Download a role for mitochondrial enzymes in inherited neoplasia and beyond

REVIEWS
A ROLE FOR MITOCHONDRIAL
ENZYMES IN INHERITED NEOPLASIA
AND BEYOND
Charis Eng*‡, Maija Kiuru§, Magali J. Fernandez* and Lauri A. Aaltonen§
Mitochondrial defects have been associated with neurological disorders, as well as cancers. Two
ubiquitously expressed mitochondrial enzymes — succinate dehydrogenase (SDH) and fumarate
hydratase (FH, fumarase) — catalyse sequential steps in the Krebs tricarboxylic-acid cycle.
Inherited heterozygous mutations in the genes encoding these enzymes cause predispositions to
two types of inherited neoplasia syndromes that do not share any component tumours.
Homozygous mutations in the same genes result in severe neurological impairment. Understanding
this link between inherited cancer syndromes and neurological disease could provide further
insights into the mechanisms by which mitochondrial deficiencies lead to tumour development.
*Clinical Cancer Genetics
Program, Human Cancer
Genetics Program,
Comprehensive Cancer
Center, and Division of
Human Genetics,
Department of Internal
Medicine, The Ohio State
University, 420 W. 12th
Avenue, Ste 690 TMRF,
Columbus, Ohio 43210,
USA.
‡
Cancer Research UK
Human Cancer Genetics
Research Group, University
of Cambridge, Cambridge
CB2 2XZ, UK.
§
Department of Medical
Genetics, Biomedicum
Helsinki, P.O.Box 63
(Haartmaninkatu 8),
Helsinki FIN-00014,
Finland.
Correspondence to C. E.
e-mail: eng-1@
medctr.osu.edu
doi:10.1038/nrc1013
NATURE REVIEWS | C ANCER
Mitochondria have traditionally been viewed as the
powerhouse organelles in eukaryotic cells, charged
with the role of energy production. Much of its
machinery — the enzymes that catalyse energy production — have been studied in this role. In general,
therefore, inherited mitochondrial defects, whether
nuclear- or mitochondrial-encoded, result in severe
clinical sequelae that affect the entire organism, or at
least affect the main organs, such as the central nervous system and the heart, which have the highest
energy consumption. The central nervous system is the
most commonly affected, and patients experience
seizures, fluctuating encephalopathy, stroke-like
episodes, migraines, dementia and ataxia. Nonetheless,
the spectrum of the clinical manifestations is heterogeneous. Some patients develop symptoms that are
characteristic of a specific syndrome, whereas others
cannot be clearly categorized.
Mitochondrial enzymes were not directly implicated in hereditary neoplasias until recently. Germline
heterozygous mutations in the autosomally encoded
mitochondrial enzyme subunits SDHD, SDHC and
SDHB cause the inherited syndromes phaeochromocytoma and paraganglioma. Mutations in FH, which
encodes fumarase, cause a predisposition to cutaneous
and uterine leiomyomas, as well as to kidney cancers.
Together, the four subunits of succinate dehydrogenase (SDH) — SDHA, SDHB, SDHC and SDHD —
comprise mitochondrial complex II (also known
as succinate:ubiquinone oxydoreductase), which is
involved in electron transport and the Krebs tricarboxylic-acid cycle (BOX 1). Complex II catalyses the
energy-dependent conversion of succinate to fumarate.
Fumarate is converted to malate by fumarase as the
next step (BOX 1).
Homozygous mutations in SDHA and FH result
in severe neurological dysfunction. This is often
explained by the fact that the developing nervous system is very sensitive to any perturbation in metabolism or homeostasis (BOX 1). There are fundamental
differences between the neoplasia and neurological
disorders, however. In individuals with the neurodegenerative disorders, there is virtually complete loss
of SDHA or FH activity, and its consequences are
experienced throughout development. The inherited
neoplastic syndromes are associated with reductions
in SDH or FH gene dosage, and complete or nearcomplete loss of enzyme function occurs only in the
somatic cells that comprise the tumour. The mitochondrial effects of reduced SDHA or FH activity
might affect apoptosis and angiogenesis induction,
which are both hallmarks of tumorigenesis.
VOLUME 3 | MARCH 2003 | 1 9 3
REVIEWS
Summary
• Germline heterozygous mutations in the autosomally encoded mitochondrial enzyme
subunits of succinate dehydrogenase (SDH), SDHB, SDHC and SDHD, are associated
with hereditary predisposition to phaeochromocytoma and paraganglioma. By
contrast, homozygous germline mutations in the catalytic active-site-bearing subunit
SDHA causes Leigh syndrome, which is characterized by severe neurological
dysfunction and seizures.
• Germline heterozygous mutations in another autosomally encoded mitochondrial
enzyme — fumarate hydratase (fumarase, FH) — are associated with hereditary
predisposition to papillary renal-cell carcinoma and leiomyomatosis, whereas
homozygous FH mutations cause neurodegeneration.
• SDH and FH catalyse sequential steps in the Krebs tricarboxylic-acid cycle, which
generates ATP — the cell’s currency of energy. SDH is a component of complex II of
the respiratory electron-transport chain.
• The hereditary neurological diseases might be explained by complete or near-complete
lack of energy generation during development, leading to free-radical formation and
mitochondrial-mediated apoptotic cell death.
• Little data exist to explain the mechanism of predisposition to cancer. Hypotheses
invoke free-radical formation, leading to activation of the HIF/angiogenesis pathway
and mitochondrial-mediated anti-apoptotic activity.
• The link between mitochondrial-associated inherited neurological disease and
inherited cancer might be exploited for uncovering novel functions and mechanisms
for mitochondrial enzymes beyond energy production, for novel gene discovery and
for clinical utility.
Mitochondrial defects and inherited cancer
MATERNALLY IMPRINTED
Genes come in pairs (alleles) —
one inherited from the father
and one from the mother.
Maternal imprinting occurs
when the maternal allele is
transcriptionally silenced,
usually because of promoter
hypermethylation. In this
situation, only the paternal allele
is expressed.
FOUNDER EFFECT
When a germline mutation
occurs in a community that has
been isolated from the outside,
or where little emigration or
immigration has occurred for
many decades or centuries, this
‘founding’ germline mutation is
established and perpetrated for
generations.
194
The first inherited neoplasia syndrome that was
attributed to germline mutations in nuclear-encoded
mitochondrial enzymes involved tumours of the
adrenal medulla and paraganglia. Phaeochromocytomas are catecholamine-secreting tumours that
are benign in ~90% of cases. Phaeochromocytomas
usually arise within the adrenal medulla but can
develop in extra-adrenal sympathetic ganglia ~10%
of the time. Extra-adrenal phaeochromocytoma are
sometimes referred to as paraganglioma. Vascular
head and neck tumours, which usually arise from the
carotid body, are referred to as head and neck paragangliomas, glomus tumours or chemodectomas.
Inherited cancer syndromes with phaeochromocytoma can be a component feature of several inherited
syndromes, each with distinct genetic aetiologies1–3
(TABLE 1).
Familial clustering of glomus tumours was first
brought to common attention when reported in 1964
(REF. 4), although the earliest report might have been in
1937. After studying such families, these and other
investigators noted that glomus tumours were
inherited from the paternal line5,6. Clinical observations
indicated that the gene that caused familial glomus
tumours was MATERNALLY IMPRINTED (paternally
expressed). Working on this premise, these investigators
mapped a gene for familial glomus tumours to 11q23
and showed that this predisposition was most likely
due to a FOUNDER EFFECT in the Dutch population6,7. This
locus was labelled PGL1. PGL1 was subsequently identified as SDHD 8, and germline mutations are associated
with familial and isolated paragangliomas and
phaeochromocytomas8–11 (FIG. 1).
| MARCH 2003 | VOLUME 3
It also became obvious that there exist familial clusterings of paragangliomas — of the head and neck
region, and elsewhere — which were not linked to PGL1
and were not maternally imprinted. Germline mutations in SDHB (PGL4) and to a much lesser extent
SDHC (PGL3) were subsequently found in some families with paragangliomas and phaeochromocytomas,
and these are not associated with imprinting11–14. PGL2
— another locus for familial paragangliomas in a large
Dutch family — has yet to be precisely identified.
Diagnosis with this cancer usually occurs when patients
are in their mid-20s (range 5–60 years)10.
Interestingly, germline heterozygous mutations of
FH, which encodes an enzyme that catalyses the next
step after SDH in aerobic metabolism, cause a
completely different autosomal-dominant syndrome
— hereditary leiomyomatosis and type 2 papillary
renal-cell carcinoma (HLRCC). The clinical and
histopathological characteristics of the renal-cell carcinomas supported a common hereditary background
for these malignancies, which are distinct from other
known inherited kidney cancer syndromes (TABLE 2).
The apparently dominantly inherited tumours are
solitary and aggressive, and usually occur early in life
(between the ages of 26 and 48 years). A key feature
of these tumours is their histopathology — all component renal carcinomas that have been examined
have a peculiar papillary histology15,16, with resemblance
to the rare papillary type 2 histology17. Phenotypic
features can vary among families. Some families only
develop cutaneous and uterine leiomyomatosis without renal cancer, which are sometimes referred to as
multiple cutaneous leiomyomatosis (MCL). The
HLRCC susceptibility gene was mapped to 1q42–q43
and germline mutations in FH were subsequently
identified15,16,18–20 (FIG. 2).
It is not clear why germline mutations in genes
that encode two enzymes that act sequentially (BOX 1),
and that are ubiquitously expressed, result in two
inherited neoplasia syndromes with non-overlapping
component tumours. Virtually all tumours that are
associated with both syndromes have been shown to
have lost the remaining wild-type allele 8,15,21–23.
Further research is required to determine how loss of
the wild-type allele occurs.
The observation that germline mutations in genes
that encode mitochondrial enzymes are associated with
inherited cancer syndromes is relatively recent. In addition, somatic and non-somatic mutations and variants
in the mitochondrial genome of sporadic neoplasias
have been described (BOXES 1 and 2). Although no precise
mechanism has been elucidated, hypotheses regarding
the mitochondrial role in both energy metabolism and
apoptosis have been proposed (BOX 3).
Mitochondrial (dys)function I: SDH
Traditionally, the mitochondria are viewed as organelles
that generate energy (typically ATP) for cellular metabolism. ATP is generated via two inter-linked mitochondrial
pathways — the electron-transport (or respiratory)
chain and the Krebs tricarboxylic-acid cycle. There are
www.nature.com/reviews/cancer
REVIEWS
five mitochondrial complexes that participate in the electron-transport chain (BOX 1). Complex II consists of SDH
— a four-subunit enzyme that straddles the inner mitochondrial membrane, such that it can easily participate in
both the electron-transport chain (membrane-associated)
and Krebs cycle (mitochondrial-matrix-associated). This
complex is anchored to the membrane by two membrane-spanning proteins — SDHC and SDHD. The
peripheral portion of this complex is made up of a flavoprotein (SDHA) and an iron–sulphur subunit (SDHB),
and projects into the mitochondrial matrix. The presence
and proper functioning of both SDHA and SDHB are
required for catalytic activity. The active site of the enzyme
is located within SDHA and contains covalently bound
FAD (BOX 1). SDH catalyses the conversion of succinate
to fumarate, which, in turn, is converted to malate by
fumarase. Fumarase is located in the mitochondrial
matrix and participates in the Krebs cycle.
Defects in the Krebs cycle have been observed to
affect the nervous system and the skeletal muscles,
resulting in early-onset symptoms when oxidative
metabolism is impaired. This reflects the dependence
of neurons and muscle cells on oxidative phosphorylation for energy requirements. Evidence also indicates
that neurodegenerative diseases, such as Huntington’s
disease, Parkinson’s disease and Friedreich ataxia, are
Box 1 | Mitochondrial complexes I–V in health and disease
The diagram shows the five complexes that are involved in the mitochondrial electron-transport chain. Complexes I–IV
are the electron-transport complexes, whereas complex V synthesizes adenosine triphosphate (ATP). Electrons are passed
down the four complexes (black arrows) to molecular oxygen and then complex V generates one ATP molecule from ADP
and inorganic phosphate (Pi). The blue arrows show where the protons are pumped to the cytosolic side of the intermembrane to generate an electrochemical gradient, and where proton movement back across complex V (the ATP
synthase) is used to drive ATP synthesis.
Complex II is made up of the four components of SDH. It not only resides within the mitochondrial membrane as an
active participant of the electron-transport chain, but also contacts the mitochondrial matrix, where it participates in the
Krebs tricarboxylic-acid cycle. In this case, it converts succinate to fumarate. Fumarate hydratase catalyses the subsequent
step in the Krebs cycle, of fumarate to malate. SDHA mutations have been associated with Leigh syndrome,
paraganglioma and phaeochromacytoma.
Mutations in genes that encode complex I proteins, such as NDUFV1, have been associated with leukodystrophy and
epilepsy47. NDUFS2 mutations have also been associated with leukodystrophy48 and cardioencephalomyopathy (CMP)49.
BCS1 mutations disrupt complex III function, and have been described in patients with tubulopathy, encephalopathy
and hepatopathy50. Nuclear DNA mutations in genes such as SURF1, SCO1, SCO2 and COX10 disrupt complex IV
function, and are associated with Leigh syndrome, cardioencephalomyopathy, hepatic failure and encephalopathy51,52.
Disorders caused by heritable mitochondrial DNA (mtDNA) mutations have broad manifestations53. In addition to Leigh
syndrome, mtDNA mutations in genes that encode mitochondrial proteins have been associated with chronic progressive
external ophthalmoplegia, Kearns–Sayre syndrome, Leber hereditary optic neuropathy, and neurogenic weakness with ataxia
and retinitis pigmentosa. Heritable mutations that affect mtDNA-encoded tRNA have been associated with mitochondrial
encephalomyopathy with lactic acidosis and stroke-like episodes, as well as myoclonic epilepsy with ragged red fibres.
Cytosol
H+
H+
H+
Cyt c
H+
c
Fe-S
FMN
Mitochondrial
matrix
NADH
CoQ
FADH
Fe-S
FAD+
and H+
SDH
NAD+ and H+
Fe-S
Cu-a
b
Cu-a3
H2O
1/2 O2
Succinate
Fumarate
Fumarate
Hydratase
ADP
and Pi
ATP
Malate
NATURE REVIEWS | C ANCER
Complex
I
II
III
IV
V
Nuclear gene
mutations
NDU FS1, 2, 4,
7,8
NDU FV1
SDHA, B, C, D
BCS1L
SURF1, SCO1,
SCO2, COX10
Phenotype
Leigh syndrome
Leukodystrophy
Epilepsy
Cardioencephalomyopathy
Hepatopathy
Leigh syndrome
Encephalopathy
Paraganglioma
Phaeochromocytoma
Leigh syndrome
Cardioencephalomyopathy
Hepatopathy
Cardioencephalomyopathy
Hepatopathy
VOLUME 3 | MARCH 2003 | 1 9 5
REVIEWS
Table 1 | Inherited cancer syndromes with phaeochromocytoma as a component
Syndrome
Other component neoplasias
Susceptibility gene(s)
Von Hippel–Lindau disease
Renal-cell carcinoma, retinal angioma, central nervous system
haemangioma/haemangioblastoma
VHL
Multiple endocrine
neoplasia type 2
Medullary thyroid carcinoma, parathyroid hyperplasia/adenoma RET
Type 1 neurofibromatosis
Plexiform neurofibroma, schwannoma, astrocytoma, juvenile
chronic myeloid leukaemia, breast carcinoma
NF1
Phaeochromocytoma–
paranganglioma syndrome
None
SDHD, SDHC, SDHB
associated with mitochondrial dysfunction. It has been
postulated that respiratory-chain dysfunction and
free-radical formation, together with mitochondrial
defects, result in cell damage and apoptosis24 (BOX 3).
Homozygous mutations in SDHA cause a markedly
different phenotype than heterozygous mutations in the
genes that encode the other three SDH subunits. SDHA
mutations cause a predisposition to Leigh syndrome — a
neurodegenerative disorder that is characterized by subacute necrotizing encephalomyelopathy during infancy25,
failure to thrive and developmental delays. Lactic acidosis, seizures, ataxia and multiple brain lesions are also
part of the phenotype, and the basal ganglia, midbrain
and brainstem are frequently involved. Mutations of
SDHA usually result in inactivation of the catalytic subunit, resulting in loss of, or markedly reduced, enzymatic
activity. Half of the reported mutations (such as A524V
and R554W) occur in the 3′ end of the gene and disrupt
a SDHD
R92Y
L95P
L139P
P81L
C.33-36delTG
R38X
C11X
IVS1+2t>g
R22X
Q36X
W5X
1
1
R70G
2
3
c.54InsC
b SDHB
4
c.123InsC
W43X
R90X
C.207-210InsC
R46G
C101Y
R27X Q59X
2
Q109X
c.381-383delG
H102L
Q121X
3
4
C.334-7delGACT
P131R
5
P198R R242H
C192R C.847del4
C249Y
C196Y
6
7
8
Figure 1 | Schematic diagram of the SDHD and SDHB genes and germline heterozygous
mutations found in the phaeochromocytoma–paraganglioma syndromes. Both
truncating and missense mutations have been associated with disease phenotypes. SDHD R92Y
and L139P are Dutch founder mutations (stars). Germline mutations in SDHB are most
commonly associated with adrenal phaeochromocytoma, whereas mutations in SDHD are most
commonly associated with extra-adrenal paraganglioma. Mutations at the 5′ end of SDHD are
also believed to be associated with adrenal phaeochromocytoma. Yellow bars represent the
length of the coding region of the gene, which is divided into exons (numbered). Green squares
denote mutations that are associated with adrenal phaeochromocytoma, and blue circles signify
mutations that are associated with extra-adrenal disorders. Each unit square or circle denotes a
single family or proband that has been reported to have that mutation.
196
| MARCH 2003 | VOLUME 3
the carboxy-terminal region of the enzyme26,27. This
region of the protein folds around the highly conserved
catalytic core, and serves as gatekeeper by opening or
closing the active site28. Such carboxy-terminal mutations
result in loss of catalytic activity without alterations in
substrate binding26–28. In other words, not only do these
mutations result in loss of function, but they also serve as
dominant negatives or ‘substrate sinks’. Post-mitotic cells,
such as those of the central nervous system, have the
highest energy requirements and oxygen consumption.
Because of the latter, these cells also have the highest
potential for superoxide production. So, loss of SDHA
function results in decreased or loss of energy production from both the electron-transport chain and the
Krebs cycle, as well as overproduction of free radicals,
and this leads to severe early-onset neurodegeneration.
Mutations away from the carboxy-terminal domain
result in partial loss of enzymatic activity, decreased
energy production and later-onset neurodegeneration.
Unlike those associated with Leigh syndrome,
phaeochromocytoma- and paraganglioma-associated
mutations in the non-catalytic subunits of complex II
are likely to result in destabilization and perhaps even
loss of structural integrity of this complex14. Mutations
that affect the anchor proteins SDHC and SDHD are
predicted to result in complex disassembly. So far, only
one SDHC mutation that alters the first amino acid has
been described — in a single family with paragangliomas12. This mutation destroys the translational start
site and so, the protein must initiate translation at the
next methionine at residue 9. This truncated protein
lacks the first few amino acids, which act as the signal
peptide/presequence, and is therefore unable to insert
itself into the mitochondrial membrane. Because this is
a heterozygous germline mutation, only half of the
SDHC peptides are affected, and, probably, no physiological sequelae occur. However, loss of the remaining
wild-type SDHC allele occurs in the paragangliomas12,
thereby resulting in complete lack of SDHC in complex
II. Because both the anchor polypeptides — SDHC
and SDHD — are required to bind and hold SDHA and
SDHB within the membrane28, only SDHD would
remain in the complex if SDHC were absent.
With the available data, it seems that three-quarters
of all individuals with phaeochromocytoma and
germline SDHD mutations possess mutations in the
5′ portion of SDHD (FIG. 1). Head and neck paragangliomas, by contrast, are associated with mutations in
the 3′ region of this gene. Interestingly, approximately
www.nature.com/reviews/cancer
REVIEWS
Table 2 | Syndromes associated with heritable epithelial renal carcinomas
Syndrome
Typical renal histological type
Other component neoplasias
Von Hippel–Lindau disease Clear-cell renal-cell carcinoma
Hereditary papillary
renal-cell carcinoma
Susceptibility
gene
Phaeochromocytoma, retinal
VHL
angioma, central nervous system
haemangioma/haemangioblastoma
Type 1 papillary renal-cell carcinoma Papillary thyroid carcinoma (rarely)
MET
Hereditary leiomyomatosis Type 2 papillary renal-cell carcinoma Uterine and cutaneous leiomyoma
renal-cell carcinoma
FH
Birt–Hogg–Dube syndrome Chromophobe renal carcinomas;
oncocytomas
Folliculin
50% of all germline SDHD mutations occur in the
first two exons and the beginning of exon 3 (up
through codon 63) (FIG. 1). Disruption of this region
would result in alterations of the presequence (signal
peptide) and inability to insert into the mitochondrial
membrane. The inability to insert into the mitochondrial membrane would also occur in the presence of
truncating mutations even after amino acid 63.
Therefore, the ultimate consequence of most naturally
occurring germline SDHD mutations is to cause disassembly of complex II, leaving only SDHC present in
the membrane and absence of its catalytic activity21.
Missense mutations that occur after codon 63 (FIG. 1)
would not be expected to cause complete disassembly of
complex II, and so are not predicted to completely inactivate the complex. For example, whereas the common
P81L and L139P mutations (FIG. 1) are expected to perturb
the α-helical structure of SDHD, even complete loss of the
α-helix around L139 does not completely inactivate the
yeast complex29. In view of these functional sequelae, we
conclude that complete disassembly of complex II causes
the predisposition to phaeochromocytoma, whereas
Fibrofolliculoma, colorectal
carcinoma (?)
partial inactivation of its catalytic activity leads to head
and neck paragangliomas.
SDHB is the iron–sulphur subunit that is required
to act together with SDHA during catalysis. SDHB
cysteine-197 serves as the ligand to the iron–sulphur
moiety. Together with proline-198 and surrounding
residues (exons 6 and 7), they form the key interface
with anchor proteins SDHC and SDHD30. So, the
phaeochromocytoma-susceptibility-associated
germline SDHB mutations in exons 6 and 7 or
truncating mutations before residue 197 (FIG. 1) are
expected to prevent the assembly of the catalytic complex31,32. The physiological consequence would be a
mixture of wild-type complex II and complexes that
only contain SDHC and SDHD28,32, resulting in loss of
enzymatic activity22.
In addition to complex II integrity and catalytic
activity, other mechanisms must also come into play in
neoplasia. The carotid body contains oxygen chemoreceptors, so it is possible that chronic hypoxia (low oxygen tension) could account for the sporadic occurrence
of glomus tumours in individuals who live at high
N64T
K187del
K187R
541delAG
H137R
R190H
G239V
Q142R
IVS5-15T->G
A74P
H153R
R58X
121delTG(s)
48delC
R8E (s)
Q4X
1
2
3
66del174
4
5
6
K187R
R190H
R190C
Nonsense
Missense
Deletion (frameshift)
Deletion (in-frame)
E319D(s)
R300X
R300X(s)
M285R(s)
7
A265T
F269C
1220delG
8
E319Q
9
1302insAAA
D283V
10
W458X
Insertion (in-frame)
Deletion (whole gene)
Splice-site mutation
Figure 2 | The FH gene and disease-associated mutations. The yellow bar represents the structure of the FH gene, with the
numbers denoting the ten exons. Heterozygous mutations in FH have been detected in 27 probands that have a predisposition to
cancer19. In patients with tumour predisposition, the mutations seem to cluster at the 5′ end of FH (mutations noted above the bar),
whereas mutations associated with FH deficiency tend to cluster in the 3′ half of the gene (p <0.0001, mutations shown below the
bar). However, both sets of mutations seem to cause very similar reductions in FH activity. In addition to germline changes, the figure
includes somatic mutation(s) detected in tumours.
NATURE REVIEWS | C ANCER
VOLUME 3 | MARCH 2003 | 1 9 7
REVIEWS
Box 2 | Somatic alterations of mitochondrial genes in sporadic cancers
In addition to energy metabolism, the mitochondrial role in apoptosis43,54 led
investigators to consider genes within the mitochondrial genome as putative sporadic
cancer-associated genes. This was a particularly attractive hypothesis for the so-called
oncocytic tumours, which are characterized by their abundance of mitochondria. The
kidney and thyroid are the two most common sites for oncocytic tumours. Because of
the size of the mitochondrial genome (16.6 kb), only one study was able to
systematically examine virtually all the coding sequence in oncocytic tumours —
specifically in thyroid tumours55. In that study, 23% of papillary thyroid carcinomas
were found to have somatic mitochondrial gene mutations. More interestingly, the
distribution and spectra of non-somatic mitochondrial gene variants seemed to differ
between individuals with thyroid cancer compared to normal controls, with complex I
variants favoured amongst cancer cases. Together, these observations indicated that
somatic mitochondrial gene mutations might be involved in thyroid tumorigenesis, and
that accumulation of mitochondrial DNA variants might be related to tumour initiation
or progression in the thyroid. This hypothesis seemed to be corroborated when
sequencing of the entire mitochondrial genome in colorectal cancers revealed 12
different somatic mutations and 88 non-somatic sequence variants56. Subsequently,
mutations and variation in the mitochondrial genome have been identified in various
solid tumours, including those of the ovaries, stomach, breast, brain and prostate57–61.
Because of the size of the mitochondrial genome, most of these analyses were limited to
parts of the genome only.
altitude. Chronic hypoxia could also explain the mechanism whereby mutations in SDHX proteins lead to
hereditary paragangliomas, and by extrapolation, to
phaeochromocytomas8,14,33. Mitochondria generate
reactive oxygen species (ROS), which activate the transcription factor hypoxia-inducible factor-1 (HIF-1).
HIF-1 is a heterodimeric protein that consists of two
subunits — HIF-1α and HIF-1β. Whereas HIF-1β is
constitutively expressed, the expression of HIF-1α is
induced by low oxygen concentrations. HIF-1 activates
the transcription of genes encoding vascular endothelial
growth factor (VEGF) and glycolytic enzymes34, which
promote angiogenesis — a necessary component of
tumour growth.
Gimenez-Roqueplo and colleagues examined HIF1 activity in phaeochromocytoma samples from a
family with germline SDHD R22X mutations, compared to eight sporadic phaeochromocytomas (which
did not contain SDHD mutations)21. The phaeochromocytoma cells from the patient with hereditary disease had lost the remaining wild-type allele, and were
disrupted in complex II electron-transfer activity.
Levels of HIF-1α, HIF-2α/EPAS1, VEGF and its
receptor VEGF-R1 were shown to be increased only in
the SDHD-related phaeochromocytoma. The sporadic phaeochromocytomas, conversely, retained full
activity of complex II. Most sporadic phaeochromocytomas contain somatic mutations and/or deletions
of the tumour suppressor von Hippel–Lindau protein
(VHL). VHL inhibits HIF-1 activity by inducing
degradation of one of its subunits23 (TABLES 1 and 2).
Interestingly, naturally occurring VHL mutations that
are associated with phaeochromocytoma have not
been shown not to impair HIF-1α ubiquitylation or
subsequent degradation35,36. So, there seem to be different downstream consequences of SDHD-related
and sporadic phaeochromocytomas.
198
| MARCH 2003 | VOLUME 3
Mitochondrial (dys)function II: FH
Heterozygous germline mutations in FH that have been
observed in HLRCC families are loss-of-function
mutations that cause either truncating or missense
changes at highly conserved residues19 (FIG. 2). All mutations (11 of 16 different germline mutations) examined
had defects in fumarase activity19. The activities were
similar to those observed in heterozygous first-degree
relatives of fumarase-deficiency patients. Interestingly,
missense mutations resulted in more severely reduced
fumarase activity than truncating mutations, including
a whole gene deletion. This is conceivable because
fumarase functions as a homotetramer37, and a missense mutation in one allele results in only 1 out of 16
fully wild-type homotetramers. There are three distinct
domains, including a five-helical α-helix and a lysase
domain, in each fumarase subunit. These domains are
characteristic for this class of protein superfamily,
which includes fumarase, aspartase, argininosuccinate
lyase, adenylosuccinate lyase and crystalline. Alphahelixes from the four subunits form a superhelical
structure in the core of the tetramer, which is a putative
active site. The lyase domain (amino acids 364–373 in
human FH) sequence motif (GSX(x2)MX(x2)KN) is
thought to be involved in catalytic activity of the protein. It is conceivable that mutations in FH might affect
multimerization of the gene product, resulting in
defective catalytic activity and affecting other putative
functions of FH.
In general, the mutations that are associated with
HLRCC occur in the 5′ of the gene, whereas, the
FH-deficiency mutations tend to occur in the 3′ end
(p <0.0001)18,19 (FIG. 2). However, there is some overlap.
For example, a homozygous R190H mutation has
been described in FH deficiency and a heterozygous
R190H mutation in HLRCC. The phenotypes are also,
at least to some extent, overlapping; heterozygous parents of patients that are recessive for fumarase deficiency are predisposed to tumours 19. One obvious
explanation for the observed difference in the mutation spectra could be the degree of fumarase deficiency involved, in that 3′ mutations result in more
severe enzymatic deficiencies. However, this simple
explanation is unlikely, because the activities seen in
the context of tumour-predisposing mutations
are very similar to those in the recessive fumarase
deficiency. The preferential mutation spectrum is
intriguing and implies that the molecular basis of FHdeficient tumorigenesis might not simply relate to the
enzymatic defect.
Considering the findings from the studies of SDHX
genes, the association of FH with HLRCC was not a
complete surprise. FH acts in the Krebs cycle, in the step
following that of the SDH complex (BOX 1). Unlike SDH,
FH has no known role in the electron-transport chain.
Given this fact, the mechanisms of tumorigenesis that
are driven by FH defects can only be speculated about. It
has been proposed that DNA mutations that result from
oxidative damage could promote tumorigenesis, but
this remains unproven19. Another suggestion has been
that blockage of the tricarboxylic-acid cycle leads to
www.nature.com/reviews/cancer
REVIEWS
feedback on oxidative metabolism, and subsequently on
the cell cycle, through angiogenic pathways19. Other
unknown functions of SDH and FH might also be
involved in tumour suppression.
Neoplasia or neurodegeneration?
The observation that homozygous loss-of-function
germline mutations in SDHA or FH lead to neurodegeneration, without cancer, whereas heterozygous
germline mutations in three out of four SDHD genes
or FH lead to cancer susceptibility is at first puzzling.
This apparent paradox deepens when we consider the
observation that, at the cellular level, loss of the
remaining wild-type allele occurs in component neoplasias that have germline heterozygous mutations in
the respective genes. In other words, why does complete or near-complete loss of enzymatic function in
SDHA and FH lead to neurodegeneration, yet a similarly complete loss of enzymatic function effected by
germline heterozygous mutation followed by somatic
loss of the wild-type allele in the transforming cells
result in cancer? When this conundrum is considered
at the organismal, organ and cellular levels, several
plausible explanations begin to emerge.
At the organismal and organ levels, homozygous
mutations in either SDHA or FH lead to complete or
near-complete loss of activity (see above) in every cell
of the body throughout development and shortly
after birth — crucial times in neurological development. Given the particularly high oxygen and energy
demands in the developing nervous system, it
becomes plausible that this chronic lack of oxygen
and energy would lead to severe neurological dysfunction. That Leigh syndrome and fumarase deficiency are consequences of a nervous system that is
starved of energy are supported by the observation
that mutations in both the mitochondrial-encoded
enzymes and non-complex-II respiratory-chain components also result in similarly devastating neurological manifestations. These genes include NDUFS4
(REF. 38) , NDUFS8 (REF. 39) and NDUFS7 (REF. 40)
(which are all part of complex I), as well as SURF1
(REF. 41) and COX10 (REF. 42), which comprise complex
IV. Heterozygous germline mutations in SDH or FH,
by contrast, only reduce gene dosages. These reduced
dosages are adequate for normal neurological development, as heterozygous individuals do not have
neurological deficits.
Box 3 | Mitochondria mediate energy-dependent and energy-independent apoptosis
Energy-independent apoptosis is mediated by oxygen
Mitochondrial membrane
(O2) free radicals. When mitochondrial function is
impaired, severe energy deficits occur and large amounts
Mitochondrial
of oxygen free radicals are generated. When
membrane
disruption
mitochondria sense the presence of oxygen free radicals
(FH, SDHX mutation)
(hypoxia), hypoxia-inducible factors (HIFs), such as
HIF-1, are activated and translocated to the nucleus
Complex II disassembly
where they induce gene expression. HIF-1 gene targets
encode proteins that promote cellular proliferation or
FH
prevent apoptosis, leading to neoplasia. This might
Procaspases
SDH
explain the cancer predisposition that are observed in
APAF1
Cytosol
patients with hereditary leiomyomatosis and type 2
Cytochrome c
papillary renal-cell carcinoma and inherited
phaeochromocytoma–paraganglioma syndrome. Some
Energy
Nucleus
depletion
HIF-1 target genes are pro-apoptotic, such as TP53,
which activates cell death. This might explain the
Caspases
neurodegeneration that is observed as a result of SDHA
or fumarase deficiency. This mechanism of neuronal-cell
Induction of
anti-apoptotic,
damage and death might lead to the manifestations of
O2 free
Apoptosis
pro-proliferative
radicals
Leigh syndrome and fumarase deficiency.
genes (TGFβ,
Conversely, it is plausible that heterozygous mutations
PDGF, EGFR)
Hypoxia
HIF
HIF
in SDHX, which leads to partial complex assembly or
complete complex II disassembly, result in sufficient
alteration in membrane composition to allow for some
resistance to apoptosis and might explain, at least partially, cancer predisposition. It is also plausible that mutant forms of
SHD that do not insert in the mitochondrial membrane have anti-apoptotic activity.
Another caspase-independent, energy-independent pathway (not shown) of apoptosis is mediated by apoptosisinducing factor (AIF) and endonuclease-G (endoG), which are believed to be released from the mitochondrial intermembrane on apoptosis induction and translocate to the cytosol and then the nucleus43,44. Once AIF reaches the nucleus,
it causes peripheral-chromatin condensation and DNA fragmentation. Unlike AIF, endoG is believed to require other
cofactors, perhaps other DNases, to mediate DNA degradation and apoptosis.
Energy-dependent apoptosis involves the activity of caspases. When the mitochondrial membrane is breached or its
composition altered, the proteins that reside between the inner and outer mitochondrial membranes, such as
procaspases, apoptosis-inducing factor (APAF1) and cytochrome c, are released. Pro-caspases are then cleaved to caspases
in an energy-dependent reaction, and these enzymes mediate apoptosis.
NATURE REVIEWS | C ANCER
VOLUME 3 | MARCH 2003 | 1 9 9
REVIEWS
For example, whereas a child with homozygous FH
mutations developed neurological disorders, the
child’s heterozygous mother developed only HLRCC,
without neurological signs or symptoms19. One explanation for this is that in cases of inherited cancer syndromes, somatic loss of the respective remaining allele
occurs later in life, and these cells then become transformed. But the mechanism by which the eventual loss
of SDH and FH activity leads to neoplasia is not
understood. One explanation could be that individuals
suffering from severe neurological dysfunction due to
homozygous SDHA or FH mutations die in childhood
or early teens, before the age in which heterozygous
SDH- and FH-related tumours develop. Although this
might be true for patients with FH, who develop
tumours in their mid-20s, the youngest ages at which
heterozygous SDH-related tumours develop can be as
young as 5 years old (average 25 years)10.
It is possible that energy depletion or free-radical
generation triggered by mitochondrial dysfunction
underlie the cellular basis of neurodegeneration and
tumorigenesis. A paradox seems to exist regarding the
ability of mitochondria and free radicals to both promote and prevent apoptosis. Partial or complete loss of
SDH or FH activity leads to energy depletion and freeradical formation. Mitochondrial-mediated apoptosis
can be energy dependent or independent (BOX 3). It is
conceivable that partial decreases in activity, as would
be the case in the heterozygous cancer-predisposing
states, would lead to chronic, low-level reductions in
energy production, which are insufficient to cause overt
symptoms but could contribute to inefficient energydependent apoptosis43,44. The subsequent loss of the
wild-type allele in the cell could lead to complete loss of
energy-dependent apoptosis.
The integrity of the mitochondrial outer membrane
is also crucial to maintaining cellular homeostasis. Once
the membrane is breached or undergoes a change in
composition, an energy-independent apoptotic cascade
occurs that involves release of cytochrome c and procaspases43,44 (BOX 3). Presumably, if SDH components disassemble or become disorganized within the membrane,
the membrane composition might be altered, and the
cell might be able to resist apoptosis. With reductions in
complex II or FH function, free-radical production also
increases and is sensed by the mitochondria as hypoxia45.
This leads to stabilization of HIF-1, its translocation to
the nucleus and activation of its target genes45. The HIF-1
gene targets include transforming growth-factor-β
(TGFβ), platelet-derived growth-factor-receptor-β
(PDGFβ) and a ligand for the epidermal growth-factor
receptor (EGFR) — all of which have oncogenic properties and induce cellular proliferation. Furthermore, elevated ROS levels that are secondary to dysfunctional
mitochondria often lead to activation of other signaltransduction pathways, such as the mitogen-activated
protein kinase (MAPK) pathway, which is central to
tumorigenesis. Together with oncogenic-pathway activation, elevated ROS levels result in resistance to apoptosis
— probably due to more effective antioxidant defences
— and also contribute to neoplasia.
200
| MARCH 2003 | VOLUME 3
What of neurodegeneration? Unfortunately, there
is little data to explain the precise mechanism that
leads to neurodegeneration in Leigh syndrome or
fumarase deficiency. It is believed that when oxidative
phosphorylation is shut down, large quantities of ROS
are produced. In this situation, peroxidation of the
mitochondrial membrane occurs and apoptosis
ensues, as a result of membrane breakdown (BOX 3).
Crosstalk between the HIF-1 and other pathways
might also occur. Reactive oxygen species such as
nitric oxide and carbon monoxide prevent HIF-1
induction and binding to DNA. In cases of severe
hypoxia or anoxia, which can occur during Leigh
syndrome or fumarase deficiency, HIF-1 can induce
expression of the pro-apoptotic genes, such as TP53
(REF. 45), which mediate neuronal-cell death. So, the
yin and yang of the ‘life-giving’ energy-producing side
and the ‘death-promoting’ side of mitochondria, gene
dosage, timing and exposure time to loss-of-function
as well as tissue sensitivities could explain the different
phenotypic manifestations of complex II or FH dysfunction. However, how the cell decides which route
to take is still unknown.
Finally, it is plausible that beyond their traditional
enzymatic functions in the electron-transport chain
and Krebs cycle, SDH and FH might have other independent functions. A good model that might shed
some light on this possibility comes from studies of the
apoptosis-inducing factor (AIF). AIF is believed to
induce nuclear apoptosis independent of the caspases
(BOX 3). In addition to its apoptotic function, AIF can
independently catalyse the reduction of cytochrome c
in the presence of NADH46, and stably bind FAD. So,
AIF can be considered to be a flavoprotein that has oxydoreductase activity — properties that are similar to
those of SDH. By analogy, it is also possible that SDH
can act independently of its energy-generating functions to mediate apoptosis directly. In this regard, given
that most heterozygous mutations in SDHB, SDHC
and SDHD prevent the entire complex from assembling within the mitochondrial membranes, or disrupt
assembly, it is possible that it is the non-membraneassociated/cytosolic, or the truncated SDH subunits,
that are released and have anti-apoptotic activities.
These hypotheses need to be empirically investigated.
Future directions
The unexpected association between two nuclearencoded mitochondrial enzymes, FH and SDH, and
human cancer should inspire further research into the
genetic, cellular and clinical aspects of mitochondrial
function. Although the traditional activities of the electron-transport chain and Krebs cycle seemed rudimentary, and are first learned at the secondary school level,
there is a clear lack of understanding about how mitochondrial deficiencies lead to tumour development.
Although SDH and FH are known to act in tandem in
electron transport and the Krebs cycle (BOX 1), and are
both ubiquitously expressed, the cancer syndromes
that develop when their function is disrupted
(phaeochromocytoma–paraganglioma syndromes and
www.nature.com/reviews/cancer
REVIEWS
HYPERBARIC OXYGEN
Oxygen that is delivered at high
tension/pressure, which is well
above atmospheric oxygen
tension at sea level (defined as
1 atm).
HLRCC/MCL) are quite different. Because of the association of mitochondria with energy production and
apoptosis, hypotheses involving free radicals, hypoxia,
angiogenesis, cell damage and apoptosis are invoked
(BOX 3). Apart from scant data, little else is known.
So can we exploit our knowledge of SDH and FH
for routine clinical application? In the case of SDH, in
addition to molecular diagnosis and predictive testing
in known phaeochromocytoma–paraganglioma families, recent data indicate that genetic analysis for
mutations in SDHX genes should be performed routinely on patients that develop phaeochromocytomas
and paragangliomas10,23. Such evidence-based changes
in medical practice will improve diagnosis, prognostic
analysis and prophylactic surgery decisions for the
patients involved and their families. Although
HLRCC/MCL families are rare, the challenge for using
routine FH clinical testing is to delineate the extent of
such germline mutations in individuals with uterine
leiomyomas. Uterine leiomyomas are one of the most
common tumours in humans, and although benign,
require a tremendous amount of health-care
resources. So, mechanistic understanding of FH’s role
in neoplasia might provide important insights into
medical management either by early diagnosis,
prevention and/or therapy.
Examination of gene–gene and gene–environment
interactions might also reveal new preventative measures.
For instance, studies have shown that individuals living at
high altitude, such as the Andes, have a higher incidence
of carotid-body paragangliomas, presumably as a result
of chronic (relative) hypoxia associated with neoplastic
1.
Latif, F. et al. Identification of the von Hippel–Lindau disease
tumor suppressor gene. Science 260, 1317–1320 (1993).
2. Mulligan, L. M. et al. Specific mutations of the RET protooncogene are related to disease phenotype in MEN 2A and
FMTC. Nature Genet. 6, 70–74 (1994).
3. Eng, C. et al. The relationship between specific RET protooncogene mutations and disease phenotype in multiple
endocrine neoplasia type 2: International RET Mutation
Consortium analysis. JAMA 276, 1575–1579 (1996).
4. Kroll, A. J., Alexander, B., Cochios, F. & Pechet, L.
Hereditary deficiencies of clotting factors VII and X
associated with carotid-body tumors. N. Engl. J. Med. 270,
6–13 (1964).
5. van der Mey, A. G., Maaswinkel-Mooy, P. D.,
Cornelisse, C. J., Schmidt, P. H. & van de Kamp, J. J.
Genomic imprinting in hereditary glomus tumours: evidence
for new genetic theory. Lancet 2, 1291–1294 (1989).
6. Heutink, P. et al. A gene subject to genomic imprinting and
responsible for hereditary paragangliomas maps to 11q23qter. Hum. Mol. Genet. 1, 7–10 (1992).
7. Heutink, P., van Schothorst, E. M. & van der Mey, A. G. L.
Further localization of the gene for hereditary
paragangliomas and evidence for linkage in unrelated
families. Eur. J. Hum. Genet. 2, 148–158 (1994).
8. Baysal, B. E. et al. Mutations in SDHD, a mitochondrial
complex II gene, in hereditary paraganglioma. Science 287,
848–851 (2000).
Identification of SDHD as the susceptiblity gene for
11q-linked hereditary paraganglioma families. SDHD
is the first nuclear gene that encodes a mitochondrial
component and that has been shown to be involved in
a cancer-susceptiblity syndrome.
9. Gimm, O., Armanios, M., Dziema, H., Neumann, H. P. H. &
Eng, C. Somatic and occult germline mutations in SDHD,
a mitochondrial complex II gene, in non-familial
pheochromocytomas. Cancer Res. 60, 6822–6825 (2000).
10. Neumann, H. P. H. et al. Germ-line mutations in
nonsyndromic pheochromocytoma. N. Engl. J. Med. 346,
1459–1466 (2002).
NATURE REVIEWS | C ANCER
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
21.
enlargement of the carotid body 33. Therefore, should
individuals who are found to carry germline SDHX
mutations be advised to migrate to lower altitudes or be
exposed to HYPERBARIC OXYGEN. Iron is necessary for the
proper functioning of the iron–sulphur moiety that is
integral to complex II, which, in turn, affects the function
of fumarase (BOX 1). Would supplementation with elemental iron prevent neoplasia in individuals with heterozygous germline SDHX or FH mutations? Along the
same lines, could addition of increased oxygen tension, in
addition to iron supplementation, also prevent disease?
Another possibility is to provide fumarate supplements to
individuals with heterozygous SDHX mutations, and
malate supplements to those with HLRCC/MCL, to
prevent cancer.
Finally, the recent associations of inherited cancer
syndromes with neurological disorders, both caused by
germline defects in the same genes, indicate an interesting
bioinformatics exercise for new cancer-predisposition
gene discovery. Germline homozygous mutations in
many of the nuclear-encoded mitochondrial components have been delineated. It would be ideal if
the gene-expression profiles of patients with SHDX/
phaeochromocytoma–paraganglioma syndromes/
Leigh syndrome and FH/HLRCC-MCL/fumarase
deficiency could be studied, to determine common and
distinct profiles in inherited neoplasia compared to
neurological disease. This might be extrapolated to sporadic neoplasias as well, given that mitochondrial
genome variation has been described both in neuromuscular disorders and somatically in some sporadic
solid tumours (BOXES 1 and 2).
Population-based study showing that 25% of unrelated
non-syndromic, non-familial phaeochromocytoma
cases are due to germline mutations in one of four
genes, including SDHD and SDHB.
Baysal, B. E. et al. Prevalence of SDHB, SDHC and SDHD
in clinic patients with head and neck paragangliomas.
J. Med. Genet. 39, 178–183 (2002).
Niemann, S. & Muller, U. Mutations in SDHC cause autosomal
dominant paraganglioma. Nature Genet. 26, 141–150 (2000).
Astuti, D. et al. Mutations in the mitochondrial complex II subunit
SDHB cause susceptibility to familial paraganglioma and
pheochromocytoma. Am. J. Hum. Genet. 69, 49–54 (2001).
Identification of SDHB as the susceptibility gene for
familial paraganglioma and/or phaeochromocytoma.
Baysal, B. E. Hereditary paraganglioma targets diverse
paranglia. J. Med. Genet. 39, 617–622 (2002).
Kiuru, M. et al. Familial cutaneous leiomyomatosis is a two-hit
condition associated with renal cell cancer of characteristic
histopathology. Am. J. Pathol. 159, 825–829 (2001).
Launonen, V. et al. Inherited susceptibility to uterine
leiomyomas and renal cell cancer. Proc. Natl Acad. Sci. USA
98, 3387–3392 (2001).
Delahunt, B. & Eble, J. N. Renal cell neoplasia. Pathology
34, 13–20 (2002).
Kiuru, M. et al. Few FH mutations in sporadic counterparts of
tumor types observed in hereditary leiomyomatosis and renal
cell cancer families. Cancer Res. 62, 4554–4557 (2002).
Tomlinson, I. P. M. T. et al. Germline mutations in the
fumarate hydratase gene predispose to dominantly inherited
uterine fibroids, skin leiomyomata and renal cell cancer.
Nature Genet. 30, 406–410 (2002).
Identification of germline heterozygous mutations in
FH that cause susceptibility to HLRCC.
Alam, N. A. et al. Localization of a gene (MCUL1) for multiple
cutaneous leiomyomata and uterine fibroids to chromosome
1q42.3-q42. Am. J. Hum. Genet. 68, 1264–1269 (2001).
Gimenez-Roqueplo, A.-P. et al. The R22X mutation of the
SDHD gene in hereditary paraganglioma abolishes
enzymatic activity of the complex II mitochondrial respiratory
22.
23.
24.
25.
26.
27.
28.
29.
30.
31.
chain and activates the hypoxia pathway. Am. J. Hum.
Genet. 69, 1186–1197 (2001).
First demonstration of loss of enzymatic activity as a
consequence of SDHD mutation and consequent
alterations in molecules in the HIF pathway.
Gimenez-Roqueplo, A. P. et al. Functional consequences of
a SDHB gene mutation in an apparently sporadic
pheochromocytoma. J. Clin. Endocrinol. Metab. 87,
4771–4774 (2002).
Maher, E. R. & Eng, C. The pressure rises: update on the
genetics of phaeochromocytoma. Hum. Mol. Genet. 11,
2347–2354 (2002).
Schapira, A. H. Mitochondrial dysfunction in neurodegenerative
disorders. Biochim. Biophys. Acta 1366, 225–233 (1998).
Leigh, D. Subacute necrotizing encephalomyelopathy in an
infant. J. Neurol. Neurosurg. Psychiatr. 14, 216–221
(1951).
Bourgeron, T. et al. Mutation of the fumarase gene in two
siblings with progressive encephalopathy and fumarase
deficiency. J. Clin. Invest. 93, 2514–2518 (1994).
Parfait, B. et al. Compound heterozygous mutations in the
flavoprotein gene of the respiratory chain complex II in a
patient with Leigh syndrome. Hum. Genet. 106, 236–243
(2000).
Ackrell, B. A. C. Progress in understanding structurefunction relationships in respiratory chain complex II. FEBS
Lett. 466, 1–5 (2000).
Oyedotun, K. S. & Lemire, B. D. The quinone-binding
sites of the Saccharomyces cerevisiae succinateubiquinone oxidoreductase. J. Biol. Chem. 276,
16936–16943 (2001).
Iverson, T. M., Luna-Chavez, C., Cecchini, G. & Rees, D. C.
Structure of the Escherichia coli fumarate reductase
complex. Science 284, 1961–1966 (1999).
Vinogradov, A. D., Ackrell, B. A. C. & Singer, T. P. On the
possible interrelations of the reactivity of soluble succinate
dehydrogenase with ferricyanide, reconstitution activity and
the Hipip iron-sulfur center. Biochem. Biophys. Res. Comm.
67, 803–809 (1975).
VOLUME 3 | MARCH 2003 | 2 0 1
REVIEWS
32. Schmidt, D. M., Saghbini, M. & Scheffler, I. E. The
C-terminus of the succinate dehydrogenase IP peptide of
the Saccharomyces cerevisiae is signficiant for assembly of
complex II. Biochemistry 31, 8442–8448 (1992).
33. Arias-Sella, J. & Valcarcel, J. Chief cell hyperplasia in the
human carotid body at high altitudes; physiologic and
pathologic significance. Hum. Pathol. 7, 361–373 (1976).
34. Chandel, N. S. et al. Mitochondrial reactive oxygen species
trigger hypoxia-induced transcription. Proc. Natl Acad. Sci.
USA 95, 11715–11720 (1998).
35. Clifford, S. C. et al. Contrasting effects on HIF-1α regulation
by disease-causing pVHL mutations correlate with patterns
of tumourigenesis in von Hippel–Lindau disease. Hum. Mol.
Genet. 10, 1029–1038 (2001).
36. Hoffman, M. A. et al. von Hippel–Lindau protein mutants
linked to type 2C VHL disease preserve the ability to
downregulate HIF. Hum. Mol. Genet. 10, 1019–1027
(2001).
37. Teipel, J. W. & Hunt, R. L. The subunit interactions of
fumarase. J. Biol. Chem. 246, 4859–4865 (1971).
38. van den Heuvel, L. et al. Demonstration of a new pathogenic
mutation in human complex I deficiency: a 5-bp duplication
in the nuclear gene encoding the 18kD (AQDQ) subunit. Am.
J. Hum. Genet. 62, 262–268 (1998).
39. Loeffen, J. et al. The first nuclear-encoded complex I
mutation in a patient with Leigh syndrome. Am. J. Hum.
Genet. 63, 1598–1608 (1998).
40. Triepels, R. H., van den Heuvel, L. & Loeffen, J. Leigh
syndrome associated with a mutation in the NDUFS7
(PSST) nuclear encoded subunit of complex I. Ann. Neurol.
45, 787–790 (1999).
41. Tiranti, V. et al. Mutations of SURF-1 in Leigh disease
associated with cytochrome c oxidase deficiency. Am. J.
Hum. Genet. 63, 1609–1621 (1998).
42. Valnot, I., von Kleist-Retzow, J. C. & Barrientos, A.
A mutation in the human heme A:farnesyltransferrase gene
(COX 10) causes cytochrome c oxidase deficiency. Hum.
Mol. Genet. 9, 1245–1249 (2000).
43. van Loo, G. et al. The role of mitochondrial factors in
apoptosis: a Russian roulette with more than one bullet. Cell
Death Differ. 9, 1031–1042 (2002).
44. Ravagnan, L., Roumier, T. & Kroemer, G. Mitochondria, the
killer organelles and their weapons. J. Cell. Physiol. 192,
131–137 (2002).
45. López-Barneo, J., Pardal, R. & Ortega-Sáez, P. Cellular
mechanisms of oxygen sensing. Annu. Rev. Physiol. 63,
259–287 (2001).
202
| MARCH 2003 | VOLUME 3
46. Miramar, M. D. et al. NADH-oxidase activity of mitochondrial
apoptosis inducing factor (AIF). J. Biol. Chem. 276,
16391–16398 (2001).
47. Scheulke, M., Smeitink, J. & Mariman, E. Mutant NDUFV1
subunit of mitochondrial complex I causes leukodystrophy
and myoclonic epilepsy. Nature Genet. 21, 260–261
(1999).
48. Benit, P. et al. Large-scale deletion and point mutations of the
nuclear NDUFV1 and NDUFS1 genes in mitochondrial complex
I deficiency. Am. J. Hum. Genet. 68, 1344–1352 (2001).
49. Loeffen, J. et al. Mutations in the complex I NDUFS2 gene of
patients with cardiomyopathy and encphalomyopathy. Am.
J. Hum. Genet. 49, 195–201 (2001).
50. de Lonlay, P. et al. A mutant mitochondrial respiratory chain
assembly protein causes complex III deficiency in patients
with tubulopathy, encephalopathy and liver failure. Nature
Genet. 29, 57–60 (2001).
51. Papadopoulou, L. C., Sue, C. M. & Davidson, M. Fatal
infantile cardioencephalomyopathy with cytochrome c
oxidase (COX) deficiency due to mutations in SCO2, a
human COX assembly gene. Nature Genet. 23, 333–337
(1999).
52. Valnot, I., Ormond, S. & Gigarel, N. Mutations of the SCO1
gene in mitochondrial cytochrome c oxidase (COX)
deficiency with neonatal-onset hepatic failure and
encephalopathy. Am. J. Hum. Genet. 67, 1104–1109
(2000).
53. Servidei, S. Mitochondrial encephalomyopathies: gene
mutation. Neuromuscul. Disord. 12, 224–229
(2001).
54. Green, D. R. & Reed, J. C. Mitochondria and apoptosis.
Science 281, 1309–1312 (1998).
55. Yeh, J. J. et al. Two-dimensional gene scanning of the
mitochondrial genome reveals somatic mutations in papillary
thyroid carcinomas and multiple sequence variants in cases
with sporadic thyroid tumors. Oncogene 19, 2060–2066
(2000).
56. Polyak, K. et al. Somatic mutations of the mitochondrial
genome in human colorectal tumours. Nature Genet. 20,
291–293 (1998).
57. Jeronimo, C. et al. Mitochondrial mutations in early state
prostate cancer and bodily fluids. Oncogene 20, 5195–5198
(2001).
58. Kirches, E. et al. High frequency of mitochondrial DNA
mutations in glioblastoma multiforme identified by direct
sequence comparison to blood samples. Int. J. Cancer 93,
534–538 (2001).
59. Liu, V. W. et al. High incidence of somatic mitochondrial
DNA mutations in human ovarian carcinomas. Cancer Res.
61, 5998–6001 (2001).
60. Maximo, V. et al. Microsatellite instability, mitochondrial DNA
large deletions, and mitochondrial DNA mutations in gastric
carcinoma. Gene Chromosom. Cancer 32, 136–143
(2001).
61. Parrella, P. et al. Detection of mitochondrial DNA mutations
in primary breast cancer and fine-needle aspirates. Cancer
Res. 61, 7623–7626 (2001).
Acknowledgements
C. E. is a recipient of the Doris Duke Distinguished Clinical
Scientist Award and is supported by grants from the National
Institutes of Health, National Cancer Institute, American Cancer
Society, US Department of Defense Breast and Prostate Cancer
Research Programs, Susan G. Komen Breast Cancer Research
Foundation and Jimmy V Golf Classic Translational Cancer
Research Award from the V Foundation. M. K. is supported by
the Finnish Cancer Society, Duodecim, Kidney Foundation, Paulo
Foundation, Maud Kuistila Foundation, Ida Montin Foundation,
Finnish Oncology Foundation, Research and Science Foundation
of Pharmos, and AstraZeneca. L. A. A. is supported by the
Finnish Cancer Society, Sigrid Juselius Foundation, Helsinki
Central Hospital and the Academy of Finland’s Center of
Excellence Award.
Online links
DATABASES
The following terms in this article are linked online to:
Cancer.gov: http://www.cancer.gov/cancer_information/
head and neck tumours | kidney cancer
LocusLink: http://www.ncbi.nih.gov/LocusLink/
AIF | BCS1 | COX10 | FH | HIF-1α | HIF-1β | NDUFS4 | NDUFS7 |
NDUFS8 | NDUFV1 | PGL1 | PGL2 | SCO1 | SCO2 | SDHA |
SDHB | SDHC | SDHD | SURF1 | VHL
OMIM: http://www.ncbi.nlm.nih.gov/Omim/
Friedrich ataxia | HLRCC | Huntingdon’s disease | Leigh syndrome |
MCL | paraganglioma | Parkinson’s disease |
phaeochromocytoma
FURTHER INFORMATION
Medical Genetics Information Resource:
www.geneclinics.org
Access to this interactive links box is free online.
www.nature.com/reviews/cancer