Our data confirm previously unrecognized levels of temporal

Our data confirm previously unrecognized levels of temporal diversity and complexity in the subclonal landscape of GBM. In all cases analyzed, genetically distinct subclones had variable serial repopulating activity in vivo. We can exclude the possibility that the presence of in vitro aberrations played a role in the clonal architecture of the mouse xenografts, given the culture conditions used and the number of in vitro cell passages. Since the in vivo readout is likely to be a functional activity of self-renewing CSCs, this suggests that the competitive self-renewal ability of tumor-propagating stem CEP-37440 in GBM varies on the basis of frequency and/or quantitative features (e.g., proliferation rates and growth factor dependence). This is in line with the principle that the extensive replicative potential of CSCs allows subclonal evolution (Greaves and Maley, 2012). Our interrogation of subclonal genetic diversity of human GBM has revealed that tumor-propagating cells in GBM are genetically heterogeneous and have a variable competitive capacity for tumor propagation in vivo. The link forged among genetic diversity, clonal architecture, and propagating activity in vivo may facilitate the characterization of mutational variants that are responsible for disease recurrence and therapeutic resistance in patients (Johnson et al., 2014).
Experimental Procedures
Author Contributions
Acknowledgments
Introduction Frontotemporal dementia (FTD) accounts for ∼50% of dementia cases before the age of 60. Up to 40% of FTD patients have a familial history (Goldman et al., 2005; van Swieten and Heutink, 2008) due to mutations in the microtubule-associated protein tau gene (MAPT), progranulin gene (GRN), or C9orf72 gene (Baker et al., 2006; Cruts et al., 2006; DeJesus-Hernandez et al., 2011; Hutton et al., 1998; Renton et al., 2011). The majority of FTD-causing mutations in GRN are predicted to result in functional null alleles, causing haploinsufficiency. Progranulin (PGRN) has neurotrophic function in vitro and in vivo. Although PGRN−/− mice are viable, they do not recapitulate all the features of FTD (Kayasuga et al., 2007). Human somatic cell reprogramming to a pluripotent state (induced pluripotent stem cells; iPSCs)(Takahashi et al., 2007a) can create human disease models in vitro using patient-derived iPSCs (Kim, 2014), including neurodegenerative diseases (Qiang et al., 2013) and, specifically, FTD (Almeida et al., 2012). Unlike in the published FTD-iPSC model that differentiated iPSCs to a mixture of neuronal cells, we evaluated cortical neuron development from FTD-patient-derived iPSCs, as FTD is characterized by selective neurodegeneration of the frontal and/or temporal cortex (Neary et al., 2005). We demonstrate that FTD-iPSCs carrying a GRNIVS1+5G > C mutation differ in their ability to generate cortical neurons from control lines (iPSCs and human embryonic stem cells; hESCs) and that genetic correction restores this differentiation defect.
Results
Discussion We demonstrate that iPSCs generated from FTD patients carrying a GRN IVS1+5G > C null mutation poorly differentiate to cortical neurons, while differentiation to motor neurons appears normal. Introduction of the GRN cDNA in the AAVS1 locus of FTD-iPSCs using ZFN-mediated homologous recombination corrected the inefficient cortical neuron generation. Loss of PGRN caused aberrant activation of the WNT signaling pathway during neuronal differentiation, and inhibition of WNT signaling partially rescued the FTD-iPSC corticogenesis defect, further substantiating the notion that the WNT signaling pathway might be a therapeutic target for FTD (Korade and Mirnics, 2011). Almeida et al. (2012) previously demonstrated that GRNS116X-iPSCs could be differentiated toward neurons with a similar efficiency as that of control iPSCs but that GRNS116X-iPSCs were more sensitive to stress. We chose to differentiate neuronal progenitors from iPSC lines specifically toward the cortical lineage, as we hypothesized that evaluation of the influence of PGRN haploinsufficiency on cortical neurons, the most affected cell type in FTD, would likely yield the most important insights into the disease. We induced cortical neuron differentiation and demonstrated that generation of TUJ1-positive cells coexpressing cortical markers such as CTIP2, FOXP2, and TBR1 was significantly decreased in FTD-iPSCs. By contrast, Almeida et al., who differentiated iPSCs sequentially as embryoid bodies and neurospheres with additional growth factors to generate a mixture of glutamatergic, GABAergic, and dopaminergic neurons, did not observe decreased cell differentiation. We hypothesize that the differences in the efficiency of neural progeny generation between the two studies might be caused by the specific subtype of neuronal cells we generated. In fact, generation of HB9/TUJ1- and ISLET-1/SMI32-positive motor neurons appeared similar when FTD-iPSC or H9-hESC lines were differentiated to motor neurons.