Stomach region stimulated determines effects on duodenal motility in rats

Abstract

Gastric electrical stimulation (GES) is used clinically to promote proximal GI emptying and motility. In acute experiments, we measured duodenal motor responses elicited by GES applied at 141 randomly chosen electrode sites on the stomach serosal surface. Overnight-fasted (H₂O available) anesthetized male rats (n = 81) received intermittent biphasic GES for 5 min (20-s-on/40-s-off cycles; I = 0.3 mA; pw = 0.2 ms; 10 Hz). A strain gauge on the serosal surface of the proximal duodenum of each animal was used to evaluate baseline motor activity and the effect of GES. Using ratios of time blocks compared with a 15-min prestimulation baseline, we evaluated the effects of the 5-min stimulation on concurrent activity, on the 10 min immediately after the stimulation, and on the 15-min period beginning with the onset of stimulation. We mapped the magnitude of the duodenal response (three different motility indices) elicited from the 141 stomach sites. Post hoc electrode site maps associated with duodenal responses suggested three zones similar to the classic regions of forestomach, corpus, and antrum. Maximal excitatory duodenal motor responses were elicited from forestomach sites, whereas inhibitory responses occurred with stimulation of the corpus. Moderate excitatory duodenal responses occurred with stimulation of the antrum. Complex, weak inhibitory/excitatory responses were produced by stimulation at boundaries between stomach regions. Patterns of GES efficacies coincided with distributions of previously mapped vagal afferents, suggesting that excitation of the duodenum is strongest when GES electrodes are situated over stomach concentrations of vagal intramuscular arrays, putative stretch receptors in the muscle wall.

INTRODUCTION

Gastric electrical stimulation (GES) is presently being investigated intensively for its therapeutic potential for treatment of various stomach disorders (e.g., gastroparesis and reflux disease) and gastric signaling dysfunctions (e.g., obesity and type II diabetes) in different animal models (1–4) and in multiple clinical trials (5–7). Outcomes from those studies have been variable, however, and underlying mechanisms of actions remain largely unknown.

It appears that one cause of the variability in GES-evoked effects might be the consequence of using different electrode placements on the stomach. Animal studies have indicated that stimulation electrodes implanted on the distal gastric antrum produced anorexigenic and orexigenic peptides in the hypothalamus (8), whereas electrodes implanted along the lesser curvature could be used to treat obesity (9). Human studies also show that recording electrodes implanted on the gastric fundus that provided closed-loop control of stimulating electrodes on the mid antrum can be used to treat obesity with the TANTALUS system(7), whereas stimulation electrodes on the human antrum near the greater curvature can apparently improve symptoms of idiopathic gastroparesis using the Enterra system (6). Thus, overall, it appears that different stimulation electrode locations on the stomach may explain variability in the effects of GES.

To understand how GES of different stomach regions might affect small intestinal motility (and potentially duodenal-gastric feedback), we generated a detailed topographic map of regional gastric effects on duodenal motility, using duodenal motor responses elicited by gastric stimulation in different stomach regions employing stimulus parameters that activate fine c-fibers (10). Furthermore, to evaluate the underlying GES mechanisms, we also compared our newly obtained functional maps with anatomical stomach afferent maps previously produced with neural tracers (11, 12).

The generation of gastric functional maps and comparisons with anatomical maps might provide foundational information for GES neuromodulation.

MATERIALS AND METHODS

Animals

Two- to four-month-old male Sprague-Dawley rats (n = 81; Envigo, Indianapolis, IN) weighing 319 g (SD = 34 g) were housed individually in an Association for Assessment and Accreditation of Laboratory Animal Care-approved temperature (22–24°C)- and humidity (40%–60%)-controlled colony room. The room was maintained on a 12-h light-dark schedule (6:00 am–6:00 pm light period; 6:00 pm–6:00 am dark period). Pelleted chow (2018 Teklad global 18% protein rodent diet) and filtered tap water were provided ad libitum. All husbandry practices conformed to the NIH Guide for the Care and Use of Laboratory Animals (8th edition) and were reviewed and approved by the Purdue University Animal Care and Use Committee. All efforts were made to minimize any suffering as well as the number of animals used.

Surgical Procedures

Animals were transferred to wire hanging cages the day before surgery and initially fasted for 18 h (2:00 PM to 8:00 AM) with free access to water before surgery (8:00 AM). Rats were then anesthetized with 5% isoflurane (Akon Animal Health, Cat. No. NDC: 59399-106-010) in an induction box and then were transferred to a SurgiSuite platform of the SomnoSuite anesthesia system (Kent Scientific SomnoSuite). A servo-controlled homoeothermic heating blanket, equipped with a rectal thermometer, was used to maintain body temperature at 36°C during the whole experimental period. The level of anesthesia was reduced to 2.5% isoflurane for the surgical procedure. After a midline laparotomy, the stomach and 3–4 cm of proximal duodenum were partially exteriorized onto saline-soaked gauze pads. Custom-made EMG bipolar patch electrodes (MicroProbes for Life Science, Gaithersburg, MD, 20879; consisting of two Pt/Ir foil plates attached to a silicone-filled mesh) were sutured to the serosal surface of the stomach with four stitches of 8-0 suture (Nylon black, S&T AG, Toberlrasstrasse 2, CH-8212 Neuhausen, Switzerland) on the corners, with a line through the two electrode contacts longitudinally aligned (Fig. 1). Each metal electrode pole contact plate area was 2 mm × 1 mm. The impedance between plates was 0.4–0.6 kΩ in normal saline and 1.0–2.0 kΩ on the surface of stomach measured with the PlexStim electrical stimulator at 1 kHz. Electrode implant locations were chosen to cover the entire stomach surface as randomly as possible over the population of animals. The fine wire leads attached to the patch electrodes were exteriorized through the midline laparotomy. Twelve animals were prepared with three GES electrodes; 36 animals were prepared with two GES electrodes; 33 rats were prepared with a single GES electrode.

Figure 1.

Image of gastric patch electrode, duodenal strain gauge, and schematic experimental setup with implants illustrated at scale. A shows a stimulation patch electrode (consisting of two plates) sutured on the serosal surface of the antrum (empty stomach owing to overnight fast) together with a strain-gauge transducer bonded on the serosal surface of the proximal duodenum. B shows a schematic of the experimental preparation superimposed on a diagram of a full stomach. Electrodes were placed across the stomach from antrum to corpus to forestomach. Electrode orientation paralleled the longitudinal muscle using the greater curvature as a landmark in the functional map study. A custom-made foil strain-gauge transducer was attached to the proximal duodenum to record duodenal motor activity. Experiments were carried out on acutely anesthetized rats.

Strain-gauge elements for the duodenal transducers were purchased from Vishay Micro-measurements (EA-06-031CE-350) and assembled into custom-made strain-gauge transducers (13) (4 × 3.5 mm, Clunbury Scientific LLC, Bloomfield Hills, MI) that were then attached to the serosal surface of the proximal duodenum (5–15 mm distal to the pyloric sphincter) using Vetbond glue (Tissue adhesive no.1469SB; 3 M). The strain-gauge sensing axis was oriented parallel to the longitudinal muscle (Fig. 1). The fine wire leads attached to the strain gauge were exteriorized through the midline laparotomy. The animal was kept in a supine position, with the abdominal midline incision covered by saline-soaked gauze pads. Physiological saline was infused into the intraperitoneal space continuously (2.0 mL/h ip) using a syringe pump (GenieTouch Kent Scientific). The animal was then covered with a blanket to help maintain body temperature, and anesthesia was reduced to 1.5% isoflurane and maintained at that level for the remainder of the experiment.

Duodenal Motility Recording and Stimulation on Stomach Surface

After electrode and strain-gauge placement, recording of duodenal motility began. Strain-gauge measurements were made using a DC bridge amplifier from MDE GmbH (Walldorf, Germany) with the sensitivity to record the small voltage output changes from the strain-gauge flexures during duodenal contractions. Once motility (the output of the strain gauge) reached a stable baseline (usually within a 1-h period after surgery) and had been stable for at least 20 min, stimulation was initiated. Stimulation was provided by a PlexStim electrical stimulator (Plexon Inc., Texas). Stimulation parameters (chosen to ensure effective c-fiber activation) were as follows: biphasic; I = 0.3 mA; pw = 0.2 ms; 10 Hz; 20 s-on-40s-off; five 1-min cycles (for a total stimulation period of 5 min). Following stimulation, recording continued, usually for another hour.

Animal Perfusion, Electrode Location Measurement Postmortem

Once the recording was complete, each animal was given a lethal dose of ketamine (Patterson Veterinary, 275 mg/kg ip) and Xylazine (Acorns Animal Health, 27.5 mg/kg ip). The locations of electrodes used in the experiment were marked with 5-0 blue suture (Monofilament Polypropylene, Surgipro II) before the electrodes were removed. To ensure that the stomach was normally distended at the time of fixation, the organ was inspected for normal distension or accommodation, and, as required, physiological saline (3.3 mL/100 g wt) that had been warmed to body temperature was slowly infused into the stomach by a gavage catheter. With the stomach normally dilated, the animal was first transcardially perfused through the vasculature with physiological saline and then with 4% paraformaldehyde in 0.1 mol/L phosphate‐buffered saline (PBS; pH 7.4). After perfusion, the distal esophagus and the proximal duodenum were transected, and the stomach was freed and removed. The organ was then opened with a longitudinal cut along the greater curvature.

Next, the ventral and dorsal stomach walls were separated by an incision along the lesser curvature, thus yielding two gastric “hemispheres” per animal. With the help of a stereomicroscope, the ventral half stomach was placed in PBS in a dissecting dish, the lumenal surface facing up. The locations of the electrodes were marked using pins to show each end of the electrode patch clearly. For documentation purposes, photographs of the stomach capturing the arrangement were taken. The x and y locations of the midpoint of each electrode in the whole mount were then measured from image print-out obtained at consistent magnification. We also measured the overall dimensions of the stomach itself so that electrode location could be expressed as normalized values relative to the antral end of the stomach contour (x) and the greater curvature of the stomach (y). The orientation of the electrode (a line through the two poles) was also measured relative to a line drawn from the top of the limiting ridge (near the LES) to the bottom point near the greater curvature where the limiting ridge reverses direction.

Motility Data Analysis

The quantification of duodenal motility was based on analysis of the strain-gauge output voltage for the 15-min window before the initiation of stimulation (baseline) and each of three different time windows after the onset of stimulation: specifically, the 5-min period with intermittent stimulation on; the 10-min period immediately following stimulation; and the entire 15-min period consisting of both the 5-min intermittent stimulation period plus the 10-min period immediately following the end of the intermittent stimulation period. The raw strain-gauge output signal was filtered (zero-phase FIR filter; fc = 0.05 Hz), rectified, and then quantified using a custom Matlab script. The three quantitative outputs were defined as follows: 1) amplitude ratio (AR) was defined as the ratio of the average rectified amplitude of strain-gauge voltage in the poststimulation period to the average amplitude in the prestimulation period; 2) frequency ratio (FR) was defined as the ratio of the number of voltage events exceeding 10% of the maximum voltage (to exclude noise events) per unit time in the poststimulation period to the number of voltage events exceeding 10% of the maximum voltage per unit time in the prestimulation period; and 3) motility index (MI) defined as the ratio of the area under the voltage curve (AUC) per unit time in the poststimulation to the area under voltage curve per unit time in the prestimulation period.

If, for example, the motility index or MI, measured as the ratio of AUC of poststimulation divided by AUC of prestimulation, was more than 1.0, it was inferred that stimulation increased duodenal motility or stimulation had an excitatory duodenal response. On the other hand, it was presumed that stimulation that decreased duodenal motility or stimulation had an inhibitory duodenal response if the MI was less than 1.0. If MI was equal to 1.0, stimulation had no effect on the duodenal response.

Functional Maps Generation

Based on data from 141 randomly placed electrodes, we were able to map the efficacy of stimulation as a function of location in these acute experiments using the various motility assessment parameters (MI, FR, and AR). Data were visualized in two ways: 1) all stomach stimulation sites were indicated using black or white circles according to the specific duodenal motility response; and 2) the map background was rendered using a blue-yellow-red look-up table based on kernel smoothing of the discrete duodenal motility responses. The data visualization was prepared using R-language for statistical computing and spatstat library (14).

Statistical Analysis Duodenal Responses of Five Distinct Stomach Regions

For post hoc statistical analysis of the effect of stimulating electrode location on the stomach wall on duodendal motor responses (analyses performed once the analyses of random electrode placements had been performed—see also discussion), we also divided the data into five distinct groups based on the location of electrode on stomach site. If both electrode plates lay within a single region of the stomach (antrum or corpus or forestomach), then the location was specified as antrum, corpus, or forestomach, respectively. If the electrode poles spanned across one region to another, the location was named as antrum/corpus or corpus/forestomach, as appropriate. Electrode locations were confirmed post mortem using the stomach photographs. The readout data were logarithmically transformed for normalization and variance stabilization and incorporated into linear models for one-way ANOVA. The mean responses for each of the stomach regions were compared using the Tukey method and graphed for visualization. The analysis was performed using R-language for statistical computing.

Spatial Distribution Correlations

To demonstrate spatial correlations between distribution of observed functional responses and distribution of afferent vagal projections in the stomach smooth muscle wall, we used Pearson’s correlation coefficient (PCC), accompanied by secondary rank-based correlation measures (Kendall τ and Spearman correlation coefficient, SCC) (15–18). In the context of spatial measurement, PCC can be viewed as a covariance computed between two image matrices Q and W, measured using identically ordered pixels.

$$\mathrm{P}\mathrm{C}\mathrm{C}=\frac{{{\displaystyle \sum }}_{i=1}^n\left(q_i-\overline{q}\right)\left(w_i-\overline{w}\right)}{\sqrt{{{\displaystyle \sum }}_{i=1}^n{\left(q_i-\overline{q}\right)}^2}\sqrt{{{\displaystyle \sum }}_{i=1}^n{\left(w_i-\overline{w}\right)}^2}}$$

where n is the total number of pixels in the region of interest, q_i is the i-th pixel intensity, and q¯ denotes the average pixel intensity in image Q. Similarly, w_i and w¯ are the corresponding intensities for image W.

PCC expresses the extent to which the variation of values in one map can be explained by the corresponding variation in the other map (15). However, PCC captures only the linear relationship between values in Q and W; therefore, we used additional measures for confirmation. The Spearman correlation is computed identically to PCC following the replacement of the original values q and w in the maps by their ranks. Finally, the Kendall correlation τ is computed as the difference between the number of concordant pairs and discordant pairs of pixels divided by the total number of pairs (18):

$$\tau =\frac{1}{n(n-1)}{\displaystyle \sum }_{i\ne j}sgn\left(q_i-q_j\right)sgn\left(w_i-w_j\right),$$

where a pair of observations (q_i, w_i) and (q_j, w_j), for i ≠ j, are considered concordant if $sgn\left(q_i-q_j\right)sgn\left(w_i-w_j\right)>0$ and discordant if otherwise. The correlation analysis of the smoothed density maps was performed using Fiji’s colocalization analysis plugin Coloc 2 (19).

Data associated with this study (20) were collected as part of the Stimulating Peripheral Activity to Relieve Conditions (SPARC) project and are available through the SPARC Data Portal (RRID:SCR_017041) under a CC-BY 4.0 license. A detailed experimental protocol is available through Protocols.io https://doi.org/10.17504/protocols.io.2irgcd6.

RESULTS

Spontaneous Proximal Duodenal Motility Patterns in Fasted Acutely Anesthetized Rats

To establish baselines, we recorded basal proximal duodenal motility in fasted acutely anesthetized rats (Fig. 2). Our recording results are similar and consistent with other observations (21, 22), which have reported spontaneous duodenal motility patterns in fasted rats consisted of grouped motor activity and intergroup motor activity. In an expanded time scale, the grouped activity was characterized by strong bursts of propagating contractions, lasting 2–4 min, involving an extended portion of the duodenum (Fig. 2C). At the same time, intergroup activity was characterized by weak, nonpropagating, irregular contractions, lasting 4–8 min, involving only a local area of the duodenum (Fig. 2C). The spontaneous duodenal activity became relatively stable approximately 1 h after surgery and lasted another 1–2.5 h when animals were maintained at stable body temperature, respiratory rate, and heart rate with enough fluid compensation by infusion (see materials and methods). Only rats with stable baselines were included in our data analysis, and rats with unstable and decaying motility patterns during the experimental periods were excluded.

Figure 2.

Representative baseline grouped and intergroup duodenal motility patterns recorded by a strain-gauge transducer in a fasted anesthetized rat. A shows originally recorded spontaneous duodenal motor activity signal (1 h) as recorded by a strain-gauge transducer attached to the proximal duodenum. B shows the rectified and filtered curve from the above originally recorded duodenal motor activity signal before quantification analysis. C shows grouped (arrow) and intergroup (arrowhead) duodenal motor activity patterns (from the beginning of A, area boxed with dashed line) on an expanded time scale.

Representative Examples of Duodenal Motility Responses Evoked by Forestomach, Corpus, and Antrum Stimulation

To evaluate the duodenal motility responses to regional GES, we compared duodenal motility before and after stimulation using the ratio indices of MI, FR, and AR. Stimulating forestomach, corpus, and antrum evoked dramatically different duodenal motility responses, as illustrated in Fig. 3, left. Each of the stimulation location was precisely determined by post mortem measurements (Fig. 3, right). In this case, stimulation of the forestomach (precise location of the case shown in top right) increased the motility index (MI = 1.18) and frequency ratio (FR = 1.28), but not amplitude ratio (AR = 0.99; Fig. 3A). Meanwhile, stimulation of the corpus (precise location of the example shown in middle right panel) decreased the motility index ratio (MI = 0.59), the frequency ratio (FR = 0.47), and the amplitude ratio (AF = 0.92; Fig. 3B). Stimulation of the antrum (precise location of the sample shown in lower right panel) increased motility index (MI = 1.36), frequency ratio (FR = 1.40), and amplitude ratio (AR = 1.27; Fig. 3C).

Figure 3.

Representative examples of different duodenal motility responses elicited by stimulating forestomach, corpus, and antrum, respectively. Each panel consists of a strain-gauge recording (top left), an integrated area under the curve measurement (lower left), and a mucosal-side photograph of the ventral stomach responsible for the illustrated recording in the respective panel. In each photograph, the pylorus/duodenum appears in the upper right corner; the (relatively translucent) forestomach occupies the left half of the image. The stimulation period is shown, and the periods of data analyzed as “15-min prestimulation” and “poststimulation = 5-min stimulation plus 10-min after-stimulation” are marked. A shows that an excitatory duodenal motility response was evoked by stimulation of the forestomach. The image on the right shows the exact location of the stimulation electrode on forestomach postmortem. B shows that an inhibitory duodenal motility response was evoked by stimulation of the corpus. The image on the right shows the exact location of the stimulation electrode on the corpus postmortem. C shows that an excitatory duodenal response was evoked by stimulation of the antrum. The image on the right shows the exact location of the stimulation electrode on the antrum postmortem. AUC, area under the voltage curve.

Functional Maps Generated by the Smoothed-Spatially Averaged Duodenal Motility Responses Corresponding to Each Stomach Stimulation Sites

Figure 4 shows the experimental results graphically. The maps incorporate two visualizations. The individual stimulation locations are shown with as either solid black dots (excitatory motility response: MI, FR, or AR > 1.0) or open dots (inhibitory motility responses: MI, FR, or AR < 1.0). Also, the color surface of each map represents the kernel-smoothed spatially averaged value of the motility responses. The outline corresponds to the average shape of whole stomach mounts prepared according to the perfusion process described in MATERIALS AND METHODS. Rows A, B, and C in Fig. 4 correspond to data from the 5-min during intermittent stimulation, the 10-min poststimulation, and the 15-min combination, respectively. From the MI-functional maps (left column of maps) and FR-functional maps (center column of maps), it can be noted that the maximal excitatory duodenal motility response was evoked by stimulation in the forestomach area (red color, and the majority of data points colored black). In contrast, an inhibitory duodenal motility response occurred with stimulation in the corpus (blue color, and the majority of data points colored white/open). A moderate excitatory duodenal motility response occurred with stimulation in the antrum (orange color, and a mixture of black and white data points). The same gross patterns are seen independent of the analysis period (5 min during intermittent stimulation, 10 min following stimulation, or 15-min combination, though with subtle differences). The AR functional maps (right column of maps) suggest a weaker dependence on stimulation location than for MI and FR responses.

Figure 4.

Distribution maps of the smoothed-spatially averaged duodenal responses to different stomach stimulation sites. The three rows A, B, and C correspond respectively to the three time periods of data analysis (5-min intermittent stimulation period; 10-min poststimulation period; 15-min combined stimulation and poststimulation period). The three columns show maps of the three quantified responses: motility index (MI), frequency ratio (FR), and amplitude ratio (AR). Ratios are relative to the 15-min prestimulation period (normalized for time). Excitatory duodenal-response bins (solid black dots, >1.0) and inhibitory duodenal-response bins (open dots, <1.0) are shown. Background colors were created by kernel smoothing of the discrete (during- and/or poststimulation vs. prestimulation) data points. The stomach outline corresponds to an average whole mount shape postprocessing. The maps include data from 141 separate stimulation sites from 81 male rats.

GES of Five Conventional Stomach Regions Produced Different Average Duodenal Motility Responses

For additional post hoc statistical comparisons, we sorted GES electrode placements on the stomach into the five different standard regions (antrum, corpus, forestomach, plus antrum/corpus, and corpus/forestomach to correspond to transition zones between the three main regions; see the methods for detailed definitions) and compared the duodenal motility responses produced by stimulating in the five regions. Figure 5 shows dot and box plots of the data, and Fig. 6 provides a statistical comparison of the mean responses. Further numerical data from the statistical analysis are included in Supplemental Tables S1–S3 (see https://doi.org/10.6084/m9.figshare.13342052). Data are arranged as in Fig. 4. Figures 5 and 6 show that, despite the variability in individual results shown in Fig. 5, there are some clear statistically significant differences between various regions of the stomach. In almost all cases, there is a statistically significant difference between forestomach stimulation and corpus stimulation (and corpus/forestomach stimulation). Forestomach stimulation, as assessed by all three responses (MI, FR, AR), is consistently excitatory, whereas corpus and corpus/forestomach stimulation is consistently inhibitory as measured by MI and FR (but not AR).

Figure 5.

Dot and box plots of duodenal response as measured by motility index, frequency ratio, and amplitude ratio to GES as a function of stomach region. As in Fig. 4, the three rows A, B, and C correspond respectively to the three time periods of data analysis (5-min intermittent stimulation period; 10-min poststimulation period; 15-min combined stimulation and poststimulation period). The three columns show data for the three quantified responses: motility index (MI), frequency ratio (FR), and amplitude ratio (AR). Ratios are relative to the 15-min prestimulation period (normalized for time). In this case we have assigned each data point to one of five stomach regions, namely, antrum (A), corpus (C), forestomach (F), and two transition regions antrum/corpus (A/C) and corpus/forestomach (C/F). The data within each graph show the distribution of responses in the form of a dot (all data) and box (mean and quartiles) plot as a function of stomach region. The dot and box plots include data from 141 separate stimulation sites from 81 male rats. GES, gastric electrical stimulation.

Figure 6.

Means comparison of duodenal response using ANOVA analysis as measured by motility index, frequency ratio, and amplitude ratio to GES as a function of stomach region. As in Fig. 4, the three rows A, B, and C correspond respectively to the three time periods of data analysis (5-min intermittent stimulation period; 10-min poststimulation period; 15-min combined stimulation and poststimulation period). The three columns show data for the three quantified responses: motility index (MI), frequency ratio (FR), and amplitude ratio (AR). Ratios are relative to the 15-min prestimulation period (normalized for time). Each data point is assigned to one of five stomach regions, namely, antrum (A), corpus (C), forestomach (F), and two transition regions antrum/corpus (A/C) and corpus/forestomach (C/F). The data within each graph show a comparison of mean responses as a function of stomach region. The dark gray diamonds show the estimated marginal means of the linear model. The light gray bars indicate the 95-percentile confidence intervals, and the arrows provide a visual means to perform comparisons. Overlapping arrows suggest that the difference is not statistically significant in terms of Tukey adjustment and α = 0.05. The charts include data from 141 separate stimulation sites from 81 male rats. GES, gastric electrical stimulation.

DISCUSSION

To our knowledge, this study is the first to evaluate the duodenal motility responses evoked by stomach stimulation with random placement of stimulation electrodes across the stomach surface from the antrum to corpus to forestomach, with stimulation parameters optimized for c-fiber activation in fasted anesthetized rats. By mapping the stimulation sites (n = 141) weighted by the magnitudes of their corresponding elicited duodenal responses, we produced topographic maps of stimulation-site efficacy. We also assessed (see Fig. 7 and Apparent Correlations between the Anatomical Maps and the Functional Maps) the nature of the GES activation mechanisms by correlating the generated topographic functional maps with previously determined topographic anatomical maps of the different vagal afferent terminals to the smooth muscle layers in stomach regions.

Figure 7.

Structural/functional correlations: 2D topographic maps of afferent vagal projections in the stomach smooth muscle wall [extracted from Wang and Powley, 2000 (12); Powley et al., 2016 (11); and publically available SPARC project datasets (29, 30)] vs. the functional map for MI from the present experiments. A shows the distribution of intraganglionic laminar endings (IGLEs) in the stomach wall, with the highest concentrations in the corpus and near the lesser curvature. B shows the distribution of intramuscular arrays (IMAs) in the longitudinal muscle of the stomach wall, with the highest densities of the terminals located along the greater curvature of the forestomach and in the antrum. C shows that the distribution of IMAs in the circular muscle of the stomach wall is concentrated in the more caudal forestomach near the lesser curvature and in the antrum. D shows the MI functional map from Fig. 4, row C (15-min analysis period). E–G quantify the colocalizations between each of the relevant neurite density maps, and the functional motility index map (green intensity represents the MI map, magenta the neurite map, and the level of white and dark represents the degree of overlap). The strong excitatory duodenal response elicited by GES in the forestomach coincides with the area of a high density of innervation of longitudinal IMAs, while the inhibitory duodenal response elicited in the corpus coincides with the lowest density of IMA innervation and the areas with the high density of IGLEs. MI, motility index.

The most notable findings obtained in these experiments are as follows. The maximal excitatory duodenal motility responses were evoked by stimulation on forestomach sites [which coincidently have the highest density of innervation of longitudinal muscle intramuscular arrays (IMAs) plus the moderate density of innervation of circular muscle IMAs; see Fig. 7]. In contrast, only moderate excitatory duodenal motor responses occurred with stimulation on antrum sites, which have concentrations—albeit less dense than those of the forestomach—of both longitudinal and circular muscle IMAs. At the same time, inhibitory duodenal motility responses occurred with stimulation on corpus sites, which have the highest density of innervation of intraganglionic laminar endings (IGLEs) and are conspicuously limited in their IMA innervation. Lastly, weak excitatory and/or inhibitory duodenal motility responses were elicited by stimulation of antrum/corpus and corpus/forestomach junctional regions, which have sparse fields of innervation of both IGLEs and IMAs.

Our findings suggest that GES-evoked excitatory duodenal motility responses might be produced by activating longitudinal muscle IMAs and/or circular muscle IMAs densely distributed in forestomach and antrum. The GES-evoked inhibitory duodenal motor responses might be caused by multiple factors that tend to cancel one another out, such as dense IGLE innervation, sparse longitudinal and circular IMAs innervation, enteroendocrine cell release of hormones in mucosa, and gastric acid secretion. GES might produce its variable duodenal effects by stimulating diverse afferent fibers distributed unevenly across stomach sites.

Underlying mechanisms of variable GES effects mainly depend on the anatomical structures activated and the neuro-circuits involved. Stimulating different stomach afferent fibers can produce different duodenal motor activity responses apparently by activating different responses. Most stomach muscle afferents are from two putative mechanoreceptors that the vagus nerve supplies to the gastric smooth muscle: intraganglionic laminar endings (IGLEs), putative tension receptors distributed across the stomach (23–25), and intramuscular arrays (IMAs), putative stretch or length receptors most densely distributed in proximal and antral stomach (26–28).

Apparent Correlations between the Anatomical Maps and the Functional Maps

To evaluate the possible correlation between available anatomical maps of the afferent vagal projections of stomach muscle wall with our newly observed functional maps of stomach stimulation-sites-dependent duodenal motor responses, we compare the maps side-by-side in Fig. 7. Three anatomical maps were generated by mapping the arbor origin density (defined as the first significant branch point of the arbor) of vagal projections to the smooth muscle wall of the stomach (generated by tracer injection into the nodose ganglion) (11, 12, 29, 30). The map in Fig. 7A (map A) shows high concentrations of IGLEs in the corpus and, to a lesser extent, in the forestomach near the lesser curvature. The map (29, 30) in Fig. 7B (map B) shows a high density of longitudinal muscle IMAs in the antrum and the forestomach along the greater curvature. In contrast, the map (29, 30) in Fig. 7C (map C) shows high concentrations of circular muscle IMAs in the antrum and the middle forestomach. The map in Fig. 7D (map D) shows the motility index (MI)-functional map from Fig. 4, row C, illustrating the areas of excitatory and inhibitory duodenal motor response to stimulation at individual gastric sites.

Comparing the anatomical and functional maps, we noted the following: 1) The maximal excitatory duodenal motility responses are readily evoked by stimulation of the forestomach, which coincidently has the highest density of innervation of longitudinal muscle IMAs (map D vs. map B) plus a moderate density of innervation of circular muscle IMAs (map D vs. map C). 2) A moderate level of excitatory duodenal motility occurs with stimulation of antrum, which has a moderate density of innervation of both longitudinal and circular IMAs (map D vs. map B and map C). 3) Inhibitory duodenal motility response occurs with stimulation of corpus, which has the highest density of innervation of IGLEs (map D vs. map A). 4) Weak excitatory and/or inhibitory duodenal motility responses are evoked by stimulation of antrum/corpus and corpus/forestomach regions, which have less innervation of IGLEs and/or IMAs (map D vs. maps A, B, and C). Our observations were confirmed by the quantification of the colocalization of the functional and anatomical maps (see the bottom of Fig. 7). The PCC between the longitudinal muscle IMA density map and the MI map is 0.52 (SCC is 0.49, and τ is 0.32). In contrast, the PCC between the IGLE density map and the MI map is −0.59 (SCC is −0.58, and τ is −0.39). There is also a moderate negative correlation between the circular muscle IMA density map and the MI map (PCC = −0.14, SCC = −0.20, τ = −0.13).

Such a structure-function hypothesis may certainly be oversimplified, considering the complexities of stomach and duodenal structures and their innervation (31). Stimulation on the stomach surface may activate multiple gastric structures and this makes GES mechanisms very complicated. Local tissues potentially affected by GES include vagal afferent fibers and efferent fibers (32, 33), sympathetic afferent and efferent fibers (34–36), gastric enteric nervous system (37, 38), gastric muscle and ICC cells (39), as well as effects on gastric endocrine and exocrine cells (40, 41). The narrow stimulation pulse width (0.2 ms) used in this study predominantly activates nerve fibers instead of activating muscle and ICC cells while the minimal stimulation amplitude (0.3 mA) may mainly activate vagal fibers instead of sympathetic fibers (42). Based on our observations and data evaluation, we provisionally concluded that the key underlying mechanism of variable duodenal motility responses produced by stimulation on different stomach sites might mainly activate different vagal afferent fibers (IMAs and IGLEs) innervating gastric sites with possible involvement of multiple gastric and enteric nervous system mechanisms.

Perspectives and Significance

Stimulation of different stomach regions produces dramatic site-specific effects on duodenal motility, apparently mediated by reflexes. Significantly, the distribution of stimulation locations with the strongest excitatory effect on duodenal motor responses correlates with highest innervation density of IMAs. Thus, functional maps can help to identify optimal GES locations for therapeutic potential to treat GI disorders.

GRANTS

This study was supported by National Institutes of Health Grants SPARC OT2:, OD023847, and DK27627 (principal investigator was T. L. Powley).

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the authors.

AUTHOR CONTRIBUTIONS

T.L.P. conceived and designed research; Z.T.T., M.W., R.J.P., X.Z., L.C., E.A.B., and J.M. performed experiments; Z.T.T., T.L.P., M.W., R.J.P., X.Z., D.M.J., L.C., and B.R. analyzed data; Z.T.T., T.L.P., M.W., R.J.P., X.Z., D.M.J., and B.R. interpreted results of experiments; Z.T.T., T.L.P., D.M.J., and B.R. prepared figures; Z.T.T., T.L.P., and D.M.J. drafted manuscript; Z.T.T., T.L.P., M.W., D.M.J., B.R., and J.M. edited and revised manuscript; Z.T.T., T.L.P., M.W., R.J.P., X.Z., D.M.J., L.C., B.R., E.A.B., and J.M. approved final version of manuscript.