|Home | About | Journals | Submit | Contact Us | Français|
The ability to generate induced pluripotent stem (iPS) cells from a patient’s somatic cells has provided a foundation for organ regeneration without the need for immune suppression. However, it has not been established that the differentiated progeny of iPS cells can effectively reverse failure of a vital organ. Here, we examined whether iPS cell–derived hepatocytes have both the functional and proliferative capabilities needed for liver regeneration in mice with fumarylacetoacetate hydrolase deficiency. To avoid biases resulting from random genomic integration, we used iPS cells generated without viruses. To exclude compensation by hepatocytes not derived from iPS cells, we generated chimeric mice in which all hepatocytes were iPS cell derived. In vivo analyses showed that iPS cells were intrinsically able to differentiate into fully mature hepatocytes that provided full liver function. The iPS cell–derived hepatocytes also replicated the unique proliferative capabilities of normal hepatocytes and were able to regenerate the liver after transplantation and two-thirds partial hepatectomy. Thus, our results establish the feasibility of using iPS cells generated in a clinically acceptable fashion for rapid and stable liver regeneration.
The discovery that overexpression of a few transcription factors suffices to reprogram somatic cells to pluripotency has paved the way for organ regeneration with autologous cells not requiring immune suppression (1). Further progress toward the clinic was recently made by generating induced pluripotent stem (iPS) cells free of viral vector and transgene sequences, a process which prevents potential hazards from mutagenesis or persistent pluripotent gene expression (2). The first evidence for the therapeutic potential of iPS cells has come from studies of the effects of iPS cell–derived hematopoietic cells, neurons, or endothelial cells in animal models of human diseases (3–5). However, whether the differentiated progeny of iPS cells are capable of regenerating a failing vital organ remains to be determined.
The liver is a vital organ that represents a promising target for cell therapy, because of its ability to functionally integrate transplanted hepatocytes. Pioneering trials in patients and studies in animal models have revealed that successful liver regeneration requires that transplanted cells have both the complex functional and unique proliferative capabilities of fully differentiated hepatocytes (6, 7). Otherwise, replacing approximately 20% or 1 × 1010 of the hepatocytes in the liver, as is required for restoration of liver function in most human liver diseases, can neither be achieved nor sustained (8).
It has recently been shown, using tetraploid embryo complementation, that iPS cells can develop into all cell types in mice (9–11). However, it is unknown whether iPS cell–derived hepatocytes have the same functional output as normal adult hepatocytes and share their ability to respond to injury by rapidly entering the cell cycle and proliferating extensively. In fact, recent reports have challenged the therapeutic potential of iPS cells by showing that they differentiate and proliferate less efficiently than ES cells (12, 13).
Here, we aimed at defining the effectiveness of iPS cells in organ regeneration by investigating the functional and proliferative capabilities of iPS cell–derived hepatocytes. To exclude biases due to viral genomic integration or persistent pluripotent gene expression, we used iPS cells free of foreign DNA (14). Instead of tetraploid embryo complementation, which is inefficient and has yet to be achieved with iPS cells free of foreign DNA, we used blastocysts deficient in fumarylacetoacetate hydrolase (FAH) to generate chimeric mice in which progressive liver repopulation with iPS cell–derived hepatocytes could be induced postnatally. Since the entire liver was composed of iPS cell–derived hepatocytes by the time the mice reached adulthood, we were able to analyze the level of liver function and the efficacy of liver regeneration afforded by differentiated iPS cell–derived hepatocytes in the absence of compensation by non–iPS cell–derived hepatocytes.
Genetic deficiency of FAH, a tyrosine-degrading enzyme expressed specifically in hepatocytes in the liver, causes the liver disease tyrosinemia type I. Unless treated with the drug 2-(2-nitro-4-fluoromethylbenzoyl)-1,3-cyclohexanedione (NTBC), many humans and all mice lacking FAH develop severe hepatocyte damage and die from liver failure in the neonatal period (15). The liver injury caused by FAH deficiency confers a selective growth advantage on wild-type hepatocytes. Because hepatocytes have stem cell–like proliferative capabilities, a relatively small number of transplanted wild-type hepatocytes can completely repopulate the FAH-deficient liver (7). However, only cells that have the same functional and proliferative capabilities as normal adult hepatocytes are effective in liver repopulation. Cells that exhibit incomplete differentiation and impaired proliferation, a prominent example being hepatocyte-like cells generated from ES cells with current in vitro differentiation protocols, fail to restore liver function in FAH-deficient mice or similar liver repopulation models (16–18).
To definitely determine the regenerative capabilities of iPS cell–derived hepatocytes, we decided to subject these cells to the rigorous testing of hepatocyte function and proliferation afforded by the FAH-deficient liver, while avoiding the limitations of current in vitro differentiation protocols. To achieve this, we generated chimeric mice by injection of iPS cells into blastocysts obtained from FAH-deficient mice. As an example of iPS cells free of foreign DNA, we used the iPS cell line 440A-3 that was generated by repeated transfection of mouse embryonic fibroblasts with plasmids expressing Oct4, Sox2, Klf4, and c-Myc (14). To obtain neonates with a stochastic contribution of iPS cells to all tissues, we matched the mouse strain background of the iPS cells and the blastocysts. Moreover, we treated the foster mothers carrying the iPS cell–injected blastocysts with NTBC to prevent damage to FAH-deficient hepatocytes and maintain the stochastic liver chimerism until initiation of NTBC withdrawal after birth (15).
After transfer of 40 iPS cell–injected, FAH-deficient blastocysts into foster mothers, a total of 24 pups were born. In 8 of these mice, we found that iPS cells contributed to the tissues making up the digits (Supplemental Figure 1; supplemental material available online with this article; doi: 10.1172/JCI43267DS1). To investigate whether the livers of these pups or the 16 neonates without chimerism in digits (referred to herein as digital chimerism) harbored iPS cell–derived hepatocytes and whether their functional and proliferative capabilities sufficed to restore liver function, we deprived the foster mothers of NTBC by gradually lowering its concentration in the drinking water (15). In the absence of NTBC from the breast milk, the 6 neonates with high levels of digital chimerism continued to thrive, while the remaining 18 neonates lost weight and NTBC treatment had to be reinstated to prevent death (Figure (Figure1A).1A). Similarly, pups derived from FAH-deficient blastocysts that were not injected with iPS cells rapidly deteriorated and required rescue from liver failure with NTBC (Figure (Figure1A). 1A).
Next, we asked whether NTBC-independent survival from FAH deficiency was due to the presence of iPS cell–derived hepatocytes. We found no evidence for Fah gene or protein expression in liver tissue samples obtained from mice with no or low levels of digital chimerism at P28 or P70 (Figure (Figure1,1, B–D). This was not due to impaired differentiation of iPS cells, because genomic DNA isolated from the livers of these mice contained only the knockout Fah allele (data not shown). This shows that iPS cell contribution can vary between individual organs in chimeric mice. In contrast to mice with no or low levels of digital chimerism, mice with high levels of digital chimerism showed various levels of liver repopulation with FAH-expressing hepatocytes, ranging from approximately 50% repopulation at P28 (Figure (Figure1E)1E) to approximately 100% repopulation at P70 (Figure (Figure1F).1F). The extent of liver repopulation observed in tissue sections correlated with Fah gene expression levels in samples representative of the whole liver. Mice fully repopulated with iPS cell–derived hepatocytes produced approximately 80% of the Fah transcription of wild-type mice (Figure (Figure1,1, B and F). We found similar Fah expression levels in livers of chimeric mice fully repopulated with hepatocytes derived from ES cells, suggesting that iPS cell–derived hepatocytes are fully differentiated (Figure (Figure1B). 1B).
As direct evidence for the functional capabilities of hepatocytes derived from iPS cells, we found that liver repopulation with these cells led to complete restoration of liver function by P22 (Figure (Figure2).2). The less than 50% liver repopulation at this time suggested that the functional output and ability to rescue liver function of iPS cell–derived hepatocytes are similar to those of normal hepatocytes. Indeed, despite continued NTBC withdrawal, mice with high levels of digital chimerism grew and had completely normal liver function parameters at P70, when approximately 100% of the hepatocytes were of iPS cell origin (Supplemental Figure 2). Liver function tests continued to be normal at P300, that is after 10 months of NTBC withdrawal (Supplemental Figure 2). As additional evidence for normal hepatocyte differentiation, iPS cell–derived hepatocytes normalized plasma amino acid levels (Supplemental Table 1) and underwent physiological polyploidization (Supplemental Figure 3). In the absence of compensation by non–iPS cell–derived hepatocytes, these results show that iPS cells can give rise to hepatocytes that are fully and stably differentiated.
Levels of liver repopulation progressed from predominantly approximately 50% at P28 to invariably approximately 100% at P70 (Figure (Figure1,1, E and F). Since we initiated NTBC withdrawal at P6, this finding indicates that the number of iPS cell–derived hepatocytes present at birth was small and that these cells continuously proliferated to form small nodules by P28 and achieve approximately 100% liver repopulation by P70. Indeed, we found that cells in the periphery of the nodules present in chimeric mice at P28 expressed both the hepatocyte-specific protein FAH and the pan-proliferation marker Ki67 (Figure (Figure3A).3A). Since final hepatocyte number and differentiation are established by P21 in mice, these observations suggest that differentiated iPS cell–derived hepatocytes responded to FAH deficiency by dividing until all FAH–deficient hepatocytes were replaced (7).
In addition to hepatocytes, FAH is expressed in proximal tubular cells of the kidney, and its deficiency causes tubulopathy (15). Correction of liver function improves tubular function by restoring liver tyrosine catabolism (19). Therefore, we analyzed plasma creatinine and urine amino acid levels in the chimeric mice to further define the functional capabilities of iPS cell–derived hepatocytes (for details see Supplemental Methods). Since mild tubulopathy persists in FAH deficiency even after whole organ liver transplantation (19), we were surprised to find not only normal creatinine excretion but also normal reabsorption of urine amino acids (Supplemental Figure 4 and Supplemental Table 2). Considering that normal proximal tubular cells are positively selected in FAH-deficient kidneys (20), we asked whether iPS cell–derived proximal tubular cells had regenerated the kidneys of the chimeric mice. Indeed, we found substantial repopulation with iPS cell–derived proximal tubular cells (Supplemental Figure 4). Together, these findings suggest that iPS cell–derived renal proximal tubular cells provide therapeutically effective function and have the ability to proliferate extensively.
The ability of differentiated hepatocytes or renal proximal tubular cells to proliferate is an important prerequisite for both achieving and sustaining a therapeutic effect with cell therapy. To determine whether differentiated iPS cell–derived hepatocytes are effective in liver regeneration after transplantation, we isolated iPS cell–derived hepatocytes from a chimeric mouse and transplanted 1 × 106 of these cells into adult FAH-deficient mice. Similar to normal hepatocytes, iPS cell–derived hepatocytes rapidly repopulated the livers of these mice so that they became NTBC independent (Figure (Figure3,3, B and C). We then challenged livers constituted of approximately 100% of iPS cell–derived hepatocytes with two-thirds partial hepatectomy (21). Like wild-type mice and chimeric mice operated on at P28 and P70, all mice repopulated with transplanted iPS cell–derived hepatocytes tolerated two-thirds partial hepatectomy normally and survived long-term (Supplemental Figure 5 and data not shown). Moreover, we found that differentiated iPS cell–derived hepatocytes responded to two-thirds partial hepatectomy with the rapid and coordinated cell cycle entry and progression characteristic of normal liver regeneration (Figure (Figure3,3, B and C).
These findings suggested that iPS cell–derived hepatocytes can proliferate as rapidly and extensively as normal hepatocytes (22). As a definitive test, we directly compared the proliferative capabilities of iPS cell–derived hepatocytes with normal hepatocytes by performing a competitive liver repopulation experiment. We mixed an equal number of hepatocytes isolated from a highly repopulated chimeric mouse and an age-matched mouse ubiquitously expressing a β-galactosidase marker gene (Rosa26; ref. 23) and transplanted the cells into FAH-deficient mice. We used FAH-deficient mice that were also immune-deficient to exclude biases potentially caused by the marker gene (24). Analysis of the number and size of nodules forming in response to NTBC withdrawal in the livers of these mice revealed that iPS cell–derived hepatocytes had the same ability to continuously divide as normal hepatocytes (Figure (Figure3D).3D). Viewed together, our results establish that fully differentiated iPS cell–derived hepatocytes have normal proliferative capabilities.
Previously, we showed that fusion with wild-type hematopoietic cells corrects the hepatocytes of FAH-deficient mice, leading to therapeutic liver repopulation (25, 26). A recent study based on analyses in chimeric mice even suggested that hepatocyte fusion contributes to the physiological polyploidization of the liver (27). To prove that the FAH-expressing hepatocytes in iPS cell chimeras resulted from direct iPS cell differentiation and not from fusion of iPS cell progeny with blastocyst-derived hepatocytes, we determined that blastocyst-specific markers were absent in the vast majority of hepatocytes isolated from a highly repopulated chimeric mouse (Figure (Figure4,4, A and B, and Supplemental Figure 7). These analyses demonstrate that iPS cells free of foreign DNA are intrinsically able to generate fully functional hepatocytes.
In our study, after selective expansion of iPS cell–derived hepatocytes by postnatal NTBC withdrawal, all hepatocytes in the livers of adult chimeric mice were derived from iPS cells. In the absence of potential functional compensation by non–iPS cell–derived hepatocytes, our molecular and in vivo analyses established that iPS cell–derived hepatocytes are fully differentiated and that their functional output is equal to that of ES cell–derived or normal hepatocytes. Moreover, challenge by two-thirds partial hepatectomy or transplantation into adult FAH-deficient mice off NTBC showed that fibroblast origin and reprogramming did not diminish the regenerative capabilities of iPS cell–derived hepatocytes. The cells both rapidly entered the cell cycle and exhibited the extensive proliferative capabilities of normal differentiated hepatocytes.
The ability of iPS cell–derived proximal tubular cells to repopulate the kidney and correct tubulopathy extends our findings on the normal functional and proliferative capabilities of differentiated iPS cell progeny to another cell type. Recent studies showed that organ regeneration by proliferation of fully differentiated cells is not exclusive to the liver or kidney but also occurs in the pancreas (28, 29) and heart (30). Our results suggest that these proliferative capabilities could also be maintained in iPS cell–derived differentiated β cells or cardiomyocytes, which may enhance their therapeutic efficacy.
In conclusion, our results show that iPS cell–derived hepatocytes have normal functional and proliferative capabilities and are thus effective in rapidly and stably regenerating a failing liver. Since we obtained these results with iPS cells that are not subject to biases from viral genomic integration, they suggest that faithful recapitulation of hepatocyte differentiation in human iPS cells free of foreign DNA in vitro will yield an effective therapy for human liver diseases.
Procedures involving mice were approved by the Institutional Animal Care Committee at the University of California San Francisco. FAH-deficient mice on the 129S4 strain background were previously described (31). FAH-deficient mice on a pure C57BL/6 background were generated by breeding with wild-type C57BL/6 mice (The Jackson Laboratory). These animals were bred with C57BL/6 R26R mice (The Jackson Laboratory) to obtain FAH-deficient, R26R homozygous positive males for blastocyst generation. Wild-type control mice were derived from the 129S4 ES cells used to generate the original FAH-deficient mice. Thus, all control mice were on the 129S4 strain background, while the blastocysts used for iPS cell injection were F1 of 129S4 and C57BL/6 mice. Foster mothers and vasectomized males were on the ICR strain background (Taconic). Immune-deficient, FAH-deficient mice lacking B, T, and natural killer cells due to disruption of Rag2 and Il2rg were on a mixed 129S4 and C57BL/6 background (24). Rosa26 mice (The Jackson Laboratory) were on a C57BL/6 background (23).
The generation of the iPS cell line 440A-3 was previously described (14). The ES cells used for blastocyst injection were derived from blastocysts generated by a C57BL/6 with 129S1 mating. iPS and ES cells were cultured on confluent layers of irradiated primary mouse embryonic fibroblasts in Dulbecco’s modified Eagle medium supplemented with 2 mM l-glutamine, 1% nonessential amino acids, 100 U/ml penicillin, 100 U/ml streptomycin (all Gibco, Invitrogen), 15% fetal calf serum (Hyclone), 1,000 U/ml leukemia inhibitory factor (ESGRO, Chemicon), and 0.1 mM β-mercaptoethanol (Sigma-Aldrich).
Six- to eight-week-old FAH-deficient 129S4 females were superovulated by intraperitoneal injection of 10 units of pregnant mare’s serum, followed by injection of 10 units of human chorionic gonadotropin (both Sigma-Aldrich) 48 hours later. Superovulated females were mated with FAH-deficient C57BL/6 males homozygous positive for R26R to generate blastocysts that were FAH deficient and R26R heterozygous and had the same 129S4, C57BL/6 strain background as the iPS cells (14). At 3.5 days after conception (dpc), females were killed and their oviducts and uteri were flushed with M2 medium to obtain blastocysts that, prior to iPS cell injection, were incubated in KSOM medium (both EmbryoMax, Chemicon) at 37°C and 5% CO2. Blastocysts were injected with 10 to 15 iPS cells and transferred into the uterine horn of pseudopregnant (2.5 dpc) 6- to 8-week-old ICR mice.
The contribution of iPS cell–derived cells to digits and livers was determined using a PCR that simultaneously amplified the wild-type and knockout Fah allele in genomic DNA (31). In addition, established R26R (32) and GFP (33) PCR protocols were used to distinguish whether cells originated from the R26R heterozygous blastocyst or from the iPS cell line, 440A-3, that carries a GFP gene within the 5ι untranslated region of Nanog (14).
Foster mothers were given NTBC (7.5 mg/ml drinking water) during pregnancy. When the neonates were 6 days old, the concentration of NTBC was gradually reduced (50%, 25%, 12.5%, and 6.25%) every 2 days before drug treatment was halted. After 8 days of complete NTBC withdrawal, blood plasma was obtained by retroorbital bleeding. Body weights were recorded 0, 13, and 16 days after initiation of NTBC withdrawal. Wild-type and FAH-deficient pregnant females and neonates used as controls received the same NTBC treatment regimen. Adult immune-deficient, FAH-deficient mice were gradually taken off NTBC for 3 periods of 4 weeks, separated by 5-day-long recovery phases on NTBC.
Combining samples from various regions of the liver generated a representative sample. Total RNA was isolated with TRIzol (Molecular Research Center). After DNase1 (Ambion) treatment, 1 μg RNA was used for first strand cDNA synthesis with TaqMan Reverse Transcription Reagent (Applied Biosystems). PCR was performed in a 7300 Real-Time PCR System using SYBR Green (both Applied Biosystems) at 50°C for 2 minutes and 95°C for 10 minutes, followed by 40 cycles at 95°C for 15 seconds and 60°C for 1 minute. Fah expression was normalized to that of Gapdh. Primer sequences (forward/reverse) were 5′-AGGCTTCTGCGACAATGC-3′/5′-TGCTGCCGAGAAGAGTAGAAG-3′ for Fah and 5′-CATGGCCTTCCGTGTTCCT-3′/5′-GCGGCACGTCAGATCCA-3′ for Gapdh.
Liver samples were fixed at 4°C in 4% paraformaldehyde or periodate-lysine–2% paraformaldehyde (PLP) for 4 hours, cryoprotected in 30% sucrose (all Sigma-Aldrich), and embedded in optimum cutting temperature compound (Tissue-Tek, Sakura Finetek) (34). Frozen sections (10 μm) were stained with rabbit anti-FAH antibody (gift from Robert Tanguay, Université Laval, Québec City, Québec, Canada) diluted 1:15,000, rat anti-BrdU antibody (Abcam) diluted 1:100, mouse anti-Ki67 antibody (BD Pharmingen) diluted 1:25, or mouse anti-phosphorylated histone H3 antibody (Cell Signaling Technology) diluted 1:25. For fluorescence microscopy, primary antibodies were detected with goat anti-rabbit antibody conjugated with Alexa Fluor 594 or Alexa Fluor 488, goat anti-rat antibody conjugated with Alexa Fluor 594 (all Molecular Probes), diluted 1:500, and the M.O.M. Fluorescein kit (Vector). Nuclear DNA was stained with 2 μg/ml DAPI (Molecular Probes). For light microscopy, FAH immunocomplexes were detected with a swine anti-rabbit antibody conjugated with alkaline phosphatase (Dako) diluted 1:40. Alkaline phosphatase was visualized with Fast Red Substrate System (Dako).
Plasma was diluted 1:4 in 0.9% sodium chloride and assayed for total bilirubin, albumin, alkaline phosphatase, and alanine aminotransferase with an ADVIA 1800 (Siemens) chemistry analyzer.
Two-thirds of the liver was surgically removed as described previously (21). One hundred μg/g body weight 5-bromo-2-deoxyuridine (BrdU, Roche) were injected intraperitoneally 40 hours after surgery. Immunostainings for BrdU incorporation and phosphorylation of histone H3 were performed on liver tissue obtained 48 hours after surgery.
Hepatocytes were isolated from mouse liver by perfusion through a 24-gauge intraportal catheter (Surflo, Terumo) with Earle’s Basic Salt Solution (EBSS, Invitrogen) plus 0.5 mM EGTA and 10 mM HEPES (both Sigma-Aldrich), followed by EBSS plus 0.1 mg/ml Liberase Blendzyme 3 (Roche). Hepatocytes were released from the liver and filtered through a 70-μm cell strainer (BD Falcon). Gravity sedimentation on ice for 30 minutes was used to enrich viable hepatocytes to approximately 95%.
Hepatocytes were isolated from a highly repopulated chimeric mouse and a Rosa26 mouse, and 5 × 105 hepatocytes from each mouse were mixed and injected intrasplenically into immune-deficient, FAH-deficient mice. iPS cell–derived and Rosa26 hepatocytes were female, while recipient mice were male. Perioperative antibiotic treatment was applied as reported (24).
An adenoviral vector expressing Cre recombinase from the cytomegalovirus promoter was obtained from the Vector Development Laboratory at Baylor College of Medicine, Houston, Texas, USA. Four hours after plating on Primaria dishes (BD Falcon), hepatocytes were incubated with the virus for 1 hour at MOIs of 4, 10, 40, and 100.
Twenty-four hours after adenoviral infection, hepatocytes were fixed with 0.5% glutaraldehyde (Sigma-Aldrich). Both plated hepatocytes and PLP-fixed liver sections were exposed to 0.8 mg/ml 5-bromo-4-chloro-3-indolyl-β-d-galactopyranoside (X-gal, Sigma-Aldrich) for 12 hours at 37°C.
The height and width of a field was measured in pixels using Openlab software (Cellular Imaging). A standard 0.01 mm micrometer (Applied Image Inc.) was used to convert pixels to μm. The number of hepatocytes per field was estimated, assuming that a mouse hepatocyte measures 20 μm in both height and width.
The number of hepatocytes present in the 2-dimensional section showing the widest diameter of a repopulating nodule was multiplied by a previously determined correction factor to estimate the total number of hepatocytes comprising the 3-dimensional nodule (35). Assuming that all hepatocytes in a nodule contribute equally to its expansion, the number of cell divisions required to reach a certain nodule size was calculated as log2 of the total number of hepatocytes per nodule.
All statistical tests were performed with the 2-tailed Student’s t test.
H. Willenbring received grant support from the California Institute for Regenerative Medicine, American Liver Foundation, and American Society of Transplantation; S. Espejel received support from the Cystinosis Research Foundation; G.R. Roll received support from a NRSA Hepatology Training Grant. The authors thank Alexis Bailey, Raymond Ng, Sonia Roll, and Montgomery Bissell for help with preparing the manuscript.
Conflict of interest: The authors have declared that no conflict of interest exists.
Citation for this article: J Clin Invest. 2010;120(9):3120–3126. doi:10.1172/JCI43267.
See the related Commentary beginning on page 3102.