Mitochondria1 DNA Mutation Associated with Aging and Degenerative Disease‘ PHILLIP NAGLEY,b IAN R: MACKAY, ALESSANDRA BAUMER, RONALD J. MAXWELL, FRANCOIS VAILLANT, ZHONG-XIONG WANG, CHUNFANG ZHANG, AND ANTHONY W. LINNANE Centre for Molecular Biology and Medicine, and Department of Biochemistry Monash University Clayton, Victoria 3168, Australia There has been sustained interest in somatic genetic error as a basis for aging. The concept was elaborated some 30 years ago’ and was directed essentially to the nuclear genome. Other authors did, however, give consideration to a possible role for mitochondria1 genomes in aging.2 Recently, we have developed a detailed proposal for mutations in mitochondrial DNA (mtDNA) as a cause of degenerative diseases and aging, based on considerations of the molecular biology and genetics of mit~chondria.~ The relevant features include the following: (i) the rate of mtDNA mutation is 10-100-fold greater than that of nuclear DNA; (ii) mtDNA is naked in lacking histones; (iii) there is very little repair activity for mtDNA mutations; (iv) intracellular energy systems have high dependence on enzymes encoded by mtDNA. Further, mtDNA would be especially vulnerable to damage due to oxygen free radicals that are especially generated in this respiratory organelle.* On the other hand, this vulnerability of mtDNA to mutagenic error is balanced by the “back-up” provided by the high number of mtDNA genomes per cell. Even so, the present assumption is that error-damaged mtDNA will accumulate in cells throughout life with, eventually, a lethal effect on the This article discusses recent data that substantiate this view. MOLECULAR BIOLOGY OF MITOCHONDRIA Mitochondria are semiautomous cell organelles. Most of the mitochondrial proteins are encoded by the nuclear chromosomes, synthesized in the cytosol, and subsequently imported into the organelle. The circular mtDNA genome, 16.6 kilobases (kb) in humans, is exclusively concerned with the oxidative phosphorylation system of the organelle.6 Mammalian mtDNA encodes 13
a This work was supported by the National Heart Foundation of Australia and the National Health and Medical Research Council of Australia. Address for correspondence: Professor P. Nagley, Department of Biochemistry, Monash University, Clayton, Victoria 3 168, Australia.
92
NAGLEY
et a/.: MITOCHONDRIAL
DNA MUTATION
93
proteins that are subunits of 3 respiratory complexes (I, 111, and IV) and ATP synthase, and specifies 22 tRNAs and 2 rRNAs of the mitochondrial protein synthesizing system. There are some 1,OOO mtDNA molecules in each somatic cell, with many fewer per individual organelle. A population of identical mtDNAs in one cell is termed homoplasmic. Cells, and individual mitochondria, can maintain a heteroplasmic state in which there is coexistence of mtDNA molecules differing in base sequence, being either of the same or different lengths; this is the basis of the tissue energy mosaic discussed below. Mitochondria1 genetics is based on the analysis of mutants that have modified mitochondrial functions because of molecular changes in the structure of mtDNA. The principles of the molecular genetics of mtDNA have been largely worked out from studies on the unicellular yeast, Saccharomyces cerevisae,’ and are closely reflected in the properties of mammalian mitochondria. The nomenclature of mutants is derived from that adopted for yeast: rho- (rho minus) mtDNA molecules contain gross deletions encompassing one or more genes; the term rhoo (rho zero) denotes that mtDNA is completely lacking from the cell; mit- (mit minus) mutations are localized changes in mtDNA sequence, base substitutions or small deletions or insertions, that affect a protein coding region; syn- (syn minus) mutations resemble the mit- type, but they primarily affect the expression of mtDNA, such as changes in tRNA or rRNA genes. MITOCHONDRIAL CYTOPATHIES Since the first recognition of human disease due to mitochondrial dysfunction some 20 years ago, various diseases with eponymous or long descriptive designations have been categorized in terms of symptomatology and, more recently, catalogued according to discrete abnormalities of the mitochondrial genome.8 In some there is clear evidence of maternal inheritance, indicative of transmission of a mitochondrial defect via the oocyte, whereas others apparently result from a mitochondrial genetic error presumably acquired during embryogenesis. The nature of the molecular defects8 include rho- deletions, particularly prevalent in myopathies (CPEO, Kearns-Sayre syndrome), mil- mutations (Leber’s optic neuropathy), and syn- mutations (encephalomyopathies: MERRF and MELAS). It is still unclear whether single molecular defects in mtDNA are solely responsible for the cellular and clinical phenotypes or whether accretion of multiple mtDNA lesions in various combinations determines particular disease pattern^.^ Also, participation of chromosomal DNA and the products of nuclear genes in the expression of mitochondrial defects must be considered.8 MITOCHONDRIAL MUTATION, AGING, AND DEGENERATIVE DISEASE
In 1989 Linnane et u I . ~developed an hypothesis on aging based on mutations affecting mtDNA. Their evidence was derived from experience with metabolic deficits resulting from mutant mtDNA in yeast (see above), and observations on diseases in man attributable to mutant mtDNA. This hypothesis elaborated some particular points. First, there is a heteroplasmic segregation of mutant mtDNA, associated with the large number of mtDNA genomes per cell. Over successive cell divisions,
94
ANNALS NEW YORK ACADEMY OF SCIENCES
during which there is replication of mtDNA in the daughter cells, the segregation of cells results in some carrying predominantly normal “wild-type’’ mtDNA, some with predominantly mutated mtDNA, and others with differing amounts of variously mutated mtDNA molecules. Second, horizontal accumulation of mtDNA mutations during the life of an individual occurs at a significantly higher rate than vertical accumulation over successive generations. Various explanations can be offered for this apparent “mtDNA purification” during the reproductive process. One notion is that during mammalian oocyte development, when there is a gross reduction in the effectively heritable number of independent mtDNA genotypes,I0 functional progeny oocytes can be formed that contain a virtually pure population of normal mtDNA genomes. In this view, oocytes that contain a high proportion of defective mtDNA genomes could be functionally selected against, and do not further participate in species reproduction. Third, a tissue energy mosaic will occur as the eventual result of the accumulation of somatic mtDNA mutations. This mosaic can be manifested as a population of cells in a given tissue ranging in bioenergetic capacity from normal to partially or grossly defective. The accumulation over time of bioenergetically defective cells was envisaged as a key factor in the process of aging.3
HISTOLOGY THE BIOENERGETIC MOSAIC A bioenergetic mosaic in tissues is schematically represented in FIGURE 1. Mitochondria accumulate a load of mutant mtDNA genomes that increases with age and becomes randomly distributed among the cells of the tissue. This random distribution of mtDNA results in some cellular progeny being fully energy competent and others completely energy deficient. This is illustrated by a histochemical study” of human cardiac tissue, from young and old individuals, stained for the key mitochondrial respiratory enzyme, cytochrome c oxidase. Nearly all specimens in early youth stain evenly and intensely across all cardiomyocytes, whereas specimens from the aged stain as a tissue mosaic with juxtaposed regions of strong, weaker, and absent staining. The study indicated an age-related loss of mitochondrial function in increasingly numerous cardiomyocytes distributed throughout the tissue.” Similar results have been found in our laboratory in young and old individuals, including humans and rats, illustrated in FIGURES 2 and 3. FIGURE 2 shows the results of staining for cytochrome c oxidase on cardiac tissue on an old individual after death from causes unrelated to mitochondrial disease; the tissue mosaic is indicated by regions of grossly depleted cytochrome c oxidase activity. FIGURE 3A shows the renal cortex from a young rat, within which all cells of the proximal tubular epithelium are reactive for cytochrome c oxidase, whereas in FIGURE 3B there is obvious loss of cytochrome c oxidase staining in much of the equivalent tissue of the aged rat. These changes are ascribed to defective mtDNA molecules that are differentially accumulated among the cells of particular tissues, and represent the tissue energy mosaic associated with aging. Such focal histological changes are characteristically demonstrable among muscle fibers of patients with gross and symptomatic mitochondrial myopathies.8 They represent in this case the widespread, but not uniform, distribution of defective mtDNA genornes that account for a much higher proportion of total mtDNA in the patient’s tissues than occurs in normal aging.
NAGLEY et al.: MITOCHONDRIAL DNA MUTATION
95
DNA ANALYSIS: AGE-RELATED MUTANT mtDNA In certain of the symptomatic mitochondria1 cytopathies there may be a sufficient load of mtDNA deletions to allow detection by a simple Southern blot analysis of tissue DNA. The much smaller load of mutant mtDNA, postulated to accumulate during aging, requires amplification of mtDNA by the polymerase chain reaction (PCR).
Nucleus
G P i
Normal fully energy competent mitochondria
I
Fully energy competent
Mitochondrlal DNA mutatlons occur and accumulate In a cell with tlme
Varylng de rees of energy de$idency
Corn let0 energy Lficiency
TISSUE ENERGY MOSAIC
FIGURE 1. Schema of generation of tissue energy mosaic during aging. Damage to mtDNA is considered to accumulate during aging. Segregation of normal and mutant mtDNA molecules will occur, largely during cell proliferation, to generate tissues in which cells may suffer varying degrees of energy deficits, extending to complete energy deficiency. For simplicity, unfilled ovals represent mitochondria containing damaged mtDNA. Individual mitochondria may themselves contain mixtures of normal and mutant mtDNA molecules. Nuclear genes are considered to be relatively unchanged during this process.
In particular, using the PCR technique, we have searched'* for a specific 4.977kb deletion in individuals not affected by mitochondria1 disease. This deletion arises from illegitimate recombination* between a pair of identical 13-bp sequences separated by almost 5 kb in human mtDNA, and is not only prevalent in myopathy
96
ANNALS NEW YORK ACADEMY OF SCIENCES
patients but also occurs in cells of the substantia nigra in Parkinson's d i ~ e a s e . ' ~ This deletion was clearly demonstrated by us1*to occur in various tissues of individuals over the age of 40 and up to the age of 87 years, but not in infants of
FIGURE 2. Cytochrome c oxidase activity in cardiomyocytes from an aged human individual. Cardiac tissue was taken, after autopsy, from a 72-year-old male with no known cardiomyopathy (motor neuron disease patient; tissue provided by Dr. H. Preston). Staining of sections is based on the use of 3,3'-diaminobenzidine (DAB), which yields an insoluble brown deposit following its reaction with added cytochrome c substrate only when the latter is oxidized by the endogenous mitochondria1 enzyme. NC, normal cardiomyocytes; DC, cytochrome c oxidase-deficient cardiomyocytes. Magnification, 475 x . (Z.-X. Wang and A. W. Linnane, unpublished observations).
3 months and younger at death (FIG. 4). A similar increase in abundance of this 4.977-kb deletion in an age-related manner was also reported by ~ t h e r s . ' ~The ~I~ deleted mtDNA molecules were estimated to represent about 0.1% of total mtDNA
NAGLEY er 01.: MITOCHONDRIAL DNA MUTATION
97
ri 3
m
a
ANNALS NEW YORK ACADEMY OF SCIENCES
98
in adult cardiac tissue,I4 but the distribution of the aberrant shortened mtDNA form among individual cells or zones of the tissues has not been reported. In the mitochondria1 myopathies, deleted mtDNA molecules may represent more than 20% or even as much as 90% of total mtDNA.8 Significantly, within the tissue mosaic regions most extensively depleted of cytochrome oxidase, the molecules of mtDNA carrying the deletions are most highly represented.I6 The technical procedures depend on whole-cell DNA extracts as templates for PCR. The 4.977-kb deletion can be readily detected by use of two oligonucleotide primers for the PCR reaction, denoted L7901 and H13631, each 20 bases long and
1
2
3
4
6
6
7
8
9 1 0 1 1
-
I0 6.76
0.77
FIGURE 4. Detection of mtDNA deletion in tissues from five human subjects of different ages. Tissues were taken at autopsy from subjects with no known mitochondrial disease. The mtDNA fragments were amplified by PCR using oligonucleotide primers L7901 and H I363 1, separated on a I% agarose gel, and photographed after staining in ethidium bromide. Lane 1contains marker DNA (lambda DNA digested with EcoRl and HindlI1); the sizes of the fragments are shown on the left side. Lanes 2 and j: heart and liver of 80-minute-old male; lanes 4 and 5: heart and adrenal gland of 3-month-old female; lanes 6 and 7: brain and left ventricle of 40-year-old male; lanes 8 and 9: left ventricle and kidney of 46-year-old female; lanes 10 and 11: kidney and psoas muscle of 50-year-old male. The sizes of the PCRamplified fragments are indicated on the right side: the 5.75-kb band derives from full-length mtDNA, and the 0.77-kb band contains the 4.977-kb deletion. (Reprinted by permission from Linnane et a/.'?)
representing sequences, respectively, of the mtDNA light and heavy strands. These primers generate a 5.75-kb product from normal full-length mtDNA. Aberrant mtDNA molecules carrying the prevalent deletion from nucleotides 8470 to 13459 on mtDNA will generate a 0.77-kb PCR product, since the deletion is close to 5 kb in its extent. Other primers hybridizing in the vicinity of the primers L7901 and H13631 are used to establish, by primer-shift analyses, the identity of the amplified PCR product as being from the inferred region of mtDNA. Finally, the cycle number is varied to adjust limits of detection. For example, in a 30-cycle PCR with primers L7901 and H 13631, the 0.77-kb deletion product is readily visible
NAGLEY et a/.: MITOCHONDRIAL DNA MUTATION
99
when the template is from adult tissues, but it is not detectable in infant tissues (FIG.4). However, the 0.77-kb band does appear at low intensity with a 60-cycle PCR with DNA templates prepared from infant tissues. IZReliable procedures are needed for quantitative analysis of the 4.977-kb mtDNA deletion to document the progressive accumulation of this mtDNA mutation with age. Our further studies on mtDNA deletions in aged humans have revealed that by use of an extended set of PCR primer pairs, a relatively large number of mtDNA deletions can be detected. For example, one 69-year-old subject without evident mitochondrial disease showed at least 10 different mtDNA deletions, of which some were tissue-specific (Zhang, Baumer, Maxwell, Linnane, and Nagley, unpublished observations). Studies on aged rats in our laboratory suggest a role for mutant mtDNA in agerelated decreases in bioenergetic functions of tissues. Results using PCR indicate the accumulation, in an age-related manner, of one particular deletion of 4.834 kb generated by illegitimate recombination across 16-bp repeat sequences lying across a broadly similar region to that involved in the human 4.977-kb deletion (Boubolas, Maxwell, Linnane, and Nagley, unpublished observations). Histochemical studies on rodent tissues equivalent to those used on human hearts should establish the validity of this aging model.
THE DISTRIBUTION OF mtDNA MUTATION: TISSUE-SPECIFIC OR GENERAL? Mutational effects on mtDNA presumably occur at random in cells among all tissues, but particular tissues may be unduly susceptible to adverse effects of mutationally damaged mtDNA. First, there may be local tissue factors that are mutagenic for mtDNA, with free radicals one obvious candidate? this, in fact provides an attractive combination of the mutational error and free radical theories of Second, there will be a greater relative susceptibility to mtDNA mutation in tissues consisting of nonreplicating versus replicating cells, since in the latter case cells replicating with a high load of mutated mtDNA may be effectively discarded, assuming that functional performance may be demanded for their propagation. In line with this, peripheral blood lymphocytes show little evidence for accumulation of mtDNA deletions as a function of somatic age.I5 Conversely, tissues consisting of cells that are slowly replicating or not dividing at all, such as those of the central nervous system (CNS), cardiac or skeletal muscle, or liver, could suffer in later life from the cumulative effects of mutations of their mtDNA. These cytopathological effects would be exacerbated if mtDNA turnover continues in the absence of cell division, since defective mtDNA molecules would continue to accumulate. Third, cells that are highly energy-dependent would be particularly susceptible to mtDNA mutation, which is borne out by the tissue distribution, mainly to CNS and muscle, of those diseases now attributable entirely to molecular defects in mtDNA. Similarly, we should look to these tissues, CNS and muscle, for primary examples of degenerative diseases of aging that may be attributable to mutant mtDNA. Cumulative mtDNA mutations could, therefore, be relevant to both a general theory of aging and the occurrence of certain tissue-specific degenerative diseases. By analogy with the identified symptomatic mitochondria1 cytopathies, which affect predominantly striated muscle and neural tissues, we could anticipate that age-related mutations in mtDNA would impact more particularly on the same tissues. For example, the mitochondrial deletion described above is demonstrable
100
ANNALS NEW YORK ACADEMY OF SCIENCES
in cells of the substantia nigra in Parkinson’s disease, and certain cardiomyopathies have been related to mutations affecting mtDNA.9 Looking further, perhaps examples of nonatherosclerotic cardiac failure in aging, “presbycardia,” could have a similar explanation.” And finally, could the known decline in athletic performance, beginning at around age 30, be based in part on subtle bioenergetic losses due to mtDNA mutation in skeletal muscle?
FUNCTIONAL CONSIDERATIONS FOR ENERGY-DEPLETED CELLS Critical to the bioenergetic loss and eventual attrition suffered by a given tissue is the ability of the cells to sustain both performance and viability in the face of a decreased or negligible energy contribution by mitochondrial oxidative phosphorylation. The maintenance of an adequate supply of intracellular ATP is not the only factor. Just as important is the ability of cells to reoxidize the load of reduced NADH derived from energy-yielding catabolism. The need for an “electron sink” is illustrated by our studies of human cells that can grow anaerobically, in the absence of mitochondrial function, but only if pyruvate is provided.’* In this example, pyruvate is efficiently reduced to lactate by the enzyme lactate dehydrogenase, regenerating oxidized NAD from the reduced NADH. The capacity of cells to survive with a greater emphasis on anaerobic glycolysis, and the availability of suitable enzymic pathways and the appropriate substrates by which excess reducing capacity can be channeled into an electron sink, will influence the threshold at which individual tissues will functionally “drop out” in the face of accumulating mtDNA damage.
CONCLUSIONS The tissue energy mosaic in aging and the accumulation of mtDNA damage provide a framework for an eventual understanding of age-related deterioration in the biological performance of cells. Future studies should clarify in molecular detail the inferred pathologies derived from mitochondrial defects both in aging, in general, and in various tissue-specific degenerative diseases. In addition, recent evidence for chemical modification of mtDNA during aging, presumably due to free radical damage,’’ provides for an integration of the mutational error and free radical concepts of aging. The likely link between mutational damage to mtDNA and decline in bioenergetic performance of tissues during aging represents a significant new application of molecular gerontology to the problems of human aging.
SUMMARY Previous theories of aging based on somatic mutation neglected mtDNA, which has a high propensity for mutational error. Knowledge of yeast mtDNA mutations and their functional effects, and of human mtDNA mutations identified in the mitochondrial cytopathies, provides for a concept of aging based on the cumulative effect of mutations affecting human mtDNA. An essential feature of this concept is heteroplasmy, representing mixtures of normal and mutant mtDNA at the cellular and mitochondrial level, resulting in a “tissue mosaic” of focal bioenergetic
NAGLEY er al.: MITOCHONDRIAL DNA MUTATION
101
deficits. Direct evidence for the concept is provided by (i) focal loss of staining for mitochondrially encoded enzymes, such as cytochrome c oxidase, in tissues of aged individuals (humans and rats) and (ii) an age-related increase in deletional mutations in m t D N A demonstrable by application of the polymerase chain reaction t o DNA templates from individuals of different ages.
REFERENCES I. 2. 3. 4. 5.
6. 7. 8. 9.
10.
II.
12. 13. 14. 15.
16. 17.
HAYFLICK. L. 1985. Theories of biological aging. Exp. Gerontol. 2 0 145-159. HARMAN, D. 1983. Free radical theory of aging: Consequences of mitochondrial aging. Age 6: 86-94. T. OZAWA& M. TANAKA. 1989. Mitochondrial DNA LINNANE. A. W.. S. MARZUKI, mutations as an important contributor to ageing and degenerative diseases. Lancet i: 642-645. 1990. Mitochondrial mutations may increase oxidative BANDY,B. & A. J . DAVISON. stress: Implications for carcinogenesis and aging? Free Radical Biol. Med. 8: 523539. MIQUEL.J . 1991. An integrated theory of aging as the result of mitochondrial-DNA mutation in differentiated cells. Arch. Gerontol. Geriatr. 12: 99-1 17. ATTARDI, G. & G. SCHATZ.1988. Biogenesis of mitochondria. Annu. Rev. Cell Biol. 4 289-333. LINNANE, A. W. & P. NACLEY.1978. Mitochondrial genetics in perspective: The development of a genetic and physical map of the yeast mitochondrial genome. Plasmid 1: 324-345. J . M. & D. C. WALLACE. 1990. Oxidative phosphorylation diseases: DisorSHOFFNER, ders of two genomes. Adv. Hum. Genet. 19: 267-330. S. SUCIYAMA, H. INO, K. OHNO,K. HATTORI,T. OBAYASHI, OZAWA, T . , M. TANAKA, K. KAWAMURA. Y. NAKANE& K. HASHIBA.1991. Patients T. ITO, H. DECUCHI, with idiopathic cardiomyopathy belong to the same mitochondrial DNA gene family of Parkinson’s disease and mitochondrial encephalomyopathy. Biochem. Biophys. Res. Commun. 1 7 2 518-525. HAuswiRTH. W. W. & P. J . LAiPis. 1985. Transmission genetics of mammalian mitochondria: A molecular model and experimental evidence. I n Achievements and Perspectives of Mitochondrial Research, Vol. 2: Biogenesis. E. Quagliariello, E. C. Slater, F. Palmieri, C. Saccone & A. M. Kroon. Eds.: 49-59. Elsevier. Amsterdam. J . 1989. Cytochrome-c-oxidase deficient cardiomyocytes in the huMULLER-HOCKER, man heart-an age-related phenomenon: A histochemical ultracytochemical study. Am. J . Pathol. 134: 1167-1 173. R. J . MAXWELL,H. PRESTON, C. ZHANG&S.MARZUKI. L I N N A N A. E , W.. A. BAUMER, 1990. Mitochondrial gene mutation: The ageing process and degenerative diseases. Biochem. Int. 22: 1067-1076. IKEBE,S., M. TANAKA, K. OHNO,W. SATO,K. HATTORI. T. KONDO,Y. MIZUNO & T. OZAWA.1990. Increase of deleted mitochondrial DNA in the striatum in Parkinson’s disease and senescence. Biochem. Biophys. Res. Commun. 170: 1044-1048. G. A. & N. ARNHEIM. 1990. Detection of a specific mitochondrial DNA CORTOPASSI, deletion in tissues of older humans. Nucleic Acids Res. 18: 6927-6933. YEN,T.-C., J.-H. Su, K.-L. KING& Y.-H. WEI. 1991. Ageing-associated 5 kb deletion in human liver mitochondrial DNA. Biochem. Biophys. Res. Commun. 178: 124131. E. A., G. KARPATI & K. E. M. HASTINGS. Deletion mutants are functionSHOUBRIDCE, ally dominant over wild-type mitochondrial genomes in skeletal muscle fiber segments in mitochondrial disease. Cell 62: 43-49. T. OBAYASHI, T. ITO, T. SATAKE,Y. HATTORI.K., M. TANAKA.S. SUCIYAMA, & T. OZAWA.1991. Age-dependent increase in deleted H A N A KJI. ,ASAI.M. NAGANO mitochondrial DNA in the human heart: Possible contributory factor to presbycardia. Am. Heart J . 121: 1735-1742.
102
ANNALS NEW YORK ACADEMY OF SCIENCES
18. VAILLANT, F., B. E. LOVELAND, P. NAGLEY & A . W. LINNANE. 1991. Some biochemical properties of human lymphoblastoid Namalwa cells grown anaerobically. Biochem. Int. 23: 571-580. 19. HAYAKAWA, M . , K . TORII,S. SUGIYAMA, M. TANAKA& T. OZAWA.1991. Ageassociated accumulation of 8-hydroxydeoxyguanosine in mitochondria1 DNA of human diaphragm. Biochem. Biophys. Res. Commun. 179 1023-1029.
a This work was supported by the National Heart Foundation of Australia and the National Health and Medical Research Council of Australia. Address for correspondence: Professor P. Nagley, Department of Biochemistry, Monash University, Clayton, Victoria 3 168, Australia.
92
NAGLEY
et a/.: MITOCHONDRIAL
DNA MUTATION
93
proteins that are subunits of 3 respiratory complexes (I, 111, and IV) and ATP synthase, and specifies 22 tRNAs and 2 rRNAs of the mitochondrial protein synthesizing system. There are some 1,OOO mtDNA molecules in each somatic cell, with many fewer per individual organelle. A population of identical mtDNAs in one cell is termed homoplasmic. Cells, and individual mitochondria, can maintain a heteroplasmic state in which there is coexistence of mtDNA molecules differing in base sequence, being either of the same or different lengths; this is the basis of the tissue energy mosaic discussed below. Mitochondria1 genetics is based on the analysis of mutants that have modified mitochondrial functions because of molecular changes in the structure of mtDNA. The principles of the molecular genetics of mtDNA have been largely worked out from studies on the unicellular yeast, Saccharomyces cerevisae,’ and are closely reflected in the properties of mammalian mitochondria. The nomenclature of mutants is derived from that adopted for yeast: rho- (rho minus) mtDNA molecules contain gross deletions encompassing one or more genes; the term rhoo (rho zero) denotes that mtDNA is completely lacking from the cell; mit- (mit minus) mutations are localized changes in mtDNA sequence, base substitutions or small deletions or insertions, that affect a protein coding region; syn- (syn minus) mutations resemble the mit- type, but they primarily affect the expression of mtDNA, such as changes in tRNA or rRNA genes. MITOCHONDRIAL CYTOPATHIES Since the first recognition of human disease due to mitochondrial dysfunction some 20 years ago, various diseases with eponymous or long descriptive designations have been categorized in terms of symptomatology and, more recently, catalogued according to discrete abnormalities of the mitochondrial genome.8 In some there is clear evidence of maternal inheritance, indicative of transmission of a mitochondrial defect via the oocyte, whereas others apparently result from a mitochondrial genetic error presumably acquired during embryogenesis. The nature of the molecular defects8 include rho- deletions, particularly prevalent in myopathies (CPEO, Kearns-Sayre syndrome), mil- mutations (Leber’s optic neuropathy), and syn- mutations (encephalomyopathies: MERRF and MELAS). It is still unclear whether single molecular defects in mtDNA are solely responsible for the cellular and clinical phenotypes or whether accretion of multiple mtDNA lesions in various combinations determines particular disease pattern^.^ Also, participation of chromosomal DNA and the products of nuclear genes in the expression of mitochondrial defects must be considered.8 MITOCHONDRIAL MUTATION, AGING, AND DEGENERATIVE DISEASE
In 1989 Linnane et u I . ~developed an hypothesis on aging based on mutations affecting mtDNA. Their evidence was derived from experience with metabolic deficits resulting from mutant mtDNA in yeast (see above), and observations on diseases in man attributable to mutant mtDNA. This hypothesis elaborated some particular points. First, there is a heteroplasmic segregation of mutant mtDNA, associated with the large number of mtDNA genomes per cell. Over successive cell divisions,
94
ANNALS NEW YORK ACADEMY OF SCIENCES
during which there is replication of mtDNA in the daughter cells, the segregation of cells results in some carrying predominantly normal “wild-type’’ mtDNA, some with predominantly mutated mtDNA, and others with differing amounts of variously mutated mtDNA molecules. Second, horizontal accumulation of mtDNA mutations during the life of an individual occurs at a significantly higher rate than vertical accumulation over successive generations. Various explanations can be offered for this apparent “mtDNA purification” during the reproductive process. One notion is that during mammalian oocyte development, when there is a gross reduction in the effectively heritable number of independent mtDNA genotypes,I0 functional progeny oocytes can be formed that contain a virtually pure population of normal mtDNA genomes. In this view, oocytes that contain a high proportion of defective mtDNA genomes could be functionally selected against, and do not further participate in species reproduction. Third, a tissue energy mosaic will occur as the eventual result of the accumulation of somatic mtDNA mutations. This mosaic can be manifested as a population of cells in a given tissue ranging in bioenergetic capacity from normal to partially or grossly defective. The accumulation over time of bioenergetically defective cells was envisaged as a key factor in the process of aging.3
HISTOLOGY THE BIOENERGETIC MOSAIC A bioenergetic mosaic in tissues is schematically represented in FIGURE 1. Mitochondria accumulate a load of mutant mtDNA genomes that increases with age and becomes randomly distributed among the cells of the tissue. This random distribution of mtDNA results in some cellular progeny being fully energy competent and others completely energy deficient. This is illustrated by a histochemical study” of human cardiac tissue, from young and old individuals, stained for the key mitochondrial respiratory enzyme, cytochrome c oxidase. Nearly all specimens in early youth stain evenly and intensely across all cardiomyocytes, whereas specimens from the aged stain as a tissue mosaic with juxtaposed regions of strong, weaker, and absent staining. The study indicated an age-related loss of mitochondrial function in increasingly numerous cardiomyocytes distributed throughout the tissue.” Similar results have been found in our laboratory in young and old individuals, including humans and rats, illustrated in FIGURES 2 and 3. FIGURE 2 shows the results of staining for cytochrome c oxidase on cardiac tissue on an old individual after death from causes unrelated to mitochondrial disease; the tissue mosaic is indicated by regions of grossly depleted cytochrome c oxidase activity. FIGURE 3A shows the renal cortex from a young rat, within which all cells of the proximal tubular epithelium are reactive for cytochrome c oxidase, whereas in FIGURE 3B there is obvious loss of cytochrome c oxidase staining in much of the equivalent tissue of the aged rat. These changes are ascribed to defective mtDNA molecules that are differentially accumulated among the cells of particular tissues, and represent the tissue energy mosaic associated with aging. Such focal histological changes are characteristically demonstrable among muscle fibers of patients with gross and symptomatic mitochondrial myopathies.8 They represent in this case the widespread, but not uniform, distribution of defective mtDNA genornes that account for a much higher proportion of total mtDNA in the patient’s tissues than occurs in normal aging.
NAGLEY et al.: MITOCHONDRIAL DNA MUTATION
95
DNA ANALYSIS: AGE-RELATED MUTANT mtDNA In certain of the symptomatic mitochondria1 cytopathies there may be a sufficient load of mtDNA deletions to allow detection by a simple Southern blot analysis of tissue DNA. The much smaller load of mutant mtDNA, postulated to accumulate during aging, requires amplification of mtDNA by the polymerase chain reaction (PCR).
Nucleus
G P i
Normal fully energy competent mitochondria
I
Fully energy competent
Mitochondrlal DNA mutatlons occur and accumulate In a cell with tlme
Varylng de rees of energy de$idency
Corn let0 energy Lficiency
TISSUE ENERGY MOSAIC
FIGURE 1. Schema of generation of tissue energy mosaic during aging. Damage to mtDNA is considered to accumulate during aging. Segregation of normal and mutant mtDNA molecules will occur, largely during cell proliferation, to generate tissues in which cells may suffer varying degrees of energy deficits, extending to complete energy deficiency. For simplicity, unfilled ovals represent mitochondria containing damaged mtDNA. Individual mitochondria may themselves contain mixtures of normal and mutant mtDNA molecules. Nuclear genes are considered to be relatively unchanged during this process.
In particular, using the PCR technique, we have searched'* for a specific 4.977kb deletion in individuals not affected by mitochondria1 disease. This deletion arises from illegitimate recombination* between a pair of identical 13-bp sequences separated by almost 5 kb in human mtDNA, and is not only prevalent in myopathy
96
ANNALS NEW YORK ACADEMY OF SCIENCES
patients but also occurs in cells of the substantia nigra in Parkinson's d i ~ e a s e . ' ~ This deletion was clearly demonstrated by us1*to occur in various tissues of individuals over the age of 40 and up to the age of 87 years, but not in infants of
FIGURE 2. Cytochrome c oxidase activity in cardiomyocytes from an aged human individual. Cardiac tissue was taken, after autopsy, from a 72-year-old male with no known cardiomyopathy (motor neuron disease patient; tissue provided by Dr. H. Preston). Staining of sections is based on the use of 3,3'-diaminobenzidine (DAB), which yields an insoluble brown deposit following its reaction with added cytochrome c substrate only when the latter is oxidized by the endogenous mitochondria1 enzyme. NC, normal cardiomyocytes; DC, cytochrome c oxidase-deficient cardiomyocytes. Magnification, 475 x . (Z.-X. Wang and A. W. Linnane, unpublished observations).
3 months and younger at death (FIG. 4). A similar increase in abundance of this 4.977-kb deletion in an age-related manner was also reported by ~ t h e r s . ' ~The ~I~ deleted mtDNA molecules were estimated to represent about 0.1% of total mtDNA
NAGLEY er 01.: MITOCHONDRIAL DNA MUTATION
97
ri 3
m
a
ANNALS NEW YORK ACADEMY OF SCIENCES
98
in adult cardiac tissue,I4 but the distribution of the aberrant shortened mtDNA form among individual cells or zones of the tissues has not been reported. In the mitochondria1 myopathies, deleted mtDNA molecules may represent more than 20% or even as much as 90% of total mtDNA.8 Significantly, within the tissue mosaic regions most extensively depleted of cytochrome oxidase, the molecules of mtDNA carrying the deletions are most highly represented.I6 The technical procedures depend on whole-cell DNA extracts as templates for PCR. The 4.977-kb deletion can be readily detected by use of two oligonucleotide primers for the PCR reaction, denoted L7901 and H13631, each 20 bases long and
1
2
3
4
6
6
7
8
9 1 0 1 1
-
I0 6.76
0.77
FIGURE 4. Detection of mtDNA deletion in tissues from five human subjects of different ages. Tissues were taken at autopsy from subjects with no known mitochondrial disease. The mtDNA fragments were amplified by PCR using oligonucleotide primers L7901 and H I363 1, separated on a I% agarose gel, and photographed after staining in ethidium bromide. Lane 1contains marker DNA (lambda DNA digested with EcoRl and HindlI1); the sizes of the fragments are shown on the left side. Lanes 2 and j: heart and liver of 80-minute-old male; lanes 4 and 5: heart and adrenal gland of 3-month-old female; lanes 6 and 7: brain and left ventricle of 40-year-old male; lanes 8 and 9: left ventricle and kidney of 46-year-old female; lanes 10 and 11: kidney and psoas muscle of 50-year-old male. The sizes of the PCRamplified fragments are indicated on the right side: the 5.75-kb band derives from full-length mtDNA, and the 0.77-kb band contains the 4.977-kb deletion. (Reprinted by permission from Linnane et a/.'?)
representing sequences, respectively, of the mtDNA light and heavy strands. These primers generate a 5.75-kb product from normal full-length mtDNA. Aberrant mtDNA molecules carrying the prevalent deletion from nucleotides 8470 to 13459 on mtDNA will generate a 0.77-kb PCR product, since the deletion is close to 5 kb in its extent. Other primers hybridizing in the vicinity of the primers L7901 and H13631 are used to establish, by primer-shift analyses, the identity of the amplified PCR product as being from the inferred region of mtDNA. Finally, the cycle number is varied to adjust limits of detection. For example, in a 30-cycle PCR with primers L7901 and H 13631, the 0.77-kb deletion product is readily visible
NAGLEY et a/.: MITOCHONDRIAL DNA MUTATION
99
when the template is from adult tissues, but it is not detectable in infant tissues (FIG.4). However, the 0.77-kb band does appear at low intensity with a 60-cycle PCR with DNA templates prepared from infant tissues. IZReliable procedures are needed for quantitative analysis of the 4.977-kb mtDNA deletion to document the progressive accumulation of this mtDNA mutation with age. Our further studies on mtDNA deletions in aged humans have revealed that by use of an extended set of PCR primer pairs, a relatively large number of mtDNA deletions can be detected. For example, one 69-year-old subject without evident mitochondrial disease showed at least 10 different mtDNA deletions, of which some were tissue-specific (Zhang, Baumer, Maxwell, Linnane, and Nagley, unpublished observations). Studies on aged rats in our laboratory suggest a role for mutant mtDNA in agerelated decreases in bioenergetic functions of tissues. Results using PCR indicate the accumulation, in an age-related manner, of one particular deletion of 4.834 kb generated by illegitimate recombination across 16-bp repeat sequences lying across a broadly similar region to that involved in the human 4.977-kb deletion (Boubolas, Maxwell, Linnane, and Nagley, unpublished observations). Histochemical studies on rodent tissues equivalent to those used on human hearts should establish the validity of this aging model.
THE DISTRIBUTION OF mtDNA MUTATION: TISSUE-SPECIFIC OR GENERAL? Mutational effects on mtDNA presumably occur at random in cells among all tissues, but particular tissues may be unduly susceptible to adverse effects of mutationally damaged mtDNA. First, there may be local tissue factors that are mutagenic for mtDNA, with free radicals one obvious candidate? this, in fact provides an attractive combination of the mutational error and free radical theories of Second, there will be a greater relative susceptibility to mtDNA mutation in tissues consisting of nonreplicating versus replicating cells, since in the latter case cells replicating with a high load of mutated mtDNA may be effectively discarded, assuming that functional performance may be demanded for their propagation. In line with this, peripheral blood lymphocytes show little evidence for accumulation of mtDNA deletions as a function of somatic age.I5 Conversely, tissues consisting of cells that are slowly replicating or not dividing at all, such as those of the central nervous system (CNS), cardiac or skeletal muscle, or liver, could suffer in later life from the cumulative effects of mutations of their mtDNA. These cytopathological effects would be exacerbated if mtDNA turnover continues in the absence of cell division, since defective mtDNA molecules would continue to accumulate. Third, cells that are highly energy-dependent would be particularly susceptible to mtDNA mutation, which is borne out by the tissue distribution, mainly to CNS and muscle, of those diseases now attributable entirely to molecular defects in mtDNA. Similarly, we should look to these tissues, CNS and muscle, for primary examples of degenerative diseases of aging that may be attributable to mutant mtDNA. Cumulative mtDNA mutations could, therefore, be relevant to both a general theory of aging and the occurrence of certain tissue-specific degenerative diseases. By analogy with the identified symptomatic mitochondria1 cytopathies, which affect predominantly striated muscle and neural tissues, we could anticipate that age-related mutations in mtDNA would impact more particularly on the same tissues. For example, the mitochondrial deletion described above is demonstrable
100
ANNALS NEW YORK ACADEMY OF SCIENCES
in cells of the substantia nigra in Parkinson’s disease, and certain cardiomyopathies have been related to mutations affecting mtDNA.9 Looking further, perhaps examples of nonatherosclerotic cardiac failure in aging, “presbycardia,” could have a similar explanation.” And finally, could the known decline in athletic performance, beginning at around age 30, be based in part on subtle bioenergetic losses due to mtDNA mutation in skeletal muscle?
FUNCTIONAL CONSIDERATIONS FOR ENERGY-DEPLETED CELLS Critical to the bioenergetic loss and eventual attrition suffered by a given tissue is the ability of the cells to sustain both performance and viability in the face of a decreased or negligible energy contribution by mitochondrial oxidative phosphorylation. The maintenance of an adequate supply of intracellular ATP is not the only factor. Just as important is the ability of cells to reoxidize the load of reduced NADH derived from energy-yielding catabolism. The need for an “electron sink” is illustrated by our studies of human cells that can grow anaerobically, in the absence of mitochondrial function, but only if pyruvate is provided.’* In this example, pyruvate is efficiently reduced to lactate by the enzyme lactate dehydrogenase, regenerating oxidized NAD from the reduced NADH. The capacity of cells to survive with a greater emphasis on anaerobic glycolysis, and the availability of suitable enzymic pathways and the appropriate substrates by which excess reducing capacity can be channeled into an electron sink, will influence the threshold at which individual tissues will functionally “drop out” in the face of accumulating mtDNA damage.
CONCLUSIONS The tissue energy mosaic in aging and the accumulation of mtDNA damage provide a framework for an eventual understanding of age-related deterioration in the biological performance of cells. Future studies should clarify in molecular detail the inferred pathologies derived from mitochondrial defects both in aging, in general, and in various tissue-specific degenerative diseases. In addition, recent evidence for chemical modification of mtDNA during aging, presumably due to free radical damage,’’ provides for an integration of the mutational error and free radical concepts of aging. The likely link between mutational damage to mtDNA and decline in bioenergetic performance of tissues during aging represents a significant new application of molecular gerontology to the problems of human aging.
SUMMARY Previous theories of aging based on somatic mutation neglected mtDNA, which has a high propensity for mutational error. Knowledge of yeast mtDNA mutations and their functional effects, and of human mtDNA mutations identified in the mitochondrial cytopathies, provides for a concept of aging based on the cumulative effect of mutations affecting human mtDNA. An essential feature of this concept is heteroplasmy, representing mixtures of normal and mutant mtDNA at the cellular and mitochondrial level, resulting in a “tissue mosaic” of focal bioenergetic
NAGLEY er al.: MITOCHONDRIAL DNA MUTATION
101
deficits. Direct evidence for the concept is provided by (i) focal loss of staining for mitochondrially encoded enzymes, such as cytochrome c oxidase, in tissues of aged individuals (humans and rats) and (ii) an age-related increase in deletional mutations in m t D N A demonstrable by application of the polymerase chain reaction t o DNA templates from individuals of different ages.
REFERENCES I. 2. 3. 4. 5.
6. 7. 8. 9.
10.
II.
12. 13. 14. 15.
16. 17.
HAYFLICK. L. 1985. Theories of biological aging. Exp. Gerontol. 2 0 145-159. HARMAN, D. 1983. Free radical theory of aging: Consequences of mitochondrial aging. Age 6: 86-94. T. OZAWA& M. TANAKA. 1989. Mitochondrial DNA LINNANE. A. W.. S. MARZUKI, mutations as an important contributor to ageing and degenerative diseases. Lancet i: 642-645. 1990. Mitochondrial mutations may increase oxidative BANDY,B. & A. J . DAVISON. stress: Implications for carcinogenesis and aging? Free Radical Biol. Med. 8: 523539. MIQUEL.J . 1991. An integrated theory of aging as the result of mitochondrial-DNA mutation in differentiated cells. Arch. Gerontol. Geriatr. 12: 99-1 17. ATTARDI, G. & G. SCHATZ.1988. Biogenesis of mitochondria. Annu. Rev. Cell Biol. 4 289-333. LINNANE, A. W. & P. NACLEY.1978. Mitochondrial genetics in perspective: The development of a genetic and physical map of the yeast mitochondrial genome. Plasmid 1: 324-345. J . M. & D. C. WALLACE. 1990. Oxidative phosphorylation diseases: DisorSHOFFNER, ders of two genomes. Adv. Hum. Genet. 19: 267-330. S. SUCIYAMA, H. INO, K. OHNO,K. HATTORI,T. OBAYASHI, OZAWA, T . , M. TANAKA, K. KAWAMURA. Y. NAKANE& K. HASHIBA.1991. Patients T. ITO, H. DECUCHI, with idiopathic cardiomyopathy belong to the same mitochondrial DNA gene family of Parkinson’s disease and mitochondrial encephalomyopathy. Biochem. Biophys. Res. Commun. 1 7 2 518-525. HAuswiRTH. W. W. & P. J . LAiPis. 1985. Transmission genetics of mammalian mitochondria: A molecular model and experimental evidence. I n Achievements and Perspectives of Mitochondrial Research, Vol. 2: Biogenesis. E. Quagliariello, E. C. Slater, F. Palmieri, C. Saccone & A. M. Kroon. Eds.: 49-59. Elsevier. Amsterdam. J . 1989. Cytochrome-c-oxidase deficient cardiomyocytes in the huMULLER-HOCKER, man heart-an age-related phenomenon: A histochemical ultracytochemical study. Am. J . Pathol. 134: 1167-1 173. R. J . MAXWELL,H. PRESTON, C. ZHANG&S.MARZUKI. L I N N A N A. E , W.. A. BAUMER, 1990. Mitochondrial gene mutation: The ageing process and degenerative diseases. Biochem. Int. 22: 1067-1076. IKEBE,S., M. TANAKA, K. OHNO,W. SATO,K. HATTORI. T. KONDO,Y. MIZUNO & T. OZAWA.1990. Increase of deleted mitochondrial DNA in the striatum in Parkinson’s disease and senescence. Biochem. Biophys. Res. Commun. 170: 1044-1048. G. A. & N. ARNHEIM. 1990. Detection of a specific mitochondrial DNA CORTOPASSI, deletion in tissues of older humans. Nucleic Acids Res. 18: 6927-6933. YEN,T.-C., J.-H. Su, K.-L. KING& Y.-H. WEI. 1991. Ageing-associated 5 kb deletion in human liver mitochondrial DNA. Biochem. Biophys. Res. Commun. 178: 124131. E. A., G. KARPATI & K. E. M. HASTINGS. Deletion mutants are functionSHOUBRIDCE, ally dominant over wild-type mitochondrial genomes in skeletal muscle fiber segments in mitochondrial disease. Cell 62: 43-49. T. OBAYASHI, T. ITO, T. SATAKE,Y. HATTORI.K., M. TANAKA.S. SUCIYAMA, & T. OZAWA.1991. Age-dependent increase in deleted H A N A KJI. ,ASAI.M. NAGANO mitochondrial DNA in the human heart: Possible contributory factor to presbycardia. Am. Heart J . 121: 1735-1742.
102
ANNALS NEW YORK ACADEMY OF SCIENCES
18. VAILLANT, F., B. E. LOVELAND, P. NAGLEY & A . W. LINNANE. 1991. Some biochemical properties of human lymphoblastoid Namalwa cells grown anaerobically. Biochem. Int. 23: 571-580. 19. HAYAKAWA, M . , K . TORII,S. SUGIYAMA, M. TANAKA& T. OZAWA.1991. Ageassociated accumulation of 8-hydroxydeoxyguanosine in mitochondria1 DNA of human diaphragm. Biochem. Biophys. Res. Commun. 179 1023-1029.
