In myenteric neurons two different receptor subtypes govern the intracellular Ca2+ stores: the inositol 1,4,5-trisphosphate (IP3) receptor (IP3R) and the ryanodine receptor (RyR). Their degree of functional overlap was determined by examining Ca2+ release in these cells through both superfusion techniques and intracellular microinjection. Microinjection of IP3 (50 μM) and cADP-ribose (cADPr, 50 μM), specific ligands for the IP3R and RyR, respectively, demonstrated mobilization of intracellular Ca2+ stores. Perfusion with cinnarizine (50 μM) or dantrolene (10 μM), antagonists of the IP3R and RyR, respectively, eliminated the Ca2+response to microinjected IP3 and cADPr. Superfusion of the neurons with 100 μM ATP, an IP3-mediated Ca2+-mobilizing agonist, caused intracellular Ca2+ increments, which were antagonized by cinnarizine, and the RyR antagonists dantrolene, procaine (5 mM), and ryanodine (1 μM). Caffeine (10 mM) was applied repetitively in Ca2+-free conditions to deplete RyR-sensitive stores; subsequent perfusion with ATP demonstrated a Ca2+ response. Conversely, caffeine caused a Ca2+ response after repetitive ATP exposures. The internal Ca2+ stores of myenteric neurons are governed by two receptor subtypes, IP3R and RyR, which share partial functional overlap.
- myenteric plexus
- inositol 1,4,5-trisphosphate
- adenosine 3′,5′-cyclic diphosphate-ribose
the concentration of free cytosolic Ca2+ is a crucial signaling mechanism in eukaryotic cells, and most cell lines have developed highly specialized processes that utilize Ca2+ mobilization to control physiological processes. In neurons, changes in the intracellular Ca2+ concentration ([Ca2+]i) are associated with the release of neurotransmitters, neuronal excitability, and expression of various genetic programs. In the enteric nervous system, cytosolic Ca2+ increments are controlled by a variety of mechanisms, resulting either from Ca2+ influx from the external milieu or by Ca2+ release from internal stores.
A number of intracellular organelles store Ca2+ in sequestered compartments within the cytosol. These stores contain concentrations of Ca2+ in excess of cytoplasmic levels and utilize various pumps and exchangers to maintain this equilibrium. Two channels predominate to enact Ca2+ release from neurons: the inositol 1,4,5-trisphosphate (IP3) receptor (IP3R) and the ryanodine receptor (RyR) (5). Both receptors have been isolated, sequenced, and reconstituted into lipid bilayers (4). Both are tetramers that exhibit significant homology, although the RyR is considerably larger (5,8, 35).
Inositol trisphosphate activates the IP3R and is generated after specific agonist-receptor interactions at the cell surface. In enteric neurons, extracellular ATP has been shown to utilize IP3 as a second messenger (2, 18). Ca2+ is able to stimulate Ca2+ release through the RyR in a process known as Ca2+-induced-Ca2+release (CICR) (17). Caffeine and other methyl xanthines also stimulate Ca2+ release at the RyR, possibly acting through reduction of the threshold for CICR at the RyR (19). Recent evidence supports the concept that cADP-ribose (cADPr), a metabolite of nicotinamide adenine dinucleotide (NAD+), is the endogenous ligand for the RyR (9). The endogenous agonist responsible for generating intracellular cADPr is unknown. The observation that the IP3R and the RyR are homologous suggests that functional interactions of these Ca2+ stores may exist.
We report evidence that cultured myenteric neurons contain both IP3- and cADPr-sensitive Ca2+ pools. Intracellular microinjection of IP3 caused increments in [Ca2+]i, which could be abolished by the IP3R antagonist cinnarizine. IP3-induced Ca2+ transients were also inhibited by the RyR antagonist dantrolene. Likewise, Ca2+ responses to microinjected cADPr were blocked by both dantrolene and cinnarizine. Superfusion of cultured myenteric neurons with ATP resulted in elevations in [Ca2+]i that were antagonized by pharmacological agents affecting either the IP3R (cinnarizine) or the RyR (ryanodine, dantrolene, or procaine). In enteric neurons, the IP3-releasable pool appears to functionally overlap the ryanodine-releasable store, yet to possess distinct components as well.
Collagenase type V, trypsin-ethylene glycol-bis(β-aminoethyl ether)-N,N,N′,N′-tetraacetic acid (EGTA), soybean trypsin inhibitor (type I-S from soybean), penicillin-streptomycin solution, ATP, ionomycin, EGTA, poly-l-lysine, dantrolene sodium, procaine hydrochloride, cinnarizine, andN-2-hydroxyethylpiperazine-N′-2-ethanesulfonic acid (HEPES) were from Sigma Chemical (St. Louis, MO). Caffeine and ryanodine were from Research Biochemical International (Natick, MA). Rat-tail collagen was from Boehringer Mannheim (Indianapolis, IN ). Fura 2-acetoxymethyl ester (AM), fura 2-free acid, and sulforhodamine B were from Molecular Probes (Eugene, OR). Hanks' balanced salt solution (HBSS), MEM amino acids solution, MEM sodium pyruvate, 100 mM, l-glutamine, fetal bovine serum, and medium 199 are from GIBCO-BRL (Grand Island, NY). NU serum was from Collaborative Research (Bedford, MA). Eppendorf Femtotips II (0.5 ± 0.2 μm) for microinjection were purchased from Brinkmann Instruments (Westbury, NY). Neonatal Duncan-Hartley guinea pigs were purchased from Simonsen Laboratories (Gilroy, CA).
All experiments were performed in standard solutions except when noted. Standard control buffer (pH 7.40) contained (in mM) 118 NaCl, 4.7 KCl, 15 NaHCO3, 11 glucose, 10 HEPES, 1.8 CaCl2, 0.9 NaH2PO4, and 0.8 MgSO4. For Ca2+-free conditions CaCl2 was removed from the buffer and 1 mM EGTA was added. Microinjection buffer (pH 7.20) contained (in mM) 120 KCl, 20 HEPES, and 1 MgCl2.
Myenteric plexus isolation.
Dispersed primary cultures of neonatal guinea pig myenteric plexus were prepared on collagen-coated coverslips and used for experimentation ondays 3–5. The Taenia coli from 1-day-old male Duncan-Hartley guinea pigs were removed and placed in HBSS solution plus 0.1% collagenase for 16–20 h at 4°C. After a 35-min incubation at 37°C, the muscularis layers of the Taenia coli were separated from the myenteric plexus using a dissecting microscope. The myenteric plexus was trypsinized for 30 min at 37°C using a 0.1% trypsin-EGTA solution, triturated with siliconized flamed Pasteur pipettes of decreasing tip diameter, and plated in collagen-coated coverslips. Cultures were exposed to complete medium 199 plus 5% NU serum and 0.0001% trypsin inhibitor. Penicillin-streptomycin solution was added for the first 24 h at a 2% concentration. Antimitotic agents were not added. The medium was changed every other day. The cultures were incubated at 37°C with 95% O2-5% CO2 and used for experimentation on days 3–5 post-plating.
Loading and cell preparation for imaging.
Cultured cells were incubated at 37°C in fresh warmed medium containing 2–3 μM fura 2-AM for 25 min. Loaded coverslips were washed and replaced in standard control buffer and then placed in a Lucite superfusion chamber. The superfusion rate of the control buffer and experimental solutions was 1 ml/min at 37°C.
A Zeiss Axiovert inverted microscope and Attofluor digital imaging system (Rockville, MD) were used for single-cell [Ca2+]i determinations. [Ca2+]i was calculated from the ratios of the fluorescence intensities of fura 2 at 334- and 380-nm wavelengths with an emission wavelength of 500 nm. Calibration of the system was performed with the following two-point standardization equation using fura 2-free acid: [Ca2+]i =K d[(R − RLo)/(RHi − R)]β, whereK d is the dissociation constant of the Ca2+-fura 2 complex (225 nm), R is F334/F380, i.e., the fluorescence at 334 nm divided by the fluorescence at 380 nm excitation, RLo is the ratio at zero Ca2+ (1 mM EGTA), RHi is the ratio at high Ca2+ (1 mM CaCl2), and β is F380 (zero Ca2+)/F380 (saturating Ca2+). Frames were not averaged. A ratio pair was taken every second.
Microinjection of indicator dyes and signaling molecules was performed using a piezoelectric cell penetrator system (Burleigh, Rochester, NY) mounted on the stage of a Zeiss Axiovert inverted microscope and Attofluor digital imaging system (Rockville, MD). The entire apparatus was additionally mounted on a vibration-free platform (Vibraplane, Kinetic Systems, Boston, MA). Femtotip pipettes of 0.5 ± 0.2 μm tip diameter were purchased. All test solutions were solubilized in standard microinjection buffer, which was twice filtered through a 0.22-micron filter. The pipettes were mounted in a universal capillary holder (Brinkmann Instruments), which was connected to a pressure injector (Pneumatic Picopump, World Precision Instruments, Sarasota, FL). After the injection tip was located on high power and positioned to approach the cells of interest, the piezoelectric step driver was used to make 1-μm jumps toward the cell, and injection under a positive pressure of 20 psi was performed. An injection was accompanied by a rapid withdrawal of the tip after injection was performed. The success of the injection was monitored by observing that the cell maintained its morphology and its ability to retain the sulforhodamine dye (excitation 560 nm, emission 630 nm). Test cells were loaded with fura 2 as described previously, and after injection immediate examinations for [Ca2+]i were made.
Results are presented as means ± SE, significance level ofP < 0.05 by Student's t-test. Results have been calculated only from those neurons having basal [Ca2+]i levels of >20 and <100 nM. Neurons that did not meet these criteria were felt to be damaged and/or leaky and were excluded from study. At the time of experiments, two criteria were used to determine that the cells of interest were neurons as opposed to glial cells: 1) morphology, where myenteric neurons at 3–5 days are compact, phase bright, and have few or no processes, and glia have a larger, dense nucleus with wide surrounding cytoplasm, and 2) KCl depolarization, as the coverslips were superfused with 55 mM KCl at the end of each experiment. KCl increases neuronal intracellular Ca2+, whereas glial cells do not respond. A neuron was judged to have responded to an agonist if [Ca2+]i doubled the baseline value. A neuron was judged to be a nonresponder if it met the criterion for being a neuron but did not respond to the agonist. Peak [Ca2+]i was measured as the highest Ca2+ concentration achieved during exposure to the agonist.
In this study, n equals the number of neurons. One guinea pig yielded 8–10 coverslips. Dissection techniques, tissue preparation, medium, and reagent vendors remained constant throughout the study. Coverslips were chosen at random, and at least three coverslips were examined for each experimental condition. Only one high-power microscope field was used from each coverslip.
Release of intracellular Ca2+ stores by IP3 and cADPr.
Responsiveness of intracellular Ca2+ stores to IP3 and cADPr was demonstrated by microinjection of agonists into cultured myenteric neurons loaded with fura 2 (Fig.1). In 15 neurons, superfused with a solution containing 1.8 mM CaCl2, microinjection of control buffer increased [Ca2+]i by 179 ± 49 nM. Leakage of extracellular Ca2+ was not responsible for [Ca2+]i increments observed with control injections; in separate experiments (n = 8), performed in nominally Ca2+-free medium, [Ca2+]i increased 159 ± 38 nM with injection of control buffer. Superfusion with 50 μM cinnarizine, an inhibitor of the IP3 receptor, or with 10 μM dantrolene, an inhibitor of the RyR, did not affect [Ca2+]i increments observed during injection of control buffer.
When cultured neurons were microinjected with buffer containing 50 μM IP3, the [Ca2+]i increased by 507 ± 153 nM (Fig. 2). Injection of 50 μM IP3, performed in Ca2+-free superfusion medium, produced similar [Ca2+]i increments of 412 ± 58 nM (n = 7 neurons). The time for Ca2+ transients to return to within 10% of baseline after injection of IP3 was a mean of 101 s. The cytosolic concentration of IP3 produced by microinjection is unknown, but based on dilution of the marker dye sulforhodamine, the injectate was calculated to ∼0.5–2% of cellular volume. Responsiveness to cADPr was demonstrated in 11 neurons by microinjection of buffer containing 50 μM cADPr. In these cells, [Ca2+]i increased by 557 ± 143 nM relative to baseline (Fig. 3). The time for Ca2+ transients to return to within 10% of baseline after injection of cADPr was a mean of 113 s. Similar increases in [Ca2+]i were observed with cultured neurons microinjected with 50 μM cADPr in an extracellular buffer that was Ca2+ free.
Responsiveness to microinjected IP3 was abolished by preincubation of neurons with 50 μM cinnarizine, an inhibitor of the IP3 receptor (Fig. 2) (10, 12). In addition, increments in [Ca2+]i were significantly, but incompletely, inhibited by 10 μM dantrolene, an antagonist of the RyR.
Ca2+ transients induced by microinjection of cADPr were abolished by preexposure of neurons to 10 μM dantrolene (Fig. 3). Responsiveness to microinjected cADPr was also significantly inhibited by cinnarizine, an antagonist of the IP3 receptor.
Fura-loaded cultured neurons were exposed to 50 μM cinnarizine, 10 μM dantrolene, 1 μM ryanodine, or 5 mM procaine for 300 s, with no alteration in basal [Ca2+]i.
[Ca2+]iresponses to extracellular ATP.
Superfusion of fura-loaded myenteric neurons with 100 μM ATP produced rapid and transient increases in [Ca2+]i(Fig. 4). In 23 neurons, [Ca2+]i increased from a basal level of 40 ± 6 nM to a peak of 143 ± 24 nM. Secondary exposure to ATP after an interval of 300 s resulted in repetitive increments in [Ca2+]i in control cells. Neurons in experimental groups were exposed to antagonists for 300 s before the second ATP exposure. In inhibitory studies, percent inhibition was calculated using the following equation: 100 − 100x, where x equals the change in Ca2+ of the second exposure divided by the change in Ca2+ of the first exposure. All inhibitory studies required comparison of the ATP+ antagonist/inhibitor response to the expected response of a second ATP exposure alone. Thus effects of desensitization were distinguished from those of test agents.
Cinnarizine recently has been reported to block Ca2+release from IP3-sensitive intracellular stores (30,31). In 23 neurons, 50 μM cinnarizine inhibited [Ca2+]i increments caused by 100 μM ATP by 45% (Fig. 4). Cinnarizine did not act by inhibition of voltage-sensitive Ca2+ entry. In 30 neurons exposed repetitively to 55 mM KCl, 50 μM cinnarizine reduced [Ca2+]i transients by 6% relative to controls.
Ryanodine, a toxic plant alkaloid, inhibits Ca2+ release from caffeine-sensitive stores in myenteric neurons (19). Ryanodine maintains the RyR in a persistently open state and does not allow refilling of affected Ca2+ stores (28). In 33 myenteric neurons, preexposure to 1 μM ryanodine reduced ATP-evoked increments in [Ca2+]i by 22% (Fig. 4). In control experiments, ryanodine had no effect on KCl-stimulated Ca2+ transients in 19 neurons.
Dantrolene sodium has been demonstrated in prior studies to reversibly inhibit caffeine-stimulated Ca2+ release in cultured myenteric neurons (19). After a 300-s pretreatment with dantrolene (10 μM), ATP-induced Ca2+ transients were reduced by 38% relative to controls (Fig. 4). Control experiments demonstrated that dantrolene did not act via voltage-sensitive Ca2+ channels in myenteric neurons, as Ca2+release stimulated by 55 mM KCl application was unchanged in 12 cells exposed to 10 μM dantrolene for 300 s.
The local anesthetic procaine has been demonstrated previously to inhibit caffeine-stimulated Ca2+ release in myenteric neurons (19). A 300-s exposure of cultured neurons to 5 mM procaine inhibited Ca2+ release in response to 100 μM ATP by 49% (Fig. 4).
Overlap of releasable intracellular Ca2+ stores.
To examine functional overlap of Ca2+ stores, which are released by activation of the IP3R or the RyR, caffeine and extracellular ATP were utilized. Caffeine is a well-characterized agonist of the RyR, which has been demonstrated in previous studies to stimulate Ca2+ release in myenteric neurons (19). ATP also stimulates Ca2+ mobilization in cultured myenteric neurons by a mechanism consistent with activation of the IP3R (18). The experiments were performed by repetitive exposure to deplete caffeine-sensitive or ATP-sensitive Ca2+ stores followed by secondary stimulation with the other agonist (Fig. 5). Experiments were performed in Ca2+-free buffer to prevent repletion of intracellular Ca2+ stores. In 30 neurons that were depleted of ATP-sensitive intracellular Ca2+ stores, a subsequent caffeine superfusion resulted in a peak [Ca2+]i increment of 156 ± 6 nM. In the second series of experiments, caffeine (10 mM) was applied repetitively to deplete caffeine-sensitive stores, followed by superfusion with ATP (100 μM). The ATP-induced residual peak [Ca2+]i increment was 103 ± 7 nM.
Intracellular Ca2+ stores in myenteric neurons also were investigated using the Ca2+ ionophore ionomycin. In Ca2+-free buffer, caffeine-sensitive stores in 24 neurons were depleted by repetitive application of 10 mM caffeine. On subsequent application of 4 μM ionomycin, increases in [Ca2+]i were observed in all cells. Similarly, in 26 neurons, exposure to ionomycin after depletion of ATP-sensitive Ca2+ stores was associated with further increases in [Ca2+]i (Fig.6). Conversely, neither caffeine nor ATP could elicit a [Ca2+]i increase after application of 4 μM ionomycin, indicating that these stores were completely depleted by ionomycin. Ionomycin-treated neurons increased [Ca2+]i in response to 55 mM KCl when Ca2+ was replaced in the buffer.
These studies demonstrate that cultured guinea pig myenteric neurons contain intracellular Ca2+ pools sensitive to IP3 and cADPr agonists, respectively, for the IP3R and the RyR. In addition, the results suggest that significant but incomplete, functional overlap exists between these two Ca2+ pools. Four observations support the latter contention: 1) increases in [Ca2+]i caused by microinjection of IP3 can be antagonized by both cinnarizine, an antagonist of the IP3R, and by dantrolene, an inhibitor of the RyR;2) Ca2+ transients stimulated by microinjected cADPr could also be inhibited by both cinnarizine and dantrolene; 3) superfusion of intact cells with ATP-induced increases in [Ca2+]i that were inhibited by antagonists of both the IP3R cinnarizine and the RyR dantrolene, ryanodine, and procaine; and 4) responsive intracellular Ca2+ pools remained after depletion of ATP-sensitive or caffeine-sensitive Ca2+ stores. Agonist-stimulated responses are distinct from those associated with the act of injection; agonist-stimulated Ca2+ transients were greater in magnitude than those accompanying cell puncture with buffer alone and were unaffected by removal of Ca2+ from the perfusate medium or by co-exposure to cinnarizine or dantrolene.
Ryanodine-sensitive Ca2+ stores have been demonstrated in many excitable cells, including myenteric neurons (3, 19). Several lines of evidence suggest that ryanodine-sensitive stores are indistinguishable from caffeine-sensitive stores, both regulated by the RyR (19, 22, 26, 33). Previous studies have demonstrated caffeine-stimulated Ca2+ release in myenteric neurons, which is inhibited by the RyR antagonists ryanodine, dantrolene, and procaine (19). Immunocytochemical studies in the myenteric neurons also confirmed the presence of ryanodine receptors in the cytosol (19). The best candidate for the endogenous ligand of the RyR appears to be cADPr, a metabolite of NAD+(8, 15, 18). The current studies demonstrate that microinjection of cADPr into myenteric neurons causes an increase in [Ca2+]i. These results are consistent with prior observations, which show that increases in cADPr concentration in other neuronal subtypes induce Ca2+ release and that the releasable store is governed by the RyR (16, 21, 23). The ability of dantrolene, a well-recognized antagonist of the RyR, to abolish the cADPr-induced Ca2+ response in these studies is consistent with a mechanism involving direct release of Ca2+ from the RyR by cADPr.
IP3-sensitive stores are recognized as another important neuronal Ca2+ source. The IP3R has been localized in neurons and has been shown to release Ca2+ in response to IP3 and Ca2+ itself (7, 8,11, 13-14, 27, 29). IP3 is the primary ligand for Ca2+ release at the IP3R (4) and is produced by phosphoinositide (PI) hydrolysis by phospholipase C (PLC) (8, 13). ATP has been shown to increase PI hydrolysis when applied extracellularly to neurons (10). We have previously reported that ATP increases [Ca2+]i in myenteric neurons and that the source is from internally derived Ca2+ sources (18). The effects of ATP on Ca2+ release in the myenteric neurons were abolished by the PLC inhibitor U-73122 (18). Our present results support these findings, inasmuch as intracellular microinjection of IP3 into myenteric neurons stimulated a significant [Ca2+]iincrease, which was blocked by antagonism of the IP3R by cinnarizine. Extracellular ATP also was able to increase [Ca2+]i in the cells, and this effect was also inhibited by IP3R blockade.
Although some cell types possess either IP3- or RyR-sensitive stores alone, there are many cell types that contain both, a group that includes neurons (4). Previously, we have reported immunocytochemical evidence for both ryanodine receptors and IP3R (data not shown) in myenteric neurons (19). Similar findings have been reported in cerebellar Purkinje cells and other neurons; both receptors are associated with the endoplasmic reticulum in the cytosol (34). Bennett et al. (3) recently reported that in the neuronal PC12 cell line, internal Ca2+ stores had a physical overlap, with a percentage of stores sensitive to IP3R agonists, RyR agonists, or both. Kozuimi et al. (20) reported in PC12 neuronal cells that the IP3- and RyR-sensitive stores were expressed throughout the cytoplasm; confocal microscopy localized specific release sites that responded to agonists for the IP3R, the RyR, or both. Antagonizing or depleting one of these storage sites blunted the response to the other agonist. Our results in the myenteric neurons are consistent with these findings, inasmuch as [Ca2+]i transients after direct microinjection of IP3 were diminished in the presence of the RyR antagonist dantrolene. Similarly, the effects of microinjected cADPr were partially blocked by the IP3R antagonist cinnarizine. Superfusion studies with extracellular ATP corroborate these findings because the ATP response was blocked by antagonism of either the IP3R or the RyR.
Although IP3R and RyR receptors demonstrate functional overlap, the overlap appears to be incomplete. Other authors have reported unique distributions, especially in neuronal dendrites and spines and evidence of functional distinctness (4, 15,34). Kozuimi et al. (20) demonstrated that although some of the IP3R stores showed a blunted response after RyR antagonism, there were other populations of IP3R release sites that continued to respond normally. Similarly, the current study has demonstrated substantial internal Ca2+ stores that were able to respond to ATP after depletion of the caffeine-sensitive stores; caffeine-sensitive stores also remained after depletion of ATP-sensitive stores. That ionomycin evoked a Ca2+ response after depletion of these stores suggests the existence of additional Ca2+ stores from nuclear, mitochondrial, or other sources (1, 25, 32).
An unexpected finding in the current study was the incomplete antagonism of ATP-stimulated Ca2+ release by cinnarizine in experiments that utilized superfusion of extracellular ATP. We have previously demonstrated complete abolition of the ATP-stimulated Ca2+ response by inhibition of PLC, and in the current study we have shown complete blockade of the IP3-stimulated Ca2+ response by cinnarizine in microinjection experiments (18). The possibility exists of additional agonist effects of ATP. There is evidence for ATP as an allosteric activator of the IP3R in studies performed using reconstituted lipid vesicles (6, 12); these studies also showed inhibitory effects of ATP at the IP3R at higher concentrations. If ATP generates IP3 through PI hydrolysis by PLC, there is also a possible stimulatory effect of diacylglycerol (DAG) produced by PLC. There is recent evidence that uncaged DAG will lead to Ca2+release, reportedly through the ryanodine system (24). One plausible explanation for the incomplete blockade of Ca2+by cinnarizine in ATP stimulation is that the generated DAG is acting in a stimulatory fashion.
In the myenteric neuron, interactions between internal Ca2+stores appear to be complex. Our data imply the existence of two internal Ca2+ store subtypes. The IP3R- and RyR-operated stores appear to have functional overlap. This redundancy ensures a pathway for Ca2+ mobilization on cell stimulation and may allow for accentuated release. Still undetermined is whether this functional overlap is due to CICR from neighboring stores or to expression of both receptor types on one store (Fig.7).
In conclusion, myenteric neurons possess internal Ca2+stores populated by two receptor subtypes: the IP3R and the RyR. Although they are of distinct receptor systems, they share a partial functional overlap in the process of generating a Ca2+ response.
This work supported by the National Institute of Diabetes and Digestive and Kidney Diseases Research Grant DK-41204 and by the Frederick A. Coller Surgical Society Research Fellowship.
Address for reprint requests and other correspondence: M. W. Mulholland, 2922 Taubman Center, 1500 East Medical Center Dr., Ann Arbor, MI 48109-0331 (E-mail:).
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
- Copyright © 2001 the American Physiological Society