Colon cancer progression is driven by APEX1-mediated upregulation of Jagged (original) (raw)
APEX1 enhances tumorigenicity in human fibroblast GM00637 and colon cancer cell lines. To evaluate the tumorigenic effects of APEX1 expression in noncancerous cells, we generated GM00637 cells (an immortalized human fibroblast cell line) stably overexpressing full-length APEX1 (referred to herein as GM00637-APEX1 cells) or the control vector. We first explored the effects of APEX1 overexpression on cell growth in low-serum (1% FBS) medium. Compared with the control cells, GM00637-APEX1 cells were significantly more visible and proliferative in low-serum medium (Supplemental Figure 1A; supplemental material available online with this article; doi:10.1172/JCI65521DS1). We next examined the capacity of APEX1 to drive anchorage-independent growth in soft agar colony formation assays. Whereas control cells induced the formation of only a few colonies in soft agar, GM00637-APEX1 cells formed numerous colonies after 4 weeks (Supplemental Figure 1B), demonstrating that APEX1 could induce a transformed phenotype. Additionally, APEX1 overexpression enhanced the migratory potential of GM00637 cells, as demonstrated by wound-healing and Transwell assays (Supplemental Figure 1, C and D).
Given that APEX1 controls cell proliferation, transformation, and migration in GM00637 cells, that APEX1 plays a role in proliferation in colon cancer cells, and that increased APEX1 levels in colon cancer are associated with poor prognosis (24), we studied its potential effects on colon cancer progression. Western blot analysis in different human colon cell lines revealed high levels of APEX1 expression in NCI-H548, NCI-H716, DLD1, KM12SM, and KM12C cells, but low levels in SW480, HT29, and NCI-H747 cells (see below).
To assess the contribution of APEX1 to the tumorigenic potential of colon cancer cells, we analyzed the effects of APEX1 depletion on the ability of these cells to proliferate and grow in an anchorage-independent manner. After initially using siRNA to transiently reduce APEX1 protein levels, we sought to increase the knockdown efficiency by using DLD1 and KM12SM colon cancer cell lines, which normally express a high level of APEX1, to produce cell lines stably expressing reduced levels of APEX1. Thus, we generated a shRNA construct encoding an APEX1-targeting shRNA and used it to establish stable APEX1-knockdown DLD1 and KM12SM cell lines (referred to herein as APEX1-shRNA/DLD1 and APEX1-shRNA/KM12SM cells, respectively). Importantly, APEX1 depletion clearly slowed cell proliferation and also significantly impaired colony formation by DLD1 and KM12SM cells in soft agar (Figure 1, A and B). Additionally, APEX1-shRNA/DLD1 and APEX1-shRNA/KM12SM cells migrated and invaded approximately 60%–80% less than control shRNA–transfected cells (Figure 1C). Angiogenesis was also robustly inhibited in APEX1-shRNA/DLD1 and APEX1-shRNA/KM12SM cells (Figure 1D). These results suggest that APEX1 regulates the tumorigenic behavior of colon cancer cells. We next explored the effect of ectopic expression of APEX1 on tumorigenicity of SW480 and HT29 colon cancer cell lines, which express low endogenous levels of APEX1. After stable transfection with an APEX1 cDNA expression construct, the resulting APEX1-overexpressing SW480 and HT29 cells (referred to herein as SW480-APEX1 and HT29-APEX1 cells, respectively) demonstrated significant increases in cellular proliferation, colony-forming activity in soft agar, invasion and migration ability, and angiogenesis (Figure 2, A–D). Together, these results indicate that the expression of endogenous APEX1 in colon cancer cells may be important for maintaining proliferative, migratory, invasive, and angiogenic capacities.
APEX1 regulates tumorigenic features in DLD1 and KM12SM human colon cancer cell lines. DLD1 and KM12SM cells were transfected with control or APEX1 shRNA. (A) Growth curves and Western blot analysis. (B) Anchorage-independent colony formation on soft agar. (C) Cell migration and invasion. Data represent the average cell number from 5 viewing fields. (D) Tube formation assays were performed by culturing HUVECs in tube formation with supernatants from the indicated cells. Results in A–D are mean ± SD (n = 3). **P < 0.01. Scale bars: 150 μm (B); 40 μm (C); 80 μm (D).
APEX1 regulates tumorigenic features in SW480 and HT29 human colon cancer cell lines. SW480 and HT29 cells were transfected with control or APEX1-expressing vector. (A) Growth curves and Western blot analysis. (B) Colony formation on soft agar after 14 days of culture. (C) Cell migration and invasion. (D) Tube formation assays were performed by culturing HUVECs in tube formation with supernatants from the indicated cells. Results in A–D are mean ± SD (n = 3). **P < 0.01. Scale bars: 150 μm (B); 40 μm (C); 80 μm (D).
APEX1 upregulates JAG1 gene expression. To examine the role of APEX1 in the regulation of gene expression and the consequences on colon cancer progression, we used cDNA microarrays to identify genes that were differentially expressed in SW480-APEX1, APEX1-shRNA/DLD1, and their control cells. We were specifically interested in genes with increased expression in SW480-APEX1 cells, but decreased expression in APEX1-shRNA/DLD1 cells. We used a 1.5-fold induction threshold in SW480-APEX1 cells and 1.5-fold reduction threshold in APEX1-shRNA/DLD1 cells and identified 81 genes meeting these criteria (Supplemental Table 1). Using ontology analysis consisting of proliferation, migration, and angiogenesis, we discovered 4 genes that satisfied these 3 biological criteria (Figure 3, A and B). Among these, JAG1 (which encodes the Notch ligand) was particularly interesting, because activation of Notch signaling is involved in human colon cancer (5, 12, 13, 33–35). We also conducted expression microarray profiling of control and GM00637-APEX1 cells. A Venn diagram comprising genes expressed with a 3-fold increase in GM00637-APEX1 cells revealed 7 common genes involved in migration and in proliferation and differentiation (Figure 3, E and F). Importantly, JAG1 was also found to be upregulated in GM00637-APEX1 cells, further supporting the possibility that JAG1 may be a downstream target of APEX1. To confirm the microarray data, expression of Jagged1 was examined using real-time RT-PCR and Western blot analyses in APEX1-shRNA/DLD1 and SW480-APEX1 cells. Whereas JAG1 mRNA and Jagged1 protein were downregulated by transfection of an APEX1 siRNA, JAG1 mRNA and Jagged1 protein were upregulated by transfection of an APEX1 expression vector (Figure 3, C and D). Additionally, we confirmed the upregulation of Jagged1 by APEX1 in GM00637 cells, obtaining similar results (Figure 3, G and H).
APEX1 upregulates Jagged1 transcription. (A) Number of common and unique APEX1 target genes in migration, proliferation, and angiogenesis of DLD1 and SW480 cells. (B) Heat map showing 4 common genes, including JAG1 (arrow), involved in the 3 tumorigenic processes in A. (C) Quantitative real-time RT-PCR for relative JAG1 mRNA from DLD1 cells transfected with control or APEX1 siRNA and from SW480 cells transfected with control or APEX1 expression vector. (D) Western blot analysis of Jagged1 expression in the indicated cells. (E) Number of common and unique APEX1 target genes involved in migration and in proliferation and differentiation of GM00637 cells. (F) Heat map showing 7 common genes, including JAG1 (arrow), involved in the 2 biological processes in E. (G) Relative expression of JAG1 mRNA in GM00637 cells transfected with control or APEX1 expression vector. (H) Western blot analysis of Jagged1 expression in the indicated GM00637 cells. Results in C and G are mean ± SD (n = 3). **P < 0.01.
APEX1 is a positive regulator of Jagged1/Notch signaling in colon cancer cells. To further corroborate the correlation between the expression levels of APEX1 and Jagged1 in colon cancer cells, we performed Western blot and real-time RT-PCR analyses on a number of different human colon cancer cells. APEX1 and Jagged1 were coexpressed in the human colon cancer cell lines. Higher expression of Jagged1 protein and JAG1 mRNA was found in human colon cancers with high APEX1 expression, including cell lines NCI-H548, NCI-H716, DLD1, KM12SM, and KM12C (Figure 4, A and B). Conversely, human colon cancer cells expressing low levels of APEX1, including the lines SW480, HT29, and NCI-H747, exhibited little Jagged1 protein and JAG1 mRNA expression. We also examined the endogenous levels of APEX1 and Jagged1 in 17 human cancer cell lines, including SNU638, AGS, SNU216, and SNU484 (gastric cancer); DMS53, H460, H1299, Calu-1, and SK-MES-1 (lung cancer); U87, U373, M059J, and M059K (glioma); and PANC-1, ASPC-1, MIAPaCa-2, and BXPC-3 (pancreatic cancer). Although APEX1 and Jagged1 were not coexpressed in some of the glioma and pancreatic cell lines, APEX1 was closely coexpressed with Jagged1 in the gastric and lung cancer cell lines (Supplemental Figure 2).
Jagged1 expression and Notch signaling are elevated in colon cancer cells expressing APEX1. (A–D) Western blot analysis for expression of the indicated proteins (A), JAG1 mRNA levels (B), relative CBF-1–Luc promoter reporter luciferase assay (C), and HES1 mRNA levels (D) in 8 colon cancer cell lines. (E, F, and I) Western blot analysis for expression of the indicated proteins (E), JAG1 mRNA levels (F), and CBF-1–Luc promoter reporter luciferase assay (I) in 4 colon cancer cell lines transfected with control or APEX1 siRNA. (G, H, and J) Western blot analysis for expression of the indicated proteins (G), JAG1 mRNA levels (H), and CBF-1–Luc promoter reporter luciferase assay (J) in SW480 and HT29 cells transfected with control or APEX1 expression vector. Results in B–D, F, and H–J are mean ± SD (n = 3). **P < 0.01.
4 Notch proteins have been described (Notch1, Notch2, Notch3, and Notch4) that serve as receptors for the specific ligands. Upon receptor-ligand interaction, Notch proteins are cleaved by γ-secretase activity, and the resulting cleaved Notch translocates to the nucleus, where it associates with the DNA-binding protein (4). Thus, we next sought to determine the cleaved forms of Notch protein in 8 colon cancer cell lines by Western blot analysis. Activated Notch3 was present at higher levels in colon cancer cell lines with high expression of APEX1 and at lower levels in colon cancer cell lines with lower expression of APEX1 (Figure 4A).
We next quantified the levels of Notch activation by following luciferase activity driven from a Notch-dependent CBF-1–responsive reporter transfected into colon cancer cell lines. The activity of the CBF-1–dependent luciferase reporter gene was higher in colon cancer cell lines expressing high versus low levels of APEX1 (Figure 4C). We also used RT-PCR to examine Notch target gene HES1 expression in colon cancer cell lines and found the same increased levels of HES1 in colon cancer cells with high APEX1 expression (Figure 4D).
In light of this evidence supporting activated Jagged1/Notch signaling in colon cancer cells with APEX1 expression, we examined whether knockdown or overexpression of APEX1 could cause abnormal Jagged1/Notch signaling pathway activity. When APEX1 was knocked down by transfection with APEX1 siRNA in colon cancer cells, including NCI-H716, DLD1, KM12SM, and KM12C, levels of Jagged1 protein and JAG1 mRNA were dramatically reduced (Figure 4, E and F). Conversely, the levels of JAG1 mRNA and Jagged1 protein were significantly increased in SW480-APEX1 and HT29-APEX1 cells (Figure 4, G and H), which suggests that APEX1 upregulated Jagged1 in various colon cancer cells by affecting gene transcription.
We next investigated Notch signaling in colon cancer cell lines. The expression of activated Notch3 in colon cancer cell lines with high expression of APEX1 was markedly reduced when APEX1 expression was decreased by siRNA (Figure 4E). On the other hand, SW480-APEX1 and HT29-APEX1 cells exhibited substantially increased cleaved Notch1 protein levels (Figure 4G), which suggests that APEX1 plays an important role in the efficient activation of Notch1 or Notch3 signaling, depending on the cellular context of colon cancer.
To examine whether the APEX1-mediated upregulation of cleaved Notch in colon cancer cells is functional, we examined whether APEX1 is capable of regulating CBF-1–Luc activity and HES1 mRNA expression. APEX1 siRNA–transfected NCI-H716, DLD1, KM12SM, and KM12C cells exhibited decreased CBF-1–Luc activity and HES1 expression (Figure 4I and Supplemental Figure 3A). Conversely, SW480 and HT29 cells overexpressing APEX1 had enhanced CBF-1 promoter activity and HES1 expression (Figure 4J and Supplemental Figure 3B). Together, these results suggest that the enhanced activity of Notch signaling in colon cancer cells is likely APEX1 dependent.
APEX1 activates Notch signaling through Jagged1 upregulation. Notch receptors and their ligands use various positive and negative feedback mechanisms to control availability and, ultimately, signaling (36, 37). However, there are limited reports examining their effects on each other’s expression. To investigate the mechanism by which APEX1 leads to Notch signaling activation in colon cancer cells, we initially explored the functional relationship between Jagged1 expression and Notch activation using pharmacologic and genetic approaches. We treated DLD1 and KM12SM cells with DAPT, which blocks Notch cleavage and activation, and analyzed the cells for Jagged1 protein. Whereas cleaved Notch3 was substantially reduced in DAPT-treated DLD1 and KM12SM cells, DAPT did not affect Jagged1 expression (Figure 5A). We then used siRNA constructs to specifically inhibit Notch3 in order to determine whether the observed Jagged1 expression in DAPT-treated cells could be replicated. Western blot analysis revealed that Notch3-specific siRNA transfection decreased Notch3 expression more than 95% compared with expression in control siRNA–transfected DLD1 and KM12SM cells. Consistent with the DAPT data, Notch3 siRNA did not significantly alter the expression of Jagged1 (Figure 5B).
APEX1 is a positive regulator of Notch signaling through Jagged1. (A) Western blot analysis for expression of the indicated proteins in DLD1 and KM12SM cells treated with DMSO or DAPT. Expression levels were quantified by densitometry and are shown as the fold decrease relative to with DMSO-treated cells. (B) Western blot analysis for expression of the indicated proteins in DLD1 and KM12SM cells transfected with the indicated siRNA. Expression levels were quantified by densitometry and expressed as fold decrease relative to untreated cells. (C) CBF-1–Luc promoter reporter luciferase assay in the indicated cells. (D and E) Western blot analysis for expression of the indicated proteins (D) and CBF-1–Luc promoter reporter luciferase assay (E) in SW480 and HT29 cells transfected with control or Jagged1 expression vector. (F and G) Western blot analysis for expression of the indicated proteins (F) and CBF-1–Luc promoter luciferase assay (G) in SW480 and HT29 cells transfected with the indicated vector or siRNA. (H) Western blot analysis for expression of the indicated proteins in DLD1 and KM12SM cells transfected with the indicated vector or shRNA. (I) CBF-1–Luc promoter reporter luciferase assay in the indicated cells. Results in A–C, E, G, and I are mean ± SD (n = 3). **P < 0.01.
We next examined whether Jagged1 knockdown or overexpression could affect Notch activity. Transfection of DLD1 and KM12SM cells with Jagged1 siRNA for 48 hours clearly decreased the expression of Jagged1 protein, and cells with reduced levels of Jagged1 had lower cleaved Notch3 levels and CBF-1–Luc activity compared with control siRNA–transfected cells (Figure 5, B and C). Furthermore, transfection of SW480 and HT29 cells with a Jagged1 expression vector increased activated Notch1 expression as well as CBF-1 promoter activities in the cells (Figure 5, D and E). Although the functional relationship between Jagged1 and Notch activation is complex and not yet fully understood, these results suggest that Notch activation is a consequence of autocrine or paracrine activation of the Notch receptor by Jagged1 in colon cancer cells.
We next asked whether Jagged1 is essential for APEX1-induced Notch activation. Levels of endogenous Jagged1 were depressed with siRNA in SW480-APEX1 and HT29-APEX1 cells, and then Notch activity was measured. As shown in Figure 5F, Jagged1 siRNA reduced the levels of cleaved Notch1 in SW480-APEX1 and HT29-APEX1 cells. Knockdown of Jagged1 also clearly suppressed CBF-1 promoter activity in these cells (Figure 5G). When wild-type Jagged1 was transiently expressed in APEX1-shRNA/DLD1 and APEX1-shRNA/KM12SM cells, both cleaved Notch3 and CBF-1 promoter activity were completely rescued in these APEX1-deficient cells (Figure 5, H and I), which suggests that Jagged1 is required for APEX1-induced Notch activation.
APEX1 promotes colon cancer progression by Jagged1. Given the importance of Jagged1/Notch signaling in colon carcinogenesis (5, 12, 34, 35, 38), the involvement of Jagged1 in APEX1-mediated increased tumorigenicity of colon cancer cells was of interest. Additionally, the transcriptional regulation of human Jagged1 has been elusive. Thus, we analyzed the effects of Jagged1 knockdown on APEX1 function. In cell proliferation assays, scrambled control siRNA–transfected SW480-APEX1 and HT29-APEX1 cells showed cell growth, which was slowed down with siRNA-mediated Jagged1 knockdown (Figure 6A). Next, we studied the effects of Jagged1 on anchorage-independent growth. As shown in Figure 6B, knockdown of Jagged1 in both SW480-APEX1 and HT29-APEX1 cells inhibited the anchorage-independent growth induced by APEX1 compared with vector-transfected cells. Furthermore, transfection of Jagged1 siRNA in SW480-APEX1 cells clearly reduced migration, invasion, and angiogenesis (Figure 6, C and D), which suggests that APEX1 promoted cancer progression through Jagged1 expression. When we transiently transfected APEX1-shRNA/DLD1 and APEX1-shRNA/KM12SM cells with a Jagged1 cDNA expression construct to overexpress Jagged1 in these cells, cellular proliferation and anchorage-independent growth were increased (Figure 7, A and B). Moreover, overexpression of Jagged1 in APEX1-shRNA/DLD1 cells also increased invasion, migration ability, and angiogenesis (Figure 7, C and D). These results indicate that Jagged1 is required for the tumorigenetic activity of APEX1 in colon cancer.
Jagged1 knockdown attenuates tumorigenicity in APEX1-overexpressed SW480 and HT29 cells. SW480-APEX1 and HT29-APEX1 cells were transfected with control or Jagged1 siRNA. (A) Cell growth. Lysates of cells were examined for Jagged1 and APEX1 expression by Western blot analysis. (B) Colony formation in soft agar after 14 days of culture. (C) Cell migration and invasion. Data represent the average cell numbers from 5 viewing fields. (D) Tube formation assays were performed by culturing HUVECs in tube formation with supernatants from the indicated cells. Results in A–D are mean ± SD (n = 3). **P < 0.01. Scale bars: 80 μm (B and D); 40 μm (C).
Jagged1 overexpression enhances tumorigenicity in APEX1-depleted DLD1 and KM12SM cells. APEX1-shRNA/DLD1 and APEX1-shRNA/KM12SM cells were transfected with control vector or Jagged1 expression vector. (A) Cell growth. Lysates of cells were examined for Jagged1 and APEX1 expression by Western blot analysis. (B) Colony formation in soft agar after 14 days of culture. (C) Cell migration and invasion. (D) Tube formation assays were performed by culturing HUVECs in tube formation with supernatants from the indicated cells. Results in A–D are mean ± SD (n = 3). **P < 0.01. Scale bars: 100 μm (B); 40 μm (C); 80 μm (D).
APEX1/Jagged1 signaling in colon cancer cells promotes tumor growth and metastatic spreading in mice. We next evaluated the effects of the APEX1/Jagged1 pathway on tumor formation and progression in vivo using mouse xenograft models. Initially, we constructed models from DLD1 and SW480 cells implanted subcutaneously in the right flank of nude mice. At 36 days after injection, all mice implanted with DLD1 control cells developed large tumors, whereas tumor growth of SW480 control cells was negligible at 36 days (data not shown). To determine whether the established tumors were affected by APEX1, we injected APEX1-shRNA/DLD1 and control DLD1 cells into nude mice and measured the tumor volume over time. Whereas control mice displayed visible tumors at 12 days after implantation, with exponential tumor growth until day 36, tumor outgrowth and size were reduced in APEX1-deficient tumor–bearing mice (Figure 8, A and B), which indicates that stable knockdown of APEX1 in colon cancer cells reduced tumor growth in vivo. To determine whether this reduced tumor growth resulted from loss of the Jagged1 effect, nude mice were injected with Jagged1- or vector-transfected APEX1-shRNA/DLD1 cells. No tumor growth of vector-transfected APEX1-shRNA/DLD1 cells was seen at day 36, whereas Jagged1-transfected cells continued to grow until day 36 (Figure 8, A and B). However, Jagged1 overexpression did not completely rescue tumor growth in APEX1-shRNA/DLD1 cells, which suggests that additional APEX1 targets also control tumor growth. We then compared the abilities of SW480-APEX1 and control cells to form tumors in mice. At 36 days after implantation, SW480-APEX1 cells gave rise to substantially larger tumors, whereas control cells did not form tumors (Figure 8, C and D). To determine whether this increased tumor growth resulted from the induction of Jagged1 expression, nude mice were injected with Jagged1 siRNA– or control siRNA–transfected SW480-APEX1 cells. Tumor growth was detected in mice receiving control siRNA–transfected cells, but not Jagged1 siRNA–transfected cells, after 36 days (Figure 8, C and D).
APEX1 regulates tumor growth through Jagged1 in a mouse model. (A and B) The indicated DLD1 cells were implanted subcutaneously in nude mice (n = 6). (A) Photographs of representative mice and tumors (each condition was independently repeated 3 times) and immunofluorescence analysis of APEX1 and Jagged1 in tumors. (B) Growth curves of mammary tumors after implantation. (C and D) The indicated SW480 cells were injected into the right flank of nude mice (n = 6). (C) Photographs of representative mice and tumors and immunofluorescence analysis of APEX1 and Jagged1 in tumors. (D) Growth curves of mammary tumors after injection. Results in B and D are mean ± SD (n = 6). Scale bars: 40 μm (A and C).
Because the enhanced or reduced lung metastasis might have been due to faster or slower primary tumor growth, respectively, we further evaluated the in vivo effects of the APEX1/Jagged1 pathway on lung metastasis of colon cancer using an experimental metastasis assay. Red fluorescent protein–based (RFP-based) noninvasive bioluminescence imaging was used to monitor the presence of tumor cells. Control or APEX1-shRNA/DLD1 cells expressing pCMV-DsRed were injected into the peritoneal cavity of nude mice. After 5 weeks, fluorescence imaging revealed multiple large lung metastases in mice injected with control DLD1 cells, but no metastases in mice injected with APEX1-shRNA/DLD1 cells (Figure 9A). Conversely, the metastatic activity of SW480 cells, assessed by their fluorescence, was markedly increased by APEX1 compared with the control (Figure 9B). Finally, we examined the role of Jagged1 in APEX1-induced metastasis. Metastatic activity was increased in the lungs of mice injected with Jagged1-overexpressing APEX1-shRNA/DLD1 cells, whereas knockdown of Jagged1 in SW480-APEX1 cells reduced the number of lung metastases (Figure 9, A and B). Taken together with our earlier findings, these observations indicate that APEX1 promotes colon cancer cell growth and metastasis in vivo through upregulated Jagged1 expression.
APEX1 regulates tumor metastasis through Jagged1 in a mouse model. The indicated DLD1-RFP (A) and SW480-RFP (B) cells were injected into the abdominal cavity of BALB/c mice (n = 6), and colon cancer metastasis to the lung was measured by fluorescence expression. Shown are representative fluorescence images at 5 (DLD1) and 8 (SW480) weeks after injection and histological analyses showing RFP fluorescence of lung tissue sections. Scale bars: 40 μm (A and B).
Involvement of EGR1 in Jagged1 upregulation by APEX1. To explore the possible direct influence of APEX1 on the Jagged1 promoter, we cloned a human Jagged1 promoter fragment (nucleotides –1473 to +14) into the pGL3 luciferase vector and cotransfected the APEX1 cDNA construct and the Jagged1 reporter construct into GM00637 cells. Transfection reporter assays indicated that Jagged1 transcriptional activity was induced significantly by APEX1 expression (Figure 10A).
APEX1 upregulates Jagged1 expression by increasing EGR1 activity. (A) Control and APEX1-overexpressing GM00637 cells were transfected with the indicated plasmids and harvested for a Jagged1 reporter assay. (B) Top: Jagged1 promoter constructs used for reporter assay. Middle: Reporter assay in GM00637 cells transfected with the indicated Jagged1 promoter fragments fused to pGL3 basic vector. Bottom: Western blot analysis of V5-APEX1 expression in GM00637 cells transfected with the indicated vectors. (C) Reporter assay in GM00637 cells with Jagged1 promoter constructs (pC) with mutations in different EGR1 regions (mE1, mE2, mE3). (D) EMSA analysis of 3 putative EGR1 consensus sites (E1, E2, E3) of the Jagged1 promoter in the indicated GM00637 cells. Unlabeled oligonucleotides were used as competitors. For supershift assays, anti-EGR1 antibody was added to the reaction mixtures prior to separating the DNA-protein complex. (E) ChIP assay for Jagged1 promoter (E2 site) in control and GM00637-APEX1 cells. (F–I) JAG1 mRNA (F and H) and Jagged1 protein (G and I) levels in GM00637-APEX1 and SW480-APEX1 cells (F and G) or in DLD1 cells (H and I) transfected with control or EGR1 siRNA. Results in A–C, F, and H are mean ± SD (n = 3). **P < 0.01.
To dissect the promoter region required for APEX1-induced Jagged1 transcription, we generated 4 fragments from the full-length Jagged1 promoter. Specific deletions of the Jagged1 promoter demonstrated that a 360-bp region from –850 to –490 contained the major APEX1-responsive element (Figure 10B). Since APEX1 enhances DNA binding of EGR1 by acting as a transcriptional coactivator (39), and given that the APEX1-responsive region of the Jagged1 promoter (position –850 to –490) contains 3 potential EGR1-binding sites, at –794 to –778, –638 to –662, and –556 to –540 (sites E1, E2, and E3, respectively; Figure 10B), we investigated whether APEX1-induced EGR1 activity contributes to the increase in Jagged1 promoter activity. To validate the functionality of these potential EGR1 binding sites, we mutated the EGR1 binding motifs in E1, E2, and E3 of the Jagged1 promoter and used these constructs in a reporter assay. Mutations in E1 and E3 did not significantly affect the ability of APEX1 to induce Jagged1 transcription activation; however, mutations in E2 markedly decreased the effect of APEX1 on Jagged1 transcription activation (Figure 10C), which suggests that E2, but not E1 or E3, is important for APEX1-induced Jagged1 transcription. To determine whether EGR1 binds directly to E2, we performed EMSAs of nuclear extracts from parent, control vector–transfected, and GM00637-APEX1 cells using isotope-labeled DNA probes containing E1, E2, and E3. Nuclear extract from APEX1-expressing cells bound specifically to the putative EGR1 binding site E2, but not to E1 or E3; moreover, competition with an unlabeled, wild-type E2 binding site oligonucleotide blocked the binding to E2 (Figure 10D). Furthermore, supershifted DNA-protein complex was observed after adding the anti-EGR1 antibody to the DNA binding reaction. To directly confirm the binding of EGR1 to the endogenous Jagged1 promoter in vivo, we performed ChIP assays in pcDNA3- and APEX1 expression vector–transfected cells. The results showed that EGR1 bound in vivo to the E2 region in pcDNA3-expressing cells, and the amount of ChIP associated with EGR1 increased in APEX1-expressing cells (Figure 10E). These results suggest that increased interaction of EGR1 with E2 in the Jagged1 promoter contributes to an APEX1-mediated increase in Jagged1 transcription.
To assess whether the ability of APEX1 to upregulate Jagged1 expression depends on its activation of EGR1, we suppressed EGR1 expression levels in GM00637-APEX1 and SW480-APEX1 cells using EGR1 siRNA and measured Jagged1 expression by real-time RT-PCR and Western blot. Knockdown of EGR1 significantly suppressed the levels of JAG1 mRNA and Jagged1 protein in GM00637-APEX1 and SW480-APEX1 cells, whereas control siRNA did not affect Jagged1 expression (Figure 10, F and G). We then tested whether endogenous EGR1 contributes to Jagged1 expression in DLD1 cells. Indeed, EGR1 siRNA transfection significantly decreased JAG1 mRNA and Jagged1 protein expression (Figure 10, H and I). These results suggest that APEX1 increases Jagged1 expression, at least in part, by activating EGR1 transcriptional activity.
APEX1 and Jagged1 expression levels correlate in human colon cancer tissues. Because APEX1 overexpression correlates with Jagged1 expression in vitro, and these components are important in colon cancer progression, we sought to determine whether there a positive correlation exists between APEX1 and Jagged1 expression in colon carcinoma tissues. To test this, we performed immunohistochemical staining to detect Jagged1 and APEX1 expression on colorectal tissue samples consisting of normal human colon tissues, colorectal adenoma, and colorectal adenocarcinomas of different grades. Tissue staining was scored on the basis of cytoplasmic membrane and cytoplasm for Jagged1 and nuclei for APEX1. APEX1 and Jagged1 proteins were weakly detected in most normal colon tissues, which had average immunohistochemical scores of 1.21 ± 0.64 (n = 25) for APEX1 and 0.48 ± 0.25 (n = 24) for Jagged1 (Figure 11, A and B). In contrast, APEX1 and Jagged1 staining in adenoma was significantly increased, to 3.64 ± 0.71 (n = 58) and 2.12 ± 0.54 (n = 41), respectively. Immunohistochemical scores of APEX1 and Jagged1 progressively increased in adenoma, grade I adenocarcinoma, and grade II adenocarcinoma (Figure 11, A and B). We also investigated APEX1 and Jagged1 expression in patient colon tumors versus normal adjacent tissues. Analyses of APEX1 and Jagged1 expression showed that colon carcinoma tissues strongly expressed APEX1 and Jagged1, whereas these proteins were expressed at low levels in normal colon tissues (Figure 11C). This observation underscores the clinical relevance of APEX1 and Jagged1 expression in colon cancer.
Correlation between APEX1 and Jagged1 expression in human colon cancer. (A) Jagged1 and APEX1 proteins in normal colon tissue, colorectal adenoma, and grade I–III colorectal adenocarcinoma are shown by immunohistochemistry with anti-Jagged1 and anti-APEX1 antibodies. Brown staining indicates positive APEX1 or Jagged1 staining. (B) APEX1 and Jagged1 expression levels, assessed by immunohistochemistry scoring (see Methods). Jagged1 expression significantly correlated with APEX1 levels (P < 0.01, Pearson correlation test). (C) Representative images of Jagged1 and APEX1 immunoreactivity in normal colon epithelium and colon adenocarcinoma (separated by dashed lines). Results in B are SEM. *P < 0.05; **P < 0.01. Scale bars: 200 μm (A and C).