Searching high and low: Cancer stem cells in the eye


  • Potential conflict of interest: Nothing to report.

Yang ZF, Ho DW, Ng MN, Lau CK, Yu WC, Ngai P, et al. Significance of CD90+ cancer stem cells in human liver cancer. Cancer Cell 2008;13:153–166. (Reprinted with permission.)


This study characterized cancer stem cells (CSCs) in hepatocellular carcinoma (HCC) cell lines, tumor specimens, and blood samples. The CD90+ cells, but not the CD90− cells, from HCC cell lines displayed tumorigenic capacity. All the tumor specimens and 91.6% of blood samples from liver cancer patients bore the CD45−CD90+ population, which could generate tumor nodules in immunodeficient mice. The CD90+CD44+ cells demonstrated a more aggressive phenotype than the CD90+CD44− counterpart and formed metastatic lesions in the lung of immunodeficient mice. CD44 blockade prevented the formation of local and metastatic tumor nodules by the CD90+ cells. Differential gene expression profiles were identified in the CD45−CD90+ and CD45−CD90− cells isolated from tissue and blood samples from liver cancer patients and controls.


The cancer stem cell (CSC) model of tumor development and progression is based on the theory, that tumors, like normal adult tissues, contain a subset of cells that both self-renew and give rise to a differentiated progeny (Fig. 1). However, these stem cells are rare and therefore hard to define and to characterize.1 The identification of markers that allow the prospective isolation of CSCs from whole tumor tissues will allow us to develop a better understanding of the important biological properties of CSCs.2 This may lead to the discovery of uniquely expressed molecules in CSCs that can be specifically targeted to improve current cancer therapy. Research on cancer stem cells in solid tumors has just begun, and the first CSCs were identified in breast cancer only a few years ago.3 In recent issues of Cancer Cell4 and HEPATOLOGY,5 Yang et al. describe the identification of cancer stem cells in hepatocellular carcinoma (HCC) cell lines, tumor specimens, and patient blood samples. The potential cancer stem cells are characterized by the expression of CD90 and CD44 (proteins mainly involved in cell-adhesion and migration).

Figure 1.

CSCs develop from either mature parenchymal cells or a pre-existing pool of necessary stem cells for normal tissue maintenance. They can give rise to a more differentiated progeny, which builds the tumor and results eventually in another clone of CSCs.

First, the authors used HCC cell lines to evaluate the tumorigenic capacity of CD90+ cells. Only cell fractions enriched in CD90+ cells were capable of transferring the disease to nude mice; in contrast, CD90− cells did not form any tumor nodules. Multimarker analysis revealed that the majority of CD90+ cells in the investigated HCC lines also expressed CD44. Serial transplantation experiments confirmed the importance of both stem cell markers for tumor growth, as only CD90+/CD44+ double-positive cells exhibited tumorigenic properties. Interestingly, inhibition of CD44 activity induced apoptosis of CD90+/CD44+ cells in vitro and inhibited local and metastatic tumor cell engraftment in vivo. Subsequently, the authors sought to validate their findings in HCC cell lines and isolated CSCs from solid human liver tumors and blood of patients with HCCs. A combination of CD90 and CD45 was used to exclude contamination with blood cells. Flow cytometry identified CD45−CD90+ cells in all tumor samples, but a few cells were also found in nontumorous tissues. CD45−CD90+ cells isolated from normal, cirrhotic and tumor liver tissue were then transplanted into severe combined immunodeficient (Scid)/beige mice to analyze their tumorigenic potential. Only CD45−CD90+ cells isolated from tumor specimens gave rise to tumor nodules in immunodeficient mice in contrast to cells from normal and cirrhotic tissue. Those findings are repetitively reported by the same group in a recently published article in Hepatology. Multimarker analysis revealed the coexpression of CD44 and other stem cell markers such as CD133 in a subset of CD45−CD90+ cells in tumor xenografts.

Although this is the first study to identify CSCs in human liver cancer, analysis of CD90 expression alone is not sufficient to define a CSC per se, because CD90+ cells were also isolated from normal, cirrhotic, and nontumor tissues, which are however not tumorigenic. The issue is even more complicated given the fact that all the investigated tumor samples contained in the SCID-mouse model tumorigenic CD90+ cells, but not all patients with HCC, develop metastasis or relapse after receiving a liver transplant.6 Therefore other markers need to be defined to discriminate a CD90+ cell from a malignant tumor stem cell. Gene expression analysis revealed high levels of CD44, Oct4, and Bmi in CD90+cells isolated from tumor tissues, genes that all have been implicated in tumorigenesis.7 However, it might not be enough to perform global gene expression analysis, although this seems be to a first logical step.8 CSCs might undergo additional genetic changes, leading to alterations in their phenotype or a different clonal evolution of certain stem cells with even different gene profiles. Even if research on CSCs does not lead to new therapeutics in the first way, the value of a better understanding of the biology of CSCs in the liver and other cancers will be enormous. At the minimum, a patient's prognosis could be better determined, as has already been shown in other tumors.3, 9