Functional Tissue Engineering of the Liver and Islets


  • Kazuo Ohashi,

    Corresponding author
    1. Institute of Advanced Biomedical Engineering and Science, Tokyo Women's Medical University, Shinjyuku-ku, Tokyo, Japan
    2. Department of Gastroenterological Surgery, Tokyo Women's Medical University, Shinjyuku-ku, Tokyo, Japan
    • Correspondence to: Kazuo Ohashi, MD, PhD, (Current address), Department of Surgery, Nara Medical University, 840 Shijo-cho, Kashihara Department of Surgery, Nara Medical University 840 Shijo-cho, Kashihara, Nara, 634-8522, Japan. Tel: +81 744-22-3051; Fax: +81 744-24-6866; E-mail:

    Search for more papers by this author
  • Teruo Okano

    1. Institute of Advanced Biomedical Engineering and Science, Tokyo Women's Medical University, Shinjyuku-ku, Tokyo, Japan
    Search for more papers by this author


Cell-based therapies by using hepatocytes and islets have recently been evaluated as a new therapeutic modality for patients with many forms of liver diseases and insulin-deficient diabetes mellitus. In most of the recently conducted clinical trials, cells have been delivered into liver vasculatures by infusing them through the portal circulation. More recently, tissue engineering–based approaches have spurred significant interests, using hepatocytes and islets in which small but functional new tissues would be created in vivo. Under circumstances in which a higher level of cell engraftment could be obtained, these approaches could provide therapeutic effects. Considerable efforts have been given to sustaining engineered tissues and maintaining their therapeutic effects. This review highlights several strategies that can achieve a higher level of cell survival for creating new functional liver and islet tissues. Anat Rec, 297:73–82. 2014. © 2013 Wiley Periodicals, Inc.


The liver has a complex architecture comprising parenchymal cells (hepatocytes) and several types of nonparenchymal cells. Hepatocytes are responsible for most of the synthetic and metabolic functions of the liver. Hepatocytes are arranged in two-dimensional panels (called hepatic plates) and have sinusoidal and canalicular surfaces (Grisham, 2009). The sinusoidal surfaces are lined by sinusoidal endothelial cells that have fenestrae measuring 100–200 nm in diameter (Braet et al., 2002). Narrow spaces (called space of Disse) formed between endothelial cells and hepatocytes allow the intensive interaction between the sinusoidal vascular channel and the surface of the hepatocytes. Canalicular surfaces are cell surfaces contacting neighboring hepatocytes. Bile canaliculi are formed in these surfaces, with functions of collecting and draining bile secreted from hepatocytes. On the basis of these characteristic architectural features of the liver, the alignment of hepatocytes and vascular endothelial cells would be important considerations in pursuing experiments in liver tissue engineering (Ohashi, 2008, Puppi et al., 2012).


In the naïve liver, hepatocytes and endothelial cells are in contact with each other, and this alignment enables cells to exchange nutrients and oxygen, and to crosstalk molecular signals. There are two distinct approaches for enhancing hepatocyte contact with vascular endothelial cells. One would be implanting hepatocytes in vivo near or within a vascularized area. The omental surface and the space under the kidney capsule are representative sites that are near hypervascularized tissues. If hepatocyte implantation will be targeted to a local area with poor vascularity (e.g., subcutaneous space), vascular network induction at the transplant site would not only enhance hepatocyte contact with the vascular endothelial cells but also provide a higher level of oxygen tension to the grafts. Local vascularization could be stimulated using devices that gradually release inducers of angiogenesis. Among the numerous potent angiogenesis inducers identified, fibroblast growth factor 1 (FGF1), FGF2, and vascular endothelial growth factor (VEGF) have succeeded in recruiting vascularized local areas in vivo that support higher hepatocyte survival (Ohashi et al., 2005a, 2007; Kedem et al., 2005; Lee et al., 2006; Yokoyama et al., 2006). Because hepatocytes have a high level of oxygen requirement for their sustained survival (Brown et al., 2007), these cells are susceptible to the process of cell death while they are exposed to low-oxygen atmospheres. In this context, the local vascularization process should be completed at the time of hepatocyte engrafting. In previous studies, a 7- to 14-day period will be required for the implanted growth factor–releasing device to establish an appropriately vascularized local area for hepatocytes (Ohashi et al., 2005, 2007; Kadem et al., 2005; Lee et al., 2006; Yokoyama et al., 2006). Studies have shown that hepatocyte survival at a poorly vascularized area was short term when growth factors are delivered at the same time as or after hepatocyte implantation (Smith et al., 2006; Chen et al., 2012).

Another approach that enables regulation of heterotypic hepatocyte–endothelial cell interaction would be the creation of a coculture system of hepatocytes and endothelial cells followed by their implantation. Several different culture configurations have been investigated, and it was largely demonstrated that a hepatocyte–endothelial cell coculture could improve hepatocyte-specific functions (Guguen-Guillouzo et al., 1983; Harimoto et al., 2002; Takayama et al., 2007; Yamada et al., 2012). Nahmias et al. (2006) reported that coculturing on a collagen gel surface with liver sinusoidal endothelial cells at a 1:1 ratio induced the emergence of a sinusoidal surface as well as the expression of low-density lipoprotein receptors. Salerno et al. (2011) have shown that hepatocytes cultured onto human umbilical endothelial cells, seeded on a porous membrane culture surface made of a polymeric blend of polyetherketone and polyurethane, exhibited polyhedral cells with well-demarcated cell-to-cell borders. Other heterotypic cell types that have been reported to support hepatocyte functions in vitro include hepatic stellate cells and bone marrow–derived mesenchyal stem cells (BMSCs) (Abu-Absi et al., 2004; Inamori et al., 2009; Gu et al., 2009). In addition, coculture with MSCs was found to minimize the cytotoxic effects induced by acute liver failure serum on hepatocytes (Shi et al., 2011), indicating the importance of heterotypic cell configuration for achieving higher clinical values in future applications.

Several technologies have enabled to create three-dimensional hepatic tissues in vitro, which are shown to be higher functionality compared with conventional culture condition (Bader et al., 1996; Kim et al., 2012; Nagamoto et al., 2012; Ota et al., 2011). By positioning endothelial and kupffer cells on top of extrinsic collagen gel coated on the hepatocyte layer, 3D culture model that reflect in vivo liver microstructure (Bader et al., 1996). Higher functionalities including high CYP1A1 activity and positive response to LPS stimulation were reported in this 3D culture system. More recently, new concept of 3D culture system was developed based on cell sheet stratifying procedures. The concept of cell sheet engineering was described in the following section. When monolithic layer of endothelial cells was stratified on top of primary cultured hepatocytes, precisely aligned layered hepatic construct could be generated, without using extrinsic extracellular matrix components (Kim et al., 2012). This layered tissue construct was found to secrete albumin at sustained level for over 30 days. In addition, this construct shows ability to uptake chemicals from culture medium and to excrete bile acid into bile canaliculi formed between hepatocytes (Kim et al., 2012).


A novel technology that allows harvesting cells in a monolithic layer format (cell sheet) has recently been used (Yang et al., 2007, 2009). This technology includes the use of temperature-responsive culture dishes that were covalently grafted with the temperature-responsive polymer poly(N-isopropylacrylamide) (PIPAAm) on their surfaces at nanometer thickness (Okano et al., 1993). This PIPAAm coating provides a slightly hydrophobic surface at the regular culture temperature (37°C), allowing conventional cell culturing. PIPAAm has a lower critical solution temperature of 32°C, and a temperature decrease to lower than 32°C results in a change from a hydrophobic to a more hydrophilic state, resulting in the natural detachment of cells from the culture surfaces. Since PIPAAm remained on the dish surfaces in the cell detachment process, the harvested cell sheets do not contain any polymer component.

Since the liver is organized into compact lamellar structures, called the liver plate, it would be ideal to establish procedures to recreate such hepatocyte structures in culture condition. Isolated hepatocytes have been shown to favorably attach and extend on PIPAAm-grafted surfaces. Plating on PIPAAm surfaces at a density of 8 × 105 to 10 × 105 cells/cm2 allows hepatocytes to reach a confluent status Lowering the culture temperature to 20°C for 15 min initiates the natural (i.e., nonchemical) detachment of hepatocytes from the culture surfaces, resulting in the harvesting of hepatocytes as a monolithic sheet (Fig. 1) (Ohashi et al., 2007). Since this temperature-dependent harvesting step does not damage intercellular connections, the hepatocyte sheets retain numerous intercellular microstructures, including gap junctions, desmosomes, and bile canaliculi, which are essential traits for demonstrating hepatocyte-specific functions at higher levels. Indeed, the hepatocyte sheets are found to express proteins and drug metabolism at higher levels than those of hepatocytes harvested from conventional culture dishes utilizing proteolytic enzymes (Ohashi et al., 2007). When the harvested hepatocyte sheets or hepatocytes are replated on a new culture dish, the hepatocyte sheets initiate the adhering process as early as 4 h, while it will take 12 h for the enzymatically harvested hepatocytes. Although indirectly, these findings indicate that adhesional proteins are well retained in the temperature-dependent harvesting process as a cell sheet format, which would be advantageous for engrafting in in vivo transplantation.

Using the hepatocyte sheets harvested from the PIPAAm dishes, our laboratory has developed a novel procedure for creating a miniature liver system in the subcutaneous space in mice (Ohashi et al., 2007). To enhance the attachment of hepatocytes and endothelial cells in vivo, a vascularized platform has been created by implanting a basic FGF (bFGF)-releasing device before cell sheet implantation. Upon the removal of the bFGF device, hepatocyte sheets are implanted into the previous location of the device, resulting in the formation of stratified miniature liver constructs made of endothelial cells and hepatocyte layers. The engineered miniature liver system was found to stably persist long term (for >200 days) with the ability to produce hepatocyte-specific proteins (Fig. 1). Notably, the subcutaneous liver possesses the functionality to take up, and then metabolize, circulating chemicals. Further, stacking the hepatocyte sheets onto each other up to four layers within the vascularized spot was found to successfully create a 3D liver system (Fig. 1). The 3D miniature liver system showed higher biological functionality as a part of the liver.


The liver is known to possess remarkable regenerative capacity under conditions of functional loss of liver volume (Michalopoulos, 2007; Fausto et al., 2009). The loss of functional liver volume will generate liver regenerative stimuli toward the compensatory regeneration mode. Experimentally, the compensatory regeneration mode in rodents can be induced by performing a surgical procedure in which a portion of the liver mass, normally two-thirds, is removed (2/3 partial hepatectomy [2/3PH]). In both rats and mice, the regenerative growth of the liver peaks at Days 1 and 2 after 2/3PH, followed by its completion and cessation by 10–14 days. Multiple factors are involved in the regeneration process in an orchestrated manner, including the association with signaling cascades involving growth factors, cytokines, and matrix remodeling (Michalopoulos, 2007; Fausto et al., 2009; Ding et al., 2010). This regenerative event is mediated through the proliferation of mature adult hepatocytes and other cell types, not by a selective subpopulation of stem cells (Michalopoulos, 2007).

A question that is raised here is whether a liver system engineered at ectopic sites could also perform regenerative growth in response to liver regenerative stimuli. Kaufmann et al. (1994) have created liver tissues in the mesenteric leaves by implantation of hepatocytes into prevascularized poly-l-lactic acid discs, and PH was performed at the same time as hepatocyte implantation. This experiment showed minimal effects of PH on the proliferation of ectopically engrafted hepatocytes. PH did not seem to induce proliferation of the ectopic hepatocytes. Since it is known that various growth factors essential for liver regeneration are rich in the portal circulation, studies have been extended to investigate if diverting portal flow into the general circulation [portacaval shunting (PCS)] could provide positive effects on the ectopic hepatocytes. A combination of PCS and PH allowed higher survival rates of hepatocytes with some increase in cell cycling (Sano et al., 1996). These evidences demonstrate the regenerative potential of the ectopically grafted hepatocytes; however, the level of proliferation of ectopic hepatocytes in comparison with hepatocytes in the naïve liver is yet to be established.

Recently, clear evidences for the regenerative growth potential of the engineered liver were unveiled in a mouse study in which persistently functional liver tissues were engineered at the kidney capsule space or subcutaneous space by using isolated and purified mature hepatocytes (Ohashi et al., 2005a,b, 2007). In this experiment, 2/3PH was conducted at 70 days after the engineering of liver tissues, without conducting PCS. The engineered liver tissues composed initially only of hepatocytes under the kidney capsule space were found to perform rapid regenerative growth at a similar speed as the naïve liver (Fig. 2). Identical values of bromodeoxyuridine (BrdU) labeling indices of hepatocytes were observed in the engineered liver tissues and naive livers during the regeneration process (Table 1). The timing of cessation of the regenerative events was also identical to the naïve livers (Ohashi et al., 2005b, 2007). Liver tissues engineered in the subcutaneous space by transplanting hepatocytes or hepatocyte sheets showed similar regeneration profiles (Fig. 2). Furthermore, an alternative liver regeneration mode (direct hyperplasia mode of liver regeneration) could also be induced to the ectopically engineered liver tissues in a chemically induced manner (Ohashi et al., 2005a). More importantly, the regenerated liver tissues stably persisted and surprisingly were able to re-regenerate under repetitive regenerative stimuli (Ohashi et al., 2005b). In a separate set of report, an ectopically engineered liver system was found to also be capable to act with an infused hepatocye growth factor (HGF) receptor agonist, followed by active regenerative proliferation (Ohashi et al., 2000, 2001).

Table 1. Regenerative growth activities of hepatocytes determined by BrdU labeling indices (LI) in the naïve liver and engineered liver tissues
Stimulation timingRegenerative stimulus and regeneration modeBrdU LI of hepatocytes in the naive liversBrdU LI of hepatocytes in the engineered livers
  1. Liver tissues were engineered under the kidney capsule space by using mature hepatocytes. Liver regenerative stimulus was induced by performing 2/3PH or administering primary mitogen 1,4-bis [2-(3,5-dichloropyridyloxy)] benzene (TCPOBOP) at Day 70 (Ohashi et al., 2005a). In some 2/3PH mice, subsequent regenerative stimulus was induced by performing 1/2PH at Day 130 (Ohashi et al., 2007). BrdU was delivered for 14 days after the induction of each regenerative stimulus.

  2. a

    P<0.01 versus control group.

First stimuli (Day 70)None (control)12.1 ± 2.512.7 ± 2.1
 2/3 PH (compensatory regeneration)95.5 ± 3.3a91.9 ± 3.3a
 TCPOBOP (direct hyperplasia)94.9 ± 2.6a88.5 ± 3.7a
Second stimuli (Day 130)None (control)10.8 ± 4.39.7 ± 3.1
 ½ PH (compensatory regeneration)82.5 ± 6.8a80.5 ± 6.1a

It would be important to discuss why PH-induced proliferative growth was obtained in this series of experiments and not in the previous studies described in the previous paragraph. A possible reason might be the difference in the timing of induction of regenerative stimuli after the tissue engineering procedure. The 2/3PH procedure was conducted at Day 0 in previous experimental series (Kaufmann et al., 1994; Sano et al., 1996) and at Day 70 in this experimental series (Ohashi et al., 2005a,b, 2007) after hepatocyte implantation. Since nonparenchymal cell components are known to be essential for progressing liver regeneration, it could be that the 70-day period allowed liver tissues (initially made only of hepatocytes) to recruit nonparenchymal cells to become hybrid liver tissues comprising both parenchymal and nonparenchymal cell types. Histological examination revealed that a significant vascularity was established in the engineered liver system during the 70-day period. The nonparenchymal composition as well as the vascular system are likely to be attributed to progressing the complex and orchestrated process of regeneration events in the engineered livers. Further enlargement of the engineered liver could theoretically be expected if continuous regenerative growth will occur owing to chronic stimuli generated by some liver disease states. Elucidation of the cellular mechanisms involved in ectopic liver regeneration is warranted for a deep understanding of the regulation of liver regeneration. These evidences taken together established that liver tissues, which could act as a part of the host naïve liver in terms of the active regeneration profile, could be engineered using isolated hepatocytes at ectopic sites that do not have access to the portal circulation.


Provided that the engineered liver tissues can be functionally maintained and can provide liver functions of a few percent of normal, many forms of liver diseases could be benefited. Representative disease include hemophilia A and B, bleeding disorders caused by a failure in the production of biologically active coagulation factors VIII or IX from the liver. It has been well established that achieving plasma clotting activity as low as 1% of normal markedly changes the phenotype of hemophilia patients from severe to moderate forms (Pipe et al., 2008). Engineering liver tissues under the kidney capsule using wild type of hepatocytes successfully provided therapeutic level of plasma clotting activity in mouse models of hemophilia A and B (Table 2) (Ohashi et al., 2005a, 2010). Increase in the plasma clotting activity was found to be proportional to the engineered tissue volume (Ohashi et al., 2005a). Researchers have also provided therapeutic evidences using other animal disease models including ascorbic-acid deficiency and fumarylacetoacetatehydrolase deficiency (Uyama et al., 2001; Hoppo et al., 2011).

Table 2. Shortened bleeding time of hemophilia A mice with engineered liver tissues determined by tail clipping test
GroupsBleeding time (min)
  1. Liver tissues were engineered under the kidney capsule space of hemophilia A mice, in either one or both kidneys, by using wild-type hepatocytes (Ohashi et al., 2005a). Tip of the mouse tail was clipped at Day 14 of the experiment. n = 3 in all the groups.

  2. a

    P < 0.01 versus sham group.

Naïve hemophilia A mice30.0 ± 0.0 (>30)
Sham operation30.0 ± 0.0 (>30)
Liver tissue in one kidney18.3 ± 2.9a
Liver tissues in two kidneys16.0 ± 2.9a
Wild-type mouse13.3 ± 2.1


These successful experimental developments in creating functional liver systems have a considerable impact on the establishment of tissue engineering approaches for diabetes mellitus (DM). Conclusive therapeutic evidences have been provided by clinical islet transplantation, in which isolated islets were infused into the portal vein toward the engraftments into the liver. Since it is known that the liver site is suboptimal, many investigators have pursued efforts for islet engraftment at alternative sites (Merani et al., 2008; Cantarelli et al., 2011). Tissue engineering strategies are poised to achieve higher engraftment, functionality, and longer persistency.


Pancreatic islets are highly organized miniature organs composed of β and α cells. The functions of islets include insulin and glucagon production and their subsequent storage in cytoplasmic granules, rapid sensing of blood glucose levels, and concomitant release of insulin or glucagon from each cell type. Christoffersson et al. (2010) have explored the potential of striated muscles as engraftment sites for islet transplantation, and found that this site could support vascularization to the islets and support islet functions. Neutrophil recruitment is one of the key factors that trigger revascularization to the transplanted islets. Since islets are cell clusters with a relatively large diameter, ranging from 50 to 400 µm, central necrosis and apoptosis likely take place during the hypoxic period between the transplantation and the completion of revascularization. To immediately establish a vascular network with the transplanted islets, several approaches based on two-stage procedures in which a vascularized area is created first, followed by implantation of islets or tissue-engineered islet tissues, have been reported. Some of the vascularization procedures that have successfully supported islet engraftments include the implantation of a bFGF-releasing device followed by its removal (Kawakami et al., 2001), insertion of a silicone tube chamber incorporated with bFGF (Hussey et al., 2009), creation of a prevascularized collagen gel sandwich device in vitro followed by its implantation (Hiscox et al., 2008), and preparation of the islet-engrafted kidney by islet implantation under the kidney capsule space followed by transplantation of the composite kidney with islets (Yamada et al., 2011).

Gene therapy–based approaches have also been empirically investigated and shown to be a promising intervention for inducing a vascular system around the islets. Adenoviral vector–mediated transduction of HGF or VEGF genes to the islets was found to successfully induce islet revascularization, resulting in a high engraftment rate under the kidney capsule. The gene transduction procedure itself is reported not to be associated with deterioration of insulin-secretion functions (Cheng et al., 2004; Fiaschi-Taesch et al., 2008). VEGF-A seems to be especially a key factor since it plays an important role in regulating islet revascularization as well as regulating the interaction between β cells and endothelial cells. The gene therapy approach has also been shown to be promising in preventing the cell death process of islets. β cells are known to be susceptible to apoptotic cell death in many situations, including cell isolation, purification, hypoxic stress, and oxidative reaction during the revascularization processes. Hui et al. (2005) have shown that adenoviral vector–mediated X-linked inhibitor protein (XIAP) gene transduction could significantly reduce the immunosuppressive drug–related islet apoptosis that occurs in vitro. Increased cellular levels of phosphorylated Camp-responsive element binding protein and reduced levels of Smac are key elements in the XIAP-induced antiapoptotic effects. More recently, dual transduction of HGF and XIAP genes successfully showed synergistic effects in preventing loss of human islets from apoptosis and improving revascularization after implantation under the kidney capsules in SCID mice (Wu et al., 2011). The transduction efficiency to the islet has been improved by using fiber-knob modified adenoviral vectors (Mukai et al., 2007; Wu et al., 2011). Cells in the mantle area of islets are highly transducible, while those in the core are difficult to transduce. Vector infection procedures that allow uniform transduction to cells throughout the islets are still awaited. Since adenoviral vectors do not generally integrate their genomes into host cell chromosomes, the transgene expression period is short term. In this context, conventional adenoviral vectors would be suited for the purpose of transient modification, such as initiating vascularization or antiapoptotic effects at the early phase after transplantation. In case long-term effects are desired, the use of high-capacity adenoviral vectors that lack all viral coding sequences or vectors incorporated with an integrase system (e.g., DNA transposon bacteriophage integrase) is recommended (Olivares et al., 2002; Rauschhuber et al., 2012).

Addition of heterotypic cells has been investigated for enhancing revascularization of islets. Endothelial progenitor cells (EPCs) are circulating progenitor cells known to enhance neovascularization in various pathologic conditions. A recent study by Kang et al. (2012) investigated islets mixed with cord blood–derived EPCs that were transplanted under the kidney capsule space, and found that this cotransplantation of islets and EPCs resulted in an acceleration of local neovascularization with higher islet engraftments. Although cotransplantation with ECs seems to be a promising approach, the same caveats remained, including the facts that the vascular system formed by the ECs tended to be leaky and the completion of vascular formation took longer than expected (Gupta et al., 2011). The former is speculated to be due to immature vessel architecture caused by poor alignment of pericytes. Successful further maturation of EC-derived vasculatures was reported by Chamberline et al. (2012) in which BMSCs were added to microtissues made of ECs. This vascular maturation was attributed by multifactorial effects provided by the BMSCs, including promotion of EC sprouting, EC proliferation, decreasing inflammatory responses, and harboring pericyte-specific phenotypes followed by their migration to surrounding areas of the vasculature.


An alternative approach with potential for achieving higher engraftment levels of pancreatic islets is the use of dispersed single islet cells that could be obtained by enzymatic dispersion of islets. Dispersed islets could be reformed either into cellular aggregates (pseudoislets) (Tsang et al., 2007), bioartificial islets composed of polymer scaffolds (Kodama et al., 2009), or a cell sheet formation (Shimizu et al., 2009; Saito et al., 2011). By regulating cellular alignment within the constructs, a uniform supply of oxygen and nutrients to individual cells could be provided.

One of our approaches for the creation of functional islet tissues is based on the bioengineering of contiguous monolayer islet cell sheets followed by their implantation (Shimizu et al., 2009; Saito et al., 2011; Ohashi et al., 2012). Single cell monolayer constructs are thin tissues (10- to 20-µm thickness) that are able to receive uniform oxygen tension when transplanted. An important feature of this approach is the preparation of culture surfaces specifically defined for islet cell culturing; that is, laminin-332 is coated onto a tissue culture plastic grafted with an optimized amount of a temperature-responsive polymer (PIPAAm). This laminin-332-coated PIPAAm surface is found to support dispersed islet cell attachment as a monolayer and the subsequent harvesting in a cell sheet format through a temperature-dependent cell detaching process. We also found that this culture surface allowed dispersed islet cells to initiate intercellular connections (e.g., tight and desmosome junctions), which will be retained in the subsequently harvested islet cell sheets. These islet cell sheets are durable for transplantation procedures. In a mouse model study, transplantation of two islet cell sheets into a subcutaneous space, using a patch style, created functional neoislet tissues. Numerous vascular networks are formed within and surrounding the tissues, allowing for rapid glucose sensing and insulin release by the neoislets, resulting in a successful reversion of the diabetic status into the normal condition (Saito et al., 2011). The persistence of the engineered neoislet tissues and their therapeutic values were confirmed in studies followed for >100 days.

The cellular composition and structure of the neoislet tissues engineered in the subcutaneous space are important issues to address. The microstructural unit of naïve islets is known to be a trilaminar plate in which endocrine cells are strategically aligned. Within the plate, one to two layers of β cells are sandwiched between layers of α cells (Bosco et al., 2010). This trilaminar plate had folded to form the cluster of islets in vivo. The subcutaneous neoislet tissues made by stratifying islet cell sheets showed a flat panel structure with three- to four-endocrine-cell thickness, which is nearly identical to the trilaminar plate microstructure of naïve islets (Fig. 3). More importantly, β and α cells were localized at the outer and inner regions within the neoislet tissues, respectively (Fig. 3). These evidences highlight the importance of the dispersed islet cell–based tissue engineering approach by creating new and functional neoislet tissues in vivo.

Figure 1.

Hepatocyte sheet–based liver tissue engineering in the subcutaneous space. A: Fabrication of hepatocyte sheets by using culture dishes grafted with a thermoresponsive polymer. Isolated hepatocytes were plated and cultured for 3 days. By lowering the culture temperature to 20°C for 15 min, the cultured hepatocytes spontaneously detached from the culture dishes. Gentle pipetting of the culture medium allowed for the complete separation of the hepatic sheet from the culture surfaces, allowing the harvesting of cells as a continuous, monolayer sheet. B: Maintenance of the functional volume of the engineered liver tissues in the subcutaneous space created by transplanting one layer (triangle) or two layers (circle) of hepatocyte sheets. Statistical intergroup differences were observed at all time points. (C–E) Histological images of the engineered liver tissues. C: Human α-1 antitrypsin (hAAT) immunostaining of the engineered liver tissue created by transplanting one layer of hepatocyte sheet. D,E: H&E staining of engineered liver tissues created by transplanting two layers (D) or four layers (E) of hepatocyte sheets in a stratified fashion. Scale bars, 50 μm. Reproduced, partly, from Ohashi et al. (2007), with permission from Nature Publishing Group.

Figure 2.

Regenerative growth of engineered liver tissues. Liver tissues were engineered under the kidney capsule space (A, B) or into the subcutaneous space (C, D). At Day 70, liver regenerative stimulus was induced in the recipients by performing 2/3PH followed by BrdU delivery for 14 days. α-1 Antitrypsin (hAAT, a maker for hepatocytes; in green) and BrdU (a marker for proliferating cells; in red) immunofluorescent staining. Numerous BrdU-positive hepatocytes were observed in the PH group, indicating that regenerative proliferation was actively induced in the ectopically engineered liver tissues. Arrows denote the kidney capsule. Scale bars, 100 μm. Reproduced, partly, from Ohashi et al. (2005a), with permission from John Wiley & Sons Inc.

Figure 3.

Histological and immunohistochemical figures of the neo-islet tissues engineered in the subcutaneous space at Day 60. (A) H&E staining, (B) immunofluorescence staining for insulin (red) and PECAM-1 (green), and immunohistochemical staining for insulin (C) and glucagon (D). Note that intense vascular networks were recognized within and around the neo-islet tissues (B, arrowheads) and that insulin-secreting and glucagon-secreting cells localized in the core and outer layer of the engineered neo-islet tissues (C and D, arrows). Scale bars, 50 μm (A and B) and 100 μm (C and D). Reproduced, partly, from Saito et al. (2011), with permission from Lippincott Williams & Wilkins.

Since cell sheet engineering technology enables to create laminarly assembled tissue constructs, complex tissues composed of layers with islet cells and other type of cells could be created as described in the “Cell sheet–based liver tissue engineering approach” section. Testicular Sertoli cells and MSCs are able to confer local immunoprotection to the islet cells. Three-dimensional islet tissue constructs with stratified layers of immunoprotective cells and islet cells will be determined in the allogenic or autoimmune DM models.

Another potential advantage of the use of dispersed islet cells is that cells can be cryopreserved in the long term with minimal loss of cellular viability. In contrast, it has been shown that islets are susceptible to cellular damages during the cryopreservation and thawing process. Our group has established the efficacy of UW solution with 10% DMSO as a freezing medium for dispersed islet cells (Ohashi et al., 2011). Cryopreserved and thawed dispersed islets were found to show minimal loss (<10%) of viability compared with those before the cryopreservation. The biological availability of these cells was demonstrated by their insulin secretion values identical to those of fresh dispersed islet cells. Indeed, islet cell sheet formation can be induced using the cryopreserved dispersed islet cells, representing a valuable cell source for islet tissue engineering.


Encapsulation technology has been extensively studied for preventing islets from immunological attacks generated by autoimmune responses to insulin-secreting cells and alloimmune responses to allogenic donor islet cells. Toward this goal, several immunoisolation approaches have been developed, including macroencapsulation, microencapsulation, and nanoencapsulation. If the use of immunosuppressive agents could be successfully minimized, islets could also be protected from damages associated with immunosuppressive agents.

Microencapsulation of islets is performed by mixing islets with polymers that form hydrogels. Polymers, including alginate, agarose, chitosan, methacrylic acid, methyl methacrylate, and polyethylene glycol (PEG), have been used for encapsulation (Vaithilingam et al., 2011). Alginate-based encapsulation was first reported by Lim et al. (1980) who were able to achieve normoglycemia by using allogenic encapsulated islets. In most experimental conditions, microencapsulation has been reported to be effective in achieving long-term survival by avoiding immunological responses. Clinical evidence has also been provided for allogenic islet transplantation (Tuch et al., 2009). Islets isolated from cadaver pancreata were encapsulated with barium alginate with a mean diameter of 340 µm, and were transplanted into the abdominal cavity of patients with type I DM. Without any immunosuppression, patients who received islet transplantation for more than two times showed detectable C-peptide levels for 6 weeks, although the encapsulated islets became necrotic and were surrounded by fibrous tissues at the 16-month observation.

Previous experimental and clinical investigations highlighted several drawbacks, including induction of inflammatory reaction that leads to fibrotic tissue capsulation, possible free shedding of cytokines and antibodies, and a large void space remaining between the islet and capsule surface (Giraldo et al., 2010). The first two issues, related with immunological and inflammatory responses, could be overcome by combining islet encapsulation and immunosuppressive agents. Significantly prolonged xeno-islets were obtained when microencapsulated islet transplantation was combined with blockade of different pathways of T-cell costimulation pathways by CTLA4-Ig and anti-CD154mAb (Safley et al., 2005). The issue of the large void space could be overcome by developing methods that allow islets to be conformally coated with polymeric layers with thickness in the range of 10–100 µm. Another technical refinement could be nanoscale encapsulation.

Recently, coating of the islet surface with PEG chains have been investigated by several groups, because PEG coating of cellular surfaces could provide a shield that will allow reduced antigenicity and higher biocompatibility, as well as confer cytoprotective effects. Various approaches for the PEGylation of the islet surface have been reported, including linking the amide group of the islet surface with the PEG-conjugated amine group of the collagen matrix (Jang et al., 2004; Lee et al., 2002), or immobilizing amphiphilic PEG-conjugated phospholipids on the islet cellular membrane (Teramura et al., 2007). These technologies enable PEG coating with a thickness in the nanometer range. Further PEG surface modifications have been conducted to confer biological functionality. One example is that coating the islet surface with PEG-urokinase successfully attenuated the inflammatory reactions between islets and whole blood (Teramura et al., 2011). It is of note that a PEG-lipid–mediated coating approach can be applied to single dispersed hepatocytes or hepatocyte sheets without causing any toxicity or deterioration of cellular functions (Tatsumi et al., 2012). To strengthen the mechanical properties of the coating, poly(vinyl alcohol) bearing thiol groups was anchored to maleimide-PEG coated on the islet surfaces (Teramura et al., 2007). At the time of this review, nanoscale encapsulated islets have yet to be used in the clinical setting. Although the long-term durability of coating materials and their biocompatibility need to be established, the nanoscale islet encapsulation approach has a high potential in modulating cellular survival and functionality in vivo.


Intensive studies have allowed the successful development of technologies for enhancing cellular survival and prolonging the functionality of engineered tissues from liver and islet engineering approaches. Some of the developments are reviewed herein include material bioengineering, cell sheet engineering, cell surface modification, and gene therapy application. As the tissue engineering of liver/islets becomes a reality through the successful demonstrations in experimental animal models, other challenges need to be further studied, including source of cells, cell cryopreservation, and appropriate regimen of immunosuppression and/or anti-inflammation. Even with these issues, the field of tissue engineering opens a new window for creating new therapeutic options for patients with various states of liver disease and DM.


The authors thank Drs. Hirofumi Shimizu, Takafumi Saito, Kazuya Ise, and Mitsukazu Gotoh (Fukushima Medical University) for islet tissue engineering projects, Drs. Takashi Yokoyama, Hiroyuki Kuge, Yoshiyuki Nakajima, Hiroyuki Naka, and Akira Yoshioka (Nara Medical University) for hemophilia mouse projects, Dr. Hiroo Iwata (Kyoto University) for vascularized platform projects, Mark A Kay (Stanford University) for tissue regeneration projects, and Drs. Kohei Tatsumi, Rie Utoh, Natsumi Watanabe, Inkyong Shim, and Ms. Kyungsook Kim (Tokyo Women's Medical University) for liver tissue engineering projects.