报告题目：Understanding tumor cell maneuvers with MADM, a genetic mosaic system

报告人： Dr. Hui Zong, Assistant professor, Department of Biology, Institute of Molecular Biology, University of Oregon, Eugene, OR

主持人：  翁杰敏 教授

报告时间：9月5日 10:30（周三）

报告地点：闵行生命科学学院大楼534报告厅

报告内容简介：Malignant tumors are  extremely  difficult to treat due to multiple molecular and cellular aberrations and their ability to undergo metastasis. To cure cancer, we must greatly increase our efforts on studying pre-cancerous lesions so we could move the first line of defense prior to full-blown malignancy. However, patient samples and conventional animal models are not suitable for such studies  since early lesions are  “invisible” due to the lack of pathological features. Furthermore, tumor-niche interaction that is critical for tumor progression has also been hard to study  due to the difficulty in distinguishing tumor from by-standing normal cells in a tumor mass. To  overcome these problems, our lab  utilizes a powerful mouse  genetic system termed MADM (Mosaic Analysis with Double Markers), which  creates sporadic GFP-labeled  tumor-suppressor mutant cells that are traceable during the entire course of tumorigenesis. Here I will discuss how our  high-resolution modeling of two  types of  brain tumors (glioma and medulloblastoma) provided mind-opening insights into the swift maneuvers of tumor cells. Malignant glioma is incurable. Diffuse infiltrative tumor cells make complete surgical resection impossible, while the resistance against chemotherapy and radiation leaves traditional cancer treatment ineffective.  The only hope for a cure of this devastating disease is targeted therapy. While most labs are focused on molecular targets, our lab’s research is focused on identifying cell-of-origin for glioma as the cellular target. The cell-of-origin for glioma has long been thought to be neural stem cells (NSCs) based on two observations. First, purified tumor cells manifest stem cell features. Second, the introduction of p53 and NF1 mutations into NSCs in mouse models led to glioma formation. However, endpoint features do not provide a reliable basis to determine the  identity of tumor initiating cells. Furthermore,  conceptually there is a critical difference between cell-of-origin and cell-of-mutation. The former is the cell type that transforms into malignancy, while the latter is the one in which initial mutations occur but may not directly transform. The best approach to resolve this problem is to monitor the entire tumorigenic process  in vivo with a single-cell resolution. We used MADM to probe into early phases of gliomagenesis, and surprisingly found  the lack of overpopulation of mutant NSCs. Among NSC-derived cell types, we only detected dramatic over-expansion of mutant oligodendrocyte precursor cells (OPCs) at pre-transforming stages. Consistently, terminal-stage tumor cells displayed salient OPC features by both histological criteria and transcriptome profile. Most importantly, introducing the same mutations directly into OPCs was sufficient for malignant transformation. Our findings suggest OPCs as the cell-of-origin in this model even when initial mutations occur in NSCs, and highlight the importance of analyzing pre-malignant stages to identify the cancer cell-of-origin3. Medulloblastoma is the most prevalent type of pediatric brain tumors, originating from cerebellar  granule neuron precursors (GNPs) upon Shh pathway hyperactivation.  It  has been puzzling to  clinicians that, although GNPs are unipotent  for the granule neuron lineage, the tumor mass often contains glial cells. Moreover, medulloblastomas are often highly vascularized even though GNPs are known not to produce angiogenic factors. Using the GNP-specific Math1-Cre,  we established a MADM model  that  generated  medulloblastoma at full penetrance. Although the majority of green cells in the tumor mass resemble GNPs, a closer look revealed a population of green astroglial cells. The fact that these astroglial cells are GFP+ suggests that they are derived from mutant GNPs, which is surprising since normal GNPs are known not to give rise to any glial cells. To investigate the molecular mechanisms that enable tumor GNPs to  generate  astraglial cells, we performed  RNA-seq analysis of tumor  GNPs in comparison to wildtype GNPs and observed the  increased expression of  a few  transcription factors that are normally restricted to a small time window during embryonic GNP lineage development. It suggests that tumorigenic transformation might lead to a reversion of the fate of some tumor GNPs to an earlier developmental window with broader  differentiation  potentials to generate glial cells. In addition to investigating the mechanisms of glial formation, we are also analyzing the functional roles of tumor glial cells, in particular their roles in inducing angiogenesis to support tumor GNPs. Overall, our MADM-based medulloblastoma model points to a community building behavior of tumor cells, through which some of them convert their fate to non-proliferative glial cells, which in turn provide cues for the establishment and maintenance of the tumor microenvironment.  In conclusion, our studies revealed how “clever” tumor cells could make their moves, explaining why cancer remains a formidable enemy to human health. To overcome this disease, it is imperative for us to continue our investigations on mechanistic events that contribute to the malignancy with the highest precision. As Sun Zi War Strategies said, “know yourself and know your enemies, you will win all the battles.”