Site-specific pathophysiology in a neonatal mouse model of gastroparesis

Abstract

Background:

Early-life events impact maturation of the gut microbiome, enteric nervous system, and gastrointestinal motility. We examined three regions of gastric tissue to determine how maternal separation and gut microbes influence the structure and motor function of specific regions of the neonatal mouse stomach.

Methods:

Germ-free and conventionally housed C57BL/6J mouse pups underwent timed maternal separation (TmSep) or nursed uninterrupted (controls) until 14 days of life. We assessed gastric emptying by quantifying the progression of gavaged fluorescein isothiocyanate (FITC)-dextran. With isolated rings of forestomach, corpus, and antrum, we measured tone and contractility by force-transduction, gastric wall thickness by light microscopy, and myenteric plexus neurochemistry by whole mount immunostaining.

Key Results:

Regional gastric sampling revealed site-specific differences in contractile patterns and myenteric plexus structure. In neonatal mice, TmSep prolonged gastric emptying. In the forestomach, TmSep increased contractile responses to carbachol, decreased muscularis externa and mucosa thickness, and increased the relative proportion of myenteric plexus nNOS+ neurons. Germ-free conditions did not appreciably alter the structure or function of the neonatal mouse stomach and did not impact the changes caused by TmSep.

Conclusions and Inferences:

A regional sampling approach facilitates site-specific investigations of murine gastric motor physiology and histology to identify site-specific alterations that may impact gastrointestinal function. Delayed gastric emptying in TmSep is associated with a thinner muscle wall, exaggerated cholinergic contractile responses, and increased proportions of inhibitory myenteric plexus nNOS+ neurons in the forestomach. Gut microbes do not profoundly affect the development the neonatal mouse stomach or the gastric pathophysiology that results from TmSep.

Graphical Abstract

1. INTRODUCTION

Gastroparesis has an estimated incidence of 0.4–3% among adults. Incidence among children is unknown due to lack of epidemiological data, but pediatric hospitalizations due to gastroparesis have increased significantly over the past decade.[1] Gastroparesis can have many underlying causes, and pathophysiology remains poorly defined. Recent recommendations list specific areas in need of further research to better understand pediatric gastroparesis. These include mechanistic studies of contractility in specific regions of the stomach.[2]

To help address this knowledge gap, we recently characterized a new mouse model of neonatal gastroparesis. Timed maternal separation (TmSep) is a model of early postnatal malnutrition in which neonatal mice become stunted and underweight with metabolomic, metagenomic, and immune system alterations that mimic features of child malnutrition.[3–5] This model also recapitulates malnutrition-induced gastrointestinal dysmotility, which is observed in small-for-gestational-age neonates,[6, 7] children with severe acute malnutrition,[8–10] and young adults with anorexia nervosa.[11–15] TmSep mouse pups have grossly dilated stomachs, thin gastric walls, and delayed emptying of gavaged fluorescein isothiocyanate (FITC)-dextran into the small bowel.[16] Delayed passage of FITC-dextran recapitulates the gold standard diagnostic finding in human gastroparesis: delayed passage of a radiolabeled standardized meal by gastric emptying scintigraphy.[1]

For the present study we sought to determine whether TmSep and microbiota alterations affect the structure and function of different regions of the neonatal mouse stomach. We harvested rings of murine forestomach, corpus and antrum for site-specific force-transduction studies. After validating our technique with adult mice, we performed regional gastric assessments of ex vivo contractility, histologic structures, and myenteric plexus cell composition in neonatal control and TmSep mice that were conventionally housed or maintained in strict germ-free conditions.

2. MATERIALS AND METHODS

2.1. Regional isolation of gastric rings

In an adaptation of a previously published technique,[17] stomach from C57BL/6J (Charles River) or Plp1-GFP glial reporter (W. Macklin, University of Colorado;[18] Supporting Figure S1) adult mice or C57BL/6J neonatal pups was dissected in Krebs buffer (in mmol/L: 120.9 NaCl, 5.9 KCl, 2.5 CaCl2, 14.4 NaHCO₃, 1.2 NaH₂PO₄, 1.2 MgCl₂, 11.5 glucose) warmed to 37.5° C and aerated with 95% O₂/5% CO₂ gas to preserve tissue viability. Pins through the esophagus and pylorus immobilized the stomach for further dissection. Additional pins were placed along the junction between the forestomach and the glandular stomach as landmarks. Full thickness rings of forestomach, corpus, and antrum were harvested in an orientation that parallels the gastric circular muscle (Figure 1A). All rings were 3 mm wide, but their circumference differed by anatomic site, age, and intervention (Supporting Table 1). Anatomic locations were confirmed by routine histology (Supporting Figures S2–S3).

FIGURE 1.

Region-specific gastric sampling enables the detection of site-specific differences in contractility and myenteric plexus organization in adult mice. A, Rings are obtained from three regions of mouse stomach based on anatomic landmarks and confirmed by routine histology (Supporting Figures S2–S3). B, Representative force-transduction tracings illustrating equilibration at baseline followed by responses to 0.1 μmol/L then 1 μmol/L then 10 μmol/L doses of carbachol. C, Contractile activity, calculated as the area under the curve relative to minimum, is increased in adult mouse forestomach relative to corpus at baseline and in response to carbachol. D, Average contraction amplitude is increased in adult mouse antrum relative to corpus in response to carbachol. E, Representative images of the myenteric plexus of forestomach, corpus, and antrum from adult Plp1-GFP mice labeled with antibody to Hu (blue) and nNOS (red). Bars represent mean + SD; N = 5 adult female mice per group; * adjusted P < 0.05.

2.2. Timed maternal separation protocol and gastric emptying assay

Male and female C57BL/6J pups (Charles River) underwent timed maternal separation (TmSep) for 4 hours on day-of-life 5, 8 hours on day-of-life 6, and 12 hours from days-of-life 7 to 13 as reported.[3] Separated pups were kept warm as a single large group inside nesting without access to food or water. Separate litters of control pups nursed ad libitum. Mice were housed in conventional (specific pathogen free) or germ-free isolators, each with irradiated rodent unpurified diet and drinking water. Gastric emptying was quantified on day-of-life 14 by gavaging non-fasted pups with 50 μL of 10 mg/mL FITC conjugated to 70 kDa dextran (Sigma).[19] Stomach and small bowel was harvested 15 minutes after FITC-dextran gavage; because neonatal mice are resistant to hypoxia,[20] decapitation was performed without carbon dioxide. Gastric emptying was calculated as a fraction of total gastrointestinal FITC.[16] Data from male and female pups are depicted as squares and circles, respectively, and were combined for analysis based on our previous report that sex differences in upper gastrointestinal motility are not yet present at 2 weeks of life.[19] The Baylor College of Medicine Institutional Animal Care and Use Committee approved all aspects of this study.

2.3. Ex vivo gastric smooth muscle contractility

Gastric contractile activity was quantified by adapting a previous protocol.[16] Gastric rings were mounted individually in organ baths filled with 25 mL Krebs solution and infused continuously with warmed 95% O₂/5% CO₂ gas. Isometric force was measured by an external force-displacement transducer connected to a PowerLab recorder (ADInstruments). After 30 minutes of equilibration to 0.5 g tension, baseline contractile activity was recorded for 5 minutes. Motor responses to increasing doses of cholinergic stimulation of 0.1, 1, and 10 μmol/L carbachol (Sigma) were recorded for 5 minutes after each dose. Contractile activity was calculated as the area under the curve relative to the minimum force over 5 minutes, as previously described.[16]

2.4. Light microscopy and whole mount immunostaining

Gastric rings were cut open into rectangular strips that were 3 mm wide and twice the rings’ circumference in length (Supporting Table 1), formalin-fixed, paraffin-embedded, and stained with hematoxylin and eosin. Muscularis externa and mucosa thickness was measured with an Eclipse 90i Ni-E microscope (Nikon Instruments, Melville, NY, USA) and Image J (NIH) as previously described.[16] From six sections per region per mouse, 30–40 measurements of well-oriented tissue were recorded. Alternatively, whole mount immunostaining was performed by gently stretching and pinning down the rectangular strips, fixing overnight in 4% formaldehyde at 4° C, and preserving in PBS with 0.3% sodium azide. Microdissection and antibody staining of neonatal mouse stomach for Hu, neuronal nitric oxide synthase (nNOS) and S100β (Supporting Table 2) was performed as previously described.[21, 22]

To visualize immunohistochemistry and GFP transgenes in whole mount preparations of adult mouse stomach, samples were fixed overnight with 4% paraformaldehyde and further dissected to remove the mucosa and submucosa. Tissues were blocked and permeabilized in solution containing 10% donkey serum, 10% BSA and 1% triton in PBS for 1 hour at room temperature on a rocking platform. Primary antibodies were diluted in the same buffer and incubated overnight at 4°C before washing tissues in PBS. Corresponding secondary antibodies were applied for 3 hours and nuclei were stained with DAPI before washing tissues in PBS and mounting onto glass slides with AquaPolymount (Polysciences). A Nikon AXR laser scanning confocal microscope (Nikon) was utilized at The Microscopy Core of the Program in Membrane Biology at Massachusetts General Hospital.

2.5. Statistics and reproducibility

All data were tested for normality using the Shapiro-Wilk test. Force-transduction data were normally distributed and were evaluated by two-way ANOVA; when the interaction between carbachol dose and anatomic site or treatment group was significant (P < 0.05), between-group differences were determined with Tukey’s multiple comparisons tests. Normally distributed independent continuous variables were evaluated by one-way ANOVA; when the global test was significant (P < 0.05), between-group differences were determined with Sidak’s multiple comparisons test. For non-normally distributed independent continuous variables, a Kruskal-Wallis test with post-hoc Dunn’s multiple comparisons tests were used. Only adjusted P values are reported. Analyses were performed using Prism 9.4.1 (GraphPad Software).

3. RESULTS

3.1. Site specificity of gastric contractility and myenteric plexus in adult mice

We first sought to determine whether baseline tone or contractile responses to cholinergic stimulation differ by region of the adult mouse stomach. Full-thickness rings of forestomach, corpus, and antrum (Figure 1A) produced distinct motor patterns (Figure 1B). Forestomach had the greatest contractile activity at baseline and responded robustly to low doses of carbachol (Figure 1C). Antral rings responded with a dramatic increase in the amplitude of the waveforms (Figure 1D), although the net force generated from baseline was not increased (Supporting Figure S4). In contrast, corpus exhibited minimal contractility at baseline and responded minimally to carbachol.

A qualitative examination of the myenteric plexus using whole mount immunostaining confirmed that the enteric nervous system can be assessed in narrow strips of gastric tissue and revealed decreasing density of ganglia from antrum to forestomach (Figure 1E). Together, these studies in adult mice demonstrate the feasibility of a regional gastric sampling approach that permits site-specific assessments of ex vivo contractile activity and myenteric plexus neurochemistry.

3.2. Region specific changes induced by maternal separation

Next, we applied these techniques to a neonatal mouse model of gastroparesis to gain insight into mechanisms that underlie delayed gastric emptying. We performed TmSep in both germ-free and conventional (specific pathogen free) conditions to determine whether the dramatically altered gut microbiota[4] contributes to pathophysiology. Stomachs from conventionally housed TmSep pups were grossly dilated (Figure 2A), confirming our previous findings.[16] Germ-free TmSep pups had similarly dilated stomachs. In accord with these gross observations, gastric emptying was delayed by 25% in both conventional and germ-free TmSep mice relative to controls (Figure 2B). Thus, TmSep pups have markedly dilated stomachs and delayed gastric emptying irrespective of the presence of their microbiota.

FIGURE 2.

Associations between structure and function of the neonatal mouse forestomach, delayed gastric emptying, and the presence of gut microbiota. A-B, TmSep pups in both conventional and germ-free conditions exhibit grossly dilated stomachs and delayed gastric emptying. C-D, Contractile activity, calculated as the area under the curve relative to minimum, and average contraction amplitude are increased in forestomach rings from TmSep pups at baseline and in response to carbachol. E, Representative images show overlays of Hu (blue), nNOS (red), and S100β (green) in myenteric ganglia from neonatal mouse forestomach. Density of forestomach myenteric Hu+ neurons appears to decrease in TmSep mice but this may be confounded by tissue stretching. F-G, TmSep thins the forestomach muscularis externa and mucosa layers in both conventionally-housed and germ-free mice. H, The relative proportion of nNOS+ to total Hu+ neurons increases 1.8-fold in forestomach of conventionally housed TmSep mice relative to controls. Bars represent mean + SD; *** adjusted P < 0.001, ** adjusted P < 0.01, * adjusted P < 0.05; N = 5–7 mice per group.

Ex vivo contractile patterns of control neonatal mouse forestomach, corpus, and antrum were similar to those observed in adult mice; namely, low doses of carbachol enhanced contractile activity in the forestomach, increased the amplitude of contractions in the antrum, and had minimal effect in the corpus (Supporting Figure S5). Thus, we focused our remaining investigations on the neonatal mouse forestomach.

Compared to controls, conventionally housed TmSep pups exhibited increased contractile activity and amplitude at baseline and in response to carbachol. Similar trends were observed in germ-free pups (Figure 2C–D). Whole mount immunostaining of Hu, nNOS, and S100β in the neonatal mouse forestomach myenteric plexus (Figure 2E) appeared to reveal decreased cell density in the TmSep groups; however, cell density measurements may be confounded by organ size and stretching.[23] Indeed, the distention observed in TmSep mice correlated with thinner muscularis externa (Figure 2F) and mucosa (Figure 2G) layers. To avoid confounding due to organ size, we quantified the relative proportion of nNOS+ to total Hu+ neurons, and this revealed a 1.8-fold increase in the proportion of nNOS+ neurons in conventionally housed TmSep pups compared to controls (Figure 2H). All together, these data highlight the feasibility of a regional sampling approach for physiologic and histologic assessment of adult and neonatal mouse stomach. In our model of neonatal gastroparesis, gastric contractile activity, myenteric plexus cell neurochemistry, and gastric wall thickness were abnormal. Gut microbes do not contribute appreciably to any of these abnormalities.

4. DISCUSSION

Mechanistic studies of contractility in specific regions of the stomach are needed to better understand the pathophysiology underlying neonatal gastroparesis.[2] Here, we used a tissue sampling approach that facilitates analyses of motor function and histopathology in the mouse forestomach, corpus, and antrum. This technique enabled us to quantify structural and physiologic differences that are associated with delayed gastric emptying in a neonatal mouse model of gastroparesis. TmSep causes gastric distention that correlates with a thinner muscle wall, exaggerated contractile responses, and increased proportions of inhibitory nNOS+ neurons in the myenteric plexus of the neonatal mouse forestomach. These findings were present in conventionally housed and germ-free TmSep mice, indicating that the altered gut microbiota[4] does not play a causal role in gastric pathophysiology caused by TmSep. Germ-free conditions also had minimal effects on healthy control mice, suggesting that gut microbes are not critical for the development of gastric structure and function in the early neonatal period.

By harvesting rings from specific regions of the stomach, we observed increased forestomach contractile activity in TmSep mice that have delayed gastric emptying. This seemingly paradoxical result is consistent with previous reports of increased duodenal contractility at baseline and after cholinergic stimulation in malnourished adult mice that have delayed small bowel transit.[16] The mechanistic basis for this paradox remains unclear. Strong but poorly coordinated contractions could be causing gastroparesis. Exaggerated cholinergic responses could be compensating for other prokinetic signals that are depleted in TmSep mice, such as serotonin.[3] Given the immune dysregulation that occurs in maternal separation[5, 24] and malnutrition,[25] we cannot exclude the possibility that muscularis macrophages and their inflammatory mediators may deplete interstitial cells of Cajal.[26] Our finding of increased proportions of nNOS+ neurons raises the possibility that TmSep increases inhibitory vagal input. In the forestomach, vagal innervation is prominent and is a key regulator of gastric motility via the release of functionally opposing neurotransmitters acetylcholine and nitric oxide.[27, 28] Nitrergic enteric neurons, which exert inhibitory effects on smooth muscle throughout the gut, mediate adaptive relaxation of the stomach to a liquid or solid meal.[29] Why the proportion of forestomach myenteric plexus nNOS+ neurons is increased by TmSep, and whether this causes the delayed gastric emptying, remain under investigation.

Previous studies found that germ-free mice exhibit colonic dysmotility.[30] We did not find any evidence of gastric dysmotility in germ-free control pups. This could be due to the stomach’s lack of proximity to the densely populated microbial communities of the large bowel or to its sparsely populated resident microbiota. Furthermore, an absence of gut bacteria did not appreciably affect the abnormal ex vivo contractility, decreased gastric wall thickness, and altered myenteric plexus neurochemistry induced by TmSep. Together, these findings suggest that early-life interactions between gut microbes and the enteric nervous system may be more important for distal, rather than proximal, intestinal tract development and function.[31]

Rodent models highlight the broad and long-term consequences of intermittent food deprivation and stress on neurodevelopment.[32] In the central nervous system, TmSep impairs prefrontal cortex development and causes behavioral and cognitive changes that can persist into adulthood.[33–37] TmSep also affects the enteric nervous system as evidenced by increased visceral hypersensitivity to colorectal distention.[38–40] Similarly, gastric emptying is prolonged by stress in the absence of food deprivation in rats subjected to chronic restraint[41] and by malnutrition in the absence of maternal separation in adult mice maintained on a low-protein low-fat diet.[16] Although our study was not designed to determine the relative contributions of malnutrition and stress to impaired neurodevelopment and muscle contractility, these could be explored by using shorter periods of separation (e.g., 1 hour rather than 12 hours per day) that would avoid limiting milk intake to the point of growth faltering, or by using alternative models of neonatal malnutrition that manipulate the maternal diet[16] rather than restrict pups’ access to milk.

In conclusion, a regional gastric sampling approach facilitates functional, physiologic, and histologic assessments following an early-life challenge. This approach is adaptable to neonatal and adult mouse models with the potential to provide insights into mechanisms underlying numerous motor and sensory disorders of the stomach.