Potential conflict of interest: Stephen C. Strom holds stock in Stemnion, LLC.
Supported in part by the National Institutes of Health (Grants N01-DK-7-0004/HHSN26700700004C and RC1DK086135 to S. C. S.) and the National PKU Alliance (R. G. and K. J. S. and Grant HD58553 to K. M. G.).
There is improved survival and partial metabolic correction of a mouse intermediate maple syrup urine disease (iMSUD) model after allogenic hepatocyte transplantation, confirming that a small number of enzyme-proficient liver-engrafted cells can improve phenotype. However, clinical shortages of suitable livers for hepatocyte isolation indicate a need for alternative cell sources. Human amnion epithelial cells (hAECs) share stem cell characteristics without the latter's safety and ethical concerns and differentiate to hepatocyte-like cells. Eight direct hepatic hAEC transplantations were performed in iMSUD mice over the first 35 days beginning at birth; animals were provided a normal protein diet and sacrificed at 35 and 100 days. Treatment at the neonatal stage is clinically relevant for MSUD and may offer a donor cell engraftment advantage. Survival was significantly extended and body weight was normalized in iMSUD mice receiving hAEC transplantations compared with untreated iMSUD mice, which were severely cachectic and died ≤28 days after birth. Branched chain α-keto acid dehydrogenase enzyme activity was significantly increased in transplanted livers. The branched chain amino acids leucine, isoleucine, valine, and alloisoleucine were significantly improved in serum and brain, as were other large neutral amino acids. Conclusion: Placental-derived stem cell transplantation lengthened survival and corrected many amino acid imbalances in a mouse model of iMSUD. This highlights the potential for their use as a viable alternative clinical therapy for MSUD and other liver-based metabolic diseases. (HEPATOLOGY 2013)
Maple syrup urine disease (MSUD; Online Mendelian Inheritance in Man #248600) is an inborn error of metabolism characterized by elevated branched chain keto acids and branched chain amino acids (BCAA; leucine, isoleucine, valine) resulting in severe brain injury and death unless treated. Mutation in any of the four branched chain α-keto acid dehydrogenase (BCKDH) subunits and varying residual enzymatic activity defines five MSUD variants.1 Higher enzyme activity translates to fewer crises and better long-term prognosis. Treatment consists of lifelong dietary BCAA restriction,2 though compliance in children and adolescents is problematic. Noncompliance is associated with catabolic crisis, frequent hospital stays, and irreversible neurological dysfunction.
Orthotopic liver transplantation (Tx) improves MSUD patient outcome.3 However, this procedure is associated with high costs, severe complications (including death), and a need for lifelong immunosuppression. Furthermore, critical donor shortages underscore a need for alternative therapies. Hepatocyte Tx has been applied as a cell therapy for metabolic liver disease. Preclinical studies have shown significant improvement in animal models, some despite low levels of cell engraftment.4-6 Clinical hepatocyte Tx has shown promise for a number of inherited metabolic diseases, such as Crigler-Najjar type 1, ornithine transcarbamylase deficiency, citrullinemia, glycogen storage disease, and others.7, 8 However, the availability of useful hepatocytes remains a limiting factor of clinical Tx.
Amniotic membrane and amnion epithelial cells are not immunogenic and have anti-inflammatory, antimicrobial, antiviral, and antifibrotic properties.9 Human amnion epithelial cells (hAECs) do not express telomerase, are not immortal, and are immunoprivileged.10, 11 They are readily obtained from placenta after live birth12 and express molecular and surface markers characteristic of stem cells.10 Amniotic membranes have been used clinically for more than a century without immunosuppression or adverse effects,13, 14 and hAECs have been transplanted, without immunosuppression, into normal volunteers or patients with lysosomal storage diseases without reported adverse effects or evidence of rejection or tumorigenicity.14-17
Methods to differentiate hAECs to hepatocyte-like cells have been described.18, 19 Importantly, undifferentiated hAECs engrafted in the livers of mice were found to display hepatic morphology and expressed mature liver genes at levels similar to adult liver. Due to their documented plasticity and special characteristics regarding safety, we examined the hypothesis that hAEC Tx could lengthen survival and correct amino acid imbalances in a mouse model of intermediate MSUD (iMSUD).
ANOVA, analysis of variance; BCAA, branched chain amino acid; BCKDH, branched chain α-keto acid dehydrogenase; hAEC, human amnion epithelial cells; iMSUD, intermediate maple syrup urine disease; LNAA, large neutral amino acid; MSUD, maple syrup urine disease; PCR, polymerase chain reaction; Tx, transplant; WT, wild-type.
Materials and Methods
All animal studies were reviewed and approved by the University of Pittsburgh Institutional Animal Care and Use Committee.
Line A iMSUD mice4, 5, 20 and the polymerase chain reaction (PCR) genotyping protocol used21 have been described. All animals were given standard (22% protein) mouse chow.
hAEC isolation has been described.12 Institutional Review Board approval for the use of human tissue was obtained from the University of Pittsburgh and Magee-Women's Hospital.
hAECs (10 × 106/mL in Hank's balanced salt solution [catalog #04-315Q; Lonza, Walkersville, MD]) were transplanted as described,22 with some modifications. Each “Tx event” consisted of two direct hepatic transdermal injections into two sites in the liver. Two Tx events (0.5 × 106 hAECs each site) during days 1-10 were administered. Twice-weekly Tx events (1 × 106 each site) were administered during days 21-35. Depending upon availability, freshly isolated or cryopreserved cells were used. Trypan blue exclusion determined hAEC viability (>90% in all cases) immediately prior to injection.
Mouse Sacrifice and Tissue Collection.
Mice were sacrificed via cervical dislocation when animals appeared moribund or at the end of the experiment (35 or 100 days). Whole blood collection (4 mm animal lancet; MEDIpoint, Mineola, NY) for serum isolation (Microtainer serum separator tubes, catalog #365959; BD, Franklin Lakes, NJ) was performed immediately before sacrifice. Liver and brain were immediately harvested after sacrifice and processed as described.5, 22 Liver was separated into portions for (1) BCKDH activity, (2) RNA expression, and (3) DNA analysis. Tissue and sera were stored at −80°C until analysis.
Survival and Body Weight.
iMSUD, iMSUD-hAEC Tx, and wild-type (WT) mouse date of birth, weekly body weight from weaning until death, and date of death were recorded. Data were analyzed using the Kaplan-Meier method and log-rank test (survival) or one-way analysis of variance (ANOVA) and post hoc Tukey test (body weight).
Human DNA Analysis.
Liver DNA from mice (iMSUD, iMSUD-hAEC Tx, WT), human liver, and hAECs were isolated using the DNeasy Blood and Tissue Kit (QIAGEN, Germantown, MD). Human DNA was quantified using the Taqman RNase P Detection Reagents (Applied Biosystems, Carlsbad, CA) following the manufacturer's instructions and compared with naïve hAECs. Data are presented as the mean ± SEM and were analyzed via one-way ANOVA and post hoc Tukey test.
Serum/Brain Amino Acids.
Amino acids from mouse serum and brain (iMSUD, iMSUD-hAE Tx, WT) were quantified as described.5 Data are presented as the mean ± SEM and were analyzed via one-way ANOVA and post hoc Tukey test. The criteria for achieving correction or partial correction have been defined.5
BCKDH Enzyme Activity.
BCKDH enzyme activity was assessed as described.20 Data were analyzed using a nonparametric Mann-Whitney test and compared with WT (100% activity).
Human RNA Expression.
Liver RNA from mouse (iMSUD, iMSUD-hAE Tx, WT), human liver, and hAECs was isolated and converted to complementary DNA as described.22 Human gene expression was assessed using Applied Biosystems TaqMan Assays-on-Demand gene expression kits (cyclophilin A [PPIA: Hs99999904_m1], BCKDHα [Hs00958109_m1], BCKDHβ [Hs00609053_m1], DBT [Hs01066445_m1], and DLD [Hs00164401_m1]) following the manufacturer's protocol. The expression of human BCKDH subunit genes from donor hAECs present in mouse liver was determined with human-specific probes and by normalization to human cyclophilin A. Results were compared with the expression in normal adult human liver (100%) and naïve hAECs. Data are presented as the mean ± SEM and were analyzed via one-way ANOVA and post hoc Tukey test.
hAECs expressing normal BCKDH were transplanted directly into the liver parenchyma of iMSUD mice as neonates and continuing until 35 days of age. Nontransplanted mice display adverse effects of MSUD prior to weaning,20 thus early treatment is clinically relevant to this disorder, and proliferating newborn livers may confer an engraftment/proliferative advantage to donor cells.
hAEC Tx Improved Survival and Normalized Body Weight.
All animals were given a normal (22% protein) diet. Untreated iMSUD mice were noticeably smaller than littermates (with normal BCKDH) at weaning (21 days), continued to lose weight, and developed muscle weakness with intermittent recumbence followed by seizures and death at or before 28 days of age (Fig. 1A,B).5, 22, 23 Animals receiving hAEC Tx displayed significantly improved survival at 35 days (100% survival; n = 11/11) and 100 days (81.2% survival; n = 9/11) compared with iMSUD controls (Fig. 1A; P < 0.001) and restored characteristics indicative of good health (i.e., smooth coat, active, bright eyes). Body weight was also normalized to WT levels (Fig. 1B). The two animals that died prior to day 100 were found dead in their cages without prodrome.
Donor hAECs Engrafted at a Low Level.
To verify hAEC engraftment, the presence of human DNA in recipient liver was determined via quantitative PCR of the human RNase P gene (Fig. 1C). Human DNA was undetectable in controls but increased significantly in transplanted mice at both time points (3.50% ± 0.26% human DNA at 35 days; 4.9% ± 0.37% human DNA at 100 days) and was significantly different at 100 days compared with 35 days (unpaired t test, P < 0.05).
Large Neutral Amino Acids in the Brain Were Significantly Improved at 100 Days with hAEC Tx.
Large neutral amino acids (LNAAs), which include the BCAAs (leucine, isoleucine, and valine), compete for transport into the brain. Elevated BCAAs disrupt other LNAA concentrations in the brain. hAEC Tx normalized leucine, isoleucine, and valine at 100 days (Fig. 2A) and, importantly, the brain showed similar corrections (Fig. 2D and Table 1), though the values still suggested residual disease. Alloisoleucine was improved post-Tx at 100 days in both serum (>65%, Fig. 2A) and brain (>80%, Fig. 2D and Table 1) to values not significantly different than WT. The ratio of total BCAA to alanine24 was decreased by >50%, indicating a significant partial correction in both serum (Fig. 2B) and brain (Fig. 2E and Table 1). Alanine was completely corrected by hAEC Tx at 35 days in both serum and brain (data not shown). Additional LNAA corrections were observed (Fig. 2C,F and Table 1). At 100 days, the dopamine precursors phenylalanine and tyrosine were normalized, and methionine was partially corrected post-Tx in the brain (Fig. 2F). Phenylalanine was also normalized in serum at 100 days (Fig. 2C), though tyrosine was not different between groups (data not shown). Metabolite changes were maintained across time points (two-tailed t test).
Table 1. Metabolic Changes in Brains of iMSUD Mice After hAEC Tx
iMSUD Versus WT
iMSUD-hAEC Versus WT
iMSUD-hAEC Versus iMSUD
An upward arrow (↑) indicates an increase; a downward arrow (↓) indicates a decrease. An equals sign (=) indicates either a correction or partial correction (see criteria in Materials and Methods).
BCKDH activity in transplanted animals increased significantly at 35 days (13.5% of WT activity; P < 0.01) and 100 days (12.5% of WT; P < 0.01) compared with iMSUD (∼5%-6% of WT; Fig. 3A). There was no significant difference between time points (two-tailed t test).
Human BCKDH Gene Expression Was Detectable in iMSUD-hAEC Tx Mice.
To verify hAEC engraftment, expression of three of the four human BCKDH subunits (BCKDHα, BCKDHβ, and DLD) in mouse liver was determined using real-time reverse-transcription PCR and normalized to human cyclophilin A (PIAA). Human DBT expression was not reported, as all animals were human DBT transgenic, and it was the gene used to rescue the classic MSUD knockout mouse and create the iMSUD model.20 Data from control animals (WT, iMSUD) and iMSUD-hAEC Tx mice were compared with both naïve hAECs and human liver (Fig. 3B-D). Using human-specific probes, it was determined that, post-Tx, human BCKDH genes were expressed at levels not significantly different from those of human liver. Expression of human BCKDHα after hAEC Tx was 80.2% ± 7.1% of human liver expression at 35 days (P < 0.01 versus hAEC) and 79.5% ± 6.8% at 100 days (P < 0.05; Fig. 3B). BCKDHβ expression was 83% ± 6.75% of human liver at 35 days and 81.1% ± 8.3% at 100 days (P < 0.001 versus hAEC; Fig. 3C). DLD was 96.7% ± 11.1% of human liver at 35 days (P < 0.001 versus hAEC) and 81% ± 17.4% at 100 days (P < 0.01; Fig. 3D). There was no difference in expression between Tx time points (two-tailed t test). As expected, human BCKDHα, BCKDHβ, and DLD expression was not detectable in WT or iMSUD control animals.
Orthotopic liver Tx has recently become a more common treatment for many inborn errors of metabolism when more conventional approaches (i.e., dietary therapies) prove insufficient. Cell Tx offers many advantages as a mode of therapy for these diseases.25 Of significant importance for patients with metabolic disease, the native liver remains intact, eliminating the need for transplanted cells to provide complete liver support. Poor cell function or rejection would only return the recipient to their pretransplanted state, and orthotopic liver Tx would remain a viable treatment option. Organ shortage is the primary barrier for clinical cell Tx, and there is a need for alternative cell sources. hAECs are easily accessible, are free of the ethical and safety concerns associated with stem cells, cryopreserve efficiently, and exhibit many beneficial and immunoprivileged characteristics.11, 26-28
Amnionic tissue was first used clinically more than a century ago; common applications include ocular surface reconstruction and skin injury repair.9, 13, 14 hAECs were transplanted for Niemann-Pick and other lysosomal storage diseases without adverse effects.16, 29 hAECs, like other placenta-derived cells, have been reported to reduce inflammation and fibrosis and can repair and preserve function in lung injury30-32 and liver cirrhosis.33-35 Differentiated dopamine-expressing hAECs were able to survive and function in the brain of a rat model of Parkinson disease, which resulted in prevention of neuron degradation.36, 37 Finally, hAECs were used to target liver and deliver a transgene to treat the metabolic disease familial hypercholesterolemia in a rabbit model.38
hAEC Tx converted iMSUD from a lethal disease to one in which mice survived long-term, recovered a normal growth rate, and showed all the characteristics of good health (smooth coat, bright eyes, normal activity). These preclinical studies employed six intrahepatic Tx during the first few weeks of life, a protocol that has been achieved clinically.39 It is preferable to perform Tx as early as possible because (1) recipients require fewer cells, and the onset of inherited metabolic disease most often occurs at birth, and (2) delays in initial treatment or poor dietary control over time have been associated with serious and permanent side effects in MSUD and other metabolic liver diseases. In our experience, and in other reports wherein hAECs were transplanted or used on patients, there are no findings of tumorigenicity, immunorejection, or adverse effects.10, 14-17, 26 Indeed, hAECs have been reported to escape immunorecognition post-Tx and have been shown to have immunomodulatory properties similar to those described in mesenchymal stem cells.9, 11, 27 Xenotransplanted human mesenchymal stem cells were able to engraft long term and undergo site-specific differentiation in sheep,40 and were further shown to improve a variety of diseases in a number of species,41 all without immunosuppression or the development of graft-versus-host disease. Accordingly, no immunosuppression was administered to test subjects despite the Tx of human cells, and there was no evidence of cell rejection. However, in the current study, the initial hAEC Tx were conducted in the early neonatal period, and this experimental design does not exclude the possibility that the iMSUD mice were induced to become tolerant to the human cells by such early exposure.42, 43
In MSUD, leucine toxicity is the primary cause of brain injury,1 and when competing with other LNAAs for brain access, leucine is able to cross the blood-brain barrier more efficiently than other amino acids.44 Pathology from MSUD occurs in multiple compartments (liver, kidney, muscle, and brain),45, 46 and in the current protocol, cell Tx was liver directed. Nonetheless, transplanted animals survived and actually thrived, allowing temporal comparison of metabolites at later time points. The improved brain LNAA levels observed in this study suggest that hepatic correction provided metabolic correction to distal organs. Liver-directed hAEC Tx resulted in a >60% correction in BCAAs (leucine, isoleucine, valine) and alloisoleucine, as well as additional LNAA normalizations (Fig. 2). Brain isoleucine was not completely corrected at 100 days despite significant improvement. Alloisoleucine, an isoleucine analogue and pathognomonic marker for MSUD,47 was improved after hAEC Tx in both serum and brain (Fig. 2A,D), whereas isoleucine was only significantly improved by 100 days. Therefore, improvements in isoleucine were not due to an enhanced conversion to alloisoleucine. Metabolic improvements were maintained across time points, suggesting stable engraftment.
Perinatal hAEC Tx was equivalent in cell number to ∼8%-12% of the total number of hepatocytes in mouse liver and resulted in a chimeric organ with 3.5%-5% human DNA (Fig. 1C); however, we cannot exclude the possibility that cell fusion between hAECs and mouse hepatocytes had occurred. This level of engraftment more than doubled liver hepatic BCKDH activity from 5%-6% to 12.5%-13.5% (Fig. 3A). This large increase in BCKDH activity relative to cell engraftment was similar to results obtained after hepatocyte Tx5, 22 and corresponds with clinical observations wherein incremental differences in enzyme activity significantly impact disease phenotype and patient outcome.1 Although human DNA was significantly different between 35 days and 100 days, the modest increase suggests that hAECs are at least maintained throughout the study or that some level of proliferation occurred between 35 days and 100 days. Human BCKDH subunit gene expression of the engrafted donor cells was not significantly different from human liver (Fig. 3B-D), suggesting that hAECs mature post-Tx.18 It is also important to note that the human BCKDH subunit expression reported here does not represent total BCKDH expression in mouse liver, but only that the human component of these chimeric livers expressed BCKDH subunits at levels similar to hepatocytes in adult human liver.
In conclusion, transplanted hAECs significantly increased BCKDH enzyme activity in a mouse model of iMSUD, resulting in longer survival, normalized body weight, and improved circulating and brain amino acids. Our study supports the concept that placental-derived cells such as hAECs may provide a safe and abundant cell source for the treatment of MSUD and other liver-based metabolic diseases.