. First, in several species including dogs, goats, ponies and humans, bilateral CBD (within the initial few days following CBD) results in significant hypoventilation and CO2 retention during air breathing together with a substantial reduction in the slope of either the normoxic or hyperoxic response to systemic hypercapnia (Bisgard et al. 1980; Pan et al. 1998; Rodman et al. 2001; Fatemian et al. 2003; Dahan et al. 2007) or to the ventilatory response to focal medullary acidosis (Hodges et al. 2005). Rats of several strains appear to be exceptions to this evidence for hyperadditive interaction as bilateral CBD in this species results in hypoventilation but without any decrement in the slope of the response to systemic hypercapnia (da Silva et al. 2011; Mouradian et al. 2012; Takakura et al. 2014). On the other hand, it is noteworthy that reductions in the slope of the ventilatory response to CO2 and/or reductions in air-breathing eupnoea observed following partial ablation of central chemosensitive regions in the awake rodent was further reduced in a markedly hyperadditive fashion following denervation or hyperoxic inhibition of the carotid body chemoreceptors (da Silva et al. 2011; Ramanantsoa et al. 2011; Takakura et al. 2014). The second type of awake preparation with findings that support hyperaddition is found in our essentially intact canine model with isolated, perfused carotid bodies. Combining previous and current findings using this preparation revealed hyperadditive effects on the central CO2 response whether the carotid chemoreceptor was exposed to a combination of hyperoxia and hypocapnia vs. normocapnic hypoxia (Blain et al. 2010) or to normoxic hypercapnia vs. normoxic hypocapnia (present study). Further support for hyperaddition is provided by experiments in the sleeping canine model utilizing transient hypocapnia. Transient hyperventilation-induced systemic hypocapnia in this model resulted in apnoea and periodic breathing with an apnoeic threshold approximately 5 mmHg below eupnoeic On the other hand, investigators using a vascularly isolated, perfused carotid body chemoreceptor preparation in the unanaesthetized goat reported additive effects of changes in CB Hybrid modelsGiven the divergent types of peripheral–central interactions reported to date across many species and preparations it is not surprising that ‘hybrid’ models have been proposed to account for these divergent findings (Nuding et al. 2009; Wilson & Day, 2013; Guyenet, 2014). The finding that different experimental models, ranging from intact humans and animals to anatomically reduced preparations, are producing these divergent findings is not surprising – especially given the markedly ‘reconfigured’ chemoreceptor networks and substantial departures from physiological chemosensitivity present in some of these models (see above). On the other hand, a proposal that the nature of peripheral–central interactions may be critically dependent upon changing ‘physiological states’ and conditions – even within the same experimental preparation – has considerable merit and is especially relevant to interaction effects on ventilatory control (Nuding et al. 2009; Wilson & Day, 2013; Guyenet, 2014). For example, Guyenet (2014) predicts that a key determinant of the type of interaction might be found in the resting membrane potentials of the central CO2-sensitive neurons, which in turn would be influenced by the relative strengths of the inhibitory vs. excitatory input from the respiratory pattern generator, carotid chemoreceptors, and various links within the chemoreceptor pathway. Accordingly, it is suggested that the net effect of these inputs might depend upon such factors as the background states, i.e. wakefulness vs. sleep vs. exercise, prevailing levels of systemic and/or central CO2 and hypoxia, whether ventilatory drive is operating above or below eupnoea, the duration of chemoreceptor stimulation, and the central ‘amplification’ of chemoreceptor input occurring at one or more sites within the medulla, pons, or hypothalamus (Reddy et al. 2005; Schultz & Li, 2007; Nuding et al. 2009; Duffin, 2010; Wilkinson et al. 2010; Wilson & Day, 2013; Dempsey et al. 2014; Guyenet, 2014). So while future studies may well uncover an effect of one or more of these potential influences on the nature of peripheral chemoreceptor interaction, we propose that current evidence (as described above) is strongly supportive of hyperadditivity as the dominant form of peripheral–central interaction. At present we must limit this proposal to the specific conditions of acute CB inhibition/stimulation (<30 min) during quiet wakefulness or sleep, with ventilation operating within a physiological range either above or below eupnoea, and when the multiple links within the chemosensory control system are anatomically intact, chemosensitivity gains are within the physiological range, and sensory inputs from the carotid chemoreceptors are driven by a wide range of O2 and/or CO2. It is unknown if long-term CB inhibition/stimulation via variations in CB AcknowledgmentsWe gratefully acknowledge the outstanding technical contributions of Theodore (Ted) Snyder and Anthony Jacques. GlossaryAdditional informationCompeting interestsNone declared. Author contributionsC.A.S., G.M.B. and J.A.D. carried out the conception and design of research; C.A.S., G.M.B. and K.S.H. performed the experiments; C.A.S., G.M.B. and K.S.H. analysed the data; C.A.S., G.M.B., K.S.H. and J.A.D. interpreted the results of the experiments; C.A.S. prepared the figures; C.A.S., G.M.B. and J.A.D. drafted the manuscript; C.A.S., G.M.B., K.S.H. and J.A.D. edited and revised the manuscript; C.A.S., G.M.B., K.S.H. and J.A.D. approved the final version of the manuscript. FundingThis work was supported by grants HL50531 and HL15469 from the National Heart, Lung, and Blood Institute of the National Institutes of Health, and a grant from the American Heart Association. References
Articles from The Journal of Physiology are provided here courtesy of The Physiological Society |
