If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
The Wallace H. Coulter Department of Biomedical Engineering, Georgia Institute of Technology and Emory University School of Medicine, Atlanta, GeorgiaBiointerfaces Institute, University of Michigan, Ann Arbor, MichiganDepartment of Biomedical Engineering, University of Michigan, Ann Arbor, Michigan
The gastrointestinal (GI) tract regulates physiologic responses in complex ways beyond facilitating nutrient entry into the circulatory system. Because of the anatomic location of the GI tract, studying in vivo physiology of the human gut, including host cell interaction with the microbiota, is limited. GI organoids derived from human stem cells are gaining interest as they recapitulate in vivo cellular phenotypes and functions. An underdeveloped capability that would further enhance the utility of these miniature models of the GI tract is to use sensors to quantitatively characterize the organoid systems with high spatiotemporal resolution. In this review, we first discuss tools to capture changes in the fluid milieu of organoid cultures both in the organoid exterior as well as the luminal side of the organoids. The subsequent section describes approaches to characterize barrier functions across the epithelial layer of the GI organoids directly or after transferring the epithelial cells to a 2-dimensional culture format in Transwells or compartmentalized microchannel devices. The final section introduces recently developed bioengineered bacterial sensors that sense intestinal inflammation-related small molecules in the lumen using lambda cI/Cro genetic elements or fluorescence as readouts. Considering the small size and cystic shape of GI organoids, sensors used in conventional macroscopic intestinal models are often not suitable, particularly for time-lapse monitoring. Unmet needs for GI organoid analysis provides many opportunities for the development of noninvasive and miniaturized biosensors.
Bioengineered sensors characterize properties of gastrointestinal organoid epithelium, mesenchyme, and bacteria in model systems. These sensors and emerging technologies also can measure biochemicals in fluids within and proximal to the organoid.
The gastrointestinal (GI) tract is a dynamic environment that digests food, absorbs nutrients, and mediates interactions between the host and its microbiota to help maintain health.
On the other hand, it is also the target site for a variety of pathogenic bacteria, viruses, and parasites. Given inherent challenges associated with in vivo human studies, a simplified, human-relevant, GI model system would significantly accelerate mechanistic understanding, GI-focused drug development, and precision medicine efforts.
Fortunately, GI organoids recently have been developed that preserve key human cell physiology while allowing easier manipulation. Although the biological understanding and manipulation of these GI organoids is progressing rapidly, biosensing methods and techniques still are lagging, giving rise to a variety of challenges and opportunities. This review provides an overview of bioengineered sensor or biosensors, which are sensors engineered specifically for GI organoid measurements. The biosensors may be relatively simple adaptations of existing sensors or implementation of new technological advances.
Before the development of human stem cell–derived GI organoids, the most widely used in vitro GI models were 2-dimensional (2D) cultures of intestinal cell lines. A versatile example is the Transwell (Corning Inc, Corning, NY) culture of Caco-2 cells. These workhorse systems are useful in having separated and basolateral compartments that allow exposure of just the apical side to bacteria,
as well as measurement of transepithelial electrical resistance (TEER). More recently, multilayer microfluidic devices with compartmentalized chambers enabled exposure of cells to dynamic mechanostimulation by fluid flow and stretch,
Although useful, easy to image, and with the benefit of decades of sensor developments for 2D cultures, these culture formats have suffered from a lack of physiological cellular composition and maturation. That is, although human stem cell–derived GI organoid cultures
to take advantage of existing biosensing methods (Table 1). In this review, we take the view that it also is important to develop novel sensing methods or use existing biosensors in new ways to allow versatile analysis of 3D GI organoids.
Table 1Bioengineered Sensors Used in In Vitro GI Models
Many methods used to analyze cellular responses in 3D GI organoid models have been adapted from tissue histology and include staining, immunofluorescence, immunoblotting, use of reverse-transcription polymerase chain reaction (RT-PCR), next-generation sequencing,
enzyme-linked immunosorbent assay, and other fluorescent probe-based assays. Commonly, these tools require fixation of the sample, allowing only a snapshot of the organoid at a certain time point. For sensing inside the lumen of cystic organoids, thin and long pin probe-type sensors have been adapted from the microchemistry and microbioreactor field. These are useful but challenging for long-term monitoring because of their invasiveness. Methods also have been adopted from the field of cell microinjection, in which dyes or sensor particles are injected into the organoid lumen and analyzed optically.
These methods are highly applicable but suffer from the need for specialized equipment and technique that reduce throughput. These technological challenges, coupled with rapid biological advances and significant biomedical needs, provide opportunities for productive collaborations between biosensor developers, biologists, and industry.
Capturing Real-Time Chemical Microenvironments
Cells in the GI tract are exposed to widely varying oxygen (O2) environments ranging from normoxic (80 mm Hg O2) to anaerobic (<0.01 mm Hg O2). In the intestines, these different oxygen environments are arranged radially. The intestinal mucosa is well perfused and oxygenated by networks of blood capillaries; however, toward the center of the lumen, the environment is almost devoid of O2, allowing obligate anaerobes to survive and be a part of the diverse gut microbiota.
When and how do these steep gradients develop? One explanation is that the gut lumen of a newborn becomes populated by aerobes and facultative anaerobes that deplete O2 to create a suitable environment for obligate anaerobes.
Despite the importance of O2, accurate spatiotemporal tracking of the in vivo O2 microenvironment is challenging because of its anatomic location.
In vitro culture systems also can create hypoxic environments through external gas environment control or owing to cellular oxygen consumption. A microfluidics-based human-microbial cross-talk (HuMiX) device
Measured O2 levels using an O2-sensitive patch and accompanying optical fibers were 38.7 mm Hg O2 (5.43% O2) in the basal perfusion microchamber, and less than 5.70 mm Hg O2 (0.88% O2) in the luminal microbial microchamber. This hypoxia in the luminal compartment was created by purging dissolved O2 from the luminal fluid with N2, which allowed co-culture of Caco-2 with both facultative and obligate commensal anaerobes.
created intestine-like tissue constructs by co-culturing Caco-2, HT29-MTX cells, and human intestinal myofibroblasts on tubular silk scaffolds. This open-tube structure enabled a probe-type optical O2 sensor to access the lumen without damaging the epithelium. O2 levels along the longitudinal axis were effectively anoxic (<0.1%), even without bacterial colonization. These unexpectedly anoxic conditions in the absence of bacteria
of electrochemical sensors, optical O2 sensors are more reliable at lower-than-ambient O2 concentrations that are most relevant to GI biology. Electrochemical sensors consume O2 and hence can produce significant measurement errors in hypoxic environments such as the intestinal lumen.
This is particularly problematic when making multiple measurements over time because the puncture site creates major inhomogeneities, such as leakage of bacteria or virus from the lumen. A promising emerging technology is the use of dispersible bead-type O2 microsensors,
in which small sensing beads are placed and imaged remotely (Figure 1A).
Both electrochemical and optical mechanisms are used in biological pH sensing. For typical quantitative acidity measurements, pH electrodes that measure electric current from hydrogen ion transport are broadly available. These systems provide easy operation and a wider detection range compared with pH-sensitive fluorophores. For long-term measurement of pH inside organoids, however, their probe-based form factor is too invasive. Moreover, the electrode requires periodic recalibration owing to significant signal drift over time. pH-sensitive fluorescent molecules
have a narrower detection range that require careful consideration in choosing the right sensor molecule for an application. Fluorescein-derived molecules such as fluorescein isothiocyanate (FITC) are pH-sensitive, but they photobleach easily; molecules synthesized more recently have improved photostability, making them a better candidate for time-lapse monitoring of pH changes in cystic organoids.
Seminaphtharhodafluor (SNARF) family pH indicators have been used for cytosolic and luminal pH measurements. These dyes have dual-absorption and dual-emission properties, making them useful for ratiometric measurement of emission intensities using confocal or 2-photon microscopy. SNARF-4F was delivered into the cells of human enteroids to measure cytosolic pH
(Figure 1B). Change in cytosolic pH provides information on the activity of sodium-hydrogen exchanger 3 (NHE3) channels because they transport intracellular H+ from cytosol to the extracellular environment and uptake extracellular Na+. With administration of forskolin, an NHE3 inhibitor, intracellular pH decreased to 6.25.
Use of the more photostable pH-sensor dyes can be challenging, particularly for luminal pH measurements. Because the lumen is enclosed by cell layers and an extracellular matrix, optical interaction can cause single-wavelength, intensity-based sensing to be highly unreliable. Even fluorescence ratiometry can be problematic because optical artifacts can have complex absorption spectra that show different scattering and absorption behavior at different wavelengths. Thus, when the sensors are used in environments that differ substantially from the calibration conditions, the measurement result will not be as reliable.
The metabolic function, such as mitochondrial respiration and glycolysis, of cells and tissue can be quantified in vitro by the combined monitoring of O2 concentration and pH. For example, the Seahorse XF Assay from Agilent Technologies (Santa Clara, CA) measures the O2 consumption rate and the extracellular acidification rate in the culture media in the presence of different substrates, stimulators, and inhibitors.
Most commonly, this machine is used with cells cultured in 2D formats in the bottom of a specially shaped microwell. A sensor probe is lowered to within 200 μm of the cultured cells, reducing the culture media volume to just a few microliters to allow cells to rapidly (within minutes) deplete oxygen or acidify the media. The sensor proximity also allows measurements to be pericellular. Although the system is very useful, there are challenges for use with GI organoids such as the mismatch between organoid size (diameters up to millimeters) and chamber gap size (200 μm) when the sensor probes are lowered. During the measurement cycle, unwanted mechanical stress and damage may be applied to the organoids as the probe comes near the bottom. The system provides only the overall metabolic status of all cells within a well and may miss spatial variations in metabolism that are expected in GI organoids.
Although the Seahorse XF assay requires specialized instruments specific for the assay, other commercial metabolism assays using plate readers are available as well. Alamar Blue Invitrogen (Carlsbad, CA) is a well-established colorimetric/fluorimetric cell viability assay; a nonfluorescent blue molecule is reduced to become a fluorescent red molecule upon entering cells. Some studies adopt this assay to measure the relative metabolism rate; potentially, calibration may enable quantitative analysis of aerobic metabolism.
Combining multiple fluorimetric assay products from Luxcel Biosciences (Little Island, Cork, Ireland) provides similar metabolism analysis to the Seahorse XF assay. For example, the MitoXpress Xtra Oxygen Consumption Assay and the pH-Xtra Glycolysis Assay (Luxcel Biosceinces, Little Island, Cork, Ireland) generates the O2 consumption rate and the extracellular acidification rate, respectively.
Calibration of these products and measurement of dissolved O2 and pH are available by measuring the intensity or lifetime of the sensors. Instead of using a plate reader, adopting advanced optical microscopy such as phosphorescence lifetime imaging microscopy could enhance spatial resolution.
In studies involving GI organoids, quantitative RT-PCR has been a useful tool for cellular messenger RNA amplification and detection.
For studying viral infection and recombination, however, the usual RT-PCR technique can provide incorrect readouts and decrease the sensitivity of the technique. Norovirus in particular, an RNA virus causing acute gastroenteritis with fast transmission between hosts, has a high mutation/recombination rate.
can circumvent the chimera formation by encapsulating each RNA strand in picoliter droplets and enable high-throughput analysis. In addition, the droplet can isolate the host–virus interaction to a small number of host cells and virus by co-encapsulating them in a single droplet. After incubation, cells are lysed by heat shock, reagents are added to the droplet, and in-drop RT-PCR is performed
microfluidic PCR may be useful for elucidating in vivo infection dynamics of human norovirus, and predicting viral evolution for the prevention and treatment of the disease.
TEER is a common quantitative method to measure the integrity of tight junctions in monolayers of epithelial and endothelial cells. TEER measurements are a strong measure of the integrity of a cell monolayer’s ability to act as a barrier and therefore has been used in conjunction with studies on the transport of drugs, chemicals, dyes, and general membrane leakage. Benefits of TEER measurements are that it is in real time, is nondestructive, often noninvasive, and allows cell cultures to be re-used for additional studies.
In general, TEER measurements are performed by growing a monolayer of cells on a semipermeable membrane with electrodes placed on each side of the membrane while in solution and then applying an AC signal to the electrodes. The current and voltage across the cell monolayer then is measured from this signal and Ohm’s law is used to calculate the electrical resistance of the system. Ideally, this resistance is composed of both the transcellular resistance and the paracellular resistance in parallel. However, along with these resistances, there are nonideal factors that need to be considered such as the resistance of the media, electrode-to-media interfacial resistance, and the support membrane resistance.
It is most common to perform TEER measurements using a Transwell insert with a semipermeable membrane with a monolayer of cells grown on it along with chopstick TEER electrodes (Figure 1D). These 2 electrodes are composed of silver and silver/silver chloride and the typical frequency of the AC electrical signal is 12.5 Hz with a current of 10 μA.
In recent years there has been great progress in growing intestinal cell monolayers. This ability to design an experimental system to model intestinal epithelial cells as a confluent polarized monolayer has been an important milestone in understanding intestinal pathology. One major difficulty of studying primary intestinal epithelial cells in Transwells with monolayer culture and TEER measurements is that it requires a large amount of viable cells.
This lack of cells has been overcome by generating intestinal cells from stem cell–derived intestinal organoid cultures. Adult stem cell–derived organoid cultures are made from dissociated primary crypts/stem cells whereas pluripotent stem cell–derived organoid cultures are made from embryonic stem cells or induced pluripotent stem cells. Both of these cultures use media containing canonical Wnt ligand, R-spondin, and Noggin to support intestinal epithelial stem cell growth. Once enough intestinal cells from these 3D spheroid/organoid cultures are generated, 2D intestinal epithelial monolayers can be created on Transwell membranes for assays. Before such intestinal epithelial cells were capable of being reproducibly grown in 2D cultures on Transwells, it was common to use immortalized cell lines such as Caco-2 or T84, which can show microvilli formation. Monolayers of these cell lines typically provide TEER values (>300 Ωcm2). This is higher compared with TEER of human small intestinal tissue (∼40 Ωcm2) measured from freshly obtained surgical samples using an Ussing chamber to immobilize the tissue.
developed 2D monolayers of intestinal epithelial cells derived from spheroids (enteroids) on semipermeable Transwell membranes. These platforms showed TEER values of approximately 395 Ωcm2 and approximately 400 Ωcm2 for rectal and ileal monolayers, respectively.
A similar study created monolayers of intestinal epithelial cells from intestinal organoids but on a natural extracellular matrix based on porcine small intestinal scaffold instead of the standard Transwell semipermeable membranes.
The values of these intestinal epithelial monolayers were approximately 40 Ωcm2, which is closely matched to the native human small intestine. Fibroblast co-cultures did not alter TEER readings, but did provide a more heterogeneous monolayer with prismatic cells and luminal cystic structures in the epithelium as shown by hematoxylin and eosin (H&E) staining. To further investigate the barrier function of the 2 different monolayer scenarios, FITC-dextran (FITC-Dex) permeability studies were performed. The FITC-Dex permeability studies showed that the fibroblasts created a more robust and consistent membrane with a permeability of approximately 1% compared with approximately 4% for the monolayer without fibroblasts, a difference that was not discernable by TEER measurements.
There also has been research on the co-culture of intestinal epithelial monolayers with human monocyte–derived macrophages to investigate the importance of the interaction of the intestinal epithelium with the mucosal immune system. Here, it was found that the presence of monocyte-derived macrophages with intestinal epithelial cells derived from differentiated enteroids increased TEER and barrier function from approximately 800 Ωcm2 to approximately 1000 Ωcm2,
suggesting a potential role of the macrophages in enhancing maturation of the intestinal epithelium and thickening the physical barrier.
TEER measurements and dye flux assays, such as the FITC-DEX assay described earlier, frequently are performed together to provide a thorough characterization of the barrier function of cell monolayers. Similar to TEER measurements, the traditional experimental configuration for dye flux assays uses a monolayer of cells grown on the semipermeable membrane of a Transwell dish. After a confluent monolayer of cells is grown on the Transwell membrane, a dye of a specific concentration and molecular size is placed in the liquid media in the top chamber of the Transwell dish. Then the solution present in the bottom chamber is sampled at various time intervals for the concentration of the dye.
By using FITC-Dex with a molecular weight of 150 kilodaltons, 1 study showed good barrier function of commercially available primary human intestinal epithelial cells from various donors, induced pluripotent stem cell–derived intestinal cells, as well as Caco-2 cells, which are commonly used for human intestinal models.
there is a growing need to convert cell monolayer-based permeability and TEER assays into a format that is directly compatible with organoids. Although there are no published works on direct TEER measurement of organoids, there have been permeability/leakage studies of human intestinal organoids (HIOs).
In this work, the barrier function of the HIOs was evaluated by microinjecting FITC-Dex with an average molecular mass of 4 kilodaltons into the lumen of the HIOs. The HIOs retained 58.8% of their fluorescent intensity over 18 hours in control cultures whereas exposure to ethylene glycol-bis(β-aminoethyl ether)-N,N,N′,N′-tetraacetic acid, a calcium chelator that disrupts tight junctions, caused a 70% loss of fluorescence intensity in just 6 hours. An additional variation of this test was performed in which 4 kilodaltons were added to the surrounding media of the HIOs to investigate barrier leakage in the basal-to-apical direction (opposite of the microinjection method). The 2 complementary experiments show that HIOs can have good paracellular epithelial barriers.
Bioengineered Bacteria for Proinflammatory Small-Molecule Sensing
During gut inflammation, increased reactive oxygen species oxidize thiosulphate (S2O32-) to produce tetrathionate (S4O62-). Thus, S4O62- is of interest as a potential marker of GI tract disease if there were convenient ways to sense this molecule. Compared with other bacteria, pathogenic Salmonella are able to uniquely thrive under conditions of high S4O62- by using S4O62- as a respiratory electron acceptor to fuel their outgrowth.
inserted S2O32-- and S4O62--sensing TCRSs from marine Shewanella species to E coli. To avoid cross-repression of S4O62- TCRS in Salmonella typhimurium by O2 and nitrate, they identified TCRS homologous to that of S typhimurium using bioinformatics. With a fluorescent reporter gene inserted, both bacterial sensors expressed fluorescence in response to the respective target species in vitro. Although the response time for these sensors is approximately an hour,
the approaches used in this work, fluorescent-sensing modality and use of bioinformatics for homolog TCRS identification, may allow the sensors to be incorporated into organoid systems for real-time responses in the future.
Another interesting small molecule is nitric oxide. Under inflammatory conditions, NO is produced in large amounts that can reach micromolar levels. Interestingly, NO has many beneficial effects at the picomolar to nanomolar range, such as increasing tissue perfusion, vasodilation, and protecting tissue against acute inflammation.
These sensor bacteria also may be useful for in vivo and GI organoid NO sensing.
Microperfusion and Integrated Biosensors for Real-Time Organoid Monitoring
Perfusion can benefit organoid cultures through application of fluid mechanical stress, supply of nutrients, removal of waste, or dynamic application of different culture media or drugs. In one example, Jin et al
loaded mouse intestinal organoids in a microfluidic device to create an array of organoids to facilitate organoid volume change measurements. The investigators re-created secretory diarrhea by adding cholera toxin to the media, which induced swelling of the organoids. This microfluidic array potentially could be used to perform multiplex drug screening. Although there are increasing numbers of microfluidic intestine culture systems, most are used for 2D cultures and there are still few examples of microperfusion of intact, 3D organoids. It already has been shown that stimulating the apical side of the intestinal epithelium in 2D fluidic cultures induced intact and differentiated intestinal epithelium with in vivo recapitulating microstructural changes.
analyzed cell signaling molecules such as interleukin 6, interleukin 8, vascular endothelial growth factor, and β-defensin that have been secreted to the basal side of the bacteria-colonized HIOs using enzyme-linked immunosorbent assays. Also, in the same work, the investigators measured bacterial translocation across HIOs by microinjecting E coli in the lumen and counting the number of bacteria in the external media over time.
These time-lapse microscale media analytics still are limited to external culture media. Adding a microperfusion component to 3D intestinal organoids will allow access to the media in the luminal compartment as well. The increasing interest and need for microperfusion coupled with a lack of solutions make this an important area of opportunity.
Dynamic environments created by microperfusion and real-time monitoring go hand-in-hand. An important consideration for future development include decreasing the cost and the development of perfusion systems and biosensors that can go directly inside the incubator where temperature and humidity is much better controlled than on a microscope stage. One recent example is a miniaturized microscope constructed from an inexpensive webcam that can be placed inside an incubator to provide real-time images of microscale tissue dynamics.
The dynamic environment in the GI tract is associated with various physiologic responses within the human body. Advances in stem cell–derived GI organoid cultures enabled human-specific studies, complementing animal models and conventional in vitro models. The development of engineered biosensors appropriate for cystic organoids still is lagging, and spatiotemporal changes in organoids and their microenvironment are difficult to monitor. Here, we reviewed sensors that had been incorporated into GI organoid studies, targeting pericellular O2, cytosolic and luminal pH, metabolism, and transepithelial barrier function. In addition, we discussed emerging biosensing techniques that may add value to understanding dynamic host–microbiota–pathogen interactions in GI organoids. Scaling down and developing new form factors of existing biosensors are crucial. Because the size of GI organoids is typically in the range of a few hundred microns to a few millimeters, adaptation of existing micro/nanoscale technologies, for example, microfluidics, to GI organoid cultures will create synergistic effects. For broad use of these capabilities, cost reduction, miniaturization, and engineering sensors that can be used inside cell culture incubators also will be important to consider.
Effects of the gut microbiota on obesity and glucose homeostasis.
Author contributions Ge-Ah Kim and Shuichi Takayama were responsible for preparing the outline, major drafting of the manuscript, and editing; and Nicholas J. Ginga was responsible for drafting and editing the section on barrier function.
Conflicts of interest The authors disclose no conflicts.
Funding This work was supported by U19AI116482 from the National Institutes of Health.