Establishment of a cell culture platform for human liver organoids and its application for lipid metabolism research

Human liver organoids (HLOs) are reliable tools to represent physiological human liver biology. However, their use is limited especially in basic sciences. One of the reasons for this would be the insufficient systematic methodology to handle HLOs, including culture system, functional assessment, and gene transduction. Here, we generated and characterized mouse L cells stably and simultaneously overexpressing R‐spondin1, hepatocyte growth factor, fibroblast growth factor (FGF) 7, and FGF10 via lentiviral transduction. The conditioned medium of the cells contributed to HLO growth as a replacement of commercially available recombinant proteins, which leads to a significant reduction of their culture cost. Proliferative and maturation phases of the cells were controlled by switching the medium to facilitate the evaluation of hepatocyte function, including insulin responsiveness and intracellular lipid accumulation. Gene expression analysis revealed that HLOs highly expressed genes involved in lipid metabolism. Importantly, HLOs secreted physiologically matured very low‐density lipoprotein, which is rarely observed in mice and in established cell lines. Efficient gene transduction into HLOs was achieved via a transient 2‐dimensional culture during viral infection. This study provides an invaluable platform for utilizing HLOs in various research fields, such as molecular biology, pharmacology, toxicology, and regenerative medicine.


INTRODUCTION
Hepatocytes, which comprise approximately 80% of the liver mass, play an essential role in systemic homeostasis, including glucose uptake, lipid and protein synthesis, lipoprotein secretion, and drug metabolism [1] Hepatic failure can be triggered by viral infections, alcohol, drugs, and metabolic disorders, leading to multiple organ failure and death [2,3] Approximately 2 million people worldwide die of liver cirrhosis or hepatocellular carcinoma, [4] partly caused by chronic liver inflammation.Despite thorough investigation, the detailed molecular mechanisms underlying liver diseases remain unclear.One reason for this is the lack of a reliable human liver model with physiological liver function.Established cell lines and animals have been used to investigate liver function and metabolism; however, these tools do not necessarily replicate human liver biology. [5]Unlike humans, mice and rats lack cholesterol ester transfer protein, resulting in low levels of blood-low-density lipoprotein (LDL) and high levels of high-density lipoprotein (HDL), which activates reverse cholesterol transport, thereby exhibiting no symptoms of atherosclerosis. [6]Commercially available primary human hepatocytes (PHHs) and human induced pluripotent stem cell (iPSC)-derived mature hepatocytes are alternatives to these models.However, both cell types are expensive, exhibit limited proliferative capacity, and cannot be prepared in large numbers in the same lot; therefore, they are not the best suited for routine or large-scale use.
Organoids are organ-mimicking in vitro 3D structures that selforganize multiple types of cells differentiated from tissue-specific stem or progenitor cells. [7]Once transplanted into animals, organoids generate functional organs in vivo and possess highly physiological characteristics, [8] making them more useful for drug development and regenerative medicine and for extrapolating data to humans than conventional biological models. [9,10]Human liver organoids (HLOs) can be isolated from body tissues [11,12] or pluripotent stem cells. [13]HLOs possess self-organizing capacity, as observed in native liver tissues, and contain functional hepatocytes when cultured in an appropriate medium. [5]However, the expansion of HLOs requires expensive cytokines, such as R-spondin-1 (RSPO1), hepatocyte growth factor (HGF), fibroblast growth factor (FGF)−7, and FGF10, which hinders the extensive use of HLOs. [14]Whether HLOs can recapitulate the hepatic physiological functions that cannot be in cancer cell lines or animals has not yet been investigated.Although methods of transient gene transfection into organoids by electroporation have been reported, [15,16] highly efficient gene transduction into HLOs to verify their physiolog-ical significance, particularly stable gene transduction, has not been fully achieved.
In this study, we established a stable cell line of mouse L cells that simultaneously overexpress RSPO1, HGF, FGF7, and FGF10 (L-RHF2 cells), and showed that HLOs are expandable with the conditioned medium (CM) of the cells.Functional analysis revealed that HLOs represent the physiological properties of functional human hepatocytes, including the secretion of very low-density lipoprotein (VLDL), which is not correctly reproduced in established human cell lines and rodent models.We also present a methodology to efficiently transduce exogenous genes into HLOs in a convenient manner.These results provide a fundamental platform for the routine use of HLOs, which could have applications in hepatic metabolism and disease studies, drug development, and regenerative medicine, which require large-scale culture of HLOs with high homogeneity.
L-RHF2 cells were established via lentiviral infection, as mentioned in the Result section. [18]Cells were cultured in DMEM supplemented with 10% FBS, 100 units mL -1 penicillin, and 100 μg mL -1 streptomycin.Each CM was prepared from supernatants seeded at 5.0 × 10 6 cells 100 mm -1 dish for 72 h, filtered with a 0.2 μm membrane filter, and stored at −80 • C until use.
The human iPSC line, TkDN4-M, supplied by Stem Cell Bank at the University of Tokyo, was maintained as colonies in feeder-free conditions on vitronectin (Gibco)-coated plates with Essential 8 medium (Gibco).The cells were passaged every three days before reaching confluence.
HepaRG cells were purchased from KAC (Japan), thawed in basal hepatic cell medium (KAC) with thawing/plating/general purpose medium supplemented with antibiotics (Thermo Fisher Scientific), and collected for RNA extraction.
PHHs were purchased from Lonza and used according to the manufacturer's instructions.Briefly, cells were thawed in a single human thawing medium (Lonza) and either harvested for RNA extraction or seeded on a collagen-coated 6-well plate with hepatocyte plating medium (Lonza).After 1 h of plating, the medium was replaced with fresh medium and the cells were harvested 4 h after plating.
All cultures were performed in a 5% CO 2 incubator at 37 • C.

Differentiation of iPSCs into iPSC-derived HLOs (iHLOs)
Human iPSCs were differentiated into iHLOs, according to a previously described protocol, [13] with modifications to the process of differentia-tion from iPSCs into the definitive endoderm (DE). [18]Briefly, 80-90% confluent iPSCs were differentiated into DE by treatment with DE1 medium (Table S1) for 24 h, followed by DE2 medium (Table S1) for 24 h, and then with DE3 medium (Table S1) for 24 h.For primitive gut differentiation, DE cells were cultured in a primitive gut spheroid differentiation medium (Table S1) for three days.Free-floating spheroids were collected and embedded in Matrigel on Nunc Multidish 4-well plates (Thermo Fisher Scientific), followed by supplementation with posterior foregut spheroid medium (Table S1), which was replaced every two days for four days.Thereafter, the medium was changed to hepatocyte differentiation medium (HDM) (Table S1) and the cells were cultured for 10 d with a medium change every three days.

2.4
Preparation of primary HLOs (pHLOs) pHLOs were established using the two-step collagenase perfusion method [19,12] with some modifications.Healthy liver tissues of surgical specimens were washed with cold Earle's Balanced Salt Solution (EBSS; Gibco) and chopped into approximately 3-mm pieces in 5 mL of EBSS with 2.5 mg mL -1 collagenase D (Roche) and 0.1 mg mL -1 DNase I (Sigma).The fragmented tissues were then transferred to a 15-mL tube and incubated in a 37 • C water bath for 40 min with shaking.Following the addition of 5 mL of basal medium (Table S1), the tissue fragments were filtered through a 70-μm nylon mesh, 10 mL of basal medium was added, and kept on ice.The residual tissues on the mesh were further digested with 5 mL of accutase (Gibco) in a 37 • C water bath for 10 min with shaking, and 5 mL of basal medium was added, filtered through a 70-μm nylon mesh, and further added to 10 mL of basal medium.
After the digested fractions were combined, pellets were harvested via centrifugation at 100 × g for 5 min.The pellets were resuspended in 10 mL basal medium and collected again via centrifugation at 100 × g for 5 min.After resuspension in 10 mL of basal medium, the cell density was determined using a hemocytometer.Cells (10,000-20,000) embedded in Matrigel were aliquoted into each well of a Nunc Multidish 4-well plate.After solidification of the Matrigel in a 5% CO 2 incubator at 37 • C for 15 min, 500 μL of expansion medium (EM) (Table S1) was added to each well.The entire medium was changed every three days.The experiments using pHLOs complied with the Declaration of Helsinki and were approved by the Human Ethical Committees of the University of Tokyo (No. 18-338) and Hiroshima University (No. E-1582).Informed consent was obtained before tissues were sampled.

HLO culture and passage
According to a previous procedure, culture and passage of HLOs (iHLOs and pHLOs) were conducted with small modifications to the method for human intestinal organoids (HIOs). [20]Specifically, HLOs embedded in Matrigel were washed with phosphate-buffered saline (PBS) and treated with TrypLE Express solution (Thermo Fisher Scientific) for 10 min at 37 • C in a water bath.The recovered organoids were disrupted by vigorous pipetting 30-40 times and collected via centrifugation at 440 × g for 3 min.After removing the supernatant, the organoids were washed with 10 mL of basal medium.After centrifugation at 440 × g for 3 min, organoids were resuspended in Matrigel with 10% expansion EM on ice.Suspension aliquots were added to the wells of Nunc Multidish 4-well plates, leaving the border of each well untouched, and solidified in a 5% CO 2 incubator at 37 • C for 15 min.EM (500 μL) was then added to each well.The entire medium was changed every three days.HLO passage was carried out every six or seven days.
The passaging ratios ranged from 1:20 to 1:24.Cells were cultured in a 5% CO 2 incubator at 37 • C. Bright-field or fluorescent images of organoids were obtained using an all-in-one fluorescence microscope (Keyence, Japan), followed by the generation of a full-focused image from a series of z-stack images.

Lentiviral production and infection
HEK293T cells seeded at 3.3 × 10 6 cells 100 mm -1 dish were transfected with a lentiviral expression plasmid (20 μg dish -1 ) with packaging (pCAG-HIVgp; 8 μg dish -1 ) and VSV-G-and Rev-expressing plasmids (pCMV-VSV-G-RSV-Rev; 8 μg dish -1 ) as per a previous study. [21]ecifically, the calcium phosphate-plasmid DNA complex was generated in a 15 mL tube by mixing 36 μL of plasmid DNA (1 μg μL -1 ) with 60 μL of 2.5 M CaCl 2 , 504 μL of distilled water, and 600 μL of 2 × HEPES-buffered saline (pH 7.05), bubbling the mixture 20-30 times using a P1000 pipette, vortexed for 10 s, and incubated in a 37 • C water bath for 30 min.Each precipitate was added dropwise evenly over a 100 mm dish of cells.After 12 h, the culture medium was replaced with a fresh medium containing 10 μM forskolin.After 24 h, the medium containing the lentiviruses was collected and filtered.Single-cell suspensions of iHLOs were prepared by digesting iHLOs with TrypLE Express solution, as performed in "HLO culture and passage."The suspensions were seeded in collagen I-coated 6-well plates and infected with medium containing 10 μg mL -1 polybrene by centrifugation (700 × g, 90 min).After centrifugation, the cells were washed twice and incubated with fresh EM.Cells were harvested using TrypLE Express solution and embedded in Matrigel to regenerate organoids.

Separation of lipoprotein fractions
Lipoprotein fractions of the culture supernatant were separated by OptiPrep (Axis-Shield) density gradient ultracentrifugation.The cells were pre-treated with or without 30 μM PD98059 for 18 h and thereafter loaded with 500 μM oleic acid (OA) together with or without 30 μM PD98059 for 5 h in serum-free medium: DMEM for HepG2 cells and hepatocyte basal medium (Lonza) for iHLOs.The culture supernatant (4 mL) was condensed to 100 μL using an Amicon Ultra-4, PLHK Ultracel-100K (Millipore) via centrifugation at 5000 × g for 20 min.

Quantitative reverse transcription-polymerase chain reaction (RT-qPCR)
Total cellular RNA was extracted using Isogen (Nippon Gene) or RNeasy Mini Kit (Qiagen).Reverse transcription was performed using High-Capacity cDNA Reverse Transcription Kit (Thermo Fisher Scientific).mRNA levels were measured by fluorescence real-time PCR using StepOnePlus (Thermo Fisher Scientific) or QuantStudio 6 (Thermo Fisher Scientific).TaqMan universal PCR master mix (Thermo Fisher Scientific) and PrimeTime qPCR Assays (probes) (Integrated DNA Technologies) or FastStart Universal SYBR Green Master (Roche), together with gene-specific primers, were used to determine the mRNA levels of each gene.18s rRNA levels were used as internal controls for normalization.Primer sequences and probes are listed in Table S2.

Wholemount staining
Organoids cultured in Matrigel were washed with PBS and treated with a cell recovery solution (BD Biosciences) for 30 min on ice.The recovered organoids were collected by centrifugation at 440 × g for 5 min.

Measurement of intracellular triacylglycerol
After the organoids were recovered from Matrigel with TrypLE Express solution, lipids were extracted by hexane in 2-propanol (3:2, v/v).The amounts of intracellular triglyceride were determined by the Triglyceride E-test Wako Kit (FUJIFILM Wako Pure Chemical) and normalized to the amounts of total cellular protein determined by the BCA Protein Assay Kit (Pierce) according to each manufacturer's instructions.

Determination of cell viability
After the organoids were recovered from Matrigel with a cell recovery solution, the recovered organoids were collected by centrifugation at 440 × g for 5 min.Following the removal of the supernatant, organoids were suspended in 300 μL of basal medium.Cell suspensions (100 μL) were mixed with 100 μL of CellTiter-Glo 3D reagent (Promega) and incubated for 30 min at R.T.; luminescence signal was measured using Lumat 3 LB9508 (Berthold), according to the manufacturer's protocol.

Statistical analysis
Data was represented as the mean ± standard error of the mean (SEM).
Data were analyzed using the Student's t-test for two groups and Tukey's test for more than three groups, which was determined using the R software.Differences were considered significant at p < 0.05 and/or p < 0.01, indicated by asterisks.All quantifications included at least three independent biological replicates.

Hepatocyte expansion and maturation phases in HLOs are exchangeable by switching the culture medium
HLOs with self-organizing capacity can be differentiated from iPSCs. [13]We previously showed that the differentiation of HIOs from iPSCs was enhanced by adding Wnt3a and FGF2 during the differentiation of iPSCs into definitive endoderm. [18]Given that iPSC differentiation into HLOs also involves definitive endoderm and primitive gut spheroids, we modified the differentiation method for iHLOs by incorporating our procedure for iPSC differentiation into definitive endoderm (Figure 1A) and succeeded in establishing iHLOs with a high differentiation capacity (Figure 1B).We found that iHLOs proliferated and were passaged continually when cultured in EM that was used for pHLOs [11] (Figure 1B).In contrast, iHLOs failed to proliferate when kept cultured in the HDM used in the final step of iHLO differentiation.
Because liver cell properties differ between the proliferative and stationary phases, [22] we performed expression analysis of several marker genes in iHLOs cultured with EM and HDM.Expression of the stem cell marker, leucine-rich repeat-containing G protein-coupled receptor 5 (LGR5), was induced after replacing HDM with EM, suggesting an increase in stem cell-like cells within iHLOs (Figure 1C).
The expression levels of the hepatocyte markers, alpha-fetoprotein (AFP), and albumin (ALB) were reduced by changing the medium, suggesting hepatocyte maturation impairment in EM.These expression patterns persisted even after several passages.The expression levels of the stellate cell marker, vimentin (VIM), and the Kupffer cell marker, CD68, were also decreased by changing the medium from HDM to EM. VIM expression decreased as the number of passages increased.The expression of the cholangiocyte marker keratin 19 (KRT19) did not change with the culture medium.Furthermore, we investigated whether the expression of marker genes was reversible by switching the medium from EM to HDM.LGR5 expression was prominently decreased, whereas AFP and ALB expression was increased with the medium change (Figure 1D).However, the expression levels of KRT19, VIM, and CD68 did not change significantly compared with those of LGR5, AFP, or ALB.Consistent with the induction of mRNA expression, AFP and ALB protein secretion was also increased (Figure 1E).Similar results were obtained when pHLOs were used (Figure S1).These observations indicate that hepatocyte growth and maturation in HLOs can be controlled bidirectionally by interchanging EM and HDM.

HLOs are expandable with CM prepared from L-RHF2 cells
A major hindrance to the routine use of HLOs is the high cost of the commercially available recombinant proteins for culture.Previously, we established mouse L cells stably overexpressing Wnt3a, RSPO1, and Noggin, together with or without HGF via lentiviral gene transduction, and found that HIOs were continuously expanded with each CM, leading to a significant reduction in their culturing costs. [23,18]though crucial growth factors differ among organoid types, we assumed that L cells stably expressing exogenous factors essential for HLO proliferation, namely RSPO1, HGF, FGF7, and FGF10 (RHF2), could be generated using a strategy similar to that used for HIO culture (Figure 2A).After the preparation of each lentivirus, L cells were infected with each virus individually, and the range of optimal expression levels was adjusted by virus concentration or dilution.Through the examination of the initial cell density followed by the culture periods, we defined the optimal culture conditions to prepare L-RHF2 CM by seeding the cells at 5.0 × 10 6 per 100 mm dish and culturing them for three days (Figures 2B and 2C; Figure S2).Eventually, we were able to obtain a CM containing approximately 2400 ng mL -1 RSPO1, 80 ng mL -1 HGF, 160 ng mL -1 FGF7, and 150 ng mL -1 FGF10, which was determined via densitometric analysis of the western blots.We noticed that the molecular weights of FGF7 and FGF10 in the CM were higher than those of standard proteins prepared using the bacterial expression system.Considering that FGF7 can be N-glycosylated, [24] we examined whether FGF7 in the CM was N-glycosylated using glycopeptidase F (also referred to as peptide N-glycosidase F), which hydrolyzes the glycosylamine linkages of asparagine-linked glycoproteins.We found that treatment with glycopeptidase F decreased the molecular weight of FGF7 in CM to a level similar to that of the standard protein (Figure S3).In addition, FGF7 and FGF10 share high sequence similarity and glycosylation sites, [25,26] suggesting that these proteins are N-glycosylated in host cells before secretion into the medium.To further examine the stability of the exogenous protein concentrations of L-RHF2 CM, we harvested the CM after passaging the cells every three days and continued the process over 20 passages (from P6 to P26).
The same volume of each CM, assured by staining the total protein with Coomassie brilliant blue G-250 (Figure 2D), was used to trace the time-course changes in protein levels.The protein levels of each factor remained constant (Figure 2D), indicating that the exogenous protein secretion capacity of the L-RHF2 cells was maintained for at least two months.
Subsequently, we investigated whether recombinant proteins can be substituted with L-RHF2 CM to expand HLOs.iHLOs cultured with recombinant proteins or 25% L-RHF2 CM were passaged at a constant dilution ratio from P1 to P3, growth rates were compared via bright-field microscopy, and viable cell numbers were determined using CellTiter-Glo 3D reagents.Unexpectedly, despite no significant difference in protein concentrations between recombinant proteins and L-RHF2 CM, L-RHF2 CM proliferated iHLOs more prominently than recombinant proteins (Figure 2E).Consistently, iHLOs cultured with L-RHF2 CM expressed LGR5 at higher levels and AFP and ALB at lower levels than those cultured with recombinant proteins at P3 (Figure S4), indicating that L-RHF2 CM can increase proliferation of LGR5+ liver stem cells better than recombinant proteins.Moreover, pHLOs could be expanded using 25% L-RHF2 CM (Figure S5).L-RHF2 CM stored at −80 • C for one year could grow iHLOs as efficiently as freshly prepared CM (Figure 2F), indicating that L-RHF2 CM can be stored for a long time, which would be advantageous for the flexible handling of the CM.We conclude that HLOs can be expanded with L-RHF2 CM, which can greatly reduce the HLO culture cost.

HLOs proliferated with L-RHF2 CM possess fundamental properties as functional hepatocytes
We examined whether HLOs expanded with L-RHF2 CM exhibited the phenotypes of mature liver cells.As the liver is responsible for anabolic pathway upregulation in response to insulin, we focused on insulin signaling activation.Since HDM contain insulin, we used William's medium E as the basal medium.After serum deprivation followed by insulin treatment, the phosphorylation levels of Akt as well as its downstream proteins, mTOR and S6K, were drastically enhanced in iHLOs (Figure 3A).Furthermore, the expression of SREBP-1 target genes, namely acetyl-CoA carboxylase, fatty acid synthase, and stearoyl-CoA desaturase, which mediate fatty acid synthesis, [27] was also augmented (Figure 3B).Meanwhile, the expression of gluconeogenesis genes, including glucose-6-phosphatase and phosphoenolpyruvate carboxykinase 1, was declined upon insulin treatment.These results indicate that insulin exerts anabolic actions in iHLOs by regulating gluconeogenesis and lipid synthesis gene expression.
Several nuclear receptors, such as liver X receptor (LXR), farnesoid X receptor (FXR), peroxisome proliferator-activated receptor (PPAR) α, and pregnane X receptor (PXR), are involved in glucose, lipid, and drug metabolism in the liver. [28]We verified whether these transcription factors are functional in iHLOs cultured with HDM.After treatment with each ligand (LXR agonist: 1 μM T0901317; FXR agonist: 1 μM GW4064; PPARα agonist: 1 μM GW7647; PXR agonist: 20 μM rifampicin), the expression of target genes of each nuclear receptor was significantly induced (Figure 3C), suggesting that these receptors expressing in iHLOs are active at the protein level.
Hepatocytes accumulate excess energy as lipid droplets consisting of neutral lipids, including triacylglycerols.Because hepatocytes have a high capacity to synthesize triacylglycerol from free fatty acids, we examined whether iHLOs could store lipid droplets within cells upon loading with OAs.After culturing with HDM for three days, iHLOs were loaded with 500 μM OA.Time-course analyses revealed that the accumulation of intracellular lipid droplets, which were stained with 4,4-difluoro-1,3,5,7,8-pentamethyl-4-bora-3a,4adiaza-s-indacene (BODIPY 493/503), increased in proportion to the culture period (Figure 3D).Quantitative analysis also confirmed the Scale bar, 200 μm.(C) Relative mRNA levels of leucine-rich repeat containing G protein-coupled receptor 5 (LGR5), alpha-fetoprotein (AFP), albumin (ALB), keratin 19 (KRT19), vimentin (VIM), and CD68 in iHLOs cultured in HDM or EM for six days just after differentiation (passage 1, P1) or passaged in EM twice every six days (P2 and P3) were determined via quantitative reverse transcription-polymerase chain reaction (RT-qPCR) and normalized to 18s.The assay was performed in n = 3 independent biological replicates (mean ± standard error of the mean [SEM]).(D) iHLOs proliferated in EM for six days were passaged and cultured with EM or HDM for three days.Relative mRNA levels of LGR5, AFP, ALB, KRT19, VIM, and CD68 were determined via RT-qPCR and normalized to 18s.The assay was performed in n = 3 independent biological replicates (mean ± SEM).**P < 0.01 (versus EM).(E) After iHLOs were cultured with EM or HDM for six days, each medium was replenished with a new medium.After 24 h of culture, media were independently harvested.The concentrations of AFP and ALB were determined using an enzyme-linked immunosorbent assay (ELISA).The assay was performed in n = 3 independent biological replicates (mean ± SEM).**P < 0.01 (versus EM).See also Figure S1.3E).Furthermore, the expression of tumor necrosis factor, a marker of inflammation, was increased by OA treatment (Figure 3F).These results suggest that iHLOs maintained with HDM after expanded with L-RHF2 CM can recapitulate the physiological metabolism and function of hepatic cells.

Lipid metabolism-related genes are highly expressed in HLOs
As mentioned above, matured HLOs showed several liver-specific functions.We next performed extensive mRNA expression analysis to further elucidate the physiological properties of HLOs.We focused on pathways involved in nutrient and drug metabolism and compared the representative gene expression profiles among PHHs, established cell lines (HepaRG, HepG2, and HuH-7 cells), and HLOs (iHLOs and pHLOs).The expression levels of lipid metabolism-related genes, which mediate fatty acid, triacylglycerol, and cholesterol synthesis, fatty acid β-oxidation, LDL uptake, and VLDL secretion in HLOs, were generally comparable to or higher than those in PHHs and established cell lines (Figure 4; Figure S6).Notably, iHLOs and pHLOs exhibited similar gene expression patterns, and many genes were expressed at higher levels when cultured in HDM than in EM.In contrast, the expression levels of genes involved in other metabolic pathways, including drug metabolism, glycolysis, gluconeogenesis, and the urea cycle, were not necessarily high, despite the increase in the expression of most genes upon changing the culture medium from EM to HDM (Figure S5).These results suggest that HLOs are useful for evaluating lipid metabolism-associated events.

HLOs secrete VLDL at physiological levels, unlike HepG2 cells
[31] We investigated whether iHLOs can secrete mature VLDL, whose density is similar to that in vivo.We confirmed that apoB secretion was significantly elevated by changing the medium from EM to HDM in iHLOs (Figure 5A).Considering that VLDL and LDL contain a single molecule of apoB, especially apoB-100, per particle, [32] the results indicate that the hepatic function of VLDL secretion was enhanced by culturing iHLOs in HDM.Subsequently, we performed iodixanol (OptiPrep) density gradient ultracentrifugation to isolate each lipoprotein fraction, and apoB-100 was analyzed in each fraction to monitor VLDL and LDL levels.Consistent with a previous report, [30] HepG2 cells rarely secreted apoB-100 in the VLDL fraction (0.950-1.006 g mL -1 ), accounting for only 1.9% of the total apoB-100 proteins (Figure 5B).Intriguingly, iHLOs could secrete high amounts of apoB-100 in the VLDL fraction, accounting for 54.9% of the total apoB-100 proteins (Figure 5C), a proportion closer to that secreted from PHHs. [33]VLDL secretion deficiency in HepG2 cells is reported to be caused by mitogen-activated protein kinase kinase (MEK)/extracellular signal-regulated kinase (ERK) signaling pathway hyperactivation. [30]e proportion of apoB-100 in the VLDL fraction of HepG2 cells increased to 27.6% after treatment with 30 μM PD98059, a cell permeable MEK inhibitor, whereas it did not increase in iHLOs after PD98059 treatment (Figures 5B and 5C).Considering that PD98059 effectively inhibited MEK phosphorylation in whole cell lysates from iHLOs (Figure 5D), activated MEK expressed in HepG2 cells but not in iHLOs is responsible for the impairment of physiological secretion of VLDL.Notably, PD98059 failed to affect the amount of apoB-100 secretion in either HepG2 cells or HLOs (Figure 5E).Collectively, iHLOs can secrete VLDL in a physiological manner similar to PHHs, and can be considered a reliable model to unveil the molecular mechanism of VLDL secretion in humans.

Transient 2D culture achieves effective lentiviral gene transduction into HLOs
Organoids are usually cultured in a 3D environment and embedded in an extracellular matrix gel, making it hard to transduce exogenous genes with high efficiencies.Previously, we developed a simple method to efficiently transduce any gene in HIOs via a transient 2D culture of HIOs, followed by their culture in a 3D environment to regenerate organoids. [18]Here, we examined whether this method for HIOs applies to HLOs.Before gene transduction, iHLOs disrupted by enzymatic digestion were cultured on collagen I-coated (3000, 1000, and 300 ng mL -1 ) and undiluted CM of L-RHF2 cells cultured at a density of (B) 2.5-20 × 10 6 cells 100 mm -1 for 72 h or (C) 5.0 × 10 6 cells 100 mm -1 for 24-120 h using an anti-RSPO1 antibody.Right graph denotes the densitometric analysis to determine RSPO1 concentration in CM. (D) Coomassie brilliant blue (CBB) staining or western blotting analysis of the time-course changes in RSPO1, HGF, FGF7, and FGF10 protein levels of undiluted CM of L-RHF2 cells cultured at a density of 5.0 × 10 6 cells 100 mm -1 for 72 h from passages 6-26 using the respective antibodies.(E) Bright-field full-focused images (Left) or cell viabilities (Right) of iHLOs cultured in 25% L-RHF2 CM or the corresponding recombinant proteins for six days after passages.Scale bar, 500 μm.Cell viability was determined using CellTiter-Glo 3D reagents (n = 4 independent biological replicates, mean ± SEM).**P < 0.01 (versus recombinant proteins).(F) Images (Left) or cell viabilities (Right) of iHLOs cultured in 25% L-RHF2 CM stored for one month or one year at −80 • C for six days after passages.Scale bar, 500 μm.Cell viability was determined using CellTiter-Glo 3D reagents (n = 4 independent biological replicates, mean ± SEM).**P < 0.01 (versus recombinant proteins).See also Figures S2-4.plates, and we confirmed the emergence of 2D-proliferative cells, as in the case of HIOs (Figure 6A).We selected Venus, a modified green fluorescent protein, [34] as the exogenous gene to be transduced, which enables monitoring transduction efficiency.Dispersed iHLOs were subjected to lentivirus-concentration-dependent infection in collagen I-coated plates, cultured in EM for three days, and re-embedded in Matrigel.After iHLOs were regenerated, almost all organoids expressed Venus even when the virus was diluted 27-fold (Figure 6B), indicating highly efficient gene transduction.We confirmed virus concentration-dependent Venus expression via western blotting analysis (Figure 6C).We also confirmed that efficient gene transduction was successful through lipofection using commercially available Lipofectamine Stem Transfection Reagent (Figure S7A, S7B, and S7C).Finally, we attempted to transduce a functional gene into iHLOs.Among genes involved in lipid metabolism, we focused on Plin2, which accelerates lipid droplet formation in the liver, potentially leading to hepatic steatosis. [35]Following the same procedure as above, we infected iHLOs with a lentiviral plasmid for murine Plin2 overexpression (pCSII-EF-Plin2-C-Flag-IRES2-Venus), the biological activity of which was shown previously in adipocytes. [17]We confirmed the high infection efficiency in iHLOs by fluorescence microscopy observation of internal ribosome entry site-driven Venus expression (Figure 6D), as well as exogenous Plin2 expression by RT-qPCR and western blotting analysis (Figures 6E and 6F).We found that iHLOs accumulated more intracellular lipids by Plin2 overexpression (Figure 6G), clearly indicating that the exogenous Plin2 functioned in iHLOs.Furthermore, to show that lipofection-based gene transduction enables the examination of exogenous gene function, we overexpressed the nuclear form of SREBP-1c into iHLOs using Lipofectamine Stem Transfection Reagent, because SREBP-1c upregulates expression of genes involved in the fatty acid synthesis and plays a critical role in hepatic lipid metabolism under both physiological and pathological conditions. [27]llowing the process used for Venus gene transduction as indicated above, we confirmed the overexpression of the SREBP-1c mRNA and the increase in gene expression of SREBP-1c target genes in iHLOs (Figure S7D).Notably, iHLOs could maintain both proliferative capacity and marker gene expression levels after being subjected to transient 2D culture (Figure S8), indicating that the fundamental properties of iHLOs were unchanged by the gene transduction process.These findings indicate that the refined gene transduction method enables unraveling the physiological functions of exogenous genes, which would widen the applications of HLOs in various research fields of life science.

DISCUSSION
The liver plays a crucial role in maintaining systemic homeostasis by regulating nutrient and xenobiotic metabolisms.As decreased liver function leads to multiple fatal diseases, the establishment of a faithful human liver model is considered vital for research.The established cancer cell lines and animal models cannot recapitulate human liver biology [5] and PHHs and iPSC-derived hepatocytes are nonproliferative and thus not necessarily suitable for routine or large-scale use.HLOs are promising alternative biological models that proliferate and exhibit physiological functions of human hepatocytes.
The liver possesses a high regenerative capacity after surgical resection or chemical damage. [36]Hepatocyte proliferation during liver regeneration is stimulated by the activation of Wnt, EGF, FGF, and HGF signaling pathways, followed by the induction of Lgr5, Oct3/4, Nanog, and cMyc expression. [37,38]EM contains factors that activate these signaling pathways, and we showed that culture with EM can expand iHLOs and pHLOs, which involves increased LGR5 expression but decreased ALB expression.In consideration of this result and the previous report that ALB expression and hepatic function decline in proliferative hepatocytes, [22] the growth of HLOs with EM mimics the in vivo hepatocyte regeneration process.HDM are made from HCM, a commercially available medium for PHH culture, supplemented with dexamethasone, oncostatin M, and HGF, and were used for the final differentiation step into hepatocytes in a previous study. [13]We found that switching the culture medium from EM to HDM halts HLO growth, accompanied by high expression of ALB and low expression of LGR5.Hepatic maturation of pHLOs with differentiation medium was also reported previously, [39] although the medium composition was somewhat different from that of HDM.Once hepatocyte proliferation is terminated, the expression of hepatocyte nuclear factor 4α and CCAAT/enhancer-binding protein α is induced, and hepatocyte maturation is associated with an increase in ALB expression. [40,41]Therefore, it is plausible that culture with HDM reflects the hepatic maturation process at least partially, which may contribute to maintaining liver integrity.AFP expression is known to be induced during liver regeneration [42] ; therefore, an increase in AFP in iHLOs and pHLOs  F I G U R E 4 Higher expression of lipid metabolism genes in HLOs compared to conventional human hepatocyte models.Heat map of the expression of genes among various types of hepatic cells (primary human hepatocytes [PHHs] cultured for 4 h, HepaRG cells, HepG2 cells, HuH-7 cells, iHLOs cultured in EM, iHLOs cultured in HDM, primary HLOs [pHLOs] cultured in EM, and pHLOs cultured in HDM).Expression levels were determined via RT-qPCR and normalized to 18s.Relative mRNA levels of each gene were represented as ratios of those in PHHs before plating.Dark color indicates undetectable expression.A red square denotes lipid metabolism-related genes.See also Figure S5.cultured with HDM after the regeneration process with EM would be physiologically relevant.
We succeeded in establishing L-RHF2 cells, and the use of CM enabled significant cost reduction (by an estimated 1/100) of HLO culture compared to the cost when commercial products are used.
We determined the concentrations of growth factors in the culture supernatant as a mixture of the respective specific bands.
Unexpectedly, HLOs proliferated more efficiently when cultured in L-RHF2 CM than when cultured with recombinant proteins.Given that the calculated concentrations of growth factors in the EM with 4-fold diluted L-RHF2 CM were somewhat lower than those with recombinant proteins, chances are high that the relative activities of the exogenous growth factors in L-RHF2 CM were higher than those of the recombinant proteins.Although FGF7 in the CM, but not the recombinant FGF7, was N-glycosylated, evidence suggests that N-glycosylation is not a determinant factor for FGF7 activity. [24]One possible reason for the difference between L-RHF2 CM and recombinant proteins might be the presence of bovine serum in the CM, which may have stabilized secreted exogenous factors.In this study, lipid metabolism-related genes in HLOs were expressed at levels equal to or higher than those in PHHs, suggesting functional maturation of lipid metabolism in HLOs.In contrast, expression of gluconeogenesis-related genes was low in HLOs, which could be due to insulin supplementation in HDM that potently inhibits gluconeogenesis. [43]Nevertheless, some cytochrome P450 (CYP) family members (CYP1A2, CYP3A4, CYP2E1), involved in drug and alcohol metabolism, were expressed at significantly lower levels in HLOs than in PHHs or HepaRG cells, which are established cell lines, but highly express CYP family members. [44]These results imply that hepatic maturation is somewhat biased, depending on the metabolic pathway.Notably, the expression profiles of iHLOs and pHLOs were similar.Given that organoids established from primary tissues reflect the maturity of the original tissues, [8] this bias would not be from the cell potential (i.e., immaturity of iHLOs) but from the composition of the medium used to induce maturation.Although dexamethasone and oncostatin M-containing medium, similar to HDM, has been used for hepatic organoid maturation, [45] developing a maturation medium superior to HDM is necessary to maximize the hepatic function of HLOs.
Lipid metabolism in HLOs is expected to be highly functional; to test this, we focused on VLDL secretion, a phenomenon that cannot be accurately replicated in mice, rats, or cancer-derived cell lines.VLDL is secreted by hepatocytes and metabolized into LDL by lipoprotein lipase and hepatic triglyceride lipase in the blood.Elevated apoB-containing lipoproteins, followed by an increase in oxidized lipoproteins, are major risk factors for the development of atherosclerosis, eventually leading to fatal cardiovascular and cerebrovascular diseases. [46,47]Accordingly, regulating VLDL production in the liver is vital.To overcome the unphysiological aspects or species differences in traditional hepatic models, HLOs can become a promising analytical tool to elucidate the unresolved molecular mechanisms of human-specific VLDL synthesis, secretion, and metabolism.
Exogenous gene transduction has been used to investigate the role of the cellular functions of genes.However, conventional transduction methods designed for 2D-cultured cells have not been applied to 3D-cultured cells, including organoids.Although transient gene transduction into HLOs has been performed by electroporation, [16] this method requires a large number of cells and special equipment, and the transduction efficiency is not high.We previously developed a method to stably transduce genes of interest into HIOs at almost 100% efficiency through a transient 2D culture of dispersed HIOs. [20]Here, we demonstrated that the method could also be applied to HLOs using a fluorescent marker gene (Venus) as well as a hepatic functional gene (Plin2).HLOs stably expressing any gene can be readily established a week after lentiviral infection, because there are no requirements for separating transduced and non-transduced cells.We also showed that transient gene transduction into iHLOs was effective with lipofection.
During these transduction processes, 2D-cultured HLO-derived cells can be handled similarly when using cell lines in 2D conditions, which would be especially beneficial to researchers who do not have experience in 3D culture.Our methods may facilitate gene editing in HLOs using the clustered regularly interspaced short palindromic repeats (CRISPR)/CRISPR-associated protein 9 system, which would further accelerate the application of HLOs in the development of disease models or disease treatment strategies for drug discovery and regenerative medicine.
In summary, we prepared a systematic methodology to handle HLOs that can be adopted in both basic and applied sciences.Specifically, L-RHF2 CM can contribute to significant cost reduction in HLO culture, enabling the use of HLOs as common biological research tools.We
The organoids were fixed and permeabilized using a Cytofix/Cytoperm Kit (BD Biosciences).The samples were incubated overnight with BODIPY (1:200) at 4 • C.After washing three times with Perm/Wash Buffer, the specimens were incubated with Alexa Fluor 633-conjugated phalloidin (1:200; Thermo Fisher Scientific).After washing thrice with Perm/Wash Buffer, the specimens were further incubated with 5 μg mL -1 4′,6-diamidino-2-phenylindole for 10 min at room temperature and washed once with Perm/Wash Buffer, and fluorescence staining was visualized with fully focused z-stack images using an LSM 800 confocal laser-scanning microscope (Carl Zeiss) or an all-inone fluorescence microscope equipped with structured illumination (Keyence, Japan).

F I G U R E 1
Changes in marker gene expression in response to switching the culture medium in induced pluripotent stem cell (iPSC)-derived human liver organoids (iHLOs).(A) Schematic procedure for liver organoid differentiation from human iPSCs.(B) Bright-field full-focused images of iHLOs cultured in the hepatocyte differentiation medium (HDM) or expansion medium (EM) before and after the medium change or cell passage.

2
Establishment of mouse L cells simultaneously overexpressing R-spondin 1 (RSPO1), hepatocyte growth factor (HGF), fibroblast growth factor (FGF)−7, and FGF10 (L-RHF2) cells and use of conditioned medium (CM) for HLO culture.(A) Schematic illustration of the establishment of L-RHF2 cells and use of CM for HLO culture.(B, C) Western blotting analysis of human RSPO1 recombinant protein samples increased amounts of intracellular triacylglycerol upon treatment with 500 μM OA (Figure

F I G U R E 3
Characterization of HLOs as functional liver cells after expansion with EM. (A, B) iHLOs were cultured with EM for six days, serum-starved in Williams' Medium E for 18 h, and treated with or without 100 nM insulin for 6 h.(A) Western blotting analysis of whole cell lysates using the indicated antibodies.(B) Relative mRNA levels of sterol regulatory element-binding protein-1 (SREBP-1) target genes

5 F I G U R E 6
Physiological very low-density lipoprotein (VLDL) secretion in HLOs.(A) After harvesting culture supernatant of HepG2 cells or iHLOs (cultured with EM or HDM) during 24 h of culture, relative apolipoprotein B (apoB) amounts were determined using enzyme-linked immunosorbent assay (ELISA) kits and normalized to viable cell numbers determined using CellTiter-Glo 3D reagents.The assay was performed in n = 4 independent biological replicates (mean ± SEM).(B-E) HepG2 cells or iHLOs were incubated with or without 30 μM PD98059 for 18 h, and then loaded with 500 μM OA for 5 h.(B, C) Each culture supernatant was subjected to OptiPrep density gradient ultracentrifugation and each lipoprotein fraction was collected.The samples together with standard human apoB were then subjected to western blotting analysis using an anti-apoB antibody.Arrows indicate apoB-100.Lower panels denote the proportion of each lipoprotein determined via densitometric analysis, according to the density of each lipoprotein fraction.(D) Western blotting analysis of the whole cell lysates of iHLOs using anti-phospho-ERK1/2 (Thr202/Tyr204), anti-ERK1/2, and anti-β-actin antibodies.(E) Western blotting analysis of the culture supernatant of HepG2 cells and iHLOs together with standard human apoB using an anti-apoB antibody.Efficient gene transduction in HLOs via a transient 2D culture.(A) A phase-constant image of 2D-cultured iHLOs in collagen I-coated 6-well plates after enzymatic digestion, vigorous pipetting, and culture for two days.Scale bar, 200 μm.(B, C) After dispersed iHLOs were seeded on collagen I-coated 6-well plates, cells were infected with dose-dependent pCSII-EF-MCS-IRES2-Venus lentiviruses and cultured in EM for two days.The cells were re-embedded in Matrigel to regenerate organoids and cultured for six days.(B) Bright-field or fluorescent full-focused provide compelling evidence of the physiological properties of HLOs through VLDL secretion studies.Although there is room for improvement in terms of a maturation medium replacement for HDM, HLOs have great potential for the elucidation of unknown physiological functions of the liver, thereby contributing to the development of effective strategies to treat or prevent liver diseases.Hiro: Investigation.Tsuyoshi Kobayashi: Resources.Hideki Ohdan: Resources.Makoto Shimizu: Project administration.Yoshio Yamauchi: Project administration.Hiroshi Kiyono: Supervision.Ryuichiro Sato: Funding acquisition; Project administration; Supervision; Writingreview & editing