Molecular perturbations restrict potential for liver repopulation of hepatocytes isolated from non–heart-beating donor rats


  • Yuta Enami,

    1. Marion Bessin Liver Research Center, Department of Medicine, Albert Einstein College of Medicine, Bronx, NY
    2. Department of Surgery, Division of General and Gastroenterological Surgery, School of Medicine, Showa University, Shinagawa-ku, Tokyo, Japan
    Search for more papers by this author
  • Brigid Joseph,

    1. Marion Bessin Liver Research Center, Department of Medicine, Albert Einstein College of Medicine, Bronx, NY
    Search for more papers by this author
  • Sriram Bandi,

    1. Marion Bessin Liver Research Center, Department of Medicine, Albert Einstein College of Medicine, Bronx, NY
    Search for more papers by this author
  • Juan Lin,

    1. Department of Epidemiology & Population Health,Albert Einstein College of Medicine, Bronx, NY
    Search for more papers by this author
  • Sanjeev Gupta

    Corresponding author
    1. Marion Bessin Liver Research Center, Department of Medicine, Albert Einstein College of Medicine, Bronx, NY
    2. Department of Pathology,Albert Einstein College of Medicine, Bronx, NY
    3. Cancer Research Center,Albert Einstein College of Medicine, Bronx, NY
    4. Diabetes Research Center,Albert Einstein College of Medicine, Bronx, NY
    5. Ruth L. and David S. Gottesman Institute for Stem Cell and Regenerative Medicine Research,Albert Einstein College of Medicine, Bronx, NY
    6. Institute for Clinical and Translational Research, Albert Einstein College of Medicine, Bronx, NY
    • Albert Einstein College of Medicine, Ullmann Bldg., Rm 625, 1300 Morris Park Avenue, Bronx, NY 10461===

    Search for more papers by this author
    • fax: 718-430-8975

  • Potential conflict of interest: Nothing to report.

  • Supported by National Institutes of Health grants R01 DK071111, R01 DK088561, P30 DK41296, and P30 CA13330.


Organs from non–heart-beating donors are attractive for use in cell therapy. Understanding the nature of molecular perturbations following reperfusion/reoxygenation will be highly significant for non–heart-beating donor cells. We studied non–heart-beating donor rats for global gene expression with Affymetrix microarrays, hepatic tissue integrity, viability of isolated hepatocytes, and engraftment and proliferation of transplanted cells in dipeptidyl peptidase IV-deficient rats. In non–heart-beating donors, liver tissue was morphologically intact for >24 hours with differential expression of 1, 95, or 372 genes, 4, 16, or 34 hours after death, respectively, compared with heart-beating donors. These differentially expressed genes constituted prominent groupings in ontological pathways of oxidative phosphorylation, adherence junctions, glycolysis/gluconeogenesis, and other discrete pathways. We successfully isolated viable hepatocytes from non–heart-beating donors, especially up to 4 hours after death, although the hepatocyte yield and viability were inferior to those of hepatocytes from heart-beating donors (P < 0.05). Similarly, although hepatocytes from non–heart-beating donors engrafted and proliferated after transplantation in recipient animals, this was inferior to hepatocytes from heart-beating donors (P < 0.05). Gene expression profiling in hepatocytes isolated from non–heart-beating donors showed far greater perturbations compared with corresponding liver tissue, including representation of pathways in focal adhesion, actin cytoskeleton, extracellular matrix–receptor interactions, multiple ligand–receptor interactions, and signaling in insulin, calcium, wnt, Jak-Stat, or other cascades. Conclusion: Liver tissue remained intact over prolonged periods after death in non–heart-beating donors, but extensive molecular perturbations following reperfusion/reoxygenation impaired the viability of isolated hepatocytes from these donors. Insights into molecular changes in hepatocytes from non–heart-beating donors offer opportunities for improving donor cell viability, which will advance the utility of non–heart-beating donor organs for cell therapy or other applications. (HEPATOLOGY 2012)

Early experiences of liver cell therapy have been promising.1 However, worsening shortages of donor organs for isolating cells is a major hurdle. Therefore, alternative sources of donor cells will be highly significant, although current alternatives are unsatisfactory. For instance, the availability of surgically resected healthy liver specimens is sporadic, such specimens are generally small, and specimens adjacent to tumors may harbor malignant cells. Living related organ donation solely for isolating cells poses risks to the donor and seems premature because the efficacy of cell therapy needs to be proved by further work. Similarly, the ability to generate hepatocytes by differentiating embryonic, fetal, or adult stem cells is in its infancy. Whether extrahepatic cells, i.e., hematopoietic, mesenchymal, or other cells, could replace liver tissue is controversial. By contrast, non–heart-beating (NHB) cadaveric donors offer a large and widely available resource. Studies of cadaveric fetal livers have indicated that highly viable hepatocytes could be isolated many hours after death and tissue anoxia.2 These fetal cells expanded extensively in culture and could be cryopreserved multiple times. This differed from adult tissues and mature hepatocytes, which are much less anoxia tolerant. Although resistance to anoxia is not an intrinsic biological property of mammalian cells, as some animal species do resist anoxia,3 there has been investigation of molecular mechanisms in tissue viability, e.g., brain, after hypoxia or anoxia.4 In contrast, molecular changes in the liver after anoxia, including hepatocytes from NHB donors, have not been determined.5-7

To understand the nature of molecular perturbations in NHB donor liver, including isolated hepatocytes from such livers in comparison with heart-beating (HB) donor liver, we studied Fischer 344 rats. The fate of transplanted cells was determined in dipeptidyl peptidase IV-deficient (DPPIV−) rats, in which syngeneic DPPIV+ cells are readily identified.8 Preconditioning of recipient rats by retrorsine, a genotoxic alkaloid, plus two-thirds partial hepatectomy (PH) induces transplanted cell proliferation to permit analysis of liver repopulation.9 These systems provided insights into molecular changes in NHB donor livers, as well as the cell therapy potential of NHB donor livers.


DPPIV, dipeptidyle peptidase IV; ECM, extracellular matrix; HB, heart-beating; LPF, low-power field; NHB, non–heart-beating; PH, partial hepatectomy; TUNEL, terminal deoxynucleotidyl transferase biotin–dUTP nick end-labeling.

Materials and Methods


F334 rats were 10-12 weeks old (National Cancer Institute, Bethesda, MD). Syngeneic DPPIV− F334 rats were 6-8 weeks old and were from the Special Animal Core of the Marion Bessin Liver Research Center (Bronx, NY). Rats were housed at the Institute for Animal Studies of Albert Einstein College of Medicine under 14/10-hour light/dark cycles with unrestricted food and water. The Animal Care and Use Committee at Albert Einstein College of Medicine approved the animal protocols in conformity with National Institutes of Health requirements.

For NHB donor rats, midline laparotomy was performed under anesthesia, and a 20-gauge cannula was placed in the portal vein, and filled with 0.2 mL of normal saline containing 200 U of heparin (NDC 63323-540-31; American Pharmaceutical Partners Inc., Schaumburg, IL). Immediately afterwards, 0.2 mL of 15% potassium chloride (KCl) was injected into the inferior vena cava. When cells were not to be isolated, KCl was injected into tail vein or heart for cardiac arrest. Cadavers were stored at 4°C. HB donor rats were controls.

Hepatocyte Isolation.

Chemicals were from Sigma Chemical Co. (St. Louis, MO). Collagenase was from Worthington Corp. (Lakewood, NJ). Hepatocytes were isolated by two-step collagenase perfusion,10 except glucose was substituted with 20 mM fructose (F0127, Sigma), according to Anundi et al.11

Liver Repopulation Assay.

At 6 and 8 weeks of age, DPPIV− rats received 30 mg/kg retrorsine intraperitoneally, and 4 weeks later, PH was performed according to the Higgins and Anderson method. Cells were transplanted 7 days after PH. Liver repopulation analysis was done 3 weeks after cell transplantation.

Cell Viability.

Cells were manually counted in a Neubauer chamber with exclusion of 0.2% trypan blue dye. For attachment to culture dishes, 3 × 104 cells/cm2 were plated in Roswell Park Memorial Institute (RPMI) 1640 medium with serum and antibiotics. For apoptosis, DNA was extracted (DNeasy Blood & Tissue Kit; Qiagen Inc., Valencia, CA), and resolved in 1% agarose gels with ethidium bromide.

Metabolic Capacity of Cells.

Cells cultured overnight in dishes at 3 × 104 cells/cm2 were incubated with 5 mM ammonium chloride in RPMI 1640 medium for 2 hours at 37°C. To 100-μL aliquots of medium, we added 0.3 mL of urease buffer reagent (U-3383; Sigma) for 20 minutes at room temperature. This was followed by 0.6 mL of phenol nitroprusside (P-6994, Sigma); 0.6 mL of alkaline hypochlorite (A-1727; Sigma), and 3.0 mL of water. The mixture was incubated for 30 minutes at room temperature. Absorbance was read at 540 nm. Standard curves were made with 10 mg/mL urea (U-5128; Sigma). The data were corrected for the number of cells attached in dishes, which was measured manually in a Neubauer chamber, after the medium was harvested.

Cell Transplantation.

For cell engraftment analysis, 1 × 107 cells in 0.5 mL serum-free RPMI 1640 were injected as a bolus into spleen isolated by subcostal laparotomy. In rats conditioned with retrorsine and PH, 0.5 × 107 cells were injected in 0.25 mL of RPMI 1640 medium via spleen.

Tissue Studies.

To examine tissue integrity under light microscopy, liver sections from HB and NHB donors were stained by hematoxylin & eosin. For apoptosis, terminal deoxynucleotidyl transferase biotin–dUTP nick end-labeling (TUNEL) was performed (ApopTag peroxidase in situ apoptosis kit; Millipore, Billerica, MA). In negative controls, terminal deoxynucleotidyl transferase was omitted. Color was developed with diaminobenzidine (DAB; K3467; Dako North America Inc., Carpinteria, CA).

Global Gene Expression.

Total cellular RNA was extracted (RNEasy Mini Kit, Qiagen), and hybridized with Rat Genome 230 2.0 arrays (Affymetrix Inc., Santa Clara, CA), according to the manufacturer. Tissue samples were from HB donors immediately after death (n = 3), and from NHB donors 4, 16, and 30 hours after death (n = 3 each, total 12 arrays). Freshly isolated hepatocytes were from HB donors or NHB donors 4 and 16 hours after death (n = 3 each, total 9 arrays).

For data analysis, microarray background was corrected at probe levels by robust multiarray normalization followed by the quintiles method for normalizing perfect-match probe intensities. Normalized raw probe signal values were log2 transformed.12 Ratios of intergroup and intragroup sums of squares (BW ratio) were computed per probe.13 Data sets were segregated for analysis with P values ≤0.01, while controlling false discovery rates to <1% and fold-differences in mean expression of ≥|2.0|. Gene lists of interest were obtained from venn diagrams (SAS software; SAS Institute Inc., Cary, NC). Genes were annotated and grouped by Database for Annotation, Visualization, and Integrated Discovery (DAVID).14, 15 Gene pathways were mapped according to Kyoto Encyclopedia of Genes and Genomes (KEGG) (http://www.

Cell Engraftment and Liver Repopulation Analysis.

Liver samples were frozen to −80°C in methylbutane. Cryosections of 5 μm were fixed in cold chloroform–acetone for DPPIV staining, as previously described.8 For morphometry, three sections per liver lobe were studied per rat with 100 consecutive fields centered on portal areas under ×100 magnification. For liver repopulation, multiple microphotographs were obtained under ×40 magnification by a Spot RT digital camera (Diagnostic Instruments Inc., Sterling Heights, MI), and transplanted cell areas were measured by ImageJ (National Cancer Institute, Bethesda, MD).

Experimental Design.

For tissue integrity, NHB donor livers were studied for 15 minutes, and 2, 4, 6, 8, 10, 16, 24, 30, and 40 hours after death compared with HB donors (n = 3-4 each, total 40). Samples were collected for global gene expression analysis (n = 3 each). Hepatocytes were isolated for cell viability studies (n = 3-6 each). Gene expression was profiled in isolated hepatocytes (n = 3 per condition). We determined engraftment capacity of hepatocytes from HB and NHB donors (n = 3-4 per condition). For cell engraftment, hepatocytes from each donor were transplanted into DPPIV− rats (n = 3-4 each, total 50), and animals were killed 7 days later. For liver repopulation, cells were transplanted from NHB (n = 4) or HB donors (n = 3) into DPPIV− rats preconditioned with retrorsine and PH (n = 3-6 per donor, total 40). These animals were killed after 3 weeks.

Statistical Analysis.

Data are shown as means ± SD. Analyses included t tests, Mann-Whitney tests, or ANOVA with Dunn's test by SigmaStat 3.1 software (SysStat Software, Point Richmond, CA). P < 0.05 was considered significant.


Integrity of NHB Liver Tissue.

Liver was morphologically intact despite several hours after death in NHB donors, including after 15 minutes and 2, 4, 6, 8, 10, 16, 24, 30, or 40 hours (Fig. 1A). Hepatic necrosis or inflammatory infiltrates were absent. Hepatocytes and bile duct cells appeared unremarkable. This was similar to hepatic morphology in HB donors. TUNEL showed limited apoptosis (Fig. 1B,C). Only 0-1 apoptotic cells were found per section under ×200 magnification, up to 24 hours after death, with slightly more apoptosis after 30 hours and 40 hours in NHB donor livers, although still only two or three, or six to eight, TUNEL+ cells were found per section, respectively. DNA laddering confirmed limited apoptosis in NHB donor livers (Fig. 1D).

Figure 1.

Integrity of liver in NHB donors. (A) Hematoxylin & eosin–stained sections from HB and NHB donors after death, as indicated, showing that tissues were intact without necrosis or autolysis. (B) TUNEL+ cells were infrequent in HB and NHB donor livers. (C) Negative control liver without TUNEL and DNAse-treated control liver with extensive TUNEL. (D) DNA laddering showing little hepatic apoptosis over time. Original magnification, (A) and (B), ×400; B, methylgreen counterstain.

These restricted morphological changes in NHB liver were reflected by gene expression profiles (Fig. 2A). Remarkably, only one gene was differentially expressed in NHB livers 4 hours after death: down-regulation of lipid synthesis regulator, stearoyl-coenzyme A desaturase 2. By contrast, gene expression changed more in NHB donor livers 16 hours and 30 hours after death, with differential expression, either up or down versus HB livers, of 95 and 372 genes, respectively. These genes were clustered in relatively few curated KEGG pathways (Fig. 2B). Further study indicated perturbations in discrete pathways, including oxidative phosphorylation, leukocyte migration, cell integrity (adherens junctions), intermediary metabolism, or circadian rhythm (Fig. 2C).

Figure 2.

Gene expression profiles in HB and NHB donor livers. (A) Differentially expressed genes in NHB livers 4, 16, and 30 hours after death versus HB livers. Only one gene was differentially expressed in NHB liver 4 hours after death. After 16 hours, 64 genes were differentially expressed, including the down-regulated gene after 4 hours, and after 30 hours, 342 genes were differentially expressed, including 30 of 64 genes (47%) differentially expressed 16 hours after death in NHB livers. (B) KEGG pathways in differentially expressed gene lists, including numbers of genes in pathways. Total numbers of genes in (B) differed from (A) as the former included one or more transcripts per gene. (C) Differentially regulated pathways in NHB livers 16 or 30 hours after death.

Functional gene groups showed similar perturbations in NHB donor livers 16 hours and 30 hours after death (Table 1). Therefore, tissue changes in NHB donor livers after death were gradual, because 12 hours elapsed from differential expression of 1 gene after 4 hours versus 95 genes after 16 hours, and another 14 hours elapsed for differential expression of 372 genes after 30 hours. However, differentially expressed gene lists in NHB donors did not include genes in apoptosis or cell death pathways, which was in agreement with tissues showing limited apoptosis.

Table 1. Representation of Major Functionally Annotated Groups in differentially Expressed Gene Lists in NHB Donor Liver Versus HB Donor Liver
Gene Ontology GroupsNHB Donor Liver 16 HoursNHB Donor Liver 30 Hours
No. of Up-Regulated Genes (%)No. of Down-Regulated Genes (%)No. of Up-Regulated Genes (%)No. of Down-Regulated Genes (%)
Cellular process  195 (61) 
Metabolic process 27(43)136 (48) 
Cellular metabolic process14 (45)26 (41)143 (44) 
Primary metabolic process14 (45) 136 (42) 
Transport10 (32) 59 (18)11 (17)
Cellular component organization and biogenesis10 (32) 55 (17) 
Localization10 (32) 68 (21) 
Biological regulation10 (32)21 (33) 17 (27)
Regulation of biological process9 (29)19 (30) 15 (24)
Developmental process8 (26)13 (21) 12 (19)
Macromolecule metabolic process9 (29)23 (37)103 (32) 
Establishment of localization10 (32) 61 (19)11 (17)
Protein metabolic process9 (29) 65 (20) 
Cellular protein metabolic process  64 (20) 
Regulation of cellular process9 (29)16 (25)  
Regulation of cellular metabolic process 13 (21)  
Biosynthetic process6 (19) 46 (14) 
Response to stress5 (16)7 (11)26 (8)8 (13)
Negative regulation of biological process5 (16)   
Immune system process5 (16)   
Secretion4 (13)   
Cell death4 (13) 23 (7) 
Leukocyte activation(10)   
Multicellular organismal development(10)11 (17)  
Biopolymer metabolic process 17 (27)  
Regulation of metabolic process 14 (22)  
Nucleobase, nucleoside, nucleotide, and nucleic acid metabolic process 13 (21)  
Transcription, DNA-dependent 11 (17)  
Transcription 11 (17)  
Regulation of transcription, DNA-dependent 11 (17)  
RNA biosynthetic process 11 (17)  
Regulation of gene expression 11 (17)  
RNA metabolic process 11 (17)  
Regulation of transcription 11 (17)  
Intracellular signaling cascade 10 (16)31 (10) 

Mapping of differentially expressed genes along functional pathways, including mitochondrial oxidative phosphorylation, transendothelial leukocyte migration, adherence junctions, and glycolysis/gluconeogenesis was consistent with depletion of energy, need for glucose production, cell–cell interaction-type events (e.g., leukocyte recruitment), and cytoskeletal alterations, in NHB donor livers after death (Supporting Figs. 1-4).

Hepatocytes From NHB Donor Livers Showed Extensive Perturbations.

The yield of hepatocytes from HB donor livers was 300 ± 92 × 106 with viability of 83 ± 2%. HB hepatocytes attached in dishes with 60%-80% efficiency. Cells showed characteristic slightly rounded and then flattened morphology over several hours.

Hepatocyte yield from NHB donor livers was lower at various times after death: 15 minutes to 1 hour, 150 ± 24 × 106 cells; 2-4 hours, 114 ± 50 × 106 cells; and 6-24 hours, 56 ± 25 × 106 cells, P < 0.05, ANOVA with Dunn's test. Cell viability was also lower, particularly beyond 4 hours after death: 15 minutes to 1 hour, 56 ± 9%; 2-4 hours, 53 ± 6%; and 6-24 hours, 34 ± 11%; P < 0.05, ANOVA with Dunn's test (Fig. 3A). Only <20% hepatocytes from NHB donors isolated 15 minutes to 2 hours after death attached in dishes. These cells remained rounded subsequently. Hepatocytes from NHB donors 6 hours or longer after death did not attach or survive in dishes. However, the functional integrity was maintained in viable cells isolated 1 hour after death from NHB donors or from HB donor livers by assays of ureagenesis. Cells in both groups were metabolically active. NHB donor cells produced 280 ± 56 μg urea and HB donor cells 312 ± 82 μg urea per mL of medium per 1 × 106 attached cells, respectively, P = n.s. This indicated generally equivalent function in those cells that attached and survived after overnight culture. Although intact tissues had not shown significant apoptosis, we found onset of apoptosis with DNA laddering in hepatocytes from NHB donors, especially cells isolated 6 hours or longer after death (Fig. 3B). When hepatocytes from HB livers were transplanted into DPPIV− rats, cells engrafted in liver, as was expected (Fig. 3C). Cells from NHB donor livers 2, 4, or 6 hours after death, also engrafted. However, cells from NHB donor livers beyond 6 hours after death engrafted in liver extremely rarely. Therefore, we restricted analysis of cell engraftment to hepatocytes from NHB donors 2, 4, or 6 hours after death. In recipients of HB donor hepatocytes, we found 154 ± 37 transplanted cells per 100 periportal areas after 7 days (Fig. 3D). By contrast, cell transplantation from NHB donors 2, 4, or 6 hours after death produced 123 ± 20, 107 ± 10, and 29 ± 5 hepatocytes per 100 periportal areas, which was 20%, 30%, and 82% less, respectively (P < 0.05, ANOVA, Dunn's method).

Figure 3.

Viability of hepatocytes isolated from HB and NHB donor livers. (A) Trypan blue dye exclusion in hepatocytes from HB and NHB donors. (B) DNA ladder analysis. Lane 1, molecular weight marker of 100 bp; lane 2, HB hepatocytes; lanes 3 and 4, NHB hepatocytes 4 and 16 hours after death. (C) Cell engraftment after 7 days in DPPIV− rats with transplanted cells in periportal areas of liver. Representative findings are from recipients of HB or NHB hepatocytes isolated 2, 4, or 6 hours after death. Original magnification, ×400; methylgreen counterstain. (D) Morphometric analysis of cell engraftment. Asterisks in (A) and (D), P < 0.05 versus HB hepatocytes, ANOVA with Dunn's test (n = 3-5 donors each with three to four recipients of cells per donor).

We examined liver repopulation ability of NHB donor cells in retrorsine–PH-conditioned DPPIV− rats. Hepatocytes from both HB and NHB donors engrafted in the liver of preconditioned rats. HB donor cells formed readily visible clusters of proliferating transplanted cells (Fig. 4A). Fewer transplanted cell clusters were observed in recipients of NHB cells, including cells from donors 1, 4, or 6 hours after death (Fig. 4B). Morphometric analysis showed 25 ± 2 transplanted cell foci per low-power field (LPF) in recipients of HB donor cells (Fig. 4C). Fewer transplanted cell foci were found in recipients of NHB hepatocytes 4 hours after death, 7 ± 0.4 per LPF, which was 3.6-fold less than HB cells (P < 0.001, t test). The size of transplanted cell foci was 9.7 ± 0.4 × 10−3 mm2 in recipients of HB cells (Fig. 4D). By contrast, the size of transplanted cell foci was 3.5 ± 0.2 × 10−3 mm2 after transplantation of cells from NHB donors 4 hours after death, which was 2.8-fold smaller (P < 0.001, Mann-Whitney rank sum test). In view of significant differences in engraftment and proliferation of NHB donor hepatocytes, we studied molecular perturbations further.

Figure 4.

Proliferation of transplanted hepatocytes from HB and NHB donors. (A,B) DPPIV+ transplanted cells in liver from HB and NHB donors 4 hours after death. Original magnification x400; methylgreen counterstain. (C,D) Morphometric quantitation of transplanted cell foci (C), and area occupied by individual transplanted cell foci (n = 3-4 recipients from each of three donors). NHB hepatocytes produced fewer transplanted cell foci, and the area occupied by these transplanted cells was smaller. Asterisks, P < 0.05 versus HB donor hepatocytes, Mann-Whitney rank sum tests.

Gene Expression Changed Profoundly in NHB Hepatocytes.

First, we compared gene expression profiles in HB donor hepatocytes and hepatocytes isolated from NHB donors 4 hours and 16 hours after death (Fig. 5). This showed extensive perturbations in NHB donor cells with differential expression of 1000-2000 transcripts (Fig. 5A,B). More KEGG pathways were represented in up-regulated genes in NHB cells. Also, more KEGG pathways were represented with increasing time after death, i.e., 16 hours versus 4 hours (P < 0.05; Fig. 5C). Therefore, to further identify gene expression differences, we categorized functionally active pathways in up-regulated and down-regulated gene lists from NHB donor cells 4 hours and 16 hours after death (Fig. 5D).

Figure 5.

Gene expression profiles in hepatocytes from NHB versus HB donors. (A,B) Up- and down-regulated genes in NHB hepatocytes 4 hours and 16 hours after death compared with HB hepatocytes. Gene lists were larger in cells compared with tissues even 4 hours after death (see Fig. 2). This difference was more pronounced in cells NHB cells 16 hours after death. (C) KEGG pathways represented in differentially expressed genes from NHB hepatocytes. The numbers of these pathways and genes in individual pathways rose in NHB hepatocytes, as time-after-death increased. (D) Pathways with largest numbers of differentially expressed genes in NHB hepatocytes 4 hours or 16 hours after death.

The data established commonalities in gene expression pathways in NHB donor cells 4 hours and 16 hours after death (Table 2). Remarkably, apoptotic pathways were again not represented in NHB donor cells. In view of similarities in gene expression changes and because inferior cell engraftment and proliferation was obvious in cells from NHB donors 4 hours after death, we restricted analysis of molecular pathways to that time. We focused particularly on pathways in focal adhesion, cell adhesion, actin cytoskeleton, and extracellular matrix (ECM)–receptor interactions, because these should have affected cell attachment, as well as cell engraftment and proliferation. This identified a number of candidate pathway genes expressed differentially that could have profoundly affected these processes (Supporting Figs. 5-7).

Table 2. Representation of Major Functionally Annotated Groups in Differentially Expressed Gene Lists in NHB Donor Hepatocytes Versus HB Donor Hepatocytes
Gene Ontology GroupsCells from NHB donor liver 4 hoursCells from NHB donor liver 16 hours
No. of Up-Regulated Genes (%)No. of Down-Regulated Genes (%)No. of Up-Regulated Genes (%)No. of Down-Regulated Genes (%)
Cellular process984 (37)732 (43) 745 (47)
Primary metabolic process 480 (28)595 (22)511 (32)
Cellular metabolic process 485 (29)595 (22)508 (32)
Macromolecule metabolic process 410 (24)512 (19)450 (29)
Biological regulation455 (17)361 (21)458 (17)346 (22)
Regulation of biological process394 (15)322 (19)390 (14)308 (20)
Biopolymer metabolic process366 (14)320 (19)380 (14)332 (21)
Regulation of cellular process345 (13)287 (17)351 (13)276 (17)
Developmental process323 (12)252 (15)316 (12)236 (15)
Protein metabolic process   240 (15)
Cellular macromolecule metabolic process   227 (14)
Cellular protein metabolic process   221 (14)
Cellular localization314 (12)208 (12)326 (12)204 (13)
Cellular component organization and biogenesis279 (11)210 (12)281 (10)208 (13)
Establishment of localization271 (10)175 (10)294 (11)175 (11)
Transport263 (10)172 (10)282 (10)169 (11)
Nucleobase, nucleoside, nucleotide and nucleic acid metabolic process258 (10)243 (14)265 (10)251 (16)
Gene expression234 (9)204 (12)248 (9)214 (14)
Regulation of metabolic process211 (8)180 (11) 176 (11)
Multicellular organismal development240 (9)179 (11)229 (8)156 (10)
Establishment of localization271 (10)175 (10)294 (11)175 (11)
RNA metabolic process188 (7)172 (10)201 (7.4)182 (12)
Anatomical structure development232 (9)168 (10)238 (9)155 (10)
Regulation of cellular metabolic process193 (7)168 (10)198 (7)170 (11)
Regulation of gene expression186 (7)162 (10)178 (6.5)162 (10)
Transcription176 (6.6)160 (9.4)174 (6.4)158 (10)
 Intracellular signaling cascade142 (5)121 (7)161 (6)123 (8)
 Response to stress105 (4)102 (6)116 (4)104 (7)
 Cell adhesion103 (4) 84 (3) 
 Cell–cell signaling90 (3) 77 (3)56 (4)
 Potassium ion transport30 (1) 23 (1) 
 Induction of programmed cell death 31 (2)26 (1)27 (2)
 Wnt receptor signaling pathway  16 (0.6) 
 Calcium-mediated signaling 10 (0.6)10 (0.4)9 (0.6)

Multiple cell signaling pathways were perturbed. Sublists of genes regulating calcium signaling, along with mapping of those genes in calcium signaling pathways, indicated changes in multiple calcium-dependent intracellular processes (Supporting Fig. 8). Major changes were obvious in insulin signaling (Supporting Fig. 9), which was consistent with energy needs. We found that wnt signaling was perturbed (Supporting Fig. 10). Among the down-regulated genes were Jak-Stat signaling (Supporting Fig. 11), which is involved in cell injury. Surprisingly, genes involved in neuroactive ligand–receptor interactions constituted the single largest pathway in NHB donor cells (Supporting Table 1). The pathway representing ubiquitin-mediated proteolysis indicated further anoxia-dependent perturbations (Supporting Table 2).


This study advanced insights into the potential of NHB donors for cells. We found no hepatic autolysis, necrosis, or significant apoptosis in NHB livers over many hours, whereas molecular perturbations were due to hepatic reperfusion/reoxygenation. This identification of molecular changes in NHB livers offers opportunities for mechanistic interventions to preserve cell viability. Because cells isolated from NHB donors did possess appropriate metabolic activity, i.e., ureagenesis, and also because hepatocytes from NHB donors did engraft and proliferate to some extent, selected cell therapy applications should be feasible, e.g., when liver repopulation is unnecessary.16

Hyperkalemia caused rapid death and avoided confounding issues, e.g., with hepatic venous outflow obstruction following thoracotomy, phrenotomy, exsanguination, or cardiac clamping.17-19 This was confirmed by essentially no gene expression changes at early times. The consequences of anoxia were apparent later with gene expression changes, including mitrochondrial inner-membrane electron transport and coupled oxidative phosphorylation, although genes in complexes I (NADH-coenzyme Q reductase, Ndufs5), II (succinate ubiquinone oxidoreductase) and IV (cytochrome c oxidase) were expressed at higher levels 30 hours after death (Supporting Fig. 1). Such relative resistance to anoxia was a good portent for isolating cells from NHB livers. Organ- and species-specific mechanisms in anoxia tolerance should be relevant for maintenance of hepatic tissue integrity after death. For instance, in anoxia-resistant deep-sea turtles, NADH and cytochrome c oxidase transcripts accumulate rapidly and persist at high levels during anoxia.20 Some tissue changes identified by gene expression were typical of anoxia/reperfusion, e.g., gradients of chemotaxis, adherence, and transendothelial leukocyte migration, as observed in heart.21 Activation of adherence junction genes likely represented attempts to conserve tissue integrity. Many hepatic changes after death, i.e., expression of glycolytic genes, of lactate dehydrogenase A to generate pyruvate, and of genes in amino acid metabolism, likely reflected efforts at metabolic homeostasis with declining energy stores.

Previously, NHB donor livers were typically evaluated within 40-60 minutes after death.5 In one study, cell isolation 8-12 hours after anoxia produced viability <15%.6 By contrast, in a mouse study, viable cells were isolated even 24 hours after death, although the cell viability assay was liver repopulation in FAH−/− mice, in which even a few viable cells can proliferate extensively.7 We found that viable cells were best isolated from NHB livers up to 4 hours after death. Damage in isolated cells was clearly due to reperfusion/reoxygenation, implying roles for energy depletion and pro-oxidant stress. Among structural changes of greatest concern were focal adhesion, cell adhesion, actin cytoskeleton, extracellular matrix component (ECM)–receptor interactions, and adherence junctions. This was not previously known. However, gene expression profiling of postmortem muscle showed that ECM–receptor interactions or calcium signaling were deleterious, decreasing the water-holding capacity of myocytes.22 No doubt actin cytoskeleton weakening should have increased damage after enzymatic liver digestion, and impaired cell adhesion was reflected by inferior attachment in dishes of NHB hepatocytes. These mechanisms were relevant because focal adhesions and ECM–receptor interactions are required for engraftment of transplanted hepatocytes.23

Little is known about intracellular signaling after hepatic anoxia. Because multiple signaling pathways were altered in NHB donor hepatocytes, these processes were likely important in cell damage. For instance, mapping of focal adhesion pathways identified changes in extracellular signal–regulated kinase-1 (ERK1). ERK-1 was strongly induced by reactive oxygen species after tissue reoxygenation and contributed to disruption of cell adhesion, focal adhesion complexes, and cytoskeleton disorganization.24 After chemical hypoxia, prevention of energy depletion or acidosis, and incorporation of glycine, was hepatoprotective, albeit without restoring actin cytoskeleton or focal adhesions,25, 26 implying further complexities.

Intracellular calcium ion fluxes following anoxia have been of significance for hepatotoxicity or toxicity in other cell types, e.g., neurons.27, 28 Therefore, we were intrigued by activation of neuroactive ligand–receptor pathways in NHB hepatocytes, especially because engagement of excitatory receptors is harmful.29 Currently, no information is available about these potential anoxia-modulating mechanisms in the liver, which offer new directions for hepatoprotection. Perturbation in insulin signaling pathways was another aspect of NHB cells, most likely due to glycolysis in the setting of energy depletion.

In NHB donor hepatocytes, changes in wnt signaling, which included canonical wnt as well as wnt/calcium pathways, was unexpected, because these pathways have not been associated with anoxia. Noncanonical wnt/calcium signaling should be of interest because it may serve roles in remodeling of actin cytoskeleton,30 which was prominent in NHB cells. Moreover, wnt/calcium signaling is associated with cell survival.31 The significance of several down-regulated genes, e.g., the Jak-Stat signaling pathway, was unclear. Certainly, in view of the pleiotropic roles of wnt and Jak-Stat signaling in cell proliferation, injury, and other mechanisms, these genes may contribute to hepatocyte viability. The contributions of proteolytic pathways served by ubiquitination, which also stood out in our gene lists, have not been examined in hypoxic tissue injury.32

The integrity of NHB donor livers after death should be encouraging for isolating viable hepatocytes from humans, although in studies with cadaveric tissues from monkeys and resected specimens from humans, hepatocyte viability was limited.7 The molecular profile of hepatocytes from NHB donor livers offers frameworks to understand reasons for the inferior viability of isolated cells, while suggesting potential directions to reverse perturbations in a pathway-specific fashion. Besides mitigating cell damage due to oxidative stress, regulation of pathways in adherence junctions, cytoskeletal integrity, extracellular matrix interactions, cell signaling, etc., by various interventions, should be of particular interest. These studies will require detailed and careful prospective manipulations that could be guided by progressive reversal of molecular changes identified here in cells from NHB donors.


Ms. Chaoying Zhang provided technical assistance. The results have been published in abstract form.33