An exploratory in vitro co-culture of enteric neurons and smooth muscle cells demonstrates neuronal contribution to muscle layer formation
Rasul Khasanov, María Ángeles Tapia-Laliena, Steven Schulte, Valentin Pavlov, Michael Boettcher, Lucas M. Wessel, Karl-Herbert Schäfer
TL;DR
Researchers explored how enteric neurons and muscle cells can work together in a lab setting to create a bioengineered intestinal muscle layer, which could help treat Short Bowel Syndrome.
Contribution
This study demonstrates for the first time that co-culturing enteric neurons and smooth muscle cells in 3D scaffolds can form functional muscle layers with synaptic-like connections.
Findings
Enteric neurons and smooth muscle cells formed structural and synaptic-like connections in 3D co-cultures.
Smooth muscle organization and spontaneous contractile activity were observed in co-cultures.
The study provides a proof-of-concept for generating innervated muscle fibers in vitro.
Abstract
Short Bowel Syndrome (SBS) is characterized by insufficient functional intestinal tissue capable of nutrient transport and absorption. Tissue engineering offers a promising strategy to restore intestinal function by reconstructing a bioengineered muscle layer. In this exploratory study, we investigated the feasibility of co-culturing rat smooth muscle cells (SMCs) with enteric nervous system (ENS) cells in layered three-dimensional (3D) scaffolds. Three culture conditions were compared: SMC monocultures, paracrine signaling through a semipermeable membrane, and direct ENS–SMC co-culture. Although some effects in this reductionist model may reflect in vitro artifacts rather than true developmental processes, our results demonstrate that ENS and SMCs can form structural and potentially functional (synaptic-like) connections. Electron microscopy and immunofluorescence revealed native-like…
Genes, proteins, chemicals, diseases, species, mutations and cell lines named across the full text — each resolved to its canonical identifier and authoritative record.
Click any figure to enlarge with its caption.
Figure 1
Figure 2
Figure 3
Figure 4
Figure 5
Figure 6
Figure 7
Figure 8- —Medizinische Fakultät Heidelberg der Universität Heidelberg (9149)
Peer Reviews
No public reviews on file for this paper yet. If you reviewed it on a platform where reviews are public (OpenReview, ICLR, NeurIPS, ICML), you can paste yours below so the community can read it here.
Videos
No videos yet. Explain this paper in a talk, walkthrough, or lecture? Add one.
Taxonomy
TopicsGastrointestinal motility and disorders · Clinical Nutrition and Gastroenterology · Tissue Engineering and Regenerative Medicine
Introduction
Short Bowel Syndrome (SBS) is a devastating situation for the patient and an expensive problem in human health care. Due to the dramatic reduction in the length of the small intestine, the remaining absorptive surface is far too limited to ensure adequate nutrient uptake. This leads to intestinal failure (IF), requiring permanent dependence on daily parenteral nutrition for years or in some cases for the rest of their life. The latter is at least in parts responsible for several severe complications such us central line-associated sepsis, liver disease, bacterial overgrowth or relapsing translocation. Due to an optimized parenteral nutrition, combined with creation of interdisciplinary therapeutical teams, the long-term survival could be improved in the last decade, but the overall mortality is still up to 27%^1–8^. Despite the intense interdisciplinary therapeutic and surgical treatment the weaning from the parenteral nutrition (PN-weaning) rate does not exceed 50–70% in IF-patients^6,9^. Intestinal transplantation seemed to be a viable option, but is now considered as “ultima ratio” for IF-patients. Five years after intestinal transplantation more than 60% of transplants are rejected^10^. Survival rate of IF-patients five years after intestinal transplantation is only 56%^11^.
Tissue engineering has the potential to provide an alternative approach for the treatment of patients with short bowel syndrome^2,12^.
The enteric nervous system (ENS) is widely recognized as a critical regulator of intestinal development and function^13,14^Clinical and fundamental research have consistently demonstrated that neuronal input is indispensable for establishing normal gut physiology. Introducing neuronal innervation into bioengineered intestinal tissue is therefore considered a key step toward achieving functional constructs^15^.
A landmark study by Workman et al. demonstrated the successful incorporation of ENS structures within human intestinal organoids, highlighting the essential role of the ENS in intestinal development. Using human embryonic and induced pluripotent stem cells (iPSCs), they generated intestinal tissue containing a functional ENS by combining PSC-derived neural crest cells (NCCs) with developing human intestinal organoids (HIOs)^16^. This elegant tissue-engineering approach recapitulated normal intestinal development and underscored the importance of neuronal innervation in engineered gut models, but did not deliver an insight of the immediate crosstalk and interference between smooth muscle cells and ENS, as it will be needed in tissue engineering processes which will lead to a construction of artificial gut wall segments in SBS patients.
Recent reviews have reinforced the need for innovative co-culture systems that integrate the ENS with intestinal organoids^17^. These studies emphasize that the ENS not only interacts with the intestinal epithelium but also plays a crucial role in coordinating the muscular layers, immune responses, and the microbiome. However, despite considerable progress, discrepancies remain due to differences in experimental systems, cell sources, and methodological approaches. Standardized protocols and simplified models are therefore urgently needed to unravel the precise mechanisms of ENS–tissue interactions.
Many published organoid systems incorporate a broad diversity of cell types, which interact in complex and often confounding ways. In contrast, our study deliberately employed a reductionist approach: we focused exclusively on the muscle layer and ENS, limiting the culture to only two defined cell populations—smooth muscle cells and enteric neurons. By minimizing the influence of additional cell types and signaling factors, we aimed to create a highly controlled environment for an exploratory investigation into the direct interactions between neuronal and muscle populations.
The aim of this exploratory study was to assess the feasibility of establishing functional communication between enteric neurons and smooth muscle cells in a bioengineered model, providing hypothesis-generating evidence for the role of neuronal innervation in the formation of functional intestinal muscle layers.
Results
Optimization of 3D culture conditions
At the first stage we tested the optimal cell culture conditions. The cell culture media showed a large impact upon the co-cultures. The best growth was observed using DMEM-F12 medium with proliferation factors (GDNF, EGF) and serum.
In this cell culture medium, both cell types (neurons and SMCs) grew similarly in the mixed culture approach, as they did in their respective individual media. Here, cells were kept in DMEM-F12 supplemented with 2% B27+ (with vitamin A, Gibco), 1% albumin (Sigma), 0.25% 2-mercaptoethanol (50 mM, Invitrogen), 0.12% glutamine (200 mM, Sigma), bFGF (20 ng/ml; Tebue), GDNF (10 ng/ml, Tebue), gentamycin and metronidazole (100 µg/ml)], 10% FCS (fetal calf serum). The culture medium was changed each 48–72 h. The co-cultures were incubated for a time period between 2 weeks and 4 months. Afterwards, cells were fixed with 4% formaldehyde for 10 min. and stored in phosphate buffered saline (PBS).
At the second stage we assessed the optimal layout order, of SMCs and ENS cells, for the best 3D cultures. The layout of the cells in the 3D co-culture influenced significantly cell proliferation. Six different layouts of the cells in the scaffold were tested (Fig. 8). The best layout to achieve a kind of gut wall appearance was found to be a high concentration of ENS cells in the middle layer and distribution of the muscle cells adjacent to the layers above and below the ENS cells (sandwich) (Fig. 8e).
In this layout, SMCs and ENS cells survived together in the same single medium and showed much better proliferation in comparison with other layouts, so that all other following experiments were performed only with this approach.
Characterization of muscle fibers formation in 3D scaffolds
On day 4 of cell cultivation, we observed proliferation of the cells and forming processes between cells in the middle layer, where ENS-cells were seeded. On the 8th day of cultivation, the amount of the proliferating cells and number of processes between the cells in the middle layer of the gel increased. After 12th days of culture, we observed a structure consisting of many cells, while the border between the matrix and the cells culture was clearly visible.
On day 16, the cell culture was stopped and co-stained with smooth muscle actin (SMA) to label the muscle cells, and ß-Tubulin III for the neurons. The staining confirmed that the observed structure consisted mostly of muscle cells, but neurons were also present in the co-culture (see Supplementary Fig. S1 online).
To further characterize the cellular composition and delineate the specific enteric nervous system (ENS) cell types present in the culture, additional immunostainings were performed. Neuronal nitric oxide synthase (nNOS) was used to identify neurons expressing nNOS, a marker for nitric oxide, a principal inhibitory neurotransmitter that modulates the activity of other neurons as well as effector cells, including smooth muscle, epithelial, and enteroendocrine cells Choline acetyltransferase (ChAT) staining was employed to detect cholinergic enteric neurons, indicative of acetylcholine, the predominant excitatory neurotransmitter within the ENS. Glial fibrillary acidic protein (GFAP) immunolabeling was used to identify enteric glial cells^18,19^.
Immunostainings for nNOS and ChAT, together with PGP 9.5 (Fig. 1a), demonstrated that the co-cultures contained distinct neuronal subtypes, including nNOS-positive and ChAT-positive neurons, integrated within the culture network.
The quantitative data show that 62% of neurons are ChAT-positive, 33% are nNOS-positive, and 25% of neurons are double-positive for both ChAT and nNOS (Fig. 1b).
Fig. 1. Immunofluorescence staining of neurons cultured in co-culture with smooth muscle cells in a three-dimensional HyStem-C hydrogel and quantitative evaluation of neuronal subpopulations.
- Immunofluorescence microscopy showing choline acetyltransferase–positive neurons (ChAT, cholinergic neurons, magenta), neuronal nitric oxide synthase–positive neurons (nNOS, nitrergic neurons, cyan), and total neurons labeled with PGP9.5 (green). Cell nuclei are counterstained with DAPI (blue). Arrows indicate neurons positive for ChAT only (magenta), nNOS only (cyan), both markers (yellow), or neither marker (green). Magnification: 630×; scale bar: 50 μm.
- Graphical representation of the percentages of ChAT-positive neurons, nNOS-positive neurons, and double-positive neurons.
GFAP-positive enteric glial cells were also detected, confirming the presence of multiple ENS-related cell types within the established co-culture^19,20^ (Fig. 2).
Fig. 2. Immunofluorescence staining of enteric nervous system (ENS) cells co-cultured with smooth muscle cells in a three-dimensional HyStem-C hydrogel, showing (a) glial fibrillary acidic protein (GFAP)–positive glial fibers (magenta), (b) smooth muscle cells stained for smooth muscle actin (SMA, cyan), (c) neuronal fibers labeled with βIII-tubulin (Tuj1, green), and (d) merged image. Cell nuclei are counterstained with DAPI (blue). Non-specific staining observed in the SMA channel indicates the presence of additional cell types, likely fibroblasts. Scale bar 50 μm. (e) higher-magnification image of a GFAP-positive glial cell (magenta). Scale bar 100 μm.
Next, immunofluorescence staining of living co - cultures was performed with antibodies against smooth muscle actin CellLight™ Actin-RFP, BacMam 2.0 (Invitrogen). The peculiarity of this experiment was that the optical layers of the resulting three-dimensional cell culture could be better investigated. The morphology of the SMCs depends strongly on the culture mode and the presence or absence of ENS cells in the culture (see Supplementary Fig. S2 online) It was evident that, in the three-dimensional matrix layer without enteric nervous system (ENS) cells, the muscle cells remained round, and their configuration did not correspond to the shape of normal muscle cells. Additionally, we observed that actin expression in these cells was reduced. In contrast, in the three-dimensional matrix layer where enteric neurons were present, muscle cells exhibited an elongated shape typical of muscle cells, and actin expression in these cells was high.
Moreover, in the collagen-based scaffold (RAFT), muscle cells proliferated but had different orientations, which could lead to uncoordinated contractions that are not compatible with a functional small intestine wall.
Muscle fibers in 3D scaffolds show contractile function
Our results showed that the co-culture of muscle and ENS cells presented a prominent contractile activity. The cultures were examined daily under a light microscope. From day 20th in vitro onwards, contractile activity of muscle fibers in the native live cultures could be observed. The contraction arose spontaneously and lasted all over the observation period. We could observe contraction activity in thin muscle fibers, as well as in thick muscle fibers (Fig. 3, see Supplementary Video S1 and Video S2 online).
Fig. 3. Contracting muscle fibers in the live cultures of muscle cells together with ENS cells in HyStem-C Hydrogel (real time): (a) thin muscle fibers, (b) thick muscle fibers. Light microscopy.
The co-location of muscle cells with ENS cells in a three-dimensional culture significantly affected the growth pattern of the cells. The same cells in the same matrix developed a completely different behavior depending on the presence or absence of enteric nerve cells. However, when the arrangement of muscle cells and the ENS was optimal, muscle cells formed muscle bundles capable of contractile activity.
Muscle fibers and neurons do interact in the 3D culture scaffolds
Using immunofluorescence staining of fixed cells, we were able to identify the interaction of SMCs and neurons in 3D. Our 3D scaffolds cultures presented muscle fibers involving muscle cells and neurons, where neurons tended to form plexus-like clusters. These clusters together formed a single neuronal network (Fig. 4a).
Fig. 4. Smooth muscle and ENS cells in 3D scaffolds (14 days) co-cultured in 3D scaffold, showing neurons in green (ß-Tubulin III), muscle cells in red (smooth muscle actin (SMA)) and nuclei in blue (DRAQ5) (a) Confocal microscopy lower Magnification, (b) Confocal microscopy 3D structure in merged channels view from the side in higher magnification (Scale bars 50 μm) and (c) Confocal microscopy, separated channels view from above in higher magnification, (d) Electron microscopy of two muscle cells in close contact within the thickness of the three-dimensional matrix. The cells have large nuclei, an elongated shape, and actin microfilaments typical for muscle cells. Scale bar 2 μm, (e) Electron microscopy of plasma membrane of a muscle cell, with arrows indicating caveolae. Scale bar 200 nm.
In higher magnification we could picture the tight growth of the muscle cells and neurons (Fig. 4b), confirming that complex neuronal networks were growing on the layer of muscle cells (Fig. 4c).
Also, when only individual muscle cells and neurons were present, the shape patterns of one-to-one contact were very similar. This brings into discussion the possibility that closely located cells influence each other’s form (see Supplementary Fig. S3 online).
Finally, we analyzed the cultures by electron microscopy. Images showed that the muscle fibers grown in this experiment had a histological structure similar to that of smooth muscle fibers in the intestine: the muscle cells were tightly adjacent to each other, had an elongated shape, large nuclei, and actin microfilaments (Fig. 4d). In addition, caveolae—small (50–100 nm) flask-shaped invaginations of the plasma membrane could be detected, indicating active cell metabolism (Fig. 4e).
Characterization of type of interaction between SMCs and ENS cells in 3D scaffolds
At the third stage we studied the way of interaction between SMCs and ENS cells in 3D.
Comparison of cultivated SMCs alone with cultures exposed to the paracrine influence of ENS cells and cultures in direct contact with ENS cells, revealed different rates of cell proliferation (Fig. 8).
On day 4th we did not observe any changes in the groups with only muscle cells or where SMCs and ENS cells were separated (paracrine interaction). In contrast, in the group where both cell types were in direct contact, we found an intricate network of nerve fibers and cells interspersed among the SMCs.
On the other hand, smooth muscle cells cultured alone within the 3D-matrix kept their rounded shape and did not elongate (Fig. 5a). In case of paracrine interaction between smooth muscle and ENS cells, a moderate proliferation of the muscle cells was observed (Fig. 5b), which resulted in the formation of muscle fibers. Cultivation of SMCs in direct contact with ENS cells showed a large amount of extensive muscle fibers with strong SMA expression (Fig. 5c).
Fig. 5. Images of 3 weeks old co-culture of ENS cells and muscle cells in different layouts: (a) SMCs alone, (b) paracrine interaction between SMCs and ENS cells, (c) direct contact between SMCs and ENS cells Scale bars 100 μm, (d) paracrine interaction between SMCs and ENS in higher magnification, (e) Direct contact between smooth muscle and ENS cells in higher magnification.– muscle cells Scale bars 20 μm, (f) 3D reconstruction of muscle fibers in confocal microscopy by the direct contact between SMCs and ENS, (j) neurons, which are intercommunicated in the neuronal net within the muscle layer by the direct contact between SMCs and ENS. Confocal microscopy, Green (ß-Tubulin III) – neurons, red (smooth muscle actin (SMA)) – muscle cells. Scale bars 50 μm.
Samples with paracrine interaction showed that the muscle fibers resulted in the growth of a few neurons (Fig. 5d). In contrast, in samples with direct contact, a large number of neurons could be visualized among the muscle fibers, appearing both as groups of cells and as single cells with long, interconnected processes (Fig. 5e).
3D reconstruction of the samples with direct contact between muscle cells and ENS cells, presented neurons located in between the muscle bundles and in close contact to the muscle cells (Fig. 5f), building a 3D-neuronal net within the muscle layers (Fig. 5j). Immunohistological staining revealed close contact between ENS and muscle cells in such a manner that neurons guided the direction of the muscle cells. Our results interpretation is that this neuronal guiding is probably crucial for the creation of functional muscle layers.
Fig. 6. Imaging of the whole thickness of the 3D scaffolds after 3 weeks of ENS cells and muscle cells co-culture in different layouts: (a) SMCs without ENS cells, (b) paracrine interaction between SMCs and ENS cells, (c) direct contact between SMCs and ENS cells. Confocal microscopy, Green (ß-Tubulin III) – neurons, red (smooth muscle actin (SMA)) – muscle cells.
Next, we performed a 3D reconstruction imaging of the entire thickness of the gel. Our observations showed that in the group with only muscle cells there was no proliferation. However, in the group with paracrine interaction (without direct contact between ENS and muscle cells), we observed a very interesting condition: muscle cells in the middle of the gel did not show any proliferation, but those on the top of the gel did proliferate. Using higher magnification, we pictured a few neuronal cells in this structure. Probably some ENS cells, which were separated from the gel on the well bottom, detattached from the well, floated with the medium, and settled on top of the gel, thereby inducing the proliferation of muscle cells there (Fig. 6).
Finally, we confirmed the presence of communication structures, such as synapses between neurons and muscle cells, staining VAMP2 (Synaptobrevin 2) and smooth muscle actin (SMA) (Fig. 7) together. These showed specific contacts between neuronal and muscle cells with many synapses located on the muscle cells.
Fig. 7. Confocal microscopy images of smooth muscle and ENS cells co-cultured in 3D scaffolds (14 days of culture). Secretory vesicles and synapses are labelled in green (Synaptobrevin 2), while muscle cells in red (smooth muscle actin (SMA)). Scale bars 20 μm.
Discussion
The primary outcome of this study was to demonstrate that co-culture systems of enteric smooth muscle cells and ENS deliver functional gut wall tissue in vitro. This would be a first milestone for further developments of translational strategies to treat SBS. Our central objective was not to provide a definitive mechanistic explanation, but rather to explore the feasibility of such interactions in a controlled experimental setting.
This work should therefore be regarded as an exploratory study, providing initial experimental evidence for the necessity of neuronal innervation in the generation of a functional muscle layer in bioengineered intestine. By deliberately adopting a simplified co-culture system limited to enteric neurons and smooth muscle cells, we were able to minimize confounding effects from additional cell types and focus specifically on the earliest signs of cellular interaction.
Tissue engineering of the small bowel is an important issue in biomedical science and could be a game changer in the treatment of patients with short bowel syndrome and intestinal failure^21,22^. In contrast to numerous other organs that are solid in consistency, the intestine is a sophisticated, multi-layered, hollow organ^23^. Each layer is composed of its own cells and performs its own functions. The muscularis externa is an important part of the intestinal wall. It is a two-layered smooth muscle structure that works in conjunction with the ENS, which includes submucosal and myenteric nerve plexus. Together, they coordinate peristalsis and regulate functions such as gut hormone secretion, epithelial growth, barrier maintenance, and host-microbe interactions. Disorders of the ENS, such as Hirschsprung’s disease, in which ganglion cells in the ganglia are missing, can lead to significant health problems^23^.
The ENS regulates all sensory and motor functions within the gastrointestinal (GI) system and can operate autonomously from the brain, why it is often referred to as " the brain in the gut”^24^. It is well known from many congenital disorders of the bowel, caused by the absence or impaired number and function of neuronal ganglia, that the muscle layer without intrinsic innervation cannot perform its function. In patients with this condition, this part of the bowel causes significant motility disorders and therefore requires surgical removal^23,25^. This highlights the importance of constructing an innervated, functional muscle layer of the gut using tissue engineering.
In the present study, we developed a method for growing ENS cells and muscle cells together in a 3D matrix to construct an innervated muscle layer capable of coordinated contractions.
In the present study, immunostaining for SMA confirmed the presence of smooth muscle cells, while β-Tubulin III labeling identified neurons within the co-culture. The detection of nNOS- and ChAT-positive neurons demonstrated a diverse neuronal composition.
Most myenteric neurons express one of the two key enzymes responsible for the synthesis of the major excitatory and inhibitory neurotransmitters in the enteric nervous system, namely choline acetyltransferase (ChAT) or neuronal nitric oxide synthase (nNOS)^26^. In the present study, quantitative analysis revealed that 62% of neurons were ChAT-positive, whereas 33% were nNOS-positive. In addition, 25% of neurons exhibited co-expression of both ChAT and nNOS.
These proportions are broadly consistent with previous analyses of the human small intestine. Beck et al. reported that, in human small intestinal whole-mount preparations co-stained with the pan-neuronal marker Hu, ChAT-positive neurons were more abundant than nNOS-positive neurons (52% versus 38%)^26^. This distribution closely mirrors the predominance of ChAT-positive neurons observed in the present study and is consistent with the fact that both enteric neurons and smooth muscle cells analyzed here were derived from the small intestine.
Notably, the proportion of neurons co-expressing ChAT and nNOS observed in this study (25%) is substantially higher than that reported by Beck et al., who identified approximately 3% of double-positive neurons in the small intestine, with 7% of neurons negative for both enzymes^26^. Several factors may account for this discrepancy, including species-specific differences, variations in tissue processing and immunolabeling sensitivity, differences in the developmental or physiological state of the tissue, and the use of three-dimensional culture conditions, which may influence neurotransmitter phenotype plasticity or marker co-expression. In addition, the co-culture system used in the present study may contain a proportion of immature or partially differentiated enteric neurons, which could transiently express both cholinergic and nitrergic markers during phenotypic maturation.
In contrast to the small intestine, Beck et al. demonstrated that the large intestine exhibits a distinct neurochemical organization, with fewer ChAT-positive neurons and a higher proportion of nNOS-positive neurons (38% versus 50%). This regional specialization further supports the interpretation that the neurochemical profile observed in the present study reflects the small intestinal origin of the enteric nervous system cells analyzed.
Taken together, these findings indicate that the relative abundance of ChAT- and nNOS-expressing neurons in this model is consistent with established patterns in the small intestine, while the increased proportion of ChAT/nNOS double-positive neurons likely reflects methodological and biological features inherent to the experimental system.
Neurons expressing neuronal nitric oxide synthase (nNOS) function primarily as inhibitory motor neurons, mediating relaxation of gastrointestinal smooth muscle through the release of nitric oxide and cotransmitters such as VIP and PACAP, which regulate motility and accommodation reflexes. In contrast, choline acetyltransferase (ChAT)–positive neurons act mainly as excitatory motor and interneurons, releasing acetylcholine as the principal neurotransmitter to stimulate muscle contraction and secretory activity within the gut wall (Furness, 2000). Enteric glial cells, an essential component of the native ENS, were also integrated within the co-culture (Yang et al.). The coexistence of inhibitory (nNOS-expressing) and excitatory (ChAT-expressing) neuronal populations together with enteric glia suggests that this model establishes a physiologically relevant and interactive ENS-like microenvironment.
Additionally, we showed that close co-cultivation of smooth muscle cells and ENS cells results in the formation of innervated muscle layers, which we believe facilitates cellular interactions and highlights that ENS cells may play an important role in the development of functional, aligned muscle fibers within tissue-engineered intestinal models.
In addition, the identification of GFAP-positive glial cells indicates that the co-culture system successfully recapitulates key cellular components of ENS.
Mesenchymal-like cells observed in the cultures are interpreted to predominantly represent immature smooth muscle cells, with only a minor contribution from fibroblasts. This interpretation is supported by immunostaining for smooth muscle markers, which revealed the presence of weakly stained cells in Fig. 2, consistent with an immature smooth muscle phenotype that has not yet fully established robust expression of smooth muscle–specific actin. This conclusion is further supported by the fact that the ENS isolation and culture protocol used for smooth muscle cell differentiation was originally developed and first published by the authors’ group in 1997^27^ and has since been continuously refined and optimized, as documented in subsequent publications^28,29^.
Using electron microscopy and specific immunofluorescence with synaptophysin staining, we analyzed the microstructure of the cultivated constructs and identified synaptic connections between ENS neurons and smooth muscle cells. Importantly, this study was designed as an exploratory investigation, with the primary aim of establishing a simplified co-culture system and assessing early signs of neuronal–muscle interaction.
In our experiment, we used a hyaluronan-based hydrogel scaffold. The choice of a hyaluronic acid-based matrix was based on its unique physicochemical and biological properties, which make it suitable for growing nervous system cells, stem cells, and other cell types in tissue engineering and regenerative medicine^30,31^. The good growth and pronounced proliferation of neurons under the conditions we created confirm that hyaluronic acid in the extracellular matrix influences neuronal development in the human neocortex. This influence enhances neuronal plasticity and promotes growth by affecting the expression of receptors on neurons^30,32^.
We compared the growth pattern of a monoculture of SMCs in Hy-Stem Gel and in RAFT collagen - based matrix using live immunofluorescence staining with CellLight™ Actin-RFP, BacMam 2.0 (Invitrogen). The collagen-based RAFT matrix provides structural support and guidance cues that influence cellular proliferation, differentiation, and migration of these cells^33,34^. Unlike collagen-based scaffolds, the Hystem-C gel initially lacks structural fibers that could provide support and guidance for muscle cells. In our experiment muscle cells alone in the Hystem-C gel do not proliferate, likely due to the absence of structural support to guide them. However, muscle cells in co-cultures with ENS cells grow aligned and form muscle fibers. Unlike collagen-based structural support and guidance (as observed in the RAFT matrix), muscle cells in Hystem-C align with the ENS cells both in direction and form. The muscle fibers from our co-cultures of muscle and ENS cells show contracting activity. The presence of synapses on these muscle fibers suggests an additional role of the ENS cell network, providing not only structural support but also functional guidance.
It is known that mechanical forces play a major role in smooth muscle layer development and orientation. In our highly reductionist co-culture system, the influence of the ENS on smooth muscle alignment could, therefore, partly reflect an artificial co-culture artifact rather than a true developmental process. Nevertheless, our data indicate that the ENS can act as a modulatory factor that promotes smoother and more organized muscle alignment in the absence of other developmental cues, which could be of particular interest for the bioengineering of contractile tissue. Furthermore, the co-development of ENS and SMCs in this system results in the formation of close structural and possibly synaptic-like connections, which may have important implications for both tissue engineering and organoid model design.
While the present data do not provide direct mechanistic insight into how the ENS contributes to smooth muscle alignment, they introduce the possibility that such interactions may exist and highlight the potential of using similar approaches for the bioengineering of organized bowel wall structures. However, further studies are required to define the molecular and functional mechanisms underlying ENS–smooth muscle interactions.
In the RAFT matrix, we observed proliferation of muscle cells. However, the orientation of these cells was not aligned, making it impossible to achieve functional, directed contractions in the muscle fibers. Similarly, in aganglionic bowel from Hirschsprung’s patients, smooth muscle layers form despite the absence of ENS. These observations support the conclusion that the ENS can perform a modulatory role in the absence of other developmental factors or cell types, which could be of particular interest for tissue engineering applications.
In previous studies it could be shown that sphincter tissue can be generated while SMCs were cultivated along with neural precursor cells^35,36^. Muscular structures created through this process contain neurons and respond to exogenous contractile/relaxant transmitters^37^. However, this achieved contractility is an autonomic contractility of single muscle cells and is not neuronally mediated. This fact has also been supported by another study in which the monoculture of muscle cells cultured without neuronal cells reacted in the form of contractions to physiologically significant substances. This reaction indicates the presence of receptors involved in agonist-mediated contraction and relaxation pathways associated with G-proteins, but not coordinated^38^.
The coordinated contractions of entire muscle fibers, cultivated exclusively from smooth muscle cells and ENS cells without the contribution of other cell types as present in organoid systems, have not previously been reported in this form. 3D - reconstruction of confocal images from co-cultures of muscle cells and ENS cells in the Hy-Stem-C matrix allowed us to observe the muscle fibers throughout their entire thickness, confirming the good alignment of muscle cells with each other and with the neural network. The visualization of muscle cells (red channel in confocal microscopy) revealed a complex network of neuronal clusters within the muscle layer, closely resembling the ganglia of the native ENS and the neuronal plexus in the muscle layers of the bowel.
Examination of the nanostructure of muscle fibers with electron microscopy showed a strong similarity to native smooth muscle fibers, including large nuclei in the muscle cells, tightly aligned cells, actin fibers, and caveolae—signs of cellular activity. Numerous synapses on the muscle cells, revealed by immunofluorescence staining with vimentin, confirmed our belief that neurons and muscle cells not only grow and develop together but also interact closely.
The importance of ENS cells in muscle fiber development was explored in our experiment by cultivating muscle cells alone in a 3D matrix, with ENS cells separated by a semipermeable membrane to investigate paracrine interactions, and in direct co-culture with ENS cells in an optimal layout. As expected, in the samples with only muscle cells in the Hy-Stem gel, there was no proliferation of the cells. However, in the samples with muscle and ENS cells in an optimal layout, we observed the growth and proliferation of muscle layers. The most interesting results were found in the paracrine group, where ENS and muscle cells were separated by a semipermeable membrane to investigate paracrine interaction. In confocal microscopy, we observed muscle cells without any proliferation as well as cells showing proliferation and the formation of muscle fibers, although these fibers were considerably weaker than in the group with direct contact between muscle and ENS cells. Importantly, the muscle fibers in the paracrine group also contained neurons.
Notably, in the paracrine group, muscle fibers with ENS cells were located on the top of the gel, whereas in the group with direct co-cultivation of ENS and muscle cells, the muscle fibers were located in the middle part of the gel, where the ENS network was initially cultivated. In the paracrine group, ENS cells were placed at the bottom of the outer well, and muscle cells were embedded in the gel within the insert, without direct contact between the two. However, both were exposed to the same medium in the well. We believe that a small amount of ENS cells, initially placed at the bottom of the well, floated during medium changes and settled on top of the gel. This allowed the ENS cells to grow and proliferate, which in turn provoked muscle cells from the gel to attach to the ENS cells and form muscle fibers.
Rather than focusing on single elements of the ENS, such as glial cells^39^ or neuronal precursors^35^, we generated contractile small-intestinal muscle fibers containing structures resembling the intermuscular neural plexus, which enables peristaltic motion in bioengineered intestine (see Supplementary Videos S1 and S2). By contrast to organoid-based systems, which include a broad pool of diverse cell types that stimulate one another and give rise to partial ENS elements^17^, our study deliberately adopted a reductionist approach. We intentionally limited the culture to only two cell types—smooth muscle cells and ENS cells—allowing us to investigate their direct interactions in a highly controlled environment while minimizing the confounding effects of additional cell types and signaling factors inevitably present in complex organoid systems.
Altogether, this exploratory study establishes proof-of-concept that a functional neuromuscular system can be generated using our approach. We observed that proper co-cultivation of enteric neurons and smooth muscle cells in 3D matrices enables the artificial creation of a functional muscle layer with contractile activity.
Importantly, the observed contractile activity should be interpreted with caution, as it may reflect an apparent rather than definitively neuron-mediated effect. The present data do not exclude the possibility that contractility arises from intrinsic smooth muscle activity or culture conditions independent of direct neuronal regulation. Accordingly, additional studies will be required to rigorously evaluate the origin and mechanisms underlying the observed contractile behavior and to determine the extent to which neuronal activity contributes to smooth muscle organization, maturation, and function.
We believe that ENS cells interact with muscle cells both at the individual cellular level and at the tissue level, and that this interaction is crucial for the functional coordination between smooth muscle and neuronal cells and for the formation of a functional muscle layer of the small intestine.
To substantiate these hypotheses, further in-depth studies are required. Future work will need to address epithelial integration and its cross-talk with the ENS in order to advance translational intestinal tissue engineering. Moreover, detailed mechanistic investigations—including quantitative characterization of contractility and precise definition of the cellular composition—will be essential. A comprehensive characterization of neuronal subtypes (including inhibitory and excitatory neurons) and enteric glia, as well as the potential involvement of interstitial cells of Cajal (ICCs), represents an important next step and is planned as part of our future investigations.
The limitations of this study stem from its purely exploratory nature, as it was focused on establishing proof-of-concept for a simplified co-culture model. The system was restricted to rodent-derived smooth muscle and ENS cells, without addressing epithelial integration. Furthermore, quantitative analysis of contractility, neuronal functionality, mechanistic insight into how the ENS interacts with smooth muscle cells, and ultrastructural characterization of synapses were beyond the scope of this work but represent important directions for future research.
A key strength of this study is its deliberate reductionist design, limiting the system to only smooth muscle cells and ENS cells. This allowed us to investigate direct neuronal–muscle interactions in a highly controlled environment, minimizing confounding effects from additional cell types typically present in organoid models. Using 3D co-culture, we demonstrated the feasibility of generating contractile muscle layers with synaptic connections, thereby providing a clear proof-of-concept and a solid foundation for future mechanistic and translational studies.
Conclusion
This exploratory study establishes proof-of-concept that the Enteric Nervous System (ENS) represents a key component in constructing a bioengineered intestinal muscle layer. Within a simplified 3D co-culture system, we demonstrated that smooth muscle cells co-developed with ENS components to form aligned, innervated muscle fibers capable of spontaneous contractile activity.
The goal of this study was to explore how the ENS can contribute to smooth muscle organization under reductionist and controlled conditions. In our co-culture system, the observed effects may partly reflect an artificial co-culture artifact rather than a true developmental process. Nevertheless, our findings provide new insights showing that developing ENS and smooth muscle cells can establish close structural and possibly functional (synaptic-like) connections in vitro.
We believe that this knowledge will be valuable for the design of next-generation tissue-engineered intestinal models. While these findings highlight the potential of ENS–smooth muscle co-cultures for intestinal tissue engineering, further studies integrating epithelial elements, assessing neuronal functionality, and providing mechanistic insight into how the ENS interacts with smooth muscle cells will be necessary to advance this approach toward translational and clinically relevant applications.
Materials and methods
Animals
This study was carried out in strict accordance with the recommendations for the care and use of laboratory animals was conducted in compliance with both institutional and national guidelines for animal care and use. All methods were performed in accordance with relevant guidelines and regulations. All experimental protocols were reviewed and approved by Core Facility Preclinical Models of the Medical Faculty Mannheim, which is a internal Veterinary Inspection Office in Mannheim, under the following authorization number: I-19/23.
Pregnant Sprague Dawley rats were purchased from JANVIER LABS. The use of animals in this research was conducted in compliance with institutional and national regulations, and the study adheres to the ARRIVE guidelines(https://arriveguidelines.org) to ensure transparent and reproducible reporting of animal experiments.
Cells isolation and scaffold
The enteric nervous system (ENS) cells were isolated from the gut of postnatal rats (Sprague–Dawley) using enzymatic digestion combined with mechanical agitation^27^. The procedure was conducted in accordance with the recommendations for the care and use of laboratory animals, following both institutional and national guidelines, and was approved by the Core Facility Preclinical Models of the Medical Faculty Mannheim. The animals were 4–6 days old with an average body weight of approximately 10–15 g and were euthanized by decapitation.
Briefly, muscle and submucous layer were separated, and the dissected muscle tissue incubated in a collagenase II solution (Worthington) for 2 h. After vortexing the tissue for 10 s., isolated pieces of myenteric plexus could be collected and stored in MEM-Hepes (Gibco, Thermo Fischer Scientific) on ice. The obtained pure myenteric plexus was incubated in trypsin EDTA (Gibco, Thermo Fischer Scientific)) for 10 min. and dissociated by trituration. In a second centrifugation step, the trypsin was re- moved and replaced with 1 ml culture medium (see Supplementary Fig. S4 online).
Rat Primary Small Intestinal Smooth Muscle Cells (SMCs) were commercially available and bought (PELOBiotech GmbH, Martinsried, Germany).
As a scaffold Hystem-C (BioTime Inc., Alameda, CA) hydrogel was used. This is a hyaluronan-based hydrogel crosslinked using thiol-reactive poly(ethylene glycol) diacrylate plus thiolated denatured porcine collagen.
Experimental design
The first step was to test the optimal culture conditions. Normally, muscle cells and ENS cells are cultivated in their own culture media. To develop artificially constructed functional muscle fibers, we need to cultivate SMCs and neurons in a single medium.
To investigate the optimal conditions for both proliferation and growth of the co-cultures muscle cells and ENS cells at the first stage we tested different microenvironmental factors such as diverse media solutions with different combinations of growth factors.
Cell culture reagents were purchased to: Neurobasal A, DMF12 media were purchased to Gibco, (Thermo Fischer Scientific)), supplemented with 2% B27+ (with vitamin A, Gibco), EGF (10 ng/mL, Tebue), bFGF (20 ng/mL; Tebue) and GDNF (10 ng/mL, Tebue), 1% albumin (Sigma-Aldrich), 0.25% 2-mercaptoethanol (50mM, Invitrogen), 0.12% glutamine (200 mM, Sigma-Aldrich), gentamycin/metronidazole (100 µg/mL) or 1% penicillin/streptomycin (PAA Laboratories)).
The following culture media conditions were tested:
- Neurobasal + BSA+Antibiotics + 2-Mercaptoethanol + glutamine + B27+.
- DMEM-F12 + BSA+ Antibiotics + 2-Mercaptoethanol + glutamine + B27+.
- Neurobasal + BSA+ Antibiotics + 2-Mercaptoethanol+ glutamine + B27 + + Factors (bFGF and GDNF).
- DMEM-F12 + BSA+ Antibiotics + 2-Mercaptoethanol+ glutamine + B27 + + Factors (bFGF and GDNF).
- DMEM-F12 + BSA+ Antibiotics + 2-Mercaptoethanol + glutamine + B27 + + Factors (bFGF and GDNF) + Serum.
Once the optimal culture condition was identified, it was used for the next stages of the study.
At the second Stage we tested the optimal layout of SMCs and ENS cells.
We used 60 000 SMCs per 100 µl and 100 000 ENS cells per 100 µl. In order to investigate the optimal layout of the cells 6 different approaches were tested, where both cell types (ENS cells and SMCs) were combined cells (Fig. 8b, c,d, e,f, j).
Fig. 8. Schematic representation of plates with inserts and different layouts of muscle and ENS cells in a three-dimensional matrix: (a) plates with insert, which allows the medium to surround and support the co-culture from all sides, (b) SMCs without ENS cells in the 3D-matrix, (c) SMCs and ENS cells equally distributed in 3D-matrix, (d) SMCs distributed in upper and lower layers, ENS cells - in the middle layer of the 3D-matrix. (e) SMCs distributed in upper and lower layers, ENS cells densely arranged in the same plane of the middle layer of the 3D-matrix. (f) SMCs and ENS cells distributed in the middle layer of the 3D-matrix. (g) mixed SMCs and ENS cells densely arranged in the same plane of the middle layer of the 3D-matrix, (h) different layouts between SMCs and isolated myenteric plexus (ENS cells). SMCs alone, with direct contact and without direct contact to isolated myenteric plexus cells.
Special 24-well plates with inserts (BRAND GMBH, Germany) were used to cultivate the cells appropriately. The inserts were placed inside the wells in such a way that they only touched the edges of the wells, while the rest was immersed in the culture medium that surrounded the well on all sides (Fig. 8a). The advantage of using insert plates is that the bottom of the insert consists of a membrane with 0.4 μm diameter pores which is permeable for the culture medium. In addition, the insert’s walls have holes, so the culture medium also comes through and gets contact with the contents of the insert. In this case, this is a matrix in which the cells are immersed.
200 µl of matrix were pipetted into the inserts for the 24-well plates. The inserts were then placed into these plates (BRAND GMBH, Germany). The matrix was allowed to polymerize for 30 min.
Next, 1.5 ml of culture medium for neurocultures were added to each well. The resulting culture was placed in a HERAcell^®^ incubator (Kendro Laboratory Products) and grown for 10–21 days at 37 °C with a 5% CO2 concentration. The culture medium was changed every 48 h.
During the process of tissue development in vitro the assessment of cellular structures was conducted microsopically on a daily basis.
We wanted to compare the differences in cultivating muscle cells in collagen-based scaffolds from RAFT (Lonza).
We planted 60 000 muscle cells pro 100 µl according to standard protocol.
We performed the analysis using 24-well RAFT™ (3D) plates according to the manufacturer’s instructions (TAP Biosystems, Lonza, Cologne, Germany). Briefly, islets were mixed into the chilled collagen solution from the RAFT™ kit, which contained 2.8 mL of 10X MEM medium (Gibco, Thermo Fischer Scientific), 22.4 mL of 2 mg/mL rat tail collagen type I, 1.624 mL of neutralizing solution, and 1.2 mL of islets for a tissue culture plate, following the RAFT™ protocol. We added 60,000 muscle cells per 100 µl and plated 240 µl of the mixed collagen solution into each well. The plate was then incubated at 37 °C for 15 min to form a hydrogel. The RAFT™ absorbents were then placed on top of the hydrogel in a laminar flow hood at room temperature for 15 min. The medium was changed every two days.
At the third stage we studied the spatial relation between SMCs and ENS cells.
To investigate the influence of the ENS upon smooth muscle cell growth, ENS and SMCs were cultured within a 3D-hydrogel in three different conditions (Fig. 8h):
- Muscle cells alone,
- Muscle cells with ENS cells, separated by a semipermeable membrane to investigate paracrine interaction,
- Muscle cells with ENS cells in the same culture- in close contact.
Immunofluorescence
Staining of live cultures
Muscle cells were stained with the live staining (CellLight™ Actin-RFP, BacMam 2.0 (Invitrogen)) following a standard protocol: 20 µl of BacMam 2.0 reagent was added in media and incubated for 16–24 h at 37 °C, followed by daily microscopy.
Staining of fixed cultures
After the culture period, the 3D gel constructs were fixed in 4% paraformaldehyde in PBS (Sigma-Aldrich, St. Louis, MO, USA) for 30 min, permeabilized with 1% Triton X-100 for 15 min, and washed three times with PBS. Samples were then blocked with 10% normal goat serum (Agilent Dako, Santa Clara, CA, USA) for 1 h at room temperature and incubated overnight at 4 °C with the corresponding primary antibodies. The next day, gels were washed in PBS and incubated for 2 h at room temperature with the following secondary antibodies: Alexa Fluor^®^ 488 (Goat Anti-Mouse, #A-10667; Molecular Probes, Invitrogen, Life Technologies, Carlsbad, CA, USA), Alexa Fluor^®^ 568 (Goat Anti-Rabbit, #A-11011; Goat Anti-Mouse, #A-11004; Molecular Probes, Invitrogen, Life Technologies, Carlsbad, CA, USA), Donkey-anti-Chicken 647 (Jackson Immuno Research), Donkey Anti-Goat Alexa Fluor^®^ 647 (Thermo Fisher Scientific, A-21447; 1:500), and Donkey Anti-Rabbit Alexa Fluor^®^ 647 (Invitrogen, A32795; 1:500). After additional PBS washes, the samples were counterstained with DRAQ5 and mounted using Dako Fluorescence Mounting Medium (S3023; Agilent Dako, Santa Clara, CA, USA).
The following primary antibodies were used:
Synaptobrevin 2 (Agilent Dako, Santa Clara, CA, USA);
PGP 9.5 (Invitrogen, CA, USA),
SMA (Agilent Dako, Santa Clara, CA, USA);
Anti-βIII Tubulin Alexa Fluor^®^ 488 Conjugate (AB15708A4; Millipore Sigma, St. Louis, MO, USA);
αSMA (Invitrogen, AB5695; Rabbit, 1:200);
nNOS (R&D Systems, MAB9502; Rabbit, 1:1000);
ChAT (Novus Biologicals, NBP1-30052; Goat, 1:250);
GFAP (Abcam, AB53554; Goat, 1:250).
All staining procedures were performed directly within the 3D co-culture constructs inside the wells. Following staining, samples were stored at 4 °C in the dark until image acquisition.
Imaging and image analysis
Live images and fluorescence images were taken using an Olympus IX 50 inverted microscope (Olympus) with a CCD camera connected to the computer with AnalySIS software (Olympus). Confocal microscopy was performed with Leica TCS SP8 confocal microscope (Leica, Germany) with photographic documentation and with the software Leica LAS X.
Electron microscopy
Regarding electron microscopy, the gels were fixed in glutaraldehyde and contrast stained with OsO4. The Samples were embedded in Epon 812 and cut into 90 nm ultrathin sections and viewed under a Zeiss EM 910 electron microscope (Carl Zeiss, Oberkochen, Germany).
Supplementary Information
Below is the link to the electronic supplementary material.
Supplementary Material 1
Supplementary Material 2
Supplementary Material 3
The reference list from the paper itself. Each links out to its DOI / PubMed record.
- 1Lotfollahzadeh, S., Taherian, M. & Anand, S. In Stat Pearls (2022).