Hepatic spheroids derived from human induced pluripotent stem cells in bio-artificial liver rescue porcine acute liver failure

Dear Editor, Acute liver failure (ALF) is a complicated disorder showing a nearly 80% mortality rate, and there is currently no effective medical solutions for ALF except liver transplantation. Bioartificial liver (BAL) system is a device that consists of a bioreactor filled with hepatocytes, which could potentially rescue ALF patients by providing partial liver function until a suitable donor liver can be found or the native liver has self-regenerated. Currently, the lack of a stable and clinically applicable hepatocyte source has impeded the application of the BAL system for ALF treatment. The scarcity of human liver donors has made primary human hepatocytes (PHHs) an impractical cell source to meet the cell number requirement of the BAL system, which is in the order of magnitude of 10. Porcine hepatocytes are unsuited for application in BAL devices because of potential risks including immunogenic response and xenozoonosis, whereas hepatoma cell lines are limited by their incomplete functions. Human induced hepatocytes (hiHeps) using lineage reprogramming have been demonstrated to rescue ALF in large animal models. However, the limited lifespan of hiHeps leaves much to be desired for a stable and continuous cell source. One potential strategy of generating hepatocytes for application with the BAL system is the use of human induced pluripotent stem cell (hiPSC) lines, which could stably and perpetually selfrenew in vitro, serving as an optimal cell source for the generation of functional cell types. Moreover, the elimination of genomic integration and background oncogenic transgene expression makes hiPSC-derived cells a safe and promising cell source for clinical applications. However, it remains unknown whether hiPSC-derived hepatocytes can be efficiently generated to fulfill the requirement of large quantities of cells for the BAL device, and applied in the BAL system for ALF treatment. To establish a strategy to generate large quantities of hepatocytes from hiPSCs, we further optimized our previously reported protocol, inducing hiPSCs to differentiate into hepatic progenitor cells (hHPCs) (Fig. 1a, b; Supplementary information, Fig. S1a). The α-fetoprotein (AFP) and albumin (ALB) positive hepatic progenitor colonies could be enriched to high purity in vitro under a chemically defined medium (hHPC expansion medium) (Fig. 1d; Supplementary information, Fig. S1b, c), and expressed hepatic progenitor markers including AFP, KRT19, HNF1B and FOXA2 (Fig. 1c; Supplementary information, Fig. S2a). RT-qPCR and RNA-Seq analysis indicated that these hHPCs expressed hepatic progenitor-specific genes at similar levels to freshly isolated fetal human hepatocytes (FHHs), but distinctly from hiPSCs (Supplementary information, Fig. S2b, c). These hHPCs could also be successfully cryopreserved and recovered (Supplementary information, Fig. S1d, e). hHPCs could be further matured into functional hepatocytes when cultured in the hHPC maturation medium (Fig. 1a). Expression of hepatic progenitor marker AFP declined as maturation progressed within 20 days (Fig. 1e), following which hiPSC-derived mature hepatocytes (hMHs) showed a similar cell morphology to PHHs (Fig. 1b). The hMHs synthesized glycogen (Supplementary information, Fig. S3a), secreted albumin and urea (Fig. 1f; Supplementary information, Fig. S3b), detoxicated ammonia (Supplementary information, Fig. S4d), and expressed key CYP450 proteins (Fig. 1g; Supplementary information, Fig. S3c, d) and key hepatic transcription factors and functional genes (Supplementary information, Fig. S4a–c) at comparable levels to PHHs. RNA-Seq analysis of hMHs indicated that the expression patterns of important hepatic physiological functional genes were similar to those in PHHs (Supplementary information, Fig. S3e). We further developed a two-step amplifying culture system to generate sufficient quantities of human hepatocytes (Supplementary information, Fig. S5a). Firstly, the hiPSC-derived hHPCs were expanded using a cascade amplification process in adherent culture in hHPC expansion medium for more than 20 passages in vitro, maintaining normal karyotype without alterations in HPC characteristics and proliferation rate (Fig. 1h; Supplementary information, Fig. S5b, c). Secondly, the hHPCs were transferred into a low-speed stirring culture system for floating culture in hHPC maturation medium to generate mature hepatic spheroids on a large scale. These mature hepatic spheroids expressed mature hepatic markers ALB and CYP3A4 (Fig. 1i), secreted albumin and urea (Supplementary information, Fig. S5d), and expressed important mature hepatic genes at comparable levels to PHHs (Supplementary information, Figs. S5e, 6). To rescue ALF pigs, hiPSC-derived hepatic spheroids (hHSs; ~1 × 10 cells) suspended in culture medium were assembled into our previously reported multilayer BAL device, where medium was exchanged with pig blood plasma (Supplementary information, Fig. S7a). Eighteen adult Bama miniature pigs were treated with D-gal to induce ALF. The pigs developed severe ALF symptoms in 24 h with significant increase in liver failure indices including alanine aminotransferase (ALT), aspartate aminotransferase (AST) and ammonia levels in the blood (Supplementary information, Fig. S7a, b and Tables S2–4). These animals were randomly assigned into three groups: a non-treated group (No-BAL, n= 6) in which no BAL was used, and two BAL-treated groups with BAL containing hHSs (hHS-BAL, n= 6), or no biomaterial (Empty-BAL, n= 6). Following 4-h treatment, the health status of the pigs in the hHS-BAL group showed apparent improvement 24 h after treatment (Supplementary information, Fig. S7c), and all six hHSBAL-treated pigs successfully recovered and survived (Fig. 1j; Supplementary information, Table S1a). A significant downregulation of ALT, AST, ammonia and total bilirubin (TBIL) levels in the blood could be detected in the ALF pigs of the hHS-BAL group (Fig. 1k; Supplementary information, Tables S2–4). In the other two groups, the pigs died within 4 days due to complications from ALF (Fig.1j; Supplementary information, Table S1b, c). The surviving pigs were sacrificed on day 7 to

Fetal human hepatocytes (FHHs) were isolated from 14-week-old human embryos obtained from three abortion donors with informed patient consent. The fetal liver tissue was digested in DMEM/F12 medium supplemented with 1 mg/ml collagenase IV (Gibco) in a 37 °C incubator for 20 min, then dissociated to obtain single cells by repeated pipette action. The single cell suspension was centrifuged at 1000 rpm for 3 min, washed 3 times with DMEM/F12 medium, then plated on Matrigel (1:50 dilution) at 1 × 10 6 cells per well of a 12-well plate in hHPC expansion medium. The adherent cells were collected after 2-h culture for detection and analysis.
Primary human hepatocytes (PHHs) were isolated from leftover human donor livers resections. Liver tissue were perfused with collagenase IV until the tissue was no longer compact. Digested tissue was then separated with Adson forceps. The single cell suspension was washed 3 times with HCM (Lonza) and collected for detection and analysis.
HepG2 cells were cultured in DMEM supplemented with 10% FBS.

Large-Scale Expansion and Maturation
hHPC expansion was conducted with initial cell number of 1 × 10 7 cells, and expanded to ~1 × 10 9 cells with continuous cell passage (Supplementary information, Fig. S5a). For cell passage, hHPCs were digested with Accutase (Merck) in a 37 °C incubator for 3-5 min, after which cells were collected and centrifuged at 1000 rpm for 3 min. Single cells were resuspended and plated at a split ratio of 1:2-1:3 onto cell culture dishes pre-coated with Matrigel (1:50 dilution).
After hHPCs were expanded to ~1 × 10 9 cells, the hHPCs were transferred into a low-speed stirring floating culture system with 1 L hHPC expansion medium for ~3 days. The medium was then replaced with hHPC maturation medium and cultured for ~20 days to generate functional hepatic spheroids. The magnetic stirring apparatus was applied at 60 rpm for both these steps.
The functional hepatic spheroids were then transferred into cell bags and transported from Beijing to Nanjing, where they were cultured overnight. The hepatic spheroids suspended in 500 ml hHPC maturation medium (without sera) were assembled into the substratum of the bioreactor. The assembled BAL device was maintained at 37 °C and 5% CO2 in an incubator throughout the treatment duration.

Bio-artificial Liver System
Chinese Bama miniature pigs (~50 kg, either sex) were purchased from the Laboratory Animal Center of the Affiliated Drum Tower Hospital of Nanjing University Medical School. All animal procedures were performed according to institutional and national guidelines and approved by the Animal Care Ethics Committee of Nanjing University and Nanjing Drum Tower Hospital.
The hiPSC-derived hepatocyte BAL support system consisted of three circuits, two blood circuits and one cell circuit (Supplementary information, Fig. S7a). The whole system included three roller pumps, a heparin pump, an infusion heater, a plasma filter (Sorin Group Italia, Mirandola, Italy), a plasma component separator (Kawasumi Laboratories Inc, Tokyo, Japan), an oxygenation device, and a multilayer radial-flow bioreactor containing polystyrene nanofiber scaffolds. The incubator of the bioreactor and oxygenation device was maintained at 37 °C.
Pigs in all groups were intravenously injected with D-galactosamine (D-gal, Sigma, G0500) (0.4 g/kg) the day before BAL treatment (day 0), and the baseline blood sample was collected. Catheters were inserted into the internal jugular vein of the pigs under continuous anesthesia by intravenous administration of propofol (10 mg/kg/h) (Diprivan; AstraZeneca, Wuxi, China), and the catheters were then connected to the BAL device. Whole blood of ALF pigs was channeled through the BAL devices for 4 h, which consisted of 3 circuits. Whole blood was perfused at a rate of 40 ml/min in the first blood circuit, following which plasma was separated via the plasma filter at a rate of 15 ml/min the second blood circuit. Finally, the plasma was channeled through the plasma component exchanger, counter current to the flow of the cell circuit for sufficient exchange, at a rate of 15 ml/min. After each BAL treatment, the culture medium was drained and hiPSC-derived hepatic spheroids were washed in hHPC maturation medium. Whole blood samples were collected at regular intervals (before treatment, 1-7 days after treatment) until completion of the study 7 days after treatment.  Table S6.

Flow Cytometry
Differentiated cells were dispersed into single-cell suspensions by digestion in Accutase at 37 °C for 3-5 min and then washed in PBS. The cell suspension was fixed in fixation/permeabilization solution buffer (BD, 554714) at 4 °C for 20 min and washed 3 times with perm/wash buffer (BD, 554723). The cells were then resuspended in perm/wash buffer with primary antibodies and incubated at 4 °C overnight. Cells were then washed 3 times with wash buffer and incubated in perm/wash buffer with secondary antibodies at for 1-2 h at 4 °C. The cells were then washed three times and analyzed using a flow cytometer (Beckman CytoFlex).

Human Albumin and Urea Detection
Human albumin was measured using the Human Albumin ELISA Quantitation kit (Bethyl Laboratory, E80-129) according to the manufacturer's instructions. Urea synthesis was measured using the QuantiChrom Urea Assay Kit (BioAssay System, BA_DIUR-500) according to the manufacturer's instructions.

Ammonia Elimination Detection
To detect the ammonia elimination abilities, cells were incubated in hHPC maturation medium supplemented with 200 μM NH4Cl for 24 h. NH4 + concentrations were detected by Ortho Clinical Diagnostics (FS5600, Johnson and Johnson Company).

PAS Staining
The PAS staining system was purchased from Sigma-Aldrich. Cultures were fixed with 4% paraformaldehyde (DingGuo) and stained according to the manufacturer's instructions.

Immunofluorescence
Cultured cells were washed with PBS and fixed in 4% paraformaldehyde at room temperature for 15 min. They were then washed 3 times with PBS for 3 min and then blocked with PBS containing 0.25% Triton X-100 and 5% normal donkey serum (Jackson ImmuneResearch Laboratories, Inc) at room temperature for 1 h or at 4 °C overnight. The samples were incubated with primary antibodies at 4 °C overnight, washed three times with PBS, and then incubated with the appropriate secondary antibodies for 1 h at room temperature in the dark. Nuclei were stained with DAPI (Roche). The primary and secondary antibodies used for immunostaining are listed in Supplementary information, Table S7.

Growth curve and doubling times
To calculate the doubling time of hHPCs in the expansion stage, cells were plated at a density of 5 × 10 5 cells per well in hHPC expansion medium and cultured in a 12-well plate coated with Matrigel (1:50). The growth rate was determined by counting the cells number using a hemocytometer as a function of time. Data from the exponential phase of growth (data points at each passage time, passage ratio 1:2 to 1:3) were used to generate an exponential growth curve.

CYP Metabolism Assay: LC-MS
The hMHs and PHHs were dissociated and suspended to measure CYP450 activities.
The cell suspension was diluted to 1 × 10 6 cells/ml in reaction buffer, which consisted of William's E medium (Gibco) with Glutamax (Gibco, 100×) and HEPES (Gibco, 100×). The cell mixture was 1:1 resuspended in a 2× concentration of substrate buffer in a 500 µl reaction system. After 15-min incubation at 37 °C in an orbital shaker (~200 rpm/min), the reactions were stopped by addition of sample aliquots to the tubes containing triple the volume of quenching solvent (methanol) and frozen at -80 °C. Isotope-labeled reference metabolites were used as internal standards for further LC-MS analysis. The metabolites were quantified using a Triple Quad 4500 using validated traditional LC-MS methods.

RNA-Seq analysis
Total RNA of hiPSCs, hHPCs, hMHs, FHHs and PHHs were isolated using the RNeasy Micro Kit (Qiagen) and then sent to a private sequencing company for detection and analysis. RNA sequencing libraries were prepared with the IlluminaTruSeq RNA Sample Preparation Kit. The fragmented and randomly primed 200-bp paired-end libraries were sequenced on the IlluminaHiSeq 4000 sequencing system.

Statistics
For most statistic evaluation, an unpaired t-test was applied to calculate statistical probability in this study. P values were calculated by two-tailed test.       ALF group. Healthy porcine liver sample served as negative control. Samples were selected randomly to represent the No-BAL ALF group (n = 2), the hHS-BAL treatment ALF group (n = 6) and the healthy porcine liver samples (n = 2). Data are presented as means ± SEM. t-test, *P < 0.05, **P < 0.01. e Long-term observation of an additional hHS-BAL-treated ALF pig. Two months after BAL treatment, this pig showed excellent survival, without signs of ALF symptoms. The blood physiological indices of this pig including ALT, AST, ammonia and TBIL levels were restored to healthy levels in 7 days, and they were subsequently maintained.

Supplementary information, Table S1
Information of ALF pigs Summary of experimental groups: Weight and survival times of ALF pigs as well as total cell number contained in their respective BAL devices in the hHS-BAL treatment group (A), non-treated group (B) or empty-BAL treated group (C).