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

This Article Abstract Full Text (PDF) Free Submit a response Alert me when this article is cited Alert me when eLetters are posted Alert me if a correction is posted Citation Map 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 HighWire Citing Articles via Google Scholar Google Scholar Articles by Li, A. Articles by Nattie, E. E. Search for Related Content PubMed PubMed Citation Articles by Li, A. Articles by Nattie, E. E.

Journal of Applied Physiology
Vol. 83, No. 2, pp. 420-428, August 1997
CONTROL OF BREATHING, CIRCULATION, AND TEMPERATURE

Focal central chemoreceptor sensitivity in the RTN studied with a CO₂ diffusion pipette in vivo

Aihua Li and Eugene E. Nattie

Department of Physiology, Dartmouth Medical School, Lebanon, New Hampshire 03756-0001

ABSTRACT

INTRODUCTION

METHODS

RESULTS

DISCUSSION

ACKNOWLEDGEMENTS

FOOTNOTES

REFERENCES

---

ABSTRACT

Li, Aihua, and Eugene E. Nattie. Focal central chemoreceptor sensitivity in the RTN studied with a CO₂ diffusion pipette in vivo. J. Appl. Physiol. 83(2): 420-428, 1997.We describe and use a CO₂ diffusion pipette to produce a quickly reversible focal acidosis in the retrotrapezoid nucleus region of the rat brain stem. No tissue injection is made. Instead, artificial cerebrospinal fluid (aCSF) equilibrated with CO₂ circulates within the micropipette, providing a source for continued CO₂ diffusion into the tissue from the pipette tip. Tissue pH electrodes show the acidosis is limited to 500 µm from the tip. In controls (aCSF equilibrated with air), 1-min pipette perfusions increased tissue pH slightly and decreased phrenic nerve amplitude. In moderate- and high-CO₂ groups (aCSF equilibrated with 50 or 100% CO₂), 1-min perfusions significantly decreased tissue pH and increased phrenic nerve amplitude in a dose-dependent manner. The responses developed and reversed within minutes. Compared with our prior use of medullary acetazolamide injections to produce a focal acidosis, in this approach the acidosis 1) arises and reverses quickly and 2) its intensity can be varied. This allows study of sensitivity and mechanism. We conclude from this initial experiment that retrotrapezoid nucleus region chemoreceptors operate within the normal physiological range of CO₂-induced tissue pH changes.

ventral medulla; control of breathing; carbon dioxide sensitivity; retrotrapezoid nucleus

---

INTRODUCTION

RESULTS OF EXPERIMENTS performed in vitro and in vivo indicate a widespread distribution of ventilatory chemoreceptor sites within the brain stem (2, 6, 7, 20, 24, 25). In brain slice preparations, acidosis excites neurons in locus ceruleus (24), the nucleus tractus solitarius (NTS) (7, 28), and the caudal raphe (25). Whether these CO₂-responsive neurons do in fact stimulate respiratory output is unproven. In neonatal rat isolated brain stem preparations, neurons near the ventral medullary surface are excited by acidosis, as is the output of this isolated in vitro system (10, 11). In experiments in vivo, microinjection (1 nl) of the carbonic anhydrase inhibitor acetazolamide (AZ) produces a focal tissue acidosis and a respiratory response. This approach has identified central chemoreception in regions near the ventral medullary surface (6), in keeping with traditional ideas of central chemoreceptor locations (14, 26), more dorsally in the regions of the locus ceruleus, in the NTS (6), in the midline caudal medullary raphe (2), and in the rostral aspect of the nucleus ambiguus or ventral respiratory group (VRG) (20).

The AZ injection technique, useful in the evaluation of central chemoreceptor locations in vivo, has limitations, the most important of which are the difficulty in controlling the intensity of the stimulus and its long duration (20). Thus one cannot evaluate threshold and sensitivity of a given chemoreceptor site or consider mechanistic hypotheses because it is virtually impossible to test multiple responses in a single animal.

To overcome these limitations, we have developed a CO₂ diffusion pipette that produces a focal acidosis without any tissue injection. The acidosis develops and reverses quickly, and the degree of acidosis can be controlled by the concentration of CO₂ perfused in the pipette. In this paper, we describe this pipette, show the size of the region of focal acidosis produced in the retrotrapezoid nucleus (RTN) of the anesthetized rat, and show the effects on phrenic activity of three different concentrations of CO₂. We test the following hypotheses. 1) Central chemoreceptors, detected in the RTN by AZ injections, respond specifically to the focal acidosis and will, therefore, also be detected by the CO₂ diffusion pipette. 2) The response of such RTN chemoreceptors occurs within a physiological range of tissue pH changes and is continuous in this stimulus range; i.e., there is no threshold. 3) The phrenic responses occur in a temporal relationship with the onset of tissue pH changes and, with brief stimulation, will recover with a time course like that of tissue pH.

---

METHODS

General preparation. These experiments were reviewed and judged acceptable by the Dartmouth College Institutional Animal Care and Use Committee. Adult male Sprague-Dawley rats (290-410 g) were used in all experiments. With the rat in an enclosed box, 2.0% halothane in oxygen was administered to initially anesthetize the animal. The trachea was cannulated, and catheters were placed in the femoral artery and vein. Urethan (550 mg/kg) and -chloralose (60 mg/kg) were injected into the femoral vein over several minutes while at the same time the inspired halothane concentration was decreased. Gallamine triethiodide (3 mg/kg) was used to paralyze the rats, and supplemental doses were administered as needed. The rats were artificially ventilated (rodent ventilator model 683, Harvard Instruments) with 100% O₂. Arterial blood pressure was monitored with a strain gauge connected to the femoral artery catheter, rectal temperature was maintained at 37°C by using a feedback system, and end-tidal PCO₂ was monitored with a CO₂ analyzer (end-tidal CO₂ monitor model Lifespan 100, Biochem) and maintained at any set value by manually controlling the ventilator rate and/or tidal volume. Bilateral thoracotomies were performed, and a positive end-expiratory pressure of 3-5 cmH₂O was maintained during the experiment.

Bilateral vagotomy was performed before the ventral medullary surface was exposed. The phrenic nerve was isolated and cut, and the central end was placed on bipolar electrodes and covered with Wacker Sil-Gel (an insulating medium). Phrenic nerve activity was amplified (model BMA 831 amplifiers, Charles Ward), integrated (Paynter filter in an MA-821 moving averager, Charles Ward), and displayed on a storage oscilloscope. Integrated phrenic nerve amplitude (PNA), arterial blood pressure, and end-tidal PCO₂ were recorded on a strip-chart recorder (model MFE 1400, Gould). An on-line program using a 12-bit processor and a 150- to 200-Hz sampling rate calculated, for experimenter-determined sequences of breaths, the amplitude of the integrated phrenic activity, phrenic burst frequency, blood pressure, and end-tidal PCO₂.

Blood pressure, the frequency of phrenic discharge, and PNA were continuously monitored during the experiment to monitor the depth of anesthesia. Increases in respiratory frequency or blood pressure that could not be attributed to a microinjection or that occurred in response to a noxious stimulus, the pinch of a hind paw, were viewed as signs that the animal needed additional anesthesia, which was given in the form of one-quarter to one-third of the initial dose.

CO₂ diffusion pipette. Figure 1 shows a diagram of the diffusion pipette, which was an outgrowth and modification of similar systems used by Schlaefke et al. (26), Issa and Remmers (10), and Erlichman and Leiter (8). Double-barreled borosilicate capillary tubing (1.5 mm OD, Frederick Haes) was pulled with vertical pipette puller (Narishige), and the tip was broken back to a total diameter of ~80-100 µm. One barrel was used for the pH electrode and the other for the diffusion pipette. A flexible and durable coated quartz capillary (160-200 µm; Polymicro Technologies) or polyimide capillary tubing (World Precision Instruments) was inserted into the diffusion pipette. Its proximal end, connected to a polyethylene-10 tubing for perfusion, passed through the perfusion port of a pipette holder (model A003-2, E. W. Wright). A second port was used for suction. In practice, suction was rarely needed because the perfusion was usually sufficient to push the fluid through the inside pipette and out via the second port of the pipette holder. The resistance to flow via this exit pathway was smaller than that of the pipette tip. The pH electrode barrel was silanized in an oven with 5% dimethyldichlorosaline (40136, Fluka) in xylene (80-100°C) overnight. The pipette was dipped in Hydrogen Ionophore II-Cocktail A (95297, Fluka) until the H⁺ ion exchanger was drawn up into the barrel a distance of 500-1,000 µm. Then the barrel was backfilled with 100 mM NaCl (pH 6.1). The pH barrel was coupled to a potentiometer (model EA 920, Orion Research) by using a platinum-iridium wire. Single-barrel pH pipettes were constructed the same way. All electrodes were calibrated in vitro before and after the experiment by using precision buffer solutions with pH values of 4, 6, 7, and 10 at 24°C. In addition, the electrode responses to the initial and final whole animal CO₂-response tests were compared as an in vivo calibration. To allow comparison of tissue pH data among animals and groups, the responses in each animal were normalized to the response of that electrode in vivo to changes in end-tidal CO₂ from 4 to 9%.

---

Fig. 1. Schematic view of tip of CO₂ diffusion pipette. Right barrel is diffusion pipette; left barrel is pH electrode. Arrows show direction of perfusion of CO₂-rich artificial cerebrospinal fluid (CSF) through polyimide capillary and then out via glass pipette. CO₂ diffusion into tissue occurs at tip.
[View Larger Version of this Image (20K GIF file)]

---

Experimental protocol. All animals were first tested for their responsiveness to inspired CO₂. Baseline end-tidal PCO₂ was set just above (5 Torr) the apneic threshold while the animal was ventilated with 100% O₂. To determine the ventilatory response to CO₂, the end-tidal CO₂ was increased in ~7-Torr steps by introducing controlled amounts of 100% CO₂ into the inspiratory circuit while keeping ventilator frequency and tidal volume constant. Responses were measured when PNA had stabilized (3-5 min at any CO₂ level). When the animal had completely recovered from the CO₂ challenge and showed stable baseline phrenic activity, the CO₂ diffusion pipette and the distant pH electrode were inserted by using a hand-operated micromanipulator. The precise locations of the pipette and pH electrodes were not ascertained until postmortem examination of medullary slices. With ventral medullary exposure, we can virtually always place the pipette tip within the RTN region on the basis of the use of surface landmarks. A 1-min period of perfusion of the pipette was then made by using artificial cerebrospinal fluid (aCSF) equilibrated with 1) air, 2) 50% CO₂, or 3) 100% CO₂. The composition of the aCSF was (in mM) 152 sodium, 3.0 potassium, 2.1 magnesium, 2.2 calcium, 26 chloride, and 25 bicarbonate. The calcium was added after the aCSF was warmed to 37°C and equilibrated with 5% CO₂. We monitored the pH of the solution to ensure that each equilibration was reliable. The effects on blood pressure, phrenic discharge frequency, and PNA were observed. When the response to between one and three such pipette perfusions was completed, another CO₂ test was performed. CO₂ responses were normalized to percentage of the maximum; CO₂ diffusion responses were normalized to percentage of baseline. Normalization to percentage of maximum does not alter the conclusions.

At the end of each experiment, the rat was killed and the brain stem was quickly and carefully removed and placed into 4% paraformaldehyde overnight. After 24 h in 30% sucrose, it was placed rostral side up in dry ice and stored at 24°C. Sections (50-µm thick) were cut in a cryostat (Cryocut model 1800, Reíchert-Jung) and examined for the location of the CO₂ diffusion pipette and the pH electrodes. Alternate sections were placed on separate slides, one for identification of fluorescence from the beads to show the location of the pipette and pH electrode tips and the other for staining with cresyl violet to identify anatomic landmark structures, e.g., the facial nucleus. We dipped the double-barreled pipette into fluorescein-fluorescent microbeads and the distal pH electrode into rhodamine-fluorescent microbeads before putting them into the brain. Fast green dye was mixed in with the solutions perfused within the pipette; the presence of fast green in the sections told us if any fluid had left the pipette. Each section was viewed through a video camera and digitized by using an image-analysis system (Image Pro). With this system we could place, on the cresyl violet-stained section, the sites of fluorescence that marked the location of the pipette tips. We then created a computer-modified image that represents two adjacent sections, one with fluorescence showing the beads that came off the pipette tip surface into the tissue to mark the location and the other with cresyl violet staining (see Fig. 4). In Fig. 2 we show diagrammatically, by using sections from an atlas (22), the locations of the pipette tips for all the experiments.

---

Fig. 4. Typical response, in terms of of integrated phrenic nerve amplitude (PNA; ) and tissue pH at pipette center (), is shown after 1-min perfusion (bar in A) with artificial CSF equilibrated with 100% CO₂. Anatomic view of retrotrapezoid nucleus (RTN) region (B) shows location of this pipette tip (within box). Pipette tip location was ascertained by presence of a small amount of fluorescent beads that were on outside of pipette tip. P, pyramids. Bar in B, 0.5 mm.
[View Larger Version of this Image (16K GIF file)]

---

---

Fig. 2. Anatomic location of tip of CO₂-diffusion pipettes is shown diagrammatically on a series of medullary cross sections modified from atlas of Paxinos and Watson (22). Nos. at bottom right of each section refer to millimeters caudal to the bregma. , Sites of pipettes perfused with artificial CSF equilibrated with 100% CO₂; , sites of pipettes perfused with artificial CSF equilibrated with 50% CO₂; , sites of pipettes perfused with artificial CSF equilibrated with air. VII, facial nucleus. Bar, 1 cm.
[View Larger Version of this Image (22K GIF file)]

---

Criteria for inclusion in the study. All data reported below were obtained in experiments that fulfilled the following criteria. 1) The CO₂ diffusion pipette site was in the region of the RTN as shown by anatomic evaluation of the fluorescent microbeads placed externally on the pipette tip. 2) There was a final response to systemic CO₂ stimulation after the injection protocol indicating that CO₂ sensitivity was maintained during the protocol. 3) Mean arterial blood pressure was at least 80 mmHg during the protocol.

Experimental design. In the control group, four 1-min CO₂ diffusions with the pipette in the RTN and the perfusion solution equilibrated with air were made in two rats. In the moderate-CO₂ group, four 1-min CO₂ diffusions with the pipette in the RTN and with the perfusion solution equilibrated with 50% CO₂ were made in four rats. In the high-CO₂ group, 14 1-min CO₂ diffusions with the pipette in the RTN and with the perfusion solution equilibrated with 100% CO₂ were made in eight rats. In all groups, tissue pH was measured at the injection center and at a distant site.

---

RESULTS

Injection locations. Figure 2 shows the location of the center of each injection included in this report. All pipette tip locations were in the region of the rat RTN (23, 27). No Evans blue dye was observed within the tissue, indicating that no fluid volume left the pipette tip.

CO₂ responses. Responses of PNA and frequency to increased end-tidal CO₂ were obtained at the beginning and the end of each experiment. The presence of an intact response at the onset of the experiment allows us to proceed with some assurance that we may detect a response to focal acidosis produced by the CO₂ diffusion pipette. A maintained response to systemic CO₂ stimulation gives us some evidence that the preparation, in terms of chemoreception, has remained functional for the duration of the experiment. In general, the response to systemic CO₂ stimulation is less at the end of the experiment. This loss of CO₂ responsivity is probably the result, in these experiments, of exposure and manipulation of the ventral medullary surface and insertion of more than one micropipette. At the baseline value for end-tidal PCO₂ of 28 Torr, the initial PNA was 68.8 ± 2.4 (SE) %, expressed as percentage of the maximum observed with end-tidal PCO₂ of 63 Torr. At the end of the experiment, the PNA values at 28 and 63 Torr were 55.9 ± 4.6 and 79.1 ± 5.1% of maximum, respectively, results very similar to those reported in a recent paper examining the effects of VRG injection of AZ on tissue pH and PNA (20). The respiratory frequency changes that occur in this preparation with CO₂ stimulation are virtually absent (see Ref. 20).

Medullary tissue pH responses. We obtained tissue pH measurements at the CO₂ diffusion pipette tip and at a second distant site with perfusion of the central pipette with aCSF equilibrated with air (4 cases in 2 rats), 50% CO₂ (4 cases in 4 rats), or 100% CO₂ (14 cases in 8 rats). In each case, the measured change in tissue pH was normalized by comparison to the measured tissue pH change that occurred with end-tidal CO₂ increased to 9%. These electrodes calibrated in vitro showed sensitivities of 54.9 ± 1.4 (SE) mV/pH before the experiment and 50.1 ± 3.5 mV/pH at the end of the experiment. For in vivo calibration, we estimated the change in tissue pH that accompanied the change in end-tidal CO₂ from 4 to 9% by assuming a tissue bicarbonate value of 22 mM, a tissue CO₂ solubility of 0.03 mM · l¹ · Torr¹, and a pK of 6.1, and we calculated tissue pH at tissue PCO₂ values of 28 and 63 Torr. At the beginning of the experiment the electrode sensitivities in vivo were 65.9 ± 8.5 mV/pH and at the end of the experiment were 66.0 ± 12.9 mV/pH.

Figure 3 shows the relationship between the peak change in tissue pH, expressed as a percentage of the maximum change observed in vivo with end-tidal CO₂ increased from 4 to 9%, and the distance of the pH electrode from the CO₂ diffusion pipette tip. At the tip of the pipette perfused for 1 min with aCSF equilibrated with 100% CO₂, the average peak decrease in tissue pH is 107 ± 18% (mean ± SE; n = 14) of that observed with the change in end-tidal CO₂ from 4 to 9%. This is approximately equivalent to 1.07 × 35 = 37.5 Torr above the baseline of 28 Torr, or 65.5 Torr. This is the end-tidal PCO₂ that would produce a tissue pH change approximately equivalent to that at the diffusion pipette tip during perfusion with aCSF equilibrated with 100% CO₂. The degree of tissue acidosis diminished with distance from the pipette tip such that at 350 µm from the tip, the tissue pH is decreased only 38.3 ± 5.8% (mean ± SE; n = 14) of that observed with the change in end-tidal CO₂ from 4 to 9%. This is approximately equivalent to 0.38 × 35 = 13.4 Torr above the baseline of 28 Torr, or 41.4 Torr. By extrapolation, at 400-500 µm from the pipette tip there is no detectable pH change.

---

Fig. 3. Peak change () in tissue pH measured at tip of CO₂ diffusion pipette and at varying distances from tip during 1-min pipette perfusion is shown as function of that distance. In each rat, pH change with pipette perfusion is normalized to pH change observed with end-tidal CO₂ increased from 4 to 9%. This latter pH change is expressed as 100%, and pH changes observed with pipette perfusion are expressed as percentage of that maximum. , Results with pipette perfused with artificial CSF equilibrated with 100% CO₂;  results with pipette perfused with artificial CSF equilibrated with 50% CO₂. Values are means ± SE. Nos. of observations shown are as follows (moving from center distally): for solid circles, 14, 2, 5, 4, and 3; and for open circles, 5, 1, 1, 1, and 1 (1 distal pH electrode was dysfunctional in this group).
[View Larger Version of this Image (15K GIF file)]

---

At the tip of the pipette perfused for 1 min with aCSF equilibrated with 50% CO₂, the average decrease in tissue pH is 51.4 ± 10.1% (mean ± SE; n = 5) of that observed with the change in end-tidal CO₂ from 4 to 9%. This is approximately equivalent to 0.51 × 35 = 17.9 Torr above the baseline of 28 Torr, or 45.9 Torr, the end-tidal PCO₂ that would produce a tissue pH change approximately equivalent to that at the diffusion pipette tip. The degree of tissue acidosis diminished with distance from the pipette tip such that at 250 µm from the tip, there was no pH change.

When the pipette was perfused for 1 min with aCSF equilibrated with air, tissue pH actually increased ~21.8 ± 2.3% of the maximum tissue pH change observed with changes in end-tidal CO₂ from 4 to 9%. This alkalosis was short lived but was present on two occasions at ~350 µm from the tip of the diffusion pipette.

Responses to RTN region CO₂ diffusion. Figure 4 shows the results from a typical single experiment with the pipette perfused for 1 min with aCSF equilibrated with 100% CO₂. The pipette tip is located within the RTN region, as shown in Fig. 4B, and the period of active pipette perfusion takes place for 1 min. Tissue pH at the pipette tip decreases during the 1 min of the perfusion and then returns toward baseline within minutes. PNA increases and returns toward baseline with a very similar time course.

The averaged results of tissue pH and PNA responses to 1 min of pipette perfusion with aCSF equilibrated with air (control; n = 4), 50% CO₂ (n = 5), or 100% CO₂ (n = 14) are shown in Fig. 5. Multifactorial analysis of variance (ANOVA) shows that the changes in tissue pH after perfusion with 50 or 100% CO₂ are significantly different from each other (P < 0.0001) and are greater than during air (P < 0.005 and P < 0.001, respectively). The changes in PNA after perfusion with 50 or 100% CO₂ are significantly different from each other (P < 0.03) and are greater than during air (P < 0.05 and P < 0.001, respectively). With 100% CO₂ perfusion, the peak tissue pH change occurs at 2.0 ± 0.4 min and the peak PNA response at 1.9 ± 0.20 min. With 50% CO₂ perfusion, the peak tissue pH change occurs at 4.4 ± 0.7 min and the peak PNA response at 3.2 ± 1.9 min. Respiratory frequency was unaffected.

---

Fig. 5. Group responses of PNA (A) and central tissue pH at pipette tip (B) are shown as function of time after 1-min pipette perfusions (solid bars) with artificial CSF equilibrated with 100% CO₂ (; n = 14), 50% CO₂ (; n = 5), and air (; n = 4). Values are means ± SE; n, no. of experiments.
[View Larger Version of this Image (14K GIF file)]

---

Group mean arterial blood pressure did not change significantly in the 15 min after perfusion for 1 min with aCSF equilibrated with air or with 50 or 100% CO₂, although in the latter set of experiments (100% CO₂), mean blood pressure increased in three animals by 8, 10, and 21 Torr in the first 3 min after the period of diffusion.

An analysis of the peak tissue pH and associated peak PNA responses to these 1-min pipette perfusions is shown in Fig. 6. Baseline conditions and the tissue pH and PNA changes with an end-tidal CO₂ of 9% produced by CO₂ inhalation are shown. This represents the situation with all central chemoreceptors stimulated. Also shown are the data from the three types of pipette perfusion, air, 50% CO₂, and 100% CO₂. There is a continuum of response; no threshold is evident. For interest, the data from simultaneous measurement of tissue pH and PNA during injection of AZ into the region of the VRG in the rat are given (20). The responses appear to fall on the same dose-response curve showing the system or PNA response to focal acidification of the RTN by CO₂ diffusion pipette or the rostral VRG by AZ injection.

---

Fig. 6. Peak change in phrenic nerve amplitude (PNA) is shown as function of tissue pH measured at CO₂ diffusion pipette tip. , Baseline situation before any pipette perfusion with end-tidal CO₂ at 4% ( at left) and situation with end-tidal CO₂ at 9% ( at right). This represents response to stimulation of all central chemoreceptors. , Responses (means ± SE) to pipette perfusion with artificial CSF equilibrated with air, 50% CO₂, or 100% CO₂. This is dose response to focal stimulation of central chemoreceptors at RTN. , Data from Nattie and Li (20) with acetazolamide injections into region of rostral aspect of ventral respiratory group. Tissue pH change is normalized and expressed as percentage of maximum, i.e., that pH change observed with end-tidal CO₂ increased from 4 to 9%. Fact that pipette perfusion with artificial CSF equilibrated with 50 or 100% CO₂ changed tissue pH by ~50 or 100% of maximum is coincidence.
[View Larger Version of this Image (12K GIF file)]

---

Responses to repeated CO₂ diffusion in RTN region. One ultimate goal is to test mechanisms of central chemosensitivity in vivo via evaluation of PNA responses to focal medullary acidification before and after manipulation of the neuronal environment in the region of the focal acidosis. To perform such an experiment requires repeated similar responses to focal acidification. In this initial description of the CO₂ pipette, we performed repeat acidification in five cases by using 100% CO₂ pipette perfusion. The tissue pH change measured at the pipette tip was virtually identical in the first and second diffusions. The PNA response, however, was significantly greater during and after the second diffusion (P < 0.02; ANOVA).

---

DISCUSSION

Major findings. This study has three new findings. 1) By using a new pipette that allows diffusion of CO₂ into the tissue from a constantly renewable source at the tip, we produced, in the RTN region, focal, brief, and reversible acidification. The presence of a PNA response to this focal acidosis, similar to that observed after an AZ-induced focal tissue pH change (6, 20), indicates that the response is specific for CO₂ or pH. 2) There is a dose-response relationship between focal RTN acidification and system PNA responses, and the tissue pH changes are within the physiological range. In the RTN, no threshold is detectable, the response is a continuum. 3) The PNA responses are temporally linked more clearly to the briefer, CO₂ pipette-induced tissue pH changes than are the longer lasting changes produced by AZ injection.

CO₂ diffusion pipette. Central chemoreceptors were initially located at sites just below the ventral medullary surface by means of focal application of acidic substances to the tissue surface (3, 14, 15, 26). The methods included application by small cotton pledgets and, subsequently, the use of a "superfusion" cannula system that allowed renewal of the CO₂ and H⁺ concentrations at the medullary surface (14, 26). Both methods delivered a sufficient amount of stimulating substance to overcome potent local blood flow clearance and tissue buffer mechanisms and result in a detectable response. Blood flow to the medulla originates at the ventral surface, and this flow can quickly wash out diffusible substances placed or injected, in limited amounts, at or near to the surface (3, 13, 15). In these earlier in vivo studies, the region affected by such surface applications was never clearly defined but is assumed to be fairly large. In simpler systems without such blood flow, the use of small pipettes of a design similar to ours in this paper have been used to localize chemoreception in the isolated neonatal rat brain stem preparation (10) and in the pulmonate snail (8).

Our goal for some time has been to produce a focal and small region of acidosis at various medullary sites to define the locations of central chemoreceptors in vivo. To do this requires a source of CO₂ or H⁺ that is renewable. Over the past few years we have done this by using a small injection of the carbonic anhydrase inhibitor AZ (2, 6, 20). The AZ, by disruption of local tissue CO₂ or H⁺ clearance and/or buffer mechanisms, produces a well-defined and focal region of tissue acidosis (radius <350 µm from the injection center). The changes in tissue pH at the center, equivalent to a 36- to 47-Torr increase in end-tidal PCO₂, are relatively severe and rather long lasting as well. This time course presumably reflects the pharmacokinetics of AZ binding to tissue carbonic anhydrase. The approach is useful to localize central chemoreceptor sites, but the long-lasting pH changes make it impossible to do other experiments designed to examine how central chemoreceptors work, and the central pH changes were relatively fixed, making it difficult to examine sensitivity.

Earlier attempts in our laboratory had used the microdialysis technique to create a tissue column of continuously acidotic tissue by diffusion of CO₂ through the dialysis membrane (unpublished observations). Breathing could be stimulated, and the effects were quickly reversible, but the large microdialysis cannula disrupted neurons within these ventral medullary regions, the function of which is necessary to maintain normal respiratory output (3, 15, 17-19), and it produced a large region of acidosis relative to the effects of AZ or the new CO₂ diffusion pipette. The approach described in this paper, a microdiffusion pipette, evolved from these microdialysis experiments and from the cannulas and pipettes reported by Schlaefke et al. (26), Issa and Remmers (10), and Erlichman and Leiter (8).

Our pipette is small in size (total tip diameter of 80-100 µm), and the region surrounding the pipette tip that becomes acidotic after a 1-min period of pipette perfusion with CO₂ solutions has a radius <400-500 µm with 100% CO₂ equilibrated in the perfusate and <250 µm with 50% CO₂ equilibrated in the perfusate. Surprisingly, the use of aCSF equilibrated with air, which dropped the CO₂ to zero in the pipette tip, alkalinized the tissue and decreased PNA. CO₂ must diffuse from the tissue into the pipette tip, producing a focal region of alkalosis that inhibits the chemoreceptors at that site. In two cases, distant pH electrodes ~350 µm from the tip of the diffusion pipette detected this focal alkalosis. In this experiment we limited the pipette perfusion to 1 min because our goal was to produce transient and reversible focal pH changes associated with changes in respiratory output. It will be of interest to observe responses to more continuous focal acidosis.

Microscopic examination of tissue at the pipette tip and surrounding region found no fast green dye, which had been included in the pipette perfusion solution, indicating that no injection was made into the tissue. The RTN acidosis produced by this pipette was sufficient to stimulate phrenic nerve output (PNA), and the changes in both tissue pH and PNA occurred quickly, within minutes of the onset of the pipette perfusion.

After a 1-min period of pipette perfusion, the perfusion was stopped, and the pipette was allowed to remain in place, with the tip still containing the perfused solution. In the case of the aCSF equilibrated with 50 or 100% CO₂, some CO₂ undoubtedly continued to diffuse out of the pipette, maintaining a tissue acidosis until the added CO₂ was cleared or buffered. This continued CO₂ diffusion accounts for the continued changes in tissue pH and PNA observed after the 1-min perfusion period ended (see Fig. 5) and for the slower return to baseline of these variables compared with the initial onset of the changes. To test this interpretation, it will be necessary to follow the 1-min 50 or 100% CO₂-equilibrated perfusion period with continued perfusion by using aCSF equilibrated with 5% CO₂. We believe that this will result in faster "off" responses, which will facilitate multiple tests of chemoreceptor function at one pipette site.

Focal central chemoreceptor stimulus. The responses of PNA to these CO₂ pipette-induced changes in RTN pH support the interpretation that the AZ injection-induced tissue pH changes were the cause of the increased PNA; they were not the result of a nonspecific AZ effect. Although to date we only know the changes in tissue pH, it may be possible to also measure tissue PCO₂ by using a microelectrode designed by Voipio and Kaila (30). This will allow us to determine in an in vivo experiment whether tissue PCO₂ or H⁺ is the chemoreceptor stimulus.

Dose response of central chemoreceptors in the RTN. Because of the short time course of the changes in tissue pH and PNA that result from the 1-min perfusion of the CO₂ pipette, we were able to reliably detect rather small changes in PNA. In the AZ experiments, PNA had to be followed over prolonged time periods. During this long time duration, experimentally induced variations in PNA had to be more robust because of the greater variability in the PNA signal over the longer time period. With the CO₂ pipette, measurements can be made over a period of a few minutes, during which time the PNA signal is usually quite stable in the anesthetized rat. Thus it is easier to detect experimentally induced small changes in PNA. For this reason we did not separate the results into "responders" vs. "nonresponders" as we have done for AZ injections. The peak tissue pH responses to pipette perfusion with 100% CO₂-equilibrated aCSF ranged from 13 to 226% of the maximum (the changes in tissue pH observed with end-tidal CO₂ of 9%), and the peak PNA responses ranged from 3 to 62% of baseline. Although there was a tendency for smaller pH changes to be associated with smaller PNA responses, this relationship was not significant (linear regression analysis). Estimated sensitivity values (peak change in PNA/peak change in tissue pH) varied widely at different pipette sites. The values (means ± SE) were 0.194 ± 0.56 (%baseline PNA/%maximum tissue pH change) with a range of 0.03-0.72. It may be reasonable in the future to consider excluding individual sites with low sensitivities, but we have not done this here.

The result of this increased sensitivity for measuring PNA changes is a remarkable finding that is shown in Fig. 6. Focal acidosis of the RTN region by the CO₂ pipette produces a detectable dose-response relationship that is present well within the physiological range of CO₂-induced tissue pH changes. No threshold was detected. This information is helpful in understanding a problem posed by the concept of a widespread distribution of central chemoreceptor locations (2, 6, 15, 17). Why is it that, in the anesthetized animal, disruption solely of the RTN region can abolish the response to CO₂ (1, 9, 15, 17-21)? What is the utility or importance of the other central chemoreceptor sites when a single, quite specific disruption can abolish the entire response? Two possibilities were proposed. One is that anesthesia magnifies the relative importance of the RTN region chemoreceptors vs. those at other sites. In support of this are data showing diminished but not absent CO₂ sensitivity when the RTN is disrupted in an unanesthetized animal (1, 9, 21). The other possibility is that chemoreceptors at different sites have differing sensitivities and/or thresholds. One site might contain "physiological" chemoreceptors sensitive to CO₂ in the physiological range (see Ref. 20 for discussion). Other sites might contain "emergency" chemoreceptors designed to respond at some threshold value above the normal physiological range as an emergency-sensing system for extreme disturbances of ventilation or perfusion. The dose-response characteristics of PNA vs. focal RTN acidification by the CO₂ pipette in the RTN region support this idea, indicating that the RTN chemoreceptors are physiological in nature. Chemoreceptors at other sites need to be tested.

Focal acidosis produced by AZ microinjections resulted in a surprisingly large effect; i.e., a stimulation of PNA by the focal AZ induced acidosis that was a very large fraction of that observed with near maximum stimulation of all central chemoreceptors by an end-tidal CO₂ of 9% (2, 6, 20). A similar result was obtained in the present experiments. Figure 6 shows a relatively large stimulation of PNA by the RTN focal acidosis produced by the diffusion pipette relative to that observed with all central chemoreceptors stimulated. One explanation put forth for the relatively large effects of focal central chemoreceptor stimulation, that it was the result of a relatively severe acidosis at the center of the AZ injections, seems to be unlikely given the results shown in Fig. 6. Instead, it appears that isolated stimulation of RTN chemoreceptors can stimulate a respiratory system output, here PNA.

Time course of tissue pH and PNA responses after VRG AZ injection and RTN CO₂ diffusion pipette. The time course of the tissue pH and PNA responses to AZ injection differ markedly from those with the CO₂ diffusion pipette. Figure 7 shows the 100% CO₂ perfusion data from this paper together with similar data from AZ injections into the rat VRG region (20). Most noteworthy is the markedly faster on and off responses with the CO₂ pipette. The CO₂ diffusion pipette off-response characteristics probably reflect the use of an air control and not a 5% CO₂ control, as discussed in CO₂ diffusion pipette in DISCUSSION.

---

Fig. 7. Time courses of PNA (A) and tissue pH responses (pH; B) to focal acidosis produced by acetazolamide injection into ventral respiratory group (; data from Ref. 20) or to CO₂ diffusion pipette in RTN () are shown. Values are means ± SE. Note prolonged tissue acidosis after acetazolamide injection, lasting well beyond time of recovery of PNA. Tissue pH response to 1 min of CO₂ pipette perfusion with artificial CSF equilibrated with 100% CO₂ develops and recovers more quickly in comparison to that after acetazolamide and has a time course similar to the PNA.
[View Larger Version of this Image (15K GIF file)]

---

In the AZ experiments, we expected that, on average, the tissue pH at the injection center and the PNA responses would have a similar time course because tissue pH was assumed to be a reasonable estimate of the chemoreceptor stimulus. The initial increase in PNA does correlate well with the initial decrease in tissue pH, but the recovery time courses differ substantially. Others have pointed out (29), in experiments with systemic carbonic anhydrase inhibition, the lack of a tight correlation between medullary tissue pH and the ventilatory response.

We have suggested (20) a number of possible explanations for this hysteresis. 1) The chemoreceptor stimulus is intracellular not extracellular (tissue) pH, as suggested for the carotid body type I cells thought responsible for sensing CO₂ and/or H⁺ (4, 16) and for central chemoreception in an invertebrate model (8). The slower PNA response to an abrupt increase in CO₂ in animals with brain stem carbonic anhydrase inhibition also supports the view that intracellular pH is sensed; with rapid changes in CO₂, carbonic anhydrase speeds the conversion of CO₂ to H⁺ (5). Lassen (12) has suggested that intracellular pH is the central chemoreceptor stimulus on the basis of an analysis of intra- and extracellular pH measurements after hypercapnia or systemic AZ administration. 2) The pH electrode at the center of the AZ injection does not reflect the tissue pH at the chemoreceptor site; i.e., the site is not at the injection center. 3) The chemoreceptor mechanism or the respiratory control system might show accommodation of some type during the period of tissue acidosis.

At present, given the slightly slower "off" response compared with the "on" response in the CO₂ diffusion pipette data (Fig. 7), it is difficult to make any conclusive comments. We speculate that with the CO₂ diffusion pipette operated with a rapid turn off by switching not to air but to 5% CO₂-equilibrated perfusion solution, the on and off responses of PNA vs. tissue pH will have identical time courses. This would support the view that the hysteresis observed in these response time courses after AZ injection might be some type of accommodation to a sustained acidosis either at the chemoreceptor or in the respiratory control system.

Repeat stimuli at a focal chemoreceptor site. We found similar RTN pH changes produced by two successive perfusions at the same RTN sites. This indicates that we can reliably change tissue pH repeatedly at a single site, which will allow tests of the effects of various substances on the sensitivity of the response at a single central chemoreceptor site in vivo. We will be able to examine certain hypotheses for the mechanism of central chemoreception in vivo. The PNA response to this initial attempt at repeated stimulation at a single chemoreceptor site showed the second response to be greater. Although we have no clear explanation for this at present, we find this observation encouraging in that in this type of preparation the maintenance of central chemosensitivity during a prolonged experiment can be problematic.

Conclusion. We conclude that focal medullary pH can be quickly changed to varying degrees and reversed with this CO₂ diffusion pipette. In the RTN, such focal acidification shows remarkable central chemoreceptor sensitivity to focal pH changes well within the normal physiological range.

---

ACKNOWLEDGEMENTS

We thank Drs. Joe Erlichman and Lee Coates for help in the development of the CO₂ diffusion pipette, which is similar to their pipette used to test central chemoreception in a pulmonate snail, and Karen Giordano for help with the pipette.

---

FOOTNOTES

   This study was supported by National Heart, Lung, and Blood Institute Research Grant HL-28066.

Address for reprint requests: E. E. Nattie, Dartmouth Medical School, Dept. of Physiology, 706E Borwell, Lebanon, NH 03756-0001 (E-mail: Eugene.Nattie{at}Dartmouth.EDU).

Received 6 January 1997; accepted in final form 18 May 1997.

---

REFERENCES

1.	Akilesh, M. R., M. Kamper, A. Li, and E. E. Nattie. Effects of unilateral lesions of the retrotrapezoid nucleus on breathing in awake rats. J. Appl. Physiol. 82: 469-479, 1997[Abstract/Free Full Text].
2.	Bernard, D. G., A. Li, and E. E. Nattie. Evidence for central chemoreception in the midline raphé. J. Appl. Physiol. 80: 108-115, 1996[Abstract/Free Full Text].
3.	Bruce, E. N., and N. S. Cherniack. Central chemoreceptors. J. Appl. Physiol. 62: 389-402, 1987[Abstract/Free Full Text].
4.	Buckler, K. J., R. D. Vaughan-Jones, C. Peers, D. Lagadic-Gossmann, and P. C. Nye. Effects of extracellular pH, PCO2 and HCO3 on intracellular pH in isolated type-I cells of the neonatal rat carotid body. J. Physiol. (Lond.) 444: 703-721, 1991[Abstract/Free Full Text].
5.	Coates, E. L., A. Li, and E. E. Nattie. Acetazolamide on the ventral medulla of the cat increases phrenic output and delays the ventilatory response to CO2. J. Physiol. (Lond.) 441: 433-451, 1991[Abstract/Free Full Text].
6.	Coates, E. L., A. Li, and E. E. Nattie. Widespread sites of brain stem ventilatory chemoreceptors. J. Appl. Physiol. 75: 5-14, 1993[Abstract/Free Full Text].
7.	Dean, J. B., D. A. Bayliss, J. T. Erickson, W. L. Lawing, and D. E. Millhorn. Depolarization and stimulation of neurons in nucleus tractus solitarii by carbon dioxide does not require chemical synaptic input. Neuroscience 36: 207-216, 1990[Medline].
8.	Erlichman, J. S., and J. C. Leiter. Central chemoreceptor stimulus in the terrestrial, pulmonate snail, Helix aspersa. Respir. Physiol. 95: 209-26, 1994[Medline].
9.	Forster, H. V., P. J. Ohtake, L. G. Pan, and T. F. Lowry. Effects on breathing of ventrolateral medullary cooling in awake goats. J. Appl. Physiol. 78: 258-265, 1995[Abstract/Free Full Text].
10.	Issa, F. G., and J. R. Remmers. Identification of a subsurface area in the ventral medulla sensitive to local changes in PCO2. J. Appl. Physiol. 72: 439-446, 1992[Abstract/Free Full Text].
11.	Kawai, A., D. Ballantyne, K. Mückenhoff, and P. Scheid. Chemosensitive medullary neurones in the brainstem-spinal cord preparation of the neonatal rat. J. Physiol. (Lond.) 492: 277-292, 1996[Abstract/Free Full Text].
12.	Lassen, N. A. Is central chemoreceptor sensitive to intracellular rather than extracellular pH? Clin. Physiol. 10: 311-319, 1990[Medline].
13.	Lipscomb, W. T., and L. L. Boyarski. Neurophysiological investigations of medullary chemosensitive areas of respiration. Respir. Physiol. 16: 362-376, 1972[Medline].
14.	Mitchell, R. A., H. H. Loeschcke, W. H. Massion, and J. W. Severinghaus. Respiratory response mediated through superficial chemosensitive areas on the medulla. J. Appl. Physiol. 18: 523-533, 1963[Abstract/Free Full Text].
15.	Nattie, E. E. Central chemoreception. In: Regulation of Breathing (2nd ed.)., edited by J. A. Dempsey, and A. I. Pack. New York: Dekker, 1995, vol. 79, p. 473-510. (Lung Biol. Health Dis. Ser.)
16.	Nattie, E. E. Invited editorial on "Adaptation to hypercapnia vs. intracellular pH in cat carotid body: responses in vitro." J. Appl. Physiol. 80: 1088-1089, 1996[Free Full Text].
17.	Nattie, E. E., and A. Li. Fluorescent location of RVLM kainate microinjections that alter the control of breathing. J. Appl. Physiol. 68: 1157-1166, 1990[Abstract/Free Full Text].
18.	Nattie, E. E., and A. Li. Lesions in the retrotrapezoid nucleus decrease ventilatory output in anesthetized or decerebrate cats. J. Appl. Physiol. 71: 1363-1375, 1991.
19.	Nattie, E. E., and A. Li. Retrotrapezoid nucleus lesions decrease phrenic activity and CO2 sensitivity in rats. Respir. Physiol. 97: 63-77, 1994[Medline].
20.	Nattie, E. E., and A. Li. Central chemoreception in the region of the ventral respiratory group of the rat. J. Appl. Physiol. 81: 1987-1995, 1996[Abstract/Free Full Text].
21.	Ohtake, P. J., H. V. Forster, L. G. Pan, T. F. Lowry, M. J. Korducki, E. A. Aaron, and E. M. Weiss. Ventilatory responses to cooling the ventrolateral medullary surface of awake and anesthetized goats. J. Appl. Physiol. 78: 247-257, 1995[Abstract/Free Full Text].
22.	Paxinos, G., and C. Watson. The Rat Brain in Stereotaxic Coordinates. New York: Academic, 1986.
23.	Pearce, R., R. Stornetta, and P. Guyenet. Retrotrapezoid nucleus in the rat. Neurosci. Lett. 101: 138-142, 1989[Medline].
24.	Pineda, J., and G. K. Aghajanian. Carbon dioxide regulates the tonic activity of locus ceruleus neurons by modulating a proton- and polyamine-sensitive inward rectifier potassium current. Neuroscience 77: 723-734, 1997[Medline].
25.	Richerson, G. B. Response to CO2 of neurons in the rostral ventral medulla in vitro. J. Neurophysiol. 73: 933-944, 1995[Abstract/Free Full Text].
26.	Schlaefke, M. E., W. R. See, and H. H. Loeschcke. Ventilatory response to alterations of H+ ion concentrations in small areas of the ventral medullary surface. Respir. Physiol. 10: 198-212, 1970[Medline].
27.	Smith, J. C., D. E. Morrison, H. H. Ellenberger, M. R. Otto, and J. L. Feldman. Brainstem projections to the major respiratory neuron populations in the medulla of the cat. J. Comp. Neurol. 281: 69-96, 1989[Medline].
28.	Southard, T. L., R.-Q. Huang, and J. B. Dean. Electrophysiological properties of neurons in chemosensitive areas of the dorsal and ventral brainstem. Soc. Neurosci. Abstr. 21: 1883, 1995.
29.	Teppema, J., F. Rochette, and M. Demedts. Effects of acetazolamide on medullary extracellular pH and PCO2 and on ventilation in peripherally denervated cats. Pflügers Arch. 415: 519-525, 1990[Medline].
30.	Voipio, J., and K. Kaila. Interstitial PCO2 and pH in rat hippocampal slices measured by means of a novel fast CO2/H+-sensitive microelectrode based on a PVC-gelled membrane. Pflügers Arch. 423: 193-201, 1993[Medline].

---

0161-7567/97 $5.00 Copyright © 1997 the American Physiological Society

This article has been cited by other articles:

				P. Haouzi and H. J. Bell Control of breathing and volitional respiratory rhythm in humans J Appl Physiol, March 1, 2009; 106(3): 904 - 910. [Abstract] [Full Text] [PDF]

				A. C. T. Takakura, T. S. Moreira, E. Colombari, G. H. West, R. L. Stornetta, and P. G. Guyenet Peripheral chemoreceptor inputs to retrotrapezoid nucleus (RTN) CO2-sensitive neurons in rats J. Physiol., April 15, 2006; 572(2): 503 - 523. [Abstract] [Full Text] [PDF]

				N. A. Ritucci, J. S. Erlichman, J. C. Leiter, and R. W. Putnam Response of membrane potential and intracellular pH to hypercapnia in neurons and astrocytes from rat retrotrapezoid nucleus Am J Physiol Regulatory Integrative Comp Physiol, September 1, 2005; 289(3): R851 - R861. [Abstract] [Full Text] [PDF]

				M. R. Hodges, L. Klum, T. Leekley, D. T. Brozoski, J. Bastasic, S. Davis, J. M. Wenninger, T. R. Feroah, L. G. Pan, and H. V. Forster Effects on breathing in awake and sleeping goats of focal acidosis in the medullary raphe J Appl Physiol, May 1, 2004; 96(5): 1815 - 1824. [Abstract] [Full Text] [PDF]

				J L Ribas-Salgueiro, S P Gaytan, R Crego, R Pasaro, and J Ribas Highly H+-sensitive neurons in the caudal ventrolateral medulla of the rat J. Physiol., May 15, 2003; 549(1): 181 - 194. [Abstract] [Full Text] [PDF]

				E. E. Nattie and A. Li CO2 dialysis in the medullary raphe of the rat increases ventilation in sleep J Appl Physiol, April 1, 2001; 90(4): 1247 - 1257. [Abstract] [Full Text] [PDF]

				J. Holleran, M. Babbie, and J. S. Erlichman Ventilatory effects of impaired glial function in a brain stem chemoreceptor region in the conscious rat J Appl Physiol, April 1, 2001; 90(4): 1539 - 1547. [Abstract] [Full Text] [PDF]

				C. S. Torgerson, M. J. Gdovin, and J. E. Remmers Sites of respiratory rhythmogenesis during development in the tadpole Am J Physiol Regulatory Integrative Comp Physiol, April 1, 2001; 280(4): R913 - R920. [Abstract] [Full Text] [PDF]

				C. S. Torgerson, M. J. Gdovin, R. Brandt, and J. E. Remmers Location of central respiratory chemoreceptors in the developing tadpole Am J Physiol Regulatory Integrative Comp Physiol, April 1, 2001; 280(4): R921 - R928. [Abstract] [Full Text] [PDF]

				A. K. Curran, R. A. Darnall, J. J. Filiano, A. Li, and E. E. Nattie Muscimol dialysis in the rostral ventral medulla reduced the CO2 response in awake and sleeping piglets J Appl Physiol, March 1, 2001; 90(3): 971 - 980. [Abstract] [Full Text] [PDF]

				E. Nattie and A. Li Muscimol dialysis in the retrotrapezoid nucleus region inhibits breathing in the awake rat J Appl Physiol, July 1, 2000; 89(1): 153 - 162. [Abstract] [Full Text] [PDF]

				D Ballantyne and P Scheid Mammalian brainstem chemosensitive neurones: linking them to respiration in vitro J. Physiol., June 15, 2000; 525(3): 567 - 577. [Abstract] [Full Text] [PDF]

				A. Li, M. Randall, and E. E. Nattie CO2 microdialysis in retrotrapezoid nucleus of the rat increases breathing in wakefulness but not in sleep J Appl Physiol, September 1, 1999; 87(3): 910 - 919. [Abstract] [Full Text] [PDF]

				E. E. Nattie, J. S. Erlichman, and A. Li Brain stem lesion size determined by DEAD red or conjugation of neurotoxin to fluorescent beads J Appl Physiol, December 1, 1998; 85(6): 2370 - 2375. [Abstract] [Full Text] [PDF]

				J. S. Erlichman, A. Li, and E. E. Nattie Ventilatory effects of glial dysfunction in a rat brain stem chemoreceptor region J Appl Physiol, November 1, 1998; 85(5): 1599 - 1604. [Abstract] [Full Text] [PDF]

This Article Abstract Full Text (PDF) Free Submit a response Alert me when this article is cited Alert me when eLetters are posted Alert me if a correction is posted Citation Map 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 HighWire Citing Articles via Google Scholar Google Scholar Articles by Li, A. Articles by Nattie, E. E. Search for Related Content PubMed PubMed Citation Articles by Li, A. Articles by Nattie, E. E.