Mitochondrial p32 is upregulated in Myc expressing brain cancers and mediates glutamine addiction

Abstract

Metabolic reprogramming is a key feature of tumorigenesis that is controlled by oncogenes. Enhanced utilization of glucose and glutamine are the best-established hallmarks of tumor metabolism. The oncogene c-Myc is one of the major players responsible for this metabolic alteration. However, the molecular mechanisms involved in Myc-induced metabolic reprogramming are not well defined. Here we identify p32, a mitochondrial protein known to play a role in the expression of mitochondrial respiratory chain complexes, as a critical player in Myc-induced glutamine addiction. We show that p32 is a direct transcriptional target of Myc and that high level of Myc in malignant brain cancers correlates with high expression of p32. Attenuation of p32 expression reduced growth rate of glioma cells expressing Myc and impaired tumor formation in vivo. Loss of p32 in glutamine addicted glioma cells induced resistance to glutamine deprivation and imparted sensitivity to glucose withdrawal. Finally, we provide evidence that p32 expression contributes to Myc-induced glutamine addiction of cancer cells. Our findings suggest that Myc promotes the expression of p32, which is required to maintain sufficient respiratory capacity to sustain glutamine metabolism in Myc transformed cells.

Figure 1. Upregulation of p32 in malignant brain tumors

Matched normal and malignant brain tissue (A), and a human brain tumor array (B) were stained with a polyclonal anti-p32 antibody. P32 positive staining for each core was quantified using Aperio software (B left panel). The array contains over 100 cases of gliomas, including both astrocytoma and medulloblastoma subtypes, and normal tissue. The two-tailed Student t test was used for statistical analysis. Significant differences are indicated using the standard Michelin Guide scale (p < 0.05 (*), significant; p < 0.01 (**), highly significant; p < 0.001 (***) extremely significant. Representative cores, at low (10x) and high (400x) magnification, of normal cerebellum and astrocytomas of different malignancy grade (GR) are indicated on the right panel of Figure 1B. (C) Increasing expression of p32/C1QBP correlates with poor survival. Shown is a survival plot for mesenchymal subtype of GBM based on increasing expression of p32/C1QBP using Affymetrix HT_HG-U133A. Plot generated using the Glioblastoma Bio Discovery Portal (http://robtcga.nci.nih.gov/#genes).

Figure 2. Correlation between p32 and Myc expression in human gliomas and glioma cell lines

(A) Correlation between Myc and p32 expression in medulloblastoma human tumors. Left panel. Heat map of p32, c-Myc and N-Myc for each medulloblastoma subgroup (c1 through c6) and an additional ATRT subgroup (atypical teratoid/rhabdoid tumors) [23], C5/c1 Medulloblastoma subgroup, characterized by a Myc activation signature [23], exhibits high expression of p32. WNT, Wingless signaling pathway, SHH Sonic Hedgehog signaling pathway. nl cbl, Normal cerebellum samples. (A)-right panel and (B): correlation between p32 and Myc expression in a medulloblastoma array ((A)-right panel) and a mixed glioma (B) as indicated by the % of p32 and Myc positive staining for each core of the arrays. Sequential slides of each array were stained separately with polyclonal anti p32 and c-Myc antibodies. The % of p32 and Myc positive staining for each core was quantified using Aperio software. The Pearson correlation coefficient (r) of linear regression was calculated using data sets deprived of samples expressing low Myc (<15% of staining) but moderate-high p32 (>15% of staining). The immunohistochemistry images (200x magnification) show representative glioma cores (red arrows in graph) exhibiting correlation between Myc and p32 expression. (C) qPCR analysis of p32 and Myc expression in established glioma cell lines (red) and patient derived glioma stem cells (blue) as compared to normal astrocytes (black). The bars shown are normalized to an internal β-actin control and represent the mean ± SEM of at least three independent experiments.

Figure 3. Myc promotes p32 expression

Left panel: qPCR analysis of target genes from total RNA isolated from MRC5 MycER cells treated with 250 nM OHT or vehicle (EtOH) for 24 hrs. Ubiquitin (Ubb) is included as a negative control while cyclin D2 (CycD2) and glutaminase (GLS1) are positive controls. The bars shown are normalized to an internal β-actin control and represent the mean ± SD of at least three independent experiments. Fold of each gene induction over control treatment is indicated. Middle panel: immunofluorescence staining of Myc and p32 24 hrs post-OHT treatment. The immunoblot on the right shows upregulation of p32 protein upon Myc induction. Tubulin was used as loading control.

Figure 4. Myc binds to the p32 promoter

(A) Schematic of the human p32 promoter sequence starting from 3 kb upstream of exon 1 to exon 6 of the c1qbp gene. Exons (ex) are represented by black boxes and the E box is indicated with a vertical bar. Horizontal bars indicate the regions amplified for scanning ChIP analysis. E2 is the p32 promoter region containing the E-box and amplified by conventional PCR (Figure 3B). (B and C) A ChIP assay was performed on SF188 cells with anti Myc antibody and IgG as a control. Precipitated chromatin was PCR-amplified using E2 primers (B). Quantitative PCRs were performed to amplify and quantify E1, ex1, in1, and in3 promoter regions. Shown are averages with standard deviations of triplicate independent experiments. Binding to amplicons is shown as a percentage of total input DNA plotted relative to the signal obtained from IgG precipitation. (D) Quantification of MycER binding to the p32 promoter after addition of OHT. MRC5 MycER cells were serum starved for 24 hrs and either treated with vehicle (EtOH) or 250nM OHT for 4 hours. Subsequently ChIP assays were conducted as described in (B). The bar graph presented is indicative of three independent experiments.

Figure 5. Attenuation of p32 expression reduces glioma cell proliferation and tumor growth

(A) Upper panels: colony formation assay of control and p32 knockdown stable cell lines. The graphs represent the mean ± SD of at least three independent colony formation assays each performed in triplicate (middle panel). Lower: Western blot analysis demonstrating p32 stable knockdown in the indicated glioma cell lines. (B) Microscopic analysis of p32 knockdown and control GBM4 and GBM5 glioma stem cell-like neuropheres. Right panel-western blot showing efficient p32 knockdown in the indicated glioma stem cell-like cells. (C) Tumor growth properties of U87 cells (left panel) and GBM4 glioma stem cell-like cells (right panel) infected with viruses encoding control or p32 kd shRNA. The graphs represent the mean ± SEM (p < 0.001) of tumor volumes as a function of time. The lower panels show pictures of tumors from each group (n = 14 or n = 11) at the end point.

Figure 6. Loss of p32 sensitizes cells to glucose withdrawal and reduces sensitivity to glutamine deprivation

(A) p32 knockdown cells are more sensitive to the glycolytic inhibitor 2-DG. SF188 p32 knockdown cells were plated in low glucose media (2.5 mM) and 3mM 2-DG. After 18 h cell viability was determined by trypan blue exclusion. Data is the average of three experiments ± SD. (B) SF188 Control or p32 knockdown cells were plated in glutamine or glucose free media. Cell viability was determined by trypan blue exclusion. Each time point is the mean of three experiments ± SD. (C) Control (left) and p32 knockdown (right) SF188 cells were plated in complete media (25mM glucose and 4mM glutamine) and subsequently cultured in either complete media or media with 2mM glucose or without glutamine. Cell viability was determined at the indicated time points by FACS analysis of PI stained cells. Upper panel-Microscopic analysis of p32 knockdown and control SF188 cells after 2 days of growth in the indicated tissue culture media. The shown result is representative of three independent experiments.

Figure 7. P32 is required for Myc-dependent glutamine metabolism

(A) Immunoblot detecting expression of p32, Myc, Erk, phospho Erk (p-Erk), Akt, and phospho-Akt (p-Akt) in lysates from Control and p32 knockdown SF188 cells. (B) Myc was knocked down in both SF188 ShRNA Control and ShRNA p32 cells. Cells were then plated in glutamine free media and 24 hrs later cell viability was determined by trypan blue exclusion. Data shown is the mean of three experiments ± SD. (C) Model for p32 role in cancer cell metabolism downstream of Myc. Myc promotes the expression of p32, which helps maintain OXPHOS and glutamine addiction. Loss of p32 results in a dependence on glucose. Loss of both Myc and p32 inhibits cell growth.