NDRG4 is required for cell cycle progression and survival in glioblastoma cells

Abstract

NDRG4 is a largely unstudied member of the predominantly tumor suppressive N-Myc downstream-regulated gene (NDRG) family. Unlike its family members NDRG1-3, which are ubiquitously expressed, NDRG4 is expressed almost exclusively in the heart and brain. Given this tissue-specific expression pattern and the established tumor suppressive roles of the NDRG family in regulating cellular proliferation, we investigated the cellular and biochemical functions of NDRG4 in the context of astrocytes and glioblastoma multiforme (GBM) cells. We show that, in contrast to NDRG2, NDRG4 expression is elevated in GBM and NDRG4 is required for the viability of primary astrocytes, established GBM cell lines, and both CD133(+) (cancer stem cell (CSC)-enriched) and CD133(-) primary GBM xenograft cells. While NDRG4 overexpression has no effect on cell viability, NDRG4 knockdown causes G(1) cell cycle arrest followed by apoptosis. The initial G(1) arrest is associated with a decrease in cyclin D1 expression and an increase in p27(Kip1) expression, and the subsequent apoptosis is associated with a decrease in the expression of XIAP and survivin. As a result of these effects on cell cycle progression and survival, NDRG4 knockdown decreases the tumorigenic capacity of established GBM cell lines and GBM CSC-enriched cells that have been implanted intracranially into immunocompromised mice. Collectively, these data indicate that NDRG4 is required for cell cycle progression and survival, thereby diverging in function from its tumor suppressive family member NDRG2 in astrocytes and GBM cells.

FIGURE 1.

Characterization of NDRG4 expression in GBM. A, representative IHC images of NDRG4 expression in normal human cortex and GBM sections. B, real-time RT-PCR analysis of two independent lots of normal primary human astrocytes (NHA 1 and 2) and cultured cells derived from three human GBM xenograft samples (GBM1, 459 cells; GBM2, T4105 cells; GBM3, T3559 cells) and subsequently separated into GBM CSC-enriched (+) and nonstem cell (−) populations by CD133 status. C, nucleocytoplasmic fractionation of U251 cells and subsequent Western analysis. Smad1 served as a positive control for leptomycin B (LMB) treatment, and lamin A/C and α-tubulin were used as controls for nuclear (nuc) and cytoplasmic (cyto) fractions, respectively. D, real-time RT-PCR analysis and Western analysis of U251 cells partially synchronized with a double thymidine block and released for the indicated amounts of time. The percentage of cells in each phase of the cell cycle is indicated. γ-Tubulin was used as a loading control for Western analysis.

FIGURE 2.

Knockdown of NDRG4, but not NDRG2, decreases cell viability in normal primary human astrocytes and GBM cells. A, Western analysis and MTS cell viability analysis of U251 control (sh-NT) and NDRG4 knockdown (sh-NDRG4-a, sh-NDRG4-b) cell lines. γ-Tubulin was used as a loading control for Western analysis. B, Western analysis and MTS cell viability analysis of U251 control (sh-NT) and NDRG2 knockdown (sh-NDRG2-a, sh-NDRG2-b) cell lines. C, real-time RT-PCR analysis and MTS cell viability analysis of control and NDRG4 knockdown normal primary human astrocytes (NHA). NDRG4 knockdown was confirmed by RT-PCR instead of Western analysis due to limited cell numbers.

FIGURE 3.

Overexpression of NDRG2, but not NDRG4, decreases U251 cell viability. A, Western analysis and MTS cell viability analysis of U251 control (Vec con) and NDRG2 overexpression cells. γ-Tubulin was used as a loading control for Western analysis. B, Western analysis and MTS cell viability analysis of U251 control (Vec con) and NDRG4 overexpression cells. Two isoforms of NDRG4 were overexpressed: the B isoform (NDRG4(B)) and the H isoform (NDRG4(H)). γ-Tubulin was used as a loading control for Western analysis.

FIGURE 4.

NDRG4 knockdown causes G₁ arrest and subsequent apoptosis in U251 cells. A, cell cycle analysis following partial synchronization with a double thymidine block. Control (sh-NT) and NDRG4 knockdown (sh-NDRG4-b) cells were analyzed at the indicated time points following release from block. B, cell cycle analysis following infection with control virus (sh-NT) or NDRG4 knockdown virus (sh-NDRG4-b) for the indicated number of days. C, Western analysis of G₁ cell cycle progression markers in control cells (sh-NT) and NDRG4 knockdown cells (sh-NDRG4-a, sh-NDRG4-b) following infection with control virus or NDRG4 knockdown virus for 1 day, 2 days, and 3 days. γ-Tubulin was used as a loading control. D, percentage of annexin V-positive cells in NDRG4 knockdown (sh-NDRG4-b) and control (sh-NT) cell populations at day 5 after infection. E, Western analysis of apoptotic markers in control cells (sh-NT) and NDRG4 knockdown cells (sh-NDRG4-a, sh-NDRG4-b) following infection with control virus or NDRG4 knockdown virus for 1 day, 2 days, and 3 days. γ-Tubulin was used as a loading control. F, caspase-3/7 activity, as determined by colorimetric assay, in knockdown (sh-NDRG4-b) and control (sh-NT) cell populations at day 5 after infection.

FIGURE 5.

NDRG4 knockdown-induced cell cycle arrest and apoptosis in U251 cells are not due to mitotic defects or DNA damage. A, Western analysis of control (sh-NT), NDRG4 knockdown (sh-NDRG4-b), and NDRG1 knockdown (sh-NDRG1-a) cells. γ-Tubulin was used as a loading control. B, flow cytometry analysis of cells that were first infected with control, NDRG1, or NDRG4 knockdown virus for 24 h and then treated with nocodazole (+Noco) or left untreated (−Noco) for 48 h. DNA content (2N, 4N, and 8N) is represented on the x axis. C, comet assays in control (sh-NT) and NDRG4 knockdown (sh-NDRG4-a, sh-NDRG4-b) cells. Cells exposed to 10 Gy of ionizing radiation (IR) were used as a positive control for DNA damage-induced “comet tails” (indicated by arrows).

FIGURE 6.

Knockdown of NDRG4 decreases viability and self-renewal of both CD133⁺ and CD133⁻ GBM cells. A, real-time RT-PCR analysis and MTS cell viability analysis of control (sh-NT) and NDRG4 knockdown (sh-NDRG4-a, sh-NDRG4-b) GBM CSC-enriched CD133⁺ cells isolated from a T4105 human GBM xenograft. Knockdown was confirmed by RT-PCR instead of Western analysis due to limited cell numbers. B, average number of neurospheres formed per well at day 4 of neurosphere formation assay. CD133⁺ cells were initially plated out at a density of 100 cells per well. Representative images of neurospheres are shown. C, percentage of wells containing neurospheres over time. CD133⁺ cells were initially plated out at densities of 100 cells per well. D, real-time RT-PCR analysis and MTS cell viability analysis of control and NDRG4 knockdown CD133⁻ GBM cells isolated from a T4302 human GBM xenograft. Knockdown was confirmed by RT-PCR instead of Western analysis because of limited cell numbers.

FIGURE 7.

Knockdown of NDRG4 decreases growth of GBM tumor xenografts in immunocompromised mice. A, percentage of mice that remained free of neurological functional impairment over time after being injected intracranially with U251 control cells (sh-NT) or NDRG4 knockdown cells (sh-NDRG4-a, sh-NDRG4-b). Neurological functional impairment included ataxia, lethargy, seizures, and inability to feed. B, percentage of mice that remained free of neurological functional impairment over time after being injected intracranially with control (sh-NT) or NDRG4 knockdown (sh-NDRG4-a) CD133⁺ GBM CSC-enriched cells. CD133⁺ cells were originally isolated from a T3559 human GBM xenograft. C, representative H&E and Ki-67 staining images of brain sections from mice injected with control and NDRG4 knockdown CD133⁺ cells.