Disease-Causing Mitochondrial Heteroplasmy Segregated Within Induced Pluripotent Stem Cell Clones Derived from a Patient with MELAS§


  • Clifford D.L. Folmes,

    1. Department of Medicine, Division of Cardiovascular Diseases, Mayo Clinic, Rochester, Minnesota
    2. Center for Regenerative Medicine, Mayo Clinic, Rochester, Minnesota
    Search for more papers by this author
  • Almudena Martinez-Fernandez,

    1. Department of Medicine, Division of Cardiovascular Diseases, Mayo Clinic, Rochester, Minnesota
    2. Center for Regenerative Medicine, Mayo Clinic, Rochester, Minnesota
    Search for more papers by this author
  • Ester Perales-Clemente,

    1. Department of Medicine, Division of Cardiovascular Diseases, Mayo Clinic, Rochester, Minnesota
    2. Center for Regenerative Medicine, Mayo Clinic, Rochester, Minnesota
    Search for more papers by this author
  • Xing Li,

    1. Department of Health Sciences Research, Division of Biomedical Statistics and Informatics, Mayo Clinic, Rochester, Minnesota
    2. Center for Regenerative Medicine, Mayo Clinic, Rochester, Minnesota
    Search for more papers by this author
  • Amber Mcdonald,

    1. Department of Laboratory Medicine and Pathology, Mayo Clinic, Rochester, Minnesota
    Search for more papers by this author
  • Devin Oglesbee,

    1. Department of Laboratory Medicine and Pathology, Mayo Clinic, Rochester, Minnesota
    Search for more papers by this author
  • Sybil C. Hrstka,

    1. Center for Regenerative Medicine, Mayo Clinic, Rochester, Minnesota
    2. Department of Medicine, Division of General Internal Medicine, Mayo Clinic, Rochester, Minnesota
    Search for more papers by this author
  • Carmen Perez-Terzic,

    1. Department of Medicine, Division of Cardiovascular Diseases, Mayo Clinic, Rochester, Minnesota
    2. Center for Regenerative Medicine, Mayo Clinic, Rochester, Minnesota
    3. Department of Physical Medicine and Rehabilitation, Mayo Clinic, Rochester, Minnesota
    Search for more papers by this author
  • Andre Terzic,

    1. Department of Medicine, Division of Cardiovascular Diseases, Mayo Clinic, Rochester, Minnesota
    2. Department of Molecular Pharmacology and Experimental Therapeutics, Mayo Clinic, Rochester, Minnesota
    3. Center for Regenerative Medicine, Mayo Clinic, Rochester, Minnesota
    4. Department of Medical Genetics, Mayo Clinic, Rochester, Minnesota, USA
    Search for more papers by this author
  • Timothy J. Nelson

    Corresponding author
    1. Department of Molecular Pharmacology and Experimental Therapeutics, Mayo Clinic, Rochester, Minnesota
    2. Center for Regenerative Medicine, Mayo Clinic, Rochester, Minnesota
    3. Department of Medicine, Division of General Internal Medicine, Mayo Clinic, Rochester, Minnesota
    4. Transplant Center, Mayo Clinic, Rochester, Minnesota
    • Department of Medicine and Transplant Center, Mayo Clinic, Stabile 5-42, 200 First St., S.W., Rochester, MN 55905, USA
    Search for more papers by this author
    • Telephone: 507-538-7515; Fax: 507-266-9936

  • Author contributions: A.M.-F., C.D.L.F., and E.P.-C.: conception and design, collection and/or assembly of data, data analysis and interpretation, manuscript writing; C.P.-T., S.H., and A.M.: collection and/or assembly of data; X.L.: data analysis and interpretation; D.O.: provision of study materials or patients; A.T. and T.J.N.: conception and design, data analysis and interpretation, manuscript writing, final approval of manuscript. C.D.L.F., A.M.-F., and E.P.-C. contributed equally to this article.

  • Disclosure of potential conflicts of interest is found at the end of this article.

  • §

    first published online in STEM CELLS EXPRESS April 3, 2013.


Mitochondrial diseases display pathological phenotypes according to the mixture of mutant versus wild-type mitochondrial DNA (mtDNA), known as heteroplasmy. We herein examined the impact of nuclear reprogramming and clonal isolation of induced pluripotent stem cells (iPSC) on mitochondrial heteroplasmy. Patient-derived dermal fibroblasts with a prototypical mitochondrial deficiency diagnosed as mitochondrial encephalomyopathy with lactic acidosis and stroke-like episodes (MELAS) demonstrated mitochondrial dysfunction with reduced oxidative reserve due to heteroplasmy at position G13513A in the ND5 subunit of complex I. Bioengineered iPSC clones acquired pluripotency with multilineage differentiation capacity and demonstrated reduction in mitochondrial density and oxygen consumption distinguishing them from the somatic source. Consistent with the cellular mosaicism of the original patient-derived fibroblasts, the MELAS-iPSC clones contained a similar range of mtDNA heteroplasmy of the disease-causing mutation with identical profiles in the remaining mtDNA. High-heteroplasmy iPSC clones were used to demonstrate that extended stem cell passaging was sufficient to purge mutant mtDNA, resulting in isogenic iPSC subclones with various degrees of disease-causing genotypes. On comparative differentiation of iPSC clones, improved cardiogenic yield was associated with iPSC clones containing lower heteroplasmy compared with isogenic clones with high heteroplasmy. Thus, mtDNA heteroplasmic segregation within patient-derived stem cell lines enables direct comparison of genotype/phenotype relationships in progenitor cells and lineage-restricted progeny, and indicates that cell fate decisions are regulated as a function of mtDNA mutation load. The novel nuclear reprogramming-based model system introduces a disease-in-a-dish tool to examine the impact of mutant genotypes for MELAS patients in bioengineered tissues and a cellular probe for molecular features of individual mitochondrial diseases. STEM Cells2013;31:1298–1308


Mitochondrial defects are associated with a spectrum of clinical manifestations from common diseases to rare genetic disorders [1–3]. Sharing metabolic similarities, maladaptive conditions are the consequence of the central role for mitochondria in energy production, reactive oxygen species generation, and regulation of life/death decisions [4]. These vital cellular functions are at risk due to lifelong accumulation of mutations in mitochondrial DNA (mtDNA). Due in part to the close proximity of oxidative stress, mtDNA suffers from a mutagenic rate that is three orders of magnitude higher than the nuclear genome [2]. Although only 13 genes of the central electron transport machinery are encoded by mtDNA, along with 22 tRNAs and 2rRNAs for mitochondrial protein synthesis, cellular dysfunction results when abnormal mitochondria accumulate and manifest as an unhealthy mixture of mutant mtDNA [1–3]. Disease causing heteroplasmy can be identified at several levels, including within the multiple copies of mtDNA per mitochondria, between healthy and diseased mitochondria within a cell, or among mosaic cellular subpopulations assembled within tissues. Without strategies to definitively alter the deterministic balance of heteroplasmy, mitochondrial medicine has been limited to palliative approaches.

The degree of mitochondrial heteroplasmy within individual stem cells is the product of a stochastic process identified as the genetic bottleneck or mitochondrial segregation [5–8]. In the embryo, this spontaneous process functions to reduce the hundreds of thousands of mitochondria per zygote to ∼200 within individual preimplantation blastomers and segregate distinctive combinations of mitochondria within individual primordial germ cells [6–7]. Additionally, segregation of mitochondrial heteroplasmy has been observed postnatally in mice varying with the specific metabolic requirements of individual tissues [9]. These innate processes collectively preserve mitochondrial diversity and genotype selection to ensure maximum variation and phenotypic selection that prevent widespread lethal accumulation of mitochondrial mutations [10, 11]. Furthermore, these dynamic processes offer an explanation for the wide spectrum of possible genotypes of individual offspring from a single maternal pool of oocytes. Notably, the deterministic balance of wild-type (Wt) versus mutant mitochondria within individual clonal populations is jointly dependent on the nature of the mutation, the interaction between nuclear and mitochondrial genomes, and the metabolic threshold tolerated by specific tissues. Therefore, strategies that ensure innate communication between mtDNA and nuclear DNA are necessary to accurately probe disease-causing pathways within the authentic cellular environment.

As mitochondrial mutational load directly affects disease outcome and is erratic between siblings or even tissues within a single individual, recent efforts have focused on embryo manipulations to predict disease severity and ultimately design therapeutic strategies that eliminate diseased mitochondria [12, 13]. An emerging strategy to advance mitochondrial medicine is to reduce disease burden by modulating the imbalance of mitochondrial heteroplasmy through stem cell manipulations. Here, we applied an embryo-independent strategy of nuclear reprogramming to bioengineer induced pluripotent stem cells (iPSC) from a patient with mitochondrial encephalomyopathy with lactic acidosis and stroke-like episodes (MELAS) that represents a prototypical metabolic derangement due to a high-heteroplasmy burden at position G13513A. Cardiovascular diseases such as cardiomyopathy and arrhythmias potentially affect MELAS patients. Individual MELAS patient-specific iPSC (M-iPSC) lines derived herein contained either the initial 50% heteroplasmy load or were devoid of mutated mtDNA at the disease-causing position in complex I corresponding to the mosaic parental source. Furthermore, iPSC clones burdened with high mutational mtDNA load were able to purge mutant mtDNA and establish a lower level of heteroplasmy with extended passaging. Tissues derived from M-iPSC lines maintained their respective genotypes throughout lineage specification and differentiation, enabling in vitro analysis of tissue-specific differentiation as a function of mtDNA heteroplasmy. Notably, this study establishes the ability to segregate distinct levels of heteroplasmy and thus disease-burdens of mtDNA within isogenic stem cell lines, providing a spectrum of patient-specific lineages to dissect etiologies of cellular dysfunction within the context of native nuclear and mitochondrial genomes.


Virus Production

pEX-QV (GAG/POL) expression vector and pMD-G packaging plasmid were used as described [14] with individual expression constructs psin-c-myc, psin-KIF4, psin-Oct4, and psin-Sox2. 293T/17 cells (ATCC#CRL-11268) were transfected using Optimem and Lipofectamine 2000 (Invitrogen, Grand Island, New York, www.invitrogen.com) and cultured overnight in Dulbecco's modified Eagle's medium (DMEM) with 10% fetal bovine serum (FBS). Virus-containing medium was harvested 72 hours after adding fresh DMEM with 10% FBS. Medium was filtered through a 0.45-μm Polyethersulfone filter before use or freezing.

iPSC Derivation and Cell Culture

Control human BJ fibroblasts (ATCC#CRL-2522) and MELAS fibroblasts that were collected according to institutional regulations (Mayo Clinic IRB 10-006969) were cultured in fibroblast medium with or without 200 μmol/l uridine, respectively. Cells (5 × 104) were plated for 6 hours before reprogramming. Medium was replaced by viral supernatant supplemented with polybrene to a final concentration of 0.6 μg/ml and incubated overnight. After 12 hours, cells were washed, and medium replaced by fibroblast medium and 72 hours later, cells were trypsinized onto inactive feeders in fibroblast medium. Medium was changed 48 hours later into human embryonic stem cell (hESC) medium (DMEM/F12 containing 20% knockout serum replacement and supplemented with 10 ng/ml βFGF and 200 μmol/l uridine). Colonies in a homogeneous single cell layer were apparent after 4 weeks and were manually picked and expanded. Established clones were transferred to matrigel-coated plates and medium switched into expansion medium (mTeSR supplemented with 200 μmol/l uridine). Differentiation was initiated by treating confluent cell cultures with collagenase IV (Invitrogen, Grand Island, New York, www.invitrogen.com) for 20 minutes, followed by gentle dislodgment from the plate surface and transfer of cell aggregates into suspension culture using ultra low attachement 6-well plates (Corning, New York, www.corning.com) in medium comprised of DMEM/F12 (Invitrogen, Grand Island, New York, www.invitrogen.com), 20% FBS (Characterized, Thermo Scientific; Hyclone, Hudson, New Hampshire, www.thermoscientific.com/hyclone), 1% glutamax (Invitrogen, Grand Island, New York, www.invitrogen.com), 1% nonessential amino acids (Cellgro, Corning, New York, www.cellgro.com), and 0.1% 2-mercaptoethanol (Sigma, St. Louis, MO, www.sigmaaldrich.com). On day 5, embryoid bodies were transferred to gelatin-coated tissue culture plates and processed for analysis up to 25 days in adherent cultures. Medium was refreshed every 2–3 days throughout the differentiation process.


Cells were fixed for 15 minutes with 3% paraformaldehyde, permeabilized with 0.1% triton-X and blocked using Superblock. The primary antibodies included monoclonal mouse anti-TRA-1-60 IgM (Millipore, Billerica, MA, www.millipore.com, 1:100), rat anti-SSEA3 (Stemgent, Cambridge, MA, www.stemgent.com, 1:100), and rabbit anti-Nanog (Stemgent, Cambridge, MA, www.stemgent.com, 1:100). Conjugated secondary antibodies (Invitrogen, Grand Island, New York, www.invitrogen.com) included Alexa fluor 568 anti-mouse IgM, Alexa fluor 488 anti-rat, and Alexa fluor 633 anti-rabbit, all used at 1:250 dilution. Nuclei were stained with 4′,6-diamidino-2-phenylindole (DAPI). Images were acquired with a Zeiss LSM 510 confocal microscope.

RNA Extraction and RT-PCR

RNA from fibroblasts, undifferentiated, and differentiated cells were extracted using Qiagen RNeasy mini kit. For real-time polymerase chain reaction (RT-PCR), cDNA was synthesized using reverse transcriptase supermix reagents (Invitrogen, Grand Island, New York, www.invitrogen.com). RT-PCR was performed using predesigned primers (Applied Biosystems, Carlsbad, CA, www.appliedbiosystems.com or IDT Integrated DNA technologies, Coralville, IA, www.idtdna.com) for UTF1 (Hs00747497_g1), GDF3 (Hs00220998_m1), DNMT3B (Hs01003405_m1), CDH1 (Hs00170423_m1), TDG1 (Hs02339499_g1), SALL4 (Hs00360675_m1), DPPA2 (Hs00414521_g1), DPPA5 (Hs00988349_g1), NANOG (Hs02387400_g1), TERT (Hs99999022_m1), TUBB3 (Hs00801390_s1*), SMA (Hs00426835_g1*), AFP (Hs00173490_m1*) and NKX2.5 (Hs00231763_m1*), GSC (Hs.PT.53a.1797922.g), HNF4A (Hs.PT.53a.24630422), KDR (Hs.PT.49a.15369695), MYH6 (Hs.PT.49a.20011601.g), TBX5 (Hs.PT.49a.2214557), and TNNI3 (Hs.PT.53a.18942460) all normalized to the GAPDH housekeeping gene (402869).

Mitochondrial Morphology

Mitochondrial content and morphology was quantified using transmission electron microscopy acquired with a JEOL 1,200 EXII electron microscope. Cells were fixed using 1% glutaraldehyde and 4% formaldehyde in 0.1 M phosphate-buffered saline (pH 7.2), stained with lead citrate and ultramicrotome sections were mounted for imaging. Between 25 and 140 mitochondria were analyzed for controls and between 62 and 190 mitochondria from MELAS samples.

Oxygen Consumption

Oxygen consumption rates were measured using a XF24 Extracelluar Flux Analyzer (Seahorse Biosciences, Billerica Massachusetts, www.seahorse.com). In brief, cells were plated into wells of a XF24 Cell Culture Microplate and maintained until 50–80% confluent. Before assay, plates were equilibrated in unbuffered XF assay medium supplemented with 25 mM glucose, 2 mM glutamax, 1 mM sodium pyruvate, 1× nonessential amino acids, and 1% FBS in the absence of CO2 for 1 hours. Mitochondrial processes were interrogated by serial addition of oligomycin (0.5 μg/ml), carbonyl cyanide 4-(trifluoromethoxy) phenylhydrazone (FCCP, 1 μM) and rotenone (0.5 μM) to establish basal respiration rates, maximal respiration (respiration in response to FCCP), and reserve capacity (maximal respiration – basal respiration). Each plotted value is the mean of at least 10 replicate wells, and are normalized to baseline oxygen consumption and total protein quantified using a Bradford assay (Biorad, Hercules, CA, www.bio-rad.com).

Short Tandem Repeat PCR

DNA was extracted from cell pellets with the QIAamp DNA Mini Kit (Qiagen, Germantown, MD, www.qiagen.com). Twelve short tandem repeat (STR) markers were selected for analysis and cell line comparison. PCR amplification of STR markers was performed in six multiplex reactions and each PCR reaction contained 0.8 mM ABI GeneAmp (a mix of dATP, dGTP, dCTP, and dTTP) (Applied Biosystems, Carlsbad, CA), 2.5 mM MgCl2, and 0.05 U/μl AmpliTaqGold DNA Polymerase (Applied Biosystems, Carlsbad, CA, www.appliedbiosystems.com). STR primer mixes were ordered from ABI PRISM Linkage Mapping Set except mycl (F- (VIC)-TGG CGA GAC TCC ATC AAA G; R- GTG TCT TCC TTT TAA GCT GCA ACA ATT TC). Primer mix concentrations were as follows: (a) 0.5 μM D7S484 + 1.0 μM D14S70, (b) 0.5 μM D13S158 + 2.0 μM D21S1252, (c) 0.3 μM mycl + 0.3 μM D10S197, (d) 1.0 μM D16S520 + 1.0 μM D2S2368, (e) 0.075 μM D15S1002 + 0.5 μM D6S441, and (f) 2.0 μM D17S250 + 1.6 μM D8S262. PCR cycling conditions consisted of a 15-second 94°C hotstart, 10 cycles of 15 seconds 94°C, 15 seconds 52°C and 30 seconds 72°C, 25 cycles of 15 seconds 89°C, 15 seconds 52°C, and 30 seconds 72°C, a hold of 10 minutes at 72°C and final cool to 4°C. Amplified products were pooled, diluted with water, and mixed with GS-400HD Size Standard (Applied Biosystems, Carlsbad, CA, www.appliedbiosystems.com) and Hi-Di deionized formamide. Products were run on the ABI 3130xl instrument using POP-7 polymer and a 36-cm capillary array (Applied Biosystems, Carlsbad, CA, www.appliedbiosystems.com).

Analysis of the G13513A Mutation

Restriction fragment length polymorphism (RFLP) analysis of the G13513A mutation in DNA from control and MELAS samples was performed according to a standardized protocol [15]. The amplified 170-bp fragment was digested with MboI, run on an 10% polyacrylamide gel and stained with SYBR Green I nucleic acid gel stain (Invitrogen, Grand Island, New York, www.invitrogen.com) or ethidium bromide, resulting in a 144-bp band corresponding to the Wt sequence and a 170-bp band corresponding to G13513A.

Next Generation Sequencing Analysis of Mitochondrial Genome

Mitochondrial genome was amplified in two fragments using the TaKaRa LA Taq PCR Kit (TAKARA Biotechnology, Otsu, Shiga, Japan, www.takara-bio.com) with the addition of DMSO and 0.6 μM of PCR primer (Mito 1-2F CATAGCACATTACAGTCAAATCCCTTCTCGTCCC, Mito 1-2R ATTGCTAGGGTGGCGCTTCCAATTAGGTGC, Mito-3F-TCATTTTTATTGCCACAACTAACCTCCTCGGACTC, Mito-3R-CGTGATGTCTTATTTAAGGGGAACGTGTGGGCTAT, Mito-1-2F ACATAGCACATTACAGTCAAATCCCTTCTCGTCCC, Mito-1-2R ATTGCTAGGGTGGCGCTTCCAATTAGGTGC). PCR cycling conditions consisted of a 2-minute 94°C hotstart, 30 cycles of 30 seconds 94°C, and 10 minutes 68°C, a hold of 15 minutes 68°C and final cool to 4°C. Amplicon was purified using the Agencourt AMPure XP kit (Beckman Coulter, Danvers, MA, www.beckmancoulter.com) and quantified using the Quant-iT PicoGreen dsDNA kit (Invitrogen, Paisley PA4 9RF, U.K.). Two purified amplicons were mixed in equimolar concentrations and mitochondrial genomes sequenced twice in separate lanes (Illumina Hi-seq 2000 using HiSeq Control Software version 1.4.8). Basecalls (RTA version were converted to fastq files (CASAVA 1.7), and aligned to the reference mitochondria genome (BWA 0.5.9-r16) [16]. The alignment produced SAM files that were converted to BAM with SamTools 0.1.7 (r510) [17] and sorted using the SortSam tool in Picard 1.47. Recalibration and local realignment was carried out using GATK 1.0.5777 [18, 19] along with indel calling and SNVMix 0.11.8-r3 [20]. Globally about 89.4% of the reads mapped to the reference genome and approximately 14 million reads per sample had a mapping quality greater than 20. Mitochondrial heteroplasmy rates in each sample were calculated as the percentage of the number of alternative reads to that of total reads (the sum of reference reads and alternative reads) on each nucleotide position. Results were filtered by heteroplasmy rates greater than 1% and a minimum of 1,000 reads at each nucleotide position. The range of total reads varied between 5,000 and 8,000.

mtDNA Copy Number

The mtDNA content of fibroblast, iPSCs, and differentiated cells was performed by multiplex RT-PCR [21], using the Eco Real-Time PCR System (Illumina, San Diego, CA, www.illumina.com).

Statistical Analysis

Data are presented as mean ± SEM. Student t test was used to evaluate two group comparisons, and ANOVA with a Bonferroni post hoc correction for three group comparisons. A value of p < .05 was considered significant.


MELAS-Derived Fibroblasts Demonstrate Metabolic Deficiencies

Dermal fibroblasts harboring the G13513A mtDNA mutation from a MELAS patient were expanded as originally isolated [22]. On electron micrographs (Fig. 1A), MELAS fibroblasts had smaller mitochondria than BJ fibroblasts (major axis: 1.09 ± 0.09 (n = 62) vs. 1.62 ± 0.22 μm (n = 26); minor axis: 0.25 ± 0.01 versus 0.55 ± 0.06 μm, p < .05; Fig. 1B). MELAS fibroblasts, compared with Wt fibroblasts, had impaired mitochondrial function including maximal respiration rate (9.4 ± 0.4 vs. 11.1 ± 0.9 pmol/minutes/μg protein, p < .05) and reserve capacity (−0.7 ± 0.2 vs. 4.6 ± 0.6 pmol/minutes/μg protein, p < .05; Fig. 1C). The mtDNA genotype of the MELAS fibroblasts was determined with RFLP analysis focused on the disease causing position (G13513A) to estimate mitochondrial heteroplasmy of the fibroblast population. MELAS fibroblasts demonstrated stable heteroplasmy with 53% Wt sequence (144 bp) and 47% G13513A sequence (170 bp) compared with a single Wt band in control fibroblasts. The ratio was not modified (in a 3-week metabolic challenge) with galactose (Fig. 1D). On clonal expansion of individual fibroblast subpopulations from the MELAS sample, ∼20% of individual clones derived from the original sample contained only Wt sequence, revealing a mosaic patient-derived fibroblasts population (Fig. 1D). The level of heteroplasmy was confirmed using next-generation sequencing of the entire mitochondrial genomes of control and MELAS fibroblasts (Fig. 1E). The 50% heteroplasmy of the disease-causing mutation at position G13513A was confirmed in MELAS sample compared with the Wt sequence of control fibroblasts. The characterized MELAS fibroblasts served as a defined somatic source to examine mitochondrial heteroplasmy on nuclear reprogramming.

Figure 1.

Mitochondrial encephalomyopathy with lactic acidosis and stroke-like episodes (MELAS)-derived dermal fibroblasts contain a mutation in G13513A and demonstrate reduced oxidative reserve capacity. (A): Mitochondrial morphology revealed by transmission electron micrographs demonstrated distinctive patterns. (B): MELAS fibroblasts demonstrated smaller mitochondria compared with BJ control fibroblasts in both the major and minor axis. (C): Mitochondrial function based on in vitro oxygen capacity in response to 0.5 μg/mL oligomyocin (a), 1 μM 4-(trifluoromethoxy) phenylhydrazone (FCCP,b), and 0.5 μM rotenone (c). (D): Genotype of the MELAS fibroblasts was determined with restriction fragment length polymorphism analysis with a focus on the disease causing position (G13513A), demonstrating 100% Wt sequence in control fibroblasts and heteroplasmy of 53% Wt sequence (144 bp) and 47% G13513A sequence (170bp) in MELAS fibroblasts that was unchanged despite 3 weeks of culturing in either glucose (Glu) or galactose (Gal) containing media. Clonal expansion of fibroblast (FB) subpopulations revealed that ∼20% of expandable clones from the original patient-derived MELAS sample-contained Wt sequences (red boxes) after 5–8 weeks of in vitro cell growth. (E): Alternative sequence (orange) reads at position G13513A was confirmed to be ∼50% heteroplasmy with next-generation sequence analysis within the MELAS sample compared with 100% Wt sequence (blue) of the control BJ fibroblasts. Values are mean ± SEM; * p < .05 versus BJ Wt. Abbreviations: FB, fibroblast; MELAS, mitochondrial encephalomyopathy with lactic acidosis and stroke-like episodes; Wt, wild-type.

Derivation of iPSC Clones from MELAS Patient Fibroblasts

MELAS fibroblasts were infected with lentiviral particles to express OCT4, SOX2, C-MYC and KLF4 stemness factors [23]. Embryonic stem cell (ESC)-like compact colonies appeared within 4 weeks of ectopic gene expression and nearly 100 clone-like clusters were picked and expanded in stem cell maintenance medium. Three M-iPSC lines (M-iPS1-3) were selected according to stem cell morphology, clonally expanded in a feeder-free system, and characterized by expression of pluripotency markers, including surface antigens SSEA-3 and Tra-1-60, as well as the nuclear transcription factor Nanog (Fig. 2A). Further characterization of patient-derived iPSC clones demonstrated upregulation of endogenous stemness genes according to quantitative RT-PCR (UTF1, GDF3, DNMT3B, CDH1, TDG1, SALL4, DPPA5, DPPA2, NANOG, and TERT) indicating acquisition of the pluripotent regulatory machinery absent from parental fibroblasts (Fig. 2B). Spontaneous in vitro differentiation utilizing a 10-day embryoid body protocol in high-serum medium converted M-iPS1-3 into differentiated tissue (M-Diff1-3). Analysis of gene expression in differentiated tissue demonstrated a significant increase in the lineage specific markers, neuronal TBB3 (ectoderm), smooth muscle SMA (mesoderm), AFP (endoderm), and NKX2.5 (cardiac), compared with pluripotent stem cells (cardiac, Fig. 2C). Examination of progeny indicated elongated and cristae-rich mitochondria acquired on differentiation, and the appearance of features of specialized tissues including cardiac sarcomeric ultrastructure with evidence of desmosomes (Fig. 2D). Each clone was profiled with genetic fingerprinting analysis comparing 12 STR markers to verify that M-iPSC lines had the same STR markers as the original MELAS patient-derived fibroblasts, which distinguished patient-derived samples from the control human fibroblast cell line (Fig. 2E). Patient-specific MELAS-derived fibroblasts can be thus successfully reprogrammed to pluripotency, producing an unlimited supply of pluripotent stem cells with the ability to produce tissue-specific lineages.

Figure 2.

Patient-specific induced pluripotent stem cells (iPSCs) derived from primary dermal fibroblasts. On nuclear reprogramming, mitochondrial encephalomyopathy with lactic acidosis and stroke-like episodes (MELAS)-derived fibroblasts acquired embryonic stem cell-like compact colonies within 4 weeks. (A): Three MELAS iPSC lines (M-iPS1-3) expressed cell surface antigens SSEA-3 and Tra-1-60 as well as the nuclear transcription factor Nanog. (B): Genetic fingerprinting established the identity of each clone originating from the MELAS cell population in contrast to the genetically distinctive profile of control BJ cell lines according to twelve short tandem repeat. Values are mean±SEM; *, p<.05 versus MELAS fibroblast, and p<.05 versus M-iPSC. (C): Patient-derived iPSC lines (denoted in blue bars) demonstrated upregulation of endogenous stemness genes according to quantitative real-time polymerase chain reaction compared with MELAS fibroblasts (denoted in red bars). (D): Spontaneous in vitro differentiation within embryoid bodies converted M-iPS1-3 into differentiated tissue (M-Diff1-3; denoted in orange bars) demonstrating a significant increase in the lineage specific markers for ectoderm, mesoderm, endoderm along with NKX2.5 gene expression for early cardiac tissue. (E): Examination of differentiated progeny using electron microscopy highlighted mature mitochondria and the appearance of specialized subcellular structures such as cardiac sarcomeres. Abbreviations: AFP, Alphafetoprotein; DAPI, 4′,6-diamidino-2-phenylindole; FB, fibroblast; iPSC, induced pluripotent stem cell; MELAS, mitochondrial encephalomyopathy with lactic acidosis and stroke-like episodes.

iPSC Mitochondrial Morphology and Function Independent of MELAS-Specific Mutations

Mitochondrial remodeling occurs during nuclear reprogramming [24–28]. The impact of a deleterious mtDNA mutation on this process has not been investigated in human disease conditions such as MELAS. As noted herein, nuclear reprogramming reverted mature mitochondria of MELAS fibroblasts (Fig. 1A) to spherical cristae poor structures in M-iPSCs (Fig. 3A, top). On differentiation, the mitochondria matured into elongated structures with mitochondrial appearance distinct from the pluripotent state and similar to somatic tissues (Fig. 3A, bottom). M-iPSC lines had similar major and minor axis lengths as control human (ESCs) and BJ-derived iPSC (Fig. 3B). M-iPSCs displayed low maximal respiration and reserve capacity, in agreement with the functional mitochondrial profile of control iPSC clones and human ESCs (Fig. 3C). iPSC clones demonstrated reduced mtDNA copy number compared with parental fibroblasts from both control and patient-derived samples (Fig. 3D). Furthermore, differentiation of iPSC clones provoked a dramatic increase of mtDNA copy number in all three M-iPSC lines (Fig. 3D), and a lengthening of the major and minor axis dimensions (Fig. 3B). Thus, nuclear reprogramming of iPSC clones results in diminished reliance on oxidative metabolism through reduction in mtDNA copy number and immature mitochondrial structure/function, suggesting that patient-derived iPSC clones in the pluripotent state are agnostic to disease-specific defects caused by complex I mutation in MELAS.

Figure 3.

Mitochondrial transformation in induced pluripotent stem cells (iPSCs) results in reduced oxygen dependence. (A): Mitochondrial acquire spherical cristae poor structures in all three M-iPSC lines (top) and elongated complex morphology in differentiated tissues (bottom). (B): Mitochondrial structures were indistinguishable between iPSC clones 1–3 with mitochondrial encephalomyopathy with lactic acidosis and stroke-like episodes (MELAS)-derived iPSC lines having similar major and minor axis lengths compared with wild-type (Wt) BJ iPSC (denoted in black and white bars) and embryonic stem cells (denoted by 95% confidence intervals highlighted in red lines). On differentiation (M-Diff), two of the three progeny acquired longer major axes. (C): Metabolic function according to in vitro cellular oxygen consumption for M-iPSCs profiles were similar to control BJ Wt iPSCs and hESC for all components of oxygen consumption. (D): M-iPSC lines demonstrated significant reduction in mitochondrial DNA (mtDNA) copy number according to polymerase chain reaction-based amplification and quantification as compared with parental MELAS fibroblasts (FB) denoted by 95% confidence interval highlighted by red line and demonstrated similar reduction compared with BJ-derived iPSCs. The mitochondrial copy number of all iPSC lines was equivalent to the low levels of hESC (denoted by 95% confidence interval highlighted by blue line). Differentiation of iPSCs provoked a dramatic reactivation of mtDNA copy number in progeny from all three MELAS-derived iPSC lines and BJ-derived control iPSCs to levels above the original MELAS fibroblasts. (E): Restriction fragment length polymorphism analysis of the G13513A complex I mutation indicated that M-iPS1 and 2 have similar levels of heteroplasmy as the starting material, while the mutation could not be detected in M-iPS3, with consistent results obtained after in vitro differentiation (M-Diff). Values are mean ± SEM; *, p < .05 versus MELAS FB; &, p < .05 versus M-iPSC; #, p < .05 versus BJ iPSC. Abbreviations: FB, fibroblast; hESC, human embryonic stem cell; iPSC, induced pluripotent stem cell; MELAS, mitochondrial encephalomyopathy with lactic acidosis and stroke-like episodes; mtDNA, mitochondrial DNA; Wt, wild-type.

Nuclear Reprogramming to Produce iPSC Clones Segregates Disease-Causing Mitochondrial Mutations

RFLP analysis of the G13513A mutation (Fig. 3E) indicated that nuclear reprogramming yielded stem cell lines with healthy and disease genotypes. Of the three qualified pluripotent clones, two MELAS iPSC clones demonstrated similar levels of heteroplasmy as the parental fibroblast population, and maintained equivalent levels following tissue-specific differentiation. In contrast, no detectable mutation at position 13,513 was observed in M-iPS3 pluripotent cells or differentiated progeny, indicating a complete segregation of mtDNA genotypes at this position. A plausible explanation for this phenomenon is that M-iPS3 clone is the result from nuclear reprograming of a Wt fibroblast present in the heterogeneous starting population. To confirm the absence of disease-causing mutation and determine if alternative mutations were acquired at other positions in the mtDNA, next-generation sequencing enabled complete coverage of mitochondrial heteroplasmic alterations with a total number of reads ranging from 16.2 million to 22 million corresponding to sequence depths of 98,000-132,000× coverage for cellular progeny (Fig. 4). Comprehensive analysis confirmed the presence of Wt and mutant sequence in two pluripotent clones (M-iPS1, 56% Wt sequence; M-iPS2, 54% Wt sequences), which was maintained in respective differentiated samples (M-Diff1, 55% Wt sequence; M-Diff2, 61% Wt sequence). Importantly, sequence analysis confirmed the absence of G13513A mutation in M-iPS3 and derived cell progeny following differentiation. In contrast to the distinctive differences at position 13513, the level of heteroplasmy across the three iPSC clones and differentiated progeny was indistinguishable at hypervariable regions within the remaining mitochondrial genome. Thus, nuclear reprogramming of clonal iPSC lines from clinical-grade MELAS-derived fibroblasts segregated disease causing mtDNA mutations within individual cell lines to reset a unique level of mutant mtDNA without disruption of nuclear or remaining mtDNA sequences.

Figure 4.

Nuclear reprogramming yields disease-free, isogenic clones from a mitochondrial encephalomyopathy with lactic acidosis and stroke-like episodes (MELAS) patient. Utilizing next-generation sequencing of the whole mitochondrial genomes of MELAS fibroblasts, M-induced pluripotent stem cells (iPSCs) (M-iPSC 1,2,3) and differentiated progeny (M-Diff 1,2,3), sequence alignments highlights the unique differences in patient-derived samples compared with control fibroblasts with mitochondrial genome labeled on the perimeter of the circle plot with colors corresponding to coding regions. The control fibroblasts (BJ Wt FB) demonstrated alternative reads (orange) in hypervariable regions but not in disease causing position of G13513A illustrated in the innermost ring of the circle plot. Sequence analysis confirmed the presence of Wt (reference sequence denoted in blue) and mutant sequence (alternative sequence denoted in orange) in patient-derived fibroblasts (MELAS FB) of nearly 50% at position 13,513 along with a unique profile of hypervariable regions illustrated in the second smallest ring of the circle plot. Two iPSC clones (M-iPS1, 56% Wt sequence; M-iPS2, 54% Wt sequences), along with their respective differentiated progeny (M-Diff1, 55% wt sequence; M-Diff2, 61% wt sequence) maintained the equivalent heteroplasmy of the parental fibroblasts. Sequence analysis confirmed the absence of G13513A mutation in M-iPS3 as well as differentiated progeny (M-Diff3) as noted in the all blue bar without orange bars at the disease-causing position without variation in the remaining hypervariable regions across the mitochondrial genome. Abbreviations: FB, fibroblast; iPSC, induced pluripotent stem cell; MELAS, mitochondrial encephalomyopathy with lactic acidosis and stroke-like episodes; Wt, wild-type.

Pluripotent Stem Cells Decrease Disease-Causing Mitochondrial Mutation Load

As fibroblasts derived from the MELAS patient contained mosaic cellular genotypes, we aimed to determine if the high mutant mtDNA burden could be reduced within individual clonal subpopulations of pluripotent stem cells with extended passaging. The M-iPS1 cell line was subcloned into 38 individual clones to confirm the homogeneous composition of cells with defined balance of mtDNA heteroplasmy, in contrast to the mosaic subpopulation of original fibroblasts (Supporting Information Fig. 1). The M-iPS2 cell line was maintained in standard feeder-free culture conditions continuously for 1-year and up to 44 passages. As expected with pluripotent stem cells, iPSC clones were capable of extended passages without jeopardizing cellular phenotype or inducing cellular senescence as is common in somatic fibroblasts. RFLP analysis of the G13513A mutation (Fig. 5A, 5B) indicated that the 50% heteroplasmy observed in passage three decreased to 37% and 7% at passage 21 and 44, respectively. Extended passaging did not alter the high levels of pluripotency markers (Fig. 5D, blue bars). On differentiation, pluripotency gene expression for DNMT3B (p < .001), SOX2 (p < .001), and NANOG (p < .001) was significantly decreased in low passaged iPSC clones and was similar for DNMT3B (p < .001), SOX2 (p < .001), and NANOG (p < .001) to that of high passage cells (Fig. 5D, orange bars). Passage 44 cell lines, with lower mtDNA heteroplasmy, demonstrated elevated cellular growth and multilayer differentiation compared with passage 21 cell lines during 25 days of differentiation in embryoid body culture conditions (Fig. 5C). Gene expression of the differentiated tissues demonstrated variability in early stage gene expression (HNF4A, GSC, KDR) with elevated expression of cardiac markers (TBX5, p = .10; TNNI3, p = .06; MYH6, p < .05) in late passage iPSC clones with lower heteroplasmy (Fig. 5E). As differentiated progeny were derived from subclones of iPSC clones containing high and low levels of mtDNA heteroplasmy, these studies decrease the inherent variability between iPSC clones and suggest that differences in heteroplasmy may contribute to cardiogenic cell fate decisions in bioengineered patient-specific pluripotent stem cells.

Figure 5.

Reduced mitochondrial DNA (mtDNA) heteroplasmy improves cardiogenesis in patient-specific induced pluripotent stem cells (iPSCs). (A): Analysis of heteroplasmy after 1-year in culture of the M-iPS2 clone compared with mitochondrial encephalomyopathy with lactic acidosis and stroke-like episodes fibroblasts until senescence. (B): Heteroplasmy in the mtDNA remains unchanged following 25 days of in vitro differentiation compared with the heteroplasmy of iPSCs. (C): Light microscopy of differentiated progeny highlighted increased 3-dimensional differentiated areas within M-Diff2 p44 compared with M-Diff2 p21. (D): Low and high passage patient-derived iPSC line (denoted in shades of dark and light blue bars respectively) demonstrated downregulation of endogenous stemness genes according to quantitative real-time polymerase chain reaction (RT-PCR) in differentiated progeny compared with undifferentiated iPSCs (denoted in orange bars). (E): RT-PCR demonstrates an increase in TBX5, TNNI3 and significant increase in MYH6 gene expression on side-by-side differentiation analysis. Values are mean ± SEM; *, p < .05 versus respective M-iPS2; #, p < .05 versus M-Diff2 p21. Abbreviations: FB, fibroblast; iPSC, induced pluripotent stem cell; MELAS, mitochondrial encephalomyopathy with lactic acidosis and stroke-like episodes.


Understanding the pathobiology of mitochondrial disease affecting the semiautonomous organelle has proven difficult despite the limited number of genes encoded by the mtDNA genome [1–3]. The stochastic behavior of mitochondrial heteroplasmy provides a challenge as the deterministic balance of healthy and diseased mitochondria spans a wide range within individual organelles, cells, and even tissues [3]. Herein, genetically indistinguishable iPSC clones at the nuclear level were produced from a MELAS patient-derived sample characterized with ∼50% heteroplasmy at the mitochondrial level within the population of cells that was composed of ∼20% of individual fibroblasts containing only Wt mtDNA. The iPSC clones achieving pluripotency criteria were subjected to comprehensive mtDNA sequence analysis to quantify the degree of heteroplasmy following nuclear reprogramming. Of the selected clonally derived M-iPSC lines, two maintained heteroplasmy at ∼50% and one contained no detectable mutant mtDNA. Subcloning the iPSC clones with ∼50% heteroplasmy demonstrated the consistency of mtDNA and lack of mosaicism at the cellular level within these clones. Extended culturing of pluripotent stem cells with high mtDNA heteroplasmy for up to 44 passages progressively purged mutant mtDNA, in contrast to minimal changes observed in clonal fibroblast lines following 20 passages. Thus, nuclear reprogramming yielded isogenic clones with distinct mitochondrial genotypes and pluripotent stem cells capable of continuously decreasing the burden of mutant mtDNA. This bioengineering technology offers a strategy for in vitro segregation of heteroplasmy and characterization of patient-specific derivation of mitochondrial defective cellular progeny.

Clinically, patients affected by MELAS have a spectrum of phenotypes and present with disease features even at modest levels of heteroplasmy compared with other mitochondrial diseases [29]. The mutation at position G13513A of the ND5 subunit of respiratory chain complex I is reported to cause metabolic disturbance leading to MELAS, as well as Leigh syndrome [30] and Wolff–Parkinson–White syndrome associated with cardiomyopathy [31]. Using real-time metabolic profiling, we demonstrated significant deficiencies in oxidative capacity within patient-derived dermal fibroblasts compared with Wt cell lines, consistent with the clinical features of the disease. RFLP, the gold-standard analysis of heteroplasmy in cell culture, was able to discriminate between the control and MELAS-derived fibroblasts and validate the presence of both Wt and mutated mtDNA. Furthermore, this analysis confirmed that control fibroblasts were indeed Wt, and the MELAS starting material expressed approximately 50% heteroplasmy as originally reported [22]. In vitro examination of this patient-derived MELAS cell line provided a well-characterized somatic source to delineate whether heteroplasmy could be segregated during nuclear reprogramming. Traditionally, cybrids have proven to be a powerful approach to analyze mtDNA within a defined and controlled nuclear genome to determine if mtDNA mutations are sufficient to cause cellular defects and model mitochondrial diseases [32]. Alternatively, transmitochondrial mouse pluripotent stem cells have been generated by fusion of previously pharmacologically treated mouse ESCs, to prevent transmission of endogenous mtDNA, with cytoplasts carrying mtDNA mutations [33]. These specialized systems have enabled mechanistic insight into mtDNA mutations with an artificial nuclear context. Herein, patient-derived somatic cells contained the disease-causing mtDNA heteroplasmy within the native context of the nuclear genome such that amplifying or dampening combinations of individual mutations or genetic variants can be individually examined. Somatic cells reprogrammed into iPSCs containing distinctive mtDNA genotypes could provide an innovative bioengineering approach, customized for individual patients, without a bias according to non-self genotypes or xenograft model systems.

Consistent with a required metabotype conversion from oxidative metabolism to glycolysis during nuclear reprogramming [24, 27, 34–36], human iPSC clones respired at a lower maximal rate than the somatic starting material. Notably, iPSC clones, independent of a healthy or diseased mtDNA mixture, demonstrated an equivalent maximal oxygen consumption and reserve capacity. Furthermore, regression of elongated somatic mitochondria produced immature structures in M-iPSC and control iPSC clones resembling mitochondria in hESCs. Thus, nuclear reprogramming is sufficient to reduce mitochondrial density and oxygen dependency consistently across iPSC clones [28], independent of original somatic cell mitochondrial function. Remarkably, the metabolic profiles of MELAS-derived iPSC clones were indistinguishable from the control iPSC clones in terms of function and morphology despite dramatic physiological consequences of mtDNA mutations in the somatic tissues. These data confirm the reduced oxidative dependence of iPSC clones [24, 27] and offers new perspective as to how the metabolic requirements of pluripotent stem cells may be agnostic to disease-defining mtDNA mutational load at this stage of development. Nuclear reprogramming may thus provide an approach to obtain a distribution of heteroplasmy in primordial stem cells and then enable dissection of the functional impact of mtDNA on tissue-specific differentiation.

The genetic bottleneck during embryogenesis may offer a mechanism that could be provoked at the time of reprogramming or during subsequent asymmetrical pluripotent stem cell divisions to generate a range of mtDNA defining healthly and diseased iPSC clones. The observed segregation of mtDNA herein may resemble the processes active in natural blastomeres that leads to embryonic dispersion of mtDNA and contributes to genetic diversity [5–11]. Clonal expansion and differentiation of daughter cells, both in embryos and herein for iPSC-derived progeny, can maintain the preset balance of heteroplasmy to generate tissues for mitochondrial disease modeling. Whether the distribution of mtDNA heteroplasmy is regulated or stochastic in pluripotent stem cells, mitochondrial pathways are dependent on the symbiotic relationship between mtDNA and nuclear DNA. Producing independent subclones that have distinctive heteroplasmic mtDNA patterns in the context of native nuclear DNA enables deconvolution of authentic disease-specific processes in mitochondrial diseases and offers potential strategies to reduce disease-causing thresholds of mutant mtDNA for production of healthy autologous cell types for regenerative applications.

Herein, an isolated M-iPSC clone demonstrated a single Wt RFLP DNA fragment of 144 bp, indicating the absence of the heteroplasmic genotype within this iPSC clone and subsequent differentiated progeny. According to the sensitivity of the RFLP assay, we were unable to document mutant mtDNA sequence in the Wt homoplasmic M-iPSC sample at the pluripotent state or within subsequent differentiated progeny. Confirmed with next-generation sequencing, the G13513A alternative read was undetectable in the M-iPSC line and/or differentiated progeny. Of interest, other polymorphisms within the mitochondrial sequence matched both the starting material and the other derived M-iPSC lines. Heteroplasmic changes in common variants during nuclear reprogramming within healthy pluripotent stem cells have been documented, establishing the plasticity of the mitochondrial genome during reprogramming [28]. Detected within early passage M-iPSCs and maintained throughout differentiated tissues, we demonstrate variable degrees of disease-causing heteroplasmy following nuclear reprogramming and expansion of pluripotent stem cells in vitro that remained stable across tissue-specific differentiation. This feature, resembling natural embryogenesis, enables genotype/phenotype designs to explore the effect of heteroplasmy within mitochondrial patient-specific iPSCs and clonogenic progeny.


Mitochondrial diseases regulated by mtDNA heteroplasmy contribute to significant unmet needs in clinical practice. Nuclear reprogramming is achievable across distinct mitochondrial diseases [37] and yet it remains to be determined how efficient the process will be to isolate isogenic iPSC clones that contain a spectrum of healthy and diseased heteroplasmy. The potential impact of this technology on mitochondrial diseases establishes new possibilities for regenerative mitochondrial biology. With the present demonstration that patient-specific iPSC clones from a patient with MELAS are able to segregate the dysfunctional heteroplasmy, the field of mitochondrial medicine now has an innovative bioengineering tool to comparatively probe the mechanisms of individual disease in the context of isogenic cell lines with or without the disease-causing mtDNA mutation. This cell-based platform unlocks the distinctive advantages of patient-specific iPSC technology for genotype/phenotype analysis allowing discovery of patient-specific drug targets, mechanisms of mtDNA mutations in the native context of nuclear DNA, and ultimately offers the potential of a curative cell-based therapeutic strategy for regenerative mitochondrial medicine.


We thank Bruce W. Eckloff, Matthew A. Bockol, Sumit Middha, and Jared M Evans in Mayo Clinic Sequencing Core for the help with mitochondrial genome sequencing and alignment analysis and Israel Perez Medina for his technical assistance. C.D.L.F. is the recipient of a Canadian Institutes of Health Research Fellowship, Mayo Clinic Center for Regenerative Medicine Fellowship, and a Marriott Individualized Medicine Career Development Award; A.M.-F. is the recipient of an American Heart Association Midwest Affiliate Postdoctoral Fellowship (AHA 11POST7600085); and E.P.-C. is the recipient of a Spanish Ministry of Education MEC Fellowship. This work was supported by the Todd and Karen Wanek Family Program for Hypoplastic Left Heart Syndrome and NIH New Innovator Award OD007015-01 (to T.J.N.) and the Marriott Heart Disease Research Program and the Leducq Foundation (to A.T.).


The authors indicate no potential conflicts of interest.