Recent studies have shown that mouse and human fibroblasts can be effectively converted to neurons 
. Disease-specific fibroblasts from Alzheimer’s patients can also be converted to neurons 
. However, it is unknown whether human glioma cells can be induced to become neurons or whether this would result in reduced proliferation rate. In this study, we showed that a combination of three transcription factors Ascl1, Brn2 and Ngn2, could efficiently convert human glioma cells to functional neurons. Similar to functional neurons, ABN iN cells derived from glioma cells expressed multiple neuronal markers and voltage-gated functional membrane channel proteins, as well as fired action potentials. During the conversion process, the proliferation of human glioma cells was greatly suppressed both in vitro and in vivo. Further the sorted iN cells did not grow over time if they were successfully induced and obtained from ABN infected human glioma cells.
Human glioma cells have relatively positive membrane potential (−50 to −10 mV), which is a property of actively proliferating cells 
. A high chloride-channel expression level is correlated with malignancy of high grade glioma, and it plays critical role in glioma migration and proliferation 
. In our study, the control glioma cells had a mean membrane potential of −58.6±1.58 mV (n
4), and expressed chloride channel-like conductance. At an early stage of induction (23 days), although the sodium currents density was too low to fire action potentials, the changes of the membrane intrinsic properties (input resistant and membrane potential) and the expression of multiple neuronal makers (Tuj1, Map2, Neurofilament, NeuN and synapsin) indicated that glioma cells were truly altered to another type of cells. Further, at later stage of induction (33 days), the sodium currents density nearly tripled to burst repetitious action potentials. Thus, the combination of three transcription factors Ascl1, Brn2 and Ngn2 indeed drove human glioma cells to become neurons from mitotic cells.
The major challenge in our glioma cell conversion system was the possibility of neurons or neural progenitors being present in the starting material. However we did not detect neurons in our starting glioma cells when screening for neuronal markers. Furthermore, we did not detect neurons after prolonged culture of glioma cells with neuron medium. To carefully exclude the presence of neural progenitors in our glioma cells, CD133-negative cells population was sorted and used as the starting cell population for the induction and the result showed there was similar conversion efficiency with unsorted cells. These results suggested that neurons were induced directly from glioma cells with high efficiency.
It should be noted that the iN cells conversion efficiency is higher (20–40%) from primary human glioma cells than that of iN cells from primary human fibroblasts (less than 5%) in other studies 
. The induction time from human glioma cells is short (less than 2 weeks) whereas the induction process from human fibroblasts to neurons takes longer. These results indicate that glioma cells might maintain more active transcriptional networks and are more easily induced to become neurons. Future studies should focus on increasing the induction efficiency and optimizing the culture conditions for neuronal maturation.
Although several studies have shown that neuronal transcription factors inhibit glioma proliferation 
, it is not known whether transcription factors induce glioma cells to become neurons to dramatically affect cell growth. In the proliferation studies, ABN-transfected glioma cells grew slowly and formed fewer colonies. Importantly, we observed that the iN cells from human glioma cells did not grow completely. Additionally, if the glioma cells were not completely induced to neurons, ABN neuronal transcription factors inhibited glioma cell growth rate (Figure S5
). After analyzing cell cycle, it seemed that glioma cells were arrested at G0/G1 phase after infection with ABN. Moreover, ABN neuronal transcription factors inhibited glioma cells growth and prolonged nude mouse survival. Future studies will be necessary to determine whether there is a difference of conversion between different grades of glioma. It will be interesting to determine whether glioma cells can be converted to neurons by direct injection of lentiviral expression viruses in vivo in an effort to reduce tumorigenicity. Although the conversion efficiency from primary human glioma cells to neurons is high (around 20–40%), it should be noted there are remaining un-converted glioma cells. The ultimate goal would be to convert all glioma cells to neurons. Therefore, further studies will be needed to increase the conversion efficiency. Also, it remains to be determined whether the converted neurons would cause side-effect, including interfering with the neural circuitry and function in the brain, or connecting to wrong target cells.
In summary, our studies indicate that human glioma cells can be induced to become neurons (iN cells), and that iN cells express multiple neuronal specific proteins and fired action potentials. During the glioma-to-neuron conversion process, the proliferation rate of glioma cells is inhibited and iN cells that originated from glioma cells could not proliferate over time. This induction approach from glioma cells to neurons may provide clues developing targeted strategies to treat gliomas.