Targeted Knockdown of Notch1 Inhibits Invasion of Human Prostate Cancer Cells Concomitant with Inhibition of Matrix Metalloproteinase-9 and Urokinase Plasminogen Activator

Notch signaling has been known for decades to developmental biologists as a key player in cell fate determination (1, 2) and tissue homeostasis by maintaining the self-renewal potential of some tissues and inducing differentiation of others (3), including formation of the prostate gland (4). In humans, notch family of transmembrane proteins consists of four receptors, Notch1 through Notch4, and five ligands, Jagged1 and Jagged2 and Delta-like ligand 1 (Dll1), Dll3, and Dll4 (5). Broadly, the modular structure of notch consists of an extracellular ligand binding domain, a hydrophobic transmembrane domain, and a notch intracellular domain. Notch receptors undergo a series of programmed proteolytic events, first by α-secretase at the extracellular surface, which leads to liberation of the extracellular fragment, and then by intramembraneous cleavage mediated by γ-secretase. Notch intracellular domain is then released from the inner surface of cell membrane and is translocated into nucleus where it activates transcription of the target genes (6). It has been shown that Notch1 receptor ligand Jagged1 is overexpressed in metastatic human prostatic tissue compared with localized prostate cancer (PCa) or benign prostatic tissue (7), implicating Notch1 in PCa progression. A recent study has shown that down-regulation of Jagged1 inhibits growth of PCa cells (8). Elevated expression of Notch1 was observed in highly metastatic PCa cell lines as compared with normal prostate epithelial cells (9). Notch1 level was reported to be elevated in malignant prostatic epithelial cells of primary and metastatic tumors of transgenic mouse model of PCa (10, 11). Recent studies show the involvement of notch signaling in cancer angiogenesis and metastasis (12–14); however, mechanisms for these effects remain unknown. In the present study, we provide evidence that Notch1 plays an important role in invasiveness of human PCa concomitant with decrease in the expression of matrix metalloproteinase 9 (MMP9) and urokinase plasminogen activator (uPA).

Cell culture. Human PCa cell lines PC3, DU-145, LNCaP, and 22Rν1 were obtained from American Type Culture Collection and cultured in RPMI 1640 supplemented with 10% fetal bovine serum and 1% penicillin-streptomycin solution. Prostate epithelial cells and their growth media were procured from Cambrex BioScience. Cells were grown in a humidified incubator containing 5% CO₂ at 37°C.

Immunohistochemistry. Human prostate tissues were obtained from the Department of Pathology, University of Wisconsin-Madison under an institutional review board approval. Immunohistochemical staining was done using an automated benchmark immunostainer (Ventana Medical Systems). Tissue sections were subjected to antigen retrieval, incubation with specific primary antibody of full-length Notch1 (Santa Cruz Biotechnology, Santa Cruz, CA), which is a rabbit polyclonal antibody raised against amino acids 20 to 150 mapping within an extracellular domain of Notch1 of human origin at a dilution of 1:50, followed by incubation with appropriate horseradish peroxidase–conjugated secondary antibody. Immunoreactive complexes were detected using 3,3′-diaminobenzidene and visualized under a Zeiss-Axiophot DM HT microscope and captured with an attached camera.

Histopathologic grading of PCa specimens. The Gleason system and the WHO grading system were used for evaluation of Notch1 expression in the prostate tissues. Prostate adenocarcinoma was first graded in Gleason patterns 2, 3, and 4. The intensity of immunoperoxidase staining for Notch1 was scored as 0 (negative), 1 (weak), 2 (moderate), and 3 (strong) based on the Gleason patterns. The primary Gleason pattern and the secondary Gleason pattern were added to arrive at a Gleason score, ranging from 6 to 8. A total of 318 core tissue samples were from 41 patients. Normal glandular tissues, high-grade prostatic intraepithelial neoplasia (HGPIN), and blood vessels were from the areas adjacent to the cancerous tissue.

Notch1 small interfering RNA transfection. Validated Notch1-specific and scrambled small interfering RNAs (siRNA) were purchased from Santa Cruz Biotechnology. The Notch1 siRNA contains the pool of three independent sequences: (a) sense 5′-CACCAGUUUGAAUGGUCAAtt-3′ and antisense 5′-UUGACCAUUCAAACUGGUGtt-3′; (b) sense 5′-CCCAUGGUACCAAUCAUGAtt-3′ and antisense 5′-UCAUGAUUGGUACCAUGGGtt-3′; and (c) sense 5′-CCAUGGUACCAAUCAUGAAtt-3′ and antisense 5′-UUCAUGAUUGGUACCAUGGtt-3′. Using electroporation (Amaxa, Inc.), cells were transiently transfected with Notch1 siRNA in concentrations ranging from 30 to 120 nmol/L and scrambled siRNA (120 nmol/L). Cells were allowed to grow further in a CO₂ incubator for 24 h and later harvested for further analysis.

Western blot analysis. Forty micrograms of protein resolved over 4% to 20% Tris-glycine gels (Invitrogen) were transferred onto nitrocellulose membranes and probed with appropriate primary antibody against Notch1 (Santa Cruz Biotech) and cleaved Notch1 (Val1744 Cell Signaling Technology). The cleaved Notch1 antibody detects endogenous levels of the cytosolic domain of Notch1 only when cleaved between Gly1743 and Val1744. This antibody does not recognize full-length Notch1 or cleaved Notch1 at other positions. Pro-MMP9 and uPA antibodies were purchased from Chemicon International, Inc., and uPAR antibody was obtained from R&D Systems. Expression levels of proteins were analyzed as described (15). Densitometric measurements of the bands in Western blot analysis were done using digitalized Scientific Software program, UN-SCAN-IT, purchased from Silk Scientific Corporation.

Gene expression analysis. Gene expression analysis was done using pathway focused Human Extracellular Matrix and Adhesion molecules oligo gene array (Superarray) containing 96 genes encoding proteins important for the attachment of cells to their surroundings. The array consisted of 96 cDNA in a 8 × 14 grid of tetraspots printed on a nylon membrane. The array was hybridized with total cellular cDNA that was reverse transcribed from total cellular RNA obtained either from PC3 cells transfected with scrambled siRNA or from Notch1 siRNA. After hybridization, membrane was developed as per manufacturer's instructions. Data were analyzed using GEArray Expression Analysis Suite software (Superarray).

In vitro chemoinvasion assay.Notch1 siRNA– or scrambled siRNA–transfected cells were resuspended in fresh culture medium and incubated in chemoinvasion chamber containing polycarbonate filter coated with Matrigel (Chemicon International) for 24 h. In the upper chamber, 30,000 cells were seeded in fetal bovine serum–free culture media and the lower chamber contained culture media containing 10% fetal bovine serum as a chemoattractant. The cells were allowed to migrate for 24 h, following which the chamber was washed with PBS and cells were visualized as per manufacturer's instruction. To quantitate the migratory cells, the invasion chamber was dipped in 10% acetic acid, and the resultant solution was spectrophotometrically read at 540 nm according to the vendor's protocol (16).

Gelatin zymography. Equal number of PC3 cells was transiently transfected with Notch1-siRNA and control siRNA for 24 h. The conditioned media were collected, concentrated using Amicon filter (Millipore) as per manufacturer's protocol, and electrophoresed (40 μg protein) under nonreducing conditions. The gelatinolytic activity of MMP9 was determined with a zymography kit (Invitrogen) as per vendor's protocol.

Luciferase assay. Cells (2 × 10⁶) were nucleofected with human MMP9 luciferase reporter plasmid (pGL3-MMP9, 1 μg), a gift from Dr. Dougles D. Boyden (M. D. Anderson Cancer Center, Houston, TX), along with 50 ng of renilla luciferase reporter plasmid pRL-TK (Promega), which was used as an internal control to normalize transfection efficiency and varying concentrations of Notch1 siRNA (30, 60, and 120 nmol/L). In parallel, cells were also transfected with empty pGL3 reporter vector and a scrambled siRNA (120 nmol/L) with no validated target sequence in the human genome. After nucleofection, 30,000 cells were distributed per well of a 24-well plate and allowed to grow for another 24 h. Dual luciferase assay reagent kit was procured from Promega and luciferase activity was measured according to the manufacturer's protocol.

MMP9 ELISA assay. Twenty-four hours posttransfection, the cell culture media were collected and used for quantifying MMP9 levels by using an MMP9-specific ELISA kit from Amersham Bioscience and following the vendor's protocol.

Semiquantitative PCR. PCR reactions were carried out using forward and reverse primer combinations for uPA (forward 5′-gtgaagaagggcgtccaaag-3′; reverse 5′-tcggcagtcaatgaggaaagt-3′), uPAR (forward 5′tctatccggagcagctgaaaa-3′; reverse 5′-cgtgtagacgcctggcttgt-3′), and glyceraldehyde-3-phosphate dehydrogenase (GAPDH; forward 5′-aatcccatcaccatcttccaggag-3; reverse 5′-gcattgctgatgatcttgaggctg-3). PCR reaction standardization kits were obtained from Epicentre. The cDNA was amplified with an initial denaturation at 94°C for 2 min followed by sequential cycles of denaturation at 94°C for 45 s, annealing at 59°C for 45 s, and extension at 72°C for 1 min for 30 cycles, with a final extension at 72°C for 7 min.

Quantitative reverse transcription-PCR. Real-time amplification of MMP9, uPA, and uPAR was done from 2 μL cDNA prepared from 2 μg of total RNA. The following primers were used: MMP9, forward 5′-atttctgccaggaccgcttctact-3′ and reverse 5′-cagtttgtatccggcaaactggct-3′; uPA, forward 5′-tcacaccaaggaagagaatggcct-3′ and reverse 5′-aatgacaaccagcaagaaagcggg-3′; uPAR, forward 5′tgtggctcatcagacatgagctgt-3′ and reverse 5′-ttgttgtggaaaccattggagccc-3′; and β-actin, forward 5′-atctggcaccacaccttctacaatgagctgcg-3 and reverse 5′-cgtcatactcctgcttgctgatccacatctgc-3. PCR reaction standardization kits were obtained from Epicentre. The cDNA was amplified with an initial denaturation at 95°C for 10 s followed by sequential cycles of denaturation at 94°C for 45 s, annealing at 55°C for 10 s, and extension at 72°C for 1 min for 30 cycles, with a final extension at 72°C for 7 min.

Statistical analysis. All measures were summarized as mean ± SE. Associations of categorical variables were evaluated using the Fisher exact test. All tests were two-sided and conducted at the α = 0.05 significance level. All statistical analyses were done with the S-plus, Professional version 6.2 (Insightful Corp.) software.

Notch1 is overexpressed in human PCa cells.Notch1 has been implicated in many malignancies but its role in prostate carcinogenesis is not well defined. Therefore, we examined the expression of Notch1 in different PCa cell lines and also in normal prostate epithelial cells. Western blot analysis showed increased expression of Notch1 and cleaved Notch1 in human PCa cell lines. A significantly higher level of Notch1 expression was observed in androgen-independent PC3, DU-145, and 22Rν1 cells and androgen-dependent LNCaP cells compared with prostate epithelial cells (Fig. 1A). Notch1 protein in 22Rν1 cells migrated faster than in other PCa cells, which could be a splice variant of the native Notch1 protein.

Notch1 expression increases in human prostate tumor specimens with increasing tumor grade. To further establish the contextual role of Notch1 in PCa, immunohistochemistry was done in human prostate tumor specimens of normal, HGPIN, and PCa representing different tumor grades. A total of 318 specimens were examined. We observed enhanced immunoreactivity of Notch1 in tumor cells compared with normal epithelial cells in adjacent areas of the same tissue (Fig. 1C, a). In addition, enhanced expression of Notch1 was also prominent in the cells surrounding blood capillaries in cancerous tissue (Fig. 1C, c). Based on the scoring patterns, a significant difference was observed in Notch1 expression between cancer and normal tissues (Table 1). The staining for Notch1 protein was moderate to strong in PCa specimens as compared with normal prostate specimens, which exhibited either none or weak to moderate staining (Table 1). In a total of 91 specimens of normal tissue obtained from adjacent regions of tumor tissues, the staining for Notch1 was weak in 33 (36%), moderate in 53 (58%), and negative in the remaining 1 (1%) specimen (Table 1). HGPIN specimens (n = 69) showed strong staining for Notch1 in 21 (30%), moderate staining in 41 (60%), and weak staining in the remaining 7 (10%) specimens (Table 1). The percentage of HGPIN specimens exhibiting strong Notch1 staining was 7-fold higher than that in normal gland. Because HGPIN has been identified as the most significant risk stage for PCa development and the expression of Notch1 protein was found to be significantly increased in both HGPIN and PCa specimens, a strong link could be suggested between the expression of Notch1 protein and development of human PCa. Staining for Notch1 was observed in epithelial cells; however, in stromal cells the staining was either occasional or negative in normal as well as in cancer specimens. The immunoreactivity of Notch1 was observed in a coarsely granular pattern in the cytoplasm of epithelial cells of normal, HGPIN, and Gleason pattern 3 (Gleason score 5-6) to Gleason pattern 4 (Gleason score 7-10) prostatic adenocarcinomas (Fig. 1B, c-f). Accumulative analysis of all PCa specimens (n = 158) suggested higher levels of Notch1, with moderate staining in 67 (42%) to strong staining in 86 (54%) specimens, weak staining in 5 (3%) specimens, and no staining in 0 (0%) specimens (Table 1). These results indicate that Notch1 is overexpressed in human PCa.

Notch1 knockdown decreases cell invasion. To address the role of Notch1 in PCa invasion, knockdown of Notch1 was achieved by transfecting PC3 cells with three independent pools of Notch1 siRNA compared with scrambled siRNA where no effect on Notch1 and cleaved Notch1 expression was observed (Fig. 2A). Because we had hypothesized that Notch1 is involved in PCa invasion, and the fact that PC3 cells represent advanced metastatic cancer, we selected this particular cell line for our studies. Notch1 siRNA dose-dependently decreased both Notch1 and cleaved Notch1 expression with maximum effect observed at a concentration of 120 nmol/L at 24 hours posttransfection (Fig. 2B). To further show the effect of Notch1 knockdown, PC3 cells were subjected to invasion assay. Notch1 knockdown cells showed only a marginal invasion through the extracellular matrix compared with cells transfected with nonspecific siRNA (Fig. 2C). Notch1 knockdown caused 40% to 80% decrease in cell invasion (Fig. 2D), suggesting an essential role of Notch1 in conferring invasive properties to PCa cells.

cDNA array identifies decreased transcripts of genes implicated in tumor invasion. To define the role of Notch1 in PC3 cell invasion, a specifically designed cDNA gene array was used. This array contained 96 cDNA fragments specifying genes with role in cell invasion and metastasis. A significant difference in gene expression of CD31, MMP9, uPA, and uPAR (Fig. 3A) was observed in Notch1 knockdown cells. To ascertain fold changes between scrambled and Notch1 siRNA knockdown cells, we performed a detailed analysis of gene expression profile by clustergram, and scatter plot indicated a 35-fold down-regulation in the expression of uPA, 18-fold in CD31, and >3-fold in the expression of MMP9 and uPAR (Table 2). These results indicate that Notch1 regulates the expression of these downstream target genes that are involved in extracellular matrix degradation, suggesting that Notch1 might play a determining role in cell invasion.

Notch1 knockdown decreases MMP9 expression. To further validate the results of microarray analysis, Notch1 knocked down cells were subjected to Western blot analysis. We observed a decrease in pro-MMP9 protein levels (Fig. 3B), suggesting that Notch1 may be involved in MMP9 activation either by enhancing its expression or by stabilizing the protein. Gelatin zymography was done to assess MMP9 activity in cultured medium from Notch1 knockdown cells and we observed a significant decrease in MMP9 activity (Fig. 3B). To further strengthen our findings, MMP9-specific ELISA was done to quantify secreted MMP9 protein levels. Results indicated a consistent decrease in MMP9 secretion compared with scrambled control (Fig. 3C), further emphasizing the role of Notch1 in regulation of MMP9. To further elucidate the involvement of Notch1 in MMP9 expression, luciferase assay was done to evaluate the effect of Notch1 on MMP9 promoter activity. Notch1 knockdown cells showed a marked 2-fold decrease in reporter activity (P < 0.01), indicating the involvement of Notch1 in MMP9 expression (Fig. 3D). Surprisingly, the decrease observed in MMP9 promoter activity was not dose dependent. This suggests that although Notch1 is involved in MMP9 transcription, it may not be an exclusive mechanism that regulates MMP9 expression and indicates the existence of a posttranslational stabilization mechanism. However, real-time reverse transcription-PCR (RT-PCR) data showed a dose-dependent decrease in mRNA expression of MMP9 in Notch1 knockdown cells, which further confirms our microarray data (Supplementary Fig. S1).

Notch1 knockdown decreases the expression of uPA and its receptor uPAR. uPA is another marker for invasion and metastasis and has been shown to be up-regulated in many malignancies including PCa. uPA regulates the conversion of pro-MMPs to their active forms, which are involved in extracellular matrix degradation. RT-PCR and real-time RT-PCR were done to confirm the microarray data, which showed down-regulation of uPA and uPAR transcripts in Notch1 knockdown cells (Fig. 4A-C). We also observed that knockdown of Notch1 in PC3 cells inhibited protein levels of the ligand uPA and its receptor uPAR (Fig. 4D). These data were in direct agreement with the microarray data and suggest the involvement of Notch1 in regulation of uPA and uPAR.

Notch1 knockdown decreases invasion and inhibits the expression of MMP9, uPA, and uPAR in another human PCa 22Rν1 cell line. To ascertain whether Notch1 knockdown produces similar effects in other PCa cells, we selected androgen-sensitive 22Rν1 cells and transiently transfected with scrambled siRNA and specific Notch1 siRNA. Knockdown of Notch1 in 22Rν1 cells significantly decreased protein expression of Notch1 and cleaved Notch1 (Fig. 5A), invasion of cells through the extracellular matrix (Fig. 5B), uPA expression (Fig. 5C), and MMP9 promoter activity (Fig. 5D) compared with cells transfected with nonspecific siRNA. RT-PCR and real-time RT-PCR were done and showed down-regulation of MMP9, uPA, and uPAR transcripts in Notch1 knockdown cells (Supplementary Fig. S5A-C). We also observed that knockdown of Notch1 in 22Rν1 cells inhibited protein levels of the ligand uPA and its receptor uPAR (Supplementary Fig. S5D). These data suggest that the effect of Notch1 knockdown in PCa cell invasion could be comprehensive because we observed similar mechanism in both the PCa androgen-sensitive 22Rν1 and androgen-insensitive PC3 cells.

Aberrant expression of Notch1 has been detected in various types of human cancers including T-cell acute lymphoblast leukemia (17), breast carcinoma (18), and brain tumor (19). Whereas some studies have shown the involvement of Notch1 in cancer progression, others have suggested antiproliferative effect of Notch1 in some cancer types (3). Notch1 has also been reported to be overexpressed in malignant phenotype including moderately differentiated adenocarcinoma of TRAMP mice (11). In spite of these studies, the role of Notch1 in prostate carcinogenesis remains poorly understood. In this study, we report overexpression of Notch1 in PCa cell lines, consistent with previous findings of Zayzafoon et al. (20). We also observed significantly elevated expression of Notch1 in human PCa tissues. Notch1 expression increased with increasing tumor grade, with specimens of Gleason pattern 3 and 4 exhibiting significantly higher percentage of strong expression. An interesting observation was a significant induction of Notch1 in vascular endothelial cells of these tissues, consistent with earlier reports (21), suggesting that Notch1 may facilitate angiogenesis of PCa cells to neighboring and distant organs. We observed that targeted disruption of Notch1 in PC3 cells resulted in significant decrease in cell invasion across artificial matrix, which mimics in vivo extracellular matrix. In cDNA array, we observed significant modulation of genes that are involved in cell invasion and angiogenesis. We observed significant decreases in the expression of CD31, uPA, uPAR, and MMP9. MMPs have been shown to be involved in extracellular matrix degradation and are overexpressed in advanced stage PCa. We examined the functional activity of MMP9 in Notch1 knockdown cells and observed concentration-dependent decrease in MMP9 expression and activity, indicating that Notch1 might directly regulate the expression of MMP9 by enhancing its promoter activity. Intracellular domain of Notch 1 vests a transactivation function, and it is possible that it might help recruit transcriptional machinery to MMP9 regulatory element to enhance its expression. Although a progressive decrease in Notch1 was observed with increasing concentration of Notch1 siRNA, the decrease in MMP9 promoter activity was not concentration dependent, in contrast to MMP9 protein expression, which rules in the possibility for the existence of a posttranslational stabilization mechanism. We suggest that Notch1 controls MMP9 expression directly by enhancing its transcriptional activity where it functions as a transcription factor. Our data correspond to a previous report by Wang et al. (22), which showed the regulation of MMP9 via Notch1 in pancreatic cancer cells. An additional mechanism that is commonly involved in promoting tumor cell invasion is the uPA-uPAR system, which is one of the most frequent alterations observed in invasive type of PCa (23). We also observed decrease in uPA and uPAR expression in Notch1 knockdown cells, suggesting that Notch1 can also regulate the expression of uPA and its receptor via a distinct gene expression mechanism. These findings suggest that Notch1 is involved in invasion and metastasis of PCa by regulating the expression of MMPs in addition to the uPA-uPAR system, which might work in synergy to enhance tumor cell invasion and metastasis. To illustrate the broader relevance of our hypothesis, we performed additional studies on androgen-sensitive human PCa 22Rν1 cells, which also showed significantly higher expression of Notch1. Our results indicate that the disruption of Notch1 in these cells also leads to a decrease in their invasiveness accompanied by a decrease in MMP9 and uPA expression and suggest the versatile role of Notch1 in PCa cell invasion. We suggest that Notch1 could be a target for intervention of human PCa.

No potential conflicts of interest were disclosed.

We thank Dr. Dougles D. Boyden for providing pGL3-MMP9-luc reporter plasmid.