Involvement of Mitochondrial DNA Sequence Variations and Respiratory Activity in Late Complications following Radiotherapy Free

Purpose: Mitochondria and ionizing radiation overlap in a number of features; for instance, both generate harmful reactive oxygen species, and that radiation can induce cell death through the intermediary of mitochondria. Because a number of genetic variations in nuclear genes are frequently associated with response to cancer treatment, the aim of this case-control study was to test the hypothesis that mitochondrial DNA (mtDNA) genetic variations can contribute to patient-to-patient variability in normal tissue response to radiotherapy.

Experimental Design: Thirty-two nasopharyngeal carcinomas patients treated with definitive radiotherapy were included. The grade (G) of s.c. and deep tissue fibrosis was scored according to the Radiation Therapy Oncology Group/European Organization for Research and Treatment of Cancer grading system. Coding and RNA mtDNA (between 611 and 15,978 bp) were sequenced, and genetic variations were scored. Mitochondrial respiratory activity was measured by resazurin reduction assay.

Results: Data showed a significantly (P = 0.003) higher number of nonsynonymous genetic variations in the radiosensitive (G₂-G₃; 16 patients) as compared with the control (G₀-G₁; 16 patients) groups. The nonsynonymous A10398G variation in the ND3 gene was significantly associated with fibrotic reaction (P = 0.01). The radiosensitive patients had a 7-fold (95% confidence interval, 1.16-51.65) higher risk of developing moderate to severe fibrosis (G₂-G₃) following radiotherapy. This was significantly correlated with lower mitochondrial respiratory activity (P = 0.001).

Conclusion: Mitochondria contribute to radiation sensitivity, and genetic variations can be associated with late reactions to radiotherapy. Predictive markers of radiosensitivity should take into account mtDNA genetic variations in addition to variations in nuclear genes. (Clin Cancer Res 2009;15(23):7352–60)

Radiation doses used to treat cancer patients are limited by the tolerance of surrounding normal tissues. These tolerance doses vary not only among tissues but also among individuals (1). Radiosensitive patients are at risk of developing severe irreversible late damage to normal tissue, months to years after radiotherapy (2). Late effects are caused by cell depletion of the tissue renewal units, leading to atrophy, fibrosis, and ischemic tissue necrosis that compromise the quality of life of cancer survivors (1). Therefore, much interest in the radiosensitivity of normal tissue has emerged and raised the possibility of developing biomarkers to predict radiation response and individualize treatment planning to improve the therapeutic gain (3, 4). Apart from germline mutations in specific genes such as ATM, NBS1, MRE11, and Ligase IV (5–7), genetic polymorphic variations in a number of nuclear genes have been associated with radiosensitivity (8–12).

In addition to reducing cell survival, ionizing radiation has been shown to simultaneously induce premature differentiation of "fibroblast" cells in culture (13). The possible consequence of this accelerated biological aging (different stages of differentiation leading to cellular senescence) is the decline in cellular reserve necessary to maintain tissue function following irradiation, and thus the expression of radiation-induced complications in cancer patients treated with radiotherapy. It has been suggested that the stage of differentiation of fibroblasts could be correlated to the degree of fibrosis, a late reaction to radiotherapy, in patients who had undergone postmastectomy irradiation (14). On the other hand, several diseases associated with senescence seem to be the direct result of cells containing dysfunctional mitochondria (15). Therefore, it seems tempting to speculate that mitochondria could influence radiosensitivity and mitochondrial DNA (mtDNA) genetic polymorphic variations and/or mutations may be associated with the development of normal tissue complications following radiotherapy.

The typical human cell has several hundred mitochondria. The entire mitochondrial genome is circular and totals 16.6 kb in length. It contains the structural genes for 13 proteins of the respiratory chain and ATP synthase, as well as mitochondrial 12S and 16S rRNAs and 22 tRNAs. Mitochondria are the sites of three important processes: energy conversion (ATP), production of reactive oxygen species (ROS), and initiation of cell death (apoptosis or necrosis). Dysfunctional mitochondria can impair energy conversion and increase ROS production. MtDNA sequence variations leading to mitochondrial respiratory dysfunction can be functionally detected in life cells by resazurin assay (16). Ionizing radiation increases oxidative mtDNA damage, and the lack of efficient repair can contribute to radiation sensitivity along with the damage induced in nuclear DNA (17). The aim of this case-control study was to test the hypothesis that genetic variations in coding and RNA mtDNA can influence radiosensitivity and may be associated with late normal tissue complications to radiotherapy. Thirty-two head and neck cancer patients treated with radiotherapy for nasopharyngeal carcinomas were included. The grade of s.c. and deep tissue fibrosis was scored according to the Radiation Therapy Oncology Group (RTOG)/European Organization for Research and Treatment of Cancer (EORTC) grading system. Coding and RNA mtDNA (between 611 and 15,978 bp) were sequenced, genetic variations were scored, and mitochondrial respiratory function was measured and compared among patients.

A total of 32 nasopharyngeal carcinoma patients were included from an ongoing prospective study to evaluate treatment outcome for head and neck cancer. Patients were selected according to the reactions of their normal tissues to radiotherapy (grade of fibrosis, see below). The King Faisal Specialist Hospital and Research Centre (KFSH&RC) institutional review board approved the study, and all patients signed an informed consent. The patients were treated in the Radiation Oncology Section at the KFSH&RC. The treatment was standardized and involved definitive radiotherapy with no surgery. Locally advanced stages also received standard neoadjuvant and concurrent chemotherapy (cisplatinum and epirubicin; ref. 18). Total radiation dose to the upper neck was 66 Gy (positive lymph nodes) delivered using 2 Gy per fraction per day over 6.5 wk. Where possible, some patients (n = 10) received a boost of two additional fractions to the nasopharynx to bring the dose received to 70 Gy in 7 wk. The radiation treatment was planned using computed tomography–based three-dimensional conformal technique detailed previously (18). The grade (G) of s.c. and deep tissue fibrosis, a late radiation-induced side effect, was scored in the neck jointly by the participating physicians at the follow-up visit according to the RTOG/EORTC grading system (19). For group comparison, the patients with no or minimal fibrotic reactions (G₀-G₁) were referred to as control (16 patients) and the patients with moderate to severe fibrosis (G₂-G₃) were referred to as the radiosensitive group (cases, 16 patients). An effort was made to balance selected patients between these two groups taking into account the total radiation dose and the chemotherapy received.

Genomic and mtDNA were extracted from cultured fibroblasts using Puregene DNA Purification Kit (Gentra System) according to the manufacturer's instruction. The method of establishing fibroblast cell cultures from punch skin biopsies from each patient was described before (9). All the coding and RNA regions between 611 and 15,978 bp of the mitochondrial genome were amplified in 20 overlapping fragments using previously described primers (20). Each fragment was separately amplified by thermal cycling (initial denaturing at 95°C for 1 min followed by 35 cycles of denaturing at 95°C for 30 s, primer annealing for 45 s at 61°C, and primer extension at 72°C for 2 min with a final extension at 72°C for 5 min) using HotStarTaq DNA polymerase (Qiagen) and 50 ng template DNA in 25 μL volume with standard reaction conditions. Amplified fragments were directly sequenced with dye termination chemistry (BigDye Terminator ver. 3.1 Cycle Sequencing kit, Applied Biosystems, Inc.), and samples were run on a sequencer (Prism 3100, Applied Biosystems Inc.). Sequencing results were aligned to the revised Cambridge reference sequence (GenBank accession number AC_000021.2), and mtDNA genetic variations were genotyped using SeqManII sequence analysis software (DNASTAR Inc.).

Resazurin is a redox-active blue dye that becomes pink when reduced. It competes with oxygen for electrons in the energy production pathway, and a change in fluorescence reflects respiration. Patients' fibroblasts were trypsinized and put into suspension in PBS at a concentration of 5 × 10⁵ cells/mL. For mitochondrial respiratory activity (MRA), triplicates of 10⁵ cells in 200 μL were incubated with 6 μmol/L resazurin (Sigma-Aldrich) for 4 h, and the fluorescence intensity resulting from resazurin reduction was measured by spectrofluorimetry. Protein concentration was measured in the remaining fibroblasts. PBS was removed and the pellet volume was estimated and lysed in 2 to 3 volumes of RIPA buffer (50 mmol/L Tris-HCl, pH 7.2, 150 mmol/L NaCl, 1% Nonidet P-40, 1% sodium deoxycholate, 0.1% SDS) containing the protease inhibitor phenylmethanesulfonyl fluoride. Lysis was carried out on ice for 20 min with frequent vortexing. The cell lysate was cleared by centrifugation and the supernatant was used as total cellular protein extract. Protein concentration was determined by a modified Bradford protein assay (BioRad). MRA was calculated by normalizing the resazurin fluorescence intensity for protein concentration and background activity, as described previously (16).

Testing for significant differences between two groups of continuous outcomes variables was carried out using Student's t-test. When the conditions for using the t-test were not satisfied (normality of distribution and equality of variance), the nonparametric Mann-Whitney rank sum test was used. The association between grade of fibrosis and mtDNA genetic variations was measured by the odds ratio and its 95% confidence interval. Significance of odds ratio was assessed by the χ² test. In case the latter was not applicable, Fisher's exact test was used. A P value of ≤0.05 was considered statistically significant. Correction for multiple comparisons was carried out using Bonferoni method, which indicates statistical significance when the P value is lower than the type I error (0.05) divided by the number of comparisons.

The age of patients at radiotherapy ranged from 19 to 69 years with a median of 51 years. There were 28 males and 4 females. Follow-up ranged from 24 to 144 months (median, 36 months). Control (G₀-G₁ fibrosis) and radiosensitive (G₂-G₃ fibrosis) patients were balanced for radiation boost and chemotherapy received (Fig. 1). Therefore, the average total doses received (with and without boost) in the control (67.00 Gy; SD, 1.79) and the radiosensitive (67.50 Gy; SD, 2.00) groups were comparable. Similarly, the number of patients without chemotherapy (6 patients) and those who received chemotherapy (10 patients) were the same in the control and the radiosensitive groups (Fig. 1).

As compared with the revised Cambridge reference sequence, a total of 622 mtDNA variations distributed over 214 locations in the coding (435 variations) and RNA (186 variations) regions, between 611 and 15,978 bp, were observed in the 32 patients. In the coding regions, there were 80 nonsynonymous single nucleotide variations, distributed over 34 locations, predicted to alter the amino acid sequences in the encoded proteins (Table 1A), and 356 synonymous single nucleotide variations distributed over 125 genetic locations (Table 1B). The RNA genes showed 186 single nucleotide variations distributed over 55 genetic locations (Table 1C). The majority of the variations were base substitution and few were single nucleotide deletion. A comparison between the control (G₀-G₁ fibrosis) and the radiosensitive (G₂-G₃ fibrosis) groups of radiotherapy patients is given in Table 2. The radiosensitive group showed a slight increase in the number of variants as compared with the control group. Statistical analysis showed significant difference in the mean numbers of nonsynonymous variations between the radiosensitive and the control groups (P = 0.003, t-test).

Table 1A and B show three mtDNA sequence variants (highlighted in bold characters) that are potentially different between the two groups of patients. A summary of the association study between these sequence variations and grade of s.c. and deep tissue fibrosis following radiotherapy is given in Table 3. A significant association (P = 0.01) was observed for the nonsynonymous, A to G substitution at nucleotide 10398 located in the NADH dehydrogenase subunit 3 gene (ND3) leading to a change in the amino acid sequence from Thr to Ala at codon 114. The G-variant was more frequent in the radiosensitive group than in the control group (odds ratio, 7.2; 95% confidence interval, 1.16-51.65). A borderline significant association (P = 0.06) was observed for the synonymous, T to C transition at nucleotide 9540 located in the cytochrome c oxidase subunit 3 gene (CO3, also known as COX3), where radiotherapy patients harboring the C-variant could be at higher risk of developing severe fibrosis (Table 3). The association observed for ND3 A10398G remains statistically significant after taking into consideration multiple comparisons using Bonferoni correction, which in this case equals 0.017.

The change in resazurin fluorescence, which reflects MRA, ranged from 11.3 to 22.1 with an average of 16.9 (SD, 2.9). Interestingly, fibroblasts of patients who developed moderate to severe fibrosis (G₂-G₃) showed a lower level of MRA (15.3; SD, 2.5) as compared with the control (G₀-G₁) group (18.5; SD, 2.3). The comparison between the radiosensitive (G₂-G₃) and the control (G₀-G₁) groups has been analyzed by box plot (Fig. 2A). Although there were variations, the radiation-sensitive group showed lower median value of MRA compared with the control. The Mann-Whitney rank test showed a statistically significant difference between the two groups (P = 0.001). Similar results were obtained with mtDNA 10389 genotype in which the variant G showed significantly lower level of MRA (P < 0.001; Fig. 2B).

Mitochondria and ionizing radiation share a common feature of generating ROS that can induce cellular damage. Furthermore, ionizing radiation can induce cell death that involves translocation of p53 to mitochondria. Both of these mechanisms can contribute to normal tissue radiosensitivity (17, 21). Therefore, in this study, we hypothesized that mitochondria contribute to radiosensitivity, and mtDNA variations can play a part in the genetic differences among patients that culminate in dissimilarity in the manifestations of side effects to cancer treatment. To our knowledge, this is the first clinical study examining mtDNA in the context of radiotherapy-fibrosis. The 32 patients included in this study had nasopharyngeal carcinoma. This cancer site is prevalent in Saudi Arabia and is ideal for this type of study because patients follow standardized treatment without surgery. Radical radiotherapy is the main treatment and it was delivered using 6 MV photon linear accelerator. The total radiation dose to the upper neck, where most radiation effects are scored, was 66 Gy given as 2 Gy per fraction. Locally advanced tumors are also treated with neoadjuvant and concurrent chemotherapy consisting of cisplatinum and epirubicin (18). Patients were carefully selected for this study during the follow-up of their disease. All patients completed at least two years of follow-up, which in our experience is largely sufficient for the appearance and the intensification of s.c. and deep tissues fibrosis following radiotherapy. Associated diseases were uncommon: three patients had diabetes (two in the control and one in the hypersensitive groups) and two patients were hypertensive. Nine patients were smokers with no clear preponderance between the control (four patients) and the radiosensitive (five patients) groups. The patients were well balanced between the radiosensitive (G₂-G₃) and the control (G₀-G₁) groups taking into account gender, age, and total radiation dose and chemotherapy received (Fig. 1). Therefore, overall no differences could be attributed to treatment-related factors.

The coding and RNA regions between 611 and 15,978 bp were directly sequenced, and genotypic variations were scored. We found 622 mtDNA single nucleotide variations in the 32 patients, 80 nonsynonymous (Table 1A), 356 synonymous (Table 1B), and 186 in the RNA regions (Table 1C). The analysis of the results indicate a higher number of genetic variations in the radiosensitive group of patients who developed moderate to severe s.c. and deep tissue fibrosis (G₂-G₃) compared with the control group who developed minimal fibrotic reactions (G₀-G₁) following radiotherapy (Table 2). The differences in the number of nonsynonymous variations between the two groups was statistically significant where the radiosensitive patients harbored a 2-fold higher number of base substitutions than the controls (P = 0.003). MRA was significantly lower in the radiosensitive group (P = 0.001). Therefore, it is possible that increased nucleotide variation in mtDNA may contribute to respiratory inefficiency through a cumulative effect of a series of polymorphisms of minor individual mutagenic potential (22). These results are encouraging and suggest that mtDNA variations can influence radiosensitivity and can contribute to late complications to radiotherapy in cancer patients. The nonsynonymous, A to G single nucleotide substitution at nucleotide 10398 leading to Thr to Ala amino acid change at codon 114 of the ND3 gene was significantly associated (P = 0.01) with higher grade of fibrosis (Table 3). The patients harboring the G-variant (Ala) were at 7.2-fold increased risk of developing severe late reaction to radiotherapy (grade of fibrosis) than patients harboring the dominant A (Thr) genotype. Interestingly, this risk factor, i.e., the G-variant (Ala) associated with increased radiosensitivity, was significantly correlated with lower mitochondrial respiratory activity (Fig. 2B). However, the probability of false discovery (false positive = 0.23) is considerable with this small number of patients studied. Furthermore, this genetic variation is rather common and constitutes about 46% of the individuals reported in the mitochondria data base.⁴

Therefore, it can be classified as a nonsynonymous polymorphism, and it has a frequency of about 41% in our patients reported here. This polymorphism along with A8701G in ATP6 gene (Table 3) were previously shown to affect mitochondrial matrix pH and intracellular calcium dynamics and suggested to play a role in the pathophysiology of the complex diseases of Parkinson's, Alzheimer's, bipolar disorder, and cancer (23). A borderline significant association (P = 0.06) was observed for the synonymous, T to C transition at nucleotide 9540 located in the CO3 gene, where radiotherapy patients harboring the C-variant may be at higher risk of developing severe fibrosis (Table 3). Interestingly, these two risk factors (9540C and 10398G variants) were present together in 31% of G₂-G₃ patients compared with 6% in G₀-G₁ (Fig. 3). Furthermore, in the G₂-G₃ group of patients, 43% of the severe G₃ fibrosis harbored the two risk factors as compared with 22% in the moderate G₂ fibrosis. This suggests an additive effect of the number of risk factors; the more risk factors the patient has, the higher the patient's radiosensitivity (9, 10).

The results presented here support the assumption that radiosensitivity is determined by multitrait genetic components, which is consistent with possible defects in mtDNA (24). Mutations and genetic polymorphic variations in a number of nuclear genes involved particularly in processing of DNA repair have been associated with radiosensitivity (5–12, 25). Mitochondria also have DNA repair albeit less efficient than nuclear DNA repair mechanisms (26). The mechanisms of mtDNA repair have attracted interest due to their association with certain aspects of “tissue” aging that ionizing radiation could accelerate. Mitochondrial disorders are increasingly recognized as the cause of a large number of progressive genetic diseases in humans (15, 27). Mutated mtDNA has been associated with Parkinson's and Alzheimer's diseases, diabetes, deafness, heart failure, optic nerve degeneration, cancer, and several progressive muscle diseases, as well as with the aging process itself (28–33). The concept of extranuclear gene mutations and/or polymorphisms being contributing factors to radiosensitivity is in agreement with the fact that oxidative and radiation-induced mtDNA damage takes place at the inner mitochondrial membrane near the sites of formation of highly reactive oxygen species. Mitochondrial DNA may be unable to counteract the damage inflicted because, in contrast to the nuclear genome, it lacks a fully operational DNA repair system. The postulated damage or loss of mtDNA will prevent the replication of the mitochondrial genome, which will decrease or prevent the “rejuvenation” of the mitochondria leading to energetic insufficiency, premature terminal differentiation, and cellular depletion, which are important steps in the development of side effects following irradiation.

In conclusion, our results indicate a possible involvement of mitochondria in radiation sensitivity. Variations among patients in mtDNA coding and RNA genes, such as the nonsynonymous 10398 A>G in ND3, may lead to suboptimum mitochondrial respiratory activity that could be contributing factors to their genetically determined variations in radiosensitivity and susceptibility to radiation-induced side effects following radiotherapy.

No potential conflicts of interest were disclosed.

We thank Dr. Mohamed Shoukri for statistical advice, Dr. Bilal Moftah for his continuous support, and Ms. Muneera Al-Buhairi and Ms. Rasha Ramadan for their technical assistance.