Use of human tissue stem cell-derived organoid cultures to model enterohepatic circulation

Abstract

The use of human tissue stem cell-derived organoids has advanced our knowledge of human physiological and pathophysiological processes that are unable to be studied using other model systems. Increased understanding of human epithelial tissues including intestine, stomach, liver, pancreas, lung, and brain have been achieved using organoids. However, it is not yet clear whether these cultures recapitulate in vivo organ-to-organ signaling or communication. In this work, we demonstrate that mature stem cell-derived intestinal and liver organoid cultures each express functional molecules that modulate bile acid uptake and recycling. These organoid cultures can be physically coupled in a Transwell system and display increased secretion of fibroblast growth factor 19 (FGF19) (intestine) and downregulation of P450 enzyme cholesterol 7 α-hydroxylase (CYP7A) (liver) in response to apical exposure of the intestine to bile acids. This work establishes that organoid cultures can be used to study and therapeutically modulate interorgan interactions and advance the development of personalized approaches to medical care.

NEW & NOTEWORTHY Interorgan signaling is a critical feature of human biology and physiology, yet has remained difficult to study due to the lack of in vitro models. Here, we demonstrate that physical coupling of ex vivo human intestine and liver epithelial organoid cultures recapitulates in vivo interorgan bile acid signaling. These results suggest that coupling of multiple organoid systems provides new models to investigate interorgan communication and advances our knowledge of human physiological and pathophysiological processes.

INTRODUCTION

Most biological processes in the human body involve biochemical communication between two or more organs. Communication plays a major role in the physiology of an organ giving single-organ model systems limited capability to accurately model in situ physiology and pathophysiology. Animal models have traditionally been used to model interorgan communication; however, animal models do not necessarily represent human physiology. Thus, generation of human systems that simulate organ-to-organ communication and facilitate experimental manipulation will help define how one organ influences the biology of another, allow investigations that are too risky to be done in humans, and aid the development of preclinical assays that have predictive power for human therapies.

Previous in vitro work investigating intestinal epithelial biology was limited to studies in short-term intestinal explants (1) or in transformed cell lines, such as the Caco-2, T84, and HT29 cultures that do not replicate the heterogeneity or all aspects of the physiology of the human intestinal epithelium. Recently, the development of intestinal organoid cultures from either 1) induced or embryonic-derived pluripotent stem cells or 2) human intestinal tissue-derived crypts or single Leucine Rich Repeat Containing G Protein-Coupled Receptor 5 (LGR5)⁺/Paneth cell units has begun to provide insights into human intestinal epithelial physiology that had not been previously appreciated. The human tissue-derived intestinal organoids (HIOs) are grown as three-dimensional (3-D) spheroids in Matrigel and can be passaged indefinitely in growth medium (2) where they maintain regional specificity and host-specific phenotypic properties (3). After the removal of growth factors, the cultures differentiate into the mature cell types found in the intestinal epithelium (4) and exhibit characteristics such as increased brush border formation that model physiological features of the intestinal mucosa (5, 6). Intestinal organoid cultures are used to model a wide array of normal intestinal biology and intestinal diseases that include many genetic diseases, intestinal-microbe interactions, intestinal regeneration, drug discovery, and cancer (7).

The establishment, validation, and wide use of liver organoid cultures to study human liver biology has lagged behind that of the intestine. Primary human hepatocytes are widely used but suffer from being nonrenewable and limited in function when cultured. Transformed or cancer liver cell lines have also been used but they lack many features of normal hepatocytes. Human liver organoids (HLOs), either derived from pluripotent stem cells or tissue localized stem cells, may offer powerful potential to provide insight into human liver function. As with the intestine, human liver cell lines have limited capability in recapitulating liver biology, structure, and function. The establishment and propagation of HLOs derived from either liver tissue bile duct cells (8) or hepatocytes (9) are similar to that of the HIO cultures. Liver stem cells are isolated from human tissue and grown in Matrigel with a growth factor-supplemented media. Manipulation of the growth factors results in the generation of hepatocyte-like cells that synthesize bile acids, albumin, and α-1-anti-trypsin while engaging in detoxification, LDL uptake, and glycogen accumulation (8, 9). The HLOs have been used to study fibrosis, metabolic diseases, infectious diseases, and liver cancer (10–12).

Bidirectional signaling between intestine and liver is important to the physiology and function of both organs. The liver produces and secretes factors such as bile acids and cholesterol into the lumen of the small intestine via the biliary tree (13–15). These factors influence the uptake of nutrients by the intestinal epithelium including digestion and absorption of fats. Glucose, amino acids, and recycled bile acids absorbed by the small intestinal epithelium are carried to the liver through the portal circulation where the liver rapidly processes, stores, recycles, or distributes them. These interactions indicate that the physiology of the small intestine and liver are intimately connected. Until now, the ability to study these interactions in an ex vivo environment that more readily recapitulates features of the human enterohepatic circulation has been limited due to the lack of good model systems. The establishment of HIO and HLO cultures offers the opportunity to determine whether these novel ex vivo model systems can be used to dissect signaling pathways and communication between the two organs and to model diseases such as cholestasis and bile acid malabsorption syndromes that result from abnormal interactions. One of the most well-characterized signaling pathways between the intestine and the liver is the enterohepatic circulation of bile acids and fibroblast growth factor 19 (FGF19) that regulates bile acid synthesis, secretion, and intestinal absorption and transport (16). In the intestine, bile acids are absorbed by ileal enterocytes. The uptake of bile acid triggers enterocyte production of FGF19. The bile acid and FGF19 are secreted into the portal system where they are transported to the liver and inhibit bile acid synthesis (Fig. 1). We have examined whether HIO and HLO cultures functionally model aspects of the bile acid recycling pathway that would be expected in the native organs. To assess whether the appropriate signaling pathways are intact, we have physically coupled the two organoid models in a Transwell system and find that the coupled system can replicate features of the in vivo physiology with respect to bile acid uptake, transcellular transport, and FGF19 synthesis. This work demonstrates that intestine-liver interorgan communication is feasible ex vivo using mature stem cell-derived organoid cultures. Models such as these will be an important step in developing personalized medicine-based approaches to treat both gastrointestinal and liver diseases.

Figure 1.

Schematic of key components of the enterohepatic bile acid synthesis and recycling pathways. Bile acids are secreted by the liver hepatocyte into the hepatobiliary tree via BSEP, a unidirectional efflux transporter, and is carried by the hepatobiliary tree to the small intestine where they function to facilitate lipid digestion and absorption. In the ileum, an apical bile acid transporter ASBT transports bile across the enterocyte apical membrane into the cytoplasm where it binds to FXR, a steroid hormone receptor, and initiates gene transcription of the steroid hormone FGF19, the intracellular bile acid transporter FABP6, and the basolateral bile acid transporter OST. FABP6 and OST facilitate transport of bile acids from the cell interior into the portal vein. FGF19 is also secreted into the portal vein. The portal vein transports both FGF19 and bile acids to the liver. In the liver, bile acids are transported into the hepatocyte via an uptake transporter, NTCP, where they can directly downregulate transcription of CYP7A1, a cytochrome P450 enzyme that is necessary for de novo bile acid synthesis. Bile acids can also affect CYP7A1 indirectly through activation of nuclear receptor transcription factor SHP that in turn can also downregulate CYP7A1. FGF19 is transported into the hepatocyte where it also downregulates the transcription of CYP7A1. Thus, the recycling of bile acids directly results in the inhibition of de novo bile acid synthesis from cholesterol precursors by the hepatocyte via multiple pathways that result in CYP7A1 downregulation. Both recycled and newly synthesized bile acids are secreted by the hepatocyte into the hepatobiliary system via BSEP to restart the cycle. Created with BioRender.com. ASBT, apical sodium-dependent bile acid transporter; BSEP, bile salt export pump; CYP7A1, P450 enzyme cholesterol 7 a-hydroxylase; FABP6, fatty acid binding protein 6; FGF19, fibroblast growth factor 19; FXR, farnesoid X receptor; NTCP, sodium-taurocholate cotransporting polypeptide; OSTα/β, organic solute transporter α/β; SHP, small heterodimer partner.

MATERIALS AND METHODS

Human Ileal Organoid Cultures

Human ileal intestinal organoid lines were established from three donors and maintained by the Baylor College of Medicine (BCM) Digestive Diseases Center Organoid Core as previously described (5). The BCM Institutional Review Board approved the study protocol to obtain tissue samples from which the HIO lines were derived and written informed consent was obtained from all donors. HIOs were grown on 24-well Transwell permeable polycarbonate 0.4 μm supports (Corning #3710, Corning, NY) for 2 days and then differentiated for 5 days by changing the media as previously described (5). Before the start of the experiment, the transepithelial electrical resistance (TEER) was measured with an EVOM voltmeter (World Precision Instruments, Sarasota, FL) with >450 Ω·cm² for ileal monolayers used as benchmarks for differentiation. All HIOs were cultured at 37°C in 5% CO₂. Chenodeoxycholic acid (CDCA), cholic acid (CA), deoxycholic acid (DCA), and lithocholic acid (LCA) were purchased from Sigma-Aldrich (St. Louis, MO). The semisynthetic bile acid analog obeticholic acid (OCA) was purchased from Abcam (Cambridge, UK). All compounds were dissolved in freshly opened ampules of DMSO (Sigma) at a concentration of 20 mg/mL and stored at −80°C until use. Dilutions were performed in intestinal organoid differentiation media with a final concentration of DMSO of 0.05% (0.01% for LCA). HIOs were treated either apically or basally as indicated with 250 or 10 µM concentrations of the bile acids for 24 h.

Human Liver Organoid Cultures

For the establishment of liver organoids, liver tissue samples were obtained from the normal margins surrounding tumor tissue from three individual donors. The BCM Institutional Review Board approved the study protocol to obtain tissue samples from which the liver organoid lines were derived. Written informed consent was obtained from all the donors. Cells were liberated from the liver tissue sample using a digestion solution [2.5 mg/mL collagenase (Roche, Basal, Switzerland), 0.1 mg/mL DNase I (Sigma), EBSS (Hyclone)] followed by Accutase (Gibco) digestion for 10 min at 37°C, passed through a 70-μm filter, and pelleted at 400 g. The recovered cells were resuspended in a 30 µL Matrigel (Corning) droplet in a 24-well tissue culture plate covered with 500 µL of liver expansion medium as previously described (8). Organoids were passed 3–5 times and then dissociated and seeded on type IV collagen (Sigma) as monolayers in a 96- or 24-well tissue culture plate for 3–5 days in liver expansion medium, followed by 5 days in expansion medium supplemented with 25 ng/mL BMP7 (Peprotech; Rocky Hill, NJ). The medium was then changed to differentiation medium for up to 15 days with media changes every 2 to 3 days (8). For the ileal liver coculture experiments, differentiated ileal monolayers on a Transwell insert were added to the 24-well liver culture and treated apically with CDCA.

Total RNA Sequencing

Transcriptional analysis (RNA-Seq) analysis was performed on two donor-derived differentiated HIO lines of ileal cultures. Total cellular RNA was extracted using Qiagen RNeasy mini Kit (Qiagen) and the integrity was assessed on a 2100 Bioanalyzer (Agilent Santa Clara, CA). Cytoplasmic and mitochondrial ribosomal RNA was removed using RiboZero rRNA removal product (Illumina). A double-stranded cDNA library was created using 500 ng of total RNA (measured by pico green). An ERCC RNA Spike-In Control Mix 1 was added to each sample according to the manufacturer’s protocol. cDNA was generated using random primers. Libraries were created from the cDNA by first blunt ending the fragments, attaching adenosine to the 3′-end, and ligating unique adapters to each end for amplification. Paired-end sequencing of the cDNA libraries was performed using an IlluminaHiSeq 2500 with 100-bp paired reads at a depth of 30 million paired-end reads per sample. The fastq files were quality trimmed using Trim Galore! (17). Trimmed reads were aligned to GRCh38 using Hisat2 and a count matrix was generated from the aligned reads using the featureCounts (18). Gene expression was performed on protein-coding regions using limma R package with voom transformation (19). The principal component analysis was generated using the R statistical programming environment and used the log₂-transformed blocking corrected trimmed mean of M-values (TMM) normalized counts per million gene expression values. The data have been deposited in National Center for Biotechnology Information (NCBI)’s Gene Expression Omnibus (GEO) and are accessible through GEO Series accession number GSE168005.

Single-Cell Sequencing

Single-cell sequencing was performed on three donor-derived differentiated HLO lines. HLO monolayers were trypsinized with TrypLE to produce single cells and used directly for library preparation. Sequencing libraries were prepared with the Fluidigm C1 microfluidic system and the Chromium single cell 3′ v3 kit. Libraries were sequenced with 150-bp paired-end sequencing on the NovaSeq 6000 platform. The sequencing data were aligned to the GRCh38 human genome with Cell Ranger version 4.0. Cells with 200–2,500 genes and less than 50% mitochondrial content were accepted for analysis (20) and clustered with the standard SCTransform workflow (Seurat version 3.2) (21). Seurat’s AddModuleScore was used to score cells for hepatocyte, cholangiocyte, and progenitor identities using previously published gene sets (9). The data have been deposited in NCBI’s Gene Expression Omnibus and are accessible through GEO Series accession number GSE166589.

Western Blot

For Western blot analysis, HIOs from two Transwell inserts were solubilized in 200-µL radioimmunoprecipitation assay buffer (RIPA) buffer containing protease inhibitors (Roche) and stored at −20°C until assayed. The 293T cells or 293T cells transfected with a farnesoid X receptor (FXR) expression plasmid [a gift from Dr. David Moore (BCM)] were used as negative and positive controls, respectively. The samples were resolved on denaturing SDS-PAGE gels (4%–20%), and Western blotting was performed with an horseradish peroxidase (HRP)-conjugated, anti-FXR antibody (Cell Signaling Technology, Danvers, MA) diluted 1:250 overnight in casein blocking buffer (Sigma). GAPDH (Abcam) was used as a loading control for normalization. Species-specific secondary antibodies conjugated to IRDye 680 or 800 were used to detect the primary antibodies. Blots were imaged using a Li-Cor Odyssey CLX imaging system (Lincoln, NE).

FGF19 Quantification

FGF19 was determined by the human FGF19 quantikine ELISA kit (R&D Systems) per the manufacturer’s instructions. To obtain results within the working range of the ELISA assay, the optimal media dilution was determined for each HIO and HLO.

qPCR

For transcriptional analysis, organoids were lysed with either TRIzol (Thermo Fisher, Waltham, MA) or Ribozol RNA extraction reagent (VWR, Radnor, PA). RNA was isolated from the Ribozol or TRIzol using a Direct-zol-96 RNA purification kit (Zymo Research; Irvine, CA). Quantitative polymerase chain reaction (qPCR) was performed using a StepOnePlus real-time PCR System (Applied Biosystems; Foster City, CA) with One-Step RT-qPCR ToughMix with carboxyrhodamine (ROX) reference dye following the manufacturer’s protocol (Quant Biosciences, Gaithersburg, MD). Primer-probe mixes were obtained from Molecular Probes (Eugene, OR). Each sample was run in duplicate for 42 cycles. Expression levels for each gene were normalized to GAPDH levels for each sample and analyzed using the 2-Δ cycle threshold (Ct) method. qPCR probes: SLC10A2, H01001557_m1; FABP6, Hs01031183_m1; SLC51A, Hs00380895_m1; GAPDH, Hs02786624_g1; CYP7A1, Hs00167982; PTPN11, Hs01590340_gH; ALB, Hs00609411_m1; ABCB11, Hs00994811_m1; NTCP, Hs00161820_m1; FXR, Hs01026590_m1.

Bile Acid Transport

Total bile acid output was determined at 24-h post administration using the Total Bile Acid Assay (Cell Biolabs, San Diego, CA) following the manufacturer’s protocol. 20 µL of media from the ileal organoid basolateral or apical compartment was analyzed.

FITC-Dextran Permeability

HIOs were incubated apically with CDCA concurrently with 4-kDa fluorescein isothiocyanate (FITC)-Dextran (Sigma-Aldrich, St. Louis, MO) diluted in differentiation media to a final concentration of 5 µg/mL. After 24 h, 50 µL of the basolateral media was collected and fluorescence quantified using a SpectraMax M5 (Molecular Devices, San Jose, CA) and compared with a standard curve.

Statistical Analysis

Statistics were calculated using Prism software (GraphPad, La Jolla, CA) or Microsoft Excel (Redmond, WA) and the significance was determined by using Student’s t tests (P < 0.05). Three individual donor lines per segment of intestine or liver were used (biological replicates) in duplicate (technical replicate), and the experiment was performed between two and six times (technical replicates). In all cases, statistical significance was calculated on the biological replicates (n = 3) following averaging of the technical replicates. For intestine/liver organoid cocultures, three HIOs were independently cocultured with two HLOs providing n = 6 biological replicates with different combinations of ileum/liver from which statistical significance was calculated.

RESULTS AND DISCUSSION

HIO Cultures Derived from Adult Stem Cells Isolated from Human Ileal Tissue Exhibit Established Features of Bile Acid Uptake, Transport, Signaling, and Transcriptional Response

Bile acid uptake and recycling is a known feature of the human ileum (16). To assess whether human ileal HIOs have key features that would predict the organoids could mimic bile acid signaling ex vivo, we used an RNA-Seq-based approach to examine the transcriptional expression of several key components of the uptake and transcriptional response to bile acids (Fig. 1): the apical sodium-dependent bile acid transporter (SLC10A2/ASBT), the intracellular bile acid binding protein (FABP6/I-BABP), the basal heterodimer organic solute transporter (SLC51A/B/OSTα/OSTβ), and the bile acid transcription factor farnesoid X receptor (NR1H4/FXR). The ileal HIO cultures examined expressed key bile acid pathway transcripts (Fig. 2A), suggesting that the components necessary for bile acid recycling and uptake are present in the organoid cultures. The nuclear receptor, FXR, the natural ligands of which are chenodeoxycholic acid (CDCA) and other bile acids (22), was expressed at the protein level based on analyzing protein extracts from several ileal lines by Western blot. Similar levels of FXR were detected in each of the individual HIO lines (Fig. 2B) normalized to the GAPDH loading control. Detection of the bile acid receptor FXR suggested that the HIO cultures could respond appropriately when exposed to apical bile acids.

Figure 2.

Ileal HIOs exhibit in vivo aspects of bile acid uptake and transport. Ileal HIOs were differentiated for 5 days on a Transwell membrane before the start of the experiment. A: the transcripts in HIO cultures from two individual patients were assessed using RNA-seq. A PCA plot shows the relative gene expression analysis between the two cultures of several key transcripts involved in ileal bile acid recycling and the enterocyte marker sucrase isomaltase (SI). Fatty acid binding protein 6 (FABP6), membrane G protein-coupled bile acid receptor 1 (TGR5), principal component analysis (PCA). B: Western blotting to measure FXR levels was performed on HIOs from three different patients. GAPDH was used as a loading control and 293T cells expressing FXR (FXR) or a control plasmid (CTL) were used as controls for the antibody. C: HIOs on Transwell membranes were incubated apically with 250-µM CDCA, CA, LCA, and DCA for 24 h. Levels of FGF19 present in the basal media were measured by ELISA. D: apical and basal FGF19 production following apical incubation with either 10-µM CDCA or OCA was assessed by ELISA. E: transcriptional levels of SLC10A2, FABP6, and SLC51A analyzed by qPCR following apical treatment with 250-µM CDCA. F: levels of total bile acid in the basal compartment of HIOs over time following apical addition of bile acid were quantified by colorimetric assay. G: transepithelial electrical resistance (TEER) in each Transwell before and 24 h following addition of CDCA. H: permeability to 4-kDA FITC-Dextran added to the apical compartment determined by quantifying the amount of fluorescence in the apical and basolateral compartment 24 h after the addition of the bile acid and the FITC-Dextran to the apical compartment. A sample without FITC-Dextran was used as a control. In all graphs, bar or points represent average ± SD (n = 3 individual patient organoid lines). *P < 0.05 according to Student’s t test compared with control samples. ASBT, apical sodium-dependent bile acid transporter; CA, cholic acid; CDCA, chenodeoxycholic acid; DCA, deoxycholic acid; FGF19, fibroblast growth factor 19; FITC, fluorescein isothiocyanate; FXR, farnesoid X receptor; HIO, human tissue-derived intestinal organoid; LCA, lithocholic acid; OCA, obeticholic acid.

Because the ileal HIOs expressed FXR, we determined whether treatment of the apical side of polarized HIOs in a Transwell insert with bile acids would activate NR1H4/FXR transcription and stimulate FGF19 synthesis and release (23). Incubation of the HIOs with CDCA, cholic acid (CA), lithocholic acid (LCA), and deoxycholic acid (DCA) resulted in a significant increase in the production of FGF19 levels in the basal compartment of the Transwell compared with vehicle treatment alone (Fig. 2C). CDCA is a potent ligand for FXR (24) and, as would be predicted, was the most effective at inducing FGF19 production (range: 20.0–27.9 ng/24 h; P < 0.05) compared with vehicle (range: 0.09–0.26 ng/24 h), whereas DCA (range: 3.3–6.7 ng/24 h), CA (range: 0.7–2.1 ng/24 h), and LCA (range: 4.2–7.3 ng/24 h) all induced statistically significant (P < 0.05) but lower levels of FGF19 compared with CDCA. As a control, duodenal and jejunal HIOs were also treated with CDCA and very little production of basolateral FGF19 was observed above the vehicle (range: vehicle 0.07–0.28 ng/24 h; CDCA 0.12–0.42 ng/24 h). These findings are in contrast to those published using the colon cancer cell line Caco-2 as a model of mammalian small intestine in which FGF19 transcripts were upregulated following bile acid treatment, although secreted functional FGF19 was not demonstrated (25). In vivo, FGF19 is secreted only on the basolateral side of the intestinal epithelium where it is transported via the portal blood flow to the liver. To assess whether bile acids induced any apical secretion of FGF19, ileal HIO cultures were treated apically with CDCA and the production of FGF19 was measured in the apical compartment. A modest but still detectable amount of FGF19 was also present in the apical HIO compartment (range: 2.3–3.0 ng/24 h; P < 0.05), but the predominant amount was secreted basally as expected (Fig. 2D). This finding is consistent with work in which short-lived human ileal explants incubated with bile acids resulted in FGF19 production (1). Addition of CDCA to the basolateral side of the HIOs resulted in no production of FGF19, a finding consistent with the function of the apical ileal sodium bile acid transporter being required for active bile acid absorption (26). Ileal HIOs also were treated both apically and basally with 10-μM 6-ethylchenodeoxycholic acid (obeticholic acid, OCA) (27), a semisynthetic bile acid mimic of CDCA that has a high-binding affinity for FXR. Preferential basolateral FGF19 secretion was observed in ileal HIOs treated apically with OCA (range: 11.6–25.3 ng/24 h; P < 0.05) compared with vehicle HIOs, similar to what was observed following 10-μM CDCA administration (Fig. 2D). Like CDCA, basolateral administration of OCA resulted in no production of FGF19. We calculated the half maximal effective concentration (EC₅₀) for CDCA (>250 μM) and OCA (10 μM) and found the values to be consistent with the higher affinity ligand-binding properties of OCA for FXR compared with those of CDCA (27). These results indicate FXR-driven transcription in the ileal HIOs is functioning physiologically in response to apical bile acid exposure by producing FGF19.

FXR activation also increases transcription of other factors that are involved in transport of bile acids across the cell and in excretion at the basal side into the portal system (Fig. 1). Transcripts that code for FABP6 and SLC51A/B/OSTα/β are regulated by FXR-dependent transcription that results from binding bile acid (28, 29). To evaluate these transcripts, CDCA-treated HIOs were assessed by qPCR and compared with vehicle-treated HIOs. Treatment of ileal HIOs with CDCA induced significant increases (P < 0.05 compared with vehicle-treated HIOs) in transcriptional expression of both FABP6 (fold-change range: 2.1–4.1) and SLC51A/OSTα (fold-change range: 2.2–3.2) (Fig. 2E). OCA treatment resulted in a similar induction of transcript expression. Evidence from short-term bile acid treatment of human ileal explant cultures also showed increased gene transcription of FGF19 as well as FABP and SLC51A/B/OST (1), similar to what is observed in the ileal HIOs, suggesting that primary human culture systems such as the HIOs will provide an improved model system to study intestinal bile acid recycling pathways compared with transformed cell lines. SLC10A2/ASBT is not thought to be positively regulated by FXR activation (30), and the examination of SLC10A2/ASBT transcripts in the HIOs following bile acid treatment did not reveal a significant increase compared with vehicle treatment (fold-change range: 1.1–1.3) (Fig. 2E). Thus, this analysis indicates that FXR activation in the HIOs by bile acids also induces transcription of key molecules required for the transport of bile acid across the cell and secretion into the portal blood flow.

In addition to synthesizing FGF19 in response to the presence of apical bile acid, the intestinal enterocyte also transports bile acid across the epithelium (Fig. 1). To determine whether the HIOs could transport bile acid following apical treatment with CDCA, the amount of CDCA in the basolateral compartment over a 24-h period was measured using a colorimetric assay. After treatment, bile acid was detected in the basolateral media that peaked by ∼10 h (Fig. 2F). To ensure that barrier function of the HIO monolayer was maintained, transepithelial electrical resistance (TEER) was measured before and 24 h following the addition of CDCA. There were no detectable changes in the TEER 24 h after CDCA incubation compared with values obtained after vehicle incubation (Fig. 2G). Permeability was assessed using concurrent apical administration of the CDCA with the nondigestible 4-kDa dextran conjugated with fluorescein isothiocyanate (FITC-dextran). FITC-dextran levels in the basal compartment were undetectable (Fig. 2H). These findings are in contrast to those obtained from treatment of the transformed Caco-2 cells with bile acids, which resulted in decreased TEER and increased dextran flux (31). These results indicate that the HIOs take up bile acid at the apical surface and are able to transport it across the cell and secrete it at the basal surface without significant alterations in tight junctional permeability.

Stem Cell-Derived HLO Cultures Respond to Intestinal Bile Acid Pathway Signals

There are thought to be two sources of proliferating “stem cells” in the liver: oval/cholangiocyte cells that reside near the bile duct tree (32) and dedifferentiation of a mature hepatocyte in a stem-like state (33). HLOs have been derived from both sources of liver proliferating cells (8, 9). Using HLOs derived from the oval cell proliferative response (8), 10X chromium single-cell sequencing was performed to examine the cellular composition of the cultures. Gene list comparison (9) (Supplemental Table S1; see https://doi.org/10.6084/m9.figshare.14605479) indicated that the HLOs were a mixture of hepatocytes, cholangiocytes, and a large number of phenotypic progenitor cells (Fig. 3A and Supplemental Fig. S1; see https://doi.org/10.6084/m9.figshare.14605434). The three cell types were widely distributed across clusters due the fact that the organoids are heterogeneous and have a spectrum of differentiating cells (Fig. 3A). To determine whether HLOs expressed appropriate molecules that would predict the ability to synthesize and recycle bile acid, the expression of the basal uptake transporters, sodium-taurocholate cotransporting polypeptide (NTCP, SLC10A1), the apical bile salt export pump (BSEP, ABCB11), and the P450 enzyme cholesterol 7 α-hydroxylase (CYP7A1), which is the rate-limiting factor for bile acid synthesis, were examined in undifferentiated or differentiated HLOs by qPCR. Other molecules important in liver bile acid synthesis were also assessed including NR1H4/FXR and small heterodimer partner 2 (PTPN11/SHP). Increases in the transcript levels for each of these, with the exception of PTPN11/SHP2, were observed using qPCR after the HLO cultures were differentiated by the manipulation of media growth factors (Fig. 3B), indicating a shift toward a hepatocyte-like phenotype that has intact pathways to respond to intestinal-recycled bile acid and FGF19 signaling. HLOs have been reported to secrete bile acid into the media (8). To determine whether the established liver organoids produced bile acids, media from HLO-differentiated cultures were assessed for bile acid levels using a colorimetric assay. All five differentiated HLO lines produced detectable levels of bile acid (Fig. 3C), indicating the potential to detect the downregulation of secretion if given the appropriate signals.

Figure 3.

Bile acid pathway components are present in liver tissue-derived organoids. Liver organoids were differentiated for 15 days on a 96-well plate coated with collagen before the start of the experiment. A: single-cell sequencing allows transcriptional distinction of hepatocyte, cholangiocyte, and progenitor populations present in the liver organoid cultures. First column identifies Uniform Manifold Approximation and Projection (UMAP) clusters, and second through fourth columns indicate relative cell scores for hepatocyte, cholangiocyte, and progenitor gene expression. B: qPCR was performed to quantify the increase in bile acid pathway transcripts following differentiation. All samples were normalized to GAPDH expression levels. C: supernatants from differentiated liver organoids were assessed for levels of total bile acids by colorimetric assay. D: liver organoids were treated overnight with either 250 μM CDCA (n = 3 donor organoid lines), 0.3 ng/μL of recombinant FGF19 (n = 2 donor organoid lines), or untreated controls (n = 3 donor lines). Bile acid pathway transcripts were assessed by qPCR and normalized to GAPDH. CDCA, chenodeoxycholic acid; FGF19, fibroblast growth factor 19; HLO, human liver organoid.

The liver responds to incoming FGF19 and bile acid signals by downregulating expression of CYP7A1, a p450 enzyme, which is critical for de novo bile acid synthesis. To assess whether HLOs could respond appropriately to signals provided by bile acids or by FGF19, HLO cultures were treated overnight with either CDCA or recombinant FGF19. Treatment with both CDCA and the recombinant FGF19 resulted in decreased transcription of CYP7A1 (Fig. 3D) as would be expected based on the known regulatory effects of these factors on CYP7A1 (34). Bile acid treatment of the HLO additionally repressed SLC10A1/NTCP levels (Fig. 3D), the expression of which is also known to be controlled by accumulating bile acids (35). Increased bile acid uptake by the liver is reported to upregulate ABCB11/BSEP (stimulates excretion of bile acids) and PTPN11/SHP2 expression (inhibits SLC10A1/NTCP transcription) (35); however, bile acid treatment of the HLOs did not significantly increase either ABCB11/BSEP or PTPN11/SHP2 expression (Fig. 3D). Because of the fact that the HLOs are organized as a monolayer lacking the three-dimensional structure of the bile canalicular hepatocyte surface that would be present in the liver organ, it is possible that some signaling pathways might require higher or more sustained exposure to bile acids to trigger their response. However, key pathways in bile acid synthesis such as the regulation of CYP7A1 seem to be largely intact in the HLOs and respond to signals appropriately. Thus, the study of liver diseases such as cholestasis, where bile flow is restricted or obstructed, will benefit from the physiological bile synthesis activity that is present in the HLOs.

Physically Coupled HIO and HLO Cultures Emulate Aspects of Enterohepatic Bile Acid Signaling

Components of the enterohepatic recirculation of bile acids appeared to be present in both the HIO and HLO cultures. To assess whether the physical coupling of the two cultures could recapitulate key aspects of the interorgan communication, the HIO was seeded on a Transwell membrane allowing apical exposure of the HIO to bile acids and FGF19 and bile acid secretion into the basal compartment. The seeded HLO cultures in the basal compartment would be positioned to receive the HIO-derived signals, thus mimicking the portal connection between the intestine and liver (Fig. 4A). Viability of both cultures was assessed by visual inspection of floating cells characteristic of cell death, TEER in the HIO transwell, and LDH assay following physical coupling and was not different from uncoupled cultures.

Figure 4.

HLOs respond physiologically when physically coupled to bile acid-treated ileal HIOs. A: schematic representation of HIOs and HLOs coupled in a Transwell system and incubated with 250-µM CDCA on the apical side of the Transwell (n = 3 HIOs and n = 2 HLOs; total 6 cocultures). B: quantification of total bile acid in the basal compartment after 24 h using a colorimetric assay. C: FGF19 secreted into the basal compartment by the HIOs in response to vehicle or CDCA treatment was quantified after 24 h by ELISA. Number at the top of the bar is the average amount of FGF19 produced. D: bile acid pathway transcripts were assessed by qPCR in the HLOs following incubation with CDCA. Bars indicate average value (n = 6) ± SD. *P < 0.05 by Student’s t test compared with HLO. CDCA, chenodeoxycholic acid; FGF19, fibroblast growth factor 19; HLOs, human liver organoid; HIOs, human tissue-derived intestinal organoids.

To determine whether the HLO could exhibit the appropriate signaling from the HIO following bile acid administration, the ileal HIOs were treated apically with CDCA. As expected (Fig. 2F), bile acid was transported across the HIO and was present in the basal compartment of the Transwell system (Fig. 4B). In addition, the HIO responded to CDCA treatment by producing increased amounts of basolateral FGF19 (Fig. 4C). The HLO responded to the presence of transported bile acid in the basolateral Transwell compartment with the expected downregulated expression of SLC10A1/NTCP (Ct range: 8.1–11.7 liver only; 11.0–14.2 liver cultured with ileum; 8.5–15.2 liver cultured with CDCA-incubated ileum) and CYP7A1 (Ct range: 5.0–30.0 liver only; 2.0–11.7 liver cultured with ileum; 1.0–2.2 liver cultured with ileum incubated with CDCA; P < 0.05 compared with liver cultured with ileum) (Fig. 4D). ABCB11/BSEP (Ct range: 8.3–12.0 liver only; 7.4–14.0 liver cultured with ileum; 6.6–13.3 liver cultured with CDCA-incubated ileum) and PTPN11/SHP2 (Ct range: 2.7–11.0 liver only; 2.3–12.1 liver cultured with ileum; 2.3–14.5 liver cultured with CDCA-incubated ileum) transcripts, although slightly downregulated, were not statistically significant (Fig. 4D). Our results support and extend the findings of Chen et al. (36), who used hepatocyte gene expression of CYP7A1 to validate cellular communication between physically coupled Caco-2 cells and primary hepatocytes grown in an integrated gut/liver microphysiological system. The effect of bile acids was not examined in their model. Thus, the ability of physically coupled HIO and HLO cultures to recapitulate key components of interorgan communication provides exciting new in vitro models in which human physiology can be dissected.

The ability of organoid model systems to emulate many properties of the native human organ has resulted in an exponential increase in understanding biology in areas of developmental biology, tissue homeostasis and regeneration, and the response of the tissue to disease. The physical coupling of ileal HIO and HLO cultures, and the resulting appropriate response of the coupled organoid platforms to bile acid treatment, provides important validation that these organoid cultures can communicate physiologically with each other ex vivo. More studies will be necessary to continue to validate their use as a model for interorgan communication and to extend the understanding of the biology of the ileal/liver communication axis. Recent work using mouse organoid cultures has demonstrated the feasibility of integrating multiple organoids from the stomach, small intestine, and liver (37). Addition of CDCA to that platform resulted in decreased CYP7A1 expression in the liver compartment. These studies, along with the establishment of coupled human organoid platforms, provide critical steps in better modeling in vivo physiology. The development of these systems has the potential to not only reduce the numbers of animals needed for preclinical studies but to recapitulate key aspects of human physiology that can be experimentally manipulated ex vivo. It is clear that there is significant potential to couple different organoid cultures to mimic integrated responses and to model interorgan interactions. Key components common to all cultures can be identified and variable aspects can also be recognized.

Several limitations may need to be overcome to continue to evolve the organoid cultures to model interorgan communications. There appear to be high levels of variability in biological readouts in both HIO and HLO cultures. However, this variability is one of the key assets of the system in that this mimics the heterogeneity that is present in the human population and will allow true biological replicates to be studied. Another hurdle that will need to be optimized is the issue of cell ratios, scaling, and vascular flow rates between the organoid cultures. In this work, these variables were not addressed in a systematic way. More in-depth studies will be required to explore these areas and to develop platforms that better model these features. Despite these limitations, the ability of these cultures to communicate with each other has tremendous promise for the development of “personalized” systems in which the effects of experimental drugs, pathogens, and genetic manipulation can be examined on an individual’s own systemic physiology. Systems such as these, in which ex vivo cultures can be derived from a donor’s own cells, have the potential to dramatically alter how medical therapies are implemented. In addition, there is potential to gain insight into previously unrecognized signaling molecules that control global pathways such as metabolism and energy regulation.

SUPPLEMENTAL DATA

Supplemental Table S1: https://doi.org/10.6084/m9.figshare.14605479.

Supplemental Fig. S1: https://doi.org/10.6084/m9.figshare.14605434.

GRANTS

Research reported in this publication was supported by the National Institutes of Health/National Center for Advancing Translational Sciences (NIH/NCATS) UH3-TR000503, U19-AI116497, P30-DK56338, PO-AI152999, U01-AI24290, P30-ES030285, and P42-ES0327725, The Cancer Prevention Institute of Texas (CPRIT) RP170005, and NIH P30 Shared Resource Grant CA125123.

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the authors.

AUTHOR CONTRIBUTIONS

S.E.B., F.B.H., S.E.C., M.E.M.T., T.S., and M.K.E. conceived and designed research; S.E.B., S.E.C., X.-L.Z., J.R.B., and M.E.M.T. performed experiments; S.E.B., M.D., S.E.C., C.B., M.R., C.C., and M.K.E. analyzed data; S.E.B., C.B., M.D., and M.K.E. interpreted results of experiments; S.E.B., S.E.C., and C.B. prepared figures; S.E.B. drafted manuscript; S.E.B., S.E.C., C.B., M.D., and M.K.E. edited and revised manuscript; S.E.B., F.B.H., S.E.C., S.A.C., C.B., X.-L.Z., J.R.B., M.R., M.E.M.T., T.S., M.D., and M.K.E. approved final version of manuscript.