ARTICLE IN PRESS

Molecular Genetics and Metabolism xxx (2010) xxxxxx

Contents lists available at ScienceDirect

Molecular Genetics and Metabolism

journal homepage: www.elsevier.com/locate/ymgme

Minireview

Current molecular diagnostic algorithm for mitochondrial disorders

Lee-Jun C. Wong *, Fernando Scaglia, Brett H. Graham, William J. Craigen
Department of Molecular and Human Genetics, Baylor College of Medicine, Houston, TX, USA

a r t i c l e

i n f o

Article history:
Received 17 January 2010
Received in revised form 25 February 2010
Accepted 26 February 2010
Available online xxxx
Keywords:
Mitochondrial disorder
mtDNA mutation
mtDNA depletion
mtDNA deletion

a b s t r a c t
Mitochondrial respiratory chain disorders (RCD) are a group of genetically and clinically heterogeneous
diseases, due in part to the biochemical complexity of mitochondrial respiration and the fact that two
genomes, one mitochondrial and one nuclear, encode the components of the respiratory chain. Because
of the large number of genes involved, attempts to classify mitochondrial RCD incorporate clinical, biochemical, and histological criteria, in addition to DNA-based molecular diagnostic testing. While molecular testing is widely viewed as denitive, conrmation of the diagnosis by molecular methods often
remains a challenge because of the large number of genes, the two genome complexity and the varying
proportions of pathogenic mitochondrial DNA (mtDNA) molecules in a patient, a concept termed heteroplasmy. The selection of genes to be analyzed depends on the family history and clinical, biochemical,
histopathological, and imaging results, as well as the availability of different tissues for analysis. Screening of common point mutations and large deletions in mtDNA is typically the rst step. In cases where
tissue-specic, recognizable clinical syndromes or characteristic RC complex deciencies and histochemical abnormalities are observed, direct sequencing of the specic causative nuclear gene(s) can be performed on white blood cell DNA. Measurement of mtDNA content in affected tissues such as muscle
and liver allows screening for mtDNA depletion syndromes. The ever-expanding list of known diseasecausing genes will undoubtedly improve diagnostic accuracy and genetic counseling.
2010 Published by Elsevier Inc.

Background
Diagnosing primary mitochondrial respiratory chain disorders
remains a major challenge to the clinician. Many primary mitochondrial diseases involve multiple organ systems, in part reecting the
dependence on energy derived from oxidative phosphorylation in
a wide variety of tissues, and this feature may be the initial indication
as to the correct diagnosis. However, the varied clinical presentations, frequent dependence on invasive testing procedures, and
enormous genetic heterogeneity all contribute to the diagnostic
uncertainty that lingers in many cases. Because of this clinical heterogeneity a number of medical specialists may be the rst to encounter these patients, including cardiologists, gastroenterologists,
neurologists and ophthalmologists, yet recognition of this group of
disorders remains uneven. There are classic clinical syndromes
with stereotypic features such as Leigh syndrome (subacute necrotizing encephalomyelopathy), MELAS (mitochondrial myopathy,
encephalopathy, lactic acidosis and stroke-like episodes), and Alpers
disease (epilepsy and liver failure). However, many patients exhibit
only non-specic features of developmental delay or regression, further hindering accurate diagnoses. Diagnostic criteria have been
proposed that attempt to take into account clinical manifestations,
* Corresponding author. Address: Department of Molecular and Human Genetics,
Baylor College of Medicine, One Baylor Plaza, NAB 2015, Houston, TX 77030, USA.
Fax: +1 713 798 8937.
E-mail address: ljwong@bcm.edu (Lee-Jun C. Wong).

enzymatic and physiologic analyses, tissue histochemical results,

levels of biochemical analytes, and DNA analysis to more reliably
characterize patients [1,2].
RCD can be caused by mutations in either the mitochondrial or
nuclear genomes. Adding to the complexity of diagnostic testing,
the mitochondrial genome is maintained in large part by proteins
encoded in the nucleus, giving rise to either primary mitochondrial
DNA (mtDNA) mutations, which may exhibit a matrilineal pattern
of inheritance, or Mendelian traits that manifest as mtDNA alterations and thus mimic primary mtDNA alterations. Furthermore, the
penetrance or severity of a primary mtDNA mutation is potentially
modied by the nuclear genome background in which it coexists or
by environmental factors. In the case of Mendelian disorders, autosomal recessive, autosomal dominant and X-linked patterns have
all been observed.
Mitochondrial disorders are likely more common than has been
historically believed, yet attempts to establish the incidence of diseases of the respiratory chain have been limited by disease heterogeneity and the lack of simple, denitive biomarkers. Studies in
adults and/or children suggest a prevalence approaching 1 in
7600 [3]. However, in a thoughtful meta-analysis, prevalence of
common mtDNA mutation was estimated to be at least 1 in 5000
[4]. Moreover, a recent study of more common mitochondrial
DNA mutations in newborns and their mothers suggest the possibility of a much higher number, in part due to an apparently high
rate of de novo mutations [5]. It remains unknown to what degree

1096-7192/$ - see front matter 2010 Published by Elsevier Inc.

doi:10.1016/j.ymgme.2010.02.024

Please cite this article in press as: L.-J.C. Wong et al., Current molecular diagnostic algorithm for mitochondrial disorders, Mol. Genet. Metab. (2010),
doi:10.1016/j.ymgme.2010.02.024

ARTICLE IN PRESS
2

Lee-Jun C. Wong et al. / Molecular Genetics and Metabolism xxx (2010) xxxxxx

these inherited and de novo mtDNA mutations detected in otherwise healthy newborns will lead to disease over the course of a
lifetime. While adults that develop signs and symptoms of mitochondrial disease are often found to harbor primary mtDNA mutations, the vast majority of pediatric patients likely represent
Mendelian inheritance of nuclear gene mutations. Since current
estimates of the number of nuclear genes that contribute to the
biogenesis and appropriate function of mitochondria approaches
1300 [6], there is a signicant need for denitive molecular testing,
given that currently clinical analysis is available for less than 50
nuclear genes known to cause RCD. Recent reviews [7,8] have described in detail the clinical presentations and overall clinical and
biochemical evaluation of patients with suspected mitochondrial
disease, which are beyond the scope of this review. This review
provides guidelines for the use of currently available DNA-based
testing in order to aid in establishing a denitive diagnosis of a
mitochondrial respiratory chain disease.

Rationale/evidence
Obtaining a denitive and specic molecular diagnosis for a patient with clinically suspected mitochondrial disease is important
for multiple reasons: (1) to allow more specic medical management and health surveillance of the patient, (2) to allow diagnostic
testing of any at risk relatives, and (3) to allow more specic genetic counseling and recurrence risk estimation for the family. While
the diagnosis of mitochondrial disease requires an integrative approach, the identication of a molecular pathogenic defect, when
possible, can overcome the ambiguities often seen with biochemical analytes, histological, and enzymological evaluations.
There is no reliable screening or diagnostic biomarker that
would be reasonably sensitive or specic to denitively establish
the diagnosis of mitochondrial disease alone. The concentration
of serum lactic acid is a commonly used screening test. Nevertheless, lactic acidosis should not be considered a prerequisite for
diagnosis because lactic acid is not a reliably sensitive or specic
marker of mitochondrial dysfunction [911]. For example, the serum lactate is often normal or mildly increased at best in some
cases of Leigh syndrome, neuropathy, ataxia, retinitis pigmentosa
(NARP) and POLG1 (DNA polymerase c catalytic subunit) deciency [12]. When properly performed, elevated lactate-to-pyruvate ratios provide a useful tool in helping to establish the
diagnosis of RCD [13]. Other biomarkers such as elevated serum
alanine level, increased tricarboxylic acid (TCA) cycle intermediates or lactic acid measured by urine organic acid analysis, and
plasma acylcarnitine elevations indicative of disrupted fatty acid
oxidation all suggest the presence of mitochondrial dysfunction,
although normal results should not preclude any work-up if there
is a high suspicion [7].
The morphological analysis of skeletal muscle can be helpful for
diagnosis but often is complicated by ambiguous results. Although
the presence of mitochondrial proliferation in the form of raggedred bers (RRF) is suggestive of a mitochondrial RCD, they are rarely
seen in early childhood, when subsarcolemmal accumulation of
mitochondria may be mild or absent. Additionally, abnormal accumulation of mitochondria may be absent in patients with proven
mitochondrial disease such as NARP. In general, normal muscle histology does not rule out the diagnosis of mitochondrial dysfunction
because patients with denite dysfunction of mitochondrial metabolism may lack ultrastructural abnormalities of the mitochondria
and conversely, mitochondrial proliferation or even cytochrome c
oxidase deciency by enzyme histochemistry may be seen in nonRC defects and in normal individuals over 60 years of age [14].
The diagnosis of RCD relies in part on enzymatic investigations
from tissue samples, and may lead to results such as partial de-

ciencies or other ambiguities which may be difcult to interpret.

Different methods for assaying mitochondrial enzymes are used
by reputable laboratories, and a systematic program to share samples and standardize methodologies has not yet been implemented. To add further complexity to this issue, biochemical
assays of the respiratory chain have low inter-laboratory reproducibility [15]. In addition, some mitochondrial groups argue in favor
of using fresh tissue for enzymatic analysis [7,8]. However, this approach may not be practical due to logistical and geographical constraints. Another confounding factor is that decreased activities of
the mitochondrial respiratory chain may be secondary to non-primary respiratory chain disorders such as pyruvate dehydrogenase
complex deciency [16]. The converse is also true; a RC defect may
trigger mitochondrial proliferation resulting in a compensatory increase in mitochondrial activities through mitochondrial accumulation, and this increase may in turn mask a RC deciency [17].
Assessing the appropriate tissue can also be challenging since certain RCD are quite tissue-specic in the abnormal RC activity. For
example, in cases associated with mutations in the SCO2 gene
(cytochrome c oxidase assembly gene) deciency, the activity of
complex IV is markedly reduced in heart and skeletal muscle,
whereas liver and broblasts have only mild COX deciency [18].
While a skin biopsy is less invasive, it is not uncommon for patients
with oxidative phosphorylation defects in skeletal muscle to have
normal RC enzyme activities in broblasts [19,20]. Furthermore,
skeletal muscle biopsy in a young infant may be limited by a lack
of normal age-matched controls.
Since molecular diagnostic techniques are frequently unequivocal and thus generally unambiguous, there is a perception of greater reliability in comparison to biochemical assays of the
mitochondrial RC. Given the similar methodologies employed by
diagnostic laboratories providing DNA-based testing and the reduced likelihood of sample quality or tissue-specic differences
inuencing the results, there is a higher emphasis placed on DNA
results in diagnostic algorithms. Screening of common point mutations and large deletions in mitochondrial DNA is usually the rst
step taken when evaluating a patient for a suspected mitochondrial
cytopathy. However, some caveats regarding molecular diagnoses
for mitochondrial disease should be considered.
Analysis of mtDNA represents a particular diagnostic challenge
in that false negative results may be due to low-level heteroplasmy
in peripheral blood samples. Moreover, the identication of some
mtDNA alterations, such as multiple mtDNA deletions or mtDNA
depletion, can indicate a Mendelian disorder because the maintenance of mtDNA integrity is highly dependent on a variety of nuclear-encoded proteins. In addition, nuclear DNA mutations likely
account for the majority of severe mitochondrial diseases presenting in infants and children. Due to the small number of known nuclear disease-causing loci, there are a limited number of denitive
DNA-based tests that can be offered in the majority of pediatric
cases of RCD. Furthermore, the interpretation of sequence variants
of no established clinical signicance remains a major challenge,
both in nuclear genes and the mtDNA genome. Functional studies
may help determine the pathogenicity of a specic sequence variant, but such tests are generally only available on a research basis.
While family studies, computational structural approaches, evolutionary comparisons, and population databases may guide the
interpretation of sequence variants, unresolvable ambiguity is frequently encountered in practice.

Current practice
Given the biological complexity of mitochondria, a molecular
diagnostic plan for individual patients must be informed by the
clinical presentation, family history, and biochemical testing

Please cite this article in press as: L.-J.C. Wong et al., Current molecular diagnostic algorithm for mitochondrial disorders, Mol. Genet. Metab. (2010),
doi:10.1016/j.ymgme.2010.02.024

ARTICLE IN PRESS
3

Lee-Jun C. Wong et al. / Molecular Genetics and Metabolism xxx (2010) xxxxxx

(Fig. 1). Based on the authors collective personal experiences

and the relevant literature, 8095% of patients with clinically
suspected primary mitochondrial disease do not have a detect-

able pathogenic mtDNA mutation. These cases are therefore presumed to harbor mutation(s) in a nuclear-encoded mitochondrial
gene [11,2125].

PATIENT WITH SUSPICION OF MITOCHONDRIAL RC DISORDERS

CLINICAL, METABOLIC, AND IMAGING EVALUATION

BLOOD
Common mtDNA mutations

Positive

Family Member Studies

Negative1

Recognizable syndrome?

Leigh

Encephalocardiomyopathy

MNGIE

EncephaloHepatopathy

Encephalonephropathy

PEO

Myopathy

PDHC genes
Complex I Subunits
Complex IAssembly Genes
COX subunits
COX Assembly Genes(e.g., SURF1)
mtDNA Genome Seq

SCO2
TAZ
mtDNA Genome Seq

TYMP

POLG1
DGUOK
MPV17
C10orf2
SCO1
BCS1L

BCS1L
COX 10
PDSS2
COQ2
RRM2B

POLG1
POLG2
SLC25A4
C10orf2
OPA1

TK2
POLG
SUCLA2
mtDNA

IF RESULTS

MUSCLE, LIVER, OR
SKIN BIOPSY2

Positive

Negative

Family Member Studies

Electron Transport Chain (ETC)

Common mtDNA mutations

Positive

Family Member Studies

Family Member Studies

Negative

mtDNA quantification
qPCR, oligo aCGH

High/
normal

Positive
Complex I, III, and IV
Deficiencies

depletion

Depletion Genes3
POLG1
DGUOK
MPV17
C10orf2
TYMP
TK2
SUCLA2
SUCLG1
RRM2B

Isolated
Complex
Deficiency

Proliferation
Positive

mtDNA Genome Seq.

Negative

Negative

Positive

mtDNA Genome Seq.

mtDNA Deletion
PDH Complex

Sequence BOTH:
Complex Subunits
& Assembly genes

Negative

Molecular Dx Undetermined

Family Member Studies

mtDNA Genome Seq.

mtDNA deletion

Positive

Negative

Molecular Dx Undetermined

Fig. 1. Testing algorithm for molecular diagnosis of patients with suspected mitochondrial disease. A generalized algorithm for molecular testing based on clinical and
biochemical information is presented. This algorithm is designed to be a general guide and is not intended to encompass every potential clinical scenario nor all possible
genetic etiologies. For the genes listed, HUGO gene nomenclature is used. MNGIE, Mitochondrial NeuroGastroIntestinal Encephalopathy syndrome; PEO, Progressive External
Ophthalmoplegia syndrome; Oligo aCGH, oligonucleotide array comparative genomic hybridization. 1If a common mtDNA mutation is not detected in the initial screen, then a
tissue biopsy for additional molecular and biochemical studies is required. If a tissue biopsy is not available, then screening appropriate nuclear genes based on a recognized
clinical pattern can be pursued. 2A muscle or liver biopsy can be used for both molecular (tissue-specic mtDNA mutations/heteroplasmy and/or mtDNA quantication) and
biochemical (ETC activity measurements) studies. A skin biopsy for broblasts may be appropriate for biochemical studies, especially in infants too young for muscle biopsy,
but is not useful for detecting mtDNA depletion or multiple deletions of mtDNA. 3POLG, DGUOK, MPV17, and C10orf2 mutations have been observed with mtDNA depletion
and hepatoencephalopathy. TYMP mutations are associated with MNGIE. TK2, SUCLA2, SUCLG1, and RRM2B have been observed with mtDNA depletion and
encephalomyopathy. Myopathy with elevated creatine kinase is a frequent feature of TK2 deciency.

Please cite this article in press as: L.-J.C. Wong et al., Current molecular diagnostic algorithm for mitochondrial disorders, Mol. Genet. Metab. (2010),
doi:10.1016/j.ymgme.2010.02.024

ARTICLE IN PRESS
4

Lee-Jun C. Wong et al. / Molecular Genetics and Metabolism xxx (2010) xxxxxx

Currently, a specic molecular diagnosis is obtained for only a

fraction of these patients. However, an increased proportion of successful molecular diagnoses have been observed for certain clinical
subsets of patients with suspected mitochondrial disease. For
example, mutations in POLG1 are responsible for a heterogeneous
group of at least six major clinical phenotypes: (1) childhood Myocerebrohepatopathy Spectrum disorders (MCHS), (2) Alpers syndrome, (3) Ataxia Neuropathy Spectrum (ANS) disorders
including Spinocerebellar Ataxia with Epilepsy (SCAE) and Mitochondrial Recessive Ataxia Syndrome without Ophthalmoplegia
(MIRAS), (4) Myoclonus, Epilepsy, Myopathy, and Sensory Ataxia
(MEMSA), (5) autosomal recessive Progressive External Ophthalmoplegia (arPEO) including Sensory Ataxic Neuropathy, Dysarthria
and Ophthalmoparesis (SANDO), and (6) autosomal dominant Progressive External Ophthalmoplegia (adPEO) [27]. In addition, they
have been detected in patients with other clinical phenotypes
associated with classical mitochondrial syndromes such as Leigh
syndrome, mitochondrial encephalomyopathy lactic acidosis and
stroke-like episodes (MELAS) syndrome. In a recent study, 15 patients with PEO were screened for mutations in POLG1, TWINKLE,
and ANT1. Mutations were identied in six patients [26]. In another
series of patients with the above spectrum of clinical features, 61 of
350 individuals had identiable mutation(s) in POLG1, making it
one of the most frequent nuclear-encoded mitochondrial disease
genes to date [27]. Given the wide phenotypic spectrum and relatively high mutation detection rate, sequencing POLG1 in patients
who present with encephalopathy, seizures, hepatopathy or PEO
should be viewed as a screening test, particularly if mtDNA screening is negative and if there are other diagnostic clues such as multiple deletions or depletion of mtDNA on skeletal muscle and the
presence of RRF in adults (Fig. 1). In a second example, the detection of SURF1 mutations in patients with Leigh syndrome and isolated cytochrome c oxidase deciency is high [28], thus, if the
clinical features are consistent with Leigh syndrome and the family
history suggests autosomal recessive inheritance, then sequencing
of SURF1 would most likely identify the causative mutation, as
shown in Fig. 1.
Study of 180 patients with cytochrome c oxidase deciency, 75
were found to have mutations in either mtDNA or a nuclear gene
[29]. In addition to cytochrome c oxidase deciency (complex
IV), abnormalities in complex I, complex II, pyruvate dehydrogenase, or various mtDNA mutations are all possible etiologies for
Leigh syndrome. In general, if the disease clearly exhibits matrilineal inheritance and the screening of common mutations is negative, sequence analysis of the whole mitochondrial genome
would be recommended.
Screening by quantitative PCR for mtDNA depletion in muscle
or liver samples may also be rewarding. In the context of mtDNA
depletion, infantile liver dysfunction, and/or encephalomyopathy
recent reports have clearly illustrated the success of identifying
mutations in nuclear genes required for mtDNA integrity, including
POLG1, MPV17, DGUOK, RRM2B, C10ORF2, SUCLA2, and SUCLG1
[27,3037]. In particular, if the encephalomyopathy is associated
with methylmalonic aciduria, testing for mutations in the SUCLA2
and SUCLG1 genes should be considered [31]. If mtDNA depletion
in skeletal muscle is associated with a myopathy that starts in infancy or childhood and follows a slowly progressive clinical course,
testing for mutations in the TK2 gene is warranted [38]. In the context of complex I + III and II + III RC deciencies, the conrmation
of skeletal muscle coenzyme Q10 deciency, which can be associated with diverse clinical phenotypes, should prompt the sequencing of genes involved with coenzyme Q10 biosynthesis (PDSS1,
PDSS2, CoQ2, CoQ9 [39], CABC1, or ADCK3 [4042]. DNA sequencing
of TYMP needs to be considered in the clinical context of Mitochondrial NeuroGastroIntestinal Encephalomyopathy (MNGIE) and an
elevated plasma or urine thymidine level [43]. These examples

emphasize the importance of obtaining appropriate biochemical

and/or clinical information before embarking on molecular testing
in order to maximize the chance of successfully obtaining a molecular diagnosis.
Given a 520% detection rate of pathogenic mtDNA mutations
in pediatric patients with clinically suspected mitochondrial disease [11,2124] and a recent study suggesting that the detectable
frequency of pathogenic mtDNA mutations in the general population may be as high as 1 in 200 [5], screening for common mtDNA
pathogenic mutations in blood is typically the rst step, because of
the ease in obtaining a test sample and wide availability of the test.
If initial mtDNA screening from a blood sample does not establish a
molecular diagnosis, then decisions for further molecular diagnostic testing (i.e., whole mtDNA genome sequencing or nuclear gene
testing such as POLG1) should be based on the specic clinical presentation, family history and results of histological and biochemical testing from appropriate tissues (e.g., skeletal muscle or skin
broblasts), as described in the next section. There are currently
a number of clinical diagnostic laboratories that offer molecular
testing of mtDNA and various subsets of nuclear-encoded mitochondrial genes (http://genetests.org).

Recommendations for evaluation of a patient with suspected

mitochondrial disease based on currently available genetic tests
Although many RCDs are clinically heterogeneous, certain clinical features can potentially point to molecular defects in specic
genes. Thus, the evaluation of a patient with suspected mitochondrial disease should start with a thorough and accurate clinical
evaluation since it could be critical in helping to decide which tests
to perform. The evaluation should include age of onset, a detailed
family history, with an emphasis on potential matrilineal inheritance or recessive inheritance, and specic items listed in Table 1.
In addition, non-invasive evaluations including brain MRI, brain
MR spectroscopy, EKG, EEG, BAERS, and VERS, should be performed
as clinically appropriate.
This initial evaluation should also include a metabolic screening
with measurement of basic blood and urine analytes such as serum
transaminases, lactate and pyruvate (if collected appropriately),
plasma amino acids, urine organic acids, plasma acylcarnitine prole, and creatine kinase. If available, CSF lactate, pyruvate, and protein concentration should also be measured. It is useful to calculate
a lactate-to-pyruvate ratio because an elevated ratio suggests the
possibility of a mitochondrial RCD.
Based on the above clinical, metabolic, and imaging studies, if a
classical mitochondrial syndrome caused by common mitochondrial DNA (mtDNA) mutations is suspected, to promptly conrm
the diagnosis, screening for the 12 most common mitochondrial
DNA (mtDNA) point mutations and mtDNA deletions should be
performed. Although a blood specimen would be the most appropriate biological specimen to screen for most of the common
mtDNA mutations, urinary sediment is the most appropriate specimen to ascertain the common MELAS m.3243A>G mutation in the
mitochondrial tRNALeu(UUR) gene [44] as there seems to be a process of negative selection against the mutant load for this particular mutation in blood over time [45]. If maternally inherited
sensorineural hearing loss due to aminoglycoside ototoxicity is
suspected, then testing the m.1555A>G in the mitochondrial 12S
ribosomal RNA gene may be considered [46].
Since, to date, POLG1 is the most frequently mutated nuclear
gene [27,32] that, similar to mtDNA defects, causes highly variable
clinical phenotypes, DNA sequence analysis of POLG1 should be
considered following normal mtDNA screening. This is especially
true if the patient presents with non-specic hypotonia, developmental delay, and intractable epilepsy, or progressive liver disease

Please cite this article in press as: L.-J.C. Wong et al., Current molecular diagnostic algorithm for mitochondrial disorders, Mol. Genet. Metab. (2010),
doi:10.1016/j.ymgme.2010.02.024

ARTICLE IN PRESS
Lee-Jun C. Wong et al. / Molecular Genetics and Metabolism xxx (2010) xxxxxx
Table 1
Possible signs and symptoms of mitochondrial disorders.
Central nervous system
Developmental delay
Hypotonia, autistic features
Dementia/encephalopathy
Headaches/migraine
Stroke/transient ischemic episodes
Ataxia
Episodic coma
Seizures
Myoclonus or myoclonic seizures
Pyramidal signs
Hemiparesis
Spasticity
Dystonia
Chorea
Neuromuscular
Peripheral neuropathy
Exercise intolerance
Muscle weakness/DMD-like
Muscle cramps
Fatigue
Ophthalmoparesis, CPEO
Ptosis
Cardiac
Cardiomyopathy
Heart block
Arrhythmia

Visceral
Gastrointestinal reux
Delayed gastric emptying
Pancreatitis
Diarrhea
Constipation
Cyclic vomiting
Pseudoobstruction
Hepatic failure
Renal tubular acidosis
Steroid resistant nephrotic syndrome
Apnea/hypoventilation
Respiratory deciency/failure

Sensory
Retinitis pigmentosa
Optic atrophy
Cataract
Sensorineural hearing loss

Other clinical
Failure to thrive
Microcephaly
Sideroblastic anemia
Pancytopenia/bone marrow
failure
Neutropenia
Metabolites/enzymes
Elevated serum/CSF lactate
High CSF lactate
Abnormal urine organic acids
Abnormal acylcarnitine prole
Elevated CPK
Elevated CSF protein
Elevated alanine
Lactate/pyruvate ratio
Elevated transaminases
Plasma thymidine
Muscle CoQ10
Electrophysiology
Abnormal BAERS
Abnormal VERS
Abnormal EEG
Abnormal EMG
Abnormal NCV
Abnormal EKG
Imaging studies/other studies
T2 brain MRI increased signal
basal ganglia
Basal ganglia calcications
Delayed myelination
Cerebellar atrophy
Stroke-like episodes
Atypical stroke
Leukodystrophy
MRS/lactate peak
Congenital brain malformation
by MRI
Muscle biopsy
Abnormal histology
Abnormal ultrastructure (EM)
Abnormal respiratory chain
enzymes
COX decient bers
Ragged-red ber

Endocrine
Diabetes
Exocrine/pancreatic deciency
Gonadal failure
Hypothyroidism Hypoparathyroidism
Adrenal deciency
Short statue

triggered by infection in an otherwise developmentally normal

child. In addition, young adults and adults with ataxia, neuropathy,
or muscle weakness may also be candidates for sequence analysis
of POLG1 if their clinical features t within the six major clinical
phenotypes detailed above.
If only a blood sample is available and screening for mtDNA
deletions and common point mutations and POLG1 sequencing
are negative, based on the results of a clinical and metabolic evaluation, specic nuclear gene or genes can be selected for DNA
sequencing. These are depicted in Fig. 1, a diagnostic algorithm

based on currently available genetic testing of known disorders.

If a nuclear gene is suspected and gene sequencing identies only
a single heterozygous pathogenic mutation, then an intragenic
deletion of the other allele should be ruled out. Intragenic deletions
can be effectively detected with the use of clinically available targeted oligonucleotide arrays for comparative genomic hybridization [47].
However, in many cases it would be more prudent to measure
mitochondrial enzymes before ordering a specic nuclear gene
test. As an example, if a child would present with Leigh syndrome
and a family history suggestive of autosomal recessive inheritance,
it would be more rewarding to measure cytochrome c oxidase

Table 2
Isolated or generalized complex deciency: representative genes to be sequenced.
Isolated complex deciency
Complex I
 mtDNA complex I subunit genes: ND1, ND2, ND3, ND4, ND4L, ND5, and
ND6
 Nuclear-encoded complex I subunit genes: NDUFS4, NDUFS3, NDUFS7,
NDUFS6, NFUFS8, NDUFS2, NDUFS1, NDUFV1, NDUFV2, NDUFA1, and
NDUFA2
 Nuclear-encoded complex I assembly genes: NDUFAF1, NDUFAF2,
NDUFAF3, NDUFAF4, C8orf38, and C20orf7
Complex II
 Nuclear-encoded genes only: SDHA, SDHB, SDHC, and SDHD
Complex III
 mtDNA complex III subunit gene: CYTB
 Nuclear-encoded complex III assembly genes: BCS1L
Complex I/III and II/III
 Complex III deciency as above, or
 Defects in CoQ10 biosynthesis pathway: PDSS1, PDSS2, CoQ2, CoQ9, and
CABC1/ADCK3
Complex IV
 mtDNA complex IV subunit genes: COI, COII, and COIII
 Nuclear-encoded complex IV assembly genes: SURF1, SCO2, SCO1,
COX10, LRPPRC, and COX15
 Nuclear-encoded complex IV subunit genes: COX4I2 and COX6B1
Complex V
 mtDNA complex V subunit genes: ATP6 and ATP8
 Nuclear-encoded complex V assembly gene: ATP12
 Nuclear-encoded complex V ancillary factor: TMEM70
Complex I, III, and IV deciencies: reduction in all complexes except complex II
Defects in mitochondrial DNA biosynthesis and maintenance of mtDNA
integrity causing mtDNA depletion and/or mtDNA multiple deletions
 POLG1: catalytic subunit of DNA polymerase gamma
 POLG2: accessory subunit of DNA polymerase gamma
 C10orf (TWINKLE): DNA helicase
 ANT1: adenine nucleotide translocase 1
Defects in genes involved in deoxynucleotide salvage pathway causing
deoxynucleotide imbalance
 DGUOK: purine nucleotide salvage
 TK2: pyrimidine nucleotide salvage
 TYMP: thymidine phosphorylase
 SUCLA2: possibly regulate nucleotide diphosphate kinase activity
 SUCLG1: possibly regulate nucleotide diphosphate kinase activity
 RRM2B: p53 induced ribonucleotide reductase subunit
 MPV17: mitochondrial inner membrane protein, exact function not
known
Defects in mitochondrial tRNA affecting mitochondrial translation
 mtDNA tRNA gene mutations
 mtDNA deletions
 Sequence analysis of whole mitochondrial genome
Defects in nuclear genes encoding mitochondrial translation factors
 EFTs, EFTu, EFG1, and PUS1
 MRPL33 and MRPS16
Defects in mitochondrial amino acyl tRNA synthetases
 DARS2: mitochondrial aspartyl-tRNA synthetase
 RARS2: mitochondrial arginyl-tRNA synthetase

Please cite this article in press as: L.-J.C. Wong et al., Current molecular diagnostic algorithm for mitochondrial disorders, Mol. Genet. Metab. (2010),
doi:10.1016/j.ymgme.2010.02.024

ARTICLE IN PRESS
6

Lee-Jun C. Wong et al. / Molecular Genetics and Metabolism xxx (2010) xxxxxx

(COX) in broblasts before pursuing DNA testing, since normal

COX activity would reliably exclude SURF1 deciency as the basis
for the disease. Moreover, if nuclear gene analyses in blood do not
yield meaningful results, a tissue-specic analysis (biopsy of muscle, liver, or skin) should be considered as the clinically appropriate
next step for histological, biochemical, and molecular analyses. The
specic clinical presentation and the results from histochemical/
electron microscopy structural analysis, respiratory chain enzyme
assays, and mtDNA copy number (depletion, over-amplication)
analysis should be used to dictate the selection of appropriate
molecular tests (Fig. 1). The biochemical nature of the RC complex
deciency and whether it is a generalized or isolated defect can
point to mtDNA depletion, mtDNA tRNA gene mutations, or complex subunit/assembly factor gene defects (Table 2). An increasing
number of genes are known to cause mtDNA depletion both in
muscle and liver, and quantication of mtDNA copy number by
real time quantitative PCR can now be viewed as a reliable screening test for this group of gene defects. If the mtDNA copy number
increases, it suggests that the molecular defect is likely to be located in the mtDNA genome, and thus, whole mitochondrial genome sequence analysis should be considered, in particular if
there is a family history suggestive of matrilineal inheritance.
When molecular defects of mtDNA are considered, sequencing
DNA derived from the affected tissue(s), if available, may provide
the greatest probability of detecting a pathogenic mutation, since
in some cases pathogenic mtDNA mutations may give low or undetectable heteroplasmy in blood [17,48].

Conclusions
Establishing a specic diagnosis in a patient with suspected
mitochondrial disease is a complex endeavor that requires the
integration of clinical assessments, family history, biochemical
testing, histopathological examination, and directed molecular
testing. Close collaboration between primary clinicians, geneticists,
pathologists, other clinical specialists, and diagnostic laboratories
with expertise in mitochondrial biochemical and molecular testing
is critical to maximize the likelihood of obtaining a correct diagnosis. With the list of nuclear-encoded genes documented to cause
mitochondrial disease ever-expanding, the complexity of potential
molecular testing increases, but so does the prospects of identifying molecular diagnoses for patients with mitochondrial disease.

Resources
http://www.Mitomap.org.
http://mamit-trna.u-strasbg.fr/.
Human DNA polymerase gamma mutation database: http://
www.niehs.nih.gov/polg.cfm.
Human mitochondrial protein database: http://bioinfo.nist.gov/
hmpd/.
Human mitochondrial genome database: http://www.genpat.uu.se/mtDB/index.html.
Database of clinically available DNA tests: http://www.genetests.org/.
References
[1] F.P. Bernier, A. Boneh, X. Dennett, C.W. Chow, M.A. Cleary, D.R. Thorburn,
Diagnostic criteria for respiratory chain disorders in adults and children,
Neurology 59 (9) (2002) 14061411.
[2] N.I. Wolf, J.A. Smeitink, Mitochondrial disorders: a proposal for consensus
diagnostic criteria in infants and children, Neurology 59 (9) (2002) 14021405.
[3] D. Skladal, J. Halliday, D.R. Thorburn, Minimum birth prevalence of
mitochondrial respiratory chain disorders in children, Brain 126 (Pt. 8)
(2003) 19051912.

[4] A.M. Schaefer, R.W. Taylor, D.M. Turnbull, P.F. Chinnery, The epidemiology of
mitochondrial disorderspast, present and future, Biochim. Biophys. Acta
1659 (23) (2004) 115120.
[5] H.R. Elliott, D.C. Samuels, J.A. Eden, C.L. Relton, P.F. Chinnery, Pathogenic
mitochondrial DNA mutations are common in the general population, Am. J.
Hum. Genet. 83 (2) (2008) 254260.
[6] D.J. Pagliarini, S.E. Calvo, B. Chang, et al., A mitochondrial protein compendium
elucidates complex I disease biology, Cell 134 (1) (2008) 112123.
[7] R.H. Haas, S. Parikh, M.J. Falk, et al., Mitochondrial disease: a practical approach
for primary care physicians, Pediatrics 120 (6) (2007) 13261333.
[8] R.H. Haas, S. Parikh, M.J. Falk, et al., The in-depth evaluation of suspected
mitochondrial disease, Mol. Genet. Metab. 94 (1) (2008) 1637.
[9] D.M. Kirby, M. Crawford, M.A. Cleary, H.H. Dahl, X. Dennett, D.R. Thorburn,
Respiratory chain complex I deciency: an underdiagnosed energy generation
disorder, Neurology 52 (6) (1999) 12551264.
[10] B.H. Robinson, Lacticacidemia, Biochim. Biophys. Acta 1182 (3) (1993) 231
244.
[11] F. Scaglia, J.A. Towbin, W.J. Craigen, et al., Clinical spectrum, morbidity, and
mortality in 113 pediatric patients with mitochondrial disease, Pediatrics 114
(4) (2004) 925931.
[12] R.H. Triepels, L.P. van den Heuvel, J.L. Loeffen, et al., Leigh syndrome associated
with a mutation in the NDUFS7 (PSST) nuclear encoded subunit of complex I,
Ann. Neurol. 45 (6) (1999) 787790.
[13] F. Poggi-Travert, D. Martin, T. Billette de Villemeur, et al., Metabolic
intermediates in lactic acidosis: compounds, samples and interpretation, J.
Inherit. Metab. Dis. 19 (4) (1996) 478488.
[14] H. Vogel, Mitochondrial myopathies and the role of the pathologist in the
molecular era, J. Neuropathol. Exp. Neurol. 60 (3) (2001) 217227.
[15] D. Chretien, P. Rustin, Mitochondrial oxidative phosphorylation: pitfalls and
tips in measuring and interpreting enzyme activities, J. Inherit. Metab. Dis. 26
(23) (2003) 189198.
[16] J. Hui, D.M. Kirby, D.R. Thorburn, A. Boneh, Decreased activities of
mitochondrial respiratory chain complexes in non-mitochondrial respiratory
chain diseases, Dev. Med. Child Neurol. 48 (2) (2006) 132136.
[17] L.J. Wong, C.L. Perng, C.H. Hsu, et al., Compensatory amplication of mtDNA in
a patient with a novel deletion/duplication and high mutant load, J. Med.
Genet. 40 (11) (2003) e125.
[18] L.C. Papadopoulou, C.M. Sue, M.M. Davidson, et al., Fatal infantile
cardioencephalomyopathy with COX deciency and mutations in SCO2, a
COX assembly gene, Nat. Genet. 23 (3) (1999) 333337.
[19] D.R. Thorburn, J. Smeitink, Diagnosis of mitochondrial disorders: clinical and
biochemical approach, J. Inherit. Metab. Dis. 24 (2) (2001) 312316.
[20] L.P. van den Heuvel, J.A. Smeitink, R.J. Rodenburg, Biochemical examination of
broblasts in the diagnosis and research of oxidative phosphorylation
(OXPHOS) defects, Mitochondrion 4 (56) (2004) 395401.
[21] M. Jaksch, S. Hofmann, S. Kleinle, et al., A systematic mutation screen of 10
nuclear and 25 mitochondrial candidate genes in 21 patients with cytochrome
c oxidase (COX) deciency shows tRNA(Ser)(UCN) mutations in a subgroup
with syndromal encephalopathy, J. Med. Genet. 35 (11) (1998) 895900.
[22] M. Jaksch, S. Kleinle, C. Scharfe, et al., Frequency of mitochondrial transfer RNA
mutations and deletions in 225 patients presenting with respiratory chain
deciencies, J. Med. Genet. 38 (10) (2001) 665673.
[23] M.H. Liang, L.J. Wong, Yield of mtDNA mutation analysis in 2000 patients, Am.
J. Med. Genet. 77 (5) (1998) 395400.
[24] R. Marotta, J. Chin, A. Quigley, et al., Diagnostic screening of mitochondrial
DNA mutations in Australian adults 19902001, Intern. Med. J. 34 (12) (2004)
1019.
[25] A.M. Schaefer, R. McFarland, E.L. Blakely, et al., Prevalence of mitochondrial
DNA disease in adults, Ann. Neurol. 63 (1) (2008) 3539.
[26] M. Naimi, S. Bannwarth, V. Procaccio, et al., Molecular analysis of ANT1,
TWINKLE and POLG in patients with multiple deletions or depletion of
mitochondrial DNA by a dHPLC-based assay, Eur. J. Hum. Genet. 14 (8) (2006)
917922.
[27] L.J. Wong, R.K. Naviaux, N. Brunetti-Pierri, et al., Molecular and clinical
genetics of mitochondrial diseases due to POLG mutations, Hum. Mutat. 29 (9)
(2008) E150E172.
[28] E.A. Shoubridge, Cytochrome c oxidase deciency, Am. J. Med. Genet. 106 (1)
(2001) 4652.
[29] M. Bohm, E. Pronicka, E. Karczmarewicz, et al., Retrospective, multicentric
study of 180 children with cytochrome c oxidase deciency, Pediatr. Res. 59
(1) (2006) 2126.
[30] A. Bourdon, L. Minai, V. Serre, et al., Mutation of RRM2B, encoding p53controlled ribonucleotide reductase (p53R2), causes severe mitochondrial
DNA depletion, Nat. Genet. 39 (6) (2007) 776780.
[31] R. Carrozzo, C. Dionisi-Vici, U. Steuerwald, et al., SUCLA2 mutations are
associated with mild methylmalonic aciduria Leigh-like encephalomyopathy,
dystonia and deafness, Brain 130 (Pt. 3) (2007) 862874.
[32] D.P. Dimmock, Q. Zhang, C. Dionisi-Vici, et al., Clinical and molecular features
of mitochondrial DNA depletion due to mutations in deoxyguanosine kinase,
Hum. Mutat. 29 (2) (2008) 330331.
[33] O. Elpeleg, C. Miller, E. Hershkovitz, et al., Deciency of the ADP-forming
succinyl-CoA synthase activity is associated with encephalomyopathy and
mitochondrial DNA depletion, Am. J. Hum. Genet. 76 (6) (2005) 10811086.
[34] M. Oskoui, G. Davidzon, J. Pascual, et al., Clinical spectrum of mitochondrial
DNA depletion due to mutations in the thymidine kinase 2 gene, Arch. Neurol.
63 (8) (2006) 11221126.

Please cite this article in press as: L.-J.C. Wong et al., Current molecular diagnostic algorithm for mitochondrial disorders, Mol. Genet. Metab. (2010),
doi:10.1016/j.ymgme.2010.02.024

ARTICLE IN PRESS
Lee-Jun C. Wong et al. / Molecular Genetics and Metabolism xxx (2010) xxxxxx
[35] E. Ostergaard, E. Christensen, E. Kristensen, et al., Deciency of the alpha
subunit of succinate-coenzyme A ligase causes fatal infantile lactic acidosis
with mitochondrial DNA depletion, Am. J. Hum. Genet. 81 (2) (2007) 383387.
[36] E. Ostergaard, F.J. Hansen, N. Sorensen, et al., Mitochondrial
encephalomyopathy with elevated methylmalonic acid is caused by SUCLA2
mutations, Brain 130 (Pt. 3) (2007) 853861.
[37] L.J. Wong, N. Brunetti-Pierri, Q. Zhang, et al., Mutations in the MPV17 gene are
responsible for rapidly progressive liver failure in infancy, Hepatology 46 (4)
(2007) 12181227.
[38] A. Saada, A. Shaag, H. Mandel, Y. Nevo, S. Eriksson, O. Elpeleg, Mutant
mitochondrial thymidine kinase in mitochondrial DNA depletion myopathy,
Nat. Genet. 29 (3) (2001) 342344.
[39] A.J. Duncan, M. Bitner-Glindzicz, B. Meunier, et al., A nonsense mutation in
COQ9 causes autosomal-recessive neonatal-onset primary coenzyme Q10
deciency: a potentially treatable form of mitochondrial disease, Am. J. Hum.
Genet. 84 (5) (2009) 558566.
[40] C. Lagier-Tourenne, M. Tazir, L.C. Lopez, et al., ADCK3, an ancestral kinase, is
mutated in a form of recessive ataxia associated with coenzyme Q10
deciency, Am. J. Hum. Genet. 82 (3) (2008) 661672.
[41] J. Mollet, A. Delahodde, V. Serre, et al., CABC1 gene mutations cause
ubiquinone deciency with cerebellar ataxia and seizures, Am. J. Hum.
Genet. 82 (3) (2008) 623630.

[42] C.M. Quinzii, S. DiMauro, M. Hirano, Human coenzyme Q10 deciency,

Neurochem. Res. 32 (45) (2007) 723727.
[43] I. Nishino, A. Spinazzola, M. Hirano, Thymidine phosphorylase gene mutations
in MNGIE, a human mitochondrial disorder, Science 283 (5402) (1999) 689
692.
[44] M.T. McDonnell, A.M. Schaefer, E.L. Blakely, et al., Noninvasive diagnosis of the
3243A > G mitochondrial DNA mutation using urinary epithelial cells, Eur. J.
Hum. Genet. 12 (9) (2004) 778781.
[45] S. Rahman, J. Poulton, D. Marchington, A. Suomalainen, Decrease of 3243 A>G
mtDNA mutation from blood in MELAS syndrome: a longitudinal study, Am. J.
Hum. Genet. 68 (1) (2001) 238240.
[46] M.X. Guan, N. Fischel-Ghodsian, G. Attardi, Nuclear background determines
biochemical phenotype in the deafness-associated mitochondrial 12S rRNA
mutation, Hum. Mol. Genet. 10 (6) (2001) 573580.
[47] L.J. Wong, D. Dimmock, M.T. Geraghty, et al., Utility of oligonucleotide arraybased comparative genomic hybridization for detection of target gene
deletions, Clin. Chem. 54 (7) (2008) 11411148.
[48] L.J. Wong, C. Wladyka, R. Mardach-Verdon, A mitochondrial DNA mutation in a
patient with an extensive family history of Duchenne muscular dystrophy,
Muscle Nerve 30 (1) (2004) 118122.

Please cite this article in press as: L.-J.C. Wong et al., Current molecular diagnostic algorithm for mitochondrial disorders, Mol. Genet. Metab. (2010),
doi:10.1016/j.ymgme.2010.02.024