Tumor progression: abnormal developmental plasticity/reprogramming?

In leukemia and in the development of carcinoma, several oncogenes / tumor suppressor genes cooperate to induce abnormal survival, proliferation, and developmental / behavioral plasticity in tumor cells, which may adopt properties of stem cells. Using both genetically modified mice and in vivo-like cell culture models, we focus on molecular mechanisms that transform hematopoietic stem cells into leukemia-initiating cells, and reprogram epithelial cells into primitive mesenchymal, stem cell-like cells during the progression of carcinoma and metastasis.

Mechanisms in hematopoietic progenitor renewal and leukemogenesis.

Figure 1 (Click to view legend)

Human leukemia involves the cooperation of mutated transcription factors/chromatin regulators with mutated/overexpressed receptor tyrosine kinases/signal transducers In erythroid cells. Progenitor renewal requires cooperation of the EpoR, c-Kit, and the glucocorticoid receptor (GR) in primary mouse erythroblasts and in immortal but diploid, and in vivo-like murine ES cell-derived erythroblasts (ESEPs). The latter are almost indistinguishable from erythroid progenitors in vivo by antigenic and functional characteristics, and were used to characterize erythropoiesis defects caused by lack of Stat5, p38 MAPK and Flk-1. We also purified hematopoietic stem cells (HSC) from multilineage mouse leukemias caused by the human chronic myelogenous leukemia (CML) oncogene BCR-ABL p210 and constitutively active Stat5, showing that they function as leukemic stem cells (LSC’s). These cells could be expanded in culture on a long-term basis without losing their ability to cause multilineage leukemia upon transplantation. In contrast, HSCs infected with BCR-ABL p185 (causing human B-ALL) initiate leukemia, but are lost during disease progression through differentiation into committed immature B-cell progenitors which function as LSCs in secondary transplants, although constituting the bulk of leukemic cells (Fig 1). We now focus on potential molecular and cellular differences between these two LSC types and normal HSCs by expression profiling and functional analysis.

Tumor progression and metastasis: Ras plus mutated genes disrupting epithelial polarity?

Figure 2 (Click to view legend)

Epithelial to mesenchymal transitions (EMT) are essential during normal development, carcinoma progression and metastasis, and may be a hallmark of carcinoma stem cells. Oncogenic Ras plus TGFβ cause EMT and metastasis in the mammary epithelial cell model (EpH4/ EpRas), while a Ras mutant hyperactivating the PI3K pathway (EpC40 cells) induces a “scattering” phenotype upon TGFβ-addition and is tumorigenic, but not metastatic. We currently focus on 4 genes identified by polysome-bound mRNA expression profiling/protein studies, i.e. ILEI (interleukin-like EMT inducer), CREG (cellular repressor of E1A regulated genes, AnxA1 (Annexin-A1), and mScrib (mouse paralog of Drosophila Scribble). ILEI causes reversible EMT, slow tumor growth and metastasis in oncogene-free EpH4 cells, but requires oncogenic Ras for these functions in cultured, in vivo-functional p19ARF-/- hepatocytes.
ILEI thus resembles CREG, AnxA1 and mScrib in their requirement for oncogenic Ras to cause EMT and metastasis upon loss or gain of function. 

Figure 3 (Click to view legend)

Surprisingly, ILEI causes EMT by intracellular mechanisms, as shown by ILEI mutant proteins – neither secreted nor extracellularly cleaved by plasmin – that cause EMT. Secreted ILEI may, however, also perform extracellular functions: 1. Non-cleaved, secreted ILEI forms homodimers and is cleaved by plasmin after binding to ECM-fibronectin. 2. A highly secreted ILEI mutant protein lacking the propeptide removed by plasmin functions as a “super ILEI” in metastasis induction. Importantly, ILEI localizes to the trans-Golgi-network in normal cells, but is redistributed to vesicles in the entire cytoplasm upon EMT induction, explaining our earlier observation that cytoplasmic localization of ILEI is a strong predictor of metastasis in human breast cancer and melanoma. CREG, a secreted, mannose 6P-receptor binding, lysosomal protein, also caused EMT and metastasis in EpC40 cells, a feature shared by Drosophila and plant (Arabidopsis) CREG paralogs (Fig 2). Finally, RNAi-mediated knockdown of AnxA1 – a Ca++-binding, membrane-associated protein regulating multiple aspects of vesicle-trafficking  – caused EMT and metastasis in EpC40 cells, but reversed EMT and abolished metastasis when overexpressed in dedifferentiated metastatic human mammary carcinoma cell lines. AnxA1 RNAi-induced EMT in EpC40 cells which required JAK/STAT3 signaling and Erk/MAPK activation. Therefore, cooperation of Ras with loss or gain of function in proteins from molecular machines essential for epithelial polarity causes EMT and metastasis, possibly by altering intracellular organelle-associated signal transduction occurring in MAPK, PI3K, TGFβR and STAT3 signaling. A variant of MDCK cells (the major cell model to study epithelial polarity) – able to undergo complete EMT in response to an estradiol-activated RafDD-ER protein (Fig. 3)  – will be used to further investigate the above mentioned hypothesis.

 
topprint