Cancer Stem Cells and the Microenvironment

∗ Disclaimer: Max S. Wicha has financial holdings and is a scientific adviser for OncoMed Pharmaceuticals, is a scientific adviser for Veristem, and has research support from Dompe.

Accumulating evidence suggests that most if not all tumors are maintained by a subpopulation of cells that display stem cell properties including self-renewal and lineage differentiation. Cancer stem cells (CSC) have been isolated from a number of human malignancies by using cell surface markers and enzymatic activity of cytoplasmic proteins. Subsequent characterization of these CSCs in mouse xenograft models revealed that these cells can mediate metastasis and contribute to treatment resistance and relapse. Furthermore, recent studies suggest that the CSCs are regulated by the components of tumor microenvironment through complex networks of cytokines and growth factors. Importantly, these components have a direct influence on CSC properties and thus may represent attractive targets for development of novel therapeutics. This chapter highlights advances in elucidating the networks between CSC and the tumor microenvironment and efforts to target these CSC regulatory networks.

Identification of Cancer Stem Cells

Embryonic and tissue-specific stem cells display two distinct properties: (1) self-renewal, or the ability of the cell to undergo several symmetric or asymmetric divisions while maintaining an undifferentiated cell pool; and (2) differentiation, or the ability to generate distinct cell types. Tissue-specific stem cells are distinguished from embryonic stem cells in that their differentiation is primarily limited to cell types of a particular organ. Tissue-specific stem cells have the capacity to self-renew as well as to differentiate into committed progenitors and terminally differentiated cells with specialized functions. There is increasing evidence that a similar hierarchy governs many human malignancies, including tumors of the hematopoietic system and solid organs. CSCs are operationally defined by their ability to initiate tumors in mice on serial passage, a demonstration of self-renewal, as well as their ability to differentiate into the non–self-renewing cells forming the tumor bulk. Human CSC assays have used immunocompromised mice whereas mouse CSCs have also been identified via transplantation studies in syngeneic hosts. In fact, these early studies on normal organ development and tumors suggested that the tumors are indeed organ-like structures resembling their normal counterparts in that they are both comprised of heterogeneous cell populations.

According to the CSC hypothesis, tumors are hierarchically organized whereby self-renewing CSCs drive tumorigenesis while differentiated cells form the tumor bulk. The CSC model fundamentally differs from the traditional or “stochastic” model of carcinogenesis in which any cell may have equal malignant potential. Based on the stochastic model, most therapeutic strategies have been selected for their ability to cause tumor shrinkage by targeting rapidly cycling cells, whereas the CSC model predicts that targeting and elimination of self-renewing cancer stem cells will be necessary to significantly improve outcome and ultimately cure patients with cancer.

Despite the fact that the heterogeneity of tumor cells has been widely acknowledged, the CSC model was difficult to validate until appropriate mouse models were developed. The development of biomarkers to identify CSCs, as well as validation of in vitro and mouse models, has facilitated the isolation and characterization of these cells from both murine and human tumors. Bonnet and Dick were the first to describe the hierarchical organization of acute myeloid leukemia (AML) and the existence of CSCs in that disease. The team isolated such cells from AML and demonstrated that a small subset of cells characterized by CD34 ⁺ /CD38 ^- phenotype that made up less than 0.01% of tumor cells could transfer human AML into NOD/SCID mice, whereas the remaining cells that lacked this phenotype failed to do so. Using similar techniques, CSCs have subsequently been isolated from many human malignancies, including those of the brain, breast, colon, prostate, pancreas, lung, and liver.

Examples of Cancer Stem Cells

Uchida and colleagues described a phenotype, CD133 ⁺ CD34 ⁻ CD45 ⁻ , that is characterized by the expression of a cell surface antigen—Prominin-1 (CD133), a five-transmembrane glycoprotein—and lack of CD34 and CD45 to isolate human central nervous system stem cells. Brain tumors displayed a cellular hierarchy reminiscent of their normal counterparts, in which CD133 ⁺ tumor cells but not CD133 ⁻ cells were able to form tumors in NOD-SCID mouse brains and neurospheres in in vitro cultures.

A subpopulation of breast cancer cells that displayed stem cell properties was characterized by expression of the cell surface markers ESA (EpCam) and CD44 in the absence of CD24 expression. As few as 200 EpCam ⁺ CD44 ⁺ /CD24 ⁻ Lin ⁻ cells were able to generate tumors in immunocompromised NOD/SCID mice, whereas 100-fold more cells that did not express these markers isolated from the same tumors were nontumorigenic. Furthermore, the tumor-initiating populations regenerated tumors that recapitulated the heterogeneity of the initial tumor. Subsequently, both normal and malignant mammary stem cells were isolated by virtue of their increased expression of aldehyde dehydrogenase (ALDH), which can be accessed by Aldefluor assay. Interestingly, CD44 ⁺ /CD24 ⁻ and ALDH ⁺ identify overlapping but not identical cell populations. Furthermore, these markers can be used to isolate CSC populations from established breast cancer cell lines, as well as primary tumor xenografts.

In the normal prostate, CD133 ⁺ cells have been shown to display properties of stem cells, whereas the CD44 ⁺ /α2β1 ^hi /CD133 ⁺ phenotype (about 0.1% of the total tumor cells) marked prostate cancer stem cells. As few as 500 cells with this phenotype were able to form tumors in mice, whereas 5 × 10 ⁵ cells that lacked this phenotype failed to generate tumors.

Distinct pulmonary epithelial cells located at the bronchioalveolar ductal junction displayed functional stem cell properties in mouse models. These cells, characterized by the Sca1 ⁺ /CD34 ⁺ /CD45 ⁻ /Pecam ⁻ phenotype, were able to generate lung epithelium after tissue damage, suggesting that these cells possess self-renewal and lineage differentiation properties. Furthermore, in K-Ras–induced mouse lung adenocarcinoma, a Sca1 ⁺ /CD34 ⁺ /CD45 ⁻ /Pecam ⁻ phenotype identified lung cancer stem cells capable of initiating tumors when transplanted into syngeneic mice. In addition, human lung cancers contain a rare population of cells that express CD133 and ALDH1. These cells are able to initiate tumors as well as to generate differentiated cells in mouse xenografts.

Tumor-initiating cells have also been isolated in colorectal tumors. Ricci-Vitiani and co-workers and O’Brien and colleagues, isolated CD133 ⁺ human colon cancer cells and injected them subcutaneously or under the renal capsule of NOD-SCID mice. Both groups demonstrated that CD133 ⁺ cells not only were capable of forming tumors, but also generated tumors that recapitulated the cellular heterogeneity of the initiated tumors. In contrast, CD133 ⁻ cells lacked these properties, suggesting that the capabilities of self-renewal and lineage differentiation are required for tumor initiation as well as generating non–self-renewing cells that constitute the tumor bulk.

Human pancreatic carcinomas contain a subpopulation of cells with a CD44 ⁺ /CD24 ⁺ /ESA ⁺ phenotype (0.2% to 0.8%) that were 100-fold more tumorigenic than those that did not express these markers. These pancreatic CSCs also displayed activation of developmental pathways such as sonic hedgehog.

Human liver cancer stem cells characterized by the expression of CD133 demonstrated self-renewal potential, increased colony-forming ability, and tumor formation in mouse xenografts. In addition, liver tumor cells that express CD90 (Thy1), glycosylphosphatidylinositol-anchored glycoprotein, and CD44 have been shown to be able to initiate tumors that are capable of metastasis in immunodeficient mice.

You're Reading a Preview

Become a Clinical Tree membership for Full access and enjoy Unlimited articles

Become membership

If you are a member. Log in here