ARTICLE IN PRESS

R.R. and

SupportLee Trasalary su

Conflictconflict

PresenteLas Veg

Accepte

Reprinttute, Be7th Flooedu.

0039-60

� 2015

http://d

Transplantation of human stemcell-derived hepatocytes in an animalmodel of acute liver failureRajesh Ramanathan, MD,a Giuseppe Pettinato, PhD,b John T. Beeston, RN,a David D. Lee, MD,c

Xuejun Wen, MD, PhD,b Martin J. Mangino, PhD,a and Robert A. Fisher, MD,d Richmond, VA,Jacksonville, FL, and Boston, MA

Introduction. Hepatocyte cell transplantation can be life-saving in patients with acute liver failure(ALF); however, primary human hepatocyte transplantation is limited by the scarcity of donorhepatocytes. We investigated the effect of stem cell-derived, hepatocyte-like cells in an animalxenotransplant model of ALF.Methods. Intraperitoneal D-galactosamine was used to develop a lethal model of ALF in the rat. Humaninduced pluripotent stem cells (iPSC), human mesenchymal stem cells, and human iPSC combined withhuman endothelial cells (iPSC + EC) were differentiated into hepatocyte-like cells and transplanted intothe spleens of athymic nude rats with ALF.Results. A reproducible lethal model of ALF was achieved with nearly 90% death within 3 days.Compared with negative controls, rats transplanted with stem cell-derived, hepatocyte-like cells wereassociated with increased survival. Human albumin was detected in the rat serum 3 days aftertransplantation in more than one-half the animals transplanted with hepatocyte-like cells. Only animalstransplanted with iPSC + EC-derived hepatocytes had serum human albumin at 14 days posttransplant.Transplanted hepatocyte-like cells homed to the injured rat liver, whereas the ECs were only detected in thespleen.Conclusion. Transplantation of stem cell-derived, hepatocyte-like cells improved survival with evidence ofin vivo human albumin production. Combining ECs may prolong cell function after transplantation.(Surgery 2015;j:j-j.)

From the Departments of Surgerya and Chemical and Life Science Engineering,b Virginia CommonwealthUniversity, Richmond, VA; the Department of Surgery,c Mayo Clinic Jacksonville, Jacksonville, FL; and theTransplant Institute,d Beth Israel Deaconess Medical Center, Boston, MA

CELL TRANSPLANTATION holds great promise as analternative to whole-organ liver transplantationfor patients suffering from some forms of acute,

G.P. contributed equally to this report.

ed by the Virginia Commonwealth University Hume-nsplant Center’s Irvin Grant for supplies and residentpport. There was no industry funding for this study.

s of interest: None of the authors have any personalof interests to report.

d at the 10th Annual Academic Surgical Congress inas, Nevada, February 3–5, 2015.

d for publication April 22, 2015.

requests: Robert A. Fisher, MD, FACS, Transplant Insti-th Israel Deaconess Medical Center, 110 Francis Street,r, Boston, MA 02215. E-mail: rafisher@bidmc.harvard.

60/$ - see front matter

Elsevier Inc. All rights reserved.

x.doi.org/10.1016/j.surg.2015.04.014

life-threatening liver disease. Despite its promiseand preliminary success in human trials, wide-spread application of cell transplantation for liverfailure has been hindered by the scarcity of humandonor livers for primary cell isolation.1,2

Stem cell-derived hepatocytes are an alternatesource for cell transplantation. Induced pluripo-tent stem cells (iPSC) and mesenchymal stem cells(MSC) are adult-derived cell sources that can beexpanded to generate large numbers of cells.These cells display the potential for multilineagedifferentiation, are devoid of ethical controversy,may be less immunogenic, and are a potentialautologous cell source enabling immunosuppres-sion minimization or cessation.3-6 The iPSC andMSC have been differentiated successfully intohepatocyte-like cells. We recently developed anovel, suspension-based, embryoid body approachfor differentiating multipotent stem cells fromdifferent sources into hepatocyte-like cells.7

SURGERY 1

ARTICLE IN PRESSSurgeryj 2015

2 Ramanathan et al

Because the transplantation of stem cell-derivedhepatocytes results in transiency in function, we hy-pothesized that incorporating endothelial cells(ECs) into clusters of stem cell-derived hepatocytesmay improve engraftment and prolong function.8

In this study, we utilized an animal xenotrans-plant model of lethal acute liver failure (ALF) toinvestigate the efficacy of human stem cell-derivedhepatocyte-like cells in reversing ALF. Using anidentical differentiation protocol, we evaluated thein vivo functionof hepatocyte-like cells derived fromhuman iPSC, human MSC, and human iPSC + EC.

METHODS

Rat model of ALF. The Institutional AnimalCare and Use Committee approved the use of ratsfor experimentation in this study. The goal for themodel of ALF was reproducible mortality within3 days of onset of liver failure. Sterile D-galactos-amine (Sigma-Aldrich, St. Louis, MO) dissolvedin Hank’s balanced salt solution (HBSS) was usedto induce ALF. Anesthesia was induced by place-ment of the rats in an isofluorane induction cham-ber. Intraperitoneal injection of 800–1,000 mg/kgof D-galactosamine was evaluated during the devel-opmental stage of this model of ALF to determinethe dose responsiveness. The model was initiallydeveloped in male Lewis rats (LEW/SsNHsd; Har-lan Laboratories, Indianapolis, IN) and repeatedand translated in male nude athymic rats(Crl:NIH-Foxn1rnu [Charles River Laboratories,Wilmington, MA] and Hsd:RH-Foxn1rnu [HarlanLaboratories]).

The Lewis rats were housed in our institutionalvivarium with 12-hour light–dark cycles andreceived standard chow and water ad libitum; theathymic nude rats were housed in a specializedbarrier facility and received irradiated chow andwater. All rats were weighed daily and signs ofhepatic encephalopathy scored using publishedneurologic criteria.9 Liver injury was quantified bymeasuring rat serum alanine transferase (ALT) bytail venipuncture (VetScan 2.0, Abaxis, Union City,CA). Animals were monitored for 14 days aftertransplantation, and when humanely killed, liverand spleen samples were retrieved. Animals werekilled at earlier time points if they met thefollowing humane criteria: moribund state, weightloss of $30% from before induction of ALF, or in-crease in ALT to >15 times baseline.

The data were analyzed to determine theoptimal dosage required to achieve a reproducibleincrease in ALT and mortality within 72–96 hoursafter induction of ALF.

Xenotransplant model. To develop the model,primary hepatocytes isolated from mouse liverswere transplanted into the spleens of athymicnude rats 16–18 hours after induction of ALF.ALF was induced in 270–350 g male, athymic nuderats with 950–975 mg/kg of sterile D-galactosaminedissolved in HBSS at a concentration of 97.5 mg/mL.

Isolation of hepatocytes. Primary mouse hepato-cytes were isolated from 6- to 8-week-old male mice(C57BL/6NHsd, Harlan Laboratories) weighing25–30 g. We used 22 mice to establish and optimizethe technique. An additional 16 mice were used toobtain donor hepatocytes for culture and trans-plantation. Under ketamine anesthesia, the infe-rior vena cava was cannulated with a 24-gaugeangiocath (Becton-Dickson, Franklin Lakes, NJ)and the liver flushed and digested retrograde. Theflush and digestion solutions were oxygenated for10 minutes before perfusion and warmed to 378C.The flush solution consisted of 6.67 g/L NaCl,0.16 g/L KCl, 2.1 g/L NaHCO3, 0.34 g/L KH2PO4,0.11 g/L MgCl2, and 0.14 g/L MgSO4. The diges-tion solution consisted of the flush solution with1.2 mmol/L calcium chloride and 1 mg/mL Type2 Collagenase (Worthington Biochemical Corp,Lakewood, ND). Both solutions were administeredfor 7 minutes each at a flow rate of 6 mL/min. Af-ter digestion, the peritoneal and diaphragmatic at-tachments of the liver were transected and the liverwas removed in a 50-mL, conical tube with sterileKrebs-Henseleit buffer (25 mmol/L NaHCO3,118 mmol/L NaCl, 4.7 mmol/L KCl, 1.2 mmol/LMgSO4, 1.2 mmol/L NaH2PO4, 10 mmol/Lglucose and 1.3 mmol/L CaCl2) at 48C. Using asterile, cotton-tipped applicator, the liver capsulewas ruptured and cells released. The solution wasstrained through a sterile nylon mesh, and theresultant solution was washed twice at 50g for 2 mi-nutes with cold Krebs-Henseleit buffer. The finalpellet was re-suspended in 5 mL of cold, hepato-cyte maintenance medium (HMM) consisting ofenriched Iscove’s Modified Dulbecco’s Medium(Invitrogen, Grand Island, NY) with 10% fetalbovine serum (Hyclone Laboratories, Logan,UT), 1 3 penicillin/streptomycin (100 U/mLpenicillin + 0.1 mg/mL streptomycin), 2 g/Lglucose, 1 g/L galactose, 0.5 g/L nicotinamide,0.110 g/L sodium pyruvate, 10 mg linoleic acid,0.5 mL of insulin–transferrin–selenium, 50 mg/Ldexamethasone, and 10 mg/L epidermal growthfactor (all from Sigma-Aldrich). The resulting sam-ple was analyzed for yield and viability using Try-pan blue exclusion. Using our isolation protocol,the mean yields were 1–5 3 107 hepatocytes per

Fig 1. Isolated mouse hepatocytes displaying strong green autofluorescence, Magnification = 20x (A) and close to 90%viability by Trypan blue staining (B).

Table I. Posttransplant survival and human albumin detection in the rat serum at 3 and 14 days after celltransplant

Variable n Survivors to 14 d Mean survival (d) Human albumin at 3 d Human albumin at 14 d

iPSC 14 10 (71%)* 11.2 ± 4.7* 58% (7/12)* 0% (0/10)iPSC + EC 8 6 (75%)* 11.2 ± 4.5* 63% (5/8)* 50% (3/6)*MSC 11 8 (73%)* 11.6 ± 4.2* 22% (2/9) 0% (0/8)Mouse hepatocytes 9 4 (44%) 8.1 ± 5.6 — —Control 6 1 (14%) 5.4 ± 4.5 (0/2) (0/1)

*P < .05 versus control.EC, Endothelial cells; iPSC, induced pluripotent stem cells; MSC, mesenchymal stem cell.

ARTICLE IN PRESSSurgeryVolume j, Number j

Ramanathan et al 3

mouse. The isolated hepatocytes showed strong au-tofluorescence and viability between 75 and 90%(Fig 1). The sample was diluted to approximately1.2 3 106 cells/mL using warm HMM. For immedi-ate transplantation, 2 3 106 cells in 500 mL ofHMM were used per rat. Nine athymic nude ratswere transplanted with 2 3 106 mouse hepatocytesper rat. Rats that were transplanted with mousehepatocytes demonstrated survival to 14 days in 4of 9 rats compared with 1 of 6 rats among the con-trols (Table I).

Cell transplantation. Under isofluorane anes-thesia, the caudal pole of the spleen in the recipientrat was eviscerated through a left subcostal incision.After ligating the caudal spleen hilum, cells andmedia were injected via the caudal spleen into thebody of the spleen, and the caudal spleen paren-chyma was ligated with 6-0 silk ties. The spleen wasplaced back in the abdomen, and the muscle andskin were closed with 4-0 Vicryl suture.

The animals were monitored for #14 days aftertransplantation. They were weighed and examineddaily. Animals were killed in accordance withpredefined humane care criteria.

Experimental groups. Cell transplantation wasperformed using hepatocyte-like cells derived from

3 cell types: iPSC, MSC, and iPSC + EC. The iPSC(WiCell Research Institute, Madison, WI), the MSC(ScienceCell, Carlsbad, CA) and the iPSC com-bined with human adipose-derived EC (Science-Cell) were differentiated into hepatocyte-like cellsusing a 4-stage, embryoid body-based, differentialprotocol in suspension developed based on thein vivo differentiation process of human hepato-cytes.7 Cells differentiated using this protocoldemonstrated in vitro hepatocyte morphology, al-bumin synthesis, urea metabolism, and sequentialmRNA expression and protein expression of thehepatocyte markers SOX-17, FOXA2, Hhex,GATA4, HNF-4a, AFP, albumin, and CK-18. Fortransplantation, 100 clusters (approximately1.2 3 106 cells) of iPSC-derived, MSC-derivedand iPSC + EC-derived hepatocyte-like cells wereinjected into the body of the spleen via the caudalspleen pole through an 18-gauge needle. The con-trol group consisted of transplantation with 500 mLof HMM vehicle.

Endpoints. The primary endpoint used wasmortality with complete survival defined as survivalto 14 days after cell transplantation. Secondaryendpoints included serum ALT, presence of hu-man albumin in the rat serum, and histologic

ARTICLE IN PRESSSurgeryj 2015

4 Ramanathan et al

evidence of human cells in the rat liver and spleen.Concentration of serum ALT in whole blood wasmeasured using VetScan 2.0 (Abaxis). The tail veinwas used to obtain blood samples before celltransplantation, at 48–72 hours after transplanta-tion, and at killing when possible. The presence ofhuman albumin in the rat serum was evaluatedusing a Human Albumin ELISA Quantitation Setthat was non-cross reactive with rat albumin(Bethyl Laboratories, Montgomery, TX).

Liver and spleen samples were recovered at timeof death and fixed with 10% neutral bufferedformalin. The cells were embedded in paraffin andsectioned with hematoxylin and eosin staining forhistologic assessment. The paraffin-embeddedslides were deparaffinized using xylene substituteand ethanol. Immunohistochemistry and immu-nofluorescence were performed on liver andspleen sections to identify the presence of humanhepatocyte markers and human ECs.

Immunohistochemistry. For detection of humanalbumin, after nonspecific binding was blockedwith 2% donkey serum (Sigma-Aldrich), the slideswere incubated with a non–cross-reactive goatantibody to human albumin primary antibody(Bethyl Laboratories; 1:500) for 60 minutes. Thesecondary antibody used was HRP-conjugateddonkey antibody to goat immunoglobulin (Ig)G(Santa Cruz Biotechnology, Dallas, TX; 1:200) for60 minutes. For detection of ECs, a non–cross-reactive mouse antibody to human platelet ECadhesion molecule (PECAM)-1 (Santa Cruz) wasused as the primary antibody. PECAM-1, alsoknown as CD31, is expressed in multiple hemato-poietic cell types and has high expression onhuman ECs. A goat antibody to mouse IgG1-HRP(AbD Serotec, Raleigh, NC) was used as the sec-ondary antibody for 60 minutes. Metal-enhanceddiaminobenzidine substrate was used to activatethe horseradish peroxidase in both cases (Thermo-Scientific, Waltham, MA). The slides were counter-stained with Mayer’s hematoxylin (Sigma-Aldrich).

Immunofluorescence. After deparaffinization, anti-gen retrieval was achieved through boiling theslides at 1008C for 10 minutes. Permeabilizationwas performed with 0.3% (v/v) Triton-X 100 for1 hour and nonspecific blocking was performedwith 0.5% (v/v) goat serum (Sigma-Aldrich) for1 hour. The slides were incubated with thefollowing primary antibodies overnight at 48C:mouse anti-human HNF-3b, goat anti-human albu-min, rabbit anti-human C-MET, and mouse anti-human PECAM-1 (Santa Cruz). The followingsecondary antibodies were used: Cy2-AffiniPuregoat anti-mouse IgG (Fc Subclass 1 Specific),

Cy2-AffiniPure goat anti-mouse IgG (Fc Subclass2 Specific), Cy3-AffiniPure donkey anti-goat IgG(H + L), and Cy5-conjugated AffiniPure goat anti-rabbit IgG (H + L; Jackson ImmunoResearch, WestGrove, PA). Nuclei were counterstained with4’6’-diamidino-2-phenylindole (DAPI) for 10 mi-nutes. Images were acquired with confocal micro-scopy using Olympus IX81.

For statistical analysis, categorical variables wereevaluated using Fisher’s or Chi-square tests andcontinuous variables were evaluated with Student ttests or analysis of variance. All statistical analyseswere performed using JMP 9.0 (Stata Corp LP, Col-lege Station, TX).

RESULTS

D-galactosamine can reproducibly induce lethalALF. Twelve immunocompetent Lewis rats and 14athymic nude rats were used to build the animalmodel of ALF. Serum ALT was used as the primarybiomarker to monitor liver failure. A dose of950–1,000 mg/kg resulted in a mean increase ofALT from 64 to 3,945 U/L 1 day after induction ofALF. Histologically, D-galactosamine causedmassive hepatic necrosis (Fig 2).

Among the 10- to 14-week-old athymic nude rats(n = 14) who achieved an initial increase in serumALT of 3,000 U/L (9/14 rats), the 3-day mortalityafter induction of ALF with 975 mg/kg of D-galac-tosamine was 89% (8/9 rats). Those that had anALT of <3,000 U/L had a 3-day mortality of 40%(Fig 3). Those with a sublethal injury demon-strated spontaneous resolution of liver failure, asdemonstrated by near normalization of ALT andimprovement in signs of hepatic encephalopathyand weight gain. In those rats with spontaneousresolution of liver failure, reversal of signs ofhepatic encephalopathy was noted approximatelyfive days after induction of ALF. Given these find-ings, 10- to 14-week-old nude athymic rats with aninitial ALT of >3,000 U/L on the first day afterinduction of ALF were used for experimentation.

Stem cell xenotransplant resulted in increasedsurvival and in vivo function. Rats transplantedwith iPSC (n = 14) had 71% survival (10/14) to14 days, those with iPSC + EC (n = 11) had 75%survival (8/11) to 14 days, and those with MSC(n = 8) had 73% survival (6/8) to 14 days (TableI). Controls (n = 6) had 14% survival (1/6) to14 days. All the transplanted animals demonstrateda greater survival to 14 days (P < .05; Fig 4).Compared with controls, there was increased sur-vival to 14 days in animals transplanted withiPSC-derived and MSC-derived hepatocyte-like

Fig 2. Representative liver section with hematoxylin and eosin staining displaying widespread hepatic necrosis noted1 day after induction of acute liver failure with 975 mg/kg of D-galactosamine. Left panel displays10 3 magnification and right panel 40 3 magnification. Box identifies magnified area. Scale bar represents 50 mm.

Fig 3. Kaplan–Meier survival curve of 10- to 14-week-oldcontrol animals. Those animals that incurred a liverinjury with an alanine aminotransferase (ALT) level of>3,000 U/L at 1 day after D-galactosamine injectionhad a 3-day mortality of 8 of 9, compared with a 3-daymortality of 2 of 5 in those with an ALT of <3,000 U/L.

Fig 4. Kaplan–Meier survival plot of animals after stemcell transplantation, indicating greater survival in allstem cell groups (P < .05). EC, Endothelial cells; iPSC,induced pluripotent stem cells; MSC, mesenchymalstem cell.

ARTICLE IN PRESSSurgeryVolume j, Number j

Ramanathan et al 5

cells (P < .05 each). Compared with controls, ani-mals transplanted with iPSC-derived, iPSC + EC-derived, and MSC-derived hepatocyte-like cellshad a greater mean survival (P < .05).

Evaluation of serum ALT levels indicated nodifference in baseline ALT between the variousgroups (Table II). Among the groups, the iPSCgroup had a lesser pretransplant ALT level thanthe MSC group (3,765 vs 5,104 U/L; P < .05),but the values were not different from the controlgroup. Day 3 ALT values include animals that sur-vived to 14 days and animals that did not. Whencomparing survivors to 14 days posttransplantand nonsurvivors, there was no difference in pre-transplant ALT (4,238 vs 4,689 U/L; P = .15), butby day 3 posttransplant, ALT was significantly lessamong those rats that survived compared with non-survivors (650 vs 3,677 U/L; P < .01). ALT values at

time points beyond 3 days are not displayed owingto low number of rats still alive at the subsequenttime points.

Analysis of rat serum for the presence of humanalbumin using a non–cross-reactive assay demon-strated human albumin in 14 of the 31 rats thatwere tested on posttransplant day 3 (Table I). Hu-man albumin was detected in the rat serum 3 daysafter transplant in 58% (7/12) of the rats trans-planted with iPSC-derived hepatocytes, 63% (5/8) of rats transplanted with iPSC/EC-derived hepa-tocytes, and 22% (2/9) of rats transplanted withMSC-derived hepatocytes. Rat serum human albu-min concentration at 3 days posttransplant rangedfrom 2.6 to 14.0 ng/mL. There was no correlationbetween albumin concentration and ALT or sur-vival. None of the controls demonstrated humanalbumin in the rat serum. At 14 days after trans-plantation, 2 of the 4 surviving rats transplanted

Table II. Comparison of rat serum alanine transferase (ALT) levels between transplant groups and betweensurvivors and nonsurvivors

Variable Baseline ALT (U/L) Pretransplant ALT (U/L) Day 3 ALT (U/L)

iPSC 53 ± 7 3,765 ± 858* 1,094 ± 1,531MSC 53 ± 7 5,104 ± 597* 2,000 ± 1,088iPSC + EC 54 ± 5 4,236 ± 858 1,686 ± 2,337Control 51 ± 6 4,621 ± 924 3,632 ± 2,139P value .90 <.01 .06Survivors 52 ± 5 4,238 ± 969 650 ± 634*Nonsurvivors 54 ± 7 4,689 ± 853 3,676 ± 1,873*P value .36 .15 <.01

*Specific values are significantly different (P < .05).EC, Endothelial cells; iPSC, induced pluripotent stem cells; MSC, mesenchymal stem cell; Survivors, survival to 14 days posttransplant.

ARTICLE IN PRESSSurgeryj 2015

6 Ramanathan et al

with iPSC + EC–derived hepatocyte-like cells dis-played serum presence of human albumin. Noother group had human albumin in the rat serum14 days after cellular transplantation.

Immunostaining to detect human albumin inthe rat liver and spleen sections was performed inall the animals. Additionally, in animals trans-planted with iPSC + EC–derived hepatocyte-likecells, immunostaining was performed for detectionof human ECs through a PECAM-1 antibody.Figure 5 is a representative image from sectionsof a representative rat transplanted withiPSC + EC–derived hepatocyte-like cells displayingthe presence of human albumin in the liver andECs in the spleen. In those animals transplantedwith iPSC and MSC-derived hepatocyte-like cells,human albumin was detected only in the rat liversand not in the spleens. In those animals trans-planted with iPSC + EC and MSC + EC–derivedhepatocyte-like cells, human PECAM-1 for ECdetection was identified only in the rat spleensand not in the livers. There was no human albuminor EC staining in any sections among the controls.Similarly, immunofluorescence staining displayedscattered presence of human hepatocyte markersHNF-3b, albumin, and C-MET in the liver sectionswith an absence of human hepatocyte markers inthe spleen (Fig 6). There was diffuse presence ofhuman PECAM-1 in the spleen tissue (Fig 6).

DISCUSSION

In a reproducible, lethal, D-galactosaminemodel of ALF, stem cell-derived, hepatocyte-likecells were capable of rescuing animals from ALF.Stem cell-derived, hepatocyte-like cells demon-strated albumin production in vivo and evidenceof engraftment in the host liver. Hepatocyte-likecells combined with ECs tended to demonstrateprolonged activity in vivo.

Hepatocyte transplantation can be an effectivebridge to solid organ transplantation or to nativeliver recovery for patients with life-threatening liverfailure. To date, hepatocyte transplantation hasbeen documented in 143 human patients world-wide, with indications including metabolic liverdisease, chronic liver failure, and ALF.1,10,11 Forpatients who are not transplant candidates owingto concurrent infections, hepatocellular carci-noma, prohibitive medical comorbidities, or unsta-ble social supports, hepatocyte transplantation canprolong life and serve as a bridge to transplanteligibility.1 Additionally, hepatocyte transplanta-tion may maximize utility of scarce donor organs.Most clinical hepatocyte transplant protocolsinvolve hepatocyte quantities ranging from1 3 104 to 1 3 109 hepatocytes per patient; theaverage hepatocyte yields from a liver are on theorder of 1–3 3 1010 viable cells from a whole liveror 7–20 3 106 hepatocytes per gram of liver tis-sue.12 Limited evidence also notes a role for theuse of cryopreserved hepatocytes, which wouldovercome the unpredictability associated withwhole-organ availability.

The full potential of primary human hepatocytetransplantation has yet to be realized owing toorgan scarcity for hepatocyte isolation. Stem cellsoffer an attractive alternative to primary humanhepatocytes for hepatocyte transplantation.Although the use of embryonic stem cells isplagued by ethical controversy, iPSC and MSCrepresent a source of multipotent stem cellscapable of differentiating into a number of targetcell types, including hepatocytes.3,13-15 Addition-ally, reports suggest that MSC may be immunopri-vileged, thereby enabling minimization oravoidance of immunosuppression.6,16 Among thebest studied sources for MSC is adult bone marrow,harboring the potential for a patient’s own bone

Fig 5. Representative liver and spleen sections with immunohistochemical staining. Background staining with hematoxylin.Scale bar represents 2.5 mm. (A) Normal human liver section (positive control) stained for human albumin with103magnification on left panel and 203magnification on right panel. Precipitate indicates presence of antibody to humanalbumin. Black arrows indicate presence of human albumin both intracellularly and extracellularly. (B) Liver section from acontrol rat that was transplanted with medium only (negative control). Section stained for human albumin with103magnificationon leftpanel and203magnificationonrightpanel. Intracellular andextracellular absenceofnon-rat crossreactive antibody to human albumin are noted. (C) Representative image of liver section from a rat transplanted with inducedpluripotent stem cells (iPSC) plus endothelial cell (EC)-derived hepatocyte-like cells with 203magnification on the left paneland 403magnification on the right panel. Immunohistochemical staining performed with non-rat cross-reactive antibody tohuman albumin. Precipitate indicates presence of antibody to human albumin. Black arrows indicate presence of intracellularhuman albumin. (D) Representative image of a spleen section from a rat transplanted with iPSC + EC-derived hepatocyte-likecells with 203magnificationon the left panel and 403magnificationon the right panel. Immunohistochemical stainingwithnon-rat cross-reactive antibody tohumanECmarker plateletECadhesionmolecule (PECAM)-1. Brown staining indicates pres-ence of antibody to human PECAM-1. White arrows indicate human PECAM-1 positive ECs within the spleen.

ARTICLE IN PRESSSurgeryVolume j, Number j

Ramanathan et al 7

Fig 6. Representative liver and spleen sections with immunofluorescence staining. Nuclear staining with DAPI. Scalebar represents 100 mm. (A) Normal human liver section (positive control) stained for human HNF-3b, human albuminand human C-MET. Image displays presences of all 3 markers. (B) Representative image of liver section from a rat trans-planted with mesenchymal stem cells (MSC) and endothelial cell (EC)-derived hepatocyte-like cells. Immunofluores-cence staining with non-rat cross-reactive antibodies to hepatocyte markers human HNF-3b, human albumin, andhuman C-MET. Image displays scattered low-frequency presence of all 3 markers. (C) Representative image of a spleensection from a rat transplanted with MSC + EC–derived hepatocyte-like cells. Immunofluorescence staining with non-ratcross reactive antibodies to EC marker human platelet EC adhesion molecule (PECAM)-1, human albumin and humanC-MET. Image displays presence of human PECAM-1 but no presence of human albumin or C-MET.

ARTICLE IN PRESSSurgeryj 2015

8 Ramanathan et al

marrow-derived MSC being used for autologouscell transplantation of differentiated hepato-cytes.17 In our study, we noted that MSC-derivedhepatocyte-like cells displayed functionality aftertransplantation, as evidenced by in vivo productionof human albumin and resulted in increased sur-vival for #14 days compared with control animals.

iPSC are generated by dedifferentiation ofterminally differentiated adult cells back to apluripotent state. Common sources include epithe-lial cells, keratinocytes, and fibroblasts.18,19 Dedif-ferentiation can be achieved through the use ofviral vectors for retrotranscription or electropora-tion with stemness factors.18,20 Although MSC area source of multipotent stem cells, iPSC are apluripotent source of cells, possessing a greaterbreadth of differentiation potential. Similar to

MSC, there exists the potential to obtain iPSCfrom a target patient for cell transplantation ofautologous iPSC-derived hepatocyte-like cells,thereby mitigating the shortage of cells for trans-plant and risks of immunosuppression and im-mune sensitization. The suspension-based,embryoid body technology used in this study repre-sents a novel scalable method that results in 0.5- to1-mm differentiated cell clusters that are stablymaintained in cell suspension.7 This technique en-ables improved transport and transplantation,obviating the need for detachment from adher-ence culture described in other reports.12 Theuse of iPSC for transplantation in this study re-sulted in in vivo production of human albuminand increased survival for #14 days comparedwith control animals.

ARTICLE IN PRESSSurgeryVolume j, Number j

Ramanathan et al 9

Animals transplanted with iPSC, MSC, andiPSC + EC–derived hepatocyte-like cells in ourstudy demonstrated the presence of human albu-min in the rat serum at 3 days, although thefrequency was less among those in the MSC group.Only the animals transplanted with iPSC + EC–derived, hepatocyte-like cells displayed a sustainedpresence of human albumin in the serum of 2 of 4surviving rats at 14 days after cell transplantation.Even though all the cell types were differentiatedinto hepatocyte-like cells using the same differen-tiation protocol and displayed in vitro hepatocytefunction before transplantation, there was a largerange in the concentration of human albumin inthe serum of the rats. The concentration of ratserum albumin did not correlate with survival orALT. The disparate presence and quantity ofhuman albumin in the rat serum may be owingto variable engraftment and functional expressionor a type II error owing to inadequate power. Poorengraftment, primary nonfunction, and transiencyof function have long been documented barriersafter cell transplantation.21 Poor function andtransiency of function have been attributed to acombination of factors including inflammatory-mediated, nonspecific destruction, cellular rejec-tion, cellular ischemia owing to poor engraftment,unfavorable local environment, and shear stress invessels.

To address the poor engraftment and promoteneovascularization, we investigated the impact ofcombining the iPSC cells with ECs before differ-entiation. Embryologic development of the fetalliver occurs in conjunction with EC proliferation;indeed, ECs constitute 19% of the total liver cellmass.22 Additionally, ECs promote hepatic differ-entiation through inhibition of the Wnt and Notchpathways.23 Han et al23 demonstrated that an ECniche modulates the Wnt and Notch pathways,thereby expanding endodermal proliferation andimproving cellular fate signaling toward hepato-cytes. In our study, we did not note an increasedrate of albumin production at 3 days among theiPSC + EC group; however, we did find that theaddition of ECs to iPSC resulted in sustained func-tion as noted by the presence of human albuminin the rat serum. The difference in sustained func-tion likely did not result in any survival benefit,because survival was likely a result of early functionof the cells transplanted and bridging to nativerecovery in our model of ALF. We found thatALT levels among eventual survivors began tonormalize by day 3, indicating that bridging bythe transplanted cells in the first 1-3 days allowednative hepatocyte regeneration. One would not

expect the transplanted to cells to decrease theobserved increase in ALT in the rats, because theALT increase reflects sustained hepatocyte injuryfrom the D-galactosamine administered previouslyto cause the ALF. The decrease, therefore, repre-sents amelioration of ongoing injury and recoveryof those previously injured hepatocytes. Investiga-tion of potential survival benefits of sustained activ-ity would be better evaluated with fibrosis oranhepatic models. Further investigation isrequired as to the precise mechanisms by whichECs improve sustained function.

Histologically, human albumin–producing cellswere noted only in the liver of the transplantedanimals and not in the spleens by immunohisto-chemistry and immunofluorescence. Human ECs,however, were only identified in the rat spleen andnot the rat liver at 14 days after transplantation.This observation could suggest that the ECs offerimproved initial engraftment in the spleen andthat the natural history of hepatocyte-like cellsinjected into spleen is to home or embolize to theliver where they are able to be engrafted, sup-ported, and subsequently function. Embolizationto the liver of cells transplanted into spleen is awell-documented phenomenon and is likely adap-tive given that the native microcirculation of theliver that contain the trophic factors necessary forhepatocyte sustenance and growth.24 Additionalinvestigation is required to identify the mecha-nisms involved in hepatocyte homing to the liver.Brulport et al8 have demonstrated 2 distinct path-ologic findings associated with human albuminstaining cells in mouse liver using probes for insitu hybridization and polymerase chain reaction.They concluded that human albumin staining inmice livers after human hepatocyte cell transplan-tation can be caused by differentiating humanhepatocyte-like cells or by apoptotic human cellfragments in mouse livers. Based on our data, thepresence of human albumin in the rat serum sug-gests that the cells seen on immunohistochemistryrepresent functional transplanted cells producinghuman albumin. Additional investigation isrequired to quantify the stage of differentiation.

The most valuable role of cell therapy in liverfailure is likely to be as a bridge therapy to rescuepatients with ALF or fulminant decompensatedchronic liver failure who face imminent death.Currently, the only widely accepted and availabletherapy is solid organ liver transplantation. Usingstem cell-derived hepatocytes, multiple potentialadvantages and possibilities exist. The use ofcryopreserved, stem cell-derived hepatocytes mayhelp to bridge patients in fulminant failure,

ARTICLE IN PRESSSurgeryj 2015

10 Ramanathan et al

whereas autologous transplantation may provide avaluable time bridge for a suitable organ donorliver to become available.2,11 Given current limita-tions with prolonged function, application as abridge therapy seems the most realistic therapy.

As such, we used an animal model of ALF that islethal in 2–3 days without intervention in thisstudy. D-galactosamine is an amine sugar thatresults in reproducible, dose-dependent, acutehepatic necrosis. Mouse hepatocytes were used todevelop a xenotransplant model in nude athymicrats to eventually test the hepatocyte-like cells ofhuman origin. They also served as a positive con-trol to ensure that the model of liver failure wassensitive enough to detect reversal of injury andresult in survival. Survival was used as the primaryendpoint to assess the efficacy of hepatocyte trans-plantation, because changes in ALT or other meta-bolic parameters could be confounded by nativeliver regeneration. Surviving animals were hu-manely killed after 14 days based on findingsfrom our initial model development with sublethalliver failure and mouse hepatocytes. In those ex-periments, maximal mortality occurred 2–4 daysafter induction of ALF and survivors demonstratedreversal of signs of hepatic encephalopathy by post-transplant day 3 with near-normalization of weightand ALT by posttransplant days 10–14. Becausethis was a model of ALF rather than liver fibrosis,the transplanted cells served as a bridge to nativerecovery. This experimental approach allowedassessment of short-term hepatocyte function.The in vivo function of the transplanted cells wasevaluated by production of human albumin inthe rat serum. The detection of human albuminin the majority of animals provides further evi-dence of direct, hepatocyte-like function as thebridge to animal recovery, rather than a cellular‘‘sink’’ effect whereby the transplanted cells servedto sequester or preferentially absorb the inflamma-tory effect of the induced liver failure. Despite ouruse of a model of ALF, the conclusions may alsoextend to chronic liver failure as well, but investi-gation using chronic liver failure models will berequired.

In conclusion, in a lethal animal model of ALF,splenic transplantation of iPSC and MSC-derivedhepatocyte-like cells displayed reversal of liverfailure and improved survival over controls. Trans-plantation of these cells demonstrated in vivofunction as noted by the production of humanalbumin, with prolonged function noted only inthose hepatocyte-like cells derived from iPSC com-bined with ECs. The hepatocyte-like cells homed tothe injured liver, and human ECs were only

identified in the spleen. Stem cells offer a prom-ising future for hepatocyte cell transplantation thatcould bridge the current gap created by organscarcity. Further studies investigating co-culturewith supporting cell types and the dynamics oftransplanted cells may help optimize stem celltransplantation.

REFERENCES

1. Fisher RA, Strom SC. Human hepatocyte transplantation:worldwide results. Transplantation 2006;82:441-9.

2. Hughes RD, Mitry RR, Dhawan A. Current status of hepato-cyte transplantation. Transplantation 2012;93:342-7.

3. Schwartz RE, Reyes M, Koodie L, Jiang Y, Blackstad M, LundT, et al. Multipotent adult progenitor cells from bonemarrow differentiate into functional hepatocyte-like cells.J Clin Invest 2002;109:1291-302.

4. Hannan NR, Segeritz CP, Touboul T, Vallier L. Productionof hepatocyte-like cells from human pluripotent stem cells.Nat Protoc 2013;8:430-7.

5. Aoi T, Yae K, Nakagawa M, Ichisaka T, Okita K, Takaha-shi K, et al. Generation of pluripotent stem cells fromadult mouse liver and stomach cells. Science 2008;321:699-702.

6. Le Blanc K, Tammik C, Rosendahl K, Zetterberg E, RingdenO. HLA expression and immunologic properties of differ-entiated and undifferentiated mesenchymal stem cells.Exp Hematol 2003;31:890-6.

7. Pettinato G, Wen X, Zhang N. Formation of well-definedembryoid bodies from dissociated human induced pluripo-tent stem cells using microfabricated cell-repellent micro-well arrays. Sci Rep 2014;4:7402.

8. Brulport M, Schormann W, Bauer A, Hermes M, Elsner C,Hammersen FJ, et al. Fate of extrahepatic human stemand precursor cells after transplantation into mouse livers.Hepatology 2007;46:861-70.

9. Cauli O, Rodrigo R, Boix J, Piedrafita B, Agusti A, Felipo V.Acute liver failure-induced death of rats is delayed or pre-vented by blocking NMDA receptors in brain. Am J PhysiolGastrointest Liver Physiol 2008;295:G503-11.

10. Puppi J, Strom SC, Hughes RD, Bansal S, Castell J, Dagher I,et al. Improving the techniques for human hepatocytetransplantation: report from a consensus meeting in Lon-don. Cell Transplant 2012;21:1-10.

11. Fox IJ, Chowdhury JR. Hepatocyte transplantation. AmJ Transplant 2004;4(Suppl 6):7-13.

12. Li AP. Human hepatocytes: isolation, cryopreservation andapplications in drug development. Chem Biol Interact2007;168:16-29.

13. Baiguera S, Jungebluth P, Mazzanti B, Macchiarini P. Mesen-chymal stromal cells for tissue-engineered tissue and organreplacements. Transpl Int 2012;25:369-82.

14. Snykers S, De Kock J, Rogiers V, Vanhaecke T. In vitro differ-entiation of embryonic and adult stem cells into hepato-cytes: state of the art. Stem Cells 2009;27:577.

15. Shi LL, Liu FP, Wang DW. Transplantation of human umbil-ical cord blood mesenchymal stem cells improves survivalrates in a rat model of acute hepatic necrosis. Am J MedSci 2011;342:212-7.

16. Krampera M, Glennie S, Dyson J, Scott D, Laylor R, Simp-son E, et al. Bone marrow mesenchymal stem cells inhibitthe response of naive and memory antigen-specific T cellsto their cognate peptide. Blood 2003;101:3722-9.

ARTICLE IN PRESSSurgeryVolume j, Number j

Ramanathan et al 11

17. Kern S, Eichler H, Stoeve J, Kluter H, Bieback K. Compara-tive analysis of mesenchymal stem cells from bone marrow,umbilical cord blood, or adipose tissue. Stem Cells 2006;24:1294-301.

18. Aasen T, Raya A, Barrero MJ, Garreta E, Consiglio A, Gon-zalez F, et al. Efficient and rapid generation of inducedpluripotent stem cells from human keratinocytes. Nat Bio-technol 2008;26:1276-84.

19. Takahashi K, Tanabe K, Ohnuki M, Narita M, Ichisaka T, To-moda K, et al. Induction of pluripotent stem cells from adulthuman fibroblasts by defined factors. Cell 2007;131:861-72.

20. Yu J, Vodyanik MA, Smuga-Otto K, Antosiewicz-Bourget J,Frane JL, Tian S, et al. Induced pluripotent stem cell linesderived from human somatic cells. Science 2007;318:1917-20.

21. Stutchfield BM, Forbes SJ, Wigmore SJ. Prospects for stemcell transplantation in the treatment of hepatic disease.Liver Transpl 2010;16:827-36.

22. Kmiec Z. Cooperation of liver cells in health and disease.Adv Anat Embryol Cell Biol 2001;161:III-XIII; 1–151.

23. Han S, Dziedzic N, Gadue P, Keller GM, Gouon-Evans V. Anendothelial cell niche induces hepatic specificationthrough dual repression of wnt and notch signaling. StemCells 2011;29:217-28.

24. Ponder KP, Gupta S, Leland F, Darlington G, FinegoldM, DeMayo J, et al. Mouse hepatocytes migrate to liverparenchyma and function indefinitely after intrasplenictransplantation. Proc Natl Acad Sci U S A 1991;88:1217-21.