European mtDNA Variants Are Associated With Differential Responses to Cisplatin, an Anticancer Drug: Implications for Drug Resistance and Side Effects

Background: Cisplatin, a powerful antitumor agent, causes formation of DNA adducts, and activation of apoptotic pathways. Presently, cisplatin resistance develops in up to 70% of patients but the underlying molecular mechanism(s) are unclear and there are no markers to determine which patients will become resistant. Mitochondria play a significant role not only in energy metabolism but also retrograde signaling (mitochondria to nucleus) that modulates inflammation, complement, and apoptosis pathways. Maternally inherited mitochondrial (mt) DNA can be classified into haplogroups representing different ethnic populations that have diverse susceptibilities to diseases and medications. Methods: Transmitochondrial cybrids, where all cell lines possess identical nuclear genomes but either the H (Southern European) or J (Northern European) mtDNA haplogroups, were treated with cisplatin and analyzed for differential responses related to viability, oxidative stress, and expression levels of genes associated with cancer, cisplatin-induced nephrotoxicity and resistance, apoptosis and signaling pathways. Results: The cisplatin-treated-J cybrids showed greater loss of cell viability along with lower levels of reactive oxygen species and mitochondrial membrane potential compared to cisplatin-treated-H cybrids. After cisplatin treatment, J cybrids showed increased gene expression of BAX, CASP3, and CYP51A, but lower levels of SFRP1 compared to untreated-J cybrids. The cisplatin-treated-H cybrids had elevated expression of CDKN1A/P21, which has a role in cisplatin toxicity, compared to untreated-H cybrids. The cisplatin-treated H had higher transcription levels of ABCC1, DHRS2/HEP27, and EFEMP1 compared to cisplatin-treated-J cybrids. Conclusions: Cybrid cell lines that contain identical nuclei but either H mtDNA mitochondria or J mtDNA mitochondria respond differently to cisplatin treatments suggesting involvement of the retrograde signaling (from mitochondria to nucleus) in the drug-induced cell death. Varying toxicities and transcription levels of the H vs. J cybrids after cisplatin treatment support the hypothesis that mtDNA variants play a role in the expression of genes affecting resistance and side effects of cisplatin.


INTRODUCTION
Cisplatin is a non-specific alkylating agent used for decades to successfully treat various cancers. However, cisplatin has dose-dependent toxicity and resistance often develops (1,2). Mitochondrial dynamics play an important role in resistance to chemotherapy and severity of side effects. Many anti-cancer medications can cause mitochondrial dysfunction and DNA damage, and cells depleted of mtDNA show increased resistance to chemotherapeutic agents. Cisplatin is a pro-apoptotic drug that damages mitochondria, contributing to the toxicities in gastrointestinal, auditory, and kidney proximal tubule cells (3)(4)(5). Cisplatin treatment of head and neck squamous cell carcinoma and Chinese hamster ovarian cells significantly increases adduct formation in mitochondrial (mt) DNA compared to nuclear (n) DNA (6). Furthermore, mtDNA is more likely to remain damaged after cisplatin exposure because mitochondria lack the nucleotide-excision repair mechanisms found in nDNA. The severity of cancer drug side effects and incidence of induced resistance to chemotherapy drugs vary amongst individuals but the mechanism(s) are still not fully understood.
Cisplatin side effects include nausea, vomiting, myelosuppression, nephrotoxicity, neurotoxicity, cognitive dysfunction retinopathy, and hearing loss. Intravenous delivery of cisplatin commonly causes mild to moderate pigmentary retinopathy, along with intra-retinal hemorrhages, exudates, and cotton wool spots (7). In addition, significant vision loss can occur with both systemic and local delivery methods of the drug (7). However, it is unclear which individuals are going to be susceptible to this toxicity. The decision to continue a drug regimen depends not only on the effectiveness in treating the malignancy, but also on a person's tolerance to the drug and the risks of end-organ damage. Thus, insight into the mechanisms of cisplatin toxicity is valuable to patient care and therefore, a model using human retinal pigment epithelial (RPE) cells was developed to study mechanisms of cytotoxic damage.
Mitochondria are unique organelles that play essential roles in ATP production, calcium homeostasis, apoptosis, and cell signaling. The mtDNA are maternally inherited and can be classified into different haplogroups based on patterns of single nucleotide polymorphisms (SNPs) that have accumulated over thousands of years. The mtDNA haplogroups represent populations from different geographic origins and subsequently, can be used to characterize different ethnic groups. Clinically, different ethnic populations show dissimilar susceptibilities to diseases and drug responses (8)(9)(10), and it has been suggested that mtDNA haplogroups may play important roles in these differences (11). Specific mtDNA haplogroups have also been associated with renal, prostate, breast, and lung cancers (12)(13)(14)(15).
In addition, somatic and germline mtDNA mutations, as well as levels of mtDNA copy numbers, have been associated with increased risk of cancer and different responses to anti-cancer drugs (16). These studies demonstrate how polymorphisms and/or variants in mtDNA can lead to significant changes at the molecular and cellular levels and can be associated with increased cancer risk.
One method to characterize the functional consequences of cells having specific mtDNA haplogroups is through transmitochondrial cybrids (cells with identical nuclei but different mtDNA). Previously, we demonstrated that H cybrids (Southern European maternal origin mtDNA) vs. J cybrids (Northern European maternal origin mtDNA) have significantly different cellular homeostasis. Although all cybrids had identical nuclei and culture conditions, cells containing the J mtDNA had increased rates of growth along with higher lactate and glycolysis levels, but showed significantly lower MT-RNA expression and ATP levels compared to the H cybrids (17,18). Interestingly, even though the J cybrids were created in a non-cancerous human retinal cell line , the behavior of these cells was characteristic of the Warburg Effect, which described cancerous cells to be more glycolytic, using less oxidative phosphorylation, and producing high amounts of lactate in the presence of oxygen (aerobic glycolysis).
Using a HeLa cybrid model, Amo et al. was able to demonstrate that resistance to cisplatin was conferred via alterations of the mtDNA within the control region and that cisplatin-resistant clones possessed shorter OriB variants within a 16184-16193 region enriched with cytosine repeats (19). Interestingly, alterations of the nuclear genome were not involved in the cisplatin resistance. These findings support the hypothesis that the SNP differences that define the different haplogroups would be important for the responses to cisplatin and as the H and J haplogroup populations possess disparate SNP patterns, then their cybrids would show different responses.
In the present study, we compared the effects of cisplatin on human RPE cell cybrids that possessed either H or J haplogroup mtDNA and found differential responses in levels of cell viability, reactive oxygen species production, mitochondrial membrane potential, and gene expression levels in pathways related to cell signaling, apoptosis, and cisplatin resistance.

Cybrid Creation
All subjects read and signed an informed consent (IRB #2003-3131) from the Institutional Review Board of the University of California, Irvine. All clinical investigations and protocols  were conducted according to the principles of the Declaration of Helsinki and approved by the appropriate investigational review boards (University of California, Irvine). Cybrids were generated as described previously (17,18). H and J cybrids were created by polyethylene glycol fusion of platelets with the Rho0 (mtDNA free) ARPE-19 cells, which had been treated by low dosage ethidium bromide as described by Miceli and Jazwinski (20). H and J cybrids were cultured to the fifth passage using DMEM-F12 containing 10% dialyzed fetal bovine serum, 100 unit/ml penicillin and 100 µg/ml streptomycin, 2.5 µg/ml fungizone, 50 µg/ml gentamycin, and 17.5 mM glucose. Figure 1 provides description of the background of the subjects used in this study. The ages for the H mtDNA subjects (n = 7 cybrids) were 30.57 ± 3.39 years old, while ages for the J mtDNA subjects (n = 7 cybrids) were 36.14 ± 5.47 years (P = 0.4) ( Figure 1A). There were 4 males and 3 females for the J haplogroup subjects and 5 males and 2 females in the H haplogroup subjects. Figures 1B,C show the haplogroup defining SNPs in the mtDNA for the J cybrids and H cybrids used in this study (see sequencing method below).

Sequencing of mtDNA From H and J Cybrids
DNA was extracted from the individual cybrids (n = 7 for H cybrids and n = 7 for J cybrids) using a kit (DNeasy Blood and Tissue Kit, Qiagen, Germantown, MD). Next Generation Sequencing (NGS) technology that sequences both strands of mtDNA independently in both directions was used to quantitate the haplogroup-defining SNPs, private SNPs (not defining haplogroups), and low frequency heteroplasmy SNPs across the entire mitochondrial genome.

NGS Sequencing
Primers were designed across the entire human mitochondrial genome. A total of 171 primer pairs were designed that overlapped each other to capture the intervening sequences. A total of 100 ng of DNA per sample was used to construct NGS libraries using the TruSeq Custom Amplicon kit (Illumina, San Diego, CA). Two independent sets of primer pools were synthesized. The two primer pools had primers that were complementary. The two primer pools would interrogate the two strands of the mitochondrial genome independently.  Our previous studies have shown that the J cybrids grow more rapidly than H cybrids (17) so the ROS levels were normalized per cell number. Simultaneously to the ROS experiments, H and J cybrids were plated in 6 well-plates (500,000 cells/plate), incubated 24 or 48 h, and treated in the identical fashion as described above. Cell viabilities were assessed by the Beckman Coulter ViCell Counter, allowing us to determine ROS levels per cell numbers. Differences in cisplatin-treated cells compared to untreated cells were analyzed (Prism, GraphPad Software Inc.) and were considered to be statistically significant when P ≤ 0.05. Experiments were analyzed in quadruplicate replicates and the entire experiment repeated three separate times.
RNA Isolation, cDNA Synthesis, and Quantitative Reverse Transcription PCR (qRT-PCR) H and J cells were plated (500,000 cells/well) and incubated for 24 h in six-well plates. Cells were treated with culture media containing either 0 or 40 µM of cisplatin for another 48 h. Trypsinized cells were pelleted and RNA isolated following the manufacturer's protocol (RNeasy Kit, Qiagen, Valencia, CA). After RNA quantification (Nanodrop 1000, Thermoscientific, Wilmington, DE), the cDNA was transcribed from 100 ng of RNA (Qiagen), and then used for quantitative reverse transcription-PCR (qRT-PCR) (StepOnePlus instrument; Applied Biosystems, Carlsbad, CA). SYBR Green-based primers were used (Qiagen). Table 2A shows the GenBank Accession numbers and functions for 23 genes that were investigated. Cancer-related genes were Type   Regulates genes that induce cell cycle arrest, apoptosis, senescence, DNA repair, or changes in metabolism. Induces apoptosis.

Statistical Analyses
Statistical analysis of the data was performed by ANOVA (GraphPad Prism, version 5.0). Newman-Keuls multiplecomparison or the two-tailed t-tests were used to compare the data within each experiment. P ≤ 0.05 was considered statistically significant. Error bars in the graphs represent standard error of the mean (SEM).

Sequencing of mtDNA From H and J Cybrids
The entire mtDNA from the J and H cybrids were sequenced using NGS technology. Figure 1A shows the age and gender of person in this study. The private SNPs are those that do not define the J or H haplogroups (non-haplogroup defining). The unique SNPs are not listed in www.MitoMap.org or other programs. Table 1A shows the SNPs in the J haplogroup cybrids. There were 11 private SNPS in the mtDNA regions of the J cybrids:  Table 1C. The m.3010G>A variant (found in all seven of the J haplogroup cybrids and three of the H haplogroup cybrids) is associated with cyclic vomiting syndrome and migraines. The other SNP variants listed are found in either the J cybrids or the H cybrids but not both.

IC50 Analyses Results
IC-50 analyses were performed to determine the concentration of cisplatin required to inhibit the cell viability by 50% (Figure 2). The Goodness of Fit values were R 2 = 0.8388 and R 2 = 0.8828 for the H and J cybrids, respectively. The IC-50 values for cybrid-H were 47.13 µM (95% confidence interval 38.62-57.53 µM) and

Reactive Oxygen Species (ROS) Production After Cisplatin Treatment
The ROS levels, measured in RFU, were compared between the H and J cybrids with and without cisplatin treatments after 48 h incubation ( Figure 3C). The cisplatin-treated-J cybrids showed significantly lower ROS compared to the untreated-J cybrids (56.79% ± 7.731 vs. 78.33% ± 4.24, P = 0.03, respectively) and also compared to cisplatin-treated-H cybrids (98.26% ± 8.66, P = 0.03). There was no difference between the cisplatin-treated-H cybrids and untreated-H cybrids after 48 h incubation (P = 0.37). Since ROS production levels had been normalized to cell viability for each condition, our findings showed that after cisplatin treatment, the J cybrid cultures showed significantly less ROS production compared to H treated cybrid cultures.
The levels of ALK, BRCA1, BMI1, CDK2, EGFR, ERBB2, ERCC1, KAT5/TIP60, and TP53 were not different from each other in the H and J cybrids under any conditions (with or without cisplatin treatment; data not shown).

Apoptosis Pathway Genes
Untreated-H and untreated-J cybrids expressed similar levels of BAX (1.01-fold, P = 0.97). After cisplatin exposure, BAX, and CASP3 levels increased in cisplatin-treated-J cybrids compared to the untreated-J cybrids (1.78 ± 0.26-fold, P = 0.05 and 3.33 ± 0.62-fold, P = 0.02, respectively). In contrast, the treated and untreated-H cybrids were not significantly different from each other (BAX, P = 0.32 and CASP3, P = 0.12). The BBC3, BCL2L13, and CASP9 gene levels were similar in H and J cybrids with or without cisplatin treatment, indicating that cisplatin did not affect these apoptotic genes. These findings indicate that in the J cybrids, cisplatin-induced apoptosis upregulation of BAX and CASP3.

Signaling Pathway Genes
The cisplatin-treated-J cybrids demonstrated lower EFEMP1 gene expression levels than cisplatin-treated-H cybrids (0.31 ± 0.08-fold, P = 0.04). The MAPK8, MAPK10, and FOXM1 expression levels were not different between H and J cybrids, nor were they affected by cisplatin treatment (data not shown).

Target Sites for Cisplatin Within the MT-Dloop and Comparison of mtDNA GG Stretches
The entire control regions of H (n = 6) and J (n = 7) cybrids were sequenced ( Table 3) and analyzed for the numbers of GG stretches, which are known to be target DNA sequences for cisplatin (28). The MT-Dloop was analyzed because it is the region controlling replication and transcription for mtDNA. In H cybrids (n = 6) and J cybrids (n = 7) there were three GGG stretches (nt16455-16457; nt16516-16518, and nt34-36), one GGGGG stretch (nt16470-16474), and one GGGGGG stretch (nt66-71) site. One H cybrid had a GGGG stretch (nt322-325) that was lacking in any of other H or J cybrids. Greatest variability was found in regions of GG stretches: nucleotides (nt) 184-185 (5/7 in J cybrids); nt228-229 (4/7 in J cybrids); nt322-323 (5/6 in H cybrids); nt513-514 (1/6 H cybrids); and nt526-527 (1/6 H cybrids). A difference in GG stretch patterns of H vs. J cybrids may potentially lead to variations in numbers of cisplatin-mtDNA adducts. However, we believe additional studies will be needed to clarify mechanisms of interaction between the mtDNA and cisplatin that might affect cellular responses.

Cell Culture Studies
Although cybrids have identical nuclei and culture conditions, the cell lines with J haplogroup mtDNA exhibit different responses to cisplatin than cybrids with H haplogroup mtDNA. The untreated-J cybrids have significantly increased rates of growth compared to untreated-H cybrids (226 vs. 100%, P = 0.001), a finding consistent with a previous study (17). After treatment with cisplatin, J cybrids show a dose-dependent decrease in cell viability with a 35% decline at 25 µM cisplatin (P = 0.044) and 58% decline at 50 µM cisplatin (P = 0.0023) compared to the untreated-J cultures. In contrast, H cybrids had non-significant 12% decrease at 25 µM and 38% decline at 50 µM (P = 0.05) compared to the untreated-H cybrids. The large decline in cell viability for J cybrids may be because cisplatin, similar to many anti-cancer drugs, is more effective on rapidly growing cells (1), which is the status of cells containing J mtDNA haplogroup patterns (17,29). Alternatively, it may be that the differential effects of cisplatin are related to J cybrids having lower oxygen consumption, ATP levels, and mitochondrial membrane potential (17,30). Interestingly, Ghelli et al. reported that Leber's Hereditary Optic Neuropathy (LHON) cell lines with the J haplogroup showed increased sensitivity to 2,5-hexanedione (2,5-HD), a toxic solvent that causes neurological and retinal pathology after exposure (31).
The m decreased significantly in cisplatin-treated-J cybrids but not in cisplatin-treated-H cybrids compared to their untreated controls. The decline in m represents early changes that can lead to downstream events such as apoptosis. The release of cytochrome C and induction of intracellular apoptosis are mediated through the voltage-dependent anion channel (VDAC) and cisplatin binds to cysteine and methionine sites of the VDAC (32). The mitochondrial size, shape, and degree of fragmentation can affect binding capacity of BAX, causing changes in mitochondrial outer membrane permeability and apoptosis. The lower m levels in cisplatin-treated-J cybrids are consistent with qRT-PCR results showing increased apoptotic gene expressions (BAX and CASP3) compared to untreated-J cybrids, while the levels in H cybrids did not vary after cisplatin treatment. Our findings show that the mtDNA variants within cells can mediate cisplatin-induced pro-apoptosis events that might contribute to the degrees of toxicity and/or resistance in different individuals.
The cisplatin-treated-H cybrids showed higher ROS levels compared to the cisplatin-treated-J cybrids, which is not completely surprising because the H cybrids utilize OXPHOS, a system that can generate endogenous ROS, while J cybrids use predominantly glycolysis (17). In addition, there are reports that cisplatin reduces mitochondrial respiration complexes I-IV activity by 15-55%, resulting in higher ROS generation in porcine proximal tubular cells (33). A similar stimulus of ROS production may occur in H cybrids because of their reliance on the OXPHOS bioenergetics.
The RPE cell line used in this study is non-cancerous (ARPE-19) and cancer cells genomes may respond differently to cisplatin treatment. One side effect of cisplatin therapy is mild to moderate pigmentary retinopathy (abnormalities of the RPE cells) that occurs in some patients but not in others. Our findings demonstrate that in vitro the human RPE cells are affected deleteriously by cisplatin treatment but depending upon the mtDNA haplogroup (H vs. J), the responses are differentially expressed. This differential response may contribute to the pigmentary retinopathy found in some patients but lacking in other subjects (7).

Differentially Expressed Genes
We looked at several genes involved with apoptosis: BAX, BBC3, BCL2L13, CASP9, and CASP3. RNA levels for BAX and CASP3 were upregulated in cisplatin-treated-J cybrids compared to untreated-J cybrids, but the H treated cybrids remained similar to untreated controls. Cisplatin-induced upregulation of CASP3, the downstream effector gene for apoptosis, has also been reported in carcinoma cells (34). Consistent with elevated apoptosis genes, RNA levels for BAX were upregulated after cisplatin treatment in J cybrids in comparison to untreated-J cybrids. BAX proteins help form porous defects in the mitochondrial outer membrane, leading to release of apoptotic factors (26). In a mouse model, cisplatin has been shown to induce BAX in renal tubular cells (35). Mice deficient of BAX show less cytochrome C release from mitochondria, lower levels of renal tubular apoptosis, and increased resistance to cisplatin  (35). Further studies into the mechanisms of BAX and CASP3 upregulation in our cybrids are needed. SFRP1 levels were not significantly changed after cisplatin treatment in H cybrids. In contrast, the SFRP1 levels were lower in cisplatin-treated-J cybrids compared to untreated-J cybrids (P = 0.05). This is significant because all nuclei and culture conditions are identical, indicating that cisplatin has different effects on cells depending on whether they contain H or J mtDNA. SFRP1 is an extracellular inhibitor of the WNT pathway and acts as a tumor suppressor. Lower SFRP1 levels, as seen in the cisplatin-treated-J cybrids, can be associated with resistance to cisplatin and poorer patient survivability (36).
The CDKN1/p21 gene levels were higher after cisplatin treatment in H cybrids compared to untreated-H cybrids (P = 0.002). In contrast, the CDKN1/p21 gene levels were not changed in the J cybrids after treatment (P = 0.12). Higher levels of CDKN1/p21 are significant because of its role in cellular responses to DNA damage and cell cycle arrest. Overexpression of p21 inhibits colony formation of tumor cells (37). In addition, Duensing et al. reported that abnormal expression of p21 is associated with chromosomal instability as seen in tumor cells (38). Usually, induction of CDKN1A mRNA level is dependent on Tp53 (39) but in our H and J cybrids, the Tp53 expression levels were not changed in response to cisplatin treatment (data not shown). This suggests that upregulation of CDKN1A/p21, as seen in cisplatin-treated-H cybrids, is independent of Tp53 expression.
There was an interesting disparity in the J and H cybrids after cisplatin treatment. The untreated-H and J cybrids had similar expression levels for ABCC1, DHRS2/HEP27, and EFEMP1 but after cisplatin-treatment, the treated-J cybrids showed lower expression levels compared to the untreated-J cybrids (P = 0.02, P = 0.02, and P = 0.04, respectively), while the cisplatin-treated-H were similar to the H-untreated. After cisplatin treatment, higher expression levels of DHRS2/HEP27 were found in the cisplatin-treated-H cybrids compared to the cisplatin-treated-J cybrids. An accumulation of DHRS2/HEP27 pre-protein in the mitochondria matrix can lead to mature DHRS2/HEP27 translocating to the nucleus, where it eventually binds and inhibits MDM2, leading to stabilization of the Tp53 pathway and indirect tumor repressor functions within the cell. Higher levels of DHRS2/HEP27 gene expression in cisplatin-treated-H cybrids could be associated with MDM2-mediated breakdown of the Tp53 gene, and ultimately stabilization of cellular homeostasis (40).
ABCC1 functions as a transporter, mediates export of drugs from the cytoplasm, and also confers resistance to anti-cancer drugs (41,42). In osteosarcoma cells, ABCC1 was the most relevant transporter associated with drug resistance, along with the ABCB1 transporter (27). After cisplatin treatment, upregulations of ABCC1 and EFEMP1 were found in the cisplatintreated-H cybrids compared to the cisplatin-treated-J cybrids. The higher ABCC1 levels in H cybrids suggest these cells may be more likely to become resistant to anticancer drugs than cells with J mtDNA. If our in vitro findings represent in vivo events, then patients with J haplogroup mtDNA might be less likely to become resistant to anticancer drugs while the H haplogroup patients are more likely. EFEMP1 is a gene associated with the production of EGF-containing fibulin-like extracellular matrix protein 1. Overexpression of this gene has been linked to in vivo and in vitro chemotherapeutic drug resistance in cases used to treat glioblastoma tumor growth. The higher EFEMP1 gene expression levels in H cybrids may be a precursor to drug resistant cybrid cells treated with cisplatin (43). Future studies should be conducted to determine if there is correlation between mtDNA haplogroup patterns and cisplatin resistance.
Based upon our findings, it is apparent that after cisplatin treatment, cells containing J haplogroups (Northern European) variants show different biological behavior and gene expression patterns compared to those cells possessing the Southern European H haplogroup mtDNA. At this time, it is not clear how the H vs. J mtDNA profiles influence the nuclear gene expression in cybrids. The effects may be not strictly related to the haplogroup mtDNA as a whole but rather the presence of specific SNPs, which in this case, are defining the J haplogroup. For example, the m10398A>G polymorphism (MT-ND3, ALA-THR) defines the J haplogroup (www.MitoMap.org). The m10398A>G has been associated with breast cancer in African-American women (44), although no association was found in another study (45). Another J haplogroup-defining SNP, m13708G>A (MT-ND5, ALA-THR), has also been associated with breast cancer (46). Even though these SNPs were located in different mtDNA encoding genes, each of these has an amino acid change from Alanine, a non-polar, neutral amino acid to the polar, neutral Threonine, which could change bonding and function of proteins. In contrast, when mtDNA from H cybrids were sequenced, there were no mtDNA variant/mutations that have been associated with human cancers (47). One can speculate that one or all of these J haplogroup-defining SNPs can contribute to changes in retrograde signaling between the mtDNA and nuclear genomes. While further work is needed to identify mechanisms and pathways, our findings support the hypothesis that an individual's mtDNA background can contribute to the response to cisplatin, which is important for the drug's efficacy, level of side effect and development of resistance.
Sequencing of the Entire mtDNA NGS analyses performed on the mtDNA from each of the H and J cybrids showed that the majority of the SNPs identified were haplogroup defining. The private SNPs (non-haplogroup defining), unique SNPs (not listed in www.MitoMap.org), and heteroplasmy SNPs were found in individual cybrids and not throughout all of the H or J cybrids. This suggests that the differential retrograde signaling between the H and J mtDNA haplogroups is due to the accumulation of the haplogroup defining SNPs rather than a single mutation or private SNP. Our findings are consistent with the sequencing results from another cybrid study that used allelic discrimination and Sanger sequencing to identify the mtDNA haplogroups (48). The advancement of NGS for mtDNA analyses allowed deep sequencing in ranges from 1,000 to 100,000 with an average depth of 30,000 so that low level heteroplasmy could be reliably identified. In addition, our method allowed for both strands of mtDNA to be independently sequenced and in both directions, which helps to distinguish between artifact and low level heteroplasmy. The mechanisms of retrograde signaling for the different mtDNA haplogroups are under investigation and likely include as of yet unidentified pathways.

Comparison of GG Stretches Within the MT-DLoop Region
Our previous report related to the MT-DLoop region showed the greatest SNP variations in J vs. H mtDNA were in (a) the nucleotides (np) 263-461 region; (b) Conserved Sequence Block 2 region (np 299-315); (c), H Strand Origin region (np 110-441); (4) Hypervariable Segment 2 region (np 57-372); and (5) the np 310-321 region that had high variability with C insertions (48). However, H and J mtDNA had similar total numbers of CpG and non-CpG methylation sites in the MT-DLoop. The mtDNA is a target for epigenetic modifications and altered methylation patterns have been associated with diseases, drug exposure and aging (49)(50)(51). However, the degree of mtDNA methylation is still controversial and some have suggested that the methylation levels are very low to absent in mtDNA (52,53). Cisplatin causes DNA damage by adduct formation at intra-strand d(GpG) crosslink sites. The rate of DNA adduct formation increases in acidic conditions (28,54) and higher numbers of GG stretches may lead to more binding of cisplatin to the mtDNA. Importantly, our previous bioenergetic studies showed that J cybrids preferentially use glycolysis and have higher levels of extracellular acidification rates (ECAR) than the cybrids with H haplogroup mtDNA (48), leading to the possibility that the H and J mtDNA MT-DLoop may possess different levels of GG sites. However, we found that within the MT-Dloop regions, the numbers of GG stretches (GG, GGG, GGGG, and GGGGG) were similar in H and J cybrids. Therefore, the environmental microenvironment may be playing a bigger role in cisplatin-related adduct formation rather than the numbers of GG stretches in the mtDNA but further work is needed to clarify this question.

ETHICS STATEMENT
This study was carried out in accordance with the recommendations from the Institutional Review Board (#2003-3131) of the University of California Irvine with written informed consent from all patients. All subjects gave written informed consent in accordance with the Declaration of Helsinki. The protocol was approved by the University of California Irvine.

AUTHOR CONTRIBUTIONS
TP designed experiments, interpreted data, wrote manuscript. LN, CL, and KT interpreted data, wrote manuscript. SC, SA, SL, SJ, MM, and DB interpreted data. MC designed experiments, interpreted data. SRA and NU designed and performed the sequencing experiments and interpreted data. MK designed experiments, Interpreted data, wrote manuscript, provided funding resources.