Regulated expression of ATF5 is required for the progression of neural progenitor cells to neurons

Abstract

An important milestone in brain development is the transition of neuroprogenitor cells to postmitotic neurons. We report that the bZIP transcription factor ATF5 plays a major regulatory role in this process. In developing brain ATF5 expression is high within ventricular zones containing neural stem and progenitor cells and is undetectable in postmitotic neurons. In attached clonal neurosphere cultures ATF5 is expressed by neural stem/progenitor cells and is undetectable in tau-positive neurons. In PC12 cell cultures nerve growth factor (NGF) dramatically downregulates endogenous ATF5 protein and transcripts, whereas exogenous ATF5 suppresses NGF-promoted neurite outgrowth. Such inhibition requires the repression of CRE sites. In contrast, loss of function conferred by dominant-negative ATF5 accelerates NGF-promoted neuritogenesis. Exogenous ATF5 also suppresses, and dominant-negative ATF5 and a small-interfering RNA targeted to ATF5 promote, neurogenesis by cultured nestin-positive telencephalic cells. These findings indicate that ATF5 blocks the differentiation of neuroprogenitor cells into neurons and must be downregulated to permit this process to occur.

Figure 1.

NGF downregulates ATF5 protein in PC12 cells: reciprocal relationship with neurite outgrowth. A, Time course of the effect of NGF treatment on ATF5 protein expression in PC12 cells. Cells were exposed to NGF for the indicated times, and 135 μg of whole cell extracts was subjected to Western immunoblotting with anti-ATF5 and, after stripping, with anti-ERK to normalize for loading. The numbers at left indicate the positions of molecular weight markers (in kDa). Comparable results were achieved in three independent experiments. B, Comparison of the kinetics of NGF-dependent downregulation of ATF5 expression and promotion of neurite outgrowth. The relative levels of ATF5 expression, as determined by densitometry and normalized to levels of ERK protein in the same sample, are reported in arbitrary units. Proportions of cells bearing neurites of length at least twice the diameter of the cell body were determined in the same cultures by scoring at least 200 cells per time point.

Figure 2.

Overexpression of ATF5 represses neurite outgrowth in PC12 cells, whereas NTAzip-ATF5 accelerates neuritogenesis. A, Detection and NGF response of PC12 cells expressing exogenous ATF5. PC12 cells were transiently transfected with pCMS-eGFP (a, b) or pCMS-eGFP expressing FLAG-tagged ATF5 (c, d). At 2 d after transfection the cultures were treated with NGF. At 5 d after transfection (i.e., after 3 d of NGF exposure) the cells were fixed and costained with rabbit anti-GFP (a, c) and mouse anti-FLAG antibody (b, d), with detection by FITC (GFP) and rhodamine-conjugated secondary antibody (FLAG-ATF5). Scale bar, 50 μm. B, Quantification of the effects of exogenous ATF5 and of NTAzip-ATF5 on NGF-promoted neurite outgrowth. PC12 cells were transiently transfected with pCMS-eGFP without insert or expressing FLAG-tagged ATF5 or FLAG-tagged NTAzip-ATF5. At 2 d after transfection the cultures were treated with NGF. Cultures were fixed at the indicated times after commencement of NGF exposure and were immunostained with anti-GFP and anti-FLAG as above. Transfected cells (positive for FLAG and/or GFP staining) were assessed for the presence or absence of neurites. The proportions of transfected cells bearing neurites are reported ± SEM; n = 3 cultures (at least 300 transfected cells assessed per culture). Comparable results were achieved in four additional independent experiments, and in all cases (including the data shown) ANOVA analysis indicated p < 0.05 at the 72 hr point of NGF treatment for eGFP versus ATF5. C, NTAzip accelerates NGF-promoted neurite outgrowth. Cultures were transfected, treated, and assessed as in B at 24 hr after NGF exposure. Values represent the mean ± SEM for the results of four independent experiments. In each experiment the data were normalized to the percentage of neurite-bearing cells transfected with pCMS-eGFP. The average percentage of such cells was 10.6 ± 3.7 (NTAzip vs eGFP, p < 0.02; Student's t distribution test).

Figure 3.

ATF5 is expressed differentially in the ventricular zones of E12–E15 rat brain. Aa, Ab, Expression of ATF5 message in developing rat brain. In situ hybridization was performed by using an ATF5 antisense probe in sagittal sections of E15 rat brain. Shown is the area around the fourth ventricle (a) and the telencephalon (b). There was no positive signal with a control ATF5 sense probe. c–f, Expression of ATF5 protein in coronal sections of E12 (c, d) and E14 (e, f) rat telencephalon. c, Staining with preimmune serum. d–f, Costaining with anti-ATF5 (red) and anti-tubulin β (class III; TUJ1 antibody; green). Arrows indicate staining of ATF5 in the ventricular zone (VZ); CX, cortex. Scale bar, 100 μm. B, High-power confocal images of reciprocal expression of ATF5 (red) and tubulin β (class III) in coronal sections of E14 rat telencephalon. Immunochemical staining was performed as in A. Images showing the ventricular zone (a) and cortex (b) are from the same section and were photographed in the same confocal z-plane section (1.3 μm). Arrowhead shows a migratory cell undergoing a transition from a progenitor to a neuron by exhibiting both ATF5 and tubulinβ (class III) staining. Colocalization was confirmed by yz and xz confocal images. Scale bar, 20 μm.

Figure 4.

Reciprocal expression of ATF5 and tubulin β (class III) in E17 rat brain. A–C, Expression of ATF5 (red) and tubulin β (class III; green) in the area of the anterior (A–C) and posterior (D–F) lateral ventricles of the E17 rat brain. Immunohistochemical staining was performed as in Figure 3 and as discussed in Materials and Methods. Scale bar, 100 μm.

Figure 5.

ATF5 is expressed in neural stem cells and progenitor cells, but not in mature neurons in attached neurosphere cultures. Attached clonal neurosphere cultures were established from the subventricular zone and hippocampal dentate gyrus of newborn mouse brain and were maintained as described in Materials and Methods. Cultures were fixed and costained as follows: A, ATF5 (red) and AC133 (green), a stem cell marker. Thick arrows show examples of nuclear staining; thin arrows show cytoplasmic staining. B, ATF5 (red) and nestin (green), a marker for neural progenitor cells. Arrows indicate nuclear staining. C, D, ATF5 (red) and NF-M(green), a marker for the neuronal lineage. Arrows show nuclear staining in C and cell body in D. E,F, ATF5 (red) and anti-tau (green), a neuronal marker. Comparable results were achieved in 10 independent experiments. Arrows show neurons at the periphery of the cultures; arrowhead shows stem and neural progenitor cells at the center of the culture. Stained cells were examined and photographed by confocal microscopy. Scale bars: A, 20 μm; B–F, 50 μm.

Figure 6.

ATF5 represses, and NTAzip-ATF5 promotes, neurite outgrowth and expression of neuronal markers in neural progenitor cells. A, Cultured E14 telencephalic cells were transiently transfected with pCMS-eGFP containing no insert (empty vector), FLAG-ATF5, or NTAzip-ATF5. At 3 d after transfection the cultures were fixed and coimmunostained for GFP and either nestin or tubulinβ (class III; TUJ1 antibody). Transfected cells (GFP ⁺) were assessed for the presence of neurite-like processes as well as for coexpression of the indicated markers. Values represent the mean ± SEM for three cultures in which at least 300 transfected cells were evaluated per culture. Comparable results were achieved in four independent experiments (ANOVA analysis of transfected cells. Total cells: nestin/eGFP alone vs nestin/ATF5, p < 0.001; TUJ1/eGFP alone vs TUJ1/ATF5, p < 0.05; nestin/eGFP alone vs nestin/NTAzip and TUJ1/eGFP alone vs TUJ1/NTAzip, no significant difference. Process-bearing cells: TUJ1/GFP alone vs TUJ1/ATF5, p < 0.05; nestin/eGFP alone vs nestin/NTAzip and TUJ1/eGFP alone vs TUJ1/NTAzip, no significant difference). B, Cultured E14 telencephalic cells were infected with retroviruses expressing eGFP or FLAG-ATF5 and eGFP. At 1 week after infection the cultures were fixed and assessed as in A as well as for NF-M expression. Comparable results were achieved in three independent experiments (ANOVA analysis. Total cells: nestin/eGFP alone vs nestin/ATF5, p < 0.001; TUJ1/GFP alone vs TUJ1/ATF5, p < 0.01; NFM/eGFP vs NFM/ATF5, p < 0.001. Process-bearing cells: TUJ1/GFP alone vs TUJ1/ATF5, p < 0.001; NFM/GFP alone vs NFM/ATF5, p < 0.01; nestin/eGFP alone vs nestin/ATF5, no significant difference. TUJI vs NFM, no significance both with eGFP alone and +ATF5). C, E14 telencephalon cells were infected with retroviruses expressing eGFP, eGFP and FLAG-ATF5, or eGFP-FLAG–NTAzip-ATF5. At 4 d after infection the cultures were fixed and evaluated as in A. Comparable results were achieved in two independent experiments (ANOVA analysis. Nestin/GFP alone vs nestin/ATF5, p < 0.001; total and process-bearing cells: TUJ1/eGFP alone vs TuJ1/ATF5, p < 0.01; TUJ1/GFP alone vs TUJ1/NTAzip, p < 0.05). D, Cultured E14 telencephalic cells were transiently transfected with pCMS-eGFP with or without ATF5 siRNA. At 4 d after transfection the cultures were fixed and coimmunostained either for GFP and TUJ1 antibody or with GFP and ATF5 antiserum. Transfected cells (GFP ⁺) were assessed for the presence of the neuronal marker tubulinβ (class III; TUJ1) or ATF5. Values represent the mean±SEM for six cultures in which atleast 300 transfected cells were evaluated perculture. Comparable results were achieved in three independent experiments (two with E14 telencephalon cells cultured with serum plus EGF and FGF2 and one experiment with only serum) (ANOVA analysis. TUJ1/eGFP alone vs TUJ1/ATF5 siRNA, p < 0.001; ATF5/GFP alone vs ATF5/ATF5 siRNA, p < 0.001). E, ATF5 suppresses NT3-promoted neuronal differentiation. E15 telencephalon cells were infected with retroviruses expressing eGFP, eGFP and FLAG-ATF5, or eGFP-FLAG–NTAzip-ATF5, all ± NT3. At 4 d after infection and maintenance of ± NT3 treatment the cultures were fixed and evaluated as in A for eGFP and TUJ1 expression. Comparable results were achieved in two independent experiments (ANOVA analysis. -NT3/eGFP alone vs +NT3/GFP alone, p < 0.001; -NT3/eGFP alone vs -NT3/ATF5, p < 0.05; +NT3/eGFP alone vs +NT3/ATF5, p < 0.001; -NT3/eGFP alone vs -NT3/NTAzip, p < 0.001; +NT3/GFP alone vs +NT3/NTAzip, no significant difference).

Figure 7.

NTAzip-ATF5 and VP16-CREB reverse ATF5-promoted repression of CRE-mediated gene expression and of neurite outgrowth. A, PC12 cells were cotransfected with pGl3-CRE-luciferase, pcDNA-LacZ, and 1 μg/culture of pCMS-eGFP expressing no insert (empty vector), FLAG-ATF5, FLAG-NTAzip-ATF5, or VP16-CREB. The cultures also were exposed to NGF for 2 d before and during the time of transfection (for a total of 3 d), during the time of transfection (1 d), or during the last hour before harvesting. At 1 d after transfection the cells were harvested and assessed for luciferase expression and LacZ activity (β-gal). Values represent mean normalized CRE-luciferase activity (in arbitrary units) ± SEM (n = 3). Comparable results were achieved in three independent experiments [Student's t distribution test. Empty vector (eGFP alone) vs VP16-CREB at all times, p < 0.001; GFP alone vs ATF5, p < 0.033 by day 3]. B, PC12 cells were cotransfected with pGl3-CRE-luciferase, pcDNA-LacZ, and the indicated combinations of pCMS-eGFPexpressing no insert(GFP),FLAG-ATF5(ATF5),FLAG-NTAzip-ATF5(AZIP), or VP16-CREB. The latter vectors were used each at 0.5 μg/culture, and empty vector was added as needed to bring the total DNA level to 1 μg/culture. Cultures were harvested 1 d later for assay of luciferase expression and LacZ activity (β-gal). Where indicated, NGF was added to the medium 1 hr before harvesting. Values represent mean normalized CRE-luciferase activity (in arbitrary units) ± SEM (n = 6), with data pooled from two independent experiments (Student's t distribution test. -NGF eGFP alone vs ATF5, p < 0.003; eGFP alone vs NTAzip, p < 0.0003; eGFP alone vs ATF5/NTAzip, no significant difference; eGFP alone vs VP16-CREB and VP16-CREB/ATF5, p < 0.0001. +NGF eGFP alone vs ATF5, p < 0.0001; eGFP alone vs NTAzip, p < 0.02; eGFP alone vs ATF5/NTAzip, p < 0.02; eGFP alone vs VP16-CREB and VP16-CREB/ATF5, p < 0.0001). C, PC12 cells were cotransfected with the indicated constructs, and NGF was added to the medium 2 d later. Transfected cells (identified for eGFP) were assessed for neurite outgrowth at the indicated times. Values represent the mean ± SEM of results for three cultures in which at least 300 transfected cells were scored per culture. Comparable results were obtained in two independent experiments (ANOVA analysis after 72 hr of NGF-treatment. eGFP alone vs ATF5, p < 0.001; eGFP alone vs NTAzip-ATF5, NTAzip/ATF5, VP16-CREB, or VP16-CREB/ATF5, no significant difference).