Mutations in mitochondrial DNA polymerase-gamma promote breast tumorigenesis

Abstract

Decreased mitochondrial oxidative phosphorylation (OXPHOS) is one of the hallmarks of cancer. To date, the identity of nuclear gene(s) responsible for decreased OXPHOS in tumors remains unknown. It is also unclear whether mutations in nuclear gene(s) responsible for decreased OXPHOS affect tumorigenesis. Polymerase-gamma (POLG) is the only DNA polymerase known to function in human mitochondria. Mutations in POLG are known to cause mitochondrial DNA (mtDNA) depletion and decreased OXPHOS, resulting in mtDNA depletion syndrome in humans. We therefore sequenced all coding exons (2-23) and flanking intron/splice junctions of POLG in breast tumors. We found that the POLG gene was mutated in 63% of breast tumors. We identified a total of 17 mutations across the POLG gene. Mutations were found in all three domains of the POLG protein, including T251I (the exonuclease domain), P587L (the linker region) and E1143G (the polymerase domain). We identified two novel mutations that include one silent (A703A) and one missense (R628Q) mutation in the evolutionarily conserved POLG linker region. In addition, we identified three novel mutations in the intronic region. Our study also revealed that mtDNA was depleted in breast tumors. Consistently, mutant POLG, when expressed in breast cancer cells, induced a depletion of mtDNA, decreased mitochondrial activity, decreased mitochondrial membrane potential, increased levels of reactive oxygen species and increased Matrigel invasion. Together, our study provides the first comprehensive analysis of the POLG gene mutation in human cancer and suggests a function for POLG (1) in decreased OXPHOS in cancers and (2) in promoting tumorigenicity.

Figure 1. POLG D1135A mutant depletes mtDNA and promotes tumorigenicity in breast cancer cells

POLG D1135A cDNA was cloned in the tetracycline inducible plasmid pTRE-Tight-BI-AcGFP1. A bicistronic promoter provided the expression of both GFP and POLG simultaneously. Transfected MCF7 Tet-on Advanced cells were treated with 1000 ng/ml doxycycline for up to 5 day and were sorted by FACS. A) GFP fluorescence was used as a guide to sort cells expressing the mutant POLG gene. Mean fluorescent intensity was determined on the FL1 channel of a FACSCalibur flow cytometer. Data represent geometric mean fluorescence intensity. B). MtDNA index in MCF7 Tet-on Advanced cells expressing POLG D1135A. The ratio of mtDNA to nuclear DNA was used as an index for measuring the mtDNA content C). DHE oxidation of MCF7 Tet-on Advanced cells containing POLG D1135A was measured. Mean fluorescence intensity of each treatment group was normalized to day 0 and expressed as fold DHE oxidation + 1 SD. D). Mitochondrial membrane potential was measured by TMRE fluorescence. Data represents mitochondrial membrane potential as a percent of control (day 0) + 1 SD. E). Mitochondrial respiratory activity was measured by the rate of resazurin reduction. F).Tumorigenicity was measured by Matrigel invasion assay.

Figure 2. POLG mutations in breast tumors and breast cancer cell lines

A) Intron/Splice variants in the POLG genome; B) Mutations in the POLG protein with amino acid change. Green and red arrows indicate the novel variants and disease-associated mutations/polymorphisms, respectively. The grey and orange boxes indicate the novel silent and missense mutations, respectively; C) The amino acid conservation at the mutant residue of R628Q, a novel missense mutation observed in the linker region.

Figure 3. Decreased mtDNA content

A) in breast tumor samples and B) in breast cancer cell lines. The ratio of mtDNA to nuclear DNA was used as an index for measuring the mtDNA content (described in material methods).

Figure 4. Breast tumor POLG mutations lead to increased tumorigenicity

Matrigel invasion of MCF7 Tet-on Advanced cells expressing representative mutations in POLG identified in primary breast tumors. Cell were treated with 1000 ng/ml doxycycline and sorted for GFP fluorescence. Cells were grown in the presence of doxycycline for 5 day and the Matrigel invasion was carried out. Data represents mean percentage of invading cells normalized to negative vector control ± 1 SEM.