The aim of this study was to characterize in vivo rat colonic motor activity in normal and inflamed states and determine its neural regulation. Circular muscle contractions were recorded by surgically implanted strain-gauge transducers. The rat colon exhibited predominantly giant migrating contractions (GMCs) whose frequency decreased distally. Only a small percentage of these GMCs propagated in the distal direction; the rest occurred randomly. Phasic contractions were present, but their amplitude was very small compared with that of GMCs. Inflammation induced by oral administration of dextran sodium sulfate suppressed the frequency of GMCs in the proximal and middle but not in the distal colon. Frequency of GMCs was suppressed by intraperitoneally administered atropine and 4-diphenylacetoxy-N-methyl-piperidine methiodide and was enhanced by N w-nitro-l-arginine methyl ester. Serotonin, tachykinin, and calcitonin gene-related peptide receptor or receptor subtype antagonists as well as guanethidine and suramin had no significant effect on the frequency of GMCs. Verapamil transiently suppressed the GMCs. In conclusion, unlike the canine and human colons, the rat colon exhibits frequent GMCs and their frequency is suppressed in inflammation. In vivo GMCs are stimulated by neural release of acetylcholine that acts on M3 receptors. Constitutive release of nitric oxide may partially suppress their frequency.
- N ω-nitro-l-arginine methyl ester
- peristaltic reflex
- enteric neurons
intact colons of humans and dogs in the conscious state generate three distinct types of contractions: rhythmic phasic, giant migrating (GMCs), and tonic contractions (33, 34). Spatial and temporal patterns of these contractions, their electrophysiological regulation, and cellular signaling pathways differ. Rhythmic phasic contractions predominate in fasting and fed states in these species. They are largely disorganized in time and space and either do not propagate or propagate over short distances of a few centimeters. These contractions, regulated by slow waves and contractile electrical complexes (9, 21, 32, 41), cause mixing and slow net distal propulsion of luminal contents (39). The GMCs are large in amplitude, occlude the lumen, and propagate uninterruptedly over long distances and hence produce mass movements (23, 39). The GMCs usually precede defecation (23,27). The myoelectric correlates of GMCs are not completely known, but it is known that they are not regulated by slow waves, because their duration is severalfold longer than that of a slow wave cycle (18, 20, 23, 25-27, 38, 39). A tonic contraction is also not regulated by slow waves for the same reason. The precise role of increase in tone in colonic motor function is not known, but it is thought that by reducing the lumen size, it may enhance the propulsive and mixing movements of the phasic contractions and GMCs.
Frequency of colonic GMCs, and hence of mass movements, is increased in ulcerative colitis patients and in experimental models of inflammation (25, 38). At the same time, the phasic contractions and tone are suppressed in colonic inflammation (25, 38). Frequent mass movements produced by GMCs and absence of mixing movements due to the suppression of phasic contractions are major factors in producing inflammatory diarrhea.
Rodents are now increasingly being used in investigations of enteric smooth muscle and neural function because of the availability of transgenic animals. However, patterns of contractions in the in vivo rat colon in the intact conscious state have not been defined, and, therefore, it is difficult to extrapolate the findings from this species to others. Most previous studies have used myoelectrical tracings consisting of short- and long-duration spike bursts to describe in vivo colonic activities in rats (10, 12). Ferre and Ruckebusch (10) recorded concurrent colonic myoelectric and contractile activities from normal conscious rats but quantitated primarily the myoelectrical activity that did not distinguish between phasic contractions and GMCs. Some investigators have used strain-gauge transducers to record colonic motor activity from anesthetized rats (28, 44).
It is known that stooling habits, and shape and consistency of stools in rats, differ from those in dogs and humans. The rats stool frequently and defecate pellets, whereas dogs and humans have formed stools once or twice a day in health. Our hypothesis is that the in vivo motor patterns of the rat colon differ from those of humans and dogs to account for the above differences in motor functions. In particular, the rat colon may generate more frequent GMCs. Accordingly, our specific aims were to 1) record in vivo colonic motor patterns from conscious rats, 2) identify specific neurotransmitters and receptors that regulate in vivo GMCs, and3) determine how GMCs and phasic contractions are modulated in dextran sodium sulfate (DSS)-induced inflammation of the colon. We also concurrently recorded ileal motor activity to determine whether changes due to inflammation differ between the two organs.
MATERIALS AND METHODS
Two commonly used strains of rats, Lewis and Sprague Dawley (300–380 g; Harlan Sprague-Dawley, Indianapolis, IN) were used to determined whether the in vivo motor patterns are strain dependent. Later on, the Lewis rats were used for inflammation experiments and the Sprague-Dawley rats for pharmacological experiments. The Sprague-Dawley rats were used for pharmacological experiments, because substantial in vitro pharmacological data are available in this strain. The rats were housed under controlled illumination (lights on from 6 AM to 6 PM) and temperature (22°C) following institutional guidelines for the care and use of laboratory animals as approved by the Animal Care Committee at the Zablocki Veterans Affairs Medical Center. The rats were fed a low-residue diet (Harlan/TekLAD, Madison, WI) ad libitum for 5 days after surgical strain-gauge transducer implantation; thereafter, they were fed standard rat chow with free access to water.
Induction of Experimental Colitis
The rats were administered a 5% solution of DSS ad libitum (molecular wt 40,000; ICN, Aurora, OH) in the drinking water for 7 consecutive days with free access to food to induce colonic inflammation. The rats also had free access to food in their cages throughout the experimental days.
Construction of Strain-Gauge Transducers
Construction of strain-gauge transducers for large animals has been described previously (1, 8); it was adapted for use in the rat. Briefly, each cannula consisted of six miniaturized strain-gauge transducers (model EA-06–031 DE-120; Measurements Group, Raleigh, NC) cut to a size 2 × 4 mm2. Teflon-coated silver-plated copper wire (model 336-FTE; Measurement Group) was soldered to the transducers, and then the gauge was affixed to a plastic backing for reinforcement. The gauge was then covered with an acrylic coating (model M-Coat D; Measurement Group), a silicone rubber coating (model 3145 RTV; Dow Corning, Midland, MI), and a nitrile rubber coating (model M-Coat B; Measurement Group). Two silicone sheets (0.25 mm; Technical Products, Decatur, GA) and silicone glue were used to sandwich the gauge, and then they were hardened in a mold for 24 h. The silicone sheets were trimmed to a final size of 5 × 6 mm2. Wires were soldered to a connecting plug (Amphenol), and gauges were calibrated using known gram weights. The whole assembly was gas sterilized before implantation of the transducers on the colon.
Surgical Implementation of Strain-Gauge Trandsucers
Rats were fasted overnight with free access to water before midline laparotomy. Anesthesia was induced by an intraperitoneal injection of pentobarbital sodium (50 mg/kg Nembutal; Abbott Laboratories, N. Chicago, IL). Rats were also given 0.2 mg of atropine (0.4 mg/ml; Elkins-Sinn, Cherry Hill, NJ) to reduce oral secretions. The strain-gauge transducers were oriented to record circular muscle contractions and sewn onto the seromuscular layer of the colon and ileum using 6-0 polypropylene sutures. Five transducers were implanted on the colon at 3.1 ± 0.4, 6.9 ± 0.7, 11.1 ± 0.9, 13.4 ± 1.1, and 16.9 ± 1.0 cm from the cecocolonic junction. The most distal gauge was above the peritoneal reflection. The transducer closest to the cecum was used to analyze frequency and amplitude of contractions from the proximal colon; the transducer closest to the peritoneal reflection, to analyze similar data from the distal colon; and the middle transducer, to analyze the data from the middle colon. The remaining two transducers along with the other three transducers were used to analyze the propagation characteristics of GMCs. An additional strain-gauge transducer was implanted onto the terminal ileum 10 cm proximal to the ileocecal junction. Wires were exteriorized through the abdominal wall and subcutaneously tunneled to between the shoulder blades where a small skin incision was made and the connecting plug was sutured to the skin. Rats were allowed to recover for 5 days during which time they were fed a low-residue diet. Regular rat chow was given during experiments.
Motility recordings and analysis.
Conscious rats were put in nonrestraint polycarbonate cages, and the cannulas were connected to a multichannel pen recorder (model O polygraph; Grass, Quincy, MA). Signals were recorded simultaneously on a personal computer using dedicated software (version 1.68; Data Q Instruments). Each rat served as its own control. The animals were recorded for 4–6 h at the same time every morning without prior overnight fasting. Control recordings were made for 2–3 days, and then recordings were made for 7 consecutive days during DSS-induced inflammation. The frequencies of migrating motor complex (MMC) activity in the ileum, GMCs in the ileum and colon, propagating GMCs, and phasic contractions in the colon were analyzed visually. Amplitude of GMCs during the last 10 min of the first and last hour of recordings was averaged. The GMCs were defined as contractions of duration >150% and amplitude >300% of that of phasic contractions at the same recording site. The GMCs were considered to propagate if they traveled over at least three consecutive transducers. The mean velocity of propagation of at least six clearly propagated GMCs was determined (see Fig. 3). All other GMCs that fell within mean ± 2 SD of this value were considered to propagate. Phasic contractions were identified by their frequency in the range of 10–13 c/min, which is the frequency of short-duration spike bursts (10). Data are expressed as the means ± SE. Differences between groups were determined by one-way repeated measures of variance followed by the Student-Newman-Keuls test; n represents the number of rats.
Sections of the distal colon from normal and DSS-treated rats were collected, and an assay evaluating the production of myeloperoxidase (MPO) was performed using an established method (4).
Materials, Rats, and Supplies
Lewis and Sprague-Dawley rats (250–300 g each) were obtained from Harlan, Indianapolis, IN. Hexamethonium chloride, verapamil hydrochloride, guanethidine,N w-nitro-l-arginine methyl ester (l-NAME), suramin sodium salt, 4-diphenylacetoxy-N-methyl-piperidine methiodide (4-DAMP),l-703,606 oxalate salt, LY-53,857 maleate, NAN-190 HBR, SDZ-205–557, methoctramine tetrahydrochloride, and pirenzipine dihydrochloride were purchased from Sigma (St. Louis, MO), MEN · 10,376[Tyr5,d-T-p6,8,9 Lys10] neurokinin (NK)A4–10, [Trp7, β-Ala8] NKA4–10, and calcitonin gene-related peptide (CGRP)8-37 were from Peninsula (Belmont, CA), and atropine sulfate was from Fujisawa (Deerfield, IL). All substances were administered intraperitoneally.
All instrumented rats recovered from surgery within several hours as evidenced by normal feeding and grooming behavior. They lost up to 10% of their body weight initially after surgery but recovered before the beginning of the recording sessions. The mean daily intake of DSS mixed with water was 32 ± 2.4 ml. The Lewis rats developed nonbloody loose stools by ∼3 days post-DSS intake. Mean weight loss after 7 days of DSS treatment was 13 ± 3%. MPO activity was significantly increased in the distal colon and ileal tissue (P < 0.05, n = 6; Table1) on the 7th day of DSS treatment.
Spontaneous In Vivo Motor Activity of the Normal Rat Colon
In the normal state, the colons of Lewis rats predominantly exhibited GMCs in the proximal, middle, and distal colons (Fig.1) at frequencies of 44.1 ± 6.0, 34.5 ± 6.1, and 16.9 ± 3.0 per hour, respectively (Fig.2). Frequency in the distal colon was significantly less than that in the proximal or the middle colon (n = 6, P < 0.05). Frequency of GMCs propagating over at least three consecutive strain-gauge transducers (Fig. 3) was 5 ± 1 per hour in the normal state.
Amplitude of phasic contractions in the proximal, middle, and distal colon was very small compared with that of GMCs (shown by arrow in Fig.1). At the sensitivity used to record GMCs in full scale, sometimes the phasic contractions seemed to be absent. The mean frequencies of phasic contractions when they were visually detectable were 12.5 ± 0.5, 10.5 ± 0.l, and 12.5 ± 0.7 per minute in the proximal, middle, and distal colons respectively (P > 0.05,n = 6).
Basal motor patterns in Sprague-Dawley rats were similar to those in Lewis rats. Frequencies of GMCs in the proximal, middle, and distal colons of these rats were 37 ± 5.2, 33 ± 3.6, and 15 ± 3.2 per hour, respectively (n = 4). There was no significant difference between the frequencies of GMCs in the Lewis and Sprague-Dawley rats. The amplitude of phasic contractions in Sprague-Dawley rats was also very small compared with that of the GMCs. Mean frequencies of phasic contractions in these rats were not different from those in Lewis rats (data not shown).
Effect of Inflammation on Colonic Motor Activity
Inflammation suppressed the frequency of GMCs in the proximal and middle but not in the distal colon (Fig. 2). In the proximal colon, decrease in GMC frequency was significant on the 7th day of DSS treatment, whereas in the middle colon, the decrease was significant on the 3rd and the 7th days of DSS treatment (Fig. 2). The frequency of propagating GMCs also decreased significantly in inflammation from 5 ± 1 per hour before inflammation to 2.3 ± 1 per hour onday 3 and to 1.4 ± 0.7 per hour on day 7 of DSS treatment (n = 6, P < 0.05 control, days 3 and 7). There was no significant difference in the amplitudes of GMCs (Table2) between the normal and the inflamed colons.
Ileal Motor Activity in Normal and Inflamed States
The terminal ileum displayed periodic phase IIIactivity in the normal state at a frequency of 1.5 ± 0.2 per hour, which significantly decreased to 0.8 ± 0.3 per hour on day 3 and to 0.5 ± 0.2 per hour (n = 6, P < 0.05) on day 7of DSS treatment (Fig. 4, Aand B).
The ileum in four of seven rats also exhibited irregularly occurring GMCs (Fig. 4 C). In the normal state, they occurred at a frequency of 0.7 ± 0.3 per hour (n = 7). Their frequencies on the 3rd and 7th days of DSS treatment, 1.3 ± 0.7 and 0.9 ± 0.5 per hour, respectively (n = 7), were not different from those in the normal state.
Neurotransmitters and Receptors Regulating Spontaneous In Vivo GMCs
Cholinergic and nicotinic receptors.
Intraperitoneal administration of 100 μg/kg atropine, a nonspecific cholinergic receptor antagonist, or 10 mg/kg hexamethonium, a nicotinic receptor antagonist, significantly inhibited the spontaneous GMCs throughout the colon (Fig. 5). Frequency of GMCs was suppressed for at least 60 min after the administration of these drugs. By contrast, 100 μg/kg pirenzepine ip, a muscarinic M1 receptor antagonist, or 100 μg/kg methoctramine ip, a muscarinic M2 receptor antagonist, had no significant effect on the frequency of GMCs (n = 4 each; data not shown). In two rats, intraperitoneal administration of 100 μg/kg 4-DAMP, an M3 receptor antagonist, had effects similar to those of atropine (data not shown).
Serotonergic, tachykinin, and CGRP receptors.
Intraperitoneal administrations of NAN-190 HBR, a 5-HT1Areceptor antagonist; LY-53857, a 5-HT2 receptor antagonist; granisetron, a 5-HT3 receptor antagonist; and SDZ 205–93, a 5-HT4 receptor antagonist (each 0.5 mg/kg,n = 4) had no significant effect on frequency of GMCs in the proximal, middle, or the distal colon (data not shown). Similarly, intraperitoneal administration of NK1 receptor antagonist, L-703,606; NK2 receptor antagonist, MEN 10,376 [Tyr5-d-Trp6,8,9-Lys10]NKA4–10; NK3 receptor antagonist, [Trp7-β-Ala8]-NKA4–10(each 1 mg/kg, n = 4), and CGRP antagonist (α-CGRP8–37; 125 μg/kg; n = 4) had no significant effect on the frequency of GMCs in the proximal, middle, or the distal colon (data not shown).
Nitrergic, Purinergic, and Sympathetic Neurons
Intraperitoneal administration of 10 mg/kg l-NAME significantly increased the frequency of GMCs in the proximal colon for at least 60 min, but it had no significant effect on them in the middle or the distal colon (n = 4, Fig.6). By contrast, 0.5 mg/kg suramin ip, a nonspecific antagonist of purinergic receptors, had no effect on GMC frequency in any part of the colon (data not shown). Guanethidine (1 mg/kg ip), which depletes norepinephrine from the sympathetic neurons, also had no significant effect on the frequency of GMCs (Fig. 6).
L-Type Ca2+ Channels
Verapamil (5 mg/kg ip, n = 4) significantly reduced the frequency of GMCs in the proximal colon for 45 min. This effect lasted for only 15 min in the middle and the distal colons. Smaller doses of verapamil (≤1 mg/kg) had no significant affect on the frequency of GMCs in any part of the colon.
Our findings show that the composition of spontaneous motor patterns in the colon of intact conscious rats differs markedly from that in the canine or the human colon. The predominant motor activity of human and dog colons consists of rhythmic phasic contractions (32-34). In the dog colon, the phasic contractions are organized as migrating and non-MMCs (35). The organization of phasic contractions as motor complexes has not yet been established in the human colon. Healthy colons of both species generate infrequent GMCs, once or twice a day (2, 18, 20, 25, 26,38). By contrast, the predominant motor activity of the normal rat colon is GMCs. However, unlike human and dog colons, a majority of GMCs in the rat colon do not propagate or propagate over distances of less than a few centimeters, which was the distance between the successive transducers in our study. It seems that the composition of colonic motor activity in rats is not strain-dependent, at least as far as the Lewis and Sprague-Dawley strains are concerned. Other investigations (28, 44) have reported previously in vivo occurrence of contractions in anesthetized Wistar and Sprague-Dawley rats that look similar to GMCs reported in our study. Fida et al. (11) also recorded GMCs from ex vivo mouse colon but termed them colonic MMCs.
Interspecies differences in the composition of motor activity have been noted also between the cat small intestine and the human and canine small intestines. In the fasting state, the canine and human small intestines generate phasic contractions organized as MMCs (42,46). By contrast, the cat small intestine does not generate MMCs but instead exhibits GMCs that usually propagate over long distances (30, 36). As reported previously by others (3,37), the ileums of rats in our study generated regular MMCs. However, in about half the rats, the ileum also generated GMCs at a frequency of ∼1 h; the GMCs are not observed so frequently in the normal human or the canine small intestine (24, 31). Scott and Tan (37) found that the frequency of GMCs in the rat jejunum is increased in food protein-induced intestinal anaphylaxis.
Amplitude of phasic contractions in the rat colon was very small compared with that of GMCs. However, the frequency of phasic contractions was not different among the proximal, middle, and distal colons. The relatively small amplitude of phasic contractions suggests that, in this species, their role in mixing and propulsion of colonic contents may be minimal compared with that of the GMCs. The propagating GMCs may cause propulsion. In support of this, Tomaru et al. (44) found that glycerol anema in anesthetized rats stimulates GMCs that propagate in the distal colon and hence cause rapid evacuation of colon contents. On the other hand, Mizuta et al. (28) reported that intravenous administration ofl-NAME stimulates random uncoordinated GMCs in the rat colon and hence delays transit.
In vivo neural regulation of rat colonic GMCs has not been reported previously. Our findings show that they are regulated by concurrent excitatory input from the cholinergic neurons and inhibitory input from the nitrergic neurons. The GMCs were inhibited by hexamethonium, indicating that the stimulus for these contractions originates from presynaptic neurons and involves at least one nicotinic synapse. The GMCs were inhibited also by atropine but not the NK1antagonist, indicating that the final neurotransmitter at the neuroeffector junction is acetylcholine, not substance P, although the two excitatory neurotransmitters are colocalized in the same neurons (6). Also, the stimulation of GMCs seems to be mediated by M3, but not M1 and M2 receptors. Atropine and 4-DAMP, but not pirenzepine and methoctramine, blocked them.
Inhibition of nitric oxide (NO) synthase (NOS) by l-NAME enhanced the frequency of GMCs in the proximal colon, indicating a constitutive release of NO to suppress their frequency. However,l-NAME had no significant effect on the frequency of GMCs in the middle and the distal colon, indicating a lack of sufficient constitutive release of NO to affect the spontaneous frequency of GMCs in these parts of the colon. These data are in agreement with the findings of Takahashi and Owyang (43) who found that the number of NOS-containing neurons, as well as NOS protein and its mRNA, is significantly greater in the proximal than in the distal rat colon. They did not investigate the middle colon, but our findings suggest that the NO neurons in the middle colon may also be fewer than those in the proximal colon.
In contrast to nitrergic neurons, the sympathetic or purinergic neurons do not seem to regulate the spontaneous GMCs in rat colon because their respective blockade had no effect on the frequency of GMCs. Hata et al. (19) also found that the descending inhibition in the proximal rat colon by balloon distension may not be mediated by ATP. The selectivity of NO in inhibiting the frequency of GMCs is supported by the findings in muscle strips that pyridoxal phosphate-6-azophenyl-2',4'-disulfonic acid (P2receptor antagonist), reactive blue (P2γ receptor antagonist), as well as MDL-12330 (adenylate cyclase inhibitor) do not affect the amplitude or frequency of GMCs, whereas l-NAME enhances them (15). By contrast, there is evidence that both NO and vasoactive intestinal polypeptide may be involved in producing descending relaxation when flat sheet muscle strips are prestretched (17). Therefore, it seems that the neural mechanisms of inhibition of spontaneous GMCs in the intact conscious state may differ from those that produce descending inhibition in in vitro preparations.
Blockade of serotonergic and tachykininergic receptor subtypes also had no significant effect on the frequency of spontaneous GMCs. The doses of antagonists we used are similar to those that are reported to be effective in vivo (16, 45). However, both of these receptor subtypes have been reported to be involved in the initiation of orad excitation (increase in tone) due to mucosal stroking or radial stretch in in vitro flat sheets of muscle preparations (7, 13,22). As noted above, the neural mechanisms of stimulation of spontaneous GMCs in the intact state may be different from those that increase the orad muscle tone in vitro.
Recent studies (14) show that the circular muscle strips prepared from the rat middle colon also generate giant contractions (GCs). However, there are important differences between the neural regulation of giant contractions in vivo and in vitro. Whereas, in vivo, the GMCs in the rat colon were blocked by atropine and hexamethonium, in vitro the GCs were not affected by these cholinergic and nicotinergic antagonists, indicating myogenic origin of in vitro GMCs. According to Huizinga et al. (21), the spontaneous contractile activity in muscle strips may be induced by stretch that opens Ca2+ channels. The inhibition of NOS byl-NAME, however, enhanced the frequency of in vitro GCs, which is similar to the inhibition observed for in vivo GMCs. The constitutive release of NO to inhibit GMCs is, therefore, preserved in muscle strips. whereas the cholinergic input may be impaired in them.
Inflammation has been reported to enhance the frequency of GMCs in canine and human colons (25, 38) in which they occur very infrequently. In the rat colon, where GMCs occur regularly, their frequency was reduced during inflammation in the proximal and the middle colons. Decrease in the frequency of GMCs during inflammation in rat colon may be due to a decrease in acetylcholine release as reported by Collins et al. (5) and Shi and Sarna (40), as well as due to alterations in excitation-contraction coupling, because the response to acetylcholine in circular muscle strips is also suppressed (14, 29, 47). Myers et al. (29) found that inflammation retards transit in the rat colon. Our data suggest that the slower transit in the inflamed rat colon may be due to the suppression of propagating GMCs.
A 5.0 mg/kg dose of verapamil inhibited GMCs differentially in the proximal, middle, and distal colons. The effect lasted for at least 45 min in the proximal colon but only 15 min in the distal colon. This suggests that the expression of L-type Ca2+ channels may be greater in the distal than in the proximal colon.
In conclusion, the predominant motor activity of the rat colon in the intact conscious state is composed of GMCs. These GMCs are stimulated by acetylcholine release from the cholinergic excitatory neurons and are partially suppressed in frequency by the constitutive release of NO from the nonadrenergic noncholinergic neurons. Serotoninergic and tachykininergic receptors and purinergic and sympathetic neurons do not seem to be involved in the spontaneous occurrence of these contractions. Inflammation reduces the frequency of rat colonic GMCs that may be the basis of slower transit in the inflamed colon. Diarrhea in DSS-induced inflammation in rats may primarily be of secretory origin. In other species, such as dog and man, the frequency of GMCs is increased in colonic inflammation, causing frequent mass movements and hence diarrhea. Frequent occurrence of GMCs in the rat colon represents a unique animal model to investigate their regulation and function.
This study was supported, in part, by National Institute of Diabetes and Digestive and Kidney Diseases Grant DK-32346 (to S. K. Sarna) and the Department of Veterans Affairs Medical Research Service (to S. K. Sarna).
Address for reprint requests and other correspondence: S. K. Sarna, Division of Gastroenterology, Dept. of Internal Medicine, University of Texas Medical Branch at Galveston, 1108 The Strand, Galveston, TX 77555-0632 (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.
October 17, 2001;10.1152/ajpgi.00114.2001
- Copyright © 2002 the American Physiological Society