Figure 1	Return to article
	Fig. 1 Egr1 overexpression and loss of Nab2 expression in prostate tumors from CR2-T-AG transgenic mice. a, Immunohistochemical analysis shows overexpression of Egr1 in a PIN lesion relative to the adjacent benign epithelium (arrowheads). Antibody reactivity was visualized using DAB substrate (brown). b, Double labeling with antibody against T antigen (green, FITC label) and Nab2 (red, Cy3 label) using immunoflourescence. Most T antigen-expressing cells in the PIN lesion are negative for Nab2 expression. An occasional cell is seen expressing both markers (arrow). c-f, Expression of Nab2 in neuro-endocrine cells from CR2-hGH transgenic mice. Double labeling showing a neuro-endocrine cell expressing hGH (green) and Nab2 (red). Images were taken through several focal planes at 1-µm intervals with a confocal microscope. At each focal plane, the two markers are co-expressed by the same cell (arrows). Four representative images are shown in c-f.
Figure 2	Return to article
	Fig. 2 Kaplan-Meier survival analysis of Egr1-deficient CR2-T-Ag mice. Differences in survival were highly significant between CR2-T-Ag/Erg1-/- (, n=40) and either CR2-T-Ag/Erg1-/- (O, n=58) or CR2-T-Ag/Erg1+/+ (, n=34) mice. P=0.0001 for CR2-T-Ag/Erg1-/- relative to CR2-T-Ag/Erg1+/- and P=0.0006 for CR2-T-Ag/Erg1-/- versus CR2-T-Ag/Erg1+/+ mice by log rank test. Differences between CR2-T-Ag/Erg1+/+ and CR2-T-Ag/Erg1+/- mice were not statistically significant (P=0.08).
Figure 3	Return to article
	Fig. 3 MRI analysis of tumor incidence and growth rate in transgenic mice. a, Plot of number of mice (%) with detectable tumor by MRI (5-µl volume cutoff) by age. , Control CR2-T-Ag/Egr1+/+ and CR2-T-Ag/Egr1+/- (CTRL) mice, n=23. , KO CR2-T-Ag/Egr1-/- mice, n=16). P<0.05 for all time points for CTRL compared with CR2-T-Ag/Egr1-/-. b, Mean tumor doubling time in transgenic mice as measured by MRI in CR2-T-Ag mice with the indicated Egr1 genotypes, c and d, Example of prostate tumor growth detected by MRI in the same mouse scanned at 23.5 and 25.5 wk.
Figure 4	Return to article
	Fig. 4 Analysis of tumor intiation in transgenic mice. a and b, Expression of T antigen detected by immunofluorescence in a tumor from a 25-wk old CR2-T-Ag/Egr+/+ mouse (a) and in the prostate of its CR2-T-Ag/Egr-/- litter-mate with PIN (b). Original magnifications: x100. c and d, Prostates from 15-wk old CR2-T-Ag/Egr+/+ (c) and CR2-T-Ag/Egr-/- (d) littermates showing examples of PIN lesions (arrows). Original magnifications: x400. e, Quantitative analysis of PIN development. Prostates from control CR2-T-Ag/Egr+/+ and CR2-T-Ag/Egr+/- (CRTL, ) or CR2-T-Ag/Egr-/- (KO, ) mice at 15, 25 or 35 wk were examined for the presence of foci of PIN and the number of PIN foci/section/mouse determined. For each point, n=4-5 mice (*) At 35 wk, most of the prostates from control mice have been taken over by tumor; thus PIN counts were not possible in these samples.
Figure 5	Return to article
	Fig. 5 Altered patterns of gene expression in CR2-T-Ag/Egr-/- tumors. a-c, In situ hybridization analysis shows evidence of TGF-ß1 expression in tumor (a, arrows) and PIN lesion (b, arrows) from a CR2-T-Ag/Egr+/+ mouse relative to the benign-looking glands labeled "B". In contrast, lesions from a CR2-T-Ag/Egr-/- mouse (T) show low expression of TGF-ß1, d-f PDGF-A expression is evident in tumor (d, arrows) and high-grade PIN lesion (e, arrows) from a CR2-T-Ag/Egr+/+ mouse, but not a tumor (T) from a CR2-T-Ag/Egr-/- mouse (f). Original magnifications: a,c,d and f, x200; b and e x400. g Quantitative RT-PCR analysis of TGB-ß1 and PDGF-A expression. The expression levels of the indicated genes were determined in tumors from CR2-T-Ag/Egr+/+ and CR2-T-Ag/Egr+/- (, CTRL) or CR2-T-Ag/Egr-/- mice (, KO). Each bar represents a tumor from a different mouse. Relative expression levels for each gene were normalized to the level of glyceraldehyde 3-phosphate dehydrogenase (GAPDH) expression, and the results are expressed relative to the expression level in one of the knockout tumors, which is set at 1. The mean values with standard errors for an experiment performed in triplicate are shown.
Figure 6	Return to article
	Fig. 6 Nab2 represses Egr1-mediated gene activation in prostate carcinoma cells. LAPC4 cells were infected with adenovirus constructs expressing Egr1 and Egr1-1293F, singly or in combination with a Nab2-expressing adenovirus. RT-PCR was used to determine the expression levels of the endogenous TGF-ß1 gene relative to the expression level of GAPDH, and then normalized to the level found in uninfected LAPC4 cells, whcih was set as 1. The means and standard errors from a typical experiment performed in triplicate are shown.
Figure 7	Return to article
	Fig. 7 Lack of Egr1 impairs tumorigenesis in the TRAMP model of prostate cancer. The incidence of microscopically evident invasive carcinoma >1 mm (a), or grossly evident tumor (b) was determined in TRAMP mice of the indicated ages and genotypes. : control TRAMP/Egr1+/+ mice, : TRAMP/Egr1-/- mice mice. The numbers of mice used for the analysis at each time point are indicated in parentheses. * P<0.05, ** P<0.01 by Fisher exact test t None of the mice examined had tumor at this point.