Morphological changes in response to retinoic acid treatment
Five neuronal-like cell lines were tested for their capacity to mimic early events of neuronal determination and differentiation taking place during development or adult neurogenesis: D283 medulloblastoma (human medulloblastoma), DAOY (human medulloblastoma), NTERA-2 (human teratocarcinoma), PC-12 (rat pheochromacytoma), SK-N-SH (human teratocarcinoma). Additionally, HeLa cells (human adenocarcinoma) were chosen as non-neuronal control cell line. To differentiate cell lines towards a neuronal phenotype, retinoic acid (10 μM) was added to the cultures.
Figure shows the morphological appearance of cell lines in response to retinoic acid treatment. Hence, after 28 days of differentiation, NTERA-2 cells morphology showed pronounced reorganization characterized by the formation of cell clusters and processes creating a complex cellular network (Fig and ). In addition, D283 cell lines adopted a flattened morphology and showed increased cell death during differentiation treatment (Fig and ). Finally, minor or no morphological responses could be detected for DAOY, HeLa, PC-12 or SK-N-SH cells following retinoic acid treatment.
Figure 1 Morphological changes in response to retinoic acid treatment. Cell lines grown in standard conditions (A-F) or treated with retinoic acid (G-L). Only NTERA-2 cells (A vs. G) and D283 cells (B vs. H) showed a morphological response to retinoic acid. PC12 (more ...)
PC-12 cells are known to undergo neuronal morphological differentiation in the presence of nerve growth factor (NGF). To confirm the potential of our PC-12 cultures to adopt a neuronal morphology upon differentiation, we maintained them four weeks in the presence of NGF (50 ng/ml). As previously described, PC12 cells underwent an extensive process outgrowth in response to NGF (Fig inset).
Induction of early neuronal marker expression in response to retinoic acid
To examine if morphological changes observed in response to retinoic acid correlated with expression of genes induced during neuronal early differentiation, expression of DCX was assessed by immunocytochemistry. Figure shows the absence of DCX expression in undifferentiated cells from the six cell lines investigated. Following differentiation however, robust expression of DCX could be detected in a large fraction of the NTERA-2 cells, especially those forming the cell clusters (Fig ). Expression could also be detected at a low level in differentiated D283 cells (Fig ). Moreover, most neuronal-differentiated NTERA-2 cells expressing DCX also induced expression of the neuronal-specific microtubule-binding protein Map2, hence revealing some degrees of neuronal maturation (Fig ). In contrast, only a few DCX-expressing differentiated D283 cells did express Map2. Finally, Map2 expression could be detected in differentiated PC12 cells and rarely in HeLa cells as well, however, DCX expression could never be detected in these cell lines.
Figure 2 Induction of DCX expression in response to retinoic acid treatment. Cell lines grown in standard conditions did not express DCX (green) (A-F), whereas following retinoic acid treatment NTERA-2 cells (G) and to a lower extent D283 cells (H) expressed DCX. (more ...)
Induction of DCX expression in response to retinoic acid
Induction of DCX mRNA expression following retinoic acid treatment was further quantified in total RNA from each cell lines either in a proliferative state or following differentiation. In agreement with observation made by immunocytochemistry, undifferentiated cell lines did not express DCX, or did only in trace amounts (Fig ). On the other hand, following retinoic acid-induced differentiation, levels of DCX mRNA in NTERA-2 cells were increased approximately 16 times. Likewise, DCX mRNA levels were approximately 12 times more elevated in differentiated D283 cells then in the non-differentiated cells. The absolute DCX mRNA levels in D283 cells, however, were roughly half of those found in NTERA-2 cells. Finally, DCX mRNA expression was not detected in the other differentiated cell lines.
To examine the induction of DCX expression in respect to other neuronal genes, we followed the expression of NeuroD1, DCX and neuron-specific enolase (NSE) at various differentiation time points, i.e. proliferative state and following 3 days, 7 days, 14 days, 21 days and 28 days of retinoic acid treatment. As shown in figure , retinoic acid treatment rapidly and transiently induced high levels of NeuroD1 expression, a neurogenic transcription factor [9
]. As NeuroD1 expression declined between day 14 and 21 of differentiation, expression of DCX mRNA was induced and further increased until the last time point at 28 days. Finally, expression of NSE mRNA, an enzyme found in mature neurons [10
], was induced only after the third week of differentiation. Hence, our differentiating NTERA-2 cultures expressed a key neurogenic factor, a neuronal precursor marker, and a mature neuronal marker in successive phases. This suggests that differentiation of NTERA-2 cells mimicked faithfully the various phases of the neuronal determination and differentiation program also induced in neural stem cells.
Western blot analysis substantiated the induction of DCX expression measured at the mRNA level. Whereas, very low levels of DCX protein expression could be measured in proliferating NTERA-2 cells, strong expression could be detected in differentiated NTERA-2 cells (Fig ). In addition, low levels of DCX protein expression could be detected in differentiated D283 cells (Fig ), in agreement with the lower abundance of DCX mRNA measured in these cells (Fig ). The presence of DCX protein could not be documented in the other cells lines, notwithstanding their differentiation status.
Expression of several neural markers in differentiating NTERA-2 cells was scrutinized in more details by immunocytochemistry at various time points. As shown in figure , within the first 7 days of exposition to retinoic acid, the expression A2B5, a neural cell surface antigen, was induced in a large proportion of the cells, in agreement with previous report [8
]. Expression of nestin, an intermediate filament expressed in neural stem cells, was found in a high percentage of NTERA-2 cells at all time points. Nestin expression, however, decreased slowly during the course of differentiation. In contrast, NG2, an antigen present on the cell surface of glial precursors, could not be detected in our NTERA-2 cultures at any time point.
Expression of neuronal markers became significant after 3 weeks of differentiation. Hence, expression of βIII-tubulin, DCX and Map2 were strongly induced and following 4 weeks of retinoic acid treatment, was found in more then 90% of the cells present in our NTERA-2 cultures. Expression of GFAP was detected in a minor fraction of the NTERA-2 cells at every time point. Finally, expression of NeuN, a marker found in more mature neurons, could not be detected using our differentiation paradigm.
In vitro model of neuronal determination
This capacity of differentiating NTERA-2 cells to induce expression of the DCX gene was further exploited to establish an in vitro
reporter system for neuronal early differentiation events. We recently reported that a genomic fragment of 3,5 kb of the human DCX gene drove expression of reporter genes in neuronal precursors and young neurons in vitro
and in vivo
]. We therefore generated stable NTERA-2 cell lines bearing the EGFP or the firefly luciferase reporter gene under the control of the DCX promoter sequences (NTERA-2DCX-EGFP
). In parallel, we also generated transfected HeLa cell lines (HeLaDCX-EGFP
) as non-neuronal controls. Integration of the reporter construct was confirmed by PCR on the genomic DNA (data not shown).
Whereas undifferentiated NTERA-2DCX-EGFP and HeLaDCX-EGFP did not express the EGFP reporter, retinoic acid treatment strongly and specifically induced the expression of EGFP in the NTERA-2DCX-EGFP cells but not in the HeLaDCX-EGFP cells (Fig ). Immunocytochemistry further confirmed that EGFP-expressing cells also expressed DCX, and frequently Map2 as well (Fig ).
Figure 3 Activation of DCX-promoter-reporter constructs in differentiating NTERA-2 cells. Upon differentiation using retinoic acid, NTERA-2DCX-EGFP clones (B and F), but not the HeLaDCX-EGFP clones, could induce the expression of the DCX-promoter-EGFP reporter (more ...)
The induction of the reporter gene in differentiating NTERA-2 cells was examined at the mRNA and protein levels. In agreement with the strong induction of the endogenous DCX mRNA expression following differentiation of NTERA-2DCX-EGFP cells, expression of the EGFP reporter mRNA was observed to be strongly induced (Fig ). In contrast, HeLaDCX-EGFP did not upregulated the expression of the EGFP reporter in response to retinoic acid treatment. The potent induction of the reporter's expression in NTERA-2 cells could also be documented by western blotting (Fig ). Again, as expected from the mRNA measurements, expression of the reporter was not induced in HeLaDCX-EGFP cells, underscoring the neuronal specificity of the reporter system.
Neurogenin 1 and neurogenin 2 (Ngn2) are transcription factors of the basic helix-loop-helix (bHLH) family expressed in the earliest steps of mammalian cortical neurogenesis. They initiate a cascade of gene activation leading to the expression of terminal neuronal differentiation genes [13
]. It was recently reported that mouse Ngn2 could bind directly on the DCX promoter and activate it [15
]. The capacity of our reporter system to respond to neurogenic factors was tested using a mouse Ngn2-encoding vector to transiently transfect the NTERA-2DCX-luci
cell were used in parallel as a non-neuronal control. As shown in figure , brief expression of mouse Ngn2 for as little as four days sufficed to activate DCX-promoter driven expression of the luciferase reporter gene in NTERA-2 cells, but not in HeLa cells.