Roles of arrest-defective protein 1(225) and hypoxia-inducible factor 1alpha in tumor growth and metastasis

Abstract

Background: Vascular endothelial growth factor A (VEGFA), a critical mediator of tumor angiogenesis, is a well-characterized target of hypoxia-inducible factor 1 (HIF-1). Murine arrest-defective protein 1A (mARD1A(225)) acetylates HIF-1alpha, triggering its degradation, and thus may play a role in decreased expression of VEGFA.

Methods: We generated Apc(Min/+)/mARD1A(225) transgenic mice and quantified growth of intestinal polyps. Human gastric MKN74 and murine melanoma B16F10 cells overexpressing mARD1A(225) were injected into mice, and tumor growth and metastasis were measured. VEGFA expression and microvessel density in tumors were assessed using immunohistochemistry. To evaluate the role of mARD1A(225) acetylation of Lys532 in HIF-1alpha, we injected B16F10-mARD1A(225) cell lines stably expressing mutant HIF-1alpha/K532R into mice and measured metastasis. All statistical tests were two-sided, and P values less than .05 were considered statistically significant.

Results: Apc(Min/+)/mARD1A(225) transgenic mice (n = 25) had statistically significantly fewer intestinal polyps than Apc(Min/+) mice (n = 21) (number of intestinal polyps per mouse: Apc(Min/+) mice vs Apc(Min/+)/mARD1A(225) transgenic mice, mean = 83.4 vs 38.0 polyps, difference = 45.4 polyps, 95% confidence interval [CI] = 41.8 to 48.6; P < .001). The growth and metastases of transplanted tumors were also statistically significantly reduced in mice injected with mARD1A(225)-overexpressing cells than in mice injected with control cells (P < .01). Moreover, overexpression of mARD1A(225) decreased VEGFA expression and microvessel density in tumor xenografts (P < .04) and Apc(Min/+) intestinal polyps (P = .001). Mutation of lysine 532 of HIF-1alpha in B16F10-mARD1A(225) cells prevented HIF-1alpha degradation and inhibited the antimetastatic effect of mARD1A(225) (P < .001).

Conclusion: mARD1A(225) may be a novel upstream target that blocks VEGFA expression and tumor-related angiogenesis.

Figure 1

Effect of transgenic overexpression of mARD1A²²⁵ on intestinal tumorigenesis in Apc^Min/+ mice. A) Representative images (top) and hematoxylin and eosin staining (bottom) of small intestinal polyps from 16-week-old Apc^Min/+ mice and Apc^Min/+/mARD1A²²⁵ transgenic mice (Tg). Magnification ×100. Scale bar = 200 μm. B) Tumor size distribution of polyps in the small intestine from Apc^Min/+ mice (n = 21 per group) and Apc^Min/+/mARD1A²²⁵ Tg (n = 25 per group). Polyps were classified according to size in millimeters. Results are means with 95% confidence intervals (error bars). *P < .001, compared with Apc^Min/+ mice, determined with the two-sided Mann–Whitney U test. C) Kaplan–Meier survival plot of Apc^Min/+ (mean = 141 days, 95% confidence interval = 131 to 151, n = 10 per group, solid line) and Apc^Min/+/mARD1A²²⁵ Tg (mean = 170 days, 95% confidence interval = 150 to 191, n = 10 per group, dotted line). *P = .005, compared with Apc^Min/+ mice, determined with log-rank tests. All statistical tests were two-sided.

Figure 2

Effect of transgenic overexpression of mARD1A²²⁵ on neovascularization and vascular endothelial growth factor A (VEGFA) expression in intestinal polyps in Apc^Min/+/mARD1A²²⁵ transgenic mice. A) Immunohistochemical analysis of mARD1A²²⁵ (top row), VEGFA (second row), and von Willebrand Factor (vWF, third row) in intestinal polyps of Apc^Min/+ mice and Apc^Min/+/mARD1A²²⁵ transgenic mice at 16 weeks of age (n = 8 mice per group). Magnification ×200. Scale bars = 100 μm. Higher magnification of the selected area (third row, yellow boxes) on vWF with vessels indicated by red arrows within the polyps (bottom row). Magnification ×400. Scale bar = 20 μm. B) The numbers of VEGFA-positive cells in polyps are presented as means with 95% confidence intervals (error bars). VEGFA-positive cells in four high-power fields (×200) from each of the eight mice per group were scored for quantification. *P = .001, compared with the number of VEGFA-positive cells in Apc^Min/+ mouse intestinal polyps, as determined with the two-sided Mann–Whitney U test. C) Concentrations of mouse VEGFA protein in serum from Apc^Min/+ mice and Apc^Min/+/mARD1A²²⁵ transgenic mice, respectively (n = 12 mice per group). VEGFA protein levels were quantified by enzyme-linked immunosorbent assay (ELISA). The experiment was performed in two replicate for each of 12 mice per group. The levels of VEGFA protein in serum are presented as means and 95% confidence intervals (error bars) that are derived from the means of the individual mice (n = 12). *P = .023, compared with the levels of VEGFA protein in Apc^Min/+ mouse serum, as determined with the two-sided Mann–Whitney U test. D) The numbers of vWF-positive vessels in polyps are presented as means and 95% confidence intervals (error bars). Vessels from five high-power fields (×400) from each of the eight mice per group were scored for quantification. *P = .001, compared with the number of vWF-positive vessels in Apc^Min/+ mouse intestinal polyps, as determined with the two-sided Mann–Whitney U test. Tg = transgenic mice.

Figure 2

Effect of transgenic overexpression of mARD1A²²⁵ on neovascularization and vascular endothelial growth factor A (VEGFA) expression in intestinal polyps in Apc^Min/+/mARD1A²²⁵ transgenic mice. A) Immunohistochemical analysis of mARD1A²²⁵ (top row), VEGFA (second row), and von Willebrand Factor (vWF, third row) in intestinal polyps of Apc^Min/+ mice and Apc^Min/+/mARD1A²²⁵ transgenic mice at 16 weeks of age (n = 8 mice per group). Magnification ×200. Scale bars = 100 μm. Higher magnification of the selected area (third row, yellow boxes) on vWF with vessels indicated by red arrows within the polyps (bottom row). Magnification ×400. Scale bar = 20 μm. B) The numbers of VEGFA-positive cells in polyps are presented as means with 95% confidence intervals (error bars). VEGFA-positive cells in four high-power fields (×200) from each of the eight mice per group were scored for quantification. *P = .001, compared with the number of VEGFA-positive cells in Apc^Min/+ mouse intestinal polyps, as determined with the two-sided Mann–Whitney U test. C) Concentrations of mouse VEGFA protein in serum from Apc^Min/+ mice and Apc^Min/+/mARD1A²²⁵ transgenic mice, respectively (n = 12 mice per group). VEGFA protein levels were quantified by enzyme-linked immunosorbent assay (ELISA). The experiment was performed in two replicate for each of 12 mice per group. The levels of VEGFA protein in serum are presented as means and 95% confidence intervals (error bars) that are derived from the means of the individual mice (n = 12). *P = .023, compared with the levels of VEGFA protein in Apc^Min/+ mouse serum, as determined with the two-sided Mann–Whitney U test. D) The numbers of vWF-positive vessels in polyps are presented as means and 95% confidence intervals (error bars). Vessels from five high-power fields (×400) from each of the eight mice per group were scored for quantification. *P = .001, compared with the number of vWF-positive vessels in Apc^Min/+ mouse intestinal polyps, as determined with the two-sided Mann–Whitney U test. Tg = transgenic mice.

Figure 3

Effect of mARD1A²²⁵ on primary tumor growth in mice. A) Representative images of primary tumor masses (left panel) and growth curves (right panel) are shown. C57BL/6 mice were injected subcutaneously with B16F10 (open diamond), B16F10-mock (open square), or B16F10-mARD1A²²⁵ cells (two clones; closed symbols). Results are means and 95% confidence intervals (error bars). *P = .003, **P < .001, compared with mock control cells, as determined with two-way repeated measures analysis of variance (ANOVA). B) Representative images of primary tumor masses (left panel) and growth curves (right panel) are shown. Nude mice were injected subcutaneously with MKN74 (open diamond), MKN74-mock (open square), or MKN74-mARD1A²²⁵ cells (two clones; closed symbols). Results are means and 95% confidence intervals (error bars). *P = .007, **P < .001, compared with mock control cells, as determined with two-way repeated measures ANOVA. C) Tumor weights (n = 8 per group) were measured 3 weeks after inoculation with B16F10, B16F10-mock, or B16F10-mARD1A²²⁵ cells. Results are means and 95% confidence intervals (error bars). *P = .021, compared with mock control cells, as determined with the two-sided Mann–Whitney U test. D) Tumor weights (n = 10 per group) were measured 3 weeks after inoculation with MKN74, MKN74-mock, and MKN74-mARD1A²²⁵ cells. Results are means and 95% confidence intervals (error bars). *P = .017, **P = .015, compared with mock control cells, as determined with the two-sided Mann–Whitney U test. A#1= B16F10-mARD1A²²⁵ #1; A#2 = B16F10-mARD1A²²⁵ #2; A#3 = MKN74-mARD1A²²⁵ #3; A#4 = MKN74-mARD1A²²⁵ #4.

Figure 4

Effect of mARD1A²²⁵ on lung nodule formation in mice. A and B) C57BL/6 mice were injected intravenously with B16F10, B16F10-mock, or B16F10-mARD1A²²⁵ cells. The incidence of lung nodules was determined 19 days after injection (n = 8 per group). A) Representative images of lungs (left). Numbers of surface tumor nodules (right). Results are means and 95% confidence intervals (CIs) (error bars) of two independent experiments performed with four mice per group. *P = .006, compared with the number of surface tumor nodules detected in B16F10-mock–injected mice, as determined with the two-sided Mann–Whitney U test. B) Hematoxylin and eosin staining of 4-μm lung tissue sections excised from mice described in (A). Magnification ×50. Scale bar = 500 μm. C) Kaplan–Meier survival plot of C57BL/6 mice injected with B16F10 (mean = 16.7 days, 95% CI = 15.7 to 17.7), B16F10-mock (mean = 18 days, 95% CI = 15.5 to 20.5), or B16F10-mARD1A²²⁵ cells (A#1, mean = 23.9 days, 95% CI = 22.5 to 25.3; A#2, mean = 22.7 days, 95% CI = 20.8 to 24.5) (n = 9 per group). Comparisons were made with log-rank tests. All statistical tests were two-sided.

Figure 5

Effect of mARD1A²²⁵ on VEGFA expression. A) B16F10, B16F10-mock, and B16F10-mARD1A²²⁵ cells were exposed to 1% O₂ (4H) or 21% O₂ (N) for 4 hours (top). MKN74-mock and MKN74-mARD1A²²⁵ cells were exposed to 1% O₂ (4H) or 21% O₂ (N) for 4 hours (bottom). HIF-1α and mARD1A²²⁵ protein levels were examined by Western blot analysis. Anti-Myc antibody was used for the detection of mARD1A²²⁵. B) B16F10, B16F10-mock, and B16F10-mARD1A²²⁵ cells were exposed to 21% O₂ (N) or 10 μM MG132 and 1% O₂ for 6 hours (6H + MG132) (top). MKN74, MKN74-mock, and MKN74-mARD1A²²⁵ cells were exposed to 21% O₂ (N) (not shown) or 10 μM MG132 and 1% O₂ for 4 hours (4H + MG132) (bottom). Cell lysates were immunoprecipitated with an anti–acetyl lysine antibody and subjected to Western blot analysis with an anti-HIF-1α antibody. The level of mARD1A²²⁵ was determined using an anti-Myc antibody. C) MKN74, MKN74-mock, and MKN74-mARD1A²²⁵ cells were exposed to 1% O₂ (48H) or 21% O₂ (N) for 48 hours. Reverse transcription–polymerase chain reaction analysis was performed to detect gene expression using specific primers for VEGFA and glyceraldehyde 3-phosphate dehydrogenase. D) VEGFA protein levels were quantified in conditioned media from control and mARD1A²²⁵-expressing cells treated with CoCl₂ using enzyme-linked immunosorbent assay ELISA. B16F10, B16F10-mock, and B16F10-mARD1A²²⁵ cells were cultured in the absence (N) or presence of 200 μM CoCl₂ for 4 hours (4H) (left). MKN74, MKN74-mock, and MKN74-mARD1A²²⁵ cells were cultured in the absence (N) or presence of 200 μM CoCl₂ for 48 hours (48H) (right). Three independent experiments were performed, each in two replicates. Results are means and 95% confidence intervals (error bars) derived from the means of individual experiments (n = 3). *P = .046, **P = .05, ***P = .021, compared with mock control cells, determined with the two-sided Mann–Whitney U test.

Figure 6

Effect of VEGFA treatment or HIF-1α/K532R overexpression on tube formation, migration, and invasion by mARD1A²²⁵. A) Human umbilical vein endothelial cells (HUVECs) were incubated for 6 hours on Matrigel with hypoxic-conditioned media from MKN74, MKN74-mock, and MKN74-mARD1A²²⁵ (A#3 and A#4) cells (top left) and MKN74-mARD1A²²⁵ cells treated with 10 ng/mL recombinant human VEGFA protein (A#3+VEGFA and A#4+VEGFA), and MKN74-mARD1A²²⁵ cells transfected with HIF-1α/K532R (A#3+K532R and A#4+K532R) (bottom left). Scale bar = 200 μm. Quantification of tube formation (right). Results are mean of three replicates from one experiment and their 95% confidence intervals (error bars) (n = 3). *P = .05, compared with the number of MKN74-mock cells; #P = .05, compared with the number of MKN74-mARD1A²²⁵ #3 or #4 cells, as determined with the two-sided Mann–Whitney U test. B) (Left) HUVECs and conditioned media were added to transwells, and migrating cells were counted. Data are averaged over three independent experiments. Scale bar = 100 μm. Quantification of migration (right). Results are mean of three replicates from one experiment and their 95% confidence intervals (error bars) (n = 3). *P = .05, compared with the number of MKN74-mock cells; #P = .05, compared with the number of MKN74-mARD1A²²⁵ #3 or #4 cells, as determined with the two-sided Mann–Whitney U test. C) Transwells were coated with Matrigel. Cells were stained with hematoxylin and eosin. Representative microscopic images illustrate HUVEC invasion (left). Scale bar = 100 μm. Quantification of invasion (right). Results are mean of three replicates from one experiment and their 95% confidence intervals (error bars) (n = 3). *P = .05, compared with the number of MKN74-mock cells; #P = .05, compared with the number of MKN74-mARD1A²²⁵ #3 or #4 cells, as determined with the two-sided Mann–Whitney U test. A#3 = MKN74-mARD1A²²⁵ #3; A#4 = MKN74-mARD1A²²⁵ #4; A#3+VEGFA = MKN74-mARD1A²²⁵ #3+ recombinant human VEGFA protein; A#4+VEGFA = MKN74-mARD1A²²⁵ #4+ recombinant human VEGFA protein; A#3+K532R = MKN74-mARD1A²²⁵ #3+HIF-1α/K532R; A#4+K532R = MKN74-mARD1A²²⁵ #4+HIF-1α/K532R.

Figure 7

Effect of mARD1A²²⁵ on neovascularization. A) Western blot analysis of HIF-1α and mARD1A²²⁵ protein expression in transplanted B16F10 (left) and MNK74 (right) tumor cells. B) Western blot analysis of VEGFA protein expression in transplanted B16F10 (left) and MNK74 (right) tumor cells. C) BALB/cSlc-nude mice were injected subcutaneously with MKN74, MKN74-mock, or two MKN74-mARD1A²²⁵ clones, A#3 and A#4. After 3 weeks, formalin-fixed primary tumor sections from each mouse were prepared and stained with an anti-VEGFA antibody. The brown color signifies VEGFA in primary tumors. Magnification ×400. Scale bar = 50 μm. A#3 = MKN74-mARD1A²²⁵ #3; A#4 = MKN74-mARD1A²²⁵ #4. One representative image of five per cell line is shown. D) Anti-CD34 staining of xenograft tumors derived from MKN74, MKN74-mock, and MKN74-mARD1A²²⁵ cells, indicative of endothelial cells (left). Numbers of CD34-positive vessels in primary tumors (right) expressed as means and 95% confidence intervals (error bars). Vessels from four or five fields from each of the five tumors per group were scored for quantification. *P = .043, **P = .021, compared with the number of CD34-positive vessels in tumors derived from MKN74-mock cells, as determined with the two-sided Mann–Whitney U test. Magnification ×200. Scale bar = 100 μm.

Figure 8

Effects of lysine 532 mutation in HIF-1α. A) HIF-1α protein levels were determined by Western blot analysis. B16F10-mock, B16F10-mARD1A²²⁵, and B16F10-mARD1A²²⁵-HIF-1α/K532R cells were cultured in the absence (N) or presence of 200 μM CoCl₂ for 24 hours (24H). Similar results were obtained with three independent experiments. B) VEGFA levels (in picograms per milliliter) were quantified using enzyme-linked immunosorbent assay. B16F10, B16F10-mock, B16F10-mARD1A²²⁵, and B16F10-mARD1A²²⁵-HIF-1α/K532R cells were cultured in the absence (N) or presence of 200 μM CoCl₂ for 4 hours (4H). Three independent experiments were performed, each in two replicates. Results are means and 95% confidence intervals (error bars) derived from the means of the individual experiments (n = 3). *P = .05 for all comparisons as determined with the two-sided Mann–Whitney U test. C) B16F10, B16F10-mock, B16F10-mARD1A²²⁵, and B16F10-mARD1A²²⁵-HIF-1α/K532R cells were grown to 80% confluence. A linear injury line was created at the center of the cultured monolayer by scraping with a sterile pipette tip. Cells were incubated in the presence of 200 μM CoCl₂ for 24 hours. B16F10, B16F10-mock, and B16F10-mARD1A²²⁵-HIF-1α/K532R cells were treated with 3 μg/mL anti-mouse VEGFA antibody. B16F10-mARD1A²²⁵ cells were treated with 25 ng/mL recombinant mouse VEGFA protein. Images of cell migration (left). Quantification of cell migration (right). Migration distance (in micrometers); P = .004, determined with the two-sided Mann–Whitney U test. Three independent experiments were performed, each in two replicates. Results are means and 95% confidence intervals (error bars) derived from the means of the individual experiments (n = 3). Magnification ×100. Scale bar = 100 μm. D) C57BL/6 mice were injected intravenously with B16F10-mock (n = 40 mice), B16F10-mARD1A²²⁵ (n = 31 mice) and B16F10-mARD1A²²⁵-HIF-1α/K532R cells (n = 11 mice). Lung nodule formation was determined 14 days after injection. Representative gross images of lungs are shown (left). Numbers of surface tumor nodules per mouse (right). Results are means and 95% confidence intervals (error bars). *P < .001, compared with the number of surface tumor nodules in B16F10-mock–injected mice, #P < .001, compared with the number of surface tumor nodules in B16F10-mARD1A²²⁵–injected mice, as determined with the two-sided Mann–Whitney U test. A#1 = B16F10-mARD1A²²⁵ #1; A#2 = B16F10-mARD1A²²⁵ #2; A#1/HIF K532R = B16F10-mARD1A²²⁵ #1-HIF-1α/K532R; B16F10 + anti-VEGFA = B16F10 + anti-mouse VEGFA antibody; Mock + anti-VEGFA = B16F10-Mock + anti-mouse VEGFA antibody; A#1/HIF K532R + anti-VEGFA = B16F10-mARD1A²²⁵ #1-HIF-1α/K532R + anti-mouse VEGFA antibody; A#1+VEGFA = B16F10-mARD1A²²⁵ #1 + recombinant mouse VEGFA protein; A#2+VEGFA = B16F10-mARD1A²²⁵ #2 + recombinant mouse VEGFA protein.

Figure 8

Effects of lysine 532 mutation in HIF-1α. A) HIF-1α protein levels were determined by Western blot analysis. B16F10-mock, B16F10-mARD1A²²⁵, and B16F10-mARD1A²²⁵-HIF-1α/K532R cells were cultured in the absence (N) or presence of 200 μM CoCl₂ for 24 hours (24H). Similar results were obtained with three independent experiments. B) VEGFA levels (in picograms per milliliter) were quantified using enzyme-linked immunosorbent assay. B16F10, B16F10-mock, B16F10-mARD1A²²⁵, and B16F10-mARD1A²²⁵-HIF-1α/K532R cells were cultured in the absence (N) or presence of 200 μM CoCl₂ for 4 hours (4H). Three independent experiments were performed, each in two replicates. Results are means and 95% confidence intervals (error bars) derived from the means of the individual experiments (n = 3). *P = .05 for all comparisons as determined with the two-sided Mann–Whitney U test. C) B16F10, B16F10-mock, B16F10-mARD1A²²⁵, and B16F10-mARD1A²²⁵-HIF-1α/K532R cells were grown to 80% confluence. A linear injury line was created at the center of the cultured monolayer by scraping with a sterile pipette tip. Cells were incubated in the presence of 200 μM CoCl₂ for 24 hours. B16F10, B16F10-mock, and B16F10-mARD1A²²⁵-HIF-1α/K532R cells were treated with 3 μg/mL anti-mouse VEGFA antibody. B16F10-mARD1A²²⁵ cells were treated with 25 ng/mL recombinant mouse VEGFA protein. Images of cell migration (left). Quantification of cell migration (right). Migration distance (in micrometers); P = .004, determined with the two-sided Mann–Whitney U test. Three independent experiments were performed, each in two replicates. Results are means and 95% confidence intervals (error bars) derived from the means of the individual experiments (n = 3). Magnification ×100. Scale bar = 100 μm. D) C57BL/6 mice were injected intravenously with B16F10-mock (n = 40 mice), B16F10-mARD1A²²⁵ (n = 31 mice) and B16F10-mARD1A²²⁵-HIF-1α/K532R cells (n = 11 mice). Lung nodule formation was determined 14 days after injection. Representative gross images of lungs are shown (left). Numbers of surface tumor nodules per mouse (right). Results are means and 95% confidence intervals (error bars). *P < .001, compared with the number of surface tumor nodules in B16F10-mock–injected mice, #P < .001, compared with the number of surface tumor nodules in B16F10-mARD1A²²⁵–injected mice, as determined with the two-sided Mann–Whitney U test. A#1 = B16F10-mARD1A²²⁵ #1; A#2 = B16F10-mARD1A²²⁵ #2; A#1/HIF K532R = B16F10-mARD1A²²⁵ #1-HIF-1α/K532R; B16F10 + anti-VEGFA = B16F10 + anti-mouse VEGFA antibody; Mock + anti-VEGFA = B16F10-Mock + anti-mouse VEGFA antibody; A#1/HIF K532R + anti-VEGFA = B16F10-mARD1A²²⁵ #1-HIF-1α/K532R + anti-mouse VEGFA antibody; A#1+VEGFA = B16F10-mARD1A²²⁵ #1 + recombinant mouse VEGFA protein; A#2+VEGFA = B16F10-mARD1A²²⁵ #2 + recombinant mouse VEGFA protein.