MINIREVIEW
The mitochondrial-lysosomal axis theory of aging
Accumulation of damaged mitochondria as a result of imperfect autophagocytosis
Ulf T. Brunk and Alexei Terman
Division of Pathology II, Faculty of Health Sciences, Linko
¨
ping University, Sweden
Cellular manifestations of aging are most pronounced in
postmitotic cells, such as neurons and cardiac myocytes.
Alterations of these cells, which are responsible for essential
functions of brain and heart, are particularly important
contributors to the overall aging process. Mitochondria and
lysosomes of postmitotic cells suffer the most remarkable
age-related alterations of all cellular organelles. Many
mitochondria undergo enlargement and structural disor-
ganization, while lysosomes, which are normally responsible
for mitochondrial turnover, gradually accumulate an unde-
gradable, polymeric, autofluorescent material called lipofu-
scin, or age pigment. We believe that these changes occur not
only due to continuous oxidative stress (causing oxidation of
mitochondrial constituents and autophagocytosed mater-
ial), but also because of the inherent inability of cells to
completely remove oxidatively damaged structures (biolo-
gical ÔgarbageÕ). A possible factor limiting the effectiveness of
mitochondial turnover is the enlargement of mitochondria
which may reflect their impaired fission. Non-autophago-
cytosed mitochondria undergo further oxidative damage,
resulting in decreasing energy production and increasing
generation of reactive oxygen species. Damaged, enlarged
and functionally disabled mitochondria gradually displace
normal ones, which cannot replicate indefinitely because of

autofluorescent, undegradable, polymeric material [7,8].
Accumulating evidence suggests that age-related damage
is an ineluctable consequence of normal oxygen metabo-
lism, which is associated with a relentless formation of
reactive oxygen species (ROS) [9–11]. Superoxide dismutase,
catalase, glutathione peroxidase, vitamin E, and other
antioxidant defense systems reduce, but do not prevent,
macromolecular damage such as single- and double
DNA-strand breaks, DNA–protein and protein–protein
cross-linking, protein fragmentation, and oxidation and
decomposition of lipids, resulting in the formation of
dangerous products such as hydroperoxides, alkyl radicals,
cyclic endoperoxides, and aldehydes [11–13]. Cellular aging
is also characterized by disturbances in protein synthesis,
decreased enzyme activity and progressive impairment of
the functions of mitochondria and other organelles [11].
Damaged and effete cellular structures may well be consid-
ered as biological ÔgarbageÕ.
In proliferating cells, oxidant-induced damage does not
accumulate substantially with age, apparently because the
process of cell division efficiently dilutes damaged structures
[14,15]. Indeed, Hydra, a primitive multicellular organism
composed only of cells that are continuously renewed by
proliferation, do not seem to show any signs of aging [16].
Similarly, age-related changes are not seen in cultured
neoplastic cells or in early passages of normal, actively
proliferating cells. However, when cell proliferation is
inhibited in confluent cultures of normal diploid cells, or
Correspondence to U. T. Brunk, Division of Pathology II, University
Hospital, SE-58185, Linko

subsequent fusion with lysosomes [21,22]. Alternatively,
the material enters lysosomes by invagination of the
membrane (microautophagy) [23] or by chaperone-medi-
ated selective autophagy [24].
Despite such continuous recycling of cellular compo-
nents, postmitotic cells age, suggesting that the recycling
mechanisms are inherently imperfect [15,25], and this may
provide an attractive explanation for many of the features of
aging. A number of early explanations of aging, such as
Orgel’s error catastrophe theory and the somatic mutation
theory, were based on the idea that aging results from the
accumulation of synthetic errors [26,27]. Adequate support
for these theories, however, could not be found, suggesting
that organisms age because they cannot completely remove
Ôbiological garbageÕ, rather than because they incorrectly
synthesize new constituents [15]. Despite functioning pro-
teasomes and lysosomes, postmitotic cells progressively
accumulate oxidatively modified proteins, other macromol-
ecules and defective organelles. Even when these defective
structures are autophagocytosed, they are not totally
degraded, and lipofuscin forms.
Here we discuss the involvement of mitochondria and
lysosomes in the aging process, based on our opinion that
structural alterations and misfunction of these two organ-
elles account for the majority of senescent changes in
postmitotic cells.
INTRALYSOSOMAL ACCUMULATION
OF LIPOFUSCIN
Deposition of lipofuscin pigment within postmitotic cells,
one of the most prominent signs of aging, has been known

peroxide (which easily diffuses throughout the cell), forming
the extremely reactive hydroxyl radical (Fenton reaction):
Fe
2 þ
þ H
2
O
2
! Fe
3 þ
þ OH
À
þ HO
Æ
Hydroxyl radicals, which have a half-life on the order of
10
)9
s, do not diffuse, making their reaction with surround-
ing biomolecules site-specific. In unsaturated fatty acids they
initiate a chain-reaction:
LH þ HO
Æ
! L
Æ
þ H
2
O
L
Æ
þ O

AGE-RELATED MITOCHONDRIAL
DAMAGE
Mitochondrial alterations in aging postmitotic cells have
been characterized extensively. Usually, mitochondrial size
varies more in old cells, as compared to corresponding
young cells, with a high proportion of large, sometimes
extremely large (ÔgiantÕ), mitochondria [5,37]. Ultrastructural
changes range from swelling and loss of cristae to complete
deterioration and homogenization of matrix and mitochon-
drial membranes [4,5]. Mutations of mitochondrial DNA
Ó FEBS 2002 Mitochondrial aging and autophagocytosis (Eur. J. Biochem. 269) 1997
(mtDNA) accumulate progressively, often involving the
sites coding for respiratory chain proteins [6,38]. Aging is
associated with decreased activity of the citric acid cycle,
beta-oxidation, and oxidative phosphorylation enzymes
[39,40]. As a result, mitochondria of aged postmitotic cells
have decreased membrane potential and produce less ATP
than the mitochondria of young cells [6,38].
Mitochondria are the primary sites of ROS generation,
which may be one reason why they are more affected by age
than other organelles [38,41–43]. Certain properties specific
to mtDNA also contribute to a high susceptibility of
mitochondria to age-related damage. mtDNA is more
vulnerable than nuclear DNA, because it is not protected by
histones, contains a much larger proportion of expressed
genes, and shows less efficient repair, at least for some types
of lesions [6,38]. Consistent with this, aging primarily affects
complexes I and IV of the electron-transport chain, encoded
by mtDNA, but not complexes II and III, which are mainly
coded by nuclear genes [44]. Obviously, mitochondrial

would make mitochondria large and, perhaps, exclude them
from normal autophagocytotic degradation?
Regulation of mitochondrial division (fission) is poorly
understood [46]. However, it is likely to be impaired by
oxidative damage to mitochondrial DNA and proteins.
Although mitochondrial fission is not absolutely dependent
on mtDNA replication (as is the case for cell division which
requires nuclear DNA replication); the amount of normal
DNA per mitochondrion apparently must not be too low.
Consistent with this, mtDNA-depleted cells usually contain
substantially enlarged mitochondria. Therefore, age-
dependent accumulation of mutations in mtDNA, perhaps
especially in the control region for replication [47], may also
diminish mitochondrial fission which latter could govern the
appearance of abnormal, large mitochondria.
Once a mitochondrion starts to enlarge (due to initial
oxidative damage and disturbed fission), it may be, as
mentioned above, less likely to be autophagocytosed and
recycled. This will lead to further mitochondrial damage by
ROS, a decrease in ATP production, and additional growth
as mitochondrial proteins synthesized by nuclear genes are
continuously imported into existing mitochondria. Gradu-
ally, more and more mitochondria become enlarged and
dysfunctional. These ÔgiantÕ mitochondria, as well as lipo-
fuscin deposits, may remain within postmitotic cells forever.
In a number of myopathies, cardiomyopathies and
neurodegenerative diseases, mutant mitochondria also
progressively accumulate and gradually replace normal
mitochondria [48,49]. Furthermore, as is the case in the
aging postmitotic cell, enlarged mitochondria are found in

. As a result, the membranes of defective mitochondria
may suffer less oxidative damage than those of normal
mitochondria, may be less targeted for autophagocytosis
and therefore preferentially survive. This hypothesis re-
quires evidence – currently lacking – that mitochondria are
selected for autophagocytosis based on the amount of
membrane damage. This may, however, be the case.
Although it was long believed that macroautophagy is a
nonselective process [22], a recent report shows that sperm
mitochondria are tagged with ubiquitin, providing for their
selective degradation after fertilization [54]. It is still not
clear, however, whether ubiqutin also may label oxidatively
damaged mitochondria.
Increased replication of defective mitochondria
A number of congenital mitochondrial neuropathies,
myopathies and cardiomyopathies begin, in young age, as
discrete heteroplasmic mtDNA mutations. It is well docu-
mented that in many such conditions, there is age-depend-
ent somatic selection which favours the replication and/or
persistence of mitochondria carrying the mutation, explain-
ing why the neurologic and muscular effects of these
diseases may appear only later in life [48,55]. Based on these
facts, it was hypothesized that defective mitochondria have
a replicative advantage over normal mitochondria [56,57].
Analogous selection for dysfunctional mitochondria may
also occur in the case of aging; Wanagat et al. recently
reported that atrophic muscle fiber segments from old rats
contain mtDNA deletions and have depressed cytochrome c
oxidase activity [57]. This suggests a clonal expansion of
defective mitochondria; however, because mitochondria of

and accelerating lipofuscin formation from this material.
Therefore, in the aging cell, lysosomal enzymes, essential for
autophagocytotic recycling of damaged cellular constituents,
are in short supply, forcing senescent cells to use their worn-
out and poorly operating macromolecules and organelles.
Consistent with this hypothesis, schematically outlined in
Fig. 2, we have found that lipofuscin loading of fibroblasts
does decrease their autophagocytotic capacity [61].
Fig. 2. Age-related accumulation of damaged mitochondria and lipofuscin inclusions within postmitotic cells. Oxidative injury to mitochondrial DNA
and proteins may disturb fission of some mitochondria, resulting in their enlargement. Large mitochondria are likely to be poorly autophago-
cytosed, which leads to further damage and enlargement. Lysosomes of aging postmitotic cells progressively accumulate lipofuscin pigment.
Apparently, lipofuscin-containing lysosomes attract large amounts of newly synthesized lysosomal enzymes which, however, fail to degrade the
undegradable pigment. Consequently, it may result in insufficient supply of lysosomal enzymes for autophagocytosis, further lessening mito-
chondrial recycling. Poorly functioning ÔgiantÕ mitochondria and lipofuscin-filled lysosomes (acting as sinks for lysosomal enzymes) gradually
displace normal organelles, eventually causing decreased ATP production and failing autophagocytosis, resulting in death of postmitotic cells.
Oxidatively-damaged mitochondria are indicated by a dotted pattern (more dots ¼ more damage). F, mitochondrial fission. APS, autophago-
some. The column charts show age-related trends in the mitochondrial production of ATP, autophagocytotic capacity (APG), lipofuscin content
(LF), and formation of reactive oxygen species (ROS).
Ó FEBS 2002 Mitochondrial aging and autophagocytosis (Eur. J. Biochem. 269) 1999
Although loaded with biological ÔgarbageÕ such as
defective mitochondria and lipofuscin deposits, postmitotic
cells do contain normal mitochondria that can divide to
compensate for decreased ATP production. This explains
why many senescent and severely damaged postmitotic cells
continue to function, and is consistent with the fact that
such cells usually are larger than corresponding young cells.
Cells cannot increase in size indefinitely, however, because
larger cells poorly move nutrients to their core. Therefore,
this compensatory mechanism finally fails and the senescent
postmitotic cell dies for lack of ATP.

injury and dysfunction play a central role in the aging of
postmitotic cells, as well as in aging of the whole organism,
considering the particular importance of such cells (inclu-
ding neurons and cardiac myocytes) for life maintenance.
Mitochondria are the main source of ROS formation, as
well as the main target for free radical attack. The
accumulation of defective mitochondria within aging cells
suggests that some are not properly autophagocytosed.
Aged mitochondria are often enlarged (apparently due to
impaired division) which may explain their inefficient
autophagocytotic removal. Moreover, macromolecular
components of autophagocytosed mitochondria and other
cellular structures undergo further oxidative modification
within lysosomes, resulting in the formation of an unde-
gradable material called lipofuscin. Heavy lipofuscin load-
ing of postmitotic cells decreases their autophagocytotic
capacity, leading to progressively less mitochondrial recyc-
ling. Consequently, mitochondrial and lysosomal age-rela-
ted alterations may amplify each other, eventually causing
profound dysfunction and death of postmitotic cells.
ACKNOWLEDGEMENTS
We thank Ms. Diane Konzen for expert linguistic help and Drs John
W. Eaton and John D. Furber for numerous helpful comments.
Supported by the Swedish Medical Research Council (grant 4481) and
the Linko
¨
ping Health University/University Hospital Aging Founda-
tion.
REFERENCES
1. Strehler, B.L. (1977) Time, Cells, and Aging. 2nd edn. Academic

such as phospholipase A2, while the latter, together with lysosomal
enzymes, may attack mitochondria, releasing cytochrome c and, as
well, induce further lysosomal breach.
2000 U. T. Brunk and A. Terman (Eur. J. Biochem. 269) Ó FEBS 2002
13. Halliwell, B. & Gutteridge, J.M.C. (1999) Free Radicals in Biology
and Medicine. 3rd ednn. Oxford University Press, New York.
14. Sheldrake, A.R. (1974) The ageing, growth and death of cells.
Nature 250, 381–385.
15. Terman, A. (2001) Garbage catastrophe theory of aging: imperfect
removal of oxidative damage?. Redox Report 6, 15–26.
16. Martinez, D.E. (1998) Mortality patterns suggest lack of senes-
cence in hydra. Exp. Gerontol. 33, 217–225.
17. Hayflick, L. (1965) The limited in vitro lifetime of human diploid
cell strains. Exp. Cell Res. 37, 614–636.
18. Campisi, J. (1996) Replicative senescence: an old lives’ tale?. Cell.
84, 497–500.
19.Sorimachi,H.,Ishiura,S.&Suzuki,K.(1997)Structureand
physiological function of calpains. Biochem. J. 328, 721–732.
20. Myung, J., Kim, K.B. & Crews, C.M. (2001) The ubiquitin-pro-
teasome pathway and proteasome inhibitors. Med. Res. Rev. 21,
245–273.
21. Mortimore, G.E., Miotto, G., Venerando, R. & Kadowaki, M.
(1996) Autophagy. Subcell. Biochem. 27, 93–135.
22. Klionsky, D.J. & Emr, S.D. (2000) Autophagy as a regulated
pathway of cellular degradation. Science 290, 1717–1721.
23. Marzella, L., Ahlberg, J. & Glaumann, H. (1981) Autophagy,
heterophagy, microautophagy and crinophagy as the means for
intracellular degradation. Virchows Arch. B. 36, 219–234.
24. Dice, J.F. (2000) Lysosomal Pathways of Protein Degradation.
Eurekah.com/Landes Bioscience, Georgetown, Texas.

pigment. Gerontology 41, 283–295.
34. Pisoni, R.L., Acker, T.L., Lisowski, K.M., Lemons, R.M. &
Thoene, J.G. (1990) A cysteine-specific lysosomal transport system
provides a major route for the delivery of thiol to human fibroblast
lysosomes: possible role in supporting lysosomal proteolysis.
J. Cell Biol. 110, 327–335.
35. Yin, D. (1996) Biochemical basis of lipofuscin, ceroid, and age
pigment-like fluorophores. Free Rad. Biol. Med. 21, 871–888.
36. Treff, W.M. (1974) Das involutionsmuster des nucleus dentatus
cerebelli. Altern (Platt, D., ed.), pp. 37–54. Schattauer, Stuttgart.
37. Sachs, H.G., Colgan, J.A. & Lazarus, M.L. (1977) Ultrastructure
of the aging myocardium: a morphometric approach. Am. J. Anat.
150, 63–71.
38. Richter, C. (1995) Oxidative damage to mitochondrial DNA
and its relationship to ageing. Int. J. Biochem. Cell Biol. 27, 647–
653.
39. Hansford, R.G. & Castro, F. (1982) Age-linked changes in the
activity of enzymes of the tricarboxylate cycle and lipid oxidation,
and of carnitine content, in muscles of the rat. Mech. Ageing Dev.
19, 191–200.
40.Perron,J.T.,Tyson,R.L.&Sutherland,G.R.(2000)Main-
tenance of tricarboxylic acid cycle kinetics in Brown–Norway
Fischer 344 rats may translate to longevity. Neurosci. Lett. 281,
91–94.
41. Chance, B., Bank, W.W. & Zhang, C. (1992) Instabilities of
metabolic regulations in aging. EXS 62, 58–63.
42. Sohal, R.S. & Brunk, U.T. (1992) Mitochondrial production
of pro-oxidants and cellular senescence. Mutat. Res. 275, 295–
304.
43. Cadenas, E. & Davies, K.J. (2000) Mitochondrial free radical

52. Kurogouchi, F., Oguchi, T., Mawatari, E., Yamaura, S., Hora,
K., Takei, M., Sekijima, Y., Ikeda, S. & Kiyosawa, K. (1998) A
case of mitochondrial cytopathy with a typical point mutation for
MELAS, presenting with severe focal-segmental glomerulo-
sclerosis as main clinical manifestation. Am. J. Nephrol. 18,
551–556.
53. de Grey, A.D. (1997) A proposed refinement of the mitochondrial
free radical theory of aging. Bioessays 19, 161–166.
54. Sutovsky, P., Moreno, R.D., Ramalho-Santos, J., Dominko, T.,
Simerly, C. & Schatten, G. (1999) Ubiquitin tag for sperm
mitochondria. Nature 402, 371–372.
55. Silvestri, G., Santorelli, F.M., Shanske, S., Whitley, C.B.,
Schimmenti, L.A., Smith, S.A. & DiMauro, S. (1994) A
new mtDNA mutation in the tRNA (Leu (UUR) gene asso-
ciated with maternally inherited cardiomyopathy. Hum. Mutat. 3,
37–43.
56. Shoubridge, E.A., Karpati, G. & Hastings, K.E. (1990) Deletion
mutants are functionally dominant over wild-type mitochondrial
genomes in skeletal muscle fiber segments in mitochondrial dis-
ease. Cell 62, 43–49.
Ó FEBS 2002 Mitochondrial aging and autophagocytosis (Eur. J. Biochem. 269) 2001
57. Wanagat, J., Cao, Z., Pathare, P. & Aiken, J.M. (2001) Mito-
chondrial DNA deletion mutations colocalize with segmental
electron transport system abnormalities, muscle fiber atrophy,
fiber splitting, and oxidative damage in sarcopenia. FASEB J. 15,
322–332.
58. Hochhauser, D. (2000) Relevance of mitochondrial DNA in
cancer. Lancet 356, 181–182.
59. Woolf, N. (1998) Pathology Basic and Systemic. W.B. Saunders,
London.