Liver Stem Cells and Hepatocellular Carcinoma

Abstract

Although the existence of cancer stem cells (CSCs) was first proposed over 40 years ago, only in the past decade have these cells been identified in hematological malignancies, and more recently in solid tumors that include liver, breast, prostate, brain, and colon. Constant proliferation of stem cells is a vital component in liver tissues. In these renewing tissues, mutations will most likely result in expansion of the altered stem cells, perpetuating and increasing the chances of additional mutations and tumor progression. However, many details about hepatocellular cancer stem cells that are important for early detection remain poorly understood, including the precise cell(s) of origin, molecular genetics, and the mechanisms responsible for the highly aggressive clinical picture of hepatocellular carcinoma (HCC). Exploration of the difference between CSCs from normal stem cells is crucial not only for the understanding of tumor biology but also for the development of specific therapies that effectively target these cells in patients. These ideas have drawn attention to control of stem cell proliferation by the transforming growth factor beta (TGF-β), Notch, Wnt, and Hedgehog pathways. Recent evidence also suggests a key role for the TGF-β signaling pathway in both hepatocellular cancer suppression and endoderm formation, suggesting a dual role for this pathway in tumor suppression as well as progression of differentiation from a stem or progenitor stage. This review provides a rationale for detecting and analyzing tumor stem cells as one of the most effective ways to treat cancers such as HCC.

One of the remarkable features of liver stem cells is that despite their large number and rapid rate of cell division during regeneration, these cells rarely acquire the age-related genetic defects associated with cancer induction or show deterioration in functional competence. This observation suggests that hepatic stem cells have evolved protective mechanisms against genetic damage. Among them is the ability of stem cells to selectively sort the old (parental) and new DNA strands when they divide, retaining only parental DNA strands. This ensures that replication-induced errors are excluded from stem cells.^1,2 Furthermore, random errors introduced into parental strands (for example, after exposure to genotoxic agents) induce protective responses such as p53-dependent stem cell apoptosis and G1 arrest by transforming growth factor (TGF-β).³ Stem cells are generally characterized by their capacity for self-renewal through asymmetrical cell division, multipotency for producing progeny in at least two lineages, long-term tissue reconstitution, and serial transplantability.⁴ For tumors containing a subpopulation of cancer stem cells (CSCs), there are at least two proposed mechanisms of CSC origin; oncogenic mutations may inactivate the constraints on normal stem cell expansion, or alternatively, oncogenic mutations in a more differentiated cell generate continual proliferation of cells that no longer enter a postmitotic differentiated state, thereby creating a pool of self-renewing cells in which further mutations can accumulate.⁵ Potentially, biologically significant pathways that modulate these stem/progenitor cells in cancer tissue could be identified through dual roles in embryonic stem cell development and tumor activation or suppression.⁶

The plasticity of such cells is reflected by recent studies in which pluripotent stem cells from embryonic or adult fibroblasts can be induced by introducing four factors, Octamer 3/4 (Oct3/4), sex determining region Y-box 2, c-Myc, and Kruppel-like factor 4 (Klf4) under embryonic stem cell culture conditions.⁷ Conceivably, alteration of stem cell/cell cycle activators such as Wnt/β-catenin, Hedgehog, Notch, and TGF-β signaling systems or targets such as β-catenin, human telomerase reverse transcriptase, cyclin-dependent kinases, Myb (myeloblastosis gene family), or c-MYC could expedite tumorigenesis or conversely, by specific targeting serve as a powerful cancer-prevention tool. In fact, we have previously demonstrated that disruption of TGF-β signaling by disruption of embryonic liver fodrin (ELF) is critical to the development of hepatocellular cancer (HCC).⁸ Furthermore, immunohistochemical and confocal analysis of HCC showed cells that label with stem cell markers and have unexpectedly lost transforming growth factor beta type II receptor (TBRII) and ELF. Expression analysis of these tumors showed marked activation of the interleukin-6 (IL-6) pathway, suggesting that HCC may develop from IL-6–driven CSCs with disrupted TGF-β signaling. Suppression of IL-6 signaling through the generation of mouse knockouts with a positive IL-6 regulator, interalpha-trypsin inhibitor heavy chain 4, resulted in a reduction of HCC in elf^+/− mice, thereby demonstrating a link between two major signaling pathways in the development of cancer and suggesting a possible therapeutic target.⁹ However, the specific stages that tissue-specific stem cells enter to develop into differentiated cells such as hepatocytes and cholangiocytes and the mechanisms of asymmetrical cell division still needs to be addressed. The details of the hierarchical biologically functional systems necessary for generating liver stem cell asymmetry will likely come from genetics and in vivo loss of function studies.

Embryonic Stem Cells, Hepatic Stem Cells, and Liver Development Are Linked by Common Pathways

During mammalian embryonic development, pluripotent embryonic stem cells originate from the blastocyst inner cell mass and give rise to somatic stem cells that further differentiate into multipotent tissue-specific stem or progenitor cells. In liver development, the liver bud is seen around embryonic day 8.5, because tissue-specific foregut endodermal stem/progenitor cells begin to proliferate and differentiate under the influence of signals from the septum transversum mesenchyme.^10,11 Multiple signaling pathways direct the early development of the liver bud from endoderm (Fig. 1, Fig. 3). Mouse and zebrafish studies indicate that the coordinated signaling of fibroblast growth factors from cardiac mesoderm (fibroblast growth factors 1, 2, and 8) and bone morphogenetic proteins (BMP) from the septum transversum mesenchyme (BMP 2, 4, 5, and 7) appear to cooperatively induce hepatoblast development from endoderm via mitogen-activated protein kinase (extracellular signal-regulated kinase 1 and 2).^12–14

Fig. 1.

TGF-β signaling via Smad proteins in liver development. TGF-β signaling in hepatocyte precursors directs development and differentiation. Liver progenitor cells (LPCs) and rodent oval cells express phenotypical markers of both (immature) hepatocytes (alpha-fetoprotein and albumin) and bile duct cells (such as biliary-type CKs, oval cell marker 6, CK7, CK14, and CK19) as well as hematopoietic markers (such as the receptor for stem cell factor, c-kit and CD34).

Fig. 3.

Interplay of TGF-β, Wnt/β-catenin, Hedgehog, and Notch in stem cell differentiation and tumorigenesis. Schematic diagram of TGF-β signaling pathway. Note interplay between Wnt/β-catenin, Hedgehog, Notch, and TGF-β signaling pathway targets and expedited tumorigenesis.

In mammals, after liver bud development, Wnt signaling promotes liver growth.^15–17 This is in contrast to hepatic induction, when Wnt signaling is initially repressed to counteract its inhibitory actions through hematopoietically expressed homeobox repression.¹⁸ The liver bud then gives rise to cells destined to become bipotential liver progenitor cells (LPCs).^10,19 LPCs initially express alphafetoprotein and albumin and later express cytokeratins (CKs) −7 and −19. Just before embryonic day 16, LPCs diverge along two cell lineages: hepatocytes (alpha-feto-protein⁺/albumin⁺) and cholangiocytes (CK19⁺).¹² Notch signaling via Jagged1 and Notch2 promotes differentiation of LPCs toward a biliary epithelial phenotype, whereas hepatocyte growth factor antagonizes biliary differentiation and promotes hepatocytic differentiation.²⁰ Modulation of biliary and hepatocyte morphogenesis during development is also prominently controlled by TGF-β signaling via Smad proteins. Heterozygous inactivation of both Smad2 and Smad3, which are receptor-Smads key in nuclear translocation of the TGF-β signal, in mice shows fetal liver hypoplasia. Interestingly, addition of hepatocyte growth factor to Smad2^+/− and Smad3^+/− mutant liver explants restores their growth in vitro, correcting integrin-β, and indicating that hepatocyte growth factor and TGF-β provide parallel growth signals in the fetal liver (Fig. 1).²¹ In adult human tissues, immature epithelial cells have been found residing in the smallest terminal branches of the biliary tree known as the Canals of Hering.²² These cells have been described as “intermediate hepatobiliary cells” and “hepatic progenitor cells” or the human equivalent of rodent “oval cells.”²³ Within this niche, hepatic progenitor cells are in direct physical continuity with hepatocytes at one membrane boundary and bile duct cells at another boundary and are considered to represent hepatic stem cells.

Identification of Long-Term Label Retaining/Progenitor Cells Through Human Liver Regeneration

In the adult liver, mature hepatocytes seldom proliferate and have a life span of over a year. After partial hepatectomy, however, proliferation of the normally quiescent hepatocytes and cholangiocytes, followed by proliferation of the hepatic stellate cells and endothelial cells, quickly restores the liver to its original mass. In the rodent model, DNA synthesis starts 12 to 16 hours after the standard partial hepatectomy and peaks at 24 to 48 hours. The original organ mass is almost restored 3 days postresection.^24,25 Serial transplantation experiments have shown that hepatocytes have a near infinite capacity to proliferate.^26–28 When mature hepatocytes and cholangiocytes are damaged or inhibited in their replication, however, a reserve compartment of hepatic progenitor cells is activated.²⁹ The activation of the stem cell compartment, referred to as a “ductular reaction” in humans and “oval cell reaction” in rodents, is observed in circumstances of prolonged necrosis, cirrhosis, and chronic inflammatory liver diseases. This process involves expansion of bipotential transit amplifying progenitor cells, which can differentiate into hepatocytes and biliary cells. Intermediate hepatocytes, with an intermediate phenotype between progenitor cells and mature hepatocytes, are seen in moderate to severe inflammatory hepatitis.²⁹ Moreover, the degree of stem cell and intermediate hepatocyte activation correlates with the degree of inflammation and fibrosis in diseases such as chronic hepatitis, hemochromatosis, and nonalcoholic steatohepatitis.^30,31

Living donor liver transplantation offers a unique window of opportunity to study liver regeneration instead of the standard partial hepatectomy performed on a diseased liver. In living donor liver transplantation, transplanted liver is subjected not only to the significant loss of tissue mass but also to hepatocyte and bile ductule injury during cold and warm ischemic time. In addition, the recipient of a live donor liver transplant routinely undergoes serial tissue biopsy to assess graft function and therefore, the availability of regenerative tissues at different time points enables us to identify cells that possess “long-term label retention” among a pool of proliferating cells.⁹ Most importantly, hepatocyte proliferation and liver regeneration is for the most part complete at 4 months. However, at this point, it is possible that the liver tissue still has an expanded population of “long-term label retaining” or stem cells that are difficult to detect in normal adult human liver. Understanding and identification of the bipotential progenitor cells may hold promises for new therapeutics in a wide range of liver disorders that include congenital metabolic diseases, end-stage liver cirrhosis, and hepatocarcinogenesis.

To search for hepatic stem cells, we previously studied five patients with monthly posttransplantation liver biopsies. The surgical procedure involves resection and transplantation of a lobe representing 55% to 60% liver mass from a donor to a recipient, which by 3 months grows to 85% of original mass.³² We hypothesized that at the end of liver regeneration, there would be an expanded population of liver progenitor/stem cells that were “long-term label retaining” and label for proteins that serve as markers for stem cell renewal as well as for liver differentiation.³³ We ultimately directed our search for cells expressing five proteins: Oct4, Nanog (Nanog homeobox), signal transducer and activator of transcription protein 3 (STAT3), TBRII, and ELF.

Both Oct4 and Nanog are expressed in embryonic and pluripotent stem cells, whereas STAT4 appears to be essential for embryonic visceral endoderm development as well as for self-renewal of pluripotent embryonic stem cells.^34–37 Both TBRII and ELF have been implicated in both early embryonic development of the foregut as well as in endodermal malignancies.^6,8 Examination of serial liver sections by immunohistochemistry revealed a cluster of two to four cells out of the entire 30,000 to 50,000 cell population of living donor liver transplanted specimens that expressed STAT3, Oct4, and Nanog as well as TGF-β signaling proteins TBRII and ELF. These cells also stained positively for both a hepatocytic cell lineage marker, albumin, and a cholangiocytic cell lineage marker, cytokeratin-19 (CK19), along with phosphorylated histone H3, a marker for active proliferation (Table 1).³³ These putative progenitor/stem cells were generally found localized in the portal tract region surrounded by a “shell” of six to seven cells expressing TBRII, ELF, and albumin, but not Nanog or Oct4, reflecting a more differentiated phenotype. These findings together with the known role of the TGF-β signaling pathway in liver development led us to hypothesize that: (1) these STAT3 ⁺ , Oct4 ⁺ , Nanog⁺, TBRII ⁺ , and ELF⁺ cells represent the progenitor/stem cell pool, which becomes activated during the regenerative process, and (2) TBRII and ELF may be involved in the initiation of hepatocyte differentiation of STAT3⁺/Oct4⁺ progenitor/stem cells. Thus, this study demonstrates a hepatic progenitor/stem cell in postembryonic human liver.

Table 1.

Phenotypic Expression of Human LPCs, Cell Lineages, HCC, and Putative Liver Stem Cells

Protein	Liver Progenitor Cells (LPCs	Cholangiocytic Lineage	Hepatocytic Lineage	HCC	Putative Liver Cancer Stem Cell	References
EpCAM (TACSTD1)	+	+	+	Diff.	+	42, 43, 44+
BMI1	+/−	ND	ND	Diff.	+ +	45, 43, 46, 47 +
NANOG	+	+	+	Diff.	+ +	43, 9#,+,%
Oct3/Oct4 (POU5F1)	+	+	+	Diff.	+ +	45, 9#, + ,% 43, 46
STAT3	+	+	+	Diff.	+	45+, 9#, + ,%
NOTCH1	+	ND	ND	Diff.	+/−	45 + , 43, 46
SMAD4	+	ND	+	+	ND	48#,+
TGFBR2	+	+	+	−	−	9#, + ,% 48#,+
ELF (SPTBN1)	+	+	+	−	−	9#, + ,% 48#,+
ABCG2	+	+	+	Diff.	+	43, 49
KIT	+/−	ND	ND	Diff.	+/−	43, 50 *, 51, 52 + , 53 +
WNT3a	+	ND	ND	Diff.	+	45
β-catenin (CTNNB1)	+	+	+	Diff.	+	54, 46
CD133 (PROM1)	+/−	ND	−	Diff.	+	46, 55, 56, 50 *
CD90 (THY1)	+	ND	ND	Diff.	+	45 *,#, +, 50*
CK18 (KRT18)	+	−	+	+	+	44+, 46, 50 *
CK8 (KRT8)	+	+	+	ND	+	50 *, 30 +
CK19 (KRT19)	diff.	+	−	Diff.	+/−	41, 50 *, 51, 57 + , 58 +
CK7 (KRT7)	+	+	+	ND	+	59 + , 57 +
AFP	+/−	−	+	Diff.	+	45, 43, 42 + , 44+, 46, 41, 50 *, 60
ALB	+	−	+	+	+	45, 43, 44+, 46, 50*, 59 +
CD44	+	ND	ND	Diff.	+	45, 46, 50 *
OV6	+	+/−	+/−	Diff.	+	54, 57 + , 61 +
CK14 (KRT14)	+	ND	ND	ND	+/−	52 + , 62 +
Ki-67 (MKI67)	ND	ND	ND	+.	+	54
SMO	ND	ND	ND	Diff.	+	46
GCTM-5	+	low	−	−	ND	63 +
MET	+/−	−	−	Diff.	ND	64 *, 65 +
HepPar1	+	−	+	ND	ND	62 + , 58 +
NCAM1	−	+	−	ND	ND	50, 66+, 57 +
CD34	+	ND	ND	Low	+/−	46, 50 *, 53, 67 +
CD45	−	ND	ND	ND	−	45, 50

Cholangiocytic and hepatocytic lineage were defined based on morphology or expression of CK19 (cholangiocytic marker) or albumin and alpha-fetoprotein (hepatocytic marker). The table key is as follows: ND, not determined/reported; +, positive expression; −, negative; ++, high/over expression; low, low expression; diff., positive differential (low or high) expression; +/−, contradictory supporting evidence. All proteins are listed as common aliases of the official human gene symbols from the HUGO Gene Nomenclature Committee (http://www.genenames.org/), which are provided in parenthesis. All sources are listed from most recent to past for reference. Strict immunohistochemical and confocal labeling was adhered to in the following studies denoted by #. Studies demonstrating serial transplantablity of cells ndiating potential “stem-ness” are denoted with *. Studies demonstrating in vivo expression of markers is denoted by +. Studies in which “stem-ness” was studied by mechanistic genetic studies are denoted by %.

Human Liver Progenitor Cells and Cancer Stem Cells

The existence of a hepatic stem cell compartment gives rise to expectations regarding the practical applications of such research. Understanding and identification of the bipotential LPCs may hold promise for new therapeutic treatments to a wide range of liver pathological conditions ranging from congenital metabolic diseases, end-stage liver cirrhosis, and hepatocarcinogenesis. Human hepatic stem cells most likely can give rise to HCC as well as cholangiocarcinomas.^38,39 In these models, a periportal population of small “primitive” oval epithelial cells proliferate either in association with or before hepatocyte multiplication. Several studies have shown a progenitor cell phenotype in a substantial number of HCCs. Detailed immunophenotyping of HCCs indicated that 28% to 50% of HCCs express markers of progenitor cells such as CK7 and CK19 (Table 1). These tumors also consist of cells that have an intermediate phenotype between progenitors and mature hepatocytes. In fact, HCCs that express hepatocyte and biliary cell markers such as albumin, CK7, and CK19 carry a significantly poorer prognosis and higher recurrence after surgical resection and liver transplantation.⁴⁰

Nevertheless, the question remains whether this immature intermediate phenotype represents progenitor cell differentiation arrest or dedifferentiation of mature hepatocytes. Histological and immunophenotyping studies favor the progenitor cell differentiation arrest model. Furthermore, 55% of small dysplastic foci (less than 1 mm in size), which represent the earliest premalignant lesions, comprise progenitor cells and intermediate hepatocytes.²⁸ Moreover, a side population of cells, with characteristics of both hepatocytic and cholangiocytic lineages, in human HCC cell lines huh7 and PLC/PRF/5 cells was found to give rise to persistently aggressive tumors on serial transplantation in immunodeficient nonobese diabetic/severe combined immunodeficient mice.⁴¹ Furthermore, our observations demonstrating the presence of TGF-β signaling components TBRII and ELF in human hepatic stem cells led us to explore the impact of these components of the TGF-β pathway on liver development and tumorigenesis. As we have reported previously, mice homozygous for elf (elf^−/−) undergo midgestational death with hypoplastic livers.⁸ Heterozygous elf^+/− mice, however, spontaneously develop tumors of the liver with an incidence of over 40%. The liver lesions include early centrilobular steatosis, dysplasia in most sections, with nuclear disarray and stratification, mitosis and apoptosis, proceeding to poorly differentiated carcinoma. We hypothesized that the interruption of the TGF-β pathway resulted in HCC through disruption of a normal pattern of cellular differentiation by hepatic progenitor/stem cells.

We examined human HCC tissue specimens from 10 individuals. In nine of the 10 HCC tissues, we observed small strongly positive clusters of three to four Oct4⁺ cells that was negative for TBRII and ELF. Cells with this phenotype were never observed during our surveys of either normal liver or of biopsies from regenerating organs. We speculate that these STAT3⁺/Oct4⁺ positive human HCC cells, which have lost TGF-β signaling proteins, have the potential to give rise to HCCs, and could thus represent a “cancer stem/progenitor cell population” (Table 1).

Indeed, multiple studies have delineated markers for LPCs and liver cancer stem cells. A review of the literature shows a number of markers expressed in putative LPCs, cells of the cholangiocytic (CK19 +) and hepatocytic lineages (albumin and alpha-fetoprotein +), HCC, and putative liver cancer stem cells. These markers are listed in detail in Table 1. The problem of identifying stem cell markers, however, lies with the ability to isolate such cells from normal liver or from HCC and the subsequent functional and molecular characterization of such cells as true stem cells that can give rise to or are contained in HCCs. Strict double and triple immunohistochemical and confocal labeling is necessary to clearly identify marker-positive stem cells. Such rigorous labeling, however, has not been consistently adhered to. Furthermore, definitive experiments with serial transplantability of marker positive cells has yet to be demonstrated. Moreover, it is difficult to identify gene products that are specifically associated with putative LPCs or with HCC. The challenge lies in defining the markers specific to these cells at varying stages of differentiation, in HCC, and the elusive liver cancer stem cell.

The Notch and Wnt Pathway and Cancer Stem Cells

Understanding the control of stem cell proliferation, however, may serve as a useful target in this challenge. The Notch signaling pathway plays an important role in stem cell self-renewal and differentiation.^3,5,68 However, other signaling pathways influence whether Notch functions as a tumor suppressor or oncogene in a particular tissue.⁶⁹ The activated intracellular form of Notch3, as well as the Notch ligand Jagged, are highly expressed in human breast cancer.^70–72 Notch-dependent transformation is associated with extracellular signal-regulated kinase activation downstream of the Ras pathway, which increases Notch messenger RNA stability and is required for transcription of the Notch target gene, Hes1.^73–75 Notch3 gene expression is increased in human breast spheroid cultures (“mammospheres”), which are enriched in stem and progenitor cells.⁷⁶ Stimulation of Notch-dependent transcription increases colony formation, whereas inhibition of Notch4 blocks branching in Matrigel, characteristics that are similar to the phenotype of Notch4 transgenic mice.^77,78

The Wnt signaling pathway also plays a crucial role in regulating stem and progenitor cell expansion.^3,79 Among the 19 members of the Wnt family, Wnt-1, Wnt-2, Wnt-3, Wnt-3a, and Wnt-10b are ligands for the canonical Wnt pathway, which activates β-catenin/transcriptional factor (TCF)-mediated transcription and induces transformation of mammary epithelial cells.^80,81 The Wnt pathway stabilizes β-catenin, a coactivator of the TCF transcription factor family.⁸² Activation of the Wnt receptor, Frizzled, in association with its co-receptor lipoprotein receptor-related protein-5/6, activates Disheveled, which results in dissociation of the tetrameric glycogen synthase kinase 3 β/β-catenin/APC/axin complex, reduction of β-catenin phosphorylation, β-catenin stabilization, and its entry into the nucleus, where it associates with and coactivates TCF. Stabilization of β-catenin can result from direct inhibition of glycogen synthase kinase 3β by protein kinase c and extracellular signal-regulated kinase, and transformation of mammary epithelial cells resulting from increased β-catenin/TCF signaling has been linked to protein kinase c alpha activation downstream of PDK1.^83–88 Conversely, glycogen synthase kinase 3β activation phosphorylates β-catenin and axin to target β-catenin for ubiquitination and proteasomal degradation. In differentiated cells, β-catenin binds to cadherins, as well as other tight-junction proteins, to form adherins junctions that are important for lumen integrity and secretory epithelial cell function.⁸¹

In gut epithelial cells, interplay between the Wnt, Hedgehog, BMP, and Notch pathways determines whether stem cells self-renew or differentiate (Fig. 2). Wnt signaling is activated in the colonic crypts and maintains cells in a proliferative state. Increased activity of the Wnt pathway leads to enlarged crypts and intestinal tumors, whereas Wnt inhibition results in loss of the stem cell compartment altogether. Wnt signaling is essential for maintenance of this stem cell compartment and regulates cellular differentiation.⁸⁹ This effect of Wnt signaling can be mimicked by stabilizing cytoplasmic/nuclear β-catenin alone, and the effects of Wnt signaling on stem cells are modulated through association with other signaling pathways including Notch, Sonic hedgehog, and TGF-β signaling.^6,90,91 Understanding of the regulation of this stem cell compartment has come mostly from examining skin, intestinal, and hematopoietic cells. Mutations and deletions of AXIN1 and AXIN2 and overexpression of β-catenin and frizzled homolog-7 have been described in 17% to 40% of human HCCs analyzed.^92–97 Not all studies, however, demonstrate a correlation between elevated nuclear β-catenin and expression of its transcriptional targets in HCC, suggesting that the expression of these target genes is also regulated by alternative signaling pathways.^98–100

Fig. 2.

Directed stem cell renewal and liver differentiation by multiple signaling pathways. Proteins direct self-renewal of embryonic stem cells (red) and then differentiation of stem cells into mature hepatocytes (blue). The mechanisms directing transformation of stem cells to cancer stem cells and hepatocellular cancer, however, remain to be elucidated.

In addition, Wnt signaling can promote stem cell renewal and rescue of the hematopoietic stem cells compartment after irradiation either through the canonical (Wnt-3a) or possibly also through the non-canonical (Wnt-5a) pathway.^101,102 In fact, repopulation of a hematopoietic stem cells compartment from CD34+ cells was augmented more than threefold through intraperitoneal injection of Wnt-5a alone.¹⁰² These effects could be mimicked through activation of the Wnt signaling pathway by a degradation-resistant β-catenin, whereas hematopoietic stem cells compartment reconstitution can be inhibited by overexpression of Axin.¹⁰¹ How the Wnt pathway promotes stem cell renewal and is involved in liver cancer stem cells, however, is still unclear. In some circumstances, Notch acts jointly with Wnt to sustain stem cell proliferation, and is essential for the differentiation of specific cell types. Moreover, Hedgehog signaling promotes differentiation and restricts crypt formation that is mediated through its effect on BMP signaling.¹⁰³

Liver Cancer Stem Cells and the TGF-β Signaling Pathway

Recent evidence also suggests a key role for the TGF-β signaling pathway in both foregut cancer suppression and normal gut endoderm development, suggesting a dual role in liver tumor suppression as well as transition of stem cells to a progenitor and fully differentiated phenotype (Fig. 3).^6,104 The TGF-β signaling pathway appears to be most prominent at the interface between development and cancer in liver and gut epithelial cells.⁶ Smad signaling is pivotal for embryonic hepatocyte proliferation, as well as in the formation of gastrointestinal cancers.^105–107 Smad activation is modulated by various receptor-interacting or Smad-interacting proteins that include ubiquitin and small ubiquitin-related modifier ligases, as well as multiple adaptor proteins that include SARA (Smad anchor for receptor activation), filamin, and ELF. ELF, a β-spectrin first isolated from foregut endodermal stem cell libraries, is crucial for the propagation of TGF-β signaling.¹⁰⁸ Specifically, ELF associates with Smad3, presenting it to the cytoplasmic domain of the TGF-β type I receptor complex; followed by heteromeric complex association with Smad4, nuclear translocation and target gene activation.⁸ Disruption of ELF in mice leads to disruption of TGF-β signaling, resulting in a phenotype similar to Smad2 ^+/−/Smad3 ^+/− mutant mice with mid-gestational death due to gastrointestinal, liver, neural, and heart defects, and loss of intrahepatic bile ducts. Interestingly, while liver lineage is established, hepatocytes are poorly formed and liver architecture is lost with an absence of primitive bile ducts as in the Smad2 ^+/− Smad3 ^+/− mutants (Fig. 2). Moreover, bile duct formation can be induced in liver explant cultures treated with TGF-β. In this model of haplo-insufficiency, the addition of TGF-β appears sufficient to drive up Smad2 and Smad3 levels and the formation of a limiting plate and bile ducts. Moreover, a TGF-β/BMP regulated protein PRAJA (Proliferation Regeneration Architecture Juxtaposition Apoptosis) is expressed in hepatoblasts and modulates ELF and Smad3, adding to the complexity of TGF-β signaling in the gut.¹⁶

Thus, emerging new data indicate that TGF-β is important for stem cell transitioning to a progenitor cell phenotype and ultimately its conversion to a fully differentiated phenotype in the liver and biliary system. Ultimately, the common mediator Smad4 and ELF heterozygotes survive to adulthood only to succumb to a gastrointestinal cancer. The elf ^+/− heterozygous mice develop HCC.^21,109–111 Moreover, inactivation of at least one of the TGF-β signaling components occurs in almost all gastrointestinal tumors.^6,105 Thus, it is conceivable that tumors arise in organs lacking crucial differentiating factors such as Smad2 with Smad3 at the progenitor to transitional stages or at the stages when stem cells divide into progenitor cells that then develop into immature epithelial cells. The data support that absence of TGF-β–driven epithelial differentiation favors carcinogenesis. Impaired TGF-β signaling may distinguish cancer stem cells from normal stem cells and give clues toward identifying a human progenitor cell pool and other functional pathways that become activated in such “cancer stem/progenitor cells.”

Prognostic Markers

Recent research efforts have focused on identifying markers exclusive to CSCs and developing targeted therapies (Table 1).^46,112 For example, cluster of differentiation 90 (CD90⁺) is a surface protein that is expressed by hepatic stem/progenitor cells during liver development, but infrequently in the adult liver. CD90⁺ cells exhibit stem cell–like properties and have been shown to propagate tumorigenicity into secondary severe combined immunodeficient/beige mice. CD45⁺CD90⁺CD44⁺ cells also may serve as a sensitive and specific marker for early tumor diagnosis in HCC and suggest a valuable possible therapeutic target.⁴³ CD133, a transmembrance glycoprotein, is also a valuable marker and is expressed in 1% to 5% of human HCC, while absent in normal liver cells.⁴⁶ However, the role of these markers in defining functionally distinct populations of cells from progenitor to differentiated hepatocytes is controversial. Part of the problem lies in their ubiquitous expression, compounded by loss of function studies that do not yield the expected loss of progenitor cell population in the multiple lineages incriminated. Similarly, loss of expression of members of the TGF-β pathway such as TBRII, ELF, and Smad4 in cells that express embryonic stem cell markers Nanog, STAT3, and Oct3/4 could represent a significant prognostic event in HCC. Whereas genetic studies in mice suggest that loss of ELF/TGF-β signaling and an increase in STAT3 contribute to the transformation of a normal hepatic stem cell to a cancer stem cell, the low numbers of stem cells and difficulties in isolation and primary culture have precluded clear delineation of stages of differentiation. There is therefore an urgent need for defining clear culture conditions to show the role of all of these markers in specific cell stages, from stem cell to a differentiated hepatocyte or HCCs, with clear endpoint assays reflecting the different stages of differentiation.

New Treatment Strategies

The failure of existing treatments for cancer has initiated the search for new methods that effectively target CSCs. CSCs are difficult to treat with conventional methods because of their chemoresistant and radioresistant properties, as well as their ability to stimulate angiogenesis. Moreover, these properties of CSCs contribute to tumor recurrence and treatment resistance in advanced cases via increased levels of BCL-2 family proteins or activation of ABC transporters.^113,114 Radioresistance has been shown to involve Wnt/β-catenin induction, promoting genomic instability and accelerating conversion of nontumorigenic cells to glioma CSCs that are able to survive radiation.¹¹⁵ Notch and Hedgehog-Gli 1 signaling pathways, involved in human CSC self-renewal and tumorigenicity, also may be responsible for CSC recurrence after radiation therapy.¹¹⁶ Successful therapies would target CSCs via inducing differentiation of CSCs into nontumorigenic cells or completely eliminating the cells via inhibition of the self-renewing stem-cell state.^117,118 An effective treatment would also include uniquely targeted agents that sensitize CSCs to radiation therapy, debulk cell mass, and disrupt angiogenesis. Inhibitors of the pathways that lead to these properties would be optimal targets for future cancer care.¹¹³

Numerous pathways involved in signaling normal stem cell fate, including self-renewal, can be manipulated to decrease the self-renewal and proliferating capabilities of CSCs. For example, the Hedgehog signaling pathway normally regulates adult neuronal progenitor fate and self-renewal.¹¹⁷ In pancreatic cancers, which often have an aberrantly overactive Hedgehog pathway, cyclopamine, a small molecule Hedgehog inhibitor, reduces aldehyde dehydrogenase, a stem/progenitor marker, and is currently in phase II trials for this disease.^119–122 Similarly, inhibition of the Notch pathway via specific gamma secretase inhibitors inhibits cancer stem cell self-renewal and decreases tumor growth.¹²³ The Wnt pathway is inhibited via small molecules, either small interfering RNAs or antibodies, that block β-catenin interaction with TCF gene activation or transcriptional coactivator and also by vita-min D.^124–127 Moreover, examinination of HCC reveals hepatic stem cells that have lost TBRII and ELF. Further analysis revealed activation of the IL-6 signaling pathway, suggesting that HCC may be induced by IL-6 activation in the setting of loss of the TGF-β signaling pathway. This reveals an important therapeutic target, IL-6, for future HCC therapy.⁹ Currently, some successful strategies involve blocking IL-6 signaling, which in effect limits self-renewal.¹²⁸

Moreover, differences in posttranscriptional and post-translational modifications in normal and cancerous cells reveal that different splicing variants of the adhesion receptor, CD44, are expressed differently in each. Hence, future therapeutic strategies can focus on targeting this cell surface marker with specified antibodies.^118,129 Differentiation of cancer cells into less aggressive, more differentiated cells has also been shown to be a successful strategy, particularly in treatment of acute promyleocytic leukemia. All-trans retinoic acid after normal chemotherapy resulted in a 90% remission and 70% cure rate.¹¹⁷ However, the response of a cancer stem cell to a differentiating agent such as retinoic acid may not be predictable and may result in inappropriate differentiation. For example, some breast cancer cells trans-differentiate into cells with the genotypical and phenotypical characteristics of blood vessels when treated with retinoic acid.¹³⁰ Differentiation therapy in HCC is an attractive strategy but will require a major effort to define the liver cancer stem cell and its differentiation pathways.

Targeting the stem cell niche is another promising therapeutic strategy. The specified microenvironment in which stem cells reside often dictates self-renewal and reproduction.¹³¹ Alteration of components of the stem cell niche can effectively change stem cell fate, as in the case of experimental parathyroid hormone induction.^131–133 Furthermore, human embryonic stem cell-derived fibroblast-like cells provide a supportive environment for stem cells through insulin-like growth factor 2.¹³⁴ Targeting insulinlike growth factor 2 therefore can manipulate the stem cell microenvironment. How this therapy may be efficacious in cancers of the liver, however, remains to be elucidated.

A clear and precise delineation of markers specific to stages of CSC formation, from cell surface markers such as CD90, CD45, and CD133 to stem cell markers such as Nanog, Oct3/4, and STAT3, as well as the precise tumor suppressive role of the TGF-β/Smad/ELF pathway, remain crucial for developing therapeutic strategies targeting CSCs in HCC. Targeting pathways that lead to the self-renewal and proliferating properties of cancer stem cells such as Hedgehog, Wnt, Notch, and IL-6, using differentiation therapy such as retinoic acid, and targeting the microenvironment of the stem cell niche could all provide powerful new therapeutic approaches to this lethal and poor prognosis cancer.

Abbreviations

BMP

bone morphogenetic protein

CD90

cluster of differentiation 90

CK

cytokeratin

CSC

cancer stem cells

ELF

embryonic liver fodrin

HCC

hepatocellular carcinoma

IL

interleukin

LPC

liver progenitor cells

Oct

Octamer

STAT3

signal transducer and activator of transcription protein 3

TBRII

transforming growth factor beta type II receptor

TCF

transcriptional factor

TGF-β

transforming growth factor beta