Digital PCR Quantitation of Muscle Mitochondrial DNA: Age, Fiber Type, and Mutation-Induced Changes

Abstract

Definitive quantitation of mitochondrial DNA (mtDNA) and mtDNA deletion mutation abundances would help clarify the role of mtDNA instability in aging. To more accurately quantify mtDNA, we applied the emerging technique of digital polymerase chain reaction to individual muscle fibers and muscle homogenates from aged rodents. Individual fiber mtDNA content correlated with fiber type and decreased with age. We adapted a digital polymerase chain reaction deletion assay that was accurate in mixing experiments to a mutation frequency of 0.03% and quantitated an age-induced increase in deletion frequency from rat muscle homogenates. Importantly, the deletion frequency measured in muscle homogenates strongly correlated with electron transport chain–deficient fiber abundance determined by histochemical analyses. These data clarify the temporal accumulation of mtDNA deletions that lead to electron chain–deficient fibers, a process culminating in muscle fiber loss.

Mitochondrial DNA (mtDNA) sequence and copy number alterations are implicated in the physiological changes of aging skeletal muscle (1). Studies of mtDNA copy number in aging muscle have found increases (2), decreases (3–5), or no change (6). Differences in these studies may arise from the small magnitude of change with age (ie, often less than twofold), differences in tissue sampling (eg, different muscles, biopsy locations, or different fiber type compositions) and differences in mtDNA quantitation approach (eg, Southern blot vs quantitative polymerase chain reaction [PCR]). These differences may be compounded by normalization to nuclear DNA copy number, which itself changes with age (7) and the reporting of fold-changes rather than absolute quantities.

With age, mtDNA deletions originate and accumulate within individual cells. Because cells contain many copies of the mitochondrial genome, deletions exist in a heteroplasmic state with wild-type molecules. When deletion abundance exceeds the phenotypic threshold (8), there is a loss of electron transport chain (ETC) activity. An example of this is the accumulation of mitochondrial DNA deletion mutations to high levels (>90% of mtDNA) in segments of aged muscle fibers, a process that results in a concomitant loss of ETC activity (COX-) (Figure 1). The loss of ETC activity results in muscle fiber apoptosis and necrosis (9). The cumulative loss of ETC-deficient muscle fibers over time contributes to age-induced frailty (9–11). Thus, accurate measurement of mtDNA copy number and mtDNA deletion abundance is critical for understanding heteroplasmy in single cells. In tissue homogenates, accurate measurement of heteroplasmy would permit estimation of the number of ETC-deficient cells, impending cell loss and functional decline (12).

Figure 1.

Mammalian mitochondrial genome, primer locations, and representative micrographs of an electron transport chain (ETC)-deficient muscle fiber. (A) The ~16,300 mammalian mitochondrial genome encodes components of the ETC as indicated. Approximate locations of the primers used in the manuscript are shown. (B) Representative example of ETC-deficient fiber (X). Enzymatic staining for cytochrome c oxidase, succinate dehydrogenase and dual histochemical staining on serial sections of Rectus femoris. The black bar denotes 100 microns.

Following the first detection of mtDNA deletions by hybridization approaches (13,14), methods have evolved toward more sensitive and accurate quantification in both single cells and tissue homogenate samples (10,15–18). Two significant challenges of accurate mtDNA deletion quantitation were succinctly described by Kraytsberg and colleagues (19): (i) the low percentage of deleted molecules and (ii) the complex spectra of mtDNA deletion breakpoints in aging tissues. The challenge of low mutation frequency has been addressed by restriction digestion to ablate wild-type mtDNA, enriching the sample for mutant targets (19–22). Although deletion breakpoints vary widely, the ubiquitous loss of major arc segments (Figure 1) (17,23–26) has led to the exploitation of this feature in quantification (16,27–29). The recent emergence of digital PCR (dPCR) has extended the sensitivity of this approach and permitted absolute quantification at lower mutation frequencies without the need for external calibrators (21,22,30,31). The advantages of dPCR over qPCR using the loss of major arc approach include absolute quantitation, improved precision, and reduced assay standard error (29).

Previous determinations of mtDNA deletion abundance within an entire muscle were inferred from counting the number of ETC-deficient fibers within a tissue sample using histological methods. In the current study, we apply two distinct digital PCR approaches to quantitate mtDNA content and mutation frequencies in single cells and tissue homogenate samples. We found that single fiber mtDNA copy number reflects fiber biochemical type and is affected by age. We contrasted indirect and direct methods of deletion measurement and found that a direct method permitted the quantification of deletion abundance with respect to both age and the underlying focal mutation accumulation, that is, ETC-deficient fiber abundance.

MATERIALS AND METHODS

Animals and Tissues

This study was carried out in accordance with the recommendations in the NIH Guide for Care and Use of Laboratory Animals and the guidelines of the Canadian Council on Animal Care. The protocols used were approved by the Institutional Animal Care and Use Committees at the University of Alberta and UCLA. Male C57Bl/6JNia mice or Fischer 344 x Brown Norway F1 hybrid rats were obtained from the NIA Aging Rodent Colony.

We chose quadriceps muscle because it contains a mixed fiber type, exhibits considerable muscle mass and fiber loss, and accrues significant ETC-deficient fibers with age (32). Quadriceps muscles were dissected from rodents at specified ages and cryo-embedded in OCT or snap frozen in liquid nitrogen. Tissues were stored at −80°C. Thirty months old rats were treated with guanidinopropionic acid (GPA) for 4 months as described (31). GPA treatment induces mitochondrial biogenesis and drives the replication of deletions in old rats. Therefore, muscle samples from GPA-treated aged rats are a useful positive control because of their elevated abundance of mtDNA deletions and ETC-deficient fibers.

Muscle Homogenate DNA Isolation

Frozen rodent quadriceps muscles were powdered using a mortar, pestle, and liquid nitrogen. Total DNA was extracted by proteinase K digestion, phenol/chloroform extraction, and ammonium acetate precipitation. Total DNA quantity and quality were measured using spectrophotometry at A230, A260, and A280 (Thermo Scientific Nanodrop 2000 Spectrophotometer) and integrity examined by gel electrophoresis.

Histochemical and Immunohistochemical Staining

Muscle fiber types were identified in 10-micron thick cryosections by immunohistochemical (IHC) staining of serial sections with an anti-myosin antibody, My-32 (Sigma-Aldrich, St. Louis, MO), to identify type I and type II fibers and succinate dehydrogenase histochemistry to differentiate type IIa from type IIb fibers. Histochemical staining for cytochrome c oxidase (COX), succinate dehydrogenase, and identification of ETC-deficient fibers was performed as previously described (11,31).

Laser Microdissection

Single muscle fiber sections were captured using a Leica LMD 7000 into slide microwells (Supplementary Figures 1 and 2). The microwells facilitated tissue digestion and confirmation of digestion in 1 µl of proteinase K digestion solution as described (10,11). Digested cells were resuspended to a total volume of 10 µl with nuclease-free water.

Quantitative PCR

Quality assessment and rough estimation of mtDNA copy number was obtained by quantitative PCR (qPCR) using 5-prime nuclease cleavage assays specific to the mouse or rat mitochondrial genome purchased from Integrated DNA Technologies (Coralville, IA) or ThermoFisher (Waltham, MA) (Supplementary Table 1). qPCR reactions included 1 µl of target DNA, 1.25 µl of 20× ND4/ND1 primer probe mix, 12.5 µl of 2× AmpliTaq Master Mix (ThermoFisher), and 10.25 µl of nuclease-free water for each 25 µl reaction. qPCR was performed with standard TaqMan cycling conditions in a BioRad CFX96 Real Time Thermocycler. Final primer and probe concentrations were 900 and 250 nM, respectively.

Digital PCR

A 5-prime nuclease cleavage assay specific for the mouse mitochondrial minor arc ND1 gene was used to quantitate total wild-type mtDNA copy number using a chip-based dPCR approach (Figure 2A). Samples were diluted to the manufacturer’s recommended target range (200 copies per µl to 2,000 copies per µl) in 25 µl reactions using QuantStudio 3D Digital PCR Master Mix v2 (ThermoFisher) and loaded onto Quantstudio 3D digital PCR 20K Chip (Version 2, ThermoFisher). Final primer and probe concentrations were 900 and 250 nM, respectively. Similar to Belmonte and colleagues (29), we did not observe differences in results between multiplex and singleplex dPCR reactions. To enhance precision of singleplex reactions analyzed for ND1 and ND4 copy number, we divided assembled reactions prior to the addition of the primer/probe.

Figure 2.

Quantitative analysis of mitochondrial genome abundance in single laser-captured muscle fiber sections by digital polymerase chain reaction (PCR). (A) A graphical representation of the 16,300 bp circular wild-type mitochondrial genome and the location of the primer sites in minor arc used to quantify mtDNA abundance by digital PCR. O_H and O_L denote the heavy and light strand origins of replication, respectively. (B) ~12,000 copies of mtDNA were partitioned into 20,000 individual wells on a nanofluidic chip. Following PCR amplification, the presence of fluorescent signal was detected in each well that received an amplifiable template of mtDNA. (C) The fraction of wells with a fluorescence signal above a mathematically determined threshold (3429.24 in C) is used to determine the sample copy number. (D) MtDNA abundance in type I, IIa, and IIb quadriceps skeletal muscle fibers identified by IHC and isolated by laser microdissection from 6- and 24-month-old mice. MtDNA copy number is normalized to the cross-sectional area of the captured fibers to account for heterogeneity in fiber cross-sectional area of skeletal muscle fibers of different types. Different letters denote significance at p less than .05.

Digital PCR cycling conditions were Taq-polymerase activation at 96°C for 10 minutes, 40 cycles of denaturation at 98°C for 30 seconds, and annealing/extension at 60°C for 2 minutes. MtDNA copy numbers per microliter and the threshold were determined using QuantStudio 3D Analysis Suite Cloud Software (Version 3, ThermoFisher). Direct quantitation of the major arc deletions by dPCR used the same cycling conditions but for 60 cycles. Quantitation of mtDNA deletions from muscle homogenates was performed on blinded samples.

Long-Extension PCR

Rat mtDNA deletions were amplified from laser microdissected muscle fibers using the rat F2485 and R322 primer set (Supplementary Table 1) purchased from IDT. Long extension PCR reactions were assembled according to the manufacturer’s instructions, Go-Taq Long PCR master mix (Promega, Madison, WI). PCR cycling conditions were polymerase activation at 95°C for 2 minutes, denaturation at 94°C for 20 seconds, and annealing at 68°C for 10 minutes, repeated for 40 cycles. PCR products were fractionated on 1% agarose gels, stained with ethidium bromide, and visualized under UV light.

Mixing Experiments

Double-stranded DNA targets representing wild-type and deletion mutation targets were generated by long extension PCR amplification of normal and ETC-deficient muscle fibers obtained by laser capture microdissection (Supplementary Figure 3 and Supplementary Table 1). Wild-type and deletion DNA targets were column purified and quantitated using spectrophotometry and digital PCR. The wild-type target spanned 14,137 bp of the rat mitochondrial genome. The deletion target spanned the same region of the mtDNA with residues between bp 6,344 and 15,002 deleted resulting in an LX-PCR product of 5,479 bp.

Statistical Analysis

Individual data and the mean ± SD are plotted. Data were analyzed by ANOVA followed by pairwise Student’s t tests. When comparing mtDNA deletion frequencies from different aged rats, mutation frequencies were log transformed to adjust for skew. Linear regression was used to examine the correlation between ETC-deficient fiber abundance and mtDNA deletion frequency.

Results

Single Muscle Fiber mtDNA Quantification by Digital PCR

We quantitated the absolute abundance of mtDNA within individual quadriceps muscle fibers of 6- and 24-month old mice with dPCR. We accounted for the age-induced tissue and fiber type heterogeneity by laser microdissection of individual skeletal muscle fibers (Supplementary Figures 1 and 2). Although type II fibers predominated in the quadriceps, type I fibers were observed in the rectus intermedius and in regions of the vastus lateralis, vastus medialis and rectus femoris closest to the femur. In contrast to the young mouse quadriceps, sections of old quadriceps showed grouping of type I and type IIA fibers (data not shown).

Absolute copy number of mtDNA within individual fibers was normalized to fiber cross-sectional area. Significant differences were observed in mtDNA copy numbers between fiber types with the highest copy numbers in type I fibers followed by type IIa and then type IIb fibers (Figure 2D and Supplementary Table 2) and this effect was independent of age. Within individual fiber types, mtDNA copy number decreased with age. In 24-month-old mouse muscle fibers, mtDNA copy number declined by 40.4% in type I fibers, 47% in type IIa, and 68.8% in type IIb fibers.

Indirect Quantitation of mtDNA Deletions

MtDNA deletion mutations accumulate with age, are large, with 4–10 kb lost and are predominantly located in the major arc region of the genome region of the genome (Figure 3A). The loss of major arc genes centered around ND4 and preservation of minor arc genes such as ND1 has enabled quantitation by comparison of the abundances between these two target sites (16,31,33). The minor arc target (eg, ND1) detects both wild-type and deletion-containing molecules and, therefore, represents total mtDNA abundance. Conversely, the major arc target detects only wild-type molecules because deletion containing molecules lack the target site (eg, ND4). In a sample comprised of only full length, intact mtDNA molecules, the expected wild-type:total copy number ratio would be 1. In contrast, a heteroplasmic sample containing a mixture of deletions (ie, lacking the major arc) and wild-type molecules, the wild-type:total copy number ratio would be less than 1. We denote this as the “indirect” approach as it quantifies deletion abundance by the relative loss of the major arc site.

Figure 3.

Indirect approach for estimation of heteroplasmy by digital polymerase chain reaction (PCR) in rat muscle homogenates. (A) Schematic depiction of a representative mtDNA deletion and PCR primer locations that detect total (mtDNA deletion and wild-type) or wild-type only mtDNA in the minor and major arc, respectively. (B) Copy number of total and wild-type mtDNA molecules from a spike-in experiment. Experimentally determined copy numbers were used to estimate sample heteroplasmy. Error bars indicate the variance in technical replicates. (C) Comparison of total and wild-type mtDNA copy number variation in rat quadriceps muscle homogenates with different mutation loads. Different letters denote significance at p less than .05.

We utilized this indirect approach on experimentally mixed samples of rat wild-type and deletion-containing molecules in fixed ratios while holding constant the total mtDNA. The abundances of wild-type and total mtDNA molecules were quantified by dPCR (Figure 3B). As the fraction of deletion-containing molecules increased, we detected the corresponding loss of the wild-type target. The percent heteroplasmy calculated from the dPCR results closely matched the known mutant fraction until the percent heteroplasmy dropped below 50%. In a control sample comprised of only wild-type mtDNA target, the measured heteroplasmy was 11%, a significant deviation from the expected 0% heteroplasmy.

Coupling this indirect approach with laser microdissection of individual skeletal muscle fibers from aged mice, we measured the difference between total and wild-type mtDNA in an ETC normal control fiber to be 3.7%, whereas ETC-deficient fibers showed greater than 80% heteroplasmy for mutant mtDNA (Supplementary Figure 3). The levels of heteroplasmy measured in the ETC-deficient and control fibers were consistent with previous findings in ETC-deficient fibers in rats (10,31) and humans (17), confirming the utility of this approach in ETC-deficient fiber segments where mtDNA deletions are relatively abundant.

We applied the indirect approach to muscle homogenates from rats of different ages and following treatment with GPA, which drives the clonal amplification of latent mtDNA deletions at old ages (31). We chose these samples because they contain a large range of ETC-deficient fiber abundances. Using the indirect deletion detection approach, there was no detectable difference in mtDNA deletion abundance between 12-month and 34-month-old rat muscles even though 34-month-old muscle contains an increased burden of ETC-deficient fibers. The indirect approach confirmed the significant mutation abundance in samples derived from 34-month-old GPA-treated rats (Figure 3C) and is consistent with the 1,200% increase in mtDNA deletion-containing ETC-deficient muscle fibers in these rats (31).

Direct Quantitation of mtDNA Deletions

The wild-type:total copy number ratio approach is an indirect assay that relies upon the relative absence of the major arc genes to infer the presence of mtDNA deletions. To directly detect deletions, we adapted an established approach that relies on a large deletion to bring two primers located outside the major arc close enough to amplify a product (Figure 4A). Limiting the PCR extension time selectively amplifies mtDNA deletions (25,34–36). To accurately quantitate the discrete mutation events, we used digital PCR to partition individual mtDNA deletion molecules into discrete PCR reactions. Using this direct approach, we quantitated a synthetic rat deletion mutation in control experiments down to a level of 0.03%. Detection of deletions was unaffected by the addition of increasing amounts of wild-type mitochondrial DNA (Figure 4B). This sensitivity was sufficient to quantitate the abundance of mtDNA deletions in muscle homogenates from normally aged rats. In rat quadriceps muscle homogenate samples, the deletion frequency increased from 0.002% to 0.121% between 12 and 34 months of age and up to 0.41% in positive control muscle homogenates from 34-month-old, GPA-treated, rats (Figure 4C).

Figure 4.

Direct detection of mtDNA deletions and quantification of mtDNA deletion frequency in tissue homogenates by digital polymerase chain reaction (PCR). (A) Schematic depiction of a representative mtDNA deletion and PCR primer locations that detect total (mtDNA deletion and wild-type) or major arc mtDNA deletions. (B) Percent mtDNA deletions measured in a spike-in experiment. MtDNA deletion and wild-type molecules were mixed as indicated and mutation frequency was measured over two ranges of mutation frequency. Error bars indicate the variance in technical replicates. (C) Digital PCR quantification of age-induced and guanidinopropionic acid-expanded mtDNA deletions in muscle homogenates using major arc deletion primers. Different letters denote significance at p less than .05.

MtDNA Deletion Frequency Predicts ETC-Deficient Fiber Number

Using the direct assay, we quantitated mtDNA deletions across the rat life span in quadriceps muscle homogenate samples. We found that deletion frequency exponentially increased with age beginning around 24 months (Figure 5A). Histological analyses showed that the ETC-deficient fiber abundance also increased exponentially from the baseline at 30 months of age and paralleled the increase in deletion frequency. To inspect the relationship between ETC-deficient fiber abundance and deletion frequency, we analyzed tissues from the same rats by both histological and molecular methods. We found a close correlation between ETC fiber abundance and mtDNA deletion frequency in contralateral muscle samples (R² = .985; Figure 5B).

Figure 5.

Integrating molecular and histologic analyses of mtDNA deletion abundance in Fischer/Brown Norway F1 rats. (A) MtDNA deletion frequency and electron transport chain (ETC)-deficient cell abundance increase exponentially with age. Data are normalized relative to the 12-month-old samples. ETC-deficient fiber abundance was determined by directly counting the COX-/SDH++ fibers from 10 µm thick cross-sections along 1 mm of rat vastus lateralis. (B) Positive correlation between direct deletion digital polymerase chain reaction mutation frequencies from individual muscle homogenates and corresponding ETC-deficient muscle fiber abundance determined by histochemical analysis. The 95% confidence band is plotted around the regression line with dashed curves and the slope of the regression line was non-zero, p less than .0001. COX = cytochrome c oxidase; SDH = succinate dehydrogenase.

Discussion

Digital PCR approaches are well suited to the analysis of mtDNA heterogeneity induced by aging. By partitioning mtDNA genomes into thousands of individual PCR reactions, dPCR facilitates absolute quantitation without the use of standard curves or normalization to targets that may be affected by the underlying biology or sampling process. Partitioning samples to single target molecules also dilutes competing targets such as nuclear DNA and allows for sensitive detection of rare targets and precision to ~10% (37).

There is conflicting data on age-induced changes in muscle mtDNA abundance using qPCR or semiquantitative Southern blot approaches (2–4,6). Homogenate approaches are unable to account for sampling or fiber type differences and include cell types other than muscle fibers that may confound copy number differences. Applying a digital PCR approach to single muscle fiber sections, we found significant age and fiber-type-induced alterations in mtDNA copy number. Absolute quantitation of mtDNA content in single fibers demonstrates that mtDNA copy number reflects the fiber-specific metabolic phenotype. Type I oxidative fibers had the highest mtDNA copy numbers. Normalization to nuclear DNA content, which may also change with age and fiber type, was also avoided by our approach. Measurements of fiber mtDNA copy number by digital PCR are comparable to those using qPCR (33,38) but show less variability between individual fibers of the same type.

Indirect mtDNA deletion quantification using a wild-type to total mtDNA ratio approach has recently been applied using dPCR (29,31). Belmonte and colleagues (29) used this approach to quantify deletion abundance down to 2.5% in mixing experiments. Similarly, our data also show that the indirect approach is accurate at higher deletion frequencies and is useful for measuring mtDNA heteroplasmy in single ETC-deficient fibers (Supplementary Figure 3). In contrast to the indirect approach, direct detection of the deletion breakpoint event facilitates quantitation of lower deletion frequencies (19,21,22,30). These approaches use restriction endonuclease digestion to discriminate between wild-type and deleted mtDNA molecules and interrogate a portion of the major arc. We adapted these direct mutation detection methods to quantify a wider range of mtDNA deletion events using primers that flank the entire major arc. The direct dPCR quantification of mtDNA deletions builds on earlier long-extension PCR approaches that discriminate wild-type from deletion mutant mtDNA by restricting PCR extension time while leveraging the advantages of dPCR (25,34–36). Digital PCR assays utilize end-point analysis of each partition to generate quantitative data and are independent of PCR efficiency. This independence from PCR efficiency permits the equivalent quantification of deletions despite size-specific differences in PCR efficiency.

MtDNA deletion abundances are known to increase with age in muscle (36,39,40) but details regarding the age at which they increase and the rate of increase have been lacking. Our quantitation of deletions using direct dPCR illuminated these kinetics in muscle aging. In rat muscle, deletion frequency increased above baseline at ~24 months of age. The exponential increase in deletion frequency preceded ETC-deficient fibers and fiber and mass loss, which accelerates at 30 months of age in the F344xBN rat strain (41). The exponential increase in deletions, beginning at 24 months of age, contrasts with other markers of aging whose increase is linear with respect to time throughout the adult life span (42). Accurate quantitation of mtDNA content and mtDNA deletion frequency may provide additional indices of mitochondrial quantity and quality that inform estimates of biological age.

We found a high degree of positive correlation between mtDNA deletion frequency and ETC-deficient fiber abundance regardless of age or treatment group. This strong correlation shows that the deletion frequency quantitated by the direct dPCR approach is a biomarker of the histological mtDNA deletion phenotype. Compared to histological analyses, the direct dPCR approach is faster, less labor and resource intensive, and suitable for screening or human muscle biopsy studies.

Supplementary Material

Supplementary data is available at The Journals of Gerontology, Series A: Biological Sciences and Medical Sciences online.

Funding

This work was supported by the American Federation for Aging Research (J.W.), the Glenn Foundation for Medical Research (J.W.), the UCLA Hartford Center of Excellence (J.W.), National Institutes of Health (AG032873 to J.W., AG030423 to J.M.A., DK060484, DK073227, DK097748 to A.L.H.), UCLA Older Americans Independence Center (AG028748 to J.W.), Ellison Medical Foundation (New Scholar Award to J.W., Senior Scholar Award to J.M.A.), UCSD/UCLA Diabetes Research Center Pilot and Feasibility Grant (DK063491 to J.W.); Laser microdissection was performed at the CNSI Advanced Light Microscopy/Spectroscopy Shared Resource Facility at UCLA, supported with funding from NIH-NCRR shared resources grant (CJX1-443835-WS-29646) and NSF Major Research Instrumentation grant (CHE-0722519) as well as at the Brain and Aging Research Building at the University of Alberta with support from the Canadian Foundation for Innovation (LOF24776 to J.M.A.); UCLA Iris Cantor Women’s Health Center–UCLA Clinical and Translational Science Institute (UL1TR000124 to A.L.H.).