Introduction

Mitochondria play important roles in energy metabolism, generation of reactive oxygen species (ROS), and apoptosis. The most well-characterized mitochondrial function is the production of adenosine triphosphate (ATP) through oxidative phosphorylation. This process is accomplished by the respiratory chain and ATP synthases, which comprise a series of protein complexes that are encoded by both nuclear (nDNA) and mitochondrial DNA (mtDNA). The human mitochondrial genome is a supercoiled, double-stranded circular molecule of 16 569 base pairs in length. It codes for 13 of the 87 total proteins that constitute the respiratory chain as well as the 12S and 16S rRNAs and 22 tRNAs required for mitochondrial protein synthesis. One feature that distinguishes the mitochondrial and nuclear genomes is their intrinsic susceptibilities to damage. mtDNA is considerably more vulnerable to mutations than nDNA due to its lack of protection by complex chromatin organization, limited repair capacity, and close proximity to the electron transport chain, which constantly generates superoxide (O₂^-) radicals. Considering that mtDNA lacks sizeable introns, most mutations occur in the coding regions and, are thus, likely to be of biological consequence.^1,2

Mutations and deletions of mtDNA have been implicated in diseases such as Leber's hereditary optic neuropathy, maternally inherited diabetes mellitus, and Leigh's syndrome. Each cell contains many mitochondria with multiple copies of mtDNA; therefore it is possible for wild-type and mutant mtDNA to coexist in a state called heteroplasmy. The proportion of mutant mtDNA within a cell can vary over time and may ultimately drift toward predominantly mutant or wild type to achieve homoplasmy. As such, the biological impact of a given mutation is variable and dependent on the proportion of mutant mtDNAs carried by the individual. This effect contributes to the diverse phenotypes observed among family members carrying the same pathogenic mtDNA mutation.³

Although the aforementioned diseases of the mtDNA are attributed to germline mutations, somatic mtDNA-related defects have been detected in other diseases, most notably in cancer. These abnormalities include altered expression and activity of electron transport subunits, decreased oxidation of NADH-linked substrates, and mtDNA mutations/deletions.⁴ Their prevalence in cancer is consistent with the intrinsic susceptibility of mtDNA to damage. In spite of numerous reports of these phenomena, the mechanisms responsible for their initiation and their roles in cancer development, drug resistance, and disease progression still remain to be elucidated. Considering the extensive role of mitochondria in ATP metabolism, free radical generation, and regulation of apoptosis, mutations in mtDNA are likely to affect cellular energy capacities, increase oxidative stress, trigger ROS-mediated damage to DNA, and alter the cellular response to apoptosis induction by anticancer agents. We hypothesized that certain chemotherapeutic agents may cause mtDNA mutations, resulting in alterations in ROS generation and changes in sensitivity to therapeutic agents. To test this hypothesis, we analyzed mtDNA mutations in primary leukemia cells isolated from patients with chronic lymphocytic leukemia (CLL) with or without prior chemotherapy and evaluated the relationship between mtDNA mutation frequency, free radical generation, and clinical response to chemotherapy.

Materials and methods

Isolation of primary leukemia (CLL) cells

Peripheral blood samples were obtained from patients with CLL at various disease states after appropriate informed consent. The clinical characteristics of the CLL patients including age, RAI stage, and treatment status are listed in Tables 3 and 4. Primary CLL cells were isolated from the blood specimens using a Ficoll density centrifugation method as previously described.⁵ The isolated leukemia cells were washed twice with phosphate-buffered saline, and a portion of each sample was frozen and stored at -80°C for isolation of DNA and sequencing. The remaining cells were resuspended in RPMI 1640 medium supplemented with 10% fetal bovine serum (Life Technologies, Inc., Grand Island, NY, USA). Cells were incubated at 37°C in a water-jacketed incubator (Nuaire, Plymouth, MN, USA) with 5% CO₂ for 24 h prior to analysis of basal superoxide production.

Table 3 - Comparison of heteroplasmic and homoplasmic mtDNA mutations and the consequent alterations in amino acids of the encoded proteins in CLL cells from patients with or without prior chemotherapy.

Full table

Table 4 - Correlation of mtDNA mutation frequency with O₂^- generation and clinical treatment status in CLL patients.

Full table

Analysis of cellular O₂^- contents

Basal cellular superoxide (O₂^-) production was assessed in primary CLL cells using hydroethidine (Molecular Probes, Eugene, OR, USA) as an O₂^--sensitive fluorescent dye, and quantified using the FL-3 H channel on a FACSCalibur flow cytometer (Becton-Dickinson, San Jose, CA, USA) as previously described.⁶ To ensure the comparability of the data from CLL cells obtained at different times, the same flow cytometer settings were used, and calibrated each time using a reference cell line (Raji, human B-cell lymphoma) for normalization of slight variation due to experimental conditions.

DNA isolation

DNA was isolated from cryopreserved CLL cells using a modified phenol:chloroform:isoamyl alcohol method.⁵ Briefly, cells were thawed on ice and suspended in a lysis buffer comprising 10 mM Tris-HCl (pH 7.8), 100 mM NaCl, 25 mM EDTA (pH 8.0), 0.5% SDS, and 0.1 mg/ml Proteinase K and placed in a 45°C water bath for 16 h. Following digestion, cell lysates were mixed with an equal volume of phenol:chloroform:isoamyl alcohol (25:24:1; Promega; Madison, WI, USA) and rotated at ambient temperature for 10 min. Samples were centrifuged at 2000 g for 15 min. The aqueous phase was transferred to a fresh tube and mixed with 150 mM NaAc, 10 mM MgCl₂, and three volumes of ice-cold ethanol and precipitated overnight at -20°C. DNA was recovered from ethanol precipitates by centrifugation at 10 000 rpm for 30 min at 4°C. DNA pellets were washed twice with 100% ethanol, dissolved in TE buffer (10 mM Tris-HCl, 1 mM EDTA, pH 7.5), and then stored at -20°C until use.

Polymerase chain reaction and DNA sequencing

The six regions of the mitochondrial genome indicated in Figure 1a were amplified under standard polymerase chain reaction (PCR) conditions for direct DNA sequencing using mtDNA-specific primers as previously described.^7,8 Briefly, amplifications were performed in 50 l reactions comprising of approximately 50 ng DNA, 0.1 mM dNTPs (Roche, Indianapolis, IN, USA), 1 PCR Gold buffer, 1.5 mM MgCl₂, 0.5 M each primer, 2.5 U AmpliTaq Gold polymerase (Applied Biosystems, Foster City, CA, USA), and nuclease-free H₂O (Promega, Madison, WI, USA). Primer sequences were as follows: 5'CACCCTATTAACCACTCACG3' sense and 5'TGAGATTAGTAGTATGGGAG3' antisense for D loop (#15-484),⁷ 5'AACATACCCATGGCCAACCT3' sense and 5'GGCAGGAGTAATCAGAGGTG3' antisense for NDI (#3304-3836),⁸ 5'TATCACCTTTCATGATACGC3' sense and 5'GACGATGGGCATGAAACTG3' antisense for COXII (#7645-8215), 5'CTGTTCGCTTCATTCATTGCC3' sense and 5'GTGGCGCTTCCAATTAGGTG3' antisense for ATPase 6 (#8539-9059), 5'GACTCCCTAAAGCCCATGTCG3' sense and 5'TTGATCAGGAGAACGTGGTTAC3' antisense for ND4 (#11403-11927), and 5'AGTCCCACCCTCACACGATTC3' sense and 5'ACTGGTTGTCCTCCGATTCAGG3' for cytochrome b (#15260-15774).⁷ All primers were synthesized by Sigma-Genosys (The Woodlands,TX, USA). PCR reactions were amplified in a Robocycler (Stratagene, LaJolla, CA, USA) using the following cycling protocol, where X represents the annealing temperature: 94°C for 3 min, X°C for 3 min, 72°C for 1 min (first cycle) and 94°C for 1 min, X°C for 1 min, 72°C for 1 min for 35 cycles. The final cycle was modified to allow for a 7 min extension at 72°C. Annealing temperatures used were as follows: 46°C for D loop, 55°C for NDI, 56°C for COXII, 56°C for ATPase 6, 58°C for ND4, and 58°C for cytochrome b. In all, 10% of each PCR product was electrophoresed along with 500 ng of a 100 bp ladder (Invitrogen, Carlsbad, CA, USA) on a 1.2% agarose (Invitrogen, Carlsbad, CA, USA) gel at 100 V for 45 min to confim product purity and correct size. The remaining portion of the PCR product was column purified using a Qiaquick PCR Purification Kit (Qiagen, Valencia, CA, USA). Cycle sequencing reactions were performed using the ABI Prism Big Dye™ Terminator Sequencing kit version 3.0 (Applied Biosystems, Foster City, CA, USA). Sequencing reaction products were analyzed on an ABI 3100 automated DNA sequencer (Applied Biosystems, Foster City, CA, USA; sequencing services performed by SeqWright, Inc., Houston, TX, USA), using the same primers used for PCR amplification. Sequencing results were aligned with the revised Cambridge reference sequence (RCRS), GenBank accession #NC_001807.4, GI: 17981852, using the pairwise BLAST alignment method (National Center for Biotechnology Information, http://www.ncbi.nlm.nih.gov). Amino-acid changes were determined using the MitoAnalyzer Tool (National Institute of Standards and Technology, http://www.cstl.nist.gov/biotech/strbase/mitoanalyzer.html). Genetic polymorphisms were identified using the MITOMAP database (http://www.mitomap.org). The sequencing chromatograms were further evaluated manually to identify heteroplasmic and homoplasmic deviations from the RCRS. Any nucleotide position with two or more significant peaks of mixed nucleotide signals (heteroplasmy) was estimated for percentage of each nucleotide, based on the area under the curve of the corresponding nucleotide peak. Cases of heteroplasmy accounting for less than 30% of the total base signal were considered insignificant and were not scored. Only those with greater than 30% heteroplasmic signal were counted and included for calculation of mutation frequency.

Figure 1.

Increased O₂^- generation in CLL cells from patients with prior chemotherapy. (a) Regions of the mitochondrial genome selected for DNA sequencing. Numbers represent nucleotide positions in the human mitochondrial genome. D loop=displacement loop, ND=NADH dehydrogenase, CO=cytochrome c oxidase, Cyt b=cytochrome b. (b) Heteroplasmic and homoplasmic mtDNA mutations detected in primary leukemia cells isolated from CLL patients T9 (3372 T C/T) and T4 (8026 A->T). Representative DNA sequencing chromatograms are shown. (c) Flow cytometric analysis of cellular O₂^- was performed as described in the Materials and methods. Representative histograms for one untreated (UT4) and one treated (T4) patient are shown. The P-value indicates the significance of the difference in O₂^- generation between the untreated and treated patient and was determined by Kolmogorov–Smirnov analysis.

Full figure and legend (280K)

Statistical analyses

Wilcoxon's signed-rank tests were performed to determine the statistical significance of differences between two different groups of CLL patients (with or without prior chemotherapy) in their age, basal cellular superoxide production, mtDNA mutation frequencies, and the consequent alterations of amino acids. The difference in superoxide generation between the primary CLL cells obtained from the same patient at two different times (before and after chemotherapy) was determined by the Kolmogorov–Smirnov test (Cell Quest Pro software, Becton-Dickinson, San Jose, CA, USA). Kendall's was used to analyze the correlation between mutation frequency and O₂ levels. A P- value of less than 0.05 was considered statistically significant.

Results

Chemotherapeutic treatment is associated with a significant increase in the frequency of heteroplasmic mtDNA mutations

It is known that mtDNA is more susceptible to damage than (nDNA) and that certain chemotherapeutic agents can accumulate in mitochondria and damage mtDNA in vitro.^1,9,10 In order to explore the possibility that mtDNA mutations may be a consequence of chemotherapeutic treatment, we obtained peripheral blood specimens from a total of 20 CLL patients, and analyzed the frequency of mtDNA mutations. A total of 10 patients had a history of prior chemotherapy with fludarabine/alkylator-based regimen, while the other 10 CLL patients were previously untreated. DNA was isolated from the primary CLL cells of each patient, and the six regions of the mitochondrial genome indicated in Figure 1a were sequenced using a PCR-based method to estimate the mtDNA mutation frequency. These regions were selected to include segments of at least one gene from each mtDNA-encoded respiratory chain complex (complexes I, III–V) as well as a portion of the noncoding D-loop region, which has been previously identified as a mutational hotspot in human cancer.^11,12,13 Figure 1b shows examples of heteroplasmic and homoplasmic mtDNA mutations detected in two of the patient samples. Detailed sequencing results for each of the untreated and previously treated patients are presented in Tables 1 and 2, respectively. The bulk of the mtDNA sequence variations were detected in the two coding regions of complex I genes (ND1 and ND4). Additionally, a great deal of nucleotide alterations was detected in the D-loop region in both untreated and previously treated CLL patients (Tables 1 and 2). This is consistent with earlier studies that identified the D loop as a mutational hotspot in human cancer.^11,12,13 Owing to the high intrinsic polymorphism of this region and the lack of matched normal cells from the same patients to serve as sequence controls, we did not consider the high frequency of nucleotide alterations in the D loop to be the consequence of drug exposure. Thus, the nucleotide substitutions detected in the D loop that were matched to known polymorphisms were excluded from our calculation of mutation frequency to avoid possible overestimation of drug-induced mutation rates.

Table 1 - MtDNAmutations detected in primary leukemia cells from 10 CLL patients (UT1–UT10) without prior chemotherapy^a .

Full table

Table 2 - MtDNA mutations detected in primary leukemia cells from 10 CLL patients (T1-T10) who had prior chemotherapy^a.

Full table

The overall mtDNA mutation frequencies and their correlation with patients' clinical parameters are summarized in Table 3. Our data revealed that the previously treated CLL patients had a significantly higher frequency of heteroplasmic mtDNA mutations in the regions we sequenced as compared to the untreated patients (11.90 10^-4 vs 2.45 10^-4, W=84.5, P=0.00799, Table 3). Mutations in the DNA sequence may or may not result in amino-acid alterations of the encoded protein, depending on the nucleotide positions and nature of the substitution. Therefore, we used the MitoAnalyzer program to further determine if the mutations observed were associated with amino-acid changes. Table 3 shows that there was a significant increase in the number of heteroplasmic amino-acid alterations of the encoded proteins in CLL cells from patient with prior chemotherapy (W=82, P=0.0125). This is logical given that the mitochondrial genome lacks sizeable introns; thus any mutation has a high probability of changing the protein sequence. No significant differences were noted in the mean homoplasmic mtDNA mutation frequencies or mean homoplasmic amino-acid alterations between the two groups in the regions analyzed (W=45.5, P=0.7383 (mutations), W=46, P=0.6701 (amino-acid alterations), Table 3). It is important to note that the increase in heteroplasmic mutations was not correlated with advanced patient age (Table 3), further suggesting that the mutations observed were the result of exposure to drugs with DNA-damaging properties, rather than a consequence of the aging process.^14,15 Overall, mutations (W=83, P=0.01147) and amino-acid alterations (W=81, P=0.01668) were significantly higher in treated patients than in patients with no history of prior chemotherapy.

Heteroplasmic mtDNA mutations are correlated with increased superoxide radical generation

As mentioned previously, the human mitochondrial genome codes for 13 subunits of the electron transport chain. Considering that the respiratory chain is the primary intracellular source of free radical generation, we wondered whether the patients with an increased frequency of heteroplasmic mtDNA mutations would generate higher levels of oxygen radicals. In order to explore this possibility, we determined the basal O₂^- production in primary CLL cells isolated from all 20 patients. Representative histograms for an untreated (UT4) and previously treated patient (T4) are shown in Figure 1c. The results indicated that CLL cells from patients receiving prior chemotherapy produced significantly higher (W=88, P=0.00288) levels of cellular superoxide (40.015.4) than CLL cells isolated from previously untreated patients (16.312.6). The data shown in Table 4 suggest a possible correlation between overall mtDNA mutation frequency and increased superoxide generation (=0.2953, z.=1.8203, P=0.0687). The P-value of 0.0687 likely reflects the small sample number and the substantial individual variation among different patient cells. Apart from mtDNA mutation, several other factors such as cellular metabolic status, the expression of ROS scavenger enzymes, and level of antioxidants such as glutathione (GSH) may also significantly affect the overall superoxide levels in the cells.¹⁶

Relationship between mtDNA mutations and clinical response to anticancer agents

In order to evaluate whether mtDNA mutation frequency was related to clinical outcomes, we analyzed the relationship between mtDNA mutations and the clinical response to alkylating agents (cyclophosphamide or chlorambucil) and fludarabine, the most commonly used drugs in the clinical treatment of CLL. Patients who were clinically refractory to alkylators or fludarabine (T2, T4, T5, T8, and T9) tended to have higher rates of mtDNA mutations with amino-acid alterations. In contrast, patients (T1, T7, and T10) who showed low mtDNA mutations with no or minimal changes in amino acids were responsive to the drug treatment in vivo. Although the number of patients is small, statistical analysis suggests a correlation between mtDNA mutations and patient responses to chemotherapy (W=22, P=0.02697). It is not clear at the present time if the resistance to alkylators or fludarabine was due to changes in mitochondrial apoptotic responses or attributed to ROS-mediated alterations in other cellular events.

Development of a heteroplasmic mtDNA mutation and increased O₂^- generation in CLL cells after chemotherapy

In order to further explore the possibility that chemotherapeutic treatment is associated with the development of mtDNA mutations in vivo, we obtained paired blood samples from the same patient (T10) before and after (6 months apart) chemotherapy with fludarabine and cyclophosphamide. Sequencing analysis revealed the presence of a new heteroplasmic mutation (G C/G) affecting the cytochrome c oxidase II (COII) gene at nucleotide position 7762 after chemotherapy (Figure 2a). This mutation was not present in CLL cells obtained from the same patient before treatment. For comparison, we also sequenced mtDNA from several pairs of sequential samples from CLL patients without chemotherapy, and did not detect the development of any new mutations during the time interval. Thus, the heteroplasmic mutation seen in the patient under chemotherapy was most likely induced by the therapeutic agents (fludarabine/cyclophosphamide), which are known to have DNA-damaging properties and cause mutations.^{16,17,18,19,20,21} Importantly, substitution of the G nucleotide at position 7762 by a C nucleotide resulted in an amino-acid change (Gln His) at codon #59 of the COII enzyme, a key component of mitochondrial electron transport complex IV. Interestingly, comparison of the cellular levels of the superoxide radical (O₂^-, a product of mitochondrial respiration) in the two CLL samples obtained before and after chemotherapy revealed a substantial increase of O₂^- production in the post-therapy CLL cells harboring the mtDNA heteroplasmic mutation at nucleotide #7762 (Figure 2b). This increase in free radical generation was highly significant (P<0.001). Taken together, these data suggested that chemotherapy may cause mtDNA mutations, leading to functional changes in the mitochondrial electron transport chain and increased free radical generation.

Figure 2.

Identification of a novel heteroplasmic mtDNA mutation and corresponding increase in O₂^- generation in CLL cells from a patient (T10 in Table 1) after chemotherapy. (a) Automated sequencing results illustrating a novel, heteroplasmic mutation at nucleotide position 7762 of the COII gene. Upper panel shows a segment of mtDNA sequence of CLL cells obtained before chemotherapy; lower panel shows the same segment of mtDNA sequence of CLL cells from the same patient 6 months after chemotherapy with a fludarabine/cyclophosphamide regimen. (b) Quantitation of cellular O₂^- generation in primary CLL cells obtained from patient T10. O₂^- levels were determined by flow cytometric analysis as described in the Materials and methods. The left and right curves correspond to CLL samples obtained before and after chemotherapy, respectively. Kolmogorov–Smirnov analysis was used to determine the statistical significance of O₂^- difference between the two samples, and the P-value is indicated on the histogram.

Full figure and legend (257K)

Discussion

Mitochondrial defects have been implicated in the development and progression of cancer for several decades. The groundbreaking work by Warburg²² suggested that a key event in cancer development involved an 'injury' to the respiratory machinery, resulting in compensatory increases in glycolytic ATP production. It is possible that some of the effects observed by Warburg were associated with mtDNA mutations. Although a direct causal effect of mtDNA mutations in oncogenesis has not yet been established, it is conceivable that these mutations could contribute to the development of a cancer phenotype by changing cellular energy capacities, increasing free radical generation leading to further DNA damage and genetic instability, and/or modulating apoptosis.

The present study focused on how the aforementioned potential consequences of mtDNA mutations could impact cancer (CLL) patients. We hypothesized that certain chemotherapeutic agents with DNA-damaging properties may cause mtDNA mutations, resulting in increased mitochondrial ROS generation. This hypothesis was based on several observations. First, the work by Talarico et al⁹ and Olivero et al¹⁰ showed that certain chemotherapeutic agents such as cisplatinum can accumulate in mitochondria and damage mtDNA in vitro. Second, a previous study in our laboratory indicated that CLL patients who received prior chemotherapy generated significantly higher levels of cellular O₂^- than those who were untreated.²³ Third, the majority of cellular O₂^- radicals are generated by leakage of electrons from the respiratory chain, especially from complexes I and III.^1,2,3,4 Mutations in mtDNA are likely to cause changes in its encoded respiratory chain complex components, leading to increased electron leakage, and enhanced superoxide radical formation. The results of our study suggest that there is a relationship between mtDNA mutations, increased ROS generation, and chemotherapeutic treatment. However, the complexities of the in vivo situation preclude us from definitively determining if the mutations are a direct consequence of chemotherapy-induced DNA damage or if the treatment-associated increases in free radical generation cause oxidative damage to mtDNA, resulting in further increases in ROS generation due to a compromised respiratory chain. It is likely that both scenarios are occurring in vivo.

As mtDNA may undergo spontaneous mutations without exposure to chemotherapeutic agents, it is possible that the CLL cells from some of the previously treated patients might have had a higher background of mtDNA mutations before drug treatment was initiated or that the mutations were somehow associated with disease progression. A previous report regarding mtDNA mutations in tumors of the head and neck demonstrated that mtDNA alterations indeed occurred in the early premalignant lesions and rose in incidence with increasing histological severity.²⁴ However, considering that untreated patients tended to have very low rates of both hetero- and homoplasmic mutations in the present study, it is more likely that the high frequencies of heteroplasmic mutations observed in treated CLL patients are chemotherapy-related. This is supported by the data from 20 CLL patient samples, which revealed a correlation between chemotherapeutic treatment, mtDNA mutations, and O₂^- levels. These observations are further substantiated by the development of a new, amino-acid-changing heteroplasmic mutation in the COII gene after chemotherapy, which was associated with a concomitant increase in O₂^- generation (Figure 2).

Since each cell contains many mitochondria with multiple copies of mtDNA, mutations in mtDNA can be hetero- or homoplasmic. It is important to note that the majority of the chemotherapy-associated mtDNA mutations detected in this study were heteroplasmic. While both homo- and heteroplasmic mutations have been frequently observed in cancer cells, the exact mechanism by which homoplasmy arises from heteroplasmic mutations is uncertain. Possible mechanisms include the selection or clonal expansion of the mutants with growth and/or survival advantage or a random segregation of the mutant.^25,26 Data obtained by computer modeling suggest that if a mtDNA mutation occurs in a tumor progenitor cell, homoplasmy could be achieved entirely by chance without selection for physiological advantage through unbiased mtDNA replication and sorting during cell division.²⁷ It is likely that drug-induced mtDNA mutations in cancer cells may initially appear in a heteroplasmic state. If the mutations occur in the coding sequence and produce functional alterations in the encoded protein, some of the mutants may become homoplasmic as a result of in vivo selection processes.²⁸ Additional studies are required to gain a better understanding of how homoplasmy can be achieved, particularly in the in vivo situation.

Our study suggests that increased ROS generation seems to be an important biological consequence of mtDNA mutations. Considering that the mitochondrial respiratory chain is a major source of intracellular superoxide radicals, it is not surprising that superoxide generation was increased in CLL cells from patients who underwent prior chemotherapy and had increased mtDNA mutations. As illustrated in Figure 3, increased ROS production in cancer cells may lead to several significant consequences. First, because ROS are chemically reactive and able to damage DNA, proteins, and lipids, increased ROS production could culminate in additional mutations in both nDNA and mtDNA and further increases in cellular oxidative stress, contributing to genetic instability and disease progression.^1,2,3 Second, because mitochondria play an essential role in apoptosis, the increase in ROS may affect the sensitivity of cancer cells to chemotherapeutic agents.²⁹ Previous studies have indicated that ROS can stimulate cell proliferation and activate prosurvival transcription factors such as NF-B.^30,31 Taking into consideration that many antiapoptotic proteins such as XIAP are under the transcriptional control of NF-B, it is plausible that this effect could contribute to drug resistance and disease progression.^32,33 Such drug-resistance mechanisms might be a factor in the chemotherapy failure observed in CLL patients.^34,35,36 The fact that the five refractory patients all had high frequencies of mtDNA mutations with amino-acid alterations in the encoded proteins (Table 4) seems to suggest a relationship between mtDNA mutations and drug resistance, although the small number of patients did not allow a formal statistical analysis. Furthermore, clinical drug resistance is likely caused by multiple factors. Mutation of mtDNA is only one of the contributing factors and may not be used alone to predict relapse. Third, the increase in ROS generation in cells with mtDNA mutations may also serve as a biochemical basis to design novel therapeutic strategies to preferentially kill cancer cells with ROS stress. A previous study in our laboratory established that CLL cells from patients with prior chemotherapy had increased ROS levels and were more sensitive to 2-methoxyestradiol (2-ME), a novel anticancer agent that causes additional superoxide accumulation by inhibiting its elimination.^6,37,38 This correlation was significant when the IC₅₀ values were plotted against their corresponding cellular superoxide levels (P<0.011).²³ We reason that cells with increased levels of superoxide may depend heavily upon the antioxidant enzyme superoxide dismutase (SOD) for elimination of this free radical. Inhibition of SOD by 2-ME would cause a severe accumulation of superoxide in these cells, leading to lethal cellular damage. These results along with the data obtained in the present study suggest that the malfunctioning respiratory chain in cancer cells could possibly be exploited therapeutically by using novel agents that work via a free radical-mediated mechanism.

Figure 3.

Schematic illustration of the possible functional consequences of mtDNA mutations induced by chemotherapy (Rx) with DNA-damaging agents. Drug-induced mtDNA mutations may initially appear as heteroplasmic mutations, and some of the mutants may eventually evolve as homoplasmic mutations. Certain mtDNA mutations will cause a functional change in the mitochondrial respiratory chain, leading to increased generation of ROS, which in turn cause further damage to mtDNA and nDNA and contribute to genetic instability. It is possible to preferentially induce apoptosis in malignant cells that have increased ROS levels, using novel agents that cause further ROS stress in the cells such as 2-ME.^6,23

Full figure and legend (99K)

In summary, our data suggest that chemotherapy with certain DNA-damaging agents may cause mtDNA mutations, which initially appear heteroplasmic, and are associated with an increase in ROS generation. It is possible that respiratory inefficiency as a consequence of mtDNA alterations may contribute to the elevated glycolytic activity (Warburg effect) and constitutive oxidative stress frequently observed in malignant cells. The increase in ROS generation, on the one hand, may contribute to genetic instability and disease progression and, on the other hand, may also provide a biochemical basis to preferentially kill cancer cells with novel anticancer agents that impose additional free radical stress to the malignant cells. Further studies are warranted to explore such a new therapeutic strategy.

References

1. Wallace DC. Mitochondrial diseases in man and mouse. Science 1999; 283: 1482–1488. | Article | PubMed | ISI | ChemPort |
2. Rasmussen LJ, Singh KK. Genetic integrity of the mitochondrial genome. In: Singh KK (ed). Mitochondrial DNA Mutations in Aging, Disease, and Cancer. New York, NY: Springer, 1998, pp 115–127.
3. Wallace DC. Diseases of the mitochondrial DNA. Annu Rev Biochem 1992; 61: 1175–1212. | Article | PubMed | ISI | ChemPort |
4. Carew JS, Mitochondrial defects in cancer. Mol Cancer 2002; 1: 9. | Article | PubMed |
5. Huang P, Robertson LE, Wright S, Plunkett W. High molecular weight DNA fragmentation: a critical event in nucleoside analogue-induced apoptosis in leukemia cells. Clin Cancer Res 1995; 9: 1005–1013.
6. Huang P, Feng L, Oldham, EA, Keating MJ, Plunkett W. Superoxide dismutase as a target for the selective killing of cancer cells. Nature 2000; 407: 390–395. | Article | PubMed | ISI | ChemPort |
7. Levin BC, Cheng H, Reeder DJ. A human mitochondrial DNA standard reference material for quality control in forensic identification, medical diagnosis and mutation detection. Genomics 1999; 55: 135–146. | Article | PubMed |
8. Wei YH, Kao SH, Lee HC. Simultaneous increase of mitochondrial DNA deletions and lipid peroxidation in human aging. Ann NY Acad Sci 1996; 786: 24–43.
9. Talarico T, Cullinan CM, Gray PJ, Webster LK, Deacon GB, Phillips DR. Nuclear and mitochondrial distribution of organoamidoplatinum(II) lesions in cisplatin-sensitive and resistant adenocarcinoma cells. Anticancer Drug Des 2001; 16: 135–141. | PubMed | ChemPort |
10. Olivero OA, Chang PK, Lopez-Larraza DM, Semino-Mora MC, Poirier MC. Preferential formation and decreased removal of cisplatin–DNA adducts in Chinese hamster ovary cell mitochondrial DNA as compared to nuclear DNA. Mutat Res 1997; 391: 79–86. | PubMed | ISI | ChemPort |
11. Sanchez-Cespedes M, Parrella P, Nomoto S, Cohen D, Xiao Y, Esteller M et al. Identification of a mononucleotide repeat as a major target for mitochondrial DNA alterations in human tumors. Cancer Res 2001; 61: 7015–7019. | PubMed | ISI | ChemPort |
12. Nomoto S, Yamashita K, Koshikawa K, Nakao A, Sidransky D. Mitochondrial D-loop mutations as clonal markers in multi-centric hepatocellular carcinoma and plasma. Clin Cancer Res 2002; 8: 481–487. | PubMed | ISI | ChemPort |
13. Miyazono F, Schneider PM, Metzger R, Warnecke-Eberz U, Baldus SE, Dienes HP et al. Mutations in the mitochondrial DNA D-Loop region occur frequently in adenocarcinoma of Barrett's esophagus. Oncogene 2002; 21: 3780–3783. | Article | PubMed | ISI | ChemPort |
14. Reynier P, Malthiery Y. Accumulation of deletions in MtDNA during tissue aging: analysis by long PCR. Biochem Biophys Res Commun 1995; 217: 59–67.
15. Kajander OA, Karhunen PJ, Jacobs HT. The relationship between somatic mtDNA rearrangements, human heart disease and aging. Hum Mol Genet 2002; 11: 317–324. | Article | PubMed | ChemPort |
16. Richardson ME, Siemann DW. DNA damage in cyclophosphamide-resistant tumor cells: the role of glutathione. Cancer Res 1995; 55: 1691–1695.
17. Berger MR, Bedford P, Zeller WJ, Kaufmann M. A comparative study of therapeutic activity, myelotoxicity and DNA damage in the bone marrow of mice after cyclophosphamide and ASTA Z 7557 (INN mafosfamide). Invest New Drugs 1984; 2: 181–186.
18. Ford CD, Warnick CT. DNA damage and repair in patients receiving high-dose cyclophosphamide and radiation. NCI Monogr 1988: 41–44.
19. Chatterjee R, Haines GA, Perera DM, Goldstone A, Morris ID. Testicular and sperm DNA damage after treatment with fludarabine for chronic lymphocytic leukaemia. Hum Reprod 2000; 15: 762–766. | Article | PubMed | ISI | ChemPort |
20. Huang P, Plunkett W. Fludarabine- and gemcitabine-induced apoptosis: incorporation of analogs into DNA is a critical event. Cancer Chemother Pharmacol 1995; 36: 181–188. | PubMed |
21. Huang P, Siciliano MJ, Plunkett W. Gene deletion, a mechanism of induced mutation by arabinosyl nucleosides. Mutat Res 1989; 210: 291–301. | PubMed | ChemPort |
22. Warburg O. On the origin of cancer cells. Science 1956; 123: 309–314. | Article | PubMed | ISI | ChemPort |
23. Zhou Y, Hileman EO, Plunkett W, Keating MJ, Huang P. Free radical stress in chronic lymphocytic leukemia cells and its role in cellular sensitivity to ROS-generating anticancer agents. Blood 2003.
24. Ha PK, Tong BC, Westra WH, Sanchez-Cespedes M, Parrella P, Zahurak M et al. Mitochondrial C-tract alteration in premalignant lesions of the head and neck: a marker for progression and clonal proliferation. Clin Cancer Res 2002; 8: 2260–2265. | PubMed | ISI | ChemPort |
25. Polyak K, Li Y, Zhu H, Lengauer C, Willson JK, Markowitz SD et al. Somatic mutations of the mitochondrial genome in human colorectal tumors. Nat Genet 1998; 20: 291–293. | Article | PubMed | ISI | ChemPort |
26. Fliss MS, Usadel H, Caballero OL, Wu L, Buta MR, Eleff SM et al. Facile detection of mitochondrial DNA mutations in tumors and bodily fluids. Science 2000; 287: 2017–2019. | Article | PubMed | ISI | ChemPort |
27. Coller HA, Khrapko K, Bodyak ND, Nekhaeva E, Herrero-Jimenez P, Thrilly WG. High frequency of homoplasmic mitochondrial DNA mutations in human tumors can be explained without selection. Nat Genet 2001; 28: 147–150. | Article | PubMed | ISI | ChemPort |
28. Coskun PE, Ruiz-Pesini E, Wallace DC. Control region mtDNA variants: Longevity, climatic adaptation, and a forensic conundrum. Proc Natl Acad Sci USA 2003; 100: 2174–2176. | Article | PubMed | ChemPort |
29. Wang X. The expanding role of mitochondria in apoptosis. Genes Dev 2001; 15: 2922–2933. | PubMed | ISI | ChemPort |
30. Liu SL, Lin X, Shi DY, Cheng J, Wu CQ, Zhang YD. Reactive oxygen species stimulate human hepatoma cell proliferation via cross-talk between PI3K/PKB and JNK signaling pathways. Arch Biochem Biophys 2002; 406: 173–182. | Article | PubMed | ISI | ChemPort |
31. Haddad JJ. Redox regulation of pro-inflammatory cytokines and IkappaB-alpha/NF-kappaB nuclear translocation and activation. Biochem Biophys Res Commun 2002; 296: 847–856.
32. Tang G, Minemoto Y, Dibling B, Purcell NH, Li Z, Karin M, Lin A. Inhibition of JNK activation through NF-kappaB target genes. Nature 2001; 414: 265–266. | Article | PubMed | ISI | ChemPort |
33. Stehlik C, de Martin R, Kumabashiri I, Schmid JA, Binder BR, Lipp J. Nuclear factor (NF)-kappaB-regulated X-chromosome-linked iap gene expression protects endothelial cells from tumor necrosis alpha-induced apoptosis. J Exp Med 1998; 188: 211–216. | Article | PubMed | ISI | ChemPort |
34. Kitada S, Zapata JM, Andreeff M, Reed JC. Bryostatin and CD40-ligand enhance apoptosis resistance and induce expression of cell survival genes in B-cell chronic lymphocytic leukaemia. Br J Haematol 1999; 106: 995–1004. | Article | PubMed | ChemPort |
35. Masdehors P, Omura S, Merle-Beral H, Mentz F, Cosset JM, Dumont J et al. Increased sensitivity of CLL-derived lymphocytes to apoptotic death activation by the proteasome-specific inhibitor lactacystin. Br J Haematol 1999; 105: 752–757. | Article | PubMed | ISI | ChemPort |
36. Delic J, Masdehors P, Omura S, Cosset JM, Dumont J, Binet JL et al. The proteasome inhibitor lactacystin induces apoptosis and sensitizes chemo- and radioresistant human chronic lymphocytic leukaemia lymphocytes to TNF-alpha-initiated apoptosis. Br J Cancer 1998; 77: 1103–1107. | PubMed | ISI | ChemPort |
37. Wood L, Lesse MR, Leblond B, Woo LW, Ganeshapillai D, Purohit A et al. Inhibition of superoxide dismutase by 2-methoxyoestradiol analagues and oestrogen derivatives: structure–activity relationships. Anticancer Drug Des 2001; 16: 209–215.
38. Golab J, Nowis D, Skrzycki M, Czeczot H, Baranczyk-Kuzma A, Wilczynski GM et al. Anti-tumor effects of photodynamic therapy are potentiated by 2-methoxyestradiol, a superoxide dismutase inhibitor. J Biol Chem 2002; 278: 407–414.

Acknowledgements

This study was supported in part by research Grants CA77339, CA85563, CA81534, and T32 EY07119.