Close this search box.

A miniaturised semi-dynamic in-vitro model of human digestion – Scientific Reports

Device design and fabrication

A technical diagram of the digestion-chip can be found in Fig. 1a, together with a description of the different elements integrated. The device comprises three circular compartments (17 mm diameter, 20 mm height) that are two incubation chambers and one reservoir: (1) chamber to emulate the oral and gastric phase, (2) chamber to emulate the intestinal phase and (3) reservoir for simulated intestinal fluids that will be added along the digestion process. As can be seen in Fig. 1, both gastric chamber and reservoir of simulated intestinal fluids are connected to the intestinal chamber by two peristaltic micro-pumps that pump simultaneously, at the same flow rate (1:1 ratio), “chyme” and the simulated intestinal fluids from the gastric chamber and the reservoir, respectively, into the intestinal chamber where intestinal phase is emulated. These devices were fabricated from PMMA sheets using CNC micro-milling (Flexicam Viper 606 with ArtCam software) and laser cutting (Widlaser LS1390 Plus with LaserWorkV6 software). The devices were assembled by bonding several layers using double-sided 3 M tape (467MP; 3 M, Saint Paul, MN, US) applying continuous pressure overnight inside the oven at 65 °C to avoid the formation of bubbles.

Figure 1
figure 1

Technical description of the digestion-chip. (a) Diagram of the digestion chambers with the integrated components. (b) Description of the different layers that constitute the device. (c) Picture of the pH and temperature sensors and the digestion-chip.

As shown in Fig. 1b, the device contains 6 layers: (i) 4 layers (5 mm thickness) that define the height of the digestion chambers and reservoir, (ii) 1 layer (3 mm thickness) that is the bottom layer of chambers and reservoir with grooves to accommodate temperature sensors and heating elements and (iii) 1 bottom layer (5 mm thickness) to incorporate the peristaltic micro-pumps that allow gastric emptying. The bottom PMMA layer contains 2 parallel channels (21.14 mm3 each channel) along with 3 chambers (50.27 mm3 each chamber) that constitute a double synchronous peristaltic pump that connect the reaction chambers and that were fabricated by micromilling using 500-µm (11,000 RPM) and 1.5 mm (6000 RPM) diameter mills (vhf camfactor AG, Lettenstraße 10, 72,119 Ammerbuch, Germany). A rectangular cavity allows to place a PDMS piece bonded to a PDMS membrane fabricated by spin coating PDMS (1:10 w/w) at 1000 rpm for 1 min on top of square glass slide. This PDMS slab contains chambers that match the chambers of the peristaltic pump. Thus, the 6 chambers can be pressurised (using a Fluigents MFCS™-EZ Pressure controller), bending the membrane, displacing the fluid between chambers. The total internal volume of each parallel peristaltic pump is ~ 193 µL. The bottom layer is not bonded but rather screwed to the PMMA layers. This fully seals the PDMS piece between the middle and bottom layers of PMMA and avoids leakage through the holes connecting the digestion chambers, reservoir, and the micro-pump channels. An image of the assembled device with all components integrated can be seen in Fig. 1c. The pump is capable of driving up to 170 µL· min−1 and allows a precise control of the gastric emptying volume with approximately 50 µL per pumping cycle. A depiction of the pumping cycles is provided in Fig. S3.

To ensure a constant mixing of the sample and the simulated digestive fluids, 2 × 5 mm magnetic stirring bars were placed at the bottom of the digestion chambers. The devices were then placed on top of a custom-made magnetic stirrer fabricated out of two pulse width modulation (PWM) controllable motors (recycled from computer cooling fans) placed below the digestion chambers. This provides efficient mixing without causing enzyme degradation by mechanical stress. Mixing is critical for these experiments, not only for enzyme–substrate reactions, but also to prevent localised fluid overheating near the heating elements (which could cause irreversible enzyme denaturation) and ensure fast pH equilibration. A picture of the custom magnetic stirrer and stirring bars can be found in the Electronic Supplementary Information (ESI) (Fig. S3).

Finally, the 3 circular compartments are closed using PDMS lids to avoid evaporation (see Fig. 1c), that were punched with Biopsy punches (Kai-Europe GmbH, Germany) to hold the pH electrodes and tubing for the pH adjustment and the gradual addition of simulated gastric fluids and enzymes. These lids are easy to remove or place back and thus enable easy sampling from the digestion chambers at any time using a standard micropipette.

Temperature control

Digital temperature sensors 2 × 2 mm (MAX30205 Human Body Temperature Sensor, Maxim IntegratedTM, San José, California, USA) were integrated at the bottom of the gastric and intestinal chambers (Fig. 1c). The sensors have high accuracy (± 0.1 °C) in the human body temperature range and have a small footprint. The temperature readings are used to modulate the power delivered to the heating elements, more precisely insulated Nichrome (NiCr) wires having high resistivity, via a proportional integral derivative (PID) control loop to maintain temperature constant at 37 °C. The PID control algorithm was used following the standard method. The error in the temperature was set as the difference from the current temperature and the set point. The proportional term was calculated proportionally to this error. The integral term was calculated considering the change in temperature achieved during this last step. The derivative term was calculated taking into account the time evolution of the temperature, i.e. the change in the error in temperature over time. The constants used to calculate and equilibrate these terms were calibrated experimentally for a fast and precise response while avoiding overshooting in temperature to avoid enzyme degradation.

pH sensing and control

For pH sensing, potentiometric Iridium oxide (IrOx) sensors were used. The electrodes were fabricated following the anodic deposition approach reported by Yamanaka et al.26, which yields hydrated IrOx films with high (super-Nernstian) pH sensitivity27. The metal oxide sensors have number of attractive characteristics for integration into microsystems, such as their simple construction and ease of miniaturization, high mechanical and chemical stability and low cross-sensitivity to various salts28. In our system the sensors were fabricated by electrochemical deposition onto a titanium wire substrate. IrOx sensors are also biocompatible and have even been shown in vivo applications29.

As reference electrode a simple single-junction Ag/AgCl reference electrodes encapsulated with 3 M KCl solutions in micropipette tips were fabricated using an approach adapted from Barlag et al.30. In our system, the Ca2+-Alginate gel on the reference electrodes acts as the salt bridge that enables a connection between the reference electrode and the samples. A detailed protocol for the electrode fabrication is provided in ESI.

The electric potential difference between the pH electrodes is registered by an electronic circuit based on the ESP32 microcontroller. The real-time pH readings are used to control the pumping of 1 M HCl or 1 M NaOH from two inexpensive, custom-made programmable syringe pumps, to adjust the pH to the values set for each phase of digestion. The custom-made syringe pumps were adapted from previous work by Martin Fischlechner (Projects | DropletKitchen—

A Printed Circuit Board (PCB) was designed to interface the syringe pumps, temperature sensors, heating elements and mixer motors with an Arduino Uno microcontroller and other extra electronic components required for the automation. A conceptual schematic of the system can be found in ESI (Fig. S4).

In-vitro digestion

Detailed information on material references and suppliers can be found in ESI. A commercial casein labelled with a red fluorescent dye BODIPY TR-X (EnzChek™ Protease Substrate, 589/617 nm)31 and lipase substrate labelled with a green fluorescent dye BODIPY-C12 (EnzChek™ Lipase Substrate, 482/515 nm)32 were used to measure the digestion rate and study the enzyme kinetics during in-vitro digestions. These reporter molecules are quenched by proximity, but become fluorescent after exposure to digestive enzymes.

The simulated digestive fluids were prepared following the INFOGEST protocol (see ESI—Table S1 for detailed recipes) and stored at − 20 °C before use. Enzyme activity of pepsin and pancreatin was determined according to the procedure described by Brodkorb et al.10. Samples of 100 µL in the static digestions and 20 µL in the semi-dynamic digestions were collected throughout the gastric and intestinal phases and enzymes were inactivated by adding Pefabloc® (5 mM final concentration in the digesta) to inhibit protease activity and Orlistat® (1 mM final concentration) to inhibit lipase activity. The digestion rate was quantified by measuring the fluorescence of the test molecules using a BioTeK® Synergy H1 microplate reader (Winnoski, VT, USA).

INFOGEST static protocol

In vitro static digestions were carried out in a thermomixer (Eppendorf™, Hamburg, Germany) or inside a single compartment of the digestion-chip. Briefly, 1 mL sample (protease and lipase substrate) was mixed 1:1 with simulated salivary fluid (SSF) and incubated for 2 min at 37 °C. Following this, the simulated gastric fluid (SGF) and pepsin (final concentration 2.000 U/mL), were mixed 1:1 with the bolus, the pH adjusted to 3.0 (using 1 M HCl) and allowed to digest for 2 h. Then, the simulated intestinal fluid (SIF) containing bile salts (10 mM) and pancreatin (final concentration of trypsin 100 U/mL) were added to the end-point of the gastric phase (chyme), the pH was increased to 7.0 (using 1 M NaOH) and the chyme was incubated for further 2 h.

Semi-dynamic protocol

The digestion-chip aims to faithfully emulate the in-vivo conditions by incorporating some dynamic features, in the gastric phase, such as gradual acidification, addition of enzymes and simulated gastric fluid, and performing different gastric emptying, as shown in Fig. 2. The semi-dynamic protocol followed here was adapted from Mulet-Cabero et al.18 and is represented in Fig. 6. Glucose (100 mg mL−1) was added to the enzyme substrate to increase the overall caloric content of the sample as the total incubation time in the gastric phase is calculated as a function of the caloric content of the sample, according to the spreadsheet provided by Mulet-Cabero et al., leading to 80 min of total gastric incubation time. Briefly, 100 mg·mL−1 glucose was added to 500 µL of substrate (protease or lipase substrates) and mixed 1:1 with SSF and then incubated for 2 min at 37 °C, in the gastric chamber. The gastric phase initiates with just 10% of the total amount of gastric secretions to be added owing to the fasting levels. Thus, 10% of simulated gastric fluids (100 µL of SGF + pepsin) was added to the oral phase in the gastric chamber and the remaining 90% (900 µL of SGF + pepsin) was gradually added using a New Era NE-2000 syringe pump (New Era Pump Systems, Inc., NY, USA) at a constant flow rate of 11.25 µL·min−1, achieving a 1:1 volume ratio with respect to the oral phase at the end of the gastric phase (80 min). The pH was gradually acidified from ≈ 7 to ≈ 2 at the end. This was achieved by a constant communication between the pumping of HCl (1 M) using a custom-made syringe pump and the readings from the pH sensors. The gradual addition of HCl contributes to an increasing enzyme activity over the gastric phase.

Figure 2
figure 2

Conceptual diagram of the semi-dynamic digestion process in the digestion-chip. The graphs describe the evolution of sample volumes and of the concentration of substrate, reaction products and enzymes.

During the incubation time in the gastric phase, sequential emptying into the intestinal chamber were performed. The number of gastric emptying is defined by the user, and in this case 4 were done, every 20 min. The gastric emptying was performed through sequential pressurisations of the chambers from the peristaltic micro-pump using the pressure controller. Chyme from the gastric chamber is mixed 1:1 with simulated intestinal fluids (SIF + bile salts + pancreatin). Here, 250 µL of chyme was mixed with 250 µL of simulated intestinal fluids after each emptying and incubated in the intestinal chamber for 2 h. Following each emptying the pH was adjusted to 7.0 adding NaOH (1 M).