Sequence-Based Polymorphisms in the Mitochondrial D-Loop and Potential SNP Predictors for Chronic Dialysis

Background The mitochondrial (mt) displacement loop (D-loop) is known to accumulate structural alterations and mutations. The aim of this study was to investigate the prevalence of single nucleotide polymorphisms (SNPs) within the D-loop among chronic dialysis patients and healthy controls. Methodology and Principal Findings We enrolled 193 chronic dialysis patients and 704 healthy controls. SNPs were identified by large scale D-loop sequencing and bioinformatic analysis. Chronic dialysis patients had lower body mass index, blood thiols, and cholesterol levels than controls. A total of 77 SNPs matched with the positions in reference of the Revised Cambridge Reference Sequence (CRS) were found in the study population. Chronic dialysis patients had a significantly higher incidence of 9 SNPs compared to controls. These include SNP5 (16108Y), SNP17 (16172Y), SNP21 (16223Y), SNP34 (16274R), SNP35 (16278Y), SNP55 (16463R), SNP56 (16519Y), SNP64 (185R), and SNP65 (189R) in D-loop of CRS. Among these SNPs with genotypes, SNP55-G, SNP56-C, and SNP64-A were 4.78, 1.47, and 5.15 times more frequent in dialysis patients compared to controls (P<0.05), respectively. When adjusting the covariates of demographics and comorbidities, SNP64-A was 5.13 times more frequent in dialysis patients compared to controls (P<0.01). Furthermore, SNP64-A was found to be 35.80, 3.48, 4.69, 5,55, and 4.67 times higher in female patients and in patients without diabetes, coronary artery disease, smoking, and hypertension in an independent significance manner (P<0.05), respectively. In patients older than 50 years or with hypertension, SNP34-A and SNP17-C were found to be 7.97 and 3.71 times more frequent (P<0.05) compared to patients younger than 50 years or those without hypertension, respectively. Conclusions and Significance The results of large-scale sequencing suggest that specific SNPs in the mtDNA D-loop are significantly associated with chronic dialysis. These SNPs can be considered as potential predictors for chronic dialysis.


Introduction
Mitochondria (mt) are organelles that are susceptible to oxidative stress. The presence of excessive amounts of reactive oxidative species (ROS) results in mitochondrial oxidative damage and inefficient repair of mtDNA [1][2][3]. This can contribute to pathophysiological processes, including aging, degenerative disease [4][5][6] and cancer [7]. In these circumstances, somatic mutations are also generated [8].
The displacement loop (D-loop) regions of mtDNA does not encode any functional proteins [9,10] and is known to accumulate mutations at a higher frequency than other regions of mtDNA in the setting of increased oxidative stress [11]. The D-loop contains the initial site of heavy chain replication and the promoters for heavy and light chain transcription. Therefore, it is responsible for the regulation of mtDNA replication and transcription [10,11]. The D-loop is highly polymorphic, and some polymorphisms are associated with aging [12][13][14][15], coronary artery disease [16], and a variety of tumors, including lung [17], colorectal [18], liver [19], gastric [20], breast [21], cervical [22], melanoma [23], head and neck [24], oral [25], and kidney [26] cancers. However, D-loop polymorphisms are not associated with prostate cancer [27,28]. Most of these D-loop studies focus on some cancer-associated single nucleotide polymorphisms (SNPs) for mtDNA, which were accompanied by poly-C tract alterations [21,24,25,29,30]. However, D-loop polymorphisms have not been systematically characterized in chronic dialysis patients.
Complications of chronic kidney disease (CKD) promote morbidity and mortality [31]. CKD patients can be classified according to kidney function along a continuum from mild renal dysfunction to irreversible kidney failure. CKD increases oxidative stress [32] which has been demonstrated to influence mtDNA content in CKD patients [33,34].
Because the D-loop region susceptible to oxidative stress, we hypothesized that specific SNP patterns in the D-loop of chronic dialysis patients may serve as potential genetic markers for chronic dialysis. To examine this hypothesis, we performed D-loop sequencing and used bioinformatic tools to identify SNPs that were associated with chronic dialysis when compared to healthy controls.

Subjects
We enrolled 704 unrelated Taiwanese of ethnic Chinese background in this study through the hospital health examination center after giving consent. Participants included 312 men and 392 women with a mean age of 51.9 years. We enrolled 193 dialysis patients from the outpatient dialysis unit of the same hospital. They were composed of 78 men and 115 women with a mean age of 49 years. Venous blood samples were collected after overnight fasting. The serum was separated using a centrifuge and stored at 280uC. DNA was isolated from leucocytes using PUREGENEH DNA Purification kit (Gentra, Minneapolis, MN, USA) and stored at 220uC . The protocol for the present study was approved by the  Committee on Human Research at Kaohsiung Chang Gung  Memorial  Hospital  (CMRPG850271,  CMRPG850272,  CMRPG850242, CMRPG850252, IRB 95-0395B) and conducted in accordance with the Declaration of Helsinki. All participants signed a written informed consent form to obtain the approval for participation in this study.

Assessment of Oxidative and Anti-oxidative Stress Capacities
Serum free thiols were determined by direct reaction of the thiols with 5,5-dithiobis(2-nitrobenzoic acid) (DTNB) to form 5thio-2-nitrobenzoic acid (TNB). The amount of thiols was calculated from the absorbance determined using the extinction coefficient of TNB (A412 = 13,600 M 21 cm 21 ). The serum thiobarbituric acid reactive substance (TBARS) concentration was assessed according to the method of Ohkawa et al. [35]. Results are expressed as micromoles of TBARS per liter. A standard curve of TBARS was obtained by hydrolysis of 1,1,3,3tetraethoxypropane (TEPP).

D-loop Sequencing
The mtDNA control region segment (relative to nucleotide (nt) regions 15911-16569 and 1-602 in the Revised Cambridge Reference Sequence (''rCRS'') [36]; NC_012920) was amplified using the forward primer L15911 (59-ACCAGTCTTG-TAAACCGGAG-39) and the reverse primer H602 (59-GCTTTGAGGAGGTAAGCTAC-39). The products were purified with gel extraction kits (Watson BioMedicals Inc.) and sequenced using primer L15911 and primer L29 (59-CTCACGG-GAGCTCTCCATGC-39) on an ABI 377XL DNA Sequencer (Applied Biosystems, Foster, CA, USA). However, due to the conversion of thymine to cytosine and the presence of homopolymeric cytosine tracts at nt16184-16193 and nt303-315 within the D-loop region of some subjects, the sequencing procedure was prematurely terminated. Therefore, we also performed reverse sequencing using 2 additional sets of primers, H81  and H528 (59-TTCGGGGTATGGGGTTAGCA-39). The polymerase chain reaction (PCR) conditions used were as follows: an initial denaturation step at 95uC for 5 min, followed by 35 cycles of denaturation at 95uC for 1 min, annealing at 60uC for 1 min, and extension at 68uC for 2 min, with a final extension of 10 min at 72uC. The PCR fragments were analyzed by electrophoresis on a 2% agarose gel and visualized by staining with ethidium bromide.

SNP Identification
DNA sequences were analyzed by using the DNASTAR software and Bio Edit Sequence Alignment Editor freeware (http://www.mbio.ncsu.edu/bioedit/bioedit.html). After multiple sequence alignments were performed, both 59 and 39 ends of the sequences were trimmed into blunt ends. The SNPs were identified by calculating each nucleotide (A, T, C, or G) for each position in the trimmed and aligned sequences by ''count if = '' in Excel software. SNP frequencies greater than 1% were selected for further investigation. The SNPs were compared to the D-loop polymorphisms in rCRS as shown in MITOMAP [37] (http:// www.mitomap.org/MITOMAP/PolymorphismsControl).

Statistical Analysis
Chi-square tests were used to compare basic characteristics between patients and controls. A sequence of analyses was adopted for SNP selection. The Chi-square tests were first used to compare distributions of SNPs between patients and controls. Nine SNPs with significant differences and with sufficient cell sizes were chosen for further analysis. These 9 SNPs were included in a logistic regression model with backward selection. Only statistically significant SNPs were selected by logistic regression. The same logistic regression selection process was also conducted for several subgroups. Lastly, the adjusted odds ratios (AOR) from selected SNPs were computed on the basis of logistic regression with additional covariates of basic demographic characteristics ( Table 1). The statistical data were expressed as mean 6 SD. A P value of less than 0.05 was considered as statistically significant.

Basic Demographic Characteristics
The study participants included 193 dialysis patients and 704 healthy controls, and their basic characteristics are shown in Table 1. Most of these characteristics were found to be significantly different, except for age groups and blood TBARS levels. The patients were 3 years younger (49.0613.9 vs. 51.9612.9) than the controls and had lower values of body mass index (BMI), blood thiols, and cholesterol levels. The mean triglyceride (TG) level was higher in patients than in controls. There was a significantly higher incidence of comorbidities of diabetes, hypertension (HT), and coronary heart disease (CHD) in dialysis patients compared to controls.

D-loop Sequencing, Alignment, and SNP Identification
There are 2 poly-C regions in the mitochondrial D-loop that stretch between nt16180-16195 [38] and nt303-315 [9]. Because the length of these mononucleotide repeats varies, they may interfere the sequence alignment processing or lead to error alignment in part. Accordingly, the sequences for these 2 repeat regions were replaced with the corresponding sequences for the reference CRS to improve the performance of sequence alignment. The sequencing data from the 59 and 39 ends of nt15911-16000 and nt486-602 were of poor quality and, therefore, were trimmed after confirmation of sequence alignment. Finally, aligned sequences were trimmed to the same length ranging from nt16000-16569 and nt1-485 for further SNP identification (Table S1 and Table S2; all D-loop trimmed sequences for cases and controls and their alignment visualization, respectively). After examining each nt for each position of the trimmed sequence, 77 SNPs with frequencies greater than 1% were identified (Table S3). The relationships between positions of the aligned sequences and D-loop in the reference CRS as well as the SNP types in the IUPAC code are listed in Table 2.

Significance Analysis for 77 Individual SNPs
The P values for 77 individual SNPs with A, G, C, and T distribution data were analyzed (Table S4). Nine SNPs were selected from 77 SNPs by Chi-square tests with significant differences and sufficient cell sizes; their genotype distributions are compared in Table 3. For each SNP, the genotype that appeared at a higher frequency in patients was selected as the   , and 189R) were T, C, C, A, C, G, C, A, and G, respectively. These 9 indicators were further added into a logistic regression by employing the backward selection method.

Backward Logistic Regression Analysis for 9 SNPs
As shown in Table 4, we identified 3 statistically significant indicators (SNP55 G, SNP56 C, and SNP64 A). Individuals with the SNP55 G increase risk of chronic dialysis by 4.78 times (OR, 95% CI = 1.26,18.09, P = 0.0212). SNP56 C or SNP64 A subjects increase risk of chronic dialysis by 1.47 (95% CI = 1.06,2.04, P = 0.0225) or 5.15 (95% CI = 2.29,11.60, P = 0.0001) times. The AORs of the 3 SNPs were further computed by adding the covariates shown in Table 1 into the logistic regression analysis. Following this, only SNP64 A remained significant (OR = 5.13, 95% CI = 1.61,16.35, P = 0.0057). Hence, SNP64 is only an independent SNP for disease as well as for the patients' basic characteristics. On the other hand, while SNP55 and SNP56 found in the backward logistic regression could only be considered as independent SNPs among the 77 SNPs, they were affected by covariates.

Stepwise Regression for Subgroups Related to Several Basic Demographic Characteristics
Similar procedures were also conducted in several subgroups (Table 5). While the frequencies of SNP55 and SNP64 were found to be significantly higher in women, only those with SNP64 A genotype had a statistically significant higher risk of chronic dialysis (AOR = 35.80, 95% CI = 3.23,396.84, P = 0.004). In subjects older than 50 years, SNP34 A genotype was significantly associated with chronic dialysis (AOR = 7.97, 95% CI = 1.25,50.94, P = 0.028). For subjects without diabetes, without CHD, no smoking habit, or without HT, SNP64 A was the independent SNP in association with chronic dialysis (AOR = 3.48, 4.69, 5.55, and 4.67, P = 0.010, 0.016, and 0.046, respectively). For subjects with history of hypertension, SNP17 C was significantly associated with chronic dialysis (AOR = 3.71, 95% CI = 1.10,12.55, P = 0.035).

Discussion
To date, most association studies of chronic dialysis focus on the nuclear genome [39][40][41][42][43] rather on mtDNA. In our previous report [9], we addressed the association between polymorphisms in the poly-C tract (D310) of the mtDNA D-loop and probability of dialysis treatment. However, we found that the poly-C tract was not significantly different in dialysis patients compared with healthy controls. In addition to the poly-C tract, SNPs are also found in the D-loop. Therefore, we decided to determine whether there was any association between chronic dialysis and SNPs in the D-loop in this study.
Using sequence alignment, we found 9 SNPs present at significantly higher frequency in dialysis patients (SNP5, 17,21,34,35,55,56,64, and 65). Among them, 3 significant indicators (SNP55 G, SNP56 C, and SNP64 A) were independently associated with a high risk of chronic dialysis. Furthermore, only women with the SNP64 A genotype were statistically significant to be associated with chronic dialysis. SNP34 A was significantly associated with chronic dialysis in subjects older than 50 years. For subjects without diabetes, CHD, or hypertension, or in nonsmokers, SNP64 A was statistically associated with chronic dialysis. Individuals with history of hypertension were significantly associated with chronic dialysis if they carried SNP17 C.
In this study, we focused solely on the question of whether individual SNPs within the D-loop were associated with chronic dialysis. However, the consideration of interdependence among SNPs was found to improve the association of genetic variations with several diseases [44,45] and cancers [46][47][48][49][50][51][52][53][54]. Therefore, we cannot exclude the possibility that some rare SNPs may still contribute to the synergistic association with chronic dialysis.
The acquisition of ROS-induced mutations in CKD may be a consequence of increased oxidative burden in patients with chronic renal failure [9,32,33,57]. For example, elevated oxidative stress in chronic peritoneal dialysis patients may lead to alterations in the mtDNA copy number in peripheral leukocytes [33]. In our current study, the mtSNPs listed in Table 3 were homoplasmic, as revealed by sequencing chromatograms (data not shown) [58][59][60]. However, we cannot exclude the possibility that a minor fraction of heteroplasmic mutations, below the level of sensitivity of the sequencing method that we used, may be present. We suggest that additional PCR/restriction fragment length polymorphism (RFLP) analysis may assist in the identification of mitochondrial heteroplasmy [61,62]. In light of this, we are unable to identify mtSNPs that are suitable as progression markers for CKD with our current data, since our sequencing method lacked sufficient sensitivity to detect ROS-induced mutations. Therefore, the biological and clinical significance of the homoplasmic mtSNPs are more suitable as potential genetic markers for chronic dialysis, rather than progression markers of CKD.
To the best of our knowledge, this is the first report of SNPs in the mtDNA D-loop showing that they are significantly associated with chronic dialysis. The study also demonstrated the relationship of SNPs with comorbidities in dialysis patients. One may postulate that the presence of these SNPs is a risk factor for the development of end-stage renal disease, and that they may be used as markers to predict the likelihood of dialysis. In the future, further studies are needed to establish the role of these SNPs in the pathophysiology of CKD and to validate their clinical application.