Study on drug absorption and toxicity process in intestinal hepatic organoid serial co culture

In order to compensate for the limitations of animal models, a new drug safety assessment model has been proposed to * and reduce existing models. In order to simulate drug absorption and metabolism in the gut liver microphysiological system (MPS) in vitro, and predict pharmacokinetics and toxic effects, scientists from the Institute of Safety Evaluation of the Chinese Academy of Food and Drug Control (National Center for Drug Safety Evaluation and Monitoring), the Chinese Academy of Medical Sciences&Peking Union Medical College, and TissUse GmbH in Germany collaborated to establish a gut liver tandem culture chip to detect the acute liver injury process after APAP (acetaminophen) overdose,And it was published in the journal Food and Chemical Toxicology in September 2024.

Intestinal organoids were established using Caco-2 and HT29-MTX-E12 cell lines, while liver organoids were established using HepG2, HUVEC-T1, PMA induced THP-1, and human hepatic stellate cells. Determine APAP concentration using high-performance liquid chromatography and fit pharmacokinetic parameters using Phoenix software through non compartmental analysis. The changes in liver injury biomarkers aspartate transaminase and alanine transaminase, as well as liver function marker albumin, indicate that these two organ chip models are stable in short-term culture within 4 days. After administration of acetaminophen (APAP), the reactive oxygen species signal increased, while the mitochondrial membrane potential decreased. Caspase-3 (caspase-3) was activated, and the p53 signal increased, indicating that excessive APAP caused toxic reactions. In the gut liver multi organ system model, we fitted toxicokinetic parameters and simulated the liver toxicity process after APAP overdose, which will contribute to the application of organ chips in drug toxicity detection.

1. Introduction

Drug induced liver injury (DILI) is a major obstacle to drug development and can lead to drug withdrawal from the market (2021; 2016). The differences between animal models and humans, as well as increasingly stringent ethical requirements, make it necessary to adopt new drug safety assessment models (2014). This microphysiological system (MPS), also known as "organ on chip," is a device constructed by combining microfluidics, microfabrication, and three-dimensional (3D) cell culture. It can reproduce physiological organ functions for animal models. The MPS platform is a miniaturized in vitro physiological environment (Shinohara et al., 2021).

This enables the simulation and precise control of chemical gradients and biomechanical forces to simulate the in vivo environment and respond. Pre defined microfluidic channels act as engineered blood vessels, reproducing physiological tissue functions in vivo (Low et al., 2020), and allowing for integration with other multi organ systems (Kulthong et al., 2020). In addition, multi organ systems can achieve real-time monitoring of tissue function (Milani et al., 2022).

The multi organ chip model connected by microchannels can further simulate human dynamic responses and internal organ interactions (Arakawa et al., 2020), and provide auxiliary evidence for the relationship between drug exposure and drug efficacy/toxicity. The intestine and liver are the main organs for absorption, distribution, metabolism, and excretion (ADME), making them crucial in pharmacokinetic (PK) studies (Vernetti et al., 2017).

Oral medication is directly absorbed into the bloodstream through the digestive tract and intestines, affecting the bioavailability and systemic dose of the medication. The intestinal chip model can simulate drug absorption and depict the ADME and blood drug concentration of drugs (Milani et al., 2022). Lee et al. (2017) established a three-dimensional gut liver chip to depict the pharmacokinetic model of acetaminophen (APAP), while Arakawa et al. (2020) constructed an gut liver model for in vitro quantitative estimation of continuous metabolism of SAN zolam. Milani et al. (2022) evaluated the amount of drug precursor mycophenolate mofetil produced due to intestinal and hepatic metabolism of mycophenolate mofetil. Therefore, similar methods can be applied to other multi organ MPS to study the pharmacological processes in the human body.

Currently, only a few examples of multi organ chip systems that simultaneously perform toxicity testing and sample identification have been reported. Liu et al. (2020) studied the absorption, metabolism, and toxicity of ginsenoside complex K using a multi organ chip system based on gut, vascular, liver, and kidney chips. Jeon et al. (2021) reported a gut liver chip that reproduced the absorption of fatty acids and subsequent lipid accumulation in the liver.

In order to explore the microphysiological system of gut liver in vitro, we developed a reproducible gut liver dual organ chip by concatenating gut cultures with 3D liver spheres constructed using four cell lines on a dual organ chip (2OC). We analyzed the function of the chip by detecting the levels of functional biomarkers and proteins, and then tested liver toxicity by adding the hepatotoxic compound acetaminophen (APAP) to the intestinal chamber in the chip to simulate liver toxicity after drug absorption. After 48 hours of incubation, drug distribution and liver toxicity signals were detected in two chambers.

2. Materials and Methods

2.1. Cell and culture cell resources and culture methods can be found in the supplementary information.

2.2. Cultivation of intestines

To construct an intestinal model on Transwell plates, Caco-2 cells were seeded into Transwell plates (Merck, PIHP01250) at a density of 1 × 10-6 cells/well. Mix Caco-2 and HT29-MTX-E12 cells in a 9:1 ratio and inoculate them into each culture plate at the same concentration. From day 0 to day 7, change the culture medium every two days, and after day 7, change it every day.

2.3. Cultivation of 3D (liver)

Based on our previous study (Sun et al., 2024), THP-1 cells were diluted to 5 × 10-5 cells/mL in RPMI 1640 medium and then treated with 25 nanomolar phorboloyl acetic acid (PMA, MedChemExpress) for 48 hours in a 6-well plate. Subsequently, THP-1 cells were cultured in the medium for 24 hours and screened using TryplE (Gibco). THP-1 cells induced by PMA were collected. 2 HepG2、 Human umbilical vein endothelial cells T1, THP-1, and human hepatic stellate cells (HHSC) were digested and collected, and the cell suspension was mixed in a ratio of 60:19:15:6. In order to form equal amounts of cells, 100 cells were seeded into a round bottomed U-shaped low adhesion 96 well plate (LV-ULA002-96UW, Shenzhen Liver Biotechnology Co., Ltd.) after dilution. Replace half of the culture medium every two days. The spherical morphology was observed and measured using a high connotation cell imaging analyzer (PerkinElmer, Opretta series).

2.4. Chip based intestinal liver co culture

HUMIMIC Chip2 24 well plate (TissUse GmbH, Germany) was prepared according to the recommended protocol (Wagner et al., 2013). On the 14th day after establishing the intestinal model, Transwell cell chimeras were added to a 24 well culture chamber, and 50 spheres were added to a 96 well culture chamber to form the intestinal liver chip model.

According to the previous method (Lin et al., 2020), co culture was performed in a 2OC microchannel cycle using 600 ul of circulating medium and 400 ul of medium located at the top of the barrier.

The culture medium in the liver chamber is changed once a day, and half of the circulating supernatant is collected and replaced. At the end of the 7-day co culture, liver organ specific functional markers were analyzed using immunofluorescence. Set the on-chip microfluidic pump to a frequency of 0.8 Hz.

2.5. Integrity of intestinal organs

Measure the transendothelial resistance (TEER) and intestinal apparent permeability coefficient (Papp) of fluorescein sodium and 40 kDa fluorescein isothiocyanate dextran at different time points to evaluate the integrity of intestinal organoids. See supplementary information for the method.

2.6. Performance evaluation of liver organoids

In the liver model, a sensitive dual color fluorescence cell viability assay was chosen to distinguish between live and dead cells. Calcium yellow green fluorescent AM (a cell permeable dye) was used as a live cell indicator, and iodized BOBO-3 (Invitrogen) was used as a dead cell indicator. The components of living cells produce strong and uniform green fluorescence (excitation/emission 488 nanometers/515 nanometers) in living cells, while the components of dead cells mainly produce nuclear red fluorescence (excitation/emission 570 nanometers/602 nanometers). Various biomarkers were measured in the supernatant obtained using the established liver organ chip model. We evaluated the levels of albumin (ALB), urea (UREA), aspartate aminotransferase (AST), and alanine aminotransferase (ALT), which are indicators of liver metabolism and injury. According to the manufacturer's instructions, the content of adenosine triphosphate (ATP) was measured using Cell Titor Glo 3D (Promega). Cholecyl lysine fluorescein (CLF) is a fluorescein labeled bile acid, and its biological behavior is very similar to that of natural bile glycine. To visualize bile ducts in a 3D liver model, we prepared 20 micromolar CLF stock solution and incubated the model at 37 ° C for 2 hours. Stain with Hoechst 33342 and incubate for 10 minutes. Cells were washed three times with phenol red free culture medium and fluorescence signals were detected at excitation/emission wavelengths of 498/517 nanometers.

2.7. Morphology and Immunostaining

As mentioned earlier (Sun et al., 2024), equal sections of intestine and liver were fixed with paraformaldehyde and dehydrated with sucrose solution. Frozen sections are cut using hematoxylin eosin (HE) staining, periodic acid Schiff (PAS) staining, or immunostaining before staining. For immunostaining, the slices were first washed three times with PBS and treated with PBS containing 0.5% Triton X-100 (Salbo) for 10 minutes. The cells were then treated with perforation and blocked with goat serum (Beyotime Biotechnology) for 30 minutes at room temperature. Add an antigen ligand to each well at a certain dilution according to the instructions and incubate overnight at 4 ℃. Use the following antigenic epitopes: anti-Occludin (91131S, CST, 1:400 dilution), anti-ZO-1 (ab221547, Abcam, 1:100 dilution), anti-MRP2 (ab172630, Abcam, 1:500 dilution), and anti-PGP (item number 25081-2-AP, ProteinTech, 1:500 dilution). Wash the cells, treat them with the corresponding fluorescent conjugated antigen site, and co localize them with DAPI (Beyotime Biotechnology). Observe the clear field morphology and sections using a high content cell imaging analyzer or microscope (Olympus IX71). For in situ immunofluorescence staining, the intestinal membrane and 3D liver model were washed three times with PBS and fixed with 4% paraformaldehyde for 30 minutes. Next, the samples were punctured in PBS containing 0.5% Triton X-100 (Salbo) for 10 minutes and blocked with goat serum (Beyotime Biotechnology) at room temperature for 30 minutes. According to the instructions, add the diluted in situ antibody to each well at a certain dilution and incubate overnight at 4 ℃. The in situ antibodies used include: anti-Occludin (91131S, CST, 1:400 dilution), anti-ZO-1 (sc-33725, Santa Cruz Biotechnology, 1:500 dilution), anti-MRP2 (ab172630, Abcam, 1:500 dilution), anti-ABCB11/BSEP (ab255605, Abcam, 1:100 dilution), anti-PGP (Cat No. 25081-2-AP, ProteinTech, 1:500 dilution), anti-CYP3A4 (MA5-17064, Thermo Fisher, 1:1000 dilution), anti-Ki67 (9449S, CST, 1:10000 dilution), anti-CD31 (3528S, CST, 1:100 dilution) Anti CD68 (ab213363, Abcam, 1:100 dilution), anti SMA (14395-1-AP, ProteinTech, 1:100 dilution), anti lytic caspase-3 (9664S, CST, 1:400 dilution), and anti p53 (2524S, CST, 1:2000 dilution). The cells were washed, incubated with fluorescent conjugated secondary antibodies, and co localized with DAPI (Beyotime Biotechnology). The secondary antibodies used include goat anti rabbit IgG H&L (Alexa Fluor) ® 488) (ab150077, Abcam, 1:1000 dilution), goat anti mouse IgG (H+L) (Alexa Fluor) ® 555 Conjugate (4417, CST, 1:1000 dilution) and goat anti mouse IgG H&L (Alexa Fluor) ® 488) (ab150113, Abcam, diluted 1:1000). The results were evaluated through high content analysis and fluorescence microscopy.

2.8. Determination of drug toxicity and cell viability

As mentioned earlier (Sun et al., 2024), Cell Titor Glo was used to detect cell viability and measure cell survival rate and cell survival rate (%). Inoculate cells into a 96 well plate at a density of 1 × 10-4 cells per well; The next day, discard the supernatant and culture it with acetaminophen (4 mmol). Use the same method to determine the survival rate of the sphere. This test is conducted using a photometer.

2.9. Simulated oral administration of aspirin in co cultured chips

Seven 2OC models were loaded, of which four microchannels were treated with acetaminophen (APAP), while the remaining circuits were not treated. To simulate the oral process and APAP induced liver toxicity, 400 μ L of 10 mmol APAP was added to the top of the insert on the second day after loading the 2OC circuit. Collect 10 microliters of supernatant at 0, 0.5, 1, 2, 4, 8, 24, 32, and 48 hours for liquid chromatography quantitative analysis. After 48 hours of administration, collect the supernatant for measurement of various biomarkers to evaluate transdermal resistance (TEER). Simultaneously collect liver spheroids for liver toxicity assessment. We also collected at least 3-5 samples for cell viability assay, and transferred the rest to a 96 well plate for preservation for fluorescence staining.

2.10. Liver toxicity assessment of co cultured chips

According to the recommended reagent kit protocol, measure cell viability and levels of albumin (ALB), aspartate aminotransferase (AST), and alanine aminotransferase (ALT) to evaluate liver damage 48 hours after drug administration. Measure mitochondrial membrane potential using a mitochondrial membrane potential detection kit (Beyotime Biotechnology, C2001S). Transfer the 3D embedded liver model of the 2OC chip to a 96 well plate. Fluorescence signals of tetramethylrhodamine ethyl ester (TMRE) and Hoechst 33342 were detected at Ex/Im=550/575 and 350/461 nanometers, respectively. Measure oxidative stress using CellROX oxidative stress reagent (Thermo Fisher Scientific, C10444). Transfer the 3D embedded liver model of the 2OC chip to a 96 well plate. Detect the fluorescence signals of CellROX and Hoechst 33342 at 485/520 nanometers and 350/461 nanometers, respectively. Use immunofluorescence staining to detect in situ Ki-67, cleaved caspase-3, and p53 signals to evaluate cell proliferation and apoptosis.

2.11. Sample quantification

Based on the previous method (Marin et al., 2019), the concentration of acetaminophen (APAP) was measured using high-performance liquid chromatography (HPLC) (Shimadzu, LC-2040C 3D). The basal chamber (liver chamber) is filled with fresh culture medium, while the parietal chamber is exposed to acetaminophen culture medium. At each time point, as described in Section 2.9, take 10 microliters of sample from each side. Select methanol as the protein precipitant, add 30 microliters of methanol to precipitate the protein, and remove the precipitate by centrifugation (13000 rpm, 4 ℃, 15 minutes). Use an inert sustainable C18 (150 mm length, inner diameter 4.6 mm, particle size 5 microns, Shimadzu) chromatographic column. Use two mobile phases: solvent A (methanol) and solvent B (distilled water), with a flow rate of 0.8 mL/min. The solvent composition is 5% A at 0-4 minutes, 5-100% A at 4.01-10 minutes, and 5% A at 10.01-15 minutes. The injection volume is 10 microliters, and detection is performed using a photodiode array (PDA) detector (wavelength 280 nanometers). Measure the concentration of acetaminophen using a standard curve. The pharmacokinetic parameters were fitted using a non compartmental model (NCA) in Phoenix WinNonlin (Certara, USA). The linear trapezoidal linear interpolation method is used as the calculation method, with the model type defined as plasma and the dose option defined as extravascular. Fit the concentration time data of acetaminophen (APAP) in the liver compartment determined by high-performance liquid chromatography to determine the area under the curve (AUC) from 0 to 48 hours. The peak concentration (Cmax) and time to peak (Tmax) are directly read from the individual concentration time curve (Gao et al., 2022; Wang et al., 2022). The signal of NAPQI was detected by ultra-high performance liquid chromatography tandem mass spectrometry (UPLC-MS) (Zhang et al., 2018). Prepare a stock solution with a concentration of 1 mg/mL by dissolving N-acetyl hydroquinone imine (NAPQI) (MCE, HY-66005) in methanol, and further dilute it with methanol to obtain a working solution of 1 μ g/mL. An analysis was conducted on the formation of NAPQI. Detection was performed using ultra-high performance liquid chromatography-mass spectrometry (UPLC-MS) equipped with an AB SCIEX LC AC system, PDA detector, and triple quadrupole 6600+instrument at 0.5, 1, 2, 4, 8, 24, 32, and 48 hours after taking acetaminophen (APAP). Chromatographic separation was achieved using a Waters ACQUITY UPLC BEH C18 column (2.1 mm x 50 mm). 1.7 micrometers). The eluent used was (A) 0.1% (volume ratio) formic acid aqueous solution and (B) methanol, separated using a gradient elution procedure, as shown in Table S1. The flow rate of ultra-high performance liquid chromatography is 0.3 mL/min, and 10 microliters of sample are used for ultra-high performance liquid chromatography-mass spectrometry analysis. The ion pair characteristics corresponding to dissociation potential, collision energy, collision energy distribution, and temperature are m/z 149.98-108.04 (30, 30, 15, and 350), corresponding to NAPQI.

2.12. statistical analysis

Continuous variables are expressed as mean ± deviation (standard deviation). Use a two tailed student t-test to study the differences between two independent samples. The data analysis was conducted using IBM SPSS Statistics 25 version (IBM Corporation). The linear regression model was performed using SPSS and GraphPad Prism 7.00 (GraphPad software). The statistical significance is set to P ≤ 0.05.

3. Results

3.1. Cultivation of organoids

Inoculate the cell mixture into Transwell culture plates at densities of 1 × 10-6, 5 × 10-5, and 2.5 × 10-5 cells/well to obtain in vitro intestinal models. The transepithelial resistance (TEER) of all groups reached 300 Ω cm2 within 21 days, with the high-density co culture group reaching 300 Ω cm2 within 12 days (Figure 1A). On the 14th day, both fluorescein sodium and fluorescein glucan met the permeability test requirements in the high-density co culture group model (Figure 1B). When the cross epithelial resistance reaches 200-1000 Ω cm2, the Caco-2 monolayer can be considered intact (Iftikhar et al., 2020). In subsequent experiments, the high-density co culture group was selected to form the intestinal barrier. The tight junction proteins Occludin and ZO-1 were stained to demonstrate their integrity under intestinal equivalent culture conditions. On the 7th and 21st day, we observed a continuous network structure of Occludin (Figure 1C), while in the slices on the 14th and 21st day, Occludin and ZO-1 showed a network structure (Figure 1D, Figure S1A), indicating that the intestinal barrier was considered intact on the 14th day. On the 14th and 21st day, the transporter proteins MRP2 and PGP were distributed on one side of the intestinal slices (Figure 1E, Figure S1B), indicating the polarity of the intestinal model. Overall, the results indicate that the intestinal equivalents obtained on day 14 have tight junctions and polarity; Therefore, on the 14th day, the intestinal barrier was considered mature and used for further experiments. In Figure 2A, the diameter of the model rapidly increased within 10 days and stabilized at approximately 400 micrometers. HE staining showed no obvious necrotic core on day 7 (Figure 2B), but there was mild necrotic core or several cell apoptosis on day 10 (Figure 2B), and severe central necrotic area on day 14 (Figure S1I). Consistent with the distribution of dead cells shown in live/dead cell staining (Figure 2C), more dead cells were detected on day 10. The total amount of ATP in the model continued to increase (Figure S1C), while the secretion of lactate dehydrogenase (LDH) did not significantly increase (Figure S1D). The model grows steadily, and as the cultivation time prolongs, the number of necrotic cells increases. Internal necrotic cells cannot be detected by detecting LDH levels in the supernatant. To avoid the formation of necrotic cell nuclei, these results indicate that the cultivation time is limited to 7 days. On the 7th day (Figure 2D) and the 10th day, the fluorescence signals of human umbilical vein endothelial cell T1 (HUVEC-T1), human alveolar macrophage THP-1, and human hepatic stellate cell HHSC were tracked using CD31, CD68, and α - SMA signals, respectively, and immunofluorescence staining was used for tracking.

Figure S1G indicates that the three cell types mentioned above still exist in the sphere on day 10.On the 7th day, the transport proteins MRP2, PGP, and BSEP were detected (Figure 2E), and the function of bile acid efflux was tracked by fluorescent probe CLF (Figure S1H). CLF is a substrate for MRP2 and BSEP (Boaglio et al., 2010), and the CLF signal also indicates the function of these transporters in organoids. Finally, PAS staining showed glycogen deposition (Figure 2B), and the increasing trend of ALB and UREA indicated that the overall synthesis metabolism function of the liver was not hindered even in emergency situations of cell death (Figure S1E-F). These results indicate that the impact of organoids on liver function remained relatively stable within 14 days.

3.2. Establishment and Functional Overview of Intestinal Liver Chip

On the fourteenth day of establishing the intestinal model, the intestine and 50 liver like organs were placed in separate compartments to form the intestine liver chip model. The interconnected compartments allow liquid to circulate in the channels and chambers, promoting co culture of the two types of organs. The 2OC system was maintained for 7 days. Figure 3 summarizes the in vitro cell viability and intestinal and liver function after establishing the 2OC model. We monitored TEER levels (Figure 3A) to evaluate the maintenance of the intestinal barrier. Despite fluctuations, TEER remained above 300, indicating that the integrity of the intestinal barrier was consistently maintained even when co cultured with liver equivalents. During the cultivation period, the levels of liver function indicators ALB (Figure 3B) and UREA (Figure 3C) in the liver chamber were measured. ALB levels increased in the first 3 days and then decreased, while UREA levels did not show significant changes. In addition, LDH, AST, and ALT levels were evaluated to determine the cellular viability of intestinal and/or hepatic organoids (Figure 3E and F). The levels of lactate dehydrogenase (LDH) in both chambers slightly increased, while with prolonged cultivation time, the levels of aspartate aminotransferase (AST) and alanine aminotransferase (ALT) increased. These results indicate changes in function and cell viability within the 2OC system, suggesting an unexpected decline in overall liver organ function and cell viability despite maintaining the integrity of intestinal organoids. Next, the immunostaining signals of liver organoids obtained from 2OC chips and U-bottom 96 well plates were compared (Figure 3G). Firstly, during the cultivation of the U-plate, the Ki67 signal was mainly observed to be distributed along the edges of the sphere and uniformly distributed inside the sphere, although its intensity was weak after 7 days of co cultivation with the 2OC chip. When cultured on a U-plate, the PGP, MRP2, and BSEP signals along the edge of the liver sphere are stronger and can also be detected inside the sphere. After co culturing with 2OC chips for 7 days, strong transporter signals were observed along the edges, and the signal intensity of PGP and MRP2 was slightly stronger than that of the signal cultured on the U-plate. Finally, from the two cultivation methods, the cytochrome P450 (CYP) 3A4 signal was dispersed in the spheres, and the signal of the 2OC model was slightly enhanced. These results describe the basic functions of liver organoids and indicate that the expression of Ki67, PGP, and MRP2 in liver spheroids in the 2OC chip is different from that in the U-plate.

3.3. Toxicity assessment of excessive acetaminophen in 2OC dual organ chip

Before testing the 2OC chip, HepG2 and Caco-2 cells were exposed to different concentrations of acetaminophen (APAP) for 48 hours. When the concentration exceeded 16 millimoles, a sharp decrease in the survival rate of Caco-2 cells was observed (Figure S2A), while when the APAP concentration exceeded 2 millimoles, the inhibition rate of HepG2 cells exceeded 20% (Figure S2A). These results indicate that strong liver toxicity may occur within the range of 2-16 millimoles of APAP, but no significant toxic effects were detected in Caco-2 cells. In liver organoids, when the concentration exceeded 2.5 millimoles, the cell survival rate decreased to below 80% (Figure S2B). After treatment with 4 millimoles of APAP for 48 hours, the survival rate of liver organoids was measured to be 70.69 ± 26.97%, consistent with the significant toxicity trend observed in HepG2 cells. Based on these results, select this dose for further testing of intestinal organs. To achieve a final concentration of 4 mM, 10 mM APAP was applied from the top side for 48 hours, and survival rate and TEER were measured to evaluate the viability and barrier function of intestinal cells. Compared with the control group, there was no significant decrease in cell viability after drug administration (Figure S2C). Although TEER showed a decreasing trend (Figure S2D), it still remained above 300 Ω cm2, meeting the requirements for intestinal barrier integrity. These results indicate that APAP has significant cytotoxicity towards liver cells, but does not impair intestinal barrier function at the detected concentrations. In the 2OC chip, the top compartment containing mature intestinal organs was added under the condition of 10 millimoles of aspirin (APAP) concentration. After 48 hours, the basal chamber showed a survival rate of 83.94% for each sphere, slightly higher than the survival rate of the U-plate (70.69%). Although no significant differences were found in transepithelial electrical resistance (TEER) (Figure 4A). Between these two groups, AST (Figure 3B) (P<0.05) and ALT (Figure 3C) levels showed an upward trend compared to the control group, while ALB levels decreased after 48 hours of treatment (Figure 3D). AST, ALT, and ALB reflect liver damage, while TEER levels do not reflect the severity of intestinal barrier damage.

Finally, we detected changes in the fluorescence signal after APAP treatment. The detection of Ki67 signal reflects changes in cell proliferation ability (Figure 4E). In the control group, Ki67 signals were mainly concentrated on the side, and scattered signals were observed inside the sphere. Compared with the control group, the Ki67 signal after treatment became less significant on the side, and the signal was mainly concentrated on the inside, which may be due to the outermost cells being exposed to acetaminophen (APAP) for a longer period of time than the inner cells, enhancing the inhibition of cell proliferation. Next, CellROX levels were detected and stronger signals were observed in the APAP group (Figure 4F). APAP induced apoptosis was observed in hepatic spheroids, and TMRE was initially used to evaluate mitochondrial membrane potential (Figure 4G). In the APAP group, the signal intensity on the lateral side decreased, while in the control group, the signal distribution was more uniform, indicating that the mitochondrial membrane potential of lateral cells inside the sphere decreased after APAP administration. After 48 hours of incubation on the 2OC chip, the expression of caspase-3 (Figure 4H) and p53 (Figure 4I) involved in the cleavage of apoptotic proteins in the acute pancreatitis (APAP) group showed a stronger signal from cleaved caspase-3 and a slight increase in p53 signal. These results indicate that acute pancreatitis induced hepatic spheroid proliferation inhibition, oxidative stress, and cell apoptosis can be detected after 48 hours of incubation in the 2OC chip.

3.4. Absorption and Metabolism of Drugs in 2OC Dual Organ Chips

Figures 5A and 5B show the evolution of drug concentration over time in each chamber. A rapid decrease in acetaminophen concentration was observed in the intestinal compartment (Figure 5A), while a corresponding increase in drug concentration was detected in the liver compartment (Figure 5B). At 24 hours, 3.93 ± 0.46 mM and 3.68 ± 0.41 mM of acetaminophen were detected in the intestinal and hepatic compartments, respectively. Although the absorption process is easy to observe, the metabolic process is not easy to simulate, as can be seen from the insignificant decrease in acetaminophen concentration, which may be due to excessive drug dosage. Even if the total drug concentration shows a downward trend (Figure 5C), the downward trend is not solely due to sampling or drug metabolism. We calculated the theoretical total amount of acetaminophen without metabolism for 48 hours by referring to the concentration measured over 32 hours. When there is no cellular metabolism, the amount of acetaminophen should be 3.16 ± 0.53 mol; However, the actual detected amount was 2.95 ± 0.20 mol. These results indicate that even in the event of drug overdose, the use of our 2OC chip can simulate the absorption and metabolism processes of drugs. next. We have detected the toxic metabolite NAPQI, which was not detected in the supernatant. In order to accurately describe the drug absorption process in the 2OC chip, we used the NCA model to calculate the drug absorption kinetic parameters of acetaminophen (APAP). The calculated Tmax (peak time), Cmax (peak concentration), and AUC (area under the curve) are 24 ± 11.31 h, 3.98 ± 0.59 mM, and 157.73 ± 14.88 h mM, respectively. However, due to the minimal decrease in APAP concentration, it is difficult to fit the pharmacokinetic parameters of drug elimination.

4. Discussion

This study utilized an in vitro gut liver multi organ chip platform to evaluate drug absorption and predict pharmacokinetics and toxic effects. We constructed an intestinal liver chip system by coupling two independent organs on a 2OC multi organ chip platform to reproduce the continuous drug absorption of orally administered acetaminophen (APAP) and the subsequent acute liver injury. In the 2OC system, the absorption rate and toxic effects of APAP are consistent with known in vivo treatment processes. These results indicate that the absorption process can be simulated in our 2OC model and toxic effects can be detected in the presence of upstream intestinal organs, consistent with previous reports. Our findings provide support for future use of multi organ coupled multi organ chip models as auxiliary models for drug toxicity prediction. Based on the ratio of intestinal epithelial cells to goblet cells in the human small intestine, Caco-2 and HT29-MTX-E12 cells were used for intestinal construction (Chen et al., 2017); Tsamandouras et al., 2017). Create liver spheroids based on the proportion of cellular components in healthy liver (Weiler Norman and Rehermann, 2004). Based on previous research, we detected the structure and liver function during a two-week co culture. The signals of CD31 (Newman, 1997), CD68 (Gao et al., 2018), and α - SMA (Chiabotto et al., 2021) markers induced by HUVEC-T1 in THP-1 and HHSC indicate cell distribution after 10 days of co culture. The accumulation of chronic liver failure (CLF) (Hopwood et al., 2006), transport proteins MRP2, BSEP, and PGP (Rendic, 2021), as well as increased levels of albumin (ALB) and urea (UREA) (Baudy et al., 2020), respectively, indicate enhanced bile transport, drug efflux function, and synthetic metabolism function. Chronic liver failure accumulates in the tubular structure of 3D spheroids (Ramaiahgari et al., 2014), indicating the formation of preliminary tubular structures after 7 days of culture. We constructed a microfluidic based gut liver chip based on our gut and liver organoids (Marin et al., 2019); Maschmeyer et al., 2015). Other gut liver models have shown that organ surface area and metabolic capacity can affect the absorption and metabolism of oral medications (Lee et al., 2017). Use aspartate aminotransferase (AST) and alanine aminotransferase (ALT) (Meunier and Larrey, 2019) to evaluate liver toxicity. After 3 days of co culture, the increasing trend indicates potential liver cell damage, while the decreasing trend of albumin (ALB) may be due to insufficient nutrient supply, drug administration, malnutrition, reduced protein intake, and liver damage (Belinskaia et al., 2020); Meunier and Larrey (2019). In this study, the CYP3A4 signal was slightly enhanced, which is consistent with previous research results, indicating that the activity of CYP3A4 is enhanced after gut liver co culture (Chen et al., 2018b; Shinohara et al., 2021).

Due to previous studies mainly focusing on the enhanced metabolic activity of liver organoids after gut liver co culture (Shinohara et al., 2021), we have focused on the transporter proteins MRP2, PGP, and BSEP, which mediate the clearance of exogenous substances from liver cells (Jetter and Kullak Ublick, 2020). Our results clearly indicate that the 2OC model can simulate the absorption of toxic hazards after oral administration, and the supernatant is easy to preserve for monitoring human biomarkers.

The reason for choosing this concentration is as follows: 1) At the tested concentration, the integrity of the intestinal barrier will not be compromised; 2) The toxic effect of this drug on the liver is higher than its toxic effect on the intestine. Although the absorption of acetaminophen (APAP) is largely influenced by gastric emptying (Toes et al., 2005), it is mainly absorbed through passive diffusion in the proximal small intestine (M'esza'ros et al., 2022). In vitro, high concentrations of APAP have been shown to reduce liver cell viability (Bell et al., 2020). At low doses, APAP exhibits limited toxic effects; However, excessive or abusive use of APAP can lead to acute and severe liver cell damage (Bunchorntavakul and Reddy, 2018). The relationship between exposure concentration and toxic effects over time can explain the risk of time-dependent toxicity and the molecular mechanisms of chemicals (Tennekes and S á nchez Bayo, 2013). The concentration of APAP is crucial for reflecting drug exposure; Therefore, we examined the transportation status of APAP (Bessems and Vermeulen, 2001; Krenkel et al., 2014; Makin and Williams (1996).

Acetaminophen (APAP) has good oral bioavailability (88%) because it can be well absorbed and reaches a peak plasma concentration within 90 minutes after ingestion (Hodgman and Garrard, 2012). In our study, the predicted Tmax values in our 2OC model (24 ± 11.31 h) were higher than those in clinical data (0.33-4 h) and other 2OC chips (6 h) (Marin et al., 2019). APAP is rapidly absorbed at therapeutic doses and slows down with increasing doses (Rosenberg et al., 1981). At therapeutic doses, the plasma half-life of APAP is between 1.5 and 2.5 hours; However, we detected drug concentration equilibrium in both chambers 24 hours later, consistent with previous research findings. Rawlins et al. (1977) measured the plasma concentrations of volunteers after oral administration of APAP (500, 1000, and 2000 milligrams), and the results showed a delay in Tmax at 2000 milligrams. Some case reports have shown that excessive use of APAP preparations can lead to delayed peak plasma concentrations and prolonged APAP absorption (Chiew et al., 2010; Graudins et al., 2010; Roberts and Buckley, 2008). Mazaleuskaya et al. (2015) concluded that due to metabolic damage caused by excessive APAP, the half-life is prolonged to 4-8 hours. A non enterotoxic liver toxicity dose was selected, with an administration concentration (acetaminophen, 4 mmol, 604.64 mg/L) of approximately four to six times the toxic plasma concentration (100-150 mg/L) and two to three times the comatose lethal plasma concentration (200-300 mg/L) (Schulz et al., 2020). Due to the toxic concentration of acetaminophen in vitro (Bell et al., 2020) far exceeding the therapeutic plasma concentration in humans (5-25 mg/L) (Schulz et al., 2020), no significant drug elimination process was observed within 48 hours in our study. The auxiliary evidence of pharmacokinetics (PK) suggests that the chip model needs further modification, especially in studying the conversion methods of in vitro and clinical data to clinical settings, in order to improve the correlation between in vitro results and clinical outcomes.

Finally, we tested the toxic effects of our device on hepatic spheroid cells. Compared with the 96 well plate, the survival rate of spherical cells in the 2OC chip is slightly improved, which specifically reflects the changes in drug toxicity with drug concentration and exposure time. Excessive intake of acetaminophen (APAP) significantly increases serum levels of aspartate aminotransferase (AST) and alanine aminotransferase (ALT) (Acheampong and Thomas, 2016); Stockman， 2008）； However, after 48 hours of exposure in the 2OC model, there was a slight increase in AST and ALT levels. Meanwhile, the downregulation of ALB indicates that the metabolic function of the liver is inhibited and severe liver damage occurs. This decline is consistent with a previous report (Wu et al., 2022). After drug delivery, the Ki67 signal (necessary for cell division) of the outermost cells is weakened, indicating that the proliferation of the outermost cells is inhibited. Initially, we detected obstructed hepatic spheroids after treatment for acute pancreatitis.

At therapeutic doses, acetaminophen is mainly converted into non-toxic metabolites through glucuronidation and sulfation. Acetaminophen is oxidized by CYP2E1 enzyme and activated into the toxic product N-acetylaspartate aminotransferase inhibitor (NAPQI). NAPQI depletes glutathione (GSH) and converts it to acetaminophen cysteine. When GSH is below 30% of normal levels, NAPQI is no longer reduced, leading to toxic reactions. Although NAPQI can be detected in animal serum and organs (Russomanno et al., 2023); Zhang et al., 2018), but it is still difficult to detect this unstable intermediate in cell lines, except in primary liver cells (Pingili et al., 2019) (Zhang et al., 2020). As mentioned earlier (Zhang et al., 2020), only a small amount of acetaminophen is converted to NAPQI, and its half-life is extremely short (Coles et al., 1988); Madsen et al., 2007). It can be rapidly and easily detoxified by binding to GSH within 130 milliseconds (Coles et al., 1988; Madsen et al., 2007), or through NADPH, dihydronicotinamide adenine dinucleotide, and reductase reduction back to acetaminophen (Zhang et al., 2020), therefore detecting NAPQI in their in vitro models is difficult (Zhang et al., 2020). APAP protein binding occurs after depletion of glutathione (GSH), which is an early event in the mechanism of liver toxicity (Jaeschke et al., 2014); McGill et al., 2013). As an alternative in vitro model, GSH depletion (Leclerc et al., 2015) and oxidative capacity (Geib et al., 2021); Leclerc et al., 2015), covalent protein binding (Geib et al., 2019, 2021); Pierce et al. (2002) and in vitro toxicity information obtained using systems biology models (Leclerc et al., 2015) may be alternative methods for tracking or suggesting the formation and accumulation of NAPQI. NAPQI binds to mitochondrial proteins to induce protein acylation, which can lead to liver cell necrosis or oxidative stress, as well as mitochondrial dysfunction, ultimately resulting in cell apoptosis and liver damage (Bessems and Vermeulen, 2001); Krenkel et al., 2014; Makin and Williams, 1996; Saito et al., 2010). During the pathogenesis of acute pancreatitis (APAP) induced liver toxicity, liver cells exhibit mitochondrial dysfunction and oxidative stress (Shan et al., 2018); Ye et al., 2018). After treatment, the decrease in TMRE signal in the liver model indicates a decrease in mitochondrial membrane potential of liver cells (especially lateral cells), which affects mitochondrial function. Mitochondrial membrane potential represents changes in mitochondrial activity and disruption of mitochondrial homeostasis (Vianello et al., 2023). In the early stages of acute liver failure, mitochondrial membrane potential decreases, leading to changes in mitochondrial morphology and affecting mitochondrial function, which is the initial signal of liver cell damage (Li et al., 2011; Umbaugh et al., 2021). Considering that oxidative stress caused by acute liver failure can impair protein folding and trigger endoplasmic reticulum stress (Ferret, 2001; Ye et al., 2018), we used CellROX to locate reactive oxygen species signals and observed a significant distribution of reactive oxygen species signals in the acute liver failure group. These results indicate that using an 2OC chip to simulate excessive acute liver failure can detect changes in mitochondrial membrane potential and cellular oxidative stress in a liver model.

Liver cell apoptosis occurs in the early stages of acute liver failure (Possamai et al., 2013). Acute liver failure (APAP induced) can lead to depletion of glutathione (GSH), and APAP triggers mitochondrial toxicity pathways and processes that convert cysteine aspartic protease 3/7 precursors into active cleaved cysteine aspartic protease (Baek et al., 2016). Cysteine aspartate protease-3 is a key executor of cell apoptosis, and its precursor enzyme form is cleaved before being converted into its active form (Fernandes Alnemri et al., 1994). The tumor suppressor gene p53 regulates cell apoptosis through transcriptional activation (Arakawa, 2005). APAP induced liver toxicity may lead to slight upregulation of p53 (Chen et al., 2018a). After APAP administration, the signal of cleaved caspase-3 protein significantly increased, and the p53 signal was slightly enhanced, indicating that APAP induced cell apoptosis signaling.

However, it cannot be denied that animal models have made significant contributions to the understanding and progress of modern medicine, especially in the past few decades, which may be irreplaceable by alternative in vitro models for the safety assessment of drugs and chemicals. Although these new models have played and will continue to play an important role in biomedical and regulatory research, allowing for the study of a series of effects on a substance and providing high sensitivity without interference from other biological phenomena (Kaplan et al., 2024). On the other hand, it should be emphasized that the 2OC model cannot yet replace existing biological models, but rather serves as a research tool that provides specific insights in advance to reduce the number of animals required for drug screening. Our results indicate that the 2OC model can use fluorescent probes to measure drug absorption and detect signal changes associated with multiple toxicities, demonstrating the potential of multi organ chips as auxiliary evidence for simultaneously detecting drug efficacy/toxicity and pharmacokinetic potential.

5. Conclusion

We have established a gut liver systemDual organ chipThe system evaluates the toxic effects of acetaminophen (APAP) by simulating its acute liver injury process after excessive use. Intestinal organoids were established by co culturing Caco-2 and HT29-MTX-E12 cells, and hepatic spheroid organoids were established using HepG2, HUVEC-T1, PMA induced THP-1 cells, and HHSC. In the multi organ chipset, there was a slight improvement in the expression of PGP, MRP2, and CYP3A4 in liver organoids. After obtaining the concentration time curves of APAP in two chambers, the NCA method was used to fit the pharmacokinetic parameters Tmax, Cmax, and AUC. APAP treatment increased the levels of liver toxicity biomarkers AST and ALT, reduced the secretion of albumin (ALB), inhibited cell proliferation, enhanced reactive oxygen species (ROS) signaling, reduced mitochondrial membrane potential, activated caspase-3, and enhanced p53 signaling. This study provides auxiliary evidence for the application of organ chips in drug toxicity research, while detecting drug content and toxicity reactions, providing potential for predicting the pharmacokinetic and toxicological properties of candidate compounds.