Nox-4-dependent nuclear H2O2 drives DNA oxidation resulting in 8-OHdG as urinary biomarker and hemangioendothelioma formation

Abstract

Hemangioendotheliomas are classified as endothelial cell tumors, which are the most common soft tissue tumors in infants. In a murine model of hemangioendothelioma, we previously showed that MCP-1 is required for its development and that the expression of MCP-1 in EOMA cells is redox sensitive. Here, we sought to identify the source of oxidants that drive hemangioendothelioma formation. Seven known isoforms exist of the catalytic subunit gp91. Only the nox-4 isoform of gp91 was present in EOMA cells, in contrast with non-tumor-forming murine endothelial cells that contained multiple forms of nox. Nox-4 knockdown markedly attenuated MCP-1 expression and hemangioendothelioma formation. We report that in EOMA cells, nox-4 is localized such that it delivers H2O2 to the nuclear compartment. Such delivery of H2O2 causes oxidative modification of DNA, which can be detected in the urine of tumor-bearing mice as 8-hydroxy-2-deoxyguanosine. Iron chelation by in vivo administration of deferoxamine improved tumor outcomes. The current state of information connects nox-4 to MCP-1 to form a major axis of control that regulates the fate of hemangioendothelioma development in vivo.

Figures

FIG. 1.

H2O2, not the superoxide form of ROS, is abundant in EOMA cells. (a) EOMA cells and non–tumor-forming murine aortic endothelial (MAE) cells were treated with 5-(and-6)-carboxy-2′,7′ dichlorofluorescein diacetate (DCF), and oxidation was detected with flow cytometry. (b) Dihydroethidium (DHE) specifically detects superoxide production with red fluorescence (570 nm). Low DHE oxidation in EOMA cells indicates that the contribution of superoxide to DCF oxidation is low. (c) Amplex red assay (Invitrogen) showed that H2O2 levels were significantly elevated in EOMA cells compared with nontumor MAE cells. **p < 0.01.

FIG. 2.

Nox-4 is the predominant homologue of gp91present in EOMA cells. Real-time PCR was done on mRNA from (a) tumor-forming EOMA cells or (b) non–tumor-forming murine aortic endothelial (MAE) cells. Both cells lines are transformed murine endothelial cell lines that grow readily in culture, but only the EOMA cells make tumors when injected into mice. (c) The abundance of nox-4 mRNA is 68-fold higher in EOMA compared with MAE cells. Error bars are too small to be visible. **p < 0.01.

FIG. 3.

Nox-4 contributes to H2O2 levels in EOMA cells. (a) Knockdown of nox-4 activity resulted in a significant decrease in DCF oxidation. (b) H2O2 levels as determined by Amplex red assay. Knockdown of nox-4 significantly decreased H2O2 levels. **p < 0.01.

FIG. 4.

Nox-4 knockdown arrested kaposiform hemangioendothelioma growth in vivo. EOMA cells were transduced with a nonreplicating lentiviral vector containing shRNA for either nox-4 or a nontargeting negative control (scramble) sequence. Puromycin selection was performed to select the clone with the most efficient knockdown of nox-4, as determined by (a) real-time PCR and (b) Western blot. Significant knockdown of nox-4 gene and protein was achieved. **p < 0.01. (c) EOMA cells stably transfected with nox-4 or control/scrambled shRNA were injected subcutaneously into 129P/3 mice. Tumors were harvested at 7 days after EOMA cell injection. Mice injected with EOMA cells transduced with nox-4 shRNA had significantly smaller tumors compared with those receiving EOMA cells with control shRNA. **p < 0.01.

FIG. 5.

MCP-1 expression in EOMA cells is nox-4 dependent. (a) Real-time PCR confirmed knockdown of the nox-4 gene. (b) Western blot demonstrates decreased levels of nox-4 protein in knockdown cells. (c) Real-time PCR for MCP-1 mRNA shows that knockdown of nox-4 attenuates MCP-1 gene expression. (d) ELISA for MCP-1 protein shows that knockdown of nox-4 attenuates MCP-1 protein expression. **p < 0.01.

FIG. 6.

Nox-4 knockdown disrupts angiogenic tube formation of EOMA cells. To determine the contribution of nox-4 to the angiogenic property of EOMA cells, an in vitro assay was performed by using EOMA cells transduced with either (a) control shRNA or (b) nox-4 shRNA and seeded on reduced growth factor Matrigel. Cells were stained with calcein-AM, and the area within the formed tubes was quantitated by using AxioVision LE software. (c) Tube formation by EOMA cells was significantly blunted in response to nox-4 knockdown. **p < 0.01. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article at www.liebertonline.com/ars).

FIG. 7.

Nox-4 contributes to EOMA cell proliferation and progression through the G2-M cell-cycle checkpoint. (a) BrdU assay showed a significant attenuated proliferation of nox-4 knockdown cells. (b) Nox-4 knockdown caused cells to accumulate in the G2-M transition phase of the cell cycle. **p < 0.01.

FIG. 8.

Nox-4 is localized in the nuclear membrane region of EOMA cells. (a) Control shRNA or (b) nox-4 shRNA. (c) Quantitative comparison of perinuclear green fluorescence intensity was done by using Zeiss AxioVision software. Analysis was performed on three cells transduced with control shRNA and compared with results from three cells transduced with nox-4 shRNA. Loss of signal in knockdown cells confirms the specificity of the antibody used for nox-4. **p < 0.01. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article at www.liebertonline.com/ars).

FIG. 9.

Nox-4 delivers H2O2 directly into the nucleus of EOMA cells. Catalase (0.1 mU/cell) tagged with Qdot was microinjected to the cytoplasm [(a) catalase; (b) DCF; (c) merged image 30 s after injection; (d) catalase; (e) DCF; (f) merged image 3 min after injection) or the nucleus of an EOMA cell (g) catalase; (h) DCF; (i) merged image 30 s after injection; (j) catalase; (k) DCF; (l) merged image 3 min after injection]. Three minutes after microinjection, catalase-injected cytoplasm (e) and nucleus (k) showed lower DCF intensity (m). Results are expressed as mean ± SD. *p < 0.05; white scale bar, lower right hand corner = 30 μm. Catalase-sensitive DCF signal in the nucleus confirms that the DCF oxidation signal noted was generated by H2O2. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article at www.liebertonline.com/ars).

FIG. 10.

Urinary excretion of oxidatively modified DNA from tumor-bearing mice and the effect of treatment with iron chelator DFO. (a) Pooled urine samples were collected on specified days from mice injected with EOMA cells. 8-Hydroxy-2'-deoxyguanosine (8-OHdG) was detected in the urine. No difference in 8-OHdG levels at day –3 and day 0, but days 3, 5, and 7 all had significantly elevated levels compared with baseline/day 0 values (**p < 0.01, ANOVA). To determine whether ferrous iron contributed to oxidative modification of DNA, tumor-bearing mice were treated with deferoxamine (DFO), an FDA-approved drug that functions as an iron chelator. Mice received daily intraperitoneal (IP) injection of either 0.1 ml of PBS (vehicle control) or DFO (100 mg/kg) beginning at day –1 relative to EOMA cell injection. KHE samples were collected at 7 days after EOMA cell injection, and calipers were used to measure (b) volume (length × width × height) and (c) mass. Mice treated with DFO had significantly smaller KHE for both volume and weight. **p < 0.01.

FIG. 11.

The Nox-4-MCP1 axis of control for hemangioendothelioma formation. Nox-4–derived hydrogen peroxide drives MCP-1 expression, which was previously shown to be a key driver of HE formation. Nox-4–derived nuclear hydrogen peroxide, through an Fe2+-dependent mechanism, causes oxidative modification of DNA, forming 8-OHdG. 8-OHdG clears through the urine of tumor-bearing mice and may be considered a biomarker candidate. Inset: Appearance of hemangioma or hemangioendothelioma tumor in children (informed consent obtained). (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article at www.liebertonline.com/ars).