Journal Logo

Basic Science

Improvement of Hepatocyte Transplantation Efficiency in the mdr2–/– Mouse Model by Glyceryl Trinitrate

Boudechiche, Lyes1; Tranchart, Hadrien1,2,3; Branchereau, Sophie2,4; Davit-Spraul, Anne5; Laïnas, Panagiotis1,2,3; Groyer-Picard, Marie-Thérèse1; Weber, Anne1,2; Hadchouel, Michelle1,2; Dagher, Ibrahim1,2,3

Author Information
doi: 10.1097/TP.0000000000000463
  • Free

Liver transplantation is considered to be the ultimate treatment for various liver diseases. Owing to the limited availability of donor livers, surgical morbidity and problems related to immunosuppression, there is growing interest for the transplantation of isolated hepatocytes. More than 70 patients have already received hepatocyte transplantation1-8; however, the extent of liver repopulation observed in humans does not reach that obtained in studies in rodent models, where transplanted hepatocytes display a proliferative advantage.9 In most cases, correction of human metabolic disorders has only been partial and transient.2,7,8 This is partly because of the poor efficacy of hepatocyte engraftment in noninjured livers.

In a study by Stephenne et al., hepatocyte transplantation resulted in metabolic control with psychomotor catch-up in a 3.5-year-old girl with argininosuccinate lyase deficiency.10 However, this treatment required three series of hepatocyte infusions over 5 months to obtain, 8 months after transplantation, argininosuccinate lyase activity reaching 3% of control levels. These results suggested that more than 50% of infused hepatocytes were lost during the whole procedure, as in rat models.11 Previous studies have shown that entry of transplanted cells into hepatic sinusoids is driven by mechanical factors, including portal-sinusoidal blood flow–related cell distributions and by the relationship between the sizes of cells and vascular structures.12 Further studies demonstrated that pretreatment of rats with adrenergic receptor blockers, such as phentolamine and splanchnic vasodilators (e.g., nitroglycerine), promoted the entry of transplanted syngenic hepatocytes into liver sinusoids and improved cell engraftment at short-term in dipeptidyl peptidase intravenous rats.13 As nitroglycerine, an NO donating drug, with an important local vasodilatory activity, is widely used in clinical protocols, we evaluated nitroglycerine efficacy in a mouse model (mdr2 knockout) of progressive familial intrahepatic cholestasis type 3,14 a condition caused by mutations in the MDR3 (ABCB4) gene encoding the phospholipids (PLs) export pump. The absence of PLs from bile causes chronic bile salt (BS)–induced damage to hepatocytes in this mouse model, resulting in a proliferative advantage for the transplanted normal syngeneic hepatocytes.15 We used this model to determine the effect of nitroglycerine on hepatocyte engraftment at long term and on the metabolic improvement in mdr2 (−/−) transplanted mice, which secrete virtually no PL in their nontransplanted state.14,15


Hepatocyte Engraftment

To determine whether infusion of glyceryl trinitrate (GTN) improved hepatocyte engraftment, we isolated hepatocytes from FVB donor mice. The cells were immediately transplanted into control (n=8) and GTN-treated mdr2 knockout animals (n=8). Hepatocyte engraftment in recipient livers was then evaluated 3 months after transplantation.

The engrafted cells expressed the mdr2 protein, which was localized at the membrane with a polarized distribution clearly visible (Fig. 1C). Because this protein is naturally expressed in the bile canalicular domain of all hepatocytes, changes in transplanted cell numbers in the recipient livers were established by morphologic methods. Transplanted hepatocytes were organized in clusters ranging from 30 to 250 hepatocytes and were equally distributed.

Hepatocyte engraftment in mdr2−/− mice. Liver sections were analyzed by immunohistochemistry 3 months after transplantation. Mdr2 protein was expressed at the bile canalicular pole of engrafted hepatocytes. Representative liver section are shown in GTN-treated (magnification 10× [A] and magnification 20× [C]) and control (magnification 10× [B]) mice. (D) Quantification of hepatocyte engraftment in GTN-treated and control mice. Data represent mean±standard error of the mean of 16 experiments. The number of mdr2 expressing cells per field was significantly increased in GTN-treated mice. GTN, glyceryl trinitrate.

The number of transplanted cells was significantly higher in GTN-treated mice than in controls: 99±20 cells per field (22% engraftment) in GTN-treated mice; 58%±12 (13% engraftment) in control mice (P=0.02) (Fig. 1D). We thus were able to confirm, in a preclinical rodent model, that greater entry of transplanted hepatocytes into hepatic sinusoids elicited by a vasodilatator results led in a statistically significant increase in hepatocyte engraftment on the long term.

Bile Analysis and Liver Fibrosis

We further investigated whether nitroglycerine improved metabolic correction in transplanted mdr2 (−/−) mice. Given that disruption of the mdr2 gene in mice leads to a complete absence of PLs from bile, we analyzed PLs concentrations in bile 3 months after transplantation to evaluate the function of the engrafted hepatocytes. PL secretion was significantly higher in GTN-treated animals than that in the control group: 18.34±2.28 nmol per min per 100 g versus 5.18±3.87 nmol per min 100 g (P<0.0001) (Fig. 2A). The rate of PL secretion directly depends on the rate of BS secretion. Therefore, the capacity to secrete PLs can be expressed as the ratio of PLs to BSs. The ratio of PLs to BSs (PL/BS×100) was also higher in GTN-treated mice: 6.76±1.31, versus 3.20±1.65 in controls (P=0.03) (Fig. 2B).

Phospholipid secretion in bile 3 months after hepatocyte transplantation in mdr2−/− mice. A, quantification of phospholipid secretion in GTN-treated and control mice after hepatocyte transplantation. B, the ratio of PL/BS was calculated. Data represent mean±standard error of the mean of 16 experiments. PL, phospholipids; BS, bile salts; GTN, glyceryl trinitrate.

We also compared the extent of liver fibrosis between these two groups of mice. It is known that in mdr2 (−/−) mice, the secretion of hydrophobic BSs in the absence of PLs results in damage of membrane and in liver pathology, which shows progression to the development of fibrosis. In our study, as shown in Figure 3, the percentage of liver fibrosis was significantly lower in GTN-treated mice than that in controls: 5.7%±2.3% vs. 12.4%±2.9% in control animals (P=0.016).

Evaluation of liver fibrosis in mdr2−/− mice. Liver sections were stained with hematoxylin-eosin for histology and fibrosis quantification in control mice (magnification, 10× [A] and magnification, 5× [C]) and in GTN-treated mice (magnification, 10× [B] and magnification, 5× [D]). E, the extent of fibrosis was quantified with Perfect Image Software. The relative amount of fibrosis was expressed as the percentage of the total section area. GTN, glyceryl trinitrate.


In this study, we show that infusion of GTN into the mouse portal vein during cell transplantation increased the number of engrafted hepatocytes, substantiating the long-term effect of sinusoidal vasodilatation on cell engraftment efficacy. As a consequence, we also observed improved metabolic correction in GTN-treated animals.

Indeed, as well as problems related to immunosuppression, poor hepatocyte engraftment efficacy represents a major limitation to allogeneic hepatocyte transplantation. It is well established that more than 50% of transplanted hepatocytes remain trapped in portal spaces and are cleared by macrophages not only in rodent models but also in nonhuman primates.11,16 It was more recently suggested that sinusoidal vasodilatation improved hepatocyte engraftment in a syngenic rat model after transplantation into the spleen. However, the effect of GTN on hepatocyte engraftment was assessed in the short term by determining the number of engrafted hepatocytes 2 weeks after transplantation.13 Given that long-term functional studies of engrafted hepatocytes are a prerequisite of clinical application, we tested the effect of GTN in a mouse model of metabolic disease. In contrast to a previous study, in which normal hepatocytes or transgenic MDR3-expressing hepatocytes were injected through the spleen into the mdr2 (−/−) mouse model,15 we chose to infuse mdr2 (+/+) hepatocytes and GTN directly into the portal vein. This approach has the advantages that it more closely resembles clinical protocols used in humans and that it avoids a local effect of GTN in the spleen.

In a preliminary study, we assessed the effect of portal GTN infusion on the distribution of Hoechst-labeled transplanted hepatocytes within normal liver parenchyma. The transplanted syngenic hepatocytes were located more distally in GTN-treated animals, whereas they remained in the periportal area in control animals (data not shown). In mdr2 (−/−) mice, however, in contrast to normal animals, chronic damage to parenchymal cells led to a continuous state of regenerative proliferation, allowing transplanted hepatocytes to proliferate within the lobule. Furthermore, in the mdr2 (−/−) mice, the hepatocytes are exposed to injury from bile acids present in the bile canaliculi, without the counterbalancing effect of PLs. Therefore, the engrafted cells can have a proliferative advantage, leading to their preferential proliferation. The difference in the initial engraftment (the step at which GTN presumably works) is enhanced by cell proliferation.

We injected four times fewer hepatocytes into the portal vein than those previously injected into the spleen in the same model by Slehria et al.13 (5×105 and 2×106 cells, respectively). Despite this, PLs secretion was increased in GTN-treated mice and reached 38% of the levels observed in mdr2 (+/+) mice.17 The PL-to- BS ratio was also increased, confirming an improvement in metabolic function.

Lack of biliary PL secretion leads to severe liver disease, characterized by inflammation of the portal tract, proliferation of the bile ducts, fibrosis, and then cirrhosis in mice and humans.18 As a consequence of metabolic improvement, we also observed that fibrosis was significantly reduced in GTN-treated mice. The mdr2 (−/−) mouse is a model of nonautonomous cell damage, in which transplanted repopulating cells are able to protect neighboring defective cells. Thus, by contrast to the fah−/− mouse, a model for tyrosinemia in humans, complete repopulation of the liver is not required to prevent the progression of cell damage. Our results support this idea and show that a simple infusion of a vasodilatator is sufficient to modify the progression of the disease.

Our results have potential clinical implications. Indeed, each fold-increase in transplanted cell engraftment would curtail the number of transplanted cells required. Glyceryl trinitrate treatment could be used for the transplantation of genetically modified autologous cells or allogeneic hepatocytes. However, the need for immunosuppressive therapy might have an impact on the role of GTN on allogenic hepatocytes engraftment. This reason associated with the increasing organ shortage led our team to specifically focus during the last decade on ex vivo genetically modified autologous hepatocytes transplantation, which can be considered as an alternative strategy. Congenic hepatocytes were used in this model to facilitate the procedure. We believe that our approach could be applied to progressive familial intrahepatic cholestasis type 3 patients, provided that hepatocyte transplantation could be carried out at an early stage, when liver pathology has not yet developed, providing sufficient time for transplanted cells to proliferate and confer protection to the remaining resident deficient hepatocytes.

Recent studies have described the isolation of adult stem-cell populations in rodents and humans.19-21 Fetal liver progenitors have been purified from murine liver22 and have been isolated from human livers.9,23 These cells are significantly smaller than adult hepatocytes (10–15 microm) and display proliferative ability in vivo.9,20 Combining the use of stem cells with portal vasodilatation should generate useful models for research on cell transplantation for metabolic diseases.

In conclusion, GTN increases hepatocyte engraftment, leading to improved metabolic function in a mouse model of liver disease. Further investigations are needed to evaluate the consequences of GTN administration on hepatocyte engraftment in large animal models, such as nonhuman primates.



Male mdr2 (−/−) mice (generous gift of RPJ Oude Elferink) were transplanted with hepatocytes from donor mice with an identical FVB background. All animals were maintained on alternating 12-hr light-dark cycles; food and water was available ad libitum. Experiments were carried out in accordance with the European Legislation on Animal Care and Experimentation.

Experimental Protocol

Hepatocytes were isolated using a standard two-step collagenase perfusion.24 The isolated cells were suspended in 199 medium (EUROBIO, Les Ulis, France), containing 10% fetal bovine serum (Gibco, Saint Aubrin, France), glutamine 2 mM and antibiotics. The isolated hepatocytes were purified by centrifugation on Percoll density gradient. Cells were then washed three times in phosphate-buffered saline (Gibco) and resuspended in saline for transplantation. Cell viability after isolation and purification was determined by Trypan blue dye exclusion and attachment to tissue culture plastic. Donor hepatocytes for transplantation were used when viability exceeded 95% and cells attached to culture dishes within 30 min.

We used 8-week-old to 10-week-old male mdr2 (−/−) mice (weight, 25–30 g). They were divided into two groups of eight animals: the first group received GTN (Nitronal, France) (GTN) diluted in normal saline, infused into the portal vein (0.5 μg/kg/min) at 0.3 mL per hr with a Harvard pump 5 min before cell injection and ending 10 min later; the second group received only normal saline without GTN. Each recipient was anesthetized, the abdomen was incised, and the left lateral lobe was immobilized gently. Saline solution at 37°C was suffused to prevent drying. Hepatocytes (5×105 cells in 200 μL of saline) were injected directly into the portal vein at 0.3 mL per hr using a 29-gauge needle (VYGON, Ecouen, France). One week after the operation, all transplanted mdr2 (−/−) mice were fed ad libitum with a diet supplemented with 0.03% cholic acid (U.A.R., VilleMoisson-Sur-Orge, France). Three months after transplantation, the animals were anesthetized, and the abdomen was incised. Bile was collected by gallbladder cannulation after distal ligation of the common bile duct for PL and BS determination. Snap-frozen liver specimens were prepared for mdr2 P-gp immunohistochemistry, and the mice were killed.

Bile Analysis

Phosphatidylcholine levels in bile were determined enzymatically with phospholipase D and choline oxidase. Total BS concentration was measured enzymatically with three α-hydroxysteroid dehydrogenase.25

Histology and Immunohistochemistry

Tissue samples were immediately frozen in liquid nitrogen and stored at −80°C or fixed in formalin and embedded in paraffin. Sections (5 μm) of paraffin-embedded tissue were stained with hematoxylin-eosin and Trichrome for histology and fibrosis quantification. For detection of mdr2 P-gp, 5-μm cryostat sections were fixed in acetone for 10 min. Endogenous peroxidase activity was blocked with a peroxidase block (Dako, Glostrup, Denmark) for 5 min. Sections were then treated with Dako biotin blocking system for 10 min and then with 10% fetal bovine serum (Invitrogen, Cergy Pontoise, France) for 1 hr. Cryostat sections were subsequently incubated with rabbit antibody (Peptide EEFEVELSDEKA; Agro-Bio, France),26 diluted to 1 μg/mL, for 4 hr at room temperature. Binding was revealed with a secondary polyclonal anti rabbit IgG HRP antibody (GE Healthcare, Vélizy-Villacoublay, France) diluted 1/100. Tissue sections were counterstained with hematoxylin. To control for background staining, sections were incubated without primary antibody. Twenty sections of each liver lobe were analyzed at random levels (magnification, 20×), corresponding at 9,000 hepatocytes counted in every lobe.

Fibrosis Quantification

For each mouse, the extent of fibrosis was quantified in 10 different fields. Each field was acquired at 10× magnification, and the fibrosis area was quantified with Perfect Image Software (Claravision Orsay, France). The relative amount of fibrosis was expressed as the percentage of the total section area. Transplanted mdr2-positive hepatocytes were counted and expressed as the number of positive cells for each field.

Statistical Analysis

Values are presented as means±SEM. Statistical significance between groups was analyzed using the Mann-Whitney U test. P values less than 0.05 were considered significant.


The authors thank Dr. Oude Elferink for the gift of mdr2 knockout mice.


1. Fisher RA, Strom SC. Human hepatocyte transplantation: worldwide results. Transplantation 2006; 82: 441.
2. Gras JM, Gerkens S, Beguin C, et al. Steroid-free, tacrolimus-basiliximab immunosuppression in pediatric liver transplantation: clinical and pharmacoeconomic study in 50 children. Liver Transpl 2008; 14: 469.
3. Fili D, Vizzini G, Biondo D, et al. Clinical burden of screening asymptomatic patients for coronary artery disease prior to liver transplantation. Am J Transplant 2009; 9: 1151.
4. Al-Hussaini A, Taylor RM, Samyn M, et al. Long-term outcome and management of hepatopulmonary syndrome in children. Pediatr Transplant 14: 276.
5. Asif S, Sedigh A, Nordstrom J, et al. Oxygen-charged HTK-F6H8 emulsion reduces ischemia-reperfusion injury in kidneys from brain-dead pigs. J Surg Res 178: 959.
6. Aghajani K, Keerthivasan S, Yu Y, et al. Generation of CD4CreER(T(2)) transgenic mice to study development of peripheral CD4-T-cells. Genesis 50: 908.
7. Gilroy RK, Coccia PF, Talmadge JE, et al. Donor immune reconstitution after liver-small bowel transplantation for multiple intestinal atresia with immunodeficiency. Blood 2004; 103: 1171.
8. Hughes RD, Mitry RR, Dhawan A. Hepatocyte transplantation in the treatment of liver diseases—future seems bright after all. Pediatr Transplant 2008; 12: 4.
9. Bailey AS, Willenbring H, Jiang S, et al. Myeloid lineage progenitors give rise to vascular endothelium. Proc Natl Acad Sci U S A 2006; 103: 13156.
10. Stephenne X, Najimi M, Sibille C, et al. Sustained engraftment and tissue enzyme activity after liver cell transplantation for argininosuccinate lyase deficiency. Gastroenterology 2006; 130: 1317.
11. Gupta S, Rajvanshi P, Bhargava KK, et al. Hepatocyte transplantation: progress toward liver repopulation. Prog Liver Dis 1996; 14: 199.
12. Cho JJ, Malhi H, Wang R, et al. Enzymatically labeled chromosomal probes for in situ identification of human cells in xenogeneic transplant models. Nat Med 2002; 8: 1033.
13. Malhi H, Annamaneni P, Slehria S, et al. Cyclophosphamide disrupts hepatic sinusoidal endothelium and improves transplanted cell engraftment in rat liver. Hepatology 2002; 36: 112.
14. Borst P, Schinkel AH, Smit JJ, et al. Classical and novel forms of multidrug resistance and the physiological functions of P-glycoproteins in mammals. Pharmacol Ther 1993; 60: 289.
15. De Vree JM, Ottenhoff R, Bosma PJ, et al. Correction of liver disease by hepatocyte transplantation in a mouse model of progressive familial intrahepatic cholestasis. Gastroenterology 2000; 119: 1720.
16. Allain JE, Mahieu-Caputo D, Loux N, et al. Allotransplantation in utero and immortalization of primate fetal hepatocytes. J Soc Biol 2001; 195: 57.
17. Van Nieuwkerk CM, Elferink RP, Groen AK, et al. Effects of ursodeoxycholate and cholate feeding on liver disease in FVB mice with a disrupted mdr2 P-glycoprotein gene. Gastroenterology 1996; 111: 165.
18. de Vree JM, Jacquemin E, Sturm E, et al. Mutations in the MDR3 gene cause progressive familial intrahepatic cholestasis. Proc Natl Acad Sci U S A 1998; 95: 282.
19. Sahin MB, Schwartz RE, Buckley SM, et al. Isolation and characterization of a novel population of progenitor cells from unmanipulated rat liver. Liver Transpl 2008; 14: 333.
20. Lippok B, Birkenbihl RP, Rivory G, et al. Expression of AtWRKY33 encoding a pathogen- or PAMP-responsive WRKY transcription factor is regulated by a composite DNA motif containing W box elements. Mol Plant Microbe Interact 2007; 20: 420.
21. Yovchev MI, Grozdanov PN, Zhou H, et al. Identification of adult hepatic progenitor cells capable of repopulating injured rat liver. Hepatology 2008; 47: 636.
22. Suzuki A, Oyama K, Fukao K, et al. Establishment of clonal colony-forming assay system for pancreatic stem/progenitor cells. Cell Transplant 2002; 11: 451.
23. Deshayes F, Lapree G, Portier A, et al. Abnormal production of the TNF-homologue APRIL increases the proliferation of human malignant glioblastoma cell lines via a specific receptor. Oncogene 2004; 23: 3005.
24. Bennoun M, Rissel M, Engelhardt N, et al. Oval cell proliferation in early stages of hepatocarcinogenesis in simian virus 40 large T transgenic mice. Am J Pathol 1993; 143: 1326.
25. Turley SD, Dietschy JM. Re-evaluation of the 3 alpha-hydroxysteroid dehydrogenase assay for total bile acids in bile. J Lipid Res 1978; 19: 924.
26. Buschman E, Arceci RJ, Croop JM, et al. mdr2 encodes P-glycoprotein expressed in the bile canalicular membrane as determined by isoform-specific antibodies. J Biol Chem 1992; 267: 18093.
Copyright © 2015 Wolters Kluwer Health, Inc. All rights reserved.