This protocol describes an in vitro model of necrotizing enterocolitis (NEC), which can be used for mechanistic studies into disease pathogenesis. It features a microfluidic chip seeded with intestinal enteroids derived from the human neonatal intestine, endothelial cells, and the intestinal microbiome of a neonate with severe NEC.
Necrotizing enterocolitis (NEC) is a severe and potentially fatal intestinal disease that has been difficult to study due to its complex pathogenesis, which remains incompletely understood. The pathophysiology of NEC includes disruption of intestinal tight junctions, increased gut barrier permeability, epithelial cell death, microbial dysbiosis, and dysregulated inflammation. Traditional tools to study NEC include animal models, cell lines, and human or mouse intestinal organoids. While studies using those model systems have improved the field's understanding of disease pathophysiology, their ability to recapitulate the complexity of human NEC is limited. An improved in vitro model of NEC using microfluidic technology, named NEC-on-a-chip, has now been developed. The NEC-on-a-chip model consists of a microfluidic device seeded with intestinal enteroids derived from a preterm neonate, co-cultured with human endothelial cells and the microbiome from an infant with severe NEC. This model is a valuable tool for mechanistic studies into the pathophysiology of NEC and a new resource for drug discovery testing for neonatal intestinal diseases. In this manuscript, a detailed description of the NEC-on-a-chip model will be provided.
Necrotizing enterocolitis (NEC) affects preterm infants, with an incidence of up to 10% in those born weighing < 1500 g1. The pathophysiology of NEC is complex and includes damage to the intestinal epithelium, disruption of intestinal tight junctions, increased gut barrier permeability, immune dysregulation, and epithelial cell death2,3. Our understanding of the mechanisms involved in the pathogenesis of NEC remains incomplete, and despite decades of research, there are still no effective targeted therapies.
A significant barrier to advancing NEC research is the limited availability and small size of primary intestinal tissue isolated from human infants. Intestinal tissue resected from infants with NEC is often necrotic and severely damaged, which complicates studies into mechanisms that precede disease onset. For example, the small intestine of infants with NEC is inundated with immune cells, and a reduced number of intestinal stem cells, decreased epithelial cell proliferation, and increased epithelial cell apoptosis are also observed4,5,6,7. This leads to difficulties in culturing intestinal epithelial cells from these samples and in isolating RNA and proteins, which can be degraded in this hostile inflammatory environment. Additionally, since the disease process is already advanced in infants with surgical NEC, mechanistic studies into factors that induce disease are unfeasible. These limitations have led to a reliance on animal models for mechanistic studies of NEC.
Animal models of NEC have been established for mice, rats, piglets, rabbits, and baboons5,8,9,11. A strength of animal models is that NEC-like intestinal disease is induced by factors associated with NEC onset in humans, including a dysbiotic microbiome, repeated episodes of hypoxia, and the absence of breast milk feeds5,8,10,11. In addition, the inflammatory response and pathologic changes observed during experimental NEC parallel human disease5,9,12. While these models mimic many of the features of human NEC, there are inherent differences between the pathophysiology of NEC in animals and humans. For example, the murine model of NEC is induced in mice born full-term, and although their intestinal development is incomplete, the pathophysiology of NEC is inherently different in this clinical context. Murine intestinal gene expression at birth is similar to a pre-viable human fetus and does not approximate that of a preterm neonate of 22-24 weeks gestation until day 14 (P14)13. This confounds the murine NEC model because intestinal injury cannot generally be induced in mice after P10. In addition, inbred strains of mice lack the immunologic14 and microbiologic diversity of human neonates15, which serves as another confounding factor. Thus, increased incorporation of primary human samples into NEC research improves the clinical relevance of studies in this field.
Studies into the mechanisms of NEC in vitro have traditionally utilized monotypic cell lines derived from adult intestinal cancer cells, such as colorectal adenocarcinoma (Caco2) and human colon adenocarcinoma (HT-29) cells16. These models are convenient but limited in physiologic relevance due to their growth from adult cancer cells, non-polarized architecture, and phenotypic changes related to repeated passages in culture. Intestinal enteroids improve upon those models since they can be grown from the crypts of intestinal tissue, differentiated into all intestinal epithelial subtypes, and form a three-dimensional (3D) villus-like structure17,18,19,20. Recently, intestinal enteroids have been combined with microfluidic technology to develop a small intestine-on-a-chip model and provide a more physiologically relevant in vitro model system21.
The initial organ-on-a-chip microfluidic devices were introduced in the early 2000s22,23,24. The first organ-on-a-chip model was the human breathing lung-on-a-chip25. This was followed by numerous single-organ models such as intestine21, liver26, kidneys27, bone marrow28, blood-brain barrier29, and heart30. These organ-on-a-chip models have been used to study acute, chronic, and rare diseases, including acute radiation syndrome,31 chronic obstructive pulmonary disease,32 and neurodegenerative diseases33. The polarized nature of the cells on these chips and the presence of two cellular compartments separated by a porous membrane allows for the modeling of complex physiologic processes such as perfusion, chemical concentration gradients, and immune cell chemotaxis34,35. These microfluidic systems thus provide a new tool for studying the pathophysiology and mechanisms of human disease.
The small intestine-on-a-chip model was described by Kasendra et al. in 2018, who utilized pediatric (ages 10-14 years old) small intestinal biopsy specimens differentiated into enteroids and cultured on a microfluidic device21. Vascular endothelial cells, continuous media flow, and stretch/relaxation were also incorporated into this model. They observed intestinal epithelial subtype differentiation, formation of 3D villus-like axes, mucus production, and small intestinal gene expression patterns21. This microfluidic model was applied to neonatal disease with the development of the NEC-on-a-chip system, which incorporates neonatal intestinal enteroids, endothelial cells, and the microbiome from a neonate with NEC36. NEC-on-a-chip recapitulates many of the critical features of human NEC, including inflammatory gene expression, loss of specialized epithelial cells, and reduced gut barrier function36. Thus, this model has numerous applications in the study of NEC, including mechanistic studies and drug discovery. In this manuscript, a detailed protocol for the performance of the NEC-on-a-chip model is provided.
Enteroids were derived from small intestinal samples from premature infants (born at 22 to 36 weeks gestation) obtained at the time of surgery for NEC or other intestinal conditions with non-inflammatory etiologies. All specimen collection and processing was performed after informed consent and approval from the Institutional Review Boards at Washington University in St. Louis (IRB Protocol numbers 201706182 and 201804040) and the University of North Carolina at Chapel Hill (IRB protocol number 21-3134).
1. Isolation and plating of crypts from human neonatal small intestine to establish enteroids
2. Neonatal intestine-on-a-chip model
NOTE: For detailed instructions on handling the microfluidic chips and using this equipment, please see the manufacturer's duodenum intestine-chip culture protocol37.
3. NEC-on-a-chip model
4. Intestinal permeability assay
NOTE: This can be performed at any step during the protocol. If initiated when chips are seeded, the intestinal permeability assay can be used to serially assess monolayer confluence. If initiated upon the addition of intestinal bacteria, this assay can be used to determine the influence of intestinal bacteria on gut epithelial monolayer integrity.
5. Immunohistochemistry
6. Isolation of RNA, preparation of cDNA, and quantitative real-time PCR
Enteroids were seeded onto the microfluidic device (Figure 1) and cultured as described above. Growth of the enteroids in cell culture matrix hydrogel prior to seeding and then the subsequent expansion of the intestinal epithelial cell monolayer after seeding the device was monitored via brightfield microscopy (Figure 2). A confluent intestinal epithelial cell monolayer formed and subsequently developed into a mature 3D villus-like structure (Figure 2). This microfluidic device seeded with an enteroid-derived neonatal intestinal epithelium and HIMECs is known as the neonatal intestine-on-a-chip model36.
The NEC-on-a-chip model was developed to recapitulate the microbial dysbiosis present during neonatal NEC. This model includes the addition of a dysbiotic microbiome from a neonate with severe NEC to the neonatal intestine-on-a-chip model. With the addition of this dysbiotic microbiome, significantly increased expression of an array of pro-inflammatory cytokines, including tumor necrosis factor alpha (TNFα; Figure 3), interleukin (IL)-1 beta (IL-1β)36, and IL-8 was detected36. This increase in proinflammatory cytokines mirrors what is observed for human NEC36.
Figure 1: Schematic representation of enteroid culture and the microfluidics platform. Crypts are isolated from a surgically resected piece of neonatal intestine, and they are seeded in cell culture matrix hydrogel and grown into intestinal enteroids. Enteroids are dissociated and seeded in the top channel of the microfluidics device coated with extracellular matrix (ECM). Endothelial cells (HIMECs) are then added to the bottom channel. Figure created with Biorender.com. Please click here to view a larger version of this figure.
Figure 2: Progression of intestinal epithelial cells grown using the neonatal intestine-on-a-chip model. Images acquired using brightfield microscopy demonstrate neonatal enteroids on day 0, an epithelial cell monolayer in the chip on day 3, and visible villus-like structures on day 7. Scale bars = 100 µm. Please click here to view a larger version of this figure.
Figure 3. NEC-on-a-chip intestinal epithelial cell TNFα mRNA expression. Comparison of TNFα mRNA levels upon incubation with media alone (control) or a dysbiotic microbiome from an infant with NEC (NEC-on-a-chip) for 24 h or 72 h. **** p < 0.0001 vs. control 24 h; ** p < 0.005 vs. control 72 h by Mann-Whitney U test. n=9 for control 24 h; n=10 for NEC-on-a-chip 24 h; n=6 for control 72 h; n=5 for NEC-on-a-chip 72 h. Data is mean ± SEM. Please click here to view a larger version of this figure.
Table 1. Media composition for crypt isolation and enteroid growth. Refer to Van Dussen et al.40 and Miyoshi et al.41 for detailed methods describing the preparation of L-WRN conditioned media. Please click here to download this Table.
Table 2. Media composition for the NEC-on-a-chip model. Please click here to download this Table.
This NEC-on-a-chip system is a powerful new tool that can be used to model the pathophysiology of NEC. This platform provides a complex microenvironment that more closely resembles the in vivo intestinal milieu than previous models by incorporating a co-culture system with continuous luminal flow and stretch. These conditions promote the development of 3D villus-like architecture lined by a highly polarized epithelium consisting of mature epithelial sub-types and tight junctions (Figure 2)36. Additionally, the apical epithelial surface is easily accessible for exposure to experimental stimuli, such as patient-derived microbiota or novel therapeutics. Finally, this model could also be used to study a variety of inflammatory and non-inflammatory intestinal diseases, as well as mechanisms of normal intestinal epithelial development, by utilizing enteroids derived from the relevant patient populations. This model allows for the maximization of data acquisition from a single patient sample, which is essential given the limited availability and size of neonatal intestinal samples.
Using this model, many of the physiologic changes in the intestinal epithelium that occur during NEC in infants were observed. The addition of the dysbiotic microbiome from a patient with severe NEC led to the upregulation of inflammatory cytokines (Figure 3), increased gut-on-a-chip barrier permeability, enhanced cell death, decreased proliferation, and the loss of mature epithelial sub-types36. Thus, future studies utilizing this model can focus on determining the mechanisms underlying the development of NEC and possible therapeutic interventions.
There are inherent limitations in implementing a complex model such as NEC-on-a-chip. This system requires specialized equipment, reagents, and training to utilize the microfluidics platform. In addition, the model requires close attention to detail with accurate preparation of the various media, proper seeding of the chip, and maintenance of sterile technique, being critical for success. Investigators must also have access to intestinal samples or enteroids from neonatal patients, which are generally only available at quaternary care centers with pediatric surgeons unless obtained through collaborators. Lastly, incorporation of additional important components in the pathophysiology of NEC, such as immune cells or hypoxia, further increases the difficulty of this model although increases the physiologic relevance.
In summary, NEC-on-a-chip is an important advance in the field of NEC research that will improve the quality of mechanistic studies and facilitate the testing of novel therapeutics for NEC. The ultimate goal of this research is to design and test targeted therapies that improve outcomes for infants with intestinal inflammation and NEC.
The authors have nothing to disclose.
This manuscript was supported by R01DK118568 (MG), R01DK124614 (MG), and R01HD105301 (MG) from the National Institutes of Health, the Chan Zuckerberg Initiative Grant 2022-316749 (MG), a Thrasher Research Fund Early Career Award (LCF), a UNC Children's Development Early Career Investigator Grant (LCF) through the generous support of donors to the University of North Carolina at Chapel Hill, and the Department of Pediatrics at the University of North Carolina at Chapel Hill.
[Leu15]-Gastrin I human | Sigma-Aldrich | G9145 | |
A 83-01 | Sigma-Aldrich | SML0788 | |
Advanced Dulbecco's Modified Eagle Medium/Ham's F-12 | Gibco | 12634010 | |
B-27 Supplement, serum free (50x) | Gibco | 17504044 | |
Basic Bio-kit | Emulate | N/A | |
BioTek Synergy 2 Multi-Mode Microplate Reader | Agilent | 7131000 | |
BRAND Methacrylate (PMMA) Cuvettes, Semi-Micro | BrandTech | 759085D | |
Cell Recovery Solution | Corning | 354270 | |
CFX Opus Real-Time PCR Systems | Bio-Rad | 12011319 | |
Chip Cradle | Emulate | N/A | |
Chip-S1 Stretchable Chip | Emulate | N/A | |
CHIR99021 | Sigma-Aldrich | SML1046 | |
Clear TC-treated Multiple Well Plates, 48 well | Corning | 3548 | |
Collagen from human placenta | Sigma-Aldrich | C5533 | |
Collagenase, Type I, powder | Gibco | 17018029 | |
Complete Human Endothelial Cell Medium with Kit | Cell Biologics | H-1168 | |
Conical Polypropylene Centrifuge Tubes, 15 mL | Fisher Scientific | 05-539-12 | |
Conical Polypropylene Centrifuge Tubes, 50mL | Fisher Scientific | 05-539-8 | |
Countess Cell Counting Chamber Slides | Invitrogen | C10283 | |
Countess II automated cell counter | Invitrogen | AMQAX1000 | |
DAPI (4',6-Diamidino-2-Phenylindole, Dilactate) | Invitrogen | D3571 | |
DAPT | Sigma-Aldrich | D5942 | |
Dextran, Cascade Blue, 3000 MW, Anionic, Lysine Fixable | Invitrogen | D7132 | Permeability dye |
Dimethyl sulfoxide (DMSO) | Sigma-Aldrich | D8418 | |
Disposable PES Filter Units, 0.2um aPES membrane | Fisher Scientific | FB12566504 | |
DMEM/F-12 | Gibco | 11320033 | |
Donkey serum | Sigma-Aldrich | D9663 | |
Dulbecco′s Modified Eagle′s Medium – high glucose | Sigma-Aldrich | D5796 | |
Dulbecco′s Phosphate Buffered Saline (DPBS) | Gibco | 14190-136 | |
EDTA, 0.5 M, pH 8.0 | Corning | 46-034-CI | |
ER-1 surface activation reagent | Emulate | ER-1 | Chip Activation Reagent 1 |
ER-2 surface activation reagent | Emulate | ER-2 | Chip Activation Reagent 2 |
Fibronectin Human Protein, Plasma | Gibco | 33016015 | |
Fisherbrand Petri Dishes with Clear Lid, 100mm | Fisher Scientific | FB0875713 | |
Gelatin-Based Coating Solution | Cell Biologics | 6950 | |
Genie Temp-Shaker 300 | Scientific Industries, Inc. | SI-G300 | |
Gentamicin | Gibco | 15750060 | |
HEPES, Liquid 1M Solution (238.3 mg/ mL) | Corning | 25-060-CI | |
Hoechst 33342, Trihydrochloride, Trihydrate | Invitrogen | H3570 | |
Human Collagen Type I | Sigma-Aldrich | CC050 | |
Human Primary Small Intestinal Microvascular Endothelial Cells | Cell Biologics | H-6054 | |
Inverted Microscope | Fisher Scientific | 03-000-013 | |
Isotemp General Purpose Deluxe Water Baths | Fisher Scientific | FSGPD10 | |
L-Glutamine | Gibco | 25030-081 | |
Luria Broth (LB) agar, Miller | Supelco | L3027 | |
L-WRN Cells | American Type Culture Collection | CRL-3276 | |
Matrigel Growth Factor Reduced Basement Membrane Matrix, LDEV-free | Corning | 356231 | Cell Culture Matrix |
N-2 Supplement (100x) | Gibco | 17502048 | |
N-acetyl-L-cysteine | Sigma-Aldrich | 1009005 | |
NAILSTAR UV LAMP | NailStar | NS-01-US | |
NanoDrop OneC Microvolume UV-Vis Spectrophotometer | Thermo Scientific | 840-274200 | |
Nicotinamide | Sigma-Aldrich | 72340 | |
Orb-HM1 Hub Module | Emulate | N/A | |
Paraformaldehyde | ThermoFisher | 047392.9L | |
Penicillin-Streptomycin | Gibco | 15140122 | |
Phosphate buffered saline (PBS) | Gibco | 10010023 | |
Pipet-Lite Multi Pipette L8-200XLS+ | Rainin | 17013805 | |
Pipette Tips TR LTS 1000µL S 768A/8 | Rainin | 17014966 | |
Pod Portable Module | Emulate | N/A | |
Premium Grade Fetal Bovine Serum (FBS)(Heat Inactivated) | Avantor Seradigm | 1500-500 | |
QuantiTect Reverse Transcription Kit | QIAGEN | 205313 | |
Recombinant Murine Epidermal Growth Factor (EGF) | PeproTech | 315-09 | |
SB 431542 | Tocris | 1614 | |
Square BioAssay Dish with Handles, not TC-treated | Corning | 431111 | |
SsoAdvanced Universal SYBR Green Supermix | Bio-Rad | 1725271 | |
Steriflip-GV Sterile Centrifuge Tube Top Filter Unit | Millipore | SE1M179M6 | |
Sterile Cell Strainers, 70um | Fisher Scientific | 22-363-548 | |
Sterile Syringes, 10mL | Fisher Scientific | 14-955-453 | |
Straight, fine, sharp point scissors | Miltex Instruments | MH5-300 | |
Thermo Scientific Sorvall X4R Pro-MD Centrifuge | Thermo Scientific | 75016052 | |
Triton X-100 | Sigma-Aldrich | T8787 | Detergent |
TRIzol Reagent | Invitrogen | 15596026 | RNA extraction reagent |
Trypan Blue Solution, 0.4% (w/v) in PBS, pH 7.5 ± 0.5 | Corning | 25-900-CI | |
TrypLE Express Enzyme (1X), no phenol red | Gibco | 12604013 | Enzymatic Dissociation Reagent |
Trypsin-EDTA solution | Sigma-Aldrich | T4174 | |
VIOS 160i CO2 Incubator, 165 L | Thermo Scientific | 13-998-252 | |
Y-27632 | Tocris | 1254 | |
Zoë-CM1 Culture Module | Emulate | N/A |