Figure 1.
Ischemia-mediated induction of ZEB1 protein in cortical neurons.
Specificity of the ZEB1 antibody used in these studies is demonstrated in Figure S1; Panel I. ZEB1 protein levels increase 8-fold in primary cultures of cortical neurons exposed to OGD. Photomicrographs of E16.5 primary cortical neuronal cultures exposed to normoxia (a,c), or OGD (b,d) for 90 minutes and immuno-stained for ZEB1 protein; (e), representative western analysis of total protein isolates, Z = ZEB1, A = β-actin, N = normoxia; f, densitometry of the bands from e, normalized to β-actin; four different replicate samples of cells (isolated on different days) were scored and the average+/−the S.E.M. is shown; Panel II. ZEB1 protein levels increase over 10-fold in the P7 rat ischemic cortex. Representative photomicrographs of immunostained coronal sections comparing the ischemic (a,c) and the contra-lateral (b,d) sides of the same sections, harvested 3 hours following administration of unilateral permanent FCI; the increase is largely confined to the nucleus of cells located mainly in the inner and outer pyramidal cell layers of the cortex; (e), representative western analysis of total cortical protein isolates; f, densitometry of the bands from (e), normalized to β-actin; N = normal brain, S = sham, C = contra-lateral, I = ischemic; Graph depicts the average of at least three separate experiments, and the average+/−the S.E.M. is shown; Panel III. Cells up-regulating ZEB1 in the ischemic cortex co-localize with the neuronal marker NeuN. Representative staining of 8 µm coronal sections from P7 rat pup brains harvested 12 hrs. post-FCI. Scale bars: panel I: a,b, 250 µm; c,d, 100 µm; panel II: a,b, 1 mm; c,d, 50 µm; panel III: 50 µm. ** = P<0.01, by Student's t-test.
Figure 2.
Primary cortical cultures derived from ZEB1 KO or wt E16.5 embryos were transfected with either a GFP expression vector (control, White Bars), or a plasmid expressing a full-length ZEB1 cDNA fused to GFP (Blue Bars).
Sixteen hours later, cells from both groups were subjected to OGD for 6 hrs, fixed under hypoxia, the nuclei were Hoechst-stained and then scored (in a blinded fashion) for presence of either a normal vs pyknotic/mis-shapen/condensed morphology. Over twice as many pyknotic nuclei were scored in ZEB KO cells as compared to their wt counterparts. Insert shows a representative western blot verifying the absence of ZEB1 protein in neurons derived from the KO animal. Lower bands are the loading control β-actin. Over-expressing ZEB1 cDNA conferred a similar increase in resistance above wt (Blue Bars) on neurons of either genotype. Four different replicate samples of cells (isolated on different days) were scored and the average+/−the S.E.M. is shown; Z, ZEB1; A. β-actin; * = P<0.05, by Student's t-test.
Figure 3.
Over-expressed ZEB1 protects primary cortical neurons from a subset of pro-death insults.
E16.5 primary cortical neurons transfected with GFP or ZEB1 fused in frame to GFP (for a 30 hr period) were subjected to the following pro-death agents for the indicated times: Veh., DMSO; H2O2, hydrogen peroxide (6 hrs); Glu, glutamic acid (18 hrs); NO, nitric oxide (20 hrs); TNF-a, tumor necrosis factor alpha (18 hrs); OGD, oxygen-glucose deprivation (12 hrs); IR, ionizing radiation (6 hrs). At indicated time points, cells were fixed, processed and their ability to resist cellular degradation/death was assessed using two distinct markers of cellular degradation or death. Representative fluorescence images from these analyses are shown in Figures S3 (nuclear morphology) and S4 (mitochondrial membrane integrity). Top graph: in every case except for NO, at a given concentration/level of the indicated treatment, and at the time points examined, on average, less than half as many nuclei in neurons over-expressing ZEB1 showed a shrunken/pyknotic/condensed morphology compared with their GFP-expressing counter-parts. Bottom graph: in every case except for nitric oxide, at a given concentration/level of the indicated treatment, and at the time points examined, on average, over twice as many ZEB1-over-expressing neurons maintained intact mitochondrial membranes compared with their GFP-expressing counterparts. Results shown above are averaged from at least three separate experiments (cultures isolated on different days)+/−the S.E.M. ** = P<0.05 by ANOVA.
Figure 4.
p63 is rapidly induced in response to ischemia.
(A) Expression profile of p63, p73 and ZEB1 protein induction in the ischemic cortex of P7 rat pups subjected to FCI. Brains were harvested and processed for immunostaining at the indicated time points post-FCI. Colorimetric staining reveals that p63 is induced the earliest, at least one hr post-FCI. Scale bar = 500 µm. (B) Both p63 and p73 induction co-localize with that of ZEB1. Double-immuno-fluorescence staining indicates that all three of these proteins co-localize to neurons in the ischemic cortex. In all cases, staining was carried out at least three times and typical results are shown. Scale bars, a., 500 µm; b., 50 µm. (C) Western analysis of the induction of p63 and p73 isoforms compared to ZEB1. Total cellular protein was isolated from normal (N), sham operated (S – 6 hr time point), or ischemic cortexes at the indicated time points following administration of FCI. All show some degree of up-regulation in this immediate time frame, except for TAp73, which after an initial rise, declines between 3 and 6 hrs post FCI. b-actin is included as a loading control. Shown is a typical result from separate analysis of three different experiments. (D) RT-PCR analysis of temporal expression of p63 isoforms in the cortex immediately following FCI. Primer pairs (see Materials and Methods) were designed to specifically amplify the isoform classes indicated but could not distinguish whether a particular 3′ splicing variant (α β or γ) harbored the TA or the DeltaN 5′ end. While p63 TA, alpha and gamma both showed a modest increase in the six-hr period immediately following FCI, the beta isoform showed a dramatic induction. Positive controls are shown in the first two lanes in the top panel and in the first three lanes in the bottom panel. Sizes of PCR products are indicated on the right.
Figure 5.
p63 binds and transactivates through a phylogenetically conserved binding site in the proximal promoter region of the ZEB1 gene in response to ischemia.
(A) Schematic of the promoter regions of human (H) mouse (M) and rat (R) ZEB1 genes. The binding site in the rat gene is shown, with the bases differing from the p53 binding site consensus sequence shown in red. (B) Specific p63 isoforms can transactivate the ZEB1 promoter. The listed p63 isoform cDNAs were co-transfected into the p63-null cell line Saos-2 along with a luciferase reporter plasmid driven by the fragments (listed in the right panel insert) of the mouse ZEB1 promoter. The luciferase activity derived from the co-transfection of TAp63α, TAp63β and DeltaNp63β isoforms was dependent on the p53 binding site, the mutation of which abrogated this activity (gray bars); insert is a western blot derived from equal amounts of protein lysates from parallel transfected cultures using a pan-anti-p63 antibody (identity of bands follows the same order, left to right, as the transfection data). (C) Equimolar amounts of protein derived from in vitro translated p63 isoform cDNAs were used in a micro-titre-based DNA-binding assay employing oligonucleotides harboring the site shown in panel (A) (see Materials and Methods). With the exception of the DeltaNγ isoform, all bound specifically to the site in the rat promoter. Competitor oligos are identified in the panel insert.
Figure 6.
OGD-mediated ZEB1 protein induction depends significantly on p63.
(A) Primary cortical cultures from either wt or ZEB1 KO mice were transduced with lentiviral constructs expressing either scrambled or p63 SiRNA (see Materials and Methods for sequences). Western blots of protein lysates prepared at indicated time points were carried out using antibodies shown at left. Densitometry of normalized (to β-actin) expression/induction profiles are depicted graphically in (B). Results are representative of 3 separate experiments, carried out on independent primary neuronal isolates (done on different days). At the 3 hr time point, SiRNA-mediated reduction in p63 protein levels reduces OGD-mediated ZEB1 protein induction an average 60% (ZEB1 graph). The reduction in p73 protein levels in wt neurons between 3 and 6 hrs following OGD (see Figure 4, panel C, and the top panel of westerns, labeled: “wt cells+scrambled SiRNA” in this figure) is virtually eliminated in the absence of ZEB1 (TAp73 graph). Interestingly, the dramatic increase in OGD-mediated ΔNp63 protein levels (seen in the top panel of westerns, labeled: “wt cells+scrambled SiRNA”) was reduced by half in a ZEB1 KO background (see text). (C) ChIP assay using the pan-anti-p63 antibody (recognizing all isoforms in the transfection analysis, Figure 5, panel B) shows that, in the cortex, within 1 hr of FCI, p63 binds to the site depicted in Figure 5, panel A in vivo. Primers used are depicted schematically (in green) in Figure 5, panel A. Tubulin, negative control ChIP using heterologous anti-Tubulin Ab; Pos. Cont., PCR product from post-IP supernatant harboring p63 binding sites (see Methods). Error bars: S.E.M. * = p<0.01; ** = p<0.001.
Figure 7.
ZEB1 represses TAp73 in the cortex in response to ischemic insult.
(A) Agarose gel of RT-PCR analysis of cDNA synthesized from total RNA isolated at the indicated time points from either Sham-operated (S) or FCI (I) P7 cortexes. The numbers in parentheses on the left (below the indicated PCR product) denote the number of PCR cycles used to achieve linear-range amplification of each target. Shown is a representative result from one of three independent experiments (each derived from a separate litter). To control for the presence of contaminating genomic DNA, a PCR-only reaction using identical conditions was carried out using these samples (not shown). (B) For each experiment, a numerical value for relative intensity was obtained by densitometry, and after normalization to the levels of β-actin (bottom row, panel a) Sham (background) values were subtracted from the corresponding Ischemic values. The average of these results from all three RT-PCR analyses are depicted graphically as the average value+/−the S.E.M. Two hours after the administration of FCI, steady-state levels of ZEB1 mRNA (black squares) are beginning to plateau, those of TAp73 (green circles) are declining, while ΔNp73 levels are increasing. (C) ChIP analysis showing increased ZEB1 protein binding to the previously characterized silencer region within intron 1 of the human p73 gene, beginning one hour after primary cortical neurons were subjected to OGD, and increasing significantly at three hours. Results shown are representative of three separate experiments. (D) Rescue of ZEB1-mediated repression of p73 protein levels. Primary cortical cultures isolated from ZEB1 knock-out late stage embryos were transduced with a lentivirus harboring either GFP (“Control Transduc”, top) or ZEB1 fused to GFP (“ZEB1 Rescue”, bottom). The top three panels in (D) were derived from a (representative) single western blot, and the bottom three panels from a different (representative) single western blot, each the product of one of three separate experiments (cultures isolated on different days). Restoration of ZEB1 re-established the OGD-mediated reduction in TAp73 protein (see Figure 4 Panel C, Figure 6, panel A).
Figure 8.
ZEB1 protein is up-regulated in the human brain in response to ischemic insult.
Panel I. Representative staining of thin sections of cortical brain tissue derived at autopsy from patients whose cause of death was diagnosed as either ‘sepsis and prematurity” (top row) or “extensive germinal matrix hemorrhage with interventricular extension and infarction” (bottom row). General cellularity and morphology is revealed through H and E staining (a,d). Adjacent sections stained for ZEB1 protein expression shows a clear up-regulation in the developing cortical region in the patient diagnosed with hemorrhagic ischemia (e), in a pattern similar to that seen in the experimentally-induced ischemic cortex of the P7 rat pup (Figure 1, panel II). Magnification of the corresponding sections demonstrates the nuclear localization and intensity of ZEB1 staining in the ischemic tissue (f), while levels in the non-ischemic brain remain at background (c). Scale bars = a,b,d,e. = 500 µm, c,f. = 50 µm. Panel II. Chi-squared analysis of the data indicates that the association of ZEB protein induction with the diagnosis of an acute ischemic episode is highly significant.